Improved Rocof Performance With Interlocking Techniques For Both High And Low Inertia Distributed Generation Lom Protection.

(1)

IMPROVED ROCOF PERFORMANCE WITH

INTERLOCKING TECHNIQUES FOR BOTH HIGH

AND LOW INERTIA DISTRIBUTED

GENERATION LOM PROTECTION

A thesis submitted to The University of Manchester for the degree of PhD

in the Faculty of Engineering and Physical Sciences

MOHD HENDRA BIN HAIRI

(2)

Table Of Contents

2 TABLE OF CONTENTS

TABLE OF CONTENTS ... 2

LIST OF FIGURES… ... 5

LIST OF TABLES….. ... 9

LIST OF ACRONYMS ... 10

ABSTRACT………… ... 11

DECLARATION…… ... 13

ACKNOWLEDGEMENT ... 15

CHAPTER 1. INTRODUCTION ... 16

1.1. Background ... 16

1.2. Power system protection issues ... 17

1.3. Issues with Loss of Mains detection, and Motivation ... 20

1.3.1. The problems of ROCOF protection ... 21

1.4. Proposed Solution: Improved ROCOF with Interlocking Techniques ... 24

1.5. Objectives of the work ... 25

1.6. Research Methodologies ... 26

1.7. Contribution of the thesis ... 27

1.8. Chapter outlines ... 28

CHAPTER 2. REVIEW OF ANTI-ISLANDING DETECTION METHODS .. 31

2.1. Introduction ... 31

2.2. Formation of an island ... 31

2.3. Current practices to prevent islanding ... 33

2.4. Anti-Islanding Detection ... 34

2.4.1. Passive methods ... 35

2.4.2. Active methods... 46

2.4.3. Methods using communication ... 52

2.5. Summary ... 57

CHAPTER 3. PROPOSED AN IMPROVED ROCOF PROTECTION ... 59

3.1. Introduction ... 59

3.2. ROCOP Method ... 59

3.2.1. Evaluation of ROCOP method in PSCAD/EMTDC ... 61

3.3. ROCOF with ROCOP Interlocking Method ... 66

(3)

Table Of Contents

3.4.1. Evaluation of VU method in PSCAD/EMTDC ... 69

3.5. ROCOF with VU Interlocking Method ... 73

3.6. Summary ... 75

CHAPTER 4. STUDIES OF ROCOF WITHOUT INTERLOCKING METHODS 77 4.1. Simulation Model in PSCAD ... 77

4.1.1. Synchronous Generator with active and reactive power (PQ) control performance ... 77

4.1.2. PV generator with active and reactive power (PQ) control performance. ... 81

4.1.3. ROCOF relay modelling ... 85

4.1.4. Factors Affecting ROCOF Relay Performance ... 86

4.2. Loss of Main Event ... 94

4.2.1. Imbalanced active power between generation and load ... 94

4.2.2. Generation-load balanced ... 97

4.3. Network Disturbances ... 98

4.3.1. Case 1: Load Switching (L3 and L4) ... 99

4.3.2. Case 2: Three-phase fault at PCC (F1 and F2) ... 100

4.4. Summary of Operating Performance of ROCOF Relay ... 101

CHAPTER 5. CASE STUDIES OF THE IMPROVED ROCOF PROTECTION 102 5.1. Introduction ... 102

5.2. Evaluation of ROCOF Protection Sensitivity and Stability ... 102

5.2.1. Sensitivity and Stability Assessment Method ... 103

5.2.2. Distribution Network Modelling ... 105

5.2.3. Test Scenarios and Discussions ... 108

5.2.4. Summary ... 120

5.3. Improved ROCOF Sensitivity & Stability with Interlocking Techniques ... 121

5.3.1. Evaluation of ROCOP interlock performance ... 122

5.3.2. Evaluation of VU interlock performance ... 141

5.3.3. Evaluation of ROCOP & VU Performance ... 159

5.4. Summary ... 180

CHAPTER 6. CONCLUSIONS AND FUTURE WORK ... 182

(4)

Table Of Contents

6.1.1. General ... 182

6.1.2. Summary of main findings ... 183

6.2. Future work ... 186

REFERENCES……… ... 187

APPENDIX A: DG and Network Model Data ... 193

APPENDIX B: Fundamentals ... 195

Synchronous generator model & control ... 195

Basic equations of a synchronous machine ... 196

Stator circuit equations ... 197

Stator self-inductances ... 197

Stator mutual inductances ... 197

Mutual inductance between stator and rotor windings ... 198

Rotor circuit equations ... 199

Stator flux linkages in dq0 components ... 200

Rotor flux linkages in dq0 components ... 200

Stator voltage equations in dqo components ... 200

Electrical power and torque ... 201

Summary of per unit equations ... 201

Active power control of grid connected synchronous generator ... 203

Governor control ... 204

Reactive power control of grid connected synchronous generator ... 207

Three-phase PV Generation model & control. ... 209

Power Control through PV Inverters ... 210

Phase locked loop (PLL) ... 211

Current control ... 213

(5)

List Of Tables

5 LIST OF FIGURES

Figure 1-1: Traditional electric power system (left) and electric power system with distributed

generation (right). ... 17

Figure 1-2: Power System Anti-Islanding Detection Schemes ... 21

Figure 2-1: Power system island ... 32

Figure 2-2: Equivalent circuits of a DG parallel with the grid ... 38

Figure 2-3: Normal operation-rotor displacement angle ... 41

Figure 2-4: Loss of mains-rotor displacement angle ... 42

Figure 2-5: Voltage vector shift ... 42

Figure 2-6: AFD output current waveform. ... 47

Figure 2-7: Impedance measurement technique ... 50

Figure 2-8: Impedance insertion method ... 51

Figure 2-9: Power Line Carrier Communications method ... 53

Figure 2-10: Transfer trip scheme ... 55

Figure 2-11: Sychrophasor-based anti-islanding detection scheme ... 57

Figure 2-12: Islanding detection characteristic using slip and acceleration ... 57

Figure 3-1: The rate of change of power interlock structure ... 61

Figure 3-2: Test network ... 61

Figure 3-3: ROCOP signal during LOM (surplus power mismatch) ... 62

Figure 3-4: ROCOP signal during LOM (deficit power mismatch) ... 63

Figure 3-5: The studied network ... 64

Figure 3-6: ROCOP signal during network disturbance a) load switching; b) starting an induction motor; and c) three phase fault ... 64

Figure 3-7: Flowchart of the ROCOF and ROCOP interlock algorithm. ... 67

Figure 3-8: Grid-connected PV with VU relay ... 70

Figure 3-9: VU response under various degree of power imbalance in a) surplus generation and b) deficit generation during LOM ... 71

Figure 3-10: The studied network ... 72

Figure 3-11: VU results under network disturbances a) load switching; b)induction motor switching and c) three phase fault ... 72

Figure 3-12: The VU interlock structure... 74

Figure 3-13: Flowchart of the ROCOF and VU interlock algorithm ... 74

Figure 4-1: Single line diagram of the grid connected SG... 78

Figure 4-2: Output waveforms of the Synchronous Generator ... 79

Figure 4-3: SG output waveforms during change in power demand ... 80

Figure 4-4: Output waveform of SG during three-phase fault ... 81

Figure 4-5: Single line diagram of the grid-connected PV ... 82

Figure 4-6: Output waveforms of the PV generator ... 82

Figure 4-7: PV output waveforms during change in power demand ... 83

Figure 4-8: Response of the PV output waveforms during three-phase fault ... 84

Figure 4-9: Single line diagram of a grid-connected synchronous generator with ROCOF relay ... 85

Figure 4-10: ROCOF model ... 86

Figure 4-11: Single line diagram of a grid-connected SG with ROCOF relay ... 87

Figure 4-12: Influence of generator inertia constant on ROCOF relay ... 88

Figure 4-13: Effect of Inertia Constant on ROCOF detection time ... 88

Figure 4-14: Effect of Load Type on frequency and ROCOF ... 90

Figure 4-15: Maximum ROCOF setting to trip for different type of load ... 90

Figure 4-16: Influence of load power factor on ROCOF ... 91

Figure 4-17: Effect of load power factor on ROCOF detection time, for a given range of ROCOF settings ... 92

Figure 4-18: The studied network ... 93

Figure 4-19: ROCOF waveform using different measuring windows durations ... 94

(6)

List Of Tables

6 Figure 4-21: Response of ROCOF Relay to LOM with active power mismatch of +/- 10% of

rated SG ... 95

Figure 4-22: Response of ROCOF Relay to LOM with active power mismatch of +/- 10% of rated PV ... 96

Figure 4-23: Response of ROCOF Relay to LOM in balanced generation and load ... 97

Figure 4-24: Response of ROCOF Relay to LOM when PV output power matched the local load ... 98

Figure 4-25: Studied network ... 99

Figure 4-26: ROCOF response during load (L3) switching ... 99

Figure 4-27: ROCOF response during load (L4) switching ... 100

Figure 4-28: Response of ROCOF during three-phase fault (F1). ... 100

Figure 4-29: Response of ROCOF during three-phase fault (F2) ... 101

Figure 5-1: Proposed testing structure ... 104

Figure 5-2: Morton Park & Pirelli Network Model with 33 nodes ... 108

Figure 5-3: Maximum allowable ROCOF setting to ensure tripping for various active power P mismatch conditions during LOM in PV ... 110

Figure 5-4: Maximum allowable ROCOF setting to ensure tripping for various reactive power Q mismatch conditions during LOM in PV ... 110

Figure 5-5: Summary of ROCOF setting in various cases of active and reactive power imbalance for PV ... 111

Figure 5-6: Maximum allowable ROCOF setting to ensure tripping for various active power P mismatch conditions during LOM in SG ... 112

Figure 5-7: Maximum allowable ROCOF setting to ensure tripping for various reactive power Q mismatch conditions during LOM in SG ... 113

Figure 5-8: Summary of ROCOF setting in various cases of active P and reactive power Q imbalance for SG ... 113

Figure 5-9: ROCOF signal during several of network disturbance and faults ... 115

Figure 5-10: Minimum allowable ROCOF setting to ensure stability for various network disturbances in PV case ... 115

Figure 5-11: ROCOF signal during several of network disturbance and faults ... 116

Figure 5-12: Minimum allowable ROCOF setting to ensure stability for various network disturbances in SG case ... 116

Figure 5-13: Minimum setting to preserve stability between PV and SG ... 117

Figure 5-14: ROCOF sensitivity and stability setting for PV ... 118

Figure 5-15: ROCOF sensitivity and stability setting for SG ... 119

Figure 5-16: The studied network ... 122

Figure 5-17: Performance of ROCOF with ROCOP interlocking for the network with SG under generation-load balanced condition ... 124

Figure 5-18: Performance of ROCOF with ROCOP interlocking for the network with SG under 10% mismatch in active power P condition ... 125

Figure 5-19: Performance of ROCOF with ROCOP interlocking for the network with SG under -10% mismatch in active power P condition ... 126

Figure 5-20: Performance of ROCOF with ROCOP interlocking for the network with SG under 10% mismatch in reactive power Q condition ... 127

Figure 5-21: Performance of ROCOF with ROCOP interlocking for the network with SG under -10% mismatch in reactive power Q condition ... 128

Figure 5-22: Performance of ROCOF with ROCOP interlocking for the network with SG under 20% mismatch in active power P condition ... 129

Figure 5-23: Performance of ROCOF with ROCOP interlocking for the network with SG under -20% mismatch in active power P condition ... 130

Figure 5-24: Performance of ROCOF with ROCOP interlocking for the network with SG under 20% mismatch in reactive power Q condition ... 130

Figure 5-25: Performance of ROCOF with ROCOP interlocking for the network with SG under -20% mismatch in reactive power Q condition ... 131

Figure 5-26: Performance of ROCOF with ROCOP interlocking for the network with SG under 30% and 40% mismatch in active power P conditions ... 132

(7)

List Of Tables

7 Figure 5-27: Performance of ROCOF with ROCOP interlocking for the network with SG under -30% and -40% mismatch in active power P conditions... 132 Figure 5-28: Performance of ROCOF with ROCOP interlocking for the network with SG under 30% and 40% mismatch in reactive power Q conditions ... 133 Figure 5-29: Performance of ROCOF with ROCOP interlocking for the network with SG under -30% and -40% mismatch in reactive power Q conditions ... 133 Figure 5-30: Summary of circuit breaker tripping time during LOM ... 135 Figure 5-31: Performance of ROCOF with ROCOP interlocking for the network with SG under network disturbances (Case 1) ... 136 Figure 5-32: Performance of ROCOF with ROCOP interlocking for the network with SG under network disturbances (Case 2) ... 137 Figure 5-33: Performance of ROCOF with ROCOP interlocking for the network with SG under network disturbances (Case 3) ... 137 Figure 5-34: Performance of ROCOF with ROCOP interlocking for the network with SG under network disturbances (Case 4) ... 138 Figure 5-35: Performance of ROCOF with ROCOP interlocking for the network with SG under network disturbances (Case 5) ... 138 Figure 5-36: Performance of ROCOF with ROCOP interlocking for the network with SG under network disturbances (Case 6) ... 139 Figure 5-37: Performance of ROCOF with ROCOP interlocking for the network with SG under network disturbances (Case 7) ... 139 Figure 5-38: Performance of ROCOF with ROCOP interlocking for the network with SG under network disturbances (Case 8) ... 140 Figure 5-39: Performance of ROCOF with VU interlocking for the network with PV under generation-load balanced condition. ... 142 Figure 5-40: Performance of ROCOF with VU interlocking for the network with PV under 10% mismatch in active power P condition ... 143 Figure 5-41: Performance of ROCOF with VU interlocking for the network with PV under -10% mismatch in active power P condition ... 144 Figure 5-42: Performance of ROCOF with VU interlocking for the network with PV under 10% mismatch in reactive power Q condition ... 145 Figure 5-43: Performance of ROCOF with VU interlocking for the network with PV under -10% mismatch in reactive power Q condition ... 146 Figure 5-44: Performance of ROCOF with VU interlocking for the network with PV under 20% mismatch in active power P condition ... 147 Figure 5-45: Performance of ROCOF with VU interlocking for the network with PV under -20% mismatch in active power P condition ... 148 Figure 5-46: Performance of ROCOF with VU interlocking for the network with PV under 20% mismatch in reactive power Q condition ... 148 Figure 5-47: Performance of ROCOF with VU interlocking for the network with PV under -20% mismatch in reactive power Q condition ... 149 Figure 5-48: Performance of ROCOF with VU interlocking for the network with PV under 30% and 40% mismatch in active power P conditions ... 150 Figure 5-49: Performance of ROCOF with VU interlocking for the network with PV under -30% and -40% mismatch in active power P conditions ... 150 Figure 5-50: Performance of ROCOF with VU interlocking for the network with PV under 30% and 40% mismatch in reactive power Q conditions ... 151 Figure 5-51: Performance of ROCOF with VU interlocking for the network with SG under -30% and -40% mismatch in reactive power Q conditions ... 151 Figure 5-52: Summary of circuit breaker tripping time during LOM ... 153 Figure 5-53: Performance of ROCOF with VU interlocking for the network with PV under network disturbances (Case 1) ... 154 Figure 5-54: Performance of ROCOF with VU interlocking for the network with PV under network disturbances (Case 2) ... 155 Figure 5-55: Performance of ROCOF with VU interlocking for the network with PV under network disturbances (Case 3) ... 155

(8)

List Of Tables

8 Figure 5-56: Performance of ROCOF with VU interlocking for the network with PV under

network disturbances (Case 4) ... 156

Figure 5-57: Performance of ROCOF with VU interlocking for the network with PV under network disturbances (Case 5) ... 156

Figure 5-58: Performance of ROCOF with VU interlocking for the network with PV under network disturbances (Case 6) ... 157

Figure 5-59: Performance of ROCOF with VU interlocking for the network with PV under network disturbances (Case 7) ... 157

Figure 5-60: Performance of ROCOF with VU interlocking for the network with PV under network disturbances (Case 8) ... 158

Figure 5-61: The studied network ... 160

Figure 5-62: Performance of ROCOF with ROCOP interlocking when switching on SG2 while SG1 in service... 161

Figure 5-63: Performance of ROCOF with ROCOP interlocking when switching on wind farm generator while SG1 in service. ... 162

Figure 5-64: Switching on SG3 generator while SG1 in service ... 162

Figure 5-65: Switching on PV1 generator while SG1 in service ... 163

Figure 5-66: Switching on PV3 generator while SG1 in service ... 164

Figure 5-67: Switching on PV2 generator while SG1 in service ... 164

Figure 5-68: Switching on PV2 generator while PV1 in service ... 165

Figure 5-69: Switching on PV3 generator while PV1 in service ... 166

Figure 5-70: Switching on SG1 generator while PV1 in service ... 166

Figure 5-71: Switching on SG2 generator while PV1 in service ... 167

Figure 5-72: Switching on wind farm generator while PV1 in service ... 168

Figure 5-73: Switching on SG3 generator while PV1 in service ... 168

Figure 5-74: Stability test results for three-phase fault (PV1 based generator) ... 169

Figure 5-75: Stability test results for single phase fault (PV1 based generator) ... 170

Figure 5-76: Stability test results for three-phase fault (SG1- based generator) ... 171

Figure 5-77: Stability test results for single phase fault (SG1-based generator) ... 172

Figure 5-78: The studied network ... 174

Figure 5-79: Performance of the ROCOF with VU interlocking for the feeder with PV under LOM condition (case 1) ... 175

Figure 5-80: Performance of the ROCOF with ROCOP interlocking for SG1 when B14 opened ... 176

Figure 5-81: Performance of the ROCOF with ROCOP interlocking for SG2 when B14 opened ... 177

Figure 5-82: Performance of the ROCOF with VU interlocking for PV4 when B14 opened . 178 Figure 5-83: Performance of the ROCOF with ROCOP interlocking for SG2 when B14 opened ... 179

Figure 5-84: Performance of the ROCOF with ROCOP interlocking for SG2 when B14 opened ... 180

(9)

List Of Tables

9 LIST OF TABLES

Table 1-1: Protection issues for Distribution Generation ... 18

Table 2-1: Interconnection protection objectives ... 33

Table 2-2: Passive Islanding Detection Methods ... 35

Table 2-3: Active Islanding Detection Methods ... 35

Table 2-4: Communication Islanding Detection Methods ... 35

Table 3-1: ROCOP threshold ... 65

Table 3-2: VU threshold ... 73

Table 4-1: ROCOF detection time ... 89

Table 4-2: ROCOF detection time ... 92

Table 5-1: Transformer Data at Morton Park & Pirelli Primary Substations ... 106

Table 5-2: Line Parameters of Morton Park & Pirelli network and Power consumption on each node ... 106

Table 5-3: Minimum performance criteria for PV. [T=trip, NT=no trip] ... 120

Table 5-4: Minimum performance criteria for SG. [T=trip, NT=no trip] ... 120

Table 5-5: Sensitivity performance. [T=trip, NT=no trip] ... 134

Table 5-6: Sensitivity performance. [T=trip, NT=no trip] ... 134

Table 5-7: Stability performance. [T=trip, NT=no trip] ... 140

Table 5-8: Sensitivity performance. [T=trip, NT=no trip] ... 152

Table 5-9: Sensitivity performance. [T=trip, NT=no trip] ... 152

Table 5-10: Stability performance. [T=trip, NT=no trip] ... 158

Table 5-11: Summary of stability results for PV ... 170

(10)

Acronyms

10 LIST OF ACRONYMS

AFD Active Frequency Drift AMR Automatic Meter Reading

COROCOF Comparison of Rate of Change of Frequency CT Current Transformer

DFIG Doubly-Fed Induction Generator DG Distributed Generator

DNO Distribution Network Operator EHV Extra High Voltage

FJ Frequency Jump

GPS Global Positioning System

HV High Voltage

LOG Loss of Grid LOM Loss Of Main LV Low Voltage MV Medium Voltage NDZ Non-Detection Zone P Active Power PF Power Factor

PLCC Power Line Carrier Communication PMU Phasor Measurement Unit

PSCAD Power System Computer Aided Design PV Photovoltaic

Q Reactive Power

ROCOF Rate of Change of Frequency ROCOP Rate of Change of Power ROCOV Rate of Change of Voltage

SCADA Supervisory Control and Data Acquisition SFS Sandia Frequency Shift

SG Synchronous Generator SMS Slip mode Frequency Shift SVS Sandia Voltage Shift THD Total Harmonic Distortion

UFP/OFP Under/Over Frequency Protection UHV Ultra-High Voltage

UVP/OVP Under/Over Voltage Protection VT Voltage Transformer

VS Vector Shift VU Voltage Unbalance

α1 ROCOF pickup value

α2 ROCOP pickup value

α3 Zero counter pickup value

β1 ROCOF pickup value

β2 VU pickup value

(11)

Abstract

11 ABSTRACT

Due to increasing concerns related to mitigating climate change, emphasis has been placed on the more-rapid deployment of lower-inertia Distributed Generation (DG), with various technological diversifications in the power system network. The increasing use of lower-inertia DGs in power systems has recently produced challenges in the protection of the DG, especially in Loss of Main protection (LOM). From the literature review, it was found that the most commonly-used protection method against LOM is Rate of Change of Frequency (ROCOF), due to its simplicity during installation, quick response and low cost. However, ROCOF relays are known to suffer from sensitivity problems such as nuisance tripping and mal-operation, caused by system-wide disturbances. Thus the motivation of this thesis is to propose an improved ROCOF protection with interlocking techniques for both high-inertia and low-inertia DGs, to ensure high levels of protection sensitivity and stability. The techniques are based on two methods as follows; Method 1: Utilizing Rate of Change of Frequency (ROCOF) as main detection and Rate of Change of Power (ROCOP;dp⁄dt_{) as an interlocking function}

for high-inertia types of DGs such as Synchronous Generators (SG) or Induction Machine generators, and Method 2: Utilizing Rate of Change of Frequency (ROCOF) as the main detection and Voltage Unbalance (VU) as an interlock function for low-inertia type DGs such as Photovoltaics (PV) and wind generators. Initially ROCOF alone was simulated and studied in simple study networks using PSCAD/EMTDC. The results show that it is difficult to achieve the correct balance between the required sensitivity and stability under various conditions of LOM and non-LOM events. The setting of the ROCOF relay is always a balancing exercise between sensitivity and stability. An improved ROCOF with ROCOP interlocking technique was then developed and tested using practical networks with SG. The results confirm that the proposed technique works effectively, and significantly reduces false operation of ROCOF relays during network disruptions. Next, the simulation of an improved ROCOF relay with VU interlock was conducted for a network with PV generators. The results indicate that the proposed method is capable of providing more stable operation during LOM conditions and non-LOM events. Finally, simulation was conducted for ROCOF with interlocking techniques for feeders with SG and PV. The results confirm the effectiveness of the proposed method even in conditions where there is a mixture of DGs connected to the same feeder. For these reasons, it is considered that the objective of this research activity

(12)

Abstract

has been achieved, and it has been shown that the proposed technique offers an improvement over existing techniques.Main contribution of the thesis is to develop an improved ROCOF with interlocking techniques to deal with both high and low inertial generations. The performance of the proposed improved ROCOF methods with interlocking techniques has been proved through simulations and analysis under both faults and disturbance conditions.

(13)

Declaration

13 DECLARATION

I declare that no portion of the work referred to in the thesis has been submitted in support of an application for another degree or qualification of this or any other university or other institute of learning.

(14)

Copyright Statement

14 COPYRIGHT STATEMENT

I. The author of this thesis (including any appendices and/or schedules to this thesis) owns certain copyright or related rights in it (the “Copyright”) and s/he has given The University of Manchester certain rights to use such Copyright, including for administrative purposes

II. Copies of this thesis, either in full or in extracts and whether in hard or electronic copy, may be made only in accordance with the Copyright, Designs and Patents Act 1988 (as amended) and regulations issued under it or, where appropriate, in accordance with licensing agreements which the University has from time to time. This page must form part of any such copies made.

III. The ownership of certain Copyright, patents, designs, trademarks and other intellectual property (the “Intellectual Property”) and any reproductions of copyright works in the thesis, for example graphs and tables (“Reproductions”), which may be described in this thesis, may not be owned by the author and may be owned by third parties. Such Intellectual Property and Reproductions cannot and must not be made available for use without the prior written permission of the owner(s) of the relevant Intellectual Property and/or Reproductions.

IV. Further information on the conditions under which disclosure, publication and commercialisation of this thesis, the Copyright and any Intellectual Property and/or Reproductions described in it may take place is available in the University IP Policy (see http://www.campus.manchester.ac.uk/medialibrary/policies/intellectual-property.pdf), in any relevant Thesis restriction declarations deposited in the University Library, The University Library’s regulations (see http://www.manchester.ac.uk/library/aboutus/regulations) and in The University’s policy on presentation of Theses.

(15)

Acknowledgement

15 ACKNOWLEDGEMENT

I would like to express my sincere gratitude to my supervisor Dr. Haiyu Li for his constant guidance and great efforts throughout my research, and for his trust in my ability to conduct the work.

I am greatly indebted to the Malaysian Ministry of Higher Education and Universiti Teknikal Malaysia Melaka (UTeM) for the financial sponsorship which covered my tuition fees and living maintenance for three years. To all of my colleagues in the research group and others in Ferranti building, School of Electrical and Electronic Engineering, I appreciate your company and thank you for providing an enjoyable working environment.

Last but not least, I wish to give my thanks to my parents, my parents-in-law, my siblings, my wife Mrs. Siti Rahimah Ab. Rahaman and my daughters Haifa Hanish and Hamra Zahra for their patience and selfless support during my research.

(16)

Chapter 1.Introduction

16 CHAPTER 1. INTRODUCTION

1.1. Background

For more than half a century, electrical power system structures have been defined by generating plants, transmission lines and distribution networks that are physically connected with each other. Electricity from generating plants is stepped up to high voltage (HV), extra high voltage (EHV), and ultra-high voltage (UHV) levels, to be transmitted over long distances through interconnecting transmission systems. The voltage from the transmission systems is then stepped down to medium voltage (MV) and c) radial distribution level, where the electric power is distributed to the loads [1]. Over the last few years, the concept of Distributed Generators (DGs) has been developed; these can be defined as small-scale, dispersed, decentralized, and on-site generation systems connected at the distribution network level [2]. DGs have gained wide acceptance by the electrical industry due to their technical, economic, and environmental benefits. To date, the installed capacities of DGs have been growing rapidly. For example, photovoltaic energy (PV) has seen a 170% increase in existing capacity, and is expected to reach 84GW in 2017 [3]. This increase is due to the fact that integration of DGs into an existing power system can offer several benefits, such as line loss reduction, reduced environmental impacts, peak shaving, increased overall energy efficiency, relief of transmission and distribution congestion, voltage support, and deferment of the investment required to upgrade existing generation, transmission, and distribution systems [4-16].

DG can come from renewable or non-renewable energy resources, using both modern and conventional technologies. DG technologies include internal combustion engines, small gas turbines, wind turbines, small combined cycle gas turbines, micro-turbines, solar photovoltaic cells, fuel cells, biomass and small geothermal generating plants [2, 12, 17].

Figure 1-1 outlines the structures of a traditional electric power system, and a power system with distributed generation. The existing distribution networks, operating for many decades, are designed and operated in radial configurations, with unidirectional

(17)

Chapter 1.Introduction

power flow from centralized generating station to customers [2].The introduction of DG at the distribution network level will alter the power flow in the system, and the distribution system can no-longer be considered to have a unidirectional power flow. Hence, the presence of the DG, especially in significant proportions, will definitely have both positive and negative impacts on a power system network. It is therefore necessary to identify and quantify any adverse impacts associated with DG connection when planning the future expansion of a distribution network, in order to achieve high security and quality of electricity supplies.

Large Generating Stations Interconnected Transmission System Radial Distribution System P o w e r f lo w d ir e c ti o n P o w e r f lo w d ir e c ti o n

Figure 1-1: Traditional electric power system (left) and electric power system with distributed generation (right).

1.2. Power system protection issues

Since existing distribution networks are based on a radial topology, the power flows are from upper voltage levels down to customers situated along the radial feeders. This has enabled a relatively standard and straightforward protection strategy. When applying over-current protection, for example, it has been possible to assume that the fault current can only flow in one direction. However, as the share of DG increases, distribution

(18)

Chapter 1.Introduction

networks are becoming more like transmission networks with bilateral power flows. Hence, the protection relay may not work correctly, and network security cannot be guaranteed. This can lead to poor system reliability, system instability, and power quality issues for customers. Table 1-1 outlines some issues related to system protection in the presence of DG [14-16, 18-29].

Table 1-1: Protection issues for Distribution Generation 1. False tripping of feeders (sympathetic tripping)

2. Nuisance tripping of production units 3. Blinding of protection

4. Increased or decreased fault levels 5. Unwanted islanding / loss of main 6. Prohibition of automatic reclosing 7. Un-synchronized re-closing

Of the challenges created by DG, listed in the Table 1-1, the protection against loss of main is the most concerning [27, 30, 31]. Islanding, or loss of main, (LOM) is a situation that occurs when part of the network is disconnected from the remainder of the power system, but remains energized by a distributed generator. DG islanding can be both unintentional and intentional. Unintentional islanding is prohibited due to the safety concerns listed below [25, 28, 32-34]:

1. Power quality

The distribution network operator (DNO) is responsible for maintaining a high level of power quality to its customers as DG is not allowed to regulate the voltage and the frequency at the PCC. The DGs may not be able to maintain the frequency, voltage balance and magnitude within required limits during islanding by means of their efficient control methods. As a results the consumer may experience poor power quality that can damage their electrical equipment’s.

2. Earthing & Protection

The neutral of an islanded network may not be earthed. Operating un-earthed networks may pose a significant safety risk to personnel and equipment. Also the fault level may

(19)

Chapter 1.Introduction

be altered during LOM and become insufficient to allow protection to operate satisfactory.

3. Out of synchronism reclosing

An automatic recloser is widely used in medium voltage overhead line and a very effective device for removing transient fault since faults on overhead lines are temporary in nature. The addition of DG into distribution networks raises a challenging issue to the conventional auto reclosing operation. The connected DG during recloser opening time (0.2s to a 2 seconds) can sustain the voltages and prevent fault arc extinction on the feeder. This will leads to unsuccessful reclosure attempts. The situation can get worst if the reclosing take place when the utility system is out of phase with the generation. This is due to a sudden high transient inrush current and large mechanical torque during out of phase reclosing which can stresses and damage to the DG unit.

4. Personnel Safety

Islanding may pose a threat for un-aware utility workers who may not realize that the line is still energized due to the DG continues to feed electrical power in the island. The risk of electrical shock to personnel is the most concerning issue.

For all the reasons described above, Loss of Main protection is usually considered necessary for the effective integration of DGs into the existing power network in order to achieve adequate safety and reliability operation in distribution network. All DG units need to be equipped with anti-islanding devices, to force them cease operation as soon as possible after the formation of islands. The current standards and recommendations, such as IEEE 1547 (US) [35] and Engineering Recommendation G59 (UK) [36], require all DGs to detect islanding and disconnect themselves from the grid within two seconds of the island state occurring, to prevent damage to the generator. The rules and guidelines are vary from country to country but often similar to the following requirement [25]:

• DG should be disconnected from the network in the case of abnormality in voltage or frequency.

• If one or more phases are disconnected from the grid supply the DG units should be rapidly disconnected from the network.

(20)

Chapter 1.Introduction

20 1.3. Issues with Loss of Mains detection, and Motivation

Any loss of main (LOM) event or unplanned island should be detected instantaneously by the anti-islanding protection. However, it is hard to accomplish due to several factors which might affect the success of the detection schemes, as follows:

• The scheme should work for any possible formations of islands. Note that there could be multiple switchers, re-closers and fuses between a distributed generator and the supply substation - opening of any one of the devices will form an island. Since each islanding formation can have a different mixture of loads and distributed generators, the behaviour of each island can be quite different. A reliable anti-islanding scheme must work for all possible islanding scenarios [32].

• The scheme should detect the islanding situation within a required time frame. The main constraint is to prevent out-of-phase reclosing of the distributed generators. A re-closer is typically programmed to re-energize its downstream system after about a 0.5 to 1 second delay. Ideally, the anti-islanding scheme must trip its DG before the reclosing takes place [32].

To date, many LOM methods have been developed [27, 31, 34]. However satisfying LOM protection solution has not yet been found. In general, these methods can be divided into three categories: 1) passive methods, 2) active methods and 3) communication methods. They are illustrated in Figure 1-2.

(21)

Chapter 1.Introduction

21 Anti-islanding

Scheme

Passive Active

Communication Local Detection

Frequency Voltage

Power Harmonic

Voltage/Freq or Phase shift Impedance mesurement

Power line Signalling

Transfer Trip

Figure 1-2: Power System Anti-Islanding Detection Schemes

Passive anti-islanding protection methods rely on the detection of abnormalities in voltage, frequency and phase angle at the DG output when islanding occurs. They are very cost effective and locally installed, but suffer from a large non-detection zone (NDZ) [37].

Active methods depend on the deliberate injection of a small signal to the grid, and the resulting reaction is then measured and compared to the pre-set threshold. These methods have the potential to affect power quality and generate instability in the grid, but the NDZ is much lower compared to passive methods [25]. Communication-based methods use communication links. Since these methods use a communications link, thus it can be un-economical to implement.

1.3.1. The problems of ROCOF protection

Although many LOM methods have been developed in the past 20 years, satisfying LOM protection is still missing. Among all the methods, ROCOF is today the most commonly employed LOM detection method, but it has suffer from large non-detection zone (NDZ). For this main reason, many researchers are trying to avoid ROCOF relay application and developed a different kind of detection such as active methods, communication method and hybrid methods in order to reduce risk of NDZ operation.

(22)

Chapter 1.Introduction

The ROCOF relay is typically rely on the measurement of rate of change of system frequency. It is also a passive methods that is, it does not disturb the system to which it is connected. These relays are favoured by industry to protect the DG against islanding, since they have good sensitivity, low cost and wide availability [38]. ROCOF operation assumes that loss of mains causes a sudden change in frequency due to the mismatch of load and generation within the formed island. A trip signal will be issued by the relay if the ROCOF exceeds the predetermined threshold In the UK, a typical ROCOF setting is 0.125Hz/s, in Northern Ireland the setting is ranges from 0.45Hz/s to 0.5Hz [39], in Denmark the setting is 2.5Hz/s and Spain is 2Hz/s [40]. The selection of the settings is strongly dependent on the strength of the system. The stronger the system, the lower the setting is. This is due to higher levels of non-synchronous generation that may result in higher levels of ROCOF values on the power system.

Unfortunately, ROCOF relays are not very reliable, and are prone to nuisance tripping during network transient and faults if the settings are not carefully chosen. Furthermore, ROCOF will not operate if there is a close match between generation and load in the island. This is due to the rate of change of frequency is very small, hence the islanding will be difficult to be detected by the ROCOF relay.

Also, there are some factors which can significantly affect ROCOF performance, including types of loads, load power factor, generator technologies such as the Synchronous Generator (SG), Doubly-Fed Induction Generator (DFIG), inverter-based systems, generator inertia constant, feeder length, degree of imbalance power mismatch, and multi-embedded generators [41].

In the case of types of loads for example, during islanding, the large variation of ROCOF occurs when the load is constants impedance load compared to constants power and constants current load type due to the large nodal voltage reduction in the islanded system.As a consequence, the relay can detect the variation of frequency more quickly. Power system inertia can be defined as the total amount of kinetic energy stored in all rotating generators and motors that are synchronously connected to the network [42]. The inertia of a rotating shaft can be represented as:

(23)

Chapter 1.Introduction

23 =

(1.1)

Where is the inertia constant in seconds, is the moment of inertia of the shaft in , is the base power, and is the nominal speed. The power balance in a conventional power system is determined by:

= − (1.2)

Where J [kgm2_{] is the total inertia of the rotating masses, [rad/s] the speed and}

[W] and [W] is the mechanical and electrical power respectively. Kinetic energy is stored in the rotating masses of all synchronously connected machines in the system. If the load in the system suddenly exceeds the generation, there will be an imbalance between and . Kinetic energy will be released and transferred as electric energy to the power system. Since the turbine governors are slow, the initial power increase is solely obtained by this inertial response. During this process the generator speed decreases and so does the electric frequency. The rate of change of electrical frequency is proportional to the inertia of the systems. The larger the inertia, the smaller the rate of change is. Typical values of power system inertia varies from country to country. For example, the inertia constant of the 60Hz power system of Japan is around 14 to 18 seconds [43]. In UK the inertia value is from 4 to 5.7 seconds [44], whereas in German is between 3 to 4 seconds [45].

The increasing penetration of wind generator and PV solar into power system which employ power electronic converters, do not contribute to power system inertia as they are mechanically decoupled from the power system frequency. As a consequence, the total inertia available to the power system is decreased, which makes the power system more vulnerable against frequency excursions. For this reason, high ROCOF is expected during LOM due to the frequency behaviour of the inverter is strongly dependent on the design of the inverter controller used, such as a Phase Locked Loop (PLL). This may lead to cascading outages of multiple DGs (blackouts).

In contrast, rotating machine-based DGs such as synchronous generator tend to have high-inertia and a longer time constant compared to inverter based DGs such as PV, and

(24)

Chapter 1.Introduction

therefore can tolerate a slower dynamic response of the frequency during LOM. As a consequence, ROCOF relays takes longer time to detect frequency variations. This is described as the relay being ‘insensitive’.

Also, the performance of LOM relays is influenced by their internal algorithms; ROCOF relays with different internal algorithms can react very differently to the same event, thus affecting relay performance in terms of stability and sensitivity [46].

The topic of this thesis is motivated by the increasing challenges for power system protection especially on ROCOF protection, mainly due to the rising energy shares from non-synchronously DG, notably wind & PV units which will impact on rate of change of frequency and frequency deviation in power system with decreasing system inertia. Many LOM methods have been developed in the past 20 years, but yet the ROCOF is still favoured by the utility to detect islanding although it has many drawbacks and difficulties as described before. Thus the motivation of the thesis is to design and develop an improved ROCOF method, with interlocking techniques to deal with both high and low inertia DG, to ensure high levels of protection system sensitivity and stability.

1.4. Proposed Solution: Improved ROCOF with Interlocking Techniques

interlocking function for high-inertia types of DGs such as Synchronous Generators (SG) or Induction Machine generators, and Method 2: Utilizing Rate of Change of Frequency (ROCOF) as the main detection and Voltage Unbalance (VU) as an interlock function for low-inertia type DGs such as Photovoltaics (PV) and wind generators. They are applied and extended to the problem of loss of main detection.

The first interlocking method constitutes a novel solution to real time loss of main detection for the protection of distributed generators that uses ROCOP. ROCOP were developed by Redfern et al.[47] in 1995. Previous work using ROCOP for LOM detection appeared in [48] and [49]. This method is based on monitoring the fluctuations

(1)

19 be altered during LOM and become insufficient to allow protection to operate satisfactory.

3. Out of synchronism reclosing

4. Personnel Safety

or frequency.

• If one or more phases are disconnected from the grid supply the DG units should be rapidly disconnected from the network.

(2)

1.3. Issues with Loss of Mains detection, and Motivation

(3)

21 Anti-islanding

Scheme

Passive Active

Communication Local Detection

Frequency Voltage

Power Harmonic

Voltage/Freq or Phase shift Impedance mesurement

Power line Signalling

Transfer Trip

Figure 1-2: Power System Anti-Islanding Detection Schemes

1.3.1. The problems of ROCOF protection

(4)

22 The ROCOF relay is typically rely on the measurement of rate of change of system frequency. It is also a passive methods that is, it does not disturb the system to which it is connected. These relays are favoured by industry to protect the DG against islanding, since they have good sensitivity, low cost and wide availability [38]. ROCOF operation assumes that loss of mains causes a sudden change in frequency due to the mismatch of load and generation within the formed island. A trip signal will be issued by the relay if the ROCOF exceeds the predetermined threshold In the UK, a typical ROCOF setting is 0.125Hz/s, in Northern Ireland the setting is ranges from 0.45Hz/s to 0.5Hz [39], in Denmark the setting is 2.5Hz/s and Spain is 2Hz/s [40]. The selection of the settings is strongly dependent on the strength of the system. The stronger the system, the lower the setting is. This is due to higher levels of non-synchronous generation that may result in higher levels of ROCOF values on the power system.

Power system inertia can be defined as the total amount of kinetic energy stored in all rotating generators and motors that are synchronously connected to the network [42]. The inertia of a rotating shaft can be represented as:

(5)

23 =

(1.1)

Where is the inertia constant in seconds, is the moment of inertia of the shaft in , is the base power, and is the nominal speed. The power balance in a conventional power system is determined by:

= − (1.2)

Where J [kgm2_{] is the total inertia of the rotating masses, [rad/s] the speed and}

In contrast, rotating machine-based DGs such as synchronous generator tend to have high-inertia and a longer time constant compared to inverter based DGs such as PV, and

(6)

24 therefore can tolerate a slower dynamic response of the frequency during LOM. As a consequence, ROCOF relays takes longer time to detect frequency variations. This is described as the relay being ‘insensitive’.

1.4. Proposed Solution: Improved ROCOF with Interlocking Techniques

In this thesis, the novel solution of the improved ROCOF using interlocking techniques are based on two parameter as follows; Method 1: Utilizing Rate of Change of Frequency (ROCOF) as main detection and Rate of Change of Power (ROCOP;dp⁄dt) as an interlocking function for high-inertia types of DGs such as Synchronous Generators (SG) or Induction Machine generators, and Method 2: Utilizing Rate of Change of Frequency (ROCOF) as the main detection and Voltage Unbalance (VU) as an interlock function for low-inertia type DGs such as Photovoltaics (PV) and wind generators. They are applied and extended to the problem of loss of main detection.

Improved Rocof Performance With Interlocking Techniques For Both High And Low Inertia Distributed Generation Lom Protection.