UNIVERSIDAD DEL PAIS VASCO EUSKAL HERRIKO UNIBERTSITATEA ESCUELA DE INGENIERÍA DE BILBAO BILBOKO INGENIERITZA ESKOLA PhD thesis OPTIMIZED CHARGING CONTROL METHOD FOR PLUG-IN ELECTRIC VEHICLES IN LV DISTRIBUTION NETWORKS PRESENTED BY Mr. Javier García Villalobos SUPERVISED BY Prof. Dr. Ms. Inmaculada Zamora Belver 2016 (cc)2016 JAVIER GARCIA VILLALOBOS (cc by-nc-nd 4.0) AKNOWLEDGEMENT Foremost, I would like to express my gratitude to my advisor Prof. Ms. Inmaculada Zamora for her patience, motivation and immense knowledge. To the department of Electrical Engineering at University of Basque Country (UPV/EHU) for all the support provided. To the whole ERES research group at Technical University of Denmark (DTU) for the opportunity to work with them. Thanks to my colleagues, for the welcome coffee breaks. Finalmente, para toda mi familia y especialmente para mi madre Rocío por su amor y cariño durante todos estos años. Mis éxitos son suyos. “Daily work— my hands’ employment, To complete is pure enjoyment! Let, oh, let me never falter! No! There is no empty dreaming: Lo! These trees, but bare poles seeming, Yet will yield both food and shelter” Goethe INDEX i INDEX OF CONTENTS 1. INTRODUCTION .................................................................................................................................. 3 BACKGROUND .................................................................................................................................... 3 AIMS OF THE THESIS............................................................................................................................. 4 STRUCTURE OF THE DOCTORAL THESIS ...................................................................................................... 4 2. LAND TRANSPORT: TOWARDS ELECTRIC MOBILITY ........................................................................... 7 INTRODUCTION................................................................................................................................... 7 ELECTRIC VEHICLES IN LAND TRANSPORT SECTOR. CHALLENGES AND OPPORTUNITIES ........................................... 7 CURRENT STATE OF ELECTRIC VEHICLES AROUND THE WORLD ......................................................................... 8 PLUG-IN ELECTRIC VEHICLE DRIVETRAIN TOPOLOGIES ................................................................................. 10 Plug-in hybrid vehicles (PHEV) ............................................................................................. 10 Battery electric vehicles (BEV) ............................................................................................. 13 Fuel cell plug-in hybrid electric vehicle ................................................................................. 13 Comparative between analyzed powertrains....................................................................... 14 MAIN COMPONENTS OF PLUG-IN ELECTRIC VEHICLES ................................................................................. 15 Batteries ............................................................................................................................. 15 Battery Management Systems (BMS) .................................................................................. 19 On-board chargers .............................................................................................................. 20 Electric motors and drives ................................................................................................... 21 CONCLUSIONS .................................................................................................................................. 21 3. PEVS IN ELECTRIC DISTRIBUTION NETWORKS .................................................................................. 25 INTRODUCTION................................................................................................................................. 25 CHARGING METHODS ......................................................................................................................... 25 Conductive charging............................................................................................................ 25 Inductive charging............................................................................................................... 29 Battery swapping ................................................................................................................ 30 IMPACT OF PEVS IN ELECTRIC DISTRIBUTION NETWORKS ............................................................................. 30 INFLUENCE OF DRIVING AND CHARGING BEHAVIOR .................................................................................... 31 CLASSIFICATION OF PEVS CHARGING STRATEGIES...................................................................................... 40 OPPORTUNITIES IN THE INTEGRATION OF PEVS ........................................................................................ 44 CONCLUSIONS .................................................................................................................................. 46 4. ACTIVE INTEGRATION OF PEVS IN ELECTRIC DISTRIBUTION NETWORKS ......................................... 51 INTRODUCTION................................................................................................................................. 51 MATHEMATICAL PROGRAMMING .......................................................................................................... 51 Single objective optimization (SOO) ..................................................................................... 52 Multi-objective optimization ............................................................................................... 53 SMART CHARGING ............................................................................................................................ 56 Centralized control architecture .......................................................................................... 57 Decentralized control architecture ...................................................................................... 66 Comparison between centralized and decentralized control approaches ............................. 80 DISTRIBUTED GENERATION AND PLUG-IN ELECTRIC VEHICLES ........................................................................ 81 Virtual power plants (VPP) .................................................................................................. 82 Electric microgrids (MG) ...................................................................................................... 83 PEVS INTEGRATION PROJECTS OVER THE WORLD ...................................................................................... 85 CONCLUSIONS .................................................................................................................................. 87 5. PROPOSED METHODOLOGY ............................................................................................................. 91 INTRODUCTION................................................................................................................................. 91 SIMULATION SETUP ........................................................................................................................... 92 GRID TOPOLOGY AND PEVS MODELLING ................................................................................................. 93 ii INDEX MODEL OF DRIVING AND CHARGING BEHAVIOR........................................................................................ 97 Proposed algorithm to model driving and charging behavior.............................................101 Simulation results ..............................................................................................................104 UNCONTROLLED CHARGING ...............................................................................................................107 SMART CHARGING APPROACHES .........................................................................................................110 Optimization of charging cost ...........................................................................................111 Optimization of PEVs load variance ...................................................................................118 Optimization of overall load variance ................................................................................122 NEW PROPOSED SMART CHARGING METHODOLOGY.................................................................................127 Algorithm methodology ....................................................................................................128 Static weight .....................................................................................................................132 Dynamic weight selection ..................................................................................................146 Voltage unbalance reduction (VUR) ...................................................................................152 Vehicle to Grid ...................................................................................................................157 DSM services provision ......................................................................................................160 Load forecast sensitivity analysis .......................................................................................161 COMPARATIVE ANALYSIS ...................................................................................................................163 SYSTEM ARCHITECTURE .....................................................................................................................168 CONCLUSIONS ...........................................................................................................................171 6. CONCLUSIONS ................................................................................................................................175 CONCLUSIONS OF THE THESIS .............................................................................................................175 FUTURE WORK................................................................................................................................177 REFERENCES.............................................................................................................................................181 ANNEX A – ADDITIONAL DATA OF PEV-PR CASES ....................................................................................193 ANNEX B – FIGURES OF MOO-NF ALGORITHM ........................................................................................197 ANNEX C – FIGURES OF MOO-WF ALGORITHM .......................................................................................201 ANNEX D – FIGURES OF LOAD FORECASTING ERROR ..............................................................................205 INDEX iii INDEX OF FIGURES FIGURE 2.1. WORLDWIDE SALES OF PLUG-IN ELECTRIC VEHICLES .................................................................................. 9 FIGURE 2.2. SERIES HYBRID POWER TRAIN CONFIGURATION ...................................................................................... 11 FIGURE 2.3. PARALLEL HYBRID DRIVETRAIN ARCHITECTURE ....................................................................................... 12 FIGURE 2.4. SCHEME OF A SERIES-PARALLEL HYBRID POWERTRAIN ............................................................................. 12 FIGURE 2.5. BATTERY ELECTRIC VEHICLE BASIC ARCHITECTURE ................................................................................... 13 FIGURE 2.6. SCHEME OF FUEL CELL PLUG-IN ELECTRIC VEHICLE .................................................................................. 14 FIGURE 2.7. EVOLUTION OF ENERGY DENSITY AND BATTERY COST FOR PEVS [11] .......................................................... 15 FIGURE 2.8. CYCLE AGING IN FUNCTION OF DOD FOR LI-ION BATTERIES ...................................................................... 16 FIGURE 2.9. DIFFERENT TYPES OF ENCAPSULATED CELLS: (A) POUCH (B) PRISMATIC AND (C) CYLINDER ................................ 18 FIGURE 2.10. CONSTANT CURRENT – CONSTANT VOLTAGE (CC-CV) METHOD FOR A LI-ION CELL ...................................... 20 FIGURE 3.1. DIFFERENT CHARGING MODES ACCORDING TO THE IEC 61851-1 STANDARD. (A) MODE 1 (B) MODE 2 (C) MODE 3 AND (D) MODE 4 .................................................................................................................................... 27 FIGURE 3.2. CCS TYPE 1 AND CCS TYPE 2 COMBO RECEPTACLES ............................................................................... 29 FIGURE 3.3. DAILY DISTANCE TRAVELLED FOR THE IRELAND DEMO REGION OF GREEN EMOTION [37] ................................ 33 FIGURE 3.4. TRIP DISTANCE DISTRIBUTION FOR THE IRELAND DEMO REGION OF GREEN EMOTION [37] .............................. 33 FIGURE 3.5. ENERGY CONSUMPTION PER KM IN NORTHERN DEMO REGION [36] ........................................................... 34 FIGURE 3.6. DISTRIBUTION OF NUMBER OF CHARGE EVENTS PER DAY FOR BEVS [28] ..................................................... 35 FIGURE 3.7. AVERAGE DAILY CHARGING FREQUENCY FOR PHEVS [15]........................................................................ 36 FIGURE 3.8. ENERGY PER CHARGE EVENT FOR DIFFERENT EVSE LOCATIONS: (A) HOUSEHOLD (B) OFFICE (C) PUBLIC PARKING AND (D) STREET ............................................................................................................................................. 37 FIGURE 3.9. ENERGY DEMAND PER CHARGE EVENT AT HOUSEHOLD LOCATION (GREEN EMOTION) ..................................... 38 FIGURE 3.10. ENERGY DEMAND PER CHARGE EVENT IN NASHVILLE (THE EV PROJECT) .................................................... 38 FIGURE 3.11. ENERGY DEMAND PER CHARGE EVENT IN SAN FRANCISCO (THE EV PROJECT) ............................................. 38 FIGURE 3.12. DISTRIBUTION OF INITIAL SOC IN THE EV PROJECT .............................................................................. 39 FIGURE 3.13. DISTRIBUTION OF FINAL SOC IN THE EV PROJECT ................................................................................ 39 FIGURE 3.14. CLASSIFICATION OF PEVS INTEGRATION METHODS ............................................................................... 41 FIGURE 3.15. EXAMPLE OF DIFFERENT TOU TARIFFS IN SPAIN .................................................................................. 42 FIGURE 3.16. ADVANTAGES AND DRAWBACKS OF DIFFERENT STRATEGIES FOR PEVS INTEGRATION .................................... 43 FIGURE 4.1. OPTIMIZATION OF A LINEAR PROBLEM WITH TWO VARIABLES AND SIX CONSTRAINTS ....................................... 53 FIGURE 4.2. PARETO FRONTIER OF A BI-OBJECTIVE OPTIMIZATION .............................................................................. 54 FIGURE 4.3. CENTRALIZED CONTROL ARCHITECTURE [65], [69] ................................................................................. 58 FIGURE 4.4. INFORMATION REQUIRED FOR THE AGGREGATOR OPERATION IN A CENTRALIZED ARCHITECTURE ........................ 58 FIGURE 4.5. INTERACTIONS BETWEEN DIFFERENT ELEMENTS IN THE CENTRALIZED CONTROL .............................................. 59 FIGURE 4.6. DECENTRALIZED CONTROL ARCHITECTURE USING PRICE SIGNALS [69] ......................................................... 67 FIGURE 4.7. FREQUENCY DROOP CONTROL WITH V2G [45] ..................................................................................... 68 FIGURE 4.8. APPLICATION OF MOBILE AGENT CONCEPT FOR PEVS .............................................................................. 78 FIGURE 4.9. TECHNOLOGIES THAT CAN BE INTEGRATED INTO A VPP ........................................................................... 82 FIGURE 4.10. ARCHITECTURE OF A MICROGRID WITH PEVS ...................................................................................... 84 FIGURE 5.1. SMART CHARGING APPROACHES ANALYZED AND PROPOSED IN THIS THESIS ................................................... 92 FIGURE 5.2. SIMULATION SETUP ........................................................................................................................ 93 FIGURE 5.3 SINGLE PHASE DIAGRAM OF THE MODELLED DISTRIBUTION NETWORK .......................................................... 93 FIGURE 5.4 GIS MAP OF THE PART OF BORUP DISTRIBUTION NETWORK USED IN THIS THESIS ............................................. 94 FIGURE 5.5 SATELLITE OVERVIEW OF THE PART OF BORUP TOWN, WHERE THE DISTRIBUTION NETWORK USED IN THIS THESIS IS LOCATED ............................................................................................................................................... 94 FIGURE 5.6. NO PEVS CASE: (A) DISTRIBUTION TRANSFORMER LOAD AND (B) LINE-NEUTRAL VOLTAGES AT NODE 613 .......... 96 FIGURE 5.7. MODEL OF THE CURRENT SOURCE FOR PEV CHARACTERIZATION................................................................ 97 FIGURE 5.8. DISTRIBUTION OF NUMBER OF VEHICLES PER HOUSE ............................................................................... 98 FIGURE 5.9. DISTRIBUTION OF DEPARTURE TIME IN WEEKDAYS .................................................................................. 99 FIGURE 5.10. DISTRIBUTION OF DEPARTURE TIME IN WEEKEND DAYS .......................................................................... 99 FIGURE 5.11. DISTRIBUTION OF ARRIVE TIME IN WEEKDAYS .................................................................................... 100 iv INDEX FIGURE 5.12. DISTRIBUTION OF ARRIVE TIME IN WEEKEND DAYS ..............................................................................100 FIGURE 5.13. DISTRIBUTION OF DISTANCE TRAVELLED IN FUNCTION OF TRAVEL TIME FOR WEEKDAYS ................................101 FIGURE 5.14. DISTRIBUTION OF DISTANCE TRAVELLED IN FUNCTION OF TRAVEL TIME FOR WEEKEND DAYS ..........................101 FIGURE 5.15. FLOW CHART OF THE PROPOSED ALGORITHM ....................................................................................102 FIGURE 5.16. INVERSE CDF OF GEV DISTRIBUTION FITTED TO DEPARTURE TIME DATA ..................................................103 FIGURE 5.17. INVERSE CDF OF WEIBULL DISTRIBUTION FITTED TO TRIP DISTANCE DATA ................................................104 FIGURE 5.18. PEVS POWER DEMAND FOR 10, 30, 50 AND 70% ............................................................................105 FIGURE 5.19. PEVS POWER DEMAND PROFILE IN CASE 2........................................................................................105 FIGURE 5.20. PEVS POWER DEMAND IN CASE 3...................................................................................................106 FIGURE 5.21. PEVS POWER DEMAND PROFILE IN CASE 4........................................................................................107 FIGURE 5.22. UNCONTROLLED CHARGING FOR 30% OF PEV-PR. (A) TOTAL LOAD OF DISTRIBUTION TRANSFORMER AND (B) LINE-NEUTRAL VOLTAGES AT NODE 613 ......................................................................................................109 FIGURE 5.23. UNCONTROLLED CHARGING FOR 30% OF PEV-PR. (A) PEVS POWER DEMAND COMPARED TO ELECTRICITY COST AND (B) EVOLUTION OF SOC OF EACH PEV .................................................................................................109 FIGURE 5.24. OPTIMIZATION OF CHARGING COST (NO V2G CASE) FOR 30% OF PEV-PR. (A) LOAD IN THE DISTRIBUTION TRANSFORMER AND (B) LINE-NEUTRAL VOLTAGES AT NODE 613 .......................................................................114 FIGURE 5.25. OPTIMIZATION OF CHARGING COST (NO V2G CASE) FOR 30% OF PEV-PR. (A) PEVS LOAD COMPARED TO ELECTRICITY COST AND (B) EVOLUTION OF SOC OF EACH PEV ..........................................................................115 FIGURE 5.26. OPTIMIZATION OF CHARGING COST (V2G CASE) FOR 30% OF PEV-PR. (A) LOAD IN THE DISTRIBUTION TRANSFORMER AND (B) VOLTAGES AT NODE 613 ..........................................................................................116 FIGURE 5.27. OPTIMIZATION OF CHARGING COST (V2G CASE) FOR 30% OF PEV-PR. (A) PEVS LOAD COMPARED TO ELECTRICITY COST AND (B) EVOLUTION OF SOC OF EACH PEV ..........................................................................116 FIGURE 5.28. PEVS POWER DEMAND AT DIFFERENT Z VALUES FOR PEV-PR OF 30% ...................................................117 FIGURE 5.29. OVERALL PEVS CHARGING COST AT DIFFERENT Z VALUES FOR PEV-PR OF 30% ........................................117 FIGURE 5.30. OPTIMIZATION OF PEVS LOAD VARIANCE (NO V2G CASE) FOR 30% OF PEV-PR. (A) LOAD IN THE DISTRIBUTION TRANSFORMER AND (B) LINE-NEUTRAL VOLTAGES AT NODE 613 .......................................................................120 FIGURE 5.31. OPTIMIZATION OF PEVS LOAD VARIANCE (NO V2G CASE) FOR 30% OF PEV-PR. (A) PEVS POWER DEMAND VERSUS ELECTRICITY COST AND (B) EVOLUTION OF THE SOC OF EACH PEV ..........................................................120 FIGURE 5.32. OPTIMIZATION OF PEVS LOAD VARIANCE (V2G) FOR 30% OF PEV-PR. (A) LOAD IN THE DISTRIBUTION TRANSFORMER AND (B) LINE-NEUTRAL VOLTAGES AT NODE 613 .......................................................................121 FIGURE 5.33. OPTIMIZATION OF PEVS LOAD VARIANCE (V2G CASE) FOR 30% OF PEV-PR. (A) PEVS POWER DEMAND VERSUS ELECTRICITY COST AND (B) EVOLUTION OF THE SOC OF EACH PEV ....................................................................122 FIGURE 5.34. OPTIMIZATION OF OVERALL LOAD VARIANCE (NO V2G CASE). (A) LOAD IN THE DISTRIBUTION TRANSFORMER AND (B) VOLTAGES AT NODE 613 ....................................................................................................................123 FIGURE 5.35. OPTIMIZATION OF OVERALL LOAD VARIANCE (NO V2G CASE). (A) PEVS POWER DEMAND VERSUS LOAD FORECAST AND (B) EVOLUTION OF THE SOC OF EACH PEV ............................................................................................124 FIGURE 5.36. OPTIMIZATION OF OVERALL LOAD VARIANCE (V2G CASE). (A) LOAD IN THE DISTRIBUTION TRANSFORMER AND (B) VOLTAGES AT NODE 613 .........................................................................................................................125 FIGURE 5.37. OPTIMIZATION OF OVERALL LOAD VARIANCE (V2G CASE). (A) PEVS POWER DEMAND VERSUS LOAD FORECAST AND (B) EVOLUTION OF THE SOC OF EACH PEV ............................................................................................125 FIGURE 5.38. OVERALL LOAD VARIANCE AND TOTAL CHARGING COST, AT DIFFERENT Z VALUES ........................................126 FIGURE 5.39. PEVS POWER DEMAND AT DIFFERENT Z VALUES ................................................................................126 FIGURE 5.40. EVOLUTION OF CHARGING COST AND PEVS POWER DEMAND VARIANCE IN FUNCTION OF WEIGHT U FOR THE MOO-NF ALGORITHM: (A) 30% OF PEV-PR AND (B) 100% OF PEV-PR .........................................................136 FIGURE 5.41. EVOLUTION OF OVERALL LOAD VARIANCE FOR THE MOO-NF ALGORITHM: (A) 30% OF PEV-PR AND (B) 100% OF PEV-PR .........................................................................................................................................137 FIGURE 5.42. MOO-NF CASE WITH U=0.2: (A) LOAD IN THE DISTRIBUTION TRANSFORMER AND (B) LINE-NEUTRAL VOLTAGES AT NODE 613 ...........................................................................................................................................137 FIGURE 5.43. MOO-NF CASE WITH U=0.2: (A) PEVS POWER DEMAND AND ELECTRICITY COST AND (B) EVOLUTION OF SOC OF EACH PEV ...........................................................................................................................................138 FIGURE 5.44. MOO-NF CASE WITH U=0.8: (A) LOAD IN THE DISTRIBUTION TRANSFORMER AND (B) LINE-NEUTRAL VOLTAGES AT NODE 613 ...........................................................................................................................................138 FIGURE 5.45. MOO-NF CASE WITH U=0.8: (A) PEVS POWER DEMAND AND ELECTRICITY COST AND (B) EVOLUTION OF SOC OF EACH PEV ...........................................................................................................................................138 FIGURE 5.46. MOO-NF CASE. PEVS POWER DEMAND FOR DIFFERENT VALUES OF U, WITH 30% OF PEVS PENETRATION RATE ..................................................................................................................................................139