Aziz Frequencyregulation 2017

Frequency regulation in wind integrated power system
AUTHOR(S)
A Aziz
PUBLICATION DATE
01-01-2017
HANDLE
10536/DRO/DU:30103451
Downloaded from Deakin University’s Figshare repository
Deakin University CRICOS Provider Code: 00113B

FREQUENCY REGULATION IN WIND
INTEGRATED POWER SYSTEM
ASMA AZIZ
SUBMITTED IN FULFILMENT OF THE REQUIREMENTS FOR THE

DEGREE OF
DOCTOR OF PHILOSOPHY
SCHOOL OF ENGINEERING
FACULTY OF SCIENCE, ENGINEERING AND BUILT ENVIRONMENT
DEAKIN UNIVERSITY
JAN, 2017
Signature Redacted by Library
Signature Redacted by Library

© Asma Aziz
2017
1
Dedicated to My Beloved Parents
2
ACKNOWLEDGMENTS
I have great pleasure in expressing my deep sense of appreciation and thankfulness
to my principal supervisor Prof. Aman MTO and Co-supervisor Prof. Alex
Stojcevski, for their invaluable guidance, continual encouragement, and support at
every stage of this work. No amount of words can suffice in expressing my sincere
thanks to them. My special gratitude to Prof. Aman MTO for his assistance and
encouragement during this research and providing me a valuable learning experience.
This work has been carried out at the Department of Electrical and Renewable Energy
Engineering, School of Engineering at Deakin University, Australia. The financial
support by Deakin University is gratefully acknowledged.
I would also like to thanks, Kate Hecimovic for providing valuable administrative
support during this candidature and G. M. Shafiullah for providing valuable feedback
during candidature proposal.
Finally, I acknowledge my deep gratitude to my loving mother for her endless prayers
and my late father who have always been a source of inspiration to me. Special
gratitude to my husband as this accomplishment was not possible without his
continuous help, fruitful technical discussions, and personal advice. I am very
thankful to my both dear brothers, my sisters-in-law, parents-in-law and all family
members for their continuous support, love, and encouragement for my academic
endeavor.
Last but not least, special thanks to my seven years old son Zayan who always
comfort and console, never complain or interfere and endured all my inattention
during last three years.
3
PUBLICATIONS
Journals
1. Asma Aziz, Amanullah MTO, Alex Stojcevski. “Full converter based wind
turbine generator system generic modeling: Variations and applicability,”
Elsevier’s Sustainable Energy Technologies and Assessments, Volume 14,
April 2016, Pages 46-62, ISSN 2213-1388. [IF=1.089, Q1, CiteScore=3.06].
2. Aziz, Asma, Shafiullah, G. M., Amanullah MTO, and Alex Stojcevski.
“Automatic Generation Control in Wind Integrated Power System: New
Perspectives & Challenges.” Int. J. of Thermal & Environmental Engineering,
Volume 12, No. 1 (2016) 27-38. [Journal Google Scholar h-index=12, i10-
index=17]
3. Aziz, Asma, Amanullah MTO, and Alex Stojcevski. "Automatic Generation
Control of Multi-Generation Power System." Journal of Power and Energy
Engineering 2.04 (2014): 312. [Journal Google Scholar h-index=8, Journal
Google impact factor 0.72]
4. Aziz, Asma, Shafiullah, G. M., Amanullah MTO, and Alex Stojcevski. “Load
Frequency Control in Wind Integrated Multigenerational Power System.”
Elsevier’s Sustainable Energy Technologies and Assessments (Received
Revision and under process for further submission).
5. Aziz, Asma, Amanullah MTO, and Alex Stojcevski. “LVRT Capability
Assessment of Grid Code Frequency Responsive VSWTG,” International
Journal of Electrical Energy, ISSN: 2301-3656, DOI: 10.18178/ijoee.
(Accepted)
6. Aziz, Asma, Amanullah MTO, and Alex Stojcevski. “Grid Code Frequency
Sensitive Active Power Control of Type 3 and Type 4 Variable Speed Wind
4
Turbine Generator for Improved Frequency Support”. Elsevier’s Sustainable
Energy Technologies and Assessments (Under review)
7. Aziz, Asma, Amanullah MTO, and Alex Stojcevski. “AGC in the presence of
Grid Code Frequency Responsive Wind Power Penetration using Intelligent
FGSPID.” Taylor Francis’s Electric Power Components and Systems. (Under
review)
8. Aziz, Asma, Amanullah MTO, and Alex Stojcevski. “Wind Integration In
Australian NEM: Issues and Mitigations” Elsevier’s Renewable Energy
Focus (Under review).
9. Aziz, Asma, Amanullah MTO, and Alex Stojcevski. “An Analysis of
Frequency Responsive Wind Plants Integration on AGC of Multigenerational
System” (To be Submitted)
Conference
1. Aziz, Asma, Shafiullah, G. M., Amanullah MTO, and Alex Stojcevski.
“Automatic Generation Control of Wind Energy System,” Proceedings of the
9th Annual International Green Energy Conference (IGEC-IX). 25th–28th
May 2014, Tianjin, China.
2. Aziz, Asma, Amanullah MTO, and Alex Stojcevski. "Multi-Generational
Power System Load Frequency Secondary Control,” 6th Asia-Pacific Power
and Energy Engineering Conference (APPEEC 2014), April 14 to 16, 2014,
Shanghai, China.
3. Aziz, Asma, Shafiullah, G.M., Amanullah MTO, and Alex Stojcevski.
“Participation of DFIG Based Wind Energy System in Load Frequency
Control of Interconnected Multigenerational Power System,” IEEE
5
Australasian Universities Power Engineering Conference, (AUPEC) 2014,
Perth, Australia.
4. Aziz, Asma Arangarajan Vinayagam, KSV Swarna, Suiyang Khoo and Alex
Stojcevski. “Power Quality Impacts in a Typical Microgrid,” International
Conference on Sustainable Energy and Environmental Engineering (SEEE
2015), Oct. 2015. DOI: 10.2991/seee-15.2015.20
5. Aziz, Asma, Arangarajan Vinayagam, Amanullah MTO, and Alex Stojcevski.
“Modelling and Comparison of Generic Type 4 WTG with EMT Type 4 WTG
Model”, 12th IEEE India International Conference, 2015 (INDICON 2015),
Delhi, India.
6. Aziz, Asma, Amanullah MTO, and Alex Stojcevski. “Modelling and Analysis
of Type 4 Wind Turbine Generator System for Utilization in Frequency
Regulation Studies”, IEEE Australasian Universities Power Engineering
Conference, (AUPEC) 2015, Wollongong, Australia.
7. Aziz, Asma, Amanullah MTO, and Alex Stojcevski. “Fuzzy Gain Scheduling
of Load Frequency Controller in the Presence of Frequency Responsive Wind
Power Penetration,” IEEE Australasian Universities Power Engineering
Conference, (AUPEC) 2016, Brisbane, Australia.
8. Aziz, Asma, Amanullah MTO, and Alex Stojcevski. “LVRT Capability
Assessment of Grid Code Frequency Responsive TYPE 3 VSWTG”, 9th
International Conference on Computer and Electrical Engineering (ICCEE
2016), Barcelona, Spain.
Awards
1. Best Paper Presenter Award-IEEE INDICON 2015
2. Travel Award, University of Queensland-AUPEC2016
6
TABLE OF CONTENTS
ACKNOWLEDGMENTS .................................................................................................... 3
PUBLICATIONS ................................................................................................................. 4
TABLE OF CONTENTS ..................................................................................................... 7
LIST OF TABLES ............................................................................................................ 12
LIST OF FIGURES ........................................................................................................... 14
ACRONYMS ...................................................................................................................... 22
NOMENCLATURE .......................................................................................................... 24
EXECUTIVE SUMMARY ................................................................................................ 30
1 CHAPTER 1 INTRODUCTION ............................................................................ 32
1.1 Current Status of Wind Energy System .............................................................................. 32
1.2. Motivations for Research ............................................................................................. 34
1.3. Objectives .................................................................................................................... 39
1.4. Significance .................................................................................................................. 42
1.5. Methodology ............................................................................................................... 43
1.6. Contributions ............................................................................................................... 45
1.6.1. Frequency Response Investigation in NEM Control Area ................................................. 46
1.6.2. Modeling & Performance Assessment of Type 4 Frequency Responsive Wind Turbine ... 47
1.6.3. Modeling and Investigation of Grid Code Compatible Active‐Power Controller for Type 3
& Type 4 VSWTGs .......................................................................................................................... 47
1.6.4. Development of Frequency Response Model for Wind Energy Impact Assessment on
Reserve Response .......................................................................................................................... 47
1.6.5. Integration Impact of Grid Code Frequency Responsive VSWTGs on AGC of
Multigenerational System .............................................................................................................. 48
1.6.6. Development of FGSPID Controller for Wind Integrated Power System Load Frequency
Control ........................................................................................................................................... 48
1.6.7. Fault Ride‐through Analysis of Frequency Responsive VSWTGs ....................................... 48
1.6.8. Review of Active Power‐Frequency Control in Wind Integrated Power System............... 48
1.7. Outline of the Thesis .................................................................................................... 49
1.8. References ................................................................................................................... 51
7
2 CHAPTER 2
WIND INTEGRATION IN AUSTRALIAN NEM ........................................................ 53
2.1. National Electricity Market (NEM) Overview ................................................................ 53
2.2. NEM Spot Market & AEMO .......................................................................................... 54
2.3. DRIVERS OF WIND ENERGY SYSTEM DEVELOPMENT IN AUSTRALIA .............................. 58
2.3.1. Renewable Energy Target (RET) ......................................................................................... 58
2.3.2. Wind Resources Availability ............................................................................................... 59
2.3.3. Cost Effective, Competitive & Established Technology ...................................................... 60
2.3.4. Low Carbon Footprint ........................................................................................................ 62
2.4. Wind Energy and Australian Market Outlook ................................................................ 63
2.5. Challenges to NEM Operation Due to Increased Wind Penetration ............................... 67
2.5.1. Merit Order Effect and Reduced Spot Prices ..................................................................... 67
2.5.2. Reduced Inertia and High Rate of Change of Frequency Issues ......................................... 70
2.5.3. FCAS Issues ......................................................................................................................... 72
2.5.4. Network Infrastructure Issues............................................................................................ 74
2.5.5. Interconnectors Performance Issues ................................................................................. 75
2.6. Wind Integrated Network Frequency Response Assessment ......................................... 77
2.6.1. Test Network ...................................................................................................................... 77
2.6.2. Low Load Scenario ............................................................................................................... 78
2.6.3. Assumptions Used in simulations ...................................................................................... 78
2.6.4. Simulation Discussions ....................................................................................................... 84
2.6.5. Linear Regression Model from Frequency Disturbances ................................................... 89
2.6.6. High Load Scenario ............................................................................................................. 93
2.7. Possible Frequency Control Solutions ........................................................................... 97
2.8. Conclusion ................................................................................................................. 105
2.9. References ................................................................................................................. 108
CHAPTER 3 POWER SYSTEM FREQUENCY CONTROL .................................... 113
3.1 Introduction ................................................................................................................... 113
3.2. Power System Inertia ...................................................................................................... 117
3.2.1. Power system Inertial Response ...................................................................................... 119
3.3. Primary Frequency Control .............................................................................................. 120
3.4. Secondary Frequency Control .......................................................................................... 124
3.4.1. Area Control Error ............................................................................................................ 125
3.5. Tertiary control ............................................................................................................... 129
3.6. Wind Energy System Participation in Frequency‐Power Control .................................. 129
8
3.6.1. Synthetic Inertial Control ................................................................................................. 130
3.6.2. Primary Control from VSWTG .......................................................................................... 132
3.6.3. AGC & Wind Energy ......................................................................................................... 138
3.6.4. AGC & Wind Energy Research: New Perspectives and Challenges .................................. 144
3.7. Conclusion ...................................................................................................................... 152
3.8. References ........................................................................................................................ 154
CHAPTER 4 FREQUENCY RESPONSIVE FULL CONVERTER BASED VSWTG:

MODELLING AND ASSESSMENT ............................................................................. 164
4.1. Introduction ............................................................................................................... 164
4.2. Type 4 VSWT Model with Synchronous Generator ...................................................... 166
4.2.1. Converter Topology ......................................................................................................... 167
4.2.2. Synchronous Generator Model ........................................................................................ 169
4.3. WECC’s Type 4 WTG Generic Model ........................................................................... 171
4.4. Proprietary Type 4 Generic Model .............................................................................. 172
4.5. Proposed Type 4 VSWTG Generic Model .................................................................... 174
4.5.1. Wind Turbine Mechanical Model ..................................................................................... 176
4.5.2. Pitch Control Model ......................................................................................................... 179
4.5.3. Electrical Control Model................................................................................................... 181
4.5.3. Active Power Control Loop Co‐ordination with Plant Controller ..................................... 183
4.6. Model Validation and Analysis ................................................................................... 184
4.6.1. Aerodynamic Model Suitability Assessment .................................................................... 185
4.6.2. Generator‐Converter Model Comparison with GE’s Generic Model and Detailed Model
..................................................................................................................................................... 195
4.6.3. Drive Train Mass Model Comparison with GE’s Generic Model & Detailed Simulation
Model ........................................................................................................................................... 197
4.6.4. Virtual Inertia Support Assessment ................................................................................. 198
4.7. Conclusion ..................................................................................................................... 203
4.7. References ................................................................................................................. 204
CHAPTER 5 DYNAMIC DEAD‐BAND DEPENDENT GRID CODE FREQUENCY

SENSITIVE ACTIVE POWER CONTROLLER ......................................................... 207
5.1. Introduction ............................................................................................................... 207
5.2. Grid Code Sensitive Frequency Controller ................................................................... 209
5.2.1. Dead‐Band: Static vs. Dynamic ........................................................................................ 209
5.2.3. Frequency Processing Model ........................................................................................... 210
5.3. Frequency Response Controller Model‐I ..................................................................... 215
9
5.4. Frequency Response Controller Model‐II .................................................................... 220
5.5. Conclusion ................................................................................................................. 223
5.6. References ................................................................................................................. 224
CHAPTER 6 GRID CODE FREQUENCY SENSITIVE VSWTG: IMPACT

ASSESSMENT ................................................................................................................ 225
6.1. Introduction ............................................................................................................... 225
6.1.1. Type 3 VSWTG Model ...................................................................................................... 226
6.1.2. Applied Commanded Reserve Power Mode ................................................................... 230
6.1.3. Wind Profile and Available Power Model ....................................................................... 232
6.2. Simulation Setup ............................................................................................................. 233
6.2.1. Simulation Assumptions ................................................................................................... 235
6.3. Discussion ......................................................................................................................... 237
6.3.1. Case I: Type 3 VSWTG ...................................................................................................... 237
6.3.2. Case II: Type 4 VSWTG ..................................................................................................... 246
6.4. Conclusion ..................................................................................................................... 253
6.4. References ...................................................................................................................... 255
CHAPTER 7 AUTOMATIC GENERATION CONTROL IN FREQUENCY

RESPONSIVE WIND PLANT INTEGRATED POWER SYSTEM ......................... 257
7.1. AGC / Load Frequency Controllers and Conventional Generation Models .................... 257
7.1.1. Classical AGC Controller ................................................................................................... 258
7.1.2. Intelligent AGC Controller ................................................................................................ 260
7.1.3. Hydro Turbine Governor Model ...................................................................................... 263
7.1.4. Gas Power Plant Model .................................................................................................... 268
7.1.5. Thermal Power Plant ........................................................................................................ 278
7.2. Performance Index ..................................................................................................... 279
7.3. Parameter Optimization Algorithm ............................................................................ 281
7.4. Control Area Simulation Models ................................................................................. 283
7.4.1. A Frequency Regulation Model Incorporating All Three Control Modes in a Single Area
Power System .............................................................................................................................. 283
7.4.2. Decentralized Three Area Interconnected Power System (TAIPS) Model ....................... 287
7.4.3. Test System for LVRT Capability Assessment ................................................................... 291
7.5. Analysis‐I ........................................................................................................................ 292
7.6. Analysis‐II ....................................................................................................................... 306
7.7. Analysis‐III ...................................................................................................................... 327
10
7.8. Analysis‐IV ...................................................................................................................... 338
7.9. References ................................................................................................................. 352
CHAPTER 8 CONCLUSION AND RECOMMENDATIONS .................................... 358
8.1. Conclusion ...................................................................................................................... 358
8.1.1. Type 4 Generic VSWTG Model for Frequency Response Studies ..................................... 359
8.1.2. Grid Code Sensitive, Frequency Processor Dependent Frequency Controller ................. 360
8.1.3. Wind Participation in Classical and Intelligent AGC in Multi‐Generational System ......... 361
8.1.4. Challenges and Possible Solution in Wind Integrated Network ....................................... 363
8.2. Recommendations .......................................................................................................... 365
APPENDIX ..................................................................................................................... 366
11
LIST OF TABLES
TABLE 2‐1 COMPARISON OF RENEWABLE ENERGY TARGET FOR SOME COUNTRIES WITH AUSTRALIA
.................................................................................................................................................. 59
TABLE 2‐2 SCENARIO 1: LOW LOAD .................................................................................................... 78
TABLE 2‐3 LOAD FLOW IN LOW LOAD SCENARIO ............................................................................... 80
TABLE 2‐4 AREA FREQUENCY RESPONSE INDICATORS WHEN GENERATOR 503 (150 MW) IS
ISOLATED AT 10S AND WIND FARM CONNECTED AT BUS 509 ................................................. 80
ISOLATED AT THE 20S, AND GENERATOR 501 (560 MW) IS ISOLATED AT 10S, 1146MW WIND
FARM CONNECTED AT BUS 508 ................................................................................................ 82
ISOLATED AT 10S AND AREA 3‐5 INTERCONNECTOR IS ISOLATED AT 20 S; WIND FARM
CONNECTED AT BUS 508 ........................................................................................................... 83
ISOLATED AT 10S AND AREA 3‐5 INTERCONNECTOR IS ISOLATED AT 20 S, 1080 MW WIND
FARM CONNECTED AT BUS 508 ................................................................................................ 84
TABLE 2‐10 SIMULATION RESULTS FOR LOW LOAD SCENARIO FOR AREA 5 GENERATOR 501 .......... 87
TABLE 2‐11 SIMULATION RESULTS FOR LOW LOAD SCENARIO FOR AREA 5 GENERATOR 502 .......... 88
TABLE 2‐12 VALUES FOR LINEAR POLYNOMIAL REGRESSION MODEL ............................................... 90
TABLE 2‐13 SCENARIO 1: HEAVY LOAD ............................................................................................... 94
TABLE 2‐14 MODIFIED BUS FOR HIGH LOAD SCENARIO ..................................................................... 94
TABLE 2‐15 GENERATION OUTPUT IN HIGH LOAD SCENARIO ............................................................ 95
TABLE 2‐16 AREA FREQUENCY RESPONSE INDICATORS WHEN WIND FARM IS ADDED (109X4 MW)96
TABLE 2‐17 AREA FREQUENCY RESPONSE INDICATORS WHEN GEN 503 IS ISOLATED AT 10 S AND
WIND FARM IS ADDED (436 MW) ............................................................................................. 96
TABLE 2‐18 AREA 5 FREQUENCY RESPONSE INDICATORS WHEN GEN 503 IS ISOLATED AT 5 S, AND
INTERCONNECTOR IS ISOLATED AT 10 S, WIND FARM ADDED (436 MW) ............................... 97
TABLE 2‐19 ANTICIPATED TECHNICAL SOLUTIONS FOR FREQUENCY IMPROVEMENT IN WIND
PENETRATED SYSTEM ............................................................................................................... 98
TABLE 2‐20 AREA 5 FREQUENCY RESPONSE INDICATORS DURING GENERATOR CONTINGENCY
(WITHOUT SYNCHRONOUS CONDENSER) ............................................................................... 101
TABLE 2‐21 AREA 5 FREQUENCY RESPONSE INDICATORS DURING GENERATOR CONTINGENCY (WITH
ONE SYNCHRONOUS CONDENSER) ......................................................................................... 101
TABLE 2‐22 AREA 5 FREQUENCY RESPONSE INDICATORS DURING GENERATOR CONTINGENCY (WITH
BOTH SYNCHRONOUS CONDENSERS) ..................................................................................... 101
TABLE 3‐1 BASIC FREQUENCY REGULATION SERVICES ..................................................................... 115
TABLE 3‐2 REGULATION SERVICES NOMENCLATURE ADOPTED BY VARIOUS COUNTRIES .............. 116
TABLE 4‐1 COMPARISON OF TYPE 4 WTG MODELS SUITABLE FOR FREQUENCY REGULATION
ANALYSIS ................................................................................................................................. 175
TABLE 6‐1 COMPARISON OF SIMULATION RESULTS TO SHOW THE EFFECT OF TORQUE CONTROLLER
RE‐TUNING AND DIFFERENT DROOP VALUES FOR 10% WIND ENERGY PENETRATION ......... 236
TABLE 6‐2 COMPARISON OF FREQUENCY RESPONSE SIMULATION RESULTS FOR TYPE 4 MODELS
UNDER 10% WIND ENERGY PENETRATION IN A CONTROLLED AREA ..................................... 252
TABLE 7‐1 SIMULATION STUDY SET UP CASES ................................................................................. 293
TABLE 7‐2 OPTIMUM VALUES OF AGC INTEGRAL CONTROLLER ...................................................... 309
12
TABLE 7‐3 SYSTEM PARAMETERS AND SYSTEM FREQUENCY RESPONSE INDICATORS OBTAINED
WITHOUT WIND INTEGRATION .............................................................................................. 309
TABLE 7‐4 WIND PLANT INTEGRATED SYSTEM DATA ....................................................................... 312
TABLE 7‐5 FACTORS AFFECTING FUZZY PID TUNING ........................................................................ 330
TABLE 7‐6 FUZZY TUNING RULES FOR KP’ ........................................................................................ 331
TABLE 7‐7 FUZZY TUNING RULES FOR KD’ ........................................................................................ 331
TABLE 7‐8 FUZZY TUNING RULES FOR ∝ ........................................................................................... 331
TABLE 7‐9 SECONDARY FREQUENCY CONTROLLER COMPARISON WITH PERFORMANCE INDEX FOR
AREA 1 UNDER 10% WIND PENETRATION AND 0.1 PER UNIT LOAD DISTURBANCE IN ALL
THREE AREAS........................................................................................................................... 336
TABLE 7‐10 TEST RESPONSE ............................................................................................................. 344
13
LIST OF FIGURES
FIGURE 1‐1 CUMULATIVE FORECAST BY REGION 2012‐2017 [2] ....................................................... 33
FIGURE 1‐2 STAKEHOLDERS VIEW’S ON FREQUENCY RESPONSIVE CAPABILITY AS A REQUIREMENT
FOR WIND ENERGY SYSTEM ...................................................................................................... 39
FIGURE 1‐3 RESEARCH METHODOLOGY ............................................................................................. 45
FIGURE 1‐4 CONTRIBUTIONS OF THE STUDY ...................................................................................... 46
FIGURE 2‐1 NEM REGIONAL BOUNDARIES, THEIR INTERCONNECTIONS AND GENERATION CAPACITY
[DATA DERIVED FROM [1], FORMULATED BY AUTHOR] ........................................................... 54
FIGURE 2‐2 AEMO INTERACTIONS WITH OTHER REGULATORS [FORMULATED BY AUTHOR] ........... 57
FIGURE 2‐3 AUSTRALIA WIND RESOURCES [16] ................................................................................. 59
FIGURE 2‐4 LEVELISED COST OF ENERGY (LCOE) GENERATION AND CAPACITY FACTOR FOR WIND
GENERATORS [17] ..................................................................................................................... 61
FIGURE 2‐5 LCOE COST COMPARISON WITH FUEL TECHNOLOGIES [22] ........................................... 62
FIGURE 2‐6 PER CAPITA WIND CAPACITY BY COUNTRY AND THE KEY US AND AUSTRALIAN
STATES/TERRITORIES [23] ......................................................................................................... 63
FIGURE 2‐7 WIND GENERATION CAPACITY SUMMARY FOR AUSTRALIA 2015 ................................... 64
FIGURE 2‐8 PERCENTAGE OF WIND CAPACITY BY STATE [MODIFIED FROM 26] ............................... 65
FIGURE 2‐9 SOUTH AUSTRALIA WIND POWER VS. DEMAND VARIATION ON 22 MAY [28] ............... 66
FIGURE 2‐10 MERIT ORDER EFFECT [33] ............................................................................................ 69
FIGURE 2‐11 MARKET PRICE VARIATIONS IN SOUTH AUSTRALIA ON 3‐4 NOV. 2014 WHEN WIND
GENERATION ENTERS SPOT MARKET [33] ................................................................................ 69
FIGURE 2‐12 14‐GENERATOR NEM MODEL ........................................................................................ 79
FIGURE 2‐13 AREA 5 GENERATOR TERMINAL FREQUENCY AFTER ISOLATION OF GEN 503 AT 10 S . 85
FIGURE 2‐14 AREA 5 GENERATOR TERMINAL FREQUENCY AFTER ISOLATION OF GEN 503 AT 10 S
AND INTERCONNECTOR ISOLATION AT 20 S ............................................................................. 86
FIGURE 2‐15 AREA 5 BUS FREQUENCY AFTER ISOLATION OF GEN 503 AT 10 S. ................................ 86
FIGURE 2‐16 AREA 5 BUS FREQUENCY AFTER ISOLATION OF GEN 502 AT 10 S. ................................ 86
FIGURE 2‐17 ROCOF OBSERVED AT GENERATOR 503 WHEN GENERATOR 502 IS ISOLATED AT 10 S 87
FIGURE 2‐18 POLYNOMIAL SURFACE VIEW FOR WIND PENETRATION VARIATION WITH FREQUENCY
NADIR POINT ............................................................................................................................. 91
FIGURE 2‐19 POLYNOMIAL SURFACE VIEW FOR MW‐LOSS VARIATION WITH FREQUENCY NADIR
POINT AND ROCOF AS OBSERVED AT GENERATOR 501 ........................................................... 91
FIGURE 2‐20 LINEAR POLYNOMIAL REGRESSION SURFACE VIEW FOR WIND PENETRATION
VARIATION WITH FREQUENCY NADIR POINT AND ROCOF AS OBSERVED AT GENERATOR 502
.................................................................................................................................................. 92
FIGURE 2‐21 LINEAR INTERPOLANT SURFACE VIEW FOR MW‐LOSS VARIATION WITH FREQUENCY
NADIR POINT AND ROCOF AS OBSERVED AT GENERATOR 502 ................................................ 92
FIGURE 2‐22 POLYNOMIAL REGRESSION SURFACE VIEW FOR ONLINE SYNCHRONOUS GENERATOR
UNITS VARIATION WITH FREQUENCY NADIR POINT AND ROCOF AS OBSERVED AT
GENERATOR 502 ....................................................................................................................... 92
FIGURE 2‐23 LINEAR POLYNOMIAL SURFACE VIEW FOR ONLINE SYNCHRONOUS GENERATOR UNIT
VARIATION WITH FREQUENCY NADIR POINT AND ROCOF AS OBSERVED AT GENERATOR 501
.................................................................................................................................................. 93
FIGURE 2‐24 LINEAR POLYNOMIAL SURFACE VIEW FOR WIND PENETRATION WITH FREQUENCY
NADIR POINT AND ROCOF AS OBSERVED AT GENERATOR 502 ................................................ 93
FIGURE 2‐25 ROCOF COMPARISON WITH/WITHOUT SYNCHRONOUS CONDENSERS ...................... 102
FIGURE 2‐26 FREQUENCY NADIR COMPARISON WITH/WITHOUT SYNCHRONOUS CONDENSERS .. 102
FIGURE 2‐27 BUSES FREQUENCY DURING GENERATOR CONTINGENCY AND NO SYNCHRONOUS
CONDENSER ADDITION ........................................................................................................... 103
14
FIGURE 2‐28 BUSES FREQUENCY DURING GENERATOR CONTINGENCY AND EQUIVALENT
SYNCHRONOUS CONDENSER ADDED ...................................................................................... 103
FIGURE 2‐29 FREQUENCY RESPONSIVE WIND TURBINE MODEL...................................................... 104
FIGURE 2‐30 FREQUENCY OBSERVED AT BUS 506 WHEN NORMAL OPERATED DFIG BASED WIND
FARM IS ADDED ...................................................................................................................... 104
FIGURE 2‐31 FREQUENCY OBSERVED AT BUS 506 WHEN FREQUENCY RESPONSIVE DFIG BASED
WIND FARM IS ADDED ............................................................................................................ 105
FIGURE 3‐1 FREQUENCY VARIATION ALONG WITH REGULATION SERVICES .................................... 114
FIGURE 3‐2 FREQUENCY CONTROL DURING CONTINGENCY EVENT [REFORMULATED FROM 4] .... 115
FIGURE 3‐3 INERTIA CONSTANT VALUES OF VARIOUS PLANTS, DATA DERIVED FROM [70] [45] .... 118
FIGURE 3‐4 FREQUENCY POWER CHARACTERISTICS THROUGH DROOP CONTROL ......................... 121
FIGURE 3‐5 PRIMARY FREQUENCY RESPONSE OF UNCONTROLLED THERMAL POWER PLANT WITH
DIFFERENT DROOP .................................................................................................................. 122
FIGURE 3‐6 POWER OUTPUT CHANGE OF UNCONTROLLED THERMAL POWER PLANT UNDER
DIFFERENT DROOP .................................................................................................................. 122
FIGURE 3‐7 SINGLE AREA PRIMARY FREQUENCY CONTROL MODEL WITH FREQUENCY DEPENDENT
LOADS [20] .............................................................................................................................. 123
FIGURE 3‐8 BASIC LOOP OF AGC CONTROL ...................................................................................... 127
FIGURE 3‐9 AGC CONTROLLED THERMAL HYDRO THERMAL POWER PLANT BASED THREE AREA
SYSTEM FREQUENCY RESPONSE UNDER 1% LOAD DISTURBANCE ......................................... 128
FIGURE 3‐10 AGC RESPONSE OF SINGLE AREA THERMAL POWER PLANT UNDER 1% LOAD
DISTURBANCE ......................................................................................................................... 128
FIGURE 3‐11 ACTIVE POWER FREQUENCY CONTROL RESEARCH STUDY LINE IN WIND BASED POWER
PLANTS .................................................................................................................................... 130
FIGURE 3‐12 VIRTUAL INERTIA MODEL ............................................................................................ 130
FIGURE 3‐13 INERTIA AND DROOP CONTROL MODEL FOR FREQUENCY RESPONSE FROM VSWTG 132
FIGURE 3‐14 REPRESENTATION OF OPTIMUM AND DELOADED CURVE FOR VSWTG ...................... 133
FIGURE 3‐15 CONTROL LOOP FOR DE‐LOADING CONTROL OF VSWTG ........................................... 134
FIGURE 3‐16 PITCH DROOP DE‐LOADING CONTROL ........................................................................ 135
FIGURE 3‐17 COEFFICIENT OF PERFORMANCE DE‐LOADING THROUGH PITCH CONTROL .............. 135
FIGURE 3‐18 PITCH ANGLE OFFSET VARIATION WITH DE‐LOADED POWER RESERVE IN VSWTG .... 135
FIGURE 3‐19 IMPACT ON SYSTEM FREQUENCY VARIATION WITH PITCH DROOP DE‐LOADED VSWTG
................................................................................................................................................ 136
FIGURE 3‐20 WIND ENERGY STUDY LINE FOR AGC .......................................................................... 139
FIGURE 3‐21 BASIC SCHEMATIC DIAGRAM FOR WIND FARM FREQUENCY CONTROL ..................... 140
FIGURE 3‐22 VSWTG BASED WIND FARM CONTROLLER SUPPORTING AGC [58] ............................. 140
FIGURE 4‐1 SCHEMATIC DIAGRAM OF FULL CONVERTER BASED VSWTG ........................................ 165
FIGURE 4‐2 DETAILED SIMULATION MODEL OF TYPE 4 VSWTG ....................................................... 166
FIGURE 4‐3 DC/DC BOOST CONVERTER & DECOUPLED VOC INVERTER INTERFACED SYNCHRONOUS
GENERATOR WIND ENERGY SYSTEM. ..................................................................................... 168
FIGURE 4‐4 SIMULATION FRAMEWORK OF SYNCHRONOUS MACHINE IN ROTOR REFERENCE‐FRAME
[15] .......................................................................................................................................... 169
FIGURE 4‐5 WECC TYPE 4 WIND TURBINE GENERATOR MODEL [3] ................................................. 172
FIGURE 4‐6 GE’S TYPE 4 WIND TURBINE GENERATOR GENERIC MODEL (REFORMULATED [5]) ...... 173
FIGURE 4‐7 PROPOSED TYPE 4 WTG MODEL .................................................................................... 174
FIGURE 4‐8 SIMPLIFIED AERODYNAMIC SIMULATION MODEL ........................................................ 178
FIGURE 4‐9 TURBINE POWER CHARACTERISTICS ............................................................................. 179
FIGURE 4‐10 PITCH CONTROLLER BLOCK ......................................................................................... 179
FIGURE 4‐11 GENERATOR‐CONVERTER MODEL [20] ........................................................................ 182
FIGURE 4‐12 ACTIVE POWER CONTROL ............................................................................................ 184
15
FIGURE 4‐13 TYPE 4 GENERIC MODEL INCORPORATING LINEAR AERODYNAMIC MODEL
PERFORMANCE CHARACTERISTICS AT CONSTANT WIND SPEED AT 8 M/S, 13 M/S, 18 M/S
AND 20 M/S ALONG WITH VARYING POWER SETPOINT ......................................................... 188
FIGURE 4‐14 TYPE 4 GENERIC MODEL INCORPORATING LINEAR AERODYNAMIC MODEL
PERFORMANCE CHARACTERISTICS UNDER VARYING WIND SPEED & VARYING POWER
SETPOINT ................................................................................................................................ 188
FIGURE 4‐15 COEFFICIENT OF POWER CALCULATION MODEL FOR WTG SYSTEM ........................... 189
FIGURE 4‐16 COEFFICIENT OF PERFORMANCE MODEL 2 FOR WTG SYSTEM [23] ........................... 189
FIGURE 4‐17 TYPE 4 GENERIC MODEL INCORPORATING DETAILED AERODYNAMIC MODEL 1
PERFORMANCE CHARACTERISTICS AT WIND SPEED OF 11 M/S, 13 M/S, 18 M/S AND 25 M/S
ALONG WITH VARIABLE POWER SETPOINT ............................................................................ 191
FIGURE 4‐18 TYPE 4 GENERIC MODEL INCORPORATING DETAILED AERODYNAMIC MODELS AT
DIFFERENT CONSTANT WIND SPEED ALONG WITH CONSTANT 0.8 POWER SETPOINT ......... 193
FIGURE 4‐19 TYPE 4 GENERIC MODEL INCORPORATING DETAILED AERODYNAMIC MODEL 1
PERFORMANCE CHARACTERISTICS ......................................................................................... 193
FIGURE 4‐20 TYPE 4 MODEL CHARACTERISTICS WITH DETAILED AERODYNAMIC MODEL 2
PERFORMANCE CHARACTERISTICS ......................................................................................... 194
FIGURE 4‐21 GE’S TYPE 4 GENERIC MODEL PERFORMANCE CHARACTERISTICS AT DIFFERENT WIND
SPEED RANGING FROM 8 M/S – 20 M/S ................................................................................. 196
FIGURE 4‐22 TYPE 4 DETAILED EMT MODEL INCORPORATING SYNCHRONOUS GENERATOR
CHARACTERISTICS AT CONSTANT WIND SPEED OF 13 M/S &18 M/S. .................................... 197
FIGURE 4‐23 UNCONTROLLED THREE AREA INTERCONNECTED SYSTEM FREQUENCY DURING 10%
LOAD CHANGE ......................................................................................................................... 199
FIGURE 4‐24 GE MODEL PERFORMANCE WITH/WITHOUT INERTIA SUPPORT ................................ 201
FIGURE 4‐25 TYPE‐4 MODEL PERFORMANCE WITH/WITHOUT INERTIA SUPPORT .......................... 203
FIGURE 5‐1 NEM MAINLAND FREQUENCY 1045 HRS. TO 1109 HRS. DURING CONTINGENCY EVENT
[1] ............................................................................................................................................ 210
FIGURE 5‐2 BASIC STRUCTURE OF GRID FREQUENCY PROCESSOR .................................................. 211
FIGURE 5‐3 AN EXAMPLE OF INPUT FREQUENCY OBTAINED FROM NEM‐14 GENERATOR MODEL
PROVIDED AS INPUT TO FREQUENCY PROCESSOR ................................................................. 213
FIGURE 5‐4 NEM‐14 GENERATOR MODEL PROCESSED FREQUENCY OBTAINED FROM GRID
FREQUENCY PROCESSOR ........................................................................................................ 213
FIGURE 5‐5 MEASURED FREQUENCY VS. GRID PROCESSOR OUTPUT FREQUENCY AT DIFFERENT
THRESHOLD VALUES ............................................................................................................... 214
FIGURE 5‐6 MEASURED FREQUENCY VS. GRID PROCESSOR OUTPUT FREQUENCY AT DIFFERENT
THRESHOLD VALUES (ZOOMED) ............................................................................................. 214
FIGURE 5‐7 PROCESSED FREQUENCY, TREND FILTER OUTPUT, LOW PASS FREQ. FILTER OUTPUT AND
DYNAMIC DEAD‐BAND OUTPUT ............................................................................................. 214
FIGURE 5‐8 PROCESSED FREQUENCY, TREND FILTER OUTPUT, LOW PASS FREQ. FILTER OUTPUT AND
DYNAMIC DEAD‐BAND OUTPUT WHEN THRESHOLD APPLIED IS >0.06 (ZOOMED) ............... 215
FIGURE 5‐9 A SCHEMATIC REPRESENTATION OF MASTER‐SLAVE WIND TURBINE CONTROLLER
INCORPORATING FREQUENCY RESPONSE CONTROLLER MODEL ........................................... 216
FIGURE 5‐10 FREQUENCY RESPONSE CONTROLLER MODEL‐I .......................................................... 217
FIGURE 5‐11 FREQUENCY GRID CODE UK [7] ................................................................................... 219
FIGURE 5‐12 ACTIVE POWER SETPOINT GENERATED WHEN EMPLOYING FREQUENCY RESPONSE
CONTROLLER‐I IN TWO DIFFERENT MODE ............................................................................. 219
FIGURE 5‐13 SYSTEM FREQUENCY USED TO DEPICT FREQUENCY RESPONSE CONTROLLER‐I
OPERATION IN TWO DIFFERENT MODE .................................................................................. 220
FIGURE 5‐14 FREQUENCY RESPONSE CONTROLLER MODEL‐II ......................................................... 220
FIGURE 5‐15 NON‐SYMMETRICAL DROOP CURVE FOR FREQUENCY POWER REGULATION ............ 221
16
FIGURE 5‐16 WTG POWER SETPOINTS WITH LIMITED POWER SETPOINT AS GRID CONTROLLER
OUTPUT SETPOINT AND CORRESPONDING GENERATED WTG POWER .................................. 222
FIGURE 6‐1 BASIC COMPONENTS OF GENERIC VSWTG MODEL ....................................................... 227
FIGURE 6‐2 GENERIC TYPE 3 VSWTG BASIC MODEL WITH AUXILIARY VIRTUAL INERTIA MODEL .... 228
FIGURE 6‐3 TYPE 3 WTG MODEL WITH ACTIVE POWER CONTROLLER FEEDING POWER SETPOINT
PAPC TO POWER CONTROLLER LOOP, PITCH CONTROLLER LOOP AND SPEED CONTROLLER
LOOP ....................................................................................................................................... 229
FIGURE 6‐4 CHARACTERISTICS CURVE OF TYPE 3 VSWTG UNDER MPPT AND RESERVE POWER MODE
................................................................................................................................................ 231
FIGURE 6‐5 WIND AND AVAILABLE POWER SIMULATION MODEL ................................................... 232
FIGURE 6‐6 GENERATED WIND SPEED AND AVAILABLE POWER ...................................................... 233
FIGURE 6‐7 SINGLE AREA HYDRO GOVERNOR TURBINE MODEL WITH INTEGRATED WIND FARM . 234
FIGURE 6‐8 UNCONTROLLED SINGLE AREA MODEL FREQUENCY DEVIATION FOR 0.1 P.U. LOAD
DISTURBANCE ......................................................................................................................... 238
FIGURE 6‐9 FREQUENCY RESPONSE OF SINGLE AREA LFC CONTROL MODEL FOR 0.1 P.U. LOAD
DISTURBANCE, REQ=0.035 FOR ALL CASES, TORQUE CONTROLLER GAINS (3, 0.6), NO RE‐
TUNING FOR TORQUE CONTROLLER ....................................................................................... 238
DISTURBANCE AND REQ =0.035 IN ALL CASES, TORQUE CONTROLLER RETUNED FOR OTHER
THREE DROOP BASED VSWTG MODELS .................................................................................. 239
DISTURBANCE, REQ =0.0556 IN CASE OF FREQUENCY CONTROLLER‐I AND II,
TORQUE CONTROLLER RETUNED ............................................................................................ 239
FIGURE 6‐12 RATE OF CHANGE OF FREQUENCY COMPARISON FOR DIFFERENT LOAD CHANGE (P.U.)
AND 10% VSWTG PENETRATION. ........................................................................................... 240
FIGURE 6‐13 FREQUENCY NADIR COMPARISON FOR DIFFERENT LOAD CHANGE (P.U.) AND 10%
VSWTG PENETRATION ............................................................................................................ 241
FIGURE 6‐14 ELECTRICAL POWER FROM TYPE 3 VSWTG WITH RETUNED TORQUE CONTROLLER IN
CONTROL AREA. DURING INITIAL FREQUENCY RESPONSE FOR 0.1 P.U. LOAD DISTURBANCE,
REQ= 0.0556 FOR FREQUENCY CONTROLLER‐I AND II ............................................................ 242
FIGURE 6‐15 INITIAL FREQUENCY RESPONSE FOR 0.1 P.U. LOAD DISTURBANCE ............................ 242
FIGURE 6‐16 GENERATED ACTIVE POWER SETPOINT FOR FREQUENCY CONTROLLER‐II IN THE CASE
OF 0.1 P.U. LOAD DISTURBANCE IN CONTROL AREA .............................................................. 244
FIGURE 6‐17 GENERATED ACTIVE POWER SETPOINT FOR FREQUENCY CONTROLLER‐I IN THE CASE
OF 0.1 P.U. LOAD DISTURBANCE IN CONTROL AREA .............................................................. 244
FIGURE 6‐18 VSWTG ROTOR SPEED VARIATIONS WITH DIFFERENT FREQUENCY DROOP MODELS
DURING FREQUENCY RESPONSE ............................................................................................. 246
FIGURE 6‐19 VSWTG ROTOR SPEED VARIATIONS WITH DIFFERENT FREQUENCY CONTROL LOOPS
[ZOOMED] ............................................................................................................................... 246
FIGURE 6‐20 FREQUENCY RESPONSE OF SINGLE AREA LFC CONTROL MODEL INTEGRATED WITH
PROPRIETARY TYPE 4 VSWTG MODEL FOR 0.1 P.U. LOAD DISTURBANCE, REQ = 0.035, NO
RETUNING APPLIED IN TORQUE CONTROLLER ....................................................................... 247
FIGURE 6‐21 GENERATED ELECTRICAL POWER FROM PROPRIETARY TYPE 4 VSWTG MODEL DURING
INITIAL FREQUENCY RESPONSE FOR 0.1 P.U. LOAD DISTURBANCE, NO RETUNING APPLIED IN
TORQUE CONTROLLER ............................................................................................................ 247
FIGURE 6‐22 PROPRIETARY TYPE 4 VSWTG MODEL ROTOR SPEED VARIATIONS WITH DIFFERENT
FREQUENCY DROOP MODELS DURING FREQUENCY RESPONSE ............................................. 248
PROPOSED TYPE 4 VSWTG MODEL FOR 0.1 P.U. LOAD DISTURBANCE, REQ=0.035 ............... 250
17
PROPOSED TYPE 4 VSWTG MODEL FOR 0.1 P.U. LOAD DISTURBANCE [ZOOMED] ................ 250
FIGURE 6‐25 GENERATED ELECTRICAL POWER FROM PROPRIETARY TYPE 4 VSWTG MODEL DURING
INITIAL FREQUENCY RESPONSE FOR 0.1 P.U. LOAD DISTURBANCE, REQ=0.035 .................... 250
FIGURE 6‐26 GRID CODE SENSITIVE ACTIVE POWER SETPOINT FROM FREQUENCY CONTROLLER‐I
DURING INITIAL FREQUENCY RESPONSE FOR 0.1 P.U. LOAD DISTURBANCE .......................... 251
FIGURE 6‐27 GRID CODE SENSITIVE ACTIVE POWER SETPOINT FROM FREQUENCY CONTROLLER‐II
DURING INITIAL FREQUENCY RESPONSE FOR 0.1 P.U. LOAD DISTURBANCE .......................... 251
FIGURE 6‐28 GENERATED ROTOR SPEED FROM PROPOSED TYPE 4 VSWTG MODEL DURING INITIAL
FREQUENCY RESPONSE FOR 0.1 P.U. LOAD DISTURBANCE, REQ=0.035 ................................ 251
FIGURE 7‐1 CONVENTIONAL FUZZY PI CONTROLLER MODEL 1 [REFORMULATED FROM 12] .......... 262
FIGURE 7‐2 CONVENTIONAL FUZZY PI CONTROLLER MODEL 2 ........................................................ 262
FIGURE 7‐3 CLASSICAL HYDRO TURBINE GOVERNOR MODEL .......................................................... 264
FIGURE 7‐4 HYDRO GOVERNOR MODEL ........................................................................................... 265
FIGURE 7‐5 FREQUENCY RESPONSE FROM HYDRO POWER PLANT ................................................. 267
FIGURE 7‐6 MECHANICAL POWER OUTPUT FROM HYDROPOWER PLANT DURING FREQUENCY
RESPONSE ............................................................................................................................... 267
FIGURE 7‐7 SPLIT SHAFT BASED GAS TURBINE ................................................................................. 268
FIGURE 7‐8 GAST MODEL ................................................................................................................. 269
FIGURE 7‐9 GAST MODEL IN THE PRESENCE OF FREQUENCY RESPONSE ......................................... 272
FIGURE 7‐10 MECHANICAL POWER FROM GAST MODEL DURING FREQUENCY RESPONSE ............ 272
FIGURE 7‐11 FREQUENCY RESPONSE UNDER AGC CONTROL .......................................................... 273
FIGURE 7‐12 SIMPLIFIED DCHP MODEL ............................................................................................ 273
FIGURE 7‐13 DCHP GOVERNOR MODEL ........................................................................................... 274
FIGURE 7‐14 POWER LIMITATION BLOCK ......................................................................................... 274
FIGURE 7‐15 POWER DISTRIBUTION BLOCK ..................................................................................... 275
FIGURE 7‐16 GAS DYNAMIC BLOCK .................................................................................................. 276
FIGURE 70‐17 DCHP FREQUENCY RESPONSE UNDER 0.01 P.U. LOAD DISTURBANCE ...................... 277
FIGURE 7‐18 DCHP POWER OUTPUT UNDER 0.01 P.U. LOAD DISTURBANCE .................................. 277
FIGURE 7‐19 SINGLE REHEAT TURBINE MODEL ................................................................................ 278
FIGURE 7‐20 SIMPLE GOVERNOR MODEL ........................................................................................ 279
FIGURE 7‐21 DIRECT SEARCH ALGORITHM ....................................................................................... 282
FIGURE 7‐22 FREQUENCY REGULATION MODEL FOR A SINGLE CONTROL AREA ............................. 283
FIGURE 7‐23 PRIMARY FCR SERVICE MODEL WITH WIND PARTICIPATION IN FCR .......................... 285
FIGURE 7‐24 SECONDARY FRR SERVICES WITH WIND PARTICIPATING IN FRR ................................ 286
FIGURE 7‐25 RR CONTROL LOOP ...................................................................................................... 286
FIGURE 7‐26 WTG INERTIA CONTROLLER ......................................................................................... 286
FIGURE 7‐27 RING TOPOLOGY FOR INTERCONNECTED CONTROL AREA ......................................... 287
FIGURE 7‐28 TAIPS MODEL WITH INTEGRATED WIND FARM .......................................................... 289
FIGURE 7‐29 WIND FARM NETWORK CONNECTION FOR LVRT TEST ............................................... 291
FIGURE 7‐‐30 CONTROL AREA FREQUENCY UNDER NORMAL DISTURBANCE AND NO FREQUENCY
CONTROL ................................................................................................................................. 293
FIGURE 7‐‐31 REGULATION RESERVE RESPONSE WITH INTEGRATED WIND PROVIDING NO
FREQUENCY RESPONSE ........................................................................................................... 294
FIGURE 7‐32 FCR CONTROL SIGNAL RECEIVED BY WIND PLANT WHEN 40% CONVENTIONAL FCR
UNITS ARE DISCONNECTED ..................................................................................................... 294
FIGURE 7‐33 FCR RESPONSE WITH 40 % WIND CONTRIBUTING TO FCR RESERVE .......................... 295
FIGURE 7‐34 POWER SET‐POINT GENERATED BY FREQUENCY CONTROLLER 2 WHEN WIND PLANT
PARTICIPATES AS FCR .............................................................................................................. 295
18
FIGURE 7‐35 GENERATED POWER WITH INERTIAL AND FREQUENCY CONTROLLER 2 ENABLED WHEN
WIND PLANT IS PARTICIPATING IN FCR .................................................................................. 296
FIGURE 7‐36 POWER REGULATION RESERVE RESPONSE FOR RANDOM HIGH LOAD WITH 40%
HYDRO REPLACED WITH DROOP‐INERTIA ENABLED FREQUENCY RESPONSIVE WIND PLANT297
FIGURE 7‐37 FCR AND FRR COMMAND SIGNAL RECEIVED BY WIND TURBINE UNIT UNDER RANDOM
HIGH LOAD CONDITION .......................................................................................................... 298
FIGURE 7‐38 RESERVE RESPONSE DURING RANDOM NORMAL LOAD VARIATIONS WHEN THE WIND
PARTICIPATES IN FRR ONLY ..................................................................................................... 298
FIGURE 7‐39 RESERVE RESPONSE DURING RANDOM HIGH LOAD VARIATIONS WHEN THE WIND
PARTICIPATES IN FRR & FCR .................................................................................................... 299
FIGURE 7‐40 FREQUENCY RESPONSE INDICATORS COMPARISON WITH WIND PLANT PARTICIPATION
IN DIFFERENT REGULATION SERVICES AND DIFFERENT LOAD DEMAND ................................ 300
FIGURE 7‐41 ELECTRICAL POWER UNDER NORMAL LOAD DISTURBANCE (INERTIAL POWER IS ZERO)
................................................................................................................................................ 301
FIGURE 7‐42 ROTOR SPEED UNDER LIMITED POWER SETPOINT ..................................................... 301
FIGURE 7‐43 WTG ELECTRICAL POWER DURING LOAD DISTURBANCE CONTINGENCY ................... 302
FIGURE 7‐44 WTG ELECTRICAL POWER DURING LOAD DISTURBANCE CONTINGENCY (ZOOMED) . 302
FIGURE 7‐45 WTG ROTOR SPEED DURING RANDOM HIGH LOAD .................................................... 303
FIGURE 7‐46 SYSTEM FREQUENCY WHEN WIND PLANT PARTICIPATES IN DIFFERENT REGULATION
SERVICES AT HIGH LOAD DISTURBANCE ................................................................................. 303
FIGURE 7‐47 SYSTEM FREQUENCY WHEN WIND PLANT PARTICIPATES IN DIFFERENT REGULATION
SERVICES AT NORMAL LOAD DISTURBANCE ........................................................................... 303
FIGURE 7‐48 SYSTEM FREQUENCY DURING A CONTINGENCY/HIGH LOAD ..................................... 304
FIGURE 7‐49 WTG ELECTRICAL POWER DURING CONTINGENCY/HIGH LOAD ................................. 304
FIGURE 7‐50 COMPARISON OF WIND PARTICIPATION IN FCR & FRR RESPONSE IN TERMS OF
SYSTEM FREQUENCY DURING A RANDOM LOAD ................................................................... 305
FIGURE 7‐51 COMPARISON OF WIND PARTICIPATION IN FCR & FRR RESPONSE IN TERMS OF WTG
POWER DURING A RANDOM LOAD ........................................................................................ 305
FIGURE 7‐52 COMPARISON OF AREA 1‐2‐3 FREQUENCY RESPONSE FOR 0.01 P.U. LOAD
DISTURBANCE IN AREA 1 ONLY ............................................................................................... 310
FIGURE 7‐53 COMPARISON OF AREA 1‐2‐3 FREQUENCY RESPONSE FOR 0.01 P.U. LOAD
DISTURBANCE IN ALL THREE AREA ......................................................................................... 310
FIGURE 7‐54 MECHANICAL POWER CHANGE OF ALL GENERATING UNIT IN INTERCONNECTED 3
AREA FOR 0.01 P.U. LOAD IN AREA 1 ONLY ............................................................................ 311
FIGURE 7‐55 MECHANICAL POWER CHANGE OF ALL GENERATING UNIT IN INTERCONNECTED 3
AREA FOR 0.01 P.U. LOAD IN ALL THREE AREA ....................................................................... 311
FIGURE 7‐56 TIE‐LINE POWER CHANGE OF ALL GENERATING UNIT IN INTERCONNECTED 3 AREA FOR
0.01 P.U. LOAD IN AREA 1 ONLY ............................................................................................. 311
FIGURE 7‐57 TIE‐LINE POWER CHANGE OF ALL GENERATING UNIT IN INTERCONNECTED 3 AREA FOR
0.01 P.U. LOAD IN ALL THREE AREA ........................................................................................ 312
FIGURE 7‐58 FREQUENCY DEVIATION IN AREA 1 FOR DIFFERENT LEVEL OF WIND PLANT FOR 0.01
P.U. LOAD IN AREA 1 CONSIDERING THE EFFECT OF THE WIND ON R AND B ........................ 313
P.U. LOAD IN AREA 1 WITHOUT CONSIDERING THE EFFECT OF THE WIND ON R AND B ....... 313
P.U. LOAD IN AREA 1 CONSIDERING THE EFFECT OF THE WIND ON R AND B ........................ 314
P.U. LOAD IN AREA 1 WITHOUT CONSIDERING THE EFFECT OF THE WIND ON R AND B ....... 314
19
FIGURE 7‐62 COMPARISON OF FREQUENCY DEVIATION WHEN ONLY CONVENTIONAL PLANT
PARTICIPATES IN AGC IN THE PRESENCE OF FREQUENCY RESPONSIVE WIND PLANT AND 0.01
P.U. LOAD DISTURBANCE. ....................................................................................................... 314
FIGURE 7‐63 COMPARISON OF FREQUENCY DEVIATION WHEN ONLY CONVENTIONAL PLANT
PARTICIPATES IN AGC IN THE PRESENCE OF FREQUENCY RESPONSIVE WIND PLANT AND 0.01
P.U. LOAD DISTURBANCE. ....................................................................................................... 315
FIGURE 7‐64 INTERCONNECTED 3‐AREA FREQUENCY RESPONSE WHEN GAS BASED AREA 2 IS
INTEGRATED WITH FREQUENCY RESPONSIVE WIND PLANT AND 0.01 P.U. LOAD
DISTURBANCE IS APPLIED IN AREA 2. GAST MODEL WITH LOWER FUEL LIMIT OF ‐0.02 IS
APPLIED FOR GAS PLANT MODEL. .......................................................................................... 315
FIGURE 7‐65 INTERCONNECTED 3‐AREA FREQUENCY RESPONSE WHEN GAS BASED AREA 2 IS
DISTURBANCE IS APPLIED IN AREA 2. GAST MODEL WITH LOWER FUEL LIMIT OF ‐1 IS APPLIED
FOR GAS PLANT MODEL. ......................................................................................................... 315
FIGURE 7‐66 INTERCONNECTED 3‐AREA FREQUENCY RESPONSE WHEN THERMAL‐BASED AREA IS
DISTURBANCE IS APPLIED IN AREA 1. ..................................................................................... 316
FIGURE 7‐67 INTERCONNECTED 3‐AREA FREQUENCY RESPONSE WHEN HYDRO BASED AREA 3 IS
FIGURE 7‐68 INTERCONNECTED 3‐AREA FREQUENCY RESPONSE WHEN DCHP BASED AREA 2 IS
DISTURBANCE IS APPLIED IN AREA 2. ONLY CONVENTIONAL GENERATING UNIT PARTICIPATES
IN AGC. LOAD LIMIT (1.2, ‐0.1) P.U. ........................................................................................ 316
DISTURBANCE IS APPLIED IN AREA 2. ONLY CONVENTIONAL GENERATING UNIT PARTICIPATES
IN AGC. LOAD LIMIT (1, ‐1) P.U. .............................................................................................. 317
FIGURE 7‐71 COMPARISON OF FREQUENCY RESPONSE FOR THERMAL‐BASED AREA 1 INTEGRATED
WITH FREQUENCY RESPONSIVE WIND PLANT WHEN BOTH OF THEM PARTICIPATES IN AGC AS
PER THE INDICATED PARTICIPATION FACTOR. EFFECT OF WIND PLANT ON R AND B HAS BEEN
CONSIDERED HERE. ................................................................................................................. 318
FIGURE 7‐72 COMPARISON OF POWER DEVIATION FOR THERMAL‐BASED AREA 1 INTEGRATED
WITH FREQUENCY RESPONSIVE WIND PLANT WHEN BOTH OF THEM PARTICIPATES IN AGC
WITH 0.4 PARTICIPATION FACTOR (PF) FOR THERMAL AND 0.6 PF FOR THE WIND. EFFECT OF
WIND PLANT ON R AND B HAS BEEN CONSIDERED HERE. ...................................................... 318
FIGURE 7‐73 COMPARISON OF FREQUENCY RESPONSE FOR THERMAL‐BASED AREA 1 INTEGRATED
WITH FREQUENCY RESPONSIVE WIND PLANT WHEN BOTH OF THEM PARTICIPATES IN AGC AS
PER THE INDICATED PARTICIPATION FACTOR. EFFECT OF WIND PLANT ON R AND B HAS NOT
BEEN CONSIDERED HERE. ....................................................................................................... 318
FIGURE 7‐74 COMPARISON OF POWER DEVIATION FOR THERMAL‐BASED AREA 1 INTEGRATED
WITH FREQUENCY RESPONSIVE WIND PLANT WHEN BOTH OF THEM PARTICIPATES IN AGC
WITH 0.4 PARTICIPATION FACTOR (PF) FOR THERMAL AND 0.6 PF FOR THE WIND. EFFECT OF
WIND PLANT ON R AND B HAS NOT BEEN CONSIDERED HERE. .............................................. 319
FIGURE 7‐75 ACE OBTAINED UNDER DIFFERENT SIMULATION SCENARIOS ..................................... 319
FIGURE 7‐76 COMPARISON OF FREQUENCY DEVIATION FOR THERMAL‐BASED AREA 1 INTEGRATED
WITH WIND PLANT. THERMALAGC‐WIND REFERS THAT ONLY THERMAL PLANT PARTICIPATES
20
IN AGC AND NORMAL OPERATING WIND PLANT IS INTEGRATED INTO SYSTEM. THE WIND (RB)
INDICATES R AND B VALUES ARE CHANGED AS PER INERTIA REDUCTION IMPACT. FRWIND
REFERS TO FREQUENCY CONTROLLER‐II BASED FREQUENCY RESPONSIVE WIND PLANT.
THERMALAGC‐FRWINDAGC REFERS TO PARTICIPATION OF BOTH UNITS IN AGC. ................ 321
FIGURE 7‐77 COMPARISON OF ROCOF FOR THERMAL‐BASED AREA 1 INTEGRATED WITH
FREQUENCY RESPONSIVE WIND PLANT. ROCOF VALUES ARE OBSERVED WHEN 0.01 P.U. OF
LOAD IS APPLIED AT 5 S IN AREA 1 ONLY. LEGENDS ARE SAME AS EXPLAINED ABOVE IN
EARLIER FIGURE. ..................................................................................................................... 322
FIGURE 7‐78 COMPARISON OF SETTLING TIME FOR THERMAL‐BASED AREA 1 INTEGRATED WITH
FREQUENCY RESPONSIVE WIND PLANT. LEGENDS ARE SAME AS EXPLAINED ABOVE. .......... 323
FIGURE 7‐79 MAIN EFFECT PLOTS FOR AGC PARAMETERS ON SYSTEM FREQUENCY DEVIATION. .. 324
FIGURE 7‐80 COMPARISON OF FREQUENCY RESPONSE INDICATORS FOR CONVENTIONAL PLANT
INTEGRATION WITH FREQUENCY RESPONSIVE WIND PLANT WHEN ONLY CONVENTIONAL
UNIT PARTICIPATES IN AGC AND 0.01 P.U. LOAD IS APPLIED. ................................................ 325
FIGURE 7‐81 FUZZY GAIN SCHEDULING (FGSPID) STRUCTURE ......................................................... 328
FIGURE 7‐82 MEMBERSHIP FUNCTIONS FOR ′ AND ′ ........................................................... 329
FIGURE 7‐83 MEMBERSHIP FUNCTIONS FOR ACE AND ∆ACE .......................................................... 329
FIGURE 7‐84 HYDRO‐TURBINE GOVERNOR RESPONSE .................................................................... 329
FIGURE 7‐85 FGSPID RULE IMPLICATION PROCESS .......................................................................... 332
FIGURE 7‐86 FGSPID GAINS SURFACE VIEW ..................................................................................... 333
FIGURE 7‐87 LFC CONTROLLER COMPARISON WITH 0.1 PER UNIT LOAD DISTURBANCE IN AREA 1
ONLY AND 90%‐10% HYDRO‐WIND AGC PARTICIPATION. ...................................................... 335
FIGURE 7‐88 LFC CONTROLLER COMPARISON WITH 0.1 PER UNIT LOAD DISTURBANCE IN ALL 3
AREAS AND 90%‐10% HYDRO‐WIND AGC PARTICIPATION IN AREA 1. ................................... 335
FIGURE 7‐89 LFC CONTROLLER COMPARISON WITH 0.1 PER UNIT LOAD DISTURBANCE IN ALL 3
AREAS AND 80%‐20% HYDRO‐WIND AGC PARTICIPATION IN AREA 1. (ZOOMED TO 40 SEC) 335
FIGURE 7‐90 THREE AREA FREQUENCY RESPONSE COMPARISON WITH 0.1 PER UNIT LOAD
DISTURBANCE AND 90%‐10% HYDRO‐WIND AGC PARTICIPATION IN AREA 1. FUZZY GAIN
SCHEDULED LFC CONTROLLER IS USED. .................................................................................. 337
FIGURE 7‐91 THREE AREA FREQUENCY RESPONSE COMPARISON WITH 0.1 PER UNIT LOAD
DISTURBANCE AND 80%‐20% HYDRO‐WIND AGC PARTICIPATION IN AREA 1. FUZZY GAIN
SCHEDULED LFC CONTROLLER IS USED. .................................................................................. 337
FIGURE 7‐92 AREA 1 FREQUENCY RESPONSE COMPARISON WITH 0.1 PER UNIT LOAD DISTURBANCE
AND DIFFERENT WIND INTEGRATION SCENARIO. .................................................................. 338
FIGURE 7‐93 FAULT RIDE‐THROUGH GRID CODES [48] .................................................................... 340
FIGURE 7‐94 SIMULATION RESULTS FOR 0.1 VNOM FOR 150 MSEC. .............................................. 342
FIGURE 7‐95 ROTOR CURRENT DURING VOLTAGE DISTURBANCE ................................................... 343
FIGURE 7‐96 SIMULATION RESULTS FOR 0.1 V FOR 450 MSEC. ....................................................... 343
FIGURE 7‐97 SIMULATION RESULTS FOR 0.1 V FOR 150 MSEC FOR NORMAL DFIG OPERATION .... 345
FIGURE 7‐98 FRT ENABLED ACTIVE POWER CONTROL BLOCK ......................................................... 346
FIGURE 7‐99 PFRT VARIATION WITH TERMINAL VOLTAGE .............................................................. 347
FIGURE 7‐100 ELECTRICAL PARAMETERS IN NORMAL VSWTG DURING VOLTAGE DISTURBANCE .. 347
FIGURE 7‐101 ELECTRICAL PARAMETERS IN FRT ENABLED VSWTG DURING VOLTAGE DISTURBANCE
................................................................................................................................................ 347
FIGURE 7‐102 PRIMARY FREQUENCY RESPONSE COMPARISON DURING 0.01 P.U. LOAD
DISTURBANCE ......................................................................................................................... 348
FIGURE 7‐103 AGC FREQUENCY RESPONSE COMPARISON DURING 0.01 P.U. LOAD DISTURBANCE
WITH AGC CONTROLLER GAIN AS P=‐37.47, I=0.0096 ............................................................ 348
FIGURE 7‐104 AGC FREQUENCY RESPONSE COMPARISON DURING 0.01 P.U. LOAD DISTURBANCE
WITH AGC CONTROLLER IN AS P= ‐ 45.63, I = 0.0133 ............................................................. 348
21
ACRONYMS
Area Control Error ACE
Automatic Generation Control AGC
Australian Energy Market Operator AEMO
Combined Cycle Gas Turbines CCGT
Doubly Fed Induction Generator DFIG
Decentralized Combined Heat Power Plant DCHP
Energy Management System EMS
Fault Ride Through FRT
Frequency Operating Standard FOS
Frequency Control Ancillary Service FCAS
Fast Frequency Response FFR
Frequency Restoration Reserve FRR
Frequency Containment Reserve FCR
Fuzzy Gain Scheduled Proportional Integral

FGSPID
Derivative
Integral Of Time Multiplied By Absolute Error ITAE
Integral of the Squared Errors ISE
Load Frequency Control LFC
Large-scale Renewable Energy Target LRET
Levelised Costs Of Energy Generation LCOG
Maximum Power Point Tracking MPPT
National Electricity Market NEM
National Transmission Network Development Plan NTNDP
Point Of Connection PCC
22
Renewable Energy Sources RES
Rate Of Change Of Frequency ROCOF
Replacement Reserve RR
Short Circuit Ratio SCR
Supervisory Control and Data acquisition SCADA
Transmission System Operator TSO
Three Area Interconnected Power System TAIPS
Variable Speed Wind Turbine System VSWTG
Wind Energy System WES
Western Electricity Coordinating Council WECC
23
NOMENCLATURE
f Frequency
Rate of change of frequency
∆f Frequency deviation
H Inertia constant of conventional

generating unit
S MVA rating of conventional

generating unit
W Kinetic energy of rotating mass
IR System Inertia
∆P MW loss from generating unit
Hz Hertz per second

s
F Frequency Nadir Point
P Total Wind generated power
P Synchronous generator power
P Inter area power flow
WP Wind penetration level
F Lowest permissible frequency as per

frequency operating standard
J Moment of inertia of a rotating mass
M Equivalent moment of inertia of all

generators and the motors connected
to the grid
P Mechanical Power
P Electrical Load
H Equivalent system inertia constant
S System MVA base
24
S Power rating of individual generating
unit
∆P Change in active power output of

synchronous generators in
conventional generating plants
D Damping constant
db Deadband
K 1/D
Angular speed
T Power system time constant
P Primary Frequency Control Power
R Speed regulation or droop of governor
R Variable droop
R Transient droop of hydro turbine

model
R Permanent droop of hydro turbine

model
M Actual power margin available at a

wind turbine
M Maximum power margin available

among all wind turbines
T Turbine time constant
P Power output from automatic

generation control
K Proportional time constant of AGC

controller
K Integral time constant of AGC

controller
NIA Sum of the actual power flows on all

tie lines or interconnectors
25
NIS Sum of the scheduled flows on all tie
lines or interconnectors
P Tie-line power
∆P Change in tie-line power between area

1 to area 2
T Synchronizing coefficient
δ Relative phase angle between control

areas
B Frequency bias setting
β System frequency response

characteristic
IME Interchange (tie line) Metering Error
T‐ω curve Torque-Speed curve
P‐ω curve Power-Speed curve
P Power set point for the de-loaded

wind turbine
ω Deloaded wind turbine rotor speed
ω Measured rotor speed for wind turbine
ω Optimum rotor speed
P Maximum mechanical power from

turbine
P Deloaded Power
K Deloading constant
K Optimum mechanical power constant
β Pitch Angle
∆β Offset pitch angle
Cp Turbine power coefficient
26
C λ , β Coefficient of power at optimum
value of tip speed ration and pitch
angle
λi Tip speed ratio at ith time step
ρ Air density
A Turbine blades swept area
R Blade length
Vwind Velocity of the incident wind
K System stiffness
M Inertia of power supply
Vout Chopper output voltage
Vin Chopper input voltage
D Duty cycle
i , i 2-phase line current in dq synchronous

frame
W Grid angular frequency as well as

synchronous reference frame speed
WL i & W L i Induced reference voltages in

synchronous frame
V and V Decoupled controller output in dq

reference frame
φ Flux
K Park’s transformation matrix
f Field winding
k d-axis damper winding
k q-axis damper winding
k Number of damper circuits
27
[R] Winding resistances diagonal matrix
in d q axis
[L] Winding self and mutual inductances

matrix in d q axis
[ω] Rotor speed matrix
T Electrical Torque
P Electrical power output
θ Pitch angle of the turbine blades
θ Commanded pitch angle
Kaero Aerodynamic gain factor
Pset Power setpoint
∆P Scheduled power setpoint for turbine
∆P Power setpoint imposed by primary

control
i Active current command
IPmax Upper limit of the active current
Imax lumped converters ratings
IQmax Maximum reactive current limit
P Available wind power
v Wind speeed
DR Derating command mode
P Reference power setpoint
P Transmission system operator

commanded power set point
ω Reference angular rotor speed
PAPC Active power setpoint to turbine

controller
P Final power command provided to

generator-converter model
28
L Wind penetration level
H Frequency responsive wind plant

inertia
T Total time delay associated with

generating unit model
p.u. Per unit
e Error
μ Membership value in fuzzy control
TW Water starting time constant
∆P Gate valve change
∆P Change in power command in terms

of speed change setting
∆ Change or deviation
T Governor time constant
T Turbine time constant
T Time constant of steam chest and inlet

volume
T Time constant of reheater
F Fraction of total turbine power

generated by high pressure chamber
T Turbine torque
V Control valve position
∆P , , ∆P , , ∆P , Power variation from primary,

secondary and tertiary control
P Inertial Power output from wind

turbine
29
EXECUTIVE SUMMARY
There is coupling between the reliability requirements of maintaining frequency and
the commercial interest in inadvertent energy supply. Modern electric power systems
are facing a critical challenge in the real‐time balancing of demand and supply due to
unpredictability and variability of wind power. Frequency response capability offered
by wind plant is not same as the primary control capability of conventional plants.
Increased penetration of wind generation in interconnected power system thus
intrinsically calls for frequency regulation analysis with present and future wind
technologies. This thesis titled “Frequency Regulation in Wind Integrated Power
System” aims to study the frequency regulation capability of wind energy system and
its integration impact on interconnected power system. Simulations are conducted
under the limitations of generic system modeling with reduced number of parameters
which is an acceptable practice for bulk power system studies like frequency
regulation of large interconnected power system.
First of all, a large interconnected network like Australian NEM is investigated to
assess current problems due to high wind penetration, and few possible solution
discussed. With full converter based wind turbine system being relatively new wind
technology, a simplified yet absolute representation of full converter based wind
turbine generator modeling is presented for frequency regulation related bulk studies.
Response capability of proposed model is verified against a proprietary model and a
detailed model.
This study also explores frequency regulation issues arising from high wind
penetration levels in power system. Given compulsory frequency response
requirement as part of the grid, power electronics based VSWTG requires modified
active power control algorithm to support regulation services. A moving average
30
frequency and the dynamic dead band dependent grid frequency processor has been
implemented with a grid code compatible active power control algorithm for
VSWTG. Grid code compatible frequency response from type-3 and proposed type-
4 VSWTG is investigated and analyzed for improvements with other available
models.
Different generation technologies are dispatched by their capability of responding to
variations in demand and fuel type. Design and modeling aspects of frequency
regulation scheme in the wind integrated multigenerational system are analyzed in
this research work. The frequency response of validated models of thermal, hydro
and gas-based power in the presence of frequency responsive wind plant is
investigated. Special focus is given to secondary frequency regulation or AGC
investigation in the wind integrated power system. Classical and intelligent AGC
controllers are analyzed in the presence of grid code compatible frequency responsive
wind power plant penetration.
Impending requisite for application of frequency responsive wind plants also calls for
fault ride through analysis of such plant. Simulation results for weak network
presented improved LVRT capability of frequency responsive VSWTGs in
comparison to normal VSWTGs. With active power output, well above zero and
turbine speed also remaining under limit during fault conditions, frequency grid code
compatible VSWTGs are found to satisfy FRT grid codes of many countries.
31
1 CHAPTER 1
INTRODUCTION
1.1 Current Status of Wind Energy System
The power grid is an enormous, intricate system and pervasive with interactions
between all the different components and across time. When any new technology gets
connected to this complex grid, its compatibility to rest of grid determined by some
basic characteristics. Renewable as a future fuel security and price offset option have
become an attractive proposition in the current global scenario. There is an eternal
growing obligation for electricity from renewable energy sources (RES). The impact
of the renewable energy sources on the grid and its contribution to the grid is still
relatively unknown with several challenges needing to be overcome before mass
deployment.
Wind power credentials as a rapidly deployable clean technology have put it at the
forefront in the fight against climate change. The Wind is the most cost competitive
renewable source of electricity generation behind hydropower. This proven and
mature technology has highly penetrated energy matrix at the global level, hence
among all available RES, maximum impact potential lies with wind power at largest
scale. There has been a tremendous increase in generated megawatts by the RES all
over the world. Wind energy has optimistic prospects as a partner in energy resolution
which can pave the way for shaping a new sustainable world for future generations.
According to a 2013 published key report [1], wind power deployment has more than
doubled, approaching 300 GW cumulative installed capacities led by China (75 GW),
the United States (60 GW) and Germany (31 GW). Wind power now provides 2.5%
of global electricity demand – and up to 30% in Denmark, 20% in Portugal, 18% in
32
Spain. Figure 1-1 depicts the cumulative wind energy market forecast for different
regions 2012-2017. Wind energy is anticipated as a major contributor for the
foreseeable future with 20% per annum growth forecast for global installed wind
energy capacity.
Figure 1-1 Cumulative Forecast by Region 2012-2017 [2]
Although most wind-power developments still attract financial support: either in the
form of capital subsidies or premium payments for the energy gap between the
generation costs of wind energy and electricity from conventional plants continues to
narrow and are close to cost-competitive with new natural gas generation due to
continuing technological innovation. The onshore wind costs fell in 2014 and, a
realistic best generation cost for onshore wind is now $71/MWh on a global basis [3-
4]. A Bloomberg report has listed the wind as cheaper electricity producing the fuel
than fossil fuels for Australia which is world’s biggest coal exporter. A new wind
farm in Australia supplies electricity at the cost of 80AUD / MWh, while it comes
out to be 143AUD / MWh for a new coal-fired power plant or 116AUD / MWh when
supplied from a new natural gas powered station in 2013 [5]. The prospective plunge
33
in LCOE of wind is expected shortly due to the emergence of low cost but high
technology wind turbine manufacturer.
1.2. Motivations for Research
The present power system is undergoing several changes in its basic structure which
are associated with the adoption of new power production technologies and rapid
integration of renewable energy sources (RES). Integration of renewable energy
resources at large has widespread impacts on power system stability & reliability.
The Wind is the most cost competitive renewable source of electricity generation
behind hydro. Proven and mature wind power technology has highly penetrated
energy matrix at the global level, hence among all available RES, maximum impact
potential lies with wind power at largest scale. In Australia, wind farms have
proliferated in the National Electricity Market (NEM), with some 3,100 MW of
installed capacity entering the market to date. However, it has significantly more
growth potential because of the greater level of as yet unutilized resources. Its cost
competitiveness will be enhanced by a reduction in the cost of turbines, particularly
through low-cost, high volume manufacturing in countries such as India and China,
and to a lesser extent by further efficiency gains through turbine technology
development. Factors that may limit the development of wind energy on a localized
basis are a lack of electricity transmission infrastructure to access remote wind
resources and the intermittency and variability of wind energy. The power output of
a wind turbine, which converts an uncertain energy flux to electricity, has an
additional source of uncertainty. Wind power density is a stochastic process that is a
function of both space and time. Wind power generation introduces uncertainty in
power system operation, and it is continuously variable and difficult to predict.
Modern electric power systems are facing a critical challenge in the real-time
34
balancing of demand and supply due to unpredictability and variability of wind
power.
The Emergence of Type 4 VSWTG Technology: After displacing dominant simple
constant speed turbines, variable speed based Wind Energy Conversion System
(WECS) technology is a pioneer in currently installed wind energy projects. Even
though doubly fed induction generators based type 3 is the predominant technology
nowadays, full converter based type 4 turbine technology is expected to become the
industry standard in medium terms. Joint Research Centre’s (JRC) 2014 report [6]
analyzing wind turbine purchase agreements (TPAs) sample consisting of countries
like China, USA, France, Germany, UK, Brazil and Rest of world identify the ending
of pre-eminence of type 3 technology due to increasing popularity of type 4 wind
turbine system. Considering the connection trends, Australian Electric Market
Operator (AEMO) also assume an increase in type 3 and type 4 wind turbines in NEM
[7].
Variable speed wind turbine generators (VSWTG) having incommensurable dynamic
characteristics than conventional synchronous generators requires precise modeling
as they no longer receive negative load treatment. WECS modeling is now a priority
research topic due to change in perspective of various stakeholders like transmission
system operators, researchers, and wind developers. The perpetual cumulative
obligation for electricity from these vastly penetrated technologically advanced wind
energy system poses an incendiary threat to stability and reliability of modern electric
grid if their dynamic performance aspects are not assessed beforehand. So the
transmission system operators (TSOs) all over the world advocate incoming WES
integrations to follow stringent grid codes in their modeling.
35
The Emergence of Wind-based Ancillary Services Requirement and
Corresponding VSWTG Modelling Modifications: Sophisticated technology has
ushered in improved forecasting of wind energy generation over varying time frames.
These enhancements have initiated the contemplation of wind energy based ancillary
services requirements for improved system scheduling & planning. Though WES
ancillary services like frequency control ancillary services (FCAS) are still in
research & development stage in most of the countries, future WES participation in
electricity economic markets is expected to contribute to enhanced market efficiency,
improved system reliability and macroeconomic benefits to all stakeholders. Even
with proven technology, wind energy generation sector still lacks in aspects like grid
integration assessment. Design and operation of power system in the presence of wind
energy are one of the major issues in wind power integration.
Renewable energy sources like wind having new production and nonlinear control
technology imply new energy management system (EMS) in modern power systems.
Talking about applying energy management in the modern power system, supervisory
control and data acquisition (SCADA), Automatic Generation Control (AGC),
system stability, security control and electrical load management make up the major
units [4]. AGC represents a remarkable economic function of EMS system for
regulating the balance and for distributing the imbalance between designated units.
The sensitivity of any power plant depends upon the response time taken to control
the frequency change due to load variation of governor control. System frequency is
regulated by injecting active power into electric grid through power plants. This
control is indispensable for the stable operation of grid ensuring continuous
adaptation of generation to demand. Under a wide range of ambient conditions,
frequency responsive power plants are expected to provide automatic power variation
36
with frequency drop within a given time frame and ramp limit. The conventional
power plants which remain connected to the grid even in the case of frequency
deviations accomplish active power-frequency control through turbine governor’s
reaction to nominal synchronous speed deviations and the respective boilers-turbine
frequency responsive controls.
Dead-band and speed droop are two significant parameters in conventional turbine
governor action during system event [8]. System frequency regulation is highly
dependent upon implemented dead-band for long-term fidelity. Governor dead band
is detrimental to minimum frequency variation needed before the governor action is
activated. The relative frequency deviation corresponding to the relative change in
power output, defined as speed droop is always positive for stable regulation. Factors
like the available headroom, the maximum-minimum power capacity of generating
units and power setpoint corresponding to enabled frequency operating mode have a
direct effect on total frequency responsive reserve amount. A generator operating at
its maximum generating capacity has almost negligible headroom and therefore
unable to provide any frequency responsive operation irrespective of governor-droop
being enabled. To get emulated synchronous generator response, variable speed wind
turbine generator (VSWTG) modeling should consider parameters like droop, dead-
band, headroom in frequency-active power control.
Integration Impacts of Wind Energy Penetration: Recent studies have found that
the wind energy integration impacts on system frequency and power fluctuation are
nonzero and become more significant at higher sizes of penetrations. For smaller
power systems with low inertia or system with low primary frequency response, a
large amount of wind integration may lead to frequency instability of the system.
Increasing wind penetrations tend to displace (decommit) conventional generation
37
making system less synchronous and more asynchronous. Present wind farms use
induction generators (double-fed or classical) which react weakly to frequency
changes leading to smaller effective inertia. Doubly Fed Induction Generators (DFIG)
without special inertial emulation do not have primary control capability. So
frequency response is degraded along with other effects (e.g., increased dead band,
sliding pressure controls, blocked governor, use of power load controllers, and
change in load frequency response). Fluctuations in the summated output of all wind
farms connected to a power system will cause the frequency to fluctuate and may
change the anticipated power output of dispatchable generation (affecting dispatch in
the next few hours and unit commitment in the next few days). Unexpected sudden
changes in the summated output of wind farms, due to either a widespread change in
wind conditions or response to a power system disturbance are contingencies that
must be assessed for their implications for reserve requirements. With increasing
wind energy penetration level, project participants involving transmission system
operator, the market regulator, manufacturer and wind plant owner presents their
perspective.
As represented in Figure 1-2 by the author, a discussion is going on at different
stakeholders’ level regarding the provision of frequency responsive capability to be
a requirement of all generators or be commercial ancillary service only. TSOs need
frequency control ancillary services (FCAS) from wind plant but doubt their
capability to provide. The market regulator believes that FCAS is easily obtainable
from other energy sources. Wind turbine manufacturer with production capability
assumes that wind FCAS is not anticipated while wind farm owners are not in favor
to provide FCAS without any monetary gain. Such dissimilar perspectives of
stakeholders’ can only be satisfied through by rigorous research studies.
38
Figure 1-2 Stakeholders View’s on Frequency Responsive Capability as a Requirement for Wind Energy
System
1.3. Objectives
A considerable number of studies are available in the literature regarding wind
penetration effects on in an electrical network. This work aims to be distinct due to
the applied models in the study. Proprietary DFIG models [9] [10] [11] were mostly
applied in previous works where some important factors like the incorporation of
continuously varying reference power setpoint are not considered. Frequency
regulation in the wind integrated power system calls for a continuously varying power
setpoint instead of a constant setpoint. With full converter based type 4 VSWTGs
gaining market share, frequency regulation capabilities are expected from them too.
So, a suitable frequency responsive model specific for frequency regulation studies
is highly desired. High penetration of wind energy system has led many of the TSOs
to formulate stringent grid codes for system stable and reliable operation. Previous
39
studies lack the frequency response analysis of grid code compatible wind energy
system on the interconnected system. Maintenance of system frequency in a large
inertia based system within utility prescribed tolerance band is an exclusive
responsibility of the conventional power plants. With increased renewable energy
targets globally, it is usual to find an area regulated by hydro generation
interconnected to another area regulated by thermal generation or gas based
generation or in a combination of both generations along with good penetration of
wind plants. Integration of wind-based generating plant to a system having fixed
governor settings brings wide deviations in system frequency, and further might lead
to system instability. Wind plants with variable output are integrated with peak load
plants like gas based plants & hydro plants rather than with baseload thermal plants.
Dynamics of different combinations of turbine – generators have different effects on
system frequency maintenance. Very little work is available in literature in case of
load frequency response in a multigenerational system with grid code frequency
responsive wind plants.
This research work is proposed to answer some of the following set objectives:
Frequency Control and Impacts of Increased Wind Penetration
• Frequency Control issues arising from low power system inertia levels have
become an important issue. What are the frequency control capabilities and
possibilities in large interconnected wind penetrated systems? A system with
the same amount of spinning reserve but different levels of inertia may exhibit
different frequency response during contingency events.
Investigation of Design Aspects of Load Frequency Scheme in Interconnected Mixed
Generational Power Systems in the Presence Of Wind Power.
40
• Electricity dispatched to the market reflects demand requirements for base
load and peak (or maximum) demand. Different generation technologies may
be grouped by their capability of responding to variations in demand and fuel
type. For generating systems of widely different characteristics, research
needs to be done to study automatic generation controllers in interconnection.
Study of Grid code compatible frequency response capability from VSWTG and the
corresponding effect on Fault Ride-through (FRT) Capabilities.
• Energy withdrawn by FRT action may increase the size of largest contingency
& thus magnitude of frequency excursions. Does frequency responsive wind
plants have improved fault ride through capabilities than normally operative
wind plants?
This work will specifically consider following aspects:
• Design and modeling of type 4 VSWTG for application in frequency
regulation studies.
• Effect of grid code implementation on VSWTGs performance and frequency
response.
• Investigate the effect inertia has on arresting frequency deviations during
frequency events.
• Investigate the effect of reserve distribution and effects of wind generation
participation in a different level of frequency control.
• Study of the characteristics of various frequency control strategies.
• Specify requirements and characteristics of electrical power systems
including thermal, hydro and wind-based power system.
• A power system simulation model for load frequency & tie-line power
variations performance verifications.
41
• Testing and comparison of conventional load frequency controllers with
fuzzy gain scheduled load frequency controllers.
• Application of the performance indicators and a direct search algorithm for
optimum AGC controller gains.
1.4. Significance
The focus of this study is to represent the effects of transmission-connected wind
generation on the power system and some of the improvements that can be made in
wind turbine generator modeling for better system frequency response.
Despite having sufficient feasibilities for wind power integration in Australian NEM,
it still lacks various technical and commercial aspects that require further
investigations to test the system adequacy for reliable and secure operation [12].
Intermittent wind power generation and asynchronous generators are two key
characteristics of wind generation technology affecting its integration in NEM grid
and spot market. Even though AEMO does not identify NEM-wide challenges as
each NEM region has a different generation mix, network configuration, and demand
characteristics, which lead to different challenges or different timing; future increased
wind power penetration will bring some adverse operational challenges in the NEM.
This research is aimed to be in line with the Australian Energy Market Operator
(AEMO) concerns published in latest wind integration-2013 [13]. The main way in
which wind generation can affect NEM frequency control is economic displacement
of synchronous generation. AEMO does not currently control inertia in any way: it is
simply an observed characteristic of the power system, much like load level. AEMO
currently operates the power system around the requirements that arise from the
inertia levels that are present. Participation in frequency control arrangements in the
42
NEM is voluntary, and AEMO’s experience is that wind generators choose not to
participate.
AEMO is investigating the extent to which fast frequency response (FFR) could
provide a substitute for synchronous inertia. Examples include synthetic inertia from
wind turbines, FFR from battery storage, or fast response generation ramping. AEMO
is assessing the capabilities and limitations of technologies that can provide an FFR
service, and whether FFR provides an adequate substitute or supplement to inertia.
Furthermore, the scale and type of response required to make a useful contribution to
the power system are unknown. At present, synthetic inertia has not been
demonstrated to be an exact substitute for mechanical inertia, but could, in
combination with fast-acting responses with the right characteristics, maintain the
frequency operating standard (FOS) with a lower level of synchronous inertia than
otherwise would be required.
There is a considerable body of work to be done to fully understand the dynamic
performance of the large interconnected power system as it goes through gradual
displacement of conventional synchronous plants by low inertia generating plants.
1.5. Methodology
The objectives set in this research study are achieved through advanced simulation
studies and detailed analyses of frequency control in the wind integrated power
system. The complete methodology has been represented in Figure 1-3. It starts with
wind integration assessment in Australian transmission network for problem
identifications and concludes finally at design & analysis of load frequency control
model for the wind-integrated system. An extensive literature review is undertaken
to gather realistic information about wind integration effects on Australian power
network infrastructure. A particular emphasis was given to the South Australia
43
network owing to highest wind penetrated state. This approach helped in developing
the framework of the proposed study to analyze the impacts of integrating wind
energy into the power grid. Frequency response studies require suitable
representation for the wind turbine generator system. Current literature provides
simple proprietary models for type 3 VSWTGs with some of the models not free and
publically available [14]. A Generic model of type 4 is available in public domain
[15], but it was found unsuitable under varying power setpoint requirements for
frequency response. Thus, a suitable model for type 4 VSWTG was developed, and
its performance was validated with a proprietary model performance. Various
unobserved factors which were not considered in previous studies were also
highlighted for generic modeling of type 3 VSWTG. The methodology used for
developing models of frequency control in wind energy interconnected systems
consists of identifying specifies and dynamic characteristics of frequency control in
large wind energy interconnected systems to incorporate them into the models
obtained with the basic theory of frequency control. Indeed, while frequency control
capabilities and possibilities are known and detailed in the literature, wind energy
system based frequency control capabilities and possibilities in large interconnected
systems need more detailed research. The frequency control model suggested in this
work will allow the reserve response simulation for all three types of frequency
control namely primary, secondary and tertiary control. Standard grid codes have
been applied to analyze VSWTGs performance when complying with these grid
codes. MATLAB’s Simulink based SimPower platform has been used for all
simulation studies.
44
Investigation of Frequency Response in NEM
Wind Integration Effect Under Low Load and High Frequency Response Problem Identifiction and
Load Scenario Possible Solutions
Development of Frequency Studies Model

Wind Integrated
Individual Generation Models Frequency Regulation Models Interconnected Power System
Models
Investigation of Major Challenges in Frequency Control for Wind Energy

System Penetrated Power System
Inertia Effect & Wind Penetration Reserve Distribution & Frequency Control
Analysis Through Design and Simulations
Frequency Regulations, Classical & Frequency Response of Type 3
Grid Code Implementation Intelligent Controllers & 4 VSWTGs
Wind Integrated Power System Frequency Response
Figure 1-3 Research Methodology
1.6. Contributions
Wind energy integration is a challenge for energy markets, the transmission systems
as well as for system operators. The collaboration with so many different participants
in power system calls for answers to many issues. An attempt has been made in this
thesis to answer a minor part of one of the big question of frequency dynamics issue.
Main contributions of this thesis are shown in Figure 1-4.
45
Contribution 1
Frequency Response Investigation in NEM Control Area
Contribution 2
Modelling & Performance Assessment Of Type 4 Frequency
Responsive Wind Turbine
Contribution 3
Modelling & Investigation of Grid Code Compatible Active‐
Power Controller for Type 4 & Type 3 VSWTGs
Major Contributions
Contribution 4
Development of Frequency Response Model for Wind Energy Impact
Assessment on Reserve Response
Contribution 5
Integration Impact of Grid code Frequency Responsive
VSWTGs on AGC of Multigenerational System
Contribution 6
Development of FGSPID Controller for Wind Integrated Power
System Load Frequency Control
Contribution 7
Fault Ride Through Analysis of Frequency Responsive VSWTGs
Contribution 8
Review of Active Power‐Frequency Control in Wind Integrated
Power System
Figure 1-4 Contributions of the Study
1.6.1. Frequency Response Investigation in NEM Control Area
Under high renewable penetration levels, it is conceivable that frequency response of
the system will be lower due to lower inertia of the system and the displacement of
primary frequency control reserves. An empirically-validated model of NEM is
applied to mimic frequency dynamics and analyze frequency indicators. Quantitative
46
influence of system inertia and contingency events on frequency response in a control
area of Australian network is evaluated for low load and high load scenario.
1.6.2. Modeling & Performance Assessment of Type 4 Frequency
Responsive Wind Turbine
Time-frame of different phenomena is always considered in modeling precision and
simulation set up of electrical phenomena. The proprietary generic model exists for
type 3 VSWTG while the public domain available generic model for type 4 is not
very suitable for frequency response studies. A generic model of type 4 VSWTG
model suitable for frequency response is developed, and its performance is analyzed
with different wind turbine generator models.
1.6.3. Modeling and Investigation of Grid Code Compatible Active-
Power Controller for Type 3 & Type 4 VSWTGs
A grid frequency processor dependent upon dynamic dead band is presented. This
grid frequency processor is applied to formulate grid code compatible frequency
controller for active power response from VSWTGs during frequency disturbances.
The performance of this grid code compatible frequency responsive type 3 and type
4 VSWTGs is compared with other frequency responsive models for improved
response.
1.6.4. Development of Frequency Response Model for Wind Energy
Impact Assessment on Reserve Response
A single area frequency control model is presented in which frequency responsive
VSWTG is integrated at primary control level and secondary control level. Reserve
response and frequency indicators are analyzed when wind farm participates in
individual control level or combined in both control level.

47
1.6.5. Integration Impact of Grid Code Frequency Responsive
VSWTGs on AGC of Multigenerational System
Dynamics of different combinations of turbine-generators have different effects on
system frequency maintenance. A Frequency regulation study at primary and
secondary level is conducted to analyze wind plant integration with generating units
of widely different characteristics like a thermal plant, gas plant and hydro plant in
an interconnected control area.
1.6.6. Development of FGSPID Controller for Wind Integrated Power
System Load Frequency Control
Intelligent load frequency controller for wind integrated power system is developed.
Fuzzy gain scheduling is applied for intelligent control and compared with other
prevalent load frequency controllers.
1.6.7. Fault Ride-through Analysis of Frequency Responsive VSWTGs
Conventionally active power is given low priority during voltage disturbance, but
with impending grid codes for FRT and frequency response requirements from
VSWTGs, assessment of grid code compatible VSWTGs for fault ride through
capability is essential.
1.6.8. Review of Active Power-Frequency Control in Wind Integrated
Power System
A comprehensive review of active power-frequency control strategies regarding
primary and secondary frequency control is also presented, and future perspective is
also provided.
48
1.7. Outline of the Thesis
Following a comprehensive literature review of existing power systems, wind energy
technology and their integration with the grid, this report presents the findings and
potential challenges found in the existing literature. A methodological approach has
been used to investigate wind penetration effect on frequency response. After a
literature review, a suitable VSWTG modeling is conducted followed by active
power-frequency response controller development. Wind integration effect is
analyzed for system frequency control and regulation reserves.
Chapter 1 describes the motivation and significance of the study. Key technology and
research contributions are discussed in this chapter. The scope and limitations of the
study are also presented in the chapter.
Chapter 2 reviews the literature associated with Australian electricity network, its
wind resources, and frequency regulation challenge of integrating wind energy with
power grids. 14-generator NEM model is used to investigate frequency response
under contingency events and wind penetration for low load and high load scenario.
Possible solutions are proposed.
Chapter 3 reviews the frequency control and presents new challenges for AGC studies
regarding the issues of wind integration with the power grid.
Chapter 4 presents the modeling of type 4 VSWTGs suitable for frequency response
studies.
Chapter 5 presents the grid code compatible active power frequency controller model.
Chapter 6 presents the investigation of grid code compatible type 4 and type 3
VSWTG model performance and comparison with other frequency responsive model.
Chapter 7 presents frequency regulation in the multigenerational system. Various
models including single area control model for reserve response assessment,
49
intelligent load frequency controller model and fault ride through assessment of type
3 and type 4 VSWTGs suitable for frequency response are discussed.
Chapter 8 concludes this work. The contributions of this work are briefly discussed,
and possibilities of further works are highlighted.
50
1.8. References
2. IEA, (2013). Wind Technology Roadmap to 2050. Available at www.iea.org.
3. GWEC, (2012). Cumulative Forecast by Region 2012-2017. Available at
www.gwec.net.
4. Milborrow David, (2015). Onshore wind is more competitive than ever. Article.
Available at www.windpowermonthly.com.
5. BREE, (2014). Asia Pacific Renewable Energy Assessment. Available at
bree.gov.au.
6. Paton, J. (2013). Australian Wind Energy Now Cheaper Than Coal, Gas. [Online]
Available at www.bloomberg.com/news/articles/2013-02-06/australia-wind-
energy-cheaper-than-coal-natural-gas-bnef-says.
7. JRC, (2014). JRC Wind Status Report. Publications Office of the European
Union. 2015.
8. Kundur P., Balu, N., and Lauby, M. Power system stability, and control. New
York: McGraw-Hill, 1994.
9. W. W. Price, J. J. Sanchez-Gasca (2006). Simplified wind turbine generator
aerodynamic models for transient stability studies, Power Systems Conference and
Exposition, 2006. PSCE '06. 2006 IEEE PES, vol., no., pp.986, 992.
10. M. Singh, M. Vyas, S. Santoso (2010). Using generic wind turbine models to
compare inertial response of wind turbine technologies, in Power and Energy
Society General Meeting, 2010 IEEE, vol., no., and pp.1-7.
11. J. Morren, S. W. H. de Haan, W. L. Kling, J. A. Ferreira (2006). Wind turbines
emulating inertia and supporting primary frequency control, IEEE Transactions on
Power Systems, vol.21, no.1, pp.433-434.
51
12. AEMO, (2017). Black System South Australia 28 September 2016. Available at
www.aemo.com.au
13. AEMO, (2013). Wind Turbine Plant Capabilities Report, 2013. Available at
www.aemo.com.au
14. IEC Standard 61400-27-1, (2015). Wind Turbines—Part 27-1: Electrical
Simulation Models—Wind Turbines. Available at www.collections.iec.ch/std.
15. UWIG. Type 4 - Generic Wind Turbine Generator Model (Phase II), Available at
www.uwig.org:8080
52
2 CHAPTER 2
WIND INTEGRATION IN AUSTRALIAN NEM
This chapter highlights the problems associated with wind integration in the large
interconnected network. Australian electric network has been selected to describe the
wind energy resources integration in it and associated problems with it especially
regarding frequency control response issues. 14-Generator NEM model for frequency
response assessment is applied for investigating wind plant penetration led problems,
and possible mitigation method are also discussed
2.1. National Electricity Market (NEM) Overview
Blessed with diverse & plentiful renewable and non-renewable energy resources,
Australia has the distinction of being world’s ninth largest energy producer country.
Australia under the banner of National Electricity Market (NEM) owns the largest
geographical interconnected power system in the world. Transmission lines and
associated infrastructure extending approximately 51,000 km from Port Douglas in
Queensland to Port Lincoln in South Australia and across the Bass Strait to Tasmania
[1]. NEM spans Australia’s eastern and south-eastern coasts and comprises five
participating states acting as price regions in its jurisdictions – Queensland, New
South Wales (including the Australian Capital Territory), Victoria, South Australia
and Tasmania. There are approximately 270 registered generators in NEM, six state‐
centered transmission networks linked each other by cross-border interconnectors
and 16 major distribution networks for collectively supplying electricity to
consumers. Two regional reference nodes in NEM are interconnected through
transmission flow-paths called as interconnectors and consist of transmission
53
infrastructure located on each side of a regional boundary, connected by a set of high-
voltage transmission lines or cables [2].
Having a total generating capacity of around 45000 MWh, an increase of 0.03% is
forecasted for total electricity consumption in NEM region over the 20-year forecast
period (from 183,258 GWh to 184,467 GWh – giving a flat trend [3]. NEM regional
reference nodes and their interconnections are represented in Figure 2.1 by the author.
Except for Bass Link, all other interconnectors are regulated, deriving revenue by
trading on the spot market.
Figure 2-1 NEM Regional Boundaries, Their Interconnections and Generation Capacity [Data Derived
from [1], Formulated by Author]
2.2. NEM Spot Market & AEMO
NEM initiated operation as a wholesale spot market for electricity in December 1998
[1]. Electricity is exchanged across five interconnected regions through NEM acting
54
as a wholesale commodity exchange. Pool or spot market matches real time
instantaneous demand with power supply through a centralized dispatch process.
Specified amounts of electricity at specified prices is offered by generators to be
supplied to market for agreed time periods. Australian Energy Market Operator
(AEMO) scrutinizes all bids and decides the deployment of specific generators to
produce electricity according to cost-efficient methodology with the dispatching of
the cheapest generator. Spare generating capacity is kept as a reserve after matching
electricity consumption with power generation. Every 5-minute target is applied to
determine a dispatch price based on highest or the marginal bid for electricity
delivery, the spot price is determined for each NEM region by averaging six dispatch
prices over every half hour period. All financial transactions related to electricity
traded in the NEM is settled based on this spot price. A maximum spot price is known
as Market Price Cap, and a minimum spot price, known as Market Floor Price is set
according to the National Electricity Rules. Market Price Cap was set at $13,800 per
megawatt hour, and Market Floor Price was set -$1,000 per megawatt hour in January
2015 [4]. Market Price Cap highly rewards generation supply for meeting demand in
need while negative Market Floor Price strongly encourages power reduction by all
generation, including the wind in case supply exceeds demand. These two price
settings are reviewed every four years by the Australian Energy Market
Commission’s Reliability Panel to safeguard the NEM reliability standard.
Market clearing prices change continuously in different regions with higher cost
generators called upon to meet higher demands which vary throughout the day. This
generator ‘merit order’ means that the last generator required to meet market demand
(called the marginal generator) sets the market clearing price. Generating plants are
paid through customer revenue recovered by AEMO through retailers. A commercial
55
tariff is paid by consumers to the retailers who manages energy purchases on behalf
of their customers. There is are financial risks connected with volatile spot prices
during the trading period which is managed by NEM participants through financial
contracts locking in a fixed electricity price for any production or consumption at a
given time in the future. Besides obligatory regional gross-pool spot energy
operation, NEM operation is also dependent upon Frequency Control Ancillary
Services (FCAS) markets that provide a security-constrained dispatch order every
five minutes for five interconnected regions. NEM’s scheduling and cost allocation
procedures mandate participation of all generators greater than 30MW in size. Both
the transmission system and energy market fall under the direction of the Australian
Energy Market Operator (AEMO).
AEMO as transmission system and retail electricity market operator is responsible
for NEM reliable and secure operation, including maintaining required electricity
reserve. AEMO was established by the Ministerial Council for Energy (MCE), now
the Standing Council for Energy and Resources (SCER), as an independent
operational entity by 1 July 2009. Operating on cost recovery basis, AEMO solely
recovers its operating costs through market participants, and network service
providers paid fees.
As presented by Author in Figure 2-2, AEMO operates alongside three major
organizations. First one is the Australian Energy Regulator (AER), which is
responsible for economic regulation and national rules compliance in the NEM.
Second is the Australian Energy Market Commission (AEMC), which makes the
rules governing the regulation of the energy markets. The third is the Standing
Council for Energy and Resources (SCER) which is responsible for developing
policies related to electricity markets. AEMC is responsible to COAG through SCER
56
while AER is responsible to Commonwealth Government. AER and AEMC interact
with each other through a memorandum of understanding. AEMO and AEMC
collaborate on important issues like NEM future security. Both AEMO and AEMC
can implement policy changes when appropriate and refer them to COAG energy
council. The NEM operates within the framework of National Electricity Rules under
joint legislation enacted by participating states. These rules are maintained and
developed by AEMC and enforced by the AER.
Figure 2-2 AEMO Interactions with Other Regulators [Formulated by Author]
AEMO controls NEM via two identical control centers in different regions overseeing
the performance of all generation connected to the NEM. In the case of any
transmission system limitations or demand variations, AEMO may issue market
notice for generation adjustment. In the case of generation deficiency, AEMO can
instruct for load cut off to some customers to maintain a balance between generation
and consumption. Industry investors and customers can access a range of planning
information documents including benchmark annual National Transmission Network
Development Plan (NTNDP) report published by AEMO.
57
2.3. DRIVERS OF WIND ENERGY SYSTEM
DEVELOPMENT IN AUSTRALIA
2.3.1. Renewable Energy Target (RET)
With increasing environmental change concerns, limited fossil fuel reserves and
increasing electrical energy consumption, renewable energy sources (RES)
development has become a global necessity. There has been a tremendous increase
in generated megawatts by these RES all over the world including Australia. There
were 43 countries with Renewable energy target (RET) in 2005 which increased to
164 countries in 2015 [5]. Australia government is supporting the green revolution
by having set up the RET. Australia’s Renewable Energy Target is practiced as
Large-Scale Target (LRET) is centered on fixed quantity and market price, and
Small-scale Scheme (SRES) centered on a variable amount and fixed price. LRET
mechanism is based on granting certificates to eligible renewable generators that
energy retailers must surrender in proportion to their electricity sales and a fixed
LRET trajectory. Renewable energy participates in the NEM in the same way as the
non-renewable generation, but eligible plants receive additional income that
effectively guarantees a certain volume of large-scale renewable energy in the
market. The new target for large-scale generation of 33,000 GWh in 2020 means
about 23.5 percent of Australia’s electricity generation will be from renewable
sources. It targets to minimize the difference between the renewable energy
generating cost and the cost of generating electricity from traditional fossil fuel
sources and contribute towards government’s emission reduction target. A
comparison of renewable energy target of Australia with some other countries is
presented in Table 2-1.
58
Table 2-1 Comparison of Renewable Energy Target for some Countries with Australia
Country Renewable energy target (share of Achieved so far (share of

electricity in 2020 unless specified) electricity in 2012 unless
specified)
Australia [6] 23.5% 53% (2015)
Brazil [6] 23% by 2030 85%
Canada [7] Province-level targets 65%
(projected national share: 70%)
China [8] 15% final energy by 2015 Tech- 21% (2013)
specific targets (e.g. 200 GW wind by
2020)
Denmark [9] 100% by 2035; 50% by wind by 2020 40% (2016)
France [10] 27% by 2020 and 32 % by 2030 23% (2015)
Germany [6] 40-45% by 2025 28.2% (2015)
India [11] 57% by 2027 33%
S-Africa [12] 21% by 2030 5%
U.K. [13] 30% by 2020 4.23%
USA [14] 30% by 2030 8.3%
EU [15] 20% by 2020 16.7% (2015)
2.3.2. Wind Resources Availability
Figure 2-3 Australia Wind Resources [16]
Wind energy generation at large scale boast of various striking distinctiveness making
it a viable player in Australian energy market. As shown in mesoscale wind map in
Figure 2-3, Australia has some of the world’s best wind resources along its south-
western, southern and south-eastern margins which lie in the path of the westerly
59
wind flow belts between latitudes 35° and 50°known as the ‘roaring 40s’. Westerly
wind belts are characterized by strong, comparatively constant wind conditions, with
average wind speeds in a surplus of 6 meters per second (m/s) and, in places, more
than 9 m/s (shown in orange to red colors).
With first time utilization for electricity generation in 1994, wind power industry has
developed and expanded rapidly over the years with an average annual growth of
69.5 percent. Being source of low to zero-emissions and low cost proven technology,
providing cheap rollout on a large scale, wind power has dominated other renewable
sources for government support under LRET scheme. Approximately 38 percent of
all renewable capacity installed since 2000 is attributed to wind energy which has
engrossed over $5 billion in investment directly in Australia since 2001.
2.3.3. Cost Effective, Competitive & Established Technology
Upfront capital costs and lifetime running costs make up leveled generation cost and
are expressed in cost per unit energy produced. The building of new onshore wind
energy leveled generation Costs lies in-between the cheapest option, Combined Cycle
Gas Turbines (CCGTs) and those of the next cheapest option, that of coal. Offshore
wind farms cost 50% more than that of onshore farms, but experts have predicted cost
reduction with further improvements in technology.
Figure 2-4 shows the relationship between levelised costs of generation (LCOE) and
capacity factor for wind generators as per Australian Energy Technology Assessment
2013 (AETA) resourced values. Distribution of wind resource in the marketable zone
is shown to have capacity factors distributed between 20-50%. An average capacity
factor of 33% makes Australian wind farms an attractive power generation option.
LCOE for wind power is shown to reduce moderately over the projection period,
60
stabilizing from 2020 onwards. The LCOE also is shown to reduce from 153$/MWh
at 20% capacity factor to 70$/MWh at 50%.
Figure 2-4 Levelised Cost of Energy (LCOE) Generation and Capacity Factor for Wind Generators [17]
In 2015, Coonooer Bridge Wind Farm near Bendigo in Victoria, Australia supplied
electricity at the cost of 81.50 AUD / MWh, while Ararat Wind Farm (80 MW) agreed
to provide power for 87 $/MWh and Hornsdale Wind Farm (100 MW) agreed to
provide power at 92 $/MW. Power supply cost came out to be 143 $/MWh for a new
coal-fired power plant or 116 $/MWh when supplied from a new natural gas powered
station in 2013[18]. Bloomberg report listed the wind as cheaper electricity producing
the fuel than fossil fuels for Australia which is world’s biggest coal exporter. The
onshore wind generation costs fell in 2014 and, now it is approximate $71/MWh on
a global basis [19-20]. The prospective plunge in LCOE of wind is expected in the
near future due to the emergence of low cost but high technology wind turbine
manufacturer.
61
2.3.4. Low Carbon Footprint
Carbon footprint refers to emissions during the lifecycle of a project consisting of
stages of plant construction; fuel extraction, processing and their transport if
applicable; and plant ultimate decommissioning and disposal. Research data wind
energy projects having very low carbon footprints that too only during the
construction stage only. In comparison to approximately 500 and 900gCO2eq/kWh
for existing gas and coal plant respectively, wind carbon estimate is only
15gCO2eq/kWh [21]. As per CSIRO’s LCOE cost comparison study for the year
2020, the wind is the lowest cost technology with the price of 58.4 $/MWh as shown
in Figure 2-5 Without a carbon price the fossil fuel technologies are lower cost than
solar. With reduced carbon footprint, wind and solar based technologies are less
expensive than the fossil fuel based technologies.
Figure 2-5 LCOE Cost Comparison with Fuel Technologies [22]
62
2.4. Wind Energy and Australian Market Outlook
Figure 2-6 Per Capita Wind Capacity by Country and the Key US and Australian States/Territories [23]
By country, Australia currently ranks 11th in the world for wind generation per capita
ahead of countries like China and France. Denmark leads the world, followed by
Sweden, Spain, and Germany [23]. At a jurisdictional level, Australia’s wind
generation is heavily skewed towards states like South Australia and Tasmania,
which have some of the highest per capita wind generation in the world alongside
leading US states like Iowa and Texas as represented in Figure 2-6. When looking at
a jurisdictional level, Iowa has the highest per capita capacity of wind and solar, with
South Australia ranked second in the world. South Australia currently has around 39
percent of its total generation supplied by the wind and solar. The South Australian
renewables experiment is more significant given most other high renewables
penetration regions and countries - Iowa, Denmark, and Germany – are much more
integrated into larger grids with complementary (dispatchable) generation
technologies. The South Australian grid is partially constrained, connected to
Victoria by two transmission lines which allow it for sourcing a maximum of around
63
20 percent of peak load from Victoria. By contrast, Denmark has interconnections
that allow it for sourcing its entire peak load from other countries.
As per Australia Renewable Agency report [24], the share of wind energy in total
electricity generation is projected to increase by 12.1 percent of Australia’s electricity
generation in 2029–30 after stabilized RET. After hydro with 5% share, wind power
has the second largest share with 2.0% of Australia’s total primary energy
consumption. Among renewable energy sources, wind power had the fastest growth;
increasing on average by 67% per year since 2000 [24]. Australia’s first large-scale
grid-connected wind farm (at Crook well, New South Wales) in 1998 comprised eight
600 kW wind turbines each with a rotor diameter of 44 meters for a combined energy
output of 4.8 MW [24]. Today most onshore wind turbine generators have a capacity
of 1.5 to 2 MW; the largest wind turbines – designed for offshore sites – have a
capacity of 5 MW and rotor blades up to 60 m long (120 m rotor diameter).
Figure 2-7 Wind Generation Capacity Summary for Australia 2015
In 2015, Renewable energy provided 14.6 percent of Australia’s electricity, enough
to provide power for the equivalent of approximately 6.7 million average. The
amount of wind power in Australia has doubled in the past five years [25]. The
contribution of wind power grew by over 20 percent in 2015, while the demand for
64
power across the country increased slightly, the first rise following six years of
decline. The author represents the wind generation capacity summary for Australia
for 2015 in Figure 2-7 while the percentage of wind capacity by the state is shown in
Figure 2-8. In 2015, NEM had around 11802 GWh of wind power generation
accounting for 33.7% of renewable generation and 4.93% of total electricity
generation [25]. By the end of 2013, renewable energy projects constituted 48 percent
of total proposed new capacity in the investment pipeline and 58 percent of total
capital expenditure [20]. Wind-powered generation projects represent the greatest
share of planned capacity for renewable projects accounting for 42 per cent of total
planned capacity and 53 per cent of the value of proposed electricity generation
projects in the investment pipeline.
Figure 2-8 Percentage of Wind Capacity by State [Modified from 26]
Wind energy is projected to be the second fastest growing energy industry after
geothermal over the outlook period to 2029–30, reflecting the relatively low base of
which it is growing and the relative maturity of the technology compared with other
renewable energy sources. It is projected to overtake hydro electricity production
within the outlook period, to become the largest renewable source of electricity
generation in Australia [24]. At the end of 2015, there were 76 operational wind farms
65
in Australia, totaling 4,187 MW in capacity [25]. Approximately 500 MW of wind
capacity is under construction until May 2016, which will come online in 2017. 8057
MW wind capacity has been approved for construction while 6653 MW is in approval
stage [27].
Figure 2-9 South Australia Wind Power vs. Demand Variation on 22 May [28]
Among all states, South Australia owns the leading installed capacity of
approximately 1.5 GW representing highest penetrations in any electricity market.
Wind and solar are becoming new base load power after the closure of South
Australia largest coal-fired generator. The Wind has accounted for up to 49 per cent
of total generation in the South Australia state in May 2016. On 22 May 2016, wind
energy in South Australia hit 100% of demand and exceeded local demand for more
than 10 hours, from 1.40am to just before midday (11.55am), with a peak of 120
percent of demand at 4.30am as shown in Figure 2-9 [28]. In May 2016, the seven
large-scale wind farms in NSW delivered a record output, with a combined capacity
factor of 56 percent, more than two of the biggest coal generators in the state. The
Woodlawn wind farm produced a capacity factor of 62 per cent while only two of the
66
big five coal generators, the Mt Piper and Bayswater coal generators, produced at a
higher capacity factor (67 and 66 per cent respectively) [29].
2.5. Challenges to NEM Operation Due to Increased Wind
Penetration
Despite having sufficient feasibilities for wind power integration in Australian NEM,
it still lacks various technical and commercial aspects that require further
investigations to test the system adequacy for reliable and secure operation.
Intermittent wind power generation and asynchronous generators are two key
characteristics of wind generation technology affecting its integration in NEM grid
and spot market. AEMO does not currently identify NEM-wide challenges. Each
NEM region has a different generation mix, network configuration, and demand
characteristics, which lead to different challenges or different timing. However, the
future increment in wind power penetration could bring some adverse operational
challenges in the NEM. These challenges could be in the form of merit order effect
variation in pricing market, reduced system inertia, reduced interconnector
limitations and retarded network limitation.
2.5.1. Merit Order Effect and Reduced Spot Prices
By 2020, additional wind generation of around 8.88 GW is expected in NEM region.
As per AEMO’s National Transmission Network Development Plan (NTNDP), NEM
will have 11.5 GW of total installed wind generation capacity [30]. The Australian
Wind Energy Forecasting System (AWEFS) is used by AEMO to predict the power
from all the wind farms and has a normalized mean absolute error of less than 1.5%
in the 5-minute look ahead [30]. Currently, maximum wind generation is forecasted
for timeframes ranging from five minutes ahead to two years ahead to include wind
67
generation in the central dispatch process. From being considered as non-scheduled
generators, wind farms in Australia are now both scheduled and semi-scheduled and
included in forecast generation to allow generation output outside the central dispatch
process to be considered in the load–generation balance [31].
Merit order effect in NEM spot market infers that the marginal generator meeting
market demand sets the market clearing price or in simple terms refers to lowering of
wholesale electricity price by subsidized generators by adding reduced short run cost
generation to total supply or merit order. Wind farms bid their output in NEM as price
takers with low marginal cost. Wind farms get a connection to the grid through user
pays scheme while their access to the market is not always guaranteed [32]. Subjected
to the available wind and any network limitation, present wind farms in Australia
always runs at full capacity. Without storage options, wind farm operators typically
dispatch electricity into the market regardless of price. They are they incentivized by
the RET for every unit of electricity they produce. Under favorable wind conditions,
the wind farms output displaces output from coal or gas plants by bidding their low
marginal cost to clear the market, lowering spot prices through the merit order effect.
When the wind is not blowing, electricity can be sourced from hydro, coal, gas plants
or other renewable energy generators. Effect of wind farm effect on merit order of
generators is shown in Figure 2-10. As per a Deloitte study [33] on merit order effect
in South Australia, increased wind penetration in spot market results in backing off
the dispatch of fossil fuel based marginal generator that would have set the marginal
price. Thus, during periods of high wind, the output of marginal generators reduces
significantly resulting in reduced traded electricity which ultimately effects the
economics of generators in the form of reduced prices received for baseload and mid-
merit plant.
68
Figure 2-10 Merit Order Effect [33]
The actual impact of generation technology on prices is driven by a combination of
the demand profile, power supply availability, network topography, gas supply
dependence and carbon pricing. Wind farms quite often oversupply the market and
hence cause downward pressure on the wholesale electricity price as shown in Figure
2-11. During transmission system constraint in the form of generation oversupply,
available wind farm output needs curtailment as all generation cannot be placed on
the market.
Figure 2-11 Market Price Variations in South Australia on 3-4 Nov. 2014 When Wind Generation Enters
Spot Market [33]
69
2.5.2. Reduced Inertia and High Rate of Change of Frequency Issues
All synchronized rotating generators and motors constitute inertia of a power system.
Higher the system inertia, lesser is the frequency volatility due to disturbance. The
amount of a synchronous generation’s inertia depends on its size and design and is
expressed in megawatt seconds (MW.s). It is very difficult to maintain frequency
within acceptable limits for low inertia based power system as it will slow down or
speed up very quickly. Like demand level, power system inertia is also only an
observed characteristic, and currently, AEMO has no control over it in any form.
AEMO presently operates the power system around the requirements that arise from
the inertia levels that are present. With synchronous generators (especially thermal
power plant) providing the majority of energy in NEM, each NEM region had
sufficient inertia adequacy without any effect on system security needs, however
impending renewable energy targets has increased the probability of reduced power
system inertia due to the displacement of conventional synchronous generation,
particularly in South Australia and Tasmania. Current system inertia in South
Australia is around 18725 MW.s, but a low inertia value of around 2000 MW.s has
also been observed [34]. Although some non-synchronous generation, like wind, also
has rotating turbines, these technologies are increasingly connected to the power
system via power electronic converters, so the mechanical movement is decoupled
from the power system. By the year 2010, system inertia would be below acceptable
levels for 30-40% of the time in Tasmania, and 30% of the time in South Australia
[30]. Victoria also experiences low inertia sometimes but can rely on inertia from
other NEM regions due to its strong interconnections.
The initial rate of change of frequency (ROCOF) after generating unit MW loss ∆ )
is given as:
70
∆
∗ 2.1
Or
∆
∗ f 2.2
Where IR is system inertia, H is inertia of conventional generating unit and S is the
MVA rating of the conventional generating unit. ROCOF management is critical to
power system frequency regulation within the frequency operating standard [31] and
for maintaining the power system in a secure operating state. The amount of inertia
required to maintain an ROCOF under different contingency is proportional to the
contingency size. Lower inertia leads to a higher ROCOF than higher inertia system.
That means the frequency changes faster following a disturbance in a power system
with less synchronous generation, and this could result in the loss of additional
generation or load to arrest the frequency deviation when it occurs. Reduced system
inertia can challenge the effectiveness of existing frequency control mechanisms,
which can reduce under high ROCOF. The higher ROCOF will also require
stabilizing control systems to respond more rapidly to contain the change. For
example, for a contingency event resulting in an ROCOF of 1 Hertz per second
(Hz/s), the frequency drop from 50 Hz to 49.5 Hz (the limit of the FOS for credible
contingencies), would take 0.5 seconds. An ROCOF of 2 Hz/s would halve this time,
meaning action would be required within 0.25 seconds to prevent the system
frequency from breaching the FOS for a credible contingency event. High ROCOF
will lead to additional tripping for the same size imbalance. Relays and protection
schemes on generators and feeders have inherent delays and so may not respond
quickly enough to high ROCOF. Critical schemes such as under frequency load
shedding (UFLS) could become compromised in maintaining the frequency operating
standards (FOS). Increased wind penetration would make management of ROCOF
71
after contingency events more challenging. Currently, an inconsistency exists in the
NER requirement for generating units to remain connected through an event where
ROCOF reaches 1 Hz/s, but there is no system standard for power system operation
whereby the ROCOF during any contingency event is to be maintained at 1 Hz/s or
better. Low levels of inertia in the power system are likely to increase the frequency
deviations in the five-minute dispatch period that are controlled by regulation FCAS
in NEM.
2.5.3. FCAS Issues
Frequency control ancillary services (FCAS) in power system is utilized to satisfy
load demand through generation supply. Frequency control ancillary services (FCAS)
is indispensable for the secure operation of large interconnected power system like
NEM. FCAS market also complements AEMO's centrally dispatched energy market.
FCAS services are purchased from market participants by AEMO for adequate power
system frequency control. These services include regulating FCAS services for
normal operating conditions and fast responding contingency FCAS services
following any contingency events. 130 MW is procured for raise FCAS services
while 120 MW is procured for FCAS lower services within a 5-minute dispatch
interval. NEM is managed with this amount of variability between 5-minute
dispatches of the rest of the generators. An accumulated time error of greater than ±
1.5 seconds may require additional support. The current calculation of NEM global
contingency FCAS requirements is determined by the dispatch algorithm and
considers both the size of the largest contingency and the power system demand. All
types of large FCAS requirement is calculated as FCAS requirement= Contingency
risk (MW change due to the generator or load loss) – Load relief (demand change due
to frequency deviation). Load relief factor is 1.5% for the mainland in NEM while it
72
is 1% for Tasmania [35]. The level of fast performing response capabilities or
ancillary service requirements necessary to manage frequency in NEM is determined
by the frequency operating standards. As per NEM frequency operating standards
[35] for normal system operation, the frequency must be maintained within the
normal operating frequency band (49.85 Hz to 50.15 Hz in both Tasmania and the
NEM mainland) for no less than 99 percent of the time. In the case of a credible
contingency event, the system frequency should not exceed the normal operating
frequency excursion band for more than five minutes on any occasion. Following
either a separation or multiple contingency events, the system frequency should not
exceed the normal operating frequency excursion band for more than ten minutes. In
the case of time delay for FCAS activation, synchronous generators inertia is the first
to slow down ROCOF and depending upon contingency sizes, even able to limit
frequency within specified band till FCAS are employed. In a power system with a
greater proportion of variable, non-scheduled generation like wind, operating the
power system over the five-minute cycle would require a greater reliance on
regulation FCAS. In a lower inertia based power system having high ROCOF,
frequency deviations will reach the threshold of the contingency frequency range
more quickly, reducing the required operational response time of stabilizing control
systems, and potentially increase the level of response services required to return to
normal operating conditions [31]. Presently, AEMO always assumes mainland inertia
levels to be high enough not to affect the calculation of mainland Frequency Control
Ancillary Service (FCAS) requirements and ROCOF management. Inertia levels
within Tasmania are much lower than the mainland, so managing Tasmanian
frequency is more technically challenging under the effects of increasing wind
generation. Historically synchronous generations have provided regulation and
73
contingency FCAS in NEM, but with the exit of base load dispatchable generators
from islanding prone South Australia region, FCAS availability at the local level will
be more challenging with increasing wind penetration. FCAS regulation service
required capacity in NEM will increase by approximately 20% by 2020 due to
increased renewable generation [36] which is likely to increase the magnitude of a
minute to minute generation deviations. In the case of insufficient FCAS availability,
NEM frequency maintenance within the required standards will be difficult for
AEMO, and NEM may collapse under big contingency events at worst scenario.
AEMO must review the procurement of Regulation ancillary services, especially in
smaller systems with high wind penetration like in, Tasmania and South Australia to
ensure ongoing control of power system frequency within required limits.
2.5.4. Network Infrastructure Issues
As per NEM rules, connection costs are part of any new generator including wind
farm construction cost while customers fund shared transmission network cost. The
cost of shared transmission system is indirectly affected by supply to the customer
who is dependent upon location and type of generators. Attributing additional shared
transmission cost resulting from increased wind integration in NEM is challenging
for AEMO regarding market benefit economy. Most of the wind farms in Australia
are in remote locations that will require significant transmission investment for
improved transmission network infrastructure otherwise around 35% and 15% of the
wind energy in Victoria and South Australia respectively could be curtailed due to
network limitations [37]. Therefore, balancing wind penetration around the NEM
necessitates robust interconnections to yield geographical diversity benefits. Network
augmentation costs may be deferred by correlating output generation of small wind
farms connected to existing distribution networks with local demand thereby
74
reducing network losses. The cost of connecting a 5 MW wind farm to the NSW 330
kV grid could fall from $2,500/kW to $150/kW to connect four 50MW wind farms
sharing the same substation [38]. This highlights the possibilities for connection costs
reduction for remote wind farms by capturing economies of scale in network
augmentation.

2.5.5. Interconnectors Performance Issues
Power import- export in NEM interconnected regions through interconnectors are
limited by transient or voltage stability due to the contingent trip of the largest
generating unit and potential thermal over-loadings. South Australia’s transmission
network is connected to the rest NEM via the Murraylink (DC link of 220 MW
transfer capability) and Heywood Interconnectors (AC-link upgraded from 460 to
600 MW). The Heywood Interconnector is the only link that provides a synchronous
connection between SA and the rest of the NEM. DC transmission line Basslink
which connects Tasmania to mainland Australia is unable to transfer FCAS between
the mainland and Tasmania under certain conditions most notably when it reaches
the maximum export limit or maximum import limit or when it is transferring power
within a ‘dead-zone’ between -50MW to 50MW. Increased wind generation in states
like South Australia and Tasmania has increased energy imports/export through
interconnectors putting more thermal stress, especially on Heywood interconnectors
during the period of high demand and low wind generation and vice versa. Significant
new wind generation can reduce existing interconnector transfer limits, particularly
under conditions of low demand and high wind speeds when wind generation forms
a large percentage of the generation mix [37]. Any contingency at interconnector
level may have high contingency impact on NEM spot market and FCAS market
operation due to high ROCOF. A combined effect of spot market price variation and
75
interconnectors dependence was observed recently during a contingency event which
occurred on 1st November 2015 when South Australia was islanded for 26 minutes
due to a transmission line tripping and Heywood interconnector being unavailable
due to upgrading works. Sudden fall of interconnector capacity resulted in FCAS
procurement within South Australian region resulting in price hikes.
76
2.6. Wind Integrated Network Frequency Response Assessment
In this section, an augmented form of 14-Gen NEM model is used as the test network
to study wind integration effect on frequency response performance in area 5 which
can be held representative of South Australia (SA). MATLAB’s Simscape has been
used as the simulation software [40]. Several case studies were performed to analyze
the technical problems due to wind farm integration in low load scenario and high
load scenario under low levels of conventional synchronous generation in operation.
Following the frequency response analysis under low and high load scenario, the
solution to improve the frequency response is assessed based on frequency controller
based wind farm and synchronous condensers. One of the key findings is that there
are minimum issues with the SA power system security management if SA is
connected to the remainder of the NEM via the Heywood Interconnector and
sufficient synchronous generation is available in the SA power system. Increased
wind penetration with synchronous generator losses is making frequency operating
standards management difficult due to increasing ROCOF and higher frequency
nadir.
2.6.1. Test Network
A 14-generator NEM model is a simplified model of the eastern and southern 50 Hz
Australian electrical networks, which was originally proposed for small-signal
stability studies [39]. The schematic diagram of the original 14-generator model of
the NEM is shown in Figure 2-12. Areas 1 to 5 represents Tasmania, New South
Wales (NSW), Victoria (VIC), Queensland (QLD) and South Australia (SA),
respectively. In the original model, there are 14 generators, 5 Static VAR
Compensators (SVCs), 59 buses and 104 lines with voltage levels ranging from 15kV
to 500kV. The excitation system (AVR) and power system stabilizer (PSS) of
77
generators are adopted from reference [39]. Also, it is assumed that all thermal and
hydro power plants have a standard steam turbine governor (i.e. IEEEG1) and hydro
turbine governor (i.e. HYGOV), respectively. Test network is studied when DFIG
based wind farms are integrated into Area 5. Area 5 has three lumped generators;
G503 at bus 503, G502 at bus 502 and G501 at bus 501. Generator 503 and 502 are
connected at 15 KV while generator 501 is connected at 20 KV. All related data are
provided in the appendix.
2.6.2. Low Load Scenario
Table 2-2 Scenario 1: Low Load
Summary for low load Converged: The load flow converged in 4 iterations
P (MW) Q (MVAR)
Total generation 15116.22 -569.215
Total PQ load 14807 1595
Total Z shunt 73.06775 -3313.59
Total ASM 0 0
Total losses 236.1522 1149.37
2.6.3. Assumptions Used in simulations
 DFIG based wind turbines are used in wind farm model.
 Aggregated generating machines including wind turbines inside a wind farm are
used.
 Test Network is modeled under light load scenario as given in Table 2-2 and
high load scenario which is given in Table 2-13.
 Generator and interconnector contingency simulations are conducted for area 5
which are held representative of South-Australia.
78
Figure 2-12 14-Generator NEM Model
79
Table 2-3 Load Flow in Low Load Scenario
Power Station/Bus No. of Units P(MW) Q(MVAR)

B101 2 0.001 -208.239
B201 4 2160 -141.833
B202 3 1380 -33.9972
B203 3 1410 30.82429
B204 4 1675.219 -141.814
B301 6 3330 84.09879
B302 2 760 50.08303
B401 3 960 -107.759
B402 2 580 -10.1451
B403 3 960 38.08328
B404 3 651 -20.0589
B501 2 560 -115.507
B502 3 540 19.93927
B503 1 150 -12.8902
Total 41 15116.22 -569.215

Case 1: Generator Contingency Analysis (Low Load Scenario)

Table 2-4 Area Frequency Response Indicators When Generator 503 (150 MW) is Isolated at 10s and Wind
Farm Connected at Bus 509
Frequency Nadir
Generator ROCOF (Hz/s) Power Flow
(Hz)
502 -0.5 49.90
501 -0.52 49.93
ROCOF changes from -0.5 to 0.4 Hz/s in 7 seconds soon after

isolation at G502 terminal. For G501, ROCOF changes as -0.52
to 0.15 in 5 seconds.
302 -0.12 49.99 100 MW export

changes to 230 MW
import from area 3-5
301 -0.12 49.98
ROCOF changes from -0.12 to 0.07 Hz/s in 1.2 seconds soon

after isolation at G503 terminal. For G301, ROCOF changes as
-0.125 to 0.07 in 1.2 seconds.
80
Wind farm integration effect on frequency response indicators namely, Rate of
Change of Frequency (ROCOF) (Hz/s) and frequency Nadir at Area 5 generator
terminals is investigated in this study. In various simulation studies, frequency
response indicators are observed at Area 5 and Area 3 generators when one of the
generators are isolated from the network, and equivalent rated DFIG based wind
farms are integrated into area 5. The simulations were performed in MATLAB, and
ROCOF results were determined by taking the derivative of the measured frequency
at generators terminal in the network. Values of ROCOF as observed under different
contingency are indicated in Tables (2-4 – 2-9).
Frequency Nadir
(Hz)
503 -2.14 49.68
501 -2.75 49.47

-2.75 to 2.33 in 0.6 seconds.
185 MW import
302 -0.43 49.91
changes to 330 MW
export from area 3-5
301 -0.43 49.99
ROCOF changes from +0.62 to -0.43 Hz/s in 0.75 seconds soon

-0.437 to 0.37 in 7 seconds.
81
Frequency Nadir
(Hz)
503 -2.58 49.70
502 -2.73 49.54
ROCOF changes from -2.58 to 0.94 Hz/s in 9 seconds soon

-2.73 to 1.29 in 0.6 seconds.
200 MW import
302 -0.28 49.90 changes to 350 MW
export from area 3-5
301 -0.3 49.90

-0.3 to 0.25 in 1 seconds.
Table 2-7 Area Frequency Response Indicators When Generator 503 (150 MW) is Isolated at the 20s, and
Generator 501 (560 MW) is Isolated at 10s, 1146MW Wind Farm Connected at Bus 508
Frequency Nadir
(Hz)
502 -2.68 49.61 Export of 500 MW to
50 MW after 1st
301 -0.37 49.97 contingency and then
75 MW after 2nd
302 -0.31 49.95 contingency.
For G502, ROCOF changes from -2.68 to 1.2 Hz/s in 0.6 seconds soon after 1st contingency.
ROCOF changes as -0.85 to 0.4 in 0.6 seconds after 2nd contingency. For G302, ROCOF
changes from -0.3 to 0.2 Hz/s in 0.9 seconds while for G301, ROCOF changes as -0.3 to 0.2
in 0.7
82
Case 2: Generator Contingency and Area 5 Separation Due to Interconnector
Contingency Analysis (Low Load Scenario)
The study investigated the separation of area 5 from rest of NEM due to
interconnector loss. The study considers the frequency stability of area 5 under the
condition of reduced inertia. Frequency response indicators are observed at area 5
generators when one of the generator and area 3-5 interconnector is isolated from the
network, and equivalent rated DFIG based wind farms are integrated into area 5.
Higher values of ROCOF more than 1 Hz/s is possible in case of big contingency. It
is therefore proposed that a higher ROCOF specification is defined for South
Australia which is reflected as area 5 in the current model. Proposed solution should,
of course, be preceded by a study involving the effect of changing ROCOF values on
generating unit characteristics.
Table 2-8 Area Frequency Response Indicators When Generator 503 (150 MW) is Isolated at 10s and area
3-5 interconnector is Isolated at 20 s; Wind Farm Connected at Bus 508
Frequency Nadir
(Hz)
501 -0.68 49.27
502 -0.65 49.28
G502 terminal ROCOF changes from -0.3 to 0.05 Hz/s in 3

seconds soon after G503 isolation, then -0.65 to 0.18 Hz/s in 3 s
after interconnector isolation. For G501, ROCOF changes as -
0.33 to 0.05 in 3 s at first isolation and it changes from -0.685 to
0.18 Hz/s in 3 s. 100 MW export from
area 3-5 till
interconnector is
302 -0.28 50.04 isolated.
301 -0.3 50.04
ROCOF changes from +0.125 to -0.05 Hz/s in 0.8 seconds. For

G301, ROCOF changes as +0.125 to -0.03 Hz/s in 1.1 seconds.
83
Table 2-9 Area Frequency Response Indicators When Generator 502 (540 MW) is Isolated at 10s and area
3-5 Interconnector is Isolated at 20 s, 1080 MW Wind Farm Connected at Bus 508
Frequency Nadir
(Hz)
503 -2.18 49.74
501 -2.76 49.5
G503 terminal ROCOF changes from -2.18 to 2.12 Hz/s in 1.7

seconds. For G501, ROCOF changes as -2.76 to 2.6 Hz/s in 0.6 s Initial import of 500
MW, changes to 100
302 -0.5 49.95 MW export and

zero after
interconnector loss.
301 -0.6 49.92
ROCOF changes from +0.6 to -0.5 Hz/s in 0.6 seconds. For

G301, ROCOF changes as +0.5 to -0.6 Hz/s in 0.6 seconds.
2.6.4. Simulation Discussions
The frequency operating standards (FOS) requires the frequency to be maintained
within the range 49.85 to 50.15 Hz for 99% of the time. This standard is valid for
both the mainland regions and the Tasmania region during periods of no contingency
events and no load events. Larger deviations are permitted within the range 49.75 to
50.25 Hz for no more than 1% of the time. For an island within the mainland regions
or an island within the Tasmania power system, the frequency operating standards
require that, during periods when there are no contingency events and no load events,
the frequencies be maintained within the ranges 49.5 to 50.5 Hz and 49.0 and 51.0
Hz respectively.
84
Figure 2-13 Area 5 Generator Terminal Frequency after Isolation of Gen 503 at 10 S
Figure 2-14 shows the frequency observed at three generator terminals when
generator 503 of 150 MW is disconnected at 10 s. An equivalently rated wind farm
is connected to the network. In the case of 150 MW contingency and 9.37% wind
penetration in area 5, frequency nadir observed at generator 502 (540 MW) is more
than generator 501 (560 MW), and frequency at both generators terminal is within
NEM normal frequency range of 49.75-50.25 Hz. With low generator contingency
and interconnector contingency, NEM frequency standard is violated as frequency
goes below 49.5 Hz as shown in Figure 2-15. These results match the South
Australian Jurisdiction which has recommended AEMO for permitting larger
frequency variations (47-52Hz) as frequency can go beyond 49.5 Hz – 50.5 Hz for
the credible loss of the Heywood Interconnector. Frequencies observed at area 5
buses is shown in Figure 2-16. Even though wind farm role is not clear in these
scenarios, frequency at bus bar 508 is observed within limits where 150 MW wind
farm and 490 MW of load is connected.
85
Figure 2-14 Area 5 Generator Terminal Frequency after Isolation of Gen 503 at 10 S and Interconnector
Isolation at 20 s
Figure 2-15 Area 5 Bus Frequency after Isolation of Gen 503 at 10 s.
Figure 2-16 Area 5 Bus frequency after Isolation of Gen 502 at 10 s.
86
Figure 2-17 ROCOF Observed at Generator 503 when Generator 502 is Isolated at 10 s
Figure 2-17 shows the frequency observed at area 5 generators when generator 502
with all units amounting to 540 MW is isolated from area 5, and an equivalent rated
wind farm (540 MW) is also integrated into the area. Other highest rated generator
has frequency drop up to 49.47 Hz for the very small duration, and then it gradually
settles around 49.9 Hz within 10 s. This simulation result again supports the call for
frequency regulation standard modification in regions like South Australia and
Tasmania. It should be noted that these simulation results have been obtained without
implementing secondary control in the area. In all the contingencies above 500 MW,
we observed ROCOF also breaching the standard of 1 Hz/s for around 1-2 s. An
example of ROCOF variation is shown in Figure 2-18. This result again supports the
call for ROCOF standard revision in some of the NEM regions.
Table 2-10 Simulation Results for Low Load Scenario for Area 5 Generator 501
Generator % WIND- ROCOF FNADIR MW LOSS (Gen.+ ACTIVE

Penetration (Hz/s) (Hz) Interconnector) SYN UNIT
G501 0 -0.001 49.91 150 5

G501 9.375 -0.685 49.27 350 5
G501 9.375 -0.52 49.93 150 5
G501 27.135 -2.75 49.47 540 3
G501 42.68 -2.76 49.5 740 3
87
Table 2-11 Simulation Results for Low Load Scenario for Area 5 Generator 502
Generator % WIND- ROCOF FNADIR MW LOSS (Gen.+ ACTIVE

Penetration (Hz/s) (Hz) Interconnector) SYN
UNIT
G502 9.375 -0.65 49.28 350 5
G502 9.375 -0.5 49.9 150 5
G502 27.86 -2.73 49.68 560 4
G502 32.87 -2.75 49.56 710 3
G502 44.14 -2.68 49.61 710 3
Figure 2-19 – 2-25 below shows the polynomial regression surface model for values
of ROCOF, Frequency Nadir Point (FNadir) variation. Table 2-10 and Table 2-11 list
their values for area 5 generators (G503, G502, G501) terminals under various
contingency conditions. We can notice a directly proportional relationship between
ROCOF, FNadir and Wind Penetration as well as between ROCOF, FNadir and MW
Loss. ROCOF and lowest FNadir increase with increasing amount of wind farm in Area
5 and increasing amount of power (MW) Loss. An inverse proportional relationship
is observed with a number of the active synchronous generator in area 5 with ROCOF
and FNadir point. Through simulations, it is observed that ROCOF in area 5 is
maintainable within 1 Hz/s range up to 400 MW generator contingency and increases
sharply with higher MW loss. Duration of this increase in ROCOF is also observed
to be within 2-second range. Increased ROCOF settles to lower ranges as indicated
in tables above. In the Australian NER, no standard is set for a maximum level of
ROCOF on the power system. Generation, on the other hand, is required by their
Access Standards to remain connected through an event where ROCOF reaches ±1
Hz/s. According to new connection standards in NEM, 1 Hz/s is minimum access
standard for ROCOF and 4 Hz/s for 0.25s is automatic access standard for ROCOF
88
where the plant must be capable of continuous, uninterrupted operation [35]. NEM
needs to have clear ROCOF standard for the correct operation of emergency
protection frequency relay that manages multiple contingency events should be based
on maximum ROCOF.
2.6.5. Linear Regression Model from Frequency Disturbances
Wind penetration level (WP) can be defined as ratio of total wind generated power
P by total generated power which includes synchronous generator power P ,
inter-area power flow P and wind power.
WP 2.3
By investigating the development of frequency deviations during a disturbance and
relating the change in frequency to other factors, a relationship can be found. To
generate a linear regression model for wind penetration level, MW loss and number
of the active synchronous generator with respect to ROCOF and frequency nadir,
following steps have been implemented:
1. Set the load level.
2. Maximum wind penetration level is estimated with direct replacement method
where wind power plant has added the equivalent of lowest synchronous
generator rating in area 5.
3. Trip the synchronous generator of lowest rating to initiate generation led
disturbance.
4. Find frequency nadir point and ROCOF value after disturbance from
simulation results. Though ROCOF can be obtained from theoretical equation
2.1 or 2.2, values expressed here are obtained as simulated df/ft soon after
disturbance.
89
5. Step 1-3 are repeated with increasing generation tripping and increased wind
penetration level. MW loss is obtained as generation tripping as well as area
5 to area 3 interconnector tripping.
6. Finally, a mathematical regression model between wind penetration (%) and
ROCOF, FNadir or MW loss and ROCOF, FNadir or Number of active
synchronous units and ROCOF, FNadir is formed.
7. If F F and ROCOF ROCOF ,
Wind penetration level in control area can be increased till F
F and ROCOF ROCOF . Then maximum
wind penetration level is achieved for a control area.
Linear polynomial regression model for variation between wind penetration, MW
loss, ROCOF, Frequency nadir observed for area 5 at generator 501 and 502 is given
below and coefficients given in Table 2-12:
F (x, y) = p00 + p10*FNadir + p01*ROCOF 2.4
Coefficients (with 95% confidence bounds):
Table 2-12 Values for Linear Polynomial Regression Model
Figure F (x, y) Generator p00 p10 p01 R-square
No.
2.19 Wind 501 -110.3 2.241 -12.54 0.8949
Penetration
level (%)
2.20 MW-Loss 501 1.022e+04 -202.4 -158.3 0.9083
2.21 Wind 502 -40.42 0.8702 -11.8 0.8348
Penetration
level (%)
90
2.22 MW Loss 502 1.562e+04 -312.1 -194.3 0.9543
2.23 No. of 502 -1.125 0.1323 0.7738 0.8283
active syn.
units
Figure 2-18 Polynomial Surface View for Wind Penetration Variation with Frequency Nadir Point
and ROCOF as Observed at Generator 501
Figure 2-19 Polynomial Surface View for MW-Loss Variation with Frequency Nadir Point and ROCOF
as Observed at Generator 501
91
Figure 2-20 Linear Polynomial Regression Surface View for Wind Penetration Variation with Frequency
Nadir Point and ROCOF as Observed at Generator 502
Figure 2-21 Linear Interpolant Surface View for MW-Loss Variation with Frequency Nadir Point and
ROCOF as Observed at Generator 502
Figure 2-22 Polynomial Regression Surface View for Online Synchronous Generator Units Variation
with Frequency Nadir Point and ROCOF as Observed at Generator 502
92
Figure 2-23 Linear Polynomial Surface View for Online Synchronous Generator Unit Variation with
Frequency Nadir Point and ROCOF as Observed at Generator 501
Figure 2-24 Linear Polynomial Surface View for Wind Penetration with Frequency Nadir Point and
ROCOF as Observed at Generator 502
2.6.6. High Load Scenario
Test network is studied when DFIG based wind farms are integrated into Area 5. Area
5 has three lumped generators; G503 at bus 503, G502 at bus 502 and G501 at bus
501. Generator 503 and 502 are connected at 15 KV while generator 501 is connected
at 20 KV. Wind farms are connected to bus 506 at 275 KV and bus 508 at 275 KV in
different simulation studies. 1.5 MW, DFIG based is connected at 575 v which is
further stepped up to 275 KV through two transformers. Summary of converged load
flow is given in Table 2-13.

93
Table 2-13 Scenario 1: Heavy Load
Summary for high load Scenario Converged: The load flow converged in 5 iterations!
P (MW) Q (MVAR)
Total generation 23035.66 3795.05
Total PQ load 22300.1 2556.5
Total Z shunt 92.17862 -84791.8
Total ASM 0 0
Total losses 643.3862 86030.31
All system data is described in the [39], and it is available electronically. All the data
will not be reproduced here in this study, but it is important to note that a few changes
have been introduced especially under high load scenario. The number of active
synchronous generators at each plant is different in both low load and high load
scenario. Some modifications in terms of parallel transmission line additions,
additional equivalent MVAR using load blocks and transformers addition were done
in low load scenario model for load flow convergence under high load scenario.
Modifications were done mainly in following bus numbers as given in Table 2-14.
Due to network complexity and longer simulation time, only three simulation test
were conducted for high load scenario.
Table 2-14 Modified Bus for High Load Scenario
B413 B412 B209 B213 B507 B509 B310 B314 B313
Table 2-15 summarizes the generation output in high load scenario, indicating the
active and reactive power output of an individual generation unit and the number of
units in service at each power plant.
94
Table 2-15 Generation Output in High Load Scenario
Power Station/Bus No. of Units P (MW) Q (MVAR)

B101 4 322.7648 161.7217
B201 6 3600 474.2422
B202 5 2500 630.3434
B203 4 1500 501.5524
B204 6 2950.2 672.6861
B301 7 4200 302.7628
B302 3 939.9 -87.7698
B401 4 1400 554.9564
B403 4 1400 212.9104
B404 6 1549.8 335.254
B501 2 600 -52.8549
B502 4 800 -47.5857
B503 4 436 -42.8808
B402 3 837 179.7116
Total 62 23035.66 3795.05
Case 1: Generator-Interconnector Contingency Analysis (High Load Scenario)
The frequency response of a big interconnected transmission system to a disturbance
caused by the loss of connected generation or load depends primarily on the inertial
response of the system, load damping and the action of governors subjected to
frequency errors. However, each of these elements is in turn dependent upon many
other factors and higher order responses. A detailed and complete analysis of the
frequency behavior of a system requires complex and time-consuming modeling of
each element of the system. In the current study, wind farm integration effect on
frequency response indicators namely, Rate of Change of Frequency (ROCOF) (Hz/s)
and frequency Nadir at area 5 generator terminals is analyzed. In the first simulation,
DFIG based wind farm is added, and frequency response indicators are observed at
area 5 generators. In the second simulation, generator G503 is isolated from the
network at 10 seconds, and the equivalent wind farm is present in the network.
ROCOF and Frequency Nadir as observed are given in Table 2-16 – 2-18 below:
95
Table 2-16 Area Frequency Response Indicators When Wind Farm is Added (109x4 MW)
Generator ROCOF (Hz/s) Frequency Nadir (Hz)
503 -1.1 49.97
502 -1.75 49.956
501 -1.64 49.95
302 -0.45 49.99
301 0.4 49.99
Power flow varies from 500 MW and settles around 420 M

Table 2-17 Area Frequency Response Indicators When Gen 503 is Isolated at 10 s and Wind Farm Is Added
(436 MW)
Generator ROCOF (Hz/s) Frequency Nadir (Hz) Power Flow
501 -1.15 49.93
502 -1.3 49.92 Changes from 500
MW to 750 MW
301 -0.11 50.023
302 -0.016 50.026
G502 terminal ROCOF changes from -1.3 to 0.56 Hz/s in 0.6 seconds soon after G503 isolation.
For G501, ROCOF changes as -1.15 to 0.6 Hz/s in 0.55 s. G301 terminal ROCOF changes from
+0.1 to -0.16 Hz/s in 0.5 seconds soon after G503 isolation. For G301, ROCOF changes as +0.1 to
-0.113 Hz/s in 0.7 s.
An improvement in ROCOF values is observed in Table 2-17 after G503 isolation.
This improvement is due to increased power export from area 3 to area 5 from 500
MW to 700 MW while frequency nadir is decreased. We also notice a drop in ROCOF
values at area 3 generator terminals. ROCOF changed from 0.1 Hz/s to -0.1 Hz/s at
96
G302 and G301 both. Lowest frequency point also moved from 49.99 to 50.02 for
both the generator. Similar results are observed in Table 2-18 when interconnector is
also isolated. Generator loss has direct effect on ROCOF changes which are
prominent during high load demand and integration of equivalent wind farm does not
present any improvement in ROCOF values. Test network being a complex
interconnected network a clear conclusion in frequency response behavior due to the
wind farm is difficult and need more investigations.
Table 2-18 Area 5 Frequency Response Indicators When Gen 503 is Isolated at 5 s, and Interconnector is
Isolated at 10 s, Wind Farm Added (436 MW)
Generator ROCOF (Hz/s) Frequency Nadir (Hz) Power Flow
502 -3.37 46.33

Increases to 750 MW
501 -2.85 45.94
after first isolation
302 -0.37 49.99 (at 5 s) and goes zero at

interconnector loss
301 -0.3 50.02 (after 10s)
G502 terminal ROCOF changes from -1.17 to 0.57 Hz/s in 0.6 seconds soon after G503 isolation.
At second contingency of interconnector loss, ROCOF changes as -3.37 to 0.14 Hz/s in 0.9 s. G501
terminal ROCOF changes from -1.09 to 0.57 Hz/s in 0.6 seconds soon after G503 isolation. At
second contingency of interconnector loss, ROCOF changes as -2.85 to 0.25 Hz/s in 0.9 s. G302
terminal ROCOF changes from 0.1 to -0.16 Hz/s in 5 seconds soon after G503 isolation. At second
contingency of interconnector loss, ROCOF changes as 0.4 to -0.37 Hz/s after 10 s.
2.7. Possible Frequency Control Solutions
Currently, AEMO aims to potentially manage the power system impacts with using
constraint equations in the central dispatch process to limit wind generation, or as a
last resort by market intervention through issuing directions to synchronous
generators to ensure sufficient levels of power system inertia are maintained to allow
adequate control of power system frequency. Establishment of improved ancillary
service requirements and control schemes for allowing adequate control of power
97
system frequency under conditions of low power system inertia is another action plan.
Besides these short-term action plans, AEMO intends to improve wind generation
modeling in its operational and planning tools and processes, to improve their
accuracy with increasing levels of wind generation. AEMO is investigating the extent
to which fast frequency response (FFR) could provide a substitute for synchronous
inertia.
Table 2-19 Anticipated Technical Solutions for Frequency Improvement in Wind Penetrated System
Technical Solution High ROCOF Insufficient FCAS Inertia Improvement
Synchronous √ √
Condenser [41]
Frequency responsive √ √ Needs field level testing
Wind Generation for technical feasibility
[42] [43]
Batteries [44] √ √ Needs field level testing
for technical feasibility
Frequency Responsive √ √ Needs field level testing
Photo Voltaic [45] for technical feasibility
Demand Response √ √
[46] [47]
Change in ROCOF Needs field level
protection setting [34] testing for technical
feasibility
New Interconnectors Partial
[34]
Increase Frequency Uncertainty in Uncertainty in
Deviation Limit [34] technical feasibility technical feasibility
Optimize FCAS Partial. Not fully
Requirements [34] address challenge
98
Some of the potential technical solutions and their reach corresponding to technical
frequency regulation challenges are listed in Table 2-19, which need advanced field
level testing and cost-benefit analysis. AEMO is assessing the capabilities and
limitations of technologies that can provide an FFR service, and whether FFR
provides an adequate substitute or supplement to inertia. The scale and type of
response required to make a useful contribution to the power system are still
unknown. There is a considerable body of work to be done to fully understand the
dynamic performance of the Australian power system as conventional synchronous
plants are gradually displaced by low inertia generating plants. Out of the many
possible frequency control solutions, two of them are briefly presented below:
 Synchronous Condensers
The synchronous condenser has been used as a conventional solution for reactive
power regulation, but they are losing value due to growth in power electronics based
reactive power compensation scheme. However, during the global trend towards
renewable energies, the synchronous condenser has been experiencing a renaissance
since last five years as frequency regulation solution. Increasing renewable sources
tends to lower the system inertia thereby introducing frequency regulation problem.
Renewable technology like wind energy systems has very low inertia comparing to
the conventional generating system, so there is a possibility of utilizing retiring
conventional generating units as synchronous condensers. The conversion of non-
profitable or deactivated power station is currently seen as the most cost-effective
option applying synchronous condensers.
Synchronous condensers are synchronous electrical machines attached to the
electricity grid. The machine will be brought up to speed with an electrical motor
attached or via a frequency converter. When the machine is synchronized with the
99
electricity grid, it will act as a motor, turned by the energy taken from the grid.
Because of the nature of the synchronous machine, reactive power can be consumed
and generated by controlling the excitation of the rotor. The generator of the
synchronous condenser with its rotating mass also always acts against a variation of
network frequency, thereby acting as support for stable system frequency. As per
KEMA report, only 1 % – 4 % of the nominal power rating is consumed during inertia
provision of approximately 1 sec. Synchronous condensers can respond immediately
if on otherwise takes less 15 min startup [28]. A synchronous condenser has the
benefit of providing a synchronous inertia response to support ROCOF management
capabilities and providing fault level and voltage support services beneficial to the
power system.
The impact of synchronous condenser effect on the frequency response of 14-
generator NEM model’s area 5 under contingency events was investigated.
Synchronous generator with the same number of units as that of isolated generating
plants is connected to the 275 KV high-voltage transmission network via a step-up
transformer. In practice, the startup and braking system of a Siemens synchronous
condenser is either based on a frequency-controlled pony motor or a starting
frequency converter, the simulation model has turbine governor system removed, and
voltage is controlled through AC4 excitation system. In the first simulation, generator
503 is isolated at 10 s and generator 501 isolated at 20 s from area 5. Total wind farm
penetration in the area is taken as 44% with 585 MW DFIG based wind farm
connected to bus 509 and 561 MW DFIG based wind farm connected to bus 507. In
the second simulation, synchronous condensers of equivalent ratings of generator 503
and generator 501 are introduced in the test network. First, only one synchronous
100
condenser connected at bus 508 is activated to analyze frequency response. Second
simulation test has both synchronous condensers activated in the network.
Table 2-20 Area 5 Frequency Response Indicators during Generator Contingency (Without Synchronous
Condenser)
Active FNadir (Hz) ROCOF (Hz/s) Power Flow

Generator
G502 49.61 -2.68 to 1.2 in 0.6 sec (at 1st Power flow changes from an
Gen. tripping). export of 500 MW to 50 MW after
first generator tripping and then
-0.85 to 0.4 in 0.6 sec. (At 2nd
the import of 75MW after second
Gen. tripping)
generator tripping.
Table 2-21 Area 5 Frequency Response Indicators during Generator Contingency (With One Synchronous
Condenser)
Active
Generator
FNadir (Hz) ROCOF (Hz/s) Power Flow
G502 Power flow changes from an

-2.35 to 0.875 in 0.6 sec
export of 500 MW to 50 MW
& (at 1st Gen. tripping)
49.67 after first generator tripping and
-0.75 to 0.24 in 0.6 sec.
Synchronous then the import of 70MW after
(At 2nd Gen. tripping)
Condenser second generator tripping.
Table 2-22 Area 5 Frequency Response Indicators during Generator Contingency (With Both Synchronous
Condensers)
Active FNadir (Hz) ROCOF (Hz/s) Power Flow

Generator
G502 Power flow changes from an

-1.68 to 0.5 in 0.8 sec (at
export of 500 MW to 50 MW
& 1st Gen. tripping)
49.72 after first generator tripping and
Synchronous -0.5 to 0.14 in 1.1 sec.
then the import of 65 MW after
Condenser (At 2nd Gen. tripping)
second generator tripping.
101
Table 2-20 – 2-22 gave the frequency response indicator values under these two
synchronous condensers application effect and represented in Figure 2-25 – 2-26. It
is observed that with ROCOF improves from -2.68 Hz/s to -1.68 Hz/s with
synchronous condensers rated equivalent to contingency MW loss.
‐0.5
ROCOF (Hz/s)
‐1
‐1.5
‐2
‐2.5
‐3
W/O Syn. Cond Syn. Cond 1 Syn. Cond 2
Figure 2-25 ROCOF Comparison with/without Synchronous Condensers
49.75
49.7
Frequency Nadir
49.65
49.6
49.55
W/O Syn. Cond Syn. Cond 1 Syn. Cond 2
Figure 2-26 Frequency Nadir Comparison with/without Synchronous Condensers
Synchronous condensers of the smaller size have a small effect on improving inertia.
Frequency observed at various buses of area 5 with and without synchronous
condenser is shown in Figure 2-27 – 2-28. Clear improvement in bus 507 and bus 508
frequency can be noticed. Frequency improvement suggests that if large generators
102
of retiring plants are converted as synchronous condensers, the appropriate level of
inertia is possible to achieve with desired ROCOF.
Figure 2-27 Buses Frequency during Generator Contingency and No Synchronous Condenser Addition
Figure 2-28 Buses Frequency during Generator Contingency and Equivalent Synchronous Condenser
Added
 Grid Code Compatible Frequency Controller based Wind Turbines
Frequency-active power control model is an auxiliary control algorithm implemented
in individual wind turbine generator control loop for providing controllable power
reserve on demand in the form of spinning reserve or power ramp rate limit to respond
103
system frequency deviations. A dynamic dead band based droop frequency grid code
compatible wind turbine as shown in Figure 2-29 has been used in this simulation
which will be discussed in detail in coming chapters.

Figure 2-29 Frequency Responsive Wind Turbine Model
Figure 2-30 Frequency Observed at Bus 506 when Normal Operated DFIG Based Wind Farm is added
Figure 2-30 shows the frequency at Bus 506 when DFIG based wind farm is added
to the bus. The Same bus is also receiving power from Generator 503 which is used
in the model with 4 active units, each with 109 MW rated capacity. Same capacity
wind farm without any supplementary droop control is added at Bus 506. It can be
seen from Figure 2-31 that maximum frequency deviation reaches up to 49.1 Hz after
simulation. In the second simulation, grid code compatible frequency controller based
wind farm is added at same Bus 506. A clear improvement in frequency deviation is
104
observed from Figure 2-32. Maximum deviation reached after the simulation is 49.97
in comparison to 49.1 in the previous case. Improvement in frequency deviation
shows that bus frequency can be improved with wind farms addition with some
modified control algorithms making them compatible with country grid code.
Figure 2-31 Frequency Observed at Bus 506 when Frequency Responsive DFIG based Wind Farm is
added
2.8. Conclusion
Based on the low load and high load scenario investigation under varying wind
penetration level, following results can be concluded for area 5 operation:
 Up to 45% wind penetration in area 5 has a negligible impact on strongly
interconnected area 3 regarding changes in ROCOF and frequency nadir.
Thus, a strongly interconnected network with power sharing can be a key to
sustainable wind energy integration.
 NEM has a minimum standard of ±1 Hz/s for one second for generators to
withstand ROCOF. With increasing wind penetration, there is a requirement
for ROCOF dependent relay operation to consider the time duration also for
which there is a change in ROCOF values.
105
 FCAS requirement likely to remain identical with the ramping period under 5
seconds if MW loss and number of the active synchronous generating unit are
same under different wind penetration level.
 ROCOF can be improved to lie under 2 HZ/s for contingency event if
synchronous condensers equivalent to MW loss are added in the network.
 Frequency responsive wind plants are likely to contribute to improving
frequency deviations, but ROCOF variation would need more assessment due
to the involvement of power electronics.
 Islanded network calls for more local regulation FCAS from local generating
units’ including wind energy participation in centrally managed automatic
generation control.
Three tired classification level can be formulated for ROCOF for wind integrated
power system like area 5:
 Green (ROCOF ≤ 1Hz/s) – FOS of 47–52 Hz can be easily met under wind
penetration as high as 40% for both low load and high load scenario.
 Orange (1Hz/s < ROCOF ≤ 4 Hz/s) – FOS of 47–52 Hz cannot be met under
wind penetration more than 20% and high contingency event for both low
load and high load scenario. With higher wind penetration and generation
loss, ROCOF remains under 4 Hz/s, but frequency nadir drops beyond set
level.
 Red (ROCOF > 4Hz/s) – FOS of 47–52 Hz will not be met. Even though with
wind penetration as high as 45% and MW loss as high as 740 MW, ROCOF
remains under 4 Hz/s. If ROCOF goes above 4 Hz/s, FOS standard will be
violated for more than a second leading to generation disconnection.
106
The focus of this research study is to challenge the designs built into the power system
and the regulatory framework within which it operates. Grid code compatible
frequency responsive wind turbine generator modeling is presented in subsequent
chapters, and corresponding impacts on the interconnected network regarding better
system frequency response are discussed.
107
2.9. References
1. AEMO, (2015). National electricity market factsheet. Available at aemo.com.au.
2. AEMC, (2014). Last Resort Planning Power - 2014 Review. Available at
www.aemc.gov.au
3. AEMO, (2016). National Electricity Forecasting Report. Available at
aemo.com.au.
4. AEMC. (2014). National Electricity Rules Version 62. Available at
aemc.gov.au/energy-rules/national-electricity-rules/current-rules.
5. IRENA, (2015). Renewable Energy Target Setting. Available at www.irena.org
6. REN21, 2016. Renewables 2016: Global Status Report. Available at
www.ren21.net.
7. National Energy Board, (2016). Canada Renewable Power Landscape, Energy
Market Analysis 2016. Available at www.publications.gc.ca.
8. Ma Tianjei, (2017). China's Ambitious New Clean Energy Targets, The
Diplomat. Available at www.thediplomat.com.
9. Laurie Guevara-Stone, (2016). A small country goes big with renewables:
Denmark’s goal to be fossil-free. Available at www.reneweconomy.com.au.
10. France in the United Kingdom, (2016). French renewable energy production up
over 23% in 2015, Energy policy/renewable energy – Communiqué issued
following the Council of Ministers’ meeting. Available at
www.uk.ambafrance.org
11. Tsvetomira Tsanova, (2015). India reached 22% of 175-GW renewable energy
goal, Renewables Now. Available at www.renewablesnow.com/news
12. Oxford Business Group, (2016). South Africa’s push for Renewables-Economic
News. Available at www.oxfordbusinessgroup.com
108
13. Commons Select Committee, Energy and Climate Change Committee (2016).
2020 renewable heat and transport targets report. Available at
www.parliament.uk/ecc
14. Office of Energy Efficiency & Renewable Energy, (2017). Federal Energy
Management Program, Achieving 30% Renewable Electricity Use by 2025.
Available at www.energy.gov.eere
15. Eurostat, (2017). Renewable energy in the EU, Share of renewables in energy
consumption in the EU still on the rise to almost 17% in 2015. Available at
http://ec.europa.eu/eurostat.
16. Geoscience Australia and ABARE, (2010). Australian Energy Resource
Assessment, Based on data from Wind lab Systems Pty Ltd., DEWHA
Renewable Energy Atlas (wind map data); Geoscience Australia. Available at
data.daff.gov.au.
17. Drew, G. (2015). Renewable Energy Superpower. [online] Beyond Zero
Emissions. Available at bze.org.au.
18. Paton, J. (2013). Australian Wind Energy Now Cheaper Than Coal, Gas.
Available at: http://www.bloomberg.com/news/articles/2013-02-06/australia-
wind-energy-cheaper-than-coal-natural-gas-bnef-says.
19. Milborrow David, “Onshore wind is more competitive than ever,” Article.
Available at www.windpowermonthly.com.
20. BREE, (2013). Electricity Generation Major Projects-2013. Available at
bree.com.au.
21. Mott MacDonald. (2010). UK Electricity Generation Costs Update. Available at
www.decc.gov.uk/assets/decc/statistics/projections/71-uk-electricitygeneration-
costs-update-.pdf.
109
22. Brinsmead T.S., J. Hayward and P. Graham (2014) Australian Electricity Market
Analysis report to 2020 and 2030, CSIRO Report No. EP141067
23. Australian Energy Council, (2016). Renewable Energy in Australia. Available
at www.energycouncil.com.au/.
24. ARENA, (2010). Australian Energy Resource Assessment, Chapter 9, Wind
Energy (pp 239-260). Available at www.arena.gov.au/about-renewable-
energy/wind-energy/.
25. CEC, (2015). Clean Energy Council Australia Report-2015. Available at
www.cleanenergycouncil.org.au/policy-advocacy/reports/clean-energy-
australia-report.html.
26. CEC, (2016). Clean Energy Council Australia Report-2016. Available at
www.cleanenergycouncil.org.au/policy-advocacy/reports/clean-energy-
australia-report.html.
27. CEC, (2016). Progress and Status of the Renewable Energy Target Report.
Available at www.cleanenergycouncil.org.au
28. Parkinson, G. (2016). Wind energy hits 100% of South Australia demand on
Sunday. [Online] Available at http://reneweconomy.com.au/2016/south-
australia-graphs-60608
29. Parkinson, G. (2016). Big blow in NSW: Wind farms produce more than coal
generators. [Online]. Available at: http://reneweconomy.com.au/2016/big-blow-
nsw-wind-farms-produce-coal-generators-91159.
30. AEMO, (2013). National Transmission Network Development Plan. 2013.
Available at www.aemo.com.au.
110
31. AEMO, (2015). Report to COAG Energy Council on Security and Reliability in
the Context of Changing Generation Mix. [online] SCER. Available at:
http://www.scer.gov.au.
32. Ceres project, (2013). Wind Farms and the SA Electricity Market Fact Sheet.
Available at: http://www.theceresproject.com.au.
33. Deloitte Access Economics, (2015). Energy markets and the implications of
renewables. Deloitte Access Economics Pty Ltd. Available at
www2.deloitte.com.
34. AEMO, (2016). Future Power System Security Program. Available at
aemo.com.au.
35. AEMO (2012), Frequency Control Ancillary Services. Available at
aemo.com.au.
36. J. Riesz, F.-S. Shiao, J. Gilmore, D. Yeowart, A. Turley and I. Rose, (2011)
Frequency Control Ancillary Service Requirements with Wind Generation -
Australian Projections.
37. AEMO (2013), Integrating Renewable Energy - Wind Integration Studies
Report. Available at aemo.com.au
38. Tanner P (2002), Wind farm grid connection technical issues, Presentation to
Transgrid Country Energy Workshop on Grid Connection of Wind Farms,
Sydney, 11 September.
39. M. Gibbard and D. Vowles, (2008). Simplified 14-Generator Model of the SE
Australian Power System. The University of Adelaide, South Australia.
40. MATLAB, (R2013). Simscape Power System Toolbox. Mathworks. Available
at mathworks.com.
111
41. GE Digital Energy, (2014). Synchronous condenser system. Available at
www.gegridsolutions.com.
42. Nicholas Miller, Kara Clark, and Miaolei Shao, (2010). Impact of frequency
responsive wind plant control on grid performance, GE presentation to MIT’s
wind week. Available at web.mit.edu/windenergy/windweek/Presentations.
43. Antonio Martinez, Kouroush Nayebi, Manoj Gupta and Yi Zhou, (2016). Wind
Power Plant Frequency Control to Support the Penetration of High Levels of
Renewable Sources, Vestas Wind Systems A/S presentation to Wind Industry
Forum.
44. PJM, (2016). Grid-scale energy storage for regulation services, PJM frequency
regulation, Available at www.pjm.com.
45. Clyde Loutan, Vahan Gevorgian, (2017). Using renewables to operate a low
carbon grid, demonstration of advanced reliability services from utility-scale
solar PV plant, California ISO, NREL and First Solar. Available at
www.caiso.com.
46. MacDonald, P. Cappers, D. Callaway, S. Kiliccote, (2012). Demand response
providing ancillary services, Proc. Grid-Interop 2012. Grid Wise Architecture
Council, Irving, TX.
47. AEMC, (2017). Demand Response Mechanism and Ancillary Services
Unbundling Rule. Available at aemc.com.au.
48. Uijlings, W. (2012). System Service Provision An independent view on the
likely costs incurred by potential System Service Providers. [online] London:
DNV KEMA. Available at www.eirgridgroup.com/site-
files/library/EirGrid/System-Service-Provision-DNV-KEMA-Report-2012.pdf.
112
CHAPTER 3
POWER SYSTEM FREQUENCY CONTROL
This chapter presents a brief literature review of active power frequency control
strategies for wind integrated power system model. An attempt has been made to
present new perspectives & challenges for secondary frequency control which need
more advanced investigation for future implementation in the electric grid.
3.1 Introduction
Load/Demand is primary independent control variable in power system while all
other controllable power system elements respond to serve this continuously varying
load. System frequency is one of the dependent variables which indicates system
energy balance and needs to be regulated for stable operation. Low frequency
indicates load increase while high frequency represents a high generation. System
frequency is regulated by injecting active power into electric grid through power
plants. This control is indispensable for the stable operation of grid ensuring
continuous adaptation of generation to demand. Frequency below a certain prescribed
level will deteriorate power quality & may eventually result in a system outage. Any
investigation of system frequency response requires frequency regulation analysis at
three different stages followed lastly by an emergency contingency regulation
control.
Figure 3-1 represents a typical frequency regulation scenario under contingency event
which leads system frequency to cross governor dead band. Under a wide range of
ambient conditions, frequency responsive power plants are expected to provide
automatic power variation with frequency drop within a given time frame and ramp
113
limit. The conventional power plants which remain connected to the grid even in the
case of frequency deviations accomplish active power-frequency control through
turbine governors reaction to nominal synchronous speed deviations and the
respective boilers-turbine frequency responsive controls.
Figure 3-1 Frequency Variation Along with Regulation Services
The normal power imbalance between generation and consumption is instantaneously
covered through a synchronous generators natural inertial response at first stage
followed by system synchronous generator’s governor response to bring system
frequency closer to setpoint. Contingency event leads to immediate governor
response for frequency regulation. A comparatively slower, centralized and
automatic secondary controller termed as automatic generator control (AGC) is
activated to bring frequency to the setpoint. Tertiary control is centralized manual
transmission system operator (TSO) controller utilized in case secondary reserves are
saturated before stabilizing the system frequency.
114
Table 3-1 Basic Frequency Regulation Services
↓Control→ Primary Secondary Tertiary
Impact Limit Deviation & Frequency Mitigate Forecasted

Automatic Frequency Restoration to Imbalance by
Stabilization Setpoint Replacing Primary
and Secondary
Reserves
Controller Position Decentralized, Power Centralized, TSO Manual by Tso

Plant Control Center
Input Frequency Frequency, Power Imbalance
Figure 3-2 Frequency Control during Contingency Event [Reformulated From 4]
Table 3-1 list the various stages involved in system frequency regulation. Basic
control scenario during contingency event is represented in Figure 3-2. Different
countries have different nomenclature for these three basic services. Table 3-2
provides the nomenclature adopted in various countries for generation demand
regulation. European Network of Transmission System Operators for Electricity
(ENTSO-E) [1] has proposed more explicit terminology for these services: Frequency
115
containment reserve services (FCR), Frequency restoration reserve services (FRR)
and Tertiary replacement reserve (RR) while some countries like Australia [2] and
U.K [3] have no explicit representation for tertiary control.
Table 3-2 Regulation Services Nomenclature Adopted by Various Countries
116
In the following next sections, the basic concept of frequency regulation stages is
described along with their presence in the wind integrated power system.
3.2. Power System Inertia
Conventional power system operation is based around traditional synchronous
generators operating in synchronism, i.e. with the same frequency (50/60 Hz), with a
lot of kinetic energy stored in their rotating masses. The angular velocity of these
synchronous generators is proportional to system frequency. Kinetic energy which is
stored in rotating masses will provide a cushion for this momentary imbalance of
power. This stored rotating energy in the synchronous generator is referred as Inertia
having the ability to oppose changes in frequency and represented by inertia constant
as:

H 3.1

W J 2πf 3.2
Where H is inertia constant in MWs/MVA denoting the time duration during which
the generating machine can supply its rated power (S ) in MVA exclusively with its
stored kinetic energy (W) in MJ or MW.s, J is the moment of inertia of rotating mass
117
in kg-m2. The change in kinetic energy leads to classical swing equation as shown in
Equation 3.3, which describes the inertial response of synchronous generator as
change in frequency following a power imbalance between the mechanical power
supplied by the generator P and P as the electric power demand.
J 2π ∗f∗ P P 3.3
On a system MVA base (S , equivalent system inertia constant, H can be
calculated as:
H ∑ H ∗ 3.4
Where H and S are the inertia rating and power rating of individual generating units
and N is the total number of generators.
Figure 3-3 Inertia Constant Values of Various Plants, Data Derived From [70] [45]
The system inertia constant depends on two factors including the number of operating
generators and the inertia constant of each of these generators. A larger H is
preferable as it signifies a higher amount of stored kinetic energy. Inertia constant per
118
unit base MVA for different power plants and synchronous condensers is represented
in Figure 3-3. The single wind turbine has inertia constant value in the range of 4-6 s
which is comparable to hydro and fossil steam based plants in similar MVA rating
range. An inversely proportional trend between inertia constant and machine rating
can also be seen for power plants.
3.2.1. Power system Inertial Response
The frequency of a power system with a large amount of stored kinetic energy tends
to be less sensitive to power imbalances than the frequency of a system with a lower
amount of stored energy. For a synchronous generator, the angular speed is tightly
coupled to the system frequency. When a power system experiences a large
disturbance, such as a sudden loss of generation, the system frequency begins to fall.
Therefore, when the system frequency drops, the generator-angular speed tracks the
falling frequency and, thus, the stored kinetic energy of the power plant is released.
This process is known as the inertial frequency response which acts as an initial
arresting mechanism to the falling system frequency and is determined by the type of
power plants in the system. There is direct coupling between the power system inertia
usually defined using the moment of inertia M, power system frequency and active
power output of synchronous generators in conventional generating plants ∆P as
given below in equation 3.5 and 3.6:
∆F s ∆P s ∆P s 3.5
∗
M 3.6
T 3.7
∗
119
M is the equivalent moment of inertia of all generators and the motors connected to
the grid [MW.s. Hz-1], H is the system equivalent inertia constant expressed in
megawatt second per megavolt ampere [MW.s / MVA], D is load damping constant
and T is defined as power system time constant and K is defined as power system
gain, equal to 1/D.
The generator inertia is, therefore, the first level of frequency control reserve, and it
has the characteristics of a differential controller [16]. Since inertia is proportional to
rotating mass, greater the inertia, the lesser will be acceleration, and the lesser will be
the frequency deviation. Higher system inertia gives a lower rate of change of
frequency and higher frequency nadir. As inertia affects how fast power system
frequency can change, it has implications for calculating frequency control ancillary
service (FCAS) requirements when considering high penetration of wind-based
power system.
3.3. Primary Frequency Control
Following the inertial frequency response of the power system, which reduces the
ROCOF, the primary frequency response stabilizes the frequency at a new value by
increasing (or decreasing) the power generation in proportion to the frequency
variations. Primary frequency control is a proportional control having system
frequency as a control signal and it independent of the rate of change of frequency.
Primary frequency control is represented in terms of change in turbine power ∆P as:
P ∆P ∆f 3.8
Where R is speed regulation or droop of governor and ∆f is frequency deviation. The
contribution of a generator to the correction of a disturbance on the network depends
mainly upon the generator droop value and the primary control reserve of the
120
generator concerned. The amount of frequency change which will cause 100%
change in the generator output is controlled through droop value. Two generators
having same primary control reserves but different droop will have different output
and different frequency offset as shown in Figure 3-4.
Figure 3-4 Frequency Power Characteristics through Droop Control
In case of low-frequency deviation (<∆f ), generator-a having lower droop will
provide more power output in comparison generator-b having higher droop. Even
having same primary control reserve, generator-a will reach maximum power limit
early, and its primary control reserve will be exhausted faster than generator-b. Both
generators will contribute equally to primary control in case of a major contingency
when frequency offset is > ∆f . As an example, 5% droop indicates that there will be
100% output power change for 5% change in frequency.
Author represents the simulated frequency output of thermal power plant based
uncontrolled area in Figure 3-5 when operating with different droop values while
respective power change of power plant is shown in Figure 3-6. Falling frequency
121
indicates more active power requirements, and we can see a plant with lower droop
will provide more output power during frequency disturbance.
Figure 3-5 Primary Frequency response of Uncontrolled Thermal Power Plant with Different Droop
Figure 3-6 Power Output Change of Uncontrolled Thermal Power Plant under Different Droop
Figure 3-7 represents the generic dynamic frequency control model for a single
control area having primary controlled power plants with rotating mass and frequency
dependent loads. ∆ is scheduled power setpoint for turbine while ∆ is the
power setpoint imposed by primary control. The total change in turbine power output
122
can be written in terms of equation 3.8. With ∆ 0, the change in frequency ∆f
with change in load ∆P can be related by equation 3.9. All the parameters in
equation are defined earlier in nomenclature.
∆f ∆P 3.9

Figure 3-7 Single Area Primary Frequency Control Model with Frequency Dependent Loads [20]
Droop values which indicate the required change in power output in response to
changes in frequency are defined in grid codes and vary in different power systems.
In a typical large power system, the power plants allocated to accomplish primary
frequency response should be capable of increasing their output real power within
about 10 seconds of the pre-defined frequency variation and be capable of
maintaining this response for a further 30 seconds. For example, in the Nordic power
system, the frequency controlled normal operation reserve must be at least 600 MW
and fully activated for frequency deviations of ±0.1 Hz. The frequency controlled
disturbance reserve must be at least 1000 MW and fully activated at 49.5 Hz. [17]. In
UCTE, 50% of the primary control reserves must be activated within 15 seconds, and
123
100% within 30 seconds. [18]. In Australian NEM, regulation raise/lower FCAS
of130/120 MW must be available for bringing frequency in restoration range (49.5 –
50.5) within 5 minutes [19].
3.4. Secondary Frequency Control
Maintenance of power system load-generation balance calls for a control system at
generation side as system load is normally uncontrollable and based upon customer’s
preferences. The secondary frequency Control or Automatic Generation Control
(AGC) represents a remarkable economic function of EMS system for energy market
dispatch of generating units and regulating frequency control ancillary service
(FCAS) dispatch. It is an important tool for regulating the balance and distributing
the imbalance between designated units. AGC makes up as part of real-time energy
market dependent upon load forecast. AGC which is slower than the two
aforementioned inertial and primary responses regulates the frequency at its rated
value. It is the IEEE working group which provided the standard definition for power
system AGC [21]. AGC, as defined by IEEE, is the regulation of the power output of
electric generators within a prescribed area in response to changes in system
frequency, tie-line loading, or the regulation of these to each other, to maintain the
scheduled system frequency and the established interchange with other areas within
predetermined limits.
The analysis and design of AGC system in a more deregulated complex environment
plays a vital role in automation & eventually reliable operation of the power system
at a satisfactory level. AGC system cost in modern power system consists of load
frequency control, economic dispatch; interchange scheduling of individual
generators in many generating units. The cost of load following imposed on the
system due to variable generation is also calculated by AGC setpoints and system
124
dispatch requirements. The sensitivity of any power plant depends upon the response
time taken by the automatic generation controller to control the frequency change due
to load variation of governor control. SCADA system periodically sends system
frequency and real power output data to AGC system which intermittently calculates
the area control error to adjust the setpoints of the respective system. If the load is
increased, AGC system increases the mechanical torque by increasing the system
flow rate to the turbine to compensate the load by increasing the output of the
generator in an area. Load-Frequency Control (LFC) is used as part of AGC to
maintain a constant frequency and to regulate tie-line flows. LFC represents the first
implementation of a higher-level control system. It has made the operation of
interconnected systems possible, and today it is still the basis of any advanced
concept for the guidance of large systems. Secondary reserves require a slower initial
response but are maintained for longer periods of time. Secondary control requires
the capability of increasing the output real power within about 30 seconds and
maintaining the response for a further 30 minutes. Secondary LFC or AGC is a central
control technique usually implemented through Proportional-Integral (PI) or Integral
controller as given in equation 3.10 below.
P K ∆ACE 3.10
3.4.1. Area Control Error
Present power systems are divided into various control areas each constituting
coherent group of generators. It is required to control and set the generation output
power according to energy market load-demand economics and area control error
(ACE). ACE is the real time representation of utility’s obligation to the regulatory
authority for maintaining net power inflow/outflow along with a small bias obligation
125
to maintain frequency. Stability of a system or requirement for more generation with
demand variation is indicated through ACE. Minimization of ACE is the main goal
of AGC which is a dynamic study where disturbance is small. For worthy AGC
design, there should not be an enormous increase in ACE signal, and ideally, it should
be zero. ACE should not be allowed to drift or vary much over a certain period. ACE
is allocated to various units of a control area whose participation in AGC depends
upon energy market economic dispatch.
ACE Control Modes
 Flat Frequency Control: When system frequency is utilized as the only
controlling parameter, AGC can be regarded as FFC. In this control, system
generation operations are not altered for any deviations from tie line schedule.
Normally used in the isolated control area.
 Flat Tie Line Control: When interconnection tie line flow is utilized as the only
controlling parameter, AGC can be regarded as FTLC. Limited to special
circumstances when control area frequency is stable and controlled through other
areas, this control mode is utilized to maintain a given net interchange power flow.
 Tie Line Bias Control: Responding to both tie-line power flow deviations and
frequency deviations, TLBC is the most widely utilized mode of system balancing
mode in big interconnected systems.
The ACE is calculated according to following equation 3.11:
ACE = (NIA − NIS) – 10B (FA − FS) - IME 3.11
Where ACE is the Area Control Error in MW. Negative values denote a condition of
under generation, and positive values denote over a generation. NIA is the algebraic
sum of the actual power flows on all tie lines or interconnectors (MW). NIS is the
algebraic sum of the scheduled flows on all tie lines or interconnectors (MW). It is
126
defined as the mutually prearranged intended net MW flow on the area tie lines,
including dynamic schedules. NIA − NIS is tie-line interchange error or flat tie-line
control which compares the net tie-line actual inflow/outflow with power scheduled
to be sold or bought. Flow into a system is defined as negative while flow out as a
positive value. B is the frequency bias setting (MW/Hz).
Frequency bias reflects the ability of any control area to respond to frequency
disturbance through various provisions like generators inertial response, governor’s
response or load shedding response. It is also the utility’s obligation to generate or
absorb power to regulate the total system frequency. FA is the actual frequency (Hz).
FS is the scheduled frequency (Hz). IME is Interchange (tie line) Metering Error which
is accounted from meters installed at the end of tie-lines. Time error is the
accumulation of frequency deviation over a defined period. When time error reaches
±10 seconds, the Interconnection Time Monitor initiates a time correction. The more
advanced formula for ACE can be found from WECC [22].
Figure 3-8 Basic Loop of AGC Control
127
The basic control loop of AGC control in shown in Figure 3-8 while Figure 3-9 gives
the AGC controlled thermal, hydro thermal power plant based three area system
frequency response under 1% load disturbance. The hydro plant has slightly lower
droop value in comparison to thermal plants. Area 1 has highest frequency bias
followed by area 2 and then area 3. The area with the highest bias will bring frequency
to settling values faster in comparison to other two areas. Figure 3-10 gives the AGC
response of single area thermal power plant under 1% load disturbance. It can be
noticed that plant with higher droop will take more time to settle frequency in
comparison to lower droop value based plants.
Figure 3-9 AGC Controlled Thermal Hydro Thermal Power Plant Based Three Area System Frequency
Response Under 1% Load Disturbance
Figure 3-10 AGC Response of Single Area Thermal Power Plant Under 1% Load Disturbance
128
3.5. Tertiary control
Variation of generators setpoints or load participation is done automatically or
manually through tertiary control. Tertiary control is applied to guarantee timely
provision of secondary control reserve and economical distribution of secondary
power control to various generators. Control may involve power connection tripping,
power redistribution, changing of load control and interconnector power variation
through change constraint equations. All tertiary control reserves are used to restore
secondary reserve.
3.6. Wind Energy System Participation in Frequency-Power
Control
Frequency control has become more challenging under conditions of low power
system inertia due to high wind energy penetration. Wind turbine active power
frequency regulation analysis has received considerable research focus in the form of
inertial control and primary frequency control. Figure 3-11 highlights various
concepts available in scientific literature applied for active power-frequency
regulation analysis either at individual wind turbine level or wind farm level [23-28].
Most of the past researchers focus on modifying individual turbine control algorithm
for analyzing primary frequency control capabilities from wind turbines. These
algorithms either focus on achieving de-loaded performance [29-32] to enhance
VSWTG’s frequency regulation capability or investigated the feasibility of constant
droop control [33] or combined inertial-droop control on frequency stability [34-38].
The basic philosophy behind these methods is briefly discussed in upcoming section.
129
Figure 3-11 Active Power Frequency Control Research Study Line in Wind Based Power Plants
3.6.1. Synthetic Inertial Control
Figure 3-12 Virtual Inertia Model
130
The inertial response has been introduced with different phrases such as synthetic
inertial response [39], inertial response [40-41], inertia emulation [36], short-term
frequency regulation [42], and temporary frequency control [30-31]. A segment of
VSWTG’s kinetic energy is transformed into electrical power through the inertial
control and transferred to the grid through converters. The basic idea behind most of
these droop based studies for emulating primary frequency response and inertial
response studies is to add an additional signal as shown in Figure 3-12, sensitive to
the frequency to the speed or torque controller which appear as reference power
setpoint in VSWTG model thereby momentary increasing the wind turbine output
power. The inertia constant H of a wind turbine is in the order of 2-6 s [36] which is
comparable to traditional generation with 2-10 s [43]. Hence the kinetic energy in the
rotor is high enough to support the frequency but must be released using control
actions in the converter active power control to mimic inertia. In the inertial response,
a part of the kinetic energy stored in the turbine is released by temporarily increasing
the output power. When energy is released from the turbine, the rotor speed
decreases. Consequently, the overproduction period has to be followed by a period of
underproduction where the turbine speed is increased, and the energy in the rotating
mass is restored.
Four virtual inertia and droop models are shown in Figure 3-13 which are further used
in coming chapters. Frequency droop model 1 [37], droop model 2 [44] and fourth
virtual inertia model from GE [45] takes absolute variation in frequency as input
while inertia droop model [36] takes absolute variation in frequency as well as the
rate of change of frequency as input. Transmission system operators (TSO) have been
working on grid code requirements for implementation of a generic inertial response
131
from WPPs [46]. Although some reports and recommendations for future grid codes
have been published [39], there are no mandatory requirements at present.
Figure 3-13 Inertia and Droop Control Model for Frequency Response from VSWTG
3.6.2. Primary Control from VSWTG
Similar to conventional generator response, droop characteristic and de-loaded wind
turbine operation have been explored for primary frequency response assessment
from wind turbines. Wind power plants are required to participate in primary
frequency control by ramping their output up and down according to a power–
frequency droop characteristic and during a certain period. De-loaded wind turbine
132
operation provides a power margin for ramping capability of wind turbines. De-
loading enables some reserve power procurement from wind farms for continuous
participation in primary load frequency control during negative frequency deviations.
De-loading through rotational over speed control [47-50] and pitch control [51-53]
has been discussed in various studies. Both methods are based on the idea of
achieving a non-optimal working point in the torque–rotor-speed curve of the turbine
as shown in Figure 3-14. Reference power setpoint to rotor converter control is
retrieved from measured rotor speed, de-loading margin, and T ω or P ω curves.
Figure 3-14 Representation of Optimum and Deloaded Curve for VSWTG
As represented in Figure 3-15 and under the assumption that maximum turbine power
and rotor speed are known, power setpoint for the de-loaded wind turbine can be
calculated as:
P P P P 3.12
P K P 3.13
%
K 1 3.14
133
P K ω 3.15
K ρ πR 3.16
Figure 3-15 Control Loop for De-loading Control of VSWTG
In pitch control de-loaded method shown in Figure 3-16 and represented in Figure 3-
17, pitch angle is regulated as a summation of pitch angle (β from pitch controller
and an offset pitch angle (∆β ; β β ∆β and ∆β K ∆f. Applied pitch offset
value (K can be dependent upon droop characteristics, active power reference or
de-loading power. De-load VSWTG by certain value X can be calculated by
increasing pitch angle as given by equation 3.17:
C λ ,β 100 X %∗C , λ ,β 3.17
Figure 3-18 gives an example of pitch angle offset variation with respect to de-loaded
power serve variation while Figure 3-19 presents the zoomed frequency nadir point
obtained under 10% and 20% pitch de-loaded VSWTG operation where a pitch offset
lookup table similar to Figure 3-18 is applied. An improvement in frequency nadir
can be with increased deloaded turbine and droop control enabled. Besides the offset
values, servo time constants are also important for this method deep analysis.
134
Figure 3-16 Pitch Droop De-loading Control
Figure 3-17 Coefficient of Performance De-loading Through Pitch Control
Figure 3-18 Pitch Angle Offset Variation with De-Loaded Power Reserve in VSWTG
135
Figure 3-19 Impact on System Frequency Variation with Pitch Droop De-loaded VSWTG
Pitch control methods are preferable in medium and high wind speed but lower in
comparison to an over-speeding method which is preferable at low wind speed. The
combination of pitch and over speed control is preferable in medium wind speed. A
hybrid over speeding de-loading control in provided in [54]. Rotor speed is also used
for this selection as over speeding technique is applied for lower rotor speed while
for rotor speed near or above rated value, pitch control is applied. Output reference
power from these control is not grid code compatible without additional algorithm
implementation.
Both static and dynamic droop control concept has been explored in literature for the
primary frequency response of individual wind turbine. Fixed power response can be
achieved by implementing static fixed droop value. The primary frequency
contribution ∆P of the VSWTG corresponding to the system load demand variation
can be computed from equation 3.8.
An important requisite for the stability of VSWTG during primary frequency
response provision is that its primary frequency power contribution should always be
sufficiently lower than the available power margin. To achieve improved stability and
to avoid reserve exhaustion, variable droop has also been studied. Variable power
136
response was shown to be achievable through dynamic droop value which can be
calculated by wind and power reserve conditions [72] or on the basis of ROCOF [71].
Under variable droop setting given by equation 3.18, optimum wind turbine operation
was proposed such that wind turbine having highest power reserve is assigned lowest
droop value so that it gets maximum participation in primary frequency response.
R R R R ∗ 3.18
Here M denotes power margin which is calculated as power margin: M P
P . So M denotes maximum power margin available among all wind turbines
while is the actual power margin available at a wind turbine.
Another study presented a control system for active power control of wind turbine
for tracking a power reference and provides a primary frequency response with a
constant and dynamic droop in the absence of any dead-band with diminishing
turbine structure loading [61]. Simulations were conducted on NREL CART 3 turbine
with a simple generator rather than a variable wind turbine generator model. In
another study, frequency active power set-point to wind turbine is provided as
M K df dP where K represents stiffness and M is inertia of power
supply where the value of M and K is provided by farm controller [56]. These
systems were studied for constant speed generator instead of variable speed
generator, and all of them were based on wind farm controller.
With commercial availability of 8 MW individual VSWTG system, individual
frequency responsive units need to research and investigation for fast primary and
secondary frequency response.
137
3.6.3. AGC & Wind Energy
The AGC problem has been extensively studied during the last four decades, and it
has been one of the most highlighted issues in the design and operation of
independent and interconnected power systems. Preponderance on renewable sources
like wind energy that different demand management as compared to a conventional
system has moved the electricity grid in a transition phase to a new model. These
sources having new production and nonlinear control technology imply new energy
management system (EMS) with Automatic Generation Control (AGC) being one of
the major units of EMS in modern power system [57]. A control area regulated by
hydro generation interconnected to another area regulated by thermal generation or
gas generation, solar generation, wind generation is a usual combination in a mixed
power system. With increasing RESs integration into the grid, it’s a necessity to
explore their impacts on AGC at the planning stage and in real time operation.
Recent investigations have indicated that large penetration of wind energy will have
an impact on modern power system frequency regulation along with AGC systems
and other control operation issues. AGC is expected to be a major player in the
management of short-term uncertainty of renewable power system output as well as
in mitigation of short-term impacts due to variable generation forecast error.
Application of several computational techniques have been researched for the
frequency regulation in the conventional power system; however, the large
integration of wind energy system have created a gap in theory and application with
just a few reports available on the frequency control design in the presence of wind
power units. Thus a lot of research is underway to review and analyze wind energy
association with AGC performance criteria, capabilities, and technologies for
138
ensuring proper system performance. Figure 3-20 illustrates the line of thoughts being
followed for studying AGC impact of wind energy penetration.
Energy Storage
Fuzzy logic
Conventional
Generation
Integration
Genetic Algorithm
Wind Automatic
Generation Control Intelligent
(Wind-AGC) Controllers
Particle swarm
LFC Controller
Figure 3-20 Wind Energy Study Line for AGC
Conventional AGC model assumes a direct physical connection for generators
setpoint change and the secondary control signal provided by the transmission system
operators. AGC implementation in wind energy system has been studied basically at
two levels: at wind farm level and wind turbine level. Figure 3-20 presents the basic
concept of AGC implementation in the wind-based power plant. To bring control area
frequency to required setpoint level, transmission system operator (TSO) sends
reference power set-points to wind farm controller which in turn dispatches
demanded power setpoints to individual wind turbines. A more detailed
representation can be seen in Figure 3-21 where TSO can provide different types of
power setpoint (balance, ramp or delta) as per system need. TSO provides dead band
and droop settings. The basic idea behind this model is to replace MPPT power
setpoint to wind turbine controller with Power setpoint from wind farm controller
while MPPT power output is sent to the controller as instantaneous available power.
In the case of no power limitation from TSO, same MPPT power set-point is provided
139
to turbine controller. Due to the time delay involved at farm controller level,
researchers are focusing on the possibility of implementing AGC control at wind
turbine level itself. This concept of AGC at turbine level has been explored in this
thesis.
Figure 3-21 Basic Schematic Diagram for Wind Farm Frequency Control
Figure 3-22 VSWTG Based Wind Farm Controller Supporting AGC [58]
140
The conventional LFC designs which are based on classical control have less
adaptability, are not very efficient. They are, therefore, not suitable for modern power
systems. The modern power system has a varied structure with large interconnections,
increased penetration of renewable energy sources and exigent demand for power
quality. It is expected that application of modern, robust, automatic and intelligent
control techniques to the LFC schemes will be more adaptive and appealing approach
for the wind-based power system to overcome impositions set up by technical
standards of energy market regulators. Incorporation of energy storage and
coordination of other conventional generation for frequency support in the wind-
integrated system is the other line of thought being followed for wind energy
incorporation in AGC. Table 3-3 provides some of the relevant studies related to
wind-AGC investigations where modifications were done at turbine level or wind
farm controller level. Besides these, there are numerous studies where wind
penetration effect on load frequency performance has been investigated.
Table 3-3: Wind – AGC: A Review
Refs. Methodology Special Findings
[59] Non-Linear differential algebraic equation Presents analytically traceable AGC
(DAE) modeling framework for studying model based on nonlinear differential
wind energy impact on power system algebraic equation (DAE) formalism but
frequency performance metrics utilizes wind energy system based on first
model differential equations and third
order DE for synchronous generators. No
wind participation in AGC was assumed.
[60] Hierarchical architecture of a supervisory Improved AGC performance and better
control system for active and reactive power integration of the wind with the grid when
control at wind farm substation. supervisory system controls active and
141
Supervisory control system active power reactive power while machine system
output is dependent upon TSO provided controls turbine power setpoints.
power setpoint.
[55] Active Power Controller was incorporated Static and dynamic droop curves were
[61] with AGC command tracking from wind implemented, and Field tested for
farm controller NREL’s 550kW 3-bladed Controls
Advanced Research Turbine (CART3)
for frequency regulation participation at
all three levels. Fast Wind power AGC
response especially at high wind speed
and AGC error are low with multiple
turbines in wind farms.
[62] DFIG based wind farm hierarchical control Implemented in DIgSILENT, the central
strategy to regulate the wind farm power wind farm control level controls the
production to the reference power ordered power production of the whole farm by
by the system operators sending out reference power signals to
each wind turbine, while the local wind
turbine control level ensures that the
reference power signal send by the central
control level is reached
[63] Power/load optimization in the presence of Centralized wind farm model and control
ambient wind, wake deficits, wake principles are used for control of a wind
meandering, the wind turbine system, the power to increase energy quality and
wind farm controller, and a network reduce the fatigue loads
operator
[64] Coordinated control strategy for the AGC Proposed control strategy have shown
between combined heat and power plants that the WPPs can actively help the AGC,
(CHPs) and WPPs to enhance the security and reduce the real-time power imbalance
and the reliability of a power system in the power system, by downregulating
142
operation in the case of a large wind power their production when CHPs are unable to
penetration provide the required response
[65] Developed model called (FESTIV) to The model allows for wind energy
mimic operator behavior using security penetration impacts analysis and gives
constraint unit commitment, economic insight to market operation regarding
dispatch and AGC reliability and costs.
[66] Wind farms incorporating energy storage Energy storage devices are used to
systems for AGC participation neutralize the uncertainty of wind power
plant output and maximize the dispatch
ability. With reduced uncertainty, wind
farms are shown to have better AGC
performance.
[67-68] DFIG based wind farm integrated with Operational impacts of the wind power
thermal energy system was used in study fast fluctuations are largely absorbed by
the thermal unit’s large mechanical and
thermal time constants as well as control
dead-bands, and the system dynamics
varies with the size of the AGC time
delay. It is estimated in this study 5% of
rated plant power can be accepted as
perturbation power by thermal plants,
without exceeding the one%frequency
deviation margin.
[69] (PSO-Fuzzy PI) Fuzzy logic load frequency Intelligent computational techniques
controller and Particle swarm optimization were applied to improve AGC controller
technique were applied performance in New England test system
with wind penetration, but no specific
wind farm frequency support was
accounted.
143
3.6.4. AGC & Wind Energy Research: New Perspectives and
Challenges
Wind energy has the potential to maintain and even exceed the dynamic growth rate
of the past several years. It remains important for researchers to develop more reliable
designs and technology to help wind integration with the electric grid and provide a
knowledge base to concerned authorities so as to maintain long-term growth and
innovation in generation, transmission and demand technologies and commercial
arrangements. This section provides some insights into frequency related issues
which need more research for implementation in the wind farm future grid code
development. Implicit implementation of these issues can lead to the development of
wind farm grid codes and implementation policy which will pave the way for wind
farm operation that is optimally balanced between energy cost and technical
performance.
Synthetic Inertia Supporting Wind Farm Participation in AGC
It is well known that AGC action is preceded by primary regulation activity in a
power system. Load frequency control (LFC) action assumes same area control error
(ACE) throughout the concerned power system, but this assumption does not
completely satisfy with wind power based system. High wind penetration in power
system induces large imbalances and brings instability and power quality issues.
With secondary regulation getting exhausted due to wind load variability, it is very
likely that primary control actions of power electronics based generation technologies
also reduce frequency regulation capabilities in case of a fault [73]. Primary
regulation is provided by power system inertia which controls the system frequency
rate of change in case of an imbalance of load and demand in the system. Power
system rotational inertia is normally time-variant and gets reduced non-uniformly

144
with respective grid topology. Primary inertial control can improve frequency nadir
but do little for improving settling time as the inertia of a system is directly related to
the mass of generating machine and respective prime movers. A high stable power
system has high inertia and low ROCOF. Due to energy economics, the number of
synchronous generators operating will be less in the case of low load demand and
high penetration of wind energy, leading to reduced total grid system inertia.
Frequency regulation and hence sensitivity and stability of the total system is largely
dependent upon total system inertia. Managing total system frequency thus stability
without an inherent inertial response will be difficult for wind penetrated power
system.
Studies indicate that power system is having equal value of spinning reserves but
having different inertia levels express varied frequency responses in case of any
contingency. Research indicates that low inertia will have worst effect for power
system having slow frequency response [74]. So modern wind-based systems are
supplementing their inertial response by adopting additional controllers and using
energy storage systems and synchronous condensers. Modern wind plants have the
ability to control active power output in response to grid frequency by providing
inertial and governor response. The wind turbine can fast store or release a large
amount of kinetic energy in the rotating mass because of the power electronic
converter control, large moment of inertia and wide rotational speed range. Wind
turbine inertial response is essentially energy neutral, meaning that the period of
increased power is followed by a period of decreased power. The inherent inertial
response by wind turbines is also not same under all operating conditions but depend
upon active controls. Wind farm operation based on the maximum power capture
curve to extract maximum wind energy may need replacement to de-loading curve to
145
save the available power as reserves for the long-term frequency control. Various
supplementary inertial controls attribute, and designs have been proposed in the
literature to achieve active frequency regulation; inertial control and droop control
being a prominent one. Inertial control acting as extra loop utilize the rate of change
of frequency while droop control utilizes frequency difference between measured
frequency and actual frequency to adjust power to reduce frequency deviations. There
is no standard to quantify the effect of supplementary frequency regulation loop on
the maximum rate of change of frequency (ROCOF).
The wind inertial control increases the system inertia, but the inertial power may
mask the load changes for a few seconds because of the considerably released kinetic
energy from the rotating mass. Therefore, the synchronous generators may delay their
responses to the frequency events. Varied decentralized and distributed control
methodologies to implement inertial and AGC response at turbine level and real scale
power plant level with high ROCOF ride-through capability is still very nascent and
needs more attention as research in this field appears to be sparse.
Intelligent AGC
Conventional optimization techniques based on various mathematical programming
methods have been previously utilized in power system operation and control, but
they are constrained to perform a search operation for a local optimum. Modern smart
grid with multigenerational integrated power system calls for computational
intelligence (CI) and evolutionary computation (EC) techniques for robust and
optimum results in AGC regulation. Usage of an intelligent configuration has been
advocated as the best solution for improving the contribution of a renewable power
plant in AGC function [69]. It has been established that overall deviations in grid
146
frequency and regulation time can be reduced by incorporating intelligent controllers
and distributed AGC in general power systems operating under restrictions and
certain operating ranges.
With the availability of highly capable variable generators, automatic response to
frequency is getting improved for wind plants. Pitch angle control and rotor speed
control have been deduced as popular frequency control techniques at turbine level
[75], but the absence of intelligent controllers in the wind penetrated system can cause
stability problems and penalties causing consequent losses in revenues. Application
of intelligent algorithms incorporating a high degree of optimization and control
strategies can better respond to existing complex variables of power system models
and supports AGC in mitigating the demand and load imbalance in the interconnected
system. Intelligent load frequency controllers will be front runners to enhanced
performance, improved tuning and adaptive competencies. A more advanced
simulation involving high-fidelity system models and control capability is needed to
analyze the full potential and impact of AGC for wind penetrated system. These needs
to be real time verified and investigated for wind-integrated system with application
of intelligent techniques.
Fault Ride-through Capable Wind Farm Effect on AGC
With increasing renewable energy penetrations in the modern power system, all
national TSOs have put stringent ride through capability limits on these power
systems. Ride through capability refers to the ability of generators to remain stable
and connected during normally cleared balanced and unbalanced electrical faults on
a transmission grid. This constraint has been enforced as mandatory to evade
substantial loss of wind energy generation in the occasion of grid faults. Earlier trend
of disconnection of wind farms poses significant damage to frequency and voltage

147
regulation leading to a blackout in the case of a major fault. So as per various national
grid codes, their current wind plants are required to limit their active power as a
function of system frequency in case of over frequency event. Australian grid code
[76] requires their wind plant to provide ride through multiple faults even though it
is not cleared during reclosing sequence of transmission lines. Even though technical
solutions exist now for overcoming fault ride through behavior, there are other rising
issues which need to be further addressed. More involvement of type 3 & 4 wind
turbine by displacing synchronous generators will result in higher ROCOF and
instability in the total system. For longer duration faults, turbine blade pitching
strategies along with turbine converter control become necessary.
FRT compliance is assessed at the point of connection (PCC) while manufacturers
specify operating range of wind turbine generators at its low voltage (LV) terminals.
Depending upon short circuit ratio (SCR) at the point of coupling (PCC) and wind
turbine type, there can be varying stress on different parts of the turbine when it is
subjected to near-to-generator disturbances. Integration of wind energy with a
network having low short circuit ratio (SCR) has increased the risk of system
instability even though with FRT capable wind turbines. Wind farms with such fault
ride through (FRT) capabilities are very likely to deteriorate frequency regulation as
frequency deviations result in voltage instability. As a result, energy withdrawn by
FRT action may increase the size of any contingency and thus the magnitude of
frequency deviations leading to total system instability.
To analyze FRT implementation impact on frequency regulation performance of
wind farms and vice versa; it is important to conduct more research & assessment of
these FRT capable wind system.
148
Spatially Distributed Wind Farm Participation in AGC
A comprehensive literature exists which deals with lots of variability and uncertainty
in wind power output [77-80]. Wind power ramping down event is not as abrupt as
sudden loss of conventional generator and not as smooth as load ramp-down daily
cycles [81]. In spite of technological development and implementation, there is a
limitation on the ability of modern wind energy-based power system in contributing
to regulating frequency under few seconds after a contingency. Most of the solution
proposed in literature mimics the traditional AGC response of synchronous
generators. In a high wind penetrated power system, with varying wind power and
load uncertainties, there is a spatial variation of frequency standards and required
regulation over control area. So for wind energy system (WES), traditional lumped
ACE input changes to distributed ACE and conventional AGC changes to spatially
differentiated frequency regulation. One of the studies indicates that wind farms
spread over the large geographical area have relatively low regulation limits [82]. At
normal operation, the power output of a wind farm can vary up to the 15% of installed
capacity within 15 minutes [83]. The spatial distribution and intermittency of wind
resources affect power system on time scale being followed by AGC and raise the
cost of wind-based electricity. Research from NREL done under German weather
conditions shows that ancillary service requirements decrease substantially for a large
number of turbines and due to the spatial diversity of wind resources. Regulation
analysis from the same report also advocates that the physical separation of wind
plants has a lesser effect on regulation burden than a number of turbines [84].
Conventionally economic dispatch has been applied for generating unit participation
in AGC, but with the emergence of the individual frequency responsive wind energy
system, the advance investigation is needed for such spatially distributed & large
149
aggregation of wind turbines on AGC under various grid codes. AGC system along
with economic dispatch of respective electric network needs to have improved
modeling to account for unique features of wind generation and mitigating its
variability effect.
Frequency Control Ancillary Service (FCAS) As AGC Output
AGC manages continuous correction of frequency deviations as regulation raise or
regulation low service as part of market-based ancillary service in the energy market.
Frequency control, one of the ancillary service requirements are calculated by
contingency size, total inertia of power system and load demand. Frequency control
ancillary service (FCAS) providers bid their services into the FCAS markets in a
similar fashion like generating units bidding into the energy market and receive
financial incentives. Payments for ancillary services include payments for availability
and delivery of the services. Service providers must be able to show their
conformance to the ancillary services specification. FCAS represents a guarantee that
power will be continued to be delivered in the case of contingency.
FCAS availability holds a very important value for high wind penetrated power
system with low inertia. Wind farms which produces electricity at low prices can gain
financial revenues by participating in the provision of regulation services. In a
deregulated environment, energy market participants are required to make
supplementary control contributions to transmission system operator (TSO) for the
stable and reliable power system. To meet control area performance criteria, AGC
participants must track and keep a record of instantaneous fluctuations in demand
load for the continuous duration so as to keep ACE within limits. For a transmission
system operator, assignment of control duties gets constricted with the presence of
150
wind system. System operator aims to minimize the payments to participants for
providing AGC ancillary service while wind participant strives to maximize their
revenues.
Australian energy market operator (AEMO) has the requirement for FCAS to ensure
the frequency to be contained within 49 to 51 Hz in the event of step changes in
supply that result from credible contingencies known as “separation event.”
According to Australia’s national electricity market (NEM), it is compulsory for
generators >=30 MW to participate in FCAS [85] though none of the wind farms
participate in it. An impact assessment of large-scale renewable energy target done
for Australian Energy Market Commission (AEMC) reported an increase in
regulation requirement from ±120MW to ±800MW and increase in frequency
regulation cost from 10 million per annum to 200 million per annum in 2020 due to
projected increase in intermittent wind energy. Causer-pays methodology is being
implemented in Australia for settling regulation costs, with generators or loads paying
for any deviation from their expected dispatch. With increased wind penetration, the
majority of the cost is anticipated to be borne solely by wind farms leading to an
escalation in regulation costs for wind generators from around $0.40 /MWh at present
to $6 - 8 /MWh in 2019-20 [86]. This increase in wind farm costs could have
significant repercussions.
North American Electric Reliability Corporation (NERC) has enacted control
performance standard (CPS) for evaluating frequency control performance in a
control area. Control areas must not be less than 95 % compliant with CPS1 and no
less than 85% compliant with CPS2 [87]. A Study by Wang & McCalley indicates
that CPS1 & CPS2 deteriorate with increasing wind penetration and this effect is
151
observed more for large interconnected systems [88]. Improved load forecasting
along with better AGC is required to improve these control performance standards.
According to European Network for Transmission System Operators for Electricity
(ENTSO-E), all TSOs must ensure regulation service which should be at least 50%
operational within 15 seconds and fully operational within 30 seconds of any
contingency exceeding ±20mHz [89]. A very fast response is needed to seize
frequency decline and control frequency nadir within first 5-10 seconds. Energy
storage technologies have fast response making them suitable for providing
frequency regulation ancillary services.
According to a report by California Energy Commission (CEC), application of energy
storage for the regulation purpose can be around two to three times as effective as
adding a combustion turbine to the system [90].
Besides technology, market policies, as well as system infrastructure, also play vital
role satisfactory AGC performance in a wind energy integrated power system. Wind
farms can boost their revenue and decrease total costs by implementing frequency
regulation technology. More research regarding technology, policies, and
infrastructure is needed for incorporating AGC in wind system to enable their active
participation in FCAS, better energy economics and compliance to control
performance standards. Proper management of FCAS can be a key factor for better
penetration of the wind in a deregulated complex power system.
3.7. Conclusion
This chapter presented present research and perspective challenges for wind
participation in frequency regulation. Effective research and implementation of these
terminologies in AGC will help wind energy market participants to make better
152
contributions to frequency control ancillary services to transmission system operator
(TSOs) and an overall more reliable power system. Therefore, extensive research is
required to review and analyze wind energy association with frequency regulation
criteria, capabilities, and technologies for ensuring cost effective, industry prone
frequency regulation system leading to a stable wind integrated large power system.
With improved turbine and intelligent controller technologies, AGC researchers have
to look deeply into several new aspects for achieving long-term economic
sustainability of wind energy.
153
3.8. References
1. ENTSO-E (2012). Operational reserve ad hoc team report, final version,
Available at www.entsoe.eu.
2. AEMO (2012), Frequency Control Ancillary Services. Available at
aemo.com.au.
3. National Grid (2012). The Grid Code, Issue 4, Revision 10, National Grid
Electricity Transmission plc, UK, January 2012.
4. B. Kirby (2007). Ancillary Services: Technical and Commercial Insights.
Available at www.science.smith.edu.
5. Verband der Netzbetreiber - VDN - e.V. beim VDEW. Transmission code 2007.
Network and system rules of the German transmission system operators.
Available at www.vde.com.
6. National Grid (2016). Mandatory Frequency Response | National Grid. [online]
Available at: http://www2.nationalgrid.com.
7. Christer Bäck (2013). Nordic implementation of Frequency Restoration Reserve
Automatic (FRR-A), Svenska kraftnät (SVK). Available at www.svk.se
8. Ignacio de la Fluente. Ancillary services in Spain: Dealing with high penetration
of RES. Red Eléctrica de España. Available at www.reshaping-res-
policy.eu/downloads/topical%20events/de-la-Fuente_Ancillary-Services-in-
Spain1.pdf.
9. EirGrid. Eirgrid grid code version 4.0, Available at www.eirgrid.com.
10. Hewicker, C., Werner, O. and Zeigler, H. (2013). Qualitative analysis of cross-
border exchange of balancing energy and operational reserves between
Netherlands and Belgium. Available at: http://www.tennet.eu.
154
11. TenneT, The Netherlands, Samenvatting uit “Summary of the Current Operating
Principles of the UCPTE” met betrekking tot de primaire en secundaire regeling,
Dec. 2003.
12. PJM (2003). USA, Manual 10: Pre-Scheduling Operations, Revision: 16.
[Online]. Available at www.pjm.com/contributions/ pjm-manuals/
pdf/m10v161.pdf.
13. PJM (2006). USA, Manual 12: Dispatching Operations, Revision: 13, May 26.
Available at www.pjm.com/contributions/pjmmanuals/pdf/m12v13.pdf.
14. PJM (2005). USA, Manual 35: Definitions and Acronyms, Revision: 05.
Available at www.pjm.com/contributions/pjm-manuals/ pdf/m35v051.pdf.
15. Dan Woodfin (2015). Methodology for Determining Minimum Ancillary
Service Requirements. Available at www.ercot.com.
16. T. Borsche, A. Ulbig, G. Andersson, (2014). A new frequency control reserve
framework based on energy-constrained units, Power Systems Computation
Conference (PSCC) 2014, pp. 1-7, 18-22.
17. Nordel (2006). Agreement regarding operation of the interconnected Nordic
power system (System Operation Agreement). Technical report. Available at
www.entsoe.eu.
18. UCTE. P1 - Policy 1: Load-Frequency Control and Performance [C]. Technical
report, 2009. Available at www.eeh.ee.ethz.ch.
19. AEMO, (2016). Future Power System Security Program. Available at
aemo.com.au.
20. Goran Andersson (2012). Dynamics and Control of Electric Power Systems,
Lecture 227-0528-00, ITET ETH. ETH Zurich. Available at
www.eeh.ee.ethz.ch.
155
21. IEEE (1970). Standard definitions of terms for automatic generation control on
electric power systems, IEEE committee report. IEEE Trans. Power App. Syst.
PAS-89:1356–64
22. Western Power (2011). Technical Rules, Available at
www.westernpower.com.au.
23. R. de Almeida and J. P. Lopes (2007). Participation of Doubly Fed Induction
Wind Generators in System Frequency Regulation. IEEE Transactions on Power
Systems, vol. 22, pp. 944–950.
24. X. Juankorena, I. Esandi, J. Lopez, and L. Marroyo (2009). Method to Enable
Variable Speed Wind Turbine Primary Regulation, in International Conference
on Power Engineering, Energy and Electrical Drives, pp. 495–500.
25. M. Liserre, R. Cárdenas, M. Molinas, and J. Rodriguez (2011). Overview of
Multi-MW Wind Turbines and Wind Parks, IEEE Transactions on Industrial
Electronics, vol. 58, no. 4, pp. 1081–1095.
26. J. Morren, S. de Haan, W. Kling, and J. Ferreira (2006). Wind Turbines
Emulating Inertia and Supporting Primary Frequency Control, IEEE
Transactions on Power Systems, vol. 21, no. 1, pp. 433 – 434.
27. Gowaid, A. El-Zawawi, and M. El-Gammal (2011). Improved Inertia and
Frequency Support from Grid-Connected DFIG Wind Farms, in Power Systems
Conference and Exposition (PSCE), 2011 IEEE/PES, pp. 1–9.
28. H. Ma and B. Chowdhury (2010). Working Towards Frequency Regulation with
Wind Plants: Combined Control Approaches, Renewable Power Generation,
IET, vol. 4, no. 4, and pp. 308–316.
29. Hansen, A. D. (2010). Evaluation of power control with different electrical and
control concepts of wind farms, Tech. rep., Project UpWind, Roskilde, Denmark.
156
30. N. R. Ullah, T. Thiringer and D. Karlsson (2008). Temporary Primary Frequency
Control Support by Variable Speed Wind Turbines— Potential and
Applications, in IEEE Transactions on Power Systems, vol. 23, no. 2, pp. 601-
612.
31. G. C. Tarnowski, P. C. Kjar, P. E. Sorensen and J. Ostergaard (2009). Variable
speed wind turbines capability for temporary over-production, IEEE Power &
Energy Society General Meeting, Calgary, AB, pp. 1-7.
32. Ramtharan G, Ekanayake JB, Jenkins N. (2007). Frequency support from doubly
fed induction generator wind turbines. IET Renew Power Gen; 1: 3–9.
33. A. Aziz, G. M. Shafiullah, A. Stojcevski and A. Mto (2014). Participation of
DFIG based wind energy system in load frequency control of interconnected
multi-generation power system. Power Engineering Conference (AUPEC).
Australasian Universities, Perth, WA, pp. 1-6.
34. A.B. Attya, T. Hartkopf (2013). Wind farms dispatching to manage the
activation of frequency support algorithms embedded in connected wind
turbines, International Journal of Electrical Power & Energy Systems, Volume
53, Pages 923-936.
35. Hughes FM, Anaya-Lara O, Jenkins N, Strbac G. (2005). Control of DFIG-based
wind generation for power network support. IEEE T Power Syst; 20: 1958–1966.
36. Morren J, de Haan SWH, Kling WL, Ferreira JA. (2006). Wind turbines
emulating inertia and supporting primary frequency control. IEEE T Power Syst;
21: 433–434.
37. Mauricio JM, Marano A, Gomez-Exposito A, Ramos JLM. (2009). Frequency
regulation contribution through variable speed wind energy conversion systems.
IEEE T Power Syst; 24: 173–180.
157
38. Gautam D, Goel L, Ayyanar R, Vittal V, Harbour T. (2011). Control strategy to
mitigate the impact of reduced inertia due to doubly fed induction generators on
large power systems. IEEE T Power Syst; 26: 214–224.
39. National Grid (2010). Requirements for System Inertia, National Grid Code
Frequency Response Working Group, UK.
40. J. Morren, J. Pierik, and S. Dehaan (2006). Inertial response of variable speed
wind turbines, Electric Power Systems Research, vol. 76, no. 11, pp. 980-987.
41. A. Mullane and M. O'Malley (2005). The Inertial Response of Induction-
Machine-Based Wind Turbines, IEEE Trans. Power Systems, vol. 20, no. 3, pp.
1496-1503.
42. O. Anaya-Lara, F. Hughes, N. Jenkins, and G. Strbac (2006). Contribution of
DFIG based wind farms to power system short-term frequency regulation, IEE
Proceedings Generation, Transmission and Distribution, vol. 153, no. 2, pp. 164-
170.
43. Kundur, P., Balu, N. and Lauby, M. (1994). Power system stability and control.
New York: McGraw-Hill.
44. Xue Yingcheng, Tai Nengling (2012). System frequency regulation in doubly
fed induction generators, International Journal of Electrical Power & Energy
Systems, Volume 43, Issue 1, Pages 977-983.
45. K. Clark, N. W. Miller, J. J. Sanchez-Gasca (2010). Modelling of GE Wind
Turbine-Generators for Grid Studies, Version 4.5. General Electric International,
Inc.
46. ENTSO-E (2012). Draft Network Code for Requirements for Grid Connection
applicable to all Generators, European Network for Transmission System
Operators for Electricity.
158
47. R. de Almeida, Castronuovo ED, Lopes JAP. (2006). Optimum generation
control in wind parks when carrying out system operator requests. IEEE
Transactions on Power Systems, 21(2):718-25.
48. Janssens NA, Lambin G, Bragard N. (2007). Active power control strategies of
DFIG wind turbines. In: 2007 IEEE power tech Lausanne; p. 516-21.
49. Kanellos FD, Hatziargyriou ND. (2008). Control of variable speed wind turbines
in islanded mode of operation. IEEE Transactions on Energy Conversion,
23(2):535-43.
50. Courtecuisse V, Robyns B, Francois B, Petit M, Deuse J. (2008). Variable speed
wind generators participation in primary frequency control. Wind Engineering,
32(3):299-318.
51. Holdsworth L, Ekanayake JB, Jenkins N. (2004). Power system frequency
response from fixed speed and doubly fed induction generator-based wind
turbines. Wind Energy; 7 (1):21-35.
52. Conroy JF, Watson R. (2008). Frequency response capability of full converter
wind turbine generators in comparison to conventional generation. IEEE
Transactions on Power Systems; 23 (2):649-56.
53. Moutis P, Loukarakis E, Papathanasiou S, Hatziargyriou ND. (2009). Primary
load frequency control from pitch-controlled wind turbines. In: 2009 IEEE
power tech Bucharest; p. 1-7.
54. P. Moutis, S. A. Papathanassiou, and N. D. Hatziargyriou (2012). Improved
load-frequency control contribution of variable speed variable pitch wind
generators," Renewable Energy, Vol. 48, pp. 514-523.
55. J. Aho, A. Buckspan, J. Laks, P. Fleming, Y. Jeong, F. Dunne, M. Churchfield,
L. Pao and K. Johnson. (2012). A Tutorial of Wind Turbine Control for
159
Supporting Grid Frequency through Active Power Control. Montreal:
Proceedings of the 2012 American Control Conference.
56. Garcia, J. (2010). Frequency Control. 20120035775.
57. Stanton N. K., Giri J. C. and Bose A. (2007). Energy management. In Power
system stability and control, ed. L. L. Grigsby. Boca Raton, FL: CRC Press
58. Poul Sorensen, Anca d. Hansen, Florinn Iov, Frede Blaabjerg and Martin H.
Donovan (2005). Wind farm models and control Strategies, Riso-R Report,
Denmark. Available at orvit.dtu.dk.
59. Garcia A. D. (2012). Models for Impact Assessment of Wind-Based Power
Generation on Frequency Control, In Control and Optimization Theory for
Electric Smart Grids, vol. 3, A. Chakraborty, and M. Ilić (Eds.), Springer-Verlag,
Berlin.
60. Amenedo L. R., Arnalte S., and Burgos J. C. (2002). Automatic generation
control of a wind farm with variable speed wind turbines. IEEE Trans. Energy
Convers. 17(2): 279-284.
61. A. Buckspan, J. Aho, P. Fleming, Y. Jeong and L. Pao (2012). Combining droop
curve concepts with control systems for wind turbine active power control, 2012
IEEE Power Electronics and Machines in Wind Applications, Denver, CO, pp.
1-8.
62. Anca D. Hansen, Poul Sørensen, Florin Iov, Frede Blaabjerg (2006). Centralised
power control of wind farm with doubly fed induction generators, Renewable
Energy, Volume 31, Issue 7, Pages 935-951, ISSN 0960-1481.
63. J. D. Grunnet, M. Soltani, T. Knudsen, M. Kragelund, T. Bak (2010). Aeolus
toolbox for dynamic wind farm model simulation and control", Proc. Eur. Wind
Energy Conf., pp. 3119-3129.
160
64. Basit, Abdul, et al. "Wind power integration into the automatic generation
control of power systems with large-scale wind power." The Journal of
Engineering (2014).
65. E. Ela, M. Milligan, and M. O'Malley (2011). A flexible power system
operations simulation model for assessing wind integration, 2011 IEEE Power
and Energy Society General Meeting. IEEE, pp. 1-8.
66. Michael Antonishen, H. Y. Han, T. K. A. Brekken, A. von Jouanne, Alex
Yokochi, D.A. Halamay, J. Song, D. B. Naviaux, J. D. Davidson, Alex Bistrika.
(2012). A Methodology to Enable Wind Farm Participation in Automatic
Generation Control Using Energy Storage Devices, IEEE Power and Energy
Society General Meeting, IEEE , vol., no., pp.1,7, 22-26.
67. Luo C. and B.T. Ooi. (2006). Frequency deviation of thermal power plants due
to wind farms”, IEEE Trans. Energy Conv., vol. 21, no. 3, pp. 708–716.
68. Luo C., Far H. G., Banakar H., Keung P. K. and Ooi B.T. (2007). Estimation of
wind penetration as limited by frequency deviation”, IEEE Trans. Energy Conv.,
vol. 22, no. 3, pp. 783–791.
69. Hassan Bevrani and Pourya Ranjbar Daneshmand (2012). Fuzzy Logic-Based
Load-Frequency Control Concerning High Penetration of Wind Turbines, EEE
Systems Journal, Vol. 6, No. 1.
70. P. Anderson and A. Fouad, Power System Control and Stability. Piscataway,
US: Wiley - IEEE Press, 2002.
71. M. Hwang, E. Muljadi, J. W. Park, P. Sørensen, and Y. C. Kang, "Dynamic
Droop–Based Inertial Control of a Doubly-Fed Induction Generator," in IEEE
Transactions on Sustainable Energy, vol. 7, no. 3, pp. 924-933, July 2016.
161
72. K. V. Vidyanandan, N. Senroy, "Primary frequency regulation by deloaded wind
turbines using variable droop," IEEE Trans. Power Syst., vol. 28, no. 2, pp. 837-
846, May 2013.
73. Miller N. W., Shao M., and. Venkataraman S (2011). California ISO (CAISO)
frequency response study, GE Energy.
74. Andreas U., Theodor S. B. and Göran A. (2014). Impact of Low Rotational
Inertia on Power System Stability and Operation, Submitted to IFAC World
Congress 2014, Capetown, South Africa
75. Juankorena X., Esandi I., Lopez J. and Marroyo L. (2009). Method to Enable
Variable Speed Wind Turbine Primary Regulation, in International Conference
on Power Engineering, Energy and Electrical Drives, pp. 495–500.
76. Western Power (2011). Technical Rules. Available at
www.westernpower.com.au.
77. Undrill J. (2010). Power and frequency control as it relates to wind-powered
generation, Lawrence Berkeley National Laboratory, Tech. Rep. LBNL-4143E
78. Martinez C., Xue S. and Martinez M. (2010). Review of the recent frequency
performance of the eastern, western and ERCOT interconnections, Lawrence
Berkeley National Laboratory, Tech. Rep. LBNL-4144E
79. Illian H. (2010). Frequency control performance measurement and requirements,
Lawrence Berkeley National Laboratory, Tech. Rep. LBNL-4145E. Available at
www.certs.lbl.gov
80. Mackin P. (2010). Dynamic simulations studies of the frequency response of the
three U.S. Interconnections with increased wind generation, Lawrence Berkeley
National Laboratory, Tech. Rep. LBNL-4146E
162
81. Garcia A. D. (2012). Models for Impact Assessment of Wind-Based Power
Generation on Frequency Control, In Control and Optimization Theory for
Electric Smart Grids, vol. 3, A. Chakraborty, and M. Ilić (Eds.), Springer-Verlag,
Berlin, 2012
82. Holttinen, H. (2005). Impact of hourly wind power variation on the system
operation in the Nordic Countries. Wind Energy, vol. 8, 2, 197–218
83. Ackermann T. (2005). Wind Power in Power System, John Willey & Sons.
84. NREL (1999). Technical report: Analysis of Wind Power Ancillary Services
Characteristics with German 250-MW Wind Data. Available at www.nrel.gov.
85. Thorncraft S. R. and Outhred H. R. (2007). Experience with Market-Based
Ancillary Services in the Australian National Electricity Market, Proceedings of
the IEEE PES General Meeting, Tampa Florida, USA, June 24-28
86. AMEC. (2011). Impact of the LRET on the costs of FCAS, NCAS and
Transmission augmentation. Available at amec.gov.au
87. NERC (2014). Violation Severity level Matrix-Reliability Standard. Available
at www.nerc.com
88. Wang, C. and McCalley, J. D. (2013). Impact of wind power on control
performance standards, International Journal of Electrical Power & Energy
Systems, 47, 225-234
89. ENTSO-E. (2004). UCTE Operational Handbook Policy 1: Load-Frequency
Control and Performance”- Appendix A1, V1. 9. Available at www.entsoe.eu
90. KEMA Inc. (2010). Research Evaluation of Wind Generation, Solar Generation,
and Storage Impact on the California Grid, Available at www.energy.ca.gov
163
CHAPTER 4
FREQUENCY RESPONSIVE FULL CONVERTER
BASED VSWTG: MODELLING AND ASSESSMENT
This chapter presents the generic modeling of type 4 full converter based wind turbine
generator system for application in frequency ancillary service investigations under
varying wind speed and varying reference power setpoint. Prevalent generic model,
manufacturer specific proprietary generic model along with detailed wind turbine
model with the synchronous generator is also provided to highlight various modeling
framework difference. Descriptions of individual sub-models of proposed generic
model are presented in detail and performance results are compared and validated
with GE’s proprietary generic model & detailed WTG model using simulations in the
MATLAB ® Power System Block set.
4.1. Introduction
After displacing dominant simple constant speed turbines, state of the art Variable
Speed Wind Turbine Generator (VSWTG) technology is a pioneer in currently
installed wind energy projects. Even though doubly fed induction generators based
type 3 is the predominant technology nowadays, full converter based type 4 turbine
technology is expected to become the industry standard in medium terms. Joint
Research Centre’s (JRC) 2014 report [1] analyzing wind turbine purchase agreements
identify the ending of pre-eminence of type 3 technology due to increasing popularity
of type 4 wind turbine system. Considering the connection trends, Australian Electric
164
Market Operator (AEMO) also assume an increase in type 3 and type 4 wind turbines
in its Wind Turbine Plant Capability Report 2013[2].
Figure 4-1 Schematic Diagram of Full Converter Based VSWTG
With relatively new wind technology, especially full converter based type 4 VSWTG
as shown in Figure 4-1, development of suitable models to analyze grid code
requirements is a significant subject matter. Not all models are suited for specific
power system phenomenon due to substantial issues. Generalized models having
reduced dataset necessities are employed to attain approximate dynamic wind farm
performance. A simplified yet absolute representation of full converter based wind
turbine generator principles is requisite for various bulk studies. Implementation of
suitable wind turbine model containing dynamic features satisfying ancillary service
requirements is the first step in such case. Very few research papers are available
dealing with generic models of full converter based WTG system which is a relatively
new technology. WECC’s generic model and GE model is the publically available
model while IEC model is not disclosed to the public. Most of the other available
research studies focus on controller development for frequency regulation utilizing
detailed model [13] [14] [15]. With no standard guidelines available, different
165
approaches have been used in adopting the generic model of type 4 WTG [3] [4] [5]
[6] [7] [8] [9], especially in initializing aerodynamic model values and application of
pitch control block. All of these models are aimed at short term simulation of the bulk
network for transient stability analysis under the assumptions of constant wind speed
and constant power setpoints. This chapter proposes a generalized model which
follows the basic structure of Western Electricity Coordinating Council (WECC)
generic type 4 wind turbine generators (WTG) model. Some modifications included
to make it suitable for frequency regulation studies. Model performance is evaluated
and validated with simulation response of GE’s proprietary generic model as well as
detailed type 4 wind turbine model.
4.2. Type 4 VSWT Model with Synchronous Generator
Figure 4-2 Detailed Simulation Model of Type 4 VSWTG
IEEE defines Type 4 wind turbine as a variable speed wind turbine with the
synchronous or asynchronous generator connected to the grid via full-scale power
166
converter [10]. Figure 4-2 shows the detail simulation model of full converter based
type 4 wind turbine generator model. These types of models are suitable for short
circuits analysis and electromagnetic transient investigations. A salient pole
synchronous generator integrated with grid via full capacity converters system is
utilized here in detail model of type 4 WTG systems. It also includes Cp calculation
based aerodynamic model and double mass based gear drive train model. Power
conditioning system consists of a diode rectifier, a DC-DC boost converter, and a
DC/AC IGBT-based pulse width modulated (PWM) inverter. Rotor speed is
completely decoupled from grid frequency through controlled frequency converter.
4.2.1. Converter Topology
Appropriate power converters with a full capacity similar to that of generators
implicitly segregate their dynamics from the main grid. A representation of converter
topology is given in Figure 4-3. With variable input supply, WES generator stator
provides 3-phase AC voltage of variable frequency and variable magnitude which is
later rectified into DC supply through 3-phase diode based Graetz bridge rectifier. A
regulated high DC voltage supply across capacitor suitable for the inverter is
subsequently achieved after being boosted through DC/DC single channel boost
converter. Steady state boost converter operation holds the following relation 4.1:
4.1
Chopper output voltage (Vout coincides with inverter dc bus voltage of inverter while
chopper input voltage (Vin is same as three phase rectifier output voltage. Input dc
voltage varies with generator speed while output dc voltage is maintained constant
higher value. The ratio of a turn on time to switching time period termed as duty cycle
(D) is attuned through a series of PI regulators for boost converter to control the
167
generator active power. IGBT-based PWM inverter using pulse width modulation
(PWM) employing 3 kHz carrier frequency then finally inverts dc voltage to essential
AC voltage and frequency ratings. This grid side converter maintains dc link voltage
by active power injection to the grid. The inductor smoothens any ripple effect in dc
link.
Figure 4-3 DC/DC Boost Converter & Decoupled VOC Inverter Interfaced Synchronous Generator
Wind Energy System.
Grid side connected IGBT inverter is controlled through voltage-oriented control
(VOC) with a decoupled controller. Closed loop current control of transformed 2-
phase line current in dq synchronous frame (i , i provides independent active and
reactive power control. Decoupled controller output can be expressed as [11]:
V K i∗ i WL i V 4.2
V K i∗ i WL i V 4.3
W is grid angular frequency as well as synchronous reference frame speed, W L i
& W L i are induced reference voltages in synchronous frame and K are PI
168
controller transfer function. Assuming balanced DC voltage, DC side voltage and
inverter AC line to neutral voltage can be related as [12] [13] [14]:
√
V m V 4.4
√
More details about converter operation can be referred from [11]
4.2.2. Synchronous Generator Model
The electrical part of the salient pole machine is represented by a fifth-order state-
space model considering the dynamics of the stator, field, and damper windings.
Computer simulation of synchronous machine model is implemented utilizing the
approach given in [15] where variables abc are transformed directly to rotor reference
frame using park’s transformation matrix Ks which has arbitrary reference frame
speed equal to rotor speed (ω = ω ) and presented in Figure 4-4.
Figure 4-4 Simulation Framework of Synchronous Machine in Rotor Reference-Frame [15]
Basic voltage and flux linkage are represented by equation 4.5 – 4.10 as given below:
v R ∗i ω ∗ φ, φ L ∗I 4.5
169
Where: fd = field winding, kd= d-axis damper winding, kq = q-axis damper winding, k
= no. of damper circuits, [R] = winding resistances diagonal matrix in d q axis
[L] = winding self and mutual inductances matrix in d q axis
[ω] = rotor speed matrix having all zeros except ω 1,2 ω ; ω (2, 1) =-ω
v = voltage vector = [V V V V V ]'
φ= d- and q- axis flux linkage vector = [φ φ φ φ φ ]'
i = current vector = [i i i i i ]'
With as base angular velocity corresponding to rated or base frequency,
manipulation of voltage and flux linkage equations gives following equations to be
used in simulation [15]:
φ V φ φ φ 4.6
φ V φ φ φ 4.7
φ V φ φ 4.8
φ e φ φ 4.9
φ V φ φ 4.10
i φ φ 4.11
i φ φ 4.12
170
i φ φ 4.13
i φ φ 4.14
i φ φ 4.15
ω T T 4.16
Under balanced operation, per unit electrical torque and electrical power output is
calculated as:
T φ ∗ i φ ∗ i 4.17
P Vi V i 4.18
4.3. WECC’s Type 4 WTG Generic Model
Implemented in Siemens PTI PSS®E and GE PSLF software, Western Electricity
Coordinating Council’s (WECC) Renewable Energy Modeling Task Force (REMTF)
developed reduced-order, positive sequence generic models [3] are appropriate for
dynamic stability evaluation of power system subsequent to system disturbances.
WECC’s generic model of type 4 wind turbines generator system pursues modular
approach where all imperative functions are characterized by individual submodels
working in unison.
Figure 4-5 represents WECC’ generic type 4 model consisting of three respective
submodels as implemented in PSLF. Generator/Converter dynamics block is
implemented with input active and reactive current command yielding real and
reactive power as an output. Second sub-block constitutes a simplified linearized
model for wind aerodynamics and two mass drive train model depicting torsional
171
mode oscillations. The electrical control model incorporating active and reactive
power control uses reference power as an input and current command as output. PSSE
model incorporates only electrical control block and generator-converter block.
Verification of this model’s performance amid unbalanced events is challenging due
to lack of any standard precept from WECC. Additionally, this model is assessed to
be incompatible for investigations associated with volatile wind speed amid dynamic
simulation. This generic model is suitable for bulk power system transient stability
analysis during the early stage of network interconnection planning but not suitable
for investigations like active power-frequency control analysis under varying wind
and changing power demand conditions.
Figure 4-5 WECC Type 4 Wind Turbine Generator Model [3]
4.4. Proprietary Type 4 Generic Model
GE’s type 4 generic model distinguish itself from WECC’s generic model in terms
of electrical control incorporating dynamic braking resistance power consumption,
172
additional torque controller and specific aerodynamic model based on two-
dimensional manufacturer specific curve given below:
C θ, λ ∑ ∑ α,θλ. 4.19
Figure 4-6 GE’s Type 4 Wind Turbine Generator Generic Model (Reformulated [5])
More details about the model can be found in reference work [5]. When power order
generated in this model is higher than the electrical power to be delivered to network,
excess energy is absorbed in braking resistance. So rotor model, as well as reference
speed generator, receives electrical power along with dynamic braking resistor
173
power. During real power priority, real current order, IPCMD is limited by a minimum
of a maximum of temperature dependent converter current and hard active current
limit. GE’s proprietary generic model which is applicable in bulk power system
disturbance studies is represented in Figure 4-6.
4.5. Proposed Type 4 VSWTG Generic Model
Figure 4-7 Proposed Type 4 WTG Model
Pursuing WECC second generation generic modeling framework, some additional
functionality like pitch control and available wind power look up has been
supplemented in proposed model making it admissible for frequency control and
associated investigation. Rotor speed is also incorporated in active power control loop
to account for fluctuating speed. The complete model is shown in Figure 4-7, and its
various sub-modules are discussed in succeeding sections. Salient aspects of WECC
174
generic WTG model and GE type 4 WTG model is distinguished with the proposed
full converter based WTG model and presented in Table 4-1.
Table 4-1 Comparison of Type 4 WTG Models Suitable for Frequency Regulation Analysis
WECC’s Generic Type 4 GE’s Proprietary Generic Modified Generic Type 4
WTG Model Type 4 WTG Model WTG Model
 Applicable for constant  Can be applied to variable  Aerodynamic model & pitch
wind speed evaluation. wind speed related model added making it
 First generation model investigations. suitable for varying winds as
excludes mechanical mass  Includes two-dimensional well as dynamic power
model, pitch model, and aerodynamic model coupled reference setpoints.
aerodynamic model while with double mass model &  Available wind power lookup
second generation model pitch model makes it suitable for primary
includes double mass drive  Normal as well as de-loaded and secondary frequency
train model. operations are possible control assessment
 Applicable to normal  The model utilizes active and investigations. Nominal
operation only. reactive power controller in value or calculated value is
 It utilizes active and reactive the electrical circuit for utilized depending upon wind
power controller in the transient analysis. speed.
electrical circuit for  No study available for  The model utilizes simplified
transient stability analysis. assessment of model for electrical control with only
 The model requires varying reference power active power controller in the
additional modifications for command / setpoints. electrical circuit for
application in frequency frequency control analysis.
regulation investigations. Rotor speed variation is an
additional component added
to the electrical control.
Detailed Model: Mainly applied for short circuit electromagnetic transient analysis over the shorter
time frame. Due to longer simulations times and complex sub-component representation, detail type
4 models are not deemed suitable for primary & secondary frequency control investigations.
175
4.5.1. Wind Turbine Mechanical Model
Wind turbine’s limitation to inertia control contribution can be contemplated through
rotor speed dynamics making the coupling imperative between mechanical drive train
model and the aerodynamic model for accurate speed dynamics reflection. The
mechanical model provides significant turbine dynamics representation associated
with WTG instantaneous shaft speed. Combination of tip-speed ratio calculation
block, power coefficient calculation block and aerodynamic power calculation block
constitute the complete wind turbine model. One mass lumped mechanical shaft
model as given by equation [16] [17] is implemented in this generic WTG model.
ω
T T Dω 4.20
Where H is the total inertia constant given by the ratio of the stored kinetic energy
(MJ) at synchronous speed to the machine rating in MVA. D is the shaft damping
constant in p.u. power / p.u. speed with a default value of 1 while ω denotes the
angular speed. This lumped model assumes constant turbine and generator angular
speed. Aerodynamic model intricacy is highly influenced by the representation of the
power coefficient Cp. Cp is the dimensionless fractional value conveying the amount
of available wind power that can be extracted by the rotor and given as the ratio of
the available wind power to the available wind kinetic power.Betz limit defines the
optimal value of performance coefficient as 59.26% which is applicable to all wind
turbine. WECC proposed a simplified aerodynamic model in generic modeling while
GE’s generic model utilize two-dimensional Cp curve model. WECC proposal of
linear aerodynamic model doesn’t provide any publically available standard
guidelines initializing the model.Initialization of pitch angle and power in linear
aerodynamic model finds scant mentions in current research literature. Suitability of
WECC simplified the linear aerodynamic model, and detailed Cp calculation based
176
model [18] [19] like the one given by 4.21 & 4.22 is assessed in this study before
making final recommendations:
C λ, θ C C θ C e + C λ 4.21
.
4.22
.
The empirical power coefficients c1 to c6 are: c1 = 0.5176, c2 = 116, c3 = 0.4, c4 =
5, c5 = 21 and c6 = 0.0068. λi is tip speed ratio at ith time step. Cp acheives a
maximum value of 0.48 for zero degree pitch angle and nominal value of lambda =
8.1. Maximum extractable wind power defined as turbine mechanical power utilizing
detailed Cp based aerodynamic is given as 4.23 .
1
Pwind = ρ ACp ( , θ V 4.23
2
Where ρ is the air density, A is the turbine blades swept area (πR2 where R is blade
length), Cp is the turbine power coefficient, λ is tip speed ratio defined as rotor
peripheral speed to the wind speed, Vwind is the velocity of the incident wind. At very
high wind speeds, the power is limited to rated power by increased pitch angle. At
very low wind speeds, an increase in pitch angle increases the value of Cp thus
increasing power captured from the wind. Simplified linear aerodynamic model [6]
expressed by 4.24 has reduced parameters leading to lesser simulation time and based
on the assumption of constant wind during initialization without involving power
coefficient curve.
A simulation model of the linear aerodynamic model is shown in Figure 4-8.
P P K ∗ θ 4.24
Where & θ is the initial mechanical power and initial pitch angle respectively.
- Kaero is the aerodynamic gain factor
- θ is the pitch angle of the turbine blades
177
The aerodynamic gain factor Kaero has been given a default value of 0.007 [16-17].
θ ∗ 1
.
The value of Theta2 supplied is 26◦. Vwindpu is the velocity of the wind in per-unit
of the rated wind velocity (13 m/s).
P _ k C _ V _ 4.25
Figure 4-8 Simplified Aerodynamic Simulation Model
Equation 4.25 is the normalized form of equation 4.23, where Pm_pu is per-unit
nominal power for particular values of ρ and A, Cp_pu is per-unitized performance
coefficient of the maximum value of Cp, Vwind_pu is per-unit wind speed in p.u. of the
rated wind speed. Kp is is less than or equal to 1 and denoted as power gain for Cp_pu=1
p.u. and Vwind_pu = 1 p.u. As seen in Figure 4-9, the power generated is proportional
to available wind power i.e.; Ppu = V for underrated wind speed and power
generated is Ppu = 1 for beyond rated wind speed. The actual mechanical power
delivered by the wind turbine to the generator is thus obtained by subtracting per unit
wind power from maximum available wind power.
178
4.5.2. Pitch Control Model
At pitch angle=27 degree At pitch angle=0 degree
Figure 4-9 Turbine Power Characteristics
Figure 4-10 Pitch Controller Block
Figure 4-9 shows the per unitized turbine output power w.r.t to mechanical power at
pitch angle = 0 degrees and pitch angle =27 degree. Pitch control enables optimum
control of aerodynamic wind power by rotor blade pitching to regulate turbine torque.
Maximum power of 1 p.u. is available at zero pitch angle while it is highly reduced
with highest pitch angle. Though pitch control is not adopted in WECC generic type
4 model, a simple variable pitch controller without gain scheduling has been proposed
179
for fluctuating wind speed adaptable WTG model. Simplified aerodynamic model in
the proposed WTG model inputs pitch angle at zero in the case of low wind speed
otherwise amplitude & rate limited pitch angle is initialized through pitch controller
in case of wind speed more than rated speed. Following equations mathematically
express pitch angle obtained through a series of the PI regulator as shown in Figure
4-10:
.
θ K ω ω . K P
P 4.26
Default speed reference calculation according to generated electrical power is done
as per equation:
ω A A P A P 4.27
Where , and are coefficients for turbine operation at maximum power.
λ= 4.28
ω v 4.29
For optimal operation and per unit base:
ω v C v 4.30
Similarly, we can write equation regarding power during optimum operation as:
,
P P , C v 4.31
,
Combining these equations, we can write following equation for three different wind
speeds:
v ,
1 C v , C v , A
C v , 1 C v , C v , A 4.32
v , A
1 C v , C v ,
180
With respect to zero pitch angle, rated wind speed will give rated rotational speed and
rated power thereby converting equation 4.30 and 4.31 as:
C 4.33
C 4.34
Substituting rated wind speed in 4.33-4.34 and other two wind speed values in
medium wind speed range in 4.32, rotor reference speed can be obtained.
Under maximum Cp of 48%, this study uses a reference speed given by:
ω 0.79131P 1.526046P 0.49188 4.35
under the assumption 0.15 p. u. P 0.75 p. u. For P 0.75 p. u, ω
1.2 and P 0.15, ω 0.689 p. u.
Pitch compensation is also introduced in conventional pitch control to vary pitch
angle during active power control so as to aid desired frequency regulator operation.
Required generation MW capacity is adjusted through the variable Pset. It is initialized
through external active power reference control model to provide governor control if
frequency regulation is objective otherwise it is kept fixed at one p.u. The Pset setting
can be used to curtail maximum output of wind turbine such that blade pitch is set to
produce P < Prated under normal situation.
4.5.3. Electrical Control Model
As discussed in the previous section of the detailed model, type 4 VSWTG has very
fast flux dynamics, and machine side converters are completely decoupled from grid
side converters. Hence the type of generator, as well as machine side converter, does
not significantly influence the total system behavior towards grid. Considering the
requirement of an appropriate WTG model for frequency regulations related analysis
rather than short circuit transient analysis, a simplified static generator model is
181
adopted in proposed model. Reduced order model reduction for the full converter is
presented in [21]. This study suggested usage of either equivalent Thevenin voltage
source model or Norton current source model representing generator and converter
which is subsequently followed by a low-pass filter that outputs reference current ip
to network according to input active current command i [21]. Under normal
operation, the active current is given priority for the availability of maximum
available active power to the grid while during voltage disturbances, the reactive
current has high priority for extra reactive current transfer to improve the grid voltage.
Figure 4-11 Generator-Converter Model [20]
During P-priority, the upper limit of the active current IPmax is dependent upon lumped
converters ratings Imax while the maximum reactive current limit IQmax is determined
based on both the lumped converter rating Imax and the active current command.
182
I I , I I I 4.36
With active current priority in frequency regulation mode, all corresponding flux
dynamics have been avoided to reflect swift power command transfers from the
electrical controller to converters. Power from Generator/Converter model is given
as P V ∗ i and corresponding equivalent framework of generator-converter and
their interface to the network is given by Figure 4-11.
Complete electrical model as proposed by WECC and GE generic model includes
reactive power control along with active power control which is achieved by
active/reactive current injection into the grid as per controller commands. With the
assumption that frequency regulation is valid only when active power is priority and
system have near nominal voltage ratings, reactive control with voltage logic blocks
has been ommitted in proposed model for simplicity. Dark blocks show the electrical
model as used in proposed model. The principle feature of the converter model is
incorporated by imposing the max. current limit of 1.1 p.u.
4.5.3. Active Power Control Loop Co-ordination with Plant Controller
Frequency control tends to be operational during normal grid operation with peak
weightage given to real power transfer through activated real power controller.
Real/active power controller exploits power error signal achieved from the difference
of electrical power and reference power setpoint as provided by plant controller.
Generator speed is also included in the loop to elucidate the shaft power oscillation
effects in generated electrical power. WTG system power order from the PI output is
passed through low pass filter and rate limited. Division of power order through
terminal voltage provides desired active current command signal i by a terminal
voltage which is further rate limited by maximum current obtained from the reactive
183
control loop. A reference power setpoint Pset with value varying from 0.5-1 is used
as plant reference in this model, but plant controller can set this value according to
demanded power from system operator as per respective operation like frequency
control, absolute power, Delta power, etc. In the simplest form, it can also be set as a
minimum of nominal WTG power and optimal aerodynamic power achieved from
maximum power point tracking [22]. Virtual inertia control and active power
frequency control can also be added further in this model. Plant controller supporting
active power control is illustrated in Figure 4-12.
Figure 4-12 Active Power Control
4.6. Model Validation and Analysis
Detailed WTG model requires small time step, and this results in large simulation
time consumption. In the current study, different simulation periods have been used
for the generic model assessment. A set of response simulations are performed to
evaluate & compare the proposed generic model in terms suitability of linear
aerodynamic model with the detailed aerodynamic model, suitability of single mass
drive train model in place of double mass drive train model and suitability of static
generator converter model in place of dynamic generator converter model. All results
are compared with proprietary and detail VSWTG model.
184
4.6.1. Aerodynamic Model Suitability Assessment
Mechanical power harvested from the wind depends upon available wind power,
machine power coefficient and ability to respond wind fluctuations. Various
numerical approximations have been proposed in the literature for determining the
coefficient of performance. Simplified aerodynamic model, as well as detailed model
aerodynamic model based on Cp calculations, is assessed for their utilization in
generic type 4 WTG model. Generic model performance characteristics are analyzed
under three major categories: variable wind speed & variable power setpoint, constant
wind speed & constant power setpoint, variable wind speed & constant power
setpoint to assess the performance regarding rotor speed, electrical power generated,
mechanical power and pitch angle.
4.6.1.1. Simplified Linear Aerodynamic Model
WECC proposal of linear aerodynamic model doesn’t provide any publically
available standard guidelines initializing the model. Initialization of pitch angle and
power in linear aerodynamic model finds scant mentions in current research
literature. According to [7], for less than nominal wind speeds, a linear relation exists
between mechanical power & wind speed while a linear relationship between pitch
angle and wind speed can be assumed for wind speeds above nominal value for
initialization of wind turbine model. Another detailed analysis of the simplified
aerodynamic model is provided in [16]. Linear aerodynamic model is dependent upon
initial mechanical power and changing pitch angle to calculate total mechanical
power delivered to the turbine. With dependence on constant wind speed only,
caution is needed in initializing pitch angle in this model to achieve limited accuracy.
185
In the current study, initial mechanical power is initialized according to available
wind power for wind speed less that rated speed while it is limited to rated mechanical
power for higher wind speed. At constant wind speeds, the active power controller
can maintain generated power at 1.0 p.u. and rotor speed at 1.2 p.u. for constant as
well as varying power setpoint which can be observed in Figure 4-13B - 4-13E. In
Figure 4-14, when the wind speed change from rated value of 13m/s to 8 m/s and then
shooting up to the 18m/s and settling at 25 m/s, pitch angle fails to gain as per set
algorithm rapidly. Slow acting pitch control is not able to follow the dynamic, varying
power setpoint and wind speed. Further simulation investigation reveals that transfer
of pitch angle without consideration of rated wind speed dependence in linear
aerodynamic model results in inadequate performance as compared to detail
aerodynamic model.
Since application of the simplified aerodynamic model is limited to constant wind
speed; dynamic studies under stochastic wind conditions like frequency regulations
call for a generic model which is fully adaptable to varying power setpoints and
varying wind conditions.
A. Varying Power Setpoint
186
B. Electrical Power at Different Wind Speeds Ranging from 8-20 m/s and Varying Power Setpoint
C. Mechanical Power at Different Wind Speeds Ranging from 8-20 m/s and Varying Power Setpoint
D. Pitch Angle at Different Wind Speeds Ranging from 8-20 m/s and Varying Power Setpoint
187
E. Rotor Speed at Different Wind Speeds Ranging from 8-20 m/s
Figure 4-13 Type 4 Generic Model Incorporating Linear Aerodynamic Model Performance
Characteristics at Constant Wind Speed at 8 m/s, 13 m/s, 18 m/s and 20 m/s Along with Varying Power
Setpoint
Figure 4-14 Type 4 Generic Model Incorporating Linear Aerodynamic Model Performance
Characteristics under Varying Wind Speed & Varying Power Setpoint
4.6.1.2. Detailed Aerodynamic Model
There are various numerical approximations proposed in literature for calculating C .
Linear aerodynamic model is then replaced with more detailed C calculations based
model in the second stage. Two different detailed aerodynamic models are tested here
in this study. Aerodynamic Model 1 is based on operation at maximum C calculation
model which is highly dependent on wind speed at the nominal speed and at
maximum C along with proper initial wind speed initialization as shown in Figure
4-15. Aerodynamic model 2 is simple C Calculation model as given in Figure 4-16.

188
Figure 4-15 Coefficient of Power Calculation Model for WTG System
These models are simulated and analyzed with constant wind speeds of 9 m/s, 13 m/s,
18 m/s & 25 m/s as well as varying wind speed from initial 8 m/s to final 20 m/s in
various time steps. Power setpoint used in pitch controller and plant controller is also
varied from 1 p.u. to 0.6 p.u. Both C calculation models give acceptable
performances but due to limitations of aerodynamic model 1 dependence upon base
wind speed and proper initial wind speed initialization put some limitation on its
applicability.
Figure 4-16 Coefficient of Performance Model 2 for WTG System [23]
189
A. Rotor Speed at Different Wind Speeds Ranging from 9-25 m/s and Varying Power Setpoint.
B. Electrical Power at Different Wind Speeds Ranging from 9-25 m/s and Varying Power Setpoint
C. Mechanical Power at Wind Speeds Ranging from 9-25 m/s
190
D. Pitch Angle at Different Wind Speeds Ranging from 9-25 m/s
Figure 4-17 Type 4 Generic Model Incorporating Detailed Aerodynamic Model 1 Performance
Characteristics at Wind Speed of 11 m/s, 13 m/s, 18 m/s and 25 m/s along with Variable Power Setpoint
Figure 4-17 shows various performance results achieved by applying aerodynamic
model 1 with an initial wind speed of 8 m/s. WTG performance results in changes
with a change in initial and base wind speed. It is also observed that this model fails
to provide proper results in mechanical power and corresponding rotor speed and
pitch angle if input wind speed is less than base wind speed and initial wind speed so
proper wind speed initialization is required in adopting aerodynamic model 1.
Figure 4-18 shows mechanical power and pitch angle variation for both aerodynamic
models under constant wind speed and constant power set-point. It can be seen from
these Figures that both aerodynamic models provide similar performance for
mechanical power. A slight variation in speed of pitch angle variation response is
observed with aerodynamic model 2 giving faster response at higher wind speeds.
Type 4 WTG with both aerodynamic models were also simulated at variable wind
speed and variable power set-points as shown in Figure 4-19 – 4-20. The close match
between mechanical power and electrical power is observed when applying
aerodynamic model 1 while the slight oscillatory response is observed in
aerodynamic model 2.
191
A. Mechanical Power Variation for Aerodynamic Model 1
B. Mechanical Power Variation for Aerodynamic Model 2
C. Pitch Angle from Aerodynamic Model 1
192
D. Pitch Angle from Aerodynamic Model 2
Figure 4-18 Type 4 Generic Model Incorporating Detailed Aerodynamic Models at Different Constant
Wind Speed along with Constant 0.8 Power Setpoint
A. Type 4 WTG Characteristics at Variable Wind and Variable Power Setpoint with Aerodynamic
Model 1
B. Type 4 Characteristics at variable wind and variable Power Setpoint [Zoomed]
Figure 4-19 Type 4 Generic Model Incorporating Detailed Aerodynamic Model 1 Performance
Characteristics
193
A. Type 4 Characteristics at Variable Wind and Variable Power Setpoint with Aerodynamic
Model 2
B. Type 4 Characteristics at variable wind and variable Power Setpoint with Aerodynamic
model 2 [Zoomed]
Figure 4-20 Type 4 Model Characteristics with Detailed Aerodynamic Model 2 Performance
Characteristics
Comparing overall performance, detailed aerodynamic model combined with single
mass drive train model show the best combination for dynamic studies like active
power frequency regulation analysis under varying wind and power setpoints. This
result contrasts with the simplified aerodynamic model as proposed by WECC
generic model for dynamic studies of longer time frames.
194
4.6.2. Generator-Converter Model Comparison with GE’s Generic
Model and Detailed Model
Results of proposed model with both types of aerodynamic models are compared with
detailed WTG model and GE’s generic model performance in next stage. It is to be
noted that detailed model includes a synchronous generator with dynamic converter
topology while GE’s model includes braking resistance absorbed power in
calculating reference speed and corresponding power order. Static generator
converter model has been used in proposed type 4 generic model. Generated electrical
power in per unit quantities obtained from proposed model with both types of
aerodynamic models under different wind conditions are comparable with other two
WTG models. It can be seen from Figure 4-21B that 1 p.u. electrical power is
achieved in 10-15 sec for wind speed equal or greater than rated wind speed from GE
WTG model while Figure 4-22A & 4-22 B show that same 1 p.u. electrical power is
achieved within 15 sec. for detailed WTG model due to faster converter dynamics.
Comparable electrical power in p.u. is also achievable from proposed type 4 model
within 15-20 sec. as seen in Figures 4-17, 4-19 & 4-20.
A. Rotor Speed
195
B. Electrical Power at Different Wind Speeds Ranging from 8-20 m/s
C. Pitch Angle
D. Mechanical Power
Figure 4-21 GE’s Type 4 Generic Model Performance Characteristics at Different Wind Speed Ranging
From 8 m/s – 20 m/s
196
A. Type 4 WTG Detailed Model Characteristics at Wind Speed= 13 m/s
B. Type 4 Detailed Model Characteristics at Wind Speed= 18 m/s
Figure 4-22 Type 4 Detailed EMT Model Incorporating Synchronous Generator Characteristics at
Constant Wind Speed of 13 m/s &18 m/s.
It can be established from the simulation results that static generator-converter model
with a detailed aerodynamic model based on pitch control block makes the best
combination to achieve desirable results under the varying wind, and varying power
reference setpoints, hence making complete model valid and suitable for power
frequency regulation analysis.
4.6.3. Drive Train Mass Model Comparison with GE’s Generic Model
& Detailed Simulation Model
Drivetrain comprising rotating parts (Turbine-Generator rotor and gearbox)
constitutes one of the mechanical sub-block of WTG system. Implementation of the
197
detailed mass model is a topic of discussion for WECS modeling. Detailed
mechanical mass model is important for analyzing transient loads while reduced order
model can be applied in bulk network analysis to speed up the simulation. Third stage
analysis was regarding mass model applicability in the generic model. Single mass
model for drive train oscillations has been used both in proposed model as well as in
GE’s model while two mass drive train model has been used in detailed` model.
Figures 4-13E, 4-17A, 4-19 & 4-20 shows a variation of rotor speed under different
wind speeds with satisfactory results in comparison to rotor speeds obtained from GE
model and detailed model. Some oscillations are observed in mechanical power at the
higher cut-off wind speed of 25 m/s when rotor speed goes up to 1.85 p.u.
With converters isolating machines from the grid, one mass model is sufficient to
analyze the performance of type 4 wind turbine system in long time frame dynamic
studies.
4.6.4. Virtual Inertia Support Assessment
Wind turbines provide an artificial inertial response through additional control feature
called virtual inertia as represented in Figure. 3-12, Chapter 3. Virtual inertia controller
introduces an additional power signal to reference power setpoint upon detection of
system frequency deviation thereby momentary increase in wind turbine output power.
Four virtual inertia models are tested for proposed type-4 model under 0.8 power
setpoint with satisfactory results. Three of these inertia models are provided in Figure
3-13 of Chapter 3. Inertia Model 1 incorporates virtual inertia model from GE with a
dead band of 1 which is denoted as inertia model in Figure. 3-13 while basic model
[24] is incorporated into inertia model 2 providing supplementary inertial control
through frequency deviation and rate of change of frequency. This model is referred
198
as inertia droop model in Figure 3-13. Inertial model 3 (referred as droop model 2 in
Figure 3-13) takes frequency deviation as input and has only droop control enabled
while inertial model 4 incorporates only inertial control with the rate of change of
frequency as input. Uncontrolled three area power system model under 10 % load
disturbance is utilized for providing test frequency deviation signal as given in Figure
4-23 for analyzing virtual inertia model responses. A constant wind speed of 13 m/s
and constant power setpoint of 0.8 p.u. is applied in the simulation. It should be noted
that wind turbine model is not feeding electrical power to the network during the
simulation.
Figure 4-24 presents the response of GE type-4 model where a momentary increase in
electrical power and mechanical power is observed with a very small decrease in rotor
speed of around 0.002pu. Results of four virtual inertia models in proposed type 4
model are compared without virtual inertia support and given in Figure 4-25. All
virtual inertia models can provide a higher electrical output during frequency
deviations with highest electrical power change of around 0.08 p.u. obtained from
applying GE wind inertia model. Control features in proposed model can maintain
rotor speed within the limit of 1.2 p.u. after a small speed drop during the frequency
event showing the sufficiency of the proposed type 4 model.
Figure 4-23 Uncontrolled Three Area Interconnected System Frequency During 10% Load Change
199
A. Electrical Power
B. Rotor Speed
C. Mechanical Power at 13 m/s Wind Speed
200
D. Mechanical Power at 13 m/s Wind Speed (Zoomed)
Figure 4-24 GE Model Performance with/without Inertia Support
A. Rotor Speed
B. Rotor speed (Zoomed)
201
C. Electrical Power
D. Electrical Power (Zoomed)
E. Mechanical Power (Zoomed)
202
F. Pitch Angle (Zoomed)
Figure 4-25 Type-4 Model performance with/without Inertia Support
4.7. Conclusion
Most of the frequency regulation research studies to date have been done utilizing an
only generic model of type 3 doubly fed induction generator. A simulation model
modified from the basic structure of WECC’s type-4 wind turbine generator has been
proposed here which can be utilized in wind penetration impact studies as well as
frequency regulation studies. The performance of this generic model is evaluated
through various simulation sets under different wind speeds and power set-points
combinations for rotor speed variations, generated electrical power, developed
mechanical power, pitch angle variation, and virtual inertia support. Simulation
results are also compared with GE’s proprietary generic WTG model and detail
model to reflect the adequacy of proposed model.
Simulation results show that detail aerodynamic model along with single mass drive
train model gives acceptable performance highlighting their adequacy and best
combination in generic models for frequency control studies.
The generator converter effect of designed model is also found to be comparable with
manufacturer specific generic model disabling dependence on propriety information.
203
4.7. References
1. JRC (2014). JRC Wind Status Report, Publications Office of the European Union.
2015
2. AEMO (2013). Wind Turbine Plant Capabilities Report. Available at
aemo.com.au
3. Pouyan Pourbeik (2014). Specification of second generation generic models for
wind turbine generators, EPRI-WECC.
4. UWIG. Type 4 - Generic Wind Turbine Generator Model (Phase II), Available at
www.uwig.org:8080.
5. K. Clark, N. W. Miller, J. J. Sanchez-Gasca (2010). Modeling of GE Wind
Turbine-Generators for Grid Studies, Version 4.5, General Electric International,
Inc.
6. W. W. Price, J. J. Sanchez-Gasca (2006). Simplified wind turbine generator
aerodynamic models for transient stability studies, Power Systems Conference
and Exposition, 2006. PSCE '06. 2006 IEEE PES, vol., no., pp.986, 992.
7. M. Singh, M. Vyas, S. Santoso (2010). Using generic wind turbine models to
compare inertial response of wind turbine technologies, in Power and Energy
Society General Meeting, 2010 IEEE, vol., no., and pp.1-7.
8. Hansen AD, Margaris ID. (2014). Type IV Wind Turbine Model. DTU Wind
Energy.
9. M. Singh and S. Santoso (2011). Dynamic Models for Wind Turbines and Wind
Power Plants.
204
10. IEEE (2009). Tutorial on Wind Generation Modeling and Controls, IEEE PES
PSCE.
11. Bin Wu, Yongqiang Lang, Navid Zargari, Samir Kouro (2011). Power
Conversion and Control of Wind Energy Systems, John Wiley & Sons.
12. R. Pena, J. C. Clare, G. M. Asher (1996). Doubly fed induction generator using
back-to-back PWM converters and its application to variable-speed wind-energy
generation, in Electric Power Applications, IEE Proceedings -, vol.143, no.3, pp.
231-241.
13. Juan W. Dixon and Boon-Teck Ooi (1988). Indirect Current Control of a Unity
Power Factor sinusoidal Current Boost Type Three-phase Rectifier, IEEE
Transactions on Industrial Electronics, Vol. 35, No. 4, pp. 508-515.
14. Marcelo Gustavo Molina and Pedro Enrique Mercado (2011). Modeling and
Control Design of Pitch-Controlled Variable Speed Wind Turbines, Wind
Turbines, Dr. Ibrahim Al-Bahadly (Ed.), InTech.
15. P. C. Krause (1986). Analysis of Electric Machinery, McGraw-Hill, Section 12.5
16. Jens Fortmann (2014). Modeling of Wind Turbines with Doubly Fed Generator
System, Springer.
17. Generic Type-3 Wind Turbine-Generator Model for Grid Studies (2006). WECC
Wind Generator Modeling Group, Version 1.1, Sept. 14.
18. Siegfried Heier (1998). Grid Integration of Wind Energy Conversion Systems,
John Wiley & Sons Ltd, ISBN 0-471-97143-X
19. J. G. Slootweg, S. W. H. de Haan, H. Polinder and W. L. Kling (2003). General
Model for Representing Variable Speed Wind Turbines in Power System
205
Dynamics Simulations, IEEE Transactions on Power Systems, Vol.18, No.1,
pp.144-151.
20. Seman, Slavomir (2011). Need for confidentiality. A converter manufacturer’s
view, Wind Integration Symposium, 16-17 June.
21. Fortmann, Jens et al. (2010). Generic simulation model for DFIG and full size
converter based wind turbines. 9th International Workshop on Large-Scale
Integration of Wind Power into Power Systems as well as on Transmission
Networks for Offshore Wind Power Plants, Quebec City.
22. Spudic, Vedrana et al. (2010). Hierarchical wind farm control for power/load
optimization. The Science of making Torque from Wind (Torque2010).
23. J. G. Slootweg, H. Polinder, W. L. Kling (2001). Dynamic modeling of a wind
turbine with direct drive synchronous generator and back to back voltage source
converter and its control, Proc. of the European Wind Energy Conference, pp.
1014-1017.
24. J. Morren, S. W. H. de Haan, W. L. Kling, J. A. Ferreira (2006). Wind turbines
emulating inertia and supporting primary frequency control, IEEE Transactions
on Power Systems, vol.21, no.1, pp.433-434.
206
CHAPTER 5
DYNAMIC DEAD-BAND DEPENDENT GRID CODE
FREQUENCY SENSITIVE ACTIVE POWER
CONTROLLER
Adding to the study line of active power frequency control in wind turbines which
was presented in Chapter 3, a grid frequency processor scheme based on dynamic
dead-band around moving averaged frequency instead of conventional static dead-
band for generation of continuously varying, frequency sensitive active power
reference setpoint for VSWTG is presented in this chapter. Grid frequency processor
algorithm based on dynamic dead-band is implemented for two different grid codes
to generate active power setpoints which are supplied to a modified speed–pitch
control loop based VSWTG model. Area frequency response along with VSWTG
electrical power support will be compared with other droop based VSWTG model in
next chapter 6 to establish the superiority of proposed frequency-active power
controller based VSWTG over other droop based VSWTG models.
5.1. Introduction
Frequency responsive active power control has advantages in the form of minimized
transmission lines overloading, improved power quality standard, power ramps
limitation can avoid power surges or reduced power unavailability. The conventional
power plants which remain connected to the grid even in the case of frequency
deviations accomplish active power-frequency control through turbine governors
reaction to nominal synchronous speed deviations and the respective boilers-turbine
frequency responsive controls. Dead-band and speed droop are two significant
207
parameters in conventional turbine governor action during system event [1]. Long-
term system frequency regulation reliability is highly dependent upon implemented
dead-band. Governor dead band is detrimental to minimum frequency variation
needed before the governor action is activated. The relative frequency deviation with
respect to the relative change in power output defined as speed droop is always
positive for stable regulation. Factors like the available head room, the maximum-
minimum power capacity of generating units and power setpoint corresponding to
enabled frequency operating mode have a direct effect on total frequency responsive
reserve amount. A generator operating at its maximum generating capacity has almost
negligible headroom and therefore unable to provide any frequency responsive
operation irrespective of governor-droop being enabled.
Some of the above-mentioned parameters are needed to be considered in VSWTG
modeling for frequency-active power control similar to that provided by synchronous
generator. Active power contribution to frequency deviation was earlier only
acknowledged in trivial islanded power systems, but the impending system inertia
reduction with increased integration of power electronics based wind power plants
has led transmission system operators (TSO) to establish new grid codes for
frequency based ancillary services from wind farms in the big interconnected
electrical grid. Given compulsory grid code frequency support, these power
electronics based VSWTG requires modified active power control algorithm to
support regulation services. Frequency-active power control model is an auxiliary
control algorithm implemented in individual WTG control loop for providing
controllable power reserve on demand in the form of spinning reserve or power ramp
rate limit to respond system frequency deviations.
208
The operation of the VSWTGs in maximum power tracking mode results in zero
spinning reserve for utilization in frequency regulation. A brief discussion about wind
turbine participation in primary frequency control through de-loading was presented
in chapter 3. Wind turbines are required to operate at reduced power output to
establish power reserve. As discussed earlier, various methods are present in literature
for the de-loading operation of wind turbines. Separate pitch control or rotor over
speeding de-loaded control are most common approach discussed in previous studies.
Owing to high dependence on available power reserves upon maximum allowable
rotor over-speed, de-loading of wind turbine finds the limited practical application.
Moreover, dependence upon wind speed measurement still needs more scientific
investigations as inaccurate measurement may highly affect wind turbine output and
turbine life.
5.2. Grid Code Sensitive Frequency Controller
5.2.1. Dead-Band: Static vs. Dynamic
Generator droop characteristic means that generating unit will inversely change power
output level in proportion to system frequency. Responding to frequency deviations,
generating units depart from their dispatch targets according to the set droop
characteristics. This droop behavior can be avoided by implementing a dead-band on
each unit. Static dead-band is thus asymmetrical buffer zone on either side of 50/60
Hz frequency that compensates for frequency noise. Droop response of generating
machine from controlling the power output of the machine is avoided when frequency
lies within the upper and lower dead-band range. Dynamic dead-band incorporates a
buffer zone on either side of moving averaged frequency signal instead of the nominal
frequency signal. Both these dead bands are represented in Figure 5.1.
209
5.2.3. Frequency Processing Model
Figure 5-1 NEM Mainland Frequency 1045 hrs. To 1109 hrs. During Contingency Event [1]
Providing grid frequency directly for winding turbine controller will generate noise
induced frequency sensitive active power setpoint. This will result in noisy output
power thus providing frequency response for any frequency error even due to grid
noise. Noisy power output will have an adverse effect on supplied power quality and
turbine life. An analysis of electrical grid dynamics can serve as the base of grid
processing system for distinguishing different types of frequency signals. Figure 5-1
shows the NEM mainland real frequency trace during 2009 contingency event when
3205 MW of total generation disconnected automatically resulting in under frequency
load shedding across the Australian NEM interconnected system [1]. Though we
could not get enough data points to represent a very clear picture here still three
different frequency signals 1, 2 and 3 as defined below can be distinguished in this
Figure. Electrical grid frequency signal can be considered as composed of three types
of signals [2]:
210
1. Low ramping signals arising due to normal trend of generation-demand and
grid dynamics over the long term of several minutes.
2. High amplitude, high-frequency range during contingency which is normally
low in occurrence.
3. Low amplitude signals in high-frequency range due to stochastic grid noise
with high occurrence rate.
Figure 5-2 Basic Structure of Grid Frequency Processor
First two signals have a very low effect on turbine life while highly occurring type 3
noise signals have ample effect on turbine lifetime and should be suppressed. A closer
look at Figure 5-1 shows that usual trend frequency signal 1 normally lies outside
static dead-band and high amplitude signal 2 always lies outside static dead-band.
Discerning these signals just by frequency through filters is not a viable option as
they are present in the whole frequency signal over the time period. Implementing a
dynamic dead band centered around the moving averaged control frequency signal
named as trend frequency signal instead of using a static dead-band in a grid
processing system would distinguish these types of signals. Given the above-
mentioned problem, a signal processing system centered around moving averaged
grid frequency signal was proposed in [2]. A modified version of the frequency
211
processor is implemented in the current study for utilization in active power setpoint
controller for VSWTGs operating under frequency responsive mode.
A grid frequency processor block which provides dynamic dead-band based
processed frequency output is shown in Figure 5-2. Electric grid frequency fluctuates
around the nominal frequency of 50 Hz or 60 Hz. Moving average of this measured
grid frequency can be termed as a trend frequency which characterizes the long-term
behavior of grid frequency. Frequency processor takes measured grid frequency and
nominal frequency as an input signal. The moving average filter (MAF) averages
some points from the measured frequency thereby producing every point in the output
frequency. Mathematically expressed as
yi ∑ yi j , 5.1

Where y is output frequency signal, M is a number of points in average [3].
MAF which can be simulated in various ways like finite impulse response filter
produces lowest noise component in the output signal by equally treating all incoming
signal. A low pass filter with a suitable time constant has been used to act as MAF to
construct trend frequency in this study. Resulting trend frequency signal is subtracted
from the filtered measured frequency and resulting absolute frequency difference is
then passed through dynamic dead-band.
Frequency responsive mode operates by replacing static dead-band and implying
dynamic dead band centered around the control trend frequency signal instead of set
frequency signal. With a suitable threshold value, the output frequency is passed out
as control trend signal if input measured frequency signal follows the control trend
signal within dynamic dead-band otherwise input measured signal is passed out as an
output signal. In this way, noise induced frequency signals are filtered out, and high
amplitude frequency signals, as well as normal frequency signal, can be passed to
212
next block for generating active power set-point as per implemented grid code
algorithm.
An example of frequency which is obtained from NEM 14-bus model and provided
as input signal to frequency processor is shown in Figure. 5-3 while output processed
frequency signal is shown in Figure. 5-4. Frequency disturbances are removed
applying moving averaged dead-band provided by TSO. An example of a measured
frequency and processed out frequency from grid processor for different threshold
values for selector switch is shown in Figure 5-5, and its zoomed version is shown in
Figure 5-6. Output processed frequency is the composition of trend frequency and
measured filtered frequency as seen in Figure 5-7 and Figure 5-8.
Figure 5-3 An Example of Input Frequency Obtained from NEM-14 Generator Model Provided as Input
to Frequency Processor
Figure 5-4 NEM-14 Generator Model Processed Frequency Obtained from Grid Frequency Processor
213
Figure 5-5 Measured Frequency vs. Grid Processor Output Frequency at Different Threshold Values
Figure 5-6 Measured Frequency vs. Grid Processor Output Frequency at Different Threshold Values
(Zoomed)
Figure 5-7 Processed Frequency, Trend Filter Output, Low Pass Freq. Filter Output and Dynamic Dead-
band Output
214
Figure 5-8 Processed Frequency, Trend Filter Output, Low Pass Freq. Filter Output and Dynamic Dead-
band Output When Threshold Applied Is >0.06 (Zoomed)
We can notice output processed frequency in black color following input measured
frequency in red color most of the time when dead-band output is greater than selected
threshold while processed frequency follows trend frequency in green color when it
is less than selected threshold value. This response changes with different threshold
limits. Output frequency follows the measured input frequency most of the time until
input frequency exceeds set static dead-band value.
As stated previously, frequency processor is dependent upon implemented threshold
value and dead-band limits which can be set as per TSO requirements. Upper Dead-
band of 0.015 which is the standard value for most of the TSO [4-5] and a lower dead-
band of 0.011 has been used in this simulation. As per NERC policy, total dead–band
applied should be limited to 0.035 [6].
5.3. Frequency Response Controller Model-I
A schematic representation of a master-slave agent based frequency response
controller model is shown in Figure 5-9. This configuration can be implemented
either at wind turbine level or wind farm level, but author recommends it for large
215
individual frequency responsive wind turbine. Three conditions can be applied for
selecting normal operation. If either wind speed is at a low level or rotor speed is less
than cut-in speed or rotor speed is more than cut-off speed, then wind turbine operates
at normal operating mode insensitive to any frequency variations. If the wind speed
is in medium or high-speed range and rotor speed lies between cut-in and cut-off
speed, the wind turbine can be made to operate at frequency control mode by setting
frequency FLAG.
Figure 5-9 A Schematic Representation of Master-Slave Wind Turbine Controller Incorporating
Frequency Response Controller Model
Though frequency response controller can be implemented at de-loaded power,
frequency response controller is modeled at available power level only in this work.
Frequency response is provided as automatic power variation according to system
216
frequency through a predefined grid code compatible droop. Master control agent
receives droop value and demand power setpoint from TSO and transfers it to slave
control agent. Slave control agent selects the turbine operating mode as per detected
wind speed, or rotor speed explained previously. Earlier studies didn’t consider load
setpoint, but total commanded power in this study is provided as a sum of frequency
response power and demand setpoint. Final power set-point to turbine controller is
limited according to the maximum power rating of the turbine. Figure 5-10 shows
frequency response controller model-I which incorporates basic frequency processor
in its structure to provide frequency response operation. This frequency controller
block can provide two types of a frequency response according to grid code as shown
in Figure 5.11. Working of this frequency response controller is explained below:
Figure 5-10 Frequency Response Controller Model-I
 Frequency Sensitive Response Mode:
By setting FLAG equal to 1, frequency sensitive response mode can be activated.
Power set-point will change proportionally to both up and down frequency deviation
from reference frequency signal. The processed output frequency is generated
through the coordination of the trend control frequency
217
signal and dynamic dead band. If measured frequency follows trend control
frequency signal within dynamic dead-band, trend frequency signal is forwarded for
generating droop power response otherwise measured frequency signal is forwarded.
VSWTG final active power setpoint is generated through an algorithm implemented
on frequency response power and power demand setpoint.
 Frequency Limited Sensitive Mode:
By setting FLAG equal to 0, limited frequency sensitive response mode can be
activated to provide a high-frequency response. There should not any power variation
corresponding to any frequency changes except when the system frequency exceeds
set an upper limit. 50.4 Hz is used in the current simulation. Insensitive mode
incorporates a static dead band around the frequency setpoint. If frequency error
remains within the static band, trend frequency signal and dynamic dead-band have
no influence on output power.
Frequency response operation from this controller can be understood by Figure 5-12
and Figure 5-13. Under frequency sensitive mode, wind frequency response
controller-I provides high active power setpoint in the case of low system frequency
and low active power setpoint in the case of high frequency. In limited frequency
response mode, frequency response controller-I provides low active power setpoint
when system frequency is more than 50.4 Hz for reduced power generation. In the
case of the frequency being lower than 50.4Hz, wind turbine follows the demand
setpoint. Wind availability can also be considered in this controller, but it is not
applied in this study. This type of frequency controller is highly compatible with grid
codes like that of U.K. and Australia which requires a full response from wind farms
218
under normal conditions and limited up/down response under high-frequency
conditions.
Figure 5-11 Frequency Grid Code UK [7]
Figure 5-12 Active Power Setpoint Generated when Employing Frequency Response Controller-I in Two
Different Mode
219
Figure 5-13 System frequency Used to Depict Frequency Response Controller-I Operation in Two
Different Mode
5.4. Frequency Response Controller Model-II
Figure 5-14 Frequency Response Controller Model-II
Model as shown in Figure 5-14 implements the setting of active power set-point as
per typical non-symmetrical droop curve very similar to Irish grid code as shown in
Figure 5-15. Wind speed and processed frequency from basic frequency processor
block are its two inputs. The available wind power in per unit is calculated as a
function of the cube of incoming wind speed:
P ρAV 5.2
220
Where P is the wind power [W], ρ is air density [kg/m3], and v is wind speed
[m/s], and A is the swept area A [m ] of rotor disk that is perpendicular to the wind
flow [8].
Figure 5-15 Non-Symmetrical Droop Curve for Frequency Power Regulation
Active power set-point is primarily constrained by the available wind and different
power commands regarding de-rating power or reserve power from TSO. Aho et al.
have provided three de-rating command mode [9] as defined below:
DR P ,P 1 DR P and DR P 5.3
In another study, three different operating modes (de-rated, absolute reserve and
spinning reserve) [10] are defined such that operating mode is de-rated or normal
where:
P P , if P P 5.4
P P if P P 5.5
These command modes can readily be implemented in proposed model.
A wind power plant (WPP) running as a spinning reserve will produce less power at
all wind speeds thereby always providing a power reserve. Grid code controller will
act as spinning reserve controller or de-rating controller to vary the WPP grid power
production as per transmission system operator request. In the case of wind farm
221
operation, turbine set-point can be multiplied by active power demand provided by
farm controller to generated frequency responsive power demand.
In the current model, filtered available wind power is multiplied with the power set-
point received from algorithms defining respective grid codes for incoming
frequency. Power setpoint is then compared with maximum and minimum load
setpoint restrictions to generate final active power setpoint. Limited set-point is
achieved when additional condition is imposed as IF Frequency Deviation < 0.996
AND Frequency Deviation > 1.004, P EQUALS TO P x 0.95 per unit of
rated power. Implementation of this restriction helps in maintaining maximum output
from a wind turbine in case of limited deviation. Even though wind farm operators
strive for maximum power output, wind-AGC especially downward regulation is
highly dependent upon grid code implementation. In the case of TSO commanded
restriction on active power output, Irish grid code imply all grid integrated generating
units to can operate at a reduced output level. All interconnected grid code compatible
wind farms are required to ensure a fast pitch control as well as other mechanical
controller to be able to participate in a frequency response services in acceptable time.
Figure 5-16 WTG Power Setpoints with Limited Power Setpoint As Grid Controller Output Setpoint and
Corresponding Generated WTG Power
222
Figure 5-16 shows various power set-point generated through grid code compatible
frequency controller-II at variable wind speed and fluctuating grid frequency. The
blue color is the power setpoint required according to the grid code while green is the
available wind power. This available power is multiplied by grid code power and
limited according to desired command mode as explained above thereby providing a
reserve power to be used in frequency deviations. We can notice an increase in
generated electrical power from 1 per unit to 1.1 per unit when the turbine is working
in reserve mode. The Same type of response can be implemented with wind turbine
working in de-loaded mode. In the current study, frequency response controller-I
works independently of wind speed measurement while frequency response
controller-II includes wind speed and corresponding available power in its algorithm.
5.5. Conclusion
This chapter presented a modeling framework for frequency dependent active power
setpoint generation in variable speed wind turbines. An individual wind turbine can
be made grid code frequency compatible by including additional active power
setpoint generator output to the modified torque control loop of the respective turbine.
Active power setpoint generator applies designed power limitation on available wind
power, rated turbine power, and TSO commanded power and provides setpoint to the
turbine. A grid frequency processor based on dynamic dead band and moving
averaged frequency filter is used to suppress noise frequency signals from passing to
the active power setpoint generator. In next chapter, the performance of frequency
controller with and without including wind power in the algorithm will be analyzed
for Type-3 and Type-4 wind turbine response.
223
5.6. References

1. AEMO. (2009). Power system operations event report: Multiple generator
disconnection and under frequency load shedding Thursday 2nd July 2009.
Website: aemo.com.au.
2. Hepner, S. and Wihler, A., (2000). Method for controlling an output of an
electrical power plant. US6118187 A.
3. Smith, S., (1997). The scientist and engineer's guide to digital signal
processing. San Diego, Calif.: California Technical Pub.
4. EIRGRID, (2013). Indication of Dynamic Model Validation Process.
5. EY, (2014). Ancillary service standards and requirements study. Perth:
Available at wa.aemo.com.au.
6. North American Electric Reliability Council, (2004). NERC OPERATING
MANUAL. NERC.
7. Manwell, J. F., McGowan, J. G., and Rogers, A. L., (2009). Wind Energy
Explained: Theory, Design and Application, Wiley.
8. Manwell, J. F., McGowan, J. G., and Rogers, A. L., Wind Energy Explained:
Theory, Design and Application, Wiley, 2009.
9. Janssens NA, Lambin G, Bragard N. (2007). Active power control strategies
of DFIG wind turbines. In: 2007 IEEE power tech Lausanne; p. 516-21.
10. Tarnowski, G.C.; Kjær, P.C.; Dalsgaard, S.; Nyborg, (2010). A. Regulation
and frequency response service capability of modern wind power plants. In
Proc. Power and Energy Society General Meeting, 2010 IEEE, 25-29 July,
Minneapolis, USA.
224
CHAPTER 6
GRID CODE FREQUENCY SENSITIVE VSWTG:
IMPACT ASSESSMENT
Second generation WECC VSWTG models were released in 2014, but their adoption
has not been up to the satisfactory level. With these existing models, plant controllers,
especially for frequency regulation, are still in research and development stage. A
wind turbine active power setpoint controller architecture based on the dynamic dead-
band dependent grid frequency processor was presented in chapter 5. Area frequency
response along with VSWTG electrical power support is investigated and compared
to establish the superiority of proposed grid code frequency-active power controller
based VSWTG model over other droop based VSWTG models in this chapter.
Application of type 3 VSWTG generic model and earlier proposed type 4 VSWTG
model in frequency regulation studies is presented in this chapter. With simplified
generator converter model implemented in investigated generic VSWTG model,
distinguishing electrical response for various droop inertia configuration is achieved.
Consequent electrical output has a remarkable effect on system frequency response
so generic modeling requires detailed investigation when implemented in frequency
regulation studies.
6.1. Introduction
Short-term dynamics have little effect on system frequency stability mechanisms. So
frequency regulation investigation can be satisfactorily achieved with generic
component modeling. The generic models are functional models appropriate for the
investigation with lower simulation period for large-scale power systems. These
225
generic models can sufficiently represent all dynamics associated with the impact of
active power variations in the time range of 50 ms to 100 min. Electromagnetic
transients are neglected in these simulation models.
6.1.1. Type 3 VSWTG Model
Different type 3 VSWTG models [1-3] exists in literature studies, but wind turbine
and wind power plants model still lack standardization and works are in progress.
Various software like PSLF, PSS-E, DlgSILENT, etc. have been used for
implementing these models. An official version of second generation WECC models
was released in 2014 but their adoption has not been up to the desired level as not
every type of VSWTG model can be applied in every study [4]. Long term system
dynamics requires simplified generic modeling with fewer parameters as detailed
WTG modeling with low simulation time step takes much longer simulation time.
Even though WTG model characterizations development has been fundamentally
accomplished, however, development of plant controllers especially regarding
frequency response control is still under developing stage.
Block diagram in Figure 6-1 shows the basic components of generic VSWTG model.
A basic schematic diagram implementing these basic modules for type 3 (Doubly fed
asynchronous generator) based WTG control structure is presented in Figure 6-2. The
model has three basic blocks: turbine model, Generator/Converter model, and
Electrical control model. Type 3 electrical control model is represented by active
current command only with all flux dynamics neglected. During frequency response,
real power has priority, hence detailed reactive power loop in not applied in the
current study. The aerodynamic model given in chapter 3 has been applied for type 3
226
VSWTG model while single mass drive train is utilized for drive train modelling in
this study.
Figure 6-1 Basic Components of Generic VSWTG Model
Turbine control system involves two control loops having speed error as input and
two control outputs: 1. Wind turbine reference power provided to the converter
control. 2. Pitch reference value. Rotor speed is controlled at optimum level as per
reference power command through torque limitation in speed control loop while
mechanical power and a corresponding shaft speed of wind turbine are controlled
through pitch controller and pitch compensation loop when rotor speed exceeds
optimum level. It should be noted that frequency responsive power setpoint is
subtracted from commanded power setpoint to achieve required power order to
generator/converter model while a fixed rated power command is applied to pitch
compensator. Pitch control enables optimum control of aerodynamic wind power by
rotor blade pitching to regulate turbine torque. Maximum power of 1.2 per unit is
available at zero pitch angle while it is highly reduced with highest pitch angle.
Similar to proposed type 4 VSWTG model in chapter 4, the pitch angle is obtained
227
through a series of the PI regulator is mathematically expressed by following
equations:
.
θ K ω ω K P P 6.1
Where angular reference speed is approximated as:
ω 0.79131P 1.526046P 0.49188 6.2
Under the assumption 0.15 p. u. P 0.75 p. u. For P
0.75 p. u. , ω 1.2 and P 0.15, ω 0.689 p. u.
Figure 6-2 Generic Type 3 VSWTG Basic Model with Auxiliary Virtual Inertia Model
Most of the past research studies have considered the basic type 3 model as shown in
Figure 6-2 and basic control loop for frequency response studies shown as an inset in
Figure 6-3. Current study applies a modified wind turbine control loop adopted from
[5] which is shown in red dotted line. Pitch compensation block takes frequency
228
responsive active power set-point PAPC generated from frequency responsive
controller instead of rated 1 p.u. reference power. Power order from torque controller
is altered by passing active power set-point PAPC through the power response rate
limiting block.
Figure 6-3 Type 3 WTG Model with Active Power Controller Feeding Power Setpoint PAPC to Power
Controller Loop, Pitch Controller Loop and Speed Controller Loop
Rotor speed error is given as input to both these controls. Final power order provided
to generator converter model is generated by adding limited active power setpoint
229
from frequency response controller block to difference between active power limited
setpoint power and power order from speed controller.
P P P P 6.3

The ramp rate is implemented by including a washout filter whose time constant is
detrimental to rate limit imposed on changes in power order. Power order provided
to converter control can also be transformed by inertia controller to exploit the
rotating masses stored energy along with frequency responsive active power
command.
6.1.2. Applied Commanded Reserve Power Mode
In reserve power mode as applied to this study, wind turbine regulates electrical
power according to frequency responsive final power order. Rotor speed and pitch
based de-loading methods were discussed in chapter 3. Prior de-loading of turbine
through C table inversion [6], over speed and under speed operation has been
suggested in many research studies [7-10] for enabling wind turbine frequency
responsive support. A combined torque-pitch control method utilizing PAPC is applied
in this study to obtain reserve power mode operation. Optimum speed is maintained
corresponding to active power set-point through the torque-speed controller. Pitch
controller initiates blade pitching in case speed increases and reduces mechanical
power thereby reducing speed. All interconnected grid code compatible wind farms
are required to ensure a fast pitch control as well as other mechanical controller to be
able to participate in a frequency response services in acceptable time.
In MPPT operating mode, turbine power set-point is determined such that:
.
P K ω , where K ,β 6.4
230
Let us assume wind turbine to be operating at MPPT mode initially such that P
P and W W and pitch angle β β 0 . When turbine switches from
MPPT mode to commanded power mode, MPPT power setpoint P changes
to P P ; VSWTG power drops below mechanical power thereby
increasing rotor speed to W . This increase in rotor speed is controlled through
increasing pitch angle β to β provided through pitch controller in coordination with
pitch compensator. Consequently, power-speed characteristic of wind turbine is
varied which is shown in red in Figure 6-4 and rotor speed settles at ω . To
guarantee a safe switching between MPPT mode and commanded mode and
safeguard turbine operation beyond rated values, final power order is limited to the
optimum power value (P K ω . The turbine will continue to run in MPPT
mode unless there is a change in wind speed or change in active power setpoint from
frequency controller.
Figure 6-4 Characteristics Curve Of Type 3 VSWTG under MPPT and Reserve Power Mode
231
It should be noted that active power setpoint from frequency controller is already
limited as per desired reserve power of 10%. To explain reserve power mode
operation, assume steady state condition represented by point a for blue curve at some
wind speed such that turbine mechanical power equal to VSWTG electrical
powerP P . In the case of frequency deviation due to generation loss,
frequency responsive VSWTG’s power will increase to compensate for this loss and
may reach VSWTG rated capacity limit shown by point b in Figure 6-4. Power
extraction due to kinetic energy will continue until rotor speed hits the minimum limit
at point d. After this point, mechanical power will be more than electrical power and
rotor will again speed up to finally settle at point e with speed W .
6.1.3. Wind Profile and Available Power Model
Figure 6-5 Wind and Available Power Simulation Model
A wind profile and corresponding available power simulation model are presented in
Figure 6-5. This wind model is used to input wind speed shown in Figure 6-6 to both
frequency controller presented in chapter 5 while available power is also applied to
232
frequency controller-II where active power set-point was dependent upon available
wind power. Wind speed very similar to Avalon wind speed data obtained from [11]
was applied to a lookup table, and final wind speed was limited by a maximum wind
speed and nominal speed through SR flip-flop logic. Electrical power is produced as
the cube of available wind, and nominal power is produced at nominal wind speed.
Figure 6-6 Generated Wind Speed and Available Power
6.2. Simulation Setup
A single area controlled model incorporating hydropower plant is used to test the
frequency responsive VSWTGs. Hydro turbine governor is modeled according to
Figure 6-7 and discussed further in Chapter 7. On System base value of 5000 MW
and assuming 0.05 p.u. extra support from VSWTG during frequency excursions,
wind-integrated system modified inertia constant, H is calculated as [12]:
233
H H 1 L H L 6.5
∑ ∗
H , where S ∑ S 6.6
Where H and S are the inertia rating and apparent power rating of individual
generating units. L is wind penetration level. For frequency responsive wind plant,
wind inertia contribution H to power system by providing ∆P extra active power
when the system is subjected to step load disturbance ∆P is calculated as:
H ; 6.7
∆ ∆
X e 1 6.8
∆ ∆
Figure 6-7 Single Area Hydro Governor Turbine Model with Integrated Wind Farm
Total time delay T associated with hydro governor turbine delay model is calculated
by delay theory originally given in [13]. Summation of governor time constant, valve
motion delay time and turbine response delay time results in total time delay at which
minimum frequency deviation occurs after system disturbance Value of T used in
the simulation is calculated as 3.7814 seconds. To derive dynamic response for the
uncontrolled system, let us assume hydro turbine governor time constant equal to zero
in Figure 6.7, we can write following equation 6.9 - 6.10.
234
∆
∆F s |∆ ∗ 6.9
,
∆f t ,
1 exp t ,
∆P 6.10
, T ,
Where K is defined earlier as . The minimum frequency occurring at T after load
disturbance can be calculated as equation 6.11, where the hydro turbine response
delayed by T covers the combined load disturbance in the presence of wind energy.
T
∆f t K 1 exp T ∆P 6.11

In case of integral controlled system with ∆P s , frequency
,
deviation can be expressed according to equation 6.12 as:
∆
∆F s 6.12
∗
,
.
,
6.2.1. Simulation Assumptions
System simulations are performed and compared with different frequency responsive
type 3 and type 4 wind plant model. A comparative study is made of frequency
response and its indicators, the effect on VSWTG electrical support and
corresponding rotor speed and hydraulic mechanical power change. Following
assumptions are used in simulation:
1. A constant load disturbance is applied.
2. All VSWTG models are provided same wind speed as shown in Figure 6.6.
3. Frequency response controller-I (Figure 5.8) and frequency response
controller-II (Figure 5.12) is implemented for VSWTG model as given in
chapter 5. Parameters are given in the appendix.
235
4. Frequency droop model 1, droop model 2 and inertia droop model as given in
Figure 3.13 of Chapter 3 are incorporated in basic torque control loop based
VSWTG model [12] which is shown as an inset in Figure 6.3.
5. Droop setting used in simulations are given in Table 6.1.
6. Equivalent droop setting of area 1 changes with wind penetration as R
. A 10% wind penetration is used in the simulation.
System parameters used in the simulation are given in the appendix. Results obtained
were assessed for effect of frequency responsive VSWTG model integration on:
 Area frequency response (Settling time, frequency nadir and rate of change of
frequency).
 VSWTG electrical output and VSWTG rotor speed variation during and after
frequency support.
Table 6-1 Comparison of Simulation Results to Show the Effect of Torque Controller Re-Tuning and
Different Droop Values for 10% Wind Energy Penetration
Controlled Uncontrolled Control Uncontrol

Torque
Area Area Area Area
Controller
Models Settles
Gain Settling Settling
(Hz) at
( PI ) Time (s) Time (s)
. .
Case A Torque Controller Re-tuned in basic droop-inertia models Only
Droop/Inertia (-271.55,
0.035 0.0315 49.96 49.981 0.4 30 30
Models -310.70)
Freq. Cont.-I 0.035 0.0315 (3, 0.6) 49.98 49.988 1.01 50 25
Freq. Cont.-II 0.035 Var. (3, 0.6) 49.99 49.994 1.1 50 30
Case B No tuning of Torque Controller in basic droop-inertia models.
Droop/Inertia
0.035 0.0315 (3, 0.6) 49.982 49.986 0.04 40 35
Models
Freq. Cont.-I 0.035 0.0315 (3, 0.6) 49.987 49.987 1.01 40 25
Freq. Cont.-II 0.035 Var. (3, 0.6) 49.991 49.995 1.1 40 30
236
Droop values changed for Hydro and Wind power plants; No tuning
Case C
applied in torque controller
Droop/Inertia
0.0556 0.05 (3, 0.6) 49.979 50.005 0.04 40 30
Models
Freq. Cont.-I 0.0556 0.05 (3, 0.6) 49.990 49.993 1.01 20 10
Freq. Cont.-II 0.0556 Var. (3, 0.6) 49.984 49.994 1.1 30 25
Droop values changed for Hydro and Wind power plants; No Tuning
Case D
Applied in Torque Controller
Droop/Inertia
0.035 0.0315 (3, 0.6) 49.982 49.995 0.04 40 20
Models
Freq. Cont.-I 0.0556 0.0315 (3, 0.6) 49.991 49.993 1.01 20 10
Freq. Cont.-II 0.0556 Var. (3, 0.6) 49.984 49.994 1.1 30 25
Case E Different Droop Values for Hydro and Wind Power Plants
Droop/Inertia
0.0556 0.0315 (3, 0.6) 49.979 50.005 0.04 40 20
Models
Freq. Cont.-I 0.0556 0.0315 (3, 0.6) 49.990 49.993 1.01 20 10
Freq. Cont.-II 0.0556 Var. (3, 0.6) 49.984 49.994 1.1 30 25
6.3. Discussion
6.3.1. Case I: Type 3 VSWTG
Summary of 5 test cases is provided below in Table 6-1 in which proposed frequency
controller based type 3 VSWTG models are compared with other droop inertia based
type 3 VSWTG model. Important points from these simulation results are discussed
in below sections.
 Effect on Control Area Frequency Response
Uncontrolled and controlled control area were simulated with different droop values.
Effect of torque re-tuning is also conducted to highlight the difference in generated
power and corresponding frequency response. Figure 6-8 gives the frequency
deviation obtained from single area uncontrolled model with 0.1 per unit load
disturbance. An example simulation result of area frequency response for controlled-
237
single area LFC model with integrated type-3 WTG is shown in Figure 6-9 where
basic type 3 VSWTG model’s torque controller is not retuned.
Figure 6-8 Uncontrolled Single Area Model Frequency Deviation for 0.1 P.U. Load Disturbance
Figure 6-9 Frequency Response of Single Area LFC Control Model for 0.1 P.U. Load Disturbance,
Req=0.035 for All Cases, Torque Controller Gains (3, 0.6), No Re-Tuning For Torque Controller
Similarly, frequency response for a control system with re-tuned torque controller is
shown in Figure 6-10 and 6-11. Effect of this re-tuning will be explained in section.
From these Figures, a clear improvement in area frequency deviation can be observed
for both controlled and uncontrolled system with proposed frequency responsive
VSWTG model integration. Respective minimum frequency and settling time for
these simulation test are provided in Table 6-1. Except for re-tuned model, improved
settling frequency in the range of 20s-30s is observed for proposed frequency
responsive VSWTG model. With changed in droop value from 0.0315 to 0.05 in all
238
models, an improvement in settling frequency is observed for both uncontrolled and
controlled system. Comparing case C and case E, the effect of system droop change
is more visible over wind droop value. With constant system droop and changed wind
energy droop from 0.05 to 0.0315, the same settling time is observed. With other
three droop controlled models, more overshoots and undershoots are observed a with
a slight increase in settling time. Though the inclusion of wind power in governing
active power setpoint has a role in system frequency response bringing more
oscillations in terms of undershoots and overshoots, yet superiority of grid code
compatible frequency response controllers is established through simulation results
regarding reduced settling time, improved ROCOF and frequency nadir point.
Figure 6-10 Frequency Response of Single Area LFC Control Model for 0.1 p.u. Load Disturbance and
Req =0.035 in All Cases, Torque Controller Retuned for Other Three Droop Based VSWTG Models

Figure 6-11 Frequency Response of Single Area LFC Control Model for 0.1 p.u. load Disturbance,
Req =0.0556 in Case of Frequency Controller-I and II, Torque Controller Retuned
239
 Effect on Rate of Change of Frequency and Frequency Nadir
The management of ROCOF is critical to maintaining power system frequency within
the frequency operating standard and to maintaining the power system in a secure
operating state. We discussed system inertia and the corresponding effect on ROCOF
in Chapter 2. Lower inertia leads to a higher ROCOF and power electronics based
wind turbines decoupled from grid do not contribute to any system inertia without
additional modifications in control loops. Even TSOs have no control over minimum
system inertia, but ROCOF has to be within certain limits as per grid code. Currently
installed wind turbines have maximum acceptance ROCOF within the range of 2–4
Hz/s for 200–300 milliseconds and 1 Hz/s is accepted in a long range of around 1
second [14].
Figure 6-12 Rate of Change of Frequency Comparison for Different Load Change (p.u.) and 10%
VSWTG Penetration.
In another comparative simulation study, the rate of change of frequency (ROCOF)
obtained under different load disturbance scenario, and 10% wind penetration is
given in Figure 6-12. Though the values obtained can not be held indicative of real
time scenario which has interaction between components of varying electrical
characteristics, a clear improvement is observed in ROCOF values for control area
when proposed frequency based controller based Type 3 VSWTG is integrated with
240
the hydro power plant. A more detailed system will be required to analyze actual
ROCOF changes.
Figure 6-13 Frequency Nadir Comparison for Different Load Change (p.u.) and 10% VSWTG
Penetration
Frequency nadir after any contingency is detrimental to primary frequency regulation
for maintaining system stability. The primary frequency reserve adequacy criterion
can be expressed as [15]:
Freq ∆ P , Primary Frequency Reserve, F Freq , 6.18
where ∆P is the maximum power loss during contingency, F (Hz) is the
maximum value for governor’s dead band. Freq is the frequency nadir after the
loss of ∆ P , and Freq denotes the minimum frequency required as per
operating standard. Wind power plant integration and its participation in frequency
regulation services can be analyzed through frequency nadir points achieved under
different combinations. Frequency nadir obtained for different frequency responsive
VSWTG models integrated into control area is shown in Figure 6-13. Frequency nadir
point as achieved under different load disturbance scenario indicates the superiority
of grid code compatible frequency controllers over normal droop based VSWTG
model.
241
 Effect on VSWTG Electrical Output
An important point to notice about other three droop controlled model is regarding
wind plant power output during and after frequency response in LFC model. With
simplified generator converter model implemented in currently investigated generic
VSWTG model, we observed distinguishing electrical response for various droop
inertia configuration implemented. Consequent electrical output has a remarkable
effect on system frequency response so care should be taken while implementing
generic models.
Figure 6-14 Electrical power from Type 3 VSWTG with Retuned Torque Controller in Control Area.
During Initial Frequency Response for 0.1 p.u. load Disturbance, Req= 0.0556 for Frequency Controller-I
and II
Figure 6-15 Initial Frequency Response for 0.1 p.u. Load Disturbance
242
Figure 6-14 shows the zoomed version of electrical power output from VSWTG with
different droop controlled models with retuned torque controller during the initial
frequency deviation in control area. Zoomed version of control area frequency is
shown in Figure 6-15. Even though a variable wind speed is applied in simulation, a
constant wind speed of 11.3 m/s is observed during approximate 40 seconds of
frequency deviation. In the case of a 0.1 p.u. load disturbance and 0.05 p.u. active
power support from VSWTG, control area sees this frequency support from a
machine with 3.4 s inertia constant. This inertia constant will change to 7.17 s if 0.1
p.u. of extra support from VSWTG is assumed. VSWTG will replace other
conventional generation (having equivalent inertia 6 s) if they provide an active
power support of approximately 0.8 p.u. during frequency excursions.
It can be observed that even with same load disturbance and same system inertia
constant for all control area, the different frequency response is obtained. Due to
anomalies in applying basic VSWTG model and auxiliary droop models, we notice a
slow response in generated electrical power. Electrical power from basic VSWTG
model with retuned torque controller finally settles at 0.4 p.u. as soon as frequency
improves otherwise power drops to 0.04 p.u. All these three droop based models
attempt to provide increased electrical power up to 1.2 p.u. soon after detecting the
frequency drop but fails to maintain net power output after frequency disturbance
correction. Drop in electrical power output during and after frequency response also
earlier confirmed in [16] and requires torque controller parameter retuning for
increased electrical power after temporary frequency deviations. Simulation tests
were conducted with retuned torque controller as given in Table 6-1 but still proposed
frequency controller based VSWTG outperforms other models electrical output. A
243
clear improvement in electrical power temporary support from WTG model
incorporating frequency controller-I and II can be observed from above Figure 6-14.
Figure 6-16 Generated Active Power Setpoint for Frequency Controller-II in the case of 0.1 p.u. Load
Disturbance in Control Area
Figure 6-17 Generated Active Power Setpoint for Frequency Controller-I in the case of 0.1 p.u. Load
Disturbance in Control Area
Following the active power setpoint generation algorithm, a constant active power
setpoint up to 1.01 p.u. is provided by frequency controller-I to ramp up the power
during the moments of frequency deviations. VSWTG model as used in this study
has 10% spinning reserve so a maximum electrical power of 1.1 p.u. is available
during frequency disturbance. Frequency controller-II also provides an increasing
power reference setpoint of around 0.85 p.u. during the initial 20 seconds of
frequency deviation and changes according to grid requirements and available power.
244
The active power setpoint provided by frequency response controller-I and II to the
VSWTG is shown in Figure 6-16 - Figure 6-17. Effect of these reference setpoint is
reflected in generated electrical power. WTG model with frequency controller-I can
maintain electrical power around 1.1 p.u. during the moments of initial frequency
deviations while WTG with frequency controller 2 has an initial power drop up to 0.3
p.u. and then starts increasing and settles at around 1.1 p.u.
 Effect on VSWTG Rotor Speed
A representative rotor speed variation is shown in Figure 6-18. The change in rotor
speed for type 3 VSWTG model during frequency response in control area has been
same for all test cases. A minimum limit of 0.75 p.u. and a maximum limit of 1.2 p.u.
has been imposed in currently investigated frequency responsive VSWTG models. It
can be observed from Figure 6-19 that all VSWTG models show reduced rotor speed
maintained approximately 0.87 p.u. during those initial moments of frequency
deviations till 20 seconds when extra electrical power from VSWTG is expected.
Some fluctuations in rotor speed can be seen after 5 seconds when frequency drop is
at peak position, and VSWTG control loop tries to increase power. Rotor speed from
basic type 3 VSWTG with other droop models finally settles at around 0.95 p.u. with
consequent reduced electrical output power to 0.4 p.u. If proper speed limits are not
imposed in droop model 1, rapid variations in rotor speed are observed till 250
seconds before settling smoothly around 0.95 p.u. In contrast to these basic VSWTG
model, more adaptive frequency responsive operation is seen in type 3 VSWTG with
frequency controller-I and II where we can notice a change in electrical power with
a corresponding change in rotor speed during the moment of another frequency
deviation at around 235 seconds. Drop in rotor speed is observed due to an imbalance
in mechanical and electromechanical torque when electrical power is increasing for
245
frequency controller-I and II, but as soon as frequency deviation settles, an increase
in rotor speed is also observed which settles at around 1.2 p.u.
Figure 6-18 VSWTG Rotor Speed Variations with Different Frequency Droop Models during Frequency
Response

Figure 6-19 VSWTG Rotor Speed Variations with Different Frequency Control Loops [Zoomed]
6.3.2. Case II: Type 4 VSWTG
Various generic models of full converter based type 4 VSWTG models were
presented in Chapter 4. The inertial response of proprietary model and proposed type
4 model were presented under constant wind speed and constant power setpoint of
0.8 p.u. Both the models were shown to provide inertial support for frequency
deviation which was provided to models through an auxiliary signal. No wind power
participation in supplied power was considered. In this section, simulation results of
earlier proposed type 4 VSWTG model and proprietary model are presented when
246
working in frequency responsive mode. Different inertia and droop models presented
earlier in chapter 3 are integrated with these models. Results are compared when grid
code sensitive active power setpoint from proposed frequency controller in Chapter
5 is provided with these models. The aim of these simulation results is to show the
improvement of frequency response and corresponding generated power supported
by rotor speed when working under grid code sensitive power setpoint.
Figure 6-20 Frequency Response of Single Area LFC Control Model Integrated with Proprietary Type 4
VSWTG Model for 0.1 p.u. Load Disturbance, Req = 0.035, No Retuning Applied in Torque Controller
Figure 6-21 Generated Electrical Power from Proprietary Type 4 VSWTG Model during Initial
Frequency Response for 0.1 p.u. Load Disturbance, No Retuning Applied in Torque Controller
Figure 6-20 presents the frequency response obtained for single control area when
model similar to proprietary type 4 VSWTG model shown in Figure 4-6 is integrated
into frequency responsive mode. No tuning of torque controller is applied in these
247
models. Recalling that frequency controller-II works according to grid code similar
to Irish code and frequency controller-I works according to British grid code. Wind
available power is considered in frequency controller-II while the frequency
controller-I is independent of available power. It should be noted again that during
the initial frequency deviation of around 40 s, wind speed is assumed constant. An
improved frequency response along with corresponding electrical power variation
and rotor speed can be observed in Figure 6-18 – Figure 6.20 when proposed
frequency controller based grid code sensitive power set-point is provided to it in
comparison with other droop inertia models.
Figure 6-22 Proprietary type 4 VSWTG Model Rotor Speed Variations with Different Frequency Droop
Models during Frequency Response
In similarity with type 3 VSWTG droop based models, a drop in electrical power to
0.04 p.u. is observed during and after frequency deviation for other three droop inertia
model as shown in Figure 6-21. Type 4 rotor speed also increases sharply above rated
value during frequency deviation and fails to recover back as seen in Figure 6-22. It
is to be noted that power setpoint from these auxiliary models was added to power
order from torque controller loop in the similar manner as shown in the inset of Figure
6-3 while a rated power of 1 p.u. is provided to pitch compensator. Increased power
support from proprietary type 4 model requires torque retuning which is difficult
248
during real time operation. Active power setpoint from proposed frequency
controllers was added in the modified loop similar to shown in Figure 4-6 of Chapter
4. Improved variation of electrical power during and after frequency deviation can be
observed from Figure 6-21. Even rotor speed remains under rated limit after the
frequency responsive operation. These simulation results confirm that by proper
modeling of power control loop, improved power response can be obtained during
frequency disturbances.
Impact assessment of previously proposed type 4 VSWTG model in chapter 4 is
represented in upcoming simulation results where all three droop based auxiliary
model and proposed frequency controller-I and II are integrated.
Under same droop settings and all similar system parameter, frequency response as
shown in Figure 6-23 is obtained for proposed type 4 VSWTG model. Minimum
nadir point is achieved with frequency controller-II while frequency response from
frequency controller-I in both modes is nearly similar to droop model 1 as represented
in zoomed version in Figure 6-24. It should be noted that other droop models are not
grid code compatible while an upper dead-band limit of 0.0315 and lower dead-band
limit of 0.001 has been imposed on frequency controllers to provide active power
setpoint. Improvement in system frequency response in case of frequency controller-
II can be matched with corresponding active power setpoint as shown in Figure 6-27
and respective electrical power shown in Figure 6-25. Electrical power output as
shown in Figure 6-25 is also found to be varying according to another active power
setpoint from frequency controller-I which is shown in Figure 6-26.
249
Figure 6-23 Frequency Response of Single Area LFC Control Model Integrated with Proposed Type 4
VSWTG Model for 0.1 p.u. Load Disturbance, Req=0.035
Figure 6-24 Frequency Response of Single Area LFC Control Model Integrated with Proposed Type 4
VSWTG Model for 0.1 p.u. Load Disturbance [Zoomed]
Figure 6-25 Generated Electrical Power from Proprietary Type 4 VSWTG Model during Initial
Frequency Response for 0.1 p.u. Load Disturbance, Req=0.035
250
Figure 6-26 Grid code Sensitive Active Power Setpoint from Frequency Controller-I during Initial
Frequency Response for 0.1 p.u. Load Disturbance
Figure 6-27 Grid Code Sensitive Active Power Setpoint from Frequency Controller-II during Initial
Frequency Response for 0.1 p.u. Load Disturbance
Figure 6-28 Generated Rotor Speed from Proposed Type 4 VSWTG Model during Initial Frequency
Response for 0.1 p.u. Load Disturbance, Req=0.035
251
Table 6-2 Comparison of Frequency Response Simulation Results for Type 4 Models Under 10% Wind
Energy Penetration in a Controlled Area
, ,
, (Hz) Settling Time (s)
(p.u Hz / p.u. MW) (p.u. Hz / p.u. MW)
Case A Proprietary Type 4 VSWTG model
Droop/Inertia models 0.035 0.0315 49.982 45
Freq. Cont.-I 0.035 0.0315 49.987 40
Freq. Cont.-II 0.035 Var 49.991 40
Freq. Cont.-I 0.0556 0.05 49.991 20
Freq. Cont.-II 0.0556 Var. 49.985 30
Case B Proposed Type 4 VSWTG model
Freq. Cont.-I 0.035 0.0315 49.977 80
Freq. Cont.-II 0.035 Var. 49.973 80
Freq. Cont.-I 0.0556 0.05 49.972 35
Freq. Cont.-II 0.0556 Var. 49.977 45
A comparison of both type 4 model is given in Table 6-2. Even though frequency
nadir point is slightly less than the proprietary model and more settling time more,
proposed type 4 model presents improved electrical power and rotor speed. Except
for droop model 2, proposed type 4 VSWTG can provide improved power response
with other inertia droop model though not according to grid code. Proposed type 4
VSWTG model can provide rotor speed within rated limit during and after frequency
disturbance as shown in Figure 6.28 while it overshoots to high speed for droop model
2 where we see a drop in electrical power. Increased in system droop value is found
effective in reducing the settling time. Basic Type 3 and Type 4 VSWTG models call
for torque controller gains re-tuning for improved participation in frequency response
while this can be avoided with frequency response controllers. These comparative
values are obtained for 10% wind penetration at 5000 MW system base value. The
252
more effective comparison requires higher wind penetration in the system, but overall
objective of this study was to show the improvement in frequency response and
electrical output under proposed frequency controllers and modified torque speed
loop of VSWTG.
6.4. Conclusion
An individual wind turbine can be made grid code frequency compatible by including
additional active power setpoint generator output to the modified torque control loop
of the respective turbine. Active power setpoint generator applies designed power
limitation on available wind power, rated turbine power, and TSO commanded power
and provides setpoint to the turbine. A grid frequency processor based on dynamic
dead band and moving averaged frequency filter is used to suppress noise frequency
signals from passing to active power setpoint generator. The simulation was carried
out for a single area control model with 10% wind penetration. Control area
frequency response, as well as integrated VSWTG's electrical and rotor speed
responses, were compared with other frequency droop based VSWTG model
responses. Following conclusions can be based on these simulations.
 Promising results regarding improved settling time, better electrical power
and rotor speed variations during frequency deviations were obtained for
proposed frequency responsive grid code compatible VSWTG model in
comparison to other common droop-inertia based model.
 Basic VSWTG model requires retuning of torque controller to provide
electrical power output of 0.4 p.u. soon after frequency disurbances while
electrical power drops to 0.04 p.u during frequency response. Proposed
frequency responsive grid code compatible VSWTG model does not requre
253
any tuning of torque controller and provides electrical power as per power
setpoint for frequency response and rotor speed being under rated limit.
 Improved system frequency is obtained when grid code frequency responsive
type 3 VSWTG is integrated in comparison to type 4 grid code frequency
responsive VSWTG.
 With suitable modeling, full converter based type 4 VSWTG can participate
in frequency response. A simulation test at 10% load disturbance with type 4
VSWTG is integrated to the system provided approximately same frequency
nadir but slightly increased settling time and increased ROCOF when
compared with type 3 VSWTG integrated system response.
 The effect of system droop change is more visible over wind droop value in
terms of frequency deviation.
 Frequency response for area 1 was determined using classical linear hydraulic
turbine model. It would be interesting to investigate the area frequency
response with advanced inelastic non-linear hydraulic turbine model
interaction with frequency responsive wind power plants.
254
6.4. References
1. Jonkman, J., Butterfield, S., Musial, W., and Scott, G. (2009). Definition of a 5-
MW Reference Wind Turbine for Offshore System Development, NREL/TP-
500-38060, Golden, CO: National Renewable Energy Laboratory.
2. Pourbeik, P.; Ellis, A.; Sanchez-Gasca, J.; Kazachkov, Y.; Muljadi, E.; Senthil,
J.; Davies, D. (2013). Generic stability models for Type 3 & 4 wind turbine
generators for WECC, Power and Energy Society General Meeting (PES), 2013
IEEE, pp.1, 5, and 21-2.
3. B. Sørensen, B. Andersen, J. Fortmann, K. Johansen, and P. Pourbeik (2011).
Overview, status, and outline of the new IEC 61400 -27 – electrical simulation
models for wind power generation, International Workshop on Large-Scale
Integration of Wind Power into Power Systems as well as on Transmission
Networks for Offshore Wind Power Farms.
4. WECC (2010). WECC Wind Power Plant Dynamic Modeling Guide, WECC
Renewable Energy Modeling Task Force. Available at renew-ne.org.
5. K. Clark, N. W. Miller, J. J. Sanchez-Gasca (2010). Modelling of GE Wind
Turbine-Generators for Grid Studies, Version 4.5. General Electric International,
Inc.
6. P. Venne and X. Guillaud (2009). Impact of wind turbine controller strategy on
deloaded operation. Integration of Wide-Scale Renewable Resources In to the
Power Delivery System, 2009 CIGRE/IEEE PES Joint Symposium, Calgary,
AB, pp. 1-1.
7. N.A. Janssens, G. Lambin, N. Bragard (2007). Active power control strategies
of DFIG wind turbines. IEEE power tech Lausanne, pp. 516–521.
255
8. Ramtharan G, Ekanayake JB, Jenkins N. (2007). Support for spinning reserve
from DFIG based wind turbines. In: 2007 international conference on industrial
and information systems. p. 111–116.
9. F.D. Kanellos, N.D. Hatziargyriou (2008). Control of variable speed wind
turbines in islanded mode of operation. IEEE Transactions on Energy
Conversion, 23 (2) (June), pp. 535–543.
10. V. Courtecuisse, B. Robyns, B. Francois, M. Petit, J. Deuse (2008). Variable
speed wind generators participation in primary frequency control. Wind
Engineering, 32 (3), pp. 299–318.
11. Wind speed data. Available at www.bom.gov.au.
12. N. R. Ullah, T. Thiringer and D. Karlsson (2008). Temporary Primary Frequency
Control Support by Variable Speed Wind Turbines— Potential and
Applications, in IEEE Transactions on Power Systems, vol. 23, no. 2, pp. 601-
612.
13. M. L. Chan, R. D. Dunlop and F. Schweppe (1972). Dynamic equivalents for
average system frequency behavior following major disturbances, IEEE Trans.
Power App. Syst., vol. PAS-91, no. 4, pp. 1637-1642.
14. AEMO (2013). Wind Turbine Plant Capabilities Report. Available at
aemo.com.au
15. H. Chavez, R. Baldick and J. Matevosyan (2015). The joint adequacy of AGC
and primary frequency response in single balancing authority systems, IEEE
Trans. Sustain. Energy, vol. 6, no. 3, pp. 959-966.

256
CHAPTER 7
AUTOMATIC GENERATION CONTROL IN
FREQUENCY RESPONSIVE WIND PLANT
INTEGRATED POWER SYSTEM
This chapter presents integration effect of frequency responsive wind plants effect on
load frequency control in a multigenerational interconnected power system, with
special emphasis on some of the challenges presented in chapter 3. Validated models
of conventional power plants are applied to analyze frequency responsive wind plant
integration impact on classical and FGSPID AGC performance. Effect of wind
participation in frequency regulation is analyzed regarding reserve response and
corresponding impacts on wind plant performance. FRT capability of frequency
responsive wind plant is also investigated in this chapter.
7.1. AGC / Load Frequency Controllers and Conventional
Generation Models
The AGC problem has been extensively studied during the last four decades, and it
has been one of the most highlighted issues in the design and operation of
independent and interconnected power systems. Load Frequency Control (LFC) is
used for many years as part of Automatic Generation Control (AGC) in power system
around the world. Maintenance of system frequency within utility prescribed
tolerance band was considered as the exclusive responsibility of the conventional
power plants. With increased renewable energy targets, globally, it is usual to find an
area regulated by hydro generation interconnected to another area regulated by
thermal generation or gas based generation or in a combination of both generations
257
along with good penetration of wind plants. Frequency response capability offered
by a wind plant is not the same as the control capability of conventional plants. While
frequency response capability offered by a conventional plant can be counted on with
good assurance once the appropriate control modes are on, the response capability of
a wind plant will always be conditional on the statistics availability of the wind.
Increased penetration of wind generation in interconnected power system thus
intrinsically calls for LFC analysis with present and future wind technologies. This
chapter presents integration effect of frequency responsive wind plants effect on load
frequency control in a multigenerational interconnected power system.
7.1.1. Classical AGC Controller
In case of free governor operation with fixed speed changer setting in an isolated
system, steady state frequency change due to load disturbance (∆P can be given as
[1]:
∆f ∆P 7.1
The free governor operation gives a reasonable performance with a frequency drop
of 4-5 % between no load and full load. A steady state frequency is reached when the
sum of online generation is equal to system load. For better frequency control some
other mean should be there, as a large change in frequency cannot be tolerated, and
the steady change in frequency should be zero. While steady state frequency can be
brought back to the scheduled value by adjusting speed changer setting, the system
could undergo intolerable dynamic frequency changes with changes in load. The
speed changer setting should be adjusted automatically by monitoring the frequency.
So, setpoint frequency is changed through a centralized controller which takes a
258
signal normally referred as area control error (ACE). An example of classical AGC
controller is integral controller given as:
∆P K ∆f dt 7.2
As an example, steady state frequency deviation ∆F under integral controller for a
thermal plant based system can be written as:
∆
∆F s 7.3
∗
∆f| s ∆F s 0
→
Where T is governor time constant, T is turbine time constant for thermal power
plant. K is power system gain and T is power system time constant as defined in
Chapter 3.
If the frequency drops by 1 Hz (∆f = -1) then the integrator calls for an increase in
power, with the call increasing at the initial rate of K p.u. MW/s. Negative polarity
must be chosen to cause a positive frequency error to give rise to a negative or
decrease command. From a control system point of view, the integrator has the merit
that it drives the ACE to zero in steady state. Of course, we are implicitly assuming
the system is stable, so the steady state is achievable. So, the gain of the controller
should be optimized to meet high accuracy requirements as the output response
depends in some manner upon the integral of actuating signal. A significant
contribution of this controller to the steady-state performance is however obvious as
the additional integration in the forward loop changes the system order and that the
error of the input is eliminated or considerably reduced, as the practical integration
may not be perfect. For integral type control:
1. If we use subcritical gain settings, we obtain the sluggish non-oscillatory
response of the control loop. This means that the integral of ∆f t , and thus
259
the time error, will be relatively large. In a practical situation, this setting is
most often used. The advantage is that the generator now will not necessarily
“chase” rapid load fluctuations, causing equipment wear.
2. As the sudden load increase sets in, the frequency starts falling off at some
exponential rate. During the first instants, the integral controller has not yet
had time to go into action, and the system response is determined by the
primary automatic load frequency control loop. After a certain time (the
shorter the time, the higher integral gain Ki), the integral controller comes
into action and eventually lifts the frequency back to its original value.
7.1.2. Intelligent AGC Controller
Industrial proportional-integral (PI) or proportional-integral – derivative (PID)
controllers parameters are tuned either manually depending upon operator’s
experience or in an automatic way depending upon process period of several cycles
of operation before getting related parameter gains. This tuning procedure is typically
not applicable in a multigenerational power system with fluctuating renewable power
output serving continuous load demand. For more than four decades, considerable
interest has been shown towards the application of different control strategy for
arriving at more efficient load frequency control for the interconnected power system.
Conventional LFC controllers are simple but may present derogatory control
performance under varying parameters and online frequency tuning. Moreover,
conventional LFC controllers are more model specific failing to provide a satisfactory
response in case of a large non-linear system with stochastic parameters [2-5]. On the
other hand, robust and dynamic LFC controllers involve complex control structures
which are not very much preferred in practices [6]. Motivated by its ability to handle
260
imperfect information and intrinsic uncertainties in the knowledge base, Fuzzy logic
controllers (FLCs) have gained wide popularity for analysis and control of nonlinear
systems [7-9]. Originally developed in 1965 by Zadeh, Fuzzy Logic is now a common
technology at the industrial level. The first application of fuzzy logic to control theory
dates back to 1975 [10]. Fuzzy logic is fundamentally management of non-numeric
and linguistic definite values ranging amid absolutely true and false values, such as
high, low, medium, and often Zero [11].
All fuzzy logic based controller technique comprises three basic components:
Fuzzification: In fuzzification, the input parameter values are fuzzified by some
predefined membership functions, and crisp values are converted to linguistic values
via membership functions. Different membership functions can take different shapes
like trapezoidal, Gaussian, and triangular. However, large membership quantity
rather than type and shape brings more effectiveness in control action regarding speed
and accuracy. Normally recommended for more complex engineering problems,
increased membership functions result in higher accuracy and longer processing time.
By contrast, lesser membership functions provide faster processing time and the
advanced probability of diversity.
Rule base interference: Output parameters are controlled through an inference
engine in this stage. A rule base table is used to apply rules to membership functions.
System knowledge is applied to design these rules by “if–then” concept. Mamdani's
inference method is the most common method that uses the max–min operation
approach.
Defuzzification: Output linguistic values are again converted into relevant crisp and
numerical values using membership functions with the range varying according to
system designer knowledge. Several de-fuzzification techniques exist, such as the
261
center of the area and mean of maxima; the latter is more common in control
applications [13].
Figure 7-1 Conventional Fuzzy PI Controller Model 1 [Reformulated from 12]
Figure 7-2 Conventional Fuzzy PI Controller Model 2
Conventional fuzzy PI controller use K and K parameter as two input and controlled
reference power signal as one output [12]. Area control error and its derivative are
fuzzified using membership functions in different ranges. The conventional fuzzy
logic PI controller is implemented in two different ways as depicted in Figure 7-1 and
Figure 7-2. In Figure 7.1, PI controller output and rate of change of controller output
are used as fuzzy controller input to determine the corresponding output. A
262
membership value for the various linguistic variables is calculated by the rule given
by:
μ e t , ce t , where e t K ACE K ACE dt 7.4
The grade of membership values is determined through the equation of the triangular
membership function given as:
| |
A x 7.5
Where A(x) is the membership grade value, ‘b’ is the width, ‘a’ is the point coordinate
at which the grade of membership is 1, and x is the input variables values.
Conventional fuzzy PI output performance is based on selected membership
functions, corresponding base, and de-fuzzification method. Output power reference
from the model in Figure 7.1 is obtained as:
∑ μu
∆P ∑ μ 7.6
In Figure 7.2, ACE and are two fuzzy input variables with five triangular
membership functions. K′ and K′ are output variables after defuzzification. Output
power reference signal obtained is given as:
∆P K ACE K ACE /s 7.7
7.1.3. Hydro Turbine Governor Model
The representation of hydraulic turbine and water column in load frequency control
study is selected based on the following assumptions:
1. The hydraulic resistance is negligible.
2. The penstock pipe is inelastic, and water is incompressible.
3. The velocity of water varies directly with the gate opening and with the
square root of the net head.
263
4. Turbine output power is proportional to the product of head and volume flow.
Figure 7-3 Classical Hydro Turbine Governor Model
The Classical model of hydro turbine and water column is represented by a single
pole and single zero. For an ideal lossless turbine, change in turbine power output to
change in gate opening is represented in Figure 7-3 and mathematically given as
∆
7.8
∆ .
Water starting time TW represents the time required to accelerate water in penstock
from standstill to the required velocity, and it varies with load. Mechanical power
change (∆ with gate valve change (∆P is given as:
∆P t 1 3e ∆P 7.9
Classic hydraulic turbine linear model being non-minimum phase system has an
important characteristic representing initial mechanical power flow opposite to the
direction of change in gate position. Under sudden gate opening, water inertia
restricts immediate change in flow while pressure across turbine is reduced resulting
in reduced output power. Final steady state output is established only at a steady flow
determined by water starting time.
Hydro Governor Model
As stated above, water inertia brings a non-minimum phase system characteristics to
hydro plant. A large transient droop with a large resetting time is essential for stable
operation of such system, and it is accomplished as transient droop gain feedback as
264
shown in Figure 7-4. This governor model provides high droop during large speed
deviation while exhibits low droop during steady state operation. A linearized model
of this turbine governor was represented in Figure 6-7 of Chapter 6. According to
[14], this speed regulating model will be stable in case of low phase lag contributions
from governor and if

1.5 ∗ 7.10

For transient droop compensator in current hydro model G s ,
transient droop and reset time is calculated according to [15] as
R 2.3 T 1.0 0.015 7.11
T 5.0 T 1.0 0.5 T 7.12
The permanent droop R governs the p.u. frequency deviation in steady state with
p.u. change in valve position. This classic linear model being non-minimum phase
system has an important characteristic representing initial mechanical power flow
opposite to the direction of change in gate position.
Figure 7-4 Hydro Governor Model
Mechanical power change (∆ with gate valve change (∆P is given as:
265
∆P t 1 3e ∆P . 7.13
Hydro turbine rotor speed after a load step ∆ is given as:
∆ω ∆P 7.14
Neglecting transfer function , gate valve position change is given as:
∆ ∆ ∆ ∆
. 7.15
Rate of change of turbine gate valve position just after the load step when ∆ 0
is given as:
∆ ∆
. 7.16
With these equations, it can be concluded that initial mechanical power surge lasts
for 1-2 seconds depending upon load step change, inertia constant and water starting
time. Water inertia restricts immediate water flow during sudden gate opening
thereby reducing pressure across turbine and power reduction. A new steady state
power flow is finally established depending upon water starting time, TW.
Model Validation
Frequency response capability of hydro plant model is evaluated for a hydro unit with
inertia constant of 3 s, load damping constant equal to 1 and droop value of 4.0 %.
Under the assumption of balanced initial load, model performance for primary
frequency response and secondary frequency response was then evaluated under 0.01
p.u. load disturbance. Damping constant of the remaining load of 0.99 p.u. is given
as:
D= ∗ 0.99 ∗ 0.0198 p.u. MW / p.u. Hz

∗
266
Now a regulation of 4.0% means that 4.0 % change in frequency causes a 100 %
change in power generation. So
0.5 p.u. MW / p.u. Hz

∗
Figure 7-5 Frequency Response from Hydro Power Plant
Figure 7-6 Mechanical Power Output from Hydropower Plant during Frequency Response
Composite system frequency response characteristic in this case is:
β D 0.5 0.0198 0.5198 p.u. MW/ p.u. Hz
Steady state frequency response can be calculated as:
∆ .
∆f 0.0192 p.u. Hz or -0.00038 Hz
.
Thus, steady state frequency obtained is 49.9996 Hz
Generation contribution to regulation is ∆P ∆f ∗ β 0.0095 p.u. MW

267
Governor droop response is considered as an uncontrolled primary response while
controlled response refers to AGC response of the unit. The dynamic response of this
model under 0.01 p.u. load disturbance calculated above is shown in Figure 7-5–7-6.
7.1.4. Gas Power Plant Model
The study of a gas based power plant in AGC is important from the viewpoint that
gas-based plants are increasingly being added to the grid system and gas turbines
construction and dynamics are very different from hydro and steam turbines. The
generators are of round-rotor type, have relatively low inertia compared to hydraulic
units, and they spin at higher speeds. A simple cycle gas turbine cycle consists of
three parts. First is the axial compression of gas (typically air). Second is the addition
of heat energy into the compressed gas by either directly firing or combusting the fuel
in the compressed air or transferring the heat through a heat exchanger into the
compressed gas. This stage is followed by expansion of the hot pressurized gases in
a turbine to produce useful work under Brayton cycle. The work required by the
compression is supplied by the turbine with the remaining being available for useful
work. The useful work developed by the turbine may be used directly as mechanical
energy or may be converted into electricity by turning a generator.
Figure 7-7 Split Shaft Based Gas Turbine
268
There exist several dynamic models of turbine-governors with varying degrees of
complexity to represent different makes and models of gas turbine units. There are
essentially two types of gas turbine designs. One is a high-speed single shaft design
with the compressor and turbine mounted on the same shaft as the alternator. Another
is a split shaft design that uses a power turbine connected via a gearbox. In the split
shaft design as shown in Figure 7-7, although there are two turbines, one is gasifier
turbine driving a compressor, and another is a free power turbine driving a generator,
and there are only one combustion section and one gasifier. OCGTs are used mainly
as peaking units due to their operational flexibility, but their relatively low efficiency
means they are typically too expensive for base load operation. CCGTs, which have
relatively high efficiency but reduced flexibility, are more suited for base load
operation. For a base load operation, the gas powered generator is usually more
expensive than coal. Two gas turbine models of varying complexity have been used
in this study for assessing their frequency reserve response.
GAST Model
Figure 7-8 GAST Model
269
The GAST model (based on the split shaft), developed by GE, USA and shown in
Figure 7-8 is one of the models applied for simulation studies for representing the
dynamic behavior of open cycle gas powered turbine governor systems. It is the
simplest representation of the gas turbine model still used in WECC and Western
Power, Australia. The model structure has been selected from the model of reference
[16-17] for simulation studies. Parameters values are in accordance with WSCC
testing programs for gas turbine unit [17].
Figure 7-8 shows three control loop involving fuel control, temperature control loop
and speed feedback loop. A Low-Value Select block is used to determine which
control loop is asking for minimum fuel. This control is then given priority and drives
the fuel valve actuator to control the combustor output. This model is deficient in
predicting accurate operation at high load level. Gas turbine units have relatively low
inertia compared to hydraulic units. They spin at higher speeds, and the generators
are of round-rotor type. Following a sudden load rejection event, a low inertia
machine can experience excessive over-speed, which can be harmful if over speed
protection fails to operate. If over speed protection does operate to arrest the speed,
it will introduce non-linearity in the frequency versus time characteristics. Therefore,
the amount of load rejected should be low to limit over-speed and therefore the
possible impact of non-linearities. The maximum opening by the governor must
permit short time load of at least 104 %. The load level 100 % describes the situation
that the unit is running on maximum load setting and 50 HZ. Model parameters values
are provided in the appendix.
Model Validation
Active power control capability of GAST model based gas turbine for frequency
response performance is evaluated for 95 MW machine with inertia constant of 6.5 s

270
and droop value of 4.2 %. External grid inertia is not considered here, and it is
assumed the system frequency is balance with 1 p.u. of load initially. Model
performance for primary frequency response and secondary frequency response was
then evaluated under 0.01 p.u. load disturbance. Primary frequency response is highly
dependent upon implemented droop and generator characteristics. Damping constant
of a remaining load of 0.99 is given as:
D= ∗ 0.99 ∗ 0.0198 p.u. MW / p.u. Hz

∗
Now a regulation of 4.2% means that 4.2 % change in frequency causes a 100 %
change in power generation. So
. 0.4761 p.u. MW / p.u. Hz

∗
Composite system frequency response characteristic in this case is:
β D 0.4761 0.0198 0.49599 p.u. MW/ p.u. Hz
∆ .
∆f 0.0201 p.u. Hz or -0.00040 Hz
.
Thus, steady state frequency obtained is 49.9996 Hz
Generation contribution to regulation is ∆P ∆f ∗ β 0.0099 p.u. MW
Dynamic response of this model under 0.01 p.u. load disturbance and calculated
above is shown in Figure 7-9 - 7-10. It can be seen from Figure 7-9 that as soon as
frequency drop occurs, governor droop instantly activates primary responsive FCR
to generate power to bring the frequency to a steady state value of 49.9996 Hz. Flat
frequency control is used for AGC response where the integral controller brings the
steady state frequency to the setpoint frequency of 50 Hz. AGC response when
implemented without system bias or network frequency control characteristics
271
β gives increased settling time to 60 sec while it is only 25 sec. when
implemented with bias as shown in Figure 7-11. β equals to the sum of
individual primary frequency responses of all participating generating units within
the power system.
β ∑ β
In current case, as only one gas turbine based unit is there in network, so
β β 0.4912 p.u. MW / p.u. Hz
Figure 7-9 GAST Model in the Presence of Frequency Response
Figure 7-10 Mechanical Power from GAST Model during Frequency Response
272
Figure 7-11 Frequency Response under AGC Control
Decentralized Combined Heat and Power Plant (DCHP) Model
Simplified combined heat and power plant model based on gas turbine is investigated
for its dynamic response in frequency regulation studies. DHCP model adopted in
this study is based on [18]. The basic configuration of gas turbine governor comprises
of four basic blocks as represented in Figure 7-12.
Figure 7-12 Simplified DCHP Model
Speed Governor Model
System frequency control is provided through governor droop model by activating
corresponding FCRs. A low pass filter and the static dead band are applied to avoid
any undesired frequency response and any frequency deviation above the dead band
273
will be changed into corresponding power command (∆Pc) through implemented
droop characteristic. This frequency-power command is passed to the power
limitations block of gas turbine which is shown in Figure 7-13.
Figure 7-13 DCHP Governor Model
Power Limitation
Figure 7-14 Power Limitation Block
Gas turbine power response is restricted under physical constraints imposed by
combustion technology through this limiter block as shown in Figure 7-14. Power
level corresponding to maximum and minimum load is applied for frequency
response limitation. The lowest value of power demand and limited power is passed
as commanded power. Highest value among this commanded power and reference
power setpoint serves as input to power limiter block which limits the commanded
power between the maximum and minimum load setpoint. Rate limiter block prevents
burner flame pulsation during ramping events like loading and unloading. This load
variation commanded power setpoint is then forwarded to power distribution block
for fuel and air flow management.
274
Power Distribution
Two combustion chamber are used in gas turbine: environment combustion chamber
(EV) and sequential environmental combustor (SEV). The environmental capacity is
60% that of SEV burners capacity. Compressed air is mixed with half of fuel and
heated in first EV chamber. Heated mixture after expansion through high-pressure
turbine moves to SEV chamber where it is mixed with remaining fuel. Efficiency and
emission level of the gas turbine is dependent upon air to fuel ratio. Respective
combustion power contribution factors (CEV & CSEV) and variable inlet guide vane
position compressor factor (CVGV) are used in the model shown in Figure 7-15 to
reflect physical characteristics of air-fuel flow and temperature allowance. CFM as
the function of base load, frequency, and the ambient temperature is considered as 1
in the current study.
Figure 7-15 Power Distribution Block
Gas Dynamics
Combustor and compressor dynamics of the gas turbine are represented by this
module shown in Figure 7-16. First order transfer function models the EV/SEV
environmental combustors dynamics while second order transfer function as
275
represented in Figure represents the compressor/VIGV unit dynamics. A limited
mechanical power output is finally obtained from this module to be passed to the
generator.
Figure 7-16 Gas Dynamic Block
Model Validation
Active power control capability of DCHP model for frequency response performance
is evaluated for 265 MW machine with inertia constant of 8.6 s and droop value of
4.0 %. External grid inertia is not considered here, and it is assumed the system
frequency is balance with 1 p.u. of load initially. Under varying load condition, it was
difficult to validate results analytically. The current study evaluates model
performance for primary frequency response under 0.01 p.u. step load disturbance
and initial DHCP reference power setup as 1 p.u. MW. Output power is limited
between 0 - 1.1 p.u. MW.
Primary frequency response is highly dependent upon reference power setpoint,
implemented droop and generator characteristics. Damping constant is taken as 0.
The primary response is provided through speed governor who provides up/down
power change command as shown in Figure to reduce frequency deviation. For a
276
regulation of 4.0% means that 4.0 % change in frequency causes a 100 % change in
power generation. So, Composite system frequency response characteristic for this
plant is β .
. 0.5 p.u. MW / p.u. Hz

∗
∆ .
∆f 0.02 p.u. Hz or -0.0004 Hz
.
Thus steady state frequency obtained is 49.999 Hz. Generation contribution is ∆
∆ ∗ β 0.01 p.u. MW. These can be observed in Figure 7-17 - 7-18.
Figure 70-17 DCHP Frequency Response under 0.01 P.U. Load Disturbance
Figure 7-18 DCHP Power Output Under 0.01 P.U. Load Disturbance
277
7.1.5. Thermal Power Plant
Figure 7-19 Single Reheat Turbine Model
Single reheat steam turbine model is shown in Figure 7-19. Steam enters the high
pressure (HP) section through the control valve. With substantial steam remaining in
inlet chest, exhaust steam is passed through reheater. The reheat steam flows into
intermediate pressure (IP) section through intercept valve. A path is provided for IP
section exhaust steam to low pressure (LP) inlet through crossover piping. Time
constants involve in each of these respective stages is shown in Figure 7.19.
Approximated transfer function model of the thermal system is applied under the
simplistic assumption of constant steam pressure and temperature and unlimited
steam supply. Neglecting crossover piping time constant (T and under the
assumption of control valve linear characteristics, simplified transfer function of the
turbine which relates turbine torque ∆T to changes in control valve (∆V ) is
expressed as:
∆
7.17
∆
∆
7.18
∆
278
Here, F refers to fraction of total turbine power generated by HP section, T refers
to time constant of inlet main volume and steam chest, T refers to time constant of
reheater. In the case of the non-reheat turbine, as applied in the current study, the
turbine is represented by single gain and single time constant. With R representing
droop, K as governor gain and T as governor time constant, speed governor is
modeled as simple transfer function represented in Figure 7-20 and expressed as:
∆
∆X s ∆P s 7.19
Figure 7-20 Simple Governor Model
7.2. Performance Index
The design of a control system is an attempt to meet a set of specifications, which
define the overall performance of the system in terms of certain measurable
quantities. Several performance measures have been introduced so far in respect of
dynamic response to step input (t , t , t ) and steady state error e to both step and
higher order inputs. These measures must be satisfied simultaneously in design, and
hence the design necessarily becomes a trial and error procedure. If, however, a single
performance index could be established based on which one may describe the
goodness of the system response, then the design procedure will become logical and
straightforward.
Furthermore, in many of the modern control schemes, the system parameters are
automatically adjusted to keep the system at an optimum level of performance under
279
varying inputs and varying conditions of operation. Such part of the system is called
adaptive control system. These systems require a performance index, which is a
function of the variable system parameters. Extremum (minimum or maximum) value
of this index then corresponds to the optimum set of parameters values. Other
desirable features of a performance index are its sensitivity, i.e. its power to clearly
distinguish between an optimum and non-optimum system, its sensitivity to
parameter variations and the ease of its analytical computation.
Several such performance indices are used in practice, the most common being the
Integral of the Squared Errors (ISE), given by:
SE ACE t dt 7.20
This criterion has good mathematical trackability properties penalizes positive and
negative errors equally. It penalizes heavily on large errors; hence a small ISE usually
results in a system with small overshoot. Since the integration is carried out over (0,
∞), a small ISE limits the effect of small error lasting for a long time and thus results
in small settling time. Also, a finite ISE implies that steady-state error is zero.
Minimization of ISE by adjusting system parameters is a good compromise between
reduction of rising time to limit the effect of the large initial error, reduction of peak
overshoot and reduction of settling time to limit the effect of small error lasting for a
long time.
Another similar index is the integral of time multiplied by absolute error (ITAE)
which exhibits the additional useful features that the large initial error (unavoidable
for a step input) is not heavily weighted. More sensitive than ISE, it is useful to those
practical applications which require very sensitive criteria [19]. ITAE produces
smaller overshoots and oscillations than ISE, and it is given as:
TAE t|ACE t |dt 7.21
280
These two performance indexes are used in the current study to investigate the
superior performance of different controllers. A control system is optimum when the
selected performance index is minimized. MATLAB’s Simulink is used in
conjunction with the optimization algorithm to determine the optimum parameters of
the AGC for the performance index considered.
7.3. Parameter Optimization Algorithm
There are two aspects of the optimization problem, namely, parameter optimization
and optimal system control. This project report deals with parameter optimization of
load frequency controllers. This means that optimum values of the controller
parameters are to be found for minimum deviations in the frequency and tie-line power
oscillations. To achieve optimal performance, various state feedback, state adaptive
optimal controllers, ANN, Fuzzy based controllers have been proposed all of which in
general suffer from complicated computational burden and memory.
A general pattern search algorithm for optimization has been utilized for obtaining
optimal integral gain for nominal input parameters for AGC controllers. In this
method, a combination of exploratory moves and heuristic pattern moves is made
iteratively. An exploratory move is performed in the vicinity of the current point
systematically to find the best point around the current point. After that, two such
points are used to make a pattern move. In the exploratory move, the current point is
perturbed in positive and negative directions along each variable one at a time, and
the best point is recorded.
The current point is changed to the best point at the end of each variable perturbation.
If the point found at the end of all variable perturbations is different than the original
point, the exploratory move is a success. Otherwise, the exploratory move is a failure.
In any case, the best point is the outcome of the exploratory move.
281
Figure 7-21 Direct Search Algorithm
In pattern move, a new point is found by jumping from the current best point x c
along a direction connecting the previous best point and the current base point. The
optimization method is summarized below in the flow chart in Figure 7-21.
282
7.4. Control Area Simulation Models
7.4.1. A Frequency Regulation Model Incorporating All Three Control
Modes in a Single Area Power System
Figure 7-22 Frequency Regulation Model for a Single Control Area
Figure 7.22 represents a hierarchical frequency regulation MATLAB implemented a
model for single area control formulated based on generating unit power reserve
classification given by European Network of Transmission System Operators for
Electricity (ENTSO-E) [20]. The single control area system dynamic is represented
by following differential equation:
∆f s ∆P s ∆P s 7.22
∗
Where T is defined as power system time constant; K , power
system gain and frequency deviation are represented by ∆f. Combined mechanical
power variation from primary, secondary and tertiary control is represented by ∆P
as:
∆P ∆P , ∆P , ∆P , 7.23
283
The load imbalance is represented by ∆P , H is the total system inertia and D the load
damping coefficient of the network. Hydro turbine governor model discussed earlier
has been used for primary control service while thermal turbine governor model also
discussed earlier is applied for secondary control services. Single area control model
consists of frequency containment reserve services (FCR) for primary frequency
regulation, frequency restoration reserve services (FRR) for secondary load
frequency regulation and tertiary replacement reserve (RR) services. Thermal turbine
unit or gas turbine unit can be implemented as RR units to represent tertiary control
dynamics. Proportional controller ∆P 1

R ∆f is utilized as local primary FCR
controllers where R is the TSO provided droop. PID control action on secondary
control command signal is implemented as a tertiary response. The secondary
controller incorporates a centralized tuned proportional-integral (PI) controller to
respond to area control error. All three control loops are shown in Figure 7-22. Single
area system representation by nonlinear equation can be given as:
x f x, u, w 7.24
Complete system model can be defined by 7m+1 state variables:
x ∆f x x … x 7.25
where m is a number of turbines-governors used in each control level and
x ∆P , ∆P , ∆P , ∆P , ∆P , ∆P , ∆P , 7.26
2m controlled inputs u ∆P , ∆P , 7.27
Disturbance w ∆P 7.28
Wind plant state variables have not been considered in above representation. This
representation can be appended depending upon of level of complexity of wind
turbine model and its participation in respective regulation services. According to
284
several of grid codes, primary control should start as early as 5 seconds, and
secondary control should start between 30-60 seconds and dispatched every 2-5
seconds. Hence, the delay associated with the secondary and tertiary controller in this
model has been set as 60 s and 120 s and is dispatched every 10 and 60 seconds
respectively. Primary control is required and modeled to act instantaneously.
Participation factor of individual units in FCR has been calculated as:

PF 7.29
Where predefined reserve power is denoted by FCR is. B is
grid code defined frequency control bandwidth half (Hz) for FCR activation, and R
is droop characteristics in Hz/per unit. Wind participation in FCR model is shown in
Figure 7-23. Wind plant participates in FRR services by taking automatic generation
control (AGC) power set-point from PI controller as represented in Figure 7-24.
Participation factor of individual units in FRR including wind plant has been
implemented as

PF 7.30

Figure 7-23 Primary FCR Service Model with Wind Participation in FCR
285
Figure 7-24 Secondary FRR Services with Wind Participating In FRR
Figure 7-25 RR Control Loop
Figure 7-26 WTG Inertia Controller
RR control loop is shown in Figure 7-25. Instantaneous power reserve from wind
plant is provided through synthetic inertia controller. Input frequency deviation to
synthetic inertia controller as implemented for this study and shown in Figure 7-26 is
provided through frequency processor block, and their output is considered as an
additional power reference point to modify the final power order of wind turbine.
When individual wind turbine participates in FCR control, final power order received
by electrical control is modified asP P P P . When wind
plant participates in an FCR providing group, it receives an extra FCR command P
286
which gets added up in final power order. In FRR participation, additional reference
point P is added. Economic dispatch for the generators is excluded in this study
and 50% wind participation is assumed.
7.4.2. Decentralized Three Area Interconnected Power System (TAIPS)
Model
TAIPS model interconnected in a ring topology as shown in Figure 7-27 has also
been considered to investigate AGC performance. It is assumed that same plant is
participating in primary and secondary control in the individual area. For example,
hydro plant is participating in primary as well as secondary control in area 3 while
gas plant operates these two services in area 2 and thermal plant operates these
services in area 1. An example of such TAIPS model is given in Figure 7-28.
Figure 7-27 Ring Topology for Interconnected Control Area
For a simple model of any area i with an aggregate governor, turbine generator unit
(turbine time constant T , governor time constant T , governor gate opening P ,
mechanical power output P , Power command in terms of speed change setting P )
overall generation demand dynamics can be written in terms of speed change in the
rotating mass with moment of inertia M and damping D as:
∆ω ∆P ∆P D ∆ω ∆P 7.31
Turbine model dynamic can be written as
287
∆P ∆P ∆P 7.32
Governor dynamic can be written as
∆P ∆P ∆ω ∆P 7.33
Each of the control areas is interconnected to another area through tie-lines.
Summated tie-line power gives the total power flow over the transmission line. The
relative phase angle between control areas which further depends upon relative speed
deviations in control area units dictates the direction of transmission line net power
flow. The tie-line power with small deviations in the angles is given as:
∆P cos δ δ ∆δ ∆δ MW. 7.34
With T defined as synchronizing coefficient, tie-line power deviation can be written
as:
∆P T ∆δ ∆δ MW 7.35
where,T cos δ δ and ∆δ 2π ∆f dt rad 7.36
Tie-line power in terms of frequency deviation is given as:
∆P t T ∆f t ∆f t MW; i, j 1,2,3. i j. 7.37
∆P t ∆P t 7.38
Total tie-line power flow between areas-i and other areas can be calculated as
∆P t ∑ ∆P t 7.39
Load frequency control (LFC) operation is accomplished through individual LFC
controller implementing tie-line bias control which implies area control error (ACE)
as a linear combination of frequency deviations and tie-line power. With wind plant
integration to an area, corresponding area control error is calculated as:
ACE ∆P t b , ∆f ; i 1,2,3. 7.40
288
Where ∆P is tie-line power deviations and ∆f is the frequency deviations.
Figure 7-28 TAIPS Model with Integrated Wind Farm
It was shown in equation 6.15-6.17 in Chapter 6; maximum frequency deviation is
dependent upon system inertia and the equivalent regulation constant. In the presence
of power electronics based wind plant increased integration, system inertia reduces
while the equivalent regulation constant increases. Now as presented in chapter 5 and
6, wind plant participation in frequency regulation is possible in current and future
scenario.
There are three equations in literature for calculating system inertia in the presence
of wind plant which has been applied in this chapter.
289
1. With H and S representing inertia rating and apparent power rating of
individual generating units, wind integrated system inertia constant, H
can be calculated as:
∑ ∗
H , where S ∑ S S 7.41
2. Assuming a fraction of frequency responsive wind power plant z
contribution to system inertia by displacing a fraction of conventional
generation z , new system inertia and droop can be expressed as
H H 1 z z 7.42
R R / 1 z z 7.43
3. Wind integrated system modified inertia constant, H using delay theory of
turbine governor [21] [22]:
H H 1 L H L 7.44
∑ ∗
H , where S ∑ S 7.45
Where H and S are the inertia rating and apparent power rating of individual
generating units. L is wind penetration level.
For frequency responsive wind plant, wind inertia contribution H to power system
by providing ∆P extra active power when the system is subjected to step load
disturbance ∆P is calculated as:
H 7.46
∆ ∆
X e 1 7.47
∆ ∆
Recalling from chapter 6, the minimum frequency occurring after load disturbance
with new system inertia and droop can be calculated as:
290
∆f t 1 exp t ∆P 7.48
T
A sensitivity analysis investigation showed that maximum frequency deviation to
changes in D is very small in comparison to R [23] so load damping constant being
very low can be assumed constant and it does not have a remarkable impact on the
exponential term.
7.4.3. Test System for LVRT Capability Assessment
A typical network connection of a wind farm with electrical grid is shown in Figure
7-29 which is used for LVRT simulation tests. Six DFIG based turbines, each rated
1.5 MW are used in 9 MW wind farm which is connected to the 25KV distribution
network. Wind farm exports power to 120 KV electrical grid through 25 kV feeder
which is 30 km long. A constant wind speed at 11 m/s is applied while power
electronics based IGBTs are used for converters. The wound rotor induction machine
with a maximum operating slip of ±20% produce rated converter voltage. The dc-link
capacitors with nominal dc voltage rating of 1100 V are used for their additional surge
capacity. The impact of pulse width modulation (PWM) switching is minimized by
incorporating a three-phase choke. Voltage disturbances are simulated at bus 120, so
that near to zero voltage is experienced at bus 120 until the line is cleared of
disturbances.
Figure 7-29 Wind Farm Network Connection for LVRT Test
291
7.5. Analysis-I
Synthetic Inertia Enabled Grid Code Compatible Wind Plant Participation
Capability in System Frequency Regulation and Its Effects on System Reserve
Response.
The main aim of this study is to show the design possibility of earlier presented grid
code compatible wind plant participation in various level of frequency regulation.
Simulation test was conducted to assess the frequency responsive wind turbine effect
on system frequency control response and the corresponding effect on wind turbine
output. Changes in reserve power response along with frequency response indicators
like the rate of change of frequency (ROCOF) and frequency nadir are analyzed under
generation disconnection condition as well as load disturbance condition. Normal
load disturbance and contingency load disruption conditions have been used in the
analysis. Normal load disturbance given as [0 0.025 0.04 0.05 0.2] p.u. and
contingency load disruption conditions given as [0 0.2 0.25 0.050 0.200] p.u. have
been used in power imbalance function given in [24] for load disturbance generation.
All these simulations are conducted under frequency controller-II based VSWTG.
This simulation model can be readily applied to any combination of frequency reserve
services along with wind plant integration to determine frequency response
indicators.
Continuous primary response with FCR availability along with timely activation is
very crucial for grid stability. Wind farms can participate as a module within an FCR
providing group and the minimum requirement for activation time can be distributed
among all units in FCR group. Following sections discuss wind integration and
corresponding control area response. Single area simulations are mainly classified
into following categories given in Table 7-1:
292
Table 7-1 Simulation Study Set Up Cases
Case FCR FRR WPP Remark

(MW) (MW) (MW)
A 330 500 0 0.0833 No Wind integration
B 330 500 133.2 0.082 Wind plant integrated to system but no
contribution to frequency control services
C 197 500 133.2 0.07 Wind plant is providing a contribution to
FCR services when 40% conventional
hydro FCR generating units are displaced
and FRR provided by thermal based
generating units alone.
D 330 300 200 0.0633 Wind plant providing contribution to FRR
services when 40% conventional thermal
FRR generating units are displaced and
FCR provided by hydro generating units
alone
E 197 300 333.3 0.05 Wind plant is providing a contribution to
FCR and FRR services with 40%
conventional generating units displaced in
respective services.
Case B
Figure 7--30 Control Area Frequency under Normal Disturbance and No Frequency Control
The control model is first tested for frequency variation and respective changes in
reserve power under normal per unit load change conditions as shown in Figure 7-30
– 7-31. Wind plant is just feeding generated power to the system without participating
293
in any frequency responsive services. FCR service is activated to regulate the system
frequency quickly, and their response is observed as a symmetric version of
frequency. FRR service response which is a resemblance of load disturbance profile
curve in red helps in restoring frequency to the nominal value. RR service is required
to be operational only in case of severe frequency deviation.
Figure 7--31 Regulation Reserve Response with Integrated Wind Providing No Frequency Response
Case C
Figure 7-32 FCR Control Signal Received by Wind Plant When 40% Conventional FCR Units are
Disconnected
As per Nordic grid code, primary Frequency Containment Reserve (FCR) should be
activated as soon as frequency reaches 49.9 Hz. The Same condition is valid for
NEM, Australia which requires regulation FCAS to be activated as soon as frequency
294
reaches 49.9 HZ. In second simulation condition with a random contingency load
disturbance condition and 40% of hydro FCR units disconnected, frequency
controller 2 based wind plant is used to participate in FCR services. When wind farm
participates in FCR group, FCR command signal as per the participation factor
defined previously is sent to wind farm. This FCR command signal as shown in
Figure 7-32 is further distributed to individual wind turbines through farm controller.
An aggregate wind farm model has been used in current study. Total FCR power
change as shown in Figure 7-33, consisting of 40% wind power and 60% thermal
power is initially responsible for frequency deviations reduction.
Figure 7-33 FCR Response with 40 % Wind Contributing to FCR Reserve
Figure 7-34 Power Set-Point Generated by Frequency Controller 2 when Wind Plant Participates as FCR
295
Various power setpoint generated during contingency load condition is presented in
Figure 7-34 where it can be noticed that during upward high frequency, power
setpoint drops to 0.6-0.59. Grid controller block generates frequency responsive
active power setpoint varying between 0.97 p.u. - 0.6 p.u. according to frequency
deviation and available power. Limited power setpoint as per grid code condition is
forwarded to turbine controller which further generates total power command for
generating electrical power. An increase in wind turbine power output is observed
when frequency increases above set dead band resulting in total power to approx.
1.01 p.u. including inertial power as shown in Figure 7-35.
Figure 7-35 Generated Power with Inertial and Frequency Controller 2 Enabled when Wind Plant is
participating in FCR
Power reserve response during the loss of 40% of the conventional generating unit
under random high load demand conditions is shown in Figure 7-36. FCR units
though provide support to arrest initial frequency overshoots but are not able to
regulate the frequency variations which activates FRR services which provide
comparatively fast response to follow the high demand change while RR services
activation is quite slow. Activation of FRR and RR service can be made faster by
controller tuning. If the wind plant is made to operate in FCR control with normal
296
droop-inertia controller enabled but no frequency responsive active power setpoint,
wind output is observed dropping to 0.35-0.4 per unit, and no support is noticed in
terms of total FCR power change. Increasing the wind turbine output, in that case,
requires changing the torque controller parameters as suggested in [25] and in the
previous chapter.
Figure 7-36 Power Regulation Reserve Response for Random High Load With 40% Hydro Replaced
With Droop-Inertia Enabled Frequency Responsive Wind Plant
Case D
Frequency responsive wind plant participation in secondary frequency control is
assessed by providing FRR power setpoint for winding turbine controller loop. Two
different sets of simulations were conducted for secondary control response
investigation under normal and contingency load disturbance. System regulatory
power change response is analyzed when 40% of wind based plant capacity
participates in secondary control with 60% of thermal based plant and 100% hydro
based FCR control and 40% Wind plant with 60% hydro based plant FCR
combination. FRR power reserve AGC signal from PI controller is distributed to all
participating units. A clear increase in FCR and FRR power reserve activation
response is observed with wind participation in these services due to the high ramping
capability of the wind plant. An example of FCR and FRR command signals received
297
by individual turbines when there is a random high load variation is shown in Figure
7-37, and corresponding power reserve change is given in Figure 7-38.
Figure 7-37 FCR and FRR Command Signal Received by Wind Turbine Unit under Random High Load
Condition
Figure 7-38 Reserve Response during Random Normal Load Variations When The Wind Participates in
FRR Only
Case E
In this case, the wind farm is required to provide more response to FCR services
command in comparison to FRR services. Comparing reserves response from Figures
7-36, 7-38 and 7-39, we can observe that FRR was comparatively fast and matches
closely with load disturbance while a derogatory FRR is achieved in the case of
incorporation of wind plant in both services as noticed in Figure 7-39. The addition
of wind plant output in both FCR and FRR services reduces considerable system
298
inertia and adds extra stress on automatic generation controller requiring parameter
tuning for better system frequency settling response. It is to be noted that PI controller
has been used in wind-based FRR service controller while integral controller has been
used in conventional plant-based FRR services when wind plant participates in FCR
services only. Proper controller selection and tuning is another important topic which
is out of scope for current study.
Figure 7-39 Reserve Response during Random High Load Variations When The Wind Participates in
FRR & FCR
Frequency nadir after any contingency is detrimental to primary FCR performance
for maintaining system stability [26]. Wind plant integration and its participation in
regulation services can be analyzed through frequency nadir points achieved under
different combinations. In the current study, highest drop in frequency is observed
during contingency load condition indicating insufficient frequency control power
reserves while no wind integration gives minimum frequency nadir. Figure 7-40
shows the minimum frequency nadir and ROCOF under varying wind conditions and
different integration scheme. Lowest frequency nadir point is obtained when the wind
is participating in FCR and FRR services both during under during contingency high
load and during normal load conditions. These dips in frequency nadir points are due
299
to generation disruption resulting in lower system inertia. 40% of conventional units
in each service are replaced with wind plant resulting in lower system inertia.
Figure 7-40 Frequency response Indicators Comparison with Wind Plant Participation in Different
Regulation Services and Different Load Demand
ROCOF is lowest without any wind integration. ROCOF is highest when wind plant
is integrated with network and participating in secondary control (FRR) and primary
control (FCR) both. When 40% wind plant is contributing to FCR along with 60%
conventional hydro plant, the rate of change of frequency (ROCOF) increases in
comparison to that of 100% conventional system FCR support without wind power
addition to the network. This ROCOF change from 0.005 Hz/s to 0.5 Hz/s is due to
change in total inertia of the system. An approximate increase of 50% in ROCOF is
observed during contingency load condition when primary FCR and secondary FRR
are not able to avoid sustained frequency deviations. The values obtained suggest that
no relay tripping is possible with around 40 % wind penetration in regulatory
services.
Impact of Wind Plant Participation in Regulation Services on Generated
Electrical Power, Rotor Speed, and System Frequency
Wind electrical power output is limited to 1.1 per unit in this study. Under normal
load disturbance and sufficient power reserves, electrical power of 1 per unit without
any inertial power is produced as shown in Figure 7-41. Effect of the drop in wind
300
speed change can be noticed with a drop in electrical power and rotor speed during
1800-2300 s which is shown in Figure 7-42. Figure 7-43 and Figure 7-45 shows the
increased electrical power output and rotor speed of wind plant during random high
load conditions and participation in regulation services. Zoomed version of electrical
power for a time duration of 500-1000 second is shown in Figure 7.44.
Figure 7-41 Electrical Power under Normal Load Disturbance (Inertial Power is Zero)
Figure 7-42 Rotor Speed under Limited Power Setpoint
With the increase of power from 1 p.u to 1.1 p.u. during time step of 550-650 sec, a
relatively small drop in rotor speed from 1.2 p.u. is obtained during the same time
duration. Under extreme frequency conditions on the downward side, inertia
controller provides 0.1 p.u. extra power to the total wind generated power as shown
in Figure 7-44. Wind turbine inertia controller gets activated if input frequency is
more than the frequency dead-band limit and inertia controller power output being
301
limited to 0.1 p.u. which is set in accordance with a change in rotor speed to avoid
any mechanical damage. With set power limitation and reference speed, wind turbine
operation is in a stable state during the whole operation. The torque-speed controller
can maintain rotor speed up to 1.2 p.u. with a minor drop of 0.01 p.u. during the
inertial response. Simulation results show that with modified control loop, the wind
turbine can implement frequency responsive power setpoint by generating frequency
responsive power and maintain rotor speed up to a maximum of 1.5 p.u. for wind
speed varying up to a maximum limit of 20 m/s. Wind turbine protection loops are
enabled at high wind and high-frequency deviations.
Figure 7-43 WTG Electrical Power during Load Disturbance Contingency
Figure 7-44 WTG Electrical Power during Load Disturbance Contingency (Zoomed)
302
Figure 7-45 WTG Rotor Speed during Random High Load
Figure 7-46 System Frequency when Wind Plant Participates in Different Regulation Services at High
Load Disturbance
Figure 7-47 System Frequency When Wind Plant Participates in Different Regulation Services at normal
Load Disturbance
303
Control area system frequency achieved with wind integration and participation in
regulation services is shown in Figure 7-46 - 7-47. With evident inertia changes, best
regulation of frequency to the setpoint is seen with wind plant participation in
individual control service like FCR services only while frequency overshoots at some
instances when the wind participates in both primary and secondary control services.
It is to be noticed that wind plant can respond to AGC signal for FRR participation if
frequency nadir goes beyond a minimum threshold and all FCR reserves are
exhausted.
Figure 7-48 System Frequency during a Contingency/High Load
Figure 7-49 WTG Electrical Power during Contingency/High Load
FRR capable wind plant closely follows the system frequency in terms of generating
power as per their capacity. It can be seen in Figure 7-49 that electrical power is 1.1
per unit power at a time step of 650 s when corresponding frequency drop is lowest
304
as shown in Figure 7-48. As soon frequency starts increasing, wind turbine adjusts its
power out as per grid code active power set-point. More clear response for wind
participation in FRR services at a random load can be noticed in Figure 7-51 where
WTG increases its output to 1.1 per unit at around 400 sec when system frequency
crosses the threshold of 49.5 Hz as shown in Figure 7-50. Wind FRR response is
seized as soon as frequency improves and starts generating as per given active power
setpoint. More overshoots and undershoots in frequency and corresponding
fluctuations in WTG power outputs are observed during a random load when the wind
participates in FCR services only or combination with FRR services.
Figure 7-50 Comparison of Wind Participation in FCR & FRR Response In Terms Of System Frequency
during a Random Load
Figure 7-51 Comparison of Wind Participation in FCR & FRR Response In Terms Of WTG Power
during a Random Load
305
7.6. Analysis-II
Frequency Responsive VSWTG Integration Effect on Multi-Generational
System AGC Control
For generating systems of widely different characteristics, research needs to be done
to study automatic generation controllers in interconnection. A steam turbine converts
the potential energy of high temperature and pressure steam into mechanical
(rotating) energy, which then is converted into electrical energy in the generator.
Compared with steam turbines, hydro turbines are easier and cheaper to control, while
gas turbine model is quite complex with respect to earlier two as this system must
take care of different control loops and actuate only one as per requirement. The
hydropower systems have the relatively large inertia of the water being used as the
source of energy which causes a significantly larger time lag in response to changes
of the prime mover torque, due to changes in gate position. Also, the response may
contain oscillating components caused by the compressibility of the water or by surge
tanks while gas turbines have relatively low inertia compared to hydraulic units. They
spin at higher speeds, and the generators are of round rotor type. Following a sudden
load rejection event, a low inertia machine can experience excessive over-speed,
which can be harmful if over speed protection fails to operate. Modern WTGs have
inertia constants which are comparable to those of conventional turbine-alternators.
Also, VSWTs can operate in a wide range of speed changes. The generator speed can
drop to as low as 0.7 p.u. speed, while conventional unit speed can only drop to as
low as 0.95 p.u. speed. Wind kinetic energy could be utilized to provide temporary
primary frequency control support to the grid in the event of a load/generation
mismatch. However, the power is limited by the operating conditions and the power
rating of the VSWT.
306
As presented in Chapter 5, an important feature of VSWT is the possibility for their
active power outputs to be controlled as required by system frequency and operators
command. VSWT can increase their output power almost instantaneously. This is an
important feature, although the steady-state active power delivered to the grid by a
VSWT depends on the mechanical energy transferred from the wind, the electric
power can be transiently controlled, to a certain extent, by resorting to the mechanical
system kinetic energy. With advances in control techniques, wind-based plants are
expected to participate in secondary frequency regulation. Wind plants with variable
output are integrated with peak load plants like gas based plants & hydro plants rather
than with baseload thermal plants. Dynamics of different combinations of turbine-
generators have different effects on system frequency maintenance.
This frequency regulation study is conducted to analyze wind plant integration with
thermal plant, gas plant and hydro plant in an interconnected control area and
discusses the tradeoff in selecting a hydro or thermal plant with comparatively low
inertia gas plant. Three area control model incorporating thermal, hydro and GAST
based gas power plant model presented earlier has been used for this investigation.
Hydro and thermal generating unit’s parameters are applied according to [27] [1] and
other system parameters are calculated as per the equations given earlier. Though
several simulations were conducted for different configurations of conventional
generating units, the interaction of thermal-gas-hydro implemented in area 1-area 2-
area 3 are discussed here.
Simulation test scenarios are constructed as follows:
1. Area 1 includes a thermal based generator which is combined as a single unit
participating in both primary and secondary regulation. Area 2 includes gas
based power plant while area 3 has a hydro based power plant.
307
2. Wind power is included in one area only.
3. Load disturbance is simulated by 0.01 p.u. step function at 5 s.
4. Level of wind penetration is adjusted as per variation of inertia reduction: 0%,
10%, 20%, 40% and 60%.
5. Normally operated wind plant and grid code compatible frequency responsive
wind plant are applied. Rated wind speed of 15.5 m/s is applied in this study.
6. The impact of wind power on AGC performance is analyzed with and without
considering the variation in frequency bias and equivalent system droop.
7. Area bias factor B is taken equal to power plant frequency response
characteristic .
8. Frequency responsive wind plant participation in Area 1 AGC is assessed and
simulated for different participation factor of thermal-wind as 0.9-0.1, 0.8-
0.2, 0.6-0.4 and 0.4-0.6.
9. Both equations of system inertia defined earlier are applied for frequency
responsive wind plant to analyze the impact of R, H, and B on system
frequency deviation.
10. A comparison of frequency deviation for GAST based gas turbine model is
made with DCHP based gas power plant. Effect on frequency response due to
the variation in fuel limit in GAST model is analyzed. DHCP operates at an
initial power of 0.75 p.u. All other values are provided in Appendix.
With Step reduction factor α = 2, termination parameter ε = 10 –4 & step size Δ = 0.1,
optimum values of integral controller gain were found as given below. The integral
of time multiplied by absolute error ITAE has been used for calculating performance
index of the system in this study and provided in Table 7-2:
308
Table 7-2 Optimum Values of AGC Integral Controller
Optimum Value KI / ITAE KI/ ITAE KI/ ITAE
Thermal-Gas-Hydro -0.3502/ 5.3857 -.0865/34.6329 -0.85/46.5892
Scenario 1: No Wind Integration
Three area control model is simulated to assess frequency response without wind
plant integration. System parameters and corresponding frequency response
indicators are given in Table 7-3.
Table 7-3 System Parameters and System Frequency Response Indicators Obtained Without Wind
Integration
Inertia D R 2H B ∆
Reduction
(p u)
%
0 0.01 4 10 100 20 0.26 0.15 0.01
Control Area / Generating Unit ∆ (Hz) Settling Time (s) ROCOF (Hz/s)
Load Disturbance of 0.01 p.u. is given at 5 s in all three areas.
1 – (Thermal) -0.063 60 0.05
2 – Gas -0.062 60 0.05
3 – Hydro -0.085 60 0.05
Load Disturbance of 0.01 p.u. is given at 5 s in Area 1 only.
1 – (Thermal) -0.027 50 0.05
2 – (Gas) -0.020 60 0.015
3 – (Hydro) -0.026 60 0.017
1 – (Thermal) -0.025 50 0.017
2 – (Gas) -0.034 70 0.05
309
3 – (Hydro) -0.036 70 0.023
1 – (Thermal) -0.034 60 0.02
2 – (Gas) -0.038 70 0.025
3 – (Hydro) -0.043 60 0.05
Figure 7-52 Comparison of Area 1-2-3 Frequency Response for 0.01 p.u. Load Disturbance in Area 1
Only
Figure 7-53 Comparison of Area 1-2-3 Frequency Response for 0.01 p.u. Load Disturbance in All Three
Area
310
Figure 7-54 Mechanical power Change of all Generating Unit in Interconnected 3 Area for 0.01 p.u. Load
in Area 1 Only
Figure 7-55 Mechanical Power Change of All Generating Unit in Interconnected 3 Area for 0.01 p.u.
Load in all Three Area
Figure 7-56 Tie-Line Power Change of All Generating Unit in Interconnected 3 Area for 0.01 p.u. Load
In Area 1 Only
311
Figure 7-57 Tie-line Power Change of All Generating Unit in Interconnected 3 Area For 0.01 p.u. Load in
All Three Area
Scenario 2: Wind plant integration without any frequency response
Type 3 based aggregate wind plant model is integrated with each conventional plant
in the different area and frequency response indicators investigated. Table 7-4
provides the value of R, B, and H with and without considering wind integration
impact. These values are applied only to wind integrated area. Equation 7.44 has been
applied to calculate system inertia for frequency responsive wind plant.
Table 7-4 Wind Plant Integrated System Data
Inertia
Frequency Responsive Wind Plant
Without Considering Wind Effect
Reduction R H B R H B H
(%)
0% 4 5 0.26 4 5 0.26 5
on R & B
10% 4.45 4.5 0.235 4 4.5 0.26 4.76
20% 5 4 0.21 4 4 0.26 4.50
40% 6.67 3 0.159 4 3 0.26 3.88
60% 10 2 0.11 4 2 0.26 3.10
312
Figure 7-58 Frequency Deviation in Area 1 for Different Level of Wind Plant for 0.01 p.u. Load in Area 1
Considering the Effect of The Wind on R and B
without Considering the Effect of The Wind on R and B
Scenario 3: Frequency responsive wind plant integration
Grid code frequency responsive wind plant presented in Chapter 5 is integrated with
a thermal plant in area 1. In the second case, same wind plant is integrated with
thermal, gas and hydro for comparing their response.
313
Considering the Effect of The Wind on R and B
without Considering the Effect of The Wind on R and B
Figure 7-62 Comparison of Frequency Deviation When Only Conventional Plant Participates In AGC in
the presence of Frequency Responsive Wind Plant and 0.01 p.u. Load Disturbance.
GAST Model with Lower Fuel Limit of -0.02 is applied for Gas Plant Model
314
Figure 7-63 Comparison of frequency deviation When Only Conventional Plant Participates in AGC in
the presence of Frequency Responsive Wind Plant and 0.01 p.u. Load Disturbance.
Figure 7-64 Interconnected 3-area Frequency Response When Gas Based Area 2 is Integrated with
Frequency Responsive Wind Plant and 0.01 p.u. Load Disturbance is Applied in area 2. GAST Model
with Lower Fuel Limit of -0.02 is applied for Gas Plant Model.
Figure 7-65 Interconnected 3-area Frequency Response When Gas Based Area 2 Is Integrated With
Frequency Responsive Wind Plant and 0.01 p.u. Load Disturbance is Applied in Area 2. GAST Model
with Lower Fuel Limit of -1 is applied for Gas Plant Model.
315
.
Figure 7-66 Interconnected 3-area Frequency Response when Thermal-Based Area is Integrated with
Frequency Responsive Wind Plant and 0.01 p.u. Load Disturbance is Applied in Area 1.
Only Conventional Generating Unit Participates in AGC.
Figure 7-67 Interconnected 3-area Frequency Response When Hydro Based Area 3 is Integrated with
Frequency Responsive Wind Plant and 0.01 p.u. load Disturbance is applied in Area 3.
Only conventional Generating Unit Participates in AGC.
Figure 7-68 Interconnected 3-area Frequency Response when DCHP Based Area 2 is integrated with
Frequency Responsive Wind Plant and 0.01 p.u. Load Disturbance is applied in Area 2. Only
Conventional Generating Unit Participates in AGC. Load Limit (1.2, -0.1) p.u.
316
Figure 7-69 Interconnected 3-area Frequency Response when DCHP Based Area 2 is Integrated with
Frequency Responsive Wind Plant and 0.01 p.u. Load Disturbance is applied in Area 2.
Only Conventional Generating Unit Participates in AGC. Load Limit (1, 0.02) p.u.
Figure 7-70 Interconnected 3-area Frequency Response when DCHP Based Area 2 is Integrated with
Frequency Responsive Wind Plant and 0.01 p.u. Load Disturbance is applied in Area 2. Only
Conventional Generating Unit Participates in AGC. Load limit (1, -1) p.u.
Scenario 4: Frequency responsive wind plant participation in AGC
Grid code frequency responsive wind plant with frequency controller-II participates
in AGC with each conventional plant under different participation factor which is
decided as per their penetration level.
317
Figure 7-71 Comparison of Frequency Response for Thermal-Based Area 1 Integrated with Frequency
Responsive Wind Plant when Both of Them Participates in AGC as Per the Indicated Participation
Factor. Effect of Wind Plant on R and B has been considered here.
Figure 7-72 Comparison of Power Deviation for Thermal-Based Area 1 Integrated with frequency
Responsive Wind Plant when Both of Them Participates in AGC with 0.4 Participation Factor (PF) for
Thermal and 0.6 PF for The Wind. Effect of Wind Plant on R and B has been considered here.
Figure 7-73 Comparison of Frequency Response for Thermal-Based Area 1 Integrated with Frequency
Responsive Wind Plant when Both of Them Participates in AGC as per the Indicated Participation
Factor. Effect of Wind Plant on R and B has not been considered here.
318
Figure 7-74 Comparison of Power Deviation for Thermal-Based Area 1 Integrated with Frequency
Responsive Wind Plant when Both of Them Participates in AGC with 0.4 Participation Factor (PF) for
Thermal and 0.6 PF for The Wind. Effect of Wind Plant on R and B has not been considered here.
Discussion:
Figure 7-75 ACE obtained Under Different Simulation Scenarios
319
Each power network has a specific characteristic based upon its components and
parameter involved still some noteworthy observations can be made from these
simulation tests:
1. Any positive values of area bias factor B along with an integral controller
guarantees that all ACEs will be brought to zero eventually [28]. Also, it has
been suggested earlier in [29] that selecting area frequency response B equal
to system bias factor β in all areas fulfills the non-interaction period between
frequency and tie-line power as well as interconnected area for low-frequency
approximation. Simulation results show that this rule remains valid for low
load scenario without any wind power integration giving minute ACE for all
three areas. In the case of wind integration with system inertia reduction;
setting B= β fails the non-interaction criteria for pool operation of the
interconnected network as shown in Figure 7-75. Under all the same
parameters, ACE in wind integrated area increases under the rule of B= β as
is evident from the Figure 7-75. With frequency responsive wind plant, ACE
decreases in comparison to normal wind plant integrated the system.
Incomplete speed droop adjustment of conventional units and additional
droop response of wind plant along with any surplus generation can lead to a
different bias factor. Any change in scheduled transmission line power needs
to be considered with wind power integration.
2. From scenario 1, Table 7-3, it can be noticed that frequency deviation is much
smaller in the absence of any wind plant and it increases when load
disturbance is given in all three areas. With same generating capacity and
under low load disturbance, highest and fastest power is provided by thermal
plant followed by gas and then by the hydro plant as shown in Figure 7-54. In
320
the case of high load disturbance, thermal plant response is followed by hydro
and then by the gas plant as shown in Figure 7-55. Viewing the tie-line power
response in Figure 7-56 - 7-57, it can be observed that highest tie-line power
deviation under low load is for thermal based area 1 followed by a gas plant
in area 2 while hydro based area 3 has an excess of power. This scenario
changes under high load in all three area where highest tie-line power
variation is observed for hydro based area 3 making it suitable as base load
plant.
Figure 7-76 Comparison of Frequency Deviation for Thermal-Based Area 1 Integrated with Wind
Plant. ThermalAGC-Wind Refers that only Thermal Plant Participates in AGC and Normal
Operating Wind Plant is integrated into System. The Wind (RB) Indicates R and B Values are changed
as Per Inertia Reduction Impact. FRWind Refers to Frequency Controller-II Based Frequency
Responsive Wind Plant. ThermalAGC-FRWINDAGC Refers to Participation of both Units in AGC.
3. With increasing amount of wind power in an interconnected network, there is
an increase in frequency deviation due to a proportional reduction in inertia
and higher amount of equivalent system droop. This can be observed in
321
Figures 7-58 - 7-61. A comparison graph for frequency deviation for thermal
based area 1 integrated with wind plant is shown in Figure 7-76 where we can
see that frequency deviation observed in area 1 under the effect of wind
increases with increasing amount of the wind and corresponding inertia
reduction. Frequency deviation is -0.025 for area 1 when no wind power is
added while it increases to -0.55 Hz for 60% normally operated wind power
penetration in the network when no effect on R and B is considered.
Frequency deviation worsens further to -1.35 if values of R and B are also
changed correspondingly. Frequency nadir is sensitive to system droop and
power generated in a control area while it shows low sensitivity to system
inertia. Time to reach frequency nadir being sensitive to system inertia
increases with reduction in inertia due to increasing wind penetration.
Figure 7-77 Comparison of ROCOF for Thermal-Based Area 1 Integrated with Frequency Responsive
Wind Plant. ROCOF values are observed when 0.01 p.u. of Load is applied at 5 s in Area 1 Only. Legends
are same as explained above in earlier Figure.
322
4. Even though only generic models are applied in this study, still some
significant observations can be made for ROCOF. Comparing the values of
rate of change of frequency which is measured at 5 s when 0.01 p.u. step load
is applied; it can be seen in Figure 7-77 that ROCOF increases with increasing
wind penetration. This effect is more prevalent when the effect of the wind on
values of R and B are considered in simulations. At higher penetration level
of 60%, ROCOF values improve to -0.4 Hz/s when frequency responsive
wind participation is more than the conventional unit in comparison to the
ROCOF value of 0.9 Hz/s when normally operated wind plant is integrated.
This behavior calls for a deeper analysis for a high wind penetrated system as
protective relays based on the rate of change of frequency are highly
vulnerable to ROCOF values. Further to this, the trend of increasing the
ROCOF values due to the consideration of variation in R and B values
changes for 60% wind penetration. We observed that ROCOF values improve
for higher wind penetration both normally operated and frequency responsive.
Figure 7-78 Comparison of Settling Time for Thermal-Based Area 1 Integrated with Frequency
Responsive Wind Plant. Legends are same as explained above.
323
5. Similar to ROCOF values, settling time also increases with penetration level
as observed in Figure 7-78. A decreasing trend in settling time is observed
when the wind participates in AGC along with the conventional unit. This
happens due to higher ramping of wind power with faster frequency response
controller. Figure 7-72 and 7-74 shows the power variation in area 1 when
thermal-wind plant participates in AGC with participation factor (PF) of 0.4-
0.6. With higher PF, wind plant with faster power response provides more
power than a thermal plant having low PF. Under low load disturbance of 0.01
p.u. in area 1 only, tie-line power support from other area is low. High
deviation in frequency (-0.9 Hz) and increased settling time (120 s) is
observed for changed values of R and B under wind impact while this
deviation is highly reduced to -0.2 Hz and settling time is reduced to 70 s if R
and B are not changed as per inertia reduction.
Figure 7-79 Main Effect Plots for AGC Parameters on System Frequency Deviation.
6. Area 1 was integrated with frequency responsive wind plant in different
penetration level and system inertia calculated according to both equations
7.42 and 7.44. R and B values were also varied as per Table 7-4. Frequency
324
deviations obtained was plotted against H, R and B. It can be observed from
Figure 7-79 that frequency deviation is least affected by system bias value
while it is highly affected by R variation.
7. Contingency reserves in Australia are divided into six categories as defined
in chapter 3 earlier. Four of these services are fast raise (R6) and fast lower
(L6) which provides a response within 6 seconds while slow raise (R60) and
slow lower (L60) provide stabilization response within 60 seconds.
Conventional generating unit frequency response was investigated when
respective control area has 10% inertia reduction due to wind integration.
Thermal plant is integrated into area 1 with 10% frequency responsive wind
plant and load disturbance of 0.01 p.u. is given in area 1 only while other areas
support the area 1 during this disturbance. Similarly, it is done for area 2 and
area 3. Mechanical power of thermal generating unit changes rapidly in
comparison to two other units.
Figure 7-80 Comparison of Frequency Response Indicators for Conventional Plant Integration with
Frequency Responsive Wind Plant when Only Conventional Unit Participates in AGC and 0.01 p.u. Load
is applied.
Comparing the individual response of conventional units from Figure 7-63 -
7-68 and 7-80, we can observe that thermal plant gives fastest and best
combination for integration with wind plant followed by gas and then by the
325
hydro plant. Change in initial opposite transient power due to gate opening at
penstock, hydro turbine governors require transient droop compensation with
long resetting time for stable speed operation and give a relatively slow
response. Thermal and gas plant with their normal speed droop gives a faster
ramping response, but their response decreases over time due to fuel
limitation of gas and steam. Their turbine output is highly dependent upon set
fuel and load limit. Thermal and gas based units though have relatively lower
efficiency, but their combination with fast, responsive units like frequency
responsive wind can reduce regulation procurement in an area. So, in the
Australian context, thermal of gas based units are more suitable for R6 FCAS
in the wind-integrated area while the hydro plant is better suited for R60
services.
8. A comparison of GAST model with DCHP plant model when integrated with
frequency responsive wind plant is conducted for frequency deviation under
0.01 p.u. load disturbance. The use of negative value for the low limit in
GAST model represents negative mechanical power. At high loading levels,
the high exhaust temperatures can damage the turbine blades. Therefore,
when the exhaust temperature exceeds a set temperature, gas turbine output
transfers to temperature control. The GAST model does not give an adequate
representation of the temperature control loop. Though GAST model has fuel
limitation, it was observed that by changing the GAST fuel limit from 0.02 to
-1 p.u., there is no effect on system frequency deviation but settling time
shows improvement in the interconnected system as shown in Figure 7-62 -
7-65. It can be seen from Figure 7-68 – 7-70 that much higher deviations are
observed for DCHP plant in comparison to simple GAST model. DCHP
326
model output is highly dependent upon power limitation and maximum-
minimum load limitation for frequency response. If the turbine is initially
below full-load (i.e. below its temperature limit), then in the event of a
frequency excursion there will be a transient overshoot in power output as the
governor acts to increase power. After a few seconds (the time constant
associated with the thermocouples measuring exhaust temperature) the
temperature control loop will take control and start to decrease turbine output
to maintain the exhaust temperature limit if the governor attempts to push it
beyond this limit. GTs, whether stand-alone or as part of a DCHP, can make
some change in mechanical output affecting the frequency response. The
increase in speed following a frequency disturbance should result in some
response by the speed governor to decrease turbine power during the first
swing in machine speed. This initial governor response, even if small, will
improve the ability of the unit or plant to remain synchronized. It is important
that a detailed representation including temperature control and maximum
rate of fuel valve closing be included in models.
7.7. Analysis-III
Development of FGSPID Based Intelligent AGC and Analysis in the presence of
Grid Code Frequency Responsive Wind Power Penetration
Gain scheduling is a distinct non-linear control technique to alter linear controller
parameters in a non-linear mode depending upon process state and scheduling
variable. With changing process operating points, scheduling variables also changes
its values. Fuzzy gain scheduling works on the principle where PID controller gains
are altered during the online process through fuzzy logic and reasoning. Many
researchers consider conventional fuzzy PI controller as gain scheduled controller,

327
but the original concept of fuzzy gain scheduling proposed in [30] differs from
conventional fuzzy PI techniques as the controller parameters obtained after de-
fuzzification are further normalized between a maximum and minimum range
depending upon stable system response. This concept is applied in the current study
and explained briefly in this section. The transfer function of PID controller is given
as:
G s K 1 T s 7.49
where T = K /K , T , K , K and K are proportional, integral and derivative
gains. T & T are integral and derivative time constants and related as: T ∝T .
So K 7.50
∝ ∝
Figure 7-81 Fuzzy Gain Scheduling (FGSPID) Structure
Proportional and integral derivative gains are assumed to lie within a prescribed
range, and gain scheduling method has been applied to adjust K , K and ∝
parameters conferring to area control error signal (ACE) and its first difference
changes, as illustrated in Figure 7-81 and represented by membership functions which
are shown in Figure 7-82 – 7-83. Two input ACE and ∆ACE with A and B being
corresponding fuzzy sets are applied while we have three outputs K , K with C and
328
D corresponding fuzzy sets while last output ∝ being held constant. Fuzzy rules are
selected on the generalized theory behind the step response of the process thereby
bringing area control error to zero and system frequency to set value.
Figure 7-82 Membership Functions for and
Figure 7-83 Membership Functions for ACE and ∆ACE
Figure 7-84 Hydro-Turbine Governor Response
329
Figure 7-84 shows the step response of hydro turbine governor system. It can be
noticed that at point A, the large rise time is desired implying large proportional and
integral gain and small derivative gain from LFC controller. Small T or small ∝ will
result in a strong integral gain. Rule around point A can be (Positive big), AND ∆ACE
is ZO (Zero), THEN K is Big, K is Small, and ∝= 2. Similarly, at point B large
overshoot is to be avoided thereby implying small LFC control signal which means
small Kp and Ki and big Kd values. Thus, the rules for other points can be described
as per Table 7-5.
Table 7-5 Factors Affecting fuzzy PID Tuning
Action Rise time Overshoot Steady state Settling System stability
error Time
Increase Kp Fast Increase Decrease Minor Deteriorate
Changes
Increase Kd Slow Decrease Minor Changes Decrease Improve
Increase 1/Ti Fast Increase Large Decrease Increase Deteriorate
Fuzzy rules applied in this study for K , K and ∝ are given in Table 7-6 – 7-8. It
should be noted from original work that Ziegler-Nichols PID tuning rule was
compared before selecting values of ∝ which are selected as varying from 2-5 as
single membership function varying from small to big. In all IF-THEN statement
rules, the predecessor portion of the rules consists of two parts joined with “AND”
operators which are inferred by Mamdani product [31]. Considering IF-THEN rules,
the predecessor portion of rules may be written as: μ μ ACE]. μ [∆ACE] where
μ , is the membership function value of the input fuzzy set A for a given value of
ACE, and μ is membership function value of the input fuzzy set B for a given value
of ∆ACE.
330
Table 7-6 Fuzzy Tuning Rules for Kp’
∆ACE
NB NM NS ZO PS PM PB
NB B B B B B B B
ACE NM S B B B B B S
NS S S B B B S S
ZO S S S B S S S
PS S S B B B S S
PM S B B B B B S
PB B B B B B B B
Table 7-7 Fuzzy Tuning Rules for Kd’
∆ACE
NB S S S S S S S
ACE NM B B S S S B B
NS B B B S B S B
ZO B B B B B S B
PS B B B S B S B
PM B B S S S B B
PB S S S S S S S
Table 7-8 Fuzzy Tuning Rules for ∝
∆ACE
NB 2 2 2 2 2 2 2
ACE NM 3 3 2 2 2 3 3
NS 4 3 3 2 3 3 4
ZO 5 4 3 3 3 4 5
PS 4 3 3 2 3 3 4
PM 3 3 2 2 2 3 3
PB 2 2 2 2 2 2 2
331
Figure 7-85 FGSPID Rule Implication Process
Implication process of fuzzy rule is shown in Figure 7-85 where values of K and
K for each rule are determined from corresponding membership functions. Input
membership functions as depicted in Figure 7-85 are such selected that sum of
membership grades for a particular input is unity, i.e., ∑ 1. Fuzzy output of
inference engine is converted to crisp by centroid method of de-fuzzification.
Following output parameters are obtained after de-fuzzification:
K ∑ μ K 7.51
K ∑ μ K 7.52
∝ ∑ μ ∝ 7.53
Where K and K are respective grade values corresponding to the grade μ for the
ith rule. Surface view of obtained parameters gains is shown in Figure 7-86.
After de-fuzzified gain parameters, post processing gain scheduling is performed for
normalizing the K and K between the range of K and K through following
equations:
K K K K ∗K 7.54
332
K K K K ∗K 7.55
K 7.56
∝
The minimum and maximum ranges for process parameter gains are normalized using
suitable normalization factors. The normalized gains are calculated as: K
0.32K , K = 0.6K , K 0.08K T , K = 0.15K T , where K is the
ultimate gain that can make process unstable under proportional feedback control and
T is the ultimate period under stability limit [32]. K is taken as 5.1278 while T is
taken as 2.72 in the current study.
Figure 7-86 FGSPID Gains Surface View
333
Comparison of Fuzzy Gain Scheduled Secondary Frequency Controller with
Various Secondary Controllers
Three area conventional LFC model is utilized with each area implementing hydro
turbine governor dynamics and frequency responsive generic VSWTG model
integrated to area 1. Three area controlled power system model discussed earlier is
utilized such that only area 1 has frequency responsive wind farm and conventional
generating unit of hydro plant represented by hydro turbine governor dynamics while
other areas have hydro plant only. Wind integrated system modified inertia
constant, H is calculated according to equation 7.44. System parameters are given
in the Appendix which is used in LFC simulations. Fuzzy gain scheduled controller
performance is evaluated by performing simulations using MATLAB Simulink
software under the following scenario.
Following load disturbance scenario is used in the current study:
1. 0.1 per unit load disturbance in area 1 only.
2. 0.1 per unit load disturbance in all three area.
3. 10% and 20% wind penetration in Area 1.
Reduced number of rules with five membership functions are applied for the
conventional fuzzy PI controller. Simulation results of fuzzy gain scheduled PID
controller and fuzzy gain scheduled I controller are compared with conventional
fuzzy LFC PI controller with fixed gain and conventional PID LFC controller as
shown in Figure 7-87 - 7-89.
334
Figure 7-87 LFC Controller Comparison with 0.1 Per Unit Load Disturbance in Area 1 only and 90%-
10% Hydro-Wind AGC Participation.
Figure 7-88 LFC Controller Comparison with 0.1 Per Unit Load Disturbance in all 3 Areas and 90%-
10% Hydro-Wind AGC Participation in Area 1.
Figure 7-89 LFC controller Comparison with 0.1 Per Unit Load Disturbance in all 3 Areas and 80%-20%
Hydro-Wind AGC Participation in Area 1. (Zoomed to 40 sec)
335
Gain-scheduled fuzzy controller presents better performance in comparison to other
two controllers when control areas are subjected to different load perturbations and
different wind penetration. These controllers are observed to produce better settling
time and reduced overshoots. In the case of different load disturbance, fuzzy gain
scheduled PID as well as fuzzy gain scheduled I controller shows better performance
than conventional PID and fuzzy controller in terms of better settling time and
reduced overshoots. Though some overshoots are observed in gain scheduled fuzzy
PID, it can be further reduced with optimized gain values which are not considered
in the current study.
Table 7-9 Secondary Frequency Controller Comparison with Performance Index for Area 1 Under 10%
Wind Penetration and 0.1 per Unit Load Disturbance in All Three Areas
Secondary-Frequency Controller ISE ITAE
FGSPID 0.0113 56.82
FGPI 0.015 68.40
PID 0.061 167.52
FGSI 0.021 81.65
Table 7-9 shows performance index values observed through simulation for area 1
with different load frequency controllers. Fuzzy gain scheduled produces the lowest
values among all four controllers for the performance index. Even though fuzzy
controllers show better performance than conventional PID, best results with lowest
ISE and ITAE are observed by using gain scheduling in the fuzzy controller.
336
Figure 7-90 Three Area Frequency Response Comparison with 0.1 Per Unit Load Disturbance and 90%-
10% Hydro-Wind AGC Participation in Area 1. Fuzzy Gain Scheduled LFC Controller is used.
Figure 7-91 Three Area Frequency Response Comparison with 0.1 Per Unit Load Disturbance and 80%-
20% Hydro-Wind AGC Participation in Area 1. Fuzzy Gain Scheduled LFC Controller is used.
Area frequency responses are compared with wind penetration varying from 10-20%
in area 1. It can be seen from Figure 7-90 – 7-91 that wind-integrated area has more
initial frequency deviation with increasing wind penetration. But with frequency
responsive wind plants, frequency settling time is lower in area 1 (< 40 seconds) in
comparison to other two areas (≅60 seconds). Area 1 frequency settles at around 40
sec. in comparison to around 60 seconds for other two area.
337
Case C. Comparison of Fuzzy Gain Scheduled based LFC Controller With and
Without Wind Participation in Secondary Control
Figure 7-92 Area 1 Frequency Response Comparison with 0.1 Per Unit Load Disturbance and Different
Wind Integration Scenario.
Figure 7-92 shows area 1 frequency response with different integrated wind plant
configurations. Highest deviation and longest settling time are observed when wind
plant is just feeding power to control area and smallest deviation with lowest settling
time is observed for control area without wind plant. Grid code responsive wind plant
with automatic generation control (AGC) participation shows the best performance
with frequency response comparable to control area without wind plant with identical
settling time (≅35 seconds). Integration of simplified droop based VSWPP [25]
produces highest frequency deviations and longer settling time ((≅60 seconds). Thus,
the superior performance of grid code frequency responsive WPP with gain scheduled
fuzzy LFC controller is easily acceptable from Figure 7-91.
7.8. Analysis-IV
LVRT Capability of Grid Code Frequency Responsive VSWTG and its Effect
on Frequency Response
338
Besides having maximum power extraction capability with variable speed operation,
type 3 and type 4 variable speed wind turbine generators (VSWTG) have four-
quadrant active and reactive power capability. However, these VSWTGs exhibit a
very sensitive response to electrical network disturbances, predominantly to voltage
sags. Voltage loss leads to magnetization loss and consequent high current flow
through stator and rotor circuits of induction generator based VSWTG. These high
inrush currents can damage power electronics based converters necessitating
disconnection or current bypassing to ride through disturbance and avoid damages.
Voltage disturbance due to grid faults leads to transient overcurrents in rotor circuit
and temporary power control loss in DFIG based VSWTG. Protecting methods are
imperative during transient fault periods to save the system from damage due to
overcurrent or dc overvoltage. Type of protective methods, application duration, and
efficiency, are widely researched topics for DFIG FRT response. [33-39]. In normal
operation, active power support is a priority. Low voltage ride through capability
(LVRT) threshold can vary from 0.8-0.9 p.u. contingent on the wind turbine design.
Normal wind turbine operation is suspended whenever the turbine voltage drops
below this threshold, and reactive current is injected into the grid. Active power
control is a lower priority during this disturbance period, and reference active power
is set to zero [40]. Voluminous literature exists for individual synthetic inertial
support and centralized controlled low-frequency response schemes [41-46] from
power electronics based wind farms during generation-demand mismatch. However,
their proficiency throughout the grid faults is ambiguous owing to the ride-through
capability of power electronics based converter topology during severe disturbances.
While frequency response competencies are typically evaluated as a consequence of
generator outage events in a network, LVRT capability of frequency sensitive type 3
339
VSWTG is evaluated in this section with preliminary findings showing better fault
ride through capability without incorporation of any additional hardware like crow
circuit.
FRT Grid Code Requirement
Figure 7-93 Fault Ride-through Grid Codes [48]
FRT requirements defined in various countries grid codes for connection and
operation of transmission systems is shown in Figure 7-92. The joint consensus of
these grid codes demands stable operation from all new generators under a range of
defined credible grid faults and time duration. As represented in Figure 7-93, voltage
drop level varies between 20%-0% of nominal voltage value while minimum fault
duration during this voltage drop varies from 100 msec (in Denmark) to 625 msec (in
Ireland, USA and Canada. The most stringent requirement for FRT is in Germany
which seeks 0% volts for 150 msec (Germany) while Ireland’s code is challenging
regarding fault duration of 100ms for 15% volt. Except for Australia, technical
standards defined in almost all these countries grid code are either at a fixed level or
a level determined by the transmission system operator. The Australian NEM
Standard for FRT [47] is classified as the automatic standard which requires voltage
ride through capability in the range of 80%-90% of nominal voltage for 10 seconds.
340
Most wind generators cannot meet this and therefore comply with a lesser minimum
standard of 0% volt for 430msec. Generating unit reactive power support to local
voltage is expected during fault period and fault recovery period. Reactive current
injection up to 100% during a fault is also desired in grid codes of countries like
Spain, Germany, and Great Britain.
The test system was investigated for different voltage disturbance range. Simulation
results for two of the most onerous cases are presented here: a 10% voltage enduring
for 150 msec & 450 msec and a 0% voltage enduring for 150 msec. Test cases were
chosen to reflect grid code requirements, as discussed previously. Any
comprehensive low voltage ride through technique must be capable of satisfying
following three points.
 The reduced voltage drop at the generator.
 Divert or cancel high rotor currents below the surge capability of the
converter.
 Produce appropriate power output during faults.
Simulation results for LVRT capability of FR-VSWTG show improvement over
conventional VSWTG in all the three aspects mentioned above.
Case A: 0 Voltage Disturbance
It was observed that frequency responsive VSWTG was unable to sustain 0V
disturbance for 450 msec. and exhibits un-stability in all parameters. On the other
hand, they can sustain 0V disturbance for 150 msec. without any instability.
Case B: 0.1 Voltage Disturbance
Figure 7-94 shows the voltage and current measured at different point of test system
when voltage disturbance of 0.1 V is induced at bus 120 for 150 msec. starting at
341
2 secs. As soon as voltage at point of common coupling drops to 0.1 V , active
power starts decreasing and it is at minimum level but above zero and sustain this
voltage disturbance thereby avoiding a shutdown. A slight variation (4-8% of
nominal dc voltage) of dc link voltage is observed at the start of disturbance and after
clearing of disturbance. Figure 7-95 shows rotor current where we can notice high
current up to 2 p.u. for very small duration around 5 msec. during and after voltage
disturbance while it is around 0.9 p.u. (reference rotor voltage is 0.9 p.u.) during the
rest of time of voltage disturbance.
Figure 7-94 Simulation Results for 0.1 Vnom for 150 msec.
342
Figure 7-95 Rotor Current during Voltage Disturbance
Figure 7-96 Simulation Results for 0.1 v for 450 msec.
343
When 0.1 V appears at PCC at bus 120, around 0.3 p.u. voltage drop is observed
at generator terminal. Figure 7-96 presents the simulation results for 0.1 V voltage
drop for 450 msec. Even though rotor currents are sustained within 0.9 p.u. during
the voltage disturbance duration, active power recovery time is more than the
previous case of 150 msec. Rotor currents increase to 1.8 p.u. for 17 msec. after
removal of voltage disturbance. Even though total currents are under 2 p.u. of
maximum converter limit, additional protection device would be beneficial for this
duration of disturbance. Swift active and reactive power restoration following fault
initiation and clearance is expected from any proposed FRT scheme to improve the
fault response of other locally connected equipment, ultimately contributing to
system frequency stability. Minimum active power during 0 voltage drop for 150
msec is 0.3-0.4 p.u. while it around 1.3 p.u. - 1.4 p.u. for 0.1 V drop for the same
duration. Active power changes from 1.5 p.u. to 0.5 p.u. during voltage disturbance
of 0.1 V for 450 msec. Respective test results are given in Table 7-10.
Table 7-10 Test Response
Voltage drop, Disturbance duration, Time to ramp up active Max. dc link voltage, V
p.u msec power, s
0.1 150 0.02 1240
0.1 450 0.35 1950
0 150 0.02 1270
0 450 Unstable unstable
Figure 7-97 shows the DFIG response during 0.1 V voltage disturbance when it
is not working in frequency controlled mode and rated 1 p.u. power is applied to pitch
compensator. Differences were observed in rotor current, power ramp rate, reactive
power change and rotor speed change during the voltage disturbance.
344
Figure 7-97 Simulation Results for 0.1 v for 150 msec for Normal DFIG Operation
Current at generator terminal is initially more than 2 p.u. and 1 p.u. during voltage
disturbance, while it is 0.9 p.u. with frequency responsive operation. Under similar
test conditions, lower active power is achieved both before and after the disturbance
(varying from 8 p.u. - 6.5 p.u.) in comparison to frequency responsive VSWTG
(delivering 9.5 p.u. before and after disturbance). Reactive power also takes more
time to settle to zero. It is also observed that rotor speed goes into super synchronous
speed mode i.e.; above rated 1.2 p.u. during the disturbance, while it moves into
subsynchronous speed mode when operating in frequency responsive mode. DC link
345
voltage remains the same in both cases. If reference power is set to zero during
voltage disturbance, normal DFIG operation fails to ride through, and imbalance rotor
current and high variation in dc link voltage are observed.
FRT Effect on System Frequency Response
Figure 7-98 FRT Enabled Active Power Control Block
To provide a realistic FRT capable VSWTG model and assess its effect on frequency
response, an FRT block is included in the generic model as shown in Figure 7-98.
This FRT block determines the reduction in active current based on terminal voltage.
During frequency response, active power control is in priority and current is limited
by maximum converter current. During FRT mode, reactive and active currents are
to be regulated to avoid converter overloading. To emulate reduction of active current
during voltage disturbance, P is reduced by multiplying by a factor P
[f V which is depicted in Figure 7-99.
To test the effect of this low voltage / FRT logic enabled VSWTG response on system
frequency response, a test voltage is provided to VSWTG model such that voltage
drops to 0.6 p.u. from 1 p.u. at 11 s and remains at this level till 11.2 s. FRT enabled
VSWTG, and normal VSWTG are integrated into a single control area dominated by
hydro power plant which was defined in chapter 6 earlier. A load disturbance of 0.01
p.u. is applied in the area at 5 s. Figure 7-100 showed the applied voltage and
346
generated current command for normally operated VSWTG while Figure 7-101
represents the same parameters along with PFRT for FRT enabled VSWTG. It can be
seen current and power rises with a decrease in voltage in the normal case while
representing realistic scenario during voltage transient; they decrease as per applied
PFRT factor in FRT enabled VSWTG.
Figure 7-99 PFRT Variation with Terminal Voltage
Figure 7-100 Electrical Parameters in Normal VSWTG during Voltage Disturbance
Figure 7-101 Electrical Parameters in FRT Enabled VSWTG during Voltage Disturbance
347
Figure 7-102 Primary Frequency Response Comparison During 0.01 p.u. load Disturbance
Figure 7-103 AGC frequency response Comparison during 0.01 p.u. Load Disturbance with AGC
Controller Gain as P=-37.47, I=0.0096
Figure 7-104 AGC Frequency Response Comparison during 0.01 p.u. Load Disturbance with AGC
Controller in as P= - 45.63, I = 0.0133
348
Figure 7-102 represents system primary frequency response when the voltage of 1
p.u. is applied to normally operated wind turbine while the voltage drops up to 0.6
p.u. is applied in FRT enabled VSWTG. A drop in frequency can be observed when
electrical power drops due to voltage disturbance. Though the frequency deviation is
small in this case, it might be significant during large amount wind power dropping
during voltage disturbance. Figure 7-103 - 7-104 shows the AGC based frequency
response, and it can be observed that frequency deviation in control area with FRT
enabled VSWTG can aggravate under different AGC gain. This significant drop in
frequency can trigger frequency deviation based relays and might further escalate the
problem. The inclusion of such FRT logic in generic modeling can be useful in
analyzing frequency response situations like South Australian blackout during 28
Sep. 2016, where the number of FRT enabled wind farms reduced their power output
during voltage disturbance and finally situation worsen to the level of complete
system blackout.
Conclusion
The imbalance between energy production and its consumption is best indicated by
frequency in a power system. Primary and secondary regulation comes into the act to
return the balance in the system. Higher wind penetration induces large variations in
system frequency and reduces regulation capability to large contingencies. In this
chapter, it is shown that frequency responsive grid code compatible wind power plant
participation in frequency control services, especially AGC can have a major role in
mitigation of short-term impacts due to variable generation. It is also shown that
application of modern robust, automatic and intelligent control techniques to the
AGC scheme will be more adaptive and appealing approach for the wind-based
349
power system to overcome impositions set up by technical standards of energy market
regulators. Following are the specific conclusions derived in this chapter:
Analysis-I: Synthetic Inertia Enabled Frequency Responsive Wind Plant
Participation Capability in System Frequency Regulation and Its Effects on
System Reserve Response.
Frequency response indicator results from this simulation study and tuning
requirements of system AGC controller suggest that wind participation in frequency
regulation is better in the case of singular services like FCR rather than providing its
services in both control regulation method like FCR and FRR. Besides system inertia
reduction, load change also directly impacts system frequency response. Synthetic
inertia boost power support from VSWTGs and with proper turbine power control
loop modeling, net power output can be limited as per the TSO and demand
requirement under favorable wind conditions. Frequency nadir point as obtained from
different simulations show that no relay tripping is possible with around 40 % wind
penetration in regulatory services.
Analysis-II: Frequency Responsive VSWTG Integration Effect on Multi-
Generational System AGC Control
It can be concluded from this study that a higher wind penetration level up to 40% is
a possibility in a strongly interconnected network and reduced load disturbance while
wind participation in AGC operation under supporting wind conditions can take this
penetration level up to 60%. System planners need to consider the proper system bias
value by inertia reduction and corresponding system equivalent droop to analyze
wind integration impact.
350
Analysis-III: FGSPID Based Intelligent AGC in the Presence of Grid Code
Frequency Responsive Wind Power Penetration
Extensive digital simulations with increased membership values and rule base have
verified the effectiveness of Fuzzy gain scheduled PID controller performance in
comparison to other LFC controllers when applied to a non-linear system containing
stochastic wind power. Gain-scheduled LFC controllers are found to provide
effective compensation for nonlinearities induced due to predictable wind power
variations in frequency control. Despite the promising results, this controller
performance can be improved with optimized gain values and the analysis extended
to analyze other power plants dynamic response characterization for more robust
controller design with improved wind plant technology integration to the system.
Analysis-IV: LVRT Capability of Grid Code Frequency Responsive VSWTG
and Its Effect on Frequency Response
Even though low priority is considered in active power control in conventional DFIG
operation, simulation results show improved LVRT response, when it is under
frequency responsive mode with reference power varying under grid frequency and
TSO limitations. Frequency responsive VSWTG can sustain 0 voltage drop for
150ms which is most stringent FRT requirement among all grid codes. A more detail
analysis must be conducted for concluding and finalizing change in active current
priority threshold in frequency responsive VSWTGs. A generic modeling for
frequency response assessment is also presented for FRT enabled VSWTG in a weak
grid.
351
7.9. References
1. D.P. Kothari and I.J. Nagrath (2003). Modern Power System Analysis, Tata
McGraw Hill, New Delhi.
2. Nanda J., Kaul B. (1978). Automatic generation control of an interconnected
power system, IEE Proceedings 125 (5) 385/ 390.
3. Elgerd O.I., Fosha C.E. (1970). Optimum megawatt frequency control of
multiarea electric energy systems, IEEE Transactions on PAS, PAS-89, 556/563.
4. Hiyama T. (1989). Applications of rule based stabilizing controller to electrical
power systems, IEE Proceedings C 136 (3), 175/ 181.
5. Dash P.K., Liew A.C., Mishra B.R. (1998). An adaptive PID stabilizer for power
systems using fuzzy logic, Electric Power Systems Research 44, 213/222.
6. C. C. Lee. (1990). Fuzzy logic in control systems: Fuzzy logic controller, Part I,
IEEE Trans. Syst., Man, Cybern, vol. SMC-20. pp. 404- 418.
7. M. Sugeno, ed. (1985). Industrial Applications of Fuzzy Control. Amsterdam,
The Netherlands: North-Holland.
8. Takagi T. and Sugeno M. (1985). Fuzzy identification of systems and its
applications to modeling and control, IEEE Trans. Syst., Man, Cybem., vol. 15
pp-116-132.
9. Zhao Z. Y., Tomizuka M. and Sagara S. (1992). A fuzzy tuner for fuzzy logic
controllers, in Proc. 1992 Amer. Control Conf., Chicago, IL, June 24-26, pp.
2268-2272.
10. Mamdani E, Assilian S. (1975). An experiment in linguistic synthesis with a
fuzzy logic controller. International Journal of Man-Machine Studies. 7(1):1-13.
352
11. Karatepe E. Syafaruddin, T. Hiyama (2009). Artificial neural network-polar
coordinated fuzzy controller based maximum power point tracking control under
partially shaded conditions, IET Renew Power Gener, 3, pp. 239–253
12. M. R. I. Sheikh, S. M. Muyeen, R. Takahashi, T. Murata and J. Tamura (2009).
Application of self-tuning FPIC to AGC for load frequency control in multi-area
power system, PowerTech, 2009 IEEE Bucharest, Bucharest, pp. 1-7.
13. M. A. S. Masoum and M. Sarvi (2005). A new fuzzy-based maximum power
point tracker for photovoltaic applications, Iranian Journal of Electrical &
Electronic Engineering, 1:28-35
14. IEEE Working Group on Prime Mover and Energy Supply Models for System
Dynamic Performance Studies, (1992). Hydraulic Turbine and Turbine Control
Models for Dynamic Studies, IEEE® Transactions on Power Systems, Vol. 7,
No. 1, pp. 167-179.
15. P. L. Dandeno, P. Kundur, and J. P. Bayne (1978). Hydraulic Unit Dynamic
Performance Under Normal and Islanding Conditions -- Analysis and
Validation, in IEEE Transactions on Power Apparatus and Systems, vol. PAS-
97, no. 6, pp. 2134-2143.
16. L. M. Hajagos, G. R. Berube (1999). Utility experience with gas turbine testing
and modeling, IEEE Symposium on frequency control requirements, New York.
17. M. Nagpal, A. Moshref, G. K. Morison and P. Kundur (2001). Experience with
Testing and Modelling of Gas Turbines, IEEE Power Engineering Society Entity
Annual Report No.7802-6672-7/01.
18. CIGRE Task Force 25 of Advisory Group 02 of Study Committee C4 (2003).
Modeling of Gas Turbines and Steam Turbines in Combined Cycle Power
Plants, CIGRE, 2003.
353
19. Stanley M. Shinners (1998). Modern Control System Theory and Design, 2nd
Edition, Wiley – Interscience.
20. European Network of Transmission System Operators for Electricity, (2012).
Operational Reserve Ad Hoc Team Report. Belgium: ENTSO-E. Website:
entsoe.eu.
21. N. R. Ullah, T. Thiringer, and D. Karlsson (2008). Temporary Primary
Frequency Control Support by Variable Speed Wind Turbines— Potential and
Applications, IEEE Transactions on Power Systems, vol. 23, no. 2, pp. 601-612.
22. M. L. Chan, R. D. Dunlop and F. Schweppe. (1972). Dynamic Equivalents for
Average System Frequency Behavior Following Major Distribances, IEEE
Transactions on Power Apparatus and Systems, vol. PAS-91, no. 4, pp. 1637-
1642.
23. H. Ahmadi, H. Ghasemi (2014). Security-Constrained Unit Commitment With
Linearized System Frequency Limit Constraints, Power Systems, IEEE
Transactions on, vol. 29, no. 4, pp. 1536–1545.
24. I. M. Cecilio, A. M. Ersdal, D. Fabozzi and N. F. Thornhill (2013). An open-
source educational toolbox for power system frequency control tuning and
optimization. Proceedings of the 4th IEEE PES Innovative Smart Grid
Technologies (ISGT) Europe Conference.
25. A. Aziz, G. M. Shafiullah, A. Stojcevski and A. Mto. (2014). Participation of
DFIG based wind energy system in load frequency control of interconnected
multi generation power system. Power Engineering Conference (AUPEC).
Australasian Universities, Perth, WA, 2014, pp. 1-6.
354
26. H. Chavez, R. Baldick and J. Matevosyan (2015). The joint adequacy of AGC
and primary frequency response in single balancing authority systems, IEEE
Trans. Sustain. Energy, vol. 6, no. 3, pp. 959-966.
27. Kundur, P., Balu, N. and Lauby, M. (1994). Power system stability and control.
New York: McGraw-Hill.
28. G. Andersson (2010). Dynamics, and control of electric power system. EEH –
Power system laboratory.
29. G. Quazza (1996). Noninteracting Controls of Interconnected Electric Power
Systems, IEEE Transactions on Power Apparatus and Systems, vol. PAS-85, no.
7, pp. 727-741.
30. Zhen-Yu Zhao, Masayoshi Tomizuka and Satoru Isaka (1993). Fuzzy gain
scheduling of a PID controllers, IEEE Trans. Syst., Man Cybern., 23 (5).
31. Mamdani, E. H. (1977). Application of fuzzy logic to approximate reasoning
using linguistic synthesis, IEEE Transactions on Computers 26(12): 1182–1191
32. Zhen-Yu Zhao, Masayoshi Tomizuka and Satoru Isaka (1993). Fuzzy gain
scheduling of a PID controllers, IEEE Trans. Syst., Man Cybern., 23 (5).
33. Z. Peng, and H. Yikang (2007). Control strategy of an active crowbar for DFIG
based wind turbine under grid voltage dips, International Conference on
Electrical Machines and Systems, pp. 259-264.
34. L. Peng, B. Francois, and Y. Li. (2009). Improved crowbar control strategy of
DFIG based wind turbines for grid fault ride-through, APEC 2009, pp. 1932-
1938, Washington, USA.
35. I. Erlich, H. Wrede, and C. Feltes (2007). Dynamic behavior of DFIG-based
wind turbines during grid faults, Power Conversion Conference – Nagoya 2007,
pp. 1195-1200, Nagoya, Japan.
355
36. J. Yang, J.E. Fletcher, and J. O'Reilly. (2009). A series dynamic resistor based
converter protection scheme for doubly-fed induction generator during various
fault conditions, IEEE PES GM 2009, Calgary, Canada..
37. J. Lopez, P. Sanchis, X. Roboam, and L. Marroyo (2007). Dynamic behavior of
the doubly-fed induction generator during three-phase voltage dips, IEEE Trans.
Energy Conv., vol. 22, no. 3, pp. 709-717.
38. A.H. Kasem, E.F. El-Saadany, and H.H. El-Tamaly (2008). An improved fault-
ride through strategy for doubly fed induction generator based wind turbines,”
IET RPG, vol. 2, no. 4, pp. 201-214.
39. H.N.D. Le, and S. Islam (2008). Substantial control strategies of DFIG wind
power system during grid transient faults, IEEE/PES T&D, Chicago, USA.
40. K. Okedu (2012). Stability enhancement of DFIG-based variable speed wind
turbine with crowbar by FACTS device as per grid requirement, International
Journal of Renewable Energy Research, vol. 2, pp. 431-439.
41. G. Lalor, A. Mullane, and M. O’Malley (2005). Frequency control and wind
turbine technologies, IEEE Trans. Power Syst., vol. 20, no. 4, pp. 19051913.
42. G. Ramtharan, J.B. Ekanayake, and N. Jenkins (2007). Frequency support from
doubly fed induction generator wind turbines, IET RPG, vol. 1, no. 1, pp. 3-9.
43. L. Holdsworth, J.B. Ekanayake, and N. Jenkins (2004). Power system frequency
response from fixed speed and doubly fed induction generator-based wind
turbines, Wind Energy, vol. 7, no. 1, pp. 21-35.
44. J. Morren, J. Pierik, and S.W.H. de Haan (2006). Inertial response of variable
speed wind turbines, Electric Power Systems Research, vol. 76, no. 11, pp. 980-
987.
356
45. M. Kayikci, and J.V. Milanovic (2009). Dynamic contribution of DFIG-based
wind plants to system frequency disturbances, IEEE Trans. Power Syst., vol. 24,
no. 2, pp. 859-867.
46. R.G. de Almeida, J.A.P. Lopes (2007). Participation of doubly fed induction
wind generators in system frequency regulation, IEEE Trans. Power Syst., vol.
22, no. 3, pp. 944-950.
47. Australian Energy Market Commission (AEMC) (2009). National Electricity
Rules Version 63. Available at aemc.gov.au
48. F. Iov, A. D. Hansen, P. Sørensen and N. A. Cutululis (2007). Mapping of grid
faults and grid codes, Tech. Rep. Risø-R-1617(EN), Risø Nat. Lab., Tech. Univ.
Denmark, Roskilde, Denmark.
357
CHAPTER 8
CONCLUSION AND RECOMMENDATIONS
Power systems around the world are undergoing the continuous shift from centrally
dispatched large scale synchronous generation towards scheduled and non-scheduled
intermittent distributed generation. This change in generation mix due to high
renewable energy penetration challenges the interconnected system designs built
under different network configuration, operational strategies and the regulatory
framework within which it operates. Though wind-based ancillary services like
frequency control ancillary services (FCAS) are still in research & development stage
in most of the countries, future WES participation is expected to contribute to
enhanced market efficiency, improved system reliability and macroeconomic
benefits to all stakeholders.
8.1. Conclusion
Design and operation of power system in presence of wind energy are one of the
major issues in wind power integration. Variable speed wind turbine generators
(VSWTG) having incommensurable dynamic characteristics than conventional
synchronous generators requires precise modeling as they no longer receive negative
load treatment. In view of increasing renewable penetration and compulsory grid
code frequency support, power electronics based VSWTG requires modified active
power control algorithm to support regulation services. This thesis mainly deals with
modelling issues encountered in wind integrated power system for frequency
regulation under the limitations of generic modelling with reduced parameters. This
thesis investigated the grid code compatible frequency response capability from
358
variable speed wind turbines in primary and secondary control. Important modelling
modifications needed for frequency response studies are presented specific to
generating units and system control area. Topics covered in this thesis are frequency
responsive VSWTG modelling especially type 4 VSWTG model, multi-generational
interconnected system AGC control, intelligent AGC control and fault ride through
behavior of frequency responsive wind plant.
8.1.1. Type 4 Generic VSWTG Model for Frequency Response Studies
With relatively new wind technology, especially full converter based type 4 WTG,
development of suitable models to analyze grid code requirements is a significant
subject matter. A simplified yet absolute representation of full converter based wind
turbine generator principles is requisite for various bulk studies. A simulation model
modified from the basic structure of WECC’s type-4 wind turbine generator has been
proposed here which can be utilized in wind penetration impact studies as well as
frequency regulation studies. The performance of this generic model is evaluated
through various simulation sets under different wind speeds and power set-points
combinations and compared with proprietary model and detailed model. Simulation
results show that detail aerodynamic model along with single mass drive train model
gives acceptable performance highlighting their adequacy and best combination in
generic models for frequency control studies. This result contrasts with the simplified
aerodynamic model as proposed by WECC generic model for dynamic studies of
longer time frames. The generator converter effect of designed model is also found
to be comparable with manufacturer specific generic model reducing dependence on
propriety information. Proposed model shows electrical power ramping to 1 p.u.
within 15-20 seconds which is nearly equal to ramping time of 15 seconds as achieved
by detailed WTG model.
359
8.1.2. Grid Code Sensitive, Frequency Processor Dependent Frequency
Controller
Transient low-frequency deviations are handled by inertial control while active power
frequency response controller is needed for high-frequency control. To emulate
synchronous generator like response; dynamic dead-band, speed droop and available
headroom is considered in variable speed wind turbine generator (VSWTG) modelling
for frequency-active power control. The individual wind turbine is made grid code
frequency compatible by including additional active power setpoint generator output
to the modified torque control loop of respective turbine. A modelling framework for
frequency dependent active power setpoint generation in variable speed wind turbines
is presented and its corresponding effects on system frequency regulation response is
investigated. Active power setpoint generator applies designed power limitation on
available wind power, rated turbine power and TSO commanded power and provides
setpoint to turbine. A grid frequency processor based on dynamic dead band and
moving averaged frequency processor is used to suppress noise frequency signals
from passing to active power setpoint generator. Incorporation of frequency processor
helps in avoiding unnecessary changes for frequency response from wind turbines.
Frequency controller is developed for grid codes very similar to British/NEM and Irish
grid. Promising results in terms of improved settling time and better electrical power
and rotor speed variations during frequency deviations were obtained for proposed
frequency responsive type 3 and type 4 VSWTG model in comparison to other
common droop-inertia based model. Proposed grid code compatible VSWTG type 3
model presents setting time reduced to 20-30 seconds from 40-50 s for other
frequency responsive models. Both VSWTG type-3 and type-4 models with
frequency controller-I can maintain electrical power around 1.1 p.u. during the
360
moments of initial frequency deviations while frequency controller 2 presents initial
power drop up to 0.3 p.u. before ramping and settling at around 1.1 p.u. Basic Type
3 and Type 4 VSWTG models calls for torque controller gains re-tuning for improved
participation in frequency response while this can be avoided with proposed
frequency response controllers.
8.1.3. Wind Participation in Classical and Intelligent AGC in Multi-
Generational System
Generic conventional generating unit models representing relevant dynamic features
are developed in MATLAB and numerically validated individually. Automatic
generation control is then investigated for interconnection of these generating
systems of widely different characteristics. Design aspects of AGC scheme for
minimization of system frequency of interconnected multi-generation power systems
with/without the presence of wind power is analyzed. A single area control model
incorporating all three-frequency regulation mode is presented. Control loop of
primary and secondary regulation with specific integration of frequency responsive
VSTG is presented. Performance of AGC for multigenerational power system is
examined, including how frequency bias setting, equivalent droop setting influences
AGC response and inadvertent tie-line interchange and corresponding effects of
inertia change. It is shown through simulation that both type 3 and type 4 VSWTG
can participate in frequency regulation services with suitable modelling
modifications. System frequency control shows better improvement (in terms of
frequency nadir and settling time) when grid code compatible type 3 VSWTG is
integrated to network instead of type 4 VSWTG integration into the network.
Simulation result show that frequency deviation is least affected by system bias value
while it is highly affected by droop variation in AGC. It is also observed that control
361
area having grid code frequency responsive wind integrated with base load plant
gives better AGC performance than the control area having wind integrated with peak
load plant. Simulation result shows that thermal plant gives fastest and best
combination for integration with wind plant followed by gas and then by hydro plant.
In Australian context, thermal or gas based units are more suitable for R6 FCAS in
wind integrated area while hydro plant is better suited at R60 services. Frequency
responsive wind plant participation in FCR and FRR services in a control area
presents improvement in power reserve activation response due to high ramping
capability. With higher participation factor, frequency responsive wind plant with
faster power response provides more power than thermal plant having low
participation factor in a control area AGC.
Modern smart grid with multigenerational integrated power system calls for
computational intelligence techniques for robust and optimum results in AGC
regulation. Fuzzy gain scheduled rule-based proportional integral and derivative
(FGSPID) AGC control system is analyzed for an improved solution for an overall
reduction in grid frequency deviations and settling time in general power systems
operating under restrictions and certain operating ranges. An improvement of settling
time from 60 s to 40 s is observed with FGSPID AGC. Gain-scheduled LFC
controllers are found to provide effective compensation for nonlinearities induced
due to predictable wind power variations in frequency control. Gain scheduling in
fuzzy controller provides lowest performance index values for ISE and ITAE in
comparison to other controllers.
362
8.1.4. Challenges and Possible Solution in Wind Integrated Network
Australia national electricity model is studied for real-time challenges associated with
the large-scale integration of wind plant. Even though NEM-wide challenges are not
identified as each NEM region has a different generation mix, network configuration,
and demand characteristics, which lead to different challenges or different timing;
future increased wind power penetration will bring some adverse operational
challenges in the NEM. 14-Gen NEM model is used to study wind integration effect
on frequency response performance under low load scenario and high load scenario
in area 5 which can be held representative of South Australia (SA). A linear
polynomial regression model for variation between wind penetrations, MW loss,
ROCOF, Frequency nadir is developed. The key finding shows that frequency
operating standard of NEM can easily be maintained for up to 40% of wind
penetration under all load scenario and ROCOF limited to less than 1 Hz/s. Frequency
operating standards management is difficult under wind penetration more than 20%
and high synchronous generator contingency due to increasing ROCOF and higher
frequency nadir. Simulation results show that integration of synchronous condensers
and frequency responsive wind plants can contribute to improvement in system
frequency response.
The results reported in this thesis can be summarized as follows:
 Frequency responsive type 4 model is developed for application in frequency
regulation studies. The accuracy of the developed models has been validated
through the performed case studies and comparison with detailed and a
proprietary model.
363
 It is possible to achieve frequency responsive improved power output from
VSWTGs through a modified torque speed loop and operating in spinning
mode.
 Assessment of grid code sensitive frequency responsive wind plants
integration under different penetration level is performed in a multi-
generational system with improved frequency response in an interconnected
system.
 Integration of wind plants into AGC system can be used in active power
imbalances where wind plant can ramp up and ramp down its power output as
per grid code requirement. Conventional generating units are still required for
system frequency regulation.
 A higher wind penetration level up to 40% is a possibility in a strongly
interconnected network and reduced load disturbance while wind
participation in AGC operation under supporting wind conditions can take
this penetration level up to 60%.
 Frequency Nadir shows improvement if grid code compatible wind plant
participates in either FCR or FRR services than participating in both services.
 Even though low priority is given to active power control in conventional
DFIG operation consider, simulation results show improved LVRT response
when it is under frequency responsive mode with reference power varying
under grid frequency and TSO limitations. Frequency responsive VSWTG
can sustain 0 voltage drop for 150ms which is most stringent FRT requirement
among all grid codes.
364
8.2. Recommendations
The objectives stated for the thesis are achieved within the scope and limitations of
generic modeling for frequency response studies. However, further works can be
done as follows:
 Frequency responsive VSWTG model has been simulated in MATLAB
software. The control strategy can be further implemented and tested in an
experimental setup to study the performance in practice.
 ROCOF values have been obtained under limited parameters of the
interconnected system. A detailed modeling including all system components
is required to quantify realistic ROCOF values.
 Aggregate models of generating units, both conventional and wind plants
have been considered in this study. This study can be extended to spatially
distributed wind farm model and more advanced grid code responsive
conventional generation units representation.
 Economic dispatch process could be implemented in AGC study to achieve
more realistic results.
 Step load disturbance has been considered in this study. Controllable loads
and demand response needs to be implemented in the interconnected system
to study fast frequency response in the presence of frequency responsive wind
plants.
 Even though technical features have been highlighted for wind integrated
power system model, the study is incomplete without economic implications.
Frequency bias criteria are shown to deviate from standard in the presence of
wind penetration. Future work should investigate system frequency bias
criteria for wind integrated power system.
365
APPENDIX
All data related to 14-Gen NEM model is taken from this reference:
M. Gibbard and D. Vowles, (2008). Simplified 14-Generator Model of the SE

Australian Power System. The University of Adelaide, South Australia.
Normal Steady State Analysis Data without Wind Integration
Light Load Case Heavy Load Case
Total Generation MW 15050 23030

Total Load MW 14810 22300
Inter-Area Flows
Area 4-Area 2 MW -200 500
Area 2-Area 1 MW 470 1134
Frequency Grid Processor Parameters
Pre filter Constant : 0.5s, Trend Generator/Filter Constant: 8s,

Upper dead-band: 0.015
Lower Dead-band: 0.001, Post Filter Constant: 0.5s, Droop: 0.0315
Droop Model Parameters
Inertia Droop Model: Low Pass Filter Constant: 0.1s,

K inertia: 2*H wind, High Pass Filter Constant: 1s, K droop: 0.0315
Droop Model 1: Low Pass Filter Constant: 0.01s, Washout Filter Constant: 6s,
Droop: 0.0315
Droop Model 2: Low Pass Filter Constant: 0.2s, K droop: 0.0315
366
VSWTG Parameters (Differences are highlighted in bold)
Type 3 VSWTG (Modified Torque

Type 3 VSWTG (Basic Torque
Loop) with Frequency Response
Loop) with droop controllers
Controller
Aerodynamic Model: Aerodynamic Model:
-Same
,
.
.
c1 = 0.5176, c2 = 116, c3 = 0.4, c4 = 5,

c5 = 21 and c6 = 0.0068.
Pitch Controller: Kpp = 150, Kip = 25, Pitch Controller: Same
Pitch Angle Limitation: 00-270 Angle Limitation: Same
Pitch Compensation: Kpc = 3, Kic = 30, Pitch Compensation: Same Gains
Inputs: PAPC, Pinp. Inputs: Pref=1, Pinp
Torque Controller: Kptrq = 3, Kitrq = Torque Controller: Kptrq: -271.5,

0.6 Kitrq : -310.7
Low Pass filter (TC) Constant: 60s Low Pass filter (TC) Constant: 5s
Power Limitation: 0.04-1.1 p.u. Power Limitation: Same
Washout Filter Constant: 1.0 Washout filter not present.
Gen/Converter Model: Gen/Converter Model:
Maximum current limitation: 1.1 Maximum current limitation: 1.1
Low Pass Filter: 0.02 Low Pass Filter: 0.02
Type 4 basic VSWTG model has all the

Type 4 VSWTG proprietary model has same parameters as that of modified
all the same parameters as that of Type loop based type 4 VSWTG model
3 VSWTG. Electrical Control: Ebst (pu): except Pitch compensator taking
0.2, Kdbr: 10 Pref=1 as input. Torque Controller:
Kptrq = 3, Kitrq = 0.6
367
LFC System parameters used in Single Area LFC simulation
H system T R
6 3.7814 0.0315, 0.05
L (Wind
K P (p.u.)
Penetration)
-3.88 1% 10%
System Parameters for LFC 3-Area model with hydro plant only used in
FGSPID
H D B
6 1 2.1
S P L
5000 MW 0.1 10%
T T T T T T
0.06 0.06 0.08
T 0.2 s T 1 s T 3.7814
R , 0.05 R 0.38 T 5 s
GAST Parameters
Governor Droop R 4%
Fuel system lag time constant 1 T1 1.5
Fuel system lag time constant 2 T2 0.1
Load limiter time constant T3 3
Load limit Lmax 1
Temperature control loop gain KT 1
Maximum value position Vmax 1
Minimum value position Vmin -0.02
Turbine damping Dt 0
368
DCHP Parameters
Parameters Description Variables

ω Reference frequency 1 p.u.
Pre-filter time constant 0.05 s
Governor
Dead-band Dead Band
R Governor droop 4%
maximum power level
FR max 1 p.u.
for frequency response
the minimum power level
FR min 0.04 p.u.
for frequency response
Rate
Power ramp rate 0.014 p.u./s
Limiter
CFM Base load function 1s
Environmental burner
CEV 0.15 s
capacity
Sequential
CSEV Environmental burner 0.25 s
capacity
Variable inlet guide vane
CVGV position compressor 0.60 s
Turbine capacity
Environmental burner
TEV 5s
capacity time constant
Sequential
TSEV Environmental burner 5s
capacity time constant
Un-damped natural
0.22 rad/s
frequency
Damping ratio of the
0.8 p.u.
compressor
Pmax Maximum power 1.1 p.u.
Pmin Minimum Power 0.04 p.u.
Lmax Maximum Load Setpoint 1 p.u.
Lmin Minimum Load Setpoint 0.05 p.u.
369
Non-Reheat Thermal Plant Parameters
K Governor Gain Constant 1.0 s

Governor T Governor Time Constant 0.08 s
R Droop 4%
K Turbine Gain Constant 1.0
Turbine
T Turbine Time Constant 0.4 s
Hydro Plant Parameters
T Filter time constant 0.05 s

T Servo time constant 0.2 s
R Permanent droop 0.05
Governor
R Temporary droop 0.38
Temporary droop
T 5s
time constant
Turbine T Water time constant 1.0
370

Aziz Frequencyregulation 2017

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Aziz Frequencyregulation 2017

Uploaded by

Copyright:

Available Formats

Frequency regulation in wind integrated power system

Downloaded from Deakin University’s Figshare repository

Deakin University CRICOS Provider Code: 00113B

SUBMITTED IN FULFILMENT OF THE REQUIREMENTS FOR THE

I have great pleasure in expressing my deep sense of appreciation and thankfulness

to my principal supervisor Prof. Aman MTO and Co-supervisor Prof. Alex

Stojcevski, for their invaluable guidance, continual encouragement, and support at

encouragement during this research and providing me a valuable learning experience.

Engineering, School of Engineering at Deakin University, Australia. The ﬁnancial

support by Deakin University is gratefully acknowledged.

during candidature proposal.

gratitude to my husband as this accomplishment was not possible without his

continuous help, fruitful technical discussions, and personal advice. I am very

thankful to my both dear brothers, my sisters-in-law, parents-in-law and all family

during last three years.

turbine generator system generic modeling: Variations and applicability,”

Elsevier’s Sustainable Energy Technologies and Assessments, Volume 14,

April 2016, Pages 46-62, ISSN 2213-1388. [IF=1.089, Q1, CiteScore=3.06].

2. Aziz, Asma, Shafiullah, G. M., Amanullah MTO, and Alex Stojcevski.

“Automatic Generation Control in Wind Integrated Power System: New

Perspectives & Challenges.” Int. J. of Thermal & Environmental Engineering,

3. Aziz, Asma, Amanullah MTO, and Alex Stojcevski. "Automatic Generation

Control of Multi-Generation Power System." Journal of Power and Energy

Engineering 2.04 (2014): 312. [Journal Google Scholar h-index=8, Journal

Google impact factor 0.72]

Frequency Control in Wind Integrated Multigenerational Power System.”

Elsevier’s Sustainable Energy Technologies and Assessments (Received

Revision and under process for further submission).

5. Aziz, Asma, Amanullah MTO, and Alex Stojcevski. “LVRT Capability

Assessment of Grid Code Frequency Responsive VSWTG,” International

Journal of Electrical Energy, ISSN: 2301-3656, DOI: 10.18178/ijoee.

Energy Technologies and Assessments (Under review)

Grid Code Frequency Responsive Wind Power Penetration using Intelligent

FGSPID.” Taylor Francis’s Electric Power Components and Systems. (Under

8. Aziz, Asma, Amanullah MTO, and Alex Stojcevski. “Wind Integration In

Australian NEM: Issues and Mitigations” Elsevier’s Renewable Energy

Focus (Under review).

9. Aziz, Asma, Amanullah MTO, and Alex Stojcevski. “An Analysis of

Frequency Responsive Wind Plants Integration on AGC of Multigenerational

System” (To be Submitted)

1. Aziz, Asma, Shafiullah, G. M., Amanullah MTO, and Alex Stojcevski.

“Automatic Generation Control of Wind Energy System,” Proceedings of the

9th Annual International Green Energy Conference (IGEC-IX). 25th–28th

May 2014, Tianjin, China.

2. Aziz, Asma, Amanullah MTO, and Alex Stojcevski. "Multi-Generational

Power System Load Frequency Secondary Control,” 6th Asia-Pacific Power

and Energy Engineering Conference (APPEEC 2014), April 14 to 16, 2014,

3. Aziz, Asma, Shafiullah, G.M., Amanullah MTO, and Alex Stojcevski.

“Participation of DFIG Based Wind Energy System in Load Frequency

Control of Interconnected Multigenerational Power System,” IEEE

Stojcevski. “Power Quality Impacts in a Typical Microgrid,” International

Conference on Sustainable Energy and Environmental Engineering (SEEE

2015), Oct. 2015. DOI: 10.2991/seee-15.2015.20

5. Aziz, Asma, Arangarajan Vinayagam, Amanullah MTO, and Alex Stojcevski.

Model”, 12th IEEE India International Conference, 2015 (INDICON 2015),

of Type 4 Wind Turbine Generator System for Utilization in Frequency

Regulation Studies”, IEEE Australasian Universities Power Engineering

Conference, (AUPEC) 2015, Wollongong, Australia.

of Load Frequency Controller in the Presence of Frequency Responsive Wind

Power Penetration,” IEEE Australasian Universities Power Engineering

Conference, (AUPEC) 2016, Brisbane, Australia.

8. Aziz, Asma, Amanullah MTO, and Alex Stojcevski. “LVRT Capability

Assessment of Grid Code Frequency Responsive TYPE 3 VSWTG”, 9th

International Conference on Computer and Electrical Engineering (ICCEE