Title Page
Abstract
Contents
List of Abbreviations 20
List of Nomenclatures 21
Chapter 1. Introduction 22
1.1. Applications of L/LC-filtered Voltage Source Converters 22
1.1.1. L-filtered VSCs 23
1.1.2. LC-filtered VSCs 26
1.2. Modulation Techniques and Control Methods for L/LC-filtered Voltage Source Converters 27
1.3. Problem Statement 29
1.4. Thesis Contributions and Organization 29
Chapter 2. Fundamental of Finite Control Set Model Predictive Control 32
2.1. Principle of FCS-MPC 32
2.1.1. Construction of Discrete System Model 33
2.1.2. State Variable Prediction and Cost Function Minimization 34
2.2. Practical Implementation in a Digital Platform 35
2.3. Advantages and Disadvantages of FCS-MPC 37
2.4. Summary 38
Chapter 3. An Enhanced Finite Control Set Model-Free Predictive Power Control to Eliminate Stagnant Current Variation Update for L-filtered VSCs 39
3.1. Introduction 40
3.2. FCS-MPPC of L-filtered VSCs 41
3.3. Conventional FCS-MFPPC of L-filtered VSCs 44
3.4. Proposed FCS-MFPPC of L-filtered VSCs 47
3.4.1. Grid Current Variation Analysis 47
3.4.2. Estimation Algorithm 49
3.5. Simulation Results 51
3.5.1. Waveform Quality Comparison 51
3.5.2. Influence of Stagnant Current Variation Update on Prediction Accuracy 54
3.6. Experiment Results 57
3.7. Summary 63
Chapter 4. Grid-Voltage Sensorless Finite Control Set Model-Free Predictive Current Control of L-filtered VSCs with Measurement Noise Suppression 64
4.1. Introduction 65
4.2. Conventional FCS-MFPCC and Its Limitations 67
4.2.1. Conventional FCS-MFPCC of L-filtered VSCs 67
4.2.2. Influence of Measurement Noise on Current Variations when Shortening the Sampling Time 71
4.2.3. Problem of Stagnant Current Variation Update 71
4.2.4. Problem with Employing Virtual Voltage Vectors in FCS-MFPCC 72
4.3. Proposed Grid-Voltage Sensorless FCS-MFPCC with Measurement Noise Suppression 73
4.3.1. Virtual Voltage Vector Construction 73
4.3.2. Stagnant Current Variation Update Elimination and Noise Suppression 74
4.3.3. Grid-Voltage Sensorless Technique 79
4.4. Simulation Results 81
4.4.1. Steady-State Performance Comparison under No Noise Condition 81
4.4.2. Influence of Measurement Noise on Control Performance 83
4.4.3. Grid-Voltage Sensorless Performance 86
4.5. Experimental Results 87
4.6. Summary 94
Chapter 5. Robust Finite Control Set Model Predictive Control of LC-filtered VSCs Against Model Parameter Mismatch and Variation 96
5.1. Introduction 97
5.2. Conventional FCS-MPC of LC-filtered VSCs 98
5.2.1. Principle of Conventional FCS-MPC 98
5.2.2. Effects of Model LC-Filter Parameter Mismatch 101
5.3. Proposed FCS-MPC of LC-filtered VSCs 105
5.3.1. Inductor Current Prediction Based on Old Current Variations 105
5.3.2. Capacitor Voltage Prediction with Least-Squares Minimization 107
5.4. Experimental Verification 110
5.5. Summary 117
Chapter 6. Robust Finite Control Set Model Predictive Control of LC-filtered VSCs with Reduced Current Sensors 118
6.1. Introduction 119
6.2. Problem in Existing Predictive Models 120
6.2.1. Conventional Predictive Model 121
6.2.2. Previous Predictive Model Obtained from Euler Method 121
6.3. Proposed Robust FCS-MPC with Reduced Current Sensors 122
6.3.1. New Predictive Model with Improved Accuracy and Reduced Current Sensors 122
6.3.2. Capacitance Estimator 124
6.3.3. Inductance Estimator 125
6.4. Simulation Results 127
6.5. Experimental Results 131
6.6. Summary 135
Chapter 7. Conclusions and Future Works 136
7.1. Conclusions 136
7.2. Future Works 138
Bibliography 139
Publications 151
Table 3.1. Switching states and voltage vectors of the L-filtered VSC 42
Table 3.2. LUT to store current variations in conventional FCS-MFPPC 46
Table 3.3. Performance comparison under load variations 60
Table 4.1. Voltage vectors of the L-filtered VSC 68
Table 4.2. LUT in conventional FCS-MFPCC 69
Table 4.3. Virtual voltage vector construction 73
Table 4.4. System parameters 81
Table 4.5. Steady-state performance comparison of 5 control methods with different sampling time values 90
Table 4.6. Steady-state performance comparison of the proposed GVS-FCS-MFPCC-V³ with FCS-MPCC-V³ under output power variations (Ts=20µs)[이미지참조] 90
Table 4.7. Comparison of execution time 93
Table 5.1. Switching states and voltage vectors of the LC-filtered VSC 99
Table 5.2. Relationship between inductor-current variations 106
Table 5.3. Stored old inductor-current variations in LUT 107
Table 5.4. System parameters 111
Table 5.5. Performance comparison of the proposed FCS-MPC with the conventional FCS-MPC with load variations 113
Table 6.1. Voltage vectors of the LC-filtered VSC 120
Table 6.2. System parameters 128
Table 6.3. Comparison of Sensor Quanity versus Robustness 134
Table 7.1. Control Strategy Selection for L-filtered VSC 138
Table 7.2. Control Strategy Selection for LC-filtered VSC 138
Figure 1-1. Two-level three-phase L/LC-filtered VSCs 22
Figure 1-2. Grid-connected PV systems (a) Single-stage system [1, 2], (b) Double-stage system 23
Figure 1-3. A battery energy storage system 23
Figure 1-4. Grid-connected wind turbine system 24
Figure 1-5. Active Power Filter System 25
Figure 1-6. STATCOM 25
Figure 1-7. A line-interactive UPS system 26
Figure 1-8. Grid-forming converter 26
Figure 1-9. Motor Drives 27
Figure 2-1. General FCS-MPC scheme 32
Figure 2-2. System model discretization 33
Figure 2-3. Impractical implementation of FCS-MPC 36
Figure 2-4. Practical implementation of FCS-MPC with two-prediction scheme 36
Figure 3-1. Two-level L-filtered VSC 42
Figure 3-2. Control block diagram of FCS-MPPC 44
Figure 3-3. Current variation measurement technique in FCS-MFPPC 45
Figure 3-4. Control block diagram of conventional FCS-MFPPC 46
Figure 3-5. Proposed estimation algorithm 51
Figure 3-6. Control block diagram of the proposed FCS-MFPPC 52
Figure 3-7. Steady-state performance with output voltage reference of 200V (a) FCS-MPPC, (b) Conventional FCS-MFPPC, and (c) Proposed FCS-MFPPC 53
Figure 3-8. Dynamic performance of (a) FCS-MPPC, (b) Conventional FCS-MFPPC, and (c) Proposed FCS-MFPPC 54
Figure 3-9. Prediction error comparison (a) with stagnant current variation update in the conventional FCS-MFPPC, (b) without stagnant current variation update in the proposed FCS-MFPPC 55
Figure 3-10. Comparison of average switching frequency and IGBT losses 56
Figure 3-11. L-filtered VSC prototype in laboratory 57
Figure 3-12. Steady-state performance with output voltage reference of 200V (a) FCS-MPPC, (b) Conventional FCS-MFPPC, and (c) Proposed FCS-MFPPC 58
Figure 3-13. Estimation performance of the proposed FCS-MFPPC 59
Figure 3-14. Prediction performance of (a) Conventional FCS-MFPPC, (b) Proposed FCS-MFPPC 60
Figure 3-15. Dynamic performance of the proposed method when the output voltage reference changes from (a) 200 to 300V, and (b) 300 to 200V 61
Figure 3-16. Dynamic performance of the proposed method when the load changes from (a) 97 to 64.28Ω, and (b) 64.28 to 97Ω 62
Figure 3-17. Experimental comparison when Pref=500W and Qref=0Var (a) FCS-MPPC with Lf and Rf, (b) FCS-MPPC with 0.8Lf and 0.8Rf, (c) (b) FCS-MPPC with...[이미지참조] 62
Figure 4-1. Two-level L-filtered VSC 67
Figure 4-2. Control block diagram of conventional FCS-MFPCC 70
Figure 4-3. Flow chart of conventional FCS-MFPCC 70
Figure 4-4. Measurement technique in conventional FCS-MFPCC and proposed method 71
Figure 4-5. Measurement twice technique in FCS-MFPCC with virtual voltage vector 72
Figure 4-6. Bode plot of the SOGI filter with different gains 78
Figure 4-7. Step response of the SOGI filter with different gains 78
Figure 4-8. Control block diagram of the proposed method 79
Figure 4-9. Flow chart of the proposed method 80
Figure 4-10. Steady-state performance with Ts=30µs under no measurement noise condition, (a) conventional FCS-MFPC, (b) I-FCS-MFPCC, (c) proposed GVS-FCS-MFPCC,...[이미지참조] 82
Figure 4-11. Steady-state performance with Ts=30µs under different measurement noise conditions, (a) conventional MFPCC, (b) I-MFPCC, (c) proposed GVS-MFPCC, (d)...[이미지참조] 84
Figure 4-12. Steady-state performance with Ts=30µs with different sampling time values and fixed measurement noise condition, (a) conventional FCS-MFPCC, (b) I-FCS-...[이미지참조] 85
Figure 4-13. Grid-voltage sensorless performance of the proposed GVS-FCS-MFPCC-V³ 87
Figure 4-14. Experimental setup for verification 87
Figure 4-15. Steady-state performance according to the sampling time (a) conventional FCS-MFPCC, (b) I-FCS-MFPCC, (c) proposed GVS-FCS-MFPCC, (d) FCS-MFPCC-V³, and...[이미지참조] 89
Figure 4-16. Steady-state performance with Ts=20µs (a) FCS-MPCC-V³ with accurate model parameters (e) proposed GVS-FCS-MFPCC-V³ with unknown model parameters[이미지참조] 91
Figure 4-17. Grid-voltage sensorless performance of the proposed GVS-FCS-MFPCC-V³ 91
Figure 4-18. Performance of the proposed GVS-FCS-MFPCC-V³ under distorted grid condition 92
Figure 4-19. Performance of the proposed GVS-FCS-MFPCC-V³ with lagging power factor of 0.7 93
Figure 4-20. Dynamic performance when the amplitude of the grid-current reference changes suddenly from 4.5A to 8A (a) FCS-MPCC-V³, (b) the proposed GVS-FCS-MFPCC-V³ 94
Figure 5-1. Two-level LC-filtered VSC 98
Figure 5-2. Conventional FCS-MPC scheme for the LC-filtered VSC 101
Figure 5-3. Flowchart of conventional FCS-MPC for the LC-filtered VSC 102
Figure 5-4. Tested Loads 102
Figure 5-5. Performance of the conventional FCS-MPC with accurate model LC-filter parameters 102
Figure 5-6. Performance of the conventional FCS-MPC with (a) -50% model inductance, (b) +50% model inductance 103
Figure 5-7. Performance of the conventional FCS-MPC with (a) -50% model capacitance, (b) +50% model capacitance 103
Figure 5-8. Performance of the conventional FCS-MPC with (a) -50% model internal resistance, (b) +50% model internal resistance 104
Figure 5-9. Performance of the conventional FCS-MPC according to various degrees of model parameter mismatch (a) THD, (b) RMSE 104
Figure 5-10. Performance of conventional FCS-MPC with +20% model inductance and +20% model capacitance 105
Figure 5-11. The proposed FCS-MPC scheme 109
Figure 5-12. Flowchart of the proposed method 110
Figure 5-13. Experiment setup 111
Figure 5-14. Capacitor voltage and load current with a linear load (a) Conventional FCS-MPC with accurate model parameters, (b) Proposed FCS-MPC with unknown model parameters 112
Figure 5-15. Capacitor voltage and load current with a nonlinear load (a) Conventional FCS-MPC with accurate model parameters, (b) Proposed FCS-MPC with unknown model parameters 112
Figure 5-16. Prediction performance with (a) Conventional FCS-MPC with accurate model parameters, (b) Proposed FCS-MPC with unknown model parameters 113
Figure 5-17. Dynamic performance of the proposed FCS-MPC when the load changes from ∞ to an R-L load of 20Ω and 15mH 113
Figure 5-18. Performance of the conventional FCS-MPC when the physical filter capacitance changes from 40µF to 20µF. 114
Figure 5-19. Performance of the proposed FCS-MPC when the physical filter capacitance changes from 40µF to 20µF. 114
Figure 5-20. Performance of the conventional FCS-MPC when the physical filter inductance changes from 2.5mH to 5mH. 115
Figure 5-21. Performance of the proposed FCS-MPC when the physical filter inductance changes from 2.5mH to 5mH. 115
Figure 5-22. Performance of the conventional FCS-MPC when the physical filter inductance and capacitance change from 2.5mH to 5mH and from 40µF to 20µF, respectively. 116
Figure 5-23. Performance of the proposed FCS-MPC when the physical filter inductance and capacitance change from 2.5mH to 5mH and from 40µF to 20µF, respectively. 116
Figure 6-1. Two-level LC-filtered VSC 120
Figure 6-2. Typical capacitor current waveform in FCS-MPC 122
Figure 6-3. Proposed FCS-MPC scheme for LC-filtered VSCs 127
Figure 6-4. Tested loads 128
Figure 6-5. Simulation result with a linear load (a) Conventional FCS-MPC with 1.2Lf and 1.2Cf, (b) Proposed FCS-MPC with online identified model parameters[이미지참조] 129
Figure 6-6. Simulation result with a nonlinear load (a) Conventional FCS-MPC with 1.2Lf and 1.2Cf, (b) Proposed FCS-MPC with online identified model parameters[이미지참조] 129
Figure 6-7. Simulated performance with different estimation methods 130
Figure 6-8. Simulated prediction error (a) Conventional predictive model (b) Previous predictive model, (c) Proposed predictive model. 131
Figure 6-9. Experiment setup 131
Figure 6-10. Steady-state performance with a linear load (a) Conventional FCS-MPC with 1.2Lf and 1.2Cf, (b) Proposed FCS-MPC with online identified parameters[이미지참조] 132
Figure 6-11. Steady-state performance with a nonlinear load (a) Conventional FCS-MPC with 1.2Lf and 1.2Cf, (b) Proposed FCS-MPC with online identified parameters[이미지참조] 132
Figure 6-12. Dynamic performance (a) Conventional FCS-MPC with 1.2Lf and 1.2Cf, (b) Proposed FCS-MPC with online identified parameters[이미지참조] 133
Figure 6-13. Prediction error comparison 133
Figure 6-14. Parameter identification performance of the proposed FCS-MPC 134