Solar Photovoltaic Power Intermittency and Implications on Power Systems [1 ed.] 1527571297, 9781527571297

Citation preview

Solar Photovoltaic Power Intermittency and Implications on Power Systems

Solar Photovoltaic Power Intermittency and Implications on Power Systems Edited by

Mohammed Albadi and Abdullah Al-Badi

Solar Photovoltaic Power Intermittency and Implications on Power Systems Edited by Mohammed Albadi and Abdullah Al-Badi This book first published 2021 Cambridge Scholars Publishing Lady Stephenson Library, Newcastle upon Tyne, NE6 2PA, UK British Library Cataloguing in Publication Data A catalogue record for this book is available from the British Library Copyright © 2021 by Mohammed Albadi, Abdullah Al-Badi and contributors All rights for this book reserved. No part of this book may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, electronic, mechanical, photocopying, recording or otherwise, without the prior permission of the copyright owner. ISBN (10): 1-5275-7129-7 ISBN (13): 978-1-5275-7129-7

TABLE OF CONTENTS

Preface ...................................................................................................... vii Chapter One ................................................................................................ 1 Solar Irradiation and its Variability Alaa I. Ibrahim Chapter Two ............................................................................................. 39 Temperature Effect on PV Output Power Variability Razzaqul Ahshan and Abdullah H. Al-Badi Chapter Three ........................................................................................... 70 Variability Analysis of PV System Output Faiza AL-Harthi, Mohammed Albadi, Rashid Al Abri and Abdullah Al Badi Chapter Four ............................................................................................. 89 Impact of Dust and External Factors on Photovoltaic System Output in Western Australia Bong Wei Li, Syed Islam and Rakibuzzaman Shah Chapter Five ........................................................................................... 108 Impact of the Integration of Large-Scale PV Power Plants on the Grid Stability and Operation Hisham M. Soliman, Abdullah Al-Badi, Hassan Yousef, Abdulsalam Elhaffar, Masoud Al-Reyami, Sultan Al-Rawahi, Maktoom Al-Hosni and Anwer Al-Harthy Chapter Six ............................................................................................. 155 Capacity Value of Photovoltaics for Estimating the Adequacy of a Power Generation System Arif Malik and Mohammed Albadi

vi

Table of Contents

Chapter Seven......................................................................................... 183 Impacts of Energy Systems’ Output Variability on Transmission and Distribution Networks Hussein A. Kazem Chapter Eight .......................................................................................... 217 The Role of Energy Storage to Reduce Variability Impacts Mostafa Bakhtvar, Hamidreza Aghay Kaboli and Amer Al-Hinai Chapter Nine........................................................................................... 264 A Nearly Net-Zero Energy Building in Oman: A Case Study Abdullah Al-Badi and Awni Shaaban

PREFACE

Solar photovoltaic (PV) systems have experienced a tremendous increase in installed capacity in the past decade. The global solar PV installed capacity increased from 5.2 GW in 2005 to 627 GW by the end of 2019. This rapid growth of solar PV system deployment is expected to continue due to several factors such as PV technology improvements and cost reduction, as well as policies and regulations promoting the use of renewable energy resources. Although solar PV power is environmentally friendly and can be used to extend the life of fossil fuel reserves, it has an intermittent nature. The implications of this intermittency need to be understood by all stakeholders. In general, there are two main impacts of intermittent renewable-based generation facilities on power system operation: variability and uncertainty. Variability of solar output can affect different operational aspects of distribution and transmission systems including power quality, losses and congestions. Intermittency of solar PV power affects the balance between supply and demand; hence the entire power system’s planning and operation. For example, when the supply-demand balance is not maintained, power system frequency deviates from steady state values; consequently, system stability and reliability are jeopardized. Although it is technically possible to integrate a large number of intermittent renewable-based facilities in power systems, higher penetration levels result in more challenges to power system stability and reliability. These challenges are quantified using intermittency integration costs. The objective of the book is to present an overview of solar PV power output intermittency and the impacts of power systems. This book contains 9 chapters. Chapter 1 covers the origin of solar irradiation variabilities and their time scales as well as different factors that affect variability such as Earth’s orbital eccentricity, geometry, rotation, and axial tilt, the effect of the atmosphere and cloud cover, the effect of the solar cycle and other longterm effects. Chapter 1 also includes modeling the solar irradiation on horizontal and tilted surfaces. Chapter 2 focuses on the effect of temperature on PV systems’ output power variability. The chapter includes the effect of ambient temperature variation on PV system design and system performances

viii

Preface

and presents the PV output variability and PV cell temperature variability for a hot and dry location. In addition, the correlation between the PV output power and the PV cell temperature is also showcased and the detailed model of PV system design, system performance analysis, and variability analysis is outlined. Chapter 3 presents a case study of PV systems’ output variability. The study used 15-minute average data recorded from 3 sites to study the smoothing effect of aggregating different PV systems’ output on both deterministic and stochastic variabilities. The implications of deterministic variabilities were demonstrated during the annular eclipse that occurred on 21 June 2020. The smoothing effect that can reduce the stochastic variabilities, which are caused by cloud fronts, is also demonstrated. Chapter 4 focuses on the impact of dust and shading on solar PV systems’ output. The chapter summarizes the research relating to the different types of dust particles that affect the efficiency of PV systems. PV system performance indices for Western Australian climatic conditions are demonstrated using experimental setups. Chapter 5 studies the impact of the integration of large-scale PV power plants on the grid stability and operation. The system steady-state stability is investigated with and without external controllers installed at the PV power plants. In addition, the chapter studies the transient stability of the power system in the presence of largescale PV stations. Chapter 6 addresses the evaluation of the capacity credit or value of renewable energy systems, particularly photovoltaic (PV) plants in a power generation system. The chapter starts by giving an overview of the various adequacy measures used to evaluate generation system reliability and then concentrates on the evaluation of the capacity value of PV plants. Chapter 7 addresses the impacts of energy systems’ output variability on transmission and distribution networks. The chapter introduces the Power Flow tool, which is used to analyze the instantaneous active and reactive power flow. The chapter includes other topics such as the integrity of multiple renewable energy resources as well as interconnection and islanding control. Chapter 8 focuses on the role of energy storage systems in reducing the negative impacts of PV systems on power systems. It includes an overview of energy storage technologies and discusses how each technology can benefit the power system to alleviate the impacts of solar PV generation. The concept of dispatchable PV-energy storage hybrid systems is also discussed. The realization of dispatchable solar PV generation can reduce the need for flexibility and reserves. A framework for dispatchable solar PV generation is proposed. The components of this framework and the tasks each should carry out are articulated. Chapter 9 presents a case study about a nearly net-zero energy building (Eco-house). The house was built on the Sultan Qaboos University (SQU) campus and

Solar Photovoltaic Power Intermittency and Implications on Power Systems

ix

provides a high level of comfort. It has two stories, is connected to the electric grid and is equipped with a 20-kW rooftop PV system. The net energy consumed in the SQU Eco-House is 945 kWh annually which is less than 3.5% of the energy demand supplied by the electrical network. Thus, the SQU Eco-House is a nearly zero energy building. We hope this mix of presentations will attract a large number of interested readers. The editors of this book would like to thank all the contributors for their continued support and hard work. We would like to thank the reviewers who provided valuable comments to improve the quality of this work. We also thank the publisher for agreeing to publish this book. Finally, we would like to thank all authors for their contributions and our families for their support. Mohammed Albadi and Abdullah Al-Badi October 2020

CHAPTER ONE SOLAR IRRADIATION AND ITS VARIABILITY ALAA I. IBRAHIM1

“The Sun will be the fuel of the future” Popular Science, 1876

Abstract Solar energy is the most prolific and accessible form of renewable energy. During the last decade, solar power has become the fastest-growing energy sector, with the worldwide solar power generation increasing by more than 30-fold, outperforming the growth rate of all other renewable energy sources combined by more than an order of magnitude. More importantly, solar energy is more equitable among countries with varying economic and development stature and has the narrowest production gap among renewable energy sources between such countries. As a perpetual source of energy, solar power is used for a wide variety of applications and can be utilized at both small and large scales. However, like all forms of natural renewable energy, solar energy has its own challenges. While new technologies and innovations are improving efficiency and reducing costs, there are variabilities and constraints that need to be understood and when possible mitigated for optimal utilization. Measured at the top of the atmosphere at Earth’s average distance from the Sun, the intrinsic electromagnetic energy output of the Sun per unit area per unit time, known as the total solar irradiance, only varies minimally over long time scales of about 11 years (due to magnetic phenomena in the Sun). On the other hand, significant variabilities at shorter time scales of minutes to months are introduced by the eccentricity of Earth's orbit, Earth’s rotation and axial tilt, the geometry of Earth’s surface, atmospheric processes, and cloud cover. 1

Department of Physics, College of Science, Sultan Qaboos University, Muscat, Oman, Email: [email protected].

2

Chapter One These dynamical variabilities change by geographic location and by date and time at the same location. Here we present a comprehensive overview and analysis of the energetics of the solar irradiation and the different variabilities and constraints associated with it as observed from Earth. We start from the fundamental principles of solar radiation then address the origin and time scale of each source of variability. The solar irradiance and the integrated hourly and daily irradiation of direct Sun light are calculated for horizontal and arbitrarily tilted surfaces. Our general treatment should be useful for photovoltaic solar power purposes as well as other techniques of harnessing solar energy. We discuss approaches that can reduce the uncertainty associated with some variabilities. We conclude that despite these variabilities, the greater parts of the world receive considerable sunshine hours annually, which make solar energy a viable option. The largest regions receiving the maximum solar irradiance and least affected by its variability are those in North Africa and the Arabian Peninsula. Keywords: Solar irradiation, irradiation variabilities, electromagnetic radiation, orbital eccentricity, cloud cover, hourly and daily solar irradiation.

1. Introduction The term solar energy refers to the radiant electromagnetic radiation from the Sun, which can be harnessed using a variety of technologies such as photovoltaics (PV; converting solar radiation into electricity through solar panels that utilize the photoelectric effect), solar heating or solar thermal energy (collecting heat by absorbing solar radiation with or without focusing), solar architecture (utilizing the illumination and heating effects of solar radiation in architectural design), and artificial photosynthesis (utilizing solar radiation in chemical processes that mimic the natural process of photosynthesis) (van der Hoeven 2011). We will be concerned here with solar energy in the context of photovoltaics’ power generation, but much of the consideration and treatment discussed is relevant and applicable to most of the aforementioned forms of solar energy harnessing. In recent years, the utilization of solar energy has been growing steadily, thanks to advancements in photovoltaic technology that improve the efficiency and reduce the cost, as well as global and local environmental, economic, and policy reforms (Shubbak 2019, van der Hoeven 2011). Over the last decade, the worldwide generation of solar power grew from 21 TWh (Terra Watt hour) in 2009, 3% of the total renewable energy generation then, to 724 TWh in 2019, 26% of the total renewable energy generation during that year (Ritchie & Roser 2020; Looney 2020), making solar energy

Solar Irradiation and its Variability

3

the fastest growing source of renewable energy. The growth rate is also significant on short timescales. The one-year growth rate between 2018 and 2019 is 24%. Furthermore, solar energy is becoming more accessible and equitable among countries of otherwise varying levels of socioeconomic development. The share of solar power generation by non-OECD2 countries, which are predominantly developing countries, has been rising consistently over the past decade as measured by the production gap between OECD (mainly high income, developed countries) and non-OECD shares relative to the worldwide total. This gap narrowed from 94% in 2009 to only 7% in 2019. See Table 1-1 for the trends and comparisons between solar energy, wind, and other renewable energy sources between 1990 and 2019. Table 1-1: Renewable Energy Power Generation Over Two Decades3 The worldwide rate, as well as the shares of the OECD and non-OECD countries, is given. The 1990, 2009, 2019 rates show the long-term growth whereas the 2018 rates are provided to show the most recent annual growth compared with 2019. 1990

2009

2018

2019

2009-2019

(OECD – (OECD – (OECD – World NonWorld World World (OECD – Non- Growth NonNon(TWh) OECD)/ (TWh) (TWh) (TWh) OECD)/World Rate OECD/World OECD/World World

Solar PV

0.4

93.8%

21.0

94.0%

582.8 13.2%

724.1 6.9%

3348.1%

Wind

3.6

99.3%

276.1

64.5%

1270.2 17.4%

1429.6 17.8%

417.8%

Other 117 Renewables

73.0%

339.6

43.3%

615

651.8 19.0%

91.9%

…

636.7

…

2468 …

2805.5 …

340.6%

Total

121

19.3%

We begin by reviewing how the solar radiation is produced, the components and model of the solar spectrum, and how the solar irradiance and integrated irradiation are measured. We then look at the origins, type, and time scale of the irradiation variability and how they affect the solar irradiance received on Earth. Hourly and daily rates of irradiation are obtained for horizontal and tilted surfaces at any day, time, and geographic location. What makes solar energy special is simply the magnitude of the power received from the Sun. Among the various natural sources of energy available, solar energy delivers more power than all other natural sources combined. The power per unit area received from the Sun at the Earth’s surface (global average) exceeds all other natural sources by a factor of 2600 2 3

Organization for Economic Co-operation and Development. Raw data obtained from (Ritchie & Roser 2020 and Looney 2020).

Chapter One

4

(Kren et al. 2017, Sellers 1965). At the top of the atmosphere this factor is 3700. Fig. 1-1a illustrates the total energy reservoirs available on Earth from

(a)

(b) Fig. 1-1: (a) Total energy reservoir available from different sources on Earth compared to the annual solar energy received by Earth and the annual global energy consumption by humans. (b) World map of the global horizontal solar irradiation, showing the long-term average of the annual and daily irradiation sums in kWh m-2. Source: IEA 2011; SolarGIS, GeoModel Solar 2013.

Solar Irradiation and its Variability

5

various sources compared to the annual solar energy received at the Earth and the human annual global energy consumption at the present time (van der Hoeven 2011). Fig. 1-1b illustrates the global average of the daily and annual solar irradiation on a horizontal surface worldwide where the effects of geographic location and environmental and weather conditions can be inferred. Still, most inhabited regions of the world receive greater than 1300 kWh m-2 annually or greater than 3.5 kWh m-2 daily.

2. Solar Radiation As a star, the Sun is the dominant body in the solar system, comprising 99% of its mass. Unlike Earth and other members of the solar system, the Sun has the right composition and conditions that allow for the production of abundant and steady energy over extremely long time spans (billions of years). The Sun is a dwarf main-sequence star with a mean volumetric radius of 695,700 km (nearly 110 times that of the Earth) and is situated at an average distance of 149.60×106 km, or 1.00 AU (astronomical unit), from Earth (Williams 2018). As a result, the Sun has an apparent angular size of 0.5° in the sky, which is similar to that of the Moon. The mass of the Sun, 1.9885×1030 kg, is made of 70.6% hydrogen, 27.4% helium, and 2% other gases (Williams 2018). The enormous gravity of the Sun, due to its mass, which keeps Earth and the other planets in orbit, also compresses and heats up the gasses and plasma inside the Sun. The temperature is highest at the core, reaching upwards of 107 K (Carroll & Ostlie 1996). Under such conditions of extreme temperature, pressure, and density, gasses are ionized into hot and dense plasma. The collisions between the nuclei of the hydrogen atoms in the core become energetic enough to fuse them together in a nuclear fusion, which produces helium nuclei and releases electromagnetic radiations, which at this stage are high-energy photons in the X-ray and gamma-ray wavelengths, and sub-atomic particles (neutrinos and positrons). As they propagate outwards, the electromagnetic radiations are first trapped by the dense layers of the Sun’s interior, undergoing millions of collisions with the dense plasma in a process known as random walk that takes up to hundreds of thousands of years and results in significant softening of the radiation to longer, less energetic wavelengths. The radiations will eventually reach the surface layer of the Sun (the photosphere), while primarily in the visible and infra-red wavelengths, with smaller components in radio, ultra-violet, X-ray, and gamma-ray wavelengths,

6

Chapter One then arrive at Earth 8 minutes later with the spectrum we are familiar with. This spectrum, as observed at the top of the atmosphere and at Earth’s surface, is shown in Fig. 1-2.

Fig. 1-2: The spectrum of direct solar radiation at the top of Earth's atmosphere (light gray) and at Earth’s surface (sea level) (dark gray). The spectrum is well described by the theoretical Planck function/model (solid black line). As the radiations of different wavelengths pass through the atmosphere, some are absorbed by the atmospheric gases and molecules (e.g., water vapor H2O and carbon dioxide CO2), resulting in specific absorption bands. Source: Wikimedia Commons 2013.

About 23% of the solar radiations that arrive at the top of Earth’s atmosphere are reflected back to space (by clouds, atmospheric particles/reflecting aerosols) and another 6% are reflected back to space after they reach the ground by bright ground surfaces such as sea ice and snow (Enteria and Akbarzadeh 2014). The atmosphere absorbs about 23% of the solar radiations (by clouds, water vapor, dust/absorbing aerosols, and ozone) and the remaining 48% are absorbed by Earth’s surface. The total absorbed by Earth’s system (surface and atmosphere) is often referred to as about 71%. The radiation absorbed by the atmosphere contributes to the diffused radiation that can also reach Earth’s surface. About 99% of the energy of solar radiation is contained in the wavelength band of the near ultraviolet, visible and near infrared regions of the solar spectrum. The makeup of the radiations that reach Earth's surface is as follows: 52%–55% is infrared above 700 nm, 42%–43% is visible light (400 nm–700 nm), and 3%–5% is ultraviolet (below 400 nm). At the top of the

Solar Irradiation and its Variability

7

atmosphere, solar radiation is 29% more intense as indicated above, with an additional 8% to the ultraviolet band (Kalogirou 2014). The solar energy arriving at Earth is quantified by the flux of electromagnetic radiation F (energy per unit area per unit time, Joule per square meter per second; J m-2 s-1) and is also referred to as the solar irradiance. The flux depends on the number of photons n received per square meter per second and the photon energy ࣟ, which can be expressed in terms of the wavelength Ȝ (or frequency Ȟ),

‫݄ܿ݊ = ߥ݄݊ = ࣟ݊ = ܨ‬/ߣ where h is Planck’s constant (6.626×10-34 J s) and c is the speed of light (2.998×108 m s-1). By expressing the wavelength in ȝm, the energy ࣟ and flux F can be obtained in units of electron volt (eV) and eV m-2 s-1, respectively, ࣟ =

1.24 1.24 ݊ , ‫= ܨ‬ ߣ ߣ

Equivalently, this gives the power received (‫ )ܪ‬per unit area (Watt per square meter, W m-2), or irradiance, at the observed wavelength. But since the solar flux varies by wavelength (as n also varies with Ȝ), see Fig. 1-2, we need to describe the irradiance per wavelength, which gives the spectral irradiance (J m-2 s-1 ȝm-1), where m-2 refers to the unit area and ȝm-1 refers to the wavelength. As a measured wavelength Ȝ is normally obtained over a finite interval of width ǻȜ (which can refer to the spectral resolution), the spectral irradiance F(Ȝ) (in units of W m-2 ȝm-1) is given by,

‫ܨ‬ሺߣሻ =

݊ሺߣሻ ࣟ ݊ሺߣሻ ݄ ܿ = ȟߣ ߣ ȟߣ

The measured spectral irradiance is satisfactorily described by Planck’s function (the black curve in Fig. 1-2), which describes an ideal emitter (commonly called the blackbody) and depends only on the surface temperature of the emitting body and the wavelength of the radiation as follows,

‫ܨ‬ሺߣሻ =

2݄ܿ ଶ ߣହ

1 ௛௖ ݁ ఒ௞்

(1) െ1

Chapter One

8

where T is temperature, and k is Boltzmann’s constant (1.38 × 10 -23 J K -1). Fitting the solar spectrum to Planck’s function gives the temperature of the surface of the Sun: 5778 K (5505°C). Integrating ‫ )ߣ(ܨ‬over the wavelength range of interest (e.g., the visible spectrum), gives us the total power density ‫( )ߣ(ܪ‬power per unit area (in units of W m-2)) or irradiance

‫ = )ߣ(ܪ‬න ‫( ߣ݀)ߣ(ܨ‬2) ఒ

If ‫ )ߣ(ܨ‬is discrete such that its value is approximately constant over a wavelength interval ȟߣ, the integration can be replaced by a summation

‫ = )ߣ(ܪ‬σఒ ‫ )ߣ(ܨ‬ȟߣ Integrating ‫ )ߣ(ܪ‬over a given period of time, over which ‫ )ߣ(ܨ‬is constant, gives the solar radiant energy per unit area (Joule per square meter, J m-2) during such period. This is referred to as the solar irradiation, solar exposure, solar insolation, or insolation.

While the solar energy is perceived as being carried by the photons of the electromagnetic radiation, these photons are quanta of discrete electric (E) and magnetic (B) fields of the radiation, described by the Poynting vector S = E × B/ȝ, where ȝ is the permeability of the medium. The Poynting vector illustrates the role of the direction of the incident radiation. In the above treatment a normal incidence is assumed but in reality, the angle of incidence on a fixed horizontal or tilted surface will change during the day due to the changing altitude of the Sun due to Earth’s rotation and from day to day during the year due to the change in the solar declination because of Earth’s orbit and axial tilt. The spectral irradiance F(Ȝ) from the Sun may also vary over an extended period of time. So, a general treatment that takes these effects into account is given in the next section.

Solar Irradiation and its Variability

9

3. Origin of Solar Irradiation Variabilities and their Time Scales As a glowing spherical body of hot gas and plasma with an average surface temperature of 5778 K, the Sun is a thermal emitter whose radiative properties are well explained by thermal and quantum physics and are confirmed by observations from Earth’s surface and space. The Sun’s total radiation flux or power density ‫ܪ‬௦௨௡ (energy per unit time per unit area, J s1 -2 m or W m-2) at its surface can be obtained by integrating ‫ )ߣ(ܨ‬over all wavelengths. From equations (1) and (2) we get ସ = 6.32×107 W m-2 ‫ܪ‬௦௨௡ = ‫ܶ ߪ = ߣ݀ )ߣ(ܨ ߣ׬‬ௌ௨௡

(3)

This is known as the Stefan-Boltzmann law, where ߪ is the StefanBoltzmann constant and ܶௌ௨௡ is the average surface temperature of the Sun. This is the power density, or irradiance, emitted by the Sun in all wavelengths of the electromagnetic spectrum, not just the visible spectrum. The total power of the Sun (energy per unit time, J s-1 or W) is obtained by account for the area of radiant surface, ଶ ܲௌ௨௡ = ‫ܪ‬௦௨௡ 4ߨܴௌ௨௡ = ߪ ܶ ସ = 3.84×1026 W

(4)

where ܴௌ௨௡ is the radius of the Sun. The solar irradiance received at the top of Earth’s atmosphere on a plane perpendicular to the Sun’s rays, ‫ܪ‬௢ୄ , see Fig. 1-3, can be obtained by conserving the power at the surface of the Sun and that at the surface of a sphere whose radius is equal to the Earth-Sun distance,

ଶ ଶ ‫ܪ‬ௌ௨௡ 4ߨܴௌ௨௡ = ‫ܪ‬௢ୄ 4ߨܴா௔௥௧௛ିௌ௨௡

‫ܪ‬௢ୄ = ‫ܪ‬௦௨௡ ቆ

ܴௌ௨௡ ܴா௔௥௧௛ିௌ௨௡

where ܴா௔௥௧௛ିௌ௨௡ is the Earth-Sun distance.

ଶ

ቇ

(5)

Chapter One

10

Fig. 1-3: The solar irradiance ‫ܪ‬௢ୄ at the top of Earth’s atmosphere on a plane perpendicular to the incident solar radiation.

The power density at the average Earth-Sun distance (1 AU) at the top of the atmosphere is called the solar constant, ଶ

‫ܪ‬ௌ஼ = ‫ܪ‬௦௨௡

ܴௌ௨௡ ቆ ቇ = 1366.80 W mିଶ 1 ‫ܷܣ‬

‫ܪ‬ௌ஼ represents the power density, or irradiance, received in all wavelengths emitted by the Sun at the top of Earth’s atmosphere. It can also be referred to as the extraterrestrial irradiance or power density. Integrating ‫ܪ‬ௌ஼ over a cross-sectional area with Earth’s mean radius (6371 km), gives the total power of 1.74×1017 W arriving at Earth (out of the original 3.84×1026 W at the Sun), which is only 4.53×10-8 % of ܲௌ௨௡ . The actual value is even less due to Earth’s curved surface. In terms of ‫ܪ‬ௌ஼ , ‫ܪ‬௢ୄ of equation (5) becomes ‫ܪ‬௢ୄ = ‫ܪ‬ௌ஼ ቆ

1 ‫ܷܣ‬ ܴா௔௥௧௛ିௌ௨௡

ଶ

ቇ

(6)

3.1 Effect of Earth’s Orbital Eccentricity The first variability in ‫ܪ‬௢ୄ given in equation (5) comes from the change in ܴா௔௥௧௛ିௌ௨௡ due to Earth’s elliptical orbit around the Sun (with eccentricity 0.0167086). As shown in Fig. 1-4, the distance between Earth and the Sun varies during the year from 147.10×106 km (0.98329 AU) during the closest approach (perihelion) on January 2-5 to 152.10×106 km (1.01670 AU) during the furthest approach (aphelion) on July 3-5. The average distance (149.60×106 km or 1.00 AU) represents the semimajor axis (Simon et al.

Solar Irradiation and its Variability

11

1994). This change in ܴா௔௥௧௛ିௌ௨௡ will cause the solar irradiance or power density to vary according to equation (5) by about 6.9% (േ3.5%) between the closest and furthest approach, with the larger irradiance (1.41 kW m-2) in early January and the smaller irradiance (1.32 kW m-2) in early July.

Fig. 1-4: Top: Earth’s distance and orbit around the Sun at the different times of the year, showing the closest approach (perihelion) and the furthest approach (aphelion) as well as the Winter and Summer solstices and the Vernal (Spring) and Autumnal (Fall) equinoxes. Summer. The day number given (dn) is for the year 2020. Bottom: The variation of the distance between the Earth and the Sun by the day number.

Chapter One

12

Because the change in the distance is periodic as can be seen from Fig. 1-4 and depends only on the day number (dn, which varies from dn = 1 on Jan. 1st to dn = 365 or 366 on Dec. 31st, depending on a leap or a common year), ܴா௔௥௧௛ିௌ௨௡ , and consequently the solar irradiance ‫ܪ‬௢ ୄ , can be expressed as a harmonic function of dn. ‫ܪ‬௢ ୄ (݀݊) = ‫ܪ‬ௌ஼ ቆ

1 ‫ܷܣ‬

ଶ

ቇ ܴா௔௥௧௛ିௌ௨௡ (݀݊)

ଶ

The ratio ൫1 ‫ܷܣ‬/ ܴா௔௥௧௛ିௌ௨௡ (݀݊)൯ is called the eccentricity correction and ܴா௔௥௧௛ିௌ௨௡ (݀݊) can be written as a Fourier expansion in terms of 1 AU with a number of coefficients (Kalogirou 2014, Spencer 1971, Iqbal 1983), which we write in the concise form ቆ

1 ‫ܷܣ‬ ܴா௔௥௧௛ିௌ௨௡ (݀݊)

ଶ

ቇ = 1 + 0.033412 cos ൬2ߨ

݀݊ െ ݀௢ ൰ 365

where ݀௢ is the day number of Earth's perihelion, or the closest approach to the Sun, which varies slightly from year to year but always occurs between January 2 and 5. For example, ݀௢ (2020) = 5 (i.e., Jan. 5th), and ݀௢ (2021) = 2 (i.e., Jan. 2nd). Hence the solar irradiance at the top of the atmosphere on a given day dn is given by ‫ܪ‬௢ୄ (݀݊) = ‫ܪ‬ௌ஼ ൤1 + 0.033412 cos ൬2ߨ

݀݊ െ ݀௢ ൰൨ 365

(7)

The variation of ‫ܪ‬௢ୄ (݀݊) with the day number through the year is shown in Fig. 1-5.

Solar Irradiation and its Variability

13

Fig. 1-5: The variability in the solar irradiance or power density ‫ܪ‬௢ ୄ at the top of Earth’s atmosphere over the course of one year. The dashed line represents the value of the solar constant ‫ܪ‬ௌ஼ .

3.2 The Effect of Earth’s Geometry, Rotation, and Axial Tilt While ‫ܪ‬௢ୄ on a given day is the same at all points at the top of the atmosphere whose planes are perpendicular to the incoming solar radiation, the curved surface of the Earth will dictate different angles for all horizontal surfaces other than the surface perpendicular to the line between the centers of the Earth and the Sun as illustrated in Fig. 1-6. Let us calculate the solar irradiance at a point P’ on a plane at the top of the atmosphere parallel to a horizontal surface at an arbitrary point P on Earth’s surface (see Fig. 1-6). The incident parallel rays in this case make an angle ߠ௭ with the perpendicular to the surface. The solar irradiance at P’ is given by ‫ܪ‬௢ (݀݊, ߠ௭ ) = ‫ܪ‬௢ୄ (݀݊) cos ߠ௭

(8)

14

Chapter One

Fig. 1-6: A parallel beam of solar radiation incident to a plane parallel to a horizontal surface at an arbitrary point P on Earth. The incident parallel rays make an angle ߠ௭ with the normal to the plane.

The solar irradiance will be at its maximum when ߠ௭ = 0 (normal incidence), will decrease as the ߠ௭ gets larger (as the point P moves further north) (and equivalently further south in the southern hemisphere), and will vanish when the plane becomes parallel to the incident solar rays (ߠ௭ = 90° ). The angle ߠ௭ is called the solar zenith angle, i.e., the angle between the zenith at a given location (the line perpendicular to the horizontal surface at P or P’, passing by Earth’s center) and the line connecting the centers of the Earth and the Sun (which defines the direction of the incident solar radiation). As shown in Fig. 1-7 and Fig. 1-8 (Top), the zenith angle will change: (a) during the day (from +90o at sunrise to െ90o at sunset, passing by its minimum at the local solar noon when the Sun reaches the highest elevation in the sky) due to the apparent motion of the Sun resulting from Earth’s rotation around its own axis, and (b) during the year from day to day as the Sun’s elevation in the sky will change for the same geographic latitude due to Earth’s axial tilt while orbiting the Sun. Changes within the day are expressed in terms of the solar hour angle ߱, which converts the local time and the local solar time to an angular value (with ߱ = 0° at solar noon, taking a positive value in the morning, and taking a negative value in the afternoon). Changes from day to day during the year are expressed in terms of the solar declination ߜ, the angle between

Solar Irradiation and its Variability

15

the plane of the equator and the line joining the centers of the Sun and the Earth, see Fig. 1-7 and Fig. 1-8. We can then write the zenith angle or cos ߠ௭ in terms of the solar hour angle ߱, the solar declination ߜ, and the geographic latitude ߶ as follows, cos ߠ௭ = sin ߜ sin ߶ + cos ߜ cos ߶ cos ߱

(9)

Fig. 1-7: While orbiting the Sun with an axial tilt, the elevation of the Sun in the sky changes from day to day during the year. As a result, the solar declination ߜ (the angle between the line connecting the centers of the Earth and the Sun and the plane of the equator) changes for each latitude.

16

Chapter One

Fig. 1-8: Top: The path of the Sun during the day on a horizontal surface, showing the solar zenith angle ߠ௭ , the hour angle ߱, the elevation or altitude ߙ, and the azimuth angle ߰. Bottom: The relation between the zenith angle ߠ௭ at an arbitrary point P and the solar declination ߜ, the geographic latitude ߶, and the hour angle ߱.

The solar declination depends only on the day number ݀݊ and its variation during the year can be written as (Iqbal 1983) ߜ(݀݊) = െ23.45௢ cos ቆ

360 (݀݊ + 10)ቇ 365

The solar hour angle ߱ is obtained from the local solar time ‫ݐ‬௅ௌ் as

Solar Irradiation and its Variability

17

߱(‫ݐ‬௅ௌ் ) = (‫ݐ‬௅ௌ் െ 12) 360° /24݄‫ݎ‬ where ‫ݐ‬௅ௌ் is obtained in terms of the local time ‫ݐ‬௅் using ‫ݐ‬௅ௌ் = ‫ݐ‬௅் + ܶ‫ܥ‬/60, with ܶ‫ = ܥ‬4(߶ െ ‫ )ܯܶܵܮ‬+ ‫ܶ݋ܧ‬. Here, LSTM is the Local Standard Time Meridian, given by ‫ = ܯܶܵܮ‬15௢ (‫ݐ‬௅் െ ܷܶ‫ )ܥ‬where UTC is the Universal Time Coordinated and EoT is the equation of time, which corrects for the eccentricity of the Earth's orbit and the Earth's axial tilt, ‫ = ܶ݋ܧ‬9.87 sin(2‫ )ܤ‬െ 7.53 cos(‫ )ܤ‬െ 1.5 sin(‫)ܤ‬, where ‫ ݊݀( = ܤ‬െ 81) 360/365 and ݀݊ is the day number (Spencer 1971). Finally, the solar irradiance on the horizontal surface can be fully obtained from the day number, latitude, and hour angle (since the solar declination is already given in terms of the day number) as ‫ܪ‬௢ (݀݊, ߶, ߱) = ‫ܪ‬௢ୄ (݀݊) (sin ߜ sin ߶ + cos ߜ cos ߶ cos ߱)

(10)

Integrating ‫ܪ‬௢ over two hour angles ߱ଵ and ߱ଶ that correspond to a time duration of one hour, centered around ߱௢ , gives the hourly solar irradiation ‫ܪ‬௢ ு௢௨௥௟௬ , which we will derive in section 4 to be ‫ܪ‬௢ ு௢௨௥௟௬ (݀݊, ߶, ߱௢ ) = ‫ܪ‬௢ୄ (݀݊) ൤sin ߜ sin ߶ +

24 ߨ sin cos ߜ cos ߶ cos ߱௢ ൨ (11) 24 ߨ

The daily irradiation ‫ܪ‬௢ ஽௔௜௟௬ can be obtained by integrate ‫ܪ‬௢ ு௢௨௥௟௬ between the sunrise and sunset hour angles, which as we will derive in section 4 is given by ‫ܪ‬௢ ஽௔௜௟௬ (݀݊, ߶, ߱௦ )

24 ߨ ‫ܪ‬௢ୄ (݀݊) ቂ ߱௦ sin ߜ sin ߶ ߨ 180 + cos ߜ cos ߶ sin ߱௦ ቃ (12) =

The hour angle at sunrise and sunset ±߱௦ can be obtained by setting ߠ௭ = 0 in equation (9), which gives cos ߱௦ =

െ sin ߜ sin ߶ cos ߜ cos ߶

,

߱௦ = cosିଵ (െ tan ߜ tan ߶)

(13)

Chapter One

18

where the hour angle of sunrise is +߱௦ and that of sunset is െ߱௦ . The length of day, or duration of sunshine under a clear sky, is therefore 2߱௦ , which can be converted from degrees to hours as follows (as Earth rotates by 360° in 24 hours) ‫ = )ݎ݄( ݕܽܦ ݂݋ ݄ݐ݃݊݁ܮ‬2߱௦ ൬ =

24 ݄‫ݎ‬ 2 ൰= ߱ 360° 15 ௦

2 cos ିଵ (െ tan ߜ tan ߶) (14) 15

3.3 Effect of the Atmosphere The propagation of solar radiation to Earth’s surface through the atmosphere results in a number of effects. As discussed earlier, the atmosphere will reflect back to space about 23% and absorb another 23% while 54% will reach Earth’s surface. Fig. 1-2 shows a reduction in the solar irradiance due to atmospheric absorption, scattering and reflection. A modification of the spectral distribution due to selective absorption and scattering by the atmospheric gases and particles also takes place. Furthermore, the atmosphere introduces a diffused or indirect component to the solar radiation reaching Earth’s surface. These effects will clearly depend on the path of the radiation through the atmosphere. A normal incidence (shortest path) will incur the least of these effects while a shallow incidence (longer path) will incur the most. To describe this mathematically, we define the air mass (AM) as ‫= ܯܣ‬

1 cos ߠ௭

(15)

where ߠ௭ is the same solar zenith angle. As shown in Fig. 1-9, at the top of the atmosphere ‫ = ܯܣ‬0 and for a normal incidence at Earth’s surface ‫= ܯܣ‬ 1 (ߠ௭ = 0) and decreases for an oblique incidence (ߠ௭ > 0).

Solar Irradiation and its Variability

19

Fig. 1-9: Illustration of the Air Mass (AM) for a simplified flat atmosphere (top) and a curved atmosphere (bottom). The expression ‫ = ܯܣ‬1/ cos ߠ௭ assumes a flat atmosphere and is only approximate.

However, since the atmosphere is rather curved and not a flat layer as implied by the simplified AM expression (equation (15)), the air mass is not simply the same as the atmospheric path length, especially when the Sun is close to the horizon. For example, at sunrise and sunset ߠ௭ is nearly 90°, which would lead to an infinite air mass while the path length is finite. An expression that takes into account the curvature of Earth and the atmosphere is given by (Kasten & Young 1989) ‫= ܯܣ‬

1 cos ߠ௭ + 0.50572(96.07995 െ ߠ௭ )ିଵ.଺ଷ଺ସ

(16)

Chapter One

20

Taking this into account, the solar irradiance due to direct sun light on a horizontal plane parallel to Earth’s surface is given by ‫ܪ‬௦ (݀݊, ߶, ߱, ‫ܪ = )ܯܣ‬௢ (݀݊, ߶, ߱) ݁ ି఑ ஺ெ

(17)

where ‫ܪ‬௢ (݀݊, ߶, ߱) is the irradiance at the top of the atmosphere given by equation (10) and ߢ is the atmospheric extinction coefficient due to Rayleigh scattering by gases in the atmosphere, scattering by particulate matter (aerosols), and molecular absorption.

3.4 Effect of Cloud Cover We now consider the impact of clouds on the solar irradiance. Cloud cover poses a considerable impact on the solar irradiance at Earth’s surface. Empirical models and time series analysis are used with meteorological data to estimate the average daily solar irradiation under cloudy skies and to forecast the next hour solar irradiance (Dazhi, Jirutitijaroen, & Walsh 2012; Nimnuan & Janjai 2012). The effect of the cloud cover will result in shortening the duration of sunshine (or length of day) on a given day. If ܵ௢ is the clear sky sunshine duration (i.e., corresponding to the daytime length in hours = 2 ߱௦ /15) and ܵ is the reduced sunshine duration due to cloud cover, we define the cloud fraction ‫ܥ‬௙ as the fraction of the daytime that the sky is obscured by clouds. ‫ܥ‬௙ can be obtained from the relation ܵ = ܵ௢ ൫1 െ ‫ܥ‬௙ ൯,

‫ݎ݋‬

‫ܥ‬௙ = (1 െ ܵΤܵ௢ )

(18)

The daily solar irradiation under a cloudy sky, ‫ܪ‬௖ ஽௔௜௟௬ , can be written as ‫ܪ‬௖ ஽௔௜௟௬ (݀݊, ߶, ߱௦ , ‫ܪ = )ܯܣ‬஽௔௜௟௬ (݀݊, ߶, ߱௦ , ‫ )ܯܣ‬ൣ1 െ ‫ܥ‬௙ (1 െ ܶ)൧ (19) where ܶ is a parameter that gives the fraction of solar radiation transmitted through the clouds (Mani et al. 1967). The cloud fraction ‫ܥ‬௙ can be obtained from direct empirical measurements of the actual clear sky duration of the day ܵ and/or meteorological data. The variability of cloud cover during the day will impact the hourly solar irradiation in a similar fashion. To account for the different degrees of cloudiness (averaged over one hour), e.g., partially cloudy skies and passing and scattered clouds, we define the Clearness Index ‫ܥ‬௜ , which is the ratio of the observed irradiance at a given time of the day to the clear-sky irradiance at the same time.

Solar Irradiation and its Variability

21

Fig. 1-10: Top: The variability of the solar irradiance at Earth’s surface due to different values of the Clearness Index ‫ܥ‬௜ . Small and large values of ‫ܥ‬௜ yield small variability while mid-range values result in the largest variability. Bottom: An illustration of the sky conditions at the different values of the Clearness Index from heavy clouds (‫ܥ‬௜ = 0 െ 0.2) on the far left to a clear sky (‫ܥ‬௜ = 0.8 െ 1) on the far right.

The impact of the cleanness index on the solar irradiance variability is illustrated in Fig. 1-10. The low and high cleanness indices (i.e., due to heavy clouds and clear skies, respectively) produce small variability in the solar irradiance whereas the moderate or mid-range cleanness index (i.e., partly cloudy skies) gives rise to large variability in the irradiance. Using historical time-series data, hourly variability in the solar irradiance can be calculated and validated for the corresponding clearness indices. Using numerical weather prediction (NWP) models, hourly solar energy forecasts can be obtained (Badescu & Dumitrescu 2014; Rangarajan, Swaminathan & Mani 1984). The hourly solar irradiation under a variable cloudy sky, ‫ܪ‬௖ ு௢௨௥௟௬ , can be written as ‫ܪ‬஼ ு௢௨௥௟௬ (݀݊, ߶, ߱௢ , ‫ܪ ݅ܥ = )ܯܣ‬௢ ு௢௨௥௟௬ (݀݊, ߶, ߱௢ , ‫)ܯܣ‬

(20)

22

Chapter One where ‫ܥ‬௜ is the hourly averaged cleanness index. Fig. 1-11 shows two scenarios of the irradiance variability with different cleanness indices. The forecast irradiance is represented by the orange line and the measured irradiance is represented by the green line. The variability forecast is illustrated by the gray area and the yellow line represents the clear sky irradiance. Very small variability is observed for the clear sky (left) with high cleanness index ‫ܥ‬௜ but larger variability is observed for the partly cloudy sky (right) with mid-range cleanness index ‫ܥ‬௜ .

Fig. 1-11: Comparison of the observed hourly irradiation (green line), forecast irradiation (orange line), forecast variability (gray area), and the clear sky irradiation (yellow line) under clear sky (left) and partly cloudy conditions (right).

3.5 Effect of the Solar Cycle We have so far examined variabilities in the solar irradiance due to Earth’s orbit, rotation, axial tilt, curved surface, and atmospheric processes. We now look at another source of variability due to the Sun itself. Careful observations showed that the intrinsic irradiance from the Sun at a specific distance outside Earth’s atmosphere, e.g., at 1 AU, i.e., the solar constant ‫ܪ‬ௌ஼ varies slightly over long periods of time (see Fig. 1-12) (Yeo et al. 2017). From sections 2 and 3, this implies a change in the intrinsic solar ଶ power at the solar surface (ܲௌ௨௡ = ‫ܪ‬௦௨௡ 4ߨܴௌ௨௡ ). We have assumed that the Sun is an ideal thermal emitter with uniform surface temperature and that its entire surface area radiates equally as described by Planck’s function. However, the observed variabilities were found to correlate with the cycle of sunspots on the surface of the Sun, which is nearly periodic with a period of ~11 years (see Fig. 1-13) (Utomo 2017). Sunspots are small dark regions on the solar surface whose sizes vary from ~16 km to ~160,000 km

Solar Irradiation and its Variability

23

(i.e., 12 times larger than the size of the Earth). They can grow and shrink in size over time as they move across the surface of the Sun. Sunspots are darker than the rest of the solar surface because they have lower temperatures due to the increased magnetic field flux at their regions, which channel the hot plasma gas away from those regions. The visibility of sunspots on the solar disk can change over the course of days to weeks due to the rotation of the Sun, which varies by solar latitude, and is about 25 days at the solar equator. The variation in the solar constant ‫ܪ‬ௌ஼ has been consistently monitored from space by radiometers aboard satellite observatories since the late 1970s. Fig. 1-12 shows the variability in ‫ܪ‬ௌ஼ during the period 1978-2015 by different observatories. During the past 40 years, the total solar irradiance has changed by only 0.12% (Finsterle 2020).

Fig. 1-12: The variability of the total solar irradiance (or the solar constant) at 1 AU at the top of Earth’s atmosphere as measured by the space observatories VIRGO, ACRIM I and II during the period 1978-2015. The green horizontal line is set at 1366 W m-2. Source: World Radiation Centre, Switzerland 2016.

24

Chapter One

Fig. 1-13: The correlation between the variability of the total solar irradiance and the solar cycle measured by the number of sunspots across the Sun between 1975 and 2015. Source: Wikimedia Commons 2013, Hansen, J., et al. 2013.

3.6 Other Long-Term Effects While previous effects vary the solar irradiance on time scales that range from hours to years, very long-term variabilities on time scales of thousands of years occur due to the accumulation of subtle changes, mainly in the eccentricity of Earth's orbit and its axial tilt (Meyers and Malinverno 2018; Laskar, J., Fienga, A., Gastineau, M. and Manche, H. 2011). This is known as the Milankovitch cycles (see Fig. 1-14), which account for very longterm climate variability such as an ice age. These changes are mainly due to small perturbations from large planets such as Jupiter and Saturn and torque on Earth from the Sun. A more elliptical orbit would result in more pronounced changes in the solar irradiance during the year, which will also vary by latitude. While Earth’s orbital eccentricity is currently near its smallest value (least elliptical and most circular), which still causes the solar irradiance to vary by 6.9% between the closest and furthest approaches, the highest eccentricity of the orbit (most elliptical) will result in a change in the solar irradiance by 47% between the closest and furthest approaches.

Solar Irradiation and its Variability

25

The cycle between the minimum and maximum eccentricity of the orbit spans about 100,000 years (see the top panel of Fig. 1-14). Similarly, a change in Earth’s axial tilt will cause the solar declination ߜ to change as well. The axial tilt varies between 22.1° and 24.5° over a period of about 41,000 years. Currently, Earth’s rotational axis is tilted by 23.45° with respect to the normal to the plane of its orbit (see the middle panel of Fig. 1-14). Finally, the Earth’s orbit around the Sun does not trace an identical ellipse each year, but shifts slightly in what is known as orbital precession, precession of the perihelion, or apsidal precession (see the bottom panel of Fig. 1-14). This is the result of a precession in the semimajor axis of the elliptical orbit within the orbital plane due to perturbations by the gravitational force from the other planets, among other effects such as anomalies in the Sun’s gravitational field due to its oblateness. This precession has a period of 19,000 to 23,000 years and results in shifting the seasons, i.e., solstices and equinoxes and the day number of Earth's perihelion (݀௢ ), throughout the year.

Fig. 1-14 Long-term variability in the ellipticity of Earth’s orbit, axial tilt, and precession of the orbit, known as Milankovitch cycles. The y-axis represents the relative magnitude of the variability within the ranges discussed in the text. The effects of these subtle changes take tens of thousands of years to be measurable and they result in significant changes in the amount of solar irradiation arriving at Earth.

Chapter One

26

4. The Hourly and Daily Solar Irradiation 4.1 Horizontal Surface The solar irradiance at a horizontal surface parallel to the ground at an arbitrary point with latitude ߶ is given by ‫ܪ‬௢ (݀݊, ߶, ߱) in equation (10) in terms of the perpendicular component ‫ܪ‬௢ ୄ (݀݊) given in equation (7). ‫ܪ‬௢ is completely determined by the day number dn, latitude ߶, hour angle ߱, and solar declination ߜ. The atmospheric correction can be applied using equation (17). Let us calculate the energy received per unit area, or irradiation, ݀‫ܪ‬௢௛ (݀݊, ߠ௭ ) in a given time interval dt (measured in hours). ݀‫ܪ‬௢ (݀݊, ߶, ߱) = ‫ܪ‬௢ୄ (݀݊) (sin ߜ sin ߶ + cos ߜ cos ߶ cos ߱) ݀‫ݐ‬ Since the local time is already expressed as an angular value (of the Sun) in terms of the hour angle ߱, we can also express dt in terms of ߱ using ݀߱ 12 ݀‫ݐ‬ = ‫= ݐ݀ ݎ݋‬ ݀߱ ߨ 24 ݄‫ ݎ‬2ߨ ‫݀ܽݎ‬ 12 ‫( )݊݀( ܪ‬sin ߜ sin ߶ + cos ߜ cos ߶ cos ߱) ݀߱ ߨ ௢ୄ

݀‫ܪ‬௢ (݀݊, ߶, ߱) =

Integrating this expression over two hour angles ߱ଵ and ߱ଶ (in degrees) that correspond to a period of time ‫ݐ‬ଶ െ ‫ݐ‬ଵ gives ‫ܪ‬௢ (݀݊, , ߶, ߱) = න

ఠమ

ఠభ

12 ‫( )݊݀( ܪ‬sin ߜ sin ߶ + cos ߜ cos ߶ cos ߱) ݀߱ ߨ ௢ୄ

‫ܪ‬௢(௧మି ௧భ) (݀݊, ߶, ߱)

ߨ 12 ‫ )݊݀( ܪ‬ቂ( ߱ଶ െ ߱ଵ ) sin ߜ sin ߶ 180 ߨ ௢ୄ + cos ߜ cos ߶ (sin ߱ଶ െ sin ߱ଵ )ቃ =

(21)

To get the hourly irradiation over an arbitrary period of one hour centered around a given hour angle ߱௢ , we integrate ݀‫ܪ‬௢ from ߱௢ െ ߨ/24 to ߱௢ + ߨ/24, which gives

Solar Irradiation and its Variability

27

‫ܪ‬௢ ு௢௨௥௟௬ (݀݊, ߶, ߱௢ ) = ‫ܪ‬௢ୄ (݀݊) ൤sin ߜ sin ߶ +

24 ߨ sin cos ߜ cos ߶ cos ߱௢ ൨ 24 ߨ

(22)

To obtain the daily irradiation ‫ܪ‬௢ ஽௔௜௟௬ , we integrate ‫ܪ‬௢ ு௢௨௥௟௬ between the sunrise and sunset hour angles ‫ܪ‬௢ ஽௔௜௟௬ (݀݊, ߶, ߱௦ ) = න

ఠೄೠ೙ೞ೐೟

‫ܪ‬௢ ு௢௨௥௟௬ (݀݊, ߶, ߱) ݀߱

ఠೄೠ೙ೝ೔ೞ೐

where ߱ௌ௨௡௥௜௦௘ and ߱ௌ௨௡௦௘௧ are given by ±߱௦ obtained earlier. The integration can be written as ‫ܪ‬௢ ஽௔௜௟௬ (݀݊, ߶, ߱௦ ) = 2 න

ఠೞ

‫ܪ‬௢ ு௢௨௥௟௬ (݀݊, ߶, ߱) ݀߱

଴

‫ܪ‬௢ ஽௔௜௟௬ (݀݊, ߶, ߱௦ )

12 ‫( )݊݀( ܪ‬sin ߜ sin ߶ ߨ ௢ୄ ଴ + cos ߜ cos ߶ cos ߱) ݀߱

= 2න

ఠೞ

‫ܪ‬௢ ஽௔௜௟௬ (݀݊, ߶, ߱௦ )

24 ߨ ‫ܪ‬௢ୄ (݀݊) ቂ ߱௦ sin ߜ sin ߶ ߨ 180 + cos ߜ cos ߶ sin ߱௦ ቃ =

(23)

The variation of the daily irradiation through the year at different latitudes is shown in Fig. 1-15.

28

Chapter One

Fig. 1-15: Top: Variation of the daily solar irradiation during the year on a horizontal surface for different latitudes. Bottom: Variation of the average yearly irradiance by latitude.

4.2 Tilted Surfaces We now consider the general case of a tilted surface in the northern hemisphere with the surface making an angle ߚ from the horizontal toward the equator (i.e., due south) and having an azimuth angle ߛ with respect to the south toward the west as shown in Fig. 1-16.

Solar Irradiation and its Variability

29

Fig. 1-16: Top: An arbitrary tilted surface in the northern hemisphere inclined at an angle ߚ from the horizontal toward the equator and at an azimuth angle ߛ with the south direction toward the west. Bottom: A surface tilted by an angle ߚ at latitude ߶ is parallel and hence equivalent (with respect to the incidence angle ߠ) to a horizontal surface at latitude ߶ െ ߚ.

4.2.1 A Surface Tilted at an angle ࢼ and Oriented South (ࢽ = ૙) Let us first consider the case with no azimuth tilt (i.e., ߛ = 0) with the surface facing the south direction. The solar radiation makes an angle ߠ with the normal to the surface as shown in Fig. 1-16 and the irradiance in this case is given by ‫ܪ‬௢ఉ (݀݊, ߶, ߱) = ‫ܪ‬௢ୄ (݀݊) cos ߠ

Chapter One

30

Since the surface at latitude ߶ tilted by an angle ߚ toward the equator is equivalent (i.e., has the same angle of incidence) to a horizontal surface at latitude (߶ െ ߚ) as shown in Fig. 1-16, the solar zenith angle (ߠ) at the latitude ߶ is equal to the zenith angle (ߠ௭ ) at the latitude (߶ െ ߚ) cos ߠ = cos ߠ௭ = sin ߜ sin(߶ െ ߚ) + cos ߜ cos(߶ െ ߚ) cos ߱ఉ where ߱ఉ is the hour angle with respect to the tilted surface. The solar irradiance on the tilted surface becomes ‫ܪ‬௢ఉ ൫݀݊, ߶, ߱ఉ ൯ = ‫ܪ‬௢ୄ (݀݊) ൫sin ߜ sin(߶ െ ߚ) + cos ߜ cos(߶ െ ߚ) cos ߱ఉ ൯ The hourly solar irradiation over an arbitrary period of one hour centered around a given hour angle ߱ఉ௢ is given by ‫ܪ‬௢ఉ ு௢௨௥௟௬ ൫݀݊, ߶, ߱ఉ௜ ൯ = ‫ܪ‬௢ୄ (݀݊) ൣsin ߜ sin(߶ െ ߚ) + cos ߜ cos(߶ െ ߚ) cos ߱ఉ௢ ൧

(24)

Fig. 1-17 shows the variation of ‫ܪ‬௢ఉ ு௢௨௥௟௬ during the day in the different months of the year for a surface tilted at 32° at latitude 30°. The hourly irradiation between two hour angles ߱ఉଵ and ߱ఉଶ can be written as ‫ܪ‬௢ఉ ு௢௨௥௟௬ ൫݀݊, ߶, ߱ఉ ൯ ߨ 12 ‫ܪ‬௢ୄ (݀݊) ቂ(߱ఉଶ െ ߱ఉଵ ) sin ߜ sin(߶ െ ߚ) = 180 ߨ + cos ߜ cos(߶ െ ߚ) (sin ߱ఉଶ െ sin ߱ఉଵ )ቃ Note that the hour angles ߱ఉ , ߱ఉଵ and ߱ఉଶ should be within the sunrise and sunset hour angles of the tilted surface (± ߱௦ఉ ), which, due to the tilt angle, will differ from those for a horizontal surface (߱௦ ). The new hour angle can be obtained from the solar zenith angle equation by setting ߠ = 90° (for sunrise). cos ߱௦ఉ =

െ sin ߜ sin(߶ െ ߚ) cos ߜ cos(߶ െ ߚ)

‫߱ ݎ݋‬௦ఉ = cosିଵ (െ tan ߜ tan(߶ െ ߚ))

Solar Irradiation and its Variability

31

Fig. 1-17: Variation of the hourly solar irradiation ‫ܪ‬௢ఉ ு௢௨௥௟௬ by time of day (in local solar time) through the year on a horizontally tilted surface (ߚ = 30°, ߛ = 0°) facing south at latitude ߶ = 30°.

To obtain the daily irradiation on the tilted surface ‫ܪ‬௢ఉ ஽௔௜௟௬ , we integrate ‫ܪ‬௢ఉ ு௢௨௥௟௬ between the sunrise and sunset angles ߱௦ఉ . Similar to the horizontal surface, we get ‫ܪ‬௢ఉ ஽௔௜௟௬ ൫݀݊, ߶, ߱௦ఉ ൯ ߨ 24 ‫ )݊݀( ܪ‬ቂ߱௦ఉ sin ߜ sin(߶ െ ߚ) = 180 ߨ ௢ୄ + cos ߜ cos(߶ െ ߚ) sin ߱௦ఉ ቃ

(25)

4.2.2 A Generally Tilted Surface in the Horizontal and Azimuth Directions We finally consider the most general case where the surface is tilted at an angle ߚ from the horizontal toward the equator and at an angle ߛ from the south toward the west as shown in Fig. 1-16. The trigonometric relation between the angle of incidence ߠ relative to the normal to the surface and the tilt angles ߚ and ߛ can be expressed in terms of ߶, ߜ, and ߱ఉఊ as follows (Kalogirou 2014), cos ߠ = sin ߜ (sin ߶ cos ߚ െ cos ߶ sin ߚ cos ߛ) + cos ߜ cos ߱ఉఊ (cos ߶ cos ߚ + sin ߶ sin ߚ cos ߛ) + cos ߜ sin ߚ sin ߛ sin ߱ఉఊ The hourly irradiation over an arbitrary period of one hour centered around a given hour angle ߱ఉఊ௢ can then be written as

Chapter One

32

‫ܪ‬௢ఉఊ ு௢௨௥௟௬ ൫݀݊, ߶, ߱ఉఊ௜ ൯ = ‫ܪ‬௢ୄ (݀݊) cos ߠ = ‫ܪ‬௢ୄ (݀݊) ൫ൣsin ߜ (sin ߶ cos ߚ െ cos ߶ sin ߚ cos ߛ) + cos ߜ cos ߱ఉఊ௜ (cos ߶ cos ߚ + sin ߶ sin ߚ cos ߛ) + cos ߜ sin ߚ sin ߛ sin ߱ఉఊ௜ ൧൯ (26)

The daily irradiation ‫ܪ‬௢ఉఊ ஽௔௜௟௬ is obtained by integrating ‫ܪ‬௢ఉఊ ு௢௨௥௟௬ between the sunrise and sunset hour angles. The hour angles for sunrise (߱௥ఉఊ ) and sunset (߱௦ఉఊ ) for this generally tilted surface will no longer be equal as in the previous cases due to the loss of symmetry with respect to the path of the Sun, introduced by the azimuth tilt ߛ. The equation of the solar zenith above can be solved algebraically by setting cos ߠ = 0 for sunrise and sunset, which gives ߱௥ఉఊ = cosିଵ ቆ ߱௦ఉఊ = cos ିଵ ቆ

െܾܽ ‫ ט‬ξܽଶ െ ܾ ଶ + 1 ቇ ܽଶ + 1

െܾܽ ± ξܽଶ െ ܾ ଶ + 1 ቇ ܽଶ + 1

where the upper signs are for ߛ > 0 (surface tilted toward the west) and the lower signs are for ߛ < 0 (surface tilted toward the east) and ܽ=

sin ߶ sin ߶ cos ߶ cos ߶ + , ܾ = tan ߜ ൬ െ ൰ tan ߚ sin ߛ tan ߛ tan ߚ sin ߛ tan ߛ

The sunrise or sunset hour angles for the tilted surfaces (߱௥ఉఊ , ߱௦ఉఊ , and ߱௦ఉ ) should always be smaller than the sunrise and sunset hour angles of the horizontal surface with no tilt (߱௦ ). To satisfy this, it is customary to indicate the values of ߱௥ఉఊ and ߱௦ఉఊ as being the minimum of (߱௦ , ߱௥ఉఊ ) and the minimum of (߱௦ , ߱௦ఉఊ ), and similarly for ߱௦ఉ . The daily irradiation ‫ܪ‬௢ఉఊ ஽௔௜௟௬ is obtained by integrating ‫ܪ‬௢ఉఊ ு௢௨௥௟௬ from sunrise to sunset. ‫ܪ‬௢ఉఊ ஽௔௜௟௬ ൫݀݊, ߶, ߱ఉఊ ൯ =න

ఠೞഁം

‫ܪ‬௢ୄ (݀݊) ൣsin ߜ (sin ߶ cos ߚ െ cos ߶ sin ߚ cos ߛ)

ఠೝഁം

+ cos ߜ cos ߱ఉఊ (cos ߶ cos ߚ + sin ߶ sin ߚ cos ߛ) + cos ߜ sin ߚ sin ߛ sin ߱ఉఊ ൧ ݀߱ఉఊ which gives

Solar Irradiation and its Variability

33

(27)

5. Summary and Conclusion While solar energy is more prolific and accessible than other forms of natural renewable energy, it comes with variabilities and constraints that vary with time, geographic location, and weather pattern. The long-term variability due to the change in the Earth-Sun distance and solar declination during the year (sections 3.1 and 3.2) is well understood and modeled. To minimize the variability due to the geographic latitude illustrated in Fig. 115 (resulting from the solar zenith angle ߠ௭ > 0), the solar panels are tilted relative to the horizontal (toward the equator) by an angle (ߚ) that is equal to the geographical latitude (߶) of the location (see section 4.2). To minimize variability due to the daily path of the Sun from east to west, solar panels are oriented facing true south (azimuth tilt ߛ = 0) in the northern hemisphere and are oriented facing true north (azimuth tilt ߛ = 180) in the southern hemisphere (see section 4.2). These horizontal and azimuth tilts result in optimal annual energy production (Gevorkian 2008). However, depending on how the solar panels are used they may be oriented at a different azimuth angle. For example, if there is more demand on power in the afternoon, the panels can be tilted to face the south-west direction. Since the maximum power is harnessed when the Sun’s rays are perpendicular to the surface of the panel (ߠ௭ = 0), utilizing solar tracking systems that allow the panel to follow the Sun in its east-west path during the day and varying elevation during the different seasons will significantly increase the efficiency of the solar panel systems compared to fixed or nontracking systems (Foster 2010). The most challenging variability is that due to cloud cover, which occurs on a short time scale of hours and minutes. Cloud cover can cause many quick changes in the solar irradiance during the day. A reliable short-term solar irradiance forecast is important to integrate solar energy sources effectively. Numerical weather prediction models and the integration of solar energy facilities can be tapped to mitigate the cloud cover variability. Innovative solutions will continue to offer better ways to minimize the uncertainties introduced by the variability in the solar irradiance. New technologies will

34

Chapter One also enhance the efficiency of solar PV systems and continue to lower the costs, thus helping to make solar energy a more reliable and efficient energy alternative. Despite the variabilities discussed, the greater parts of the world receive sufficient sunshine hours that make solar power a viable option. Fig. 1-18 shows the annular sunshine duration of the world where the majority of the populated areas receive more than 2400 sunshine hours per year, with the highest regional hours received by North Africa and the Arabian Peninsula, which makes the utilization of solar energy a prime choice in those regions.

Fig. 1-18: Annual sunshine duration (or sunshine hours) map of the world. It is a general indicator of the cloudiness of a location. Source: The Hague Center for Strategic Studies 2020.

Solar Irradiation and its Variability

35

List of Abbreviations and Symbols AM AU ߙ B ߚ

Air mass (dimensionless) Astronomical Unit (149.60×106 km) Solar elevation angle or altitude (degrees) Magnetic flux density (Tesla or T) Tilt angle of a horizontal plane on Earth’s surface at latitude ߶ toward the equator (degrees) c Speed of light (2.998×108 m s-1) ‫ܥ‬௙ Cloud fraction (fraction of the daytime the sky is obscured by clouds) (hours) ‫ܥ‬௜ Hourly averaged cleanness index (dimensionless) dn Day number of the year ݀௢ Day number of Earth's perihelion ߜ Solar declination angle (degrees) E Electric field (volts per meter, V/m) EoT Equation of time (mm:ss) ࣟ Photon energy (J) F Flux of electromagnetic radiation (energy per unit area per unit time, J m-2 s-1) ߛ Azimuth angle of a plane relative to the south direction toward the west (degrees) h Planck’s constant (6.626×10-34 J s) ‫ܪ‬ Total power density or irradiance (power per unit area, W m-2) Solar irradiance on a horizontal surface at the top of the atmosphere ‫ܪ‬௢ (W m-2) ‫ܪ‬௢ୄ Solar irradiance at the top of the atmosphere on a plane perpendicular to the Sun’s rays (J s-1 m-2 or W m-2) ‫ܪ‬௦ Solar irradiance on a horizontal surface at Earth’s surface with the perpendicular to the surface making an angle ߠ௭ (zenith angle) with the Sun’s rays (W m-2) ‫ܪ‬ௌ஼ Solar Constant (1366.80 W m-2) ‫ܪ‬௦௨௡ Sun’s total radiation flux or power density at its surface (J s-1 m-2 or W m-2) ‫ܪ‬௖ ஽௔௜௟௬ Daily solar irradiation under a cloudy sky (W m-2) ‫ܪ‬௖ ு௢௨௥௟௬ Hourly solar irradiation under a variable cloudy sky (W m-2) ‫ܪ‬௢ ு௢௨௥௟௬ Hourly solar irradiation on a horizontal surface at the top of the atmosphere (W m-2) ‫ܪ‬௢ ஽௔௜௟௬ Daily solar irradiation on a horizontal surface at the top of the atmosphere (W m-2)

Chapter One

36

Solar irradiance on a surface at the top of the atmosphere tilted by angle ߚ relative to the horizontal toward the equator (W m-2) ‫ܪ‬௢ఉ ு௢௨௥௟௬ Hourly solar irradiation on a surface at the top of the atmosphere tilted by angle ߚ relative to the horizontal toward the equator (W m-2) ‫ܪ‬௢ఉ ஽௔௜௟௬ Daily solar irradiation on a surface at the top of the atmosphere tilted by angle ߚ relative to the horizontal toward the equator (W m-2) ‫ܪ‬௢ఉఊ ு௢௨௥௟௬ Hourly solar irradiation on a surface at the top of the atmosphere tilted at an angle ߚ relative to the horizontal toward the equator and at an angle ߛ from the south toward the west (W m-2) ‫ܪ‬௢ఉఊ ஽௔௜௟௬ Daily solar irradiation on a surface at the top of the atmosphere tilted by angle ߚ relative to the horizontal toward the equator and at an angle ߛ from the south toward the west (W m-2) k Boltzmann’s constant (1.38 × 10 -23 J K -1) ߢ Atmospheric extinction coefficient at Earth’s surface (dimensionless) LSTM Local Standard Time Meridian (hh:mm:ss) Ȝ Photon wavelength (m) n Number of photons received per square meter per second (m-2 s-1) Ȟ Photon frequency (Hz or s-1) Total power of the Sun (J s-1 or W) ܲௌ௨௡ ߶ Geographic latitude (degrees) ߰ Solar azimuth angle (degrees) ܴௌ௨௡ Radius of the Sun (m) ܴா௔௥௧௛ିௌ௨௡ Earth-Sun distance S Poynting vector (W m-2) ܵ௢ Clear sky sunshine duration (hours) ܵ Reduced sunshine duration due to cloud cover (hours) ߪ Stefan-Boltzmann’s constant (5.67 × 10-8 W m-2 K-4) T Temperature (Kelvin K or °C) ܶௌ௨௡ Sun average surface temperature (5778 K) ‫ݐ‬௅ௌ் Local solar time (hh:mm:ss) ‫ݐ‬௅் Local time (hh:mm:ss) Zenith angle (degrees) ߠ௭ UTC Universal Time Coordinated (hh:mm:ss) ȝ Permeability (Henry per meter H/m, permeability of vacuum ȝ0 = 4ʌ × 10í7 H/m) ߱ Solar hour angle (degrees) ߱௦ Sunrise (+߱௦ ) and sunset (െ߱௦ ) hour angle for a horizontal surface parallel to Earth’s surface (degrees) ‫ܪ‬௢ఉ

Solar Irradiation and its Variability

߱ఉ ߱௦ఉ ߱ఉఊ ߱௥ఉఊ ߱௦ఉఊ

37

Solar hour angle for a surface tilted by angle ߚ relative to the horizontal toward the equator (degrees) Sunrise (+߱௦ఉ ) and sunset (െ߱௦ఉ ) hour angles for a surface tilted by angle ߚ relative to the horizontal toward the equator (degrees) Solar hour angle for a surface tilted by angle ߚ relative to the horizontal toward the equator and at an angle ߛ from the south toward the west (degrees) Sunrise hour angle for a surface tilted by angle ߚ relative to the horizontal toward the equator and at an angle ߛ from the south toward the west (degrees) Sunset hour angle for a surface tilted by angle ߚ relative to the horizontal toward the equator and at an angle ߛ from the south toward the west (degrees)

References Badescu, V., and A. Dumitrescu. 2014. “New types of simple non-linear models to compute solar global irradiance from cloud cover amount”. Journal of Atmospheric and Solar-Terrestrial Physics 117, 54-70. Carroll, B., and D. Ostlie. 1996, An Introduction to Modern Astrophysics, 2nd Edition. Cambridge: Cambridge University Press. Dazhi, Y., P. Jirutitijaroen, and W. Walsh. 2012. “Hourly solar irradiance time series forecasting using cloud cover index”. Solar Energy 86(12), 3531-3543. Enteria, Napoleon and Aliakbar Akbarzadeh. 2014, Solar Energy Sciences and Engineering Applications. London: Taylor & Francis Group. Finsterle, Wolfgang. 2020. “Solar Constant: Construction of a Composite Total Solar Irradiance (TSI) Time-Series from 1978 to the Present”. Accessed September 7, 2020. https://www.pmodwrc.ch/en/researchdevelopment/solar-physics/. Foster, Robert. 2010. Solar Energy. Boca Raton: Taylor and Francis Group. Gevorkian, Peter. 2008. Solar Power in Building Design. United States of America: McGraw-Hill. Hansen, J., et al. 2013. “Assessing ‘‘Dangerous Climate Change’’: Required Reduction of Carbon Emissions to Protect Young People, Future Generations and Nature” PLOS ONE 8, 12, e81648 Iqbal, M. 1983. An Introduction to Solar Radiation. Toronto: Academic Press. Kalogirou, Soteris A. 2014. Solar Energy Engineering: Processes and Systems, 2nd edition. Oxford: Academic Press.

38

Chapter One Kasten, F., and A. T. Young. 1989. “Revised optical air mass tables and approximation formula.” Applied Optics 28, 4735-4738. Kren, A. C., P. Pilewskie, and O. Coddington. 2017. “Where does Earth's atmosphere get its energy?” Journal of Space Weather and Space Climate 7, A10. Laskar, J., A. Fienga, M. Gastineau, and H. Manche. 2011, “La2010: A new orbital solution for the long-term motion of the Earth.” Astronomy & Astrophysics 532, A89. Looney, Bernard. 2020. Statistical Review of World Energy. London: BP Publications. Meyers, R. Stephen and Alberto Malinverno. 2018. “Proterozoic Milankovitch cycles and the history of the solar system.” PNAS 115(25), 6363-6368. Nimnuan, P., and S. Janjai. 2012. “An approach for estimating average daily global solar radiation from cloud cover in Thailand”. Procedia Engineering 32, 399-406. Rangarajan, S., M. Swaminathan, and A. Mani. 1984. “Computation of solar radiation from observations of cloud cover”. Solar Energy 32(4), 553556. Ritchie, Hannah and Max Roser. 2020. “Our World in Data: Renewable Energy”. Accessed August 8, 2020. https://ourworldindata.org/renewable-energy. Sellers, W. D. 1965. Physical Climatology. Chicago: University of Chicago Press. Shubbak, M. 2019. “Advances in solar photovoltaics: Technology review and patent trends”. Renewable and Sustainable Energy Reviews 115, 109383. Spencer, J. W. 1971. “Fourier series representation of the position of the Sun.” Applied Optics 10, 2569-2571. Utomo, Y. S. 2017. “Correlation analysis of solar constant, solar activity and cosmic ray.” Journal of Physics: Conference Series 817 012045. van der Hoeven, Maria. 2011. Solar Energy Perspectives. Paris: IEA Publications. Williams, D. R. 2018. “NASA Space Science Data Coordinated Archive (NSSDCA)”. Accessed August 12, 2020. https://nssdc.gsfc.nasa.gov. Yeo, Kok Leng et al. 2017. “Solar Irradiance Variability is Caused by the Magnetic Activity on the Solar Surface.” Phys. Rev. Lett. 119, 091102.

CHAPTER TWO TEMPERATURE EFFECT ON PV OUTPUT POWER VARIABILITY RAZZAQUL AHSHAN1 AND ABDULLAH H. AL-BADI1

Abstract Temperature, solar irradiance, cloud, dust, wind, shading, and solar eclipse are the factors that affect the PV system design and its output generation. In weather that is hot and dry, the temperature is one of the crucial variables that influence the PV system design and its energy performance significantly. The design outcomes can lead to an oversized or undersized PV system if the temperature effect is not considered in the design process. The temperature change can also affect PV system performances, such as capacity factors and system efficiency. In addition, a wide range of temperature variation can lead to a significant PV output power variation. This chapter provides evidence about the effect of ambient temperature variation in PV system design and the system performances. This chapter also presents the PV output variability and PV cell temperature variability for a hot and dry location. In addition, the correlation between the PV output power and the PV cell temperature is also showcased as part of the variability analysis. The detailed model of the PV system design, system performance analysis, and variability analysis is outlined in this chapter. Keywords: Solar PV, temperature variability, power variability, variability modeling, system design, derating factor.

1

Sultan Qaboos University, Dept. of Electrical and Computer Engineering, Muscat, Oman, Email (Razzaqul Ahshan): [email protected]

40

Chapter Two

Nomenclature PV PDC Eload Df

PSh Nday

Tcell TA Ssite SNOCT TNOCT

Pmodule ȕ D f ,Temp

Photovoltaic DC power generated by the solar PV panel in kW Energy demand by the load in a year in kWh/year Derating factor due to the temperature and the other components in the system Peak sun hour Number of days in a year PV cell temperature Ambient temperature in 0C Average solar irradiance in kW/m2 Solar irradiance at Standard Testing Condition (STC) in kW/m2 Normal operating cell temperature of the module in 0C Percentage reduction in maximum power output of a module Reduction in maximum power for per 0C temperature variation of the module in %/0C. Derating factor due to the temperature variation

Df

Derating factor due to the nameplate of the PV module, inverters, module discrepancy, DC and AC wiring, and shading Total derating factor

A PV

Area required for PV panel installation in m2

Į SSTC nmodule

Module efficiency Solar irradiance in kW/m2 at STC Number of modules

Amodule

Area of each module, m2

Vmo

Module output voltage

Vmo,STC

PO , array

Module output voltage at STC Reduction in module output voltage for per 0C temperature variation in %/0C PV array output

Prated , PV

Rated power of the PV system in kW

CF Eaeo HT EPF Fspf Fsf Gti

Capacity factor Actual energy output of the PV system in kWh Total number of hours in a period (day, month, or year) Energy performance factor System performance factor Sun factor Total solar irradiation (kWh/m2)

D f , comp

N

Temperature Effect on PV Output Power Variability

K PV

Total number of sun hours in a period (day, month, or year) PV module efficiency

K Tref

PV module efficiency at the reference temperature

Tref k

Reference temperature Time step (minute, hour, or day) Ramp rate of the PV cell temperature

Hsun,T

Rrate ,Tcell

Tcell k Tcell k 1

Cell temperature at the kth time step

tk

Time at the kth time step

t k 1

Time at the (k+1) time step

Rrate, P

Ramp rate of the PV output power

Pk

PV output power at the kth time step in kW

41

Cell temperature at the time step (k+1)

Pk 1

PV output power at the time step (k+1) in kW

V Rrate,T

Standard deviation of the ramp rates data for the PV cell temperature PV cell temperate ramp rate for the ith sample

cell

R



rate ,Tcell i

PRrate,T

Mean of the ramp rates of the PV cell temperature

n Rrate, P

Number of data Standard deviation of the ramp rates data for the PV output power

 Rrate,P i

PV output power ramp rate for the ith sample

PRrate,P

Mean of the ramp rates of the PV output power

r

Correlation coefficient

cell

Introduction With the improvement in technologies and a decrease in their costs, renewable energy system installations and applications are increasing rapidly. Photovoltaic (PV) technology is currently one of the fastestgrowing technologies to extract solar power. PV installation has escalated over the last decade worldwide, which is evident from Fig. 2-1 [1]. However, with the integration of a large number of PV systems, the output power variability for a PV power plant appears to be an issue for power system operation and planning due to the potential impact on reliable power delivery, and operational and grid stability [2]-[3]. Various factors such as temperature, solar irradiance, cloud, dust, wind, shading, size, solar eclipse,

42

Chapter Two

and the number and spatial dispersion of PV systems have a significant influence on PV output variability [6]-[7]. The effect of all the aforementioned factors on PV output variability depends on the sites, where PV systems are installed.

Fig. 2-1: Worldwide cumulative installed capacity of PV systems

Solar irradiation has a direct effect on the output power variability of a PV system [8]. This is because of the wide range of variation in solar irradiation over the total sunshine period in a day. Such a high irradiance variability effect on the PV output power can be reduced by aggregating PV systems that are installed in neighboring sites [9]. Several researchers have concluded that increasing the number of neighborhood sites and spatial dispersion reduces solar irradiance variability and hence, PV generation [10]-[11]. Cloud also creates PV output power variability because the cloud acts as a barrier for the direct components of the solar irradiation [12]. The PV output power can significantly drop depending upon the clearness index or cloud density, as reported in [13]. This effect becomes significant for the countries that have a more extended period of the rainy season or a very intermittent cloud formation. Dust is another environmental factor that has an impact on PV power output variability, which depends upon the amount of dust accumulation on the panel [14]. The dust accumulation, in general, is slower, having a slower rate of decrease in PV output power. However, a 5% PV output power reduction was observed for daylong dust accumulation in the Saudi Arabian environment [15]. Different cleaning techniques such as water, air, module vibration, and a combination of them have been investigated for automated cleaning of the dust accumulation on PV systems to minimize the reduction in PV power generation [15]. A yearly 10% reduction in energy production is reported due to dust accumulation for a desert type PV system in the Omani environment [16].

Temperature Effect on PV Output Power Variability

43

Temperature is one of the environmental conditions that has a direct effect on the PV energy production. In designing a PV system, it is essential to estimate the annual energy production by the system. Many researchers have applied the thermal model to estimate the PV cell temperature during the design process to assess energy production in a year or month. In [17], the effect of different temperature elements such as cell temperature, surface temperature, back end temperature, and ambient temperature for a PV array is studied considering tropical conditions. It has found that each of these temperature elements has a direct or indirect effect on the PV generation. The authors of [18] have studied the temperature effect on the electrical efficiency of the PV system and its output power. This study has reported an inversely proportional relationship between the electrical efficiency of the PV system and the module or cell temperature. Research in [19] indicated that the effect of temperature on the module degradation is much higher than the cumulative irradiation effect, which results in the fill factor loss and short circuit current loss. A comparative study with and without the effect of temperature on the PV systems has been presented in [20]. It has revealed that depending upon the temperature magnitude; the PV output power can decrease 12 to 15%. The effect of temperature on PV module performance has also been investigated using different temperature prediction models for rack-mounted building-integrated PV systems [21]. A study presented in [22] reports that an increase in module temperature between 0 and 240C reduces the module output power by 0-9.09%. The authors of [23] have presented the field test performance analysis of a 10 MW PV power plant, which indicates that as the temperature increases, the PV conversion efficiency decreases, and hence the performance ratio decreases. A recent study in [24] has highlighted the importance of utilizing passive heat dissipation techniques for PV module temperature regulation. It has revealed that the passive cooling techniques can reduce PV module temperature for a range of 1.50 to 340C depending upon the approaches applied. The cell performance is highly sensitive to cell temperature, which depends on ambient temperature, solar irradiance, cell materials and module encapsulation absorption [25]-[26]. Increasing the module temperature will normally decrease the cells’ band gap, resulting in the absorbing of longer wavelength photons and a drop in PV energy generation [27]-[28]. The influence of increased temperature on the power production of roof integrated PV panels was discussed in [29]. The temperature has more effect on integrated PV compared with free standing PV panels. They found that the difference in the energy production ranges between 3 and 4% for cold, moderate and warm climates. For a hot climate, the difference is more than

Chapter Two

44

5%. Good passive ventilation or nonstandard encapsulation materials could be used to reduce the temperature influence. The literature review has revealed that the ambient temperature variation has a significant effect on the PV module cell temperature, which results in a significant reduction in system conversion efficiency. Thus, including the effect of temperature variation in the PV system’s design stage allows the designer to ensure the annual energy production as per the demand or load requirement. However, the change in PV cell temperature results in a reduction in the efficiency and output power of the PV module randomly. Such a phenomenon can result in an adverse effect on the PV output variability if the magnitude and the range of temperature variation are significant during the sunshine period. Thus, the PV output power variability of multiple array or PV plants can be significant, which may pose a challenge to the power system operation and planning. The behavioral study of the PV output power variability due to the temperature variation, which remains unfold, is one of the vital interests of this chapter. Moreover, this chapter presents the detailed design of a PV system for a typical villa style home in Muscat, Oman, and the system performance in terms of the performance indices such as capacity factor, energy performance factor, system efficiency, system performance factor, and sun factor. The following section presents a methodology that describes models of designing a PV system, models for system performance indices, and models for PV output variability. This is followed by a presentation and analysis of the detailed temperature data for a site that is hot and dry. Then, the PV system design outcomes, the performances of the designed system, and the variability analysis of the PV output are discussed in the results and discussion section. This study is then summarized in the concluding section.

Methodology Design of a PV system The output DC power of the PV panel is calculated as [30] PDC

Eload D f PSh N day

(1)

where PDC is the dc power generated by the solar PV panel in kW, Eload is the energy demand by the load in a year in kWh/year, Df is the derating

Temperature Effect on PV Output Power Variability

45

P

factor due to the temperature and the other components in the system, Sh is the peak sun hour, and Nday is the number of days in a year. The PV cell temperature is calculated as [31]

Tcell

§T  200 C · TA  ¨ NOCT ¸¸ u Ssite ¨ SNOCT © ¹

(2)

where TA is the ambient temperature in 0C, Ssite is the average solar irradiance in kW/m2, SNOCT is the solar irradiance at the Standard Testing Condition (STC) in kW/m2, and TNOCT is the normal operating cell temperature of the module in 0C, which is considered as per the module data sheet. The percentage reduction in maximum power output of the module due to the temperature variation is determined as [31]

Pmodule



E u Tcell  250 C



(3)

where ȕ is the reduction in maximum power for per 0C temperature variation of the module in %/0C. Thus, the derating factor, temperature variation is expressed as D f ,Temp

D f ,Temp

1  Pmodule

due to the

(4)

The derating factor due to the other system components, Df,comp can be considered based on the available literature [30]. Such derating factors include the DC rating on the nameplate of the PV module, inverters, module discrepancy, DC and AC wiring, and shading. The total derating factor, Df is determined as Df

D f ,Temp u D f ,comp

(5)

The area required for the PV panel installation is determined as

APV

PDC

D SSTC

(6)

where APV is the area required for PV panel installation in m2, Į is the

Chapter Two

46

module efficiency that can be obtained from the data sheet and SSTC is the solar irradiance in kW/m2 at STC. The number of modules required in the PV panel is determined as

nmodule

APV Amodule

(7)

To design the number of strings required and the number of modules to be connected in a string, the effect of the temperature has been considered in order to compute the module output voltage at the maximum power point. The new module output voltage has been computed as [32]





Vmo ,STC u 1  N Tcell  250 C

Vmo



(8)

where Vmo,STC is the module output voltage at STC, and N is the reduction in module output voltage for per 0C temperature variation in %/0C. The output of the PV array is calculated as [39] PO , array

§S Prated , PV D f ,comp ¨¨ site © SSTC

· ¸¸ ª¬1  E Tcell  Tcell ,STC º¼ ¹





(9)

Using equations (2), (3), and (4), the equation (9) can be expressed as PO , array

§S Prated , PV D f ,comp ¨¨ site © SSTC

· ¸¸ D f ,Temp ¹

(10)

The PV array output in equation (10) can be expressed using equation (5) as PO ,array

§S Prated , PV D f ¨¨ site © SSTC

· ¸¸ ¹

(11)

However, the PV array output can be modeled without considering the effect of temperature as PO , array

§S Prated , PV D f ,comp ¨¨ site © SSTC

· ¸¸ ¹

(12)

Temperature Effect on PV Output Power Variability

47

Equations (11) and (12) are used to determine the PV output power with and without the temperature effect. Equation (11) indicates the PV array output that varies due to both the temperature and the solar irradiance variation, while equation (12) reveals that the PV output power variation is only due to the irradiance variation.

System Performance Indices Model The capacity factor is defined as the ratio of the actual energy output of a PV system to the energy output at the rated power of the system over a period. It reveals the productivity of a particular PV system at a specific site. The capacity factor (CF) is calculated as [33] CF

Eaeo Prated , PV u H T

(13)

where, Eaeo is the actual energy output of the PV system in kWh, Prated,PV is the rated power of the PV system in kW, and HT is the total number of hours in a period (day, month, or year). The energy performance factor (EPF) is defined as the ratio of the system performance factor (Fspf) factor to the sun factor (Fsf) [5].

EPF

Fspf Fsf

(14)

The system performance factor and the sun factor can be expressed as given in (15) and (16), respectively. Fspf

Fsf

Eaeo Prated , PV u H sun ,T Gti S STC u H sun ,T

(15)

(16)

where, Gti and Hsun,T are the total solar irradiation (kWh/m2) and the total number of sun hours in a period (day, month, or year). The PV module efficiency, K PV as a function of the temperature is expressed as [34]-[35]

Chapter Two

48

KTref ª¬1  E Tcell  Tref º¼

KPV

(17)

K E where, Tref is the PV module efficiency at the reference temperature, is the temperature coefficient of power, and Tref is the reference temperature.

System Output Variability Model Ramp rate statistics can be utilized to analyze the PV output power variability. The ramp rate of the PV cell temperature can be calculated as

Tcell k 1  Tcell k

Rrate,Tcell

tk 1  tk

where, k is the time step (minute, hour, or day), the PV cell temperature,

(18) Rrate,Tcell

is the ramp rate of

Tcell k is the cell temperature at the kth time step,

Tcell k 1 is the cell temperature at the time step (k+1), tk is the time at the kth time step, and t k 1 is the time at the (k+1) time step. And, the ramp rate of the PV output power can be determined as

Pk 1  Pk tk 1  tk

Rrate, P

(19)

Rrate, P

is the ramp rate of the PV output power, Pk is the PV output power at the kth time step, and Pk 1 is the PV output power at the time step (k+1). The standard deviation of the ramp rate data for the PV cell temperature

V Rrate,T

cell

is determined as [4]

V Rrate,T



¦in 1 Rrate,Tcell

P i

n 1

cell

(20)

Rrate ,Tcell



2

Temperature Effect on PV Output Power Variability

R

rate ,Tcell

where,

49

is the PV cell temperate ramp rate for the i

th

i

sample,

PRrate,T

is the mean of the ramp rates of the PV cell temperature, and n is the amount of data. cell

Similarly, the standard deviation of the ramp rate data for the PV output power,

Rrate , P

is determined as



¦ in 1  Rrate, P i  P Rrate ,P

V Rrate ,P

R



2

n 1

(21)



PR

rate , P i where, is the PV output power ramp rate for the ith sample, rate ,P is the mean of the ramp rates of the PV output power, and n is the amount of data.

The association between the output power of the PV system and the PV cell temperature is an important aspect for variability analysis. The correlation coefficient between them is determined using (22) as [4]

r



¦ in 1  Rrate, P i  P Rrate ,P



¦ in 1  Rrate, P i  P Rrate ,P



2

 R

rate ,Tcell



P i

¦ in 1 Rrate,Tcell

Rrate ,Tcell

P i



Rrate ,Tcell



2

(22)

The coefficient of determination is calculated by taking the square of the correlation coefficient, r.

Temperature Data Analysis The output power and the efficiency of the PV system directly depend on the ambient temperature of the site, where the PV systems are located. In this analysis, a hot and dry climate such as Muscat, Oman is considered. The measured temperature data are obtained from the echo-house located at the Sultan Qaboos University (SQU), Muscat, Oman.

50

Chapter Two

Fig. 2-2: Hourly temperature profile for each month in the year 2018

Fig. 2-2 demonstrates the hourly temperature variation in each month for the year 2018. A wide range of variation in temperature can be seen in each month. The hottest months are April, May, June, July, August, and September, while the coldest months are January and December. The rest of the months have moderate temperatures. Regardless of the months, it is seen that there is a sharp change in temperature in each hour, which can influence the PV conversion efficiency and hence the PV output. Fig. 2-3 shows the monthly average, maximum, and minimum temperatures for each month. It is clear that the range of temperature variation is nearly the same for every month, regardless of hot, cold or moderate temperatures. Therefore, the months of January (coldest), July (hottest), and November (moderate temperatures) are selected for analyzing PV system performances and system variability analysis.

Temperature Effect on PV Output Power Variability

51

Fig. 2-3: Monthly temperature variation for the year 2018

Results and Discussion A villa style home load in Muscat, Oman, is selected for designing a PV system since the Sultanate of Oman is currently integrating a large number of rooftop grid-connected PV systems to the distribution network. The analysis of PV system design, system performances, and output power variability is explained in this section based on a PV system for home applications in Oman. However, given the models in the methodology section, this analysis can be performed for specific systems and sites, including, large, or small, ground or rooftop mounted PV plants located in hot, cold, or moderate temperature regions. The load consumption for the selected home is found to be 102 kWh/day or 37.23 MWh/year, as obtained from the Muscat Electricity Distribution Company in Oman.

System Design Analysis Two cases are considered in designing the PV system for the selected home. Case I designs the PV system without considering the effect of temperature, and Case II designs the PV system considering the effect of temperature.

Case I: In this case, the PV system is designed without considering the effect of temperature on the PV cells, which considers that the derating D

factor, f ,Temp is 1 due to temperature. Considering the derating factors, Df,comp for the other components is 0.7895, and the peak sun hour is 5.45 hours, and the size of the PV system to support the home load of 37.23 MWh/year was found to be 23.7 kW. The area required for this PV array was calculated and found to be 140 m2, which requires 71 modules.

Chapter Two

52

Case II: In this case, the PV system is designed considering the effect of D

temperature on the PV cells. The average derating factor, f ,Temp due to temperature, is calculated as 0.8749. Considering the derating factors, Df,comp for the other components is 0.7895, and the peak sun hour is 5.45 hours, and the size of the PV system to support the home load of 37.23 MWh/year was found to be 27 kW. The area required for this PV array was calculated and found to be 158 m2, which requires 81 modules. The design outcomes reveal that without considering the temperature effect, the PV system can be undersized, which cannot meet the expected energy demand for the loads.

System Performance Analysis The system performance is analyzed for three different months (January, July, and November) of 2018 and they are selected based on the temperature profile discussed in the preceding section. The system performance indices such as the system output power, the system capacity factor, the energy performance factor, and the efficiency of the system are discussed with and without the effect of temperature. Moreover, the sun factor and the system performance factor are also described. (a)

(b)

Fig. 2-4: Hourly PV output power performance with and without the temperature effect for January 2018

Fig. 2-4(a) shows the hourly output power production of the PV system for January, which is considered to be the coldest month in a year. The hourly production varies due to the variation in solar irradiation. Moreover, it is observed that there is a reduction in the output power in each hour due to the variation of temperature. Fig. 2-4(b) clearly illustrates a 5.07% reduction in peak power production on January 13. Fig. 2-5 illustrates the daily variation in capacity factors of the PV system for January. The daily

Temperature Effect on PV Output Power Variability

53

capacity factors with and without the temperature effect are within the limit that is consistent with the literature. However, the daily capacity factors without the effect of temperature are higher, which is misleading. The actual capacity factors that account for the effect of temperature are lower due to the reduction in efficiency of the PV system. Fig. 2-5(b) demonstrates the energy performance of the PV system with and without the temperature effect. Since temperature affects the efficiency, the energy performance factors decrease with an increase in temperature. The highest temperature has been found on days 27 and 28 of this month, having the lowest energy performance factor. The range of the energy performance factor for January is found to be 0.73–0.76.

Fig. 2-5: Capacity factor and energy performance factor for January 2018: (a) Daily variation in the capacity factor with and without the effect of temperature; (b) Daily energy performance factor with and without the temperature effect

Fig. 2-6(a) illustrates the daily efficiency variation along with the cell temperature variation in each day for January 2018. It is observed that with the increase in cell temperature, the efficiency of the PV system decreases, and vice versa. The highest efficiency is found on day 9 at the lowest cell temperature condition. The highest efficiency results in the system energy

54

Chapter Two

performance factor (Fig. 2-5(b)) being higher. However, the system performance factor, as shown in Fig. 2-6(b), is lowest on this day due to the lower sun factor that is influenced by the total solar irradiation available on the day. One of the reasons for the lower sun factor is the presence of cloudy days.

Fig. 2-6: (a) Cell temperature and efficiency variation in each day for January; (b) Daily variation in sun factor and system performance factor for January

Fig. 2-7(a) presents hourly PV output power generation for July, which is the hottest month in a year. The hourly production varies due to the variation in solar irradiation. The hourly power production is higher compared to January due to the higher solar irradiation. However, it is also observed that there is a significant reduction in the output power generation in each hour due to the rise in ambient temperature. Fig. 2-7(b) clearly illustrates that a 14.25% reduction in peak power production on July 13 of the same year is due to the rise in PV cell temperature.

Temperature Effect on PV Output Power Variability

(a)

55

(b)

Fig. 2-7: Hourly PV output power performance during July 2018

Fig. 2-8: Capacity factor and energy performance factor for July 2018: (a) Daily variation in capacity factors with and without the effect of temperature; (b) Daily variation of the energy performance factor with and without the temperature effect

Fig. 2-8(a) explains the capacity factor variation of the PV system in each day for July 2018. The daily capacity factors with and without the

56

Chapter Two

temperature effect are within the limit that is consistent with the literature. The actual capacity factors that account for the effect of temperature are lower due to the reduction in efficiency of the PV system. However, the daily capacity factors are higher compared to those of January because of the higher solar irradiation and the sun factor in July. Fig. 2-8(b) demonstrates the effect of higher ambient temperature on the energy performance factor of the PV system. The energy performance factors increase with a decrease in temperature and vice versa. Due to the higher cell temperature during this month, the range of the daily energy performance factors (0.67–0.71) is lower compared to the range of the daily energy performance factors (0.73–0.76) for January.

Fig. 2-9: (a) Cell temperature and efficiency variation in each day for July; (b) Daily variation in sun factor and system performance factor for July Fig. 2-9(a) demonstrates the daily variation in system efficiency and the PV cell temperatures for July 2018. With the increase in cell temperatures, the efficiency of the PV system decreases, and vice versa. The reduction in efficiency is significant in July compared to the efficiency reduction in January. The lowest efficiency is found on day 26 at the highest cell temperature condition. However, the lowest

Temperature Effect on PV Output Power Variability

57

system performance factor is observed on day 16, as can be seen from Fig. 2-9(b). This lowest system performance factor results from the fact that the sun factor is low, and it is influenced by the total solar irradiation available on the day. The lower solar irradiation presence on day 16 is confirmed by the daily power output presented in Fig. 2-7(a). (a)

(b)

Fig. 2-10: Hourly PV output power performance for November 2018

The hourly PV output power generation for November 2018 is presented in Fig. 2-10(a). November is found to be a month of moderate temperature for the selected location. The moderate temperature is referred to here as not high, like the temperature in July and not low like the temperature in January. The daily power production for this month varies within the window of power production between January and July. Since the temperature for this month is neither extremely high (like July) nor low (like January), an 8.6% reduction (Fig. 2-10(b)) in peak power generation has been noticed on November 13 of the same year. Such a reduction occurs because of increasing the cell temperature during the PV system operation. Fig. 2-11(a) reveals the capacity factor variation of the PV system in each day for November 2018. With the effect of ambient temperature on the PV cell, the daily capacity factors reduce. The lowest capacity factor is found to be 13.64, and the highest capacity factor is found to be 15.92. Fig. 211(b) demonstrates the effect of moderate ambient temperature on the energy performance of the PV system. The range of energy performance factors for this month is 0.72–0.74. This energy performance factor is consistent with the energy performance factor obtained for the hottest (0.67–0.71) and the coldest (0.73–0.76) months.

58

Chapter Two

Fig. 2-11: Capacity factor and energy performance factor for November 2018: (a) Daily variation in capacity factors with and without the effect of temperature; (b) Daily variation in energy performance factors with and without the temperature effect

Fig. 2-12(a) demonstrates the daily variation in system efficiency and the cell temperature for November 2018. During the first half of November, the ambient temperature is higher compared to the second half of the month. It increases the cell operating temperature in the first half and causes a decrease in the operating cell temperature during the second half of the month. The daily efficiency variation also follows a similar pattern, as can be seen from Fig. 2-12(a). The lowest efficiency is found on day 2 at the highest cell temperature condition. However, the lowest system performance factor is found on day 11, as can be seen from Fig. 2-12(b). This lowest system performance factor results due to the lower value of the sun factor, which is affected by the total solar irradiation available on the day. The lower solar irradiation presence on day 11 is confirmed by the daily power output shown in Fig. 2-10(a).

Temperature Effect on PV Output Power Variability

59

Fig. 2-12: (a) Cell temperature and efficiency variation in each day for November 2018; (b) Daily variation in sun factor and system performance factor for November 2018

Table 2-I illustrates the performance comparison of a PV system for three different months, such as the coldest, the hottest, and the moderate temperature months. The comparison is provided by considering the system performance variation with the variation of temperature in these three months. In July, the monthly average solar irradiance is higher compared to January and November. This leads to an expectation of a better energy performance by the PV system. However, the range of the energy performance factor for this month is lower compared to the other two months. This is because of the lower efficiency of the PV system that results from the higher ambient temperature and cell temperature. The temperature effect is also evident from the reduction of the PV output power, and the amount of the reduction in July is the highest among the three months. However, the capacity factor is higher in July due to the more extended sunshine hours and the higher solar irradiance, which is also evident from the daily sun factor and the system performance factor. The daily capacity factor varies between 0.1021 and 0.22 for January, July, and August, which is expected for the other months in the year. This range of capacity factors

Chapter Two

60

is almost the same as that claimed in [36], where the range of the capacity factors was found to be 0.11 to 0.23 in Oman. It also indicated that the annual average capacity factor for a typical PV plant was found to be 0.1484 in Morocco, 0.1569 in India, 0.1058 in Norway, and 0.122 in Serbia. Table 2-I: Comparison of variables and indices related to the system performance among the coldest, hottest and moderate temperature months Performance variable/indices

January (Coldest month)

July (Hottest month)

November (Moderate temp. month)

Hourly ambient temperature (oC) Daily average cell temperature (oC) Day peak PV output reduction (%) Daily capacity factor Daily energy performance factor Daily efficiency (%) Daily sun factor Daily system performance factor Monthly average solar irradiance (W/m2/day)

14.64–32.29

27.97–46.33

17.11–35.59

32.04–36.52

46.73–56.70

36.52–42.72

5.07

14.25

8.6

0.1021–0.1607

0.137–0.22

0.1364–0.1592

0.737–0.7624

0.679–0.713

0.72–0.74

16.26–16.58 0.2921–0.464 0.2227–0.3502

14.81–15.53 0.3845–0.611 0.2637–0.4238

15.82–16.26 0.4262–0.5042 0.3116–0.3638

396

509

422

Nevertheless, in January, the range of energy performance factors is higher compared to July and November. This is because of the higher efficiency of the PV system with the minimal effect of ambient and cell temperature. The temperature effect is also evident from the reduction of the PV output power; however, the effect is not that significant, as it has been observed for the hottest month. The capacity factor is comparatively low for January due to the lower solar irradiance, sun factor and system performance factor. The temperature and solar irradiance profiles for November fall between the temperature and solar irradiance profiles of January and July. As a result, the performance variables and indices for November remain between the performance variables and indices of January and July.

Temperature Effect on PV Output Power Variability

61

Table 2-II: Similarity in temperature profile for the twelve different months Hourly ambient temperature (oC) Monthly average temperature (oC)

January 14.64–32.29

December 14.59–30.52

February 15.94–34.33

21.15

21.49

23.44 September

April

27.5045.63

August 27.0144.84

26.8544.40

24.0544.28

36.56

33.93

32.69

32.61

July

May

June

Hourly ambient temperature (oC)

27.9746.3

27.944.96

Monthly average temperature (oC)

36.06

35.81

Hourly ambient temperature (oC) Monthly average temperature (oC)

November 17.11–35.59

March 19.11–40.53

October 24.11–41.64

26.39

27.21

31.62

Table 2-II illustrates the similarity in temperature profile for the twelve different months of the year. It reveals that the temperature characteristic for January is similar to that of December and February. Thus, the system performance indices for December and February can be considered similar to those in January with no significant variation. Similarly, the temperature profile for July is comparable to that of April, May, June, August, and September. Therefore, the system performance indices for July are considered analogous to those of April, May, June, August, and September. The same assumption is applicable for March, October, and November. It is important to note that the highest reduction of daily peak PV output power occurs during July. Half of the year follows temperature characteristics like July. Therefore, a significant reduction of power output can result due to the effect of temperature during this period.

System Variability Analysis The system variability analysis is composed of the variability analysis of the PV cell temperature and the PV output power. First, a ramp rate analysis is carried out both for the cell temperature and the PV output power for three

62

Chapter Two

different months. The ramp rate analysis is performed to investigate the individual variability characteristics of these parameters (cell temperature and PV output power) for the selected months. Then, the correlations of the PV output power and the PV cell temperature are evaluated to investigate the associations between them. Finally, the coefficient of determination is determined to demonstrate the level of variability of PV output power due to the variation of the PV cell temperature.

Ramp Rates of PV Cell temperature The effect of the ambient temperature variation has the ultimate effect on the PV cell temperature, and hence the output power variation of the PV system. Thus, the ramp rates for the PV cell temperature are an important aspect. One-minute ramp rates for the PV cell temperature are determined using equation (18). The ramp rate data of the PV cell temperature are fitted to the normal distribution function to obtain its variability characteristics for the coldest, hottest, and moderate temperature months. Fig. 2-13 illustrates the one-minute ramp rate variability for three different months in a year. It is found that the PV cell temperature variability is maximum in January, which is about 4.58%. Such variability results are due to the intermittent clouds during this month. The variability of the PV cell temperature is not maximum; however, it is high in July at 4.19%. The rise in temperature and humidity in July can cause such high variability in the cell temperature. On the other hand, the PV cell temperature ramp rate variability is lower, which is 2.53% in November. It is essential to mention that November is found to be a month of moderate temperature and has more clear days.

Fig. 2-13: Ramp rate variability of PV cell temperatures

Temperature Effect on PV Output Power Variability

63

Ramp Rates of PV Output Power The analysis of the ramp rate variability of the PV output power is performed for three different months in a year. One-minute ramp rates for the PV output power are determined using equation (19). The ramp rate data of the output power are fitted to the normal distribution function to obtain and compare the variability characteristics for the coldest, hottest, and moderate temperature months. Fig. 2-14 demonstrates the probability distribution function of the ramp rates of PV output power. The highest variability in PV output power is found in January, and the lowest PV output variability is found in November. The PV power ramp rate variability in July is lower than that in January; however, it is more than that in November. The PV output power variability in January is 1.13 times higher than it is in July and 1.8 times higher than in November. The power variability in January results from the intermittent clouds, and this is consistent with the temperature variability of the same month. Power variability in July occurs due to the rise in ambient temperature, which results in a decrease in PV system efficiency. Moreover, there is almost zero existence of clouds in July. Power variability is lowest in November because of the bright days and moderate temperature levels.

Fig. 2-14: Ramp rate variability of PV output power

Correlation of PV output Power and Cell Temperature In order to verify the interrelation between the PV output power and the PV cell temperature, the correlation between them is examined. Moreover, the determination of the coefficient is investigated to

Chapter Two

64

observe the variability in PV output power due to the variation in PV cell temperature. Table 2-III shows the correlation coefficient and the coefficient of determination between the output power and the cell temperature for three different months in a year. The coefficients reveal a strong relationship between the PV output power and the cell temperature for each month. However, the lowest coefficient is found in January, and the highest coefficient is in November. The lowest coefficient in January comes from the fact that this month has days with intermittent clouds, which leads to the higher PV output variability during this month, and this is evident from Fig. 2-14. The coefficient of determination for January indicates that 83.17% of the PV output variability results due to the change in cell temperature in this month. On the other hand, the highest coefficient in November results due to the bright days, leading to the lower PV output variability, and this is also shown in Fig. 2-14. The determination of the coefficient for November indicates that 94.09% of the PV output variability occurs because of the variation in cell temperature in this month. Table 2-III: Correlation coefficient between PV output and cell temperature Months

January

July

November

Correlation coefficient (%)

91.2

95.8

97.4

Coefficient of determination (%)

83.17

91.77

94.86

Conclusions This chapter presents a detailed analysis of the site temperature effect on PV system designs, system performances, and variability of the PV output power. This analysis is performed based on a detailed model presented in this chapter. It also presents a statistical analysis of the correlation between the PV cell temperature and the PV output power. The analysis was focused on sites, which have hot and dry climatic conditions. One-year weather and solar irradiation data were measured and collected for a site in Muscat, Oman, and utilized for this analysis. The measured data were recorded every 20 seconds.

Temperature Effect on PV Output Power Variability

65

One-year hourly load data were analyzed for each month in a year. They reveal that July is the hottest month with a range of temperature variation of 27.96–46.330C, and January is found to be the coldest month with a range of temperature variation of 14.64–32.290C. Moreover, the moderate range of temperature variation is found in November to be 17.11–35.990C. They also reveal that April, May, June, August, and September have a similar temperature profile to July, which indicates that the PV system installed in such a location has to operate with an extreme temperature condition for half of the time in a year. The design outcomes show that to meet a 37.23 MWh/year load for a typical villa type home, a 27 kW PV system is required if the effect of temperature is taken into consideration. This system requires 81 modules, and the area required to install the PV modules was calculated to be 158 m2. However, without considering the temperature effect, the system size ends up with 23 kW, which needs 71 modules to be installed in a 140 m2 area. Although the latter size is encouraging in terms of cost-effectiveness and space requirement, this system may not produce the expected energy required by the load due to the derating effect of the high ambient temperature. The lowest range of energy performance factors is found in the hottest month of July, which is 0.679–0.713. The daily energy performance factor varies between 0.679 and 0.7624 for January, July, and August, which is expected for the other months in the year. This range of energy performance factors is almost the same as that claimed in [16], where the range of the performance ratio was found to be 0.37–0.96 (0.67 average) for a desert type PV system in Oman. The range of the performance ratio in Southern Algeria was reported to be 0.67–0.86 [32], whereas the annual average performance factor in northern Ghana was reported to be 0.7 [38]. In July, the range of PV cell temperature variation is highest, causing a decrease in PV system efficiency to the lowest range of 14.81–15.53%. It also reveals that the day peak PV output power reduction is 14.25%, which is highest among the selected three months. However, the range of the capacity factor is still high compared to January and November because of the higher sun factors. The least reduction in the day peak power is observed in January, and it is 5.07%. The range of PV system efficiency in January is highest because of the lower impact of the ambient or cell temperature. This results in the highest range of energy performance factors in January. However, the range of the daily capacity factors for January is the lowest because of the lowest range of sun factors. It is clearly shown that the ambient temperature or the PV cell temperature has a significant effect on

66

Chapter Two

the PV system efficiency and hence the energy performance of the system. One-minute ramp rate variability of the PV output power and the PV cell temperature is carried out and compared using the monthly variability mode. It reveals that the PV output power variability in January, July, and November is 4.28%, 3.8%, and 2.37%, respectively. In addition, the PV cell temperature variability in January, July, and November is 4.58%, 4.19%, and 2.53%, respectively. Both the PV output power and the cell temperature variability are highest in January because of the presence of intermittent clouds. In July, the PV power and cell temperature variability are close to the highest value, and this is because of the rise in temperature. The least variability is found in November, which has bright days with lower temperature. Such high variability in PV output power especially in January and July can cause a major challenge for power system operation and planning. It is because the utility owner needs to accommodate such a ramp rate variation in order to maintain a reliable and stable operation of the power network that has significant PV penetration. In addition, the correlation coefficient reveals a strong relationship between the PV output power and cell temperature. The coefficient of determination indicates that the PV cell temperature causes the highest PV output power variability in July and November. To improve the PV efficiency, the generated heat can be utilized by a thermoelectric generator to produce electricity. Thermoelectric (TE) devices have many advantages such as a solid-state operation, a long life, vast scalability and being environmentally friendly. The feasibility of using PV-TE was discussed in [40].

References 1. 2. 3.

4. 5.

“2019 Renewable Energy Global Status Report”, Accessed March 1, 2020, http://www.ren21.net/gsr_2019_full_report_en. F. Bai, R. Yan, and T. K. Saha. 2017. “Variability study of a utilityscale PV plant in the Fringe of Grid, Australia.” IEEE Innovative Smart Grid Technologies – Asia (ISGT-Asia), Dec. 2017. D. Palmer, E. Koubli, I. Cole, T. Betts, and R. Gottschalg. 2017. “Comparison of solar radiation and PV generation variability: system dispersion in the UK.” IET Renewable Power Generation 11(5), 550557. A. Buttler, F. Dinkel, S. Franz, and H. Spliethoff. 2016. “Variability of wind and solar power – An assessment of the current situation in the European Union based on the year 2014.” Energy 106, 147-161. H. Ibrahim, and N. Anani. 2017. “Variations of PV module parameters

Temperature Effect on PV Output Power Variability

6.

7.

8. 9. 10. 11. 12. 13.

14.

15. 16. 17.

67

with irradiance and temperature.” Energy Procedia 134, 276-285. S. V. Raygani, R. Sharma, and T. K. Saha. 2014. “Analysing the PV output variability and its mitigation through aggregation in Queensland, Australia.” IEEE PES General Meeting, 2014, National Harbor, MD, USA. A. R. Amelia, Y. M. Irwan, W. Z. Leow, M. Irwanto, I. Safwati, and M. Zhafarina. 2016. “Investigation of the Effect of Temperature on Photovoltaic (PV) Panel Output Performance.” International Journal on Advanced Science Engineering and Information Technology 6, 682688. Thomas E. Hoff, and Richard Perez. 2012. “Modeling PV fleet output variability.” Solar Energy 86, 2177-2189. M. Lave, and J. Kleiss. 2010. “Solar variability of four sites across the state of Colorado.” Renewable Energy 35, 2867-2873. J. Remund, C. Calhau, L. Perret, and D. Marcel. 2015. “Characterization of the spatiotemporal variations and ramp rates of solar radiation and PV”. Report, IEAPVPS T14-05:2015I. L. Frearson, P. Rodden, J. Blackwell, et al. 2015. “Investigating the impact of solar radiation on grid stability with dispersed PV generation”. 31st EUPVSEC, 2015, 2981-2987. R. Perez, and T. E. Hoff. 2013. “Mitigating short-term PV output intermittency.” 28th EUPVSEC, 3719-3726. M. Suri, T. Cebecauer, A. Skoczek, R. Marais, C. Mushwana, and J. Reinecke. 2014. “Cloud cover impact on photovoltaic power production in South Africa.” In Proc. of the 2nd South African Solar Energy Conference, Port Elizabeth, South Africa. A. Bonkaney, S. Madougou, and R. Adamou. 2017. “Impacts of Cloud Cover and Dust on the Performance of Photovoltaic Module in Niamey.” Journal of Renewable Energy Volume 2017, Article ID 9107502, 8 pages. A. S. Alghamdi, A. S. Bahaj, L. S. Blunden, and Y. Wu. 2019. “Dust Removal from Solar PV Modules by Automated Cleaning Systems.” Energies 12, 2923. A. H. Al-Badi. 2018. “Measured performance evaluation of a 1.4 kW grid connected desert type PV in Oman.” Energy for Sustainable Development 47, 107-113. M. E. Ya’acob, H. Hizam, M. A. M. Radzi, and M. Z. A. A. Kadir. 2013. “Field Measurement of PV Array Temperature for Tracking and Concentrating 1 kWp Generators Installed in Malaysia.” International Journal of Photoenergy Volume 2013, Article ID 502503, 8 pages.

68

Chapter Two

18. S. Dubey, J. N. Sarvaiya, and B. Seshadri. 2013. “Temperature Dependent Photovoltaic (PV) Efficiency and Its Effect on PV Production in the World-A Review.” Energy Procedia 33, 311-321. 19. J. Singh, J. Belmont, and G. TamizhMani. 2013. “Degradation Analysis of 1900 PV modules in a Hot-Dry Climate: Results after 12 to 18 years of field exposure.” IEEE 39th Photovoltaic Specialists Conference (PVSC), 2013, Tampa, FL, USA. 20. M. Bayrakci, Y. Choi, and J. R. S. Brownson. 2014. “Temperature Dependent Power Modeling of Photovoltaics.” Energy Procedia 57, 745-754. 21. M. D’Orazio, C. Di Perna, and E. Di Giuseppe. 2014. “Experimental operating cell temperature assessment of BIPV with different installation configurations on roofs under Mediterranean climate.” Renewable Energy 68, 378-396. 22. A. K. Tripathi, M. Aruna, and Ch. S. N. Murthy. 2017. “Output Power Loss of Photovoltaic Panel Due to Dust and Temperature.” International Journal of Renewable Energy Research 7(1), 439-442. 23. M. R. Behiri, M. A. Mohamed, M. J. Al shammri, and H. Rezk. 2019. “Energy Performance Analysis of On-Grid Solar Photovoltaic System – a Practical Case Study.” International Journal of Renewable Energy Research 9(3), 1292-1301. 24. A. M. Elbreki, M. A. Alghoul, K. Sopian, and T. Hussein. 2017. “Towards adopting passive heat dissipation approaches for temperature regulation of PV module as a sustainable solution.” Renewable and Sustainable Energy Reviews 69, 961-1017. 25. S. A. M. Saida, G. Hassana, Husam M. Walwil, and N. Al-Aqeeli. 2018. “The effect of environmental factors and dust accumulation on photovoltaic modules and dust-accumulation mitigation strategies.” Renewable and Sustainable Energy Reviews 82, 743-760. 26. M. M. Rahman, M. Hasanuzzaman, and R. A. Rahim. 2016. “Effects of operational conditions on the energy efficiency of photovoltaic modules operating in Malaysia.” Journal of Clean Production 143, 912-24. 27. K. Kant, A. Shukla, A. Sharma, and P. H. Biwole. 2016. “Thermal response of poly-crystalline silicon photovoltaic panels: numerical simulation and experimental study.” Solar Energy 134, 147-55. 28. K. L. Raviraj, H. R. Purushothama, A. Manjunatha, and Mishra V Kumar. 2016. “Thermal design and evaluation of cooling configuration of solar photovoltaic panel.” Indian Journal of Science & Technology 9(45).

Temperature Effect on PV Output Power Variability

69

29. T. Poulek, M. Matuška, E. Libra, and J. Sedláßcek Kachalouski. 2018. “Influence of increased temperature on energy production of roof integrated PV panels.” Energy & Buildings 166, 418-425. 30. R. Ahshan, R. Al-Abri, Al-Zakwani, and N. Ambu-saidi. 2019. “Solar PV System Design for a Sports Stadium.” IEEE 10th GCC Conference and Exhibition, 19-23 April, 2019, Kuwait. 31. M. C. A. García, and J. L. Balenzategui. 2004. “Estimation of photovoltaic module yearly temperature and performance based on Nominal Operation Cell Temperature calculations.” Renewable Energy 29, 1997-2010. 32. G. M. Masters. 2013. Renewable and Efficient Electric Power Systems. Wiley, IEEE Press. 33. R. Ahshan, A. H. Al-Badi, N. Hosseinzadeh, and M. Shafiq. 2018. “Small Wind Turbine Systems for Application in Oman.” IEEE 5th International Conference on Electric Power and Energy Conversion Systems, April 2018, Kitakyushu, Japan. 34. E. Skoplaki, and J. A. Palyvos. 2009. “On the temperature dependence of photovoltaic module electrical performance: A review of efficiency/power correlations.” Solar Energy 83(5), 614-624. 35. D. L. Evans, and L. W. Florschuetz. 1977. “Cost studies on terrestrial photovoltaic power systems with sunlight concentration.” Solar Energy 19(3), 255-262. 36. A. H. Al-Badi. 2019. “Performance assessment of 20.4 kW eco-house grid-connected PV plant in Oman.” International Journal of Sustainable Engineering, DOI: 10.1080/19397038.2019.1658824. 37. A. Necaibia, A. Bouraiou, A. Ziane, N. Sahouane, S. Hassani, M. Mostefaoui, R. Dabou, and S. Mouhadjer. 2018. “Analytical Assessment of the Outdoor Performance and Efficiency of Grid-tied Photovoltaic System under Hot Dry Climate in the South of Algeria.” Energy Conversion and Management 171, 778-786. 38. D. Mensah, O. Yamoah, and S. Adaramola. 2019. “Performance Evaluation of a Utility-scale Grid-tied Solar Photovoltaic (PV) Installation in Ghana.” Energy for Sustainable Development 48, 82-87. 39. https://www.homerenergy.com/products/pro/docs/latest/how_homer_ calculates_the_pv_array_power_output.html. Accessed, March 28, 2020. 40. G. Li, S. Shittu, T. M. O. Diallo, M. Yu, X. Zhao, and J. Ji. 2018. “A review of solar photovoltaic-thermoelectric hybrid system for electricity generation.” Energy 158, 41-58.

CHAPTER THREE VARIABILITY ANALYSIS OF PV SYSTEM OUTPUT FAIZA AL-HARTHI1,2, MOHAMMED ALBADI2, RASHID AL ABRI2, AND ABDULLAH AL BADI2 Abstract The wide-scale integration of solar PV power might bring concerns regarding PV output variability and its challenges to transmission and distribution networks. Aggregating the output of different PV systems can reduce the negative implications of such variability on the grid. This chapter presents a case study of PV system output variability on a distribution network. The study uses 15-minute data obtained from three PV systems connected to the Petroleum Development Oman (PDO) – Mina Al-Fahal (MAF) distribution network. PV system output variability is classified into two categories: deterministic and stochastic. The deterministic variability is linked to clear sky irradiation variability while the stochastic one is linked to intermittent cloud coverage. Although the deterministic variability is known ahead of time, it may have severe implications on PV system output as demonstrated during the annular eclipse that occurred on 21 June 2020. As cloud fronts do not cover different PV plants simultaneously, the aggregated PV system output is less variable than that from a single PV system. This study demonstrates that the smoothing effect is dependent on the number as well as the geographical diversity of PV systems. Keywords: Solar PV output variability, deterministic variability, stochastic 1

Petroleum Development Oman, Muscat, Oman, Email (Faiza Al Harthy): [email protected] 2 Sultan Qaboos University, Dept. of Electrical and Computer Engineering, Muscat, Oman, Email (Mohammed Albadi): [email protected]

Variability Analysis of PV System Output

71

variability, annular eclipse.

Introduction The output of a solar PV system is variable due to the variability of solar irradiation received by the solar panels. The amount of solar irradiation varies with the rotation of the earth around its axis and around the sun in the ecliptic plane. This movement is precisely known. The rising and setting of the sun can cause 10–13% of changes in the solar PV output during 15 minutes for a single axis tracking solar PV system (Mills 2009). However, the effect of the meteorological conditions, such as cloud coverage, on a solar collector is intermittent. Hence, understanding the meteorological conditions while analyzing the solar PV output variability is important to quantify the potential impacts of large solar PV systems (Mills 2009). Variability is quantified as changes from one averaging interval to the next one. In Hoff and Perez (2010), the authors define the PV system variability as the number of power output changes within a defined time. It is the change in output, rather than the output itself, for different time scales. Therefore, PV output variability has a magnitude and sign (Hoff and Perez 2010). There are two categories of solar irradiance and solar PV output variabilities: deterministic and stochastic. Deterministic variability is linked to the position of the sun with respect to solar collectors while the stochastic variability is linked to cloud coverage. The deterministic component can be identified by the daily, monthly and yearly changes perceived at a specific PV system location. The deterministic PV system output variability of different PV systems placed in a similar geographical location is highly correlated. Therefore, aggregating the output of different PV systems located in the same or adjacent geographical areas has a very limited smoothing effect (Albadi 2016; Van Haaren, Morjaria, and Fthenakis 2014; Frearson et al. 2015; Parent et al. 2019; Hoff 2012; Hoff and Perez 2010). Stochastic variabilities are due to changes in weather parameters such as sizes and types of clouds, aerosols, wind speed, temperature and humidity. Moreover, the main source of the stochastic solar irradiance variability is passing clouds. They can cause significant changes in PV system output; consequently, they can affect power system operation (Albadi 2016; Van Haaren, Morjaria, and Fthenakis 2014; Frearson et al. 2015; Parent et al. 2019; Hoff 2012; Hoff and Perez 2010). These changes occur over short time periods, seconds to hours, depending on weather conditions (Albadi 2016; Van Haaren, Morjaria, and Fthenakis 2014; Frearson et al. 2015;

Chapter Three

72

Parent et al. 2019; Hoff 2012; Hoff and Perez 2010). The highest variations in solar PV output power generally happen at the highest irradiance level, i.e., around noon time. This time does not necessarily concur with peak-load periods of the electrical system (Albadi 2016; Van Haaren, Morjaria, and Fthenakis 2014; Frearson et al. 2015; Parent et al. 2019; Hoff 2012; Hoff and Perez 2010). The critical concern when analyzing solar variability is the required ramping capacities, which are a function of the change in solar PV output power (ǻP) over two successive periods of time as shown in equation (1)

ǻP = P(t+ǻt) – P(t)

(1)

The value ǻP represents the change (rise and fall) in the kW of a PV system’s output power during a specific time interval (ǻt) such as 5, 15, and 30 minutes. In general, the variability of different solar PV systems’ power output is less than that of a single site. Although, passing clouds can cause high and rapid fluctuations at a single site, aggregating the output from different sites induces a smoothing effect and results in reducing the fluctuations’ severity (Albadi 2016; Van Haaren, Morjaria, and Fthenakis 2014; Frearson et al. 2015; Parent et al. 2019; Hoff 2012; Hoff and Perez 2010). The planners and operators are concerned about the potential impacts of solar PV output variability. This variability requires modification of the planning, scheduling, and operating plans to maintain system security and reliability. Hence, quantifying rapid ramps is important to understand PV systems’ output variability. This variability will have a direct impact on the balance between generation and loads.

Solar PV Car Park Project in the PDO-MAF The Petroleum Development Oman (PDO) commissioned Phase I of the Mina Al-Fahal (MAF) Solar Car Park Project in January 2018. The PDOMAF solar PV power is used to supply PDO offices and buildings. The total capacity of the Phase 1 project is 6.202 MWp and it is expected to produce 9.706 GWh annually. The total installed capacity is expected to reach 8.852 MWp after the commissioning of the currently ongoing Phase II of the Solar Car Park. The locations of PDO-MAF solar car park projects are shown in Fig. 3.1 (PDO 2018a, 2018b).

Variability Analysis of PV System Output

73

Fig. 3.1: Phases I & II PDO-MAF Solar Car Park Project Locations

Solar Car Park Project Locations The PDO-MAF Solar PV car park is located at longitude 58.52ၨ E, latitude 23.63ၨ N and at a height of 10 m above sea level. This location has a high level of solar irradiation throughout the year.

Fig. 3.2: PDO MAF PV Geographical Location

Chapter Three

74

Plant Layout Phase I of the PDO-MAF Solar Car Park Project covers about 37,577 m². The plant area of Phase I is divided into three blocks: Bait Al-Bushra Area (BAB), Bait Mina Al-Fahal Main Area (BMF Main) and Bait Mina AlFahal overflow Area (BMF overflow). Tables 3.1 and 3.2 illustrate the PV system details. The commissioned Phase I project includes 81 inverters and 969 strings while the Phase II project will include 36 inverters and 414 strings. Fig. 3.3 shows the Phase I Solar Car Park Project’s conceptual diagram that includes 12 strings connected in parallel to a single inverter, via a DC combiner box. Each string includes 20 solar panels connected in series. Each inverter connects the PV array to a Low Voltage (LV) switchgear. The LV switchgear is connected to the Medium Voltage (MV) PDO-MAF distribution system through a 415/11kV step-up transformer (PDO 2018b, 2018c). Table 3.1: PDO-MAF Solar Car Park System Data for Phases I & II Phase I: Existing

Phase II: Future Total

HLD & Overflow Area

BSM & Comp Building

Total

Phase I+ Phase II

4,980

19,380

3,700

4,580

8,280

27,660

396

249

969

185

229

414

1,383

27

33

21

81

16

20

36

117

DC Combiner Box

27

33

21

81

16

20

36

117

Total Array Power kWp

2,074

2,534

1,594

6,202

1,184

1,466

2,650

8,852

Total Array Power kWac

1,620

1,980

1,260

4,860

960

1,200

2,160

7,020

BAB

BMF Main

BMF Overflow

Number of Modules

6,480

7,920

Number of Strings

324

Number of Inverters

System Parameters

Table 3.2: Parameters of the PV Modules and Inverters of the PDOMAF Solar Car Park Phase I project Technology PV module Tilt angle Inverters

Standard Silicon, Poly-crystalline 320 WP-DC, 46.22VOC-DC 3ၨ 60kWAC, 1000V, 50Hz

Variability Analysis of PV System Output

75

Fig. 3.3: Conceptual Diagram of Phase 1 of the PDO-MAF Solar Car Park Project

Deterministic Variability – Solar Eclipse A solar eclipse happens when the moon passes between the sun and the earth. During a solar eclipse, the moon blocks the light of the sun from reaching the earth. The shade of the moon is then cast on the earth. For an annular solar eclipse to happen, three conditions must occur simultaneously: 1) it must be the time of the new moon, 2) the earth, the moon and the sun are aligned in a straight line, and 3) the sun is located at its farthest distance from the earth. During a solar eclipse, the sun will appear as a ring around

76

Chapter Three

the moon. The eclipse may last from a few seconds to three hours from beginning to end. Oman’s sky observed an annular eclipse on 21 June 2020, an occasion that might only be seen again after 83 years in 2103. The annular eclipse of the sun on 21 June 2020 lasted 3 hours and 5 minutes in Muscat. It started at 8:14 am and reached a peak of 98% at 9:39 am before it ended at 11:19 am. Fig. 3.4 illustrates the solar eclipse times in Muscat local time.

Fig. 3.4: Solar Eclipse Times in Muscat Local Time3

The impacts of a solar eclipse on climate are summarized in Anderson (1999). Several researchers evaluated the potential impacts of solar eclipses on PV system performance. In Weniger et al. (2014), the authors studied the probabilities of a large drop in the aggregated solar PV system output during the eclipse day. The impact of the solar eclipse on 21 June 2020 at the PDO-MAF Solar PV system has been assessed and compared with a clear sky day. Fig. 3.5 is the extracted plot of the eclipse’s impact on the output of the BMF Main PV system on 21 June 2020. In this chapter, measured data are presented in terms of the normalized power output in per unit (pu). The normalized power output is defined as the ratio of the actual output from inverters (kWAC) to the connected capacity (kWDC). The output started descending at 8:14 a.m. with the beginning of the eclipse until it reached its lowest value around 9:39 am, which was the peak of the eclipse. The PV system output returned to its normal performance at a ramping rate higher than that occurring during clear sky mornings/afternoons. The maximum observed 15-minute average drop in the PV system output was 95% (0.594 pu) compared with the value during a clear day. It is worth noting that the expected maximum drop is 98%, similar to the maximum percentage of the solar disc coverage area.

3

http://time.unitarium.com/events/eclipse/062020/ (Jun 2020).

Variability Analysis of PV System Output

77

0.90 0.80 0.70

P (pu)

0.60 0.50 0.40 0.30 0.20 0.10 0.00 19:12

0:00

4:48

9:36

14:24

Power (pu) (Clear Day) 19/06/2020

19:12

0:00 Time Stamp

Power (pu) (Eclips Day) 21/06/2020

Fig. 3.5: Normalized Solar PV Power Output during Solar Eclipse Day versus a Clear Non-eclipse Day

The same trends were observed for the BAB and the BMF overflow PV systems. Aggregating the output of the 3 systems does not produce any smoothing effect. These large changes in the PV systems’ output pose a challenge to the grid stability especially if the power grid supply mix has a large share of solar PV power. It is essential to have enough storage and spinning reserve capacities to maintain system stability during such events.

Stochastic Variability of a Single Inverter Output During a clear day, PV system output power is highly predictable; therefore, power system reliability can be maintained during morning/afternoon ramping up/down periods. However, with passing clouds, PV system output power fluctuates at a higher ramping rate. This section presents the variability of the 15-minute solar PV output data for a 60-kW inverter as shown in Figs. 3.6–3.8. To quantify output fluctuation, the normalized fluctuation of output data is presented as histograms.

Chapter Three

P (pu)

78

1.00 0.90 0.80 0.70 0.60 0.50 0.40 0.30 0.20 0.10 0.00 12:00 AM

4:48 AM

9:36 AM

2:24 PM

7:12 PM

12:00 AM

Time Satmp Clear day_26 Dec 2019

Cloudy day_28 Dec 2019

Fig. 3.6: Normalized PV System Output from one Inverter at BAB 0.60 0.40

ȴ P (pu)

0.20 0.00 -0.200:00 -0.40

4:48

9:36

14:24

19:12

0:00 Time Stamp

-0.60 -0.80

Fig. 3.7: Output power changes from a Single Inverter at BAB during a cloudy day on 26 December 2018

Variability Analysis of PV System Output

79

Fig. 3.8: Histogram of output power changes from one inverter at BAB

During the clear sky day, the peak power output reached 0.706 pu compared to 0.771 pu during a partially cloudy day. From Fig. 3.8, it can be noticed that the changes in PV output power of the single inverter case mostly ranged from -0.24 to 0.20 pu while the minimum and maximum changes were -0.57 and 0.64 pu, respectively. From the above analysis of the single 60 kW PV inverter case, it could be clearly shown that the level of variability is very high.

Stochastic Variability of Multiple Inverters’ Aggregated output Figs. 3.9 and 3.10 present the aggregated PV system output (P) and changes (ǻP) for different inverters: 9 and 18. When clouds passed, solar irradiation decreased drastically. Fluctuations in solar irradiance led to variations in solar PV system output.

Chapter Three

80

1.00

Clear day_26 Dec 2019

0.80

9 Inverters_28 Dec 2019

P (pu)

0.60 0.40 0.20 0.00 12:00 AM

4:48 AM

9:36 AM

2:24 PM

Stamp 7:12 PMTime 12:00 AM

-0.20 Fig. 3.9: Normalized PV system output from multiple inverters at BAB 9 Inverters

0.60

18 Inverters

0.40

ȴ P (pu)

0.20 0.00 12:00 AM -0.20 -0.40

4:48 AM

9:36 AM

2:24 PM

7:12 PM

12:00 AM

Time Stamp

-0.60 -0.80

Fig. 3.10: Output power changes for multiple inverters at BAB on 28 December 2019

Variability Analysis of PV System Output

81

Fig. 3.11 : Histogram of output power changes for 9 inverters at BAB

Fig. 3.12: Histogram of output power changes for 18 inverters at BAB

The histograms shown in Figs. 3.11 and 3.12 show that the system of 18 inverters had a relatively smoother output compared with the system of 9 inverters. Most PV output variations for the 18-inverter case ranged from -

Chapter Three

82

0.15 to 0.10 pu. The minimum and maximum recorded changes were -0.56 and 0.43 pu, respectively. For the 9-inverter case, most PV output variations were between -0.20 and 0.17 pu whereas, the minimum and maximum variations reached -0.66 and 0.44 pu, respectively. It is clearly seen that the smoothing effect of combining the output of multiple inverters rises with an increase in the number of inverters within the same location.

Stochastic Variability of a Single Site Aggregated Output

1.00

Clear day _26 Dec 2019

0.90 0.80

P (pu)

0.70 0.60 0.50 0.40 0.30 0.20 0.10 0.00 12:00 AM

4:48 AM

9:36 AM

2:24 PM

7:12 PM

12:00 AM

Time Stamp Fig. 3.13: Normalized output of the solar PV system at BAB

Fig. 3.13 presents the aggregated output of the full solar PV system located at BAB. The output power changes for the system are shown in Fig. 3.14.

Variability Analysis of PV System Output

83

0.30 0.25 0.20

ȴ P (pu)

0.15 0.10 0.05 0.00 12:00 AM -0.05

4:48 AM

9:36 AM

2:24 PM

7:12 PM 12:00 AM Time Stamp

-0.10 -0.15 -0.20 Fig. 3.14: Output power changes for the solar PV system at BAB on 28 December 2019

Fig. 3.15: Histogram of Output power changes of the solar PV system at BAB

As shown by the histograms of output power changes at BAB in Figs. 3.14 and 3.15, 91.6% of the changes were within the range of -0.33 to 0.15 pu. The maximum change of the PV output power was 0.28.

Chapter Three

84

Stochastic Variability of Different Sites’ Aggregated Output In this section, the aggregated PV system output (P) and changes (ǻP) were studied using the 15-minute recorded measurements and considering four cases: the single inverter output, the aggregate of all inverters in one location, the aggregate of all inverters in two locations, and the aggregate of all inverters in all three locations. The three locations are sited within a geographical distance of 5 km. Fig. 3.16 demonstrates that output ramp rates decrease as the solar PV system size increases. It is worth noting that the solar PV power output of all inverters is highly correlated. Although the three sites are located within 5 km, aggregating the output of the three locations has a positive impact on reducing the variability. 0.9

One Location Two Location Three Locations One Inverter

0.8 0.7

Power (pu)

0.6 0.5 0.4 0.3 0.2 0.1 0 -0.1

0:00

4:48

9:36

14:24

19:12 Time Stamp 0:00

Fig. 3.16: Aggregated output of different PV systems on 28 December 2019

As shown in Figs. 3.6–3.19, the PV output variability from a single inverter and a single location is higher than at aggregate locations. This variability is driven by intermittent cloud cover over the solar PV system sites. Though, if the individual site is considered to be separate PV systems within the same distribution network then their aggregated output would deliver a smoother solar PV output. In this case, the smoothing effect of the aggregated output is limited due to the short distance between the three systems (5 km) as well as the long time-scale used in the analysis (15-minute average values).

Variability Analysis of PV System Output

85

One Location 0.60

Two Location Three Location

0.40

One Inverter

ȴ P (pu)

0.20 0.00 0:00

4:48

9:36

14:24

19:12

0:00

-0.20 -0.40

Time Stamp

-0.60 -0.80 Fig. 3.17: Aggregated output changes for different PV systems

Fig. 3.18: Histrogram of aggregated output changes for two locations (BAB & BMF)

Chapter Three

86

Fig. 3.19: Histrogram of aggregated output changes for the three locations (BAB, BMF & BMF OF)

A summary of the analysis of PV output variability under various cases is shown in Table 3.3. The range of ǻP, defined as the difference between the positive and negative extreme values, is reduced from 1.21 pu for a single inverter to 1.10 and 0.99 for 9 and 18 inverters, respectively. The range of ǻP for a single location reaches 0.42 pu. The smoothing effect does not improve much when the output of two and three locations was aggregated. This is attributed to the limited distance between the sites as well as the long time-scale of available measured data. Table 3.3: Various cases of the analysis of PV output variability under various cases Cases 1 2 3 4 5 6

Single Inverter 9 Inverters 18 Inverters Single Location (27 Inverters) Two locations (60 Inverters) Three locations (81 Inverters)

ǻP (extreme negative) -0.57 -0.66 -0.56 -0.14 -0.23 -0.21

ǻP (extreme positive) 0.64 0.44 0.43 0.28 0.20 0.19

Range 1.21 1.10 0.99 0.42 0.43 0.40

Variability Analysis of PV System Output

87

Summary This chapter presented a case study of PV system output variability that includes both deterministic and stochastic variabilities. The variability of solar PV system output can be reduced by aggregating the output of different inverters. Intermittent cloud coverage results in an intermittent output. However, as cloud fronts do not cover different PV plants simultaneously, the aggregated PV system output is less variable than that from a single PV system (invertor). The smoothing effect is dependent on the number of inverters as well as the distance between sites. The deterministic variability of solar PV systems’ output is linked to clear sky irradiation, which is known ahead of time. Deterministic variabilities include morning/afternoon ramping up/down as well as other deterministic events such as solar eclipses. The chapter presented the impact of the annular solar eclipse that happened on 21 June 2020 on the PV systems’ output. Aggregating the output of different PV systems has a negligible smoothing effect, especially if the distances between different PV systems are limited.

Works Cited Albadi, M. H. 2016. “Solar PV Power Intermittency and its Impacts on Power Systems – An Overview.” Review of. Sola 2012 (2014): 2018. Anderson, Jay. 1999. “Meteorological changes during a solar eclipse.” Review of. Weather 54(7), 207-15. Frearson, Lyndon, Paul Rodden, Josh Backwell, and Mikaila Thwaites. 2015. “Investigating the impact of solar radiation variability on grid stability with dispersed PV generation.” Review of. 31st EU PVSEC: 2989-95. Hoff, Thomas E., and Richard Perez. 2012. “Modeling PV fleet output variability.” Review of. Solar energy 86(8), 2177-89. Hoff, Thomas E., and Richard Perez. 2010. “Quantifying PV power output variability.” Review of. Solar energy 84(10), 1782-93. Mills, Andrew. 2009. “Understanding variability and uncertainty of photovoltaics for integration with the electric power system.” Review of. Parent, Laurène, Delphine Riu, Tuan Quoc Tran, and Thai-Phuong Do. 2019. “Method to characterize variability of photovoltaics power output.” PDO. 2018. “PDO Pathways to Sustainability.” “Renewable Energy in PDO.”

88

Chapter Three

PDO. 2018. “PDO begins transition from hydrocarbons to energy “ Review of. Oman Daily Observer (6 Aug. 2018). PDO. 2018. “PDO Inaugurates Solar Car Park.” https://www.pdo.co.om/en/news/press-releases/Pages/PDO%20 Inaugurates%20Solar%20Car%20Park.aspx (3 Jan. 2018). PDO. 2018. “PDO: Up with the Sun.” Review of. AL Manhal (01.2018). Van Haaren, Rob, Mahesh Morjaria, and Vasilis Fthenakis. 2014. “Empirical assessment of shortဨterm variability from utilityဨscale solar PV plants.” Review of. Progress in Photovoltaics: Research and Applications 22(5), 548-59. Weniger, Johannes, Joseph Bergner, Tjarko Tjaden, and Volker Quaschning. 2014. “Einfluss der Sonnenfinsternis im März 2015 auf die Solarstromerzeugung in Deutschland.” Review of. URL http://pvspeicher. htw-berlin. de/sonnenfinsternis/ (accessed on 201701-16).

CHAPTER FOUR IMPACT OF DUST AND EXTERNAL FACTORS ON PHOTOVOLTAIC SYSTEM OUTPUT IN WESTERN AUSTRALIA BONG WEI LI,1 SYED ISLAM2 AND RAKIBUZZAMAN SHAH2

Abstract Solar energy has become a crucial energy resource with incremental fossil fuel prices and the declining cost of solar panels, converters, and connectors. A variety of solar cell types such as monocrystalline, polycrystalline, APEX (another variant of thin-film type), and amorphous silicon are widely used in photovoltaic (PV) panels. The shading from trees, buildings, and other objects, including dust accumulation, may significantly affect PV panels' power output. The partial shading may cause a reduction of power in proportion to the area of shading. This chapter reports a study on monitoring the impact of shading and various dust particles of various grades and weights on solar panels in the Western Australian climatic condition. The performance assessment of different panels was obtained using simulation software PVSOL for natural vegetation and building shading at 15 m and 25 m, respectively. Experimental studies have been carried out for various dust particles such as talcum powder, sand, and pollen with multiple weights (5 g, 10 g, and 15 g). This full chapter includes a literature overview, methodologies, experimental settings and results, and data analytics to exhibit the impact 1 School of Electrical Engineering Computing and Mathematical Sciences, Curtin University, WA 6102, Australia, [email protected]. 2 School of Engineering, Information Technology and Physical Sciences, Federation University Australia, Mt Helen, VIC 3353, Australia, [email protected]; [email protected].

90

Chapter Four

of shading and dust. This is in line with the results of spectral transmittance curves for different dust density samples, which essentially cover the wavelength range of the PV modules. KEYWORDS: Dust, experimental results, external factors, partial shading, PV performance.

4.1 Introduction The demand for PV systems has increased significantly due to the sustained price reduction of PV panels, power electronic devices, government subsidies as well as enhanced reliability and efficiency. As reported by the International Energy Agency (IEA), the capacity of a newly installed system was at a rate of 100 MW per day in 2013, which has significantly increased over the last five years (Koentges 2014, 1-250). Solar radiance intensity and temperature are the two main factors affecting the PV system. These are very much site-specific environmental factors affecting the operation of the PV system. Generally, the PV system is often located in complex topographies. Over time, the energy output from PV modules would decrease due to humidity, the thermal cycle, UV radiation, and moisture. These common causes may lead to permanent degradation of the PV panels. Besides these common causes, environmental factors such as dust accumulation and shading could significantly reduce the performance of the PV modules as reported in Munoz (2011, 2264-2274). With the increase in global pollution, dust particles from the air tend to deposit on PV surfaces more frequently. Therefore disrupting the performance of the PV panels. The dust accumulation in PV panels may cause a substantial impact on large-scale PV plants with significant losses of generation and revenue. Solar farm owners pay the maintenance cost for cleaning panels in order to prevent losses due to power degradation and maximize their income. The cost-benefit analysis of cleaning then depends on the number of income gains compared to the maintenance costs. An investigation conducted in Kuwait City has concluded that sand accumulation would cause a 17% reduction in overall power efficiency (Elminir 2006, 3192-3203). On the other hand, there have been studies concerning the dust effect in dry and rainless regions as the place has high solar intensity available for energy generation. The work in Monto Mani (2010, 3124-3131) reported a deterioration of PV panel performance up to 30% with tilt angle of 26% due to the dust deposition for four months in Dhahran, Saudi Arabia.

Impact of Dust and External Factors on Photovoltaic System Output in Western Australia

91

The work in Tanesab (2019, 347-354) investigated the effect of dust with different morphologies on the performance degradation of various PV technologies consisting of polycrystalline silicon, monocrystalline silicon, and amorphous silicon. Dust collected from two different locations, Babuin, Indonesia, and Perth, Australia, was coated artificially onto the surface of the glass of the PV modules. The analysis revealed that dust from Babuin was dominated by larger sized and porous particles that are lighter than ones from Perth. Dust from Perth with angular and diagonal shapes had a better optical property than that from Babuin featuring elliptical and spheroid particles. As a result, the transmittance values of the two types of dust tend to balance out. For an equivalent amount, the effect of dust from Babuin and Perth on each PV technology's performance degradation was similar. Furthermore, the spectral curves of the two types of dust transmittance were fairly flat with respect to the wavelength range of these technologies. Consequently, dust from either Babuin or Perth has similar effects on the three PV technologies. The work in Yan (2013, 102-108) has assessed the dust accumulation performance under Brisbane's climatic conditions for different tilt angles and orientation. It is found that the theoretical model has a good agreement with the field measurement data from the 1.2 MW PV system at the University of Queensland. However, only a certain type of dust has been considered for the analysis with three months of data. Furthermore, the partial shading could cause significant performance reduction in PV systems. The PV panel exhibits one peak under the uniform irradiance condition. When the partial shading takes place, multiple peaks appear in the system due to a bypass diode. The experimental study conducted in Liu (2019, 1014-1024) suggested that the maximum power of the PV system may drop at a constant rate with the increment of the shading thickness. It reported that the annual energy loss of 10–20% occurs due to the partial shading in PV panels (Teo 2018, 112). However, several great studies have been conducted with dust accumulation and shading effects on the PV system. Nonetheless, very few studies have considered the real-life conditions.

4.2 Effect of Dust on PV Panels The deposition of dust particles has remained a significant problem for the PV system. Research has proved that dust particle accumulation deteriorates solar cells' performance, leading to a noticeable loss of generated power output (El-Shobokshy 2007, 505-511). The work in El-

92

Chapter Four

Shobokshy (2007, 505-511) has used a solar simulator (three 1 kW lamps) with the physical properties of known materials placed at a distance of 3 m to explore the effect of the dust on PV panels. Dust particles of different sizes which are commonly found in atmospheric dust such as limestone, cement, and carbon, are considered in this work. It should be worth noting that cement and carbon have the lowest standard deviation due to a uniform size compared to others. The experiment shows that the performance of PV cells is primarily affected by finer dust on the surface and less affected by the deposition of dust with rough particles. It has been noticed that 73 g of cement particles could cause a drop of 20% in the short circuit current. Carbon particles have shown the worst performance of efficiency drop in PV panels (Adinoy 2013, 633-636). From the experiment, it is also evident that the material types in dust also affect the performance (carbon particles absorb solar radiation more effectively) of the PV panels. It is concluded from the study that the effect of dust on performance can no longer be restricted to the exposure period and would be affected mainly by particle mass or size, respectively. This experiment precisely analyses the particle size distribution, which is useful in analysing small particle effects on the panels. However, it should be worth noting that the results are obtained in an ideal laboratory environment using constant irradiance from the solar PV emulator. Furthermore, dust accumulation has been a significant issue in regions with dust storms (Gholami 2018, 346-352). It is mostly found in areas where there will be loose sediments like sand, silt or clay, and a lack of protective cover from vegetation such as the mining sites or deserts. Prior studies have used six modules exposed to outdoor conditions where the power is measured daily, including current, voltage readings, and temperature. It can be observed that besides the dust that settles on the module, there are also atmospheric microscopic particles present, which may reduce the light reaching to the module, thereby, the output from the panels. During the test period (i.e., 12 months), it was observed that power is reduced continuously by 20% due to accumulation (Gholami 2018, 346352). However, in November, the modules' power output was seen at its highest values due to the rainfall in the region. After the cleaning routine was conducted, the modules' power outputs remained to the expected level. The performance ratio was used to get overall results for all the platforms. From the experiment, it is evident that polycrystalline has the best performance under all the test conditions. Using the solar tracker, the efficiency with temperature is inversely proportional to the temperature of the solar panels. The polycrystalline cells' hourly temperature profiles revealed that the polycrystalline cells store more heat than the

Impact of Dust and External Factors on Photovoltaic System Output in Western Australia

93

monocrystalline PVs. Moreover, the dust particles with different and irregular shapes influenced the shading effect and power output of the panels. It should be worth noting that the dust accumulation on PV panels does not solely depend on the duration of module exposure but also on the frequency and intensity of dust (Gholami 2018, 346-352). It is also noticed that a higher than 50% reduction of power was recorded in the module that was not cleaned for an extended period of 6 months. This experiment has been conducted in dusty areas where natural dust can affect the performance of panels over an extended period.

4.3 Effect of Shading on Performance The performance of PV panels is profoundly affected by the shading effects from tree cover, clouds, nearby buildings, and other obstructions (Sathyanarayana 2015, 1-10). The work in Sathyanarayana (2015, 1-10) investigates the performance of PV panels affected by the shading effect mainly on parameters such as the power output, fill factor, and efficiency of the panels. The measurement of shading could be challenging due to the variation of shadow with the sun’s position throughout the year and the day. Shading effects can also depend on the severity and area of the shade as reported in Sathyanarayana (2015, 1-10). The experiment in Sathyanarayana (2015, 1-10) focuses on the effect of uniform and nonuniform shading on PV panels. It makes use of an adjustable mounted stand on butter papers for uniform shades and physical objects for nonuniform shades. Indices such as power output, fill factor, and efficiency have been recorded. This has resulted in the I-V characteristics for uniform shading decreasing with the increment of the shading percentage. Consecutively, the power output is also reduced with an increase in shading (Ramyar 2017, 2855-2864). Subsequently, the short-circuit current has a linear relationship with the irradiance received as well as the percentage of shading. It is observed that if one whole area of the panel is covered with non-uniform shading, then the power output would be zero for that area. On the other hand, power output is directly proportional to the cell area for partially covered situations. In the non-uniform shading case, both efficiency and fill factor depend on effective cell brightness and reduce the output of the panel. The cell receiving lower irradiance may act as a load. Non-uniform shading conditions should be avoided to minimize the damage of the cell.

Chapter Four

94

4.4 Methodology The study has been conducted in Perth, Western Australia. It is situated between latitude 31.950 South and longitude 115.850 East. It has four seasons: summer (December to February), autumn (March to May), winter (June to August), and spring (September to November). According to Australian long-term meteorological data, the average maximum temperature in Perth is 31.70 in February, while the average minimum temperature is 7.60 in July. Perth has more sunny days compared to other state and territory capitals in Australia. The annual average solar irradiance (the average radiation received on a surface) on the horizontal surface is 159.67 W/m2/day while it increases to 163.25 W/m2/day for a tilted surface (the tilt angle is equal to latitude). The details of these data are given in Table 4.1. Table 4.1: Climate condition of Perth over 22 years (Teo 2018, 1-12) Parameters Horizontal irradiance (W/m2/day)

Annual average 159.67

Irradiance on tilted surface (MJ/m2/day)

163.25

0

Maximum temperature ( C) Minimum temperature (0C) Rainfall (mm) Wind speed (m/s) Relative humidity (%)

24.70 12.80 61.00 2.81 70.20

4.4.1 Simulation Model The PVSOL simulation package is widely used for the design of PV panel orientation (PV Sol Premium 2018). It has numerous functions which allow the users to specify the location-specific climate data, solar irradiance, and PV module data for simulation studies. It has a 3-D visualization for shading assessment in PV panels for a geographical area. The shading from buildings and trees in the Western Australian condition has been assessed using PVSOL. The analysis has been done at 12 solar noon with an azimuth angle of 1080 and an elevation angle of 187.10. For the simulation, the building and the semi-permeable vegetation have fixed sets of values for width and distance while the heights are maintained at 15 m, 20 m, and 25 m, respectively. Fig. 4-1 shows the shading effect design

Impact of Dust and External Factors on Photovoltaic System Output in Western Australia

95

under 3-D visualization in PVSOL. The following parameters given in Table 4.2 are used for the shading analysis.

4.4.2 PV Panels Four different PV panels are used in this work for the simulation studies. These are: polycrystalline, monocrystalline, amorphous, and APEX. The parameters used for the simulation are given in Table 4.3. The three panels as given in Fig. 4.2 have been used for the simulation studies. The panels are made of polycrystalline being 110 cm in length and 50 cm in width. The specifications of the experimental PV panels are given in Table 4.4. Methods commonly applied to monitor and evaluate the electrical performance of a module are the current voltage (I-V) and power voltage (P-V) curve analysis as discussed in Trigo-gonzalez (2019, 303-312), Bouguerra (2020, 1-10). These curves may present the values of electrical parameters of a module including maximum power output (Pmax), maximum output current (Imax), maximum output voltage (Vmax), open circuit voltage (Voc), and short circuit current (Isc). To compare the effect of dust on the PV module performance, the measured electrical parameters are first transposed to equivalent values in standard test conditions (STC) (referring to radiation of 1000 W/m2, module temperature 25°C, and air mass 1.5). In this work, the IEC 60891 method (IEC 60891, 2009) (IEC 60891:2009 2009) was used to transpose between the real operating condition (ROC) and STC for the used PV technologies covered in the experiment. The key equations employed in this method are – ீమ

‫ܫ‬ଶ = ‫ܫ‬ଵ + ‫ܫ‬ௌ஼ . ቀ

ீభ

െ 1ቁ + ߙሺܶଶ െ ܶଵ ሻ

ܸ2 = ܸ1 െ ܴ‫ ݏ‬ሺ‫ܫ‬2 െ ‫ܫ‬1 ሻ െ ߢ‫ܫ‬2 ሺܶ2 െ ܶ1 ሻ + ߚሺܶ2 െ ܶ1 ሻ

(4.1) (4.2)

96

Chapter Four

Fig. 4.1: Shading effect design in PVSOL

Table 4.2: Parameter for tree shading Parameters Tree height Trunk diameter Diameter Degree of tree transmittance Distance between PV module and tree

Value 15 m 0.3 m 2m 50% 15.93 m

Based on the I-V curve produced by (4.1) and (4.2), Pmax2, Voc2, and Isc2 are obtained and applied to get the fill factor as in (4.3).

‫ܨܨ‬ଶ =

௉೘ೌೣమ ௏೚೎మ ூೞ೎మ

(4.3)

where subscripts 1 and 2: ROC and STC, respectively, G is the in-plain irradiance (W/m2), T is the module temperature (0C), Į is the current temperature co-efficient (A/0C), ȕ is the voltage temperature co-efficient (V/0C), Rs is the internal resistance, k is the curve co-efficient factor, and Pmax is the maximum power (W). The Pmax is the parameter used by the manufacturer to indicate the performance of any module. Therefore, it can be used to quantify the degradation of the PV module as given in 4.4. Along with Pmax, the efficiency of the PV module (given in 4.5) is used in this work to identify the degradation of the PV module(s).

Impact of Dust and External Factors on Photovoltaic System Output in Western Australia

‫ܥ‬ௗ௨௦௧ ሺ%ሻ = ߟ=

௉೘ೌೣ ௟௢௦௦௘௦ ௕௬ ௗ௨௦௧ ்௢௧௔௟ ௉೘ೌೣ ௟௢௦௦௘௦

× 100

௏೚೎ ூೞ೎ ஺௥௘௔ ௢௙ ௠௢ௗ௨௟௘‫כ‬௜௥௥௔ௗ௜௔௡௖௘

97

(4.4) (4.5)

4.4.3 Dust Sample Four different types of dust particles are collected and used in this work to match the effect of natural dust particles on the PV panel. Black sand, silica sand, powder, and pollen grains are used. The choice of these dust particles is made based on the particles commonly found in the Western Australian climatic condition. The dust particles are put into small 2 g bags and weighed. The dust particles are evenly distributed with a thickness of 1 μm to 100 μm. The overall weights are varied to assess the different dust accumulations on PV panels (see Fig. 4.3). The most dependable and effective way to compare the dust impact on the PV panels is to design an experiment where the dust particles are spread on the panels and measurements are taken on a daily basis. The power output of these modules with the presence of dust particles is then compared to the power output of the clean modules. The main panels are wiped and cleaned with a wet cloth to prevent any substances from sticking to the cells thus interfering with the results. Since the same type of panels with the same technologies is used, this is ideal for a comparison of performance under different dust weights.

Fig. 4.2: Experimental modules with dust

98

Chapter Four

The panels are then left to dry out for about 30 minutes until the water is entirely evaporated. The artificial dust is then spread evenly across the specific area. Both the ammeter and voltmeter are calibrated to their initial state to ensure the values are at zero state before the start of the experiment. A temperature sensor is used to get the temperature values for the solar cells by placing it above a certain point on the PV panels. The specifications of the temperature sensor and the ammeter and voltmeter used for this experiment are given in Fig. 4.5. The experiments are conducted at the optimum time settings ranging from 12 pm to 4 pm which are ideal for solar efficiency. The readings are taken at 5-minute intervals for the first hour then at 10-minute intervals for the subsequent hours. Some of the recorded data are given in Table 4.6.

(a) Type 1: Black sand:

(b) Type 2: Silica sand:

(c) Type 1: White powder

(d) Pollen:

Fig. 4.3: Dust samples used in the experiment

Impact of Dust and External Factors on Photovoltaic System Output in Western Australia

99

Table 4.3: PV parameters used for the simulation Parameters

Polycrystalline

Monocrystalline

Amorphous

APEX

Power (W) Opencircuit voltage (V) Shortcircuit current (A) Nominal temperature (0C) Title angle (0) NOCT

59.25 21

59.25 19.89

59.25 20.1

59.25 19.76

3.86

3.9

3.9

3.93

49.20

45.25

53.25

35.52

32

32

32

32

470 ± 20

450 ± 1.80

460 ± 1.80

470 ± 1.80

Table 4.4: PV module parameters (MSX60) Parameters Peak power (W) Open-circuit voltage (V) Short-circuit current (A) Max. power voltage (V) Max. power current (A) No. of cells Nominal temperature (0C)

Annual average 59.20 21.00 3.86 17.20 3.59 36 49.20

Table 4.5: Equipment specification Specification CHY temperature sensor range and accuracy Ammeter DC current range and accuracy Voltmeter DC voltage range and accuracy

Value -300 C to 5500 C; ± 2% 300 mV–1000 V; 0.25%+1 decimal 30 mA–10 A; 1.5%+2 decimal

100

Chapter Four

Table 4.6: Power at different radiation levels for the clean panel (calculated based on the measurement) Radiation level (W/m2) 650 700 750 800 850 900

Value (W) 130.35 135.02 137.23 140.38 143.57 146.78

4.5 Results and Discussion 4.5.1 Simulation Results The effect of shading on various solar cell technologies under Western Australian climatic conditions is discussed in this section. The reduction of the irradiance reception for the 15 m shading object is given in Fig. 4.4. From the results given in Fig. 4.4, it is evident that the irradiance reduction is low for APEX compared to the other solar PV technologies. Fig. 4.5 shows the reduction of the irradiance reception for the 25 m shading object. From the results, it is worth noting that both APEX and amorphous exhibit similar characteristics. Moreover, it should be noted that amorphous provides a better output performance when the cell numbers are high. If one cell is shaded, the majority of the cells would be operating near to their open-circuit voltages and flattening the entire string's output performance. It can be concluded that as the proportion of shading increases, the power dissipated by the shaded cells would also increase. Amorphous is highly efficient in shading compared to the others, i.e., monocrystalline, APEX, and polycrystalline. It can withstand higher temperatures without the output being affected. It can also perform at low light conditions as it absorbs a broader group of the visible light spectrum due to solar triple-junction cell technology (see Fig. 4.6).

Impact of Dust and External Factors on Photovoltaic System Output in Western Australia

101

Fig. 4.4: Reduction of irradiance reception at 15 m shading

Fig. 4.5: Reduction of irradiance reception at 25 m shading

Fig. 4.6: Triple layer in amorphous cell technology (Solar Facts and Advice 2013)

102

Chapter Four

4.5.2 Experimental Results The experiment mainly focuses on assessing the effect of different types of dust samples of various weights at varying insolation levels. The power outputs of the PV cells were gained for four different dust samples with varying weights by obtaining the voltage and current measurements of the cell. It is observed that the solar insolation is high between 12.00 and 3.00 pm, mostly with an insolation of 800 W/m2 or higher which drops as the daylight settles. In the summer, it appears that there is a longer duration of sunlight in the day where the minimum and maximum cell temperatures are around 35ႏ and 76ႏ, respectively. The maximum solar irradiation varies around 1023 W/m2 and the minimum is around 661 W/m2. The final output power is taken from the average of two repeated voltage and current measurements during the experiment. The voltage and current are varied by the amount of dust on the panels and affecting the power output value as given in Table 5.5. It should be noted that power ratings specified by the module manufacturers are obtained under controlled conditions (i.e., standard conditions). Therefore, the output of the PV modules differs under real conditions. It can be confirmed that when the PV surface is clean, it shows the higher amount of power output compared to a module with the presence of the dust. Similarly, throughout the duration of testing, the power output of PV panels decreased with the accumulation of dust to a large extent. It should be noted that dust particles act as external resistance and obstruct the light penetration into the solar panels. A comparative study of dust samples is carried out, and the results are summarised in Fig. 4.7 at a radiation level of 800 W/m2 with different weights of dust samples.

Fig. 4.7: Average power output for different dust particles with weight (5g, 10g, 15g)

Impact of Dust and External Factors on Photovoltaic System Output in Western Australia

103

From the results given in Fig. 4.7, it is evident that the weights of Sands 1 and 2 do not have a very significant impact on the output power of the PV system. It should be worth noting that the output power was significantly reduced with the accumulation of the fine particles on the solar modules. Moreover, the power output was significantly reduced due to the accumulation of 15 g fine dust particles. Fig. 4.8 shows how the efficiency of the studied panels changes under different irradiance conditions subjected to various dust accumulations. From the results given in Fig. 4.8, it is worth noting that the panel's efficiency reduced with the irradiance due to the temperature effect. The fine particle accumulation of 10 and 15 g weights exhibits the lowest efficiency for the studied panels. An efficiency of 0.5% is seen for such dust accumulation. It should be worth noting that the system exhibits higher efficiency under a 970 W/m2 irradiance compared to lower irradiance levels in the presence of 5 g of Sand 1 on the panel. Fig. 4.9 shows the average efficiency of the studied panels with different dust accumulations. From the results, it is evident that the highest maximum average efficiency of the studied panels is 22%. The average efficiency dropped to 18.5% for the pollen accumulation on the panel. A significant reduction in average efficiency can be observed for fine particle accumulation. The average efficiency varies between 11 and 15% for various types of sand accumulation on the panel with different weights.

104

Chapter Four

Fig. 4.8: Efficiency vs. irradiance for various dust particles

Fig. 4.9: Average efficiency with different dust particles

Impact of Dust and External Factors on Photovoltaic System Output in Western Australia

105

4.5 Conclusions This chapter summarizes the research relating to the various types of dust particles contributing to the PV panel’s efficiency. The work has also studied the shading effect on different PV panels. The effect of dust on PV panels' performance indices such as the current, voltage, power, and efficiency is recorded for Western Australian climatic conditions and discussed. Several previous works have focused on the effect of dust on the PV panel's performance, but the type of dust composition and weight composition was not assessed. In this work, the PV modules are exposed to real outdoor conditions at Curtin University, Perth, Australia, to assess the implication of various dust particles on the performance of the PV panels. Simulation studies were carried out in Western Australian climatic conditions to assess the impact of shading on various PV technologies. From the simulation results, it is evident that the percentage of shading is lower in the amorphous at shades of higher object height, followed by APEX. This result is expected as the cell technology has performed well under the low light area. It is due to the cell structure in this technology, which is different from the others (e.g., monocrystalline, polycrystalline). As stated above, it has a triple-layer structure which is optimized to capture the light for a full solar spectrum. Therefore, the amorphous PV panels are less subjected to shading effects compared with other cell technologies. According to the experimental study, a drop in output power was observed when dust particles of various weights and types were deposited on the PV panels. The average power output of the PV panels could reduce to 9.2%, 17.8%, and 68.6%, respectively, for dust deposits of various types with 15 g of weight. It can be noted that the increment of any amount of dust deposition may decrease the overall efficiency of the PV panels. On the contrary, the efficiency of the PV panels decreased with the addition of dust deposits. The lowest average efficiency of the studied PV panels could be observed for the deposition of fine dust particles, such as powder. A reduction of 21% of PV power output is expected compared to the clean panels for a 0.5% deposition of fine particles on the panels. These simulation studies have covered the shading effects of different PV panels achieved using the PVSOL simulation program conducted over 365 days. The data are precise and can be used by researchers to improve the

106

Chapter Four

technologies of PV panels. The dust effects on panels are also beneficial in investigating and improving panels to be more resistant to them. In future studies, it would be beneficial to investigate inclined panels and the effect that gravity and the surface have on the dust particles. The experiments can be conducted with and without the addition of a bypass diode for shading. There will be a benefit from the study of shading and the impact that affects overall efficiency. Consequently, experimental results, correlations, and simulations relying on mathematical modelling that cover different factors seem to be the possible future research results.

References Adinoy, M. J., and S. M. Said. 2013. “Effect of dust accumulation on the power outputs of solar photovoltaic modules.” Renewable Energy 60, 633-636. Bouguerra, S., et al. 2020. “The impact of PV panel positioning and degradation on the PV inverter lifetime and reliability.” IEEE Journal of Emerging and Selected Topics in Power Electronics 99, 1-12. El-Shobokshy, M. S., and F. M. Hussian. 2007. “Effect of dust with different physical properties on the performance of photovoltaic cell.” Solar Energy 53, 505-511. Elminir, H. K., A. E. Ghitas, R. H. Hamid, F. El-Hussainy, M. M. Beheary, and K. M. Abdel-Moneim. 2006. “Effect of dust on the transparent cover of solar collectors.” Energy Conversion and Management 47, 3192-3203. Gholami, A., I. Khazaee, S. Eslami, et al. 2018. “Experimental investigation of dust deposition effects on photovoltaic output performance.” Solar Energy 159, 346-352. Koentges, M., et al. 2014. “Review of failures of photovoltaic modules.” IEA PVPS Task, 13, 1-250. Liu, G., W. Yu, and L. Zhu. 2019. “Experimental based supervised learning approach towards condition monitoring of PV array mismatch.” IET Renewable Power Generations 13, 1014-1024. Monto Mani, R. P. 2010. “Impact of dust on solar photovoltaic (PV) performance: Research status, challenges and recommendations.” Renewable and Sustainable Energy Reviews 14, 3124-3131. Munoz, M., M. C. Alonso-Garcia, N. Vela, and F. Chenlo. 2011. “Early degradation of silicon PV modules and guaranty conditions.” Solar Energy 85, 2264-2274.

Impact of Dust and External Factors on Photovoltaic System Output in Western Australia

107

Ramyar, A., H. Iman-Eini, and S. Farhangi. 2017. “Global maximum power point tracking method for photovoltaic array under partial shading condition.” IEEE Trans. Industrial Electronics 64, 2855-2864. Sathyanarayana, R. B., P. Lakshmi Sagar, and P. S. Girish Kumar. 2015. “Effect of Shading on the Performance of Solar PV Panel.” Energy and Power 5, 1-8. Tanesab, J., D. Parlevliet, J. Whale, and T. Urmee. 2019. “The effect of dust with different morphologies on the performance degradation of photovoltaic modules.” Solar Energy 31, 347-354. Teo, J. C., R. H. G. Tan, V. H. Mot, et al. 2018. “Impact of partial shading on the PV characteristics and maximum power of photovoltaic string.” Energies 18, 1-12. Trigo-gonzalez, M., et al. 2019. “Hourly PV production estimation by means of an exportable multiple linear regression model.” Renewable Energy 135, 303-312. Photovoltaic devices – Procedures for temperature and irradiance corrections to measured I-V characteristics. IEC 60891:2009. PV Sol Premium. 2018. The Design and Simulation Software for Photovoltaic System. 2013. “What is Amorphous Silicon? Why is it so Interesting Now?” Solar Facts and Advice. Yan, R., T. Saha, P. Meredith, and S. Goodwin. 2013. “Analysis of yearlong performance of differently tilted photovoltaic system in Brisbane, Australia.” Energy Conversion and Management 74, 102108.

CHAPTER FIVE IMPACT OF THE INTEGRATION OF LARGESCALE PV POWER PLANTS ON THE GRID STABILITY AND OPERATION HISHAM M. SOLIMAN1, ABDULLAH AL-BADI1, HASSAN YOUSEF1, ABDULSALAM ELHAFFAR1, MASOUD AL-REYAMI2, SULTAN AL-RAWAHI2, MAKTOOM AL-HOSNI2 AND ANWER AL-HARTHY2

Abstract Due to the depletion of fossil energy sources used in power stations and its resulting environmental pollution, Oman is planning to install two large photovoltaic (PV) power stations, one of 500 MW in Ibri by 2022 and the other of 1000 MW in Manah by 2023. These PV stations will be connected to the main grid by voltage source converters (VSC). The PV stations inject power to the main grid. If there is a power imbalance between the PV’s power and the power demand, instability occurs. Therefore, a controller is required for the PV station. The main grid is also subject to spontaneous disturbances (load changes, temporary faults, etc.) that may grow to cause instability and as a consequence a great loss to the national economy. Therefore, studying system stability/stabilization is of paramount importance. In this chapter, steady-state stability is investigated with and without external controllers installed at Ibri and Manah PV stations. The system 1

Sultan Qaboos University, Dept. of Electrical and Computer Engineering, Muscat, Oman, Email (Abdullah Al-Badi): [email protected] 2 Oman Electricity Transmission Company, Muscat, Oman

Impact of the Integration of Large-Scale PV Power Plants on the Grid Stability and Operation

109

performance without controllers shows slow dynamics. To design the controllers, a mathematical model describing system dynamics is first developed. The model assumes that the injected power from Ibri and Manah PV power stations to the Main Interconnected System (MIS) can be represented by constant impedance loads. In other words, the PV stations see the main grid as constant impedance loads. The control objective is to design a decentralized tracking controller for the two distributed generations (DG) at Ibri and Manah. Unlike centralized control, the decentralized scheme is selected to avoid the one-point failure of the hub computer (controller), delays, and cost. Each DG utilizes a voltage source converter. The performance of the designed controllers is tested using the MATLAB R2017b-64bit framework. The simulation results verify the performance of the designed controller when the system is subjected to step changes in the reference signals. The second objective of this chapter is to study the transient stability of the MIS in the presence of large- scale PV stations. Towards this purpose, the entire grid is modelled using DlgSILENT software for the year 2022 (with the presence of the 500 MW Ibri PV station) and the year 2023 (with the presence of both the 500 MW Ibri and 1000 MW Manah PV stations). Different types of large disturbances are considered such as a three-phase short circuit at the largest power station (Sur power station), partial shading of the PV plant, large load shedding, and permanent outage of part of the PV power station. For all the fault scenarios considered in this study, the rotor angle stability, system frequency and voltage stability are not affected. Keywords: steady-state stability, mathematical modelling, decentralized tracking controller, transient stability, DlgSILENT, rotor angle stability, frequency stability, voltage stability

Introduction: Description of the main interconnected system (MIS) of Oman The Oman power system consists of three major power system grids (1) the Main Interconnected-Transmission System (MIS) in the north, (2) Dhofar Power System in the south, and (3) Petroleum Development Oman (PDO) Grid in between. Refer to Fig. 5-1.

110

Chapter Five

Fig. 5-1: Oman Map3

The OETC transmission system in Northern Oman (Main Interconnected System) has three operating voltages (400 kV, 220 kV and 132 kV) whereas 3 Oman, Map No. 3730 Rev. 4, January 2004, UNITED NATIONS. https://www.un.org/Depts/Cartographic/map/profile/oman.pdf

Impact of the Integration of Large-Scale PV Power Plants on the Grid Stability and Operation

111

in Dofar governorate it has only one operating voltage (132 kV). The Northern (MIS) and Dhofar grids are currently operated by the Oman Electricity Transmission Company (OETC). The PDO and MIS grids are interconnected through the Nahada-Nizwa 132 kV Over Head Lines (OHL). Also in the south, the Dhofar system is at present connected to the PDO system through the 132 kV OHL. The transmission system extends across the whole north of Oman. The OETC system is connected with the GCC (Gulf Cooperation Council) grid system through the UAE (United Arab Emirates) grid station via a 220kV system and was energized in 2011. Currently, the system is running in emergency cases and can transfer up to 800 MW. The OETC is connected to the PDO network at 132 kV, working only in emergency, and can transfer power up to 60 MW. The OETC is supplied with electricity generated from eleven gas-based power stations located at Rusail, Manah, Al Kamil, Barak (3 power stations), Sohar (3 power stations), Sur and Ibri. In 2019, the Main Interconnected System’s (MIS) peak demand reached 6536.75 MW with system losses of 1.43% (Oman Electricity Transmission Company 2020, 64-85). The single line diagram for part of the OETC network is presented in Figure 5-2. OPWP (Oman Power and Water Procurement) announced an ambitious development program for renewable energy projects that will be connected to the MIS starting with the 500 MW, Ibri II Solar IPP, that is supposed to be online by 2022. Another 1000 MW Solar project, in Manah, will be connected to the system by 2023 followed by additional solar and wind projects in 2023 and 2024. It was stated that by 2030 the contribution of renewable energy to the electricity generation sector will reach 30% (Berdikeeva 2019). The intermittency of PV power output is one of the main concerns for the grid operation. If there is a power imbalance between the PV’s input power (stochastic or intermittent due to solar insolation variation) and its output power, instability occurs (due to the zero inertia of the PV station). In addition, large-scale PV has no inertia and the integration of such power generation will reduce the effective inertia of the system. For any system disturbance, the machines must adjust the relative angles of their rotors after each incident to meet the condition imposed by the power transfer. Another challenge is the lack of the system’s reactive power which could lead to voltage instability (Shaha, Mithulananthana, Bansal and Ramachandaramurth 2015, 1423-1443).

112

Chapter Five

Fig. 5-2: Single line diagram for part of the OETC

Impact of the Integration of Large-Scale PV Power Plants on the Grid Stability and Operation

113

The main requirements in an electricity system are supply security, improved system efficiency, affordable energy cost, improved system power quality, high system reliability, low environmental impact and the effective integration of available sustainable energy resources. Adding Distributed Generation (DG) can manage electricity demand and generate clean on-site power that will increase the reliability of the system and reduce the system losses. Moreover, since most DG technologies are renewable; such as wind, photovoltaic, geothermal or hydroelectric, or clean energy sources such as fuel cells, cogeneration or biomass, significant environmental benefit can be achieved. Increased deployment of DG technologies will have a beneficial impact on the government’s energy policy goals, which are to reduce emissions, ensure a reliable energy supply, and promote competitive markets. The world market for solar photovoltaic (PV) systems has significantly increased recently, module prices are decreasing and emerging markets are increasing. In 2018 the global PV capacity reached more than 500 GW (REN21 2019, chap. 3). The drivers for this increase are the need to reduce gas emissions and diversify energy sources, and also for energy efficiency, lower capital costs and a shorter construction time (Jenkins, Allan, Crossley, Kirschen and Strbac 2000, chap. 1). Introducing high levels of grid tie PV and taking into account the planned utility PV units, it becomes a necessity to study the impact of PV insertion on the network system. Photovoltaic generating units connected to the power network may have several impacts on the power system such as voltage instability, transient stability and frequency stability.

Objectives of the Chapter To investigate the steady-state stability of the MIS with the penetration of one PV station and two PV stations. To investigate the transient stability of the MIS with the penetration of one PV station and two PV stations.

Power System Stability The most important limits that could restrict the power transfer in a grid are thermal limit, voltage limit and stability limit. Reliability of the power system is the capability of continuously delivering energy at acceptable standards both in normal and disturbed conditions. Whereas security can be

114

Chapter Five

defined as the ability to withstand a disturbance without losing the service, stability is the tendency of the power network to develop restoring forces that are equal to or more than the disturbing forces. Stability is divided into steady-state stability and transient stability. Steady-state stability is the ability of the power system to regain synchronism after slow and small disturbances, such as a gradual load change. Transient stability deals with a large and sudden disturbance such as a fault and sudden load changes.

Modeling of the PV System Power flow studies A power flow study is needed to determine the operation point required for a stability study. The PV system can be represented by a negative active power load which has a fixed value of power level and zero Q (PQ bus). A utility scale PV system can be modeled like conventional generators (PV bus) (Kundur et al. 2004, 1387-1401)

Steady-state stability Multi PV generators can be aggregated into a single unit having an MVA rating equal to the summation of individual units. The solar farm can be modeled similar to a synchronous generator with appropriate MVAr limits, that mean as a PQ bus or a PV bus (Shaha, Mithulananthana, Bansal and Ramachandaramurth 2015, 1423-143)

Transient Studies The central station PV system model, proposed by the Western Electric Coordinating Council (WECC), will be used to evaluate the transient stability of the OMAN MIS with large-scale PV stations (Ellis, Behnke and Elliott 2013, 8)

Impact on Power System Stability There are three stability studies: (1) voltage stability, (2) rotor angle stability, and (3) frequency stability.

Impact of the Integration of Large-Scale PV Power Plants on the Grid Stability and Operation

115

Voltage stability This refers to the ability of a power system to maintain steady voltages at all buses in the system after being subjected to a disturbance. Depending on the PV penetration level, the system could experience both beneficial and side effects for both steady-state stability and transient stability owing to their characteristics that differ from conventional generation resources (Kundur et al. 2004, 1387-1401). Short-term voltage stability is the ability of the system to control the voltages following large disturbances such as faults or loss of generation. This mainly involves the fast-acting load components capable of recovery (e.g., induction motors) after a major disturbance. Instability may occur owing to the severity of the disturbance of a control system that is too slow. Long-term voltage stability is the ability of the system to control the voltages following small disturbances such as incremental changes in the system load. It deals with slower acting equipment such as transformer tap changers or generator over-excitation limiters.

Rotor angle stability Rotor angle stability refers to the ability of synchronous generators of a power system to remain in synchronism after being subjected to a disturbance. It depends on the ability to restore equilibrium between the mechanical power input and the electrical power output of each generator in the system. Instability that may result occurs in the form of increasing angular swings of some generators leading to their loss of synchronism with other generators. Transient stability (large disturbance) This is the ability of the system to maintain synchronism during large disturbances. It deals with the grid operation subject to severe disturbances like faults, circuit switching, and sudden changes of load. The time frame is generally 3-5 seconds following the disturbance and might extend to 10-20 seconds (Kundur et al. 2004, 1387-1401). The following factors will affect the transient stability; the inertia of the system, angle differences of the bus voltages (a large injection of power from the PV will increase the angle), and the type of disturbances and their location with respect to the PV system and generating units. There are several methods that can be utilized to assess transient stability such as the

116

Chapter Five

equal area criterion, the extended equal area criterion, the single machine equivalent method, Lyapunov’s theory, and integration methods. Small signal stability (small disturbance) This is the ability of the power system to maintain synchronism under small disturbances. The disturbances are considered to be sufficiently small so that the linearization of system equations can be done. The time frame of interest in small-disturbance stability studies is in the order of 10 to 20 seconds following a disturbance (Moradi, Karimi and Ghartemani, “Robust Decentralized Control for Islanded Operation of Two Radially Connected DG Systems”, 2272-2277). The instability may result in two forms: An increase in the rotor angle owing to insufficient synchronizing torque; and Rotor oscillations of increasing amplitude owing to insufficient damping torque. The nature of the system response depends on a number of factors such as, the strength of the transmission system, the type of generator excitation controls, and the initial operating point.

Frequency stability The frequency stability is related to the ability of a power system to maintain steady frequency after being subjected to a disturbance resulting in an unbalance between the load and generation.

Steady-State Stability This part of the chapter studies the dynamic (steady-state, small-disturbance) stability of the Ibri and Manah PV power stations connected with the OETC. The stability is investigated with and without controllers installed at Ibri and Manah PV stations. To design the controllers, a mathematical model describing system dynamics is first developed. The model assumes that the injected power from Ibri and Manah PV power stations to the OETC can be represented by a constant impedance load. The control objective is to design a decentralized (tracking) controller for the islanded operation of the two distributed generations (DG), Ibri and Manah. Each DG utilizes a voltage

Impact of the Integration of Large-Scale PV Power Plants on the Grid Stability and Operation

117

source converter (VDSC). The performance of the designed controllers is tested on the MATLAB framework. Distributed generation (DG) units such as photovoltaic (PV) arrays are used to reduce the cost of the energy process and the environmental problems. In this part of the chapter, the description of the study system and its mathematical model are presented. The development of the proposed decentralized control to enhance the steady-state stability is given. The section is concluded by simulation results and observations.

Study system description Figure 5.3 shows a single-line diagram of Ibri and Manah power stations connected with the OETC.

Fig. 5-3: Ibri and Manah PV connection to the OETC

The load flow study for the OETC integrated with the PV stations is carried out by DlgSILENT. The study gives the amount of injected power from the PV stations to the OETC, as shown in Fig. 5-3. Note that solar panels will not produce reactive power as they are DC generators. However, by using a power converter and converting the same to AC power, reactive power will be drawn depending on load and power system parameters. The OETC can be represented as constant impedance loads as shown in Figure 5-4. Note that the OETC can also be modeled by an equivalent generator in series with a Thevenin reactance (obtained from the fault levels at the PV generators). However, we selected the former representation

118

Chapter Five

because of the difficulties in finding the parameters of the equivalent generator.

Fig. 5-4: The OETC represented by constant impedance loads (the power converters are removed for simplicity)

The OETC is represented by two constant impedance loads Z1 and Z2 seen at PV1 and PV2 respectively. System Data The system data for both Ibri and Manah power stations are obtained from the MIS DlgSILENT model and from OETC Engineers. Table 5-1 illustrates the line data and Table 5-2 shows the transformer data for the network presented in Fig. 5-5. Using the data given in Table 5-1 and Table 5-2, the data of Fig. 5-4 are determined as tabulated in Table 5-3. The model of the Ibri-Manah Line with parameters in per unit (Sbase 100 MVA, Vbase 400 kV) is presented in Figure 5-6.

Fig. 5-5: Connection of PV stations to the MIS

Impact of the Integration of Large-Scale PV Power Plants on the Grid Stability and Operation

Fig. 5-6: Ibri-Manah line

119

400/220kV 500MVA Ibri IPP TX

Name 0.044149

p.u.

0.0003

x (Sbase)

p.u.

1.766

2.56

2.56

kA

Irated

r (Sbase)

3

220kV Ibri IPP BB

220kV Ibri Solar BB

220kV OHL Ibri IPP-Ibri Solar

Table 5-2: Transformer data

0.15

Reactor

400kV Ibri IPP BB

400kV OHL Ibri IPP-Manah Solar (REC)

245

km

Reactor

Busbar

Busbar

Length

400kV Manah Solar BB

Terminal j

Terminal i

400kV OHL Ibri IPP-Manah Solar

Name

Table 5-1: Line data

120

-0.015

p.u.

b (Sbase)

0.00016

1.98E06

0.00323

p.u.

r (Sbase)

0

p.u.

r0 (Sbase)

0.00199

2.42E-05

0.039473

p.u.

x (Sbase)

Chapter Five

0.03532

p.u.

x0 (Sbase)

0.005547

0.001058

1.728505

p.u.

b (Sbase)

0.3

%

0

0.000602

0.983873

p.u.

b0 (Sbase)

0

kW

No Load Los.

0.007958

8.34E-05

0.136147

p.u.

x0 (Sbase)

No Load Cur.

0.0007

2.27E05

0.037

p.u.

r0 (Sbase)

Impact of the Integration of Large-Scale PV Power Plants on the Grid Stability and Operation

121

Table 5-3: Data of Fig. 5-4 i 1

Vi 227/11.480 kV

SPVi (500+j0) MVA

Pi 457.87 MW

2

413.94/10.520 kV

(1000+j0) MVA

1029.76 MW

Qi 194.87 MVAR 159.55 MVAR

Note that the reactors in Fig. 5-5 are used to compensate for the charging current in the 400 kV line thus overvoltage is avoided in light loads. So, the line is represented by the short line model, i.e., the capacitance is neglected. Since ܼ௣௨,௡௘௪ = ܼ௣௨,௢௟ௗ ‫כ‬ (5-1)

ௌ್ೌೞ೐,೙೐ೢ ௌ್ೌೞ೐,೚೗೏

‫(כ‬

௏್ೌೞ೐,೚೗೏ ଶ

௏್ೌೞ೐,೙೐ೢ

)

Hence, ܼ௣௨,௡௘௪ = 484*(220/400)2=146.4 0.29Ÿ(referred to the 400 kV ct.).

pu

Æ

.00199*146.4

=

ÆTotal reactance=Xs=133.65ŸÆLs=425.6 mH (Fig. 5-6) Using Table 5-1, the total resistance:0.00323pu*1600Ÿ+0.00016pu*164.4ŸÆRs =5.2Ÿ (Fig. 5-7) Transformer & filter at Ibri PV generation: 2x335 MVA 33kV/220 kV tr. The following assumptions are used. X୲୰ଵ = 16% (600MVA, 50Hz, 16.5/400kV െ from Digsilent) Referring to Sbase=100MVAÆ ܺ௧௥ଵ = 0.16 ‫ כ‬100‫ܣܸܯ‬/600‫ = ܣܸܯ‬0.0267pu Substituting ૙. ૚૜૟૚‫ܕ‬۶ (۴ܑ܏. 5 െ ૠ)

Zୠୟୱୣ = 1600ÆX୲୰ଵ = 42.7200π, ‫ܚܜۺ‬૚ =

ଡ଼

Assuming = 70ÆR ୲୰ଵ = ૙. ૟૚૙૜π (۴ܑ܏. 5 െ ૠ) ୖ

The filter data for Ibri PV power station are neglected.

Chapter Five

122

Transformer & filter at Manah PV generation Similarly,L୲୰ଶ =

଴.ଵଷ଺ଵ୫ୌ ଶ ୲୰ ୧୬ ୮ୟ୰ୟ୪୪ୣ୪

= ૙. ૙૟ૡ‫ܕ‬۶, R ୲୰ଶ =

૙. ૜૙૞π (۴ܑ܏. 5 െ ૠ)

଴.଺ଵ଴ଷπ ଶ ୲୰ ୧୬ ୮ୟ୰ୟ୪୪ୣ୪

=

The filter data for Manah PV are neglected.

Constant impedance model of the MIS Using the power injected into the grid by the PV stations, the OETC can be modeled by constant impedance as depicted in Fig. 5-7. Each PV station is represented by a three-phase, controlled voltage source and a series RL branch. Each load is modeled by an equivalent parallel RLC network. The line connecting the Ibri to the Manah PV station is represented by lumped series RL elements.

Fig. 5-7: Equivalent circuit for the Ibri (DG1) and Manah PV (DG2) stations connected to the OETC

Assumptions: ஼

- The load circuit should have a quality factor Q = Rට ൑1.8 ௅

- The delivered reactive power to the load is calculated by the following equation: Q = |ܸ ଶ |(

ଵ

௪௅

െ ‫)ܥݓ‬

(5-2)

- The inductor quality factor =120=w0 L/RL From the above data, the impedance models can be derived for the OETC seen at Ibri (Z1) and at Manah (Z2).

Impact of the Integration of Large-Scale PV Power Plants on the Grid Stability and Operation

123

The OETC model Z1 (seen by the Ibri PV station, referred to as the 400 kV ct.): Rଵ = 3( 227KV/1.73)ଶ /(457.87 MW) = 113.16ȳ(in the 220kV circuit) Hence Rଵ = 113.16 ‫( כ‬413.94/227)ଶ = ૜ૠ૟. ૛ૡ૝૛ ȳ (referred to as the 400 KV circuit) ஼

Since the quality factor for the load = Q = 1.8 = Rξ , substituting Rଵ = ௅ 376.2842ÆCଵ /Lଵ = 2.2883 × 10ିହ ଵ

Since ܳ = |ܸ ଶ |( െ ‫ )ܥݓ‬Æsubstituting Qଵ = 194.87/3 MVAR/ ௪௅ phase, V = 413.94/1.73 Æ Lଵ = ૞ૢ૚‫ܕ‬۶, Cଵ = ૚૜. ૞μ۴. Since the inductor quality factor = 120 = w଴ L/R ୐ , substituting w଴ = 314, Lଵ = .591ÆR ୐ଵ = ૚. ૞૝૟૝ષ. Similarly, the OETC impedance model Z2, seen by Manah PV station, can be obtained. A summary of the calculations is presented in Table 5-4 and Figure 5-8. Table 5-4: Electrical Parameters of the equivalent circuit of the study system (Fig. 5-5) DGs

Filter+transformer Rt(Ÿ), Lt(mH) DG1 0.6103, 0.1361 DG2 0.305, 0.068 Line: Rs = 5.2Ÿ, Ls = 0.4256 H

Load parameters R(Ÿ), L(H), Rl(ȍ ), C(μF) 376.2842, 0.591, 1.5464, 13.5 166.29,0.307, 0.8, 35.96

Fig. 5-8: Equivalent circuit for the study system

Chapter Five

124

Mathematical model of the study system The proposed decentralized tracking control is developed based on a linearized model of the study system of Fig. 1-7 in a synchronously rotating dq frame. The controller is based on the fundamental frequency component of the system. The study system of Fig. 5-7 is virtually portioned into two subsystems. The mathematical model of subsystem 1, in the abc frame, is ݅௧ଵ,௔௕௖ = ݅௦,௔௕௖ + ‫ܥ‬ଵ ‫ݒ‬ሶ ଵ,௔௕௖ + ݅௅ଵ,௔௕௖ + ‫ݒ‬௧ଵ,௔௕௖ = ‫ܮ‬௧ଵ

‫ݒ‬ଵ,௔௕௖ ܴଵ

݀݅௧ଵ,௔௕௖ + ܴ௧ଵ ݅௧ଵ,௔௕௖ + ‫ݒ‬ଵ,௔௕௖ ݀‫ݐ‬

‫ݒ‬ଵ,௔௕௖ = ‫ܮ‬ଵ

݀݅௅ଵ,௔௕௖ + ܴ௟ଵ ݅௅ଵ,௔௕௖ ݀‫ݐ‬

‫ݒ‬ଵ,௔௕௖ = ‫ܮ‬௦ (5-3)

ௗ௜ೞ,ೌ್೎ ௗ௧

+ ܴ௦ ݅௦,௔௕௖ + ‫ݒ‬ଶ,௔௕௖

Note that R1 is part of Z1 (the OETC constant impedance load) seen by PV1-Ibri. Other parameters in equation (5-1) are given in Table 5-3, and Fig. 5-8. Where ‫ݔ‬௔௕௖ is a 3x1 vector, and assuming a three-wire system, (5-3) is transformed to the synchronously rotating dq frame (Moradi, Karimi and Ghartemani, “Robust Decentralized Control for Islanded Operation of Two Radially Connected DG Systems”, 2272-2277).

݂ௗ௤

ଶ

cos (ߠ െ ߨ)

ସ

cos (ߠ െ ߨ) ‫ې‬ ଷ ସ ‫ۑ‬ െsin (ߠ െ ߨ) െsin (ߠ െ ߨ)‫ۑ‬ ଷ ଷ ‫ۑ‬ ଵ ଵ ‫ے‬ ξଶ ξଶ

‫ߠݏ݋ܿ ۍ‬ ଶ‫ێ‬ = ‫ێ‬െ‫ߠ݊݅ݏ‬ ଷ ‫ ێ‬ଵ ‫ ۏ‬ξଶ

ଷ ଶ

௧

(5-4)

where ߠ(‫׬ = )ݐ‬଴ ‫ ߬݀)߬(ݓ‬+ ߠ଴ is the phase angle and w is the angular frequency of the crystal oscillator internal to DG. The angle ߠ(‫ )ݐ‬is used to transform from the abc frame to the dq frame. Based on (5-3) and (5-4), the mathematical model of subsystem 1 in the dq frame is

Impact of the Integration of Large-Scale PV Power Plants on the Grid Stability and Operation

ሶ ܸଵ,ௗ௤ =

125

ܸଵ,ௗ௤ 1 1 1 ‫ܫ‬௧ଵ,ௗ௤ െ ‫ܫ‬௦,ௗ௤ െ ‫ܫ‬௅ଵ,ௗ௤ െ െ ݆‫ܸݓ‬ଵ,ௗ௤ ‫ܥ‬ଵ ܴଵ ‫ܥ‬ଵ ‫ܥ‬ଵ ‫ܥ‬ଵ ሶ ‫ܫ‬௧ଵ,ௗ௤ = ሶ ‫ܫ‬௅ଵ,ௗ௤ =

ଵ ܸ ௅೟భ ௧ଵ,ௗ௤

െ

ோ೟భ

‫ܫ‬ ௅೟భ ௧ଵ,ௗ௤

െ

ଵ ܸ ௅೟భ ଵ,ௗ௤

െ ݆‫ܫݓ‬௧ଵ,ௗ௤

1 ܴ௟ଵ ܸ െ ‫ܫ‬ െ ݆‫ܫݓ‬௅ଵ,ௗ௤ ‫ܮ‬ଵ ଵ,ௗ௤ ‫ܮ‬ଵ ௅ଵ,ௗ௤ ሶ ‫ܫ‬௦,ௗ௤ =

(5-5)

ଵ ܸ ௅ೞ ଵ,ௗ௤

െ

ோೞ

‫ܫ‬ ௅ೞ ௦,ௗ௤

െ

ଵ ܸ ௅ೞ ଶ,ௗ௤

െ ݆‫ܫݓ‬௦,ௗ௤

Similarly, the dq-frame-based models of subsystem 2, are also developed and used to construct the state space model of the overall system ‫ݔ‬ሶ = ‫ ݔܣ‬+ ‫ݑܤ‬ ൠ ‫ݔܥ = ݕ‬

(5-6)

where ‫ ݔ‬ᇱ = [ܸଵ,ௗ , ܸଵ,௤ , ‫ܫ‬௧ଵ,ௗ , ‫ܫ‬௧ଵ,௤ , ‫ܫ‬௅ଵ,ௗ , ‫ܫ‬௅ଵ,௤ , ‫ܫ‬௦,ௗ , ‫ܫ‬௦,௤, ܸଶ,ௗ , ܸଶ,௤ , ‫ܫ‬௧ଶ,ௗ , ‫ܫ‬௧ଶ,௤ , ‫ܫ‬௅ଶ,ௗ , ‫ܫ‬௅ଶ,௤ ] ‫ݑ‬ᇱ = ൣܸ௧ଵ,ௗ , ܸ௧ଵ,௤ , ܸ௧ଶ,ௗ , ܸ௧ଶ,௤ ൧, ‫ ݕ‬ᇱ = [ܸଵ,ௗ , ܸଵ,௤ , ܸଶ,ௗ , ܸଶ,௤ ] with (. )ᇱ meaning the transpose of (.), Vt1,dq, Vt2,dq, It1,dq, It2,dq, IL1,dq, IL2,dq, Is,dq, V1,dq, and V2,dq are, respectively, the space-vector representations of vt1,abc, vt2,abc, it1,abc, it2,abc, iL1,abc, iL2,abc, is,abc, v1,abc, and v2,abc in the dq frame. The matrices ‫ܣ‬, ‫ ܥ ݀݊ܽ ܤ‬are of compatible dimensions, i.e., ‫א ܣ‬ ܴଵସ×ଵସ , ‫ܴ א ܤ‬ଵସ×ସ , ‫ܴ א ܥ‬ସ×ଵସ and ‫ܣ‬ ‫ = ܣ‬൤ ଵଵ ‫ܣ‬ଶଵ

‫ܣ‬ଵଶ ‫ܤ‬ ൨,‫ = ܤ‬൤ ଵ 0 ‫ܣ‬ଶଶ

‫ڭ‬ ‫ڭ‬

0 ൨ = [‫ܤ‬ଵ‫כ‬ ‫ܤ‬ଶ

From (5-3), the submatrices are given as

‫ܥ‬ଵ ‫ܤ ڭ‬ଶ‫] כ‬, ‫ = ܥ‬൥ ‫ڮ‬ 0

0 ‫כܥ‬ ‫ ڮ‬൩ = ൤ ଵ‫ כ‬൨ ‫ܥ‬ଶ ‫ܥ‬ଶ

‫‪Chapter Five‬‬

‫ې ‪0‬‬ ‫ۑ‬ ‫ۑ ‪െ1‬‬ ‫ۑ ‪ଵ‬ܥ‬ ‫ۑ‬ ‫ۑ ‪0‬‬ ‫ۑ‬ ‫ۑ ‪0‬‬ ‫‪,‬ۑ‬ ‫ۑ ‪0‬‬ ‫ۑ‬ ‫ۑ‬ ‫ۑ ‪0‬‬ ‫ۑ‬ ‫ۑ ݓ‬ ‫ۑ‬ ‫ۑ ‪ܴ௦‬‬ ‫‪െ‬‬ ‫ے ‪௦‬ܮ‬

‫‪126‬‬

‫‪1‬‬ ‫‪ଵ‬ܥ‬

‫‪1‬‬ ‫‪ଵ‬ܥ‬

‫ݓ‬

‫‪0‬‬

‫‪െ1‬‬ ‫‪ଵ‬ܥ ‪ܴଵ‬‬

‫‪ܴ௧ଵ‬‬ ‫‪௧ଵ‬ܮ‬

‫‪െ1‬‬ ‫‪ଵ‬ܥ‬

‫‪0‬‬

‫‪െ1‬‬ ‫‪ଵ‬ܥ‬

‫‪0‬‬

‫‪െ1‬‬ ‫‪ଵ‬ܥ‬

‫‪0‬‬

‫‪0‬‬

‫‪0‬‬

‫‪0‬‬

‫ݓ‬

‫‪0‬‬

‫‪0‬‬

‫‪0‬‬

‫‪ܴ௧ଵ‬‬ ‫‪௧ଵ‬ܮ‬

‫‪0‬‬

‫ݓ‬

‫‪ܴ௟ଵ‬‬ ‫‪ଵ‬ܮ‬

‫‪0‬‬

‫‪ܴ௟ଵ‬‬ ‫‪ଵ‬ܮ‬

‫‪0‬‬

‫‪െ‬‬

‫‪0‬‬

‫ݓ‪െ‬‬

‫‪െ1‬‬ ‫‪௧ଵ‬ܮ‬

‫‪0‬‬

‫‪0‬‬

‫‪0‬‬

‫ݓ‪െ‬‬

‫‪0‬‬

‫‪0‬‬

‫‪1‬‬ ‫‪ଵ‬ܮ‬

‫‪െ‬‬

‫‪0‬‬

‫‪0‬‬

‫‪0‬‬

‫‪0‬‬

‫‪0‬‬

‫ݓ‪െ‬‬

‫‪0‬‬

‫‪0‬‬

‫‪0‬‬

‫‪0‬‬

‫‪1‬‬ ‫‪௦‬ܮ‬

‫‪ܴ௦‬‬ ‫‪௦‬ܮ‬

‫‪െ‬‬

‫‪0‬‬ ‫‪0‬‬ ‫‪0‬‬ ‫‪0‬‬ ‫‪0‬‬ ‫‪0‬‬

‫‪0‬‬ ‫ې‪0‬‬ ‫ۑ‬ ‫ۑ‪0‬‬ ‫ۑ‪0‬‬ ‫ۑ‪0‬‬ ‫‪ଶଵ‬ܣ ‪ ,‬ۑ‪0‬‬ ‫ۑ‬ ‫ۑ‪0 0‬‬ ‫ۑ‬ ‫ۑ‪0 0‬‬ ‫ے‬ ‫‪ଶ‬ܥ‪1/‬‬ ‫‪0‬‬ ‫ې‬ ‫‪0‬‬ ‫‪ଶ‬ܥ‪1/‬‬ ‫ۑ‬ ‫‪0‬‬ ‫ۑ ‪0‬‬ ‫‪0‬‬ ‫ۑ ‪0‬‬ ‫‪0‬‬ ‫ۑ ‪0‬‬ ‫‪0‬‬ ‫ے ‪0‬‬

‫‪െ‬‬

‫‪0‬‬ ‫‪0‬‬ ‫‪0‬‬ ‫‪0‬‬ ‫‪0‬‬ ‫‪0‬‬

‫‪െ‬‬

‫‪0‬‬ ‫‪0‬‬ ‫‪0‬‬ ‫‪0‬‬ ‫‪0‬‬ ‫‪0‬‬

‫‪0‬‬ ‫‪0‬‬ ‫‪0‬‬ ‫‪0‬‬ ‫‪0‬‬ ‫‪0‬‬

‫‪0 0‬‬

‫‪0‬‬

‫‪0 0‬‬ ‫‪0‬‬ ‫‪0‬‬ ‫‪0‬‬ ‫‪0‬‬ ‫‪0‬‬ ‫‪0‬‬

‫‪0‬‬ ‫‪0‬‬ ‫‪0‬‬ ‫‪0‬‬ ‫‪0‬‬ ‫‪0‬‬

‫‪െ1‬‬ ‫‪௦‬ܮ‬ ‫‪0‬‬ ‫‪0‬‬ ‫‪0‬‬ ‫‪0‬‬ ‫‪0‬‬ ‫‪0‬‬

‫‪0‬‬ ‫‪0‬ۍ‬ ‫ێ‬ ‫‪0‬ێ‬ ‫‪0‬ێ‬ ‫‪0‬ێ‬ ‫‪ 0‬ێ = ‪ଵଶ‬ܣ‬ ‫‪െ1‬ێ‬ ‫ܮێ‬ ‫‪ ௦‬ێ‬ ‫‪0‬ێ‬ ‫ۏ‬ ‫‪0 0 0‬‬ ‫ۍ‬ ‫‪0 0 0‬‬ ‫ێ‬ ‫‪0 0 0‬ێ =‬ ‫‪0 0 0‬ێ‬ ‫‪0 0 0‬ێ‬ ‫‪0 0 0‬ۏ‬

‫‪െ1‬‬ ‫ۍ‬ ‫‪ଵ‬ܥ ‪ܴଵ‬ێ‬ ‫ݓ‪ െ‬ێ‬ ‫ێ‬ ‫‪ െ1‬ێ‬ ‫ܮێ‬ ‫‪௧ଵ‬‬ ‫ێ‬ ‫‪ 0‬ێ‬ ‫ێ=‬ ‫‪ 1‬ێ‬ ‫‪ଵ‬ܮ ێ‬ ‫ێ‬ ‫‪ 0‬ێ‬ ‫‪ 1‬ێ‬ ‫ێ‬ ‫‪௦‬ܮ ێ‬ ‫‪ 0‬ێ‬ ‫ۏ‬

‫‪ଵଵ‬ܣ‬

Impact of the Integration of Large-Scale PV Power Plants on the Grid Stability and Operation

‫ܣ‬ଶଶ =

െ1 ‫ۍ‬ ‫ܴێ‬ଶ ‫ܥ‬ଶ ‫ ێ‬െ‫ݓ‬ ‫ێ‬ ‫ ێ‬െ1 ‫ێ‬ ‫ܮ ێ‬௧ଶ ‫ێ‬ ‫ێ‬ ‫ێ‬ ‫ێ‬ ‫ێ‬ ‫ێ‬ ‫ۏ‬

0 1 ‫ܮ‬ଶ 0

‫ݓ‬

1 ‫ܥ‬ଶ

െ1 ܴଶ ‫ܥ‬ଶ

0

1 ‫ܥ‬ଶ

ܴ௧ଶ ‫ܮ‬௧ଶ

‫ݓ‬

0

ܴ௧ଶ ‫ܮ‬௧ଶ

0

0 െ1 ‫ܮ‬௧ଶ

െ‫ݓ‬

0

0

0

1 ‫ܮ‬ଶ

0

0

0 ‫ ۍ‬0 ‫ێ‬1/‫ܮ‬ ௧ଵ ‫ێ‬ 0 ‫ێ‬ ‫ܤ‬1 = ‫ ێ‬0 ‫ ێ‬0 ‫ ێ‬0 ‫ ۏ‬0 ‫ܥ‬ଵ = ቂ

െ

െ1 ‫ܥ‬ଶ

0

െ

0

െ

0 0 0 ‫ې‬ ‫ ۍ‬0 0 ‫ۑ‬ ‫ۑ‬ ‫ێ‬ 1/‫ܮ‬௧ଶ 1/‫ܮ‬௧ଵ ‫ۑ‬ , ‫ܤ‬ଶ = ‫ێ‬ 0 ‫ۑ‬ ‫ ێ‬0 ‫ ێ‬0 0 ‫ۑ‬ ‫ ۏ‬0 0 ‫ۑ‬ 0 ‫ے‬

ܴ௟ଶ ‫ܮ‬ଶ

െ‫ݓ‬

127

0 ‫ې‬ ‫ۑ‬ െ1 ‫ۑ‬ ‫ܥ‬ଶ ‫ۑ‬ ‫ۑ‬ 0 ‫ۑ‬ ‫ۑ‬ 0 ‫ۑ‬ ‫ۑ‬ ‫ۑ‬ ‫ۑ ݓ‬ ‫ۑ‬ ܴ௟ଶ ‫ۑ‬ െ ‫ܮ‬ଶ ‫ے‬

0 0 ‫ې‬ ‫ۑ‬ 0 ‫ۑ‬ 1/‫ܮ‬௧ଶ ‫ۑ‬ 0 ‫ۑ‬ 0 ‫ے‬

1 0 0 0 0 0 0 0 1 0 0 0 0 0 ቃ,‫ = ܥ‬ቂ ቃ 0 1 0 0 0 0 0 0 ଶ 0 1 0 0 0 0

Substituting the values from Table 5-4, and Fig. 5-8, the PV-OETC study system matrices are:

A

74074 0 0 0 0 0 0 0 0 0 º  74074  74074 ª  196.86 314.16 « 314.14  196.86 0 74074 0 0 0 0 0 0 0 0 »»  74074  74074 « « 7347.5 0 0 0 0 0 0 0 0 0 0 0 »  4484.2 314.16 » « 0 0 0 0 0 0 0 0 0 0 »  7347.5  314.16  4484.2 « 0 « 1.692 0 0 0 0 0 0 0 0 0 0 0 »  2.6166 314.16 » « 1.692 0 0 0 0 0 0 0 0 0 0 »  314.16 2.6166 « 0 « 2.3496 0 » 0 0 0 0 0 0 0 0 0  12.218 314.16  2.3496 » « 2.3496 0 0 0 0 0 0 0 0 0 »  314.16  12.218  2.3496 « 0 « 0 0 0 0 0 0 27809 0 27809 0 0 »  167.23 314.16  27809 » « « 0 0 0 0 0 0 0 27809  314.16  167.23 0 27809 0  27809» » « 0 0 0 0 0 0 0 0 14706 0 4485 . 3 314 . 16 0 0   » « « 0 0 0 0 0 0 0 0 0 0 0 0 »  14706  314.16 » « 0 0 0 0 0 0 0  2.6059 314.16 » 3.2573 0 0 0 « 0 «¬ 0  314.16 2.6059 »¼ 0 0 0 0 0 0 0 0 3.2573 0 0

(5-7)

Chapter Five

128

0 ‫ ۍ‬0 ‫ێ‬7347.5 ‫ێ‬ 0 ‫ܤ‬1 = ‫ێ‬ ‫ ێ‬0 ‫ ێ‬0 ‫ ێ‬0 ‫ ۏ‬0

0 0 0 ‫ې‬ ‫ ۍ‬0 0 ‫ۑ‬ ‫ۑ‬ ‫ێ‬ 7347.5 ‫ۑ‬ 7347.5 , ‫ܤ‬ଶ = ‫ێ‬ 0 ‫ۑ‬ ‫ ێ‬0 ‫ ێ‬0 0 ‫ۑ‬ ‫ ۏ‬0 0 ‫ۑ‬ 0 ‫ے‬

0 0 ‫ې‬ ‫ۑ‬ 0 ‫ۑ‬ 7347.5‫ۑ‬ 0 ‫ۑ‬ 0 ‫ے‬

Using the MATLAB command, the open-loop eigenvalues of the matrix A in (5-6) are obtained as given in Table 5-5. Table 5-5: Open-loop eigenvalues െ2339.3 െ2339.3 െ2280.5 െ128.55 െ14.156 െ3.5119 െ3.0798

± ݆ 23551 ± ݆ 23551 ± ݆ 20108 ± ݆ 20229 ± ݆ 314.16 ± ݆ 314.16 ± ݆ 314.16

Although the system is stable, it has a small degree of stability (the real part of the rightmost eigenvalue = 3.0798). This results in a slow dynamic response. It also has non-zero steady state error for the unit step input, as presented in Fig. 5-9.

Impact of the Integration of Large-Scale PV Power Plants on the Grid Stability and Operation

129

(a)

(b) Fig. 5-9: Unit step response without controllers: (a) Vd1, Vq1, (b) Vd2, Vq2

It is evident that Vd2, Vq2 do not track the reference. To rectify this situation, a controller has to be designed to achieve a fast response and tracks the input with zero steady state error.

Chapter Five

130

Design of Decentralized tracking controller In control system design, the output has to follow the input. If the input is constant, the control problem is termed a regulator problem. If the input is time varying, it is called a tracking problem. In this section, a decentralized tracking controller for the system (equation 5-6) is designed. Designed controllers are decentralized in the sense that they use only the local information of their subsystem. Centralized vs. decentralized control can be summarized as follows. The benefit of using decentralized control via local subsystem information is avoiding using a hub computer (controller) whose failure will cause the loss of the global system stability. It also avoids a costly communication network, and its associated delay to transmit the information of the whole system to the centralized controller. It is first shown that the system of (5-6) can be represented by an interconnected composite system consisting of two subsystems. Then, it will be shown that each subsystem can be controlled by using only local controllers about each subsystem (Shaha, Mithulananthana, Bansal and Ramachandaramurth 2015, 1423-1443). The system (5-6) can be written as ‫ݔ‬ሶ ‫ܣ‬ ൤ ଵ ൨ = ൤ ଵଵ ‫ܣ‬ଶଵ ‫ݔ‬ሶ ଶ

‫ܣ‬ଵଶ ‫ݔ‬ଵ ‫ܤ‬ ൨ቂ ቃ + ൤ ଵ 0 ‫ܣ‬ଶଶ ‫ݔ‬ଶ

0 ‫ݑ‬ଵ ‫ݕ‬ଵ ‫ܥ‬ ൨ቂ ቃ,ቂ ቃ = ൤ ଵ ‫ܤ‬ଶ ‫ݑ‬ଶ ‫ݕ‬ଶ 0

0 ‫ݔ‬ଵ ൨ቂ ቃ ‫ܥ‬ଶ ‫ݔ‬ଶ

(5-8)

where ᇱ

‫ݔ‬ଵ = ൣܸଵ,ௗ , ܸଵ,௤ , ‫ܫ‬௧ଵ,ௗ , ‫ܫ‬௧ଵ,௤ , ‫ܫ‬௅ଵ,ௗ , ‫ܫ‬௅ଵ,௤ , ‫ܫ‬௦,ௗ , ‫ܫ‬௦,௤ ൧ , ‫ݔ‬ଶ = [ܸଶ,ௗ , ܸଶ,௤ , ‫ܫ‬௧ଶ,ௗ , ‫ܫ‬௧ଶ,௤ , ‫ܫ‬௅ଶ,ௗ , ‫ܫ‬௅ଶ,௤ ]ᇱ ᇱ

‫ݑ‬ଵ = ൣ ܸ௧ଵ,ௗ , ܸ௧ଵ,௤ ]ᇱ , ‫ݑ‬ଶ = [ ܸ௧ଶ,ௗ , ܸ௧ଶ,௤ ൧ , ‫ݕ‬ଵ = [ܸଵ,ௗ , ܸଵ,௤ ]ᇱ , ‫ݕ‬ଶ = [ܸଶ,ௗ , ܸଶ,௤ ]ᇱ Remarks: Remark I: The system shown in equation (5-8) is an interconnected composite system consisting of the following two subsystems: ‫ݔ‬ଵሶ = ‫ܣ‬ଵଵ ‫ݔ‬ଵ + ‫ܤ‬ଵ ‫ݑ‬ଵ + ‫ܣ‬ଵଶ ‫ݔ‬ଶ , ‫ݕ‬ଵ = ‫ܥ‬ଵ ‫ݔ‬ଵ ‫ݔ‬ଶሶ = ‫ܣ‬ଶଶ ‫ݔ‬ଶ + ‫ܤ‬ଶ ‫ݑ‬ଶ + ‫ܣ‬ଶଵ ‫ݔ‬ଵ , ‫ݕ‬ଶ = ‫ܥ‬ଶ ‫ݔ‬ଶ where ‫ܣ‬ଵଶ , ‫ܣ‬ଶଵ are termed the interconnection matrices.

(5-9) (5-10)

Impact of the Integration of Large-Scale PV Power Plants on the Grid Stability and Operation

131

Remark II: It can be shown that subsystems presented in equations (5-9) and (5-10), i.e., (C1, A11, B1) and (C2, A22, B2), are controllable and observable for the system parameters given in Table 5-4. Therefore, according to the results of (Kundur et al. 2004, 1387-1401), the composite system (5-8) is stabilizable by using only local controllers about the subsystems, for nonzero interconnection gain matrices A12 and A21. This implies that the decentralized control strategy can be used to control each subsystem. From the above analysis, the global system (5-6) can alternatively be written as ‫ݔ‬ሶ = ‫ ݔܣ‬+ ‫ܤ‬ଵ‫ݑ כ‬ଵ + ‫ܤ‬ଶ‫ݑ כ‬ଶ ൠ ‫ݕ‬ଵ = ‫ܥ‬ଵ‫ ݔ כ‬, ‫ݕ‬ଶ = ‫ܥ‬ଶ‫ݔ כ‬ where ‫ܤ‬ଵ‫ = כ‬ቂ

(5-10)

0 ‫ܤ‬ଵ ቃ , ‫ܤ‬ଶ‫ = כ‬൤ ൨ , ‫ܥ‬ଵ‫ܥ[ = כ‬ଵ 0], ‫ܥ‬ଶ‫[ = כ‬0 ‫ܥ‬ଶ ]. ‫ܤ‬ଶ 0

Eqn. (5-10) can be rewritten as ‫ݔ‬ሶ = ‫ ݔܣ‬+ σଶ௜ୀଵ ‫ܤ‬௜‫ݑ כ‬௜ , ‫ݕ‬௜ = ‫ܥ‬௜‫ݔ כ‬, ݁௜ = ‫ݕ‬௥௘௙,௜ – ‫ݕ‬௜ , ݅ = 1,2 (5-11) where ‫ݑ‬௜ and ‫ݕ‬௜ are the inputs and outputs of control agent i, and ݁௜ is the tracking error between the reference signal ‫ݕ‬௥௘௙,௜ and the output ‫ݕ‬௜ of the control agent , ݅ = 1,2.

Controller Design Requirements for the Decentralized reference tracking problem A decentralized controller for the plant (5-12) must have the following desired features. 1) A closed-loop system that is asymptotically stable. 2) Steady-state asymptotic tracking for all set-points ‫ݕ‬௥௘௙,ଵ , ‫ݕ‬௥௘௙,ଶ (3) The controller action should be “fast”. Two decentralized controller designs are proposed: (i) Output feedback with integral control, and (ii) State feedback with integral control.

Chapter Five

132

Design 1: decentralized output feedback + integral Controller For the output to track the input of the plant (5-12), it is necessary that the decentralized controller should include the decentralized servocompensator (integral control) ߚపሶ = 0ߚ௜ + ൫‫ݕ‬௥௘௙,௜ െ ‫ݕ‬௜ ൯, ݅ = 1,2

(5-12)

where ߚ௜ ‫ܴ א‬ଶ , ݅ = 1,2 , together with a decentralized compensator with the structure ‫ܭ = ݑ‬௉ ‫ ݕ‬+ ‫ܭ‬ூ ߚ

(5-13)

where ‫ܭ‬௉,ଵ ‫ܭ‬௉ = ൤ 0

0 ‫ܭ‬ூ,ଵ ൨ , ‫ܭ‬ூ = ൤ ‫ܭ‬௉,ଶ 0

0 ൨, ‫ܭ‬ூ,ଶ

so that the controlled closed-loop system is described by ‫ݔ‬ ‫ݔ‬ሶ ‫ ܣ‬+ ‫ܭܤ‬௉ ‫ܥ‬ 0 ൤ ሶ ൨ = ‫ܣ‬௖ ቂߚቃ + ቂ ቃ ‫ݕ‬௥௘௙ , ‫ܣ‬௖ = ቂ ߚ ‫ܫ‬ െ‫ܥ‬

‫ܭܤ‬ூ ቃ , ‫ܥ[ = ݕ‬ 0

‫ݔ‬ 0] ቂߚቃ (5-14)

The proposed control design is based on shifting the closed-loop, rightmost eigenvalues (i.e., having the largest real part, or being the least stable) to the left as much as possible. This approach results in the fastest response that can be obtained. In this case, the controller parameters (5-13) are designed for (5-8) to achieve the fast response by minimizing the performance index ‫ = ܬ‬max ቀ‫ߣ ݈ܽ݁ݎ‬௝ (‫ܣ‬௖ )ቁ , ‫ߣ׊‬௝

(5-15)

where ߣ௝ (. ) = ݁݅݃݁݊‫( ݂݋ ݏ݁ݑ݈ܽݒ‬. ). The control design problem becomes, therefore, a min-max nonlinear optimization problem whose solution yields the controller parameters. To avoid sticking into a local minimum (resulting in a slow response), the optimization problem is solved using particle swarm optimization (probabilistic optimization). The decentralized output feedback PI servo controller, obtained by using the MATLAB command “particleswarm”, is given by

Impact of the Integration of Large-Scale PV Power Plants on the Grid Stability and Operation

‫ܭ‬௉ = 80.091 െ40.765 0 0

െ175.69 92.036 ൦ 0 0

െ157.28 െ723.32 ‫ܭ‬ூ = ൦ 0 0

0 0 0 0 ൪ െ0.26857 1.5302 19.57 െ13.529

133

(5-16)

601.5 0 0 98.799 0 0 ൪ 0 135.23 െ125.3 0 76.132 166.28

Using this controller, the best achievable degree of stability for the closedloop system matrix ‫ܣ‬௖ is obtained as 7.6614. Design 2: decentralized state feedback + integral Controller In this case (5-13) is replaced by ‫ܭ = ݑ‬௉ ‫ ݔ‬+ ‫ܭ‬ூ ߚ

(5-18)

and the closed-loop matrix (5-14) is replaced by ‫ܣ‬௖ = ቂ

‫ ܣ‬+ ‫ܭܤ‬௉ െ‫ܥ‬

‫ܭܤ‬ூ ቃ 0

Similarly, as before, the min-max minimization problem is solved using the MATLAB command “particleswarm”. The obtained decentralized state feedback servo controller for the study system (5-7) is given by KP=

1.6݁6 1.38݁5 ‫ܭ‬ூ = ൦ 0 0

39191 0 0 1.83݁6 0 0 ൪ 0 1.015݁6 െ6.25݁5 0 20291 1.374݁6

Using this controller, the best achievable degree of stability for the closedloop system matrix ‫ܣ‬௖ is obtained as 79.519.

134

Chapter Five

Simulation results The unit step response using design 1 is shown in Figure 5.10.

(a)

(b) Fig. 5-10: Unit step response with controller 1 of: (a) Vd1, Vq1, (b) Vd2, Vq2

It is evident that the system with the proposed decentralized controllers can track the reference with zero steady state error.

Impact of the Integration of Large-Scale PV Power Plants on the Grid Stability and Operation

135

The unit step response using design 2 is shown in Fig. 5-11.

(a)

(b) Fig. 5-11: Unit step response with controller 2 of: (a) Vd1, Vq1, (b) Vd2, Vq2

It is clear that controller 2 is much faster than design 1. It tracks the unit step input within 0.1 seconds. The price is that more sensors are needed to measure all the states. Note that an observer could have been designed which uses only the available measured states to estimate the missing ones. However, this results in increased system dynamics and delays.

136

Chapter Five

Transient Stability Recently, tremendous changes in the infrastructure of power systems have been witnessed around the world. The infrastructural changes include the increasing integration of renewable energy and promoting the establishment of microgrids to decrease the dependence in classical power production. The study of transient stability in which the system is subject to severe disturbances e.g., faults become a crucial issue, and its importance, are going to increase with the large penetration of Distributed Energy Resources (DERs) or Inverter-Based Resources (IBRs). Power system transient stability study covers three types: (i) rotor angle stability, (ii) voltage stability, and (iii) frequency stability (Kundur et al. 2004, 1387-1401). These performance characteristics must be investigated for existing power systems integrated with large PV/wind farms. The voltage stability can be either static or dynamic. This study is limited to dynamic voltage stability. For the Oman Main Interconnected System (MIS), two PV farms of 500 MW and 1000 MW will be installed at Ibri and Manah respectively. The PV model developed by the Renewable Energy Modeling Task Force (REMTF) of the Western Electric Coordinating Council (WECC) (Ellis, Behnke, Elliott 2013, 8) will be incorporated into the MIS of Oman at the predefined locations in order to study and assess the transient stability of the entire grid under different fault scenarios, solar shading, partial outage of PV generation, and load shedding. The transient stability study will only consider the Ibri PV station connection to the MIS, since Manah PV will be connected later in 2023. According to (Oman Electricity Transmission Company 2020, 64-85), the PV output power at the peak (3.00–4.00 pm during summer) is expected to be about 30% of the total PV installed capacity. Therefore, in this study, the output power produced from Ibri PV is assumed to be 150 MW. The simulation is carried out using DlgSILENT with the PV model described in (Ellis, Behnke, Elliott 2013, 8). The following scenarios are investigated: 1- Three-phase to ground fault at the 220 kV Sur power station Bus Bar (BB) for a duration of 100 ms. This fault is selected at Sur power station BB because it is the biggest station in the MIS of Oman; consequently it represents the severest case. The fault duration is so selected because modern circuit breakers clear faults within 4 cycles (80 ms), for a 50 Hz system. 2- Partial shading at Ibri PV plant that resulted in the loss of PV generation of 100 MW for 15 s duration.

Impact of the Integration of Large-Scale PV Power Plants on the Grid Stability and Operation

137

3- Permanent outage of 25 MW of PV power at Ibri. 4- Load shedding of 150 MW.

Three-phase to ground fault at the 220 kV Sur power station for a duration of 100 ms In this fault scenario, a three-phase to ground fault at the 220 kV Sur power station BB is assumed to occur for a duration of 100 ms. The generator Sur ST (5-2) IPP is considered as the reference machine. The results of this fault scenario, namely the rotor angle stability, the voltage stability and the frequency at different buses are shown in Fig. 5-12. The system becomes stable after clearing the fault.

138

(a)

Chapter Five

Impact of the Integration of Large-Scale PV Power Plants on the Grid Stability and Operation

(b)

139

140

Chapter Five

(c) Fig. 5-12: Simulation results due to three-phase to ground fault at Sur power station: (a) Rotor angle at selected generators, (b) Voltage at selected buses and (c) Frequency at selected buses.

Impact of the Integration of Large-Scale PV Power Plants on the Grid Stability and Operation

141

Partial shading at Ibri PV station for 15 s duration In this fault scenario, partial clouding is assumed at Ibri PV station that results in the loss of PV generation of 100 MW. This type of disturbance is assumed to occur at 10 s and lasts for 15 s. Simulation results are presented in Fig. 5-13. The system becomes stable after the disturbance.

142

Chapter Five

(a)

Impact of the Integration of Large-Scale PV Power Plants on the Grid Stability and Operation

(b)

143

144

Chapter Five

(c) Fig. 5-13: Simulation results due to 100 MW power loss (partial shading at Ibri PV station): (a) Rotor angle at selected generators, (b) Voltage at selected buses and (c) Frequency at selected buses

Impact of the Integration of Large-Scale PV Power Plants on the Grid Stability and Operation

145

Permanent outage of 400 MW of PV power at Ibri In this fault scenario, a permanent loss of 400 MW of power generation from Ibri PV station is assumed to occur at a time of 1 s. The results of this disturbance are shown in Fig. 5-14. The system performance showed a stable rotor angle, bus voltages and frequencies.

(a)

-10.00

0.00

10.00

20.00

30.00

[deg]

40.00

0.00

5.00

10.00

Sur GT(1-5) IPP: Rotor angle with reference to reference machine angle Sur ST(1-2) IPP: Rotor angle with reference to reference machine angle Sur ST3 IPP: Rotor angle with reference to reference machine angle Sohar-1 ST1 PS: Rotor angle with reference to reference machine angle Sohar-2 GT1 PS: Rotor angle with reference to reference machine angle Sohar-2 GT2 PS: Rotor angle with reference to reference machine angle Sohar-2 ST1 PS: Rotor angle with reference to reference machine angle Ibri GT(1-4) IPP: Rotor angle with reference to reference machine angle Ibri ST(1-2) IPP: Rotor angle with reference to reference machine angle

15.00

20.00

25.00

[s]

30.00

146 Chapter Five

(b)

0.96

0.98

1.00

1.02

1.04

[p.u.]

1.06

0.00

5.00

132kV Airport Height BB: Voltage, Magnitude 132kV Al Filaj BB: Voltage, Magnitude 132kV Amerat BB: Voltage, Magnitude 132kV Bousher BB: Voltage, Magnitude 132kV Ghala BB: Voltage, Magnitude 132kV Ghala Heights BB: Voltage, Magnitude 132kV Khoudh BB: Voltage, Magnitude 132kV MSQ-1 BB: Voltage, Magnitude 132kV Madinat Barka BB: Voltage, Magnitude 132kV Misfah BB: Voltage, Magnitude 400kV Jahloot BB: Voltage, Magnitude 132kV Mobella BB: Voltage, Magnitude 400kV Qabel BB: Voltage, Magnitude 132kV Rusail BB: Voltage, Magnitude 132kV Rusail Industrial BB: Voltage, Magnitude 220kV Airport Height BB: Voltage, Magnitude 220kV Al Hamriyah BB: Voltage, Magnitude 220kV Misfah BB: Voltage, Magnitude 400kV Sur PS BB: Voltage, Magnitude 220kV Sur BB: Voltage, Magnitude 400kV Misfah BB: Voltage, Magnitude

10.00

15.00

20.00

25.00

29.552 s 0.973 p.u. [s]

30.00

Impact of the Integration of Large-Scale PV Power Plants on the Grid Stability and Operation 147

49.82

49.86

49.90

49.94

49.98

[Hz]

50.02

0.00

5.00

132kV Airport Height BB: Electrical Frequency 132kV Al Filaj BB: Electrical Frequency 132kV Amerat BB: Electrical Frequency 132kV Bousher BB: Electrical Frequency 132kV Ghala BB: Electrical Frequency 132kV Ghala Heights BB: Electrical Frequency 132kV MSQ-1 BB: Electrical Frequency 132kV Madinat Barka BB: Electrical Frequency 132kV Mobella BB: Electrical Frequency 220kV Airport Height BB: Electrical Frequency 220kV Al Hamriyah BB: Electrical Frequency 220kV Misfah BB: Electrical Frequency 220kV Mobella BB: Electrical Frequency 400kV Misfah BB: Electrical Frequency

10.00

15.00

20.00

25.00

[s]

29.312 s 49.846 Hz

30.00

148 Chapter Five

(c)

Fig. 5-14: Simulation results due to a 400 MW PV permanent power loss (a) Rotor angle at Ibri generators, (b) Voltage at Ibri IPP 220 kV bus and (c) Frequency at Ibri IPP 220kV bus

Impact of the Integration of Large-Scale PV Power Plants on the Grid Stability and Operation

149

Load Shedding of 150 MW In this scenario, shedding of the 150 MW load at Blue city is assumed. The results of this type of disturbance are shown in Fig. 5-15. The system becomes stable after the disturbance.

150

(a)

Chapter Five

Impact of the Integration of Large-Scale PV Power Plants on the Grid Stability and Operation

(b)

151

152

Chapter Five

(c) Fig. 5-15: Simulation results due to 150 MW load shedding: (a) Rotor angle at selected generators, (b) Voltage at selected buses, and (c) Frequency at selected buses

Impact of the Integration of Large-Scale PV Power Plants on the Grid Stability and Operation

153

From the simulation results of the above scenarios, it can be concluded that the transient stability in terms of frequency, angle and voltage magnitude will not be affected under the following types of large disturbances: threephase to ground faults at Sur power station; partial clouding on PV stations; and permanent loss of large loads.

Conclusions In this chapter, the decentralized reference tracking, control approach is used to design a control strategy for the operation of the two PV generation stations at Ibri and Manah connected to the OETC. Each PV unit is connected to the PCC through the power electronics converter. The connection of two PV units constructs an interconnected composite system consisting of two subsystems. It is shown that the system can be controlled by the local controllers of each individual subsystem. Then, the particle swarm optimization is used to design a voltage controller for each subsystem. Two PI (proportional-integral) controllers’ designs are suggested. The first uses output feedback with integral control, whereas, the second uses state feedback with integral control. The latter provides a much faster response than the former. Based on simulation case studies in MATLAB/Simulink, the performance of the proposed control strategy is also carried out in this chapter. The simulation results verify the performance of the designed controller when the system is subjected to step changes in the reference signals. To assess the transient stability of the main interconnected system (MIS) with the penetration of large PV power, the overall system was modeled using DlgSILENT software for the capability statement-year of 2021 (with the presence of Ibri PV station). Different types of large disturbances are considered such as the three-phase short circuit at the largest power station (Sur power station), partial shading of the PV plant, large load shedding, and permanent outage of part of the PV power station. For all the fault scenarios considered in this study, the rotor angle stability, frequency stability and voltage stability are not affected.

References Oman Electricity Transmission Company. “Five-Year Annual Transmission Capability Statement (2020–2024)”. Oman Electricity Transmission Company, April, 2020.

154

Chapter Five

https://www.omangrid.com/en/Report/5%20Year%20Annual%20Tran smission%20Capability%20Statement%2020202024.pdf?csrt=2965107877622248160. Berdikeeva, Saltanat. “Renewables Take Center Stage in Oman’s Energy Goals”. Inside Arabia, September 25, 2020. https://insidearabia.com/renewables-take-center-stage-in-omansenergy-goals/. Shaha, R., N. Mithulananthana, R. C. Bansal, and V. Ramachandaramurth. 2015. “A review of key power system stability challenges for large-scale PV Integration”. Renewable and Sustainable Energy Review 41, 1423143. https://doi.org/10.1016/j.rser.2014.09.027. REN21. 2019. Renewables 2019 Global Status Report. Paris: REN21 Secretariat. https://www.ren21.net/wp-content/uploads/2019/05/gsr_ 2019_full_report_en.pdf. Jenkins, N., R. Allan, P. Crossley, D. Kirschen, and G. Strbac. 2000. Embedded Generation. London, U.K.: The Institution of Engineering and Technology. Kundur, P., J. Paserba, V. Ajjarapu, G. Andersson, A. Bose, C. Cañizares, N. Hatziargyriou, D. Hill, A. Stankovic, C. Taylor, T. V. Custem, and V. Vittal. 2004. “Definition and classification of power system stability”. IEEE Trans. Power System 3, 1387-1401. DOI: 10.1109/TPWRS.2004.825981. Moradi, R., H. Karimi, and M. Ghartemani. 2010. “Robust Decentralized Control for Islanded Operation of Two Radially Connected DG Systems”. In IEEE International Symposium on Industrial Electronics, 2272-2277. Bari, Italy: IEEE. DOI:10.1109/ISIE.2010.5637651. Kundur, Prabha. 1994. Power system stability and control. New York: McGraw-Hill Education. Ellis, A., M. Behnke, and R. Elliott. 2013. Generic Solar Photovoltaic System Dynamic Simulation Model Specification. New Mexico: Sandia Report, Sandia National Laboratories Albuquerque. https://prodng.sandia.gov/techlib-noauth/access-control.cgi/2013/138876.pdf.

CHAPTER SIX CAPACITY VALUE OF PHOTOVOLTAICS FOR ESTIMATING THE ADEQUACY OF A POWER GENERATION SYSTEM ARIF MALIK1 AND MOHAMMED ALBADI1

Abstract In generation expansion planning and reliability studies, it is of utmost importance to assess the capacity value of a renewable energy system because such a system is dependent on the availability of resources and hence its rated capacity cannot be considered firm when needed. This chapter addresses the evaluation of the capacity credit or value of renewable energy systems, particularly photovoltaic (PV) plants in a power generation system. The chapter starts by giving an overview of the various adequacy measures used to evaluate generation system reliability and then concentrates on the evaluation of the capacity value of PV plants. Because of the random nature of load demand, the intermittency of renewable energy systems, and unplanned outages of generating units, the probabilistic methods of assessment are preferred over deterministic methods in reliability studies. One of the primary parameters used in the static capacity evaluation is the probability of finding the generating unit on the forced outage at some distant time in the future. A two-state generating unit model is discussed in this regard, which forms the basis for calculating the capacity outage probability table. This table is used in conjunction with the load model to estimate the loss-of-load probability (LOLP) and the loss-of-load expectation (LOLE) indices. These indices are the basis for calculating the capacity credit of any generating unit. There are several methods to estimate capacity credit, and one of the most common and accurate methods is the effective load carrying capability (ELCC). A three-step procedure for the 1 Sultan Qaboos University, Dept. of Electrical and Computer Engineering, Muscat, Oman, Email (Arif Malik): [email protected]

156

Chapter Six

calculation of the capacity credit for the PV system using the ELCC technique is implemented through a small system of three conventional generating units and a PV plant. The number of PV plants is then increased in the generation system to show the effect on capacity credits. This chapter is written with a minimum of technical language and assumptions as to the readers’ prior knowledge. In this way, it can serve as a tutorial on capacity credit that is appropriate for undergraduate students and useful for postgraduate students who want to pursue research in this area. Keywords: Solar PV; Capacity Credit; Adequacy measures; Loss-of-load expectation; LOLP; Effective Load Carrying Capability

Introduction Capacity credit or value, sometimes called effective capacity, is defined as the amount of additional load that can be served due to the addition of the generating unit while maintaining the existing levels of reliability (Simoglou et al. 2013). Alternatively, it is the amount of conventional generation that the alternate energy source would replace without an appreciable change in system performance (Khallat and Rahman 1988). It is important to determine the capacity value of generation resources to ensure that the system reliability is taken care of during power generation planning (Keane et al. 2011). Power generation from renewable resources, such as solar power, varies from hour to hour according to prevailing weather conditions. This is different from fossil-fuel generators, which can usually be dispatched according to their operators’ preferences (which, in a competitive market, means that they can choose to generate when spot prices are sufficiently high). As the penetration of variable renewable generation technologies increases, the variable nature of their output becomes a serious concern in a power system, which affects the system reliability. The reliability of the power system topic involves consideration of the security and adequacy of the system. The concept of security is related to the ability of the system to respond to disturbances arising within that system. Therefore, security measures are related to operational planning. The concept of adequacy, on the other hand, is related to the existence of sufficient facilities (generation, transmission, and distribution) within the system to satisfy the consumer demand in the future. Adequacy, therefore, is concerned with static conditions and does not include system disturbances (Billinton and Allan 1996). Thus, in planning the generation system, the question arises as to how much capacity value or credit to attribute to a given renewable energy

Capacity Value of Photovoltaics for Estimating the Adequacy of a Power 157 Generation System

system. The objective of this chapter is to provide an overview of how to estimate the adequacy of the power generation system and to find the capacity value of photovoltaic solar generation. After the introduction, section 2 reviews the reliability metrics used for assessing the generation adequacy in the electric power industry. The assessment of adequacy measures based on probabilistic analysis recognizes the random outages of generating units beside load variation and the random nature of renewable generation. Therefore, section 3 discusses the basic two-state generating unit model. This model is fundamental to estimate the probability that the load will not be supplied by the generation system in the future. Section 4 concentrates on the evaluation of loss-of-load probability (LOLP) or loss-of-load expectation (LOLE). With an example of a threeunit generating system and an hourly load of 24 hours, it shows how to calculate LOLE using the analytical method of convolution. The estimation of LOLE is the basis to calculate the capacity credit of any generating unit. Section 5 examines the three-step method to evaluate the effective load carrying capability (ELCC) of solar PV. The example of a three-unit generating system is then extended to include PV, and the three-step method follows to assess the ELCC of a given PV plant. Section 6 concludes the chapter.

Adequacy Measures The various adequacy measures or generation reliability indices used in the electric power industry can generally be grouped into two broad categories: (a) deterministic indices; and (b) probabilistic indices. Probabilistic indices permit the quantitative evaluation of system alternatives by taking directly into consideration the parameters that influence reliability, such as the capacities of individual generating units and the unavailability or forced outage rate (FOR) of each unit. These indices can also take into account the inherent uncertainties in the output of renewable energy resources. While deterministic indices are more limited, they are popular because their calculation is simple and requires little or no data, and because acceptable values of the indices are well established and benchmarked against historical experience. Deterministic indices are quite useful when the generation system is mainly conventional thermal generation. The following are the most common deterministic and probabilistic indices to find the generating system adequacy (Buehring, Huber, and De Souza 1984, NERC 2018):

Chapter Six

158

A.

B.

Deterministic Indices A.1 Reserve Margins (RM) A.2 Largest Unit (LU) A.3 Dry Year Probabilistic Indices B.1 Loss-of-Load Probability (LOLP) B.2 Expected Unserved Energy (EUE) B.3 Loss-of-Energy Probability (LOEP) B.4 Expected Loss of Load (XLOL) B.5 Frequency and Duration of Failures to Meet the Load B.6 Effective Load Carrying Capability.

A.1 Reserve Margin: The reserve margin (RM) is a measure of the generating capacity that is available over and above the peak load in the system. For example, if the installed capacity in the system is 12,000 MW and the peak demand is 10,000 MW, then the reserve margin is 20%. Mathematically, ܴ‫= ܯ‬

(ܶ‫ ݕݐ݅ܿܽ݌ܽܿ ݈݈݀݁ܽݐݏ݊݅ ݈ܽݐ݋‬െ ‫)݀݊ܽ݉݁݀ ݇ܽ݁݌ ݈ܽݑ݊݊ܣ‬ × 100% ‫݀݊ܽ݉݁݀ ݇ܽ݁݌ ݈ܽݑ݊݊ܣ‬

A.2 Largest Unit: The loss of the largest generating unit (LU) method is a reliability measure that provides a degree of sophistication over the standard percentage reserve method by reflecting the effect of unit size on the reserve requirements. Mathematically, ‫= ܷܮ‬

(ܶ‫ ݕݐ݅ܿܽ݌ܽܿ ݈݈݀݁ܽݐݏ݊݅ ݈ܽݐ݋‬െ ‫)݀݊ܽ݉݁݀ ݇ܽ݁݌ ݈ܽݑ݊݊ܣ‬ ‫ݐ݅݊ݑ ݐݏ݁݃ݎ݈ܽ ݂݋ ݕݐ݅ܿܽ݌ܽܥ‬

LU is expressed as a multiple of the largest unit capacity. A.3 Dry Year: Dry year is not an index but is instead a criterion. This criterion is used for hydro dominated systems. It is the margin available for a defined “poor” hydro condition. B.1 Loss-of-Load Probability (LOLP): LOLP is a reliability index that indicates the probability that some portion of the load will not be satisfied by the available generating capacity. LOLP is a dimensionless quantity and can be computed on the basis of either daily peak loads or hourly peak loads. A more commonly used index is a loss-of-load expectation (LOLE), which is the summation of LOLP over the period. LOLE is either given in hours/year or days/year and becomes a more meaningful index. While calculating an annual LOLE index, if hourly LOLPs are used then, LOLE

Capacity Value of Photovoltaics for Estimating the Adequacy of a Power 159 Generation System

has units in hours/year, sometimes referred to as loss-of-load hours (LOLH). If daily peaks, also called “daily risks,” are used in computing LOLP, then LOLE is given in days/year. For a given set of conventional generators, the LOLE of the system without the PV plant is calculated as: ்

‫ = ܧܮܱܮ‬෍ ܲ(‫ܩ‬௜ < ‫ܮ‬௜ ) ௜ୀଵ

where T is the total number of hours of study, Gi represents the available conventional capacity in an hour i, and Li is the amount of load. P(Gi < Li) indicates the probability of available generating capacity being less than demand, which is the LOLP in each hour. Adding these LOLPs together gives the LOLE. The calculated LOLE represents the original reliability level of the system. The majority of entities in North America conducting resource adequacy studies primarily use the LOLE metric given in days/year to establish a single resource adequacy on the criterion (2018). Target LOLE levels are typically set in the USA and Europe for long-range planning. For example, the target LOLE frequently used in the USA for large interconnected systems is one day in ten years or 0.1 day/year; in European countries, the corresponding standard varies from one day in fifteen years to one day in two and a half years (Buehring, Huber, and De Souza 1984). B.2 Expected Unserved Energy (EUE): The EUE is the summation of the expected number of megawatt-hours (MWh) of demand that will not be served in a given time period as a result of demand exceeding the available capacity across all hours. The EUE is an energy-centric index which considers the magnitude and duration for all hours of the time period, calculated in MWh (NERC 2018). This index is quite useful for renewable energy resources in the system. B.3 Loss-of-Energy Probability (LOEP): The EUE above has units of energy, and so is specific to the system studied and its current state of development. The LOEP normalizes this index to the total energy demand so that it can be used for comparison with other years or systems. ‫= ܲܧܱܮ‬

‫ܧܷܧ‬ ܶ‫ݐ݊݁݉݁ݎ݅ݑݍ݁ݎ ݕ݃ݎ݁݊݁ ݈ܽݑ݊݊ܽ ݈ܽݐ݋‬

B.4 Expected Loss of Load (XLOL): This attempts to quantify (in MW) the expected magnitude of the unsupplied load given that a failure has occurred. In terms of the above indices:

Chapter Six

160

ܺ‫= ܮܱܮ‬

‫ܧܷܧ‬ ‫ܧܮܱܮ‬

B.5 Frequency and Duration of failures to meet the load: It is possible to quantify the frequency (F) of generating capacity shortage events (as the probability-weighted average), and the duration (D) as the estimated duration of these events using hourly load curves. Physically it is more meaningful to planners and customers – but computationally intensive. ‫ܲܮܱܮ = ܦ × ܨ‬ B.6 Effective Load Carrying Capability (ELCC): ELCC is an index designed to measure the worth of a generating unit to a utility system in terms of reliability (Billinton and Allan 1996, Buehring, Huber, and De Souza 1984). The ELCC concept is explained in Fig. 6-1, which plots annual LOLE versus the annual peak load for a specific generation system both before and after a new unit is added. The original generating system has a LOLE shown by point A, which meets the design level of reliability indicated (0.1 day/year). However, due to load growth, the LOLE will increase to point B, which is unacceptable as it cannot maintain the desired 0.1 day/year LOLE criterion. At this point, a new generating unit will be added to keep the LOLE within the design limit. The addition of a new unit, however, shifts the curve to the right, as shown in Fig. 6-1, and the LOLE decreases from 0.12 day/year (point B) to 0.08 day/year (point C), which is below the desired 0.1 day/year criterion. The ELCC is defined as the difference, measured along the horizontal axis at the design criterion, between the two LOLE versus peak load curves (the difference between points E and D, in MW). This value is a measure of a newly added unit contribution to system capability. As Fig. 6-1 shows, this measure is a function of the generating unit’s characteristics (size, forced outage rate, maintenance requirements) as well as the characteristics of the power system in which it is operating.

Capacity Value of Photovoltaics for Estimating the Adequacy of a Power 161 Generation System

Fig. 6-1: Effective load carrying capability of a unit

Conventional Generating Unit Modelling The probabilistic nature of a generating unit’s outage is essential to assess the probabilistic evaluation of adequacy measures. In the reliability evaluation context, a conventional generating unit can be represented by a basic two-state model. The generating unit can move from an operating (up) state to a failure (down) state at failure rate (Ȝ) and return to an operating state at repair rate (ȝ) after the unit is repaired. Fig. 6-2 shows the two-state model of a generating unit. The two basic parameters of the generating unit used in a static capacity evaluation are the unit availability (A) and the unit unavailability (U), also called the forced outage rate (FOR). The unit availability (A) is defined as the probability of a unit being in the operating state. The unit unavailability (U) is defined as the probability of a unit being in the failure state (Billinton and Allan 1996).

Fig. 6-2: Two-state model of a generating unit of up and down

The two system states and their associated transitions can be shown chronologically on a time graph. The mean values of up and downtimes can be used to give the average performance of this two-state system (Billinton and Allan 1992). This is shown in Fig. 6-3.

Chapter Six

162

Fig. 6-3: Time graph of a two-state model

In Fig. 6-3, the period T is the unit cycle time and is equal to the sum of the mean time to failure (M) and the mean time to repair (R). This cycle time is defined as the mean time between failures (T). The following relationships can, therefore, be defined as: ‫ = ܯ‬1/ߣ

(6-7)

ܴ = 1/ߤ

(6-8)

ଵ

ܶ = =‫ܯ‬൅ܴ

(6-9)

௙

Where f is the cycle frequency. Therefore, ܷ݊ܽ‫= ܷ = )ܴܱܨ( ݕݐ݈ܾ݈݅݅ܽ݅ܽݒ‬

ఒ ఒାఓ

=

ோ ெାோ

=

ோ ்

σሾ஽௢௪௡ ௧௜௠௘ሿ

= σሾ஽௢௪௡௧௜௠௘ሿାσሾ௎௣

௧௜௠௘ሿ

(6-10) ‫ = )ܣ( ݕݐ݈ܾ݈݅݅ܽ݅ܽݒܣ‬1 െ ‫= ܴܱܨ‬

ఓ ఒାఓ

=

ெ ெାோ

=

ெ ்

σሾ௎௣ ௧௜௠௘ሿ

= σሾ஽௢௪௡௧௜௠௘ሿାσሾ௎௣

௧௜௠௘ሿ

(6-11) In addition, sometimes there are partial outages where the unit cannot operate at its full capacity. For such situations, the equivalent forced outage rates (EFOR) are used. The formula to calculate the equivalent forced outage rate is straightforward and derived below. The total period hours consist of: (a) service hours (SH), (b) scheduled outage hours (SOH), (c) forced outage hours (FOH), and (d) reserve shutdown hours (RSH). ܲ‫ ܪܵ = ܪ‬൅ ܱܵ‫ ܪ‬൅ ‫ ܪܱܨ‬൅ ܴܵ‫ܪ‬

(6-12)

Service hours include the hours when the unit generates energy even it has a partial outage. The available hours (AH) are then: ‫ ܪܵ = ܪܣ‬൅ ܴܵ‫ܪ‬

(6-13)

Capacity Value of Photovoltaics for Estimating the Adequacy of a Power 163 Generation System

The equivalent forced outage hours (EFOH) and equivalent forced outage rate (EFOR) can then be found by the following two equations: ‫ = ܪܱܨܧ‬σ[(‫ ݎ݁ܲ( × )ݏݎݑ݋݄ ݁݃ܽݐݑ݋ ݈ܽ݅ݐݎܽ݌ ݀݁ܿݎ݋ܨ‬െ ‫])݊݋݅ݐܿݑ݀݁ݎ ݕݐ݅ܿܽ݌ܽܿ ݂݋ ݁ݖ݅ݏ ݐ݅݊ݑ‬ ‫= ܴܱܨܧ‬

ிைுାாிைு ௌுାிைு

(6-14) (6-15)

where (FOH) are those hours when the unit is completely unavailable. In the literature, multistate unit models are also used instead of a two-state model. In the case of generating units with relatively long operating cycles such as base-load plants, the EFOR is an adequate estimator of the probability that the unit under similar conditions will not be available for service in the future. However, if there are fewer operating hours and many more start-ups and shut-downs with chances of start-up failures, a four-state model is proposed (see, for example (Billinton and Allan 1996)) to estimate unit unavailability at some time in the future.

Evaluation of Loss-of-Load Expectation (LOLE) The loss-of-load expectation (LOLE) is the main calculation to estimate the capacity credits of generating units. Generally, there are two basic approaches that can be used to evaluate the LOLE of a system: these are the Monte Carlo Simulation and the Analytical Method (NERC 2018). A. Monte Carlo Simulation: The Monte Carlo Simulation is a probabilistic statistical technique to simulate the actual process and random behavior of the system – treated as a series of experiments or trials. Monte Carlo simulation approaches could be “non-sequential” and “sequential”. A non-sequential simulation process, also called the state sampling approach, takes only a snapshot of the system state at various times. A sequential Monte Carlo simulation moves through time chronologically or sequentially, recognizing the fact that the status of a generating unit is not independent of its status in adjacent hours. In both “non-sequential” and “sequential” Monte Carlo simulations, the sequence of calculations is repeated by selecting new random numbers. The simulations are repeated several thousand times. Results from all the trials are used to construct probabilistic representations to achieve an acceptable level of statistical convergence. The degree of statistical convergence of a reliability index is measured by the standard deviation of the estimate of the reliability. Annual indices covering the period of interest are calculated as the average of the

Chapter Six

164

accumulated trial data until the variance is equal to or smaller than the selected convergence criteria. A Monte-Carlo based approach is based on the mathematical process of random sampling from the generating units’ availability and demand distributions and repeating the process to determine how many times there is a loss of load. The following steps can be used to estimate LOLE (NERC 2018): 1.

Assume Gjk is the available capacity of the jth generating unit in the kth trial, and m is the number of generating units in the system; then ௠

ܵ‫ = ݕݐ݅ܿܽ݌ܽܥ ݈ܾ݈݁ܽ݅ܽݒܣ ݉݁ݐݏݕ‬෍ ‫ܩ‬௝௞ ௝ୀଵ

2.

(6-16)

Assume Li is the load in the hour i and ENSki is the energy not served in the kth iteration in the hour i, then ௠

‫ܵܰܧ‬௞௜ = ݉ܽ‫ ݔ‬ቐ0, ‫ܮ‬௜ െ ෍ ‫ܩ‬௝௞ ௝ୀଵ

(6-17)

If Li is less than the system’s available capacity, the above equation will equal 0. 3.

If Li is greater than the system’s available capacity set Iki = 1, otherwise 0; where Iki is a Boolean variable representing whether there is energy not served in the hour i, in the kth iteration using the following definition: ‫ܫ‬௞௜ = max ൜

4.

0, ݂݅ ‫ܵܰܧ‬௞௜ = 0 1, ݂݅ ‫ܵܰܧ‬௞௜ ് 0

Assume N is the number of trials; then LOLPi is the count of the times that load is greater than generation availability divided by the number of trials; ே

‫ܲܮܱܮ‬௜ =

1 ෍ ‫ܫ‬௞௜ ܰ ௞ୀଵ

Capacity Value of Photovoltaics for Estimating the Adequacy of a Power 165 Generation System

5.

Loss-of-load expectation (LOLE) is the summation of loss-of-load probability in each hour i over the total period, T. If the period is for one year, the T is 8760 hours, and the unit of LOLE is in hours/year. ்

‫ = ܧܮܱܮ‬෍ ‫ܲܮܱܮ‬௜ (݄‫ݏݎ‬/‫)ݎܽ݁ݕ‬ ௜

If LOLE analysis includes the intertie transmission limits between the subareas, the problem becomes multidimensional and is modeled as a probabilistic flow network. Monte Carlo simulations are more suitable to handle such complex problem. However, as there are different answers for each simulation many runs are necessary to generate reliable statistics (about 5000 or more) which makes it very computationally expensive. B. Analytical Method: The computation time using the analytical method is much faster than the Monte Carlo approach. The analytical method for the calculation of the LOLE index requires the following three steps (NERC 2018): 1. A load model that describes the expected system load with an uncertainty representation to capture the variation of the demand associated with the weather and other factors; 2. A capacity model that describes the random behavior of the capacity outages and the energy generation of the intermittent resources; and 3. A mathematical model to compute the reliability indices probabilistically associated with the combination of the load and the capacity models. LOLE calculation methodology is explained with the help of a small system of three generating units. The methodology is first explained with the enumeration method and then extended to the discrete numerical convolution method. C. Analytical Method (Enumeration Technique): The LOLE method employs the fundamental data of unit capacity, random outage rate, and load demand. Consider an example of a three-unit generating system of a total capacity of 350 MW, as shown in Table 6-1 (Malik and Albadi 2020).

Chapter Six

166

Table 6-1: Generating system of three units Units

Capacity (MW)

FOR

Availability

Unit A

50

0.05

0.95

Unit B

100

0.07

0.93

Unit C

200

0.10

0.98

System

350

For these three units, there are eight combinations (or events or states) of generating units for on outage and in service. It is assumed that events are independent, which means that the outage of one unit has no effect on the outage of another unit. Then using the basic probability theory of independent events, the probability of occurrence of each state or event can be estimated. Table 6-2 shows the outage states and the probability of each state. Table 6-2: Outage State Estimation On Outage

In-Service

Probability

None

MW on Outage 0

A, B, C

A

50

B, C

B

100

A, C

C

200

A, B

A, B

150

C

A, C

250

B

B, C

300

A

A, B, C

350

None

0.95×0.93×0.98 = 0.86583 0.05×0.93×0.98 = 0.04557 0.95×0.07×0.98 = 0.06517 0.95×0.93×0.02 = 0.01767 0.05×0.07×0.98 = 0.00343 0.05×0.93×0.02 = 0.00093 0.95×0.07×0.02 = 0.00133 0.05×0.07×0.02 = 0.00007 Total = 1.00000

Table 6-2 above can be ordered in a monotonically increasing order of megawatts of an outage as presented in Table 6-3 below:

Capacity Value of Photovoltaics for Estimating the Adequacy of a Power 167 Generation System

Table 6-3: Rearrangement of Table 6-2 in ascending order of MW on outage MW on Outage 0 50 100 150 200 250 300 350

MW in Service 350 300 250 200 150 100 50 0

Probability 0.86583 0.04557 0.06517 0.00343 0.01767 0.00093 0.00133 0.00007

Consider evaluating the probability of not being able to supply a 220 MW load demand. If a less than 220 MW capacity is in service, a 220 MW load cannot be served. Therefore, the load cannot be served if more than 130 MW (350-220) of capacity is on the outage. According to data from Table 6-3, the probability of more than 130 MW on outage is: 0.00343+0.01767+0.00093+0.00133+0.00007 = 0.02343 Hence, the probability of not meeting the load demand of 220 MW is 0.02343. Since the computation of the probability of not meeting load demand requires an evaluation of more than X MW on outage, where X = (Capacity í load), Table 6-3 can be written as a cumulative outage table, as presented in Table 6-4 below. Table 6-4: Capacity outage probability X MW or more on outage 0 50 100 150 200 250 300 350

Probability of X MW or more on outage 1.00000 0.13417 0.08860 0.02343 0.02000 0.00233 0.00140 0.00007

168

Chapter Six

In Table 6-4, the entry of 0 MW or more on outage is the summation of Table 6-3 from 0 to 350 MW. The entry of 50 MW or more is the summation of Table 6-3 from 50 to 350 MW and so on and so forth. Thus, to evaluate the probability of not meeting a 220 MW load, the required capacity on outage should be more than 130 MW on outage. From Table 6-4, the probability of not meeting load demand is 0.02343. The procedure described above requires the enumeration of 2N states, where N is the number of generating units in the system, and 2 represents the two possible states of each generating unit (either available or on FOR). For three units, the total states are eight states (23 = 8). If there are 30 units in the system, this would mean an enumeration of more than a billion capacity states. This is computationally impractical, so a convolution process is developed. D. Analytical Method (Convolution Technique): The technique described here is called discrete numerical convolution. There are other approximate methods to achieve the convolution process (NERC 2018). The objective is to construct a table of the probability of X MW or more on outage. Assume that the table has already been constructed for several generating units and that a new unit has to be added. Let: P(X) = Cumulative probability of X MW on outage for the existing table P/(X) = Cumulative probability of X MW on outage for the new table p = Availability of the new unit q = Forced outage rate C = Capacity of the unit. Consider that we have a cumulative probability of some value of X MW or more on outage before the unit is “added”. In order to have a cumulative probability of X MW or more on outage after the unit is “added”, the original system could have a cumulative probability of X MW or more on outage and the additional unit could have 0 MW on outage with some probability p. Or, the additional unit could have C MW on outage with some probability q and the original system could have (X-C) MW or more on outage. Table 6-5 illustrates both these scenarios.

Capacity Value of Photovoltaics for Estimating the Adequacy of a Power 169 Generation System

Table 6-5: Two events of unit addition and the corresponding cumulative probability of the original system Case 1 2

Original System MW or more Cumulative on Outage Probability X P(X) X-C P(X-C)

Unit to be added MW on Probability Outage 0 p C q

The cumulative probability of X MW capacity or more on outage after the unit’s outage capacity C has been added to it, is: ܲ/ (ܺ) = ‫ )ܺ(ܲ × ݌‬+ ‫ ܺ(ܲ × ݍ‬െ ‫)ܥ‬ Eq. 6-21 is the recursive convolution formula for the cumulative probability of X MW or more on outage. For initialization: ܲ(݊݁݃ܽ‫(ܲ = )݁ݒ݅ݐ‬0) = 1.0 The case for which Eq. (6-21) was derived was based on a two-state characterization of a generating unit; the unit of capacity C is available with probability p or unavailable (zero capacity) with probability q. The two-state characterization typically uses the equivalent random outage rate (see Eq. (6-15)). In some cases, when a high degree of accuracy is required, a multistate characterization of a generating unit is used. Due to their partial outages, derating states of generating units are taken care of by multistate representation. The outage state for multistate representation can be described as: Ci = MW out of service for capacity (derating) state i qi = MW out of service for capacity (derating) state i. Then Eq. (6-22) can be generalized to: ே /

ܲ (ܺ) = ෍ ‫ݍ‬௜ × ܲ(ܺ െ ‫ܥ‬௜ ) ௜ୀଵ

and for initialization the same Eq. (6-22) as before.

Chapter Six

170

Thus, the multistate representation is not conceptually different from the two-state equation. However, the computational time increases proportionally to the number of generating unit probability states used. The computational procedure for solving either Eq. (6-21) or (6-23) begins by establishing a probability megawatt step size. S The step size is the MW increment of the probability table. Typically, the step size is calculated as the total capacity of a power system divided by 1000 (or 5000 in very detailed cases). For example, a 5000-MW capacity power system would have a 5-MW step size. A 92-MW generating unit would be represented as a 90-MW unit with a 2-MW truncation error, and a 109-MW unit would be represented as a 110-MW unit with a 1-MW truncation error. Most computer algorithms ignore the truncation error because a sum approaches zero for a system with many units. The procedure begins by initializing the cumulative outage table for no generating units. The cumulative outage table is initialized to unity for 0 MW or more on outage and 0 for all other MW values or more on outage. Equation (6-21) or (6-23) is then applied repetitively for each generating unit. As an example of the application of the recursive convolution formula for the cumulative capacity outage probability table, consider the system with three generating units described earlier. The step size chosen is 50 MW as it is divisible by all three units without truncation error. Table 6-6 shows the procedure. Units are convolved in one by one by applying Eq. (6-21) recursively. The result of the last column when all three units are convolved in is called the Capacity Outage Probability Table (COPT). This COPT is the same as in Table 6-4. Table 6-6. Capacity outage probability table using the convolution algorithm MW or more on outage

Initialize Table

Unit A convolved (50 MW, 5% FOR)

Unit B convolved (100 MW, 7% FOR)

Final results Unit C convolved (200 MW, 2% FOR)

0

1.0

50

0

100

0

1×0.95+1×0.05=1 .0 0×0.95+1×0.05=0 .05 0×0.95+0×0.05=0

150

0

1×0.93+1×0.07 =1.0 0.05×0.93+1×0.0 7=0.1165 0×0.93+1×0.07 =0.07 0×0.93+0.05×0.0 7=0.0035

1×0.98+1×0.02 =1.00000 0.1165×0.98+1×0 .02=0.13417 0.07×0.98+1×0.0 2 =0.08860 0.0035×0.98+1×0 .02=0.02343

Capacity Value of Photovoltaics for Estimating the Adequacy of a Power 171 Generation System 200

0

250

0

300

0

350

0

400

0

0×0.93+0.05×0 =0

0×0.98+1×0.02 =0.02000 0×0.98+0.1165×0 .02=0.00233 0×0.98+0.07×0.0 2 =0.00140 0×0.98+0.0035×0 .02=0.00007 0×0.98+0×0.02 =0

E. LOLE Calculation: The COPT made using the convolution algorithm is applied in conjunction with the hourly load time series to compute the hourly LOLH/LOLE. To illustrate the concept, consider a day with hourly loads served by the three-unit generating system discussed above. The hourly loads are indicated in columns 1 and 2 of Table 6-7 and shown in Fig. 6-4 (Malik and Albadi 2020). Column 3 of Table 6-7 is obtained by subtracting the hourly load from the total generating system capacity of 350 MW. Using the results from Table 6-6, the probability of not meeting the load each hour (LOLP of each hour) can be computed as presented in column 4 of Table 6-7. It may be noted that the LOLP reading may not be the same as the probability of X MW or more on outage of Table 6-6. See, for example, the 150 MW load of the first hour that can be served if 200 MW capacity is on outage but not if more than 200 MW capacity is on outage; therefore, instead of 0.02 probability, the next lower level outage probability of 0.00233 is used in Table 6-7. Table 6-7: Hourly index of LOLP and LOLE Hour

Load (MW)

0-1 1-2 2-3 3-4 4-5 5-6 6-7 7-8 8-9 9-10 10-11 11-12

150 150 150 160 160 180 200 230 230 250 250 260

Capacity more than X MW on outage (MW) 200 200 200 190 190 170 150 120 120 100 100 90

Prob. of not meeting load (LOLP) 0.00233 0.00233 0.00233 0.02000 0.02000 0.02000 0.02000 0.02343 0.02343 0.02343 0.02343 0.08860

Chapter Six

172 270 280 280 270 260 250 250 260 260 240 200 160

12-13 13-14 14-15 15-16 16-17 17-18 18-19 19-20 20-21 21-22 22-23 23-24

80 70 70 80 90 100 100 90 90 110 150 190 LOLE (hrs/day)

0.08860 0.08860 0.08860 0.08860 0.08860 0.02343 0.02343 0.08860 0.08860 0.02343 0.02000 0.02000 0.99980

300 250

Load (MW)

200 150 100 50 0 0

4

8

12

Hour

16

20

24

Fig. 6-4: Hourly demand of a three-unit generating system

Summing the LOLP for each hour yields a daily number of 0.99980 hour/day, which is roughly about 1 hour/day. This is interpreted as an expected value. Although the daily index is used as an example, the utility industry generally uses an annual index (LOLE), which is the summation of daily probabilities, or often termed “daily risks” over the entire year. Typical values used in developed countries are between 0.1 to 1.0 day/year depending on the required reliability of service.

Capacity Credit Versus Capacity Factor The capacity credit of a generating unit is often confused with its capacity factor. The capacity factor of a generating unit is the ratio of the actual

Capacity Value of Photovoltaics for Estimating the Adequacy of a Power 173 Generation System

electrical energy output of that unit over a given period to the maximum possible electrical energy output over that period. Mathematically, ‫)ܨܥ( ݎ݋ݐܿܽܨ ݕݐ݅ܿܽ݌ܽܥ‬ ‫݀݋݅ݎ݁݌ ܽ ݊݅ ݀݁ݐܽݎ݁݊݁݃ )݄ܹܯ( ݕ݃ݎ݁݊ܧ‬ = ܴܽ‫݀݋݅ݎ݁݌ ݄݁ݐ ݊݅ ݏݎ݄ ݂݋ ݎܾ݁݉ݑ݊ × )ܹܯ( ݕݐ݅ܿܽ݌ܽܥ ݀݁ݐ‬ × 100%

(6-24)

Usually, the period considered is one year, but the capacity factor of generating units can be calculated on a monthly or seasonal basis as well. Although the capacity factor of a unit is also affected by its unavailability due to its planned or forced outage it is highly dependent on whether it is needed or not in the system to generate due to variability of the demand. Therefore, the capacity factor of base-load units, which are cheaper to operate, is much higher than for peaking units that may be required only a few hours in a week. Capacity Credit (or value), on the other hand, is a term used in power system reliability and planning and is the proportion of rated capacity that is considered “firm” when planning reserves and margins to meet peak demands. In simple terms, it is the portion of the rated capacity that can be considered guaranteed to be there should it be needed. Capacity credit, therefore, is related to the reliability of the unit. It may be noted that the unit may be available for generation but cannot produce electrical energy, as the energy resource is not there.

Evaluation of Capacity Credits Methods for evaluating the capacity credit of renewable resources can be classified according to whether they are based on the reliability index or on approximation techniques. The reliability-based methods include equivalent conventional power (ECP), effective load carrying capability (ELCC), and equivalent firm capacity (EFC), whereas approximation methods include Garver’s ELCC approximation, the Z method, and capacity factor-based methods (Madaeni, Sioshansi, and Denholm 2012). The reliability-based methods use power system reliability evaluation techniques based on Loss-of-Load Probability (LOLP) and Loss-of-Load Expectation (LOLE). Essentially, the three reliability-based methods are very similar and differ only on how the capacity credit of a PV system is measured. If the capacity credit of PV is measured against a conventional generating unit that can be replaced while maintaining the same system

Chapter Six

174

reliability, it is called ECP. If the capacity credit of PV is measured against the amount by which the system’s loads can increase (when the PV generator is added to the system) while maintaining the same system reliability as before, it is called ELCC. If the capacity credit of PV is measured against a fully reliable generating unit (i.e., a unit with an EFOR of 0%) that can be replaced while maintaining the same system reliability level it is called EFC. We will discuss the ELCC method in detail.

A. ELCC Method of Capacity Credit: The three steps used to calculate the ELCC of a PV plant are as follows: 1.

For a given set of conventional generating units, the LOLEBase of the system without the PV plant is calculated using equation (6-3) and repeated here. ்

‫ܧܮܱܮ‬஻௔௦௘ = ෍ ܲ(‫ܩ‬௜ < ‫ܮ‬௜ ) ௜ୀଵ

2.

The time series for the PV plant power output is treated as a negative load and combined with the load time series, resulting in a load time series net of PV power. LOLEPV, is then calculated as: ்

‫ܧܮܱܮ‬௉௏ = ෍ ܲ(‫ܩ‬௜ < ‫ܮ‬௜ െ ‫ܥ‬௜ ) ௜ୀଵ

where Ci denotes the output of the PV plant in the hour i. LOLEPV will now be lower (and therefore better) than the target LOLE as obtained in the first step. 3.

The load data are then increased by a constant load ¨L across all hours using an iterative process, and the LOLE¨L is recalculated at each step until the target LOLEBase is reached. LOLE¨L is: ்

‫ܧܮܱܮ‬ο௅ = ෍ ܲ(‫ܩ‬௜ < ‫ܮ‬௜ െ ‫ܥ‬௜ + ο‫)ܮ‬ ௜ୀଵ

‫ܧܮܱܮ‬ο௅ ൎ ‫ܧܮܱܮ‬஻௔௦௘

Capacity Value of Photovoltaics for Estimating the Adequacy of a Power 175 Generation System

The increase in load ¨L that achieves the reliability target is the ELCC or capacity value of the PV plant. ‫ܥܥܮܧ‬௉௏ = ο‫ܮ‬

B. Capacity Credit Numerical Example: Once again, we refer back to our three-unit generating system of Table 6-1 and the same hourly loads shown in Fig. 6-4 (Malik and Albadi 2020). We see the calculation for the three steps discussed earlier. 1.

For this system, step 1 of ELCC, i.e., LOLEBase is already computed to be about 1 hour/day (see Table 6-7).

2.

Let us suppose that we want to add a PV power plant, which can produce the maximum output power of 22 MW under the site condition. The hourly load and the output of the PV plant are given in columns 2 and 3 of Table 6-8, respectively. As can be observed from column 3, the output of the PV plant is zero megawatts at night. If we add all the hourly PV outputs of column 3, this gives 185 MWh of energy in 24 hours. Therefore, by applying equation (6-24), the daily capacity factor of this power plant is 35.04%. Column 4 shows the net load after subtracting the PV output power (also shown in Fig. 6-5). Column 5 shows the capacity of more than X MW on outage for each hour, and column 6 shows the corresponding LOLP. The summation of LOLP for each hour shows the LOLEPV after the PV plant is “added.” It may be noted that the new LOLEPV is lower than the LOLEBase, which should be the case.

3.

Step 3 is to increase the PV modified load with some constant load ¨L and recalculate LOLE¨L. This is an iterative process until LOLE¨L becomes equal to or close to LOLEBase. When ¨L is made equal to 10 MW, LOLE10MW = 0.99450 hrs/day and when ¨L = 11MW, LOLE11MW = 1.19001 hrs/day. This is shown in Table 6-9. In fact, a slight increase in ¨L from 10 MW will make LOLE10MW+ = 1.19001 hrs/day. This is because the system is very small, with only eight discrete values of LOLP. For a larger system, the accuracy will be much higher. Therefore, we can conclude that the ELCC or capacity credit of 22 MW of PV is about 10 MW, which is 45.45%.

Chapter Six

176

Table 6-8: Hourly index of LOLP and LOLE after a PV plant of 22 MW is added Hour

0-1 1-2 2-3 3-4 4-5 5-6 6-7 7-8 8-9 9-10 10-11 11-12 12-13 13-14 14-15 15-16 16-17 17-18 18-19 19-20 20-21 21-22 22-23 23-24 PV Energy =

Load (MW)

PV Output (MW)

150 150 150 160 160 180 200 230 230 250 250 260 270 280 280 270 260 250 250 260 260 240 200 160 185 MWh

0 0 0 0 0 0 0 0 5 10 20 22 22 22 22 22 20 15 5 0 0 0 0 0

New Load (Net of PV Output) (MW) 150 150 150 160 160 180 200 230 225 240 230 238 248 258 258 248 240 235 245 260 260 240 200 160

Capacity Prob. of more not than X meeting MW on load outage (LOLP) (MW) 200 0.00233 200 0.00233 200 0.00233 190 0.02000 190 0.02000 170 0.02000 150 0.02000 120 0.02343 125 0.02343 110 0.02343 120 0.02343 112 0.02343 102 0.02343 92 0.08860 92 0.08860 102 0.02343 110 0.02343 115 0.02343 105 0.02343 90 0.08860 90 0.08860 110 0.02343 150 0.02000 190 0.02000 LOLEPV 0.73912 (hrs/day)

Capacity Value of Photovoltaics for Estimating the Adequacy of a Power 177 Generation System

300 Load (MW)

250 200 150

Original Load

PV Modified Load

100 50 0 0

4

8

12 Hour

16

20

24

Fig. 6-5: Hourly base load and net of PV demand of a three-unit generating system

Table 6-9: Hourly index of LOLP and LOLE after adding ¨L in the PV modified load Hour

0-1 1-2 2-3 3-4 4-5 5-6 6-7 7-8 8-9 9-10 10-11 11-12 12-13 13-14

New Load (Net of PV Output) (MW) 150 150 150 160 160 180 200 230 225 240 230 238 248 258

¨L=10 MW

160 160 160 170 170 190 210 240 235 250 240 248 258 268

Prob. of not meeting load (LOLP10 MW) 0.02000 0.02000 0.02000 0.02000 0.02000 0.02000 0.02343 0.02343 0.02343 0.02343 0.02343 0.02343 0.08860 0.08860

¨L=11 MW

161 161 161 171 171 191 211 241 236 251 241 249 259 269

Prob. of not meeting load (LOLP11 MW) 0.02000 0.02000 0.02000 0.02000 0.02000 0.02000 0.02343 0.02343 0.02343 0.08860 0.02343 0.02343 0.08860 0.08860

Chapter Six

178

14-15 15-16 16-17 17-18 18-19 19-20 20-21 21-22 22-23 23-24

258 248 240 235 245 260 260 240 200 160 LOLE¨L (hrs/day)

268 258 250 245 255 270 270 250 210 170

0.08860 0.08860 0.02343 0.02343 0.08860 0.08860 0.08860 0.02343 0.02343 0.02000 0.99450

269 259 251 246 256 271 271 251 211 171

0.08860 0.08860 0.08860 0.02343 0.08860 0.08860 0.08860 0.08860 0.02343 0.02000 1.19001

C. Effect on Capacity Credit with Increased PV Penetration: The incremental value of capacity of PV plants starts decreasing as the PV penetration level is increased in a power system. To demonstrate the effect, we refer back to our three-unit generating system and the same hourly loads shown in Fig. 6-4. We assume that we have ten PV plants of 22 MW each, and each one has the same hourly generation capability, as shown in column 3 of Table 6-8. Table 6-10 shows the effect of increasing the number of PV plants on capacity credits. Columns 1 and 2 of Table 6-10 are the number of PV plants and the corresponding PV capacity in the system. Column 3 is the LOLE of the system after “adding” the PV plants or in other words subtracting the PV generation from the baseload. Note that without any PV plant in the system, the base LOLE is 0.99980 hour/day or approximately 1 hour/day. We assume that 1 hour/day is our target LOLE on which the capacity credit of PV plants is going to be measured. Columns 4 and 5 show the constant incremental load and its corresponding LOLE that take the system LOLE lower than 1 hour/day, and columns 6 and 7 show the constant incremental load and its corresponding LOLE that take the system LOLE higher than 1 hour/day. Column 8 shows the attributed capacity credit by taking that constant incremental load that meets the reliability criteria of 1 hour/day. Finally, column 9 shows the percentage capacity credits. It can be seen from the table that the capacity credit decreases when more PV plants are added. Figure 6-6 shows the system LOLE and the capacity credit versus the number of PV plants. It can be seen that the system LOLE decreases sharply initially when the number of PV plants is increased. However, with a further increase in the number of plants, the incremental decrease in LOLE is

Capacity Value of Photovoltaics for Estimating the Adequacy of a Power 179 Generation System

reduced or even diminishes. On the other hand, the total capacity credit from PV plants increases with the number of plants. Still, this increase is not uniform as the number of PV plants rose to a certain level; the further growth in the PV system’s capacity credit will not be there. It can be noticed from Table 6-10 and Fig. 6-6 that when the numbers of PV plants are 6, 7, and 8, the capacity credit remains the same at 50 MW, then when the 9th plant is added, the capacity credit jumps to 60 MW. This is because for three conventional generating units, only eight discrete values of LOLP are used; hence, when calculating the ELCC, a sudden jump in ¨L occurs that satisfies equation (6-28). If the numbers of conventional units are large, the number of LOLP states will increase significantly and give a smooth curve for capacity credits. When bar charts are smoothed using regression, this provides the best fit of the trend line with high R2. From the trend line, it is clear that at a certain point, adding PV plants to the power system will not bring any incremental capacity credit. Fig. 6-7 shows the reduction in the PV system’s capacity credit with an increase in the number of plants. Table 6-10: Capacity credit vs. PV penetration level No. of PV plants (col 1)

PV Capacity (MW) (col 2)

LOLE (hour/day) (col 3)

¨L1 (MW) (col 4)

LOLE¨L1 (hour/day) (col 4)

¨L2 (MW) (col 5)

LOLE¨L2 (hour/day) (col 6)

Capacity Credit (MW) (col 7)

Capacity Credit (%) (col 9)

1 2 3 4 5 6 7 8 9 10

22 44 66 88 110 132 154 176 198 220

0.73912 0.66179 0.59849 0.58134 0.54257 0.45079 0.43312 0.42847 0.42661 0.39926

10 20 36 40 45 50 50 50 60 60

0.99450 0.92933 0.99793 0.86759 0.96216 0.94501 0.94158 0.85323 0.95852 0.95759

11 21 37 41 46 51 51 51 61 61

1.19001 1.06310 1.12827 1.09593 1.02733 1.15081 1.08221 1.01153 1.00409 1.00409

10 20 36 40 45 50 50 50 60 60

45.5% 45.5% 54.5% 45.5% 40.9% 37.9% 32.5% 28.4% 30.3% 27.3%

Chapter Six

0.8

Capacity Credit

LOLE

Poly. (Capacity Credit)

Poly. (LOLE)

LOLE (hours/day)

0.7

70

R² = 0.9584

60

0.6

50 R² = 0.9778

0.5

40

0.4 30

0.3

20

0.2

10

0.1 0

0 1

2

3

4

5

6

7

8

9

10

Number of PV plants Fig. 6-6: LOLE and capacity credit versus number of PV plants Capacity Credit (%)

Expon. (Capacity Credit (%))

Capacity Credit (%)

60.0% 50.0% R² = 0.7705

40.0% 30.0% 20.0% 10.0% 0.0% 1

2

3

4

5

6

7

8

Number of PV plants Fig. 6-7: Percentage capacity credit versus the number of PV plants

9

10

Capacity Credit (MW)

180

Capacity Value of Photovoltaics for Estimating the Adequacy of a Power 181 Generation System

Conclusions This chapter addressed the evaluation of the capacity credit of PV plants. The chapter has given an overview of the various adequacy measures used to evaluate generation system reliability. The Monte Carlo simulation and analytical methods of estimation of the two most important reliability indices of LOLP and LOLE are discussed. The analytical method of convolution is discussed in more detail using a three-unit generating system to construct a capacity outage probability table (COPT). The COPT is then employed with the hourly loads of one day to show how hourly LOLPs are calculated and their summation to estimate the LOLE. These indices are the basis for calculating the capacity credit of any generating unit. There are several methods to estimate capacity credit, and the most common and accurate method of effective load carrying capability (ELCC) is discussed. The three-step procedure for calculating the capacity credit for the PV system using the ELCC technique is implemented through a small system of three conventional generating units and a PV plant. It has been shown that a PV plant of 22 MW in a 350 MW system of three generating units has a capacity value of about 10 MW, and the percentage capacity credit is 45.45%. The capacity factor, on the other hand, is 35.04%. The effect on capacity credit due to an increase in the PV penetration level in the generation system is also discussed with the same example. It has been shown that with the rise of PV penetration in the generation system, the incremental value of PV capacity diminishes. Although a lot of literature already exists on capacity credit, the way the material is presented here, using a small computation example has a tutorial value and can serve as a stand-alone lesson on capacity credit that is appropriate for undergraduate students and useful for postgraduate students who want to pursue research in this area.

References Billinton, R., and R. Allan. 1996. Reliability Evaluation of Power Systems. Springer Science+ Business Media, LLC. Billinton, Roy, and Ronald Norman Allan. 1992. Reliability evaluation of engineering systems. Springer. Buehring, W., C. Huber, and J. De Souza. 1984. "Expansion Planning for Electrical Generating Systems – A Guidebook." In Technical Report Series No. 241. Vienna: IAEA. Keane, Andrew, Michael Milligan, Chris J. Dent, Bernhard Hasche, Claudine D'Annunzio, Ken Dragoon, Hannele Holttinen, Nader

182

Chapter Six

Samaan, Lennart Soder, and Mark O'Malley. 2011. "Capacity value of wind power." IEEE Transactions on Power Systems 26(2), 564-572. Khallat, M. A., and Saifur Rahman. 1988. "A model for capacity credit evaluation of grid-connected photovoltaic systems with fuel cell support." IEEE Transactions on Power Systems 3(3), 1270-1276. Madaeni, Seyed Hossein, Ramteen Sioshansi, and Paul Denholm. 2012. "Comparison of capacity value methods for Photovoltaics in the Western United States." National Renewable Energy Lab. (NREL), Golden, CO (United States). Malik, Arif S., and Mohammed H. Albadi. 2020. "A Tutorial for Evaluating Capacity Credit of PV Plants Based on Effective Load Carrying Capability." 2020 5th International Conference on Renewable Energies for Developing Countries (REDEC). NERC. 2018. "Probabilistic Adequacy and Measures." Atlanta, GA, USA: North American Electric Reliability Corporation (NERC). Simoglou, Christos K., Pandelis N. Biskas, Emmanouil A. Bakirtzis, Anneta N. Matenli, Athanasios I. Petridis, and Anastasios G. Bakirtzis. 2013. "Evaluation of the capacity credit of RES: The Greek case." PowerTech (POWERTECH) Grenoble: IEEE.

CHAPTER SEVEN IMPACTS OF ENERGY SYSTEMS’ OUTPUT VARIABILITY ON TRANSMISSION AND DISTRIBUTION NETWORKS HUSSEIN A. KAZEM1

Abstract The integrity of renewable energy resources with a conventional grid/network/utility for distributed generation require the locating of a small-scale power generation technology close to the served loads. The advance toward on-site distributed generation has been quickened due to the deregulation and upgrading of the utility business and the possibility of alternative energy sources. Distributed generation technologies can boost system reliability, improve power quality, defray utility capital investment, and reduce energy costs. There are four significant issues identified with distributed generation that are covered in this chapter: equipment and control, the impact on transmission and distribution systems, the interconnection standard, and financial assessment. The anti-islanding requirements of grid-connected photovoltaic PV systems will be discussed as well. The impact of the output variability of renewable energy technologies, particularly PV systems, on the distribution network will be covered in this chapter. Keywords: integrating of renewable energy technology; distribution networks; power flow; interconnection.

1

Sohar University, Suhar, Oman, Email: [email protected]

184

Chapter Seven

Principles of power injection While utilizing conversion technology for electrical energy to renewable and sustainable energy systems, two different types of systems must be considered: rotatory and stationary. Alternating current is mainly provided from the rotatory type. Synchronous and induction generators are the primary drivers for gas and diesel turbines, wind, and hydropower energy sources. Direct current generators are the rotatory type; however, they are not typically utilized on the grounds of their staggering expense, cumbersome size, and maintenance needs. Also, some renewable energy technologies produce DC power such as photovoltaics. Direct current is usually provided by the stationary type. PV systems are the principle sustainable power sources to be discussed in this section. This section will discuss energy converting technology, power electronic converters and power flow.

Energy converting technology Throughout the years, non-renewable sources of energy have been, and remain, the world’s number 1 source of energy. According to a BP review (Dudley 2015), fossil fuels satisfy about 86% of the world’s energy consumption as of 2014, with nuclear energy contributing 4.4%, leaving the shares of hydropower and other renewables at about 6.8% and 2.5%, respectively, i.e., less than 10% of the total world’s energy requirements are satisfied by renewable sources. Comparing these statistics to Fig. 7-1 (Perez, Richard 2009) paints a clear picture.

Impacts of Energy Systems’ Output Variability on Transmission and Distribution Networks

185

Fig. 7-1: Comparing renewable and non-renewable energy sources

In light of the way that renewable and sustainable power sources can have an important contribution in providing electrical power and energy while saving the earth, and reduce pollution, researchers, analysts and businesses centered their examinations on the hybrid power plants by utilizing renewable power sources (i.e., solar, wind, hydro, etc.) with non-renewable sources (i.e., oil, gas, coal, etc.) (Jagoda, Kalinga, Robert Lonseth, Adam Lonseth 2011) (Owen 2006). Recently, in many countries the hybrid energy system, either grid-connected or off grid-connected, has become an attractive option for many utilities, customers and applications due to many reasons, for example oil price fluctuations, cost reductions and improvements in renewable energy technologies, etc. (H.A. Kazem, Khatib, and Sopian 2013) (Farret 2006). Fig. 7-2 illustrates different types of energy sources connected to hybrid systems. Sometimes storage is needed and at other times storage is not required (ElNozahy, M. S. 2013). However, some of these sources are renewable and some are nonrenewable. Also, some generate DC power, for example solar photovoltaics, and others generate AC power, for example wind turbines. Since, the utilities use AC power, the DC generated ones need to be converted into AC power (Miqdam Tariq Chaichan and Kazem 2018). This could be done through a DC-AC converter/inverter, which is simply a

186

Chapter Seven

circuit converting DC into AC power with a specific voltage type (i.e., single or three-phase) and frequency. Also, it is worth mentioning that the inverter is categorized mainly into a pulse width modulation inverter, a square wave inverter and a single-phase inverter with voltage cancellation.

Fig. 7-2: Principle of alternative energy conversion technology and power injected to the grid

Power electronic converters Power electronics had many definitions, but simply it represents the control of power using semiconductor electronic devices. The subject is multidisciplinary as from one side it contains power, electronics and control and from the other side it contains high, medium, and low power conversion (Kassakian, John G., Martin F. Schlecht 2000). The power electronic converters are mainly four electrical circuits: AC-AC, AC-DC, DC-DC, and DC-AC converters, which are the AC voltage controller, rectifier, DC chopper and inverter, respectively (Rashid 1993). Fig. 7-3 shows different types of power electronic converters.

Impacts of Energy Systems’ Output Variability on Transmission and Distribution Networks

187

Fig. 7-3 Power electronics converters.

The AC voltage controller is used to control AC voltage according to the load needs. This converter has many applications, for example, control of three-phase induction motors, industrial heating, lighting, etc. Fig. 7-4a shows AC controllers with different arrangements based on the number of phases, waveshape, etc. (Lander 1987). The rectifier converter is used to convert AC voltage into DC. It is one of the most important converters in industry, used to control DC loads like DC motors. The rectifier is also used with the inverter in Variable Frequency Drive (VFD) to control AC loads like induction motors. Fig. 74b shows rectifiers with different arrangements based on the number of phases, waveshape, etc. Rectifiers are classified according to phases into single-phase and three-phase, and according to the used semiconductor devices into controlled, uncontrolled, and half controlled rectifiers.

Chapter Seven

188

AC Voltage controller Circuits

Single Phase

Bidirectional

R Load RL Load

Rectifier Circuits

Single Phase

Three Phase

Unidirectional Delta

Star

Full wave

Three Phase

Half wave

Delta

Star

Uncontrolled Uncontrolled

R Load RL Load

Unidirectional Bidirectional R Load

controlled

Full wave

Half controlled R Load

RL Loa

(a)

Half wave

controlled

R Load R Load

RL Load

RL Load

RL Load

(b)

Fig. 7-4 AC controller and rectifier converters.

The DC chopper converts DC voltage to another DC voltage to control the load as shown in Fig. 7-5a. It is a simple circuit converter used for controlled (i.e., transistor, thyristor, etc.) and uncontrolled (diode) semiconductor devices to chop the voltages. There are different types of choppers (Buck, Boost, Buck-Boost, Cuk, full-bridge, etc.) (Bradley 1995). Inverter circuits are used to convert the DC voltage into AC voltage with the desired voltage magnitude and frequency as shown in Fig. 7-5b.

Impacts of Energy Systems’ Output Variability on Transmission and Distribution Networks

189

(a)

(b) Fig. 7-5 DC chopper and inverter circuits.

These circuits are used to convert the generated power from different energy technologies before injection to the grid.

Power Flow The electrical power system generates, transmits and distributes the electrical power P and Q to the customers with a certain quality and standard. However, four factors are used to evaluate the system quality: constant voltage and frequency, reliability, and purity of the sinusoidal waveforms. The variation of input energy (i.e., fuel, solar irradiation, wind speed, etc.) and load is reflected in the power flow P and Q. The active power P is related to the power angle and frequency while, the reactive power Q is

190

Chapter Seven

related to voltage. Fig. 7-6 illustrates the principles of power flow control. Fig. 7-6 shows the power electronic converters used to connect renewable energy systems and technology for energy conversion and injection to the grid. The block diagram shows the voltage source inverter connected to the

Fig. 7-6 Power injected to the grid from renewable energy sources.

grid. The current is injected to the grid through inductance between the inverter output voltage V1 and the grid connected voltage Vs. A feedback of the phase currents and voltages is taken as negative feedback to be compared with the desired values to control the active and reactive power flow. The calculated error (comparison results) is used to control the pulse width modulation (PWM), which controls the inverter semiconductor

Impacts of Energy Systems’ Output Variability on Transmission and Distribution Networks

191

switches. The operation of a three-phase PWM converter depends on the PWM scheme used. Different PWM strategies, such as sinusoidal PWM (SPWM) and various space-vector modulation (SVM) schemes can be used in rectifier circuits. However, the difference between these schemes is primarily the choice of zero vectors. The simplest PWM scheme is used, where three sine waves modulating signals are compared with a triangular carrier waveform. Assume Vr1 is equal to Vs1. In this case the line-to-line voltage between Vs1 and Vc1 will lead to the voltage Vs1. However, when Vc1 leads Vs1 by 90o, the current I1 will be in phase with Vs1. This will make the active power P flow from the supply to the load. In another case, when Vs1 and Vc1 are in phase, the current I1 will lag the voltage Vs1 by 90o and this will cause a reactive power flow.

Power injection from grid-connected Photovoltaic systems The environmentally-friendly and clean technologies, particularly solar energy, are being used as an appropriate solution to overcome environmental difficulties and to address stable energy saving crises (Tahri, Tahri, and Oozeki 2018). In the last two decades, solar and wind power have begun to take their place in many countries of the world in the processing of electric power. In the last decade, solar energy has begun to compete for a convenient location as a source of clean energy and a pillar of sustainable development in many countries of the world (Demirbas 2016). Solar energy is available all over the world for most of the year, and it is free, clean, and a hundred-fold more accessible than the world needs (Humada et al. 2018). The use of solar energy has been promoted in various applications (Amin, Lung, and Sopian 2009), including the thermal heating of water for domestic and industrial purposes (Ahmad Fudholi et al. 2014), the heating of air for comfortable conditions (A. Fudholi et al. 2014), ventilation using the Trombe wall (Wang, Dengjia, Liang Hu, Hu Du, Yanfeng Liu, Jianxiang Huang, Yanchao Xu 2020) and solar distillation (A.H.A. Al-Waeli, Sopian, Chaichan, et al. 2017) (Miqdam T. Chaichan and Kazem 2015). In the production of electricity, there are several distinctive and suitable methods for the production of electric power such as the solar chimney (Kong, Jing, Jianlei Niu 2020), concentrated power stations (Keyif, Enes, Michael Hornung 2020), and photovoltaic cells (A. H. A. Al-Waeli, Sopian, Kazem, et al. 2017) (Hussein A. Kazem and Chaichan 2016) (H. A. Kazem et al. 2013). Today, photovoltaic cells have improved in terms of power generation and electrical efficiency. In modern products, this efficiency has been

192

Chapter Seven

improved to 40% by using new and innovative techniques (Green, Martin A., Ewan D. Dunlop, Dean H. Levi, Jochen HohlဨEbinger, Masahiro Yoshita 2019). Thus, the participation of these cells in the production of global power has increased (Bahattab et al. 2016). The great advances in photovoltaic technologies and the low cost of electricity production have achieved a price of less than 2 cents/kWh. As the progress of this technology continues, PV will become the cheapest source of electricity in various countries of the world. Currently, this technique (PV) is used to provide electricity for road lighting (A. H. A. Al-Waeli, Kazem, Sopian, et al. 2017), car parking lights (Ali H. A. Al-Waeli et al. 2015), communication towers (Chaichan, Miqdam T., Hussein A. Kazem, Ali A. Kazem, Khalil I. Abaas 2015), and watering pumps in remote areas away from the grid (A. H. A. Al-Waeli, El-Din, Al-Kabi, et al. 2017). Photovoltaic cells are flexible in their high composition, as they can be installed and erected near the seashores or in valleys, as well as in the high mountains (Hussein A. Kazem, Chaichan, and Yousif 2019) (Al Busaidi et al. 2016). Mostly, these cells need to be fully exposed to the sun's rays throughout the day to sunset. The installation of photovoltaic cells abroad is subject to different climatic conditions, which means that they are exposed to external weather factors such as solar radiation intensity, ambient air temperature, relative humidity, and wind (Ali H. A. Al-Waeli, Sopian, et al. 2019). Each of these factors has an effect that is different from other factors and the outcome of their different effects will determine the improvement or degradation of the solar cell’s performance in a given location. For example, the performance of photovoltaic cells is reduced by high solar radiation intensity because the bulk of it goes to heat the cell, which reduces the voltage difference produced and results in lower power generation (Ali H. A. Al-Waeli, Chaichan, Kazem, et al. 2019) (Ali H. A. Al-Waeli, Chaichan, Sopian, et al. 2019) (Ali H. A. Al-Waeli et al. 2018). According to (Khatib, Sopian, and Kazem 2013), increasing the temperature of photovoltaic cells by 1 degree reduces the photovoltaic capacity by 0.5-0.6%. Wind speed causes the natural cooling of cells resulting in improved performance (Xydis 2013). The accumulation of dust and dirt on the surface of the solar cell reduces the radiation reaching it, thereby reducing its productivity. The wind acts as a natural cleaner that removes part of the accumulated dust, which improves the cell's performance. However, high wind speeds cause the excitation and rise of sand and dust particles and transport them over long distances, helping them to accumulate on the cells. The dust suspended in the air reduces the energy produced as it disrupts the solar cells and reduces the incoming

Impacts of Energy Systems’ Output Variability on Transmission and Distribution Networks

193

radiation (Miqdam T. Chaichan, Abass, and Kazem 2018) (H. A. Kazem et al. 2017). The size and performance of any photovoltaic system therefore depend heavily on metrological data. In other words, the performance of any electrical system can be considered a variable depending on the location of the installation of the system, so it differs from one region to another depending on the nature of the climate of this region. The availability of good data of climate parameters such as solar radiation and ambient temperature is important to identify and predict the total energy available for use by a photovoltaic system (Yousif et al. 2019). Photovoltaic systems can be categorized in several ways, such as the type of material manufactured or connected to the network. In this latter class, photovoltaic cells are divided into two types of photovoltaic cells: standalone and grid connected as shown in Figs. 7-7a and 7-7b. For the photovoltaic systems connected to the grid, these can be classified into two groups: integrated PV systems (BiPV) and (DGPV) generation and distribution systems. BiPV systems provide a specific load for use (in homes or applications) and supply the excess power to the network. DGPV systems supply all the electrical power produced to the grid without the processing of a home or application. This type of station and networkrelated system can be solely considered as an energy source.

194

Chapter Seven

(a)

(b) Fig. 7-7: (a) Standalone, and (b) grid-connected PV systems

There can sometimes be a collaboration of power supplied from more than one energy source, such as a PV array and wind turbines, and/or with diesel engines or a storage unit (batteries) as shown in Fig. 7-8. Grid-connected PV systems can also be classified into two types: with or without storage for electricity; the first system has the potential to provide the critical load capacity when the network is at peak load or when the network power is interrupted. In both systems, reflectors are used to maintain the stability of the electricity transmitted through the network. These transformers are an important part that cannot be neglected when planning such systems. Most of the transformers operating in photovoltaic cells connected to the network are highly efficient. A deviation from the contact point between the utility and the PV generator may result in

Impacts of Energy Systems’ Output Variability on Transmission and Distribution Networks

195

significant effects as this deviation may cause deformation or instability of the ligaments and will sometimes result in fluctuations in the output of the photovoltaic system. The function of the inverter is not limited to switching from AC to DC only, it has a major role in synchronizing the electricity generated by the PV cells in frequency and phase with the current of the grid. The invertor also plays a major role in protecting the PV system and the network. The connection between the PV panels and the network is closed by using the fuse switches that cut the connection on both sides as the levels of instability increase in some of them. In the event of a frequent lack of electricity supply by the PV system, an independent hybrid system is then used to ensure proper load processing in the event of a shortage (Hussein A. Kazem and Khatib 2013) (Ibrahim et al. 2014) (Kaundinya, Balachandra, and Ravindranath 2009) (Verma, Midtgård, and Sætre 2011).

Fig. 7-8: Integrated PV/Wind/Diesel generator system

196

Chapter Seven

Switching photovoltaic systems or hybrid systems connected to the grid can have several benefits. In this case, the excess generated power is sold for the network, and if the system is unable to process the desired electrical load, the electricity supplied via the network is used to meet the load requirements. Sometimes the PV system is unable to meet the full load requirements, especially in the case of cooling needs for convenience purposes. In the summer, electricity supplied to the grid will only compensate for some electricity costs, thus reducing the cost of running the application. This helps to reduce electricity bills and avoid surcharges imposed by utility companies during peak demand hours. Figs. 7-9a and 79b show simple block schemes for GCPV systems with and without battery storage, respectively (Hussein A. Kazem et al. 2020).

Fig. 7-9: Block diagram of a) GCPV with a battery-bank, b) GCPV without a battery-bank

Many researchers have conducted huge research works to assess the performance of GCPV systems and determine the impact on an interconnected power system of variable factors such as general efficiency, life cycle costs, energy cost, and the energy recovery period. Eltawil and Zhao (2010) conducted a critical review study to demonstrate the importance of grid-related PV systems and potential problems at high load levels and ways to prevent the disruption of these photovoltaic systems. The study focused on ways to improve conversion efficiency and overall harmonic tension. This study was concerned with reflectors having a conversion efficiency exceeding 90% and also with maintaining the total harmonic deformation (THD) of harmonics less than 5%. The study concluded that the use of transformers in the power factor of the unit

Impacts of Energy Systems’ Output Variability on Transmission and Distribution Networks

197

should be given attention and focus and should be used to address such problems as excessive scaling in the PV generator for the inverters. In Turkey, Celik (2006) conducted an economic technical evaluation of a photovoltaic system connected to the network and having battery storage. The researchers presented an analysis of weather data for the area where the system was constructed. The authors used two different scalable methods; first, when electricity production is equal to demand, and second, to obtain equal electricity generation and demand in the month of the year with the highest demand. In this study, the electricity cost per kilowatthour of the GCPV system increased by more than 3 to 4 times the electricity supplied to the grid. The researchers attributed the result to the high level of taxes imposed by the government of the country on electricity. Sidrach-de-C. M. and López L. M. (1998) evaluated the values of the GCPV systems established in Spain, and the study used a 2.0 kW installed system. The researchers found that the average daily supply of the network was up to 7.4 kWh, and that the system can achieve between 4.1% and 8.0% of the average monthly value of electricity demand. The system used a single-phase reflector to connect the system with the network. The study concluded that the efficiency of the inverter is important in such systems to prevent the waste of energy. In the United Arab Emirates, Al-Sabounchi, Yalyali, and Al-Thani (2013) designed and evaluated the performance of the GCPV system in the country's hot weather conditions for most of the year in a study lasting one year. Because the studied system was established outdoors, the researchers showed that the dust deposited on the surface of the cells caused a significant reduction in their performance, and this factor should be considered as a concern in the design and implementation of photovoltaic cell systems in desertic areas. The researchers found that the worst performance degradation in the system was in July due to the impact of dust on the PV panels in conjunction with a high rise in air temperature and relative humidity. The researchers explained that the rate of recklessness in the electricity produced by the system reached about 27.0%. Sulaiman et al. (2012) developed an approach to the design of new systems of GCPV using different types of photovoltaic cells and reflectors. The design work was carried out according to European standard specifications and the aim was to achieve the production of electric power

198

Chapter Seven

with optimum conversion efficiency. Beser et al. (2010) provided a singlephase multi-level reflector for a grid-connected solar system. The proposed system was tested in practice to verify the validity of the design. The practical results showed a decrease in voltage and output current. Kim, Jeon, and Hong (2009) compared two types of GCPV systems taking into account the operation status and photoelectric performance. Both systems were installed in Daegu Metropolitan City. In this study, researchers were interested in the many air parameter effects such as the solar radiation intensity, ambient temperature, and tilt angle of the photovoltaic cells. The two systems were compared by calculating energy efficiency and the electric power generated on a monthly basis. The study proved that the efficiency of electricity generation for a full year was about 10.8% for the two systems studied. Also, that 80% of the costs incurred on the two systems came from the photovoltaic cells and inverters. In Kuwait, Al-Hasan, Ghoneim, and Abdullah (2004) conducted an assessment of a grid-connected solar system (GCPV) and continued the pattern of electric load with the use of weather data. The researchers concluded that GCPV systems caused the peak power load to drop and also reduced the monthly demand for electricity. Also, in Kuwait, AlOtaibi et al. (Al-Otaibi et al. 2015) evaluated the output of two PV systems connected to the grid. The two systems are composed of CIGS cells and equipped with automatic cleaning systems. These cleaning systems were used to reduce the accumulation of dust and pollutants on the performance of both systems. The researchers concluded that the photovoltaic systems studied had a performance rate of up to 70.0%. In Greece, Protogeropoulos et al. (C. Protogeropoulos, I. Klonaris, C. Petrocheilos, I. Charitos 2010) evaluated the performance of several different photovoltaic techniques in a networked project. The techniques in the system studied are monocrystalline/amorphous silicon, polycrystalline silicon, thin film CdTe and CIGS thin films. The researchers concluded that photovoltaic systems made up of monocrystalline/amorphous and polycrystalline cells connected to the grid would be more efficient than the remaining technologies. In Kosovo, Komoni et al. (2014) studied the performance of two photovoltaic cells connected to the grid. The first system consisted of monocrystalline cells and the second system of polycrystalline cells. The researchers concluded that the photovoltaic system that used polycrystalline cells generated more energy than the system that used monocrystalline cells. In India, Tripathi et al. (2014) analyzed the productivity of two photovoltaic

Impacts of Energy Systems’ Output Variability on Transmission and Distribution Networks

199

cells connected to the grid; the first operated with polycrystalline cells and the second with non-crystalline silicon cells (A-C). The results of the study showed that the photovoltaic system consisting of polycrystalline cells yielded a higher productivity than the amorphous silicon cell system. In Malaysia, Humada et al. (2016) [62] studied the electrical efficiency of two grid-connected photovoltaic systems in the tropical climatic conditions of Malaysia; one of the two systems consisted of monocrystalline cells and the other of copper-indium-delineated (CIS) cells. The researchers found that the second system generated more energy with a higher efficiency ratio than the monocrystalline cell system.

Instantaneous active and reactive power flow and control approach In the event that the correct measure of electrical power produced by the renewable and sustainable energy source is exchanged to the electrical grid/network, the power injected into the network and consumed by the load is produced by a similar source. However, if the Q is not supplied by the voltage source inverter and the AC load is observing the P, the “power factor PF” seen by the network may drop as far as possible to that permitted by the utility and perhaps constrain penalties and overcharges to the customer. In order to solve this issue, the voltage source inverter should supply P and, in the meantime, compensate the Q power and control PF inside the impediments. This section discussed the P and Q power flow and control approach. Continued operation at the nominal frequency is the best means of ensuring that the system’s real power balance is maintained, i.e., that the mechanical power input to the system generators matches the total electrical demand made up of the consumer load plus the transmission losses. “Automatic generation control (AGC)” is a technique used to control the power flow P and adjust the frequency with the change in the load (Elgerd 1982) (Alwaeli 2003). The purpose of AGC is to apply, as and when necessary, supplementary regulation to all available generators and energy sources so as to maintain a desired schedule based on optimum economy within the constraints of maintaining the necessary reserve of generation to be realized in an emergency. The AGC send signals to generator units under its control in realizing system generation changes. AGC design and performance are

Chapter Seven

200

related to the unit response and type of generating unit (renewable and non-renewable) (Elgerd, Olle I. 1970)(Chan, Wah-Chun 1981). Automatic generation control may be analyzed by implementing two types of techniques, continuous and discrete. In the past, AGC for an interconnected system used a continuous time optimization technique. However, the discrete time optimization technique has become the main AGC approach used today. Many studies (Bohn, E. V. 1972) (Tripathy, S. C., T. S. Bhatti, C. S. Jha, O. P. Malik 1984) (Kothari, M. L., J. Nanda, D. P. Kothari 1989) have been carried on the AGC using the discrete time technique implemented on simplified models, i.e., single area or two areas. However, due to the use of different energy sources in the same network and the interconnection of different countries’ power systems, this technique became more complicated but more effective, fast and efficient. Consider the interconnected power system shown in Figure 7-10. This system contains two areas (two thermal generators). Each of these generators represented by transfer functions contains a steam turbine, governor, etc. Equation 7.1 illustrates the state variable form as follows:

 = A X + B U + E ' PD X

(7.1)

Where U, X and ' PD are the input, state, and disturbance vectors, where for area 1 the state vector is: [X]T = [ ' F1

' Pg1 ' PR1 ' XE1 ]

(7.2)

And for area 2 the state vector is: [X]T = [ ' F2

' Pg2 ' PR2 ' XE2 ]

(7.3)

Impacts of Energy Systems’ Output Variability on Transmission and Distribution Networks

B1

' PD1

1 R1

- KI1 S

'

DB1 1 1+Tgv1

+

1 1+Tt1

'

XE1

Kp1 1+STp1

+

+

- KP1

F1

-

-

+

201

' PR1

+

' PG1 2 T12 S

-

A12 A12

B2

- KI2 S

' F2

DB2 +

-

1 1+Tgv2

1 1+Tt2

+

- KP2 1 R2

'

XE2

' PR2

' PG2

Kp2 1+STp2 ' PD2

Fig. 7-10: Two thermal area interconnected power system

The tie line is added to the system model as a state variable in the equation. In this case, nine state vectors define the state vector for a two Thermal-Thermal system namely [X]T = [ ' Ptie1 (7.4)

' F1 ' Pg1 ' PR1 ' XE1 ' F2 ' Pg2 ' PR2 ' XE2]

Equation 7.4 presented the variables in the state vector as differential equations. Different control systems are used in the power systems. However, the design of these control systems is not only related to the control parameters; it is also important to secure the stability and quality of the system. Different controllers are used (integral, differential, proportional, mix, fuzzy logic, genetic algorithm, etc.). Also, modern optimization approaches have been used.

Chapter Seven

202

The main aim of AGC is to make the steady state error in power and frequency due to the change in supply and load equal to zero. The controller used the interconnected area error signal as a feedback in the closed loop system. Equation 7.4 is augmented by two additional state variables X10 and X11, to design a classical controller as follows: X 10 =

³

ACE 1 dt + ACE1

and

X 11 =

³

ACE 2 dt + ACE2

In this case, the state vector of the system is defined by eleven state variables as: [X]T =[ ' Ptie1

 10 X  11 ] X

' F1 ' Pg1 ' PR1 ' XE1 ' F2 ' Pg2 ' XE2 ' PR2 (7.5)

Akagi, Hirofumi, Yoshihira Kanazawa (1984) developed a theory to be applied to renewable energy power injected by Watanabe (Watanabe, Edson H., Richard M. Stephan 1993) (Barbosa, P. G., L. G. B. Rolim, E. H. Watanabe 1998), used to design the control system for active and reactive power. The generated active power of photovoltaic systems is injected to the grid at several points in the distribution system. However, some negative consequences for the distribution system occur because of the renewable energy injected power. The intermittent nature of solar energy affects photovoltaic productivity, which consequently has an impact on the reliability, quality, and availability of the distributed system. As a result of this intermittent behavior and the active energy injection, it was found that a reliable energy prediction throughout the distribution system from different points became more complicated. In the case of high-power production from a renewable energy system (i.e., photovoltaic or wind energy systems), the distribution system could be out of control and the risk of an increase in voltage and reverse energy circulation became high (Kolhe and Rasul 2020). As an example of active and reactive power control for renewable energy technology connected to the grid, consider Figure 7-11. This shows a wind turbine generator connected to the grid via a voltage source inverter, which contains a rectifier and inverter linked through the capacitor. The rectifier converts the AC power generated from the wind turbine and the voltage is smoothed using the capacitor to produce DC power. The rectifier’s unregulated voltage became the input to the inverter, which in turn

Impacts of Energy Systems’ Output Variability on Transmission and Distribution Networks

203

converted the DC power Pg into AC power P. However, the PWM was used to control the inverter operation to control the active and reactive power injected to the grid.

Fig. 7-11: Wind turbine generator connected to the grid

Integrity of multiple renewable energy resources The combination of different renewable and sustainable energy sources (i.e., photovoltaic, wind turbine, etc.) can be viewed as a feature of DG, where DG generators are in the range of 1 kW to 10 MW. At the point of interconnection with DS, these little, generation technologies can shape another sort of power system, the microgrid. The microgrid idea is a group of micro sources and loads working as a solitary controllable system that can give heat and power to the neighborhood. The electrical connection of loads and sources should be possible through an AC and DC link, or an HFAC link.

DC Link An example of the integrity of multiple renewable energy resources is a group of small hydroelectric systems which drives induction generators. It is important to control voltage at the load link bus to provide power of a certain quality to the load. However, according to the power control strategy, the controlled current source has to adjust the voltage source on the busbar.

204

Chapter Seven

(a)

(b)

Impacts of Energy Systems’ Output Variability on Transmission and Distribution Networks

205

(c) Fig. 7-12: Energy integration (a) DC Link, (b) AC Link, (c) HFAC Link

Figure 7-12 illustrates the energy integration and different configurations using DC, AC and HFAC links. Thyristors and gate turn-off thyristors (GTOs) are still used in these systems for their high current and voltage rating capacity, reliability, cost, control, etc. The DC link is the oldest type used in this kind of system as shown in Figure 7-12a. The direct connection of DC supply to the load and battery is an example. The DC link has a proven utilization in many applications as for example, AC-DC converters, HVDC, controlled links, etc. Figure 7-12a shows the case for induction and synchronous generators (rotatory), where two converters are used to feed the DC link. However, in a stationary generator, a freewheeling diode is necessary to prevent back-feeding to the supply.

AC Link The AC link is another option, which works at 50 or 60 Hz as shown in Figure 7-10b. The busbar could be an islanded operation local grid or public grid. However, standards and utility requirements need to be flowing in this interconnection. In Figure 7-12b, the AC link feeds the grid directly from rotatory generation. The DC-AC inverter is needed for the case of stationary generation.

206

Chapter Seven

HFAC Link The HFAC system is an interface between a standalone operation and a utility grid. It provides small dimensions, fix fault protection and work in power quality functions. Many applications in industry use high frequencies (i.e., 400 Hz), for example buses, boats, submarines, etc. Also, some HFAC systems are used in electric vehicles (Kiel, n.d.) using an AC controller. In this case, two stages of AC-DC and DC-AC converters are used as shown in Figure 7-12c.

Interconnection and islanding control A critical issue identified in microgrids is islanding. This condition happens when the microgrid keeps on empowering part of the main network after that part has been isolated from the primary network. Islanding is important for the microgrid and occurs when the microgrid continues providing energy to a section of the main grid after it gets isolated. Islanding occurring intentionally is a concern for the grid because the source is part of the grid but not controlled by the utility. Utility equipment and workers could be affected by this islanding. Also, reconnection needs more arrangements made in terms of synchronization, etc. For all these reasons, it is important for the utility to avoid intentional islanding. As an example, let’s consider the case in Figure 7-13. A PV microgrid is connected to the utility through the transformer and reclosed. At the same time, the microgrid is connected to the local load. If for whatever reason the switch is opened, then the microgrid will continue to supply the load with energy. This is called an intentional island, because the load is supposed to be isolated from the utility but the microgrid still energizes the load. However, for these types of cases there should be a protective relay and sensors to work and send signals to disconnect the microgrid.

Impacts of Energy Systems’ Output Variability on Transmission and Distribution Networks

207

Fig. 7-13: Microgrid and islanding problem

Distributed generation power injection and control A microgrid can be actualized by joining sources and loads, taking into consideration international islanding, and utilizing accessible waste heat. There is a solitary point of association with the network called the “point of common coupling (PCC)”. In the microgrid, some feeders can have delicate burdens that require local generation. International islanding from the network is given by static switches that can isolate them in under a cycle. At the point when the microgrid is associated, power from neighborhood generation can be coordinated to the feeder with noncritical loads or be sold to the network whenever concurred or permitted by “net metering”. This section discusses distributed generation control and power injection. Figure 7-14 shows the structure of the distributed generation system control. It contains subsystems linked to each other and each one can coordinate lower-level units and be self-coordinated by an upper-level unit. The actuator (first layer), the most time-demanding control, is related to the thyristor in the power electronic converter. The process of the task is fast (high-speed). However, the layer takes into consideration the pulse width modulation, harmonics, protection, space vector, etc.

208

Chapter Seven

Fig. 7-14: Distributed generation system control structure

Conclusions In this chapter the impact of the variability of renewable energy system output on the transmission and distribution system is discussed. The effect of environmental parameters has been explained. However, power electronic converters, used to control the power injected to the grid are presented. It is found that the control of active and reactive power is faster and more reliable compared with conventional system control methods. On the other hand, the power quality problem has raised harmonics as an example, which need to be considered and solved using filters.

References Akagi, Hirofumi, Yoshihira Kanazawa, and Akira Nabae. 1984. “Instantaneous Reactive Power Compensators Comprising Switching Devices without Energy Storage Components.” IEEE Transactions on Industry Applications 3, 625-30. Al-Hasan, A. Y., A. A. Ghoneim, and A. H. Abdullah. 2004. “Optimizing Electrical Load Pattern in Kuwait Using Grid Connected Photovoltaic Systems.” Energy Conversion and Management 45(4), 483-94. https://doi.org/10.1016/S0196-8904(03)00163-8. Al-Otaibi, A., A. Al-Qattan, F. Fairouz, and A. Al-Mulla. 2015. “Performance Evaluation of Photovoltaic Systems on Kuwaiti Schools’ Rooftop.” Energy Conversion and Management 95, 110-19. https://doi.org/10.1016/j.enconman.2015.02.039.

Impacts of Energy Systems’ Output Variability on Transmission and Distribution Networks

209

Al-Sabounchi, Ammar M., Saeed A. Yalyali, and Hamda A. Al-Thani. 2013. “Design and Performance Evaluation of a Photovoltaic GridConnected System in Hot Weather Conditions.” Renewable Energy 53, 71-78. https://doi.org/10.1016/j.renene.2012.10.039. Al-Waeli, A. H. A., A. M. K. El-Din, A. H. K. Al-Kabi, A. Al-Mamari, H. A. Kazem, and M. T. Chaichan. 2017. “Optimum Design and Evaluation of Solar Water Pumping System for Rural Areas.” International Journal of Renewable Energy Research 7(1). Al-Waeli, A. H. A., H. A. Kazem, K. Sopian, and M. T. Chaichan. 2017. “Techno-Economical Assessment of Grid Connected PV/T Using Nanoparticles and Water as Base-Fluid Systems in Malaysia.” International Journal of Sustainable Energy. https://doi.org/10.1080/14786451.2017.1323900. Al-Waeli, A. H. A., K. Sopian, M. T. Chaichan, H. A. Kazem, A. Ibrahim, S. Mat, and M. H. Ruslan. 2017. “Evaluation of the Nanofluid and Nano-PCM Based Photovoltaic Thermal (PVT) System: An Experimental Study.” Energy Conversion and Management 151. https://doi.org/10.1016/j.enconman.2017.09.032. Al-Waeli, A. H. A., K. Sopian, H. A. Kazem, and M. T. Chaichan. 2017. “Photovoltaic/Thermal (PV/T) Systems: Status and Future Prospects.” Renewable and Sustainable Energy Reviews 77. https://doi.org/10.1016/j.rser.2017.03.126. Al-Waeli, Ali H. A., Miqdam T. Chaichan, Hussein A. Kazem, and K. Sopian. 2019. “Evaluation and Analysis of Nanofluid and Surfactant Impact on Photovoltaic-Thermal Systems.” Case Studies in Thermal Engineering 13 (March). https://doi.org/10.1016/j.csite.2019.100392. Al-Waeli, Ali H. A., Miqdam T. Chaichan, Hussein A. Kazem, K. Sopian, Adnan Ibrahim, Sohif Mat, and Mohd Hafidz Ruslan. 2018. “Comparison Study of Indoor/Outdoor Experiments of a Photovoltaic Thermal PV/T System Containing SiC Nanofluid as a Coolant.” Energy 151, 33-44. https://doi.org/10.1016/j.energy.2018.03.040. Al-Waeli, Ali H. A., Miqdam T. Chaichan, K. Sopian, Hussein A. Kazem, Hameed B. Mahood, and Anees A. Khadom. 2019. “Modeling and Experimental Validation of a PVT System Using Nanofluid Coolant and Nano-PCM.” Solar Energy 177 (January), 178-91. https://doi.org/10.1016/j.solener.2018.11.016. Al-Waeli, Ali H. A., Atma H. K. Al-Kabi Moanis M. Kamal El-Din, Asma Al-Mamari, Hussein A. Kazem, and Miqdam T. Chaichan. 2015. “A Photovoltaic Application in Car Parking Lights with Recycled Batteries: A Techno-Economic Study.” Australian Journal of Basic and Applied Sciences 9(36), 43-45.

210

Chapter Seven

Al-Waeli, Ali H. A., K. Sopian, Jabar H. Yousif, Hussein A. Kazem, John Boland, and Miqdam T. Chaichan. 2019. “Artificial Neural Network Modeling and Analysis of Photovoltaic/Thermal System Based on the Experimental Study.” Energy Conversion and Management 186 (February), 368-79. https://doi.org/10.1016/j.enconman.2019.02.066. Alwaeli, H. A. K. 2003. “Effect of Speed-Governor Deadband on {AGC} Analysis for Two Areas Thermal-Thermal System.” Sciences & Technologie B 20(1), 33-37. Amin, Nowshad, Chin Wen Lung, and Kamaruzzaman Sopian. 2009. “A Practical Field Study of Various Solar Cells on Their Performance in Malaysia.” Renewable Energy 34(8), 1939-46. https://doi.org/10.1016/j.renene.2008.12.005. Bahattab, Mohammed A., Ibrahim A. Alhomoudi, Mohammed I. Alhussaini, Mark Mirza, Jan Hegmann, Walther Glaubitt, and Peer Löbmann. 2016. “Anti-Soiling Surfaces for PV Applications Prepared by Sol-Gel Processing: Comparison of Laboratory Testing and Outdoor Exposure.” Solar Energy Materials and Solar Cells 157, 42228. https://doi.org/10.1016/j.solmat.2016.07.004. Barbosa, P. G., L. G. B. Rolim, E. H. Watanabe, and R. Hanitsch. 1998. “Control Strategy for Grid-Connected DC-AC Converters with Load Power Factor Correction.” In IEE Proceedings-Generation, Transmission and Distribution, 487-92. Beser, Ersoy, Birol Arifoglu, Sabri Camur, and Esra Kandemir Beser. 2010. “A Grid-Connected Photovoltaic Power Conversion System with Single-Phase Multilevel Inverter.” Solar Energy 84(12), 2056-67. https://doi.org/10.1016/j.solener.2010.09.011. Bohn, E. V., and Samir M. Miniesy. 1972. “Optimum Load-Frequency Sampled-Data Control with Randomly Varying System Disturbances.” IEEE Transactions on Power Apparatus and Systems 5, 1916-23. Bradley, David Allan. 1995. Power Electronics. CRC Press. Busaidi, A. S. Al, H. A. Kazem, A. H. Al-Badi, and M. Farooq Khan. 2016. “A Review of Optimum Sizing of Hybrid PV-Wind Renewable Energy Systems in Oman.” Renewable and Sustainable Energy Reviews 53. https://doi.org/10.1016/j.rser.2015.08.039. C. Protogeropoulos, I. Klonaris, C. Petrocheilos, I. Charitos, and I. Martinac. 2010. “Performance Evaluation of Different PV Module Technologies in a Grid-Connected Pilot Project in Greece.” In 25th European Photovoltaic Solar Energy Conference and Exhibition/5th World Conference on Photovoltaic Energy Conversion, 6-10 September 2010, Valencia, Spain, 4601-6. Celik, Ali Naci. 2006. “Present Status of Photovoltaic Energy in Turkey

Impacts of Energy Systems’ Output Variability on Transmission and Distribution Networks

211

and Life Cycle Techno-Economic Analysis of a Grid-Connected Photovoltaic-House.” Renewable and Sustainable Energy Reviews 10 (4), 370-87. https://doi.org/10.1016/j.rser.2004.09.007. Chaichan, Miqdam T., Hussein A. Kazem, Ali A. Kazem, Khalil I. Abaas, and Kadhim A. H. Al-Asadi. 2015. “The Effect of Environmental Conditions on Concentrated Solar System in Desertec Weathers.” International Journal of Scientific and Engineering Research 6(5), 850-56. Chaichan, Miqdam T., and Hussein A. Kazem. 2015. “Water Solar Distiller Productivity Enhancement Using Concentrating Solar Water Heater and Phase Change Material (PCM).” Case Studies in Thermal Engineering 5, 151-59. https://doi.org/10.1016/j.csite.2015.03.009. Chaichan, Miqdam T., Khaleel I. Abass, and Hussein A. Kazem. 2018. “Dust and Pollution Deposition Impact on a Solar Chimney Performance.” International Research Journal of Advanced Engineering and Science 3(1), 127-32. Chaichan, Miqdam Tariq, and Hussein A. Kazem. 2018. Generating Electricity Using Photovoltaic Solar Plants in Iraq. Generating Electricity Using Photovoltaic Solar Plants in Iraq. https://doi.org/10.1007/978-3-319-75031-6. Chan, Wah-Chun, and Yuan-Yih Hsu. 1981. “Automatic Generation Control of Interconnected Power Systems Using Variable-Structure Controllers.” In Generation, Transmission and Distribution, 269-79. Demirbas, Ayhan. 2016. Waste Energy for Life Cycle Assessment. https://doi.org/10.1007/978-3-319-40551-3. Dudley, Bob. 2015. “BP Statistical Review of World Energy 2016.” London, UK. Elgerd, Olle I., and Charles E. Fosha. 1970. “Optimum MegawattFrequency Control of Multiarea Electric Energy Systems.” IEEE Transactions on Power Apparatus and Systems 4, 556-63. Elgerd, Olle Ingemar. 1982. Electric Energy Systems Theory: An Introduction. ElNozahy, M. S., and M. M. A. Salama. 2013. “Technical Impacts of Grid-Connected Photovoltaic Systems on Electrical Networks.” Journal of Renewable and Sustainable Energy 5(3), 032702. Eltawil, Mohamed A., and Zhengming Zhao. 2010. “Grid-Connected Photovoltaic Power Systems: Technical and Potential Problems – A Review.” Renewable and Sustainable Energy Reviews 14(1), 112-29. https://doi.org/10.1016/j.rser.2009.07.015. Farret, Felix A. 2006. Integration of Alternative Sources of Energy. John Wiley & Sons, Inc.

212

Chapter Seven

Fudholi, A., K. Sopian, M. H. Yazdi, M. H. Ruslan, M. Gabbasa, and H. A. Kazem. 2014. “Performance Analysis of Solar Drying System for Red Chili.” Solar Energy 99. https://doi.org/10.1016/j.solener.2013.10.019. Fudholi, Ahmad, Kamaruzzaman Sopian, Mohammad H. Yazdi, Mohd Hafidz Ruslan, Adnan Ibrahim, and Hussein A. Kazem. 2014. “Performance Analysis of Photovoltaic Thermal (PVT) Water Collectors.” Energy Conversion and Management 78, 641-51. https://doi.org/10.1016/j.enconman.2013.11.017. Green, Martin A., Ewan D. Dunlop, Dean H. Levi, Jochen HohlဨEbinger, Masahiro Yoshita, and Anita W. Y. HoဨBaillie. 2019. “Solar Cell Efficiency Tables (Version 54).” Progress in Photovoltaics: Research and Applications 27(7), 565-75. Humada, Ali M., Ashty M. Aaref, Hussein M. Hamada, Mohd Herwan Sulaiman, Nowshad Amin, and Saad Mekhilef. 2018. “Modeling and Characterization of a Grid-Connected Photovoltaic System under Tropical Climate Conditions.” Renewable and Sustainable Energy Reviews 82 (July 2015), 2094-2105. https://doi.org/10.1016/j.rser.2017.08.053. Humada, Ali M., Mojgan Hojabri, Hussein M. Hamada, Fahmi B. Samsuri, and Mushtaq N. Ahmed. 2016. “Performance Evaluation of Two PV Technologies (c-Si and CIS) for Building Integrated Photovoltaic Based on Tropical Climate Condition: A Case Study in Malaysia.” Energy and Buildings 119, 233-41. https://doi.org/10.1016/j.enbuild.2016.03.052. Ibrahim, Adnan, Ahmad Fudholi, Kamaruzzaman Sopian, Mohd Yusof Othman, and Mohd Hafidz Ruslan. 2014. “Efficiencies and Improvement Potential of Building Integrated Photovoltaic Thermal (BIPVT) System.” Energy Conversion and Management 77, 527-34. https://doi.org/10.1016/j.enconman.2013.10.033. Jagoda, Kalinga, Robert Lonseth, Adam Lonseth, and Tom Jackman. 2011. “Development and Commercialization of Renewable Energy Technologies in Canada: An Innovation System Perspective.” Renewable Energy 36(4), 1266-71. Kassakian, John G., Martin F. Schlecht, and George C. Verghese. 2000. Principles of Power Electronics. Graphis. Kaundinya, Deepak Paramashivan, P. Balachandra, and N. H. Ravindranath. 2009. “Grid-Connected versus Stand-Alone Energy Systems for Decentralized Power – A Review of Literature.” Renewable and Sustainable Energy Reviews 13(8), 2041-50. https://doi.org/10.1016/j.rser.2009.02.002.

Impacts of Energy Systems’ Output Variability on Transmission and Distribution Networks

213

Kazem, H. A., S. Q. Ali, A. H. A. Alwaeli, K. Mani, and M. Tariq. 2013. “Life-Cycle Cost Analysis and Optimization of Health Clinic PV System for a Rural Area in Oman.” In Lecture Notes in Engineering and Computer Science. Vol. 2 LNECS. Kazem, H. A., M. T. Chaichan, A. H. Alwaeli, and K. Mani. 2017. Effect of Shadows on the Performance of Solar Photovoltaic. Mediterranean Green Buildings and Renewable Energy: Selected Papers from the World Renewable Energy Network’s Med Green Forum. https://doi.org/10.1007/978-3-319-30746-6_27. Kazem, H. A., T. Khatib, and K. Sopian. 2013. “Sizing of a Standalone Photovoltaic/Battery System at Minimum Cost for Remote Housing Electrification in Sohar, Oman.” Energy and Buildings 61. https://doi.org/10.1016/j.enbuild.2013.02.011. Kazem, Hussein A., Miqdam T. Chaichan, Ali H. A. Al-Waeli, and K. Sopian. 2020. “Evaluation of Aging and Performance of GridConnected Photovoltaic System Northern Oman: Seven Years’ Experimental Study.” Solar Energy 207 (September), 1247-58. https://doi.org/10.1016/j.solener.2020.07.061. Kazem, Hussein A., and Tamer Khatib. 2013. “Techno-Economical Assessment of Grid Connected Photovoltaic Power Systems Productivity in Sohar, Oman.” Sustainable Energy Technologies and Assessments 3, 61-65. https://doi.org/10.1016/j.seta.2013.06.002. Kazem, Hussein A., and Miqdam T. Chaichan. 2016. “Design and Analysis of Standalone Solar Cells in the Desert of Oman.” Journal of Scientific and Engineering Research 3(4), 62-72. Kazem, Hussein A., Miqdam T. Chaichan, and Jabar H. Yousif. 2019. “Evaluation of Oscillatory Flow Photovoltaic/Thermal System in Oman” 6(1). Keyif, Enes, Michael Hornung, and Wanshan Zhu. 2020. “Optimal Configurations and Operations of Concentrating Solar Power Plants under New Market Trends.” Applied Energy 270, 115080. Khatib, Tamer, Kamaruzzaman Sopian, and Hussein A. Kazem. 2013. “Actual Performance and Characteristic of a Grid Connected Photovoltaic Power System in the Tropics: A Short Term Evaluation.” Energy Conversion and Management 71, 115-19. https://doi.org/10.1016/j.enconman.2013.03.030. Kiel, Christian-albrechts-universität. n.d. “Dimensioning of a Current Source Inverter for the Feed-in of Electrical Energy from Fuel Cells to the Mains.” Fuel Cell, 1-7. Kim, Ju Young, Gyu Yeob Jeon, and Won Hwa Hong. 2009. “The Performance and Economical Analysis of Grid-Connected Photovoltaic

214

Chapter Seven

Systems in Daegu, Korea.” Applied Energy 86(2), 265-72. https://doi.org/10.1016/j.apenergy.2008.04.006. Kolhe, Mohan Lal, and M. J. M. A. Rasul. 2020. “3-Phase Grid-Connected Building Integrated Photovoltaic System with Reactive Power Control Capability.” Renewable Energy 154, 1065-75. https://doi.org/10.1016/j.renene.2020.03.075. Komoni, Vjollca, Isuf Krasniqi, Arben Lekaj, and Ilir Gashi. 2014. “Performance Analysis of 3.9 KW Grid Connected Photovoltaic Systems in Kosova.” IREC 2014 – 5th International Renewable Energy Congress, 0-5. https://doi.org/10.1109/IREC.2014.6826947. Kong, Jing, Jianlei Niu, and Chengwang Lei. 2020. “A CFD Based Approach for Determining the Optimum Inclination Angle of a RoofTop Solar Chimney for Building Ventilation.” Solar Energy 198, 55569. Kothari, M. L., J. Nanda, D. P. Kothari, and D. Das. 1989. “Discrete-Mode Automatic Generation Control of a Two-Area Reheat Thermal System with New Area Control Error.” IEEE Transactions on Power Systems 4(2), 730-38. Lander, Cyril W. 1987. Power Electronics. McGraw-Hill. Owen, Anthony D. 2006. “Renewable Energy: Externality Costs as Market Barriers.” Energy Policy 34(5), 632-42. Perez, Richard, and Marc Perez. 2009. “A Fundamental Look at Energy Reserves for the Planet.” Rashid, Muhammad H. 1993. Power Electronics: Circuits, Devices, and Applications. Prentice Hall. Sidrach-de-C. M. and López L. M. 1998. “Evaluation of a Grid-Connected Photovoltaic System in Southern Spain.” Renewable Energy 15, 199699. https://www.britannica.com/science/renewable-energy. Sulaiman, Shahril Irwan, Titik Khawa Abdul Rahman, Ismail Musirin, Sulaiman Shaari, and Kamaruzzaman Sopian. 2012. “An Intelligent Method for Sizing Optimization in Grid-Connected Photovoltaic System.” Solar Energy 86(7), 2067-82. https://doi.org/10.1016/j.solener.2012.04.009. Tahri, Fatima, Ali Tahri, and Takashi Oozeki. 2018. “Performance Evaluation of Grid-Connected Photovoltaic Systems Based on Two Photovoltaic Module Technologies under Tropical Climate Conditions.” Energy Conversion and Management 165 (March), 24452. https://doi.org/10.1016/j.enconman.2018.03.065. Tripathi, Brijesh, Pankaj Yadav, Siddharth Rathod, and Manoj Kumar. 2014. “Performance Analysis and Comparison of Two Silicon Material Based Photovoltaic Technologies under Actual Climatic Conditions in

Impacts of Energy Systems’ Output Variability on Transmission and Distribution Networks

215

Western India.” Energy Conversion and Management 80, 97-102. https://doi.org/10.1016/j.enconman.2014.01.013. Tripathy, S. C., T. S. Bhatti, C. S. Jha, O. P. Malik, and G. S. Hope. 1984. “Sampled Data Automatic Generation Control Analysis with Reheat Steam Turbines and Governor Dead-Band Effects.” IEEE Transactions on Power Apparatus and Systems 5, 1045-51. Verma, Deepak, Ole Morten Midtgård, and Tor O. Sætre. 2011. “Review of Photovoltaic Status in a European (EU) Perspective.” Conference Record of the IEEE Photovoltaic Specialists Conference, 003292-97. https://doi.org/10.1109/PVSC.2011.6186641. Wang, Dengjia, Liang Hu, Hu Du, Yanfeng Liu, Jianxiang Huang, Yanchao Xu, and Jiaping Liu. 2020. “Classification, Experimental Assessment, Modeling Methods and Evaluation Metrics of Trombe Walls.” Renewable and Sustainable Energy Reviews 124, 109815. Watanabe, Edson H., Richard M. Stephan, and Mauricio Aredes. 1993. “New Concepts of Instantaneous Active and Reactive Powers in Electrical Systems with Generic Loads.” IEEE Transactions on Power Delivery 8(2), 697-703. Xydis, G. 2013. “The Wind Chill Temperature Effect on a Largeဨscale PV Plant – an Exergy Approach.” Progress in Photovoltaics: Research and Applications 21(8), 1611-24. Yousif, Jabar H., Haitham A. Al-Balushi, Hussein A. Kazem, and Miqdam T. Chaichan. 2019. “Analysis and Forecasting of Weather Conditions in Oman for Renewable Energy Applications.” Case Studies in Thermal Engineering 13 (October 2018), 100355. https://doi.org/10.1016/j.csite.2018.11.006.

Nomenclature I

subscript referring to area I (I = 1, n)

' Ptie incremental change in the tie line power ' Fi incremental change in the frequency deviation of area I ' Pci incremental change in the speed changer position of area I Tri Ri

EI t T [B] G

reheat time constant of area I speed regulation due to the governor action of area I natural area frequency response characteristic of area I Time sampling interval input matrix area power angle

216

Chapter Seven

wo J

angular frequency of natural sinusoidal oscillation, radian cost function incremental load change of area I

' Pdi Kri

' XEi T1, T3 Tw Bi KI K [A] [E] x F(x) N 1, N2

reheat coefficient of area I incremental change in the governor valve position of area I time constants of the steam governor water starting time frequency bias setting of area I integral gain sampling instant system matrix disturbance matrix state vector nonlinear function of x furrier series coefficient associated with x and sx respectively

CHAPTER EIGHT THE ROLE OF ENERGY STORAGE TO REDUCE VARIABILITY IMPACTS MOSTAFA BAKHTVAR1, HAMIDREZA AGHAY KABOLI1 AND AMER AL-HINAI1,2 ABSTRACT The power generated by solar PV is intermittent, variable, limited when the sky is overcast, unavailable at nights, and asynchronous. Replacing the power generated by conventional generators with solar PV generation raises apprehensions on power adequacy, flexibility and reactive power scarcity, network capacity limitations, fault behavior, frequency support, and power quality, among others. This chapter focuses on the role of energy storage systems in reducing the negative impacts of PV systems on power systems. It presents an overview of energy storage technologies and how each technology can benefit the power system to alleviate the impacts of solar PV generation. The concept of dispatchable PV-Energy storage hybrid systems is also discussed in this chapter. The realization of dispatchable solar PV generation can reduce the need for flexibility and reserve. A framework for a dispatchable solar PV generation is proposed. The components of this framework and the tasks each should carry out are articulated. Keywords: Energy Storage; Variability; Solar PV Generation; Dispatchable; Flexibility

1

Sultan Qaboos University, Sustainable Energy Research Center, Muscat, Oman Sultan Qaboos University, Dept. of Electrical and Computer Engineering, Muscat, Oman 2

218

Chapter Eight

Introduction There is an ever-growing penetration of renewables in power systems (REN21 2019). Renewables are attractive sources of energy for electric power generation as they can potentially reduce greenhouse gas emissions and benefit from zero fuel cost. The power produced from renewable energy sources, in particular solar and wind, is replacing power from conventional generation units. The main drivers of this shift are the priority of dispatch of renewables and their low marginal cost (Bakhtvar 2016). Therefore, in normal operating conditions, renewables inject the maximum available power to the grid, if they do not result in breaching any constraints in the power system. A prominent feature of wind and solar energy resources, i.e., wind speed (kinetic energy of air mass) and solar irradiance (radiant energy), is their intermittency. Accordingly, solar and wind generation are inherently nondispatchable. A dispatchable power generation unit can be turned on and off and its power output can be ramped up and down in a given time per the request of the system operator. Theoretically, the power output of solar and wind generation units can be changed from zero to the maximum power available by controlling the power-electronic-based converters used in these generation units (Ye, et al. 2017). Owing to the variability of the instantaneous available wind and solar energy, the maximum available power from solar and wind generation units is also variable (Willis 2004). Conventionally, the provision of sufficient inertia, spinning reserve, and ramping capability by conventional power plants together with their associated controls enabled the power system to deal with the demand variability. However, the integration of wind and solar generation into the power systems adds to the variability of the instantaneous net demand, such that both the frequency and depth of ramping events may increase (Cochran, et al. 2014). Accommodating the high penetration of renewables entails an even greater spinning reserve and flexibility to ensure the security and reliability of the power system (Lannoye, Flynn and O’Malley 2015). Moreover, the seasonal and daily variations of renewables may not correlate with the demand; this means that a large power generation capacity (conventional) should be in place to compensate for a power deficit only in small fractions of the year. A large part of the solar PV installed capacity may be embedded in the LV distribution network and owned by prosumers (Wirth 2019). Prosumers are electricity customers that can both produce and consume electricity. On one

The Role of Energy Storage to Reduce Variability Impacts

219

hand, prosumers can reduce the net demand in distribution feeders with local electricity generation. On the other hand, the power system operator has very limited (or none) control over the injected power by prosumers; this is a source of concern, particularly, under abnormal operating conditions. For example, during an eclipse, a sharp change of power generated by solar PV cells is anticipated. Fig. 8-1 shows the solar generation in Germany for 20 March 2015. A solar eclipse occurred on this specific day. The phenomenon started at 9:26 AM, it reached 80% peak sun covering at 10:34 AM and ended at 11:46 AM. The eclipse imposed an 11.5 GW change of solar power generation in less than 75 minutes. Such a large and sharp change of power not only jeopardizes power system adequacy but also endangers the power system stability. As implied, the integration of renewable generation in power systems, despite its near-zero marginal cost, drives up the cost of power system operation (Troy, Denny and O’Malley 2010, Doherty and O’Malley 2005). 25

Power (GW)

20 15 10 5 0 0:00

6:00

12:00 Time

18:00

0:00

Fig. 8-1: Solar generation on 20 March 2015 for Germany (adapted from (Open Power System Data n.d.))

Storage devices have been an active area of research and development in the past 40 years. Not only have various technologies been proposed, but there has also been a major focus on improving the capacity, efficiency, lifetime and most importantly reducing the cost of the existing storage technologies. Therefore, bulk storage of electrical energy which was deemed impossible in the past, due to its associated costs and technical

220

Chapter Eight

challenges, nowadays is an economically viable and achievable solution. Since storage devices are capable of supplying and absorbing both active and reactive power on demand and as fast as a few milliseconds, they can be employed for various purposes in power systems including energy arbitrage, time-shifting, peak shaving, load following, renewable power smoothing, renewable power flattening, and frequency and voltage support (Ortega Manjavacas 2017, Palizban and Kauhaniemi 2016). A well-chosen deployment and control of storage devices can enhance the power quality and reliability of power systems and/or be used for making revenue (O’Dwyer, et al. 2018, Bakhtvar, Cabrera, et al. 2017). Storage of solar PV generation is an attractive solution that enables the alleviation of the variability and intermittency of solar PV generation. It can reduce the operational costs of power systems (Habibi 2001) by avoiding part of the extra cost of the flexibility and spinning reserve that is imposed by the solar PV generation (Mills, et al. 2013). Storage devices convert the electric energy produced by solar PVs into various forms of energy such as electrochemical, electromagnetic, kinetic, heat, and potential energy. Accordingly, solar power/energy generated by PVs is stored when it is abundant and released at instances when there is a deficit of power/energy. The energy capacity and charge and discharge rate are important features of storage devices. For a storage device to be effective, it must have an adequate energy capacity and a sufficiently high charge and discharge rate. This ensures that the abundant solar generation can be stored, and the power deficits can be fulfilled. The suitability of the capacity and charge and discharge rate depends on the installed solar PV capacity, historic solar irradiance, and temperature measurements at the installation site and the actual purpose of the energy storage system. For example, compared to time-shifting, the smoothing of solar PV output requires a faster charge and discharge rate and a smaller capacity since smoothing requires treating the fast and large power output changes that last for a short period (Traube, et al. 2013). Accordingly, the ramping capability of the energy storage system must also be considered. Fig. 8-2 shows the cumulative probability of the power gradient (5 minute sampling rate) for 56 utility and distribution scale solar PV systems in Texas over 1 year. It can be noted that an energy storage system must be able to change its power output at the rate of 0.04 p.u./min (normalized to the installed PV capacity) to be able to cover approximately 99% of power gradients.

The Role of Energy Storage to Reduce Variability Impacts

221

Fig. 8-2: Cumulative probability of solar PV power gradient for 56 locations in Texas over 1 year (adapted from (National Renewable Energy Laboratory n.d.))

Storage Technologies Various technologies have been developed for storing electrical energy. A common classification of these technologies divides them into 5 categories, i.e., mechanical, electrochemical, chemical, electrical, and thermal. Various types of energy storage systems are shown in Fig. 8-3. Some of the energy storage systems shown in Fig. 8-3 are described in this section.

222

Chapter Eight

Fig. 8-3: Various types of energy storage system (adapted from (Luo, et al. 2015, Cheng, Sami and Wu 2017))

The Role of Energy Storage to Reduce Variability Impacts

223

Electrochemical Electrochemical storage refers to those storage technologies that are based on the storage and release of electrons through chemical reactions triggered by applying a potential difference. Various types of batteries fall under electrochemical storage technologies.

Liquid and solid-state batteries Lithium-ion batteries are one of the most common liquid-state battery technologies. Lithium-ion batteries work based on the oxidation potential of chemical species. The tendency of chemical species to lose electrons is referred to as oxidation potential (Afzal Ahmadi 2009). Lithium and fluorine have the highest and lowest oxidation potential, respectively (Gaffney and Marley 2018). The former has only one electron in its valence layer and tends to lose it. The high oxidation potential of lithium makes it a suitable choice for battery industry applications. Fig. 8-4 schematically shows a lithium-ion battery. In lithium-ion batteries, lithium is present in one of the electrodes (cathode) of the battery within a metal oxide structure. The other electrode (anode) is commonly made of graphite (Deng 2015). The electrodes are submerged in an electrolyte solution. The electrolyte solution is a conductive liquid and allows electrical charges to move freely. To ensure that the electrodes of the battery are electrically isolated from each other and the electrolyte does not cause a short circuit between them, an ion-exchange membrane is used to separate the electrodes from each other. The ion-exchange membrane only allows ions to pass through it and prevents non-ion particles such as electrons. The electrodes of the battery are connected to its terminals through collectors commonly made of copper and/or aluminum. When a sufficiently large reverse voltage is applied to the terminals of the battery, the weak bonds of lithium with the metal oxide brake. The electron in the valence layer of the released lithium is separated and attracted to the collector (which has a positive voltage) and leaves behind the now positively charged lithium ions. The electrons travel through the external circuit and reach the graphite electrode. Similarly, the lithium ions move towards the graphite electrode (which has a negative voltage) through the path provided by the electrolyte and reach the graphite electrode after crossing the ion-exchange membrane. Due to the loose bond between the layered structure of graphite, the now joint lithium ion and electrons get trapped in it. A Li-ion battery is fully charged when most of the lithium ions and electrons move to the graphite electrode side. At this condition, lithium is unstable. Moreover, the absence of the shared electron of lithium in the

224

Chapter Eight

metal oxide electrode will give it a positive charge which attracts the electrons. Therefore, once the reverse voltage is removed, the electrons in the valence layer of lithium which is trapped in graphite tend to join back with the metal oxide. Since there is no external path, the only way for this to happen is through the electrolyte, however, the ion-exchange membrane prevents this from happening. In this situation, if an external path is provided between the two electrodes, the electrons will flow in this path and reach the metal oxide side. The flow of electrons produces electric current and can power up electrical equipment. Meanwhile, the lithium ions left behind are also attracted back to the metal oxide side and re-bond with the free electrons and reform the lithium metal oxide structure.

Fig. 8-4: Lithium-ion battery schematic (adapted from (Deng 2015))

Electrolytes are commonly unstable at the operational potential of the graphite electrode during charging. In the very first charge cycle of the lithium-ion battery, the lithium ions that pass through the ion-exchange membrane will also carry molecules of the electrolyte solvent. These will react with electrons accumulated in the graphite and form a layer of solid electrolyte interphase (SEI) that only allows the lithium ions to pass through it and prevents the electrons from further direct contact with the liquid electrolyte, therefore slowing down the degradation of the electrolyte. However, some electrons can still pass through the SEI and reach the electrolyte where they react with the lithium and solvent molecules and thicken the SEI layer gradually (Pinson and Bazant 2013). This process is sped up when the battery is operated. The consumption of electrolytes and

The Role of Energy Storage to Reduce Variability Impacts

225

lithium in this process degrades the battery over time. Moreover, the operation of the lithium-ion battery cells produces heat. If the heat is not managed efficiently, it raises the temperature of the battery and eventually leads to a thermal runaway. Thermal runaway occurs when the battery reaches a temperature at which the materials inside the battery start breaking in an exothermic process (Galushkin, Yazvinskaya and Galushkin 2018). If the battery is not able to dissipate the heat as fast as it was generated, the breaking of material will continue in an uncontrolled fashion and the flammable electrolyte may also catch fire. Therefore, the presence of a heat management system is vital for lithium-ion batteries. A newer generation of batteries has been named solid-state batteries. Fig. 85 shows a schematic of a solid-state battery. As implied, solid-state batteries have a similar structure and working principle to conventional lithium-ion batteries; however, the liquid electrolyte is replaced with solid electrolyte. The use of solid electrolyte alleviates some of the issues of the conventional lithium-ion batteries. Some of the advantages of solid-state batteries are (Kim, et al. 2015, Takada 2013): -

-

They are safer, as solid electrolyte is not flammable. The use of solid electrolyte enables employing a thinner ionexchange membrane which allows a reduction in the battery size hence improved energy density. The battery can be charged at a higher rate (although the discharge rate is low due to the low ionic conductivity of the electrolyte). The degradation of the battery is slowed down when solid electrolyte is used. This means that solid-state batteries can go through a larger number of charge and discharge cycles before they need to be replaced. The operational temperature range is generally larger, and the batteries can also be used at temperatures well below zero.

Solid-state batteries are more expensive compared to conventional lithiumion batteries (IRENA 2017). However, solid-state batteries are currently a hot research topic. Particularly, the development of a solid electrolyte that is as conductive as liquid electrolyte is a key enabler. The fast advancements in this area bring hope that the cost of solid-state batteries will drop in the future, and they will become a competitive storage solution.

226

Chapter Eight

Fig. 8-5: Solid-state battery schematic (adapted from (Kim, et al. 2015))

Flow Batteries Flow batteries (redox flow batteries) work on the principles of oxidation and reduction reactions of any suitable substances that can be electrically charged and discharged such as vanadium, iron chromium, zinc bromine, and zinc iron (Weber, et al. 2011). Error! Reference source not found. Fig. 8-6 schematically shows a flow battery. Redox flow batteries store energy in electrolyte liquids that can be stored in separate tanks named the catholyte and anolyte tanks. The electrolyte liquid in the catholyte tank is positively charged with cathode particles while the electrolyte liquid in the anolyte tank is negatively charged with anode particles. These solutions are pumped into the exchange chamber of the battery. Inside the exchange chamber, an ion-exchange membrane ensures that the catholyte and anolyte solutions are not mixed and remain separate. The membrane only allows ions to move from one side of the chamber to the other. Electrodes (e.g., porous graphite) are provided on each side of the chamber, such that the catholyte and anolyte come into contact with the electrodes. Collectors connect the electrodes to the terminals of the battery. When a large enough reverse voltage is applied to the terminals of the battery, electrons are separated from the catholyte (during an oxidation reaction triggered by the reverse voltage) and leave behind ions. The separated electrons travel through the membrane to the anolyte side and participate in a reduction

The Role of Energy Storage to Reduce Variability Impacts

227

reaction with the anode particles, lowering the electric potential of the anolyte side. Consequently, the ions that were left behind in the catholyte side travel to the anolyte side through the ion-exchange membrane and balance the cell. Conversely, when no voltage or a sufficiently small voltage is applied, an oxidation reaction occurs in the anolyte side and the released electrons are attracted back to the catholyte side through the provided external path. This flow of electrons produces an electric current that can be used to run electrical equipment or inject power to the grid.

Fig. 8-6: Flow battery schematic (adapted from (Qi and Koenig 2017))

The size and content of the electrolyte tanks define the capacity of a redox flow battery while the exchange chamber determines the power rate of these storage devices. By increasing these factors, the power and energy capacity characteristics of the flow batteries can be increased in an almost decoupled fashion. Since the exchange chamber is separated from the storage tank, flow batteries are very flexible for installation and operation, particularly for stationary applications. They can be operated both as a battery when the electrolytes are circulated in a closed loop and as a fuel cell when the electrolytes are circulated in an open loop, i.e., charged externally (SkyllasKazacos 2010). Flow batteries are excellent options where space limitation is not an issue. They can provide a large energy capacity and much power

228

Chapter Eight

with a relatively long lifetime. Therefore, despite the higher capital cost, their levelized cost of storage is significantly lower than that of solid-state batteries (Pawel 2014).

Electrical Those storage solutions that are based on the accumulation of electrons on a conductor and the establishment of electric and magnetic fields without the need of a chemical reaction are categorized as electrical storage technologies. The main members of this category are supercapacitors and superconducting magnetic energy storage systems.

Supercapacitors Conventional capacitors are made of two conductive plates separated by insulator (dielectric) material. When voltage is applied to the plates of the capacitor, electrons from the plate with high potential will move to the opposite plate and accumulate on it through the external path. This results in two plates with opposite charge. The flow of electrons continues until the potential difference between the plates becomes equal to that of the voltage source. At this stage, the capacitor is fully charged. Since the dielectric material does not allow the electrons to move from the negatively charged plate to the positively charged plate, an electric field is formed. The strength of the electric field is directly and reversely proportional to the potential difference between the plates and their distance, respectively. To ensure that the dielectric material does not break, the potential difference between the plates should always be kept below the breaking voltage of the dielectric material. The capacity of conventional capacitors is usually in the range of several hundred micro-farads. Such a small capacity is evidently not suitable for relatively large power applications. Supercapacitors replace the dielectric material with electrolyte and an ionexchange membrane. Fig. 8-7 schematically shows a supercapacitor. When a voltage source is connected across the terminals of a supercapacitor, like conventional capacitors, electrons are accumulated on the plate with higher potential, resulting with two plates with opposite charge. Within the electrolyte, the ion-exchange membrane ensures that electron transfer does not occur between the plates and by the electrolyte. However, the cations and anions of the electrolyte can pass through the ion-exchange membrane. Therefore, due to Coulomb force, the cations are attracted to the negatively charged plate and the anions are attracted to the opposite plate which is now

The Role of Energy Storage to Reduce Variability Impacts

229

positively charged. The attracted cations and anions form layers of opposite electric charge around the plates but separated from the plate by a molecule of electrolyte solvent. These layers are referred to as the Helmholtz double layers (Ander, et al. 2016). The distance between the electrodes and the cations and anions is in the range of a few angstroms. This is a much smaller distance compared to the distance between the charged plates in a conventional capacitor (Pattahil, Sivakumar and Sonia 2017). Therefore, a much larger capacitance is the result. Moreover, some of the cations and anions get adsorbed in the structure of the electrodes which includes active redox material, and create a pseudocapacitance effect (Ander, et al. 2016). The pseudocapacitance adds up to the capacitance effect from the two double layers to create an even larger equivalent capacitance in supercapacitors. It should be noted that pseudocapacitance is an electrochemical form of storage. Therefore, supercapacitors exhibit both electrical and electrochemical storage behavior.

Fig. 8-7: Schematic of supercapacitor energy storage with porous electrodes to increase the surface area (Dai, et al. 2016)

Despite their large capacitance, supercapacitors commonly allow for a smaller voltage compared to conventional capacitors (electrolytes break at lower voltage). However, they can be connected to each other to yield a supercapacitor with not only high capacity but also suitability for high voltages. Unlike batteries, the operation of supercapacitors does not primarily involve chemical reactions. Therefore, the number of charge and discharge cycles that a supercapacitor can survive through is substantially larger than that of batteries (Zakeri and Syri 2015). Despite their much

230

Chapter Eight

smaller energy density, supercapacitors can be charged and discharged at a higher rate compared to batteries (IRENA 2017). Therefore, they are a suitable solution for applications where frequent charges and discharges are required in which the time length of every charge and discharge is not too long.

SMES Superconducting magnetic energy storage devices store energy in a magnetic field. The energy stored in the magnetic field around a solenoid is ଵ given by ‫ܫܮ = ܧ‬ଶ . Based on this equation, the solenoid energy storage ଶ capacity can be increased by increasing two parameters, the inductance (‫)ܮ‬ of the solenoid and the current (‫ )ܫ‬flowing in it. The former has a linear relationship with the stored energy while the latter exhibits a quadrature relationship. To increase the inductance, three approaches can be taken, increase the length, cross-section area and/or permeability of the magnetic core of the solenoid. To increase the current in the solenoid, at a specific voltage, it is necessary to reduce its electrical resistance. To this end, superconducting magnetic energy storage devices employ superconducting materials. Superconducting materials are those that exhibit a sharp drop in resistance when the temperature is below a specific value, i.e., critical temperature. Superconducting magnetic energy storage is essentially a solenoid made of superconducting material. The solenoid is encapsulated in a vacuum cylinder where it is cooled down to a temperature below the critical temperature by using a cryocooler that is run using liquid helium or liquid nitrogen, depending on the critical temperature (liquid helium for up to 4K and liquid nitrogen for up to 77K). Fig. 8-8 schematically shows a superconductor magnetic energy storage system. When DC voltage is applied to the solenoid ends, current starts flowing in a path with near-zero resistance, allowing for a large amount of energy to be stored in the magnetic field of the solenoid. The near-zero resistance also means that the solenoid will have high efficiency and a low self-discharge rate. In fact, efficiency in the range of 95% can be achieved (Luo, et al. 2015) in superconducting magnetic energy storage devices. However, maintaining the superconductor in the superconducting state, i.e., cooling the superconductor, requires energy, hence, the overall efficiency of the superconducting magnetic energy storage is affected.

The Role of Energy Storage to Reduce Variability Impacts

231

Fig. 8-8: Schematic of superconducting magnetic energy storage (Mutarraf, et al. 2018)

Superconducting magnetic energy storage can provide a very fast response and high power rating (Chen, et al. 2009). In fact, the voltage tolerance of the superconducting material and the current rating of the AC/DC converters used to charge and discharge the storage device limits its power rating. Since it stores energy in the magnetic field and does not involve any moving parts or chemical reactions, it has a long lifetime in the range of 20 to 30 years (Zhu, et al. 2018) with a virtually infinite number of charge and discharge cycles. Not only is this a suitable solution for grid stabilization and pulsed power applications but it can also be used for canceling renewables’ power fluctuations (Panda and Penthia 2017, Ali, Wu and Dougal 2010). However, the main burden for the large deployment of gridscale superconducting magnetic energy storage is its high capital cost such that currently (IRENA 2017), these devices are not an economically viable solution for grid-scale storage applications. The main drivers for the capital cost of superconducting magnetic energy storage are superconductors (~30%) and power electronics (~60%) (Mukherjee and Rao 2019). With the future discovery of cheaper superconductors (particularly high temperature) and advancements in manufacturing these materials and the required power electronics, the capital cost of superconducting magnetic energy storage devices will decrease.

232

Chapter Eight

Mechanical Mechanical storage technologies are probably the most tangible storage solutions. The storage of energy with this technology occurs through the conversion of electric energy to kinetic or potential energy. Flywheels, compressed air energy storage, and pumped hydro storage are some of the well-known mechanical storage technologies.

Flywheel Flywheels store electric energy in the form of kinetic energy. The kinetic ଵ energy of a rotating mass is given by ‫ ݒ ݉ = ܭ‬ଶ . Based on this equation, ଶ there are two options for increasing the kinetic energy of the rotating mass, i.e., increasing the mass (݉) and increasing the speed (‫)ݒ‬. The former is equivalent to increasing the inertia and increases the stored energy linearly. The latter has a quadratic relationship with the stored energy. Therefore, increasing the speed of the rotation of the mass significantly increases its kinetic energy; flywheels work based on this concept. A schematic of a flywheel energy storage system is shown in Fig. 8-9. A flywheel energy storage system consists of a low mass rim made of carbon fiber or composite materials. The mass of the rim is optimized to ensure the satisfactory strength of the rim. Using forged light metal (e.g., aluminum), the rim is coupled to the rotor of an electrical machine (commonly a brushless permanent magnet machine) that acts both as a motor and a generator. In order to reduce friction and wear, the rim is levitated using a magnetic lift system (which consumes electrical energy). The setup is encapsulated in a sealed vacuum cylinder. The vacuum cylinder ensures that friction is minimized. It also provides protection for the setup against environmental factors that can deteriorate it, hence, the prolonged lifetime of the flywheel. When the flywheel is operating in charging mode, the electrical machine runs in motor mode to rotate the rim. When fully charged, the rim will rotate at extremely high speed in the range of 8000-16000 rpm (in commercial flywheels) and 100000 rpm (in experimental setups) (S. M. Mousavi, et al. 2017, Östergård 2011). At the time of discharging, the electrical power supply to the electrical machine is disconnected and the machine is run in generator mode converting the kinetic energy of the rim to electrical energy. During the discharge process, the rim slows down gradually and loses its kinetic energy.

The Role of Energy Storage to Reduce Variability Impacts

233

Fig. 8-9: Schematic of flywheel energy storage system (Soomro, et al. 2019)

The kinetic energy of the rim depends on the radius of the rim, its rotational speed, and mass. For example, storage capacity up to 36 kWh (per module) has been reported by (Beacon Power 2019). Several flywheel modules can be clustered to form a grid-scale capacity energy storage system. Flywheels have a very fast response time (instantaneous in the case of synchronous flywheels and approximately 10 ms in the case of non-synchronous flywheels (Engelmann, vor dem Esche and Tudi 2017)) and relatively high rated power which can be quickly ramped (~190 kW/s). The round trip efficiency of flywheels is in the range of 85% (Beacon Power 2019). However, their self-discharge rate is very high, particularly if mechanical bearings are used (Kalaiselvam and Parameshwaran 2014). Due to the minimized mechanical friction and the fact that exposure of the components to a deteriorating environment is limited in flywheels, this energy storage equipment is durable and can withstand a significantly large number of cycles (൐ 100000) (IRENA 2017). Accordingly, flywheels are a promising solution for frequency and voltage regulation, and ride through, smoothing applications at the grid level. The use of flywheels for grid-scale energy storage is a relatively new approach. Currently, the capital cost of flywheels is very high per kWh of

234

Chapter Eight

capacity installed. However, it is expected that in the future the cost of flywheel projects will reduce (IRENA 2017). This is mainly due to technology maturing, advancement in bearings, production of superior material for flywheel rims, and development of better machines (IRENA 2017). Moreover, with the ever-growing penetration of renewables, flywheels are prone to widespread employment in the power system. Consequently, economy of scale will make flywheels more economically accessible.

Compressed Air Storage Compressed air storage plants store electric energy by compressing air and storing it in a reservoir. A compressed air storage power plant consists of three main components, the compressor and generator, the compressed air reservoir, and the cooling and heating mechanism. Fig. 8-10 schematically shows a compressed air energy storage system. A large volume of cleaned and dried air is compressed using abundant/cheap electrical energy. The compression of air reduces its volume; however, substantial heat is produced during the compression process (Abdin and Khalilpour 2019). The heat needs to be removed before storing the high-pressure air. Heat exchangers are used to extract the heat. In diabatic compressed air storage plants, the extracted heat can either be released in the environment or used for heating applications. In adiabatic compressed air storage plants, the extracted heat is stored in highly efficient storage tanks for later use in the discharge process of the stored air (Abdin and Khalilpour 2019). Once the heat is removed, the high-pressure air is stored in a reservoir, e.g., salt caverns and depleted gas reservoirs (or air cylinders and tanks (Liu, et al. 2014)). The reservoir can store high-pressure air for a long time without a major drop in pressure or leakage (Salameh 2014). When electrical energy is expensive/scarce, the compressed air storage plant can be discharged to generate electricity. During the discharge process, the high-pressure air is released from the reservoir. The drop in pressure of the air leaving the reservoir leads to an increase in the air speed and a decrease in its temperature. The low temperature/frozen air can damage the compressed air energy storage plant. Therefore, it is necessary to reheat the released air. In diabatic compressed air storage plants, reheating occurs in a combustion chamber where fossil fuels are burnt to produce heat. Recovered heat from an adjacent gas turbine may also be used in diabatic plants. However, in adiabatic compressed air storage plants, the stored heat during the air compression step is used in heat exchangers to reheat the air released from the reservoir. The high-speed heated air is put through a turbine that is

The Role of Energy Storage to Reduce Variability Impacts

235

coupled to a synchronous generator. The rotation of the turbine and subsequently the generator produces electricity.

Fig. 8-10: Schematic of an adiabatic compressed air energy storage (He, Luo, et al. 2017)

Diabatic compressed air storage plants have a round trip efficiency range of less than 45% (adiabatic CAES can reach up to 70% efficiency) which is considerably lower than that of batteries (Elmegaard and Brix 2011). Despite their lower efficiency, they have a longer lifetime with a significantly larger number of charge and discharge cycles compared to batteries (Zakeri and Syri 2015). Employing the well-understood synchronous generators in compressed air storage plants means that these plants can inherently contribute to the power system inertia and synchronizing and damping torque. These are particularly important for arresting frequency excursions and limit the rate of change of frequency in trip events (Bakhtvar, Vittal, et al. 2017). Moreover, they can support voltage in the power system. Accordingly, compressed air storage plants can improve the stability and reliability of the power system (He and Wang 2018). Currently, there are two grid-scale compressed air storage plants sized 290 MW and 110 MW that are operational in Germany and the USA, respectively.

236

Chapter Eight

Since the most recent grid-scale plant dates to the early 1990s, estimating the cost of a compressed air storage plant is challenging. The international renewable energy agency (IRENA) reports a capital cost in the range of 55 USD/kWh for new installations and estimates it will reduce to approximately 45 USD/kWh by the end of the next decade (IRENA 2017).

Pumped Hydro Storage Pumped hydro storage power plants are one of the oldest grid-scale power storage solutions. These power plants store electric energy in the form of potential energy. The principle behind pumped hydro storage power plants is simple. Water is pumped to a higher altitude and later released downwards when the energy is needed. The potential energy of the mass of water (݉) pumped to a specific height (݄) is given by ܷ = ݄݉݃. Therefore, the volume of water and the height to which the water is pumped equally affect the stored energy. Pumped hydro storage power plants consist of two reservoirs with a significant height difference. The two reservoirs are connected by pipes or channels named the penstock. The penstock allows for the transfer of water between the two reservoirs. Pumped hydro storage power plants include a powerhouse at the lower reservoir (sea, river, or lake) where electrical machines and turbines are housed. A schematic of a pumped hydro storage plant is depicted in Fig. 8-11. At times when electricity is abundant/cheap, an electrical motor consumes electrical energy to run the turbine that is coupled to it in pumping mode. The water is pumped from the lower reservoir to the higher reservoir; therefore, its potential energy is increased. Inversely, when electrical energy is expensive/scarce, the water from the higher reservoir is released towards the lower reservoir through the provided gates and water flows down the penstock. The potential energy of water is converted to kinetic energy as it loses height by moving down the penstock. In the powerhouse, the water coming through the penstock hits the blades of the turbine and makes it rotate. The rotating turbine makes a generator rotate and produce electrical power. The motor and generator used in the pumped hydro storage power plant could be either dedicated machines or reversible electrical machines. Newer designs of pumped hydro storage power plants employ adjustable speed machines which run as an induction machine (asynchronous) in motoring mode and a synchronous machine in generating mode (Mohanpurkar, et al. 2018, Donalek 1995). Therefore, they provide a wide range of not only the discharging rate, but also the charging rate.

The Role of Energy Storage to Reduce Variability Impacts

237

Fig. 8-11: Schematic of the pumped hydro storage plant (Luo, et al. 2015)

Pumped hydro storage power plants are mature in technology. They were traditionally used to store energy produced by baseload generators at night time and release the stored energy at peak demand intervals (with an efficiency of ~80%) to reduce the necessity for operating expensive peaker generators (Hafemeister 2013, Rehman, Al-Hadhrami and Alam 2015). Pumped hydro storage power plants can also be employed for supporting system frequency and contributing to the inertia and synchronizing and damping torque (Dursun and Alboyaci 2010). Despite the fact that pumped hydro storage power plants are one of the cheapest solutions for grid-scale electric energy storage, they have specific geographical requirements; this is a burden for the construction of new conventional pumped hydro storage power plants. In countries where the elevation change is generally low, it might be harder to find a suitable site for pumped hydro storage power plants. Moreover, the need for large reservoirs means that they bring environmental concerns (Immendoerfer, et al. 2017). All in all, the total installed pumped hydro storage power plant capacity globally is still expected to continue rising. Particularly in China, 50 GW of new pumped hydro storage capacity is estimated to be installed by 2030 (since 2018) (Rogner and Troja 2018).

Chemical Chemical electric storage refers to the production of a substance through chemical reactions that often require a substantial amount of energy provided through electric power. The newly produced substance can be later used to reverse the chemical reaction (for example in a burner or fuel cell) during which energy is released. The production of hydrogen (Breeze 2018),

Chapter Eight

238

methanol, and natural gas is an example of chemical electric energy storage. In this chapter, the power-to-gas (carbon methanation) concept is described as one of the promising storage solutions. The production of methanol also follows a similar concept to natural gas production.

Power-to-gas and Methanol Production The power-to-gas storage system is one of the promising options for storing electrical energy. The power-to-gas technology provides both spatial and temporal balancing solutions for the high penetration of renewable energy into the power system. This energy storage system, essentially links the gas network and the power system to each other by the conversion of electrical energy into gas through two major steps: -

Water electrolysis; and The Sabatier reaction.

Fig. 8-12 shows a schematic diagram of a power-to-gas storage system. The hydrogen (H2) consumed in a power-to-gas storage system is produced by the electrolysis of water (Breeze 2018). The produced hydrogen in the previous step is then mixed with carbon dioxide (CO2). By inputting the heat produced by electric energy to this mixture, the so-called Sabatier reaction is triggered, yielding methane (CH4). As implied, renewable energy can be used to produce natural gas using a power-to-gas storage system. The produced natural gas may then be injected into the gas network where it can be stored (by increasing the pressure) and transferred in bulk over a long distance (with the aid of gas pressure stations). The generated gas is a versatile and sustainable energy carrier that can be used for cooling and heating purposes or reconversion to electricity. Power-to-gas storage facilities are the largest providers of energy storage capacity that have enhanced harvesting renewables in Germany (Jentsch, Trost and Sterner 2014).

The Role of Energy Storage to Reduce Variability Impacts

239

Fig. 8-12: Schematic of the power-to-gas storage system (Liu, Sun and Harrison 2019)

Thermal Thermal storage systems are another type of energy storage system that enables storing the electrical energy by cooling or heating a relatively high heat capacity material known as the storage medium. There are three different types of thermal energy storage systems: -

Sensible heat storage; Latent heat storage; and Thermo-chemical heat storage.

In sensible heat storage systems, the thermal energy is stored by cooling or heating a solid or liquid storage medium such as rocks, molten salts, sand, ceramic, and water (Cabeza, et al. 2015). The cheapest option for the storage medium in thermal storage systems is water. In latent heat storage systems

240

Chapter Eight

phase change materials (PCMs) are used to store heat. And finally, in thermo-chemical heat storage systems, chemical reactions are involved for the storage of heat (IRENA 2013). The stored thermal energy in a thermal energy storage system can be used for satisfying cooling and/or heating demands as well as generating electric power. Thermal energy storage systems can be particularly utilized in industrial processes and buildings. Statistics show that almost half of the energy consumed in Europe corresponds to the cooling and heating of buildings and industry. Thus, thermal energy storage can help to shift the energy demand and supply cooling or heating energy requirements using the stored thermal energy on a daily, weekly, or monthly basis (depending on the thermal energy storage technology employed). The time shift of the electrical energy demand for heating and cooling achieved by thermal energy storage can help to reduce the energy cost, energy consumption, peak demand, and CO2 emissions. Moreover, instead of curtailing (due to technical constraints or simply the absence of sufficient demand), it is possible to use the excess renewable energy to produce heat and store it in a thermal energy storage facility; hence, the share of renewables will be increased in the energy basket (IRENA 2013).

Role of Storage Electrical energy storage is often known as the “holy grail” of the power grid. Electrical energy storage is a critical solution for the aging power grid and an effective tool for realizing very high penetration of renewable energy into the power grid (for example 100%). It can act as a link between the needs of end-users and the electric utility (IstvaĔ Táczi 2016, O’Dwyer, Ryan and Flynn 2017). Energy storage systems can be integrated at various levels in the power system. The integration of electrical energy storage into the power grid is shown in Fig. 8-13. As it is shown in this figure electrical energy storage has various potential applications.

The Role of Energy Storage to Reduce Variability Impacts

241

Fig. 8-13: Electrical storage system integration into the power grid (adapted from (IRENA 2015))

Electrical energy storage facilitates high penetration of renewable energy, reduction of electricity cost, participation in demand response programs, and peak demand shifting, and enhances the power quality and reliability of the power grid. The benefits of electrical storage to the power systems are broad and cover from the power generation level, transmission, distribution, energy services, and renewable energy down to the end-user level. Some of the important applications of electrical energy storage for power systems are depicted in Fig. 8-14. As shown in Fig. 8-13, storage devices appear at both the solar PV system (distributed) and grid level (centralized). However, they are used for different purposes. The adoption of storage devices at the solar PV system level is a long-standing solution in standalone solar PV systems that seek to deliver electricity to satisfy the local demand even at times when sunlight is not available, e.g., nighttime. Storage devices can be spotted in gridconnected solar-PV systems as well. In such systems, the storage device is mainly used for two different purposes: energy arbitrage and minimizing the dependence on the electricity grid. The latter seeks to reduce the need for importing electricity from the grid. To this end, the excess solar power is stored and later released when solar power is not sufficient to meet the local demand. However, energy arbitrage is a more sophisticated task. It is particularly interesting in power systems in which the difference between the feed-in-tariff and the electricity price is variable. In such power systems,

242

Chapter Eight

Fig. 8-14: Energy storage system application in power systems (adapted from (Chen, et al. 2009))

solar energy can be stored when the feed-in-tariff is lower than the electricity price and released when the electricity price is higher than the solar energy feed-in-tariff and/or when sunlight is not available. Compared to the solar PV system level, storage devices used at the grid level are generally of larger energy capacity and charge and discharge rate compared to that of the solar PV system level storage solutions. Their role could

The Role of Energy Storage to Reduce Variability Impacts

243

involve responding to the variations of the net demand in the network and/or satisfying the peak demand. Energy storage can help to reduce the impacts of variable solar PV generation on the power system. Many benefits can be listed for energy storage, in this regard. The Electric Power Research Institute (EPRI) summarizes the benefits of energy storage systems for renewable energy integration under 5 titles (EPRI 2004). These are further discussed.

Transmission (and distribution) curtailment Grid-scale (multi-megawatt) solar PV power plants may be in remote areas far from the demand centers. Often the capacity of the transmission lines for transferring the power is limited or the transmission lines are simply not available in these locations. Moreover, large solar PV generation capacity may be embedded in the distribution system where the distribution lines and transformers were not originally sized for handling solar PV generation. Transferring the power injected by solar PV systems may require upgrading the existing transmission and distribution lines and transformers and/or building new power transfer infrastructure. This will ensure that the equipment limits are observed even when these power sources inject power at the rated value. However, solar PV generation only reaches its maximum value for a short period during a year. For instance, in the year 2006, for 56 utility and distribution scale solar PV systems in Texas, the average number of time instances (5 minute resolution sampling) at which solar PV system power output was greater than the 90% of installed capacity was 167 (out of 105,120 time instances) (National Renewable Energy Laboratory n.d.). Therefore, it might not be economically viable to make such large investments to upgrade the grid. Similarly, curtailing the solar generation will affect its cost-efficiency. A benefit of energy storage is to reduce the need for increasing the network capacity. This is achieved by storing solar power during periods in which sufficient transfer capacity is not available. Once transfer capacity becomes available, the energy storage system can be discharged. Accordingly, the need for investment in power transfer infrastructure can be minimized without having to curtail solar PV generation.

Time-shifting Time-shifting and load following are possibly the most obvious benefits of energy storage to solar PV generation. Since producing electric power by

244

Chapter Eight

solar PV is dependent on the solar irradiance that the PV cells receive, the power produced by these sources is significantly reduced under overcast and dusty conditions and is zero at nighttime. Energy storage systems can store solar power when it is available and release it when solar generation is not available or simply insufficient to fulfill the local demand. This is essential to pledge continuity of supply, particularly for standalone systems where connection to the grid is not available. Clearly, the energy storage system should be sized and operated optimally to enable fulfilling the demand and maximize the benefit gained from the solar PV generation in place and the energy storage system. It is worth mentioning that energy storage systems can also be used for energy arbitrage in grid-connected systems to further justify their capital cost. Accordingly, energy will be stored at the times of low electricity price and released when the electricity price is high.

Forecast Hedging Advances in forecasting techniques have improved the accuracy of the renewables’ forecast and substantially reduced the forecasting error. However, perfect forecasting has not been achieved yet and uncertainty of renewables remains. This means that the solar PV generation may be more/less than the forecast value used by the system operator for carrying out the unit commitment and economic dispatch. Energy storage can help in managing the excess or deficit power by charging when solar generation is higher than the forecast value and discharging when it is lower. Hence, power adequacy can be maintained despite the uncertainty of solar PV generation.

Frequency Support Frequency events are those that involve a fast decrease/increase in system frequency. Frequency events can lead to a blackout if not managed properly and sufficiently quickly. Fast injection/consumption of power is key for arresting the frequency excursion. Based on the swing equation, (1), any change in the power balance is transferred to the rotor speed of synchronous machines in the power system and consequently the frequency at the center of inertia (Bakhtvar, Vittal, et al. 2017). 2‫ ݀ ܪ‬ଶ ߜ = ܶ௠ െ ܶ௘ ߱ோ ݀‫ ݐ‬ଶ

(1)

The Role of Energy Storage to Reduce Variability Impacts

245

The connection of conventional generators to the grid is stiff and allows them to inherently react to the changes in the frequency based on their inertia and synchronizing torque coefficient. In contrast, solar PV generation does not involve a rotating mass and is also decoupled from the power grid by the power-electronic-based converters. This means that solar PV generation does not contribute to the inertia in the power system. Although virtual inertia can theoretically be provided by solar PV generation using its power-electronic-based converters, it is not a practical solution as it essentially requires a rigorous control mechanism and operation at points lower than the maximum power point (MPP). A benefit of energy storage systems is their capability to use the stored energy and available capacity for providing both actual and virtual inertia over short and long time durations (depending on the technology) (Ortega Manjavacas 2017). By monitoring the system frequency variations, storage devices can respond to the variations of the net demand. Such monitoring can be achieved either by directly measuring the local frequency in the power system or through the inherent response of storage devices. Accordingly, the flexibility provided by the storage devices can support frequency and reduce the need for a conventional spinning reserve and must run unit (due to inertia) constraints.

Fluctuation suppression The power output of solar PV generation is intermittent. Not only do power system operators require the renewable power injected to the grid to be stabilized but also local consumption of solar generation mandates it to be smoothened. One of the benefits of energy storage systems is their use for smoothing the power output of solar PV generation. Accordingly, the energy storage will need to constantly change between the charge and discharge mode to cancel out the power gradients that are larger than a certain range. Such an energy storage system should withstand a significantly large number of cycles during the lifetime of the solar PV power plant (Li, et al. 2017, Wang, Ciobotaru and Agelidis 2014). The effect of smoothing of solar PV output also manifests in the solar PV power plant behavior during transient events and can potentially improve system stability (Pothisoonthorn and Ngamroo 2016).

Applications of Electric Energy Storage As explained in earlier sections, electric storage systems come with different technologies and forms of storage. All of them are beneficial to the

246

Chapter Eight

power system; however, one that is suitable to a specific application might not suit other applications. This stems from the fact that the capabilities of all storage solutions are not equal, moreover, some technologies might only be economically practical for certain applications. Therefore, a range of storage solutions may be required for power systems to alleviate negative impacts of the variable solar PV generation. Error! Reference source not found. Fig. 8-15 shows the typical power rating and energy capacity for various energy storage technologies. In addition to the power rating and capacity, the response time of the electrical storage system is a major player in defining its suitability for a specific application.

Fig. 8-15: Positioning of Energy Storage Technologies (Akhil, et al. 2016)

Table 8-1 shows the typical general properties, including the response time and lifetime cycles, of several electric energy storage systems. It is evident that a technology such as pumped hydro storage has a totally different application to that of flywheels. Considering the response time, energy capacity, lifetime cycles, power rating and energy capacity, in general, three areas of application are defined for electrical energy storage systems, i.e., power quality, grid support and load shifting and bulk power management. Power quality requires a fast response (from several milliseconds to a few seconds) for several minutes. Grid support usually needs a fast and large response to ensure the stability and reliability of the power system for the short duration of abnormal events (a few seconds to several minutes). For

The Role of Energy Storage to Reduce Variability Impacts

247

load shifting a large capacity with an adequate power rating is necessary. Usually, response times longer than that needed for grid support can be tolerated for load shifting. Bulk energy management, as the name implies, incorporates a large energy capacity and power rating. It is a type of storage that can be scheduled several minutes to hours ahead and therefore its response time does not necessarily need to be very fast. Error! Reference source not found. Fig. 8-15 shows the storage technologies that can be used for each of these applications. Table 8-1: General properties of electrical energy storage technologies (adapted from (IRENA 2017, Abdi, et al. 2017, Aneke and Wang 2016))

248

Chapter Eight

Dispatchable Solar PV Fig. 8-16 shows the measured solar irradiance on a tilted surface in Varennes, Quebec, Canada, for two different days. The soar irradiance might not be quite smooth during the day and may fluctuate frequently. As mentioned in earlier sections, due to the intermittency of the solar irradiance, the maximum power that can be produced at each instance is variable. Therefore, despite the theoretic ability of the power-electronicbased converters, the power output of a solar PV plant is inherently nondispatchable. The intermittency of solar generation can be partly alleviated with the aid of aggregation. The aggregation of solar PV generation refers to considering the output of several solar PV generation units that are geographically dispersed and connected at various locations in the grid as one single power plant. Fig. 8-17 conceptually, compares the one day power output of a 1 MW solar PV power plant with the aggregated power output of twenty 0.05 MW solar PV power plants distributed strategically. As noted, aggregation greatly smoothens the power output of solar PV generation, however, the fluctuations remain, hence the non-dispatchability aspect of solar PV generation.

The Role of Energy Storage to Reduce Variability Impacts

Tilted Irradiance (W/m2)

1400

249

Date: 26/2/2015

1050

700

350

0 0:00

6:00

12:00

18:00

0:00

18:00

0:00

Time

Tilted Irradiance (W/m2)

1400

Date: 30/12/2014

1050

700

350

0 0:00

6:00

12:00

Time Fig. 8-16: Tilted solar irradiance in Varennes, Quebec, Canada (adapted from (Natural Resources Canada 2016))

250

Chapter Eight

Fig. 8-17: Effect of aggregation on solar PV generation (adapted from (Placer 2016))

The energy storage solution can be employed for smoothing the power output of solar PV generation and alleviating its non-dispatchability. A dispatchable solar PV power plant can guarantee to maintain a specific power output level over a certain length of time. Such a power plant can participate in the interval-ahead electricity market in a similar fashion to conventional units and get dispatched. This means, the power plant operator will need to bid in the market, several dispatch intervals ahead of occurrence. The power plant operator is bound to realize any committed power, otherwise it will be penalized based on market rules and/or agreements with the system operator. Dispatchable solar PV generation not only reduces the requirement for flexibility and reserve, compared to the current practice of renewable generation, but can also participate in the reserve market using part of the stored energy and any uncommitted power. This is a major leap towards an even higher penetration of renewables in power systems since it directly targets one of the main disadvantages of renewables. Provision of dispatchable solar PV generation is a complex task. Unlike conventional non-dispatchable solar PV generation which is relatively simple, dispatchable solar PV generation requires interaction and coordination between several components. Fig. 8-18 shows a conceptual architecture for a dispatchable solar PV power plant. The roles of forecasting and prediction, the energy storage system, dispatching and control components are essential for achieving dispatchable solar PV generation. These are briefly discussed.

The Role of Energy Storage to Reduce Variability Impacts

-

251

Forecasting and prediction: forecasting plays a major role in the dispatchable solar PV power plant. Such a tool should be in place to forecast parameters such as solar irradiance, temperature and clouding for the several hours ahead. The forecast window should slide forward as the time passes. This will ensure that the forecast is constantly updated based on the latest available data. The forecast parameters along with other important factors such as dusting need to be processed using a solar PV generation model to yield the estimated maximum power output of the power plant at each instance for several dispatch intervals ahead. It is evident that the output of forecasting and prediction should be the lower confidence limit of the forecast at a sufficient confidence level to enable reliable dispatchable solar PV generation.

Fig. 8-18: Dispatchable solar PV power plant components

-

Energy storage system: storage is the main enabler of dispatchable solar PV generation. Two types of storage are conceived, a storage device whose main role is to provide energy storage (storage device 1) and a storage device which is aimed for meeting large power setpoints for short durations (storage device 2). The former, must be capable of consuming and supplying power for the duration of the

Chapter Eight

252

-

-

dispatch intervals and based on the set-points. The latter must change its output fast and frequently to cancel the deviations of power from the committed value. In that regard it should be able to handle many charge and discharge cycles before it needs to be replaced. Dispatching: the dispatching unit is essentially a high-level control mechanism of the dispatchable solar PV generation. The predicted maximum power output and state of the charge and health of the energy storage system are input to the dispatching unit. Appropriate set-points are found for the solar PV and the energy storage device. In addition to the equipment and operational constraints, the main consideration of the dispatching unit is to ensure that the set-points found lead to a constant power output from the dispatchable solar PV power plant for the duration of the dispatch intervals ahead. The setpoints shall be used by the power plant operator to bid in the electricity market. The market decisions are then forwarded to the control unit for implementation. The solar PV power plant is committed to adhere to the market decisions. It is evident that the committed power is less or equal to the power plant operator’s bid in the electricity market. Control: the low-level control of the solar PV system and the energy storage system is accomplished by the control unit. It essentially deals with the real-time operation of the dispatchable solar PV power plant. The control unit implements the set-points provided by the dispatching unit (based on market decisions). It monitors the power output of the power plant and if there are any deviations from the committed power, in the first place it tries to compensate by using the storage device that is aimed for a large power output over short durations, i.e., storage device 2. The set-point of the other energy storage device, i.e., storage device 1, is only overridden when the power output of storage device 2 is not sufficient to compensate the power deficiency due to solar PV generation. However, there might still be rare cases where the energy storage system may not be able to fully compensate for the deficiencies (Teleke, et al. 2010). In general, the control unit should minimize any deviations in the state of charge of storage device 1 from the expected value found by the dispatching unit for the end of the current dispatch interval. The control unit should also facilitate direct instructions by the system operator to ensure the controllability of the power plant, particularly under abnormal conditions.

The Role of Energy Storage to Reduce Variability Impacts

253

To demonstrate the concept of a dispatchable solar PV power plant, a hypothetical dispatchable solar PV power plant consisting of 30 MW solar PV and a 5 MWh battery energy storage system (BESS) with a 2C charge and discharge rate was simulated using 10 minute resolution historic solar generation data of a solar power plant in Texas, USA (National Renewable Energy Laboratory n.d.) instead of a power forecast. It was assumed that this power plant is grid connected and will participate in the intraday continuous electricity market. Therefore, it should maintain its output constant for the duration of every settlement period (30 minutes in the simulated case). Fig. 8-19 shows the power output of the solar PV and battery energy storage system (BESS) for 2 different days, i.e., high and low variability solar irradiance days. The negative values in this figure correspond to the consumption of power. Comparing these two days in the high variability day, BESS’ storage capability is used more to compensate for the variation in the power output from solar PV. The resulting net power output of the dispatchable solar PV power plant in the two days of interest is illustrated in Fig. 8-20. As noted, using BESS, the dispatchable solar PV power plant can maintain its power output constant for the required 30 minutes. This enables the power plant to participate in the intraday continuous electricity market and get dispatched.

Chapter Eight

254

High Variability

Power (MW)

30 20 10 0 -10 0

24

48 72 96 120 Time Step (10 Minutes) BESS

144

Solar PV

Low Variability 30

Power (MW)

20

10

0

-10 0

24

48

72

96

120

Time Step (10 Minutes) BESS

Solar PV

Fig. 8-19: Solar PV and ESS power output for two days

144

The Role of Energy Storage to Reduce Variability Impacts

High Variability

Power (MW)

30

20

10

0 0

24

48

72

96

120

144

120

144

Time Step (10 Minutes) Low Variability

Power (MW)

30

20

10

0 0

24

48

72

96

Time Step (10 Minutes) Fig. 8-20: Dispatchable Solar PV power plant output for two days

255

256

Chapter Eight

Conclusion The integration of solar PV generation and in general variable generation introduces new challenges to the power system. The intermittency of solar PV generation makes it an inherently non-dispatchable power source. In the first place, inertia and frequency regulation capability are absent from the solar PV generation plants which means that these power supply units do not support the power system frequency. In the second place, the uncertainty of solar generation adds to the requirement for flexibility and spinning reserve. Moreover, the continuity of the power supply despite the daily and seasonal variations of solar PV generation mandate oversizing or backup power supply to be in place. This chapter focused on the role of energy storage systems in reducing the negative impacts of PV systems on power systems. The basics of operation of various energy storage technologies that fall under the categories of mechanical, electrical, electrochemical, chemical, and thermal were discussed. An overview of the benefits of energy storage solutions in reducing the negative impacts of solar PV generation was given. It was mentioned that the main applications of energy storage solutions are power quality, grid support, load shifting and bulk power management. While all storage solutions are beneficial for suppressing solar PV generation impacts, not every storage technology is suitable for all applications. The suitability of each technology for different applications was suggested based on its energy capacity, response time, power rating and lifetime cycles. It is concluded that when it comes to reducing the negative impacts of solar PV generation, there is no silver bullet. In fact, several storage solutions need to be used. In this regard, the cost of each storage technology should be considered so that the economic viability of solar PV generation is not jeopardized. For example, high-power supercapacitors could be used for suppressing intermittency, batteries for achieving load following (and constant power), frequency regulation and time-shifting on a small scale and pumped hydro storage can be utilized for time-shifting on a large scale. Moreover, the concept of dispatchable PV-energy storage hybrid systems was discussed and the roles of each component in such a system were articulated in this chapter. With the aid of a simulation of a hypothetical dispatchable solar PV power plant, it was shown that using BESS enables the dispatchable solar PV power plant to maintain its power output constant for the required duration (30 minutes in this chapter). Hence, the dispatchable solar PV power plant can get dispatched by the system operator. Dispatchable solar PV generation is a major leap towards an even higher penetration of renewables in power systems as it reduces the

The Role of Energy Storage to Reduce Variability Impacts

257

requirement for flexibility and reserve, compared to the current practice of renewable generation.

References Abdi, Hamdi, Behnam Mohammadi-ivatloo, Saeid Javadi, Amir Reza Khodaei, and Ehsan Dehnavi. 2017. “Energy Storage Systems.” In Distributed Generation Systems: Design, Operation and Grid Integration, edited by G. Gharehpetian and M. Mousavi Agah, 333-367. Oxford, UK: Butterworth-Heinemann. Abdin, Zainul, and Kaveh Rajab Khalilpour. 2019. “Single and Polystorage Technologies for Renewable-Based Hybrid Energy Systems.” In Polygeneration with Polystorage for Chemical and Energy Hubs, edited by Kaveh Rajab Khalilpour, 77-131. Academic Press. Afzal Ahmadi, Ali. 2009. General Chemistry. Nay Publisher. Akhil, A., G. Huff, A. Currier, B. Kaun, D. Rastler, S. Chen, A. Cotter, D. Bradshaw, and W. Gauntlett. 2016. “DOE/EPRI Electricity Storage Handbook in Collaboration with NRECA.” Tech. rep., Sandia National Laboratories. Ali, M. H., B. Wu, and R. A. Dougal. 2010. “An Overview of SMES Applications in Power and Energy Systems.” IEEE Transactions on Sustainable Energy 1, 38-47. Ander, Eider Goikolea, Jon Andoni Barrena, and Roman Mysyk. 2016. “Review on supercapacitors: Technologies and materials.” Renewable and Sustainable Energy Reviews 58, 1189-1206. Aneke, Mathew, and Meihong Wang. 2016. “Energy storage technologies and real life applications – A state of the art review.” Applied Energy 179, 350-377. Bakhtvar, M. 2016. “Power System Planning for High Renewables Penetration: Voltage Stability and System Frequency Aspects.” Ph.D. dissertation, Ireland: University College Dublin. Bakhtvar, M., C. A. Cabrera, G. Buttitta, O. Neu, and A. Keane. 2017. “A study of operation strategy of small scale heat storage devices in residential distribution feeders.” 2017 IEEE PES Innovative Smart Grid Technologies Conference Europe (ISGT-Europe), 1-6. Bakhtvar, M., E. Vittal, K. Zheng, and A. Keane. 2017. “Synchronizing Torque Impacts on Rotor Speed in Power Systems.” IEEE Transactions on Power Systems 32, 1927-1935. Beacon Power. 2019. Beacon Power Flywheel Energy Storage Systems. Accessed July 9, 2019. https://beaconpower.com.

258

Chapter Eight

Breeze, Paul. 2018. Power System Energy Storage Technologies. Academic Press. Cabeza, L. F., I. Martorell, L. Miró, A. I. Fernández, and C. Barreneche. 2015. “Introduction to thermal energy storage (TES) systems.” In Advances in Thermal Energy Storage Systems, edited by Luisa F. Cabeza, 1-28. Woodhead Publishing. Chen, Haisheng, Thang Ngoc Cong, Wei Yang, Chunqing Tan, Yongliang Li, and Yulong Ding. 2009. “Progress in electrical energy storage system: A critical review.” Progress in Natural Science 19, 291-312. Cheng, Meng, Saif Sabah Sami, and Jianzhong Wu. 2017. “Benefits of using virtual energy storage system for power system frequency response.” Applied Energy 194, 376-385. Cochran, Jaquelin, Mackay Miller, Owen Zinaman, Michael Milligan, Doug Arent, Bryan Palmintier, Mark O’Malley, et al. 2014. “Flexibility in 21st century power systems.” Tech. rep., National Renewable Energy Lab. (NREL), Golden, CO (United States). Dai, K., X. Wang, Y. Yin, C. Hao, and Z. You. 2016. “Voltage Fluctuation in a Supercapacitor During a High-g Impact.” Scientific Reports 6, 1-10. Deng, Da. 2015. “Li-ion batteries: basics, progress, and challenges.” Energy Science & Engineering 3, 385-418. Doherty, R., and M. O’Malley. 2005. “A new approach to quantify reserve demand in systems with significant installed wind capacity.” IEEE Transactions on Power Systems 20, 587-595. Donalek, P. 1995. “Application of adjustable-speed machines in conventional and pumped-storage hydro projects.” Tech. rep., Electric Power Research Institute. Dursun, Bahtiyar, and Bora Alboyaci. 2010. “The contribution of windhydro pumped storage systems in meeting Turkey’s electric energy demand.” Renewable and Sustainable Energy Reviews 14, 1979-1988. Elmegaard, Brian, and Wiebke Brix. 2011. “Efficiency of Compressed Air Energy Storage.” 24th International Conference on Efficiency, Cost, Optimization, Simulation and Environmental Impact of Energy Systems, 1-12. Engelmann, T., R. vor dem Esche, and R. Tudi. 2017. “Fast Response Flywheel Energy Storage Technology for Virtual Power Plants and Microgrids.” 2017 Electrical Energy Storage Applications and Technologies, 1-11. EPRI. 2004. “EPRI-DOE handbook supplement of energy storage for grid connected wind generation applications.” Tech. rep., Electric Power Research Institute and U.S. Department of Energy.

The Role of Energy Storage to Reduce Variability Impacts

259

Gaffney, Jeffrey S., and Nancy A. Marley. 2018. General Chemistry for Engineers. Elsevier. Galushkin, N., N. Yazvinskaya, and D. Galushkin. 2018. “Mechanism of Thermal Runaway in Lithium-Ion Cells.” Journal of the Electrochemical Society 165: A1303-A1308. Habibi, M. S. 2001. “Model for impact of storage on spinning reserve requirements and distributed generation.” Proceedings of the 33rd Southeastern Symposium on System Theory (Cat. No.01EX460), 161165. Hafemeister, David. 2013. Physics of societal issues: calculations on national security, environment, and energy. Springer Science & Business Media. He, Wei, and Jihong Wang. 2018. “Optimal selection of air expansion machine in Compressed Air Energy Storage: A review.” Renewable and Sustainable Energy Reviews 87, 77-95. He, Wei, Xing Luo, David Evans, Jonathan Busby, Seamus Garvey, Daniel Parkes, and Jihong Wang. 2017. “Exergy storage of compressed air in cavern and cavern volume estimation of the large-scale compressed air energy storage system.” Applied Energy 208, 745-757. Immendoerfer, Andrea, Ingela Tietze, Heidi Hottenroth, and Tobias Viere. 2017. “Life-cycle impacts of pumped hydropower storage and battery storage.” International Journal of Energy and Environmental Engineering 8, 231-245. IRENA. 2017. “Electricity Storage and Renewables: Costs and Markets to 2030.” Tech. rep., International Renewable Energy Agency. IRENA. 2015. “Renewables and Electricity Storage: A Technology Roadmap for REmap 2030.” Tech. rep., International Renewable Energy Agency. IRENA. 2013. “Thermal Energy Storage: Technology Brief.” Tech. rep., International Renewable Energy Agency. IstvaĔ Táczi, Gábor Szörényi. 2016. “Overview of the Energy Storage Possibilities to Support the Electrical Power System.” Tech. rep., Energy Regulators Regional Association (ERRA). Jentsch, Mareike, Tobias Trost, and Michael Sterner. 2014. “Optimal Use of Power-to-Gas Energy Storage Systems in an 85% Renewable Energy Scenario.” Energy Procedia 46, 254-261. Kalaiselvam, S., and R. Parameshwaran. 2014. “Energy Storage.” In Thermal Energy Storage Technologies for Sustainability, edited by S. Kalaiselvam and R. Parameshwaran, 21-56. Boston: Academic Press. Kim, Joo Gon, Byungrak Son, Santanu Mukherjee, Nicholas Schuppert, Alex Bates, Osung Kwon, Moon Jong Choi, Hyun Yeol Chung, and Sam

260

Chapter Eight

Park. 2015. “A review of lithium and non-lithium based solid state batteries.” Journal of Power Sources 282, 299-322. Lannoye, E., D. Flynn, and M. O’Malley. 2015. “Transmission, Variable Generation, and Power System Flexibility.” IEEE Transactions on Power Systems 30, 57-66. Li, P., R. Dargaville, Y. Cao, D. Li, and J. Xia. 2017. “Storage Aided System Property Enhancing and Hybrid Robust Smoothing for LargeScale PV Systems.” IEEE Transactions on Smart Grid 8, 2871-2879. Liu, Jinchao, Xinjing Zhang, Yujie Xu, Zongyan Chen, Haisheng Chen, and Chunqing Tan. 2014. “Economic analysis of using above ground gas storage devices for compressed air energy storage system.” Journal of Thermal Science 23, 535-543. Liu, Jing, Wei Sun, and Gareth P. Harrison. 2019. “Optimal Low-Carbon Economic Environmental Dispatch of Hybrid Electricity-Natural Gas Energy Systems Considering P2G.” Energies 12(7), 1355. Luo, Xing, Jihong Wang, Mark Dooner, and Jonathan Clarke. 2015. “Overview of current development in electrical energy storage technologies and the application potential in power system operation.” Applied Energy 137, 511-536. Mills, A., A. Botterud, J. Wu, Z. Zhou, B. Hodge, and M. Heaney. 2013. “Integrating Solar PV in Utility System Operations.” Tech. rep., Argonne National Laboratory. Mohanpurkar, Manish, Abdelhamid Ouroua, Rob Hovsapian, Yusheng Luo, Mohit Singh, Eduard Muljadi, Vahan Gevorgian, and Peter Donalek. 2018. “Real-time co-simulation of adjustable-speed pumped storage hydro for transient stability analysis.” Electric Power Systems Research 154, 276-286. Mukherjee, Poulomi, and V. V. Rao. 2019. “Superconducting magnetic energy storage for stabilizing grid integrated with wind power generation systems.” Journal of Modern Power Systems and Clean Energy 7, 400-411. Mutarraf, Muhammad Umair, Yacine Terriche, Kamran Ali Khan Niazi, Juan C. Vasquez, and Josep M. Guerrero. 2018. “Energy Storage Systems for Shipboard Microgrids – A Review.” Energies 11(12), 3492. National Renewable Energy Laboratory. n.d. Grid Data and Tools. Accessed August 13, 2018. https://nrel.gov/grid/data-tools.html. Natural Resources Canada. 2016. High-resolution solar radiation datasets. Accessed July 9, 2019. https://www.nrcan.gc.ca/energy/renewableelectricity/solar-photovoltaic/18409.

The Role of Energy Storage to Reduce Variability Impacts

261

O’Dwyer, C., L. Ryan, and D. Flynn. 2017. “Efficient Large-Scale Energy Storage Dispatch: Challenges in Future High Renewable Systems.” IEEE Transactions on Power Systems 32, 3439-3450. O’Dwyer, C., M. Anwar, J. Dillon, M. Bakhtvar, G. Buttitta, C. AdradeCabrera, K. Baltputnis, et al. 2018. “D3.6 Cost benefit analysis of SETS and alternative local small-scale storage options.” Tech. rep., EU Horizon 2020 RealValue. n.d. Open Power System Data. Accessed April 6, 2020. https://open-powersystem-data.org. Ortega Manjavacas, Alvaro. 2017. “Dynamic modeling, control and stability analysis of VSC-based energy storage systems.” Ph.D. dissertation, Ireland: University College Dublin. Östergård, Rickard. 2011. “Flywheel energy storage: a conceptual study.” Master’s thesis, Upsala Universitat, Sweden. Palizban, O., and K. Kauhaniemi. 2016. “Energy storage systems in modern grids – Matrix of technologies and applications.” Journal of Energy Storage 6, 248-259. Panda, Anup Kumar, and Trilochan Penthia. 2017. “Design and modeling of SMES based SAPF for pulsed power load demands.” International Journal of Electrical Power & Energy Systems 92, 114-124. Pattahil, P., N. Sivakumar, and T. Sonia. 2017. “Capacitor to Supercapacitor: an Introduction.” In Nanostructured Ceramic Oxides for Supercapacitor Applications, edited by A. Balakrishnan and K. Subramanian, 323-365. CRC Press. Pawel, Ilja. 2014. “The Cost of Storage – How to Calculate the Levelized Cost of Stored Energy (LCOE) and Applications to Renewable Energy Generation.” Energy Procedia 46, 68-77. Pinson, M., and M. Bazant. 2013. “Theory of SEI Formation in Rechargeable Batteries: Capacity Fade, Accelerated Aging and Lifetime Prediction.” Journal of the Electrochemical Society 160, A243-A250. Placer, Neil. 2016. “Finding Solar Energy’s ‘Sweet Spot’.” Renewable Energy World. Accessed July 9, 2019. https://www.renewableenergyworld.com/articles/2016/10/findingsolar-energy-s-sweet-spot.html. Pothisoonthorn, S., and I. Ngamroo. 2016. “Active power modulation control of hybrid PV generator-battery for power swing stabilization.” 2016 13th International Conference on Electrical Engineering/Electronics, Computer, Telecommunications and Information Technology (ECTICON), 1-6.

262

Chapter Eight

Qi, Zhaoxiang, and Gary M. Koenig. 2017. “Review Article: Flow battery systems with solid electroactive materials.” Journal of Vacuum Science & Technology B 35, 1-27. Rehman, Shafiqur, Luai M. Al-Hadhrami, and Md. Mahbub Alam. 2015. “Pumped hydro energy storage system: A technological review.” Renewable and Sustainable Energy Reviews 44, 586-598. REN21. 2019. “Renewables 2019 Global Status Report.” Tech. rep., REN21. Rogner, M., and N. Troja. 2018. “The world’s water battery: pumped hydropower storage and the clean energy transition.” Tech. rep., International hydro power association. Mousavi G, S. M., Faramarz Faraji, Abbas Majazi, and Kamal Al-Haddad. 2017. “A comprehensive review of Flywheel Energy Storage System technology.” Renewable and Sustainable Energy Reviews 67, 477-490. Salameh, Ziyad. 2014. “Energy Storage.” In Renewable Energy System Design, edited by Ziyad Salameh, 201-298. Academic Press. Skyllas-Kazacos, M. 2010. “Electro-chemical energy storage technologies for wind energy systems.” In Stand-Alone and Hybrid Wind Energy Systems, edited by J. K. Kaldellis, 323-365. Woodhead Publishing. Soomro, Abid, Mustafa E. Amiryar, Daniel Naknoo, and Keith R. Pullen. 2019. “Performance and Loss Analysis of Squirrel Cage Induction Machine Based Flywheel Energy Storage System.” Applied Sciences 9(21), 4537. Takada, Kazunori. 2013. “Progress and prospective of solid-state lithium batteries.” Acta Materialia 61, 759-770. Teleke, S., M. E. Baran, S. Bhattacharya, and A. Q. Huang. 2010. “RuleBased Control of Battery Energy Storage for Dispatching Intermittent Renewable Sources.” IEEE Transactions on Sustainable Energy 1, 117124. Traube, J., F. Lu, D. Maksimovic, J. Mossoba, M. Kromer, P. Faill, S. Katz, B. Borowy, S. Nichols, and L. Casey. 2013. “Mitigation of Solar Irradiance Intermittency in Photovoltaic Power Systems With Integrated Electric-Vehicle Charging Functionality.” IEEE Transactions on Power Electronics 28, 3058-3067. Troy, N., E. Denny, and M. O’Malley. 2010. “Base-Load Cycling on a System With Significant Wind Penetration.” IEEE Transactions on Power Systems 25, 1088-1097. Wang, G., M. Ciobotaru, and V. G. Agelidis. 2014. “Power Smoothing of Large Solar PV Plant Using Hybrid Energy Storage.” IEEE Transactions on Sustainable Energy 5, 834-842.

The Role of Energy Storage to Reduce Variability Impacts

263

Weber, Adam Z., Matthew M. Mench, Jeremy P. Meyers, Philip N. Ross, Jeffrey T. Gostick, and Qinghua Liu. 2011. “Redox flow batteries: a review.” Journal of Applied Electrochemistry 41, 1137-1164. Willis, H. Lee. 2004. Power Distribution Planning Reference Book. Second. CRC Press. Wirth, H. 2019. Recent Facts about Photovoltaics in Germany. Tech. rep., Fraunhofer Institute for Solar Energy Systems. Ye, H., J. Wang, Y. Ge, J. Li, and Z. Li. 2017. “Robust Integration of HighLevel Dispatchable Renewables in Power System Operation.” IEEE Transactions on Sustainable Energy 8, 826-835. Zakeri, Behnam, and Sanna Syri. 2015. “Electrical energy storage systems: A comparative life cycle cost analysis.” Renewable and Sustainable Energy Reviews 42, 569-596. Zhu, Jiahui, Panpan Chen, Chenghong Gu, Hongjie Zhang, Jianwei Li, Huiming Zhang, Ming Qiu, Jianlin Li, Weijia Yuan, and Ignacio Hernando Gil. 2018. “Techno-economic analysis of MJ class high temperature Superconducting Magnetic Energy Storage (SMES) systems applied to renewable power grids.” Global Energy Interconnection 1, 172-178.

CHAPTER NINE A NEARLY NET-ZERO ENERGY BUILDING IN OMAN: A CASE STUDY ABDULLAH AL-BADI1 AND AWNI SHAABAN1

Abstract Buildings usually consume a large portion of energy and are responsible for one-third of the global greenhouse gas emissions. Population growth, increasing levels of wealth and migration to cities will increase the energy consumption in buildings. A reduction in the building energy requirement can be achieved through policies, energy-efficient appliances and lighting as well as by the passive and active design of the buildings. Some of the key policies that can be considered are building codes, appliance standards, feed-in tariffs, energy audits and tax exemption. Net-zero energy buildings (NZEBs) and zero energy buildings (ZEBs) will play a major role in reducing energy consumption and in mitigating climate change. The NZEB can be described as a building that has zero balance between the imported energy from the grid and the exported energy to the grid on an annual basis, or a building that has zero carbon emissions on an annual basis. This can be fulfilled by different combinations of passive and active measures, and by the utilization of renewable energy sources. This chapter aims to present a case study about a nearly net-zero energy building (SQU Eco-House). The house was built on Sultan Qaboos University (SQU) campus and provides a high level of comfort. It has two stories, is connected to the electric grid and equipped with 20 kW rooftop photovoltaic system.

1 Sultan Qaboos University, Dept. of Electrical and Computer Engineering, Muscat, Oman, Email (Abdullah Al Badi): [email protected]

A Nearly Net-Zero Energy Building in Oman

265

Keywords: Net-zero energy building, Eco House, passive design, rooftop PV system, renewable energy.

Introduction Energy consumption is continuing to grow globally; by 2040 the energy demand is expected to grow by 30% compared to the current consumption level, the demand for electricity will grow by 60% and the total population will exceed 9 billion (International Energy Agency 2017, 30). The increase in energy demand is due to the increasing population, climate change, economic growth, and urbanization. The buildings and construction sector accounted for 36% of final energy use and 39% of carbon emissions in the world in 2018 (World Green Building Council 2019, 9). In Oman, buildings consumed more than 80% of the total electrical energy generated by the main system [3] (Authority of Electricity Regulation 2018, 14) as presented in Fig. 9-1. To reduce the buildings’ consumption of electrical energy, several energy-saving technologies have been proposed. Furthermore, many policies and initiatives had been proposed to promote Net-Zero Energy Buildings (NZEBs) and Zero Energy Buildings (ZEBs). Accordingly, the buildings can play a major role in reducing energy consumption and mitigating the impacts of climate change through proper building design, construction and operation. Substantial efforts have been made to establish building policies and to implement energy conservation technologies. The European Union aims to reduce its domestic greenhouse gas emissions by 80% in 2050 compared to 1990 levels (Thomas, Hermelink, Schimschar, Grözinger et al. 2011, I). Energy efficient design minimizes energy consumption in buildings by using natural measures to improve comfort conditions. Passive strategies mainly include advanced building envelopes, passive cooling and thermal storage (Cao, Dai, Liu 2016, 198-213). Passive design approaches to buildings play a major role in reducing the building energy requirement as the design affects the cooling, ventilation and lighting loads. A zero-energy building (ZEB) is an innovative idea for a high-performance structure and represents a building with a net energy consumption of zero annually. A Zero-energy building concept was introduced to reduce energy consumption through efficiency gains and the building energy needs can be provided by renewable energy source technologies (Visa, Moldovan, Comsit, Duta 2014, 72-78). This topic has received more attention in recent years and become part of policies on energy-efficient buildings in the USA and the European Union. NZEB includes three types of energy effective measures namely, passive design, service systems and renewable energy generation as shown in Fig. 9-2 (Deng, Wang, Dai 2014, 1-16). A proper passive design for a building may

Chapter Nine

266

include the best orientation, optimizes the placement of windows with low thermal transmittance and U-values, low air infiltration, a high level of insulation in walls and roofs, and window shadings. The service system with the appliances should have high energy efficiency and should create a comfortable environment for the occupants. To offset the energy consumption, renewable energy resources should be utilized. The supply of renewable energy can be either integrated into the building (on-site) or provided to the building from other places (off-site). Ministry of Defense 1% Hotels/Tourism 1%

Government 11%

Agricultural & Fisheries 1%

Residential 46%

Commercial 24%

Industrial 16%

Fig. 9-1: Main Interconnected System Supply by Tariff Category 2018

A Nearly Net-Zero Energy Building in Oman

267

NZEB

Passive design

Renewable Energy

Service System

Envelope

Lighting

Wind

Orientation

DHW

Solar

shade

HVAC

Biomass

.......

......

......

Fig. 9-2: Elements of the net-zero energy building

Concept of zero-energy buildings Zero-energy buildings were introduced to reduce future building energy consumption and to utilize renewable energy resources for building energy needs, and as a result, the negative impact on the climate will be mitigated. Several definitions and calculation methodologies were proposed for ZEBs that include several aspects such as balancing periods, type of balance, metrics, connection with the grid and utilizing renewable energy sources (Marszal, Heiselberg, Bourrelle, Musall, et al. 2011, 971-979). The balancing periods can be monthly, seasonally, annually or even an entire life cycle. About balance, there are two types (a) during the design stage, building energy demand versus renewable sources, and (b) during the operation phase, imported energy from the grid versus energy feed-in. For the off-grid connected buildings, the first balance is used (Torcellini, Pless, Deru, and Crawley 2006, 1-15). Moreover, renewable energy options can be either on-site generation (solar) or off-site generation (biomass). Off-site

268

Chapter Nine

supply uses renewable energy sources available off-site to generate energy on-site such as using biomass, biodiesel and wood pellets. Furthermore, the owner can opt to purchase green energy. A net-zero energy building (NZEB) refers to a ZEB building that is connected to the grid whereas, the other types which are not connected to the grid are called standalone/autonomous ZEB (Al-Saadia and Shaabana 2019, 112-299). The grid will provide electrical energy when there is no renewable energy available, and the building will export the excess energy generation to the grid. A stand-alone (self-sufficient) building can store sufficient renewable energy to be used when needed since there is no access to an external grid. In the European Union, the recast Directive on Energy Performance of Buildings (Directive 2010/31/EU) stated that all new buildings should be nearly ZEB by 31 December 2020, while the new public buildings should fulfill this requirement by 2018 (European Parliament, and Council, 2010, 1-23). In the USA, the Energy Independence and Security Act of 2007 (EISA2007) indicated that 50% of new commercial buildings by 2040 and all new commercial buildings by 2050 should be ZEB (Drury, Shanti, and Torcellini 2009, 18-25). Furthermore, the California Public Utilities Commission has placed an energy action plan to attain net-zero energy for all new residential buildings by 2020 and net-zero for all new commercial buildings by 2030 (Drury, Shanti, and Torcellini 2009, 18-25). Several sustainable building design standards are available which provide different ranking criteria to evaluate zero energy building and/or energy efficiencies such as Ecohomes (BRE, UK), PassivHaus (Germany), and LEED (USA) (Wang, Gwilliam, Jones 2009, 1215-1222). Nevertheless, there are no design guidelines provided for achieving ZEB. To reduce energy consumption, the building should not only have been passively designed, but also be a building that balances its energy requirements with active techniques and renewable energy resource technologies. It has been stated in (Kylili, Fokaides 2015, 86-95) that the new ZEBs can reduce their energy consumption by around two-thirds on average compared to the energy consumption of present buildings. There are four main types of NZEBs: net-zero site energy, net-zero source energy, net-zero energy emissions and net-zero energy costs. The National Renewable Energy Laboratory has come up with the following definitions for the NZEBs (Torcellini, Pless, Deru, and Crawley 2006, 1-15): “Net Zero Site Energy: A site ZEB produces at least as much energy as it uses in a year when accounted for at the site.”

A Nearly Net-Zero Energy Building in Oman

269

“Net Zero Source Energy: A source ZEB produces at least as much energy as it uses in a year when accounted for at the source. Source energy refers to the primary energy used to generate and deliver energy to the site. To calculate a building’s total source energy, imported and exported energy is multiplied by the appropriate site-to-source conversion multipliers.” “Net Zero Energy Costs: In a cost ZEB, the amount of money the utility pays the building owner for the energy the building exports to the grid is at least equal to the amount the owner pays the utility for the energy services and energy used over the year.” “Net Zero Energy Emissions: A net-zero emissions building produces at least as much emissions-free renewable energy as it uses from emissionsproducing energy sources.”

SQU-Eco House The SQU eco-house project aimed to build a nearly net-zero energy house on available land on the Sultan Qaboos University campus. The house consists of two stories with a total area of 300 m2, has one guest room, a living room, a family room, a dining room, a kitchen, three toilets and four bedrooms. The eco-house uses electricity from the grid to supplement the on-site renewable energy source (PV and solar thermal). Domestic hot water is provided by the solar water heater. The house is equipped with 20 kW roof-top PV panels which were installed on a metal structure tilted at 23.5degrees south above the flat roof. The target is to export to the grid energy equal to that imported from the grid, on an annual basis. PV OPTIMUM SLOPE Solar Altitude at noon in a given location (Local Meridian) was calculated as follows: ࡭࢔ = ૢ૙ െ (ࡸ െ ࢾ)

(9-1)

where, An: altitude at noon, L: latitude, and į: solar declination. The corresponding PV slope was calculated by the optimum slope (S) of a solar panel which produces the maximum electric power yield when the solar impact is at a normal incidence to the panel and is given by: ࡿ = ૢ૙ െ ࡭࢔

(9-2)

Chapter Nine

270

The values of monthly (An) of (S) were calculated for Muscat and listed in Table 9-1. Table 9-1: Optimum slope for the maximum solar impact Optimum slope for max solar impact

23.5

Noon altitude 90

15 May/30 July

20

86.5

3.5

15 April/30 Aug.

10

76.5

13.5

21 March/23 Sept.

0

66.5

23.5

28 Feb./15 Oct.

-10

56.5

33.5

28 Jan./15 Nov.

-20

46.5

43.5

-23.5

43

47

Date 22 June

22 Dec.

Declination

0

The optimum slopes for the extreme and middle months are plotted in Fig. 9-3 and it can be seen that: x The optimum slope for the hottest month is 0.0 when solar noon is 90.00. x The optimum slope of the mildest month is 23.50 when solar noon is 66.60. x The optimum slope of the coldest winter month is 47.00 when the noon altitude is 43.00.

Fig. 9-3: Optimum slopes for the extreme and middle months

A Nearly Net-Zero Energy Building in Oman

271

However, the optimum slope for a given month does not get a normal incidence in the other months. The study concluded with a compromised optimum annual slope of 23.5o.

Constraints of the Muscat local climate The climate of Muscat is the maritime desert type that combines high temperatures, with alternating humid and dry and dusty conditions, and very little rain. It has a long overheated period (21March to 21October), with a short comfort period (21 November to 21 March), and there is no cold period. Ample solar radiation and clear skies in Oman would contribute to the high power yield of the PV panels. On the other hand, the intense heat and the high ambient air temperature have a negative effect on the durability and efficiency of the PV system. This problem was addressed by introducing an innovative concept of the “Double-roof system” that performed the following: -

introduced shaded spaces between the PV panels; and scooped the local prevailing wind to produce cooling breezes beneath the PV panels.

Double-roof system The double-roof system comprises an upper lightweight shell which wraps around and shades the whole roof. The system is composed of three elements: the upper shell, the cavity, and the main roof proper as shown in Fig. 9-4. The upper shell comprises the PV panels and its support structure that is lifted 60 cm above the roof leaving an open cavity space. The upper shell provides shading to the main roof and the slope of the PV panel helps to scoop the air to cool the PV panels and the main roof, shown in Fig. 9-5. The cost-effectiveness of the PV system is justified due to the innovative configuration of the double-shell roof that allows the following: 1. The improved efficiency of the PV system due to the cooling breezes that eliminate overheating. 2. Saving on the running costs of air-conditioning due to the effect of roof shading that reduces heat flow through the roof.

Chapter Nine

272

Fig. 9-4: The double-roof system

a-Single-shell building

b-Double-shell building

Fig. 9-5: Single- and double-shell buildings

Ventilation of the roof cavity The prevailing wind in Muscat is NE-N-NW. The PV panels are sloped to the south with their back edges lifted upward to face the prevailing wind

A Nearly Net-Zero Energy Building in Oman

273

and scoop a cooling breeze beneath the PV panels as shown in Fig. 9-6. This concept was developed from the traditional air scope in the Gulf countries.

Fig. 9-6: Prevailing wind and scoop cooling breeze beneath the PV panels

Further cooling air currents are brought in from the open roof edges to encourage air breezes that further cool the space between the roof and the PV panels. This is achieved by open gaps between the roof and the outer double-wall that are designed to stop thermal brides between the walls and the roof. The gaps are covered by date palm wood panels with slots that allow free air infiltration as shown in Fig. 9-7.

Fig. 9-7: Gap covering between the roof and the outer double-wall

Geometric form configuration The configuration and grouping of the PV modules had to meet the following constraints: 1. The most compact layout to maximize the benefit of the available roof space. 2. Avoidance of over shading that would cause electric short circuits that interrupt power flow and may damage the modules. The vertical shadow angle (VS) is the projection of the altitude angle on a plane normal to the wall, and the horizontal shadow angle (HS) is the angle between the vertical plane of the sun and the wall's normal. They were calculated for the Muscat location at latitude 23.45° N:

Chapter Nine

274

‫ܪ‬ௌ = ܼ െ ܼ௪ ܸௌ = arctan (‫ܣ݊ܽݐ‬/cos (‫ܪ‬ௌ ))

(9.3) (9.4)

where ܼ is the solar azimuth, ‫ ܣ‬is the solar altitude, and ܼ௪ is the wall azimuth. The ‫ ܣ‬and ܼ values were obtained from the Muscat solar chart shown in Fig. 9-8 and were used to calculate ‫ܪ‬ௌ and ܸௌ values. These were used to obtain “Profile Angles” which are the plots of ܸௌ in the plane normal to the panel axis. They were used to plot the possible overshadowing of the panels on each other. The critical hour was considered to be 9:00 am on 22 December for the design of the photovoltaic panels.

Fig. 9-8: Muscat solar chart

Overshadowing model tests The results of the above geometric design were further verified by model tests in the artificial sky (Heliodor) tests. Models of adjacent PV modules were tested for the following critical dates: 22 June, 21 March/23 September, and 22 December that were indicated in Table 1. The optimum spacing that provides the compactness of a panel grouping, but without overshadowing is shown in Fig. 9-9. The gaps between the PV modules cause solar penetration to the roof that reduces shading efficiency. The gaps,

A Nearly Net-Zero Energy Building in Oman

275

therefore, were covered by wooden screens to complement the roof shading, and also to be used as maintenance walkways as shown in Fig. 9-10.

Fig. 9-9: The optimum spacing between panels

Fig. 9-10: Wooden screens

Construction methods The main double-roof system carried a subsidiary supporting structure for the PV modules that provides the required slope of 23.5 degrees. The main supporting system had to meet the structural design constraints of dead load and at the same time consider the problems of wind uplifting. It was constructed of I-beams laid on intermediate parapets on the roof as shown in Fig. 9-11.

276

Chapter Nine

Fig. 9-11: The PV supporting structure

Double-shell system External surfaces of buildings get heated due to the combination of air temperature and added heat due to solar radiation. The concept of the double shell is to remove the heat load of solar radiation and bring external building surfaces to shade air temperature. The air currents in the interim space are maximized to avoid the stagnation of hot air. A secondary light shell wraps around the house providing the full shade of all external surfaces. Moreover, the outer shell must be of lightweight material.

Natural ventilation Natural ventilation alone is possible for four months of the year, with no air conditioning required. This was done by a lower opening to inlet the cool breeze, and an upper opening to expel hot air through outlets as depicted in Fig. 9-12. The ventilation is also enhanced by the double volume height of the family room that increases the length of the thermal stack and enhances thermal pressure.

A Nearly Net-Zero Energy Building in Oman

277

Fig.9-12: House’s natural ventilation

Solar Photovoltaic modules The electrical energy is generated by using 20 kW modules (80 modules) which are tilted at 23.5o south. The rated power for each module is 255 watts. Every 20 modules are connected in series. A 20 kW inverter is used to convert DC to AC power to be used by loads in the house as well as to feed excess power back to the grid. A two-way meter is used to measure the amount of electrical energy taken from the grid or fed back to the grid.

Water conservation and treatment The SQU Eco-House uses a compact water treatment unit from Kingspan Klargester. The unit provides a reliable, environmentally safe solution for the reuse of the produced effluents. The treated water unit is used for the irrigation system.

Energy generation and Consumption The energy consumed by the different loads at a certain period is presented in Fig. 9-13. The total power produced by the PV at this instant of time is 14.2 kW; 4.1 kW is fed back to the grid and the rest of the power (10.1 kW) is consumed in the house.

278

Chapter Nine

Fig. 9-13: Energy consumed by different loads

House Occupancies One person is living in the house. However, several experimental research projects are in process in the house which require the running of several AC units most of the time.

Energy balance Based on the Energy Performance Building Directive (EPBD) recast, a nearly zero-energy building (European Parliament, and Council, 2010, 123) (a) Should have high energy performance; (b) energy needed should be covered to a very substantial extent from renewable sources; (c) the remaining energy required should be nearly zero or very small.

A Nearly Net-Zero Energy Building in Oman

279

All these criteria are satisfied by the SQU-Eco house as discussed previously. The energy balance for one year can be found by comparing Figs. 9-2 and 9-3. The net energy = 945 kWh annually which is less than 3.5% of the energy demand supplied by the electrical network. On 9 May 2017 The load of the house during the summer and the amount of energy fed back to the grid, since the installation of the PV system, on 9 May 2017 are presented in Fig. 9-14. ‫ ݁ܿ݊݅ݏ ܸܲ ݕܾ ݊݋݅ݐܿݑ݀݋ݎܲ ݏݏ݋ݎܩ‬2014 = 52618 ܹ݄݇ ‫ ݁ܿ݊݅ݏ ݁ݏݑ݋ܪ ݄݁ݐ ݕܾ ݊݋݅ݐ݌݉ݑݏ݊݋ܥ ݏݏ݋ݎܩ‬2014 = 63873 ܹ݄݇ On 4 May 2018 The load of the house during the summer and the amount of energy fed back to the grid on 4 May 2018 are presented in Fig. 9-15. ‫ ݁ܿ݊݅ݏ ܸܲ ݕܾ ݊݋݅ݐܿݑ݀݋ݎܲ ݏݏ݋ݎܩ‬2014 = 78702ܹ݄݇ ‫ ݁ܿ݊݅ݏ ݁ݏݑ݋ܪ ݄݁ݐ ݕܾ ݊݋݅ݐ݌݉ݑݏ݊݋ܥ ݏݏ݋ݎܩ‬2014 = 90902 ܹ݄݇ Gross production by the PV for one year (May 2017 to May 2018) can be calculated as: ‫ = ܸܲ ݕܾ ݊݋݅ݐܿݑ݀݋ݎܲ ݏݏ݋ݎܩ ݈ܽݑ݊݊ܣ‬78702 െ 52618 = 26084 ܹ݄݇ Gross consumption by the house for one year (May 2017 to May 2018) can be calculated as: ‫ = ݁ݏݑ݋ܪ ݄݁ݐ ݕܾ ݊݋݅ݐ݌݉ݑݏ݊݋ܥ ݏݏ݋ݎܩ ݈ܽݑ݊݊ܣ‬90902 െ 63873 = 27029 ܹ݄݇ ܰ݁‫ = ݕ݃ݎ݁݊ܧ ݐ‬27029 െ 26084 ܹ݄݇ = 945 ܹ݄݇ % of ܰ݁‫= ݕ݃ݎ݁݊ܧ ݐ‬

ଽସହ ଶ଻଴ଶଽ

= 3.4%.

280

Chapter Nine

Fig. 9-14: The load of the house during summer 2017 and the amount of energy fed back to the grid

A Nearly Net-Zero Energy Building in Oman

281

Fig. 9-15: The load of the house during summer 2018 and the amount of energy fed back to the grid

Conclusions The SQU-Eco House was designed and built to achieve a nearly zero energy building. The house is equipped with 80 roof-top photovoltaic panels connected to produce a target of 20 kW. An inverter is used to convert DC to AC power compatible with the local grid. During the day some of the

282

Chapter Nine

generated power is consumed in the house, and the surplus is exported to the grid, while at night it is imported from the grid. The target is to export to the grid energy that is equal to that imported from the grid, on an annual basis. The net energy consumed in the SQU Eco-House = 945 kWh annually which is less than 3.5% of the energy demand supplied by the electrical network. Thus, the SQU Eco-House is a nearly zero energy building. This house can be further developed to be a smart building. Governments should adopt financial instruments and incentives to support energy efficiency investments in buildings. These can include education and training, research and development, demonstration projects, awareness, low-interest loans for retrofitting, tax reduction, grants and building regulations.

Acknowledgements The authors would like to acknowledge The Research Council, Oman and Sultan Qaboos University for providing the financial support for the construction of the Eco House and providing all the required equipment, sensors and tools to do the research.

References International Energy Agency. 2017. World Energy Outlook 2017. Paris: IEA. https://www.iea.org/weo2017/. World Green Building Council. 2019. Global Status Report for Buildings and Construction, Towards a zero-emissions, efficient and resilient buildings and construction sector. London: World Green Building Council. https://www.worldgbc.org/news-media/2019-global-statusreport-buildings-and-construction. Authority of Electricity Regulation. 2018. Authority of Electricity Regulation Annual Report 2018. Oman: Authority of Electricity Regulation. https://www.apsr.om/downloadsdocs/annualreports/AnnualReportEnglish2018.pdf. Boermans Thomas, Andreas Hermelink, Sven Schimschar, Jan Grözinger et al. 2011. Principles for nearly zero-energy buildings, Paving the way for Effective Implementation of Policy Requirements. Brussels: Building Performance Institute Europe. https://www.researchgate.net/publication/271200709_Principles_for_n early_zero-energy_buildings_Paving_the_way_for_effective_ implementation_of_policy_requirements.

A Nearly Net-Zero Energy Building in Oman

283

Xiaodong Cao, Xilei Dai, and Junjie Liu. 2016. “Building energyconsumption status worldwide and the state-of-the-art technologies for zero-energy buildings during the past decade”. Energy and Buildings 128, 198-213. http://dx.doi.org/10.1016/j.enbuild.2016.06.089. Ion Visa, Macedon Moldovan, Mihai Comsit, and Anca Duta. 2014. “Improving the renewable energy mix in a building toward the nearly zero energy status.” Energy and Buildings 68, 72-78. https://doi.org/10.1016/j.enbuild.2013.09.023. S. Deng, R. Z. Wang, and Y. J. Dai. 2014. “How to evaluate performance of net zero energy building – A literature research”. Energy 71, 1-16. https://doi.org/10.1016/j.energy.2014.05.007. A. J. Marszal, P. Heiselberg, J. Bourrelle, E. Musall, et al. 2011. “Zero Energy Building – A review of definitions and calculation methodologies”. Energy and Buildings 43(4), 971-979. https://doi.org/10.1016/j.enbuild.2010.12.022. P. Torcellini, S. Pless, M. Deru, and D. Crawley. “Zero Energy Buildings: A Critical Look at the Definition”. In ACEEE Summer Study on Energy Efficiency in Buildings”. Golden, Colo.: National Renewable Energy Laboratory, 2006. https://www.nrel.gov/docs/fy06osti/39833.pdf . Al-Saadia, Saleh and Shaabana, Awni. 2019. “Zero energy building (ZEB) in a cooling dominated climate of Oman: Design and energy performance analysis”. Renewable and Sustainable Energy Reviews 112, 299-316. https://doi.org/10.1016/j.rser.2019.05.049. European Parliament, and Council. Directive 2010/31/EU of the European Parliament and of the Council of 19 May 2010 on the energy performance of buildings (recast). EU Institutions: Official Journal of the European Union. https://eur-lex.europa.eu/legal-content/en/TXT/? uri=CELEX%3A32010L0031. Drury, Crawley, Pless Shanti, and Paul Torcellini. 2009. “Getting to net zero”. ASHRAE Journal 51(9), 18-25. https://www.researchgate.net/publication/241962259_Getting_to_Net_ Zero. Wang, Liping, Julie Gwilliam, and Phil Jones. 2009. “Case study of zero energy house design in UK”. Energy and Buildings 41(11), 1215-1222. https://doi.org/10.1016/j.enbuild.2009.07.001. Kylili, Angeliki, and Paris Fokaides. 2015. “European Smart Cities: The Role of Zero Energy Buildings”. Sustainable Cities and Society 15, 8695. DOI: 10.1016/j.scs.2014.12.003.