WIDE AREA MONITORING, PROTECTION AND CONTROL IN THE FUTURE GREAT BRITAIN POWER SYSTEM A thesis submitted to the University of Manchester For the degree of Doctor of Philosophy In the Faculty of Engineering and Physical Sciences 2012 By Deyu Cai School of Electrical and Electronic Engineering Preface List of Contents List of Contents ............................................................................................................... 1 List of Figures .................................................................................................................. 4 List of Tables ................................................................................................................. 10 Abstract .......................................................................................................................... 13 List of Abbreviations .................................................................................................... 14 Declaration ..................................................................................................................... 15 Copyright Statement ..................................................................................................... 16 Acknowledgements........................................................................................................ 18 Chapter 1 Introduction ................................................................................................. 19 1.1 Research Background ........................................................................................... 19 1.1.1 Power system blackouts ................................................................................. 19 1.1.2 Wide area monitoring, protection and control ............................................... 22 1.2 Objectives of the Research .................................................................................... 24 1.3 Thesis Structure..................................................................................................... 26 1.4 Main Contributions of This Research ................................................................... 28 Chapter 2 Synchronized Measurement Technology .................................................. 29 2.1 Introduction ........................................................................................................... 29 2.2 Phasor Measurement Unit ..................................................................................... 31 2.3 Data Concentrator ................................................................................................. 34 2.4 Synchrophasor Standard ....................................................................................... 35 2.5 Summary ............................................................................................................... 35 Chapter 3 Applications and Benefits of Synchronized Measurement Technology . 36 3.1 Introduction ........................................................................................................... 36 3.2 Off-line Applications of SMT ............................................................................... 36 3.2.1 Post-disturbance analysis ............................................................................... 36 3.2.2 Benchmarking, validation and fine-tuning of system models........................ 38 3.3 On-line Applications of SMT ............................................................................... 39 3.3.1 Wide area phase angular and power flow monitoring ................................... 39 3.3.2 Wide area frequency monitoring.................................................................... 41 2.3.3 Wide area voltage monitoring ........................................................................ 42 3.3.4 Inter-area oscillation monitoring .................................................................... 44 3.3.5 Power system restoration ............................................................................... 45 3.3.6 Improved state estimation .............................................................................. 47 3.3.7 Dynamic rating of overhead transmission lines ............................................. 48 2.3.8Intelligent controlled islanding ....................................................................... 49 2.3.9 Adaptive under-frequency load shedding ...................................................... 50 3.4 Conclusions ........................................................................................................... 51 Chapter 4 Architecture of a WAMPAC System ........................................................ 52 4.1 Introduction ........................................................................................................... 52 4.2 Architecture of a WAMPAC System .................................................................... 52 4.3 Communication Networks of WAMPAC System ................................................ 55 4.3.1 Available communication media for WAMPAC ........................................... 55 4.3.2 Communication protocols and format for phasor data transmission ............. 57 4.3.3 Communication latency ................................................................................. 59 4.4 Architecture of future GB WAMPAC system ...................................................... 59 4.5 Conclusions ........................................................................................................... 61 -1- Preface Chapter 5 The Roadmap to the Future GB WAMPAC System ............................... 62 5.1 Introduction ........................................................................................................... 62 5.2 The Roadmap to a WAMPAC System ................................................................. 63 5.3 The Future GB Power System .............................................................................. 65 5.4 GB’s strategy for WAMPAC ................................................................................ 68 5.4.1 Short term strategy ......................................................................................... 68 5.4.2 Long term strategy ......................................................................................... 71 5.5 Conclusions ........................................................................................................... 81 Chapter 6 The Physical Nature of Inter-area Oscillations in Electrical Power Systems ........................................................................................................................... 82 6.1 Introduction ........................................................................................................... 82 6.2 Nonlinear Simulations........................................................................................... 82 6.2.1 Local oscillatory mode in Area 1 ................................................................... 83 6.2.2 Local oscillatory mode in Area 2 ................................................................... 84 6.2.3 Inter-area oscillatory mode ............................................................................ 85 6.2.4 Large disturbance ........................................................................................... 87 6.3 Modal Analysis ..................................................................................................... 88 6.3.1 Dynamic system representation ..................................................................... 89 6.3.2 System linearization for modal analysis ........................................................ 90 6.3.3 Eigenvalues and eigenvectors ........................................................................ 92 6.3.4 Eigenvalues and small signal stability ........................................................... 93 6.3.5 Participation factors ....................................................................................... 94 6.3.6 Modal analysis for inter-area oscillation study .............................................. 95 6.4 The origin of lightly damped/unstable inter-area oscillations ............................ 100 6.5 Conclusions ......................................................................................................... 102 Chapter 7 Inter-area Oscillation Monitoring Using Newton-Type Algorithm ..... 104 7. 1 Introduction ........................................................................................................ 104 7.2 Signal Model Representation .............................................................................. 105 7.3 Newton Type Algorithm Derivation ................................................................... 108 7.4 Computer Simulated Tests .................................................................................. 110 7.4.1 Static tests .................................................................................................... 111 7.4.2 Noise tests .................................................................................................... 113 7.4.3 Dynamic tests ............................................................................................... 116 7.5 Dynamic Simulation of a Multi-machine System ............................................... 118 7.6 Real-life Conditions Tests................................................................................... 122 7.7 Conclusions ......................................................................................................... 124 Chapter 8 The Application of Power Electronic Devices for Damping Inter-area oscillations .................................................................................................................... 126 8.1 Introduction ......................................................................................................... 126 8.2 Modal Analysis for Control ................................................................................ 127 8.2.1 Transfer functions ........................................................................................ 127 8.2.2 Residue based damping controller design .................................................... 128 8.3 Inter-area Oscillation Damping Control with HVDC ......................................... 131 8.3.1 HVDC transmission system modelling ........................................................ 132 8.3.2 System performance without HVDC damping controller............................ 139 8.3.3 Residue based HVDC damping controller design ....................................... 140 8.3.4 System performance with a HVDC damping controller .............................. 143 8.4 Inter-area Oscillation Control with TCSC .......................................................... 146 8.4.1 TCSC modelling .......................................................................................... 146 -2- Preface 8.4.2 System performance without TCSC damping controller ............................. 148 8.4.3 Residue based TCSC damping controller design ......................................... 149 8.4.4 System performance with TCSC damping controller .................................. 151 8.5 Inter-area Oscillation Control with SVC............................................................. 154 8.5.1 SVC modelling ............................................................................................. 154 8.5.2 System performance without SVC damping controller ............................... 155 8.5.3 Residue based SVC damping controller design ........................................... 157 8.5.4 System performance with SVC damping controller .................................... 158 8.6 Conclusions ......................................................................................................... 162 Chapter 9 Wide Area Monitoring and Control System (WAMCS) in a Future GB Power System............................................................................................................... 163 9.1 Introduction ......................................................................................................... 163 9.2 Assessment of the Inter-area Oscillations in GB Power System ........................ 163 9.2.1 GB power system modelling ........................................................................ 163 9.2.2 Inter-area oscillations study in GB power system........................................ 165 9.3 Wide Area Monitoring and Control System (WAMCS) in a Future GB Power System ....................................................................................................................... 170 9.3.1 Architecture of WAMCS in GB power system ........................................... 170 9.3.2 Inter-area oscillation monitoring using NTA ............................................... 172 9.3.3 Wide area inter-area oscillation damping control with HVDC in GB ......... 174 9.3.4 The impact of time delays on wide-area control .......................................... 177 9.3.5 Interaction between PSSs and HVDC damping control system .................. 180 9.4 Conclusions ......................................................................................................... 182 Chapter 10 Conclusions and Future Work............................................................... 184 10.1 Conclusions ....................................................................................................... 184 10.2 Future Work ...................................................................................................... 187 References .................................................................................................................... 190 Appendices ................................................................................................................... 196 10.1 Appendix A ....................................................................................................... 196 10.2 Appendix B ....................................................................................................... 199 10.3 Appendix C ....................................................................................................... 205 10.4 Appendix D ....................................................................................................... 210 10.5 Appendix E ....................................................................................................... 213 10.6 Appendix F ........................................................................................................ 217 10.7 Appendix G ....................................................................................................... 221 10.7.1 Published journal papers ............................................................................ 221 10.7.2 Submitted journal papers ........................................................................... 221 10.7.3 Published conference papers ...................................................................... 221 -3- Preface List of Figures Figure 1.1: Statistics of blackouts: customers affected ................................................... 19 Figure 1.2: Line of separation from the European grid................................................... 20 Figure 1.3: A Generalized WAMPAC system ................................................................ 23 Figure 2.1: Phasor representation of a sinusoidal signal ................................................. 30 Figure 2.2: Synchronized phasor measurement in remote substations ........................... 30 Figure 2.3: A functional block diagram of a typical PMU ............................................. 31 Figure 2.4: State estimation v.s. PMU measurements .................................................... 32 Figure 2.5: Two standalone PMUs ................................................................................. 32 Figure 2.6: Two integrated PMUs................................................................................... 33 Figure 2.7: Data concentrator in a WAMPAC system.................................................... 34 Figure 2.8: SEL Data Concentrator................................................................................. 34 Figure 3.1: The reconstruction of system frequencies after a large disturbance in WECC, 14th June, 2004 ............................................................................................................... 37 Figure 3.2: Comparison of the recorded system response to the 10th August, 1996 disturbance in USA with the simulation results .............................................................. 38 Figure 3.3: RTDMS – a wide area visualization platform for the North American power system.............................................................................................................................. 40 Figure 3.4: Disturbance localization using PMUs and the triangulation method ........... 42 Figure 3.5: Estimation of the Thevenin equivalent with local measurements ................ 43 Figure 3.6: T and Thevenin representation of transmission corridor .............................. 44 Figure 3.7: Oscillations observed by two PMUs in the European grid........................... 45 Figure 3.8: PMU measurements from three areas during reclosing attempts: UCTE, 4 November 2006 ............................................................................................................... 46 Figure 3.9: PMUs installed at both ends of a transmission line ...................................... 48 Figure 3.10: Visualization of the real time mentoring of the thermal condition of an overhead transmission corridor in APG .......................................................................... 49 Figure 4.1: The general architecture of a typical WAMPAC system ............................. 53 Figure 4.2: The combination of different communication media for a WAMPAC system ......................................................................................................................................... 57 Figure 4.3: Generic schematic of the dataflow in WAMPAC system ............................ 58 Figure 4.4: A simple demonstration of communication latency in WAMPAC system .. 59 Figure 4.5: A general architecture of future GB WAMPAC system .............................. 60 Figure 5.1: WAMPAC application tree with full “smart fruits” ..................................... 63 Figure 5.2: Roadmap for deploying PMU applications .................................................. 64 Figure 5.3: UK energy target 2020 and 2030.................................................................. 66 Figure 5.4: Power transfers across the boundary between Scotland and England at peak load condition.................................................................................................................. 66 Figure 5.5: New TCSC and HVDC in GB transmission networks ................................. 67 Figure 5.6: Locations of off-shore wind farms in GB power system (2020-2030) ........ 67 Figure 5.7: ‘Logical filter one’ – The number of PMUs required .................................. 69 Figure 5.8: ‘Logical filter two’ – Commercial availability of PMU applications .......... 69 Figure 5.9: ‘Logical filter three’ – Necessity of PMU application for investors ............ 69 Figure 5.10: Global power angle and frequency monitoring system .............................. 70 Figure 5.11: Real time monitoring system over inter-tie corridors................................. 71 -4- Preface Figure 5.12: PMU placements in SHPTN ....................................................................... 72 Figure 5.13: PMU placements at the boundary between Scotland and England ............ 73 Figure 5.14: PMU placements in central England .......................................................... 74 Figure 5.15: PMU placements in the South of England ................................................. 74 Figure 5.16: A PMU installed in the Torness substation ................................................ 75 Figure 5.17: Three PMUs installed in the Crekey Beck, Keadby and Grimsby West substations ....................................................................................................................... 75 Figure 5.18: One PMU installed in the Sizewell substation ........................................... 76 Figure 5.19: Two PMUs installed in the North Wales substations of Wylfa and Stanah76 Figure 5.20: A PMU installed in the Alverdiscott substation ......................................... 76 Figure 5.21: PMU placements for the future GB wide area monitoring system............. 77 Figure 5.22: A SMT-based adaptive UFLS scheme in the future GB power system ..... 78 Figure 5.23: Inter-area oscillation damping control with power electronic devices in GB power system................................................................................................................... 79 Figure 5.24: Closed loop inter-area oscillation control using HVDC............................. 80 Figure 5.25: Closed loop inter-area oscillation control using TCSC .............................. 81 Figure 6.1: A typical two-area system ............................................................................ 83 Figure 6.2: Generator rotor speed responses to the disturbances occurred in area 1 ...... 84 Figure 6.3: Generator rotor speed responses to the disturbances occurred in area 2 ...... 85 Figure 6.4: Generator rotor speed oscillations dominated by inter-area mode ............... 85 Figure 6.5: System frequency responses in inter-area mode........................................... 86 Figure 6.6: Oscillatory active power flow on transmission line 3 .................................. 87 Figure 6.7: Oscillatory active power flow on transmission line 1 .................................. 87 Figure 6.8: Responses of the Generator rotor speeds to the large disturbance ............... 88 Figure 6.9: Active power transfer over the tie line after the disturbance ........................ 88 Figure 6.10: Oscillatory modes in the two-area system .................................................. 97 Figure 6.11: Right eigenvector (mode shape) of inter-area mode .................................. 99 Figure 6.12: The effect of inter-area power flow on system oscillatory mode ............. 101 Figure 7.1: A global block diagram of the test procedure............................................. 111 Figure 7.2: Computer generated test signal .................................................................. 112 Figure 7.3: Estimation results of the magnitude of DC component, A0 ....................... 112 Figure 7.4: Estimation results of the magnitude of oscillatory component A1 ............. 112 Figure 7.5: Estimation results of the damping factor of oscillatory component, σ ..... 112 Figure 7.6: Estimation results of the frequency of oscillatory component, f. ............... 113 Figure 7.7: Estimation results of the phase angle of oscillatory component, φ ........... 113 Figure 7.8: Computer generated test signal and estimated signal ................................. 113 Figure 7.9: Algorithm tracking capabilities in presence of noise ................................. 115 Figure 7.10: Computer generated signal with step change of σ .................................. 116 Figure 7.11: Estimation results of the magnitude of DC component, A0, for different Tdw ................................................................................................................................. 116 Figure 7.12: Estimation results of the magnitude of oscillatory component, A1, for different Tdw .................................................................................................................. 117 Figure 7.13: Estimation results of damping factor, σ , for different Tdw ....................... 117 Figure 7.14: Estimation results of the frequency of oscillatory component, f, for different Tdw .................................................................................................................. 117 -5- Preface Figure 7.15: Estimation results of the phase angle of oscillatory component, φ , for different Tdw .................................................................................................................. 117 Figure 7.16: Dynamic testing: comparison between the actual and estimated signal... 118 Figure 7.17: A block diagram of the testing procedure based on dynamic simulation of a multi-machine test system ............................................................................................. 118 Figure 7.18: Two-area test system with HVDC link .................................................... 119 Figure 7.19: Generator rotor speed changes after the disturbance ................................ 119 Figure 7.20: Inter-area oscillatory mode shape estimated by NTA .............................. 120 Figure 7.21: Oscillatory active power (on line 2) after the disturbance ....................... 121 Figure 7.22: Estimated frequency of the inter-area oscillatory mode by NTA and Prony method ........................................................................................................................... 121 Figure 7.23: Estimated damping factor of the inter-area oscillatory mode by NTA and Prony method ................................................................................................................ 121 Figure 7.24: Estimated frequency of the inter-area oscillatory mode by reduced-order Prony method and NTA ................................................................................................ 122 Figure 7.25: Estimated damping factor of the inter-area oscillatory mode by reducedorder Prony method and NTA ....................................................................................... 122 Figure 7.26: Estimated oscillatory active power from estimated parameters ............... 122 Figure 7.27: Oscillatory voltage phase angle difference between Glasgow and London ....................................................................................................................................... 123 Figure 7.28: Estimated frequency of inter-area oscillatory mode in the GB network .. 123 Figure 7.29: Estimated damping ratio of inter-area oscillatory mode in the GB network ....................................................................................................................................... 124 Figure 7.30: Oscillatory signal based on estimated oscillatory parameters .................. 124 Figure 8.1: Closed loop system with feedback control ................................................. 129 Figure 8.2: The structure of a feedback damping control ............................................. 129 Figure 8.3: The shift of an eigenvalue caused a feedback damping control ................. 130 Figure 8.4: A two-area system with HVDC .................................................................. 131 Figure 8.5: Two types of power converters .................................................................. 132 Figure 8.6: Rectifier equivalent circuit ......................................................................... 133 Figure 8.7: Inverter equivalent circuit........................................................................... 134 Figure 8.8: Monopolar HVDC link ............................................................................... 135 Figure 8.9: Bipolar HVDC link..................................................................................... 135 Figure 8.10: Homopolar HVDC link ............................................................................ 136 Figure 8.11: Equivalent circuit of HVDC link .............................................................. 136 Figure 8.12: Voltage profile of the equivalent circuit of HVDC link ........................... 136 Figure 8.13: Basic control scheme of HVDC system ................................................... 138 Figure 8.14: Inverter’s β control for constant voltage .................................................. 138 Figure 8.15: Rectifier’s α control for constant current ................................................. 138 Figure 8.16: Oscillatory modes in the two-area system with HVDC ........................... 139 Figure 8.17: Generator rotor speed responses to the small disturbances without HVDC damping control ............................................................................................................ 140 Figure 8.18: System frequency responses to the small disturbances without HVDC damping control ............................................................................................................ 140 Figure 8.19: Rectifier control with supplementary damping control ............................ 141 Figure 8.20: Feedback control with multiple input signals ........................................... 141 -6- Preface Figure 8.21: An illustration of the estimation of the phase angle of the residue .......... 142 Figure 8.22: An estimation of the residue’s phase angle for HVDC damping control design ............................................................................................................................ 143 Figure 8.23: Oscillatory modes versus the gain of HVDC damping controller ............ 144 Figure 8.24: Generator rotor speed responses to the small disturbances with and without HVDC damping control ................................................................................................ 144 Figure 8.25: Active power flow (line 3) responses to the small disturbances with and without HVDC damping control ................................................................................... 145 Figure 8.26: Generator rotor speed responses to a three-phase fault with and without HVDC damping control ................................................................................................ 145 Figure 8.27: Active power flow (line 3) responses to a three-phase fault with and without HVDC damping control ................................................................................... 146 Figure 8.28: Typical two-area system with TCSC ........................................................ 146 Figure 8.29: A structure of a typical TCSC .................................................................. 147 Figure 8.30: An ideal model of TCSC for power system stability study ...................... 147 Figure 8.31: An equivalent circuit of the transmission corridor with TCSC ................ 147 Figure 8.32: A block diagram of TCSC control............................................................ 148 Figure 8.33: Oscillatory modes in the two-area system with TCSC ............................. 148 Figure 8.34: Generator rotor speed responses to the small disturbances without TCSC damping control ............................................................................................................ 149 Figure 8.35: System frequency responses to the small disturbances without TCSC damping control ............................................................................................................ 149 Figure 8.36: A block diagram TCSC supplementary damping control ........................ 150 Figure 8.37: An estimation of the residue’s phase angle for TCSC damping control design ............................................................................................................................ 150 Figure 8.38: Oscillatory modes versus the gain of TCSC damping controller ............. 151 Figure 8.39: Generator rotor speed responses to the small disturbances with and without TCSC damping control ................................................................................................. 152 Figure 8.40: Active power flow (line 3) responses to small disturbance with and without TCSC damping control ................................................................................................. 152 Figure 8.41: Generator rotor speed responses to a three-phase fault with and without TCSC damping control ................................................................................................. 153 Figure 8.42: Active power flow (line 3) responses to a three-phase fault with and without TCSC control ................................................................................................... 153 Figure 8.43: Modified two-area system with SVC ....................................................... 154 Figure 8.44: A structure of a typical SVC .................................................................... 154 Figure 8.45: An ideal model of SVC ............................................................................ 155 Figure 8.46: Block diagram of SVC control ................................................................. 155 Figure 8.47: Oscillatory modes in the two-area system with SVC ............................... 156 Figure 8.48: Generator rotor speed responses to the small disturbances without SVC damping control ............................................................................................................ 156 Figure 8.49: Oscillatory voltage angle difference between bus3 and bus5 caused by the disturbances ................................................................................................................... 157 Figure 8.50: A block diagram of a SVC damping controller ........................................ 157 Figure 8.51: An estimation of the residue’s phase angle for SVC damping control design ....................................................................................................................................... 158 -7- Preface Figure 8.52: Oscillatory modes versus the gain of SVC damping controller ............... 159 Figure 8.53: Generator rotor speed responses to the small disturbances with and without SVC damping control.................................................................................................... 160 Figure 8.54: Active power flow (line 3) responses to small disturbance with and without SVC damping control.................................................................................................... 160 Figure 8.55: Generator rotor speed responses to a three-phase fault with and without SVC damping control.................................................................................................... 161 Figure 8.56: Active power flow (line 3) responses to a three-phase fault with and without SVC damping control ...................................................................................... 161 Figure 9.1: 500 kV HVDC links and 400 kV Series Compensators that are installed in the GB power system (vision 2015).............................................................................. 164 Figure 9.2: Inter-area oscillation monitoring SPTN and NGETN ................................ 165 Figure 9.3: Three large generators selected in Scotland for monitoring inter-area oscillations .................................................................................................................... 166 Figure 9.4: Three large generators selected in central England for monitoring inter-area oscillations .................................................................................................................... 166 Figure 9.5: Two large generators selected in the South of England for monitoring interarea oscillations ............................................................................................................. 167 Figure 9.6: The oscillatory inter-area power flow on the Harker-Hutton line after a large disturbance .................................................................................................................... 167 Figure 9.7: The system frequency variations caused by a large disturbance ................ 168 Figure 9.8: The oscillatory inter-area power flow caused by a large disturbance in the original and PSSs-reduced GB system.......................................................................... 168 Figure 9.9: System frequency variations caused by the large disturbance in the PSSsreduced system .............................................................................................................. 169 Figure 9.10: System frequency variations measured in substations HUER and DEES in the PSSs-reduced system .............................................................................................. 169 Figure 9.11: Architecture of the wide-area inter-area oscillation monitoring and control system in GB ................................................................................................................. 172 Figure 9.12: Inter-area oscillation mode identified by NTA in GB system .................. 173 Figure 9.13: Estimated inter-area oscillation mode shape of GB system ..................... 173 Figure 9.14: Schematic diagram of HVDC damping system........................................ 174 Figure 9.15: Responses of the generator rotor speeds (PEHE, LOAN, HUER and EGGB) to the large disturbance with and without the wide area HVDC damping control ....... 175 Figure 9.16: Responses of the generator rotor speeds (WBUR, RUGE, SIZE and DUNG) to the large disturbance with and without the wide area HVDC damping control ....... 176 Figure 9.17: The response of the inter-area power flow on the Harker-Hutton line to the large disturbance, with and without HVDC damping control ...................................... 176 Figure 9.18: Inter-area oscillation mode identified by NTA in the GB system, with and without HVDC damping control ................................................................................... 177 Figure 9.19: An illustration of the time delay involved in the data transmission in GB WAMCS........................................................................................................................ 178 Figure 9.20: A block diagram of HVDC damping controllers (with time delay) ......... 178 Figure 9.21: The effect of time delay in the wide area damping controllers (50 milliseconds to 200 milliseconds) .......................................................................... 179 -8- Preface Figure 9.22: The effect of time delay in the wide area damping controllers (300 milliseconds to 700 milliseconds)......................................................................... 179 Figure 9.23: Responses of the generator rotor speeds (PEHE, LOAN, HUER and EGGB) to the large disturbance (all PSSs in service) ................................................................ 180 Figure 9.24: Responses of the generator rotor speeds (WBUR, RUGE, SIZE and DUNG) to the large disturbance (all PSSs in service) ................................................................ 181 Figure 9.25: The influence of HVDC damping control on the inter-area power flow in complete GB system ..................................................................................................... 182 Figure A-1: A single line diagram of two-area system ................................................. 196 Figure A-2: Block diagram of static exciter of G1, G2 and G3 and G4 ....................... 197 Figure A-3: Block diagram of speed governor of G1, G2, G3 and G4 ......................... 197 Figure B-1: A single line diagram of the two-area system with HVDC ....................... 199 Figure B-2: Block diagram of DC exciter of G1, G3 and G4 ....................................... 200 Figure B-3: Block diagram of static exciter of G2........................................................ 200 Figure B-4: Block diagram of speed governor of G1, G2, G3 and G4 ......................... 201 Figure B-5: Rectifier operation condition (DIgSILENT interface) .............................. 202 Figure B-6: Rectifier’s α control for constant current................................................... 202 Figure B-7: Inverter operation condition (DIgSILENT interface) ................................ 203 Figure B-8: inverter’s β control for constant voltage .................................................... 203 Figure C-1: A single line diagram of the two-area system with HVDC ....................... 205 Figure C-2: Block diagram of static exciter of G1, G2 G3 and G4 .............................. 205 Figure C-3: Block diagram of speed governor of G1, G2, G3 and G4 ......................... 206 Figure C-4: Rectifier operation condition (DIgSILENT interface) .............................. 207 Figure C-5: Rectifier’s α control for constant current................................................... 207 Figure C-6: Inverter operation condition (DIgSILENT interface) ................................ 208 Figure C-7: inverter’s β control for constant voltage .................................................... 208 Figure C-8: Rectifier control with supplementary damping control ............................. 209 Figure D-1: A single line diagram of the two-area system with TCSC ........................ 210 Figure D-2: Block diagram of static exciter of G1, G2 and G3 and G4 ....................... 210 Figure D-3: Block diagram of speed governor for G1, G2, G3 and G4 ....................... 211 Figure D-4: TCSC with supplementary damping control. ............................................ 211 Figure E-1: A single line diagram of the two-area system with SVC ........................... 213 Figure E-2: Block diagram of static exciter of G1, G2 and G3 and G4 ........................ 213 Figure E-3: Block diagram of speed governor of G1, G2, G3 and G4 ......................... 214 Figure E-4: SVC composite (DIgSILENT interface).................................................... 215 Figure E-5: Primary voltage control of SVC ................................................................ 215 Figure E-6: SVC supplementary damping controller.................................................... 216 Figure F-1: Rectifier operation condition (DIgSILENT interface) ............................... 217 Figure F-2: Rectifier’s α control for constant current ................................................... 217 Figure F-3: Inverter operation condition (DIgSILENT interface) ................................ 218 Figure F-4: inverter’s β control for constant voltage .................................................... 218 Figure F-5: HVDC supplementary damping controller in GB full model .................... 219 -9- Preface List of Tables Table 2.1: Summary of available standalone and integraed PMUs ................................ 33 Table 5.1: Three scenarios for meeting the 2020 UK renewable targets ........................ 66 Table 6.1: Eigenvalues of the two-area system............................................................... 96 Table 6.2: Oscillation modes in the two-area system ..................................................... 97 Table 6.3: Right eigenvector for eigenvalue 28 (associated with inter-area mode) ....... 98 Table 6.4: Participation vector for Eigenvalue 28 (associated with inter-area mode) .. 100 Table 6.5: The effect of inter-area power flow on inter-area mode .............................. 101 Table 6.6: Effect of inter-tie line impedance on inter-area mode ................................. 102 Table 7.1: Sensitivity analysis for random noise .......................................................... 114 Table 7.2: Sensitivity analysis for sampling for frequency .......................................... 115 Table 7.3: Sensitivity analysis for data window size .................................................... 115 Table A-1: Synchronous machine parameters of G1, G2 and G3 and G4 .................... 196 Table A-2: Power generation conditions of G1, G2 and G3 and G4 ............................ 196 Table A-3: Parameters of static exciter ......................................................................... 197 Table A-4: Parameters of speed governor..................................................................... 197 Table A-5: Transformer parameters .............................................................................. 197 Table A-6: AC transmission line parameters ................................................................ 198 Table A-7: Load data .................................................................................................... 198 Table A-8: Shunt capacitors.......................................................................................... 198 Table B-1: Synchronous machine parameters of G1, G2 and G3 and G4 .................... 199 Table B-2: Power generation conditions of G1, G2 and G3 and G4 ............................ 199 Table B-3: Parameters of DC exciter ............................................................................ 200 Table B-4: Parameters of static exciter ......................................................................... 201 Table B-5: Parameters of speed governor ..................................................................... 201 Table B-6: Transformer parameters .............................................................................. 201 Table B-7: AC transmission line parameters ................................................................ 201 Table B-8: DC transmission line parameters ................................................................ 202 Table B-9: Parameters of rectifier control .................................................................... 202 Table B-10: Parameters of inverter control................................................................... 203 Table B-11: Load data................................................................................................... 203 Table B-12: Shunt capacitors ........................................................................................ 204 Table C-1: Synchronous machine parameters of G1, G2 and G3 and G4 .................... 205 Table C-2: Power generation conditions of G1, G2 and G3 and G4 ............................ 205 Table C-3: Parameters of static exciter ......................................................................... 206 Table C-4: Parameters of speed governor ..................................................................... 206 Table C-5: Transformer parameters .............................................................................. 206 Table C-6: AC transmission line parameters ................................................................ 206 Table C-7: DC transmission line parameters ................................................................ 207 Table C-8: Parameters of rectifier control .................................................................... 207 Table C-9: Parameters of inverter control..................................................................... 208 Table C-10: Parameters of HVDC supplementary control ........................................... 209 Table C-11: Load data................................................................................................... 209 Table C-12: Shunt capacitors ........................................................................................ 209 Table D-1: Synchronous machine parameters of G1, G2 and G3 and G4 .................... 210 -10- Preface Table D-2: Power generation conditions of G1, G2 and G3 and G4 ............................ 210 Table D-3: Parameters of Static Exciter........................................................................ 211 Table D-4: Parameters of Governor .............................................................................. 211 Table D-5: The capacitor and reactor of TCSC under steady state............................... 211 Table D-6: Parameters of TCSC supplementary damping control ............................... 212 Table D-7: Transformer parameters .............................................................................. 212 Table D-8: Transmission line parameters ..................................................................... 212 Table D-9: Load data .................................................................................................... 212 Table D-10: Shunt capacitors ........................................................................................ 212 Table E-1: Synchronous machine parameters of G1, G2 and G3 and G4 .................... 213 Table E-2: Power generation conditions of G1, G2 and G3 and G4............................. 213 Table E-3: Parameters of static exciter ......................................................................... 214 Table E-4: Parameters of speed governor ..................................................................... 214 Table E-5: Transformer parameters .............................................................................. 214 Table E-6: Transmission line parameters...................................................................... 214 Table E-7: Load data ..................................................................................................... 215 Table E-8: Shunt capacitors .......................................................................................... 215 Table E-9: Parameters of the primary voltage control of SVC ..................................... 215 Table E-10: Parameters of the primary voltage control of SVC ................................... 216 Table F-1: DC transmission line parameters................................................................. 217 Table F-2: Parameters of rectifier control ..................................................................... 217 Table F-3: Parameters of rectifier control ..................................................................... 218 Table F-4: Parameters of the HVDC damping controller ............................................. 219 Table F-5: A list of the PSS in service .......................................................................... 220 -11- Preface -12- Preface Abstract The growing issue of power-grid congestion and a global increase in disturbances have emphasized the need to enhance electrical power networks using Wide Area Monitoring, Protection, and Control (WAMPAC). This is a cost-effective solution for improving power system planning and operation. In addition to these existing issues, the Great Britain (GB) power system is facing significant changes, in terms of both power transmission technology and the nature of the generation mix, that will cause the operation of the future GB power system to become more unpredictable and complex. Therefore, developing a WAMPAC system will be essential to enhance the stability and optimise the operation of the future GB power system. The main objectives of the research presented in this thesis are to design a GB WAMPAC system and develop solutions to overcome the challenges that will be involved in the initial stage of the GB WAMPAC project. As Synchronized Measurement Technology (SMT) is the most essential element and enabler of WAMPAC, this thesis first provides a study of SMT and its applications. This study also reviews the state of the art of these SMT applications, and worldwide experience with the operation of WAMPAC in terms of system architecture, communication technologies and data management. After the basic study of WAMPAC, this thesis presents a new methodology for designing a roadmap that will ensure the future GB WAMPAC system will be developed in a logical and economic manner. This methodology takes into account the international experience with WAMPAC project management and the practical challenges faced in the future GB power system. With this new methodology, the GB strategies for the development of WAMPAC are devised. Two major SMT applications are then developed that can form main parts of the proposed future GB WAMPAC system. These applications are developed to enhance the small signal stability of the future GB power system. 1. Wide Area Inter-area Oscillation Monitoring using Newton Type Algorithm. 2. Wide Area Inter-area Oscillation Control using Power Electronic Devices. Finally, the operation of a proposed GB WAMPAC system is demonstrated using the DIgSILENT software package. The proposed real time applications are tested and evaluated using dynamic simulations of a full GB power system model. In addition, some key factors that will influence the operation of the future GB WAMPAC system will be analyzed and discussed. -13- Preface List of Abbreviations WAMS WAMCS WAMPAC GPS IED SMT PMU DC ISO EMS SCADA SIPS SPS WECC EIPP ADSL VPN TCP UDP IP SE PE RTDMS FNET VIP UVLS UFLS KF FFT LS NTA HVDC CSC VSC TCSC SVC AVR TG PSS SPTN SHETN NGETN Wide Area Monitoring System Wide Area Monitoring, Control System Wide Area Monitoring, Protection and Control Global Positioning System Intelligent |Electronic Device Synchronized Measurement Technology Phasor Measurement Unit Data Concentrator Independent System Operator Energy Management System Supervisory Control and Data Acquisition System Integrity Protection Scheme System Protection Scheme Western Electricity Coordinating Council Eastern Interconnection Phasor Project Advanced Digital Signal Link Virtual Private Networks Transport Control Protocol User Datagram Protocol Internet protocol State Estimation Parameter Estimation Real Time Dynamic Monitoring System Frequency Monitoring Network Voltage Instability Predictor Under-Voltage Load Shedding Under-Frequency Load Shedding Kalman Filter Fast Fourier Transform Least Square Newton Type Algorithm High Voltage Direct Current Current Source Converter Voltage Source Converter Thyristor Controlled Series Compensator Static Var Compensator Automatic Voltage Regulator Turbine Governor Power System Stabilizer Scottish Power Transmission Network Scottish Hydro-Electric Transmission Network National Grid Electricity Transmission Network -14- Preface Declaration 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. -15- Preface 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 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, trade marks 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://documents.manchester.ac.uk/DocuInfo.aspx?DocID=487), in any relevant Thesis restriction declarations deposited in the University Library, The University Library’s regulations http://www.manchester.ac.uk/library/aboutus/regulations) University’s policy on Presentation of Theses. -16- (see and in The Preface To my mother Yumei, my father Baotian, my girl friend Laura and Professor Terzija -17- Preface Acknowledgements The Research and Development (R&D) project, “Wide Area Monitoring, Protection and Control (WAMPAC) in the Future GB Power System”, presented in this thesis is funded by the National Grid Ltd (UK), Scottish Power Ltd and Scottish Southern Electricity Ltd.. This R&D project started in April 2008 under the direct supervision of Professor Vladimir Terzija at the School of Electrical and Electronic Engineering, the University of Manchester. First of all, I would like to express my great appreciation to my Ph.D. supervisor, Professor Terzija, for giving me this great opportunity to be his Ph.D. student. I have no doubt to claim that this Ph.D. has hugely changed my life, and will benefit me and my family in the rest of my life. Professor Terzija has been constantly involved in my research, and provided me with strong guidance and support all through the project. I wish to express my gratitude for his constant help. I also would like to thank my Ph.D. advisor, Professor Peter Crossley and external project supervisor, John Fitch and Mark Osborn, for their sound advice and kind help. Thanks to all my friends, in particular, Gustavo Valverde, Jairo Quiros, Pawel Regulski, Peter Wall and Gary Preston for their help and deep friendship. Lastly, my parents and my girlfriend, Laura Guo, have unwaveringly believed in my efforts. Their constant support provides me with confidence and courage to ultimately complete my thesis. -18- Chapter 1 Introduction Chapter 1 Introduction 1.1 Research Background 1.1.1 Power system blackouts Despite large-scale power system blackouts being very low probability events, their study is of great interest, due to the immense costs and consequences of such events for customers, societies and industries [1]. In previous decades, due to economic pressure from electricity markets and environmental constraints, power system operators have been forced to operate power transmission systems in highly stressed conditions closer to the system limits than ever before [1]. In this same period, the number, and size, of large-scale power system blackouts has increased. For example, the US-Canada blackout on August 14, 2003 [2] and the Italy blackout on September 28, 2003 [3] involved more than 100 million customers. Figure 1.1 presents the consequences, in terms of customers affected, of significant blackouts. Customers Affected 19 65 ,N E US 19 67 ,N E 19 US 77 , N De ew c. Yo 19 rk 94 ,W Ju e ly st 19 US 96 Au ,W g. 19 es 96 tU ,W S es t er 20 n 03 US ,U SCa na da 20 03 ,I ta 20 ly 03 ,S we de n 20 03 ,C hi le 20 04 G re ec e 60,000,000 50,000,000 40,000,000 30,000,000 20,000,000 10,000,000 0 Figure 1.1: Statistics of blackouts: customers affected [1]. It is rare for large-scale power system blackouts to be directly caused by a single large disturbance. However, a single large disturbance in a stressed system may cause a series, or cascade, of unplanned and unexpected sequential events. These events will incrementally increase the stress on the system and force it into a more vulnerable state of operation. If proper protection and control actions are not taken quickly and properly (e.g. load shedding, reactive power support and controlled islanding), then the system -19- Chapter 1 Introduction may experience further cascading events and separate into unplanned islands, or even completely collapse [4][5][6]. Figure 1.2: Line of separation from the European grid [3]. For example, on September 28th 2003 a sequence of events, which would lead to the separation of Italy from the interconnected European Power System, was triggered by the tripping of a Swiss 380 kV transmission line. The line between Mettlen and Lavorgo (marked as 1 in Figure 1.2), was tripped off at 03:01 due to a permanent fault. This tripping meant that other transmission lines began to carry the power that was previously transferred over the tripped line. This caused a second Swiss 380 kV transmission line, between Sils and Soazza (marked as 2 in Figure 1.2), to trip at 03:25 due to overload. Combined with insufficient reserve in Italy, the loss of these lines meant that the levels of overload on the remaining interconnections into Italy quickly became intolerable. This led to the automatic and, almost, simultaneous tripping of remaining transmission lines, with the effect that the Italian system was isolated from the European network. The Italian power system lost a large amount of the active power imported from its European neighbours (about 25% of the country’s total load). Such a large loss of active power caused a sudden frequency drop of approximately 1 Hz to occur in Italy. Furthermore, this significant loss of power caused multiple Italian generators to trip for various reasons, e.g. under-frequency relay operation, high temperature of exhaust gases. Despite additional load shedding, the frequency continued to decrease and the system collapsed in three minutes [3]. -20- Chapter 1 Introduction Whilst it is impossible to develop a solution that completely eliminates the possibility of a blackout, several measures can be implemented to minimize the probability of a blackout occurring. For many years, Energy Management System (EMS) has been used for the on-line monitoring of system conditions and assessment of system security. Traditional EMS uses measurements with a low-refresh rate (several seconds to one minute), from a Supervisory Control and Data Acquisition (SCADA) system, to estimate the system operating condition and to perform off-line system stability studies [7]. The EMS can provide sufficient information and support for normal steady-state system operation and to plan the system response to slow changes in the operating conditions. However, EMS is not capable of capturing system dynamics, particularly when the system is subjected to large disturbances. In addition, off-line studies cannot be used to fully anticipate all of the conditions faced by operators. These unplanned contingencies have the potential to initiate a cascade of events that will lead to a system blackout. System Integrity Protection Schemes (SIPSs) are designed to preserve system integrity after a large disturbance, and restore the system to the normal state when the system is in an emergency condition [4] [5]. Traditionally, these schemes use the results of offline studies to determine their actions [7]. These studies are based on the pre-calculated system behaviour for the assumed operational state of the system. In addition, as traditional SIPSs only use local or regional (within a power utility) measurements they lack awareness of the operating conditions in the neighbouring power systems [4] [5]. Consequently, the traditional SIPS may not be sufficient to ensure proper control of any system instabilities that may occur. The US-Canadian and Italian blackouts provided very strong evidence that the lack of a real time dynamic wide area monitoring system and the lack of real-time optimal and centralized protection and control schemes across a large interconnected network, was the root cause of these large scale blackouts [2] [3]. Furthermore, attempts to avert climate change through the introduction of renewable energy policies will force radical changes in future power systems. The most significant of these is that a large percentage of electrical energy will be generated using renewable resources (wind, solar and tidal). For example, in the GB power system, the target is for approximately 45% of electricity to be generated using renewable resources by 2030 [8]. This will prove problematic as electricity generation using renewable resources is -21- Chapter 1 Introduction highly influenced by climatic conditions and the resulting intermittent nature of the renewable resources. This will make the operation of future power systems more variable and unpredictable. In addition, renewable energy generation and transmission requires the support of power electronic technologies such as HVDC and SVC. The use of these technologies will introduce further complexity and uncertainty into power systems [9] [10]. The introduction of further variation, complexity and unpredictability to power systems will dramatically increase the likelihood of large scale power system blackouts. This will prove a serious problem, as current systems already find themselves increasingly vulnerable to such blackouts in the absence of a real time wide area monitoring system. The development of a real time wide area monitoring system is essential if the future changes in our power systems do not change the current trend toward an increase in the number and size of large-scale blackouts. A real time wide area monitoring system will allow the introduction of optimal, real time protection and control schemes that can be used to counter the growing threat of large scale blackouts. With the breakthrough made in the field of Synchronized Measurement Technology (SMT) and the availability of high-speed communication channels it is now possible to implement a practical real time wide area monitoring system. These technological developments, combined with financial support from governments, will allow the emergence of real time, wide area monitoring, protection and control systems that will be able to ensure the security of future power systems in the face of an increasingly unstable operational environment. 1.1.2 Wide area monitoring, protection and control Wide Area Monitoring, Protection, and Control (WAMPAC) involves the use of wide area synchronized measurements, reliable and high bandwidth communication networks and advanced centralized protection and control schemes [6]. SMT and related applications are the essential element, and enabler, of WAMPAC. Presently, Phasor Measurement Units (PMUs) are the most accurate and advanced synchronized measurement technology available. They provide voltage and current phasors and frequency information synchronized with high precision to a common time reference, the Global Positioning System (GPS). The measurement functions of a PMU are based on numerical algorithms. These algorithms must be both computationally efficient and suitable for real-time applications, particularly when the measurements are used to support dynamic-response applications [6]. -22- Chapter 1 Introduction Figure 1.3 shows the main components and structure of a generalized WAMPAC system. In this system, the necessary synchronized voltage and current phasors are produced by PMUs. The measurement data from these PMUs is transmitted through a Wide-Area Network (WAN) and aggregated at one, or more Data Concentrators (DCs). The aggregate data is then stored locally in the DC before being transmitted to the various Application Software or Servers (ASS) of the different utilities. The main task performed by the DCs is alignment of the received PMU data; however, the opportunity also exists to perform additional pre-processing tasks before forwarding the data to ASS [6]. Utility’s WAN PMU_1 Utility’s WAN PMU_n PMU_1 PMU_n Figure 1.3: A Generalized WAMPAC system [6]. The necessity for WAMPAC has gained worldwide acceptance [6], and a number of WAMPAC systems have been established, or initialized, in different power utilities throughout the world. For example, a Real Time Dynamic Monitoring System (RTDMS) has been implemented in the Eastern North American bulk power system. A wide area inter-area oscillation monitoring and control system was established by China South Power Grid [11]. Other countries, such as Switzerland, Sweden, Denmark, Austria, and Japan, have developed SMT based applications to improve power system stability [12]. -23- Chapter 1 Introduction 1.2 Objectives of the Research Over the coming decades, the biggest change of the GB power system is that a large percentage of electrical energy will be generated from off-shore wind farms. Majority of the off-shore wind farms will connect to the power grid with back-to-back HVDC link; hence, they will not provide inertia to the system. Coupled with replacement of the conventional coal generators with the new wind farms, the inertia of the future UK power system will be reduced. In addition, a number of Power System Stabilizers (PSSs) installed on the synchronous generators will be out of service when conventional coal generators will be replaced by wind farms. The reduction of system inertia and PSSs will reduce the small signal stability of power systems, e.g. lightly or unstable damped inter-area oscillations. Therefore, the main objectives of this research are to propose a WAMPAC system to improve the inter-area oscillatory stability in the future GB power system and develop solutions to overcome the challenges that are involved in the initial stage of the GB WAMPAC project. However, the thesis will not be focused on the assessment of future GB systems with wind farms, e.g. the system inertia reduction caused by the high penetration of wind farms connected to the grid over power electronic devices. The capital investment and logistical effort necessary for the implementation of a WAMPAC system, particularly for a network with the size and complexity of the GB power system, mean that it is infeasible to implement it in a single step. Therefore a well planned roadmap is necessary to ensure the proposed WAMPAC system is developed in a logical and economic manner. The design of the roadmap should seek to ensure that the various elements of WAMPAC functionality become available as soon as is reasonably possible, to maximize the benefit offered to the system operator. This requirement will lead to the roadmap being separated into multiple stages. Each stage will focus upon making certain elements of functionality (e.g. system monitoring applications) available to the operators, whilst ensuring that the actions taken during this stage are not short term and will lead to needless future redundancy and waste. This means that each stage will serve as a base for the implementation of more complex functionality at a future stage. -24- Chapter 1 Introduction The development of such a roadmap is a significant challenge, and as such it constitutes the primary task for this research project. As different power utilities have different requirements of WAMPAC, a generic methodology for determining the roadmap is required. This methodology is based upon an assessment of worldwide experience with building WAMPAC systems and the analysis of operational challenge of power systems. Using the roadmap developed from this research, a proposed future GB WAMPAC system will be created. To support the implement of the proposed WAMPAC system in the future GB power system, essential WAMPAC applications and algorithms for improving the inter-area oscillatory stability in the future GB power system are needed to be developed in this research. In addition, these WAMPAC applications, algorithms and the WAMPAC system’s architecture designed for serving these applications should be tested in a software package before being implemented in a real system. Therefore, another main aim of this research is to develop a platform in the DIgSILENT PowerFactory software package for the evaluation of the proposed future GB WAMPAC system. Using system models constructed in this software package, the key factors that will influence the operation of the future GB WAMPAC can be analyzed. The results obtained through these simulations will serve as a base for future work, e.g. demonstration of the operation of the GB WAMPAC system in a Real Time Digital Simulator (RTDS). The main objectives of this research can be summarized as follows: 1) Review of existing WAMPAC solutions, from applications, system architecture and technology point of view. 2) Developing a generic methodology for determining the roadmap for the development of the future GB WAMPAC system. 3) Developing new WAMPAC applications and essential algorithms for improving the inter-area oscillatory stability in the future GB power system. 4) Establishing a testing platform in DIgSILENT PowerFactory for demonstrating and evaluating the operation of the GB Wide Area Monitoring and Control System (WAMCS). -25- Chapter 1 Introduction 1.3 Thesis Structure Chapter 1 – Introduction This chapter briefly introduces the history of power system blackouts and summarises the root causes of these blackouts. Based on the evaluation of existing solutions for power system monitoring, protection and control (EMS and SIPSs), and the anticipation of developments in future power systems, it is determined that a SMT-based WAMPAC system is the only technology that could be used to reliably manage the next generation of power systems. The aims of this research are detailed in this chapter along with a list of the main contributions. Chapter 2 –Synchronized Measurement Technology This chapter briefly introduces the concept and technology behind synchronized phasor measurements. This includes the hardware and functionality of SMT devices such as Phasor measurement Unit (PMU) and Data Concentrator (DC). It will conclude with the details of major commercial PMUs and DCs. Chapter 3 – Applications and Benefits of Synchronized Measurement Technology A study of the major applications of SMT will be provided in this chapter as well as an evaluation of the state of the art and worldwide experience with these applications. Chapter 4 – Architecture of a WAMPAC System This chapter introduces the architecture of a typical WAMPAC system. This covers the core components of a WAMPAC system, i.e. the measurement devices, communication technologies, and their connectivity. The architecture for a future GB WAMPAC system is then constructed based on international experience with the operation of WAMPAC. Chapter 5 – The Roadmap to the Future UK WAMPAC System In this chapter, a methodology for designing a roadmap for the GB WAMPAC project is introduced. This methodology takes into account the international experience with WAMPAC project management and the practical challenges faced in a future GB network. With this methodology, the GB’s strategies (both short term and long term) for the development of a GB WAMPAC system are devised. -26- Chapter 1 Introduction Chapter 6 – The Physical Nature of Inter-area Oscillations The strategy for a GB WAMPAC project has highlighted that a real time inter-area oscillation monitoring and control system is an important SMT application for the future GB system. As such, a fundamental study of inter-area oscillations is provided in this chapter. Two classical methods are introduced to investigate the nature of inter-area oscillations i.e., nonlinear simulations and modal analysis. Chapter 7 – Inter-area Oscillation Monitoring Using Newton Type Algorithm In this Chapter, a new inter-area oscillation monitoring method developed for the short term strategy of the GB WAMPAC project is presented. The core of this novel Wide Area Monitoring System (WAMS) application is a nonlinear numerical algorithm, Newton Type Algorithm (NTA) that processes real time oscillatory signals to estimate the dominant inter-area oscillatory mode. Two data sets are tested using the new algorithm, one based on simulated models and the other based on real-life data records. Chapter 8 – The Application of Power Electronic Devices for Damping Inter-area oscillations In this Chapter, several closed loop control schemes that use power electronics devices i.e. HVDC, TCSC and SVC to stabilize inter-area oscillations are presented. A modal analysis based linear control theory is used for tuning the parameters of the damping controllers; and then these damping controllers are tested through dynamic simulation in a typical two-area system model. Chapter 9 – Wide Area Monitoring and Control System in a future GB Power System In this Chapter, a proposed Wide Area Monitoring and Control System (WAMCS) for the future GB power system is presented. This WAMCS is designed to enhance the small signal stability of the future GB power system, i.e., improve the damping of the inter-area oscillatory mode between Scotland and England. The operation of the WAMCS will be demonstrated in the DIgSILENT software package. Some key factors that will influence the operation of the future GB WAMCS will be discussed, including the time delay involved in the wide area data transmission, and the reactions between the new wide area control system and conventional Power System Stabilizers (PSSs). -27- Chapter 1 Introduction Chapter 10 – Conclusions and Future Work The last chapter presents the major conclusions of this research and suggests further directions for future research. 1.4 Main Contributions of This Research 1) Review of the state of the art of SMT applications and the worldwide experience of the operation of WAMPAC systems. 2) Construction of an architecture prototype for the future GB WAMPAC system, based on the international experience with WAMPAC and the likely characteristics of the future GB power system. 3) Introduction of a methodology for designing a roadmap to implement WAMPAC in the future GB power system. 4) Proposal of the UK’s strategies (short term and long term) to guide the development of the future GB WAMPAC system. 5) Development of a novel nonlinear numerical algorithm, Newton Type Algorithm (NTA), for identifying dominant inter-area oscillation mode. 6) Modelling of Power electronic devices, HVDC, TCSC and SVC. 7) Proposal of a detailed procedure for the design of a wide area inter-area oscillation damping control system using power electronic devices, i.e. HVDC, TCSC and SVC. 8) Establishment of a testing platform in the DIgSILENT software package for demonstrating and evaluating the operation of the GB Wide Area Monitoring and Control System (WAMCS). -28- Chapter 2 Synchronized Measurement Technology Chapter 2 Synchronized Measurement Technology 2.1 Introduction The voltage phase angles at the buses in an electrical power transmission network have always been of special interest to power system operators. It is well-known that active power flow over a transmission line is nearly proportional to the sine of the angle difference between the voltages at the two terminals of the transmission line. As many of the planning and operational considerations in electrical power systems are directly concerned with the active power flow, measuring voltage angle differences across the transmission line has been of concern for many years [14]. Consider a pure sinusoidal signal as shown in Figure 2.1. If the observation of the signal begins at the time t = 0 s, the signal can be represented by a complex number with a magnitude equal to the Root Mean Square (RMS) value of the signal and with a phase angle. In a digital measuring system, the samples of the waveform in one period are collected, and then the fundamental frequency component of the signal can be calculated by using the following equation [14]: 2 N X = xk e − j 2 kπ / N ∑ N k =1 (2.1) where N is the total number of samples in one period, X is the phasor representing the sinusoidal signal, and x k is kth sample of the signal. In a real measurement system, the signal is continuously sampled and each time a new sample is acquired a new phasor is produced as the data window is moved to include the new sample. The most efficient method for dealing with continuous monitoring of the input waveforms is to use a recursive form of the phasor equation [15]. -29- Chapter 2 Synchronized Measurement Technology φ Xm φ Figure 2.1: Phasor representation of a sinusoidal signal [16]. For the evaluation of the performance of a real power system, the positive sequence voltages and currents are far more useful than the single phase quantities. Positivesequence voltages of a network constitute the state vector of a power system, and it is of fundamental importance in all forms of power system analysis. The positive sequence phasor can be computed according to its definition: X1 = 1 ( X a + αX b + α 2 X c ) 3 (2.2) As illustrated in Figure 2.2, when the voltages and currents in different substations are measured and converted to positive sequence phasors in this way, their phasors can be put in the same phasor diagram. Substation B Substation A At different locations Phasor A Phasor B Figure 2.2: Synchronized phasor measurement in remote substations [16]. In a large power system, if all the substations are equipped with synchronized phasor measurement devices, the real time system operating condition can be directly measured rather than estimated; including the system frequency, the voltage phasor of each bus and the power flow between substations. The first paper to identify the importance of synchronized phasor measurement technology was published in 1983 [17]. At this time -30- Chapter 2 Synchronized Measurement Technology the Global Positioning System (GPS) [18] was beginning to be fully deployed and offered a source for the necessary synchronising clock signal. The timing pulse offered by GPS is accurate to within 1 microsecond at any location on earth. It became clear that GPS offered the most effective way of synchronizing power system measurements over long distances. 2.2 Phasor Measurement Unit At present, phasor measurement units (PMUs) are the most accurate and advanced synchronized phasor measurement equipment. Figure 2.3 gives a functional block diagram of a typical PMU. The GPS receiver provides the 1 pulse-per-second (pps) signal, and a time tag consisting of the year, day, hour, minute, and second. The l-pps signal is usually divided by a phase-locked oscillator into the number of pulses per second required for the sampling of the analogue signals. The analogue signals are derived from three-phase voltage and current transformers with appropriate anti-aliasing filtering. The microprocessor calculates the positive sequence voltage and current phasors, and determines the timing message from the GPS, along with the sample number at the beginning of a window [16]. Figure 2.3: A functional block diagram of a typical PMU [16]. If enough PMUs can be installed across a large power transmission network, the real time system operating condition can be directly measured by PMUs. In addition, as the PMUs have a high data reporting rate, the system dynamics can be captured when the system is subjected to disturbances [19]. Figure 2.4 compares the voltage angle difference between two substations obtained using PMU measurements and traditional state estimation. This comparison demonstrates clearly that a real time monitoring -31- Chapter 2 Synchronized Measurement Technology system made up of PMUs will provide much more precise and dynamic system operation information than the traditional state estimation. Figure 2.4: State estimation v.s. PMU measurements [16]. In a real implementation of synchronized measurement technology, there are two sorts of PMUs, standalone PMUs and integrated PMUs. A standalone PMU is a device that performs dedicated, high accuracy, time stamped, precision synchronized measurement tasks in a standalone device. Figure 2.5 shows two commercial standalone PMUs. Macrodyne 1690 ABB RE8521 Figure 2.5: Two standalone PMUs [16]. An integrated PMU is an Intelligent Electronic Device (IED), in which synchronized precision measurement tasks are integrated, such as a digital relay, a digital meter, a digital fault recorder. Figure 2.6 shows two integrated PMUs. As the cost of realising synchronized measurement functionality in an IED has become much affordable, and will continue to fall, IEDs with synchronized measurement capability will become more prevalent in the years to come. These integrated PMUs will be installed in large quantities in power systems, mainly for their primary functionalities (protective relaying, -32- Chapter 2 Synchronized Measurement Technology electricity usage metering, fault event recordings, etc.) rather than to offer pure monitoring. However, whatever the primary task of these integrated PMUs the result of their deployment will be that a substantial number of IEDs in a power system will be capable of offering synchronized measurements [7]. GE N60 SEL 421 Figure 2.6: Two integrated PMUs [16]. Table 2.1 gives a summary of available standalone and integrated PMUs, including manufacturers, model, and the number of inputs and outputs [16]. Table 2.1: Summary of available standalone and integraed PMUs Manufacturer ABB Model RES521 Arbiter 1133A Arbiter GE Macrodyne SEL SEL SEL 933A N60 1690 SEL-311 SEL-421 SEL-451 SEL AMETEK_RIS QualiTrol REASON REASON SEL-734 TR2000 Q9 RPV304 RPV310 Other functions No Revenue Meter, Power Quality Portable Power Quality Meter Relay No Relay Relay Relay Revenue Meter DTR No DFR DFR Internal GPS receiver Yes Number of input channels (positive sequence phasor outputs) 18(6) Yes 9(3) No Yes Yes No No No 9(3) 16 (5) 30 (10) 12 (4) 12 (4) 12 (4) No Yes Optional No No 12 (4) 32 (10) 16 (5) 16 (5) 64 (21) -33- Chapter 2 Synchronized Measurement Technology 2.3 Data Concentrator …… ASS Z ASS A DC WAN …… SMU_1 SMU_n Figure 2.7: Data concentrator in a WAMPAC system [16]. As shown in Figure 2.7, Data Concentrator (DC) is another critical ‘building block’ of a SMT-based WAMPAC system. The DC collects the synchronized phasor data and aligns that data into a single data packet for each unique time stamp; it then forwards this data to different SMT applications. The data concentrator may also include other functions, such as system event detection and archiving, data reprocessing for various applications and data calibrations. Figure 2.8: SEL Data Concentrator [16]. Figure 2.8 gives a representative view of a commercially available DC from Schweitzer Engineering Laboratories (SEL). In addition, other power system device manufacturers and power utilities have also developed their own DCs, such as PSGuard 830 of ABB, -34- Chapter 2 Synchronized Measurement Technology the DC developed by Bonneville Power Administration (BPA) and the “Super-DC” developed by the Tennessee Valley Authority (VTA) in the USA. 2.4 Synchrophasor Standard Actually, as shown in Figure 2.8, in a large-scale SMT-based WAMPAC system there will be PMUs from different vendors, as the deployment of the system typically involves multiple entities. Therefore, a WAMPAC system requires a consistent performance from all of the PMUs installed, to meet its application requirements. To ensure a uniform performance among the PMUs from different manufacturers, establishing a standard is an essential step for initializing a WAMPAC system. The IEEE PES Power System Relaying Committee has published the IEEE C37.118-2005 standard to replace the original IEEE 1344 standard for synchrophasor measurements. It addresses the definition of a synchronized phasor, time synchronization, application of time tags, a method for verifying measurement compliance with the standard, and message formats for communication. However, the Standard has not addressed the dynamic performance requirement of a PMU. Also, the Standard does not specify the method with which conformance tests should be conducted [6]. 2.5 Summary In this chapter, the concept and technology of synchronized phasor measurement has been breifly introducted. In [14] the algorithms behind synchrnized phasor measurement, including the phasor estimation with nominal frequency inputs and the phasor estimation with off-nominal frequency inputs are detailed. The hardware and functions of PMUs and DCs have also been introduced, as well as the major commercial PMUs and DCs. To ensure a consistent performance among the PMUs from different manufacturers in a WAMPAC system, synchrophasor measurement standards were developed. So far, the IEEE standard C37.118-2005 has replaced the original IEEE standard 1344. However, more detailed standard conformity testing must be developed, based on a clearly defined test requirement, setup, and procedure. It is also important that qualified test entities are established, to ensure that the tests themselves are executed consistently. -35- Chapter 3 Applications and Benefits of Synchronized Measurement Technology Chapter 3 Applications and Benefits of Synchronized Measurement Technology 3.1 Introduction Synchronized Measurement Technology (SMT) is the most essential element and enabler of WAMPAC. All of the advanced functions in a WAMPAC system are developed based on synchronized measurements. Therefore, in this chapter, a study of the major applications of SMT will be provided as well as an evaluation of the state of the art and worldwide experience of these applications. From this study, power utilities that expect to develop WAMPAC systems will be able to have a clear understanding of how SMT can be used to help improving the stability of modern electrical power system operation, and how SMT will be essential to the future power systems. Access to this information will allow power utilities to be more confident when developing their own WAMPAC systems. Generally, SMT applications are divided into two groups, off-line applications and online applications. Off-line applications are usually used to improve and validate the system model, to make system planning and off-line stability studies more reliable. Online applications are used to assist system operators in real time, by offering on-line monitoring, protection and control. This study attempts to provide a comprehensive analysis of all major applications that are being developed. However, the lack of maturity of some applications and continuing development of others means that some applications are not fully covered by this study. 3.2 Off-line Applications of SMT 3.2.1 Post-disturbance analysis The main objective of post-disturbance analysis is investigating the system dynamics during large disturbances and analyzing the system sequential events caused by those disturbances. To achieve this, power system analysts collect and assemble the data recordings from various data recorders that are at different and remote positions throughout the entire network. However, the considerable amount of the data that has -36- Chapter 3 Applications and Benefits of Synchronized Measurement Technology been used for many years is not synchronized. Therefore, it has been extremely difficult and time-consuming to reconstruct this data on the same time axis. This reconstruction is a prerequisite for understanding the sequential events that have occurred during and after the disturbance [20] [6]. Application of SMT allows all of the data gathered during the system disturbance to be time tagged based on the same synchronizing GPS signal. Therefore, it is much easier to reconstruct the sequence of events after the disturbance has occurred. Simplifying the reconstruction process will allow the time spent analyzing the vast amount of data to be reduced from months, to days, or even hours [1]. Figure 3.1 presents one example from the Western Electricity Coordinating Council (WECC) of using PMU data for the reconstruction of the sequence of events that occurred after a large disturbance, this reconstruction was performed using the post disturbance analysis software, Power System Outlook (PSO). PSO was developed by Southern California Edison (SCE) Corporation [21]; it has been used by 11 states on the west coast of the USA as well as two Provinces in both Canada and Mexico. It helps power system analysts and operators to understand the dynamic characteristics of the system by producing visualizations of the real data recorded during disturbances. 06/14/04 Event at 07:40 Pacific Time (06/14/04 at 14:40 GMT ) 60.025 59.820 59.615 59.410 59.205 59.000 14:40:34.00 14:40:44.00 14:40:54.00 14:41:04.00 14:41:14.00 14:41:24.00 VINC JDAY DEVR MALN BGCR COLS ALAM BE23 SONG BE50 KRMR SYLM DEV2 MPLV ANTP KEEL VLLY CPJK MAGN SUML LUGO SLAT GC50 SCE1 14:41:34.00 Pacific Time Figure 3.1: The reconstruction of system frequencies after a large disturbance in WECC, 14th June, 2004 [16]. -37- Chapter 3 Applications and Benefits of Synchronized Measurement Technology 3.2.2 Benchmarking, validation and fine-tuning of system models When benchmarking and validating system models the main tasks are the identification of the potential errors in the power system model, and the calculation of accurate values for any erroneous parameters. Accurate and reliable system models are essential for reliable and secure system operation, planning, and delivery of efficient and robust power system control. Inaccurate system models may cause power system operators to make conservative or incorrect decisions of system planning and operation; this could result in inefficient utilization of assets, and even system blackouts. Figure 3.2 shows a typical example of the mismatch between the actual system response and the simulated response to the same disturbance [22] in WECC. It was the mismatch that caused system planners to make an incorrect decision and led the system to blackout after the occurrence of a large disturbance. Figure 3.2: Comparison of the recorded system response to the 10th August, 1996 disturbance in USA with the simulation results [22]. Here, power system model validation is not limited to the steady-state models; it also includes the dynamic models. In the aspect of the validation of static model parameters, which includes synchronous machine inner resistance and reactance, transformer tap position and transmission line impedance etc, the algorithms are well developed and commercially available which is commonly referred to as Parameter Estimation (PE) [20]. The emergence of precise SMT has significantly improved the performance of PE -38- Chapter 3 Applications and Benefits of Synchronized Measurement Technology algorithms, as the estimated phasor values in the PE algorithms can be directly replaced by measured phasor values. For example, with a PMU installed at both ends of a long transmission line, the transmission line impedance can be directly calculated using the voltage and current phasors [20]. In comparison to dealing with a static model, the benchmarking and fine-tuning of dynamic system models is much harder and more complex. This is because it requires careful evaluation of the actual system response to the events. Dynamic model validation is usually achieved by comparing the data recorded during system events with the response of system models to the same disturbances. When differences are encountered, the dynamic model parameters are tuned until a corresponding response is obtained [20]. The high data reporting rate offered by PMUs is capable of capturing the system dynamics, with a sufficient number of PMUs installed across a large power network the dynamic characteristics of the system can be precisely profiled by using the synchronized recordings. The procedures and principles that should be observed when validating power system models are described in [22] [23] [24] and [25], alongside algorithms that allow wide area synchronized measurements to be used for the finetuning of dynamic system models. 3.3 On-line Applications of SMT 3.3.1 Wide area phase angular and power flow monitoring Since PMUs can directly measure the phase angle differences across power transmission lines, they have an inherent advantage when system operators want to monitor the real time power transfer stress on the power transmission network. This real time monitoring allows the system operators a greater degree of confidence when managing critical transmission corridors; allowing operation closer to the real stability limit of the corridor, whilst still maintaining a safe security level. As a consequence, the marginal cost between two power generation areas will be reduced, as the constraints on the transmission of the electrical power that is produced by the generators with low generation costs will be relaxed. Furthermore, confident operation of transmission lines closer to their stability margins will also reduce the need for investment in transmission line reinforcement. -39- Chapter 3 Applications and Benefits of Synchronized Measurement Technology In addition, in large interconnected power networks, a real time phase angle monitoring system, consisting of PMUs, can provide clear understanding of the entire operational situation to the system operators. The August 14, 2003 blackout that occurred in the United States and Canada demonstrated the necessity to implement such a monitoring system throughout bulk interconnected power networks, because it was the lack of awareness of the neighbouring grids’ operational status that allowed the cascading blackouts to occur. After this blackout, the provision of a real time wide area monitoring system for all transmission owners and regional transmission operators was recommended [2]. Figure 3.3: RTDMS – a wide area visualization platform for the North American power system [26]. Figure 3.3 presents one application of wide area phase angular monitoring in USA, i.e, Real Time Dynamic Monitoring System (RTDMS) that was developed for the North American bulk power system. Assisted by this visualization tool, the system operators will make optimum and confident decisions when operating the system, since both the local and global operational situations are clearly observed. The RTDMS visualization tool also offers: 1) Voltage Magnitude monitoring, 2) System Frequencies monitoring, 4) -40- Chapter 3 Applications and Benefits of Synchronized Measurement Technology Real and Reactive Power Flow across Monitored Lines, 5) a Summarized Information Display [26]. 3.3.2 Wide area frequency monitoring Power system frequency is one of the most valuable information for on-line assessment of system stability, since the system frequency is the direct measure of the balance between generation and demand. During large disturbances, in particular, the system frequency is rapidly varying and very different in different parts of a bulk system. In the next generation of power systems, one of the biggest changes will be the high integration of renewable generation resources. This will reduce the system’s ability to provide frequency control services, because generation from renewable sources tends to be less controllable than conventional synchronous generators. Therefore, power system operators require an accurate wide-area frequency measurement system, and adaptive emergency frequency control scheme for remaining system frequency stability after power system being subjected to a large disturbance e.g. a sudden outage of large generator. The high data reporting rate offered by PMUs has afforded an opportunity for power system operators to obtain accurate measurements of the dynamic system frequency. If the entire power network is monitored by a synchronized frequency measurement system, then the dynamic frequency behaviour of the system can be precisely captured. The most important application of such wide area frequency information is the analysis of system disturbances (e.g. outage of a large generator), which includes the identification of disturbance locations and the estimation of the magnitude of disturbances. The results of such an analysis serve as the preliminaries for the power system emergency load shedding scheme [27]. The Triangulation method, which was traditionally used to detect the epicentre of earthquakes, has proven a simple and effective way of determining the location of a disturbance, such as the outage of a large generator [27] [28] [29]. This method has been utilized in the Frequency Monitoring Network (FNET) that is in use in the United States, and the disturbance locations estimated by the method were very close to the actual positions of the disturbances. The estimation results for a disturbance that occurred on August 4th, 2004 are presented in Figure 3.4. The red dot represents the estimated disturbance location, while the green square is the actual location. The same -41- Chapter 3 Applications and Benefits of Synchronized Measurement Technology algorithm for disturbance location was also tested in a small power network in South Korea, and the results were also acceptable [30]. Figure 3.4: Disturbance localization using PMUs and the triangulation method [27]. Load shedding is a key action when trying to help a system recover from extreme under-frequency conditions. The ability to immediately produce an accurate estimate of the power imbalance between the generation and demand after a disturbance will be a quite useful input for an adaptive load shedding scheme. The active power imbalance (ΔP) is proportional to the rate of frequency change (df/dt), the system inertia serves as the constant of proportionality in this relationship. The wide area frequency measurement system provides real time frequency and the rate of the change of frequency, based on which the magnitude of the disturbance can be estimated. In the case presented in Figure 3.4, the estimated tripped generation was 786 MW, while the actual tripped generation was around 870 MW [27]. 2.3.3 Wide area voltage monitoring Voltage instability is a well known problem for power networks that have long transmission lines with heavy power transfer, such as WECC [1]. Conventional approaches for preventing voltage instability, such as Voltage Instability Predictor (VIP) and Under-Voltage Load Shedding (UVLS) only use local voltage and current measurements. With these local measurements, an approximate two-bus equivalent of a -42- Chapter 3 Applications and Benefits of Synchronized Measurement Technology complex network and the generator at the remote end can be estimated, as shown in Figure 3.5. i v i Z th E th v ZL Figure 3.5: Estimation of the Thevenin equivalent with local measurements [31]. The Thevenin equivalent circuit and the maximum power transfer principle can be used to estimate the voltage collapse point of the equivalent network [32]. When the voltage of the local bus is close to the estimated collapse point a certain amount of load is disconnected, with a certain time-delay, until the voltage rises above the threshold [31]. However, the drawback of these conventional approaches is that a single set of local measurements do not contain enough information to directly compute the parameters of the Thevenin equivalent accurately, this inaccuracy may lead to a large error in the estimation of the voltage collapse point. Real time voltage monitoring software, supported by a network of PMUs that are installed at both end substations of a complex transmission corridor, can overcome the shortcomings of conventional VIP. In the first stage of the application, the T-equivalent of the transmission corridor can be directly computed using the voltage and current phasors that are measured by the installed PMUs, as presented in Figure 3.6. In addition, the source impedance is calculated by using dynamic data collected from the PMUs at the sending end of the corridor. Furthermore, the dynamic parameters of the load, such as the coefficient of voltage-dependence, are estimated by using the PMU data collected at the receiving end of the corridor. In the next stage, the T-equivalent of the transmission corridor is combined with the source impedance and the dynamic load model, calculated in the first stage. Once the real time combined model of the critical transmission corridor is available, the voltage stability analysis can be directly carried out [31]. ABB has developed a PMU-based voltage stability application for real time -43- Chapter 3 Applications and Benefits of Synchronized Measurement Technology voltage-stability monitoring and assessment, based on the methodology described above [33]. This application has been implemented in the Croatia power network [12]. v1 i2 i1 i1 ZT / 2 Zg E th ZT / 2 Z sh v1 v2 i2 v2 ZL Figure 3.6: T and Thevenin representation of transmission corridor [31]. 3.3.4 Inter-area oscillation monitoring To reduce the CO2 emissions produced from conventional generators (coal, oil and gas), electrical power engineers have been integrating as much renewable based generation (Hydro, wind and photovoltaic) as possible. The geographical location of renewable resources is fixed, and as such they are distributed throughout different power utilities. Previously isolated power networks have to be connected to form a larger power grid to share the renewable energy resources. Consequently, this has resulted in long distance power transfers between adjacent power grids; which carries the potential risk of lightly damped inter-area oscillations in a large interconnected power system. The inter area oscillation mode is used to define the perturbations associated with a group of generators in one area are swinging against a group of generators in another area. A large system usually involves several inter area oscillation modes. Most analyses of inter-area oscillations are performed off line. They use system models that are limited to individual power utilities, and do not consider the complete interconnected power system [1]. The typical frequency range of inter-area oscillations is from 0.1 Hz to 0.8 Hz [34]. Hence, it is extremely difficult to capture inter area oscillations by using conventional EMS, due to its low refresh rate. A wide area monitoring system (WAMS) that consists of PMUs can offer a great opportunity to monitor the dynamic behaviour of power systems, and identify the inter-area oscillatory modes, as presented in Figure 3.7 [1]. The high data reporting rate of PMUs and the -44- Chapter 3 Applications and Benefits of Synchronized Measurement Technology availability of fast communication links are the primary enablers of this opportunity to monitor the inter-area oscillations. So far, a considerable amount of work has been done by different power utilities in developing techniques to identify the dominant inter-area oscillation modes. For example, the Prony Method has been used in the southern power grid of China and the West coast power grid of the USA [35][36]. The Nordic power system has used an adaptive Kalman Filter (KF) to track the inter-area oscillations [37] [38], and the Fast Fourier Transform (FFT) has been used in the Japanese system for estimating the oscillatory parameters [39]. Figure 3.7: Oscillations observed by two PMUs in the European grid [1]. 3.3.5 Power system restoration It has to be accepted that some of power system blackouts are unavoidable because of its nature [14]. It then becomes essential that strategies for restoring power after a blackout with minimum delays and minimum cost should be made. Quick restoration of power is extremely important as it can significantly minimize user inconvenience due to electrical power outage and the cost of blackouts. The existing procedures for power restorations are made in the form of written manuals that prescribe steps to be followed during restoration, along with appropriate checkpoints along the way in order to verify that everything goes according to the plan [40] [41]. However, these guidelines are based on assumed system conditions, which -45- Chapter 3 Applications and Benefits of Synchronized Measurement Technology may not be the same as those encountered at the present status [1]. During power system restoration, power system operators often encounter excessive standing phase angle difference across the breakers that connect adjacent power grids. Incorrectly closing such a circuit breaker could shock the power system, causing high currents, voltage drops and severe equipment damage [1]. In addition, unsuccessful attempts to reclose the tripped power lines, rather than taking other actions, could cost valuable time when the power system is operating in the emergency state [3]. The main value of using PMUs in power system restoration schemes is the ability to provide the operator with real-time information about the phase angles in relevant parts of the interconnecting grids. This ability helps the operator to know if reclosing a circuit breaker can be done without affecting the stability of the system. When the angle differences across the breakers are within acceptable bounds, the system operator can safely reconnect the adjacent areas immediately. As a consequence, the time spent on power system restoration will be significantly reduced. Figure 3.8: PMU measurements from three areas during reclosing attempts: UCTE, 4 November 2006 [42]. -46- Chapter 3 Applications and Benefits of Synchronized Measurement Technology Figure 3.8 shows the PMU measurements recorded during the reclosing attempts for the power lines between two areas, including the successful reclosing between those two areas and a third area. Seven attempts to connect these zones failed, as the operators did not use the PMU data to assist the restoration. If the PMU data had been used during the restoration these unsuccessful reclosing attempts could have been avoided [42]. 3.3.6 Improved state estimation Power system State Estimation (SE) is one of the most important online applications needed for the Energy Management System (EMS) and operational security assessments. The state estimator produces the optimal solution of the system state (voltage magnitudes and angles) using network models and estimation methods, e.g. the weighted least square method and Kalman filter, the performance of which is based on the system models and redundant measurements (voltage and current magnitudes) in the system. A brief account of the benefits of using PMUs to support SE is provided below. Phasor-Based Linear State Estimation Synchronized and accurate phasor measurements can significantly improve the overall quality of state estimation. This improvement would be seen both at the bus where phasors are introduced, and at adjacent buses in which the states can be directly calculated using the transmission line parameters, and voltage and current phasors. As the number of PMUs is increased the observability will also increase, until complete observability is achieved and the estimation method will become linear. A linear estimation method would have a lower computational burden and greater accuracy than a non-linear method. Unifying Neighbouring State Estimations When state estimations of neighbouring systems perform independently they do not have a common reference bus. Therefore, the results cannot be used to form a single estimator for the entire network. However, if a PMU is installed at the reference bus in each of the neighbouring systems the voltage angle difference between the reference-buses can be directly measured. These angle differences can be used to create a common reference point for the state estimators across the entire interconnected system. -47- Chapter 3 Applications and Benefits of Synchronized Measurement Technology Real-Time Network Model Parameter Estimation In the traditional Parameter Estimation (PE), the system line parameters are estimated from available telemetries; a high level of redundancy and accuracy in the measurements used is essential for PE. Synchronized phasor measurements can be used to calculate the actual line impedances and charging admittances, as the voltage and current phasors are directly measured at both ends of the transmission line, as presented in Figure 3.9. i I ik g ik + jb ik I ki g g + jb si si sk k + jb sk Figure 3.9: PMUs installed at both ends of a transmission line [6]. 3.3.7 Dynamic rating of overhead transmission lines In electrical power systems, overhead transmission line capacity is limited by the performance of the conductor at high temperatures, and by safety standards that specify the minimum allowable ground clearance. A number of the overhead transmission lines in power systems are designed using conservative criteria. These conservatively designed lines will have excess capacity that may not be fully exploited. There is lack of practical technology that enables real time monitoring and dynamic rating of overhead transmission lines. Such technology will allow system operators to fully utilize the available transmission infrastructure with confidence [1]. This will be of particular benefit as the conservative use of overhead transmission lines may lead to unnecessary investment in new transmission lines. With PMUs installed at both terminals of an overhead line, the synchronized phasor measurements allow the actual impedance and shunt admittance to be calculated, from which the line resistance can be calculated. Based on the known characteristics of the conductor material, the conductor temperature can be estimated in real time from the line resistance. The advantage of using a PMU-based method for monitoring a line is the low cost and relative ease of installation, as only a pair of PMUs is needed for each line. -48- Chapter 3 Applications and Benefits of Synchronized Measurement Technology Figure 3.10: Visualization of the real time mentoring of the thermal condition of an overhead transmission corridor in APG [12]. A commercial product that offers this application, “Line Thermal Monitoring” from ABB, has been installed at two locations in Europe, one in the Austrian Power Grid (APG) and another in Switzerland [12]. Figure 3.10 shows the overhead line monitoring system in APG. It provides power system operators with real time information about the network stress caused by heavy power transfer. Upon the detection of an extraordinary status, the monitoring system alerts the operator by giving an early warning or emergency alarm [12]. 2.3.8Intelligent controlled islanding Power system islanding is a measure of last resort that is employed when the system operation is in an extremely stressed state (thermal limits, phase angles across system beyond limits, voltage or frequency excursions beyond planned thresholds) and further disturbance propagation that will result in uncontrolled system separation is unavoidable. System islanding is conventionally accomplished by using a System Integrity Protection Scheme (SIPS), also known as remedial System Protection Schemes (SPS) [4] [5]. These schemes are designed based on extensive planning studies for a range of scenarios that cover various loading levels, topologies, contingencies, etc. However, in many practical situations the prevailing system conditions are quite different from those upon which the SIPS settings are based. Consequently, the performance of the system will not be optimal for the prevailing system state. -49- Chapter 3 Applications and Benefits of Synchronized Measurement Technology Wide area real time synchronized measurements provide important information about the prevailing system conditions that can be used to improve the action of SIPS. These synchronized measurements may be used to improve conventional system islanding schemes in the following areas [14]. 1. Using real time data provided by PMUs to detect whether a power system is approaching an unstable state, e.g. the phase angles across the system beyond limits, and if network islanding is needed to prevent a blackout. 2. Determine optimal islanding boundaries according to the prevailing system conditions. For example, establish which groups of generators will separate due to the loss of synchronism and how to optimally balance generation and demand in each island. 2.3.9 Adaptive under-frequency load shedding Under-frequency load shedding (UFLS) is a common practice that is used by most power systems to help prevent extreme frequency drops when a large disturbance occurs, such as a sudden outage of a large generator [43]. Nowadays, the existing UFLS protection scheme adopted by power utilities are predominantly deterministic, not taking into account the actual system state and topology, operating point, and the nature or magnitude of the disturbance. The frequency measurements are performed locally in distribution substations, when the frequency drops below the preset point the designated load feeder will trip [44]. As a consequence, the existing UFLS protection scheme very often disconnect more or less load, than is required. The availability of synchronized wide area measurements provides a great opportunity for developing an adaptive and efficient UFLS protection scheme. The adaptive UFLS protection scheme is based on real time wide area measurements of system frequency and the rate of change of system frequency. With known system inertia constant (H) the magnitude of the imbalance between generation and demands can be immediately estimated after the disturbance [45]. In addition, the wide area monitoring system can also provide other auxiliary information to support the adaptive UFLS protection scheme, for example, real time active power outputs from renewable resources, power generation reserve from synchronous generators and the power flow conditions over the tie-lines that connect adjacent power grids. This real time wide area information will -50- Chapter 3 Applications and Benefits of Synchronized Measurement Technology allow system operators to be more confident when optimizing the UFLS protection scheme. 3.4 Conclusions In this chapter, the major applications and benefits of Synchronized Measurement Technology (SMT) have been introduced in terms of off-line and on-line applications. From worldwide experience, off-line applications are usually considered to be the first step when developing a WAMPAC system; these applications include 1) accurate post disturbance analysis and 2) system model validation and benchmarking. Concerning online applications, most power systems have utilized synchronized phasor data to develop wide area real time monitoring systems; these monitoring systems include real time power flow (power angle), voltage magnitude monitoring, system frequency monitoring and inter-area oscillation monitoring. In addition, SMT offers opportunities for improving the quality of conventional power system state estimation and protection. However, other applications, such as real time power system restoration, smart controlled system islanding and adaptive load shedding, are still in the research stage, significant long term effort will be necessary before these research applications are ready for practical implementation. -51- Chapter 4 Architecture of a WAMPAC System Chapter 4 Architecture of a WAMPAC System 4.1 Introduction In Chapter 2 and 3, the Synchronized Measurement Technology (SMT) and its major applications and benefits were introduced. With this information, power system operators become confident that SMT will significantly improve the stability of future power systems, and that a SMT-based WAMPAC system will be the only way for monitoring and managing the future power systems. In this Chapter, the architecture of a WAMPAC system will be introduced. This introduction covers the core components of a WAMPAC system, e.g. measurement devices, data concentrators and communication technologies, and their connectivity. A prototype of architecture for the future UK WAMPAC system will be constructed based on the current international experience and practice with the operation of WAMPAC. 4.2 Architecture of a WAMPAC System Designing a large-scale WAMPAC system is extremely difficult; meticulous work need to be done to overcome the unique challenges that will exist for each individual WAMPAC system. First of all, a large-scale WAMPAC system involves a large number of entities (e.g. power utilities, Independent System Operators (ISOs), regional organizations). Each entity has its own specific needs that they wish to have satisfied by the PMU applications. Secondly, the WAMPAC system needs to support a large range of applications to satisfy the participants, thus it must accommodate a diverse range of technical requirements, such as data reporting rate, reliability of communication network. Thirdly, as more power utilities become aware that SMT is essential for the next generation of power systems, a number of new clients will join the WAMPAC system and the number of PMUs will increase. Therefore, the system must be scalable and flexible, so that it is capable of accommodating increasing quantities of PMU data and the requirements of the new WAMPAC participants and clients [1]. As many PMU applications are still in the research and development stage, the communication standards that will serve the WAMPAC system are not fully developed; and the data requirements for supporting each application are still not clearly defined [1]. -52- Chapter 4 Architecture of a WAMPAC System Therefore, there still hasn’t been a developed and generic architecture of WAMPAC system. However, based on the current experience of the operation of existing WAMPAC systems, a general architecture for a typical WAMPAC system has been defined; this is shown in Figure 4.1. Although this general architecture doesn’t provide many details about how a real WAMPAC system works, it presents the essential components involved in a real WAMPAC system. The results of this work will help new WAMPAC systems be built logically and economically. Communication WAN/LAN networks Utility’s applications Communication networks WAN/LAN Utility’s DC Utility’s applications Utility’s DC ` ` SCADA/ EMS Communication networks WAN/LAN SCADA/ EMS ISO’s Super DC ` Data server SCADA/ EMS Event archive Short term archive Wide area real time protection and control Wide area real time monitoring Data Archiving Long term archive Figure 4.1: The general architecture of a typical WAMPAC system. This general WAMPAC system architecture presented in Figure 4.1 was mainly based on the experience of the real WAMPAC system operating within the Eastern Interconnection Phasor Project (EIPP) community in the USA [46] [47]. Figure 4.1 shows a four–layer architecture that is typical in the operation of WAMPAC system, 1) Synchronized Phasor data acquisition 2) Synchronized data collection 3) Data services 4) -53- Chapter 4 Architecture of a WAMPAC System Synchronized measurement applications. The main equipment and functionality within each layer are briefly described as follows. Layer 1, Synchronized Phasor Data Acquisition – The primary function of Layer 1 is to acquire synchronized phasor data. In this layer, PMUs measure three-phase instantaneous voltage and current magnitudes and produce positive-sequence, voltage and current phasors for the fundamental frequency. Layer 2, Synchronized Data Collection – In the Layer 2, the time-stamped phasor measurements produced in Layer 1 are collected by Data Concentrators (DCs) and then put into a single data packet for each unique time stamp. In a large WAMPAC system, covering a number of different power utilities, each participant has its own DC. A centralized “Super DC” is then needed for the Independent System Operators (ISOs) to receive the data streams from these local DCs and other unattached PMUs, and package them together for broadcast to the WAMPAC centre and other participants. Layer 3, Data Services – The data server in Layer 3 is primarily responsible for ensuring that the data supplied to the applications in Layer 4 is suitable for their purposes. For example, the inter-area oscillation monitoring toolbox only needs the synchronized data from the PMUs at the ends of the long transmission corridor, but with a very high data-reporting rate. A linear state estimator may need data from hundreds of PMUs but a much slower data-reporting rate. The data server must also provide data processing services, e.g. error filtering, noise filtering and synchronization checks. Layer 4, Synchronized Measurement Applications – Layer 4 consists of the WAMPAC centre in which the major real time wide area SMT applications perform; these include: • Large–scale data archiving (short term archive, long term archive, event archive); • Improved EMS (improved state estimation, etc); • Real time wide area monitoring (power angle, voltage, frequency); • Real time wide area control (inter-area oscillation damping control, etc); • Real time wide area protection (under-frequency load shedding, intelligent controlled system islanding, etc). -54- Chapter 4 Architecture of a WAMPAC System 4.3 Communication Networks of WAMPAC System A future WAMPAC system requires transmission of huge amounts of data between power utilities, organizations and clients. Therefore, it is essential to establish a secure, reliable communication network. Some of the important issues involved in developing such a communication network are discussed in this section. 4.3.1 Available communication media for WAMPAC 1. Telephone Lines – At present, telephone lines are still the backbone of utility communication. The main advantages of using telephone lines for synchronized data transmission between substations are that they are easy to set up and economical to use [48]. A major drawback of this medium is its narrow bandwidth (<=56bps). However, advancements in digital communication technologies, such as Advanced Digital Signal Link (ADSL), have allowed the bandwidth of telephone lines to be increased significantly [49]. Usually, the telephone communication and internet infrastructure are owned by communication corporations not power utilities. Therefore, the reliability of using a telephone network as the backbone of the communication network for a WAMPAC system can not be guaranteed. 2. Power Lines – Power Line communication is a fast growing technology serving a number of current applications in power system protection and control. However, PL also has some disadvantages, such as bandwidth limit, propagation delay and induced electromagnetic interference. Furthermore, Power Line is subject to lightning, switching surges and power network reconfigurations [50]. 3. Satellite Communications – The major advantage of satellite communications, when compared to phone lines and power lines, is that there is no physical connection between the signal source and the destination. A satellite receives microwave signals from ground stations on the Earth’s surface, and transmits them at a different frequency to other ground stations, potentially thousands of miles away. This would prove quite suitable for data transmission between remote substations [50]. -55- Chapter 4 Architecture of a WAMPAC System 4. Microwave links – Point to point microwave link communication is highly reliable and easy to set up. Furthermore, it is capable of carrying many communication channels with a variety of information. However, Microwave communication technology is affected by electromagnetic interference from power transmission lines, wind turbines and cell phone transmission towers, etc. Microwave communication is also affected by heavy moisture, snow, rain and other bad weather conditions [50]. 5. Fibre-optic links – Many utilities have been using fibre-optic cables for power system protection communications, and the demand for this medium is growing very quickly. The advantages of using fibre-optic links are their immunity to electromagnetic and atmospheric interference and the massive bandwidth available. However, the financial cost of building a dedicated fibre-optic communication network for power system protection and control is extremely high [48]. To build a cost effective communication network for a WAMPAC system, the selection of communication media should take a comprehensive set of factors into account. For example, the financial cost to build a new communication links, the reliability and bandwidth of the current communication networks and the requirements of communication technology for different phasor data applications. The different demands placed upon the communication network by each application will be the key factors in determining the nature of the optimum communication network. For example, the acceptable level of PMU data transmission delay for a PMUintegrated State Estimator is several seconds, and the short-term absence of PMU data, does not influence the performance of the State Estimator. Therefore, power lines, phone lines and internet-based Virtual Private Networks (VPN) would be adequate to serve this application. However, a communication network that is supporting wide area real time power system protection and control applications should be extremely reliable and have minimum communication latency. In this case only a dedicated fibre-optic network would be satisfactory. China South Power Gird has developed a real time wide area monitoring and control system to monitor and damp inter-area oscillations, for which a pure fibre-optic dedicated communication network is being commissioned [51]. -56- Chapter 4 Architecture of a WAMPAC System Figure 4.2: The combination of different communication media for a WAMPAC system. To accommodate the communication needs of different applications, whilst limiting the cost of the communication network, the physical communication layer of a real WAMPAC system should be a combination of different communication media, as shown in Figure 4.2. Power lines, microwave and internet/VPN are suitable connections for PMUs that serve only real time system operation monitoring and state estimation. Phone lines could be used to transfer the event data recorded by PMUs for post disturbance analysis, and a dedicated fibre-optic communication network would be required for PMUs that serve wide area real time protection and control schemes. 4.3.2 Communication protocols and format for phasor data transmission Generally, communication protocol is a rule that defines the format and the order of messages exchanged between two or more communicating entities, as well as the actions taken on the transmission of the messages. The most frequently used message transport protocols for power system protection and control are Transport Control Protocol (TCP), User Datagram Protocol (UDP) and Internet protocol (IP). TCP/IP provides a highly reliable connection over unreliable networks, using checksums, congestion control, and automatic resending of bad or missing data. UDP/IP -57- Chapter 4 Architecture of a WAMPAC System is a protocol that provides low-latency communication across networks. However, it doesn’t provide any error-control or resending of missing or bad data, which brings more latency [52]. Concerning the phasor data format, the IEEE 1344 Standard only defined the phasor data formats for PMU to PMU communications. There were extended data formats based on the IEEE 1344 Standard, such as “PDC-stream”. This was defined and has been widely used, for data transmission beyond PMUs, i.e. from PMUs to DCs, by EIPP [48]. At present, the new phasor data protocol IEEE C37.118 defines the data format for this sort of data transmission beyond PMUs. PMU and DC vendors are using the IEEE C37.118 and its associated standard to replace IEEE 1344 for PMU manufacturing. Figure 4.3 presents a schematic of the synchronized data flow in a typical WAMPAC system. DCs collect time-stamped phasor measurements from PMUs. They then integrate all of the valid PMU measurements into a single data packet with a unique timestamp. These packets are encoded using the IEEE C.37.118 protocol or “PDCstream” and then streamed to the phasor data applications [53]. WAN WAN WAN/LAN WAN/LAN Figure 4.3: Generic schematic of the dataflow in WAMPAC system. -58- Chapter 4 Architecture of a WAMPAC System 4.3.3 Communication latency In a WAMPAC system, communication latency defines how long a packet of phasor measurements takes to be transferred from one point to another remote point. As shown in Figure 4.4, if the time stamp marked by PMU/sender is T1 and the time when the DC/receiver receives the data with time stamp T1 is T2, then the difference between T1 and T2 is the communication latency between the PMU/sender and DC/receiver. WAN/LAN WAN/LAN Figure 4.4: A simple demonstration of communication latency in WAMPAC system. The main factors that cause communication latency includes raw data (voltage and current instantaneous measurements) buffering for phasor calculation in the PMU, distant data transmission over communication links, and signal processing in the central station. The approximate communication latency in a WAMPAC system can be estimated by using expression (4.1): Tt = T f + T p + L +θ R (4.1) Where Tt is the total communication delay, Tf is the fixed delay associated with transducers used, phasor calculation, data concentration and multiplexing, Tp is the link propagation delay associated with different communication media, L is the amount of data transmitted, R is the data rate of the link, and θ is the random time delay component [48]. 4.4 Architecture of future GB WAMPAC system The future GB WAMPAC system will serve three interconnected power transmission networks, as shown in Figure 4.5, Scottish Power Transmission Network (SPTN), Scottish Hydro-Electric Transmission Network (SHETN) and National Grid Electricity -59- Chapter 4 Architecture of a WAMPAC System Transmission Network (NGETN). In this system, each power grid has its own DC. In Each power grid, the DC collects synchronized measurements from PMUs and processes the data to support the regional SMT applications; whilst, the DC also sends the collected data to the “Super-DC” to support the wide area real time applications. A “Super-DC” will be installed in the National Grid monitoring and control centre in Wokingham. The “Super-DC” collects the synchronized data from the three regional DCs and process the data for the wide area real time applications, whilst’, the “SuperDC” broadcasts the collected data back to the three regional DCs. With such SMTbased WAMPAC system, each power utility will not only supervise its own regional system operation but also monitor the operational status of the entire GB power network. WAN WAN Figure 4.5: A general architecture of future GB WAMPAC system. As introduced in Sections 4.2-4.3, a future GB WAMPAC system will use a number of different communication technologies for the real time data. For example, in the Scotland, microwave and satellite communication technologies offer good solutions for communication between remote substations. The power line technology and wideband phone lines are competent to support the hybrid state estimation with phasor measurements, whilst the wide area real time protection and control schemes will be served by a dedicated fibre-optic communication network. -60- Chapter 4 Architecture of a WAMPAC System 4.5 Conclusions This Chapter has introduced the architecture of a typical WAMPAC system. A Four– layer architecture is considered to be the most accepted approach for serving WAMPAC, based on the current international experience with WAMPAC operation. PMUs, DCs, various communication links and application software are the essential components in a typical WAMPAC system. As a large scale WAMPAC system, such as the future GB WAMPAC system, usually involves a number of power utilities; a Super-DC is the key component supporting wide area real time applications. The different SMT applications required by the power utilities have different demands, in terms of data management, reliability and security on the communication networks. Therefore, a WAMPAC system must accommodate a diverse range of technical requirements and increasing quantities of PMU data and the requirements of the new WAMPAC participants and clients. -61- Chapter 5 The Roadmap to the Future GB WAMPAC System Chapter 5 The Roadmap to the Future GB WAMPAC System 5.1 Introduction The previous Chapters stressed that the implementation of a WAMPAC system is the only choice to guarantee the reliable and secure operation of the future power systems. With this necessity in mind, this Chapter seeks to develop a roadmap that details the key steps that must be taken to implement WAMPAC in the future GB power system. Currently, the reliability of the electric power system in the GB is generally well, and very few stability problems have been encountered in recent years. As such there is no urgent demand for the introduction of WAMPAC to solve specific problems [54]. This has meant that the priorities for the deployment of WAMPAC applications in the GB electric power system have not yet to be determined. However, in the coming years the GB power system is entering a time of great change, e.g. the increasing integration of intermittent renewable generation and complex electricity transmission technology (HVDC and TCSC) that will make the system ever more unpredictable and difficult to manage. This time of change offers a good opportunity for the initialisation of GB WAMPAC project. The roadmap of GB WAMPAC project, and therefore this Chapter, is separated into several sections. The first of these is devoted to summarizing the current international experience with the development of WAMPAC systems. Based on the current international experience, a generic methodology for determining a roadmap that details the key steps that must be taken to implement a WAMPAC system will be introduced. It is then necessary to know the characteristics and problems that will be present in the future GB power system. After that a roadmap for the GB WAMPAC project will be developed by combining the knowledge gathered from the generic methodology for roadmap development with the prototype of the future GB power system. -62- Chapter 5 The Roadmap to the Future GB WAMPAC System 5.2 The Roadmap to a WAMPAC System At this time, a number of power utilities in the world are developing some forms of WAMPAC system. However, each power utility has its own specific needs, in terms of both system operation and planning, that have been addressed with the introduction of WAMPAC. Therefore, each utility has developed individual roadmaps to guide their WAMPAC projects, and as such, there is currently no generic roadmap for the development of a WAMPAC system. In addition, as many WAMPAC applications are still in the research stage of their development the requirements for their practical implementation have not been clearly defined [55] so the development of a generic roadmap is further complicated. However, it is generally accepted that the development of a WAMPAC system requires long term effort; it is impossible to achieve all the WAMPAC applications in the short term. As Figure 5.1 shows, WAMPAC applications can be grouped into three ranks based on the level of challenge involved in their development, i.e. low, medium and high [55]. The features considered in this ranking process include the difficulty of researching practical solutions for delivering these WAMPAC applications, the communication technology and data management requirements of the solution, and the limits of financial investment, etc. Those applications with a low challenge level are termed “low hanging fruits”; these can be implemented at the initial stage of a WAMPAC project [55]. Figure 5.1: WAMPAC application tree with full “smart fruits”. -63- Chapter 5 The Roadmap to the Future GB WAMPAC System This three tiered ranking of WAMPAC applications means that it would be logical to develop a WAMPAC project for the delivery of full “smart fruits” in three stages, as shown in Figure 5.2, the initial stage (1-3 years), the developing stage (3-5 years) and the developed stage (5-10 years). Figure 5.2: Roadmap for deploying PMU applications. In the initial stage of a WAMPAC project the main objective is the use of synchronized phasor measurements to afford system operators greater visibility and understanding of the status of a power system. The best applications to implement at this stage of the project would be those with fully developed, commercially available solutions (e.g. such as real time power angle (power flow) monitoring, voltage stability monitoring and inter-area oscillation monitoring). An additional benefit of these applications is that they require only limited support in terms of the number of PMUs that must be installed. This is beneficial as limitations on the available finance means that the number of PMUs deployed in the initial stage will be modest (around 5). In the second stage, with the increase of the number of PMUs being commissioned in the system, the system operators will have a deeper understanding of the power system dynamics. For example, wide area synchronized fault recordings will significantly improve system model validation. Consequently, system planning, emergency protection and control schemes will be optimized. The conventional state estimator could be upgraded to the faster and more accurate linear state estimator in this stage if 25-30% substations have a PMU installed. At the developed stage, having developed significant operational experience with PMUs operating over the course of several years, the system operators and planners will have -64- Chapter 5 The Roadmap to the Future GB WAMPAC System become more confident in their use of using synchronized phasor data. This confidence and experience will afford the opportunity to develop sophisticated closed loop control and protection schemes. In Summary, a generic roadmap for development of a WAMPAC system should meet the following requirements: 1) The roadmap must have the evolutionary character, i.e. it must assume a stepwise development of a WAMPAC system. 2) The evolutionary character must be expressed in development of two main strategies for development of a WAMPAC system: a) the short-term and b) the long term strategy. 3) The short term strategy must be based on the currently most critical system needs, which as such will determine the minimum requirements to the WAMPAC system (e.g. the minimal number of PMUs, the minimum performance of the communication infrastructure). The WAMPAC applications for the short term strategy must be well developed and tested. 4) The long term strategy must be based on the expected future system challenges (risks), which as such must be clearly explored and further used as an input for development of the future WAMPAC system. 5.3 The Future GB Power System The renewable energy policy against global warming is bringing significant change to the electrical power systems in the GB. The biggest change is the high integration of wind resources in electricity generation. Legislative targets in the UK require that 30% of electricity will be generated using wind resources by 2020. This target will be increased to 38% for 2030. Correspondingly, the proportion of power generation supplied by conventional coal power plants will be reduced to 20% by 2020, and 10% by 2030; the impact of these targets on the future generation mix in the UK can be seen in Figure 5.3 [8]. Table 5.1 shows three scenarios that would provide the targeted UK wind capacity for 2020 [8]. -65- Chapter 5 The Roadmap to the Future GB WAMPAC System Figure 5.3: UK energy target 2020 and 2030 [8]. Table 5.1: Three scenarios for meeting the 2020 UK renewable targets [8]. Wind Capacity Scenario 1 20.9GW in England and Wales 11.4GW Scotland Scenario 2 24.3GW in England and Wales 8GW Scotland Scenario 3 25.7GW in England and Wales 6.6GW Scotland In these scenarios, the minimum installed capacity of wind turbines in Scotland is 6.6 GW. The majority of the power generated by these wind turbines will be transferred to the remote load centres in Southern England. Therefore, this large increase in the generation capacity in Scotland will cause a large increase in the power flow between Scotland and England [8], as shown in Figure 5.4. Figure 5.4: Power transfers across the boundary between Scotland and England at peak load condition [8]. Obviously, the increasing levels of power transfer between Scotland and England will require substantial power transmission line reinforcements and the deployment of new technologies to facilitate the transmission of “wind power” from Scotland to England. To strengthen the power transfer capability at the interface, the National Grid Electricity Transmission (NGET) and Scottish Power Electricity Network (SPEN) are planning the installation of Thyristor Controlled Series Capacitors (TCSC) into the inter-tie lines and -66- Chapter 5 The Roadmap to the Future GB WAMPAC System new submarine High Voltage Direct Current (HVDC) links between Scotland and the North of Wales. Figure 5.5 outlines the proposed locations of the TCSC and the new HVDC link, which is the prototype of the GB future electric power transmission network (vision 2020/2030) [10]. Figure 5.5: New TCSC and HVDC in GB transmission networks [10] Figure 5.6: Locations of off-shore wind farms in GB power system (2020-2030) [8]. -67- Chapter 5 The Roadmap to the Future GB WAMPAC System The locations of the off-shore wind farms in the future GB power system are shown in Figure 5.6. These off-shore wind farms will connect to the GB power grid via back-toback HVDC links. This transmission technology will not allow the wind turbines to provide inertia to the system [56] [9]. This shortcoming, coupled with the increasing replacement of conventional generators with wind turbines, will cause the inertia of the future UK power system to be approximately 15-20% lower than it is currently [57]. The utilization of complex power transmission technology, such as TCSC and HVDC, and the increasing levels of intermittent and low inertia renewable generation that will be in use by 2020 and 2030, will cause the GB power system operate in a more complex and dynamic way. This shift will render the power system more difficult to monitor and control. Therefore, there is an essential need for the provision of WAMPAC across the entire power system. This will enable improved visibility and understanding of the system operational conditions as well as more effective protection and control schemes. 5.4 GB’s strategy for WAMPAC 5.4.1 Short term strategy As introduced in the previous sections, the generic roadmap for WAMPAC development and the characteristics of the future GB power system provides a strong position from which to determine the initial motivations of the GB WAMPAC project. The applications employed in the initial stage of the GB WAMPAC project should meet three essential requirements. 1) The number of PMUs employed at the initial stage should be modest (around 5). This is to accommodate the limitations in the availability of financial investment and Data Concentrator (DC) capacity. 2) The PMU applications employed at this stage should be fully developed and commercially available. 3) The functionality offered by the PMU applications employed at this stage should offer solutions to immediate system issues. This methodology for determining the prioritization of PMU applications at the initial stage of GB WAMPAC project are illustrated in Figures 5.7-5.9. The three essential requirements are represented as ‘logical filters’ and only those applications that successfully pass all three filters will be considered for implementation during the initial stage of introducing WAMPAC to the GB power transmission networks. -68- Chapter 5 The Roadmap to the Future GB WAMPAC System Figure 5.7: ‘Logical filter one’ – The number of PMUs required. Figure 5.8: ‘Logical filter two’ – Commercial availability of PMU applications. Figure 5.9: ‘Logical filter three’ – Necessity of PMU application for investors. -69- Chapter 5 The Roadmap to the Future GB WAMPAC System With this methodology, the GB WAMPAC project may be initialized in one of the following ways: ⑴ Wide area power angle and frequency monitoring system As Figure 5.10 shows, this initial project consists of 6 PMUs, located in Scotland, Wales, Northern England, Central England, Southern England and Western England. This simple system will provide system operators real time global system frequency and power flow monitoring. This is a very user friendly application that has been widely used by many power utilities in the initial stage of a WAMPAC project. Figure 5.10: Global power angle and frequency monitoring system. ⑵ Real time monitoring over inter-tie transmission lines This system consists of 6 PMUs, one installed at both ends and the mid-points of the two transmission corridors between Scotland and England. In the future, there will be a huge increase in the power transmission over these transmission lines. This application will help system operators overcome the challenges posed by increased power transmission by allowing improved awareness of the real time operating conditions of the inter-tie lines, e.g. thermal stability and small signal stability (inter-area oscillations). -70- Chapter 5 The Roadmap to the Future GB WAMPAC System The proposed PMU placements for this application are presented in the Figure 5.11. They are Stratheven, Eccles, Harker, Stella West, Penwortham and Thornton. Figure 5.11: Real time monitoring system over inter-tie corridors. 5.4.2 Long term strategy Long term development of the GB WAMPAC system will be driven by the new operational risks that emerge as the system continues to develop. The characteristics and potential risks that are likely to be present in the future GB power can be summarized as follows: ① Wind turbines are not as stable in producing electricity as conventional synchronous generators (coal, gas and oil). The increasing replacement of conventional generators with wind turbines will undermine the GB power system’s capability to balance power generation and demand. As a consequence, the system frequency will change dramatically after large disturbances occur. Furthermore, as electricity generation from renewable resources is highly influenced by climatic conditions, the power flow patterns and generation dispatch schemes of the future GB power system will change quite frequently. ② The majority of the new wind farms, particularly offshore wind farms, will connect to the power grid via back-to-back Voltage Source Converters [56]. Hence, they will not provide inertia to the system. Coupled with the increasing replacement of conventional -71- Chapter 5 The Roadmap to the Future GB WAMPAC System generators with wind turbines the inertia of the future UK power system will be largely reduced. Low inertia systems will experience extremely fast frequency deviations and lightly damped low frequency oscillations after disturbances. Based on these characteristics and potential risks, the long term strategy for a UK WAMPAC system should focus upon (1) Providing a real time wide area monitoring system, (2) Improving system frequency stability, and (3) Enhancing system damping for system small signal stability. For these purposes, three long term SMT applications are proposed. ⑴ Real time wide area monitoring system With the proportion of total electricity generation coming from renewable resources increasing, the future GB power system’s operation will become more dynamic and unpredictable; correspondingly the power flow pattern and generation dispatch scheme will change quite frequently. The conventional State Estimator may no longer be capable of providing a wide area, fast and accurate monitoring system for the next generation of GB system operation. In this case, the development of a SMT-based real time Wide Area Monitoring System (WAMS) is a necessity. Figures 5.12- 5.15 present the PMU placements across the GB electricity transmission network that would serve as a prototype for the future GB WAMS. Figure 5.12: PMU placements in SHPTN. -72- Chapter 5 The Roadmap to the Future GB WAMPAC System As Figure 5.12 presents, four substations in Scotland will have PMUs installed. These are Beauly, Kintore, Errochy and Tealing. As seen from Figure 5.12, these substations are the big substations that have the maximum number of branches in the Scottish Hydro Power Transmission Network (SHPTN). In the Scottish Power Transmission Network (SPTN) and the North part of National Grid Transmission Network (NGTN), there are seven substations where PMUs will be installed. They are Stratheven, Eccles, Harker, Stella West, Penwortham and Thornton and Auchencrosh. As shown in Figure 5.13, six PMUs will be used to monitor the transmission corridor between Scotland and England, as this is a key corridor in the GB power system and a very large increase in the power transmission across it. The PMU installed in Auchencrosh will be used to monitor the electricity transmission between Northern Ireland and GB. Figure 5.13: PMU placements at the boundary between Scotland and England. In central England, there are eight substations where PMU will be installed. They are Deeside, Daines, Cottam, Derakelow, Feckenham, Pembroke, East Cladon and Pelham. As seen from Figure 5.14, all these substations are big substations of the 400 kV transmission networks (represented by blue lines). The installation of PMUs into the 275kV transmission network (represented by red lines) is not recommended as it is unlikely that the operation of this highly meshed network will be improved through the use of PMU data. -73- Chapter 5 The Roadmap to the Future GB WAMPAC System Figure 5.14: PMU placements in central England. Figure 5.15 presents the PMU placements in the South of England. The PMU installed at the Sellindge substation will be used to monitor the power transmission between the UK and France, the other five PMUs are installed at the Indian Queens, Exeter, Melksham, Bramley and Lovedean substations for the monitoring of the southern 400kv transmission network. Figure 5.15: PMU placements in the South of England. So far, the proposed PMU placements for monitoring the GB electricity transmission networks have been presented. In addition, around 30% of the electricity in the future GB power system will be produced by wind farms; this constitutes a significant amount of relatively unstable electricity generation that will bring more dynamics to the system. Therefore, the future GB WAMS should also monitor the operation of wind farms, -74- Chapter 5 The Roadmap to the Future GB WAMPAC System particularly large off-shore wind farms. This monitoring could be delivered by installing PMUs at the substations shown in Figures 5.16-5.20 that will allow the real time operational information of the large wind farms to be transmitted to the WAMPAC control centre. With this real time information, more economical and efficient power generation dispatch schemes can be developed. Figure 5.16: A PMU installed in the Torness substation. Figure 5.17: Three PMUs installed in the Crekey Beck, Keadby and Grimsby West substations. -75- Chapter 5 The Roadmap to the Future GB WAMPAC System Figure 5.18: One PMU installed in the Sizewell substation. Figure 5.19: Two PMUs installed in the North Wales substations of Wylfa and Stanah. Figure 5.20: A PMU installed in the Alverdiscott substation. -76- Chapter 5 The Roadmap to the Future GB WAMPAC System The PMUs discussed above will form the prototype WAMS for the future GB power transmission network shown in Figure 5.21. With access to the real time monitoring, the power system operators will have a clear view of the operational status of the entire transmission network. In addition, the conventional State Estimator could be updated significantly due to the integration of a significant amount of precise synchronized data. Figure 5.21: PMU placements for the future GB wide area monitoring system. -77- Chapter 5 The Roadmap to the Future GB WAMPAC System ⑵ Wide area adaptive under-frequency load shedding Under-frequency load shedding (UFLS) is a common practice for most power systems. It is used to help prevent large frequency declines when a large disturbance occurs such as the sudden outage of a large generator [4] [58]. Nowadays, the plans adopted by power utilities are predominantly deterministic, not taking into account the actual system state and topology, operating point, or the nature and magnitude of the disturbance. The frequency measurements are performed locally in distribution substations, when the frequency drops below the preset point the assigned load feeder will trip [44]. As a consequence, the existing UFLS plans very often disconnect more or less load than is required. Figure 5.22: A SMT-based adaptive UFLS scheme in the future GB power system. The availability of synchronized wide area measurements provides a great opportunity for developing an adaptive and efficient UFLS scheme. A conceptual view of an adaptive UFLS protection scheme for the future GB power system is presented in Figure 5.22. The adaptive UFLS protection scheme is based on the real time wide area system frequency measurements and the rate of change of system frequencies. With known system inertia constant (H) the magnitude of the imbalance between generation and demand can be immediately estimated after the disturbance [45]. In addition, The wide are monitoring system can also provide other auxiliary information to support the -78- Chapter 5 The Roadmap to the Future GB WAMPAC System adaptive UFLS protection scheme, for example, real time active power outputs from wind farms, the power generation reserves of synchronous generators and power flow condition over the tie-line that connects the generation area and the demand area. With access to so many real time wide area measurements, system operators would be more confident when optimizing UFLS protection schemes allowing these schemes to be more efficient and effective. ⑶ Wide-area inter-area oscillation damping control with power electronic devices As mentioned above, the huge increase in the power transmission over the transmission corridors between Scotland and England, coupled with reduced system inertia, will lead to the system experiencing lightly damped inter-area oscillations. The Thyristor Controlled Series Capacitor (TCSC) and new submarine High Voltage Direct Current (HVDC) link installed in the inter-tie transmission corridors can be used to improve the damping of the inter-area oscillatory mode between Scotland and England if their controllers are set properly. Figure 5.23: Inter-area oscillation damping control with power electronic devices in GB power system. Figure 5.23 provides a conceptual view of the wide area inter-area oscillation monitoring and control system in the future GB power system. In order to monitor the inter-area oscillations accurately, six PMUs will be installed over the critical AC -79- Chapter 5 The Roadmap to the Future GB WAMPAC System transmission lines between Scotland and England. The real time inter-area oscillation signals, such as power angle difference and system frequency across these transmission lines will be captured by these PMUs. With the real time information of the inter-area oscillations the control centre calculates the parameters of the controllers of the power electronic devices to modulate power flow. A general procedure for real time inter-area oscillation closed-loop control with HVDC is presented in Figure 5.24. Six PMUs will be installed across the inter-tie transmission lines to allow monitoring of the inter-area oscillations, as shown in Figure 5.23. Measurements of the oscillatory signals, such as active power flow, power angle and system frequency difference, across the inter-tie transmission lines can be used in the centralized control centre to generate dynamic firing angles for the converters of the HVDC transmission system. In such a way, the DC power flow will be dynamically changed to stabilize the inter-area oscillations. Figure 5.24: Closed loop inter-area oscillation control using HVDC. Figure 5.25 presents an alternative procedure that uses TCSC, for real time inter-area oscillation closed-loop control. As for the HVDC closed loop control scheme, the widearea centralized damping controller will use the real time inter-area oscillatory signals captured by the six PMUs that are monitoring the inter-tie transmission lines. This controller will determine a suitable firing angle that will allow the capacity of the TCSC to vary dynamically to modulate the power flow for damping inter-area oscillations. -80- Chapter 5 The Roadmap to the Future GB WAMPAC System Figure 5.25: Closed loop inter-area oscillation control using TCSC. 5.5 Conclusions In this Chapter, the methodology of designing a roadmap to guide the GB WAMPAC project has been introduced. This methodology takes into account the international experience with WAMPAC project management and the practical challenges faced in the future GB electric power network. Based on this methodology, the GB’s strategy for the development of WAMPAC is devised. The GB WAMPAC strategy is divided into a short term strategy and a long term strategy. In the short term, only a few (5-6) PMUs will be deployed in the GB power system. This is due to the uncertainties related to the project and the limitations of financial investment. These PMUs can be distributed across the entire UK transmission network to form a wide area power angle and frequency monitoring system. In addition, these PMUs can also be used to monitor the power flow and inter-area oscillations between Scotland and England. In the long term, with the increase in the number of PMUs deployed in the UK power system, the conventional state estimator will be upgraded into a new generation of real time wide area monitoring using the methodology for PMU placement introduced. In addition, a wide area adaptive load shedding scheme is proposed. This scheme will help GB power system to overcome the challenge of reduced frequency stability that will be introduced by the high integration of renewable resources. The most ambitious element of the long term strategy for the UK WAMPAC system is the use of HVDC and TCSC to damp the inter-area oscillations between Scotland and England. Successful development of such an application would constitute a great achievement in improving the operation of power systems. This is because it will allow the conventional EMS based open loop control that is currently in use to be upgraded to a sophisticated wide-area closed loop control system. -81- Chapter 6 The Physical Nature of Inter-area Oscillations in Electrical Power Systems Chapter 6 The Physical Nature of Inter-area Oscillations in Electrical Power Systems 6.1 Introduction In Chapter 5, the proposed GB WAMPAC strategies addressed that the real time interarea oscillation monitoring and control are probably the most anticipated SMT applications in the future GB power system. In the short term strategy of the GB’s WAMPAC project, it is recommended that several PMUs will be installed across the transmission corridor between Scotland and England, to monitor the inter-area oscillations. In the long term strategy, the new power transmission technologies deployed in the future UK power system, such as HVDC and TCSC, will be used to enhance the system’s small signal stability. To support this effort toward real time wide area inter-area oscillation monitoring and control, a fundamental study of inter-area oscillations will be provided in this Chapter. First of all, nonlinear simulations will be used to present the physical nature of electrometrical oscillations in the time domain. In the nonlinear simulations, different disturbances will be simulated to initiate different oscillatory modes. From the system response to these disturbances, the physical phenomenon of power oscillations can be clearly observed. Having recognised the underlying physical phenomena, a mathematical analysis technique, modal analysis, that is necessary to predict system performance will be introduced. Modal analysis is performed in the frequency domain, which is based on the linearized model of the system around an operational equilibrium. From the outputs of the modal analysis, such as eigenvalues, eigenvectors and participation factors, the system dynamic characteristic can be accurately analyzed. 6.2 Nonlinear Simulations To investigate the physical nature of low frequency electromechanical power oscillations in the time domain, a model of a typical two-area power system was created. This two-area system was created by Ontario Hydro for a research report commissioned by the Canadian Electrical Association [59] [60]. This system was designed to exhibit -82- Chapter 6 The Physical Nature of Inter-area Oscillations in Electrical Power Systems the different types of oscillations that occur in an interconnected system [59] [60]. This two-area system can be considered as a useful tool for the study of the electromechanical oscillations in the GB power system. For this purpose Area 1 represents the Scottish power system and Area 2 represents the English power system. A single line diagram of the two-area system is shown in Figure 6.1, and the full set of the system parameters i.e., the generator, transformer and transmission line parameters, as well as the controller settings of the Automatic Voltage Regulator (AVR) and Turbine Governor (TG) are given in Appendix A. Figure 6.1: A typical two-area system. This simple model shows the electromechanical oscillations that are inherent in the twoarea system. There are three possible electromechanical modes of oscillations in this system. There are two local modes, one in which generator 1 swings against generator 2, and another in which generator 3 swings against generator 4. In addition, there is also one inter-area mode, in which the generators in area 1 swing against the generators in area 2. In this Section, nonlinear simulations will be used to give an insight into the nature of these different types of electromechanical oscillations. In these nonlinear simulations, the different modes of oscillation are initiated using a range of different disturbances. 6.2.1 Local oscillatory mode in Area 1 To investigate the nature of the local mode in Area 1, changes in the mechanical torque of the generators in that area were simulated. To properly investigate the behaviour of the local mode in Area 1 it is important to minimise the excitation of the inter area mode during these simulations. To achieve this goal, equal and opposite step changes in the mechanical torque of the two generators in Area 1 were simulated simultaneously. For example, a change of -0.01 p.u. in the mechanical torque of generator 1 is simulated then a corresponding change of 0.01 p.u. is made in the mechanical torque of generator -83- Chapter 6 The Physical Nature of Inter-area Oscillations in Electrical Power Systems 2. The response of the generators, in terms of speed, to this pair of small disturbances in Area 1 is presented in Figure 6.2. rotor speed [p.u.] 1.0002 G1 1.0001 inter-area mode observed G2 1 0.9999 0.9998 ' local mode 1' observed 0 2 4 Area 1 6 8 10 time [s] rotor speed [p.u.] 1.0001 G3 G4 1 0.9999 Area 2 0 2 4 6 8 10 time [s] Figure 6.2: Generator rotor speed responses to the disturbances occurred in area 1. In Area 1, the rotor speed changes of generator 1 and 2 were in anti-phase i.e. generator 1 oscillated against generator 2 in the local mode. This local mode dominated the oscillation for approximately 7s, at which time the generators began to swing together in the inter-area mode. The generators in Area 2 experienced oscillations with lower amplitude than those seen in Area 1. These oscillations were in phase with one another and are driven by the inter-area mode, the local mode in Area 2 was not observed here. These simulation results show that the frequency of the local oscillation mode in area 1 is approximately 1 Hz. 6.2.2 Local oscillatory mode in Area 2 In this Section, the same method used in Section 6.2.1, to excite the local mode in Area 1, is used to excite the local mode in area 2. An equal and opposite step change of the mechanical torque of the generators in Area 2 was simulated. The change in the mechanical torque of generator 3 was -0.01 p.u. and the change in the mechanical torque of generator 4 was 0.01 p.u.. The generator rotor speed responses to the small disturbances that occurred in area 2 are shown in Figure 6.3. -84- Chapter 6 The Physical Nature of Inter-area Oscillations in Electrical Power Systems speed [p.u.] 1 1 1 G1 1 0 2 G2 4 Area 1 6 8 10 time [s] 1.0002 speed [p.u.] G3 G4 inter-area mode observed 1.0001 1 0.9999 0.9998 'local mode 2' observed 0 2 4 Area 2 6 8 10 time [s] Figure 6.3: Generator rotor speed responses to the disturbances occurred in area 2. For a small disturbance in Area 2, generator 3 immediately began to swing against generator 4; this local mode dominated the response for about 5s, after which time the inter-area oscillatory mode began to dominate. The generators in Area 1 were driven by the inter-area mode and moved together with oscillations of much lower amplitude than those seen in Area 2. The frequency of the local mode in Area 2 was approximately 1 Hz. 6.2.3 Inter-area oscillatory mode The inter-area mode can be directly provoked by changing the mechanical torque of one generator in each of the different areas. In this case, the mechanical torque of generator 2 was increased by 0.01 p.u. whilst the mechanical torque of generator 4 was reduced by 0.01 p.u.. The generator speed responses to these small disturbances are shown in Figure 6.4. speed [p.u.] 1.0004 G1 G2 G3 G4 1.0002 1 0.9998 0 2 4 6 8 10 time [s] Figure 6.4: Generator rotor speed oscillations dominated by inter-area mode. -85- Chapter 6 The Physical Nature of Inter-area Oscillations in Electrical Power Systems As Figure 6.4 presents, the inter-area mode dominated the response of the generator rotor speeds to these disturbances. The generators in Area 1 began to swing against the generators in Area 2 immediately after the disturbances, and the magnitudes of the speed change of the generators in Area 2 were larger than the magnitudes of the speed change of the generators in Area 1. Initially the oscillations in Area 1 were strongly influenced by the local mode. This is evident as for the first 4 seconds the generators in Area 1 oscillated against one another whilst also moving together in the inter-area mode. The frequency of the inter-area mode was approximately 0.5 Hz. The inter-area mode was not damped by any external control and the amplitude of the inter-area oscillation was seen to increase. For obtaining more information about inter-area oscillations, the responses of the system frequency and the inter-area active power flow to the disturbances were analyzed. Figure 6.5 presents the system frequency response to the disturbances measured in Area 1 (bus 3) and Area 2 (bus 5). The oscillations in the frequency deviation in Area 1 were approximately in anti-phase to the oscillations in the frequency deviation in Area 2; which is consistent with the changes seen in the generator rotor Frequency [Hz] speeds in Figure 6.4. fre-bus3 50.002 fre-bus5 50 49.998 49.996 0 2 4 6 8 time [s] Figure 6.5: System frequency responses in inter-area mode. 10 Figure 6.6 shows the active power flow over line 3 after the disturbances. The oscillatory power flow on line 3, is purely driven by the inter-area mode; with no influence from the local modes. This occurs because the physical mechanism behind electromechanical oscillations is the active power exchange between the generators that are involved in the oscillatory mode. Therefore, as line 3, like all of the inter-tie lines, only carries power between the two areas, then only the inter area oscillations, and not the local mode oscillations, will be seen on these lines. -86- Chapter 6 The Physical Nature of Inter-area Oscillations in Electrical Power Systems 210 P [MW] P-line3 205 200 0 2 4 6 8 10 time [s] Figure 6.6: Oscillatory active power flow on transmission line 3. The power exchange driving the local modes in Areas 1 and 2 occurs along lines 1 and 8 respectively. Therefore, the power flow associated with these local oscillation modes will only be seen on these lines. To demonstrate this, the power flow on line 1 during the disturbances is shown in Figure 6.7. The variation of the power flow on line 1 shows the difference between the power flow that supports the local mode and the power flow that supports the inter-area mode. This can be seen by comparing the power flow during the first four seconds after the disturbances, where the local mode dominates, with the power flow during the next six seconds, where the inter-area mode dominates. 700 P [MW] inter-area mode observed 695 'local mode 1' observed 690 0 2 4 P-line1 6 8 10 time [s] Figure 6.7: Oscillatory active power flow on transmission line 1. 6.2.4 Large disturbance To further examine the characteristics of the inter-area oscillations, a three-phase shortcircuit fault was simulated on bus 4, the mid point of the inter-area lines that connect the two areas. A short circuit represents a much larger disturbance to the system than the small mechanical power changes simulated in the previous Sections and as such will offer greater insight into the behaviour of the oscillatory modes. The transient fault occurred at 0.1s and was cleared after at 0.2s. The response of the rotor speed of each generator to the disturbance is presented in Figure 6.8, and the active power transfer over one of the inter-tie line (line 3) is shown in Figure 6.9. -87- Chapter 6 The Physical Nature of Inter-area Oscillations in Electrical Power Systems rotor speed [p.u.] 1.005 G1 G2 4 6 G3 G4 1 0.995 0 2 8 10 time [s] Figure 6.8: Responses of the Generator rotor speeds to the large disturbance. 300 P [MW] P-line3 200 100 0 0 2 4 6 8 10 time [s] Figure 6.9: Active power transfer over the tie line after the disturbance. As seen from the Figures 6.8 and 6.9, after the system recovered from the transient fault, the generators in Area 1 started to oscillate against the generators in Area 2 in inter-area mode around the new system equilibrium point. The inter-area mode was clearly visible in the generator rotor speed responses and the oscillatory active power flow on the intertie line. 6.3 Modal Analysis In Section 6.2, a number of nonlinear simulations were performed to show the physical nature behind power system oscillations. In the nonlinear simulations the transients were induced using small disturbances (e.g. a 1% change of mechanical torque), and the system responses were essentially linear. Although the initial system response to the three-phase transient fault was nonlinear, the response quickly settled into post fault oscillations that are essentially linear around the post fault equilibrium point. This means that, for a study of electromechanical oscillations, the system model can be linearized around the steady state point. The linearization of the system model provides an excellent opportunity for modal analysis [59]. Modal analysis can be used to perform -88- Chapter 6 The Physical Nature of Inter-area Oscillations in Electrical Power Systems a wide range of tasks, such as determining the oscillatory modes, the sources of the oscillatory modes, and the parameters needed for designing oscillation controllers. In this Section, the modal analysis will be introduced. This analysis tool will then be used to explain the physical phenomenon of inter-area oscillations presented in the Section 6.2. 6.3.1 Dynamic system representation In power system modelling, the dynamic behaviour of an electrical power system can be described by a group of first order nonlinear differential equations with the following form [34]: • x = f (x, u) (6.1) where ⎡ x1 ⎤ ⎢x ⎥ x= ⎢ 2⎥ ⎢... ⎥ ⎢ ⎥ ⎣xn ⎦ ⎡ f1 ⎤ ⎢f ⎥ f = ⎢ 2⎥ ⎢... ⎥ ⎢ ⎥ ⎣ fn ⎦ ⎡u1 ⎤ ⎢u ⎥ u = ⎢ 2⎥ ⎢... ⎥ ⎢ ⎥ ⎣u r ⎦ The column vector x is the state vector, its elements xi are the state variables, and n is the number of system states. The column vector u is the vector of inputs to the system, and r is the number of inputs. f is the vector of the differential equations. The outputs of the system can be expressed in terms of the state variables and the input variables with the following form: y = g(x, u) where ⎡ y1 ⎤ ⎡g1 ⎤ ⎢y ⎥ ⎢g ⎥ 2 ⎥ ⎢ and y= g=⎢ 2⎥ ⎢... ⎥ ⎢... ⎥ ⎢ ⎥ ⎢ ⎥ ⎣ ym ⎦ ⎣g m ⎦ -89- (6.2) Chapter 6 The Physical Nature of Inter-area Oscillations in Electrical Power Systems The column vector y is the vector of outputs, g is the vector of nonlinear functions for calculating the system outputs, m is the number of system outputs. 6.3.2 System linearization for modal analysis If there are a set of state variables x0 and a set of inputs u 0 with which all the derivatives • • • x1 , x 2 ,…, x n are simultaneously zero, as represented by the equation (6.3), we can say that the system is at a steady state [34]. • x 0 = f( x 0 , u 0 ) = 0 (6.3) If a stable system is disturbed from steady state by a small disturbance, i.e. Δ x and Δ u , it will eventually come to rest at a new steady state. This transition will still satisfy equation (6.1), hence we have • x = f[( x 0 + Δx), (u 0 + Δu )] (6.4) As the disturbance is assumed to be very small, the nonlinear equation f ( x , u ) can be approximated using a Taylor series expansion in terms of Δx and Δ u . If we only consider the first order terms of this expansion we will have: • • • xi = x i 0 + Δ x i = f i [(x 0 + Δx), (u 0 + Δu)] = f i (x 0 , u 0 ) + ∂f i ∂f ∂f ∂f Δx1 + ... + i Δxn + i Δu1 + ... + i Δu r ∂x1 ∂xn ∂u1 ∂u r (6.5) • Since x i 0 = f i ( x 0 , u 0 ) we have • Δ xi = ∂f ∂f i ∂f ∂f Δx1 + ... + i Δxn + i Δu1 + ... + i Δu r i = 1, 2 ,..., n ∂x1 ∂xn ∂u1 ∂u r in the same way based on Equation 6.2, we have -90- (6.6) Chapter 6 The Physical Nature of Inter-area Oscillations in Electrical Power Systems Δy j = ∂g j ∂x1 Δx1 + ... + ∂g j ∂xn Δxn + ∂g j ∂u1 Δu1 + ... + ∂g j ∂ur Δur j = 1, 2,..., m (6.7) Therefore, the linearized system state equations around the equilibrium point are given as [34]: • Δ x = AΔx + BΔu (6.8) Δy = CΔx + DΔu (6.9) where ∂f ⎤ ⎡ ∂f 1 ... 1 ⎥ ⎢ ⎢ ∂x1 ∂x n ⎥ ⎥ A = ⎢... ⎢ ⎥ ⎢ ∂f n ... ∂f n ⎥ ⎢ ∂x ⎥ ⎣ 1 ∂x n ⎦ ⎡ ∂f 1 ∂f 1 ⎤ ... ⎢ ⎥ u ∂ ∂u r ⎥ 1 ⎢ ⎥ B = ⎢... ⎢ ⎥ ⎢ ∂f n ... ∂f n ⎥ ⎢ ∂u ∂u ⎥ r ⎦ ⎣ 1 ⎡ ∂g1 ∂g1 ⎤ ⎢ ∂x ... ∂x ⎥ n ⎥ ⎢ 1 ⎥ C = ⎢... ⎥ ⎢ ⎢ ∂g m ... ∂g m ⎥ ⎢⎣ ∂x1 ∂x n ⎥⎦ ⎡ ∂g1 ∂g1 ⎤ ⎢ ∂u ... ∂u ⎥ r ⎥ ⎢ 1 D = ⎢... ⎥ ⎥ ⎢ ⎢ ∂g m ... ∂g m ⎥ ⎢⎣ ∂u1 ∂ur ⎥⎦ ∆x is the state vector of length n ∆u is input disturbance vector of length r ∆y is the output vector of length m A is the state matrix of size n × n B is the input matrix of size n × r C is the output matrix of size m × n D is a matrix of size m × r , which defines the proportion of the input which directly influences the output, ∆y. The small signal stability of the system can be analyzed by using (6.8) and (6.9). In other words, the system’s small signal stability can be determined by solving the state equations. -91- Chapter 6 The Physical Nature of Inter-area Oscillations in Electrical Power Systems 6.3.3 Eigenvalues and eigenvectors The following equation is defined as the characteristic equation of matrix A [34]. det(λI − A) = 0 (6.10) The n solutions of the characteristic equation λ = [λ1 , λ2 ,..., λn ] are the eigenvalues of A. The eigenvalues may be real or complex. If matrix A is real, then the complex eigenvalues occur in conjugate pairs. For any eigenvalue, λ i , a n-column vector, φi , that satisfies Aφ i = λi φ i i = 1,..., n (6.11) is the right eigenvector of the A matrix associated with the eigenvalue λi . The right eigenvector φi has the form: ⎡φ1i ⎤ ⎢φ ⎥ φ i = ⎢ 2i ⎥ ⎢... ⎥ ⎢ ⎥ ⎣φ ni ⎦ (6.12) The right eigenvector describes how each mode of oscillation is distributed among the system states. In other words, it indicates on which system variables the mode is more observable [61]. The magnitudes of the elements of φ i give the extent of the behaviours of the n state variables in the ith mode, the angles of elements give the phase displacements of the state variables with regard to the mode. Thus, the right eigenvector is called mode shape. Similarly, if a n-row vector ψ i that satisfies: ψ i A = λi ψ i i = 1,..., n -92- (6.13) Chapter 6 The Physical Nature of Inter-area Oscillations in Electrical Power Systems It is called the left eigenvector of the A matrix associated with the eigenvalue λi . The left eigenvector has the form: ψ i = [ψ i1 ψ i 2 ...ψ in ] (6.14) The left eigenvector ψ i can be used to identify which combination of state variables displays only the ith mode. Thus, the kth element of the right eigenvector φ i measures the activity of the variable x k in the ith mode, and the kth element of the left eigenvector ψ i weights the contribution of this activity to the ith mode. 6.3.4 Eigenvalues and small signal stability The free motion (with inputs remaining constant) of a power system around an operating point, after a small disturbance, can be described by equation (6.15): Δx& = A Δx (6.15) In the time domain the response of the ith state variable can be described by Δ x (t ) = n ∑ φ ψ Δx (0) e λ i =1 i i it (6.16) where λi is the ith eigenvalue, φi is the ith right eigenvector, ψ i is the ith left eigenvector and Δx(0) is the initial state of the state vector Δx . Reference [34] presents the detailed derivation of equation (6.16). The expression (6.16) defines the free motion of a power system in terms of the n eigenvalues, and the right and left eigenvectors of the system, from this equation the following properties of the system response can be determined using eigenvalues [34]. (1) A real eigenvalue corresponds to a non-oscillatory mode. A negative real eigenvalue represents a decaying mode, whereas a real positive eigenvalue represents an unstable mode. -93- Chapter 6 The Physical Nature of Inter-area Oscillations in Electrical Power Systems (2) Complex eigenvalues always occur in conjugate pairs, and each pair corresponds to an oscillatory mode. For example, a pair of complex eigenvalues λi = σ i ± jωi (6.17) The real part of the complex eigenvalue gives the damping factor of this oscillatory mode, and the imaginary part gives the frequency of this oscillatory mode. A negative real part represents a damped oscillation whereas a positive real part represents an oscillation with increasing amplitude. The frequency of oscillatory mode is f = ω (Hz) 2π (6.18) A practical measure for the assessment of the damping of oscillations is the damping ratio, ζ , defined as [34]: ζ = −σ σ 2 + (2πf ) 2 × 100% (6.19) The oscillatory modes with a damping ratio less than 3% are the critical modes that must be improved [34]. When designing damping controls, a stability margin should be taken into account due to the uncertainties of system operation. Thus, a damping ratio of at least 5% should be the objective of the control design [59]. 6.3.5 Participation factors One problem in using right and left eigenvector individually for identifying the relationship between the states and the modes is that the elements of the eigenvectors are dependent on the units and scaling associated with the state variables. As a solution to this problem, a matrix called the participation matrix (P), which combines the right and left eigenvectors, is used as the measure of the association between the state variables and the modes. -94- Chapter 6 The Physical Nature of Inter-area Oscillations in Electrical Power Systems P = [p1 p 2 ...p n ] (6.20) ⎡ p1i ⎤ ⎡φ1iψ i1 ⎤ ⎢ p ⎥ ⎢φ ψ ⎥ p i = ⎢ 2i ⎥ = ⎢ 2i i 2 ⎥ ⎢... ⎥ ⎢... ⎥ ⎢ ⎥ ⎢ ⎥ ⎣ p ni ⎦ ⎣φ niψ in ⎦ (6.21) with where φ ki is the kth entry of the right eigenvector φi . ψ ik is the kth entry of the left eigenvector ψ i . The element p ki = φ kiψ ik is defined as a participation factor. The product of φ ki and ψ ik measures the net participation p ki of the kth state to ith mode. This is because φ ki measures the activity of xk in the ith mode and ψ ik weights the contribution of this activity to the mode. Calculated in this way, participation factors can be used as a measure for identifying the source of an oscillatory mode. 6.3.6 Modal analysis for inter-area oscillation study In this example, we will use the modal analysis to analyze the inter-area oscillations in typical two-area system. The test system used for this modal analysis is the same system which was used in Section 6.2 for the nonlinear simulations. Concerning system modelling for modal analysis, there are twelve states in each synchronous generator [59] [34]: (1) Six states for modelling synchronous generator δ - rotor angle, ω -rotor speed, ψ fd -flux of the excitation system, ψ 1d - flux of the damper winding on d-axis, ψ 1q - flux of the damper winding 1 on q-axis, ψ 2 q - flux of the damper winding 2 on q-axis. -95- Chapter 6 The Physical Nature of Inter-area Oscillations in Electrical Power Systems (2) Three states for modelling Automatic Voltage Regulator (AVR); vt -voltage transducer state, v ph -phase shift state and va -amplifier state. (3) Three states for modelling the Turbine Governor (TG); tg ser -governor servo state, tg hp -high pressure turbine state and tg rh -reheat stage state. In steady state, a classic modal analysis is executed. All the eigenvalues corresponding to the system state matrix are presented in Table 6.1. There are 48 eigenvalues in the system, the same number as the system state variables. As seen in Table 6.1 there are 8 pairs of complex eigenvalues (shaded), which implies that 8 oscillatory modes exist in the system. No. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 Table 6.1: Eigenvalues of the two-area system. Real part Imaginary Part No. Real part Imaginary Part 0 0 25 -10.15465 0 -100.0867 0 26 -5.587707 0 -100.0726 0 27 -5.641435 0 -100.0439 0 28 0.02665295 3.375048 -100.0426 0 29 0.02665295 -3.375048 -37.08784 0 30 -3.539402 0 -37.02837 0 31 -1.668693 1.61424 -35.27151 0 32 -1.668693 -1.61424 -34.43338 0 33 -0.9635185 1.056211 -32.24181 0 34 -0.9635185 -1.056211 -32.37108 0 35 -1.937562 0 -26.28765 0 36 -1.933964 0 -25.23291 0 37 -1.81263 0 -18.74816 0 38 -1.675042 0 -18.78778 0 39 -0.3302784 0.9980652 -17.36206 0 40 -0.3302784 -0.9980652 -16.44011 0 41 -1.108677 0 -0.473808 6.257372 42 -0.4201453 0.644802 -0.473808 -6.257372 43 -0.4201453 -0.644802 -0.4706583 6.30193 44 -0.4050254 0.6372564 -0.4706583 -6.30193 45 -0.4050254 -0.6372564 -10.21898 0 46 -0.1901362 0 -10.20276 0 47 -0.195777 0 -10.15399 0 48 -0.1958579 0 -96- Chapter 6 The Physical Nature of Inter-area Oscillations in Electrical Power Systems The oscillatory modes and their information are presented in Table 6.2. There are three electromechanical oscillatory modes, two of which have a satisfactory damping ratio (1 and 2) (>5%), and a third that is unstable (shaded). The other five oscillatory modes are excitation and governor control modes. As shown in Table 6.2, the frequency of the two local modes is 1Hz, and the frequency of the inter-area mode is 0.54Hz. Table 6.2: Oscillation modes in the two-area system. No of oscillatory mode No of Eigenvalue (λ) 1 2 3 4 5 6 7 8 18,19 20,21 28,29 31,32 33,34 39,40 42,43 44,45 Real imaginary Frequency (Hz) Damping ratio (%) -0.4706584 ±6.30193 1.002983 -0.4738081 ±6.257372 0.9958917 0.0266525 ±3.375049 0.5371557 -1.668693 ±1.61424 0.2569142 -0.3302784 ±0.9980652 0.1681012 -0.9635178 ±1.056211 0.158847 -0.4201452 ±0.6448019 0.1026234 -0.4050253 ±0.6372563 0.1014225 7.447738 7.550383 -0.78966 71.87365 67.39455 31.41638 54.59233 53.64027 The layouts of the eigenvalues that are associated with the oscillatory modes are presented in a complex panel (see Figure 6.10). Here, only the eigenvalues that have positive imaginary part are presented, as the complex eigenvalues always occur in conjugate pairs. 7 6 local modes imaginary 5 damping ratio=5% 4 3 2 1 0 -2 inter-area modes governor modes exciter modes exciter modes -1.5 -1 -0.5 0 real Figure 6.10: Oscillatory modes in the two-area system. -97- 0.5 Chapter 6 The Physical Nature of Inter-area Oscillations in Electrical Power Systems The eigenvalues and the system state matrix are now used to calculate the eigenvectors and participation factors that correspond to the inter-area oscillatory modes. The elements of the right eigenvector corresponding to the 28th eigenvalue, which is associated with the inter-area mode, are given in Table 6.3. Table 6.3: Right eigenvector for eigenvalue 28 (associated with inter-area mode). State State No. Magnitudes Phase angles No. Magnitudes Phase angles variables variables 1 G1 δ 0.000 107.24 25 G3 δ 0.514 -4.48 ω ω 2 G1 0.001 -124.43 26 G3 0.005 90.82 0.027 43.75 27 G3 ψ 1d 0.022 121.14 3 G1 ψ 1d 4 G1 ψ 1q 0.025 162.25 28 G3 ψ 1q 0.054 150.66 5 G1 ψ fd 0.023 74.92 29 G3 ψ fd 0.017 85.03 6 G1 ψ 2 q 0.017 145.31 30 G3 ψ 2 q 0.037 133.71 7 G1 v ph 0.002 101.16 31 G3 v ph 0.000 111.40 8 G1 va 0.052 9.01 32 G3 va 0.015 19.25 9 G1 vt 0.809 164.68 33 G3 vt 0.229 174.92 10 G1 tg ser 0.045 36.97 34 G3 tg ser 0.223 -107.78 11 G1 tg hp 0.023 -22.05 35 G3 tg hp 0.113 -166.80 12 G1 tg rh 0.001 -108.20 36 G3 tg rh 0.007 107.04 13 14 0.032 0.001 -24.28 -129.78 37 38 -4.94 90.90 0.034 44.97 39 G4 δ G4 ω G4 ψ 1d 0.472 0.004 15 G2 δ G2 ω G2 ψ 1d 0.022 108.72 16 G2 ψ 1q 0.034 154.12 40 G4 ψ 1q 0.054 150.55 17 G2 ψ fd 0.030 71.89 41 G4 ψ fd 0.02 81.26 18 G2 ψ 2 q 0.023 137.17 42 G4 ψ 2 q 0.037 133.61 19 G2 v ph 0.002 98.91 43 G4 v ph 0.001 105.22 20 G2 va 0.065 6.76 44 G4 va 0.023 13.07 21 G2 vt 1.000 162.42 45 G4 vt 0.347 168.73 22 G2 tg ser 0.029 31.62 46 G4 tg ser 0.201 -107.71 23 G2 tg hp 0.015 -27.40 47 G4 tg hp 0.102 -166.72 24 G2 tg rh 0.001 -113.55 48 G4 tg rh 0.006 107.12 -98- Chapter 6 The Physical Nature of Inter-area Oscillations in Electrical Power Systems Figure 6.11 gives the compass plot of the right eigenvector elements associated with generator speeds. As seen in the compass plot, the speed changes in Area 1 are approximately in anti-phase with the speed changes in Area 2. Furthermore, the magnitudes of the right eigenvector components are larger in Area 2, than they are in Area 1. These characteristics of the inter-area oscillatory mode given by the right eigenvector are confirmed by the nonlinear simulation results (see Figure 6.4). 90 120 0.004 G3 60 0.003 G4 150 0.005 30 0.002 0.001 180 0 G2 G1 210 330 240 300 270 Figure 6.11: Right eigenvector (mode shape) of inter-area mode. To obtain more information about the inter-area oscillatory mode, the participation factors are calculated, as shown in Table 6.4. The participation factors have all been normalized, so that the largest element is 1.00. This allows the source of the oscillatory mode to be identified. The states associated with the rotor speed and phase angle of the generators dominates the inter-area oscillatory mode. The participation factors associated with the rotor angles and generator speeds are much higher than those participation factors associated with the other state variables. However, since generator 1 is considered as the reference machine, the participation factor associated with the rotor angle of generator 1 is zero. -99- Chapter 6 The Physical Nature of Inter-area Oscillations in Electrical Power Systems Table 6.4: Participation vector for Eigenvalue 28 (associated with inter-area mode). No. State variables Magnitudes Phase angles No. State variables Magnitudes Phase angles 0.000 166.68 25 0.00 0.563 -176.82 26 0.496 -177.31 3 G1 ψ 1d 0.009 145.82 27 0.007 42.94 4 G1 ψ 1q 0.011 41.45 28 0.025 -148.78 5 G1 ψ fd 0.113 -126.54 29 0.082 63.31 6 G1 ψ 2 q 0.012 41.45 30 δ G3 ω G3 ψ 1d G3 ψ 1q G3 ψ fd G3 ψ 2 q 1.00 2 δ ω G1 0.027 -148.78 7 G1 v ph 0.038 -2.88 31 G3 v ph 0.011 -172.91 8 G1 va 0.025 -27.22 32 G3 va 0.007 162.76 9 G1 vt 0.005 13.82 33 G3 vt 0.001 -156.20 10 G1 tg ser G1 tg hp 0.009 36.95 34 G3 0.044 71.97 0.024 -40.67 35 tg ser G3 tg hp 0.118 -5.65 0.005 -109.60 36 G3 -74.59 0.102 144.67 37 0.59 3.32 0.374 -175.00 38 tg rh G4 δ G4 ω 0.024 14 tg rh G2 δ G2 ω 0.257 -173.44 15 G2 ψ 1d 0.011 148.39 39 G4 ψ 1d 0.007 32.04 16 G2 ψ 1q 0.017 46.94 40 G4 ψ 1q 0.028 -135.88 17 G2 ψ 0.14 -128.22 41 G4 ψ fd 0.091 61.05 18 G2 ψ 2 q 0.019 46.94 42 G4 ψ 2 q 0.031 -135.88 19 G2 v ph 0.045 -3.78 43 G4 v ph 0.016 -177.58 20 G2 va 0.029 -28.12 44 G4 va 0.01 158.09 21 G2 vt 0.006 12.92 45 G4 vt 0.002 -160.87 22 G2 0.004 34.25 46 G4 75.30 0.011 -43.37 47 tg ser G4 tg hp 0.028 23 tg ser G2 tg hp 0.075 -2.32 24 G2 0.002 -112.31 48 G4 0.015 -71.25 1 11 12 13 G1 G1 fd tg rh G3 tg rh 6.4 The origin of lightly damped/unstable inter-area oscillations Actually, a number of factors influence the inter-area oscillatory mode, such as the type of excitation control, power flow conditions, transmission network structure and load characteristics [59] [60]. Since the primary task of excitation control is to ensure voltage stability or transient stability, once the controller is installed its parameters are rarely changed [4]. The loads in a power system are distributed all across the entire power network; this means that it is impossible and not practical to introduce control measures that use load management to enhance small signal stability. Therefore, this work focuses upon analyzing the effect of power flow and transmission network parameters on the inter-area oscillatory mode. -100- Chapter 6 The Physical Nature of Inter-area Oscillations in Electrical Power Systems To investigate the effect of the inter-area active power flow on the inter-area mode, modal analysis was executed for a range of inter-area active power flow conditions. The frequency and damping ratio of the inter-area mode for the different inter-area power flow conditions are given in Table 6.5. As seen from Table 6.5, the frequency and damping ratio of the inter-area mode are reduced as the inter-area active power flow is increased. Figure 6.12 shows how eigenvalues of the system’s oscillatory modes change with the power flow condition. The arrows represent the direction of the movement of each eigenvalue that occurs as the inter-area power flow is increased. Table 6.5: The effect of inter-area power flow on inter-area mode. Active power flow (area 1 to area 2, MW) 100 Load in area 2(MW) Frequency Damping ratio (%) 1467 0.587 1.235 150 1517 0.586 0.855 200 1567 0.583 0.540 250 1617 0.577 0.262 300 1667 0.567 -0.012 350 1717 0.554 -0.345 400 1767 0.537 -0.789 7 6 local modes imaginary 5 damping ratio=5% 4 3 inter-area mode 2 governor mode exciter modes 1 0 -1.8 exciter modes -1.6 -1.4 -1.2 -1 -0.8 -0.6 -0.4 -0.2 0 real Figure 6.12: The effect of inter-area power flow on system oscillatory mode. 0.2 In addition, the effect of the impedance of the inter-tie lines on the inter-area mode was investigated. The inter-tie lines’ impedance was varied by changing the number of tie lines. Under the same inter-area power flow conditions but with a different number of inter-tie lines in service, modal analysis was once more applied to the system. The frequencies and damping ratios of the inter-area mode in the different cases are given in -101- Chapter 6 The Physical Nature of Inter-area Oscillations in Electrical Power Systems Table 6.6. In Table 6.6, “0” denotes the scenario in which all four of the inter-tie lines were in service; whereas “-1” denotes the scenario in which one inter-tie line (line 6) was out of service due to maintenance. As seen from the results of modal analysis in Table 6.6, the frequency and damping ratio of the inter-area mode were reduced as the impedance was increased. Table 6.6: Effect of inter-tie line impedance on inter-area mode. 0 Power flow (area 1 to area 2, MW) 100 0.587 1.235 -1 100 0.524 0.634 0 150 0.586 0.855 -1 150 0.522 0.416 0 200 0.583 0.540 -1 200 0.515 0.210 0 250 0.577 0.262 -1 250 0.503 -0.026 Operation scenario Frequency Damping ratio (%) 0 300 0.567 -0.012 -1 300 0.487 -0.372 0 350 0.554 -0.345 -1 350 0.463 -2.853 0 400 0.537 -0.789 -1 400 0.428 -2.140 From above analyses, it is clear to conclude that increasing long-distance power transfers and impedance of the transmission corridor causes the inter-area oscillation to become lightly damped or even unstable. Here, the increased impedance of a long transmission corridor is considered as the consequence of the outage of a transmission line due to maintenance. 6.5 Conclusions In this chapter, the physical nature of power system oscillations has been investigated in both the time domain and the frequency domain. In the time domain, a series of nonlinear simulations were carried out. By using different types of disturbances, the local modes and the inter-area mode were excited, and the characteristic of each oscillatory mode was clearly observed. The generator speed and system frequency response to disturbances are the essential signals for confirming different modes of oscillations. Whilst the active power flows across different transmission corridors can be monitored to identify oscillation modes (oscillatory frequency and damping factor). -102- Chapter 6 The Physical Nature of Inter-area Oscillations in Electrical Power Systems In terms of system operation monitoring, these oscillatory signals, such as system frequencies, active power flows and voltage angles between different buses, can be directly captured by PMUs installed in substations. To understand the physical phenomenon of the inter-area oscillation accurately, modal analysis was applied to the two-area system. Using modal analysis, the information behind the physical phenomena of inter-area oscillations, such as oscillatory frequency, damping factor, mode shape and the source of the oscillatory mode, was obtained. As observed from the modal analysis, the inter-area mode is influenced by inter-area active power flow conditions and transmission network parameters. A significant increase of inter-area active power flow and the loss of an inter-tie line will result in previously stable inter-area oscillations becoming lightly damped or unstable. Therefore, a real time monitoring and warning system is a key requirement to detect if a power system is experiencing lightly damped or unstable inter-area oscillations. -103- Chapter 7 Inter-area Oscillations Monitoring using Newton-Type Algorithm Chapter 7 Inter-area Oscillation Monitoring Using Newton-Type Algorithm 7. 1 Introduction Low frequency electromechanical oscillations commonly occur in electrical power systems, as a consequence of various disturbances and are closely related to generators’ rotor speed oscillations. As presented in Chapter 6, the oscillations associated with a group of generators in one area swinging against a group of generators in another area are called inter-area oscillations; and the frequency of inter-area oscillation is typically in the range of 0.1-0.8 Hz [34]. The observation model of the inter-area oscillations, y(t), can be represented through the following nonlinear parameter model with three unknown model parameters: y (t ) = A ⋅ e (σ + j 2πf ) t (7.1) where A is the amplitude of the oscillation, σ is the damping factor and f is the oscillation frequency. The magnitude of the oscillations decays if σ is negative, whereas it increases if σ is positive. A practical measure for the assessment of oscillations is the damping ratio, ζ , defined as: ζ = −σ σ + ( 2πf ) 2 2 × 100% (7.2) In order to ensure a sufficient stability margin in the system, the damping ratio ζ should be greater than 5% [59]. Poorly damped, or unstable oscillations ( ζ smaller than 5%) are a risk since they can lead to undesirable system conditions such as instabilities, cascading events, or ultimately a catastrophic system blackout. A typical example is the August 1996 blackout of the western US/Canada interconnected power system [62]. In large interconnected power systems, poorly damped or unstable inter-area oscillations usually occur between power grids which are weakly connected. Here, the weak connection refers to two adjacent power grids connected over high impedance -104- Chapter 7 Inter-area Oscillation Monitoring Using Newton-Type Algorithm transmission lines, or a transmission corridor that was subjected to a sudden, and substantial change in power flow [59]. To mitigate such catastrophic scenarios, system operators require reliable, high quality monitoring to establish the presence and emergence of inter-area oscillations. A SMT-based Wide Area Monitoring System (WAMS), provide a great opportunity to monitor inter-area oscillations and inform the operator when the system experiences poorly damped or unstable inter-area oscillations. Considerable research has been undertaken into developing real-time techniques for identifying power oscillation modes. The Prony method is one of the most widely used methods for identification of oscillation modes [63], however, in practical applications noise elimination and determination of the model order are still a challenge [64], [65], [66]. The Kalman Filter (KF) has been also utilized for real-time detection of the dominant oscillation modes [67]. References [68], [69], [70] and [71] present algorithms associated with the Least Square (LS) method based applications. Fourier Transform based solutions are proposed in [65] and [72]. This Chapter introduces and tests a new numerical algorithm, Newton Type Algorithm (NTA), for the real-time estimation of the dominant inter-area oscillation mode [73]. It has been demonstrated that the NTA provides a robust estimation technique for processing input signals corrupted by random noise, and in addition the algorithm can successfully process multiple signals. This is very important for obtaining the inter-area oscillation mode shape (to detect inter-area oscillations between groups of generators), which is very important to ensure stability for future bulk interconnected power systems. The determination of the inter-area oscillation mode shape is a prerequisite for a advanced control scheme e.g. intelligent controlled islanding. 7.2 Signal Model Representation The free motion of a power system around an operating point, after a small disturbance, can be described by the following matrix equation [34]: x& = Ax where A is the n × n system state matrix and x is an (7.3) n ×1 system state vector. The solution of (7.3) for the i-th state variable, xi , in terms of the eigenvalues -105- Chapter 7 Inter-area Oscillations Monitoring using Newton-Type Algorithm λk = σ k ± jωk (the oscillatory mode) of the system state matrix can be represented as follows: n x i (t ) = A0 + ∑ Ak ⋅ e σ k ⋅t ⋅ sin(ω k t + φ k ) (7.4) k =1 where A0 is a DC component, Ak is the magnitude of the k-th oscillatory component, σ k and ω k = 2π f k are the damping factor and the oscillatory frequency, respectively and φ k is the phase angle of the component. The DC component is included in the signal model because the signal processed can oscillate around a non-zero value, this will be shown in the algorithm testing and processing of real data records. Given a system disturbance, the observation of the response of the i-th state variable xi can be mathematically modelled through the following general nonlinear equation: y (t ) = h(x(t ), t ) + ξ (t ) (7.5) in which ξ (t ) is a zero mean Gaussian random noise, x(t ) is a suitable selected timevarying parameter vector and h(x(t ), t ) is the suitable parameter model of the processed signal. In this particular case, the model is: n h(x(t ), t ) = A0 (t ) + ∑ Ak (t ) ⋅ eσ k (t )t ⋅ sin(ωk (t )t + φk (t )) (7.6) k =1 The parameter model (7.6) is obviously highly nonlinear, so to estimate the unknown model parameters, the nonlinear estimation techniques must be used. Let us define the following 4n +1 vector of the unknown model parameters, x(t ) , for the signal model (7.6): x ( t ) = [ A0 ( t ) A1 ( t ) σ 1 ( t ) ω1 ( t ) φ1 ( t )... Ak ( t ) σ k ( t ) ω k ( t ) φ k ( t )... An ( t ) σ n ( t ) ω n ( t ) φ n ( t )] (7.7) From the sampled values of the input signal, the following discrete representation of the above signal model (7.6) can be created: -106- Chapter 7 Inter-area Oscillation Monitoring Using Newton-Type Algorithm y i = h( x i , t i ) + ξ i (7.8) where: n h(xi , ti ) = A0,i + ∑ Ak ,i e σ k , i ⋅t i k =1 ⋅ sin(ωk ,iti + φk ,i ) i = 1,2,... (7.9) In (9) A0 ,i , Ak ,i , σ k ,i , ω k ,i , φk,i , ξ k,i and t i refer to the values of A0 (t ) , Ak (t ) , σ k (t ) , ω k (t ) , φk (t ) , ξk (t ) and t at the discrete time index i. If the signal observed ( y (t ) ) is uniformly sampled with the sampling frequency fs=1/T Hz during a finite period of time (defined as a data window), the value of t at a discrete time index i is given by ti = iT. From m samples taken during a data window, with the size Tdw = m × T , a set of the following m nonlinear equations with 4n + 1 unknown parameters can be determined: y i = h( x i , t i ) + ξ i (7.10) where n h(xi , ti ) = A0, i + ∑ Ak , i ⋅ e σ k , i ⋅t i k =1 ⋅ sin(ωk , i ti + φk , i ) i = 1,2,..., m (7.11) In the case of m samples belonging to the data window at ti = m × T , (10) can be expressed in the following vector form: y = h ( x) + ξ (7.12) where y = [ y (1),... y ( m )]T is an m×1 measurement vector, h ( x ) = [ h ( x1 , t1 ),...h( x m , t m )]T is an m × 1 vector of nonlinear functions given by (7.11) and ξ = [ξ1 ,..., ξ m ]T is an m×1 error vector. The vector equation (7.12) represents a set of m nonlinear equations with 4n + 1 unknowns. It can be solved if m ≥ 4n + 1 using different numerical approaches. This Chapter demonstrates the use of the NTA for solving the problem stated in (7.12). -107- Chapter 7 Inter-area Oscillations Monitoring using Newton-Type Algorithm The solution (the estimates of the unknown model parameters) can be obtained iteratively, or through a sequential update of estimates obtained from a single data window. This is presented in the next Section. 7.3 Newton Type Algorithm Derivation For simplicity, the error vector ξ in (7.12) is temporarily neglected and this expression has been modified to the following matrix equation: F ( x) = h ( x) − y = 0 (7.13) where F(x) is an m×1 vector of a set of suitable selected nonlinear equations (the processed signal model) and 0 is a m×1 zero vector. The parameter vector x can be iteratively calculated by solving the nonlinear matrix equation (7.13). The initial guess xi must be used to initialize the iterative procedure, as presented later. Therefore, in the first iteration the following can be assumed: F (x) ≠ 0 (7.14) There exists an unknown correction vector Δx i , which has to be determined so that the following holds: F ( x i + Δx i ) = 0 (7.15) From the Taylor series expansion of F(x) , linearized in the neighbourhood of xi, the following expression is obtained: F(xi + Δxi ) ≅ F(xi ) + J (xi )Δxi (7.16) where J ( x i ) is an m × (4n + 1) Jacobian matrix containing the partial derivatives ∂ Fp / ∂ xk (p = 1,...,m and k = 1,…,4n+1). The Jacobian matrix J ( x i ) is defined as: -108- Chapter 7 Inter-area Oscillation Monitoring Using Newton-Type Algorithm ⎡ ∂F1 ∂F1 ∂F1 ∂F1 ∂F1 ∂F1 ∂F1 ∂F1 ∂F1 ∂F1 ∂F1 ∂F1 ∂F1 ⎤ ⎥ ⎢ ∂A ∂A ∂σ ∂ω ∂φ .. ∂A ∂σ ∂ω ∂φ .. ∂A ∂σ ∂ω ∂φ k k k k n n n n ⎥ ⎢ 0 1 1 1 1 J ( xi ) = ⎢ ... ... ... ⎥ ⎥ ⎢ ⎢ ∂Fm ∂Fm ∂Fm ∂Fm ∂Fm .. ∂Fm ∂Fm ∂Fm ∂Fm .. ∂Fm ∂Fm ∂Fm ∂Fm ⎥ ⎢⎣ ∂A0 ∂A1 ∂σ1 ∂ω1 ∂φ1 ∂Ak ∂σ k ∂ωk ∂φk ∂An ∂σ n ∂ωn ∂φn ⎥⎦ (7.17) where ∂F p ∂A 0 = 1 (7.18) ∂Fp ∂Ak = eσ k t ⋅ sin(ωk t + φk ) (7.19) ∂Fp ∂σ k = t ⋅ Ak ⋅ eσ kt ⋅ sin(ωk t + φk ) (7.20) ∂Fp ∂ωk = t ⋅ Ak ⋅ eσ k t ⋅ cos(ωk t + φk ) (7.21) ∂Fp ∂ φk = Ak ⋅ eσ k t ⋅ cos(ωk t + φk ) (7.22) (k=1,…,n), where n is the number of oscillatory components existing in the estimation model. For simplicity the notation for J(x i ) and F ( x i ) can be simplified to J i and Fi , respectively. Taking into consideration the vector ξ , (7.16) can be rewritten as: J i Δxi ≅ −Fi = −[h(xi ) − y + ξ ] = y − h(xi ) − ξ (7.23) Taking into account the error that occurred by neglecting the higher order terms in the Taylor series expansion, it further follows that: J i Δx i = y − h(x i ) + ς where ς is an m ×1 (7.24) error vector that includes the error vector ξ and the errors produced by neglecting the higher order terms in Taylor’s series expansion. By minimizing the sum of the square of the errors in (7.24), the following unknown correction vector Δxi can be obtained: Δx i = ( J Ti J i ) −1 J Ti [ y − h ( x i )] = J i# [ y − h ( x i )] -109- (7.25) Chapter 7 Inter-area Oscillations Monitoring using Newton-Type Algorithm where J i# is the left pseudo-inverse of J i . By adding the corrective vector to the original guess, xi, one obtains: x i +1 = x i + Δx i = x i + J i# [ y − h ( x i )] (7.26) The above equation is essential to the NTA. Typically, the iterative procedure should stop when all of the unknown parameters in the (i+1)th iteration do not differ from those in the i-th iteration by more than a specific tolerance, ε, that is defined in advance by the user, this condition indicates that the optimal solution has been reached. The algorithm presented in this thesis, like any other nonlinear estimator, may prematurely converge to a local minimum. Therefore, it is important to make sure that the initial estimate lies close to the true solution. If the initial estimate from which the iteration begins is far away from the global minimum, the method may reach one of the local minima or, in the worst case, totally diverge. To find an appropriate initial starting point for the NTA, the Fast Fourier Transform (FFT) algorithm is used. The FFT is used to find the number of dominant oscillatory components, n, in the samples captured in the initial data window; after that the Least Square (LS) method [74] is used to estimate the parameters of each oscillatory component, A0 , Ak , σ k , ωk = 2πf k , φk , k = 1,..., n . Furthermore, the algorithm requires appropriate choice of the following algorithm parameters: data window size Tdw and the sampling frequency f s . In the next Section the algorithm’s sensitivity to the selection of data window size and sampling frequency will be discussed. 7.4 Computer Simulated Tests In this Section, the NTA is tested by means of the input signals obtained through computer simulations. In Figure 7.1 the block diagram of the test procedure is given. -110- Chapter 7 Inter-area Oscillation Monitoring Using Newton-Type Algorithm xk ' xk ' xk δk Figure 7.1: A global block diagram of the test procedure. As shown in Figure 1, Block 1 generates groups of test signals with the signal parameters x k defined in the INPUT Block. Block 2 represents the NTA applied for the estimation of unknown parameters. The results of the estimation xk ' (saved in Block 3) are compared to the actual values of x k and then the error vector δ k = x k − x k ' is analyzed in Block 4. The results of the error analysis are saved in Block 5. The computer simulation tests comprise Static tests, Noise tests and Dynamic tests. 7.4.1 Static tests The test signals were generated in advance using known parameters. The first representative signal generated for the static tests is described by the following equation: y (t ) = A0 + A1 ⋅ e σ 1 t ⋅ sin( 2πf 1t + φ1 ) (7.27) where A0 = 10 , A1 = 5 , σ 1 = −0 .2 , f1 = 0.5 and φ1 = 60 . This signal consists of a DC component and a single damped oscillatory component, as presented in Figure 7.2. In this test, the sampling frequency fs = 100 Hz and data window size Tdw = 2 s. Figures 7.3-7.7 present the estimation results of the unknown parameters. Figure 7.8 shows the comparison between the original signal and the signal that was constructed with estimated parameters. The estimation errors are practically negligible in all cases, leading to the conclusion that the algorithm is suitable for the processing of steady state signals. -111- Chapter 7 Inter-area Oscillations Monitoring using Newton-Type Algorithm y(t) 15 10 5 0 2 4 6 8 10 time [s] Figure 7.2: Computer generated test signal. A0 Actual Esti. 10 9.8 2 3 4 5 6 time [s] 7 8 9 10 Figure 7.3: Estimation results of the magnitude of DC component, A0. Actual Esti. A1 5.2 5 4.8 2 3 4 5 6 time [s] 7 8 9 10 damping factor Figure 7.4: Estimation results of the magnitude of oscillatory component A1. Actual Esti. -0.15 -0.2 -0.25 2 3 4 5 6 time [s] 7 8 9 10 Figure 7.5: Estimation results of the damping factor of oscillatory component, σ . -112- Chapter 7 Inter-area Oscillation Monitoring Using Newton-Type Algorithm 0.52 f [Hz] Actual Esti. 0.5 0.48 2 3 4 5 6 time [s] 7 8 9 10 phase angle [deg] Figure 7.6: Estimation results of the frequency of oscillatory component, f. 60.2 60 Actual Esti. 59.8 2 3 4 5 6 time [s] 7 8 9 10 Figure 7.7: Estimation results of the phase angle of oscillatory component, φ . Actual Esti. y(t) 15 10 5 0 1 2 3 4 time [s] 5 6 7 8 Figure 7.8: Computer generated test signal and estimated signal. 7.4.2 Noise tests In this Subsection, the NTA algorithm is tested using computer generated test signals corrupted by a white zero-mean Gaussian random noise, which was superimposed on to the magnitude of the input test signal. The random noise standard deviations are selected to obtain a prescribed value of the Signal-to-Noise-Ratio (SNR), defined as follows: SNR = 20 log -113- A0 2σ (dB) (7.28) Chapter 7 Inter-area Oscillations Monitoring using Newton-Type Algorithm where A0 is the magnitude of the DC component, and σ is the noise standard deviation. Here, three series of noise tests are performed. In all cases the mean error vector δ k of the estimation results for the unknown parameters is defined as: n δk = ∑x i =1 k − xk ' n xk ×100% (7.29) where x k represents the actual unknown parameters, and xk ' is the estimated values of the unknown parameters, n is the number of estimation results. 7.4.2.1 Sensitivity to noise To test the sensitivity of the NTA to random noise, signals with different white noise level (i.e. different SNRs) were processed. The SNR was varied in the range 30-90 dB. In all tests the sampling frequency used was f s = 100Hz and the data window size was Tdw = 3s . The results of the sensitivity analysis are presented in Table 7.1. The results suggest that larger errors are obtained for a higher noise level. Table 7.1: Sensitivity analysis for random noise. Mean error in the estimation results (%) Parameters SNR=30 dB SNR= 50dB SNR=70 dB SNR=90 dB 0.6088 0.31985 0.0826 0.0317 A0 A1 58.0023 25.62634 14.8779 5.0705 σ1 39.5772 17.74586 8.7892 3.0651 f1 (Hz) 3.7504 1.55972 0.5360 0.2177 φ1 (deg) 92.0565 36.4835 13.2939 5.2441 7.4.2.2 Sensitivity to sampling frequency The second test considers the impact of the data sampling frequency on the quality of the NTA estimation. The test signal was corrupted by an additive white noise with SNR = 50 dB . The data window size was set to Tdw = 3s . The results (summarized in Table 7.2) indicate that for a higher sampling frequency better estimation results are obtained. However, in practice, the improved accuracy obtained by increasing the sampling frequency would need to be balanced against the processing time requirements. -114- Chapter 7 Inter-area Oscillation Monitoring Using Newton-Type Algorithm Table 7.2: Sensitivity analysis for sampling for frequency. Mean error in the estimation results (%) fs =100 Hz fs =200 Hz fs =500 Hz Parameters fs =50 Hz fs =1000 Hz A0 0.38549 0.31985 0.17782 0.1412 0.10415 A1 39.34897 25.62634 18.96515 14.54942 12.31657 σ1 22.08241 17.74586 14.90845 9.40634 8.12717 f1 (Hz) 2.60918 1.55972 1.00475 0.52258 0.50971 φ1 (deg) 62.29959 36.4835 23.87996 12.32148 11.9796 7.4.2.3 Sensitivity to data window size The impact of the data window size, Tdw , on the quality of the estimation was also studied with a test signal corrupted by a random noise with SNR = 50 dB and sampled with f s = 100Hz . The results obtained are presented in Table 7.3 and indicate that the selection of the data window size must be carefully considered, particularly since very short/long data windows (e.g. 3, or 6 seconds) produce unsatisfactory estimates. Table 7.3: Sensitivity analysis for data window size. Parameter A0 Mean error in the estimation results (%) Tdw =3 s Tdw =4 s Tdw =5 s Tdw =6 s 0.31985 0.25163 0.21066 0.19179 A1 25.62634 24.22365 7.45586 7.06324 σ1 17.74586 12.88686 6.91787 5.09057 1.55972 0.75748 0.58151 0.43645 36.4835 15.44149 11.7549 7.3249 f 1 (Hz) φ1 (deg) In Figure 7.9, both the estimated and processed noisy signals are presented ( SNR = 50 dB , f s = 100 Hz , Tdw = 3s ), suitably demonstrating that the algorithm is capable of processing noisy signals. Actual Esti. y(t) 15 10 5 0 1 2 3 4 5 6 7 time [s] Figure 7.9: Algorithm tracking capabilities in presence of noise. -115- Chapter 7 Inter-area Oscillations Monitoring using Newton-Type Algorithm 7.4.3 Dynamic tests In this Section the results from the dynamic testing will be presented. Here the NTA properties during step changes of signal parameters are investigated. The test signal defined in (27) had a step change of the damping factor σ (from -0.2 to 0) of the oscillatory mode at t = 5 s. In this test a sampling frequency of f s = 100Hz was selected, whereas the data window size was changed in the range of 2-4 seconds. Figure 7.10 illustrates the original test signal. As the damping factor of the oscillatory component approaches zero at t=5 s, the magnitude of the signal remains constant (undamped oscillations). Figures 7.11-15 present the estimation results of the unknown parameters for different data window sizes. The fastest convergence was obtained with the shortest window size ( Tdw = 2 s ), but with a large estimation deviation during convergence period. As the data window size was increased accuracy improved, but with slower convergence. Figure 7.16 provides a comparison between the real test signal and the estimated signal ( Tdw = 2 s ).A very high level of correlation can be observed, confirming the algorithm accuracy. y(t) 15 10 5 0 2 6 8 10 time [s] Figure 7.10: Computer generated signal with step change of σ . 12 4 Actual Tdw=2s Tdw=3s Tdw=4s A0 10 8 6 2 3 4 5 6 7 8 9 10 time [s] Figure 7.11: Estimation results of the magnitude of DC component, A0, for different Tdw. -116- Chapter 7 Inter-area Oscillation Monitoring Using Newton-Type Algorithm Actual A1 15 Tdw=2s Tdw=3s Tdw=4s 10 5 0 2 3 4 5 6 7 8 9 10 Figure 7.12: Estimation results of the magnitude of oscillatory component, A1, for different Tdw. 1.5 damping factor Actual Tdw=2s Tdw=3s Tdw=4s 1 0.5 0 -0.5 2 3 4 5 6 time [s] 7 8 9 10 Figure 7.13: Estimation results of damping factor, σ , for different Tdw. f [Hz] 0.5 0.4 Actual Tdw=2s 0.3 Tdw=3s Tdw=4s 2 3 4 5 6 time [s] 7 8 9 10 Figure 7.14: Estimation results of the frequency of oscillatory component, f, for different Tdw. phase angle [deg] 400 Actual Tdw=2s 200 Tdw=3s Tdw=4s 0 2 3 4 5 6 time [s] 7 8 9 10 Figure 7.15: Estimation results of the phase angle of oscillatory component, φ , for different Tdw. -117- Chapter 7 Inter-area Oscillations Monitoring using Newton-Type Algorithm Actual y(t) 15 Esti. 10 5 0 1 2 3 4 5 6 7 8 time [s] Figure 7.16: Dynamic testing: comparison between the actual and estimated signal. 7.5 Dynamic Simulation of a Multi-machine System More realistic testing of the new algorithm was carried out using a dynamic simulation of a multi-machine power system. Figure 7.17 presents a block diagram of this test procedure. The power system was modelled and simulated using the simulation software ‘DIGSILENT PowerFactory’ [75] (see Block 1). Block 2 represents the oscillatory signal obtained through the DIGSILENT simulation of a large disturbance in the test system. The dominant inter-area oscillatory mode estimated by the NTA (represented by Block 3) is stored in Block 4 and compared with the results obtained using the Eigenvalue analysis performed by Block 5. λ =σ ± jω λ'=σ'±jω' δ Figure 7.17: A block diagram of the testing procedure based on dynamic simulation of a multimachine test system. The simulated two-area power system consists of two active networks in which each one has two generators (see Figure 7.18). The two active areas are connected via four AC transmission lines and one HVDC transmission link. The full set of the system parameters are given in Appendix B. -118- Chapter 7 Inter-area Oscillation Monitoring Using Newton-Type Algorithm Figure 7.18: Two-area test system with HVDC link. To provoke inter-area oscillations, at t = 2 s a transient three-phase short circuit fault was simulated in the middle of the line 5. This transient fault lasted for 0.1 s. Consequently, after the disturbance, the two groups of generators started to oscillate against each other, as seen in the generator rotor speed changes presented in Figure 7.19. rotor speed (p.u.) 1.006 G1 G2 G3 G4 1.004 1.002 1 0.998 0.996 0.994 0 5 10 15 time [s] 20 25 30 Figure 7.19: Generator rotor speed changes after the disturbance. The generator rotor speed variations are simultaneously analysed using the NTA. These oscillatory signals were processed with a sampling frequency f s = 100Hz and a data window size Tdw = 5s . In order to obtain a clear mode shape of the inter-area oscillation mode and to avoid the influence of the other oscillation modes (e.g. local modes and control modes), only the oscillatory signals in the range of 10-30 s are considered. In Figure 7.20, the estimated magnitudes and phase angles of the dominant oscillatory component of each processed signal are presented in a polar diagram. The estimation results were updated every two seconds and the polar diagram provides a very clear view of the shape of the dominant inter-area oscillatory mode. The phase angles of the oscillatory speeds of the generators in area 1 are in anti-phase with respect to the oscillatory speeds of the generators in area 2. -119- Chapter 7 Inter-area Oscillations Monitoring using Newton-Type Algorithm 90 120 60 G2 150 30 G1 180 0.001 0.002 0.003 0 G3 210 330 G4 240 300 270 Figure 7.20: Inter-area oscillatory mode shape estimated by NTA. For the fault described above, the oscillatory active power flow over one of the AC inter-tie lines (line 2) is presented in Figure 7.21. This oscillatory signal (from 5 s to 30 s) was processed with the sampling frequency f s = 100Hz and the data window size Tdw = 5s . In Figures 22-23, the estimated dominant inter-area oscillation mode (damping factor and frequency) obtained from the oscillatory active power flow is presented. The estimation results obtained by using the NTA were compared with the Prony method and Eigenvalue analysis (performed under steady state). In the Prony analysis, an estimation model with 21 orders was used for better fitting the original oscillatory signal. As seen from Figures 7.22-23, the results obtained using the NTA are practically the same as the results obtained using the classical Eigenvalue analysis. Furthermore, it is demonstrated that the NTA algorithm delivered more accurate and stable estimation results, compared to the Prony method. This is because the Prony method is based on a linear approximation technique and the oscillation modes are calculated by solving independent approximated polynomial equations with the data captured by individual data windows. In the case of the NTA method, the oscillation unknown parameters are calculated by the continuous iteration procedure based on a set of partial differential equations. In the iteration procedure, the estimation results obtained from the previous sliding data window are the initial points for the next iteration using the data from the next data window. In order to eliminate the effects of the weaker high order components in the estimation model used for Prony method, the sampling frequency of original oscillatory signal was -120- Chapter 7 Inter-area Oscillation Monitoring Using Newton-Type Algorithm reduced to f s = 5 Hz (data window size Tdw = 5s ); the Prony estimation model was reduced to have 3 orders, i.e., one DC component and one oscillatory component. The dominant inter-area oscillation mode (damping factor and frequency) estimated by the Prony method are shown in Figures 7.24-25. As seen from these plots, the quality of the estimation of the dominant inter-area oscillation mode was significantly improved. The estimation results obtained from NTA algorithm are still more accurate and stable than the results obtained from Prony method. Figure 7.26 presents the reconstructed signal based on the NTA estimated parameters. P (MW) 200 100 0 0 5 10 15 time [s] 20 25 30 Figure 7.21: Oscillatory active power (on line 2) after the disturbance. frequency [Hz] 0.54 0.52 0.5 10 Prony NTA Eigen 20 25 30 time [s] Figure 7.22: Estimated frequency of the inter-area oscillatory mode by NTA and Prony method. damping factor 15 0 -0.1 -0.2 10 Prony 15 NTA Eigen 20 25 30 time [s] Figure 7.23: Estimated damping factor of the inter-area oscillatory mode by NTA and Prony method. -121- Chapter 7 Inter-area Oscillations Monitoring using Newton-Type Algorithm frequency (Hz) 0.53 Reduced-order Prony NTA Eigen 0.52 0.51 10 15 20 25 30 time [s] Figure 7.24: Estimated frequency of the inter-area oscillatory mode by reduced-order Prony damping factor method and NTA. 0 Reduced-order Prony NTA Eigen -0.05 -0.1 10 15 20 25 30 time [s] Figure 7.25: Estimated damping factor of the inter-area oscillatory mode by reduced-order Prony method and NTA. P (MW) 200 100 Actual 0 0 5 10 15 time [s] 20 Esti. 25 30 Figure 7.26: Estimated oscillatory active power from estimated parameters. 7.6 Real-life Conditions Tests This Section considers an example of the off-line assessment of the inter-area oscillations in the GB power network using NTA. On November 8th 2010 at 07:24:00 the sudden disconnection of a large generating unit caused a significant mismatch between the active power generated and the active power consumed by the system. The phase angle difference measured between two remote locations in the network (Glasgow and London) is shown Figure 7.27, along with the system response through damped inter-area oscillations. The angular differences were recorded by the FlexNet Wide Area Monitoring System installed in the GB network [76]. -122- Chapter 7 Inter-area Oscillation Monitoring Using Newton-Type Algorithm A post-mortem analysis of the event was performed using the NTA with the sampling frequency f s = 100 Hz and the data window size Tdw = 5 s , which estimates the dominant oscillatory frequency and the damping ratio defined to the values illustrated in Figures 7.28-7.29. The estimation results from NTA were compared to those of the Prony method. For the Prony analysis, the sampling frequency was reduced to f s = 5 Hz and the data window size was Tdw = 5s . Here, the Prony estimation model has 5 orders, i.e., one DC component and two oscillatory components. As seen from Figures 7.28-7.29, both methods delivered similar estimation results of the dominant inter-area oscillation mode; however, the results from the NTA more accurately present the dynamic characteristic of the system. This can be attributed to the NTA estimating the oscillation parameters based on the real oscillatory signal rather than a decimated signal. The results have been validated through a comparison between the real and the estimated signal as given in Figure 7.30. From this exercise it can be concluded that the NTA is capable of tracking phase angle [deg.] the unknown inter-area oscillations parameters. 59 58 57 56 0 2 4 6 8 10 time [s] frequency [Hz] Figure 7.27: Oscillatory voltage phase angle difference between Glasgow and London. Reduced-order Prony NTA 0.6 0.5 0.4 0.3 5 6 7 8 9 10 time [s] Figure 7.28: Estimated frequency of inter-area oscillatory mode in the GB network. -123- damping ratio [%] Chapter 7 Inter-area Oscillations Monitoring using Newton-Type Algorithm 40 20 0 Reduced-order Prony 5 6 7 8 NTA 9 10 time [s] Figure 7.29: Estimated damping ratio of inter-area oscillatory mode in the GB network. 59 Actual Esti. y(t) 58 57 56 0 1 2 3 4 5 t (s) Figure 7.30: Oscillatory signal based on estimated oscillatory parameters. 7.7 Conclusions This Chapter has presented an NTA algorithm developed for the estimation of inter-area oscillatory modes. In the static tests, the NTA has been shown to reliably identify the oscillatory mode in a computer generated signal. The robust nature of the NTA has been further demonstrated to perform the same duty in the presence of white noise, if the sampling frequency and data window size are selected properly. The dynamic capability of the NTA has also been successfully demonstrated by capturing the dynamic oscillatory mode when the parameters of the test signal experienced a step change. The performance of the NTA to estimate inter-area oscillatory modes has been shown to not only corresponded to the results of the Eigenvalue analysis, but it was also more accurate and stable than the well know conventional Prony method for both laboratory tests and real condition tests. The NTA enables the inter-area oscillation mode shape to be observed by processing the generator rotor speeds, which is an essential application for confirming the swing modes between different parts of a large system. This work has not been possible from existing oscillation monitoring methods. The NTA was tested on real recorded data and -124- Chapter 7 Inter-area Oscillation Monitoring Using Newton-Type Algorithm successfully extracted the dynamic oscillation parameters from the oscillations observed in the GB power system following a large disturbance in 2010. In summary, the research so far suggests that the NTA is a powerful tool for analyzing PMU data captured from Wide area monitoring and could be a key constituent of future WAM applications. -125- Chapter 8 The use of Power Electronic Devices for Damping Inter-area oscillations Chapter 8 The Application of Power Electronic Devices for Damping Inter-area oscillations 8.1 Introduction Presently, using Power System Stabilizers (PSSs) is the most cost-effective approach for low frequency power oscillations control [34] [77] [78] [79]. The PSS is used to add additional damping torque to the generator rotor by controlling the generator’s Automatic Voltage Regulator (AVR). The inputs to PSSs can be generator rotor speed, electrical power or terminal bus frequency [80]. These PSSs are effective in stabilizing local modes, and if accurately tuned may also be effective in stabilizing inter-area modes. However, the effectiveness of PSSs in damping inter-area oscillations is limited because inter-area modes are not as highly controllable and observable in the generator’s local signals [81]. In interconnected power systems, lightly damped or unstable inter-area oscillations usually occur between power grids which are weakly connected. Here, the weak connection refers to the stressed transmission lines carrying heavy power flow. In modern power transmission technology, Flexible AC Transmission Systems (FACTS) devices and High Voltage Direct Current (HVDC) system are the competitive solutions to enhance the power transfer capability of the currently stressed transmission lines. Once these power electronic devices have been installed in the inter-tie lines the additional inter-area oscillation damping controllers become available [80]. In this Chapter, a wide area inter-area oscillation control scheme designed for the future GB power system is presented. As the long term GB’s WAMPAC strategy proposes, the HVDC and TCSC installed in the transmission corridor between Scotland and England can be used to control the inter-area oscillations. In this Chapter, a process for designing inter-area oscillation damping controller using power electronic devices is presented. The typical two-area system will be used to illustrate the damping control design process. In the two-area system, each generator is equipped with a high gain static excited exciter and speed governor. There is no PSS installed in the system; thus the -126- Chapter 8 The use of Power Electronic Devices for Damping Inter-area oscillations system is representative of the sort that may benefit from the use of power electronic devices for damping inter-area oscillations. 8.2 Modal Analysis for Control Chapter 6 introduced the process of using modal analysis for analyzing the physical nature of the inter-area oscillations. Although electrical power systems are essentially nonlinear systems, we have seen that the oscillation information around an operating point can be accurately described with a linearized system model [80]. Therefore, we can use linear control theory, and the system dynamic information obtained from modal analysis, to control inter-area oscillations. In this section, the details behind the use of modal analysis for inter-area oscillation control are introduced. 8.2.1 Transfer functions As presented in Chapter 6, the power system dynamics around an equilibrium point can be described using the following set of state equations [34]: • Δ x = AΔx + B Δu (8.1) Δy = CΔx + DΔu (8.2) ∆x is the state vector of length n ∆u is input disturbance vector of length r ∆y is the output vector of length m n× n B is the input matrix of size n×r C is the output matrix of size m× n D is a matrix of size m× r defines the proportion of input which directly influences the A is the state matrix of size output, ∆y. The frequency domain representation of these state equations can be defined by taking their Laplace transforms, as follows: -127- Chapter 8 The use of Power Electronic Devices for Damping Inter-area oscillations sΔ x ( s ) − Δ x (0) = A Δ x ( s ) + B Δ u ( s ) (8.3) Δy ( s ) = CΔx( s ) + DΔu ( s ) (8.4) By rearranging Equation (8.3), we have: ( sI − A ) Δ x ( s ) = Δ x (0) + B Δ u ( s ) (8.5) Δx(s) = (sI − A) −1[Δx(0) + BΔu(s)] (8.6) Thus: Substituting Δx(s) into equation (8.4) allows the system outputs to be defined as: Δy(s) = C(sI − A)−1[Δx(0) + BΔu(s)] + DΔu(s) (8.7) Since we are only concerned with the transfer function between the system inputs and outputs, the element Δx(0) can be assumed to be zero. In addition, if we assume that the system outputs, y (s) , are not a direct function of the inputs, Δu( s ) , (i.e., D=0) then the open loop transfer function of the system is: G (s) = Δy ( s) = C ( s I − A) − 1 B Δu( s) (8.8) 8.2.2 Residue based damping controller design If we only concern the transfer function between the system single input Δu k (s ) and single output Δy j (s ) , (8.8) can be changed to: G(s) = Δy j (s) Δu k (s) n Ri i =1 s − λi = c j ( s I − A) − 1 b k = ∑ (8.9) where Ri = c j φi ψ i b k -128- (8.10) Chapter 8 The use of Power Electronic Devices for Damping Inter-area oscillations is the residue of the transfer function between the input Δuk (s ) and the output Δy j (s ) . c j is the jth row vector of the C matrix and b k is the kth column vector of the B matrix. φi is the right eigenvector associated with the eigenvalue λi , and ψ i is the left eigenvector associated with the eigenvalue λi . A residue describes the sensitivity of the corresponding eigenvalue, λi , to a feedback control. Figure 8.1 represents the open loop transfer function G ( s) and a positive feedback control transfer function H(s). For the original open loop control, Δuk (s ) is a single input and Δy j (s ) is a single output; whereas for the feedback control, Δy j (s ) is used as an input signal. The output of the feedback control will be used to modify the input of the original open loop control. When feedback control is applied, the eigenvalues of the original system are changed according to the rule defined in (8.11) [82]; Δλi is the shift of the eigenvalue λi caused by the closed loop feedback control. Δλi = Ri H (λi ) Δuk (s ) + (8.11) Δ y j (s ) G(s) + H(s) Figure 8.1: Closed loop system with feedback control. Figure 8.2 gives a block diagram of a typical feedback damping controller. The controller consists of an amplification block, a low-pass filter, a washout filter and several compensation blocks [83]. umax kd 1 1 + sTm sTw 1 + sTw 1 + sTlead 1 + sTlag 1 + sTlead 1 + sTlag umin Figure 8.2: The structure of a feedback damping control. -129- Chapter 8 The use of Power Electronic Devices for Damping Inter-area oscillations The transfer function of the damping control, H ( s ) , is given by: ⎛ 1 H ( s ) = k d ⎜⎜ ⎝ 1 + sTm ⎞ ⎛ sTw ⎟⎜ ⎟ ⎜ 1 + sT w ⎠⎝ ⎞ ⎛ 1 + sTlead ⎟⎜ ⎟ ⎜ 1 + sT lag ⎠⎝ ⎞ ⎟ ⎟ ⎠ N (8.12) where k d is a positive constant gain, Tm is the time constant of the low pass filter (typically 0.1s) [83] and Tw is the washout time constant (typically 3s-10s) [84]. Tlead and Tlag are the lead and lag time constants for the N phase-compensation blocks. As shown in equation (8.11), the distance of the shift of the eigenvalue λi that is caused by the feedback damping control is proportional to magnitude of the corresponding residue and the gain of the feedback damping control. The direction of this shift of the eigenvalue λi depends on the residue’s phase angle and the phase shift across the feedback control transfer function, H ( s ) . jω Ri Δλi = kd Ri ϕ H (s ) θR λi λinew σ Figure 8.3: The shift of an eigenvalue caused a feedback damping control. An ideal feedback damping control move the selected eigenvalue, λ i , directly into the left stable area (damping ratio>5%), moving parallel with the real axis. In other words, the residue’s phase angle, θR , and the phase shift across transfer function of the feedback control, ϕ H ( s ) , should satisfy the relationship θ R + ϕ H ( s ) = 180 0 . The compensation angle, ϕ H ( s ) , is predominantly determined by the phase compensation blocks. The necessary parameter values for Tlead and Tlag , which will properly define the phase compensation blocks, can be calculated using the following equations: ϕ H(s) = 180o − θR -130- (8.13) Chapter 8 The use of Power Electronic Devices for Damping Inter-area oscillations ϕH(s) ≤ 60o ⎧⎪1 N=⎨ ⎪⎩2 α = Tlead Tlag Tlag = (8.14) ϕH(s) ≤ 120o 1 − sin( = 1 + sin( ϕ H (s) N ϕ H (s) N ) (8.15) ) 1 (8.16) ω α Tlead = αTlag (8.17) where θ R is the phase angle of the residue, R i , and ω (rad/sec) is the oscillation frequency of the oscillatory mode to be modified. 8.3 Inter-area Oscillation Damping Control with HVDC In this section, the process for designing a HVDC supplementary controller for damping inter-area oscillations is presented. For illustrating the application of HVDC damping control, a typical two-area system with an HVDC link, shown in Figure 8.4, was constructed using ‘DIgSILENT PowerFactory’. The details of this test system are given in ‘Appendix C’. Bus 2 G1 Bus 6 Tr1 Bus 3 Line 1 REC Line 3 Line 2 Area 1 Line 8 Cap 3 Bus 1 Line 5 Line 4 Line 6 Load 1 G2 Tr2 Tr3 INV Bus 5 DC link Cap 4 Bus 7 Line 7 Area 2 Load 2 Cap 1 Bus 4 Cap 2 Figure 8.4: A two-area system with HVDC. -131- Tr4 G4 G3 Chapter 8 The use of Power Electronic Devices for Damping Inter-area oscillations 8.3.1 HVDC transmission system modelling Generally, the modelling of HVDC transmission systems for power flow and system stability studies consists of three main parts: 1) Power converter modelling, 2) HVDC transmission network modelling, 3) HVDC control system modelling. In a HVDC transmission system, power electronic converters are required to convert electric power from the AC side to the DC side and then back from the DC side to the AC side. The DC power flow, transferred over the DC transmission lines, is managed by the HVDC control system by adjusting the firing angles of the power converters. The details of HVDC system modelling are presented in the following sections. 8.3.1.1 Power converter modelling A HVDC system uses power electronic converters to convert electric power from the AC system to the DC system and vice versa. There are two types of power converters, Current Source Converters (CSC) and Voltage Source Converters (VSC). The fundamental design of these two converter technologies is shown in Figure 8.5. Figure 8.5: Two types of power converters [85]. Modern HVDC transmission technology either uses CSC or VSC as power converters. The proper selection of converter technology in a HVDC transmission system should be based on a wide range of factors; a number of which are addressed in the detailed comparison of CSC and VSC technologies given in [85]. The capacity of the current VSC technology is currently limited to 250 MW, due to practical limitations of the electronic switches. This limitation means that CSC technology will be more suitable for use in the HVDC links that are expected to be installed between Scotland and central England, as shown in Figure 5.5. This is because the power transferred over the new HVDC links will be approximate 2000 MW [86] during heavy load condition. -132- Chapter 8 The use of Power Electronic Devices for Damping Inter-area oscillations Therefore, in the discussion of modelling presented in this section only the CSC based technology is only considered. A detailed introduction and discussion of CSC converter theory and performance is presented in [34]. We directly move to the introduction of the equivalent circuits of CSC power converters. A power converter has two operation modes, rectifier mode and inverter mode. The operation mode of a converter is determined by the firing angle, α, of the controlled valves. With a firing angle of 0°<α<90° a converter operates in rectifier mode; whereas when the firing angle is 90°<α<180° a converter operates in inverter mode. A rectifier converts electric power from the AC system to the DC system; whereas an inverter converts electric power from the DC system to the AC system. Figure 8.6 shows the equivalent circuit of a rectifier with one six-pulse bridge. Rcr Id E LL Vdor cos α Vdr Figure 8.6: Rectifier equivalent circuit [34]. V dor = 3 2 π TE LL Vdr = Vdor cosα − Rcr I d Rcr = 3 π X cr where: Vdor is the ideal open circuit direct voltage of rectifier ELL is the line-line voltage on converter side T is the transformer ratio α is the ignition delay (firing) angle -133- (8.19) (8.20) (8.21) Chapter 8 The use of Power Electronic Devices for Damping Inter-area oscillations Rcr is the commutation resistance of rectifier Xcr is the commutation reactance of rectifier Vdr is the dc output voltage of rectifier Id is the direct current Figure 8.7 presents the equivalent circuit of an inverter with one six-pulse bridge. The inverter operation may be described in terms of α; defined in the same way as for the rectifier but with an ignition (firing) delay angle of between 90° and 180°. However, the common practice is to describe an inverter using the ignition advance angle β (β=π-α), instead of α. Rci E LL Id Vdi Vdoi cos β Figure 8.7: Inverter equivalent circuit [34]. V doi = 3 2 π TE LL Vdi = Vdoi cos β + Rc I d Rci = 3 π X ci where Vdoi is the ideal open circuit direct voltage of inverter ELL is the line-line voltage on converter side β=π-α is the ignition advance angle of inverter Rci is the commutation resistance of inverter Xci is the commutation reactance of inverter -134- (8.22) (8.23) (8.24) Chapter 8 The use of Power Electronic Devices for Damping Inter-area oscillations Vdi is the dc output voltage of inverter Id is the direct voltage and current 8.3.1.2 HVDC transmission network modelling Generally, there are three types of HVDC links: Monopolar links, Bipolar links and Homopolar links, these are shown in Figures 8.8-8.10. The characteristics of these links are introduced in [34] and [85]. As shown in Figure 8.8, the Monopolar link uses a main power transfer conductor and another metallic conductor as the return path. Such a system has the advantage of low investment costs and offers a reliable return path, particularly for underground/underwater cable systems [34]. As the UK power system is planning to use submarine cable technology to implement the new HVDC links between Scotland and central England, shown in Figure 5.5, this section only studies the modelling of a Monopolar HVDC transmission system. Figure 8.8: Monopolar HVDC link [34]. Figure 8.9: Bipolar HVDC link [34]. -135- Chapter 8 The use of Power Electronic Devices for Damping Inter-area oscillations Figure 8.10: Homopolar HVDC link [34]. Figure 8.8 shows the three main parts of a Monopolar HVDC transmission network: the rectifier at the power sending terminal, the inverter at the power receiving terminal and the DC transmission line. Combining this configuration with the equivalent circuits of power converters allows the equivalent circuit of a Monopolar HVDC link to be formed. This equivalent circuit is shown in Figure 8.11 alongside a typical voltage profile of the equivalent circuit in Figure 8.12. Rcr Vdor cos α RL Rci Id Vdr Vdi Vdoi cos β Figure 8.11: Equivalent circuit of HVDC link [34]. Vdor cos α Vdr Vdi Vdoi cos β Figure 8.12: Voltage profile of the equivalent circuit of HVDC link [34]. -136- Chapter 8 The use of Power Electronic Devices for Damping Inter-area oscillations The direct current flow in the equivalent circuit of a closed HVDC circuit can be described by: Id = V dor cos α − V doi cos β Rci + R L + Rci (8.25) The sending power at the rectifier terminal is: Pdr = Vdr I d (8.26) The receiving power at the inverter terminal is: Pdi = Vdi I d = Pdr − R L I d2 (8.27) 8.3.1.3 HVDC control system modelling From the given equivalent circuit of a HVDC link (see Figure 8.11), it can be seen that the voltages at the two terminals of the link, and the current across the DC line, can be controlled by the converters’ internal voltages ( Vdor cosα ) and ( Vdoi cos β ). Thus, if the open circuit voltages of rectifier and converter are assumed to be constant, by controlling the firing angle of the rectifier and converter the power flow across the DC link can be controlled. The HVDC control scheme developed in this section is a simple form of control. In practical implementation, there are a number of different, and more complex, control schemes for managing DC power flow; a detailed discussion of these are given in [34]and [85]. A simplified HVDC control scheme is given in Figure 8.13. In this scheme the amount of active power to be transferred over the DC link is first determined by the system operators. Then, the direct voltage at the power receiving terminal is held at 1.0 p.u. through ‘β control’ of the inverter (see Figure 8.14). In ‘β control’ of the inverter, the input signal is the error between the voltage reference, Vref , and the real time voltage measurement, Vmeas . This error is processed using a Proportional-Integral (PI) regulator to calculate a new value for the firing angle. -137- Chapter 8 The use of Power Electronic Devices for Damping Inter-area oscillations With the predetermined power transfer and a constant voltage at the inverter terminal, the direct current through the DC link can be directly calculated. Using this calculated value as a current reference, the rectifier’s ‘α control’ acts to hold the direct current at the reference value by adjusting the firing angle α (see Figure 8.15). β control α control I ref Vref = 1.0 p.u. Figure 8.13: Basic control scheme of HVDC system. β max kp 1 1 + sT f Vref Vmeas ki s β β min Figure 8.14: Inverter’s β control for constant voltage [80]. α max kp 1 1 + sT f I ref I meas ki s α min Figure 8.15: Rectifier’s α control for constant current [80]. -138- α Chapter 8 The use of Power Electronic Devices for Damping Inter-area oscillations 8.3.2 System performance without HVDC damping controller At a steady state, modal analysis was used to find the system’s oscillatory modes. The eigenvalues associated with all of the oscillatory modes are shown in Figure 8.16. As seen from Figure 8.16, there are 8 oscillatory modes in this system and they are all well damped except for the inter-area oscillatory mode ( λint er = −0.0128 ± j3.212 , damping ratio = 0.3%). 7 6 local modes imaginary 5 damping ratio=5% 4 3 inter-area mode 2 exciter modes 1 0 -2 governor modes exciter modes -1.5 -1 -0.5 0 0.5 real Figure 8.16: Oscillatory modes in the two-area system with HVDC. A nonlinear simulation was performed to show that the inter-area oscillatory mode under which the generators in the areas swing against each other after disturbances. For exciting the inter-area mode, the mechanical torque of generator 2 was increased by 0.01 p.u. at 0s, the mechanical torque of generator 4 was simultaneously reduced by 0.01 p.u.. Figure 8.17 shows the generator rotor speed responses to this pair of small disturbances. The shape of the inter-area mode can be clearly observed in the responses, i.e. the change of the generator rotor speeds in area 1 are always in anti-phase with the change of the generator rotor speeds in area 2. Figure 8.18 shows the frequency response at two locations in the system; one is measured in area 1 (at bus 3) and the other is measured in area 2 (at bus 5). As seen from Figures 8.17-8.18, the inter-area mode is not damped. -139- Chapter 8 The use of Power Electronic Devices for Damping Inter-area oscillations 1.0003 G1 speed [p.u.] 1.0002 G2 G3 G4 1.0001 1 0.9999 0.9998 0 5 10 time [s] 15 20 Figure 8.17: Generator rotor speed responses to the small disturbances without HVDC damping control. Frequency [Hz] fre-bus3 fre-bus5 50.005 50 49.995 0 5 10 time [s] 15 20 Figure 8.18: System frequency responses to the small disturbances without HVDC damping control. 8.3.3 Residue based HVDC damping controller design Figure 8.19 gives a block diagram of a supplementary damping controller for HVDC. The input of the supplementary HVDC damping controller is the frequency difference between area 1 and area 2; one frequency is measured at bus 3 and the other one is measured at bus 5. The output of this damping controller is used to modify the current reference of the rectifier’s ‘α control’. Since the DC voltage at the inverter terminal is held constant by the inverter’s ‘β control’, modifying the current reference of the rectifier control allows corresponding changes to be made to the active power transfer over the DC line. Actually, the input signal selection for the damping controller is not unique. Other oscillatory signals, such as the power angle and active power across the AC inter-tie lines; and the voltage magnitude at the ends of the AC inter-tie lines, can also be used. The inter-area oscillation is associated with the differences between the rotor speed changes of different groups of generators; these differences can be clearly observed by -140- Chapter 8 The use of Power Electronic Devices for Damping Inter-area oscillations monitoring system frequencies at different locations (see Figure 8.18). Therefore, the frequency difference between the two areas is a logical choice for the input signal of the damping controller. I ref α max kp 1 1 + sT f I meas ki s α α min u max 1 1 + sTm kd 1 + sTlead 1 + sTlag sTw 1 + sTw 1 + sTlead 1 + sTlag umin Figure 8.19: Rectifier control with supplementary damping control. Figure 8.20 represents such closed loop system, in which the feedback control has two input signals, Δy j (s ) and Δyl (s) . Then based on the equation (8.10), the residue of the open loop transfer function between the system input Δuk (s ) and the system outputs Δy j (s ) and Δyl (s) can be represented by equation (8.28). Δuk (s ) + + Δ y j (s ) G(s) H(s) Δyl (s ) + − Figure 8.20: Feedback control with multiple input signals. Ri = c j φi ψi bk − cl φi ψi bk (8.28) where c j and cl is the jth and lth row vector of the C matrix and b k is the kth column vector of the B matrix. φi is the right eigenvector associated with the eigenvalue λi , and ψ i is the left eigenvector associated with the eigenvalue λi . -141- Chapter 8 The use of Power Electronic Devices for Damping Inter-area oscillations As introduced in Section 8.2.2, the phasor angle of the shift of the selected eigenvalue λi caused by the feedback control should be equal to the sum of θR and ϕ H ( s ) (see Fig. 8.3). Therefore, if the phase shift across the transfer function of the feedback damping control, at the frequency associated with λi is zero (no phase compensation block is used in the feedback control), and then the feedback control will move the eigenvalue λi in the same direction as the residue. Hence, the corresponding movement of the eigenvalue λi caused by such feedback control can be described by the following equations: Δλi = Ri H (λi ) = Δσ + jΔω (8.29) Δω ) Δσ (8.30) θ R = ac tan( where Δλi is the shift of the eigenvalue λi caused by the transfer function of the feedback control H (λi ) and the residue Ri . Δσ and Δω represent the change of the real part and imaginary part of Δλi , respectively. This approach for estimating the phase angle of the residue is illustrated in Figure 8.21. 2 λishift 2 Δλi Δω 8 6 λi λ θ R = ac tan Δω Δσ Δσ Figure 8.21: An illustration of the estimation of the phase angle of the residue. In this case, the method shown in Figure 8.21 was used to estimate the phase angle of the residue. First of all, a simple feedback damping controller was integrated into the system. This damping controller only consisted of a low pass filter and a washout filter; no phase compensation blocks are included. Modal analysis was applied for both k d = 0 and k d = 100 (see Figure 8.14); for both of these gain values the locations of the eigenvalues that are associated with the inter-area mode are presented in Figure 8.22. With the shift of the eigenvalue, the phase angle of the residue was estimated to be 68.48 0. -142- Chapter 8 The use of Power Electronic Devices for Damping Inter-area oscillations 3.218 3.217 λint er = −0.01105942 + j3.216411 K d = 100 imaginary 3.216 3.215 3.214 3.213 3.212 3.211 θ shift = 68.480 Kd = 0 λint er = −0.0127485 + j 3.212157 -0.0125 -0.012 -0.0115 -0.011 real Figure 8.22: An estimation of the residue’s phase angle for HVDC damping control design. Therefore, to implement an ideal feed back controller (according to equation (8.14)) phase compensation blocks should be included to introduce a phase shift of 111.52 0 at 0.5 Hz (the frequency of inter-area mode). Using equations (8.15-8.19) suitable parameter values for the tuning of a set of phase compensation blocks can be found; in this case they are Tlead = 0.1s , T lag = 1 .03 s . The transfer function of the modified feedback damping controller is represented by the following equation: ⎛ 1 ⎞ ⎛ 10s ⎞ ⎛ 1 + 0.1s ⎞ ⎛ 1 + 0.1s ⎞ H (s) = K d ⎜ ⎟⎜ ⎟⎜ ⎟⎜ ⎟ ⎝ 1 + 0.1s ⎠ ⎝ 1 + 10s ⎠ ⎝ 1 + 1.03s ⎠ ⎝ 1 + 1.03s ⎠ (8.31) 8.3.4 System performance with a HVDC damping controller Having integrated the HVDC damping controller into the system, the modal analysis was repeated and 9 oscillatory modes were found in the system. Analysis of the participation factors shows that the new oscillatory mode is a HVDC control mode. Figure 8.23 shows the eigenvalues of the system’s oscillatory modes for different controller gains. The inter-area oscillatory mode and HVDC control mode are quite sensitive to the gain of the HVDC feedback damping control; whereas the other modes are relatively insensitive. As the gain of the HVDC damping controller increases the poorly damped inter-area oscillatory mode moved directly to the left; whereas the HVDC control mode moved directly to the right. When the damping controller’s gain increased to 1600, the inter-area mode started to move back to the right; this means that the damping ratio of the inter-area mode starts to decrease. -143- Chapter 8 The use of Power Electronic Devices for Damping Inter-area oscillations 7 local modes 6 damping ration=5% imaginary 5 kd=400 kd=800 kd=1400 HVDC control mode 4 3 inter- area mode kd=1600 2 governor mode 1 0 -4 exciter modes -3.5 -3 -2.5 -2 exciter modes -1.5 -1 -0.5 0 0.5 real Figure 8.23: Oscillatory modes versus the gain of HVDC damping controller. Nonlinear time-domain simulations were used to check the ability of the HVDC damping controller to properly damp inter-area oscillations. At 0 s, the mechanical torque of generator 2 was increased by 0.01 p.u.; simultaneously the mechanical torque of generator 4 was reduced by 0.01 p.u.. Figure 8.24 presents the responses of generator rotor speeds to the pair of disturbances with and without the HVDC damping controller. In addition, Figure 8.25 presents the responses of the inter-area power flow (line 3) to these small disturbances with different damping controller gains. These results show that the larger controller gain is used the larger damping is obtained. 1.0003 1.0003 no control 1.0002 rotor speed [p.u.] rotor speed [p.u.] no control kd=1600 1.0001 1 0.9999 G1 0.9998 0 5 10 15 1.0002 kd=1600 1.0001 1 0.9999 G2 0.9998 20 0 5 time [s] 10 1.0003 20 1.0003 no control 1.0002 no control rotor speed [p.u.] rotor speed [p.u.] 15 time [s] kd=1600 1.0001 1 0.9999 G3 0.9998 0 5 10 15 1.0002 kd=1600 1.0001 1 0.9999 G4 0.9998 20 time [s] 0 5 10 15 20 time [s] Figure 8.24: Generator rotor speed responses to the small disturbances with and without HVDC damping control. -144- Chapter 8 The use of Power Electronic Devices for Damping Inter-area oscillations 160 without control with control kd=400 with control kd=800 with control kd=1600 P [MW] 155 150 145 0 5 10 time [s] 15 20 Figure 8.25: Active power flow (line 3) responses to the small disturbances with and without HVDC damping control. The behaviour of the HVDC damping controller during large disturbances was also tested. At 1 s a permanent three-phase short-circuit fault was simulated at the mid-point of line 6, after 100 ms the faulted line was disconnected. The response of the generator rotor speeds to the three-phase fault with, and without a HVDC damping controller are presented in Figure 8.26. As seen from the simulation results, the HVDC damping controller has shown its robustness in damping inter-area oscillations after the system is subject to a large disturbance. In addition, Figure 8.27 presents the responses of the inter-area power flow (line 3) to the large disturbance with different damping controller gains. As before, when the controller gain is increased the damping is also increased; no control 1.005 rotor speed [p.u.] rotor speed [p.u.] this allows the system to be stabilized more quickly from the large disturbance. kd=1600 1 G1 0.995 0 5 10 15 no control 1.005 kd=1600 1 G2 0.995 20 0 5 time [s] 1.015 15 20 1.015 no control G3 1.01 rotor speed [p.u.] rotor speed [p.u.] 10 time [s] kd=1600 1.005 1 0.995 no control G4 1.01 kd=1600 1.005 1 0.995 0 5 10 15 20 time [s] 0 5 10 15 20 time [s] Figure 8.26: Generator rotor speed responses to a three-phase fault with and without HVDC damping control. -145- Chapter 8 The use of Power Electronic Devices for Damping Inter-area oscillations 400 without control with control kd=400 with control kd=800 300 P [MW] with control kd=1600 200 100 0 -100 0 5 10 time [s] 15 20 Figure 8.27: Active power flow (line 3) responses to a three-phase fault with and without HVDC damping control. 8.4 Inter-area Oscillation Control with TCSC In this section, we will present a process for designing a controller that uses TCSC to damp inter-area oscillations. The system used for illustrating this application of TCSC supplementary damping control is given in Figure 8.28. It is similar to the two-area system used in the previous section (see Figure 8.4). In the mid-point of the additional AC line there is a TCSC that provides 40% compensation of the line reactance. This test system was modelled using ‘DIgSILENT PowerFactory’ and further details are given in Appendix D. Figure 8.28: Typical two-area system with TCSC. 8.4.1 TCSC modelling A Thyristor Controlled Series Capacitor (TCSC) is an impedance compensator. It is added in series to an AC transmission line for the purpose of increasing the power transfer capability of that transmission line, and controlling power flow [87] [88]. A -146- Chapter 8 The use of Power Electronic Devices for Damping Inter-area oscillations typical TCSC consists of a series capacitor bank, C, in parallel with a thyristorcontrolled reactor, L, as shown in Figure 8.29. Figure 8.29: A structure of a typical TCSC. For power flow and power system stability studies, the effect of the thyristor operation can be neglected in the simulations. Therefore, for these studies, a TCSC can be represented using the ideal model shown in Figure 8.30; that consists of a fixed capacitor in parallel with a variable reactor. Figure 8.30: An ideal model of TCSC for power system stability study. Using this ideal model the equivalent circuit of a transmission corridor with TCSC shown in Figure 8.31 is constructed. In the equivalent circuit the resistance of the transmission line is neglected. As the equivalent reactance of “line 1” can now be smoothly changed, by adjusting the reactance of the TCSC, the active power flow across the two transmission lines can be dynamically regulated. xTCSC Pline1 xline1 xline 2 VS Pline 2 VR Figure 8.31: An equivalent circuit of the transmission corridor with TCSC. -147- Chapter 8 The use of Power Electronic Devices for Damping Inter-area oscillations Figure 8.32 gives a block diagram of basic TCSC control. Based on off line power flow study (see Figure 8.31), the reactance of the variable reactor is calculated to obtain an expected power flow condition (see Figure 8.30). In addition, TCSC can also play an important role in improving the damping of low frequency power oscillations. This can be achieved by adding a damping control that provides a damping signal to change the reactance dynamically. xmax xC ⋅ xTCSC xC + xTCSC xTCSC 1 1 + sT f xreactor xmin Figure 8.32: A block diagram of TCSC control. 8.4.2 System performance without TCSC damping controller At a steady state, modal analysis was used to find the system’s oscillatory modes. The eigenvalues associated with all of the oscillatory modes are presented in Figure 8.33. These eigenvalues show that the system is stable, as all of the eigenvalues’ real parts are negative. However, the damping ratio of the inter-area oscillatory mode (λinter-area = 0.0203107±j4.09283) is only 0.4%, which does not satisfy the requirements of practical power system operation. 7 6 local modes imaginary 5 damping ratio=5% 4 inter-area modes 3 2 1 0 -2 governor mode exciter modes -1.5 -1 exciter modes -0.5 0 0.5 real Figure 8.33: Oscillatory modes in the two-area system with TCSC. -148- Chapter 8 The use of Power Electronic Devices for Damping Inter-area oscillations After the modal analysis, a nonlinear simulation was executed in the time domain to show the inter-area oscillatory mode under which the generators in the areas swing against each other after disturbances. For exciting the inter-area mode, the mechanical torque of generator 2 was increased by 0.01 p.u. at 0s, the mechanical torque of generator 4 was reduced simultaneously by 0.01 p.u.. Figure 8.34 shows the generator rotor speed responses to this pair of small disturbances. The shape of the inter-area mode can be clearly observed in the responses, i.e. the change of the generator rotor speeds in area 1 are always in anti-phase with the change of the generator rotor speeds in area 2. Figure 8.35 shows the frequency response at two locations in the system; one is measured in area 1 (at bus 3) and the other is measured in area 2 (at bus 5). 1.0002 speed [p.u.] G1 G2 G3 G4 1.0001 1 0.9999 0.9998 0 5 10 15 20 time [s] Figure 8.34: Generator rotor speed responses to the small disturbances without TCSC damping control. Frequency [Hz] 50.006 fre-bus3 fre-bus5 50.004 50.002 50 49.998 49.996 0 5 10 time [s] 15 20 Figure 8.35: System frequency responses to the small disturbances without TCSC damping control. 8.4.3 Residue based TCSC damping controller design Figure 8.36 gives the block diagram of a TCSC supplementary damping controller. The input of the damping controller is the frequency difference across the AC transmission corridor; one is measured in area 1 (at bus 3) and the other is measured in area 2 (at bus -149- Chapter 8 The use of Power Electronic Devices for Damping Inter-area oscillations 5). The output of this damping controller is used to modify the reactance reference of the TCSC. xmax xC ⋅ xTCSC xC + xTCSC + xTCSC + 1 1 + sT f xreactor xmin u max Freq area 1 + 1 1 + sTm kd - 1 + sTlead 1 + sTlag sTw 1 + sTw Freq area 2 1 + sTlead 1 + sTlag N stages Limit u min Figure 8.36: A block diagram TCSC supplementary damping control. In this case, the method shown in Figure 8.21 is used to estimate the phase angle of the residue of the inputs and the output of the system’s open loop transfer function. First of all, a feedback damping controller is integrated into the system. This damping controller only consists of a low pass filter and a washout filter; there are no phase compensation blocks. Modal analysis was applied for two gain values, k d = 0 and k d = 100 . Figure 8.37 presents the locations of the eigenvalues associated with the inter-area mode obtained from this modal analysis. Using the shift of the eigenvalue, the phase angle of the residue was estimated to -42.417 0. 4.0932 4.093 imaginary 4.0928 4.0926 k d = 0 λint er = −0.02031065 + j 4.09283 θ shift = −42.417 0 4.0924 4.0922 4.092 4.0918 4.0916 λinter = −0.01913188 + j4.09175 k = 100 d -0.0204 -0.0202 -0.02 -0.0198 -0.0196 -0.0194 -0.0192 -0.019 real Figure 8.37: An estimation of the residue’s phase angle for TCSC damping control design. In this case, a negative feedback control is used to produce an ideal feed back controller, based on equation (8.14) the phase compensation blocks of the feedback control should produce a phase shift of 42.417° at 0.65 Hz (the frequency of the inter-area mode). The parameters values that produce the necessary compensation were calculated by using -150- Chapter 8 The use of Power Electronic Devices for Damping Inter-area oscillations equations (8.15-8.19); these values are Tlead = 0.11s , Tlag = 0 .556 s . Hence, we have a transfer function for the negative feedback damping controller: ⎛ 1 ⎞ ⎛ 10 s ⎞ ⎛ 1 + 0.22 s ⎞ ⎟⎟ ⎜⎜ ⎟⎟ ⎜⎜ ⎟⎟ H (s) = Kd ⎜⎜ ⎝ 1 + 0.1s ⎠ ⎝ 1 + 10 s ⎠ ⎝ 1 + 1.38 s ⎠ (8.32) 8.4.4 System performance with TCSC damping controller Modal analysis was again applied to the two-area system with the addition of the new TCSC damping controller; the oscillatory modes are shown in Figure 8.38. There were now 9 oscillatory modes identified in the system. Using participation factor analysis this new oscillatory mode was confirmed to be a TCSC control mode. Figure 8.38 shows the changes that occur in system’s oscillatory modes when the gain of TCSC damping controller is varied. As the gain of the damping controller is increased the poorly damped inter-area oscillatory mode moves to the left; whereas the TCSC control mode moves to the right. When the damping controller’s gain, kd , is increased to approximately 22000 the inter-area mode begin to move back to the right; this means the damping ratio of the inter-are mode has begun to decrease. 7 local modes 6 damping ratio =5% imaginary 5 4 Kd=22000 Kd=18000 Kd=12000 Kd=6000 TCSC control mode 3 2 governor mode exciter modes 1 inter-area mode exciter modes 0 -4 -3.5 -3 -2.5 -2 -1.5 -1 -0.5 0 0.5 real Figure 8.38: Oscillatory modes versus the gain of TCSC damping controller. Nonlinear simulations were used to demonstrate that the ability of the TCSC damping controller to move the inter-area mode eigenvalue corresponds to an ability to damp the inter-area oscillation. At 0 s, the mechanical torque of generator 2 was increased by 0.01 p.u.; simultaneously the mechanical torque of generator 4 was reduced by 0.01 p.u.. Figure 8.39 presents the responses of generator rotor speeds to these small disturbances -151- Chapter 8 The use of Power Electronic Devices for Damping Inter-area oscillations with and without the TCSC damping control. These results confirmed the TCSC damping controller’s ability in damping inter-area oscillations. In addition, Figure 8.40 presents the responses of the inter-area power flow (line 3) to these small disturbances with different damping controller gains. These results show that the larger controller gain was used the larger damping was obtained. 1.0002 no control rotor speed [p.u.] rotor speed [p.u.] 1.0002 kd=22000 1.0001 1 no control kd=22000 1.0001 1 G1 0.9999 0 5 10 15 0.9999 20 G2 0 5 time [s] 15 20 1.0003 no control 1.0002 rotor speed [p.u.] rotor speed [p.u.] 1.0003 kd=22000 1.0001 1 0.9999 0.9998 10 time [s] G3 0 5 10 15 no control 1.0002 kd=22000 1.0001 1 0.9999 G4 20 0.9998 0 5 time [s] 10 15 20 time [s] Figure 8.39: Generator rotor speed responses to the small disturbances with and without TCSC damping control. 112 111 110 P [MW] 109 108 107 without control with control kd=6000 106 with control kd=12000 105 104 with control kd=22000 0 5 10 time [s] 15 20 Figure 8.40: Active power flow (line 3) responses to small disturbance with and without TCSC damping control. Furthermore, the damping effect of the TCSC damping controller was also verified for a large disturbance. At 1 s a permanent three-phase short-circuit fault was simulated at the mid-point of line 6, after 100 ms the faulted line was disconnected. The response of the -152- Chapter 8 The use of Power Electronic Devices for Damping Inter-area oscillations generator rotor speeds to the three-phase fault with, and without a TCSC damping controller are presented in Figure 8.41. As seen from the simulation results, the TCSC damping controller has shown its robustness in damping inter-area oscillations after the system is subject to a large disturbance. In addition, Figure 8.42 presents the responses of the inter-area power flow (line 3) to the large disturbance with different damping controller gains. As before, when the controller gain is increased the damping is also increased; this allows the system to be stabilized more quickly from the large disturbance. 1.003 no control rotor speed [p.u.] rotor speed [p.u.] 1.003 kd=22000 1.002 1.001 1 G1 0.999 0 5 10 15 no control kd=22000 1.002 1.001 1 G2 0.999 20 0 5 no control 1.004 kd=22000 1.002 1 0.998 G3 0 5 10 15 20 time [s] rotor speed [p.u.] rotor speed [p.u.] time [s] 10 15 no control 1.004 kd=22000 1.002 1 0.998 20 G4 0 time [s] 5 10 15 20 time [s] Figure 8.41: Generator rotor speed responses to a three-phase fault with and without TCSC damping control. 110 100 90 P [MW] 80 70 60 without control 50 with control kd=6000 with control kd=12000 40 with control kd=22000 30 20 0 5 10 time [s] 15 20 Figure 8.42: Active power flow (line 3) responses to a three-phase fault with and without TCSC control. -153- Chapter 8 The use of Power Electronic Devices for Damping Inter-area oscillations 8.5 Inter-area Oscillation Control with SVC In this section, a fundamental study of using SVC for damping inter-area oscillations will be presented. The system used for testing the SVC damping control is shown in Figure 8.43. There is a SVC connecting to Bus 5. The SVC acts to hold the voltage magnitude of Bus 5 close to 1.0 p.u.. This test system was modelled using ‘DIgSILENT PowerFactory’ and details of it are given in Appendix E. Figure 8.43: Modified two-area system with SVC. 8.5.1 SVC modelling A Static Var Compensator (SVC) is a shunt-connected static reactive power compensation component that is originally designed to ensure that the bus voltage is held close to a set value. Here, the term ‘static’ is used to indicate that an SVC, unlike synchronous machines, has no rotating components [34] [88]. Figure 8.44 shows the structure of a typical SVC: it consists of a shunt capacitor bank and a thyristor controlled shunt reactor. Figure 8.44: A structure of a typical SVC. -154- Chapter 8 The use of Power Electronic Devices for Damping Inter-area oscillations For power flow, and power system stability studies, a SVC can be modelled using the circuit diagram shown in Figure 8.45. This model represents an ideal SVC; which consists of a fixed capacitor connected in parallel with a variable reactor. By controlling the reactance value of the reactor, the SVC releases or absorbs reactive power dynamically to ensure the voltage stays in an acceptable range. Figure 8.45: An ideal model of SVC. Figure 8.46 gives a block diagram of SVC control. Under steady states, the SVC controller uses the difference in magnitude between the reference voltage and the measured voltage to determine the adjustment of the reactive power (Q) output of the SVC. These adjustments act to hold the local voltage close to the reference voltage; however, SVC can also be used to improve the system’s small signal stability if the parameters of the controller are set properly. Vref VSVC Qmax Qfixed k 1 1 + sT f 1 + sTa 1 + sTb QSVC Qmin Figure 8.46: Block diagram of SVC control [34]. 8.5.2 System performance without SVC damping controller At a steady state, modal analysis was applied to find the system’s oscillatory modes. The eigenvalues associated with all of these oscillatory modes are presented in Figure -155- Chapter 8 The use of Power Electronic Devices for Damping Inter-area oscillations 8.47. As shown in Figure 8.47, the system is unstable as the there is an unstable interarea oscillatory mode (λinter-area = 0.02996345 ± j3.416078) in the system. 7 6 local modes imaginary 5 damping ratio= 5% 4 inter-area mode 3 2 governor mode exciter modes 1 0 -2 exciter modes -1.5 -1 -0.5 0 0.5 real Figure 8.47: Oscillatory modes in the two-area system with SVC. A nonlinear simulation was used to demonstrate that the inter-area mode, identified by the modal analysis, would cause the system to become unstable. At 0s, the mechanical torque of generator 2 was increased by 0.01 p.u., simultaneously the mechanical torque of generator 4 was reduced by 0.01 p.u.. Figure 8.48 shows the generator rotor speed responses to this pair of small disturbances. Figure 8.49 shows the voltage angle difference across the inter-are transmission corridor (line 3 and line 5) during the disturbances. As seen from the simulation results, the inter-area mode was not damped. G1 speed [p.u.] 1.0004 G2 G3 G4 1.0002 1 0.9998 0.9996 0 5 10 time [s] 15 20 Figure 8.48: Generator rotor speed responses to the small disturbances without SVC damping control. -156- phase angle [deg] Chapter 8 The use of Power Electronic Devices for Damping Inter-area oscillations phase angle (between bus3&bus5) 26 25 24 23 0 5 10 time [s] 15 20 Figure 8.49: Oscillatory voltage angle difference between bus3 and bus5 caused by the disturbances. 8.5.3 Residue based SVC damping controller design Figure 8.50 gives a block diagram of a SVC supplementary damping controller. In this case, the voltage angle difference between bus3 and bus5 (see Figure 8.49) was selected to be the input of the damping controller. The output of this damping controller was used to modify the local voltage reference of the primary SVC control. Vref VSVC Qmax Qfixed 1 1 + sT f 1 + sTa 1 + sTb k QSVC Qmin u max kd 1 1 + sTm sTw 1 + sTw 1 + sTlead 1 + sTlag 1 + sTlead 1 + sTlag u min Figure 8.50: A block diagram of a SVC damping controller. For the SVC damping control design, the method presented in Figure 8.20 was used to estimate the phase angle of the residue of the system’s open loop transfer function between the inputs and the output. First of all, a feedback damping controller was integrated into the system; this damping controller only consisted of a low pass filter and a washout filter, no phase compensation blocks were used at this time. Modal analysis was performed twice, once with k d = 0 and again with k d = 0.01 . Figure 8.51 presents the locations of the eigenvalues associated with the inter-area mode obtained from the modal analysis. With the shift of the eigenvalue, the phase angle of the residue was estimated to be -46.095°. -157- Chapter 8 The use of Power Electronic Devices for Damping Inter-area oscillations 3.42 k d = 0 λinter −area = 0.02996345 + j 3.4116078 3.415 θ shift = −46.095 0 imaginary 3.41 3.405 3.4 3.395 kd = 0.01 3.39 λinter −area = 0.05412006 + j 3.39098 3.385 3.38 0.025 0.03 0.035 0.04 0.045 real 0.05 0.055 0.06 Figure 8.51: An estimation of the residue’s phase angle for SVC damping control design. In this case, we used negative feedback control. To implement an ideal feedback controller, according to equation (8.12), the phase compensation blocks of the feedback control should introduce a phase shift of 46.095 0 at 0.54 Hz (the frequency of the interarea mode). The parameters values that deliver such a phase shift were and Tlead = 0.22 s , Tlag = 1.38 s , these were determined using equations (8.13-8.17). Hence, the transfer function of the negative feedback damping controller is: ⎛ 1 ⎞ ⎛ 10 s ⎞ ⎛ 1 + 0.22 s ⎞ ⎟⎟ ⎜⎜ ⎟⎟ ⎜⎜ ⎟⎟ H (s) = K d ⎜⎜ + s + s + s 1 0 . 1 1 10 1 1 . 38 ⎝ ⎠⎝ ⎠⎝ ⎠ (8.33) 8.5.4 System performance with SVC damping controller Modal analysis was again performed for the two-area system, but the new SVC damping controller was used. There were now 9 oscillatory modes in the system. Participation factor analysis confirms that this new oscillatory mode was a SVC control mode. Figure 8.52 shows the changes that occurred in the system’s oscillatory modes when the gain of the SVC damping controller was varied. As seen in Figure 8.52, when the gain of the damping controller is increased the unstable inter-area oscillatory mode moves to the left; whereas the SVC control mode moves to the right. Furthermore, two exciter modes move to the left as well as the interarea mode. However, when the damping controller’s gain, kd , is increased to -158- Chapter 8 The use of Power Electronic Devices for Damping Inter-area oscillations approximately 1.0, the two exciter modes begin to move back to the right, and the damping factor of the SVC control mode becomes less than the damping factor of the inter-area mode. Therefore, to obtain an optimum set of parameters for the existing system controllers the gain of the SVC damping controller was set to kd = 1.0 . 7 6 local modes imaginary 5 4 kd =1.2 kd =1.0 3 inter-area mode kd =0.8 2 0 -4 exciter modes kd =0.8 kd =1.0 kd =0.8 kd =1.0 -3.5 -3 -2.5 kd =0.2 kd =0.5 SVC control mode 1 damping ratio=5% governor mode exciter modes kd=1.0 -2 -1.5 -1 -0.5 0 0.5 real Figure 8.52: Oscillatory modes versus the gain of SVC damping controller. To confirm that the SVC damping controller can prevent the system from becoming unstable, by damping the inter-area oscillations, nonlinear simulations were used. At 0s, the mechanical torque of generator 2 was increased by 0.01 p.u.; simultaneously the mechanical torque of generator 4 was reduced by 0.01 p.u.. Figure 8.53 presents the responses of generator rotor speeds to these small disturbances with and without the SVC damping controller. These simulation results confirmed the SVC supplementary controller’s ability in damping inter-area oscillations. In addition, Figure 8.54 presents the responses of the inter-area power flow (line 3) to these small disturbances with different damping controller gains. These results show that the larger controller gain was used the larger damping was obtained. -159- Chapter 8 The use of Power Electronic Devices for Damping Inter-area oscillations 1.0003 no control 1.0003 G1 rotor speed [p.u.] rotor speed [p.u.] 1.0004 kd=1 1.0002 1.0001 1 0.9999 0 5 10 15 no control 1.0002 kd=1 1.0001 1 0.9999 20 G2 0 5 no control 1.0005 G3 kd=1 1 0.9995 0 5 10 10 15 20 time [s] rotor speed [p.u.] rotor speed [p.u.] time [s] 15 20 no control 1.0005 G4 kd=1 1 0.9995 0 5 time [s] 10 15 20 time [s] Figure 8.53: Generator rotor speed responses to the small disturbances with and without SVC damping control. 215 P [MW] 210 205 200 without control 195 with control kd=0.2 with control kd=0.5 with control kd=1.0 190 0 5 10 time [s] 15 20 Figure 8.54: Active power flow (line 3) responses to small disturbance with and without SVC damping control. The robustness of the SVC damping controller to large disturbances was also tested. At 1 s a permanent three-phase short-circuit fault was simulated at the mid-point of line 6, after 100 ms the faulted line was disconnected. The response of the generator rotor speeds to the three-phase fault with and without a SVC damping controller are presented in Figure 8.55. As seen from the simulation results, the SVC damping controller has shown its robustness in damping inter-area oscillations after the system is subject to a large disturbance. In addition, Figure 8.56 presents the responses of the -160- Chapter 8 The use of Power Electronic Devices for Damping Inter-area oscillations inter-area power flow (line 3) to the large disturbance with different damping controller gains. As before, when the controller gain was increased the damping was also increased; this allowed the system to be stabilized more quickly from the large no control G1 1.005 rotor speed [p.u.] rotor speed [p.u.] disturbance. kd=1 1 0.995 0 5 10 15 kd=1 1 0.995 20 no control G2 1.005 0 5 time [s] 15 20 time [s] 1.02 1.02 no control G3 rotor speed [p.u.] rotor speed [p.u.] 10 kd=1 1.01 1 0.99 no control G4 kd=1 1.01 1 0.99 0 5 10 15 20 0 5 time [s] 10 15 20 time [s] Figure 8.55: Generator rotor speed responses to a three-phase fault with and without SVC damping control. 350 300 250 P [MW] 200 150 100 without control 50 with control kd=0.2 with control kd=0.5 0 -50 with control kd=1.0 0 5 10 time [s] 15 20 Figure 8.56: Active power flow (line 3) responses to a three-phase fault with and without SVC damping control. -161- Chapter 8 The use of Power Electronic Devices for Damping Inter-area oscillations 8.6 Conclusions In this Chapter, a process for designing inter-area oscillation damping controller using power electronic devices is presented. The results of the damping analyses performed here, in both the frequency and time domain, demonstrate that the proper control of HVDC, TCSC and SVC installed in the inter-area power transmission corridor can significantly improve the damping of the inter-area oscillation mode. The conventional modal analysis carried out in the frequency domain can be used to find the optimum set of parameters for the system controllers, i.e. the operation point with maximum damping. Nonlinear simulations can be used to assist the conventional eigenvalue analysis. These simulations confirm the ability of the damping controllers, designed in the frequency domain, to provide additional damping to the system robustly. The selection of the input signals that feed the damping controllers is not unique; it is dependent on the observability of the inter-area oscillation mode extracted from the selected input signals. In this chapter, the frequency difference and voltage angle difference across the inter-area power transmission corridor were tested as the input signals for the damping controllers. In practical implementation, these wide area signals can be successfully captured by PMUs and sent to the centralized control centre via high speed communication network. The robustness of the damping controllers was only tested through a single large disturbance at one operating condition. However, the damping controllers designed based on the conventional linear control theory often only work within a limited operating range. For real power system operation, variable operating conditions and different contingencies may reduce the effect of the existing damping controllers, and allow lightly damped or even unstable inter-area oscillations to exist. Therefore, for real implementations, a large number of simulations should be performed to confirm the robustness of the damping controllers at variable operating conditions and for different contingencies. -162- Chapter 9 Wide Area Monitoring and Control System (WAMCS) in the Future GB Power System Chapter 9 Wide Area Monitoring and Control System (WAMCS) in a Future GB Power System 9.1 Introduction In this Chapter, a proposed Wide Area Monitoring and Control System (WAMCS) for a future GB power system is presented. This WAMCS is designed to enhance the system’s small signal stability, i.e. monitor and improve the damping of the inter-area oscillatory mode between Scotland and England. The motivation to establish such a system was introduced in Chapter 5 (Strategies of GB WAMPAC project). This WAMCS combines the applications of Wide Area Monitoring (WAM) and Wide Area Control (WAC). It consists of two real time applications, 1. real time inter-area oscillation monitoring using Newton Type Algorithm (NTA), 2. real time inter-area oscillation damping control using HVDC. These two applications were introduced and tested using a typical two-area system in Chapters 7 and 8. In this Chapter, the operation of the proposed WAMCS will be demonstrated using a full model of the future GB power system (vision 2015). This model was created in the DIgSILENT PowerFactory software package. In addition to the demonstration of the major applications, some key factors that will influence the operation of the wide area inter-area oscillation damping control scheme will be tested and discussed in this Chapter. These factors include the time delay involved in wide area data transmission, and the reactions between the additional HVDC damping controller and conventional Power System Stabilizers (PSSs). 9.2 Assessment of the Inter-area Oscillations in GB Power System 9.2.1 GB power system modelling A GB power system model (vision 2015) used for the GB WAMPAC study was constructed in DIgSILENT PowerFactory. This system consists of three areas, the Scottish Power Transmission Network (SPTN), the Scottish Hydro-Electric -163- Chapter 9 Wide Area Monitoring and Control System (WAMCS) in the Future GB Power System Transmission Network (SHETN) and the National Grid Electricity Transmission Network (NGETN). There are 201 synchronous generators in the system. Of the 201 synchronous generators, 196 are equipped with AVRs, 121 are equipped with Governors and 98 are equipped with PSSs. 20 wind farms consisting of fixed speed wind turbines are modelled as groups of induction machines, and the 170 wind farms consisting of doubly feed wind turbines are modelled as static generators with AVR. In addition, 38 substations in NGETN have dynamically-controlled SVCs intended for the purpose of maintaining the bus voltages at an acceptable level. All of the 627 loads in the system model are modelled as constant impedance loads for all operation conditions. Figure 9.1: 500 kV HVDC links and 400 kV Series Compensators that are installed in the GB power system (vision 2015). As part of vision 2015, two Monopolar 500 kV CSC-HVDC links are installed between the 400 kV Substations at Hunterston and Deeside, to enhance the power transfer capability of the transmission corridors between Scotland and England. In addition, eight Series Compensators (SCs) are installed into the 400 kV AC transmission lines connecting SHETL and NGETL. At heavy load condition, these two HVDC links -164- Chapter 9 Wide Area Monitoring and Control System (WAMCS) in the Future GB Power System transfer 1900 MW of active power from Southern Scotland to Northern Wales, whilst 2400 MW of active power is transferred over the AC transmission lines. Figure 9.1 presents the layouts of the 500 kV HVDC links and 400 kV SCs. The two CSC-HVDC links use the control system introduced in Chapter 8, i.e., constant current control for the rectifier (power sending terminal) and constant voltage control for the inverter (power receiving terminal). The block diagrams and parameters of the HVDC control system are included in Appendix F. 9.2.2 Inter-area oscillations study in GB power system In order to excite the inter-area oscillatory mode in the GB power system, a large disturbance was initiated in the power transmission corridor between SPTN and NGETN. This disturbance was a permanent three-phase short-circuit fault simulated at the mid-point of one of the transmission lines connecting the substations at Torness and Eccles. This fault was simulated at 1s and after 100 ms the faulted line was disconnected. The inter-area power flow on one of the transmission lines connecting the Harker and Hutton substations was used for the inter-area oscillation damping assessment. Figure 9.2 gives the locations of the faulted transmission line and the line selected for monitoring the inter-area oscillations. Figure 9.2: Inter-area oscillation monitoring SPTN and NGETN. -165- Chapter 9 Wide Area Monitoring and Control System (WAMCS) in the Future GB Power System To observe the inter-area oscillations associated with the generators in Scotland swinging against the generators in England the system frequencies across the GB power transmission network were monitored. These system frequencies were measured at the terminals of some of the large generators (>500 MVA) that are operating at different locations in GB. The locations of the large generators selected for the monitoring of inter-area oscillations are presented in Figures 9.3-9.5. Figure 9.3: Three large generators selected in Scotland for monitoring inter-area oscillations. Figure 9.4: Three large generators selected in central England for monitoring inter-area oscillations. -166- Chapter 9 Wide Area Monitoring and Control System (WAMCS) in the Future GB Power System Figure 9.5: Two large generators selected in the South of England for monitoring inter-area oscillations. Figure 9.6 presents the oscillatory inter-area power flow (over the Harker-Hutton line) caused by the large disturbance. As seen from the response of the inter-area power flow to the disturbance, the inter-area oscillatory mode is well damped in the GB power system. Figure 9.7 shows the system frequency responses to the large disturbance. The frequencies measured in Scotland (PEHE, LOAN and HUER) were closely coupled and swinging around the frequencies measured in England. 1300 line Harker-Hutton 1200 1100 P [WM] 1000 900 800 700 600 500 0 2 4 6 time [s] 8 10 12 Figure 9.6: The oscillatory inter-area power flow on the Harker-Hutton line after a large disturbance. -167- Chapter 9 Wide Area Monitoring and Control System (WAMCS) in the Future GB Power System PEHE system frequencies [Hz] 50.3 LOAN HUER 50.2 EGGB WBUR RUGE SIZE 50.1 DUNG 50 49.9 0 1 2 3 4 time [s] 5 6 7 8 Figure 9.7: The system frequency variations caused by a large disturbance. To create more oscillatory behaviour for observing the inter-area oscillations between Scotland and England, 68 conventional PSSs were removed from service. Appendix F lists the PSSs that are in the service. The same disturbance (a three-phase short-circuit fault on the Torness-Eccles line occurring at 1 s and cleared after 100ms) was simulated to provoke a lightly damped inter-area oscillation. Figure 9.8 presents the oscillatory inter-area power flow transferred over the Harker-Hutton line. Comparing this response to that of the original system, it was seen that the damping of the inter-area oscillation mode was significantly reduced. P (line Harker-Hutton) [WM] 1300 original system PSSs-reduced system 1200 1100 1000 900 800 700 600 500 0 5 10 time [s] 15 20 Figure 9.8: The oscillatory inter-area power flow caused by a large disturbance in the original and PSSs-reduced GB system. -168- Chapter 9 Wide Area Monitoring and Control System (WAMCS) in the Future GB Power System Figure 9.9 presents the system frequency variations caused by the large disturbance in the PSSs-reduced system. As seen from the simulation results, the inter-area mode dominated the system frequency variations, i.e. the system frequency variations in Scotland were nearly in anti-phase with the frequency variations in England. The interarea oscillation associated with the generators in Scotland swinging against the generators in England was clearly observed, even though the inter-area oscillation was also influenced by local oscillation modes at the beginning of the dynamic period. Figure 9.10 presents the system frequency variations measured from the substations Hunterston (HUNT) and Deeside (DEES). The inter-area oscillation mode was also clearly observed i.e., the frequency variation measured in the HUER substation was more or less anti-phase to the frequency variation measured in the DEES substation. PEHE LOAN HUER system frequencies [Hz] 50.3 EGGB 50.2 WBUR RUGE SIZE 50.1 DUNG 50 49.9 49.8 0 5 10 time [s] 15 20 Figure 9.9: System frequency variations caused by the large disturbance in the PSSs-reduced system. HUER DEES frequency [Hz] 50.2 50.1 50 49.9 49.8 0 5 10 time [s] 15 20 Figure 9.10: System frequency variations measured in substations HUER and DEES in the PSSs-reduced system. -169- Chapter 9 Wide Area Monitoring and Control System (WAMCS) in the Future GB Power System As seen from these simulation results (Figures 9.8-9.10), after a number of PSSs were removed from service, the damping of the inter-area oscillatory mode was significantly reduced. This operational condition will probably emerge in the future as the system inertia will be significant reduced due to the integration of a large mount of offshore wind farms [57]. Because offshore wind farms will connect to the power grid using back-to-back Voltage Source Converters [63], they will not provide inertia to the system. Coupled with the increasing replacement of conventional generators with the offshore wind farms the inertia of the future UK power system will be largely reduced. A number of power electronic devices such as HVDC and TCSC will be installed to strengthen the Scotland-England interconnection (see Figure 9.1), there is an essential need to establish a wide-area inter-area oscillation control system to extract maximum benefit from these power electronic devices and improve the system damping. 9.3 Wide Area Monitoring and Control System (WAMCS) in a Future GB Power System 9.3.1 Architecture of WAMCS in GB power system Figure 9.11 presents architecture of a proposed WAMCS in a future GB power system. The GB WAMCS has two major applications: 1. real time inter-area oscillation monitoring using the Newton Type Algorithm (NTA), 2. real time inter-area oscillation damping control with HVDC. The WAMCS consists of three parts: ⑴ a synchronized data acquisition system, ⑵ a real time monitoring and control centre and ⑶ a real time control execution system [89]. ⑴ Synchronized data acquisition system There are twelve PMUs in the synchronized data acquisition system. Two PMUs are installed in the 400 kV substations Harker (HARK) and Hutton (HUTT). The inter-area active power flow over one of the transmission lines connecting the two substations can be captured by the two PMUs. The NTA is used to process the active power flow for the real time inter-area oscillation damping assessment. Eight power stations (See Figures 9.3-9.5) are equipped with PMUs to measure system frequencies. The mode shape of the inter-area oscillation mode can be obtained by processing the system frequencies with NTA. -170- Chapter 9 Wide Area Monitoring and Control System (WAMCS) in the Future GB Power System Another two PMUs are installed in the substations/HVDC converter stations, Hunterston (HUER) and Deeside (DEES) to monitor the DC power flow over the HVDC link. Furthermore, the frequency difference between the two substations is used as the input signals for the wide-area HVDC damping control system. The synchronized data is directly transmitted to the monitoring and control centre through a dedicated fibre optic communication network. To further reduce the effect of the time delay involved in the data transmission in WAMCS, it is recommended that a Data Concentrator (DC) is not applied. ⑵ Real time monitoring and control centre The monitoring and control centre is the core of the WAMCS. It receives synchronized data streams and processes the data to serve real time applications. For real time interarea oscillation damping assessment, the wide-area oscillatory signals are collected in WAMCS centre. This real time data is then processed by the NTA-based small signal stability assessment software. The software alerts system operators by issuing an alarm when the oscillatory damping is unsatisfactory. Then system operators will not allow the inter-area transmission lines to carry more active power flow. For the real time interarea oscillation damping control scheme, the WAMCS centre uses the real time interarea oscillatory signals to calculate the parameters of the firing angle control of the HVDC system, then the DC power flow is modulated to stabilise the inter-area oscillations. ⑶ Real time control execution system The real time control execution system consists of several execution units installed at the AC/DC and DC/AC converter stations of the HVDC link. An execution unit is a GPS-synchronized device, which receives commands regarding the adjustments to the HVDC converter firing angles from WAMCS centre and transmits these commands to the power converters. In addition, the start/stop of the whole damping control system is controlled by the execution units. -171- Chapter 9 Wide Area Monitoring and Control System (WAMCS) in the Future GB Power System Figure 9.11: Architecture of the wide-area inter-area oscillation monitoring and control system in GB. 9.3.2 Inter-area oscillation monitoring using NTA For the assessment of the small signal stability, the lightly damped oscillatory active power flow over the AC transmission line HARK-HUTT (see Figure 9.8) was processed using NTA. This oscillatory signal (from 3 s to 20 s) was processed with a sampling frequency f s = 100Hz and a data window size Tdw = 5s . Figure 9.12 presents the dominant inter-area oscillatory information estimated by the NTA. The frequency of the inter-area mode is approximately 0.6 Hz, and the damping ratio is approximately 5 %. To show the inter-area oscillation mode shape, the system frequency variations caused by the disturbance (see Figure 9.9) were simultaneously processed using NTA. These oscillatory signals were processed with a sampling frequency f s = 100Hz and a data window size Tdw = 3s . To obtain a clear mode shape of the inter-area mode, and remove the influence of the other oscillatory modes, only the oscillatory signals in the range of -172- Chapter 9 Wide Area Monitoring and Control System (WAMCS) in the Future GB Power System 6 s-20 s were processed. In Figure 9.13, the magnitudes and phase angles of the dominant oscillatory component of each oscillatory signal are presented in a polar diagram. Here, the estimation results were updated every one second. The polar diagram provides a very clear view of the shape of the inter-area oscillatory mode in the GB system. The phase angles of the change of system frequencies (measured at PEHE, LOAN and HUER) in Scotland are nearly in anti-phase with the change of system 0.65 0.6 0.55 Damping ratio (%) Frequency (Hz) frequencies in England. 6 8 10 12 14 16 18 20 14 16 18 20 t [s] 10 5 0 6 8 10 12 t [s] Figure 9.12: Inter-area oscillation mode identified by NTA in GB system. 90 0.4 120 60 150 30 0.3 0.2 0.1 180 0 330 210 PEHE LOAN HUER EGGB WBUR RUGE SIZE DUNG 300 240 270 Figure 9.13: Estimated inter-area oscillation mode shape of GB system. -173- Chapter 9 Wide Area Monitoring and Control System (WAMCS) in the Future GB Power System 9.3.3 Wide area inter-area oscillation damping control with HVDC in GB The damping ratio of the inter-area mode was reduced to approximately 5% after a number of PSSs were removed from service (see Figure 9.12). With this level of damping it is essential to develop an additional control system that uses the two HVDC links to improve the damping of the inter-area mode. The fundamentals of HVDC supplementary control design were introduced in Chapter 8. Figure 9.14 gives a schematic diagram of the HVDC damping control system. The frequency oscillation between the HUER and DEES substations allowed clear observation of the inter-area mode (see Figure 9.10). Therefore, the difference between the two frequencies was used as the input signals for the HVDC damping control system. The signals, Δ I s, generated by the damping controllers were added to the current references of the rectifier controllers to dynamically modulate the DC power. Expression (9.1) gives a transfer function of the HVDC damping controllers. In this case, the parameters of the HVDC supplementary damping controllers were determined by trial and error method. I max kd 1 1 + sTm sTw 1 + sTw 1 + sTlead 1 + sTlag 1 + sTlead 1 + sTlag I meas I max f DEES kd 1 1 + sTm sTw 1 + sTw 1 + sTlead 1 + sTlag kp ΔI I min f HUNT I ref 1 + sTlead 1 + sTlag I ref ki s kp ΔI I min I meas ki s α max 1 1 + sT f α min α max 1 1 + sT f α min Figure 9.14: Schematic diagram of HVDC damping system. ⎛ 1 ⎞ ⎛ 10s ⎞ ⎛ 1 + 0.05s ⎞ ⎛ 1 + 0.05s ⎞ H ( s) = 100⎜ ⎟⎜ ⎟⎜ ⎟⎜ ⎟ ⎝ 1 + 0.1s ⎠ ⎝ 1 + 10s ⎠ ⎝ 1 + 1.8s ⎠ ⎝ 1 + 1.8s ⎠ -174- (9.1) Chapter 9 Wide Area Monitoring and Control System (WAMCS) in the Future GB Power System To check the abilities of the HVDC damping controllers to improve the damping of the inter-area mode, a same disturbance was used as before. At 1 s, a permanent three-phase short-circuit fault was simulated at the mid-point of one of the transmission lines connecting substations Torness and Eccles; after 100 ms the faulted line tripped. Figures 9.15-9.16 present the responses of the generator rotor speeds to the large disturbance with and without the additional HVDC damping control system. Figure 9.17 shows the response of the inter-area active power flow over the AC transmission line Harker–Hutton to the large disturbance. As confirmed by these simulation results, the HVDC damping controllers allow the system to be stabilized much more quickly after the inception of the large disturbance. 1.005 no control rotor speed [p.u.] rotor speed [p.u.] 1.005 with control 1 PEHE 0.995 0 5 10 15 with control 1 LOAN 0.995 20 no control 0 5 time [s] 10 15 20 time [s] 1.002 no control rotor speed [p.u.] rotor speed [p.u.] 1.005 with control 1 HUER no control with control 1.001 1 0.999 EGGB 0.995 0 5 10 15 20 time [s] 0 5 10 15 20 time [s] Figure 9.15: Responses of the generator rotor speeds (PEHE, LOAN, HUER and EGGB) to the large disturbance with and without the wide area HVDC damping control. -175- Chapter 9 Wide Area Monitoring and Control System (WAMCS) in the Future GB Power System 1.002 rotor speed [p.u.] rotor speed [p.u.] no control with control 1.001 1 WBUR 0.999 0 5 10 15 with control 1.001 1 RUGE 0.999 20 no control 1.002 0 5 time [s] 10 15 20 time [s] 1.002 no control no control rotor speed [p.u.] rotor speed [p.u.] 1.002 with control 1.001 1 0.999 0.998 SIZE 0 5 10 15 with control 1.001 1 0.999 20 DUNG 0 5 time [s] 10 15 20 time [s] Figure 9.16: Responses of the generator rotor speeds (WBUR, RUGE, SIZE and DUNG) to the large disturbance with and without the wide area HVDC damping control. P (line Harker-Hutton) [WM] 1600 without HVDC damping control with HVDC damping control 1400 1200 1000 800 600 400 200 0 5 10 time [s] 15 20 Figure 9.17: The response of the inter-area power flow on the Harker-Hutton line to the large disturbance, with and without HVDC damping control. The oscillatory power flow (from 3 s to 10 s) over the Harker-Hutton line was processed using NTA with the sampling frequency f s = 100Hz and the data window size Tdw = 3s . The estimated inter-area oscillation mode, shown in Figure 9.18 has confirmed that the -176- Chapter 9 Wide Area Monitoring and Control System (WAMCS) in the Future GB Power System damping of the inter-area oscillatory mode has been significantly improved by integrating the new HVDC damping control system. Frequency (Hz) 0.65 0.6 0.55 Damping ratio (%) 0.5 without HVDC damping control with HVDC damping control 6 6.5 7 7.5 8 time [s] 8.5 9 9.5 10 8.5 9 9.5 10 40 without HVDC damping control with HVDC damping control 30 20 10 0 6 6.5 7 7.5 8 time [s] Figure 9.18: Inter-area oscillation mode identified by NTA in the GB system, with and without HVDC damping control. 9.3.4 The impact of time delays on wide-area control Wide area closed loop control systems use remote signals as feedback input signals. The dependence of the control system on wide area signals makes the time delay involved in their transmission of significant concern. The time delay of the data transmission can vary from tens to hundreds of milliseconds. It depends on the communication distance, protocols and time consumed by numerical calculations [90]. In this Section, the effect of time delay in the wide area damping control system is tested. The delays involved in the transmission of a wide area signal can be defined as follows, with reference to Figure 9.19. A packet of measurements (real time oscillatory signal) produced by a PMU is tagged with a time stamp defined as ‘T1’, and then the measurements are transmitted to the WAMCS centre. With this real time oscillatory signal, the WAMCS centre calculates the corresponding commands for the HVDC converter control and sends these commands to the remote HVDC rectifier station. The time when the execution units in the HVDC converter stations receive these commands is defined as ‘T2’. The difference between ‘T2’ and ‘T1’ is defined as the time delay of the data transmission in the wide area control system. -177- Chapter 9 Wide Area Monitoring and Control System (WAMCS) in the Future GB Power System Figure 9.19: An illustration of the time delay involved in the data transmission in GB WAMCS. Figure 9.20 gives a block diagram of HVDC damping controllers, in which the time delay of the transmission of the input signals is taken into account. The element e − sT represents the effect of time delay; T represents the time delay in seconds [91]. In these tests, the time delays in different signal transmission channels are assumed to be constant. I max f HUER kd e − sT 1 1 + sTl sTw 1 + sTw 1 + sTlead 1 + sTlag 1 + sTlead 1 + sTlag ΔI I min I max e − sT kd f DEES 1 1 + sTl sTw 1 + sTw 1 + sTlead 1 + sTlag 1 + sTlead 1 + sTlag ΔI I min Figure 9.20: A block diagram of HVDC damping controllers (with time delay). Figure 9.21 presents the effects of different time delays in the wide area control system, ranging from 50 milliseconds to 200 milliseconds. The effect of wide area HVDC control was gradually reduced when the delay time increased. In addition, when the time delay increased to 150 ms, a high frequency oscillatory component appeared [91]. -178- Chapter 9 Wide Area Monitoring and Control System (WAMCS) in the Future GB Power System P (line Harker-Hutton) [WM] 800 790 780 770 760 no latency latency=50ms latency=100ms 750 latency=150ms 740 730 latency=200ms 10 12 14 16 18 20 time [s] Figure 9.21: The effect of time delay in the wide area damping controllers (50 milliseconds to 200 milliseconds). Figure 9.22 presents the effect of the time delay in wide area control system, from 300 milliseconds to 700 milliseconds. As seen from the simulations, when the time delay increased to 300 ms, the high frequency oscillatory component disappeared [91]; and the inter-area oscillation became unstable as the time delay approached 700 ms. 1300 latency=300ms P (line Harker-Hutton) [WM] 1200 latency=400ms latency=500ms 1100 latency=600ms 1000 latency=700ms 900 800 700 600 500 0 5 10 time [s] 15 20 Figure 9.22: The effect of time delay in the wide area damping controllers (300 milliseconds to 700 milliseconds). -179- Chapter 9 Wide Area Monitoring and Control System (WAMCS) in the Future GB Power System 9.3.5 Interaction between PSSs and HVDC damping control system As seen from the simulation results (Figures 9.15-9.17), the HVDC damping control system has shown its powerful capability to stabilize inter-area oscillations when a number of PSSs are removed from the GB power system. This capability exists because the additional damping control system has very strong influence on the dynamic behaviour of the system, this strong influence could also have unforeseen negative effects. To ensure that the new HVDC damping control system does not have negative effects on the GB power system, the interaction between the PSSs that were removed form service and the new HVDC damping controllers was investigated. With all of the PSSs in service the same simulations was used, with and without the HVDC damping control. In these tests, the effect of time delay was not taken into account. Figures 9.23-9.24 present the responses of the generator rotor speeds to the large disturbance. 1.005 no control rotor speed [p.u.] rotor speed [p.u.] 1.005 with control 1 PEHE 0.995 0 5 10 with control 1 LOAN 0.995 15 no control 0 5 no control 1.005 with control 1 0.995 HUER 0 5 15 time [s] rotor speed [p.u.] rotor speed [p.u.] time [s] 10 10 15 time [s] no control 1.001 with control 1 0.999 EGGB 0 5 10 15 time [s] Figure 9.23: Responses of the generator rotor speeds (PEHE, LOAN, HUER and EGGB) to the large disturbance (all PSSs in service). -180- Chapter 9 Wide Area Monitoring and Control System (WAMCS) in the Future GB Power System no control rotor speed [p.u.] rotor speed [p.u.] 1.0015 with control 1.001 1.0005 1 WBUR 0.9995 0 5 10 no control 1.002 with control 1.001 1 0.999 15 RUGE 0 5 time [s] 10 15 time [s] no control rotor speed [p.u.] rotor speed [p.u.] 1.0015 with control 1.001 1.0005 1 0.9995 SIZE 0 5 10 15 time [s] no control 1.0015 with control 1.001 1.0005 1 0.9995 DUNG 0 5 10 15 time [s] Figure 9.24: Responses of the generator rotor speeds (WBUR, RUGE, SIZE and DUNG) to the large disturbance (all PSSs in service). As seen from the simulations results, the new wide area damping control system has changed the dynamics of the original system (all PSSs in service). However, the responses of the system are still well damped. The new HVDC damping control system slightly improves the damping of the generators in Scotland (PEHE, LOAN and HUER). The new HVDC damping control system did not introduce negative effects to other the generators. The oscillatory inter-area active power flows (over the Harker-Hutton line) caused by the disturbance in the scenario with, and without HVDC damping control are shown in Figure 9.25. The HVDC additional damping control has changed the response of the inter-area power flow during disturbances. A larger deviation of the inter-area power flow was introduced by the HVDC damping control system; the system with HVDC damping control took longer time to stabilize at a new steady state after the large disturbances. -181- Chapter 9 Wide Area Monitoring and Control System (WAMCS) in the Future GB Power System P (line Harker-Hutton) [WM] 1300 without HVDC damping control with HVDC damping control 1200 1100 1000 900 800 700 600 500 0 2 4 6 time [s] 8 10 12 Figure 9.25: The influence of HVDC damping control on the inter-area power flow in complete GB system. 9.4 Conclusions In this Chapter, a proposed Wide Area Monitoring and Control System (WAMCS), designed for enhancing the small signal stability of a future GB power system, has been presented. A permanent three-phase fault was introduced to one of the inter-tie lines between Scotland and England, to investigate the inter-area oscillations associated with the generators in Scotland swinging against the generators in England. As seen from the system response (inter-area active power flow) to the disturbances, the inter-area mode was well damped in the original GB power system. However, after a number of PSSs were removed from service, the damping of the inter-area mode was significantly reduced. To improve the small signal stability of the PSSs-reduced GB power system, two real time SMT applications were introduced and tested through dynamic simulation with a full GB system model. The NTA has successfully extracted the inter-area oscillatory mode from the lightly damped oscillatory power flow transferred between Scotland and England. The inter-area mode shape has been obtained through NTA processing of the system frequencies across the GB power system. With this application, system operators will be more confident to utilize the inter-area transmission lines, and to ensure the system stay at secure level when they increase the inter-area power flow. The wide area -182- Chapter 9 Wide Area Monitoring and Control System (WAMCS) in the Future GB Power System HVDC damping control system has shown that by dynamically modulating the DC power flow it can be powerful tool for improving the damping of the inter-area oscillatory mode. With this application, the power transfer capacity of the inter-area transmission lines can be increased. The influence of the time delay on the wide area control system has been discussed. The effect of the wide area HVDC control system was gradually reduced when the time delay increased. When the time delay increased to a certain range (approximately 150 ms-200 ms), a high frequency oscillatory component was introduced into the interarea oscillations. Finally, the interactions between the HVDC wide area damping control system and the conventional PSSs were investigated. The new wide area damping control system has changed the dynamics of the original system. However, the responses of the system to a large disturbance were still well damped. The damping of some of the generators has been further improved by the wide area damping control system; and it did not introduce obvious negative effects to other the generators. -183- Chapter 10 Conclusions and future work Chapter 10 Conclusions and Future Work 10.1 Conclusions This thesis describes key work needed to overcome the challenges involved in the development of the future GB WAMPAC system. This includes the provision of a thorough study of Synchronized Measurement Technology (SMT), the development of a roadmap to guide the GB WAMPAC project, and the proposal and evaluation of SMT applications and algorithms in the DIgSILENT simulator. SMT is the most essential element and enabler of WAMPAC. To develop a roadmap for the practical implementation of WAMPAC in the GB network, the state of the art and worldwide experience with SMT applications are provided. This thorough study of SMT helps power system operators become more confident that the deployment of SMT will significantly improve the stability of future power systems, and that a SMT-based WAMPAC system will be the only choice for monitoring and managing the future power systems. A proposed architecture of the future GB WAMPAC system, which is presented in this thesis, was based on international experience with existing WAMPAC systems and includes the core building blocks of a WAMPAC system (PMUs, DCs and communication technologies). The future GB WAMPAC system will predominantly serve three power utilities, SPTN, SHETN and NGETN. Therefore, each power utility has a single DC to serve regional applications, and one Super-DC that collect synchronized data from the three regional DCs and support wide area applications. WAMPAC systems involve significant financial investment and long project lead times. Therefore, to ensure the efficient development of the GB WAMPAC system, a roadmap that guides the GB WAMPAC project is required. In this thesis, a methodology of designing a roadmap to guide the GB WAMPAC project is introduced. This methodology takes into account the international experience with WAMPAC project management and the practical challenges faced in the future GB electric power network. Based on this methodology, the roadmap to the future GB WAMPAC system has been devised. -184- Chapter 10 Conclusions and future work The roadmap is divided into a short term and a long term strategy. In the short term, a few (5-6) PMUs will be deployed in the GB power system. This is due to the uncertainties related to the project and the financial limitations. These PMUs will be distributed across the entire UK transmission network to form a wide area power angle and frequency monitoring system, or the transmission corridor between Scotland and England to monitor inter-area oscillations. In the long term, the methodology introduced for increasing the number of PMUs deployed in the GB power system will allow a new generation of real time wide area monitoring system to become available. Furthermore, several SMT-based smart protection and control schemes have been proposed. For example, a wide area inter-area oscillation control system is introduced. This new control system uses wide area oscillatory signals as input signals, and produces dynamic parameters for the control of HVDC or TCSC to stabilize inter-area oscillations. This application has the potential to allow the conventional EMS-based open loop control to be upgraded to a sophisticated wide-area closed loop control system. The study performed to develop the GB WAMPAC roadmap highlighted that an application that can perform real time inter-area oscillation monitoring and control would be a valuable component of the future GB WAMPAC system. As a result of this, the thesis moves to focus on using WAMPAC to enhance the small signal stability i.e., inter-area oscillation stability. For this purpose, a fundamental study of inter-area oscillations is provided. Two classical methods were used to investigate the nature of inter-area oscillations i.e., nonlinear simulations and modal analysis. Different types of disturbance were used in the nonlinear simulations to observe the physical characteristics of local modes and the inter-area mode. To accurately understand the physical phenomenon behind inter-area oscillations, i.e. oscillation frequency, mode and source, modal analysis was introduced. From this modal analysis, it was found that the inter-area oscillatory mode is strong influenced by the inter-area power flow and transmission network parameters. This means that the large increase in inter-area active power flow caused by permanent faults may lead to inter-area oscillations becoming unstable, particularly due to the increased loading of future power systems. Therefore, a real time monitoring and warning system will become increasingly desirable to detect if a lightly damped/unstable inter-area oscillation exists in the system. The core of the research described in this thesis is to establish a real time inter-area oscillation monitoring platform for estimating the dominant inter-area oscillatory mode -185- Chapter 10 Conclusions and future work (frequency and damping) using the real time oscillatory signals captured by PMUs. This thesis presented a novel NTA for the real time estimation of the dominant inter-area oscillation mode. Two data sets were tested using the new algorithm; one data set was taken from simulated models and the other from real-life data records. These tests have shown that the NTA can reliably identify the oscillatory mode in static conditions for both data sets and in the presence of white noise. The dynamic capability of the NTA has also been successfully demonstrated by capturing the dynamic oscillatory mode when the parameters of the test signal experienced a step change. In both cases the NTA results were more accurate and stable than these using conventional Prony methods. The NTA-based real time inter-area oscillation monitoring can be considered as a suitable SMT application for deployment during the short-term phase of the GB WAMPAC strategy proposed in this thesis. As part of the long term GB WAMPAC strategy, this monitoring application could be extended into a real time wide area closed loop control application. Detailing how this oscillation control application could be created formed a key part of this thesis. Modal analysis-based linear control theory was used to design a wide area damping control system that uses HVDC, TCSC and SVC installed in an inter-area power transmission corridor, to stabilize inter-area oscillations. Conventional modal analysis in the frequency domain was used to find the optimum set of parameters for the system controllers, i.e. the operation point with maximum damping. Nonlinear simulations were used to confirm the new controllers’ ability to damp the inter-area oscillations. The final part of this thesis proposes a Wide Area Monitoring and Control System (WAMCS), designed to enhance the small signal stability of a future GB power system. The operation of this WAMCS was evaluated using a full GB power system model (vision 2015). When this system was in a vulnerable state, e.g. several PSS equipped generators were removed from service; the WAMCS developed here was capable of significantly increasing the damping of the lightly damped inter-area mode in the system. This loss of PSS control, and the associated lightly damped inter-area oscillation, will occur as the number of synchronous generators will be replaced by offshore wind generation that cannot provide inertia to the system. Two real applications i.e. inter-area oscillation monitoring with New Type Algorithm (NTA) and inter-area oscillation damping control with HVDC, were tested using dynamic simulation with a full GB power system model. The NTA has successfully -186- Chapter 10 Conclusions and future work extracted the inter-area oscillatory mode from a lightly damped inter-area oscillatory power flow between Scotland and England. In addition, the inter-area mode shape was obtained through NTA processing of the frequencies across the GB power system. A wide area damping control system was added to the submarine HVDC link between the SPTN and NGETN, to improve the damping of the inter-area mode. The input signal to these HVDC supplementary damping control system is the frequency difference between Scotland and England. This new damping control system modulated the DC power flow to stabilize the inter-area oscillations. Simulation results have shown that the damping of the inter-area mode has been significantly improved by the inclusion of the supplementary damping control system. One of most important issues in wide area control is the effect of any time delays. Here, the influence of time delay in the range of 50 ms to 500 ms was investigated. It has been concluded that the effect of wide area HVDC control system in damping inter-area oscillations is gradually reduced when the delay time is increased. A time delay in a certain range (approximately 150 ms-200 ms) introduces a high frequency oscillatory component into the inter-area oscillations. The additional HVDC damping control system has significantly changed the system’s dynamics and the interactions between this control system and the existing PSSs cannot be ignored. Simulation results have shown that the additional HVDC damping control system increased the damping of some generators and didn’t introduce a negative effect on the other generators. However, in a practical implementation the parameters of the PSSs should be tuned to coordinate with the HVDC damping controllers to ensure optimal system control. 10.2 Future Work The research described in this thesis represents the initial stage of tailoring a WAMPAC system for the needs of a future GB power system. Significant further research is needed before the proposed GB WAMCS can be implemented in the real GB power system. This thesis presents several wide area control schemes that can use individual power electronic devices i.e. HVDC, TCSC and SVC to damp inter-area oscillations. Here, -187- Chapter 10 Conclusions and future work only the use of HVDC in damping inter-area oscillations has been analyzed in a full GB power system model. In the future GB power system, several TCSC and SVC will be installed in the transmission corridors between Scotland and England to enhance the power transfer capability of the existing transmission lines. The potential opportunity to use TCSC and SVC to damp inter-area oscillations should also be evaluated in the full GB power system model. Coordination between all of the controllers in a power system is essential if system damping is to be optimized. Achieving this coordination should form part of any future work. This will require the development of optimised methods for the coordinated design and tuning of controllers; these methods should be evaluated in the full GB power system model. In this thesis, the wide area damping control schemes were designed using conventional linear control theory. Therefore, these control schemes often only work within a limited operating range. During power system operation, variable operating conditions and different contingencies can reduce the effect of the damping controllers, and allow lightly damped or even unstable inter-area oscillations to exist. Therefore, for practical implementations, a robust control design method is needed to ensure that the damping controllers are viable under a sufficiently wide range of operating conditions. A large number of simulations should be performed in the full GB power system model to confirm the robustness of the controllers for variable operating conditions and for different contingencies. As this thesis presents, the inevitable time delay of the wide area signal transmission is the main operational challenge for the implementation of a real time wide area control system. It has been shown that the time delay can reduce the damping effect of the wide area HVDC damping control system on inter-area oscillations, even allowing unstable system operation to occur. The time delay is usually caused by a number of variables, such as measurement processing times, the bandwidth of the communication medium and geographic distance, the time delay of in different channels may be different and variable. Therefore, another main aspect of the future work is developing a method for compensating the negative effect introduced by time delay. The time delay compensation method should be robust and flexible to deal with the variables; furthermore, the compensation method should also concern the issues related to the reliability of the communication network, GPS communication challenges. -188- synchronization and other Chapter 10 Conclusions and future work This research has developed the essential applications and algorithms of wide area monitoring and control for improving the inter-area oscillatory stability of the future GB power system. As the other essential part of the future GB WAMPAC system, SMTbased system protection schemes should also be developed to optimize the existing system emergency protection schemes, e.g. intelligent controlled islanding and wide area adaptive under-frequency load shedding. In the future work, the essential algorithms required for supporting these wide area protection schemes should be developed; and the performance of these wide area protection schemes i.e., the key factors that influence the performance of wide area protection schemes should be evaluated in full GB system model. -189- References References [1] D. Novosel, “CIEE Phasor Measurement Application Study,” project report, for California Energy Commission, KEMA Inc., 2008. [2] “Final Report on the August 14, 2003 Blackout in the United States and Canada: Causes and Recommendations,” US-Canada Power System Outage Task Force, April 2004. [3] “Final Report of the Investigation Committee on the 28 September 2003 Blackout in Italy,” TUCE, April 2004. [4] “System protection schemes in power networks,” CIGRE working group file, Task Force 38.0219. [5] “Defence plan against extreme contingencies,” CIGRE working group file, Task Force C2.02.24, April 2007. [6] V. Terzija, G. Valerde, D. Cai, P. Regulski, V. Madani, J. Fitch, S. Skok, Miroslav M. Begovic and A. Phadke, “Wide-Area Monitoring, Protection and Control of Future Electric Power Networks”, IEEE Proceedings, vol. 99, No.1, pp. 80-93, January. 2011 [7] Y. Hu, V. Madani, R. M. Moraes, and D. Novosel, "Requirements of Large-Scale Wide Area Monitoring, Protection and Control Systems," 10th Annual Conference, Fault and Disturbance Analysis, Georgia Tech, April 2007. [8] A. Hiortns, “National Grid 2020-2030 Networks”, National Grid presentation, April, 2009. [9] N. Tleis, “Technical challenges in delivering Vision 2020,” Presentation, National Grid, April, 2009. [10] C. Hor, J. Finn, G. Thumm, S. Mortimer, “Introducing series compensation in the UK transmission network,” 9th IET International Conference on AC and DC Power Transmission, March, 2010. [11] J.H. Shi, P. Li, X.C. Wu, J.T. Wu, C. Lu, Y. Zhang, Y.K. Zhao, J. Hu, “Implementation of an adaptive continuous real-time control system based on WAMS” Monitoring of Power System Dynamics Performance Conference, Saint Petersburg 28-30 April 2008. [12] “Wide area monitoring and control for transmission capability enhancement”, working group report, CIGRE, C4.601, August 2007. [13] J. Fitch, “Strategy document of UK WAMPAC”, Project report on UK WAMPAC R&D project, National Grid Ltd, July 2009. [14] A.G. Phadke and J.S.Thorp, “Synchronized Phasor Measurements and Their Applications,” Springer, USA, 2008. [15] A.G. Phadke, “Synchronized phasor measurements in power systems,” IEEE journals, Computer applications in power systems, vol.6, no.2, pp.10-15, April 1993. [16] V. Terzija, P. Crossley, D. Novosel, D. Karlsson, H. Li, presentation of WAMPAC course, Manchester, UK, July 2007. [17] A. G. Phadke, J.S. Thorp and M.G. Adamiak, “A new measurement technique for tracking voltage phasors, local system frequency, and rate of change of frequency”, IEEE Transactions on PAS. vol. 102, no. 5, pp 1025–1038, May 1983. [18] http://en.wikipedia.org/wiki/Gps. -190- References [19] IEEE standard for Synchrophasor measurement, C37.118, 2005. [20] D. Novosel, K. Vu, V. Centeno, S. Skok, M. Begovic, “Benefits of SynchronizedMeasurement Technology for Power-Grid Applications,” in Proc. 40th Annual Hawaii International Conference, System Sciences, Hawaii, 2007, pp. 118-126. [21] B.Bhargava and A.Salazar, "Use of Synchronized Phasor Measurement system for monitoring power system stability and system dynamics in real-time," IEEE Power and Energy Society General Meeting - Conversion and Delivery of Electrical Energy in the 21st Century, 2008. [22] E. Allen, D. Kosterev and P. Pourbeik, “Validation of power system models,” IEEE General Meeting, Power and Energy Society, Indianapolis, USA, 25-29 July 2010. [23] S.M. Rovnyak, “Power system model validation study using wide-area frequency data,” IEEE general Meeting, Power and Energy society, Indianapolis, USA, 25-29 July 2010. [24] J. Ma, D. Han, W.J. Sheng, R.M. He, C.Y. Yue and J. Zhang, “Wide area measurements-based model validation and its application,” IET, Generation, Transmission & Distribution, Vol. 2, No. 6, pp. 906-916, November 2008. [25] Z. Huang, B. Yang and D. Kosterev, “Benchmarking of planning models using recorded dynamics,” IEEE/PES, Power Systems Conference, Richland, 15-18 March 2009. [26] Y. Makarov, “Western Interconnection Phasor Based Projects” Overview – Working Document, PNNL, April 2006. [27] Z. Zhong, C.C. Xu, B. J. Billian, L. Zhang, S.J. S. Tsai, R. W. Conners, V. A. Centeno, A. G. Phadke, Y. Liu “Power system frequency monitoring network (FNET) implementation,” IEEE Trans, power systems, Vol. 20, No. 4, pp.1914 – 1921, November 2005. [28] J. N. Bank, R. M. Gardner, J. K. Wang, A. J. Arana, Y. Liu “Generator Trip Identification Using Wide-Area Measurements and Historical Data Analysis,” in Proc. PSCE conference, Virginia, USA, pp.1677-1683, October 2006. [29] A. M. Oritiz, T. K. Bravo, “Earthquake Location-regional triangulation with real data,” http://www.scieds.com/spinet. [30] Y. Liu “Synchronized Frequency Measurements & Applications,” Presentation in Virginia Tech, May 2007. http://phasors.pnl.gov/Meetings/2007_may/presentations/synch_freq_meas.pdf. [31] M. Zima, M. Larsson, P. Korba, C. Rehtanz and G. Andersson, “Design Aspect for Wide-Area Monitoring and Control System,” IEEE Proceedings, Vol. 93, No. 5, pp. 980-996, May 2005. [32] T.V. Custem and C. Vournas, “Voltage Stability of Electric Power Systems,” 2nd printing edition, Springer, USA, Oct. 2007. [33] M. Larsson, C. Rehtanz and J. Bertsch, “Monitoring and operation of transmission corridors,” PowerTech Conference, Bologna, Italy, 23-26 June 2003. [34] P. Kundur, “Power System Stability and Control,” McGarw-Hill, Inc., 1993. [35] J. He, C. Lu, X. Wu, P. Li and J.Wu, “Design and experiment of wide area HVDC supplementary damping controller considering time delay in China southern power grid”, IET Generation, Transmission & Distribution, Vol. 3, No. 1, pp.17-25, January 2009. -191- References [36] J. Hauer, “Application of Prony analysis to the determination of modal content and equivalent models for measured power system response”, IEEE Trans., Power system, Vol. 6, No. 6, pp. 1062 - 1068, August 1991. [37] P. Korba, K. Uhlen, “Wide-area monitoring of electromechanical oscillations in the Nordic power system: practical experience”, IET, Generation, Transmission and Distribution, Vol. 4, No.4, pp.1116-1126, August 2009. [38] P. Korba, M. Larson, C. Rehtanz, “Detection of Oscillations in Power System Using Kalman Filtering Techniques”, Proc. IEEE Conference on Control Applications, vol.1, pp.183-188 August 2003. [39] N. Kakimoto, M. Sugumi, T. makino and K. Tomiyama, ”Monitoring of Inter Area Oscillation Mode by synchronized Phasor Measurement” IEEE trans. on Power Systems, vol. 21, pp. 260-268, January 2006. [40] A. Bretas, “Robust Electric Power Infrastructures. Response and Recovery during Catastrophic Failures,” Ph.D. dissertation, Virginia Tech, 2001. [41] M. M. Adibi, “Power System Restoration: Methodologies and Implementation Strategies,” Ed., Published in Jun 2000 by Wiley US. [42] D. Novosel, V. Madani, B. Bhargava, V. Khoi and J. Cole, “Dawn of the grid synchronization”, IEEE Power and Energy Magazine, Vol. 6, No. 1, pp. 49-60, February 2008. February 2008. [43] W. Taylor et al., “Northwest power pool transient stability and load shedding controls for generation-load imbalances,” IEEE Trans. On Power Apparatus and Systems., Vol. PAS-100, No. 7, pp.3486-3495, July 1981. [44] Westinghouse, “Applied Protective Relaying”, Westinghouse Electric Corporation, Newark, N.J., 1976. [45] V. Terzija, “Adaptive underfrequency load shedding based on the magnitude of the disturbance estimation”, IEEE Trans. on Power Systems, Vol. 21, No. 3, pp. 1260– 12 [46] M. Parashar, J.Dyer and T. Bilke, “EIPP real-time dynamics monitoring system,” http://certs.lbl.gov/pdf/eipp-rt.pdf. [47] C. Martinez, M. Parashar, J. Dyer, J. Coroas, “Phasor Data Requirements for Real Time Wide-Area Monitoring, Control and Protection Applications”, EIPP White PaperFinal Draft, January 16, 2005. http://www.phasorrtdms.com/downloads/research/PhasorDataRequirements-WhitePaper012605.pdf [48] B. Naduvathuparambil, M. C. Valenti, and A. Feliachi, “Communication delays in wide area measurement systems,” Procs. of the Thirty-Fourth South-eastern Symposium on System Theory, November. 2002. [49] http://en.wikipedia.org/wiki/Asymmetric_digital_subscriber_line [50] A. Khatib, “Inter-based Wide Area Measurement Applications in Deregulated Power Systems” Ph.D. dissertation, Virginia Tech, July 2002. [51] J.H.Shi, P.Li, X.C.Wu, J.T.Wu, C.Lu, Y.Zhang, Y.K.Zhao, J.Hu, “Implementation of an adaptive continuous real-time control system based on WAMS”, CIGRE meeting on Monitoring of Power System Dynamics Performance, Saint Petersburg, Russia 28-30 April 2008. [52] P. Myrda, E. Gunther, “EIPP Data Management Task Team Architecture” Procs. of the 40th Hawaii International Conference on System Sciences, January 2007. -192- References [53] J.Y.Cai, Z.Huang, J.Hauer and K.Martin, “Current status and experiences of WAMS implementation in North America,” IEEE/PES Transmission and Distribution Conference and Exhibition: Asia and Pacific, Dalian, China, 2005. [54] D. H. Wilson, “wide Area Monitoring systems in the UK: Operational Experience and Systems Development”, CIGRE meeting on Monitoring of Power Dynamics Performance, Saint Petersburg, Russia, 28-30 April 2008. [55] “Phasor Measurement Application Study”, PIER Final Project Report for California Energy Commission, KEMA Inc. [56] N. Tleis, “GB Grid Code Requirements for HVDC,” Presentation in HVDC seminar, at the University of Manchester, September, 2008. [57] A. Carter, “WAMPAC I Projecting Meeting”, Wokingham, 2009. [58] W. Taylor et al., “Northwest power pool transient stability and load shedding controls for generation-load imbalances,” IEEE Trans. Power App. Syst., vol. PAS100, no. 7, pp.3486-3495, Jul.1981. [59] G. Rogers, “Power system oscillations”, Kluwer, 2000. [60] M. Klein, G. J. Rogers, P. Kundur, “A fundamental study of inter-area oscillations”, IEEE Trans. Power systems, vol. 6, pp. 914-921, Aug. 1991. [61] R. Sadikovi’c, “Use of FACTS Devices for Power Flow Control and Damping of Oscillations in Power Systems,” PhD Dissertation, Swiss Federal Institute of Technology Zurich. 2006. [62] C.W. Taylor, “Improving grid behavior”, IEEE Spectrum, June 1999, pp. 40-45. [63] J. Hauer, C. Demecure and L. Scharf, “Initial results in Prony analysis of power system response signals,” IEEE Trans. Power Systems, vol. 5, no.1, pp.80-89, Feb. 1990. [64] J. Xiao, X. Xie, Y. Han, and J. Wu, “Dynamic tracking of low-frequency oscillations with improved Prony method in wide-area measurement system”, in Proc. IEEE Power Engineering Society General Meeting, Denver, 2004, vol. 1, pp. 1104-1109. [65] P.O’Shea, “A high-resolution spectral analysis algorithm for power system disturbance monitoring”, IEEE Trans, Power system, vol. 17, no.5, pp. 676-680, Aug. 2002. [66] J.C.-H. Peng and N.-K.C. Nair “Adaptive sampling scheme for monitoring oscillations using Prony analysis”, IET, Generation, Transmission & Distribution, vol. 3, no.12, pp. 1052-1060, July, 2009. [67] P. Korba, “Real-time monitoring of electromechanical oscillations in power systems: first findings”, IET, Generation, Transmission & Distribution, vol. 1, no.1, pp. 8088, Jan. 2007. [68] R.W. Wies, J.W. Pierre and D.J. Trudnoswki, “Use of Least-mean squares adaptive filtering technique for estimating low-frequency electromechanical modes in power systems”, in Proc. IEEE American Control Conference, Anchorage, 2002, vol. 6, pp. 4867 - 4873. [69] N. Zhou, J. W. Pierre, D. J Trudnowski, R. T. Guttromson, “Robust RLS methods for online estimation of power system electromechanical modes”, IEEE Trans, Power systems. vol. 22, no.3, pp. 1240-1249, Aug. 2007. -193- References [70] N. Zhou, D. J. Trudnowski, J. W. Pierre and W.A. Mittelstadt, “Electromechanical mode online estimation using regularized robust RLS methods”, IEEE Trans, Power Systems, vol. 23, no.4, pp. 1670-1680, Nov. 2008. [71] R. W. Wies, J. W. Pierre and D. J. Trudnowski, “Use of ARMA block processing for estimating stationary low-frequency electromechanical modes of power systems ”, IEEE Trans. Power systems, vol. 18, no.1, pp. 167-173. Feb. 2003. [72] N. Kakimoto, M. Sugumi, T. Makino and K. Tomiyama, “Monitoring of inter-area oscillation mode by synchronized phasor measurement”, IEEE Trans, Power systems, vol. 21, no.1, pp. 260-268, Feb. 2006. [73] V. Terzija, M. B. Milenko, B. D. Kovacevic, “Voltage phasor and local system frequency estimation using Newton type algorithm”, IEEE Trans. power delivery, vol. 9, no.3, pp.1368-1374, July 1994. [74] V. Terzija, M. Djuric, B. Kovacevic, “A new self-tuning algorithm for the frequency estimation of distorted signals”, IEEE Trans. power delivery, vol. 10, no.4, pp. 1779-1785, Oct. 1995. of DIGSILENT Powerfactory”, available at: [75] “Introduction http://www.digsilent.de/Software/DIgSILENT_PowerFactory/. [76] V. Terzija, P. Regulski, L. P. Kunjumuhammed, B. C.P al, G. Burt, I. Abdulhadi, T. Babnik, M. Osborne, W. Hung, “FlexNet Wide Area Monitoring System”, in Proc. IEEE Conf. Power and energy society general meeting, Detroit, 2011. [77] E.V.Larsen, D.A.Swann, "Applying Power System Stabilizers Part I: General Concepts," IEEE Transactions on Power Apparatus and Systems, vol. PAS-100, no.6, pp 3017-3024. [78] E.V.Larsen, D.A.Swann, "Applying Power System Stabilizers Part II: Performance Objectives and Tuning Concepts" IEEE Transactions on Power Apparatus and Systems, Volume: PAS-100, no. 6, pp 3025-3033. [79] E.V.Larsen, D.A.Swann, "Applying Power System Stabilizers Part III: Practical Considerations ," IEEE Transactions on Power Apparatus and Systems, vol. PAS100 , no.6, pp 3034-3046. [80] G. Rogers, “Power system oscillations”, Kluwer, 2000. [81] M.Klein, G.J. Rogers, S. Moorty and P. Kundur, “Analytical investigation of factors influencing power system stabilizers performance,” IEEE Trans. on power systems, vol. EC-7, pp 382-388, September 1992. [82] M. E. Aboul-Ela, A. A. Salam, J. D. McCalley and A. A. Fouad, “Damping Controller Design for Power System Oscillations Using Global Signals”, IEEE, Trans. on Power Systems, Vol. 11, No.2, May 1996, 767-773. [83] R. Sadikovi’c, “Use of FACTS devices for power flow control and damping of oscillations in power systems,” Ph.D thesis, Swiss Federal Institute of Technology Zurich, 2006. [84] S. K. Yee, “Coordinated tuning of power system damping controllers for robust stabilization of the system”, Ph.D thesis, School of IEEE, The University of Manchester, Oct. 2005. [85] V. K. Sood, “HVDC and FACTS Controllers”, Kluwer Academic Publishers, Boston, 2004. [86] National Grid full DIGSILENT model. -194- References [87] “Thyristor controlled series compensation”, CIGRE working group file, ‘Working Group 14.18’, 2000. [88] K. Keerthivasan, V. S. Deve, J. Jerome and R. Ramanujan, “Modelling of SVC and TCSC for power system dynamic simulation”, Proc. The 7th international Conf. Power engineering, Singapore, 2005, vol. 2, pp. 696-700. [89] J.H.Shi, P. Li, X.C. Wu, J.T. Wu, C. Lu, Y. Zhang, Y.K, Zhao and J. Hu, “Implementation of an adaptive continuous real-time control system based on WAMS”, Conf. Monitoring of power system dynamics performance, Saint Petersburg, 28-30 April 2008. [90] B. Nadu., M. C. Valenti, A. Feliachi, “Communication delays in wide area measurement systems,” in Proc. 34th Southeastern Symposium on System Theory, 2002, pp. 118-122. [91] J. He, C. Lu, X. Jin, P. Li, “Analysis of time delay effects on wide area damping control,” IEEE Asia Pacific Conference on Circuits and Systems, 2008. -195- Appendices Appendices 10.1 Appendix A Figure A-1: A single line diagram of two-area system. Table A-1: Synchronous machine parameters of G1, G2 and G3 and G4. No. of Generator 1 2 3 4 Rotor type Round Rotor Round Rotor Round Rotor Round Rotor Inertia time constant H (rated to Sgn) Mechanical damping Stator resistance ra Stator leakage reactance xl Synchronous reactance xd d-axis Synchronous reactance xq q-axis Transient reactance xd’ d-axis Transient reactance xq’ q-axis Sub-transient reactance xd’’ d-axis Sub-transient reactance xq’’ q-axis Transient time constant Td0’ d-axis Transient time constant Tq0’ q-axis Sub-transient time constant Td0’’ daxis Sub-transient time constant Tq0’’ qaxis 6.5 0 0.0025 0.2 1.8 1.7 0.3 0.55 0.25 0.25 8 0.4 0.03 6.5 0 0.0025 0.2 1.8 1.7 0.3 0.55 0.25 0.25 8 0.4 0.03 6.5 0 0.0025 0.2 1.8 1.7 0.3 0.55 0.25 0.25 8 0.4 0.03 6.5 0 0.0025 0.2 1.8 1.7 0.3 0.55 0.25 0.25 8 0.4 0.03 0.05 0.05 0.05 0.05 Table A-2: Power generation conditions of G1, G2 and G3 and G4. Gen 1 2 3 4 Bus type SL PV PV PV Rated power (MVA) 900 900 900 900 Nominal voltage (L-L kV) 20 20 20 20 Active power output 750 700 719 700 -196- Reactive power output 185 235 176 202 Terminal voltage (p.u.) 1.03 1.01 1.03 1.01 Appendices Vref Vmeas + 1 1 + sTr - Vmax K 1 + sTe 1 + sTb 1 + sTa Vmin Figure A-2: Block diagram of static exciter of G1, G2 and G3 and G4. 1 2 3 4 5 6 7 Table A-3: Parameters of static exciter. 0.01 Transducer filter time constant Tr Voltage regulator gain K 150 Voltage regulator time constant Te 0.05 Transient gain reduction time constant Ta 1 Transient gain reduction time constant Tb 10 Maximum voltage regulator output Vmax 3 Minimum voltage output Vmin -3 second p.u. second Second Second p.u. p.u. Tmax 1. 0 K ωrotor Pref 1 1 + sTs 1 + sT3 1 + sTc 1 + sT4 1 + sT5 Tmin Figure A-3: Block diagram of speed governor of G1, G2, G3 and G4. 1 2 3 4 5 6 7 8 Table A-4: Parameters of speed governor. Governor gain K 50 p.u. Servo time constant Ts 0.1 second Transient gain time constant T3 0 Second HP turbine time constant Tc 0.5 Second Time constant to set HP ratio T4 1.25 Second Reheat time constant T5 5 Second Maximum power Pmax 1 p.u. Minimum power Pmin 0 p.u. Table A-5: Transformer parameters. No. of Transformer Rated power (MVA) Rated voltage (HV) Rated voltage (LV) Short circuit voltage (positive sequence %) Short circuit voltage (zero sequence. %) Winding connection (HV) Winding connection (LV) -197- 1 900 230 20 15 3 YN YN 2 900 230 20 15 3 YN YN 3 900 230 20 15 3 YN YN 4 900 230 20 15 3 YN YN Appendices Table A-6: AC transmission line parameters. No. of line 1 2 3 4 5 6 Rated voltage (L-L kV) 230 230 230 230 230 230 Length (km) 25 10 110 110 110 110 0.0529 0.0529 0.0529 0.0529 0.0529 0.0529 Resistance (Ohm/km) Reactance (Ohm/km) 0.529 0.529 0.529 0.529 0.529 0.529 3.3075 3.3075 3.3075 3.3075 3.3075 3.3075 Susceptance (µs/km) 7 230 10 8 230 25 0.0529 0.0529 0.529 0.529 3.3075 3.3075 Table A-7: Load data. Load Active power (MW) Reactive power (MVar) Load 1 967 100 Load 2 1667 100 The active load is modeled as 50% constant and 50% constant impedance. The reactive load is modeled as constant impedance. No. of Capacitor 1 2 3 Nominal voltage (kV) 230 230 230 Table A-8: Shunt capacitors. Connected Reactive power bus output (MVar) 3 215.34 4 246.73 5 378.38 -198- Minimum output (MVar) Maximum output (MVar) 100 100 50 600 500 500 Appendices 10.2 Appendix B Figure B-1: A single line diagram of the two-area system with HVDC. Table B-1: Synchronous machine parameters of G1, G2 and G3 and G4. No. of Generator 1 2 3 4 Rotor type Round Rotor Round Rotor Round Rotor Round Rotor Inertia time constant H (rated to Sgn) Mechanical damping Stator resistance ra Stator leakage reactance xl Synchronous reactance xd d-axis Synchronous reactance xq q-axis Transient reactance xd’ d-axis Transient reactance xq’ q-axis Sub-transient reactance xd’’ d-axis Sub-transient reactance xq’’ q-axis Transient time constant Td0’ d-axis Transient time constant Tq0’ q-axis Sub-transient time constant Td0’’ daxis Sub-transient time constant Tq0’’ qaxis 6.5 0 0.0025 0.2 1.8 1.7 0.3 0.55 0.25 0.25 8 0.4 0.03 6.5 0 0.0025 0.2 1.8 1.7 0.3 0.55 0.25 0.25 8 0.4 0.03 6.5 0 0.0025 0.2 1.8 1.7 0.3 0.55 0.25 0.25 8 0.4 0.03 6.5 0 0.0025 0.2 1.8 1.7 0.3 0.55 0.25 0.25 8 0.4 0.03 0.05 0.05 0.05 0.05 Table B-2: Power generation conditions of G1, G2 and G3 and G4. Gen 1 2 3 4 Bus type SL PV PV PV Rated power (MVA) 900 900 900 900 Nominal voltage (L-L kV) 20 20 20 20 Active power output 750 700 719 700 -199- Reactive power output 185 235 176 202 Terminal voltage (p.u.) 1.03 1.01 1.03 1.01 Appendices Vref Vmeas 1 1 + sTr Vmax 1 + sTc 1 + sTb 1 sTe Ka 1 + sTa Vmin Ke f ( E1 , Se1 , E2 , Se 2 ) Kf 1 + sT f Figure B-2: Block diagram of DC exciter of G1, G3 and G4. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 Table B-3: Parameters of DC exciter. Transducer filter time constant Tr 0.02 Voltage regulator gain Ka 20 Voltage regulator time constant Ta 0.005 Filter derivative time constant Tc 3.1 Filter delay time constant Tb 40 Exciter time constant Te 0.05 Exciter constant Ke 0.5 Stabilization path gain Kf 0.01 Stabilization path time constant Tf 0.3 Saturation factor 1E1 3.9 Saturation factor 2Se1 0.0001 Saturation factor 3E2 5.2 Saturation factor 4Se2 0.001 Maximum voltage regulator output Vmax 5 Minimum voltage regulator output Vmin -5 Vref Vmeas 1 1 + sTr second p.u. second second second second p.u. p.u. second p.u. p.u. p.u. p.u. p.u. p.u. Vmax 1 + sTb 1 + sTa K 1 + sTe Vmin Figure B-3: Block diagram of static exciter of G2. -200- Appendices 1 2 3 4 5 6 7 Table B-4: Parameters of static exciter. Transducer filter time constant Tr 0.01 Voltage regulator gain K 200 Voltage regulator time constant Te 0.05 Transient gain reduction time constant Ta 1 Transient gain reduction time constant Tb 10 Maximum voltage regulator output Vmax 3 Minimum voltage output Vmin -3 second p.u. second Second Second p.u. p.u. Tmax 1.0 K ωrotor Pref 1 1 + sTs 1 + sT3 1 + sTc 1 + sT4 1 + sT5 Tmin Figure B-4: Block diagram of speed governor of G1, G2, G3 and G4. 1 2 3 4 5 6 7 8 Table B-5: Parameters of speed governor. Governor gain K 50 p.u. Servo time constant Ts 0.1 second Transient gain time constant T3 0 Second HP turbine time constant Tc 0.5 Second Time constant to set HP ratio T4 1.25 Second Reheat time constant T5 5 Second Maximum power Pmax 1 p.u. Minimum power Pmin 0 p.u. Table B-6: Transformer parameters. No. of Transformer 1 2 Rated power (MVA) 900 900 Rated voltage (HV) 230 230 Rated voltage (LV) 20 20 Short circuit voltage (positive Sequence %) 15 15 Short circuit voltage (zero Sequence %) 3 3 Winding connection (HV) YN YN Winding connection (LV) YN YN 3 900 230 20 15 3 YN YN Table B-7: AC transmission line parameters. No. of line 1 2 3 4 5 6 Rated voltage (L-L kV) 230 230 230 230 230 230 Length (km) 25 10 110 110 110 110 0.0529 0.0529 0.0529 0.0529 0.0529 0.0529 Resistance (Ohm/km) Reactance (Ohm/km) 0.529 0.529 0.529 0.529 0.529 0.529 3.3075 3.3075 3.3075 3.3075 3.3075 3.3075 Susceptance (µs/km) -201- 4 900 230 20 15 3 YN YN 7 230 10 8 230 25 0.0529 0.0529 0.529 0.529 3.3075 3.3075 Appendices Table B-8: DC transmission line parameters. DC transmission line 1 Rated DC voltage (kV) 56 Rated current (kA) 3.6 Length (km) 200 Overhead line Line type 0.0075 Resistance (Ohm/km) Inductance (mH/km) 1 Figure B-5: Rectifier operation condition (DIgSILENT interface). α max kp 1 1 + sT f I ref I meas ki s α min Figure B-6: Rectifier’s α control for constant current. 1 2 3 4 5 6 Table B-9: Parameters of rectifier control. Current reference for current control 1.8 kA Proportional gain Kp 0.001 p.u. Output gain Ko 1.0 p.u. Rectifier time constant Tf 0.001 second Integral gain Ki 0.01 p.u. 5 Minimum firing angle αmin degree 90 Maximum firing angle αmax degree -202- α Appendices Figure B-7: Inverter operation condition (DIgSILENT interface). β max kp 1 1 + sT f Vref Vmeas ki s β β min Figure B-8: inverter’s β control for constant voltage. 1 2 3 4 5 6 7 Table B-10: Parameters of inverter control. Voltage reference for voltage control 1.0 p.u. Proportional gain Kp 0.001 p.u. Output gain Ko 1 p.u. Rectifier time constant Tf 0.001 second Integral gain Ki 0.01 p.u. 30 Minimum advance firing angle βmin degree 90 Maximum advance firing angle βmax degree Table B-11: Load data. Load Active power (MW) Reactive power (MVar) Load 1 967 100 Load 2 1667 100 The active load is modeled as 50% constant and 50% constant impedance. The reactive load is modeled as constant impedance. -203- Appendices No. of Capacitor 1 2 3 Nominal voltage (kV) 230 230 230 4 230 Table B-12: Shunt capacitors. Connected bus Reactive power output (MVar) 3 189.88 5 341.61 Rectifier AC 118.67 terminal Rectifier AC 122.00 terminal -204- Minimum output (MVar) 100 50 0 Maximum output (MVar) 600 500 125 0 125 Appendices 10.3 Appendix C Figure C-1: A single line diagram of the two-area system with HVDC. Table C-1: Synchronous machine parameters of G1, G2 and G3 and G4. No. of Generator 1 2 3 4 Rotor type Round Rotor Round Rotor Round Rotor Round Rotor Inertia time constant H (rated to Sgn) Mechanical damping Stator resistance ra Stator leakage reactance xl Synchronous reactance xd d-axis Synchronous reactance xq q-axis Transient reactance xd’ d-axis Transient reactance xq’ q-axis Sub-transient reactance xd’’ d-axis Sub-transient reactance xq’’ q-axis Transient time constant Td0’ d-axis Transient time constant Tq0’ q-axis Sub-transient time constant Td0’’ daxis Sub-transient time constant Tq0’’ qaxis 6.5 0 0.0025 0.2 1.8 1.7 0.3 0.55 0.25 0.25 8 0.4 0.03 6.5 0 0.0025 0.2 1.8 1.7 0.3 0.55 0.25 0.25 8 0.4 0.03 6.5 0 0.0025 0.2 1.8 1.7 0.3 0.55 0.25 0.25 8 0.4 0.03 6.5 0 0.0025 0.2 1.8 1.7 0.3 0.55 0.25 0.25 8 0.4 0.03 0.05 0.05 0.05 0.05 Table C-2: Power generation conditions of G1, G2 and G3 and G4. Gen 1 2 3 4 Bus type SL PV PV PV Rated power (MVA) 900 900 900 900 Nominal voltage (L-L kV) 20 20 20 20 Active power output 750 700 719 700 Reactive power output 185 235 176 202 Vref Vmeas 1 1 + sTr Vmax 1 + sTb 1 + sTa K 1 + sTe Vmin Figure C-2: Block diagram of static exciter of G1, G2 G3 and G4. -205- Terminal voltage (p.u.) 1.03 1.01 1.03 1.01 Appendices 1 2 3 4 5 6 7 Table C-3: Parameters of static exciter. Transducer filter time constant Tr 0.01 Voltage regulator gain K 200 Voltage regulator time constant Te 0.05 Transient gain reduction time constant Ta 1 Transient gain reduction time constant Tb 10 Maximum voltage regulator output Vmax 3 Minimum voltage output Vmin -3 second p.u. second Second Second p.u. p.u. Tmax 1.0 K ωrotor Pref 1 1 + sTs 1 + sT3 1 + sTc 1 + sT4 1 + sT5 Tmin Figure C-3: Block diagram of speed governor of G1, G2, G3 and G4. 1 2 3 4 5 6 7 8 Table C-4: Parameters of speed governor. Governor gain K 50 p.u. Servo time constant Ts 0.1 second Transient gain time constant T3 0 Second HP turbine time constant Tc 0.5 Second Time constant to set HP ratio T4 1.25 Second Reheat time constant T5 5 Second Maximum power Pmax 1 p.u. Minimum power Pmin 0 p.u. Table C-5: Transformer parameters. No. of Transformer 1 2 Rated power (MVA) 900 900 Rated voltage (HV) 230 230 Rated voltage (LV) 20 20 Short circuit voltage (positive Sequence %) 15 15 Short circuit voltage (zero Sequence %) 3 3 Winding connection (HV) YN YN Winding connection (LV) YN YN No. of line 1 2 3 4 5 6 7 8 Rated voltage (L-L kV) 230 230 230 230 230 230 230 230 3 900 230 20 15 3 YN YN Table C-6: AC transmission line parameters. Length Resistance Reactance (km) (Ohm/km) (Ohm/km) 0.0529 25 0.529 0.0529 10 0.529 0.0529 110 0.529 0.0529 110 0.529 0.0529 110 0.529 0.0529 110 0.529 0.0529 10 0.529 0.0529 25 0.529 -206- 4 900 230 20 15 3 YN YN Susceptance (µs/km) 3.3075 3.3075 3.3075 3.3075 3.3075 3.3075 3.3075 3.3075 Appendices Table C-7: DC transmission line parameters. DC transmission line 1 Rated DC voltage (kV) 56 Rated current (kA) 3.6 Length (km) 200 Overhead line Line type 0.0075 Resistance (Ohm/km) Inductance (mH/km) 1 Figure C-4: Rectifier operation condition (DIgSILENT interface). α max kp 1 1 + sT f I ref I meas ki s α min Figure C-5: Rectifier’s α control for constant current. 1 1 2 3 4 5 6 Table C-8: Parameters of rectifier control. Current reference for current control 1.0 p.u. Proportional gain Kp 0.001 p.u. Output gain Ko 1.0 p.u. Rectifier time constant Tf 0.001 second Integral gain Ki 0.01 p.u. 5 Minimum firing angle αmin degree 90 Maximum firing angle αmax degree -207- α Appendices Figure C-6: Inverter operation condition (DIgSILENT interface). β max kp 1 1 + sT f Vref Vmeas ki s β β min Figure C-7: inverter’s β control for constant voltage. 1 2 3 4 5 6 7 Table C-9: Parameters of inverter control. Voltage reference for voltage control 1.0 Proportional gain Kp 0 Output gain Ko 1 Rectifier time constant Tf 0.001 Integral gain Ki 0.01 30 Minimum advance firing angle βmin 90 Maximum advance firing angle βmax -208- p.u. p.u. p.u. second p.u. degree degree Appendices I ref + I meas - α max kp + 1 1 + sT f + - ki s α (rectifier) α min umax Freq area 1 + 1 1 + sTm kd - 1 + sTlead 1 + sTlag sTw 1 + sTw 1 + sTlead 1 + sTlag Limit umin Freq area 2 Figure C-8: Rectifier control with supplementary damping control. 1 2 3 4 5 6 7 Table C-10: Parameters of HVDC supplementary control. Damping control gain Kd 1600 p.u. Low pass filter time constant Tm 0.1 second. Washout time constant Tw 10 second Lead phase compensation time constant Tlead 0.1 second 1.03 second Lag phase compensation time constant Tlag -5 Minimum damping controller output umin second 5 Maximum damping controller output umax second Table C-11: Load data. Load Active power (MW) Reactive power (MVar) Load 1 967 100 Load 2 1667 100 The active load is modeled as 50% constant and 50% constant impedance. The reactive load is modeled as constant impedance. No. of Capacitor 1 2 3 Nominal voltage (kV) 230 230 230 4 230 Table C-12: Shunt capacitors. Connected bus Reactive power output (MVar) 3 189.88 5 341.61 Rectifier AC 118.67 terminal Inverter AC 122.00 terminal -209- Minimum output (MVar) 100 50 0 Maximum output (MVar) 600 500 125 0 125 Appendices 10.4 Appendix D Figure D-1: A single line diagram of the two-area system with TCSC. Table D-1: Synchronous machine parameters of G1, G2 and G3 and G4. No. of Generator 1 2 3 4 Rotor type Round Rotor Round Rotor Round Rotor Round Rotor Inertia time constant H (rated to Sgn) Mechanical damping Stator resistance ra Stator leakage reactance xl Synchronous reactance xd d-axis Synchronous reactance xq q-axis Transient reactance xd’ d-axis Transient reactance xq’ q-axis Sub-transient reactance xd’’ d-axis Sub-transient reactance xq’’ q-axis Transient time constant Td0’ d-axis Transient time constant Tq0’ q-axis Sub-transient time constant Td0’’ daxis Sub-transient time constant Tq0’’ qaxis 6.5 0 0.0025 0.2 1.8 1.7 0.3 0.55 0.25 0.25 8 0.4 0.03 6.5 0 0.0025 0.2 1.8 1.7 0.3 0.55 0.25 0.25 8 0.4 0.03 6.5 0 0.0025 0.2 1.8 1.7 0.3 0.55 0.25 0.25 8 0.4 0.03 6.5 0 0.0025 0.2 1.8 1.7 0.3 0.55 0.25 0.25 8 0.4 0.03 0.05 0.05 0.05 0.05 Table D-2: Power generation conditions of G1, G2 and G3 and G4. Gen 1 2 3 4 Bus type SL PV PV PV Rated power (MVA) 900 900 900 900 Nominal voltage (L-L kV) 20 20 20 20 Active power output 750 700 719 700 reactive power output 185 235 176 202 Terminal voltage (p.u.) 1.03 1.01 1.03 1.01 Vmax Vref 1 1 + sTr 1 + sTb 1 + sTa Vmeas K 1 + sTe Vmin Figure D-2: Block diagram of static exciter of G1, G2 and G3 and G4. -210- Appendices Table D-3: Parameters of Static Exciter. Transducer filter time constant Tr 0.01 Voltage regulator gain K 150 Voltage regulator time constant Te 0.05 Transient gain reduction time constant Ta 1 Transient gain reduction time constant Tb 10 Maximum voltage regulator output Vmax 3 Minimum voltage output Vmin -3 1 2 3 4 5 6 7 second p.u. second Second Second p.u. p.u. Tmax 1.0 1 1 + sTs K ωrotor 1 + sT3 1 + sTc 1 + sT4 1 + sT5 Pref Tmin Figure D-3: Block diagram of speed governor for G1, G2, G3 and G4. 1 2 3 4 5 6 7 8 Table D-4: Parameters of Governor. Governor gain K 50 p.u. Servo time constant Ts 0.1 second Transient gain time constant T3 0 Second HP turbine time constant Tc 0.5 Second Time constant to set HP ratio T4 1.25 Second Reheat time constant T5 5 Second Maximum power Pmax 1 p.u. Minimum power Pmin 0 p.u. Table D-5: The capacitor and reactor of TCSC under steady state. 1 Series capacitor 20 Ohm 2 Series reactor (in parallel with the capacitor) 35 Ohm. xmax xC ⋅ xTCSC xC + xTCSC xTCSC 1 1 + sT f xreactor xmin u max kd sTw 1 + sTw 1 1 + sTm 1 + sTlead 1 + sTlag u min Figure D-4: TCSC with supplementary damping control. -211- Appendices Table D-6: Parameters of TCSC supplementary damping control. 1 Damping control gain Kd 22000 p.u. 2 Low pass filter time constant Tm 0.1 second. 3 Washout time constant Tw 10 second 0.11 4 Lead phase compensation time constant Tlead second 0.556 second 5 Lag phase compensation time constant Tlag -10 6 Minimum damping controller output umin Ohm 10 7 Maximum damping controller output umax Ohm Table D-7: Transformer parameters. No. of Transformer 1 Rated power (MVA) 900 Rated voltage (HV) 230 Rated voltage (LV) 20 Short circuit voltage (pos. Sequ. %) 15 Short circuit voltage (zero Sequ. %) 3 Winding connection (HV) YN Winding connection (LV) YN No. of line 1 2 3 4 5 6 7 8 9 10 Rated voltage (L-L kV) 230 230 230 230 230 230 230 230 230 230 2 900 230 20 15 3 YN YN 3 900 230 20 15 3 YN YN Table D-8: Transmission line parameters. Length Resistance Reactance (km) (Ohm/km) (Ohm/km) 0.0529 25 0.529 0.0529 10 0.529 0.0529 110 0.529 0.0529 110 0.529 0.0529 110 0.529 0.0529 110 0.529 0.0529 10 0.529 0.0529 25 0.529 0.0529 110 0.529 0.0529 110 0.529 4 900 230 20 15 3 YN YN Susceptance (µs/km) 3.3075 3.3075 3.3075 3.3075 3.3075 3.3075 3.3075 3.3075 3.3075 3.3075 Table D-9: Load data. Load Active power (MW) Reactive power (MVar) Load 1 967 100 Load 2 1667 100 The active load is modeled as 50% constant and 50% constant impedance. The reactive load is modeled as constant impedance. No. of Capacitor 1 2 Nominal voltage (kV) 230 230 Table D-10: Shunt capacitors. Connected Reactive power bus output (MVar) 3 193.37 5 341.78 -212- Maximum output (MVar) 600 500 Appendices 10.5 Appendix E Figure E-1: A single line diagram of the two-area system with SVC. Table E-1: Synchronous machine parameters of G1, G2 and G3 and G4. No. of Generator 1 2 3 4 Rotor type Round Rotor Round Rotor Round Rotor Round Rotor Inertia time constant H (rated to Sgn) Mechanical damping Stator resistance ra Stator leakage reactance xl Synchronous reactance xd d-axis Synchronous reactance xq q-axis Transient reactance xd’ d-axis Transient reactance xq’ q-axis Sub-transient reactance xd’’ d-axis Sub-transient reactance xq’’ q-axis Transient time constant Td0’ d-axis Transient time constant Tq0’ q-axis Sub-transient time constant Td0’’ daxis Sub-transient time constant Tq0’’ qaxis 6.5 0 0.0025 0.2 1.8 1.7 0.3 0.55 0.25 0.25 8 0.4 0.03 6.5 0 0.0025 0.2 1.8 1.7 0.3 0.55 0.25 0.25 8 0.4 0.03 6.5 0 0.0025 0.2 1.8 1.7 0.3 0.55 0.25 0.25 8 0.4 0.03 6.5 0 0.0025 0.2 1.8 1.7 0.3 0.55 0.25 0.25 8 0.4 0.03 0.05 0.05 0.05 0.05 Table E-2: Power generation conditions of G1, G2 and G3 and G4. Gen 1 2 3 4 Bus type SL PV PV PV Rated power (MVA) 900 900 900 900 Nominal voltage (L-L kV) 20 20 20 20 Active power output 750 700 719 700 Reactive power output 185 235 176 202 Terminal voltage (p.u.) 1.03 1.01 1.03 1.01 Vmax Vref 1 1 + sTr 1 + sTb 1 + sTa Vmeas K 1 + sTe Vmin Figure E-2: Block diagram of static exciter of G1, G2 and G3 and G4. -213- Appendices 1 2 3 4 5 6 7 Table E-3: Parameters of static exciter. Transducer filter time constant Tr 0.01 Voltage regulator gain K 150 Voltage regulator time constant Te 0.05 Transient gain reduction time constant Ta 1 Transient gain reduction time constant Tb 10 Maximum voltage regulator output Vmax 3 Minimum voltage output Vmin -3 second p.u. second Second Second p.u. p.u. Tmax 1.0 K ωrotor Pref 1 1 + sTs 1 + sT3 1 + sTc 1 + sT4 1 + sT5 Tmin Figure E-3: Block diagram of speed governor of G1, G2, G3 and G4. 1 2 3 4 5 6 7 8 Table E-4: Parameters of speed governor. 50 p.u. Governor gain K Servo time constant Ts 0.1 second Transient gain time constant T3 0 Second HP turbine time constant Tc 0.5 Second Time constant to set HP ratio T4 1.25 Second Reheat time constant T5 5 Second Maximum power Pmax 1 p.u. Minimum power Pmin 0 p.u. Table E-5: Transformer parameters. No. of Transformer Rated power (MVA) Rated voltage (HV) Rated voltage (LV) Short circuit voltage (pos. Sequ. %) Short circuit voltage (zero Sequ. %) Winding connection (HV) Winding connection (LV) 1 900 230 20 15 3 YN YN 2 900 230 20 15 3 YN YN 3 900 230 20 15 3 YN YN 4 900 230 20 15 3 YN YN Table E-6: Transmission line parameters. No. of line 1 2 3 4 5 6 Rated voltage (L-L kV) 230 230 230 230 230 230 Length (km) 25 10 110 110 110 110 0.0529 0.0529 0.0529 0.0529 0.0529 0.0529 Resistance (Ohm/km) Reactance (Ohm/km) 0.529 0.529 0.529 0.529 0.529 0.529 3.3075 3.3075 3.3075 3.3075 3.3075 3.3075 Susceptance (µs/km) -214- 7 230 10 8 230 25 0.0529 0.0529 0.529 0.529 3.3075 3.3075 Appendices Table E-7: Load data. Load Active power (MW) Reactive power (MVar) Load 1 967 100 Load 2 1667 100 The active load is modeled as 50% constant and 50% constant impedance. The reactive load is modeled as constant impedance. No. of Capacitor 1 2 Table E-8: Shunt capacitors. Connected Reactive power bus output (MVar) 3 197.11 5 217.53 Nominal voltage (kV) 230 230 Maximum output (MVar) 600 500 Figure E-4: SVC composite (DIgSILENT interface). Vref VSVC Qmax Qfixed k 1 1 + sT f 1 + sTa 1 + sTb Qmin Figure E-5: Primary voltage control of SVC. Table E-9: Parameters of the primary voltage control of SVC. 10 p.u. 1 Proportional gain Kp 3 Lead phase compensation time constant Tlead 0.65 second 0.2 second 4 Lag phase compensation time constant Tlead 6 SVCr time constant Tf 0.03 second -100 Ohm 7 Minimum reactive power output Qmin 400 Ohm 8 Maximum reactive power output Qmax -215- QSVC Appendices Vref VSVC - + - Qmax Qfixed k + 1 + sTa 1 + sTb 1 1 + sT f + Qmin u max Phase angle 1 + kd - 1 1 + sTm sTw 1 + sTw 1 + sTlead 1 + sTlag u min Phase angle 2 Figure E-6: SVC supplementary damping controller. Table E-10: Parameters of the primary voltage control of SVC. 1 Voltage reference for voltage control 1.0 p.u. 2 Damping control gain Kd 1.0 p.u. 3 Low pass filter time constant Tm 0.1 second. 4 Washout time constant Tw 10 second 0.22 5 Lead phase compensation time constant Tlead second 1.38 second 6 Lag phase compensation time constant Tlag -10 p.u. 7 Minimum damping controller output umin 10 p.u. 8 Maximum damping controller output umax -216- QSVC Appendices 10.6 Appendix F Table F-1: DC transmission line parameters. DC transmission line 1 Rated DC voltage (kV) 500 Rated current (kA) 3.6 Length (km) 200 Cable Line type 0.0075 Resistance (Ohm/km) Inductance (mH/km) 1 Figure F-1: Rectifier operation condition (DIgSILENT interface). α max kp 1 1 + sT f I ref ki s I meas α min Figure F-2: Rectifier’s α control for constant current. 1 2 3 4 5 6 Table F-2: Parameters of rectifier control. Proportional gain Kp 0 p.u. Output gain Ko 1 p.u. Rectifier time constant Tf 0.001 second Integral gain Ki 0.01 p.u. 5 Minimum firing angle αmin degree 90 Maximum firing angle αmax degree -217- α Appendices Figure F-3: Inverter operation condition (DIgSILENT interface). β max kp 1 1 + sT f Vref Vmeas ki s β min Figure F-4: inverter’s β control for constant voltage. 1 2 3 4 5 6 Table F-3: Parameters of rectifier control. Proportional gain Kp 0 p.u. Output gain Ko 1 p.u. Rectifier time constant Tf 0.001 second Integral gain Ki 0.01 p.u. 90 Minimum advance firing angle βmin degree Maximum advance firing angle βmax 150 degree -218- β Appendices I ref α max kp 1 1 + sT f I meas ki s α α min umax 1 1 + sTm kd 1 + sTlead 1 + sTlag sTw 1 + sTw 1 + sTlead 1 + sTlag umin Figure F-5: HVDC supplementary damping controller in GB full model. Table F-4: Parameters of the HVDC damping controller. 1 2 3 4 5 6 7 Damping control gain Kd 100 p.u. Low pass filter time constant Tm 0.1 second. Washout time constant Tw 10 second Lead phase compensation time constant Tlead 0.05 second 1.8 second Lag phase compensation time constant Tlag -4 p.u. Minimum damping controller output umin 4 Maximum damping controller output umax p.u. -219- Appendices Table F-5: A list of the PSS in service. No. of PSS PSS name 1 GRMPSS Power gird generator name SPTL BPGR01_1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 NGET NGET NGET NGET NGET NGET NGET NGET NGET NGET NGET NGET NGET NGET NGET NGET NGET NGET NGET NGET NGET NGET NGET NGET NGET NGET NGET NGET NGET ABBPSS ABBPSS ABBPSS ABBPSS ABBPSS ABBPSS ABBPSS RYHPSS RYHPSS RYHPSS RYHPSS RYHPSS RYHPSS ST4PSS ST4PSS ST4PSS ST4PSS ST4PSS ST4PSS ST4PSS ZVTPSS ZVTPSS ZVTPSS ZVTPSS ZVTPSS ZVTPSS ZVTPSS ZVTPSS ZVTPSS -220- DEES8A_1 DEES8B_1 DEES8S_1 KILL8A_1 KILL8B_1 KILL8C_1 KILL8F_1 KILL8D_1 KILL8E_1 RYEH8A_1 RYEH8B_1 RYEH8C_1 RYEH8S_1 BAGB8B_1 SHOS8A_1 SHOT8A_1 SHOT8B_1 SHOT8S_1 WBUR8A_1 WBUR8B_1 LITB8S_1 PEMB8A_1 PEMB8B_1 PEMB8C_1 PEMB8D_1 PEMB8E_1 SEAB8A_1 SEAB8B_1 SEAB8S_1 Appendices 10.7 Appendix G 10.7.1 Published journal papers Paper 1. V. Terzija, D. Cai, V. Stanojevic, G. Strbac, “Frequency and Power Components Estimation from Instantaneous Power Signal,” IEEE Transactions on Instrumentations and Measurements. Vol. 60, no. 1, pp. 3640-3649, June 2011. Paper 2. V. Terzija, G. Valerde, D. Cai, P. Regulski, V. Madani, J. Fitch, S. Skok, Miroslav M. Begovic and A. Phadke, “Wide-Area Monitoring, Protection and Control of Future Electric Power Networks”, IEEE Proceedings, vol. 99, No.1, pp. 80-93, January. 2011. Paper 3. S. Chakrabarti, E. Kyriakides, B. Tianshu; D. Cai, V. Terzija, “Measurements get together,” IEEE Power and Energy Magazine, vol.7, no.1, pp.41-49, January 2009. 10.7.2 Submitted journal papers Paper 4. D. Cai, M. Osborne, V. Terzija, “A Smart Application for Inter-area Oscillations Monitoring based on the Newton-Type Algorithm”, submitted to IEEE transactions on Smart Grid. Paper 5. D. Cai, M. Osborne, V. Terzija, “Wide Area Inter-area Oscillations Monitoring based on fast Nonlinear Estimation Algorithm”, submitted to Special Issue on Planning and Operation of Transmission Grid with Applications to Smart Grid, IEEE Transactions on Smart Grid. 10.7.3 Published conference papers Paper 6. V. Terzija, D. Cai, J. Fitch, “Monitoring of inter-area oscillations in power systems with renewable energy resources using Prony method,” The 20th International Conference and Exhibition on Electricity Distribution - Part 2, CIRED, 2009. Paper 7. V. Terzija, D. Cai, K. Mustafa “Power imbalance estimation in distribution networks with renewable energy resources,” The 20th International Conference and Exhibition on Electricity Distribution - Part 2, CIRED, 2009. Paper 8. G. Valerde, D. Cai, J. Fitch, V. Terzija, “Enhanced state estimation with realtime updated network parameters using SMT,” IEEE Power & Energy Society General Meeting, 2009. -221-