MODELING, SIMULATION, AND APPLICATION OF A DOUBLYFED RELUCTANCE ELECTRICAL MACHINE
Zhiqing Wu
Doctor of Philosophy in Engineering
Brushless doublyfed (dualwinding, mixedpole) machines have received renewed attention in the last few years for use in adjustablespeed drives and variable speed generator systems. This class of machine has two threephase stator windings wound for different pole numbers and a cage or a reluctance rotor. It appears to be highly attractive due to its structural simplicity, high efficiency, lower manufacturing cost, and its ability to operate in induction and synchronous machine modes with the possibility of sub and supersynchronous speeds operations. This dissertation presents an accurate model of the doublyfed reluctance machine, which considers the coreloss of the machine. The qd equivalent circuits with series coreloss resistances or shunt coreloss resistances are given. The analysis and performance characterizations of a few systems with this machine are set forth. These systems are: (1) Standalone doublyfed reluctance generator system with capacitive excitation in both the power and control windings. (2) Doublyfed synchronous generator system with DC current excitation in control windings and threephase impedance load or a loaded threephase diode rectifier in power windings. (3) Doublyfed synchronous reluctance generator systems with controlled DC output voltage. DCDC buck converter or boost converter is used as DC voltage regulator. (4) Fieldorientation doublyfed motor control system that can run in a wide speed range. IP controller, Inputoutput linearization controller, voltagecontrolled voltage source PWM inverter, and currentcontrolled voltage source PWM inverter are used in two control schemes. Dynamic simulation, steadystate calculation, and experimental measurement are used to reveal its potential application.
VITA
Mr. Zhiqing Wu was born in Guangdong, People’s Republic of China, on December 14, 1964. He attended Zhengzhou Light Industrial Institute of Technology and received a degree of Bachelor of Science in Electrical Engineering in July 1985. He entered the graduate school of Guangdong University of Technology in September 1987 and received a Master of Science degree in Electrical Engineering in July 1989. He started his Ph.D. program at Tennessee Technological University in January 1994 and received a Doctor of Philosophy Degree in Engineering in December 1998. From 19851987, he worked as an electrical engineer in New Star Industry Company, GuangZhou, China. From 19891993, he joined Guangdong Province Technology Exchange Center as a project engineer in GuangZhou, China. He worked as a Research Assistant while he was a graduate student at Tennessee Technological University.
MODELING, SIMULATION, AND APPLICATION OF A DOUBLYFED RELUCTANCE ELECTRICAL MACHINE
______________________________________
A Dissertation
Presented to the Faculty of the Graduate School
Tennessee Technological University by Zhiqing Wu
_______________________
In Partial Fulfillment
of the Requirements for the Degree DOCTOR OF PHILOSOPHY
Engineering
_________________________
December 1998
CERTIFICATE OF APPROVAL OF DISSERTATION
MODELING, SIMULATION, AND APPLICATION OF A DOUBLYFED RELUCTANCE ELECTRICAL MACHINE
by Zhiqing Wu Graduate Advisory Committee: ________________________ Chairperson ________________________ Member ________________________ Member ________________________ Member ________________________ Member ________________________ Member ________________________ Member ________ date ________ date ________ date ________ date ________ date ________ date ________ date Approved for the Faculty:
Dean of Graduate Studies Date
DEDICATION
This dissertation is dedicated to my parents and sisters
ACKNOWLEDGMENTS
I would like to express my appreciation to Dr. Olorunfemi Ojo, my major professor and chairman of the advisory committee, for his valuable direction and assistance. I would also like to thank Dr. Marie B. Ventrice a member of my advisor committee for her encouragement and assistance during this work. I also want to thank my other advisory committee members: Dr. A. Chandrasekaran, Dr. Brian M. O’Connor, Dr. Ahmad Smaili, Dr. Charles E. Hickman, and Dr. Ghadir Radman for their assistance. I express my gratitude to Dr. P. K. Rajan and Dr. Ken Purdy for their assistance. I am grateful to Obasohan I. Omozusi, John Cox, Sandy Garrison, Helen Knott, and Keith Jones for their support. I acknowledge the financial support of the Center for Electric Power at Tennessee Technological University for this research.
TABLE OF CONTENTS
Page
LIST OF SYMBOLS ............................................................................................. viii LIST OF TABLES................................................................................................... xi
............................................................................... Stator and Rotor Structure of the Experimental Machine .....................3.......................................................................................................................................45 3......................48 3.............................2.........11..... Research Motivation .1 1........................................................ Introduction .72 4......................... Calculation of Electromagnetic Torque ....................3....24 3...........................................22 3...............27 3.................41 3..............................13 3........... 73 ...............................................................1 1........... ManleyRowe Relationships.........38 3.......... Steadystate Equivalent Circuits .................................2.............................................. 3..............4.. Machine Model .....13...................... Transient and Oscillatory Behavior ............ Calculation of Machine Parameters ..............................6.....7.....................................1......................................................................................................................33 3...........LIST OF FIGURES ...............................................................57 Chapter Page 3.................................... Including Saturation Effects ........................22 3....................xii Chapter 1..............................12..........................5.......68 3.............................62 3................................. Proposed Research.10.......................... Modeling of a DoublyFed Reluctance Machine ................................. Synchronous motor .......... Research Background ......... Performance Characteristics of DoublyFed Reluctance Generator ...................10 2................................8...................1...............................................................................4 1................................................................ Some Design Aspects.............. Calculation of Inductance Using Winding Function Theory........ Literature Survey ..........9...................... Conclusion . Introduction ......
.... 141 6..86 4. Introduction..................................... Generator with Impedance Load...2...............................................................2....2...............1.... Performance of a Doublyfed Synchronous Reluctance Generator with Controlled DC Voltage................................ 162 ..... 154 Chapter Page 6............................5...........74 4......................... 128 6................ 105 5..... Conclusion ..............................104 5.... Steadystate Results .............................................5.6................... Simulation of Generator System ...........................................................................5..................................79 4..... Introduction................................1........................ 130 6.................................4......... Introduction............... Description and Modeling of Generator Systems.......................................... 161 7..4.......................................................................1....................................... Steadystate Characteristics of the Generator Systems ...........................3........................................................3................................ Introduction..3........ 160 7............ 102 5.... 129 6....4............................129 6................................7............................................................................... Field Orientation Principle ... Model of Generator Loads............ Machine Model .. 104 5.................................... Conclusion ..................................92 4.................... 125 5................................................2............... Synchronous Operation of a Doubly Fed Reluctance Generator... Conclusion .......................................................................................1....................................... Power Capability and Parametric Analysis ......................................... Simulation of SelfExcitation Process............................ 114 5.....................4.................. Field Orientation Control of a Doubly Fed Synchronous Reluctance Machine................................................... Generator with Rectifier Load ................................... Steadystate Generator Model ............. 161 7......................95 4......................................................................73 4.....
....................7.......................... 217 8.......................................................6........................ 220 APPENDIX .................... The Description of Two Control Schemes .................................................................................................................................... q = Pole pair number of the power and control windings..............4....... Conclusion ...................... 274 LIST OF SYMBOLS p1................................... respectively Pr = Pole pair number of the salient rotor structure rp.........2............................... Simulation Results Below the Base Speed ...................................... Ias = Power winding phase ‘A’ current and control winding phase ‘a’ current. Nonlinear Controller Design.............................................................................. 219 REFERENCES ............. respectively IAp...... 203 7................................... 173 7.......1....10........ 199 7....................... 188 7............... 197 7.................... 228 VITA.............................7.............................................. 180 7.....8..................................................................... rs = Perphase winding resistance of the power and control windings........9................................... respectively .............. Simulation Results Above the Base Speed.................... Suggestion for Further Work . Conclusion ....................... 215 8...... Voltage Source SPWMInverter .. Design of the IP Controller..........................3............................................... Conclusion and Suggestion for Further Work........5.............................................................. 175 7..... 217 8......... Current Control in Voltage Source Inverters.......
ICr = Phase currents flowing into the threephase diode rectifier Sav. R’ms = Perphase coreloss resistances for the power winding and control windings. ωrm = Electrical and mechanical angular rotor speed θp = ωpt. IBr. respectively Rmp. θs = ωst . Sbv. l = Radius of the rotor and the effective length of the machine. perphase for the power and control windings. θr = ωrt Np.IAo = Phase ‘A’ load current IAr. Ns = Effective turns perpole. respectively . ωs = Angular frequency of the power and control winding supply voltages. SbI. Vds = Control winding terminal magnitudes in the qd reference frame V’qs. Vas = Power winding phase ‘A’ voltage and control winding phase ‘a’ voltage. Scv = Voltage switching functions of the three legs of the rectifier SaI. respectively r. Vdp = Power winding terminal voltage magnitudes in the qd reference frame Vqs. respectively λAp. respectively ωr. ScI = Current switching function of the rectifier Id = Filter inductor current VAp. respectively Vqp. Vqds = Qd complex form terminal voltage magnitudes of the power and control winding. Idp = Q and d power winding currents. respectively V’qds = Referred qd complex form control winding voltage ωp. λas = Power winding phase ‘A’ flux linkage and control winding phase ‘a’ flux linkage. V’ds = Referred control winding qd terminal voltage Vqdp. respectively α = Rotor pitch factor Iqp.
respectively Iqo. λds = Q and d control winding flux linkage. respectively Lp. λdp = Q and d power winding flux linkage. Ids = Q and d control winding currents. Ido = Q and d load currents. respectively I’qds = Referred qd complex form control winding current Iqdo = Complex form load current Io = Load current of the rectifier circuit λqp. Laa = Selfinductance of the power and control windings. λdm = Q and d magnetizing flux linkage. respectively λ’qs. respectively Iqdp. respectively . respectively I’qs. respectively λm = Magnitude of the magnetizing flux linkage Cp. respectively λqm. between power and control winding. k = A. LAB. j = a. λ’ds = Referred q and d control winding flux linkage. Lm = Referred control winding selfinductance and mutual inductance.k. respectively LBC. LAC = Mutual inductances between phases of the power windings Lbc. Cq = Perphase shunt capacitor connected across the power and control winding terminals C’q = Referred perphase shunt capacitor connected to the control windings LAA. I’ds = Referred q and d control winding currents. respectively Lm12 = Mutual inductance between the power and control windings L’s. B. = Mutual inductances between the power and control winding phases L1. L2 = Perphase leakage inductance of the power and control winding. c.Iqs. b. Lab. Lac = Mutual inductnces between phases of the control windings Lj. C. Ls = Selfinductance of the power and control winding. respectively λqs. Iqds = Qd complex form power and control winding current.
... respectively p = d/dt D = duty ratio LIST OF TABLES Table Page 7....... 194 .........Llp = Lp – Lm = Perphase effective leakage inductance of the control winding L’ls = L’s – Lm = Perphase effectively referred leakage inductance of the control windings Ro..............................2........................ Lo = Load resistance and inductance..........1.. Ld = Rectifier filter capacitor and inductance.......................... respectively Vcd = Rectifier filter capacitor voltage Id = Rectifier inductor current Cd........ State sequence of each sector .. 191 7.................. Inverter states and space vector .
........5.............................. Twoaxis dynamic model for the cage rotor machine................................................3..........3 1..2 1......................2.............. Schematic of BDFMASD................................ Doublyfed Reluctance Machines .........7 1.......4......5 1...............8 .... Steadystate equivalent circuit .....................................LIST OF FIGURES Figure Page 1...... Qd equivalent circuit of salient rotor machine ...............................1.......................
6.........................................47 3....28 3.................12..29 3.........28 3......... Conventional slip power recovery system.....................................1..............44 Figure Page 3......................6. Effective leakage factors for values of αm and ratio of effective turns per pole of the power winding.......................4..............5............................51 3......... Machine parameters using steadystate equivalent circuit with shunt coreloss resistances...... Machine parameters using steadystate equivalent circuits with series coreloss resistances .42 3............................ Doublyfed reluctant synchronous generator systems ...... Fundamental components of the winding functions . Qd equivalent circuit with shunt coreloss resistances ..............40 3...................4..1............... Regulated DC power generation systems using doublyfed synchronous reluctance machine ..........................5.........................1..............................................................................................11...16 2.............17 2........9...............7...2..................................................2.................23 3.14 2..........15 2. Qd complex equivalent circuit of doubly fed reluctance machine ..... Nps = Np/Ns .10.........................8..............56 ................................................................................ Winding function for 6pole phase A .. Steadystate equivalent circuit with series coreloss resistances ......... Winding function for 2pole phase a............ Qd equivalent circuits with series coreloss resistances .....3................ Steadystate equivalent circuit with shunt coreloss resistances .......... Approximate steadystate equivalent circuit for the parameter calculation.......3..47 3...............9 2.........52 3. Stator and rotor structure of the experimental machine ........................... Selfexcited doublyfed reluctance generator systems..20 3........ Field orientation control systems for a doublyfed synchronous reluctance machine..........
.............71 4......... Measured waveforms of generator system feeding a rectifier load .........................................66 3.............................93 4. Noload transient of rotor speed and torque when inverter frequency = 15 Hz....................................2........................4.......... Selfexcitation process of doublyfed reluctance generator ..........................................88 Figure Page 4.13.....60 3......14..............67 3......... Schematic diagram of the doublyfed synchronous reluctance motor with DC excitation ..............15..................... Effective resistance RL ........87 4............................6............ Schematic diagram of standalone generator systems ............79 4..........67 3..............................................3.. Experimental waveforms showing machine oscillatory motion with inverter frequency=15Hz ......... Measured and calculated results of generator system feeding a loaded rectifier (ωrm = 825 rev/min and ωrm = 750 rev/min)........................................................................90 4............. Parametric characteristics of generator feeding a resistive load rotor speed = 900 rev/min ................75 4....8................. Doublyfed synchronous reluctance motor characteristics ............................................. Noload transient of rotor speed and torque when inverter frequency = 10 Hz............ 101 ..66 3................................................16.. Airgap power contribution of power (Pp) and control (Ps) windings normalized with developed mechanical power as a function of control winding frequency .. Experimental waveforms showing machine oscillatory motion with inverter frequency=10Hz .19..............................................................1.................18........................................7....89 4.............9..........3.. Measured and calculated results of doublyfed reluctance generator system feeding impedance loads .............................. load resistance 20Ω) (10ms/div) ................................................. Calculated generator characteristics of selfexciting generator system feeding impedance load....................68 3...............17.5......................... Measured generator waveforms (rotor speed =850 rev/min......................91 4....
........ 126 5....................................... Experimental waveforms of selfexcitation process . Schematic diagram of the doublyfed reluctance machine with DC excitation ........3................ 121 5................ Schematic diagram of the doublyfed synchronous generator systems ...................9................ Calculated and measured characteristics of synchronous generator system with impedance load for different rotor speeds .............7.................................................... Calculated and measured characteristics of synchronous generator system with impedance load for a constant rotor speed and different control winding excitation currents .......................11........10..............4........ 130 6... 115 5........... 127 6........................................................10................ Schematic diagram of the doublyfed synchronous generator system with dcdc buck converter .................................5...................................... Complexform equivalent circuit of doublyfed synchronous reluctance machine with series coreloss resistances ...............6.......4. Complexform equivalent circuit of doublyfed synchronous reluctance machine with shunt coreloss resistances . Voltage deexcitation phenomena in the generator feeding RL load ...................... Calculated synchronous generator characteristics of generator system feeding impedance load.............................................................. Doublyfed synchronous generator systems..... 114 5.8.........2....... 113 5............................................... 122 5. 131 6...... Measured and calculated steadystate performance curves of synchronous generator system with a threephase rectifier load ......2........ 123 5...... 138 ..........................11............. 124 Figure Page 5.........1. Buck DCDC converter ................................. Measured waveforms of the synchronous generator system feeding a threephase rectifier load (5msec/div) ...3.......... 105 5...... 103 5.............. Simulated starting transient of the synchronous doublyfed reluctance generator system feeding an impedance load ....................................................... 102 4.....1... Measured waveforms of synchronous generator system with impedance load (5msec/div)........................ 120 5.........................
.......11........... Equivalent circuit representation of a leadacid battery.........................................1(a) ... The steadystate characteristics of the generator system with DCDC boost converter in the Figure 6....................... maximum output torque operation ...4........................................10............................................................................7...................... 153 6............ 152 6............2.........16................................8..... 157 Figure Page 6...............................................6..........................................................3.....5. Electrical excitation transient for the doublyfed synchronous reluctance generator charging a leadacid battery .... Measured and calculated steadystate characteristics of the doublyfed synchronous reluctance generator feeding an impedance load.................................. 152 6........ 150 6.............1(a) ...................... 164 7..14.................................................................................... Boost DCDC converter ................................................. 149 6........ Starting transient of the generator system shown in Figure 6.. Generator deexcitation due to reduced DCDC converter duty ratio .....6.........13........2(a) ..1(a) ... Control Scheme I: Fieldorientation control of doubly fed synchronous .....1..... Measured waveform of the generator with DCDC boost converter in the Figure 6......................17............. Measured waveform of the generator with DCDC buck converter in the Figure 6....... The steadystate characteristics of the generator system with DCDC buck converter in the Figure 6..........1(b) ................... Performance curves for extended speed range................................ 159 7........... 140 6.. 155 6............................... The steadystate characteristics of the generator system with DCDC buck converter in the Figure 6..... 156 6... Control winding connection ....12........................ Steadystate waveforms of the generator system feeding impedance load ........................... The steadystate characteristics of the generator system with DCDC boost converter in the Figure 6...... Measured steadystate waveforms of the generator charging a 24V battery .. 148 6.1(a) ....................1(b) ..................9...... 139 6.......................... 158 6... 151 6..... 172 7.................................15..
.... 198 7..............................12...................... 211 7....... 185 7....................... 202 7...4.............. Block diagram of a currentcontrolled VSI .................................. 207 Figure Page 7....... Schematic of the space vector PWM ................ Control Scheme II: Fieldorientation control of doublyfed synchronous reluctance machine with currentcontrolled voltage source inverter (VSI) 174 7.... Block diagram of IP speed control system................................... 192 7..... Profiles of the phase current and voltage during the simulation process ..20...........6..9....................... Integral proportinal linear and decouping controller structure................ Simulation of the currentcontrolled VSI............................................ Dynamic response to a twodirection trapezoid speed command ................10......... 206 7..........................16 Profiles of the power winding phase current and voltage during the simulation process ........................ 201 7....reluctance machine with voltagecontrolled voltage source inverter (VSI)..........................15............ 209 7...................7....... Response to change of load torque ................ Dynamic response to a twodirection trapezoid speed command.................................... 188 7.................... Dynamic response to a wide speed range of 2............. 208 7.................................5........................................... 203 7............ Dynamic response to a stepchange of speed .........................11.. Simulation of the voltagecontrolled PWMVSI...........17..............18............................................................... 175 7........ 196 7............19. Dynamic response to a trapezoid speed command . 212 7. Circuit diagram of a threephase VSI ........... Profiles of the power winding phase current and voltage during the simulation process .......... Profiles of the power winding phase current and voltage during the simulation process ..8.................................................... 174 7..13................14...........5 times base speed ............................. 197 7......21....
..........for control scheme I ...........22............................................... Profiles of the power winding phase current and voltage during the simulation process for control scheme I ............ 215 ............ 214 7.................................
brushless doublyfed induction machine has a special cage construction to support the two airgap fields produced by the two sets of stator windings. The stator windings of the doubly fed machines are shown in Figure 1. Separate stator windings are generally undesirable except added flexibility of design or operation is necessary. In Figure 1. mixedpole) machines have received renewed attention in the last few years for use in adjustablespeed drives where efficiency optimization and energy conservation are desirable.1(a). while other threephase windings usually used for speed and power flow control are called the control windings. ABC and abc. The threephase stator windings that carry the load are called the power windings. In general. In Figure 1.1(b). the resulting windings can be divided into two sets of stator windings with different pole numbers when viewed from two sets of leads. by following special rule of winding connection. When these two sets of leads are powered simultaneously from two independent sources. two stator windings are separated from each other. This class of machine has two threephase stator windings wound for different pole numbers and a cage or a reluctance rotor. 1 .CHAPTER 1 RESEARCH BACKGROUND 1. Two sets of stator windings can be connected together or physically separated. the terminal currents will not affect each other due to the symmetric nature of the winding.1.1 Introduction Brushless doublyfed (dualwinding.
originated one of those ingenious techniques developed to overcome fixed speed limitation of the induction motor before the frequency conversion was developed using power electronics. called Hunt motor [1]. on the other hand. Based on the ideas of the Hunt A B C a b c (a) (b) (c) (d) Figure 1. (c) Salientpole rotor. (a) The same group stator windings. Doublyfed reluctance machines. The concept of a single machine with two sets of polyphase stator windings. .1(d)). has either a simple salient laminated pole (shown in Figure 1.1. resulted from the incorporation of two effective machines in a single magnetic circuit.The doublefed reluctance machine. (b) The separate group stator windings. which do not couple directly but interact via a specially formed rotor. (d) Axiallylaminated rotor.1(c)) or axially laminated rotor structure with no cage windings (shown in Figure 1. The first machine to use this concept.
rotor polepair number. the rotor speed ωr will be adjustable by increasing or decreasing the current frequency of the control windings. Because the speed of the selfcascaded machine can be controllable without brushgear by adjustable resistance.2. the socalled selfcascaded induction machine [2] was made by overcoming the structural problems in design. Where ω1 is the angular frequency of the power winding currents. This interest is promoted by the demonstrated adjustable speed drive (ASD) capability in which one of the stator windings is supplied by a converter of a rating significantly smaller than that of the machine. while ω2 is the angular frequency of the control winding currents. motor. This configuration. Broadway and his associates [411] had extensively investigated this type of machine two decades ago.Converter Bidirectional Fractional rating Control Windings Power Windings BDFM Stator Figure 1. bidirectional power flow has revived interest in the selfcascaded induction machine [1213]. The advent of power electronics converters capable of adjustablefrequency. adjustablevoltage.2. shown schematically in Figure 1. Its basic adjustable speed principle is based on this equation: ωr = (ω1 + ω2)/Pr. and ω1 are fixed. some significant industrial use has been found over several years [3]. Schematic of BDFM ASD. If Pr. . is now referred to as a brushless doubly fed machine (BDFM).
by using an accurate model. The model can be used to investigate the performance of machines and show the effects of certain trends in machine design. (d) The design and analysis of the rotor structure of brushless doublyfed machines with cagerotor and reluctance rotor [4245]. Recent research has developed a few models of brushless doublyfed machine with cage and reluctance rotors.1. (b) The investigation of doublyfed machine as a motor in adjustable speed drives systems (ASD) [2732]. The simulation. will provide adequate representation of full performance for control and stability. (c) The investigation of doublyfed machine as a generator in variable speed.1 Modeling and Analysis of DoublyFed Reluctance Machines An accurate model is very important for design and application engineers. constantfrequency powers generation and in standalone application [3341].2. . intensive investigations have been undertaken. The research can be divided into four areas: (a) The development of dynamic and steady state models for performance evaluation and design of doublyfed machines [1526]. 1.2 Literature Survey To explore the potential of brushless doublyfed machine and improve its design to sufficiently obtain its expected advantages.
except for the extra series impedance KpXpq and KqXpd in the primary and secondary circuits. Xpq is a referred value of mutual reactance between the 2ppole and 2qpole component windings. a steady state equivalent circuit is given to predict machine's performance in the synchronous mode of operation and to investigate the characteristics of the drive system in steady state conditions. and different .811]. Xp and Xq represent the selfreactances of 2ppole and 2qpole component windings. a common feature of all the above analytical work is the assumption that the machine is equivalent to two magnetically separate wound rotor motors. Vp. This circuit is similar to that of a conventional induction machine. Vq. and iq are the voltages and currents of the 2ppole and 2qpole component windings. The steadystate equivalent circuit was shown in Figure 1.3 [5]. β is the relative displacement between the 2qpole component winding and the rotor at a zero instant in time. In early research [4.3. Sq and Sd are slip values of primary and secondary windings. ip. To investigate the characteristics of doublyfed machines based on the steady state model. respectively.rp vp ip xp kpxpq kqxpq xpq x’q r’q/spsq i’qejβ v’qejβ/ spsq Figure 1. Steadystate equivalent circuit. The constants Kp and Kq are dependent on the magnetic properties of the rotor.
iqr. Vd2. and id6 represent 6pole stator winding qd voltages and currents. It only considers the fundamental mutual inductance and transforms the equations to a twoaxis rotor reference frame. Lm2. To reduce the complexity of the detailed model so that it is suitable for BDFM drive system dynamic studies. The twoaxis model and the detailed model have been successfully used to predict the effects of certain trend in machine design [17. Lm6. respectively. λd6.pole numbers of the rotors which are connected electrically and mounted on a common shaft. the twoaxis model [20] was developed from the detailed model. it is not adequate for detailed machine and drive system design. Vd6.4[20] where Vq6. However. and λd2. 2pole) and rotor winding. Ll6 and Ll2 are the leakage inductances of 6pole and 2pole stator windings. and Lrm represent the magnetizing inductances of stator windings (6pole. Vqr. Vdr. a detailed. Although this approach is appropriate for conceptual understanding. The qd flux linkages of stator windings with 6pole and 2pole are λq6. r2. represent the resistances of stator windings and rotor winding. these models do not include the expressions of the machine parameters so that the application of the models to predict different running modes and configurations were limited.r6. Recently. M6 and M2 are mutual inductances between stator windings (6pole. and rr . 2pole) and rotor winding. iq6. This model was developed for the brushless doublyfed induction machine with the cagerotor. λq2. and idr are rotor winding qd voltages and currents.18] and to develop closed speed control systems [21]. Llr is leakage inductance of rotor winding. dynamic model has appeared in the literature [19]. The twoaxis model equivalent circuit of brushless doublyfed machine with cagerotor was shown in Figure 1. and id2 represent 2pole stator winding qd voltages and currents. while Vq2. iq2.
r6 Vq6 3ωλd6 . . Twoaxis dynamic model for the cage rotor machine.+ r2 Vq2 L l2 L2m M62=0 M2 iq2 r6 Vd6 3ωλ q6 L l6 .+ L6m Ll2 M62 =0 M2 V d2 id2 L2m Figure 1.4.+ M6 Llr Lrm rr idr Vdr id6 r2 ωλq2 .+ L l6 L6m M6 Llr Lrm rr iqr Vqr iq6 ωλ d2 .
. One has presented rigorous analytical model [22. For the brushless doublyfed reluctance machine having either salient or axially laminated rotor structure.5[26]. which can readily represent the effect of saturation in cagerotor machine. a transient model is presented in [25. 26]. which are dominant in the operation of doublyfed reluctance machines. It is developed from the concept of winding functions and the principles of dq arbitrary reference frame transformation.5. 23] based on generalized harmonic theory [46]. Recent work has led to more progress. another [24] described a timestepping finiteelement model. Qd equivalent circuit of salient rotor machine.Figure 1. The qd equivalent circuit of this machine is shown in Figure 1. This model ignores the influence of magnetic saturation and core loss.
since brushless doublyfed reluctance machines have both field and armature windings on the stator and reluctance rotor does not carry currents.6. brushes and rotor copper loss are completely eliminated.2. has reduced the required converter rating. Besides.6). but high cost and bulky size of woundfield induction machines and the maintenance required for the sliprings. 1.2 The Application of Brushless DoublyFed Machines as Motors and Generators The conventional slip power recovery system employing woundfield induction machines (shown in Figure 1.Power Grid Stator windings Rotor windings BDFM PWM Converter Figure 1. Furthermore. Therefore the system has a simpler and more reliable structure. recent research have explored the potential of applying the brushless doublyfed machine to variable speed generating systems. Conventional slip power recovery system. have unfortunately limited their applications. less maintenance cost and higher efficiency than the conventional system. slip power recovery system with brushless reluctance machine has drawn much attention recently. . To avoid the disadvantages of the system. It has the same advantage of substantially reducing the inverter power rating. such as wind power generation.
the concept and implementation of field orientation control of brushless doublyfed reluctance machine for variable speed drive and generating system are presented. (2) With a rotor structure optimized for minimum core loss. .In papers [2831].and supersynchronous speeds operations. so variable speed drive and generator operation with decoupled active/reactive power control can be achieved. To apply this system to restricted applications where accessibility to the rotor shaft is prohibited. (3) Controllable powerfactor and low harmonic distortion of the utility supply. the literature [30] presents a sensorless control scheme. a doublefed reluctance machine may have better efficiency than an induction machine of the same rating. Some expected advantages of ASD or VSG system with this machine are listed below: (1) Its ability to operate in induction and synchronous machine modes with the possibility of sub. and compatibility with existent production line [14].3 Research Motivation The brushless doubly fed reluctance machine with simple saliency on the rotor is considered. lower manufacturing cost. (4) Robust machine construction. The stator flux orientation is employed to achieve decoupled control of torque and active/reactive power through the secondary currents. high efficiency. It will further enhance the reliability and reduce the cost of the drive. 1. It appears to be highly attractive due to its structural simplicity.
relatively little attention has been paid to their use as standalone generators. a model [24] considering the effect of saturation has been presented for the application of the cagerotor structure. The idea of using doublyfed reluctance machines as generators was inspired by the fact that these machines can run at high speeds where the efficiencies of primemovers (turbines) are relatively high. The reluctance machine with saliency rotor has more iron loss and higher saturation than cagerotor machine. there is no need for special starting arrangement. the dynamic model developed for saliency rotor machine ignores these effects. (b) While a lot of work is being done on the analysis and control of doublyfed reluctance machines used in adjustablespeed drive application. Hence we need to look for a new model whose equations consider the general case which includes the influence of sequences of the stator windings and the saturation of the airgap flux linkage on self and mutual inductance. These advantages look very attractive for some applications such as . applied to one or both stator windings. Also. so it is necessary to develop a model including the effect of iron loss and saturation. At present. is increased. since the reluctance generator is run with a prime mover. the machine will naturally saturate. The performance analysis is less accurate if the model neglects the effects of iron loss and saturation. To solve this problem. When the voltage.(5) Operation as an induction motor in the event of converter failure. We hope that this generalization can provide a more accurate insight into the machine design and performance analysis. Literature survey yields some important information of this machine: (a) In recent years various efforts have been made to establish proper dynamic model and explore the design variations of the doublyfed brushless machine with both cagerotor and reluctance rotor.
With the use of PWM converter and digital signal processor. . The motivation for this work was inspired from the above discussion. We hope the field orientation control of the doublyfed reluctance machine in the synchronous mode can also achieve the same performance level as the induction machine. The field orientation control strategy is now used in induction motors for achieving precise and fast dynamic speed and/or torque responses. marine generators for gasturbine drives and electrical vehicle generator systems.aeroplane power systems. One important characteristic of doublyfed reluctance machine is its capability of operating in synchronous mode with one set of stator windings connected to a DC source. military and power system due to its low cost. high reliability and flexible control methods. Doublyfed reluctance machine has been shown to be effective in variablespeed drive and generating system with the two sets of windings connected to AC sources. we hope that the proposed research will promote the doublyfed reluctance machines to find more applications in industry. it is significant to explore the potential of this machine as stand alone generator and use it to develop new DC power generation schemes. Little attention has been paid to its synchronous drive performance. Hence.
In this equivalent circuit. The concept of the qd harmonic balance is used to determine voltage equations. threephase winding construction from the other set of terminals (A. which is shown in Figure 2. When viewed from one set of terminals (a. the parameters of the experimental machine will be measured. respectively. b.1(a) and (c). Modeling of a doublyfed reluctance machine This topic will deal with the modeling and analysis of the doublyfed reluctance machines with simple salient or laminated rotor structure having 2Pr poles and two stator windings with pole numbers given as 2p1 and 2q. Its stator winding arrangement and rotor structure are shown in Figure 1. The resulting model includes saturation effects and core losses. C). The inductance parameter. An experiment machine. The stator windings are distributed in 36 slots. Lp. Based on this model. c). it is a 6pole winding construction and a two pole. Rcp and Rcs represent the core loss resistances of power windings and control windings.1.CHAPTER 2 PROPOSED RESEARCH The research work will include these topics: 1. The electromagnetic torque and the contributions of the control and power windings to the mechanical output power is determined using the ManleyRowe powerfrequency relationships. . respectively. The influence of the sequences of stator currents and machine windings are included in the analysis permitting the elucidation of the different possible modes of operations. will be used for this research. supplied by the Advanced Motor Development Center of Emerson Motor Co. Ls. and Lm change with the airgap flux linkage. B.
2. called power winding. A qd equivalent circuit model accounting for saturation of the magnetic paths and saturationdependent core and harmonic loss resistance will be used to predict experimental results. while the power winding (2pole) will be connected with impedance load and rectifier load.Figure 2. 2. called control winding. . and the load (impedance or rectifier) are connected to 2pole winding. The performance characteristics of a doublyfed selfexcited reluctance generator Proposed selfexcited generators are shown in Figure 2. the results of steadystate calculation and dynamic simulation will be compared with the experimental results from this machine. 3. Synchronous operation of a doublyfed reluctance generator Proposed synchronous generator system is shown in Figure 1. The experimental measurement and computer simulation will show the steady state and dynamic performance characteristics of this generator system. Qd complex equivalent circuit of doubly fed reluctance machine. For this generator system. A three phase capacitor banks will be connected to 6pole winding. To validate the modeling approach.3. control winding (6pole) will be fed with direct current source.1.
The operation performance of the doublyfed synchronous reluctance generator as a controllable power source for DC loads will be investigated. Figure 2. Its performance characteristics will be investigated by experimental measurement and computer simulation. Selfexcited doublyfed reluctance generator systems. and harmonic losses will be included in the dynamic and steady state models of generator system. core. where a battery is used as . a threephase rectifier and a threephase capacitor bank.3(a) shows the generator run by a source of mechanical power and feeding an impedance load. The system consists of a doubly fed reluctance machine.Figure 2. System schematic diagram of feeding an impedance load and feeding a battery was shown in Figure 2. a DCDC buck converter.3.2. Saturation effects.
This generator scheme is envisaged for standalone applications requiring regulated dc voltage or current and in charging battery in electrical automotive applications. starting source.3(b). Doublyfed reluctant synchronous generator systems. The generator scheme for battery charging is shown in Figure 2. The dynamic and steady state models of generator topologies will be set forth and used to calculate steady state performance characteristics. The power winding will deliver the dc power to the load.(b) with boost converter.(a) (b) Figure 2. (a) With DCDC buck converter. Its performance characteristics including steadystate . The excitation process will be established by connecting the generated dc voltage to the machine's control windings through the buck converter. with the control windings acting as the vehicle for excitation.3.
(a) (b) Figure 2. (b) feeding battery. . Regulated DC power generation systems using doublyfed synchronous reluctant machine. (a) Feeding an impedance.4.
1) to (2.Iqp Ids ] (2. The airgap flux is regulated using Iqs while Idp controls the instantaneous torque of doublyfed synchronous reluctance motor. Iqs = Is . A fieldoriented doublyfed synchronous reluctance machine will emulate a separatelyexcited dc . which has been yielded from the derivation above.2) (2.5) where voltage equations of power winding and control winding are (2.5 ( p1 + q ) Lm [ Idp Iqs .and dynamic will be shown through the computer simulation and experimental measurement. Because of Ids = 0. Field orientation control of a doublyfed reluctance machine in the synchronous mode The dynamic equations of the doublyfed synchronous reluctance machine in the rotor reference frame are: Vqp = rp Iqp + p λqp + ωr λdp Vdp = rp Idp + p λ dp . By connecting the control winding to a DC current source Is. 4. for decoupling the field control from the torque control.5 ( p1 + q ) Lm Idp Iqs. The field orientation principle defines conditions. torque equation can be expressed as Te = 1.5). and torque equation is (2.4). the following equations are obtained: Vqs = (2/3)Vdc.ωr λ qp Vqs = rs Iqs + p λ qs Vds = rs Ids + p λ ds Te = 1. .4) (2. Iqs and Ids are the qd control winding currents while Iqp and Idp are qd power winding current. Lm is mutual inductance.1) (2.3) (2. Vds = 0 Ids = 0 .
motor. the voltages of power winding are kept at rated value while the control winding current. This is called constantpower. constant torque will be hold below the base speed while the fieldweakening operation conditions must be used to achieve constant power output above the base speed. . is equal to rated value. Iqs will be kept constant to maintain the constant magnetizing flux while Idp will be used to control the instantaneous torque with setting Iqp = 0. Above the base speed. which will provide an adjustable frequency voltage source for the power windings. its design procedure will also be given. a novel nonlinear controller is used and its principle as well as design procedure will be described. To investigate the performance of this machine running in a wide speed range. Figure 2. Its steady state characteristics will be calculated and dynamic performance will be demonstrated using computer simulation program. The space vector pulse width modulation (SVPWM) technique and hysterestic current control technique will be used to control voltage source inverter. In control scheme I. IP controller with noovershoot performance is used in control scheme II.5 shows the block diagram of two control systems. To maintain the constant output power and maximize the output torque. Iqp and Idp need to be regulated at the same time. Iqs. The control windings will be supplied by a DC current source. fieldweakening operation. the voltages of the primary winding rise linearly with speed. Below the base speed.
ω r + Kdd Vdd + + 1 . The research work will make contribution in these areas: . Idp. Idp.5. Iqs.Iqp .ω r IP Current Control I* dp I* Ap IBp ICp I* Bp I* Cp ωb ωb I* qp T (θr) 1 θr ! FieldWeakening Control (b) Figure 2. Field orientation control systems for a doublyfed synchronous reluctance machine. (a) Control scheme I. (b) control scheme II. ωr Nonlinear Controller linearization Vdp Vqp T ( θr) Power Windings IAp IBp Control Windings Is ωr* + I* qp . ωr T(θr) Idp θr ! (a) Power Windings Control Windings Is ωr VDC VSI PWM IAp DFM ωr* + .I qp Kqq + VSI PWM ICp DFM ωr Vqq + Iqp.
(1) A dynamic model of doublyfed reluctance machine considering the effects of core loss and saturation will be presented. . (3) Two field orientation control schemes for a doublyfed synchronous reluctance machine will be investigated for highperformance operation both in the constant torque and constant power region. (2) The performance characteristics of doublyfed selfexcited reluctance generator and doublyfed synchronous reluctance generator will be investigated and analysis methods will be presented. It will provide a more accurate model for the performance evaluation and the design of doubly fed reluctance machine. It has potential for use in standalone applications and in electric automobiles. The simulation results will supply valuable information for realization and application of the control systems. An innovative DC power generator scheme using doubly fed reluctance machine with salient rotor structure will be investigated.
2 Stator and Rotor Structure of the Experimental Machine The model set forth in this chapter applies to doublyfed reluctance machines with simple salient laminated rotor having 2Pr poles and two stator windings with pole numbers given as 2p1 and 2q. The resulting model.CHAPTER 3 MODELING OF A DOUBLYFED RELUCTANCE MACHINE 3. The electromagnetic torque and the contributions of the control and power windings to the mechanical output power are determined using the ManleyRowe powerfrequency relationships. q. including saturation effects and core losses. respectively. represent the pole22 . the concept of winding functions is used to derive the machine inductances. The influence of the sequences of stator currents and machine windings are included in the analysis permitting the elucidation of the different possible modes of operations. First. respectively. p1.1 Introduction This chapter mainly deals with the modeling and analysis of the doublyfed reluctance machine. 3. Conditions for the development of average electromagnetic torque are developed which also give insight into the machine design criteria. and Pr. is used to unveil the inherent oscillatory instability of the machine and to predict the steadystate performance and dynamic characteristics for its motor or generator systems in the following chapters.
The experimental machine used for this research has stator winding arrangement and rotor structure shown in Figure 3. Stator and rotor structure of the experimental machine. B.1. b. control windings.pair number of the power windings. these two symmetrical sets of threephase windings are electrically independent of each other.1(a) shows the stator winding connection diagram of the machine distributed in 36 slots which when viewed from one set of terminals (a. and C). is a threephase.1. . a nontriplen twopole. g2 g1 θ m + π (2n − α ) / 2 Pr θ m θ m + π ( 2n + α ) / 2Pr (a) (b) Figure 3. (b) Rotor structure. and c). From the other set of terminals (A. For any threephase balance ac power supply. and rotor. 6pole winding construction. Figure 3. threephase winding construction is observed. (a) Stator windings.
and the airgap length is assumed to take a constant value g1 over the rotor arc and g2 elsewhere as shown in Figure 3. classical assumptions are made in order to obtain closedform equations for machine inductances. The permeabilities of the stator and rotor iron parts are assumed to be infinite. for the 2p1 and 2q windings where K3 and K4 are either +1 or 1. they are defined as k3ωp. the stator winding distributions are approximated by their fundamental components.5) (3.3 Machine Model In the analysis that follows. The stator windings are supplied with threephase voltage vectors VABC and Vabc. k4 ωs .7) .7. therefore. The stator voltage differential equations expressed in terms of the phase currents and flux linkages are.1) (3.2) (3.3.6) T Vabc = [Va Vb Vc ] (3. expressed as [27] VABC = rp IABC + pλABC Vabc = rsIabc + pλabc where λABC = L ABC I ABC + L ABCabc I abc λ abc = L abc I abc + L abcABC I ABC T VABC = [VA VB VC ] T I ABC = [I A I B IC ] (3. Saturation effects will be included in section 3.4) (3.1(b). respectively.3) (3. Since the frequencies of the supply voltages can be positive or negative (positive or negative current sequence).
respectively.8) LAC $ LBC " " LCC " # LABC ' LAA =% % LBA % & LCA LAB LBB LCB Lab Lbb Lcb (3.15) (3. currents.14) (3. and flux linkages of each set of threephase stator windings are transformed to their respective qdn synchronous reference frame equations using the matrix transformation T(θp) and T(θs). which are defined as follows: Vqdnp = T(θp)VABC Vqdns = T(θs)Vabc Iqdnp = T(θp)IABC Iqdns = T(θs)Iabc λqdnp = T(θp)λABC λqdns = T(θs)λabc and (3. and the derivative d/dt is given as p.11) The perphase resistances of the stator windings are rp and rs.9) ' Laa Labc = % % Lba % & Lca Lac $ Lbc " " Lcc " # ' LAa =% % LBa % & LCa LAb LBb LCb LAc $ LBc " " LCc " # (3.17) .12) (3. (3. The phase voltages.13) (3.16) (3.10) L ABCabc = L T abcABC .IT abc = [ I a I b I c ] (3.
21) into Equations (3.20) θ s = ! K 4ω s dt + θ so (3.18) (3.19) θp = ! K 3ωp dt + θ po ' %cos θ s 2% T (θ s ) = % sin θ s 3% % 1 % & 2 cos(θ s − 2π 2π $ ) cos(θ s + ) " 3 3 2π 2π " )" sin(θ s − ) sin(θ s + 3 3 " 1 1 " " 2 2 # (3.25) (3.24) (3.' %cos θp 2% T( θp ) = % sin θp 3% % 1 % 2 & cos( θp − 2π 2π $ ) cos( θp + ) " 3 3 2π 2π " sin(θ p − ) sin( θp + )" 3 3 " 1 1 " " 2 2 # (3. the resulting qdn voltage and flux linkage equations are given as Vqdnp = rpIqdnp + pλqdnp + K3 ϖpλqdnp Vqdns = rpIqdns + pλqdns + K4 ϖsλqdns λqdnp = LppIqdnp + LpsIqdns λqdns = LssIqdn + LspIqdnp where ' 0 1 0$ ' 0 1 0$ ϖ p = ωp %− 1 0 0" .21) where θpo and θso are initial angles of the synchronous reference frames.123.183. ϖ s = ωs % − 1 0 0" % " % " % % & 0 0 0" # & 0 0 0" # (3. Substituting Equations (3.23) (3.22) (3.26) .17).
θrm) and Nj(φ. θ j ) N i ( φ. The angle. θrm) is called the winding function and represents. as Ni(φ. the mutual inductance between and winding "i" and "j" in any machine (mean airgap radius is constant) is calculated by L ij = µ 0 rl ! g −1 ( φ.27) L sp = LT ps = T( θ p ) L ABCabc T( θs ) −1 ' L mqq % = % L mdq % L mnq & L qds L dds L nds (3.' Lqqp % Lpp = T (θ p ) LABCT (θ p ) −1 = % Ldqp % Lnqp & Lqdp Lddp Lndp Lqnp $ " Ldnp " Lnnp " # L mqd L mdd L mnd L mqn $ " L mdn " L mnn " # (3. and the inverse gaplength is represented by g1(φ. the winding is only a function of the stator .θrm). θ rm ) N j ( φ.29) 3. φ. Lsp and Lss are calculated using the winding function approach [48. According to this method. The winding functions of winding "i" and "j" are given. If this winding is located on the stator.28) L ss = T ( θ s ) L abc T( θ s ) −1 ' L qqs % = % L dqs % L nqs & L qns $ " L dns " . θ i )dφ .49]. θrm). motor stack length is l. defines the angular position along the stator inner diameter while the angular position of the rotor with respect to the stator reference is θrm. in effect. the MMF distribution along the airgap for an unit current in winding "i" or "j". 0 2π The average airgap radius is r.4 Calculation of Inductance Using Winding Function Theory The qdn machine inductance matrices Lpp. L nns " # (3. The term Ni(φ. respectively. θrm) or Nj(φ.
Winding function for 2pole phase a.50].4. Although the winding functions for the mixedpole machine have substantial space harmonic contents as shown in [14. Winding functions representing the six and two pole windings are drawn in Figures 3. Na(φ) 2π φ Figure 3.peripheral angle φ while if the winding is located on the rotor the winding must be expressed as a function of both φ and the mechanical position of the rotor θrm. .2 and 3.3. Winding function for 6pole phase A. have the greatest effect on energy conversion. respectively. NA(φ) 2π φ Figure 3. they are represented here by their fundamental components since only these components.3.2. as shown in Figure 3.
the qd components of phase winding function NA. NB and NC are expressed as N qp = N p cos( p (φ − K1φ p )) N dp = N p sin( p(φ − K1φ p )) .30) (3.34) (3. 3 (3.35) By using the synchronous reference frame T(θp).NA(φ) Na(φ) 2π φ Figure 3. Fundamental components of the winding functions.37) .36) (3. These components of the phase windings accounting for the winding sequences are expressed as N A = N p cos( p(φ − φ p )) N B = N p cos( p (φ − φ p ) − K1 N C = N p cos( p(φ − φ p ) + K1 2π ) 3 2π ) 3 (3.4.33) N a = N s cos(q (φ − φq )) N b = N s cos(q (φ − φ q ) − K 2 N c = N s cos( q(φ − φq ) + K 2 2π ) 3 2π ) .32) (3. (3.31) (3.
The fundamental winding distribution factors are defined as Kωp and Kωq. N1 and N2 are the numbers of series connected turns perphase of Cp and Cq circuits for the power and control windings.41). respectively. Nb and Nc are expressed as N qs = N s cos(q (φ − K 2φq )) N ds = N s sin( q(φ − K 2φq )) .43) . N s = 4 Cqπq N 2 K ωq . The initial angular displacements between the fundamental components of the winding functions and the stator reference are φp and φq.40) (3. and Lps are obtained.39) (3.θ m ) = % % π (2n + α ) ≤ φ ≤ θ rm + π (2n + 2 − α ) % g 2 θ rm + 2 Pr 2 Pr & (3. expressions for inductances comprising Lpp. 2. Lss.303.. The airgap function from Figure 3. With Equations (3.41) In Equation (3. 3.1(b) is expressed as π ' θ rm + (2n − α ) ≤ φ ≤ θ rm + π (2n + α ) % g1 2 Pr 2 Pr % g (φ . K 2 = % &1 counterclockwise Np = 4 C pπp N1Kωp .29).By using the synchronous reference frame T(θs). 1. They are expressed as 3* 2 Lqqp = + ( µ 0 r! ! g −1 (φ .38) (3.2Pr1 and the rotor pole pitch is α.42) substituted in Equations (3. (3..θ rm ) N qp dφ .273. 4. as well as Np and Ns are defined as clockwise ' −1 K1 . respectively. the signs of the winding sequence. K1 and K2.42) where n = 0.2) 0 2π (3. the qd components of phase winding function Na.
50) . θ rm ) N dp N ds dφ . θ rm ) N qs 2 .3* L qdm = + (µ 0 r! ! g −1 (φ.52) (3.44) (3.49) (3. .48) (3.54) (3.3* L dqm = + (µ 0 r! ! g −1 (φ. θ rm ) N qp N dp dφ .51) (3.2) 0 .3* L qqm = + (µ 0 r! ! g −1 (φ. ) 0 3* L qdp = L dqp = + (µ 0 r! ! g −1 (φ.3* 2 Lddp = + ( µ 0 r! ! g −1 (φ .2) 0 .3* 2 dφ L qqs = + (µ 0 r! ! g −1 (φ.46) .2) 0 3* L qds = L dqs = + (µ 0 r! ! g −1 (φ.θ rm ) N dp dφ 2 .2) 0 L qns = L nqs = L dns = L nds = L nns = 0 2π 2π 2π (3. 2π 2π 2π 2π (3. ) 0 .2) 0 . θ rm ) N qs N ds dφ .2) 0 L qnp = L nqp = L dnp = L ndp = L nnp = 0 2π 2π (3. θ rm ) N qp N qs dφ .3* 2 dφ L dds = + (µ 0 rl ! g −1 (φ.47) (3. θ rm ) N qp N ds dφ . All inductances of the zerosequence winding components are zero. θ rm ) N ds .55) The expressions for these inductances are given in Appendix 3A.2) 0 L qnm = L nqm = L dnm = L ndm = L nnm = 0 .53) (3. θ rm ) N dp N qs dφ .45) (3.3* L ddm = + (µ 0 r! ! g −1 (φ.
and ωrm . are satisfied as expressed in Condition I and Condition II. = odd = 2 z + 1 Pr Pr Lddp = Lqqp = 3µ 0 rl 2 . q. The constant qd inductance components make sure that the requirements of the electricmechanical energy conversion and reducing the pulsating torque are satisfied. To reduce the pulsating torque. if the combination of the polepair numbers: p1. the selfinduced voltages must also have the same frequency as those of the currents. g1 g 2 ) Lqqs = Ldds = Lmqq = − Lmdd = 3µ0 rlpr N s N p . the frequency of the speed voltage in a given phase must be the same as the frequency of the current flowing in the same phase. By checking qd component expressions of inductance Lp. we can see that all qd components are equal to constant. and the relationship of frequencies among ωp. and Lps in Appendix 3A. and Pr. Condition I ( K1 p1 + K 2 q )ω rm = K 3ω p + K 4ω s K1 p1 + K 2 q K p − K 2q = even = 2m. 1 1 = odd = 2n + 1 Pr Pr 2 K1 p1 2 K 2q = odd = 2k + 1.1 1 * N s πα + + ( + ( 2 .1 1 * N pπα + + ( + ( 2 . Ls. ωs. g1 g 2 ) 3µ 0 rl 2 . 1 Condition II ( K1 p1 − K 2 q )ω rm = K 3ω p − K 4ω s .To realize the electricmechanical energy conversion.1 ' ( K1 p1 − K 2 q )πα $ 1 * + − ( " ( ⋅ sin % (K1 p1 + K 2q ) + g g 2 Pr 2 ) & # .
Pr by considering different combinations of values of K1. g1 g 2 ) & 2 Pr # where m and n are positive or negative integers.TL (3.5 Calculation of Electromagnetic Torque The mechanical equation of motion for the machine is expressed as Jpθ rm = Te . Condition I or Condition II. = odd = 2 z + 1 Pr Pr Lddp = Lqqp = 3µ 0 rl 2 . K2. 1 1 = even = 2m Pr Pr 2 K1 p1 2 K 2q = odd = 2k + 1. K3 and K4. g1 g 2 ) Lqqs = Ldds = 3µ 0 rl 2 . Different operating modes are established from Condition I and Condition II for possible values of p1. we can prove that all of them are constants.1 1 * + ( N pπα + + ( 2 . By substituting the constraint condition.56) . it is found that the rotor speed is given as (p1 ± q) ωrm = ω p ± ω s. g1 g 2 ) Lmqq = Lmdd = 3µ 0 rlPr N s N p . q.1 1 * + ( N s πα + + ( 2 . into the expressions of qd inductance elements in Appendix 3A.1 ' ( K p − K 2 q)πα $ 1 * + − ( ⋅ sin % 1 1 " ( + (K1 p1 + K 2q ) . 3.K1 p1 + K 2 q K p − K 2q = odd = 2n + 1. In general.
57) If linearity is assumed (infinite permeability). L nns " # The developed electromagnetic torque becomes Te = 1 I qdnp 2 + I qdnp [ ] T d dθ rm d dθ rm [L ][I ] + 1 [I ] 2 pp qdnp qdns T d dθ rm [Lss ][I qdns ] (3. the coenergy is equal to the stored magnetic energy given as ω co = 1 [I ABC ]T [LABC ][I ABC ] + 1 [I abc ]T [Labc ][I abc ] 2 2 T + [I ABC ] [LABCabc ][I abc ] .15) and relationships between qd inductance elements and abc inductance elements expressed as follows ' L AA =% % L BA % & LCA ' L Aa = % L Ba % % L Ca & ' L aa = % L ba % % & L ca L AB L BB LCB L Ab L Bb L Cb L ab L bb L cb L ABC ' Lqqp L AC $ " −1 % L BC " = T (θ p ) % Ldqp % Lnqp LCC " # & Lqdp Lddp Lndp Lqnp $ " Ldnp " T (θ p ) Lnnp " # L mqd L mdd L mnd L mqn $ " L mdn " T ( θ s ) L mnn " # L ABCabc ' L mqq L Ac $ −1 % " L Bc = T ( θ p ) % L mdq " % L mnq L Cc # " & ' L qqs L ac $ −1 % " L bc = T ( θ s ) % L dqs " % L nqs L cc " # & L qds L dds L nds L abc L qns $ " L dns " T( θ s ) . the moment of inertia of the rotor and connected load is J and Te is the developed electromagnetic torque.58) Equation (3.58) is further expressed in terms of qdn quantities using Equations (3.59) [ ] T [L ][I ] ps qdns . & dθ rm # (3. The electromagnetic torque is calculated from the magnetic coenergy ωco as ' dω $ Te = % co " . (3.123.where the load torque is TL.
The torque is time varying in general since the inductances are time varying and Iqdns. the qdn components of inductance [Lpp]'. L'nnm " # All the derivative components of inductance [Lpp]. and [Lps] are listed in Appendix 3B. [Lss]' and [Lps]' can be obtained. [Lss]. Iqdnp are constant quantities during steadystate operation conditions. By substituting the constraint conditions under Condition I or Condition II into the expressions in the Appendix 3B. The results are d [ L pp ] = [ L pp ]' = [0] dθ rm d [ L ss ] = [ L ss ]' = [0] dθ rm ' 0 % d [ L ps ] = [ L ps ]' = % L'dqm dθ rm % 0 & L'qdm 0 0 0$ " 0" 0" # where .where ' L'qqp % d [ L pp ] = [ L pp ] ' = % L'dqp dθ rm % L' & nqp ' L'qqs % d [ L ss ] = [ L ss ] ' = % L'dqs dθ rm % L' & nqs ' L'qqm % d ' [ L ps ] = [ L ps ] = % L'dqm dθ rm % L' & nqm L'qdp L'ddp L'ndp L'qds L'dds L'nds L'qdm L'ddm L'ndm L'qnp $ " L'dnp " L'nnp " # L'qns $ " L'dns " L'nns " # L'qnm $ " L'dnm " .
3) or ' 0 2* 0]+ ( (K1 p1 − K 2 q ) % %− Lqqm .59) will become ' 0 2* 0]+ ( (K1 p1 + K 2 q ) % %− Lqqm . Under the Condition I or Condition II. .60) . Equation (3. finally an average torque under Condition I and condition II are given respectively as Under Condition I Te = 3 (K1 p1 + K 2q ) λdp I qp − λqp I dp 2 [ ] (3.3) % & 0 Lddm 0 0 Te = [ I qp I dp 0$ ' I qs $ % " 0" " % I ds " 0" #% &0" # 2* = + ((K1 p1 + K 2 q ) (I qp Lddm I ds − I qs Lqqm I dp ) .2 2 L'qdm = L'dqm = − Lqqm ( K1 p1 + K 2 q) = Lddm ( K1 p1 + K 2 q ) 3 3 under Condition I and L'qdm = − L'dqm = 2 2 Lqqm ( K1 p1 − K 2 q) = Lddm ( K1 p1 − K 2 q ) 3 3 under Condition II . which can be proven using their expressions and constraint conditions in Appendix 3B.3) % & 0 Lddm 0 0 Te = [ I qp I dp 0$ ' I qs $ % " 0" " % I ds " 0" #% &0" # 2* = + ((K1 p1 − K 2 q ) (I qp Lddm I ds − I qs Lqqm I dp ) .3) The power winding qaxis and daxis flux linkage are expressed as λqp = LqqmIqp + LqdmIqs λdp = LddmIdp + LqdmIds . By using these two equations and considering the equality of Iqqm = Iddm.
64) (3. s) Ns * ([L ps ].Np * ( I 'qds = + + N ( I qds . .223.65) where N =+ s + Np .63) λ qdnp = L pp I qdnp + L'ps I 'qdns λ'qdns = L'ss I 'qdns + L'ps I qdnp (3.62) (3. * ( rs ( ) * ( [L ss ] ( ) * ( Vqds ( ) 2 2 rs' [ ] L'ss N =+ s + Np . If we refer the state variables of the control windings to the power windings. [L'ps ] = ++ N ( . p ) .Under Condition II Te = 3 (K1 p1 − K 2q ) λdp I qp − λqp I dp . Equations (3. 2 [ ] (3. N ' =+ s Vqds + Np .25) become V qdnp = r p I qdnp + pλ qdnp + K 3ϖ p λ qdnp ' ' Vqdns = rs' I qdns + pλ'qdns + K 4ϖ s λ'qdns (3.61) It is usual in electric machine analysis to refer all state variables to one set of windings.
663. q = 3 are generated from Equations (3. Simplification of these equations result in Equations (3.6 Some Design Aspects Here certain design criteria arising from the derived model equations are briefly discussed. The selection of the number of poles for the stator windings and the rotor is based on constraint conditions in Condition I and Condition II. The "effective" leakage inductances of two stator windings are the differences between the selfinductances and the magnetizing inductances. p1 = 1.69) It can be easily verified that the pole numbers of our test machine with Pr = 2.67) when n = 1 and m = 1. Figure 3.663.6 shows the graphs of the "effective" leakage factors Kp and Ks defined as Kp ≈ L qqp − L qqm L qqm = παm N ps − 1 sin( παm) .3.68) (3.69) which explicitly calculate the two stator pole numbers given the number of rotor poles: Condition I K1 p1 = ( 2m + 2n + 1) Pr 2 (3.66) K2q = ( 2m − 2n − 1) Pr 2 (3.67) Condition II K1 p1 = K2q = ( 2m + 2n + 1) Pr 2 ( 2n − 2m + 1) Pr 2 (3.
The leakage inductances of the power and control windings of the experimental machine are relatively high. N sp = Ns . the pole arc factor selected for any design must take into consideration the saturation effect. this is partly because the machine design was not optimized and moreover. the magnetizing inductance of this salientpole structure is theoretically limited by the practical feasible polearc of the rotor that ensures that the rotor and stator teeth under normal operating conditions are not in deep saturation.5 that the leakage factors are small when the turn number per pole of the two stator windings are equal with αm less than 0. Np It is observed from Figure 3. Ultimately.5. However.Ks ≈ Np Ns L qqs − L qqm L qqm = παm N sp − 1 sin( παm) where N ps = . it is expected that a doubly fed reluctance machine with an axially laminated or multiple flux barrier rotor structure have potentially smaller “leakage” inductances similar to those of . the effect of the space harmonic components of the stator windings and the curvature of the rotor pole.
5 0 0. .5 0.2 αm (b) 0 . (b) Control windings.3 0.5.5 2 1 .5 Nps=0.1 0.0 Kp 0 Nps=0.4 0 .0. (a) Power windings.5 Nps=1. Effective leakage factors for values of αm and ratio of effective turns per pole of the power winding. Nps = Np/Ns.0 0 0 0 .5 Ks 1 0 .6 Figure 3.5 Nps=1.2 αm (a) 0.4 0.
there are many small regions called domains. since they tend to flow in closed paths within the core of the machine.54] has been divided into two components: hysteresis loss and eddy current loss. There are two ways to represent this part of losses. using shunt loss resistances or series loss resistances. So the hysteresis loss in an iron core of the machine is the energy required to accomplishing the reorientation of domains during each cycle of the ac current applied to the core. all the atoms are aligned with their magnetic fields pointing in the same direction. 3.threephase synchronous or threephase induction machine. Using lamination stator and insulating resin between each lamination are the effect way to limit the eddy current and its loss. How to optimize the design of this kind of machines for achieving best performance still have a lot of work to do. Once the domains are aligned. Within the iron of the machine. core loss [53. The fact that turning domains in the iron requires energy leads to a common type of energy loss. some of them will remain aligned until a source of external energy is supplied to change them. These are shown in the equivalent circuit given . Noload test also reveals that the core loss is significant. It is clear that these significant effects must be included in the model to give accurate performance predictions. Eddy current is proportional to the size of the paths they follow within the core.7 Including Saturation Effects Experimental waveforms of winding currents and airgap flux linkage show that the machine has significant spaceharmonic current components and is highly saturated at moderate supply voltage levels. Eddy current losses are caused by induced electrical currents called eddy currents. In each domain. Traditionally.
(a) qaxis equivalent circuit. Qd equivalent circuit with shunt coreloss resistances.6.7. (b) daxis equivalent circuit. In Figure 3.(a) (b) Figure 3. in Figures 3. the resulting voltage equations including the shunt core loss resistances are expressed as Vqp Tp = rp Tp I qp + pλ qp − ω p λ dp Vdp Tp = rp Tp I dp + pλ dp + ω p λ qp ' Vqs Ts = rs' Ts I 'qs + pλ'qs − ωs λ'ds .6.6 and 3.
7. 2 Ns In Figure 3. L s = + N ( L s .' Vds Ts = rs' Ts I 'ds + pλ'ds + ωs λ'qs where λ qp = L p I qp + L m I 'qs λ dp = L p I dp + L m I 'ds λ'qs = L's I 'qs + L m I qp λ'ds = L's I 'ds + L m I dp Tp = Rmp rp + Rmp 2 . I 'ds = I ds Np Ns Lm = 3 Np L m12 .Np * . s) . s) ' Vqs = Vqs 2 Np Ns Np Ns ' . Ts = ' Rms ' r + Rms ' s . the resulting voltage equations including the series core loss resistances are expressed as Vqp = ( rp + Rmp ) I qp + pλqp − ω p λdp Vdp = ( rp + Rmp ) I dp + pλdp + ω p λqp ' ' ' Vqs = ( rs' + Rms ) I qs + pλ'qs − ω s λ'ds ' ' ' Vds = ( rs' + Rms ) I ds + pλ'ds + ω s λ'qs .Np * ' ( + ( rs' = + + N ( rs . Vds = Vds Np Ns I 'qs = I qs .
Where λ qp = L p I qp + L m I 'qs λ dp = L p I dp + L m I 'ds λ'qs = L's I 'qs + L m I qp . Qd equivalent circuits with series coreloss resistances. (b) daxis equivalent circuit. (a) qaxis equivalent circuit.7.(a) (b) Figure 3.
72) λ qdp = L p I qdp + L m I 'qds . but the physical meaning is more clear. s 2 * ( Ls ( ) 2 ' Vqs = Vqs Np Ns Np Ns ' .6 is more complicated and need accompanying computational burden. we find that the equivalent circuit in Figure 3.7 is simpler for the equation derivation and parameter calculation.λ'ds = L's I 'ds + L m I dp .71) (3.70) (3. 2 Ns These two equivalent circuits are interchangeably used in the steadystate calculation and dynamic simulation. the same results are obtained using any one of them.8 SteadyState Equivalent Circuits The complex form equations for the qd equivalent circuits in Figure 3. Vds = Vds Np Ns I 'qs = I qs . L's = + p ( +N ) .Np rs' = + +N . can be expressed as Vqdp = (rp + Rmp ) I qdp + pλ qdp + jω p λ qdp ' ' ' Vqds = (rs' + Rms ) I qds + pλ'qds + jω s λ'qds (3. while Figure 3. By comparing two equivalent circuits. 3. I 'ds = I ds Np Ns Lm = 3 Np L m12 . s * N ( rs .7.
we can obtain the equations Vqdp = ( rp + R mp ) I qdp + jω p ( L p − L m ) I qdp + jω p L m ( I qdp + I 'qds ) ' Vqds (3. I ABC power winding phase currents ' Vqds = T( θ s ) I 'abc . This is the situation considering the series core loss resistances. pλ qdp and pλ'qds . Using the transformations of T(θs) and T(θp) and rewriting the Equations (3. (3. .73) Under the steadystate. will equal to zero. s s According to these two equations.723. VABC power winding phase voltages I qdp = T( θ p ) I ABC.8.73) into (3.703. the derivation items.743.75). the steadystate equivalent circuit is shown in Figure 3.71). the steadystate equations for phase A and a of the power winding and control winding are expressed as VA = ( rp + R mp ) I A + jω p ( L p − L m ) I A + jω p L m ( I A + I 'a ) ' Va ( r ' + R 'ms ) ' = s I a + jω p ( L's − L m ) I 'a + jω p L m ( I A + I 'a ) . I 'abc control winding phase currents . I abc control winding phase voltages I 'qds = T( θs ) I 'abc .75) Vqdp = T ( θ p ) VABC. substitute Equations (3.74) s where s= ωs ωp = ( rs' + R 'ms ) ' I qds + jω p ( L's − L m ) I 'qds + jω p L m ( I qdp + I 'qds ) s (3.λ'qds = L s I 'qds + L m I qdp .
Using the same way.Figure 3. Steadystate equivalent circuit with series coreloss resistances. Steadystate equivalent circuit with shunt coreloss resistances. Figure 3.8. we can derive the steadystate equations and equivalent circuit for the situation considering the shunt coreloss resistances. The steadystate equations are expressed as Tp VA = Tp I A + jω p ( L p − L m ) I A + jω p L m ( I A + I 'a ) ' Ts Va T r' = s s I 'a + jω p ( L's − L m ) I 'a + jω p L m ( I A + I 'a ) s s .9.
and open circuit's voltage for control winding. the rotor is clamped in the locked position.9. After collecting these test date. Ts = R 'ms rs' + R 'ms . control winding is connected to adjustable voltage source while the power winding has no connection with any supply. when the parameters of the power windings are tested. The steadystate equivalent circuit with shunt coreloss resistance is shown in Figure 3. For example. and input power for power winding. . current. the parameters can be calculated for two kinds of steadystate equivalent circuits.Tp = R mp rp + R mp . Measurements are made of the voltage. and since rotor is blocked. To test the parameters for the control winding. According to the angular frequency relationship among power winding. 3. The control winding has no connection with supply.9 Calculation of Machine Parameters The inductance and resistance values for the steadystate equivalent circuit can be found by a set of blocked rotor tests. and an adjustable voltage source with 60Hz frequency is connected to the power winding and adjusted from 0 to the lineline rated voltage. control winding and rotor ω p − ωs = ω r .
and s = ωs =1 . Equivalent Circuit with Series Coreloss Resistances (considering turn ratio) 1. voltage. IA R mp = R p − rp Lp = 1 ωe  Z p 2 − ( rp + R mp ) 2 Lm = 2. Ia = 0 For the given input power. A.we can obtain ωr = 0. can be directly measured by using the multiple function meters. ωe I A .9). The other parameters’ calculation is based on the equivalent circuits with series coreloss resistance (Figure 3.8) or with shunt coreloss resistance (Figure 3. and current for power windings. rp and rs.  Zp  = VA = ( rp + R mp ) 2 + ( ωe L p ) 2 IA Rp = P3φ 3I 2 A = rp + R mp ωe L m = We can obtain. Opencircuit power windings. ' Va . and opencircuit voltage for control windings. ωp Winding resistances.Opencircuit control winding. ωp = ωs . IA = 0. 1 Va' .
We can obtain ' Rms = Rs' − rs' L's = 1 ωe '  Z s'  2 − ( rs' + Rms )2 . (3.10(a)(e).76) The calculation results and approximate polynomial equations for these parameters are plotted in Figures 3.217 ' Rms = 7.0373 L's = −0.085λ m + 0.0906λ m + 0.55λ2 m + 0.0164 L p = −0. These equations are given as L m = −0.44λ2 m + 0.09λ m + 0. By using these equations. .  Z s'  = Va' ' = ( rs' + Rms ) 2 + (ω e L's ) 2 ' Ia Rs' = P3φ '2 3I a ' = rs' + Rms . and current for control winding and opencircuit voltage for power winding.25λm + 1.05λ2 m − 1. voltage.58 .0385 Rmp = −4.24λ2 m − 4. the calculation results can be used to look for approximate polynomial equations relating the machine parameters to the airgap flux linkage magnitude.47λm + 1.36λ2 m + 0.For the given input power.
(e) Magnetizing inductance.(a) (b) (c) (d) (e) Figure 3.10. (d) Series coreloss resistance of the control winding. (b) Series coreloss resistance of the power winding. Machine parameters using steadystate equivalent circuits with series core loss resistances. . (c) Selfinductance of the control winding. (a) Selfinductance of the power winding.
The approximated .11.(a) (b) Figure 3. Equivalent Circuit with Shunt Coreloss Resistances (considering turn ratio) 1. and current IA. (a) For power windings. Ia = 0 To calculate parameters for the power winding. Approximate steadystate equivalent circuit for the parameter calculation. B. voltage VA.Opencircuit control winding. (b) for control windings. its input power P3φ. and control winding opencircuit voltage Va are measured.
X oA = XpR2 mp 2 R2 mp + X p . Rop and XoA can be calculated by the equations R op = P3φ 3I 2 A − rp X oA . and X m = ωe L m hence. obtained from Figure 3. a) .VA = + +I .equivalent circuit.Va' * + rp + mp p ( + j ( + = 2 2 ( 2 2 + I' ( + + Xp Rmp + Xp Rmp .Va' * + ( = rp + Rmp // jX p . Opencircuit control winding voltage will be ' Va = IA R mp jX m R mp + jX p . is used to derive the expressions of the parameters. Input reactance of the power windings is . ) 2 = rp + R op + jX oA where R op = R mp X 2 p 2 R2 mp + X p ( ) . X p = ω e L p + I' ( . a) 2 X p Rmp R X2 * .P3φ * + ( ( −+ 2 ) .11(a). A 2 . 3I A * ( ( ) 2 .9 and shown in Figure 3. ' Va = IA R mp X m 2 R2 mp + X p .
' Rms X s'2 * X s' Rms + + I' ( + rs + R '2 + X '2 ( ( =+ ( + j R '2 + X '2 ms s ) ms s . its input power P3φ. To calculate parameters for the control winding. and magnetizing inductance Lm are expressed as R mp = 2 X2 oA + R op R op R op R mp X oA ωe Xp X oA Lp = Lm = 2. R 'os and X 'oa can be calculated by the equations .11(b) is used to derive the expressions of the parameters. The approximated equivalent circuit shown in Figure 3. and current Ia. a) . voltage Va. a) ' '2 . Input reactance of the control winding is .Opencircuit power winding.coreloss resistance Rmp. selfinductance Lp.Va' * . and power winding opencircuit voltage VA are measured. X s = ω e Ls .Va' * ' ' ' ' ' + + I' ( ( = rs + Rms // jXs . IA = 0 ' 1 Va ωe I A . 2 = rs' + R 'os + jX 'oa ( ) where R 'os = R 'ms X 's2 2 '2 R 'ms + Xs . X 'oa = ' '2 Xs R ms 2 '2 R 'ms + Xs .
V' = + 'a +I . By using these equations. the calculation results can be used to look for approximate polynomial equations relating the control winding's parameters to the airgap flux linkage magnitude. Hence. VA I 'a = R 'ms X m 2 '2 R 'ms + Xs coreloss resistance R 'ms . and magnetizing inductance L m are expressed as R 'ms = 2 2 X 'oa + R 'os R 'os R 'os R 'ms X 'oa ω e ' Xs L's = Lm = 1 VA ωe I 'a X 'oa . a P * ( − + 3φ + 3I '2 ( . These equations are given as .R 'os = P3φ 3I 'a2 2 − rs' X 'oa . a ) * ( ( ) 2 opencircuit power winding voltage will be VA = I 'a R 'ms jX m R 'ms + jX 'ms . and X m = ωe L m . selfinductance L's .
(c) The magnetizing inductance. (d) The shunt coreloss resistances for the power winding and control winding. .12.(a) (b) (c) (d) Figure 3. (b) Selfinductance of the control winding. Machine parameters using steadystate equivalent circuit with shunt core loss resistances. (a) Selfinductance of the power winding.
in addition to the energy conservation law.091λ m + 0.2 . and timeinvariant circuits is generally not applicable to circuits or systems with timevarying or nonlinear inductances (frequency converters). By comparing two groups of the machine parameters calculated from two different steadystate equivalent circuits.2λm + 88.085λ m + 0. control winding selfinductance L's . 3.12(a)(d). . R mp ) are smaller than the shunt core loss resistances. The calculation results and approximate polynomial equations for these parameters are plotted in Figures 3.10 ManleyRowe Relationships The concept of power conservation at each frequency which governs the operation of a multifrequency. governed by further principles which control the distribution of the energy exchanged between the nonlinear or timevarying inductances and the remaining part of the circuit or system.552λ2 m + 0.35λ2 m + 0. but series core loss resistances ( R 'ms . dissipationless.77) Rmp = −5022λ2 m + 1865λm + 132 ' Rms = −8218λ2 m + 2014. the powers delivered to or drawn from these nonlinear or timevarying inductances by the individual spectral frequency components of the operating signals are.145λ m + 0.0385 (3.0327 L's = −0. and magnetizing inductance L m are very close.0164 L p = −1. In such circuits.44λ2 m + 0. we can see that power winding selfinductance Lp.L m = −0. linear.
.% ∞ ∞ ∞ ∞ (3. .82) . These relations were applied to the first time by Penfield [59] and more recently by Russell and Pickup to elucidate the real power transfer mechanism from the stator circuit to the rotor circuit and the shaft [60].78) ' nP( mω 1 + nω 2 ) $ "=0 mω 1 + nω 2 # m = −∞ n = 0 & . The complexvariable dq voltage equations of the doubly fed machine are given as Vqdp = rpIqdp + pλqdp + jωpλqdp Vqds = rsIqds + pλqds + jωsλqds .Power relations for nonlinear resistive and storage elements excited by independent harmonic signals of different frequencies were published by ManleyRowe [57] and Pantell [58].783. The ManleyRowe average real power/frequency relationships for a dissipationless circuit with nonlinear or timevarying inductances excited by two sources of independent angular frequencies (ω1 and ω2) are expressed as [5760] ' mP( mω 1 + nω 2 ) $ "=0 mω 1 + nω 2 # m = 0 n = −∞& . (3. Equations (3.79) where P(mω1+nω2) is the average real power flow due to a signal with an angular frequency of (mω1+nω2).80) (3.% (3.81) The apparent input power equations for the power and control winding circuits are given as 3 3 * * * * Vqdp I qdp = rp I qdp I qdp + pλ qdp I qdp + jω p λ qdp I qdp 2 2 [ ] (3.79) do not guarantee that the different frequencies they generate exist in the circuit under consideration and they give no information about the frequency components generated by the nonlinear or timevarying inductances (which are frequency converters) in the circuit.
Assuming that power input into the timevarying inductances is positive and output power is negative.3 3 * * * * Vqds I qds = rs I qds I qds + pλ qds I qds + jω s λ qds I qds . The independent two angular frequency are ω1 and zω2 and the dependent angular frequency is ω1 + zω2.84) (3. These relationships are only applicable to the input and output real powers (to and from the airgap) of the timevarying mutual inductances between the power and control windings.85) The ManleyRowe real power/frequency relationships are now applied to determine the active power distribution of the doubly fed reluctance machines.83) and are expressed as 3 P( ω1 ) = Re al[ jω1λ qdp I* qdp ] 2 3 P( ω2 ) = Re al[ jω2 λ qds I* qds ] . which is the angular frequency of the airgap power converted to the developed mechanical power. From Equations .783. the contribution of the control winding circuit is P(zω2) and P(ω1+ω2) is the active power converted to mechanical power to produce the electromechanical torque.86) (3.79) results in P( ω1 ) P( ω1 + zω2 ) + =0 ω1 ω1 + zω2 P( zω2 ) P( ω1 + zω2 ) + =0 zω2 ω1 + zω2 (3. use of Equations (3. 2 2 [ ] (3. 2 (3.87) where P(ω1) is the real power contributed by the power winding circuit to the airgap.823. the real powers sent by the power and control windings across the airgap to the shaft at steadystate are from equations (3.83) Also.
88) (3. if the control winding circuit is used to affect a small change of shaft speed around the synchronous speed of the power winding frequency in such applications as pumps and compressors. the following observations are deduced P( zω2 ) zω2 = =σ ω1 P( ω1 ) P( ω1 ) 1 P(zω2 ) σ . a relatively small rated inverter is required (compared to the power winding power requirement) confirming the experimental observation in [30].89) which are deduced from Equations (3.89) The ratio of the contributions of the control and power windings to the developed mechanical power is the same as the ratio of their source frequencies. Hence.(3. Airgap power contribution of power (Pp) and control (Ps) windings normalized with developed mechanical power as a function of control winding frequency.883.87). . = = P( ω1 + zω2 ) 1 + σ P( ω1 + zω2 ) 1 + σ (3.863.13. the Figure 3. . However. in highperformance drive applications where extended speed range of operation is desirable.
we have Te = P(ω1 + zω 2 ) ω rm = P(ω1 ) ω r ω1 ω rm = P (ω1 ) ω1 ( p1 + q ) = P ( zω 2 ) ω r P ( zω 2 ) = ( p1 + q) . = ( p1 + q) Re al jλqds I qds ( p1 + q) Re al jλqdp I qdp 2 2 [ ] [ ] (3. the electromagnetic torque is given as Te = 3 3 * * .13 and Equation (3.85).843. Figure 3. the relative active power contribution of the power winding decreases to the point when it equals that of the control winding when ω2 = ω1 after which frequency the control winding active power contribution predominates.89) show the contributions of the control and power winding circuits to the developed mechanical power. ωrm. zω 2 ω rm zω 2 (3.power requirement of the inverter feeding the control winding circuit may indeed exceed that of the power winding circuit when ω2 is greater than ω1.91) .863. The electromagnetic torque of the machine is defined as the ratio of the developed mechanical power to the shaft speed. the real power transferred from the power and control winding circuits across the airgap are given by equations (3. Hence. From Equations (3. Hence.90) Since power is invariant to reference frame transformations. We can see that with increase of control winding frequency.87). a doubly fed reluctance motor drive requiring a largespeed operation range is likely to be very expensive in view of the increase of inverter active power rating.
93) (3. λqm.99) λqm ' ' 1 λ $ 1 1 $ 'λ =% + ' + " % qs + qp " ' % L!p Lls Lm # " % " & & Lls L!p # −1 (3. dynamic response.3.98) 3 (K1 p1 + K 2q ) λdp I qp − λqp I dp 2 −1 [ ] (3. λdp. 1 ) Te = (3. θ p + θs − θ r = 0 (3.101).1 * " r = Te − TL J+ + p +q( (ω . λ’ds.923. λ’qs.102) To obtain Equations (3.100) λdm ' ' 1 λ $ 1 1 $ 'λ =% + ' + " % qs + dp " ' % L!p Lls Lm " % Lls L!p " & # & # (3.96) (3. flux linkage λdm.92) (3.100) and (3.95) (3.102) with assumed power and control winding voltage sources.97) ω p + ω s = ω r = ( p1 + q )ω rm " −ω λ Vqp = rp I qp + λ qp p dp " −ω λ Vdp = rp I dp + λ dp p qp ' " ' − ω λ' Vqs = rs' I qs + λ qs s ds ' " ' + ω λ' Vds = rs' I ds + λ ds s qs .11 Transient and Oscillatory Behavior The starting transient. and waveforms of the machine operating as a motor can be simulated using Equations (3. and λqp are defined as ' λ dm = Lm ( I dp + I ds ) ' λ qm = Lm ( I qp + I qs ) .101) 2 2 λ2 m = λqm + λdm (3.94) (3.
We define the inductance leakage as L p − Lm = Llp . Iqp. According these expressions. L p L's − L2 m can be transferred as ' ' L p L's − L2 m = Lls Llp + Lm ( Llp + Lls ) . I’ds. the current Idp. By substituting the expressions of Idp.λ qp = L p I qp + L m I 'qs λ dp = L p I dp + L m I 'ds λ'qs = L's I 'qs + L m I qp λ'ds = L's I 'ds + L m I dp . Hence. L's − Lm = L'ls . Iqp. we can obtain λ qm = Lm Llp Lm L'ls λ'qs + λ qp ' ' ' Llp Lls + Lm ( Llp + Lls ) Llp Lls + Lm ( Llp + L'ls ) . and I’qs into the expressions of λdm and λqm. and I’qs can be represented by the flux linkage and are expressed as follows I qp = λ qp L's − Lm λ'qs L p L's − L2 m I dp = λ dp L's − Lm λ'ds L p L's − L2 m ' = I qs λ'qs L p − Lm λ qp L p L's − L2 m ' = I ds λ'ds L p − Lm λ dp L p L's − L2 m . I’ds.
Tb = . However. Equations (3100) and (3101) are obtained from above derivation. When these constraints are applied to Equations (3. D = L p L's − L2 m.1003. 3.= 1 1 ' 1 1 ⋅ λ + ⋅ λ 1 1 1 L'ls qs 1 1 1 Llp qp + + ' + + ' Lm Llp Lls Lm Llp Lls and λ dm = Lm Llp Lm L'ls λ'ds + λ dp ' ' ' Llp Lls + Lm ( Llp + Lls ) Llp Lls + Lm ( Llp + L'ls ) = 1 1 ' 1 1 ⋅ λ + ⋅ λ . Equation (3.93. 1 1 1 L'ls ds 1 1 1 Llp dp + + + + Lm Llp L'ls Lm Llp L'ls Finally.95. it is required to determine either the reference frame angle of the control or power winding circuit and to properly account for saturation effect.103) results from which ωp can be determined ωp = − ( ωA + ω B ) L !s λ qp + L !p λ qs (3.102). These are achieved by the alignment of the airgap flux linkage on the qaxis such that the daxis airgap flux linkage and its derivative are forced to be zero every time. 3.103) where ω A = Vdp L !s + Vds L !p − L !p ω r λ qs ω B = λ dp (Ta − L2 !s Tb ) Ta = L !p rs' L m − L'!s [ rp L's + rs' L p ] + D D rp L m DL !p .
when the frequency of the inverter is ramped to 15 Hz and kept constant at that frequency as shown in Figure 3. Figures 3. the rotor experiences a bounded oscillation around the average speed. It is significant to note that after reaching a steadystate operating condition.16(a) and 3. the control windings are shorted after which the inverter connected to the control winding circuit has its frequency ramped linearly and levels off at 10Hz. a new operating speed is found with a lower average speed and a different frequency of rotor speed oscillation. During the initial starting period. The lower sideband current components of the control and power windings give rise to a negative damping torque leading to the machine oscillatory rotor motion. After some time. A jump phenomenon has occurred.Figures 3.15 show the simulated noload starting transient of the experimental machine in which the control winding circuit is supplied with a threephase inverter with a constant Volts/Hz control scheme.14 and 3.17(b) show the significant power spectra of the power winding current confirming the presence of the sideband current components when the control winding frequency is 15Hz. These simulation results have been confirmed in the laboratory.17(a) present experimental power winding current waveforms when inverter frequency is 10HZ and 15HZ. the rotor speed oscillatory component is continuously increasing.16(b) and Figure 3. Figure 3. .15 (showing only the steadystate waveform). However.
14. rotor speed. (b) The electromagnetic torque vs. (a) The rotor speed. rotor speed.) (a) (b) Figure 3.15.5 Wrm*100(RPM) 10 5 0 5 10 time(Sec.m) (a) (b) Figure 3. (b) Electromagnetic torque vs.5 5 5.5 4 4. Noload transient of rotor speed and torque when inverter frequency = 15 Hz. (a) The rotor speed of noload.12 10 Wrm*100(RPM) 8 6 4 2 0 0 1 2 3 4 5 time(Sec.) Te(N. . 13 12 11 10 9 8 3 13 12 11 10 9 8 15 Wrm*100(RPM) 3. Noload transient of rotor speed and torque when inverter frequency = 10 Hz.
(b) Power spectra of the power winding current.16. . Experimental waveforms showing machine oscillatory motion with inverter frequency=10Hz. (a) Phase 'a' power winding current. (b) Power spectra of the power winding current.(a) (b) Figure 3. Experimental waveforms showing machine oscillatory motion with inverter frequency=15Hz. (a) Phase 'a' power winding current. (a) (b) Figure 3.17.
107) Power Windings VAp IAp VBp IBp VCp ICp Control Windings Ias Ibs Vas Vbs Vcs + Ics V dc Figure 3. the doublyfed reluctance machine is running in synchronous operation condition.105) (3. Schematic diagram of the doublyfed synchronous reluctance motor with DC excitation.18. .104) (3.3.106) (3. The power winding and control winding qd equations of the doublyfed synchronous reluctance machine in the rotor reference frame are: Vqp = R p I qp + pλqp + ω p λdp Vdp = R p I dp + pλdp − ω p λqp ' ' Vqs = Rs' I qs + pλ'qs ' ' Vds = Rs' I ds + pλ'ds (3.18. The detailed information about the synchronous operation is referred to in Chapters 5 and 7.12 Synchronous Motor With the power windings connected to a balanced threephase voltage source having a frequency of ωp and the control winding connected to a DC voltage source as shown in Figure 3.
110) (3.where ' λqp = L p I qp + Lm I qs ' λdp = L p I dp + Lm I ds ' λ'qs = L's I qs + Lm I qp ' λ'ds = L's I ds + Lm I dp .109) (3.115) into (3.116) Substituting equation (3.116) yields 2 Ca ⋅ I qp + Cb ⋅ I qp + Cc = 0 (3. (3.115) (3.114) and (3.113) Vqp = R p I qp − ω p L p I dp Vdp = R p I dp − ω p ( Lp I qp + Lm I s' ) The peak value of power winding voltage is expressed as 2 2 2 Vm = Vqp + Vdp (3.114) (3. (3. Is = Ns ' N I qs = s I s' Np Np The following equations are obtained .112) Under the steadystate condition and considering the following conditions: ' I ds = 0.2) (1.117) Where 2 2 Ca = ω 2 p Lp + Rp .111) Torque equation is expressed as 3* ' ' Te = + (( p1 + q ) Lm ( I dp I qs − I qp I ds ) .108) (3.2N p * Te I dp = + + 3N ( ( ( p + q) L I s ) 1 m s .
The .117) and expressed as I qp = and I qp = − Cb + Cb2 − 4CaCc 2Ca (3.119) Only Equation (3. The phase voltage and current of power windings are I pa = 2 2 I dp + I qp 2 2 2 Vdp + Vqp V pa = 2 .118) − Cb − Cb2 − 4Ca Cc 2Ca . (3.' Cb = 2ω 2 p Lm L p I s 2 Cc = (ω p L p I dp ) 2 + ( R p I dp − ω p Lm I s' ) 2 − Vm Then Iqp can be solved using equation (3. So By applying a given constant load to the shaft of the synchronous motor and varying the control winding current from underexcitation to overexcitation and recording the power winding current at each step.118) is used in the calculation because the solution of Equation (3. the curves of Figure 3. The active power and apparent power as well as power factor can be computed using following equations: Po = 3 ( I qpVqp + I dpVdp ) 2 So = 3V pa I pa pf = Po .119) leads to poorer power factor results.19(a) are obtained.
m Iq p (A) 2 4 T e=2N. (a) Motor power factor vs. Figure 3. (c) power winding current vs. As load is applied. and 3).8 2 3 Te = 8 N. the power factor is plotted against the dc control winding current for various given loads.19(c) show the curves of power winding input power against the control winding current.m 2 4 Is (A ) 6 8 8 .m 2 3 Te = 4 N.19. 2.m 4 Te=4N.m.2 0 200 Te = 4 N.m 0 2 4 Is (A ) 6 8 0 0 2 4 Is (A ) 6 8 (a) 8 (b) 0 1 6 Ipa (A ) Te = 8 N. Doublyfed synchronous reluctance motor characteristics.m Te = 2 N.10 0 0 0 2 4 Is (A) 6 8 (b) (d) Figure 3. increase the control winding current.19(b).m Te = 8 N. (b) motor power vs. not only does the power winding current rise.power winding phase current is plotted against the dc control winding current for 2 N.m Te = 2 N. power winding current.6 Pf 0 . 4 N.m. control winding current.m 0 .4 400 Te = 4 N. respectively. but is also necessary to 1 800 1 0 .m 6 2 3 2 1 Te=8N.m Te = 2 N. control winding current. As shown in Figure 3. and 8 N. (d) qaxis current of power windings vs. control winding current.m load torque values. Note that both sets of curves show that a slightly increased control winding current is required to produce unite power factor as the load is increased (points 1. Note that slight input power increment .m P o (W ) 600 0 .
which consider the coreloss of the machine. They also will be used in the following chapters to investigate the steadystate characteristics and dynamic performance of some application systems with the doublyfed reluctance machine. and 3). . The machine inherent parameters are obtained by using the steadystate equivalent circuits and experimental test. where we can see that the unity power factor is achieved when the maximum qaxis current of power windings is obtained. The qd equivalent circuits with series coreloss resistance or shunt coreloss resistance are given. The ManleyRowe powerfrequency relationships are used to determine the relative contribution of each stator winding circuit to the developed mechanical power. These parameters and the model are successfully used to exposure the inherent oscillatory instability of the doubly fed reluctance machine in the motoring mode by computer simulation and experiment. Figure 3. 2. the higher the magnitude of maximum qaxis current of power windings (see points 1.19(d) shows the cures of qaxis current of power windings against the control winding current. the higher the load torque.for each load value is necessary to balance the losses of increment of control winding and power winding as increment.13 Conclusion This chapter presents an accurate model of the doublyfed reluctance machine. We also note that. 3.
there is no need for a special starting arrangement. (5) The dualwinding machine can be use as a generator for windpower transfer. for example) and reactive power provided to the load using a DC/AC PWM inverter in the secondary winding as the rotor speed ωr varies.CHAPTER 4 PERFORMANCE CHARACTERISTICS OF DOUBLYFED RELUCTANCE GENERATOR 4. With 73 . an automatically regulated load frequency is achieved if two stator windings are connected in series. (2) As the reluctance generator is run with a primemover. So the dualwinding reluctance generators have a better controllability than a squirrelcage induction generator that required a rectifierDC/ACPWM inverter system to control the load voltage with no facility to augment the power provided by the shaft to meet excessive load demand. the frequency ωp of the generated voltage in power winding can be controlled by regulating ωs and active power (from battery source.1 Introduction The idea of using doublyfed reluctance machines as generator was inspired by the fact [50. solar system. (3) In standalone generator applications with regulated turbine speed. 40. (4) According to the frequency relationship (ωp + ωs = ωr) between the power winding and control winding. 36]: (1) These machines can run at high speeds where the efficiencies of primemover (turbines) are relatively high.
1a. In Figure 4.3 give the derivation of the generator load model and the generator steadystate model. Conclusions are drawn in Section 4.wind as the power source. jθ p .5. Section 4. power winding is connected to a rectifier that feeds an impedance load and control winding has the same connection as the system in Figure 4. 4.1b. the frequency and current in the control windings are manipulated using currentregulated voltagesource. Two standalone generator systems considered in this chapter are shown schematically in Figure 4. The phase voltage and current for phase A of the power winding circuit are defined as VAp = V p cos(θ p + α pv ) = Re al[Vqdp e I Ap = I p cos(θ p + α pv ) = Re al[ I qdp e jθ p ] (4.1a. Calculated.2 and 4. pulsewidthmodulated (VSIPWM) inverters to track the power speed profile of the wind turbine for maximum power capture. 3capacitors are connected across the power and control windings to provide reactive power to the generator.1. This chapter explores the use of a dualwinding reluctance machine as an autonomous generator system in which reactive power is supplied to sustain the load. the generator feeds balanced 3phase load impedance.1) ] .6.2 Model of Generator loads The phase voltages and phase currents are balanced.4 and 4. measured generator steadystate characteristic curves and waveforms are included in Section 4. In Figure 4. The dynamical simulation results of the selfexcitation and deexcitation are given in Section 4.7.
1(a) are expressed as pVas = − I as Cq (4. The phase voltage equations of the load and capacitor circuits from Figure 4. (b) with a loaded 3phase diode rectifier (Bottom). the phasea voltage and current of the control winding are giving as Vas = Vs cos(θ s + α sv ) = Re al[Vqds e jθ s ] (4. (a)With impedance load (Top). Schematic diagram of standalone generator systems.Figure 4.2) I as = I s cos(θ s + α sI ) = Re al[ I qds e jθ s ].3) . Similarly.1.
the phaseA capacitor voltage. +N ( .4) (4.pVAP = − 1 ( I AP + I Ao ) CP (4. I Br = SbI I d . s) N * ' I qds = + s ( I qds . p) 2 . voltage and current equations of the rectifier and connected load are pVAP = − 1 ( I AP + I Ar ) Cq (4.1) and (4.9) I Ar = S aI I d . . I cr = S cI I d (4.6) pVqdp = 1 I qdp − jω pVqdp Cp (4. the following complexform equations are obtained: ' pVqds =− ' I qds ' Cq ' − jω sVqds (4.2) into Equations (4.8) where N * ' Cq = + s ( Cq .Np * ' Vqds =+ +N ( (Vqdp .11) (4.1(b).7) Vqdp = Lo pI qdp + Ro I qdp + jω p I qdo Lo (4.5) VAP = Lo pI Ao + Ro I Ao . Substituting Equations (4.10) Vd = S avVAp + SbvVBp + ScvVVp pVcd = 1 (Id − Io ) Cd (4.3)(4.13) .12) pI o = 1 (Vcd − Ro I o ) Lo (4.5) and using harmonic balance principles. p) From Figure 4. +N ( .
18) Vd = Re al (Vqdp S qdv ) (4. (4.19) and (4.20): .14) Assuming that the filter inductor current and filter capacitor voltage are DC quantities.15) (4. S av = A cos(θ p + α pv ) SbI = A cos(θ p − β + α sv ).15)(4.16) (4.20) We will show that the loaded rectifier (based on the fundamental component analysis) appears like a resistor (RL) connected across the capacitor Cq (see Figure 4.2). S cv = A cos(θ p + β + α pv ) where A= 2 3 π .19) where S qdI = Ae jα sv .2).17) ScI = A cos(θ p + β + α sv ).11) and applying harmonic balance technique. S qdv = Ae jα pv .9)(4. (4. the switching functions of the rectifier are approximated by their fundamental components and given as S aI = A cos(θ p + α sv ). 3 With Equations (4. and (4.21) The following equations are obtained from Equations (4. the following equations result: pVqdp = 1 ( I qdp − S qdI I d ) + jω pVqdp Cp (4. (4. Ld (4. Sbv = A cos(θ p − β + α pv ) (4. β= 2π .17) substituted in Equations (4.pI d = 1 (Vd − Vcd ) . The effective resistance is given as RL = π2 12 Ro .1).
Finally. the expression of the effective resistant RL is obtained in Equation (4. Now we can obtain RL = Ro .21). Vd = Vcd . Hence. S dI = A sin α sv S qv = A cos α pv . Vd = S qvVqp + S dvVdp = RL ( S qv S qI + S dv S dI ) I d and Vd = Vcd = Ro I o = Ro I L . Under steadystate condition. Vcd = Ro I o Vqp = I qr RL = S qI I d RL Vdv = I dr RL = S dI I d RL . I d = I o . S qv S qI + S dv S dI Because αsv are approximately equal to αpv we can assume S qv = S qI and S dv = S dI to obtain: S qv S qI + S dv S dI = A2 = 12 π2 . S dv = A sin α pv I qr = S qI I d . .S qI = A cos α sv . I dr = S dI I d Vd = S qvVqp + S dvVdp .
26) pVqp = 1 (− I qp − I qo ) − ω pVdp Cp 1 (− I dp − I do ) + ω pVqp Cp pVdp = (4. 4.22) (4.2.3 SteadyState Generator Model The generator system in Figures 4.27) .Rectifier IAP Ld IAr Id Vd Vcd Ro Io Lo Cp IAr RL Figure 4.1(a) and 4.1(b) can be described by following qd equations: Vqp = R p I qp + pλqp + ω p λdp Vdp = R p I dp + pλdp − ω p λqp ' ' Vqs = Rs' I qs + pλ'qs − ω s λ'ds ' ' Vds = Rs' I ds + pλ'ds + ω s λ'qs (4.25) (4.24) (4. Effective resistance RL.23) (4.
35).39) .38) I dp = (4.30) 1 ' ' I qs + ω sVds Cq ' pVds =− 1 ' ' I ds − ω sVqs Cq (4.36) (4. Since some parameters of machine are dependent on the airgap flux. the steadystate system equations are easily solved if the current state variables are replaced with flux linkages.34) (4.31) where ' λqp = L p I qp + Lm I qs ' λds = L p I ds + Lm I ds ' λ'qs = L's I qs + Lm I qp ' λ'ds = L's I ds + Lm I dp (4.37) R p = rp + Rcp ' Rs' = rs' + Rcs .Lo pI qo + Ro I qo + ω p Lo I do = Vqp Lo pI do + Ro I do − ω p Lo I qo = Vdp ' pVqs =− (4. By using Equations (4.28) (4.34) and (4.35) (4.33) (4. the current states can be expressed with flux linkages as I qp = (λqp L's − Lm λ'qs ) ( L p L's − L2 m) (λdp L's − Lm λ'ds ) ( L p L's − L2 m) (4.32) (4.29) (4.
43) (4.47) pI qo = Vqp Lo − Ro I qo Lo − ω p I do (4. Lo (4. (4.50) (4.50)(4.23)(4.45) (4.48) pI do = Vdp Lo − Ro I do − ω p I qo .42)(4.' I qs = (λ'qs L p − λqp Lm ) ( L p L's − L2 m) (λ'ds L p − λdp Lm ) ( L p L's − L2 m) (4.53): pλqdp = Vqdp + jω p λqdp + Tss1λqdp + Bss1λ'qds ' pλ'qds = Vqds − jω s λ'qds + Tss 2 λ'qds + Bss 2λqdsp (4.53) C p pVqdp = (Ts1λqdp + Bs1λ'qds − I qdo ) + jω p C pVqdp Lo pI qdo = (Vqdp − Ro I qdo ) + jω p Lo I qdo .46) pVqp = 1 (Ts1λqp + Bs1λ'qs − I qo ) − ω pVdp Cp 1 (Ts1λdp + Bs1λ'ds − I do ) + ω pVqp Cp pVdp = (4.38)(4.49) are expressed as Equations (4.44) (4.40) ' = I ds .41) into Equations (4.31).41) Substituting Equations (4.42) (4.51) (4.49) The complex form for Equations (4. the following equations with flux linkages as state variables are obtained: pλqp = Vqp − ω p λdp + Tss1λqp + Bss1λ'qs pλdp = Vdp + ω p λqp + Tss1λdp + Bss1λ'ds ' pλ'qs = Vqs + ω s λ'ds + Tss 2λ'qs + Bss 2 λqp ' pλ'ds = Vds − ω s λ'qs + Tss 2λ'ds + Bss 2 λds (4.52) (4.
55) (4.61) are obtained from Equations (4.56) (4.58) (4. L p L's − L2 m Bs1 = Lm L p L's − L2 m Ts 2 = − Lp Lp L − L ' s 2 m . (4. Bss1 = Tss 2 = − L L −L ' p s 2 m .61) .' ' CqVqds = (Ts 2λ'qds + Bs 2 λqdp ) − jω sVqds .54) are equal to zero.58)(4.59) 0 = (Ts1λqdp + Bs1λ'qds − I qdo ) + jω pC pVqdp 0 = (Vqdp − Ro I qdo ) + jω p Lo I qdo ' 0 = (Ts 2λ'qds + Bs 2λqdp ) − jω sVqds where Ts1 = − L's . (4.59) and expressed as I qdo = Vqdp ( Ro − jω p Lo ) (Ts 2λ'qds + Bs 2 λqdp ) (4. L p L's − L2 m Tss1 = − R p L's L L −L ' p s 1 Rs Lp 2 m . B s2 = Lm L p L's − L2 m R p Lm L p L's − L2 m Rs' Lm .54) In the steadystate.50)(4.60) ' Vqds = jω s C q . all differential items of Equations (4.1(a): 0 = Vqdp + jω p λqdp + Tss1λqdp + Bss1λ'qds ' 0 = Vqds − jω s λ'qds + Tss 2λ'qds + Bss 2 λqdsp (4. Bss1 = Equations (5.57) (4.60)(5. The following equations are obtained and used to describe the steadystate performance of the generator system in Figure 4.
K 22 " # & qds # &0# (4.63) and Iqdo with Equation (4.60) into Equation (4.66) 0 = K 21λqdp + K 22λ'qds .65)(4.1(a): 0 = K11λqdp + K12 λ'qds (4.66) is ' K11 %K & 21 K12 $ 'λqdp $ '0$ % ' "=% " .60)(4.61) into Equation (4.65) (4.65)(4. (4. jω s C q K12 = Tss 2 − jω s + Ts 2 .Substituting Equations (4.66)) of the generator system in Figure 4.63) A= Ts1 ( Ro − jω p Lo ) 1 − jω p C p ( Ro − jω p Lo ) .62) K11 = Bss 2 + Bs 2 .64)) Now we obtain a complexform equation system (Equations (4.56) gives 0 = K11λqdp + K12 λ'qds where (4.60) in Equation (4.55) yields 0 = K 21λqdp + K 22λ'qds where K 21 = A + jω p + Tss1 .67) . jω s C q Substituting Equation (4. K 22 = B + Bss1 . The matrix form of Equations (4.57) yields Vqdp = Aλqdp + Bλ'qds where (4. B= Bs1 ( Ro − jω p Lo ) 1 − jω p C p ( Ro − jω p Lo ) Replacing Vqdp with Equation (4.
∆I = 0 .70) Ro = . feasible solutions are obtained only when the determinant of K matrix is set equal to zero. and C1 − C10 are listed in Appendix 4B. (4.and K 21 with corresponding expressions in Equation (4. K 22. respectively.69) ∆R and ∆I. the ∆R and ∆I both must be equal to zero. To satisfy the condition of Equation (4. That is ∆ = K11 K 22 − k12 K 21 = 0 . λqdm = Lm ( I qdp + I qds (4. (4.68) and Appendix 4A gives the expressions of ∆R and ∆I.38) and (4.As Equation (4. and C1 − C10 are determined by machine parameters and the values of the electrical components and rotor speed.68). K12. la − qa . The complexform mutual airgap flux linkage is given as ' ).39) into a complexform gives . Equations (4.70)(4.71) are obtained: 8 7 6 5 4 3 2 C1ω 9 p + C2ω p + C3ω p + C4ω p + C5ω p + C6ω p + C7ω p + C8ω p + C9ω p + C10 = 0 4 3 2 saω 5 p + t aω p + u aω p + vaω p + xaω p + ya 2 laω 3 p + maω p + naω p + qa (4.68) By replacing K11 . a complex form equation is obtained and can be separated into real part and imaginary part ∆ = ∆ R + j∆ I = 0 . represent real part and imaginary part of Equation (4. That is ∆R = 0 . The expressions of sa − ya .67) is singular. (4.72) Combining Equations (4.71) The quantities sa − ya .69). la − qa . By using the condition ∆ R = 0 .
76): 'E %K & 11 F $ 'λqdp $ 'λqdm $ % ' "=% " K12 " # & λqds # & 0 # .77) λ'qds = (4.734. Cp. L'ls = L's − Lm : The perphase effective leakage inductance of the control winding.78) With given values of ωr. (4.74) into Equation (4. the machine parameters are computed from their approximated . Lm Llp L'ls F =( Llp = L p − Lm : The perphase effective leakage inductance of the power winding. (4.78) from Equation (4.73) and the complexform from Equations (4. (4.76) Now we can obtain Equations (4.74) Substituting Equations (4.I qp = (λqdp L's − Lm λ'qds ) ( L p L's − L2 m) (4.40) and (4.75) to obtain a matrix equation (4.77) and (4. and varying the airgap flux linkage from zero to some reasonable values.62) and Equation (4. Using Equation (4.76) λqdp = K12λqdm EK12 − FK11 − K11λqdm EK12 − FK11 .75) E=( 1 1 1 + + ' ) −1 ( Llp ) −1 Lm Llp Lls 1 1 1 + + ) −1 ( L'ls ) −1 . Cq.72) yields λqdm = Eλqdp + Fλqds where (4.41) is expressed as ' I qds = (λ'qds L p − λqdp Lm ) ( L p L's − L2 m) .
The Matlab program used to solve these nonlinear system equations is listed at appendix 4C.3(b) shows that the higher the rotor speeds are. la − qa . Then the generator loses output power when the load resistor reaches its minimum limited value that is approximately equal to 12 ohms for this experimental generator system.2 and 4. Vqdp.4 SteadyState Results The mathematical derivation given in Sections 4. Vqds ./min corresponding to rotor angular electrical frequencies of 50Hz and 55Hz respectively.77). Cq = 90µF and Cp = 166µF are chosen.1(b).3(a) that output power is inversely proportional to the load resistor when the load resistor is bigger than some values. 4. Figure 4. (4. Then they are inserted in Equations (4. The maximum power points are clearly seen in this figure. and (4.58).3 are used to obtain the steadystate results of the generator systems shown in Figure 4.63). . We can see from Figure 4. and C1 − C10 can also be determined. the larger the output power and load voltage.equations and the quantities of sa − ya .61). The state variables λqdp . The results are verified experimentally by running the rotor of the generator at 750 and 825 rev.71) to determine corresponding load impedance (or load resistance) and the angular frequencies of the power and control ' winding currents.3 gives measured and calculated steadystate performance characteristics of the generator feeding an impedance load (generator system in Figure 4.70) and (4. (4. however the output power sharply decreases when the load resistor continuously reduces in the range.78).1(a)). (4.1(a) and Figure 4. and Iqdo are then calculated using Equations (4. which has the values smaller than those values. Figure 4. λ'qds .
(a) 1 0.6 (Wrm = 825 RPM) (b) 0.8 P o w e r Fa c t o r 0.4 0. (a) Output power against perphase load resistance.2 0 0 (Wrm = 750 RPM) 50 100 150 Ro(Oh ms ) 200 250 (c) (d) Figure 4. (b) load . Measured and calculated results of doublyfed reluctance generator system feeding impedance loads.3.
4. (c) load voltage against load current. (a) (b) Figure 4. . Measured generator waveforms (rotor speed =850 rev/min. load resistance 20Ω) (10ms/div). (a) Generator lineline voltage (Top)(50V/div) and Generator phase current (Bottom)(2A/div). (d) power factor against load resistance. (b) control winding lineline voltage (Top)(50V/div) and Phase current in Cq (Bottom)(5A/div).voltage against output power.
9 1 (b) Lo = 0. generator power factor.0133H 20 0 Lo = 0.0398H 20 0 15 0 10 0 50 Lo = 0. we can see that power factor will decrease with increasing load resistance and there is little affect with different rotor speeds. Figure 4.7 Lo = 0.0133H 10 0 50 55 Lo = 0.3(c).0398H 50 0 20 40  Z 60 80 100 0 0 20 40  Z 60 80 100 (a) 25 0 Lo = 0.3(d). It is observed from Figure 4.0265H 0 .0265H 15 0 Lo = 0.0265H Lo = 0. The little discrepancies between measured and calculated results may be due to the high sensitivity of the machine performance to magnetic saturation and the presence of significant current spaceharmonic components evidenced in the waveforms shown in Figure 4. In Figure 4. Increasing inductance value causes the increment of outout voltage.6 0 20 40 P o w e r Fa c to r 0 .0133H 0 0 2 4 Io (A) 6 8 10 0 .0133H 60  Z 80 100 (c) (d) .4. in which the maximum load current points are easily obtained.0398H Lo = 0.3(a)(c) that there is good correlation between measurement and calculation results.The relationships between the load voltages and currents are illustrated in Figure 4.8 Lo = 0.0398H Vo (V) 0 . and 65 25 0 Lo = 0.5 shows the calculated generator characteristics of selfexciting generator system feeding an impedance load with three different inductance values.0265H Vo (V) 60 fp (HZ) Lo = 0.
load current. load impedance. (Cp=65uf.0265H 15 0 0 10 0 0 Lo = 0.5. (b) load voltage vs.0398H Lo = 0. load impedance. (a) (b) . Calculated generator characteristics of selfexciting generator system feeding impedance load. (c) load voltage vs. Wrm = 1800RPM) (a) Power winding frequency vs. (e) output power vs.25 0 0 20 0 0 P o (W) Lo = 0.0133H 50 0 0 0 20 40 60 80 100 (e)  Z Figure 4. load impedance. (d) generator power factor vs. Cq =45uf. load impedance.
. (a) Load voltage against output power.8 P o w e r Fa c t o r 0. (b) generator voltage against output power.6 (Wrm = 825 RPM) 0.7. (a) (b) Figure 4. (d) power factor against load resistance.2 0 0 (Wrm = 750 RPM) 100 200 Ro(Oh ms ) 300 (c) (d) Figure 4. (a) Generator lineline voltage (Top)(50V/div) and generator phase current (Bottom)(1A/div).1 0. (c) output voltage against load current.6. (b) output rectifier voltage (Top)(50V/div) and Input rectifier current (bottom)(1A/div).4 0. Measured waveforms of generator system feeding a rectifier load. Measured and calculated results of generator system feeding a loaded rectifier (ωrm = 825 rev/min and ωrm = 750 rev/min).
the larger the output power.1(b) are distributed to the power and control windings.6(a)(d) are very similar to those in Figures 4.6 shows measured and calculated performance characteristics with the generator feeding a rectifier having a resistive load.1(a) and Figure 4. The correspondence experiment and calculation results are fairly good in view of the harmonics imposed on the generator voltages and currents due to the switching diodes shown in Figure 4.output power.5(a). respectively. and (e). as shown in Figures 4.4(a)(c). Two rotor speeds: 725rev/min and 825rev/min. are choose for the steadystate calculation and experimental measurement.5(b).5 Power Capability and Parametric Analysis The active power supplied by the source of mechanical power to the generator systems in Figure 4. The curves shown in Figures 4. (d). The higher the load inductance. we note that the frequency of power winding voltage and current increases with increasing the impedance value of load. The distributions of the active power across the airgaps determined by the ManleyRowe power/frequency relationships are given as .5 (b). 4. We can see that the higher the rotor speed is. load voltage and load current. as shown in Figure 4.7. because a rectifier feeding a resistive load is equivalent to a resistive load on the fundamental component basis. Figure 4. the higher the output maximun voltage. The maximum powers and load currents relating to two rotor speeds can be obtained in these figures. In Figure 4.
the following relationships are derivable from Equation 4. It is evident from equation ωrm = ωr/(p1+q) that P (ωp +ωs) is the input mechanical shaft power. Assuming that power input to airgap is positive and power received from the airgap is negative. = ω p + ωs 1 + σ (a) (b) . Note that P (ωp +ωs) is the power input to the airgap by a source with angular frequency given by ωp +ωs.79: P(ω s ) ω s = =σ P(ω p ) ω p P(ω p ) P(ω p + ω s ) = ωp 1 .79) ωs + ω p + ωs =0 .P(ω p ) ωp P(ω s ) + P (ω p + ω s ) ω p + ωs P(ω p + ω s ) =0 (4.
the ratio of power transferred across the airgap to the power winding to meet load demand.8. it is inferred that smaller the angular frequency of the control winding. °: Cpq = 0. Hence. the frequency of the inverter source voltage . +: Cpq = 0.8. power winding core and copper losses is directly proportional to the ratio of the power winding frequency and angular rotor speeds. (c) load voltage against relative load frequency. (: Cpq = 0.6. Parametric characteristics of generator feeding a resistive load rotor speed = 900 rev/min.9). If an inverter is connected to the control winding. Furthermore. the higher the percentage of shaft power that can be utilized to meet load demand.4. In standalone generator applications operated by regulated turbines requiring regulated load frequency. (b) load voltage against load power.(c) Figure 4. ×: Cpq = 0. (a) Relative control winding frequency against load frequency. capacitors Cq and Cp must be chosen to make ωp to be slightly lower than the rotor angular speed.
The system are described using following equations: ωp + ωs =(p1+q) ωrm (4.81) Vqp = Rqp I qp + pλqp + ω p λdp Vdp = Rdp I dp + pλdp − ω p λqp ' ' ' Vqs = Rqs I qs + pλ'qs − ω sp λ'ds (4.6. Figure 4. which are 0. 0. We also can see that the maximum load voltage curve is obtained when the capacitive ratio is 0. the maximum frequency of the power winding current is different and directly relative to the capacitive ratios. The relationship between load voltage and power winding frequency is shown in Figure 4.80) (4. where we can see that maximum output power curve is obtained when the capacitive ratio is 0.6. 4. which are shown in Figures 4. The ' steadystate analysis chose Cq=90µF (Cq = 360µF ) and rotor speed ωrm=900 rev/min. The voltage and the output power does not increase monotonically with the capacitive ratios.6.8.8(a) shows that the ratios of capacity affect the regular range of power winding speed ωp and control winding speed ωs.must also be very low to ensure that most of the active power is sent to the power winding circuit. 0.82) (4. are selected.4.8(b) shows the relationship between output power and load voltage.8 shows how the performance characteristics of the experimental machine are influenced by the selection of the capacitor Cp and Cq. and 0.1(a)) is simulated. ' then different ratios C pq = C p / Cq . Hence.83) .8(b) and 4.8(c).9. Figure 4. Figure 4.6 Simulation of SelfExcitation Process The electrical starting transient of the dualwinding reluctance generator system with an impedance load (shown in Figure 4.8(c).
91) (4.93) (4.99) .94) (4.92) (4.95) (4.89) 1 ' ' I qs + ω sVds Cq 1 ' ' I ds − ω sVqs Cq ' pVds =− (4.98) Llpq = L pq − Lmq Llpd = L pd − Lmd L'lsq = L'sq − Lmq L'lsd = L'sd − Lmd λqm = Lmmq % ' λqp % Llpq & + λ'qs $ " L'lsq # " (4.86) Lo pI qo + Ro I qo + ω p Lo I do = Vqp Lo pI do + Ro I do − ω p Lo I qo = Vdp ' pVqs =− (4.88) (4.97) (4.96) (4.87) (4.90) where ' ' λqp = Lqp I qp + Lqm I qs = Llpq I qp + Lmq ( I qp + I qs ) ' ' λdp = Ldp I ds + Ldm I ds = Llpd I dp + Lmd ( I dp + I ds ) ' ' ' λ'qs = L'qs I qs + Lqm I qp = L'lsq I qs + Lmq ( I qp + I qs ) ' ' ' λ'ds = L'ds I ds + Ldm I dp = L'lsd I ds + Lmd ( I dp + I ds ) (4.84) (4.' ' ' Vds = Rds I ds + pλ'ds + ω s λ'qs (4.85) pVqp = 1 (− I qp − I qo ) − ω pVdp Cp 1 (− I dp − I do ) + ω pVqp Cp pVdp = (4.
25λmd + 1.05λ2 md − 1.116) When the condition (4.109) Lmq = −0.81)(4.47λmd + 1.085λmd + 0.114) Lmd = −0.113) (4. The parameter variations are accounted for by choosing a reference frame speed for the power winding (ωp).108) (4. such that the total magnetizing flux linkage is aligned with qaxis.0373 L'sq = −0.103) (4.110) (4. the following Equations (4.80) given the rotor speed.0385 Rmpd = −4.105) (4.47 λmq + 1.36λ2 mq + 0.36λ2 md + 0.0906λmq + 0.0906λmd + 0.0373 L'sd = −0.085λmq + 0.115) (4. When the conditions expressed in Equations (4. These conditions are expressed as λdm = 0 pλdm = 0 .58 (4.118) are obtained .55λ2 mq + 0. If the magnetizing path is unsaturated.09λmd + 0.55λ2 md + 0.112) (4. The daxis magnetizing flux linkage and its derivative then become identically equal to zero.25λmq + 1.100) R pq = rp + Rmpq (4.58 (4. all daxis machine parameters become constant while all qaxis parameters are dependent on the qaxis magnetizing linkage.117) and (4.44λ2 mq + 0.100).115)(4.44λ2 md + 0.05λ2 mq − 1.217 ' Rmsq = 7.24λ2 mq − 4. The effect of magnetizing fluxlinkage saturation on the machine parameters must be accounted for and a means must be found to determine either ωp or ωs from Equation (4. all the machine parameters are constant but vary with magnetizing flux linkage under saturated condition.104) Rmpq = −4.115) is enforced in Equations (4.107) (4.94).111) (4. (4.0164 L pd = −0.101) (4.217 ' Rmsd = 7.24λ2 md − 4.102) R pd = rp + Rmpd ' ' Rmsq = rs' + Rmsq ' ' Rmsd = rs' + Rmsd (4.0385 .λdm = Lmmd % ' λdp % & Llpd + λ'ds $ " L'lsd " # (4.116) are used in Equation (4.0164 L pq = −0.106) (4.09λmq + 0.
118) Because the machine parameters are dependent on the magnetizing flux linkage.117) pλdp = − Llpd L'lsd (4.91)(4.125) (4.119)(4.94) yield following equation: − R pq I qp = Tppq λqp + B ppq λ'qs − R pd I dp = Tppd λdp + B ppd λ'ds where (4.80)(4.90) yield pλqp = Vqp + Tssq1λqp − ω p λdp + Bssq1λ'qs pλdp = Vdp + Tssd 1λdp + ω p λqp + Bssd 1λ'ds ' pλ'qs = Vsq + Tssq 2λ'qs − (ω p − ω r )λ'ds + Bssq 2 λqp ' pλ'ds = Vds + Tssd 2λ'ds + (ω p − ω r )λ'qs + Bssd 2λdp (4.126) (4. Bssq1 = R pq Lmq L pq L'sq − L2 mq Tssd 1 = − R pd L'sd L pd L − L ' sd 2 md .120) Tssq1 = − R pq L'sq L L −L ' pq sq 2 mq .119) (4.123) (4. To realize this purpose. Bssq 2 = Bssd 2 = ' Rsq Lmq L pq L'sq − L2 mq .122) into Equations (4. ' Lsd − Rsd .λdp = − Llpd L'lsd λ'ds pλ'ds . it is necessary to use flux linkages as state variables in generator system equations. The derivation from Equations (4. the following derivations are done.122) ' '[ − Rsq I qs = Tssq λ'qs + Bssq λqp ' − Rsd I ds = Tssd λ'ds + Bssd λdp where Tssq 2 = Tssd 2 = ' − Rsq Lqp L pq L'sq − L2 mq .127) (4. Bssd1 = R pd Lmd L pd L'sd − L2 md (4. ' L pd Lsd − L2 md ' Rsd Lmd ' L pd Lsd − L2 md Substituting Equations (4.121) (4. (4.128) pVqp = γ 1Tsq1λqp + Bsq1γ 1λ'qs − ω pVdp − γ 1I qo pVdp = γ 1Tsd 1λdp + Bsd 1γ 1λ'ds + ω pVqp − γ 1I do .124) (4.
123)(4. The simulation results given in Figure 4. Llpd Equations (4. respectively. ωp = − where (ω A + ω b ) L λqp + Llpd λ'qs ' lsd (4.129) (4.133). which brings the generator to a stable 0perating condition.pI qo = pI do = R 1 Vqp − o I qo − ω p I do Lo Lo R 1 Vdp − o I do + ω p I qo Lo Lo (4.132) and Equation (4. Bsq1 = Lmq L pq L'sq − L2 mq Tsd 1 = Tsq 2 = Tsd 2 = − L'sd .131) (4.124) and (4. L pd L'sd − L2 md − Lqp L L −L ' pq sq 2 mq Bsd 1 = Bsq 2 = Bsd 2 = Lmd L pd L'sd − L2 md Lmq L pq L'sq − L2 mq Lmd L pd L'sd − L2 md . Figure 4. Cq Substituting Equations (4. flux linkages and machine parameters. The selfexcitation process of the generator feeding an impedance load was simulated with values of Cp and Cq Selected as 168µF and 360µF.132) where Tsq1 = − L'sq L L −L ' pq sq 2 mq .116) yields Equation (4.130) ' ' pVqs = γ 2Tsq 2λ'qs + Bsq 2γ 2λqp − (ω p − ω r )Vds ' ' pVds = γ 2Tds 2λ'ds + Bsd 2γ 2λdp + (ω p − ω r )Vqs (4. The simulation results are similar to the . L pd L'sd − L2 md γ1 = 1 .126) into Equation (4. − Lsd .133) are used for the simulation of the electrical starting transient of the dualwinding reluctance generator. Cp γ2 = 1 . and a constant rotor speed of 1500 rev/min.10 gives experimental results for the selfexcitation process corresponding to the simulation results. Saturation effect limits the growth of the magnetizing flux linkage.9 display the growth of the generator terminal voltage and current as the magnetizing flux linkage builds up. in which ωp is expressed in terms of ωr.133) ' ω A = (Vdp L'lsd + Vds Llpd − Llpd ω r λ'qs ) ω B = λdp (Tssd1 L'lsd + Bssd 2 Llpd − 2 L'lsd Bssd 1 − Llpd ω r λ'qs ) .
10(a) shows the dynamic process of airgap flux linkage when the load impedance is reduced . Figure 4. (c) magnetizing flux linkage magnitude. (b) power winding phase current. This dynamic process is displayed in Figure 4.9.10.(a) (b) (c) Figure 4. experimental results. When the generator is critically loaded such as when the load impedance is reduced beyond a certain threshold value. which proves that mathematical model used in the simulation program effectively reflects the true situation. We also simulate the deexcitation phenomenon of the generator. Selfexcitation process of doublyfed reluctance generator. the load voltages collapse (deexcite) due to a rapidly reducing airgap flux linkage. (a)Power winding phase voltage.
The collapse process of the voltage and current is shown in Figures 4.11(b) and (c).10. (The simulation program is listed in Appendix 4C) Comment [GS1]: 4. A qd model of the generator is proposed that accounts for the core and harmonic losses and the influence of magnetic path saturation on the machine self and mutual inductances. (b) power winding phase current. This should find utility in the accurate prediction of the dynamic and transient performance of the generator and in the design optimization of standalone . (a) Power winding phase voltage.(a) (b) Figure 4. beyond a certain threshold value.7 Conclusions This chapter sets forth the analysis and performance prediction of a standalone dual winding reluctance generator with capacitive excitation in both the power and control windings. Experimental waveforms of selfexcitation process.
(b) power winding phase current.Vpa(V) Ipa(A) LAMM(Wb) T(sec. (c) power winding phase voltage. Measured performance characteristics compare favorably with the analysis results. doublyfed reluctance generators. (a) Airgap flux linkage.11.) Figure 4. Voltage deexcitation phenomena in the generator feeding RL load. .
1 Introduction A doublyfed reluctance machine can realize its synchronous operation when its control windings are supplied by a direct current (DC) source. the performance of doublyfed synchronous generator is investigated when feeding an impedance load and a rectifier load. pump. The machine can be working in the synchronous mode as generator or motor. refrigeration and airconditioning applications. The rotor electrical angular frequency is proportional to AC supply angular frequency of the power windings in the motoring mode based on the angular frequency relationship ωp = (p1 +q)ωrm. The frequency of load voltage generated in power winding is directly dependent on the rotor speed in the generating mode. 104 This chapter is .14].CHAPTER 5 SYNCHRONOUS OPERATION OF A DOUBLYFED RELUCTANCE GENERATOR 5. slip rings and possibility of operating the load at leading or unity power factor by controlling the excitation current. It is anticipated that the machine will also find utility as a medium to high frequency generator in standalone applications such as in airplane power systems and marine generator for gas turbine drives [7. In this chapter. The obvious advantages of this generator system include the absence of brushes. Doublyfed synchronous reluctance motors show potential in fan.
Section 5.1) (5. Schematic diagram of the doublyfed reluctance machine with DC excitation. organized as follow.3.2 gives the derivation of the machine voltage and torque equations using the concept of qd harmonic balance technique. including the effect of magnetic saturation. Finally the concluding remarks are contained in Section 5. The operation of the machine in standalone generator mode feeding an impedance load is set forth in Section 5. 5.5.1. core and harmonic losses.Power Windings VAp IAp VBp IBp VCp ICp Control Windings Ias Ibs Vas Vbs Vcs + Ics V dc Figure 5.4. The generator connected to a threephase rectifier feeding a load is analyzed in Section 5.2) .1. This section also contains comparison of experimental and simulation results. the voltage equations of the control windings are expressed as Vas = Rs I as + pλas Vbs = Rs I bs + pλbs (5.2 Machine Model With the power windings connected to a balanced threephase voltage source having a frequency of ωp and the control winding connected to a DC voltage source as shown in Figure 5.
j = Lm12 cos(θ r + kβ ). Lbc are the mutualinductances between the phasewindings of control windings. The flux linkages of the control winding are also given as: (5. c and j = A. β = 2π / 3.3) λas = Laa I as + Lab I bs + Lac I cs + LaA I Ap + LaB I Bp + I aC I Cp λbs = Lba I as + Lbb I bs + Lbc I cs + LbA I Ap + LbB I Bp + I bC I Cp (5.4) (5. Lbb . Iap.7) k = 0 for LAa = LBc = LCb k = 1 for LAb = LBa = LCc k = 2 for LAc = LCa = LBb Ls = 3 Lm12 + Lls 2 .2 (5.5) (5. Lab.1.j where i = a. The inductances are given by the following expression [25]: Laa = Lbb = Lcc = Lls + Lms Lab = Lac = Lbc = − Lms / / 2 LAA = LBB = LCC = Llp + Lmp LAB = LAC = LBC = − Lmsp / 2 Li .Ibs . The mutualinductances between the control windings and power windings are given by Li. Lcc are the selfinductances of the control windings carrying currents Ias . B. b. Ibp and Icp are the currents flowing in the power windings.6) λas = Lca I as + Lcb I bs + Lcc I cs + LcA I Ap + LcB I Bp + I cC I Cp Where Laa.Vcs = Rs I cs + pλcs . respectively. C. Lac .Ics . k = 0.
q. and Pr are the pole pair number of power windings.1). respectively.. r is the radius of stator.3 and 3. NA and Na are the winding functions of power windings and control windings (see Figures 3. flux linkages. 3 µ0 is magnetic (5.4). can be transformed into a complexform signal (fqd) and zero sequence signal (fn) using the following equations where θ is the reference frame angle: f qd = 2 − jθ e [ f a + afb + a 2 f c ] 3 2 fn = [ f a + fb + f c ] . and rotor. etc.9) a=e . permeability. l is the length of stator lamination.8) It can also be shown that the original signals can be recovered from the transformed quantities using the following expressions: f a = Re[ f qd e jθ ] + 1 fn 2 1 fb = Re[a 2 f qd e jθ ] + fn 2 1 f c = Re[af qd e jθ ] + fn 2 j 2π 3 (5. Three real signals (fa. fc) which could be threephase balanced voltages. currents. p1. 2 Lm12 is given by Lm12 = −2 Pr µ 0 rlN A N a g1 (q − p1 ) sin[( p1 + q) πα 2 pr ] where. g1 is airgap width and α is polearc width (see Figure 3.Lp = 3 Lm12 + Llp . fb. . control windings.
Vcs = Re[aVqds ] I cs = Re[ aI qds ] .3). the zero sequence current (Inp) and voltage (Vnp) are zero. which are given as V Ap = Re[Vqdp e I Ap = Re[ I qdp e jθ p ]. These phase voltages and currents are transformed into complexform equations in the stationary reference frame. VBp = Re[ a 2Vqdp e ].8) to transform the voltages and currents of the power windings with balanced threephase voltage supply to a reference frame with an angular angle θp (or angular speed of ωp). we can see Vdc = Vas − Vbs 0 = Vbs − Vcs (5. jθ p ]. (5.16) (5. The actual phase voltages and currents can be expressed in terms of the transformed complexform current (Iqdp) by using Equation (5.Using Equation (5. I bs = Re[ a 2 I qds ].18) Vas = Re[Vqds ]. which are 2 Vqds = Vdc .13) (5. 3 2 2 (5.1. 3 2 Vns = [Van + Vbn + Vcn ] 3 I ns = 0 (5.1)(5. Vbs = Re[ a 2Vqds ]. the following derivation are done (detailed derivation is listed in Appendix 5A): . I as = Re[ I qds ]. ].14) I as + I bs + I cs = 0 . VCp = Re[aVqdp e jθ p ] ]. Control Winding Mathematical Model By substituting Equation (5.9).11) jθ p I Bp = Re[ a 2 I qdp e jθ p I Cp = Re[aI qdp e jθ p From Figure 5.7) to Equations (5.12) (5.10) (5.17) (5.15) 2 3 3 I qds = [ I as + j ( I bs − I cs )].
23) (5.24) .20) 1 + e jβ + e − jβ = 0 the following expression of Vas is obtained Vas = Rs I as + p{Ls I as + 3 j (θ −θ ) * Lm12 Re( I dqp e p r )} .18) into equation (5.22) into qaxis and daxis equations (in the synchronous reference frame) gives 2 3 Vdc = Rs I qs + p ( Ls I qs + Lm12 I qp ) 3 2 0 = Rs I ds + p( Ls I ds + 3 Lm12 I dp ) . 3 3 Defining β = 2π .15)(5.21) yields the qd complexform voltage equations (in the stationary reference frame) for the control windings given as 2 3 j (θ −θ ) Vqds = Vdc = Rs I qds + p[ Ls I qds + Lm12 I qdp e p r ] .19) (5.10)(5.θ x = ( p1 + q )θ rm and using the following relationships: 3 1 Re al ( A) ⋅ Re al ( B ) = [Re al ( A ⋅ B ) + Re al ( A ⋅ B ∗ )] 2 (5. 3 2 (5. 2 (5.21) Substituting equation (5. 2 (5.22) Extending equation (5.11) and (5.Vas = Rs I as + pλ as = Rs I as + p[ Laa I as + Lab I bs + Lac I cs + LaA I Ap + LaB I Bp + I aC I Cp ] 3 = Rs I as + p{( Lm12 + Lls ) I as + Lm12 cos[( p1 + q )θ rm ]I As 2 + cos[( p1 + q)θ rm − 2π 2π ]I Bs + cos[( p1 + q )θ rm + ]I Cs } .
30) .26) Using Equation (5.25) λ Ap = LAA I Ap + LAB I Bp + LAC I Cp + LAa I as + LAb I bs + LAc I cs .Power Winding Mathematical Model The phase ‘A’ voltage of balanced threephase power windings is also given as (The detailed derivation is obtained in Appendix 5B): VAp = R p I Ap + pλ Ap where the flux linkage is given in Equation (5.27) = ( R p I Ap + L p pI Ap ) + Lm12 p{Re(e jθ x ) I as + Re(e jθ x e − jβ ) I bs + Re(e jθ x e ) I cs } .26) and then substituting Equation (5.25) yield VAp = R p I Ap + p[ LAA I Ap + LAB I Bp + LAC I Cp + LAa I as + LAb I bs + LAc I cs ] (5.10)(5.11) which are substituted in Equation (5. 2 Its qaxis and daxis form equations from Equation (5.29) Vdp = R p I dp + ω p ( L p I qp + 3 3 Lm12 I qs ) + p( L p I qd + Lm12 I ds ) . (5.27) and using harmonic balance principle yield the dq complexform equation of the power windings given as Vqdp = R p I qdp + jω p [ Lp I qdp + 3 j (θ −θ ) Lm12 I qds e r p ] 2 (5.26) into Equation (5.28) + p[ L p I qdp + 3 j (θ −θ ) Lm12 I qds e r p ] . 2 2 (5.7) to Equation (5.28) are obtained (in the synchronous reference frame ωp) Vqp = R p I qp − ω p ( L p I dp + 3 3 Lm12 I ds ) + p( L p I qp + Lm12 I qs ) 2 2 (5. jβ Replacing the VAp and IAp with the expression (5.26): (5.
38) .34) are obtained as ' ' Vqp = R p I qp − ω p ( L p I dp + Lm I ds ) + p ( Lp I qp + Lm I qs ) ' ' Vdp = R p I dp + ω p ( L p I qp + Lm I qs ) + p( L p I d + Lm I ds ) (5. and Te is the electromagnetic torque.31) (5.36) ' ' Vqdp = R p I qdp + jω p [ L p I qdp + Lm I qds ] + p[ L p I qdp + Lm I qds ] (5. The electromagnetic torque is calculated from the magnetic coenergy.32) (5.33) (5.34) 2 ' ' Vdc = Rs' I qs + p( L's I qs + Lm I qp ) 3 ' ' 0 = R s' I ds + p ( L's I ds + Lm I dp ) . we chose the power windings as reference windings. It can be shown that the electromagnetic torque is given as (see Appendix 5C) Te = 3 j (θ −θ ) * ( p1 + q ) Lm12 Re[ jI qdp I qds e r p ] .To combine the mathematical models of control windings and power windings in the same system. The qd complexform equations of power windings and control windings in the synchronous reference frame are given as Equations (5.35) and (5.35) ' ' ' Vqds = Rs' I qds + p[ L's I qds + Lm I qdp ] . the moment of inertia of the rotor and connected load is J.36) Torque Equation The mechanical equation of motion is expressed as Jp 2θ rm = Te − TL (5. 2 (5.31)(5. (5. the system Equations (5. After control windings are referred to the power windings using the winding turnsratio.37) where the load torque is TL.
Np * ' r =+ ( Ls +N ( ( rs .22) and (5.Considering the turns ratio between the power windings and control windings and since θr = θp. Ls = + +N ( . s) .Np * . 2 Ns . we use magnetizing flux ' dependent series or shunt coreloss resistances: Rmp and Rms to represents these losses. Considering these factors and using the flux linkage variables. s) ' s 2 Lm = 3 Np Lm12 . the timeinvariant electromagnetic torque equation is Te = 3 * ' ( p1 + q) Lm Re[ jI qdp I qds ] 2 (5. s) 2 . 2 Since the core losses in the machine are significant.Np * ' Rms =+ ( Rms +N ( . Ts = ' Rms ' r + Rms ' s .39) 3 ' ' = ( p1 + q) Lm ( I dp I qs − I qp I ds ) .28) of the power windings and control windings are expressed as With shunt coreloss resistances VqdpTp = rpTp I qdp + jω p λqdp + pλqdp ' ' Vqds = rs'Ts I qds + pλ'qds ' λqdp = L p I qdp + Lm I qds ' λ'qds = L's I qds + Lm I qdp where Tp = Rmp rp + Rmp 2 . the complexform Equations (5.
s) 3 Np Lm12 . rp Vqdp Rmp Lp.Lm jωpλqdp Ls. Ls = + +N ( . . + Iqdp Lm . s) . s) 2 .2 and Figure 5. Iqds . Np ' Vqds = 2 Np Vdc . rs .' I qds = Ns I qds .Np * .2. 3 Ns With series coreloss resistances Vqdp = ( rp + Rmp ) I qdp + jω p λqdp + pλqdp ' ' ' Vqds = (rs' + Rms ) I qds + pλ'qds ' λqdp = L p I qdp + Lm I qds ' λ'qds = L's I qds + Lm I qdp where .Np * =+ +N ( ( Rms .3.Lm . Rms Figure 5.Np * ' rs' = + ( Ls +N ( ( rs . Complexform equivalent circuit of doublyfed synchronous reluctance machine with shunt coreloss resistances. Vqds . Lm = 2 Ns ' ms 2 2 R . The corresponding complexform equivalent circuits with shunt or series loss resistors are shown in Figure 5.
.3 Complexform equivalent circuit of doublyfed synchronous reluctance machine with series coreloss resistances.Lm Iqdp Vqdp jωpλqdp + Lm Ls.r p+ Rmp Lp.40) pVdp = − 1 ( I dp + I do ) − ω pVqp Cp .4 shows the schematic diagram of the machine operating as a standalone generator feeding a threephase impedance RL load. . A threephase capacitor bank is connected in parallel to the power windings to provide generator reactive power to sustain the load. Figure 5.Lm r s + Rms .3 Generator with Impedance Load Figure 5. Iqds Vqds . The generator system with impedance load in Figure 5.4(a) can be described by the following qd equations using the synchronous reference frame angle θp. . 5. Load Model pVqp = − 1 ( I qp + I qo ) + ω pVdp Cp (5.
' Rs' = rs' + Rcs .42) Vdp = R p I dp + ω p ( L p I qp + L I ) + p( L p I d + L I ) ' m qs ' m ds 2 ' ' ' Vdc = Rs' I qs + p ( L's I qs + Lm I qp ) 3 (5.43) 0 = R I + p( L I + Lm I dp ) ' ' ds ds ' ' s ds where R p = rp + Rcp .3) ' ' ) + p( L p I qp + Lm I qs ) Vqp = R p I qp − ω p ( L p I dp + Lm I ds (5.4. (a) With impedance. (b) with a loaded rectifier.41) Lo dI do + Ro I do − ω p Lo I qo = Vdp dt Machine Model with series coreloss resistances (Figure 5. Lo dI qo dt + Ro I qo + ω p Lo I do = Vqp (5.Figure 5. Doublyfed synchronous generator systems.
' ds ' ' s ds Since the parameters of the machine are dependent on the airgap flux linkage.44) and (5.44) λ dp = L p I dp + L I ' m ds ' + Lm I qp λ'qs = L's I qs (5.45) λ = L I + Lm I dp . It is necessary to express Equations (5.45).' λ qp = L p I qp + Lm I qs (5.43). we obtain the following equations.40)(5.48) pλ dp = Vdp − ω p λ qp + Tss1λ dp + Bss1λ ' ds . I dp = L's λ dp − Lm λ'ds L p L's − L2 m L p λ'ds − Lm λ dp L p L's − L2 m ' I qs = . From Equations (5.47) pI do = γ LVdp + ω p I qo − γ L Ro I do pλ qp = Vqp + ω p λ dp + Tss1λ qp + Bss1λ'qs (5. we can obtain I qp = L's λ qp − Lm λ'qs L p L's − L2 m L p λ'qs − Lm λ qp L p L's − L2 m . Substituting these expressions into Equations (5.43) in terms of flux linkages.46) pVdp = γ cTs1λ dp + γ c B λ − γ c I do + ω pVqp ' s1 ds pI qo = γ LVqp − ω p I do − γ L Ro I qo (5.40)(5. pVqp = γ cTs1λ qp + γ c Bs1λ'qs − γ c I qo − ω pV pd (5. ' I ds = .
2 pλ'qs = Vdc + Tss 2 λ'qs + Bss 2 λ qp 3 (5.50) (5. Bss1 = r p Lm L L p − L2 m ' s Tss 2 = − rs1 L's .51) (5. Lo The qd complexform equations of the generator system are expressed as pVqdp = γ c Ts1λ qdp + γ c Bs1λ'qds − γ c I qdo + jω pVqdp pI qdo = γ LVqdp + jω p I qdo − γ L Ro I qdo pλ qdp = Vqdp − jω p λ qdp + Tss1λ qdp + Bss1λ'qds 2 pλ'qds = Vdc + Tss 2λ'qds + Bss 2 λqdp .49) pλ = Tss 2 λ + Bss 2 λ dp ' ds ' ds where Ts1 = − L's . 3 (5.53) Under the steadystate operation. the derivatives of the state variables are identically equal to zero leading to the following system equation: .52) (5. Bs 2 = Tss1 = − rp L's L Lp − L ' s 2 m . L's L p − L2 m − Lp L Lp − L ' s 2 m B s1 = Lm L's L p − L2 m Lm L L p − L2 m ' s Ts 2 = . L's L p − L2 m Bss 2 = rs1 Lm L's L p − L2 m γc = 1 . Cp γL = 1 .
0906λm + 0. The relationship between these parameters and airgap flux linkage magnitude are empirically determined as Lm = −0.58) .57) (5.24λ2 m − 4.05λ2 m − 1. 2 " Vdc " 3 # (5. L'ls " # Lmm = [ 1 1 1 + + ] −1 .0385 Rmp = −4.25λm + 1.56) If we use the machine model with the shunt loss resistor (see Figure 5. (5.44λ2 m + 0.47λm + 1.09λm + 0.55λ2 m + 0.54) This equation is nonlinear since the machine inductances and core loss resistances vary with the magnitude of the airgap flux linkage as shown in Figure 3.0373 L's = −0. Llp Lm L'ls λ m = λ qdm .11.085λm + 0.2).0164 L p = −0.55) where the airgap flux linkage is given as λ qdm = Lmm % ' λ qdp % Llp & + λ'qds $ ".36λ2 m + 0.58 (5.−γc γ c Bs1 ' γ cTs1 % 0 0 ( jω p − γ L Ro ) % %(Ts1 − jω p ) Bss1 0 % B T 0 ss 2 ss 2 & jω p $ 'λqdp $ ' % λ' " % γL " " % qds " = % 1 " % I qdo " % " %− "% 0 #& %Vqdp # " % & 0 $ 0 " " 0 " .57) (5.217 ' Rms = 7. similar derivation is done and the complexform equations of generator system in terms of flux linkage are given as pVqdp = γ cTo λqdp + γ c Bo λ'qds − γ c I qdo + ( jω p − Tc )Vqdp pI qdo = γ LVqdp + jω p I qdo − γ L Ro I qdo pλqdp = TpVqdp − jω p λqdp + Tpp λqdp + B pp λ'qds (5.
35λ2 m + 0.60) Tpp = − rp L'sTp ∆ rs' L pTs ∆ L'sTp ∆ . (5. Bo = LmTp ∆ .145λ m + 0. To investigate the steadystate performance characteristics of the machine. the machine’s parameters are determined as follows: L m = −0.552λ2 m + 0.0327 L's = −0.44λ2 m + 0. various . B pp = rp LmTp ∆ rs' LmTs ∆ Tpp = − .2 . the following matrix equation is obtained γ c Bo ' γ cTo % 0 0 % %(Tpp − jω p ) B pp % Bss Tss & where −γc ( jω p − Tc )$ 'λqdp $ ' " % λ' " % γL ( jω p − γ L Ro ) " % qds " = % " % I qdo " % Tp 0 " %− "% 0 0 #% &Vqdp " # % & 0 $ 0 " " 0 " 2 " Vdc " 3 # (5.0385 rmp = −5022λ2 m + 1865λ m + 132 ' rms = −8218λ2 m + 2014.091λ m + 0.2λ m + 88.56) are numerically solved for fixed values of Cp = 45µF. Equations (5.59) Under the steadystate operation. Bss = To = − .085λ m + 0. For the machine model with the shunt loss resistor is used.' pλ'qds = TsVqds + Tss λ'qds + Bss λqdp .0164 L p = −1.54)(5.
5. and load impedance.values of rotor speed.5(a).5 A ) 250 ( Idc = 4.5 gives the generator characteristics when fixed Idc = 4. load voltage and load current. Calculated and measured characteristics of synchronous generator system with impedance load for different rotor speeds. It is observed from these figures that the higher the rotor speeds. load current.5A and choosing three different rotor speeds.5 A ) Vo(V) 150 100 50 0 0 Vo(V) + : 900 RPM * : 1350 RPM o : 1800 RPM 200 400 600 800 1000 + : 900 RPM * : 1350 RPM 200 o : 1800 RPM 150 100 50 0 0 0. the higher the output power. In Figure 5.5(b). . the curves show the steadystate relationship between the output power and load voltage while the steadystate relationship between the load voltage and load current is displayed in Figure 5. (b) load voltage vs. Figure 5.5 3 Po(W) Io(A) (a) (b) Figure 5.5 2 2. (a) Load voltage vs. output power. control winding dc current. 300 250 200 300 ( Idc = 4.5 1 1.
7(a) show the steadystate relationship between the load voltage and output power while the steadystate relationship between the load voltage and load current is displayed in Figure 5. the waveforms of voltage and current are not pure sinusoidal waveforms because of existing harmonic components.5A. In either case: a low resistive load or a high resistive load. These figures show that the measured and calculated generator characteristics correlate very well.5A/div). Top: generator current (0. and 6A. Three different dc current values. The discrepancies between calculations and experimental results are due to the high sensitivity of the . (a) High output power. It is observed in Figure 5. (b) Low output power.7(b). are chosen to do the experimental measurement and steadystate calculation.7 that increasing the excitation current increases the maximum attainable power and also the noload generator voltage.6. Measured waveforms of synchronous generator system with impedance load (5msec/div). Figure 5.(a) RL = 21Ω (b) RL = 400Ω Figure 5. The influence of the dc excitation current on the voltage regulation of the generator is shown in Figure 5.Top:generator current (1A/div). The curves in Figure 5. 3A. Bottom: generator voltage (50V/div). 4. Bottom:generator voltage (50V/div).6 shows the waveforms of the generator when the rotor speed is 900rpm and feeding a low resistive load (RL = 20ohms) or a high resistive load (RL = 400 ohms).7.
350 300 250
350
( Wrm = 1800 RPM )
( Wrm = 1800 RPM )
300 250
Vo(V)
200
Vo(V)
400 600 800 1000 1200
*:3A 150 + : 4.5 A o:6A
100 50 0 0 200
200
*:3A
150 + : 4.5 A 100 50 0 0 1 2 3 4
o:6A
Po(W)
Io(A)
(a)
1 0.8 P ow er Factor 0.6
(b)
[Wrm = 1800 PRM] ( Idc = 3 A, 4.5 A, 6 A)
0.4 0.2 0 0
50
100 Ro
150
200
250
(c)
Figure 5.7. Calculated and measured characteristics of synchronous generator system with impedance load
for a constant rotor speed and different control winding excitation currents. (a) Load voltage vs. output power, (b) phase load voltage vs. load current, (c) power factor vs. load resistance.
operating points to the airgap flux linkage and nonnegligible generator harmonic currents seen in Figure 5.6. In Figure 5.7(c), we can see the power factor will not be affected by varying the load resistance or control winding current. Figure 5.8 demonstrates calculated synchronous generator characteristics of generator system feeding an impedance load with three inductance volue. Increasing the load
8 00
18 0 16 0
6 00 P o (W) Vo ( V)
14 0 12 0 10 0 80 60 3 00 40 0
4 00
2 00
Lo = 0.0796H Lo = 0.0531H Lo = 0.0265H
Lo = 0.0796H Lo = 0.0531H Lo = 0.0265H 2 Io (A)
0 0
1 00 Z
2 00
1
3
(a)
0 .8
1 80
(b)
1 60 1 40 Vo (V) 1 20 1 00 80
Lo = 0.0796H Lo = 0.0531H
Lo = 0.0265H
P o w e r Fa c to r 0 .6
0 .4
Lo = 0.0796H
0 .2
Lo = 0.0531H
0 0 1 00 Z 2 00 3 00
60
Lo = 0.0265H
40
0
1 00 Z
2 00
3 00
(c) (d) Figure 5.8. Calculated synchronous generator characteristics of generator system feeding impedance load. (Wrm = 1800RPM, Idc = 4.5A). (a) Load power vs. load impedance, (b) load voltage vs. load current, (c) generator power factor vs. load impedance, (d) load voltage vs. impedance.
inductance value decreases the maximun output power, generator power factor, output current, and output voltage as shown in Figures 5.8(a)(d). These reductions are obvious in lower impedance values, however, these reduction gradually disappear with increasing impedance load values. The starting electrical transients of the generator system feeding an impedance load are simulated using Equations (5.50)(5.53) or (5.57)(5.59). Figure 5.9 displays
the simulation results. The generator is fitted with an excitation capacitor bank with a perphase value of 45µF and is driven at a constant speed of 1800rpm. Saturation effects are accounted for by updating the inductances and core loss resistances based on the instantaneous airgap flux linkage magnitude. There is a gradual growth of the load voltage and current magnitudes, which are respectively illustrated in Figures 5.9(a) and 5.9(b), as the airgap flux linkage is built up until they all reach a steadystate operating condition.
V pa(V)
T(sec.)
(a)
Ipa (A)
T(sec.)
(b)
Figure 5.9. Simulated starting transient of the synchronous doublyfed reluctance generator system feeding
an impedance load. (a) Phase ‘A’ terminal voltage, (b) phase ‘A’ generator current.
5.4 Generator with Rectifier Load
The model of the generatorrectifier system is shown in Figure 5.4(b). The model of the rectifier load connected to power winding of the generator has been derived in Section 4.2. This section also shows that a threephase rectifier with an impedance load in steadystate behaves like a wyeconnected threephase resistive load (RL) connected across the excitation capacitor. The relationship between RL and Ro is expressed as
RL =
π2
12
Ro .
Consequently, the analytical approach used for the generator feeding impedance can be used for the generator with rectifier load. The rectifier voltage Vd and current Id can be found using the following equations
pVqdp = 1 [ I qdp − S qdI I d ] + jω pVqdp Co
Vd = Re al[Vqdp S qdv ] . Figure 5.11 gives the measured and calculated performance characteristics of the generator feeding a rectifier that has a high impedance load (RL=400 ohms) or a low impedance load (RL = 5 ohms). Three different dc current values, 3A, 4.5A, and 6A, are used in the steadystate calculation and the experimental measurement. The curves of load voltage vs. output power is shown in Figure 5.11(a), where we can see that increasing control winding dc current increases the maximum attainable power. The load voltage and load current also increase with increasing the control winding dc current, which is observed from the curves of load voltage vs. load current shown in Figure
5.11(b). Figure 5.11(c) shows us that the power factor will not be affected by varying the control winding current or rectifier resistance load. The difference between calculation and measured results are due to these factors: profound commutation overlap, waveform distortions and high value of stator leakage inductance. Figure 5.10 shows the measured voltage and current waveforms of the generator feeding a threephase rectifier feeding a high impedance load or low impedance. In either case, waveform distortions are clearly seen in Figures 5.10(a) and 5.10(b). Profound commutation overlap is found in the waveform of the generator voltage as shown in Figure 5.10(b). Like the generator feeding an impedance load, there is a maximum output power corresponding to each rotor speed and control excitation current.
(a) RL = 400Ω
(b) RL = 5Ω
Figure 5.10. Measured waveforms of the synchronous generator system feeding a three
phase rectifier load (5msec/div). (a) High impedance load, Top: generator current (0.2A/div); Bottom: generator voltage (50V/div), (b) low impedance load, Top: generator current (1A/div); Bottom: generator voltage (20V/div).
11. (b) Load voltage vs.6 (b) [Wrm = 1800 PRM] ( I dc = 3 A. output power.5 A.4 0. . load current.8 P ow e r Fa cto r 0.5 A o:6A 300 200 100 100 +:3A * : 4.2 0 0 1 00 2 00 Ro (re c tifie r l o a d ) 3 00 (c) Figure 5. 6 A) 0. 4.5 A o:6A 0 0 200 0 0 1 2 3 4 5 6 Po(W) Io(A) (a) 1 0.400 400 ( Wrm = 1800 RPM ) 300 Vo(V) 200 Vo(V) ( Wrm = 1800 RPM ) 400 600 800 1000 1200 +:3A * : 4. Measured and calculated steadystate performance curves of synchronous generator system with a threephase rectifier load. (a) Load voltage vs. (c) power factor vs. rectifier load resistance.
5 Conclusion In this chapter. The discrepancies between calculations and experimental results are due to the airgap flux linkage and the nonnegligible generator harmonic currents. The generator operates essentially like a cylindrical rotor synchronous generator in which the frequency of the generated voltage is directly related to the rotor speed and the pole numbers of the windings. Hence this machine is suitable as a medium to high frequency generator. we have presented the modeling and analysis of the doublyfed synchronous reluctance generator with a dc control winding excitation. The model equations derived in this chapter can be used with profit to calculate the transient and steadystate performance of the doublyfed synchronous reluctance motor fed with either a variable or constant frequency supply.5. The severe commutation overlaps in the operation of the rectifier also contribute to the differences of calculation and experimental results. . Steadystate calculation results compare fairly well with experimental results.
The output of that is further processed by either a DCDC buck or boost converter. Section 6. This kind of generator systems can be realized by connecting the generated DC voltage from the power windings to machine control windings through a DCDC converter such as buck or boost converter.1.2 Description and Modeling of Generator Systems Figures 6. 6.2 gives a description of the generator systems investigated. In this chapter. Conclusions are contained in Section 6.2 show the schematic diagram of the synchronous reluctance generator systems considered in this chapter. in addition to the derivation of the models of the machine.4.3.1 Introduction The attention in this chapter is focused on the operation of the doublyfed synchronous reluctance generator as a controllable power source for DC loads and for use in battery charging. In Section 6. shunt capacitors.CHAPTER 6 THE PERFORMANCE OF A DOUBLYFED SYNCHRONOUS RELUCTANCE GENERATOR WITH CONTROLLED DC OUTPUT VOLTAGE 6. In Figure 6. The simulation of the 129 excitation process for the generator system with exciting source from the power windings is discussed in Section 6. Using the generated DC voltage at the power winding to excite the control winding circuit brings these advantages of eliminating the need for an independent excitation arrangement and the absence of sliprings and brushes. the generator is connected to a threephase diode rectifier. The simulation of the starting transients and steadystate waveforms of using this generator system for battery charging is also included in this section. the AC output power from the power windings of the doublyfed synchronous reluctance generator is rectified with a threephase rectifier and is further processed by either a DCDC buck or boost converter for output power or load voltage regulation. the steadystate calculation and experimental results are compared and discussed.1 and 6.3. The three . threephase diode rectifier and buck and boost DCDC converters.
Figure 6. (a)With buck DCDC converter.1. (b) with DCDC boost converter. Deltaconnected capacitors are connected across the power winding terminals to provide reactive power and to enhance the generator real power output capability. Schematic diagram of the doublyfed synchronous generator systems. Controllable DC voltage or power is obtained by connecting the filtered output of the . phase control windings are connected to a controllable source of DC current or voltage source while the threephase power windings are connected to a threephase diode rectifier.
(b) feeding a battery. which also acts as the source to the DCDC converter. Schematic diagram of the doublyfed synchronous generator system with dc dc buck converter.2. excites the generator through the . then the generator generates output voltage whose filtered rectified dc voltage feeds the impedance load and at the same time activates the dcdc buck converter. In Figure 6. At last.2(b) in which the filtered dc voltage is directly connected to the battery. The generator scheme for battery charging is shown in Figure 6. For this application.2(a). (a) Feeding an impedance load.(a) (b) Figure 6. the generator is excited by feeding the control winding from a battery at first. the battery source is disconnected when a steadystate operating condition is obtained.1(a). The DCDC converter controls the output power or voltage using a constant frequency. pulsewide modulation control (PWM) scheme that varies the turnon time (the duty ratio) of the transistor. the battery.2 are further application of the generator system with DCDC buck converter in Figure 6. rectifier to either a DCDC buck or boost converter. The generator systems shown in Figure 6.
Np * ' rs' = + +N ( ( rs . Ls = + +N ( ( Ls .2.Np * . Ts = ' Rms ' r + Rms ' s . 2 Ns Ns I qds . the complexform qd equations of the generator shunt coreloss resistances Rmp and Rms in the synchronous reference frame rotating with angular speed ωp (ωp is the angular speed of the generated voltage) are given as VqdpTp = rpTp I qdp + jω p λqdp + pλqdp ' ' Vqds = rs'Ts I qds + pλ'qds ' λqdp = L p I qdp + Lm I qds ' λ'qds = L's I qds + Lm I qdp (6. s) ' Vqds = 2 ' I qds = 2 Np Vdc 3 Ns . The model equations of the generator schemes shown in Figures 6.1 and 6. Np . which has the ' .1) (6.4) where Tp = Rmp rp + Rmp 2 . s) 2 Lm = 3 Np Lm12 .control windings. the generator sends charging current to the battery whose value is determined by the converter dutyratio and battery open circuit voltage.3) (6. Synchronous Reluctance Generator If we use the complexform equivalent circuit shown in Figure 5.Np * ' Rms =+ +N ( ( Rms . When DC output voltage is sufficiently built up.2 are presented in the following subsections.2) (6. s) .
s) .6) (6.5) (6. λqdp is the complex qd flux linkage for the power windings. resisters Rmp and Rms Vqdp = ( rp + Rmp ) I qdp + jω p λqdp + pλqdp ' ' ' Vqds = (rs' + Rms ) I qds + pλ'qds ' λqdp = L p I qdp + Lm I qds ' λ'qds = L's I qds + Lm I qdp (6. the ' and λ'qds .Np * =+ +N ( ( Rms . per pole turn numbers of the power and control windings. The quantity Vs is the input voltage to the control winding while Np and Ns are the effective perphase. Np ' Vqds = In these equations. Ls = + +N ( ( Ls .Np * . respectively.3.or the model equations of the generator can be described by the following equations if we use the complexform equivalent circuit shown in Figure 5. s) 2 Np Vdc . 3 Ns 2 Ns I qds . respectively. The measured self and magnetizing inductances of the windings and coreloss resistances of the machine used for this work are shown in Figure 3.11 (series loss resistors) or Figure 3.7) (6. 2 Ns ' I qds = R ' ms . which has the series loss ' . referred control winding complexform current and flux linkage are I qds respectively.Np * ' rs' = + +N ( ( rs . s) 2 2 3 Np Lm = Lm12 .17 . Vqdp. and Iqdp are the complexform generator terminal voltage produced by the power windings and current flowing through them.8) where . p=d/dt.
The relationship between these parameters and airgap flux linkage magnitude are empirically determined in equation (3.(shunt loss resistors). µ is the commutation angle. Shunt Capacitor. Rectifier and Load The complexform qd equations of the shunt capacitors (Co) connected across the power windings is given as pVqdp = 1 (− I qdp − S qdI I d ) + jω pVqdp . 2 π µ Sqdv and SqdI are. Io is the load current with the filter inductor current and filter capacitor voltage represented . Co (6.11) pI d = pVco = 1 (Vd − Vco ) Ld 1 (Id − Io − Ic ) Cd (6. cos( ) . respectively.10) (6. the equations describing the inputoutput voltages and currents of the rectifier. filter elements and the load are defined as * Vd = Re al (Vqdp S qdv ) (6.87) (shunt loss resistors).13) where S qdI = AI e j (σ +θ ) . the complexform qd voltage and current switching functions of the rectifier. Vd is the rectifier output voltage. AI = 2 3 S qdv = Av e jσ AV = 2 3 π sin( µ / 2) /( µ / 2).86) (series loss resistors) or equation (3.12) pI o = 1 (Vco − Ro I o ) Lo (6.9) Also.
Ld is the filter inductor with the impedance load represented as Ro and Lo. The initial angle between the power and control winding axes is σ and the power factor angle of the generator at the power winding terminals is θ. The switching function when the transistor and diode are not conducting at the same time is S3. Vco is the output filter capacitor voltage. DCDC Buck Converter The model of the DCDC buck converter feeding a resistive load or battery shown in Figure 6.15) (6. The quantity Cd is the filter capacitor.2 are derived in this subsection [8182].as Id and Vco. The output voltage and power are controlled by varying the turnon time of the transistor d1T that is regulated by changing the magnitude of the reference voltage compared to a constant frequency sawtooth waveform.3 (a) gives a typical converter inductor current and converter switching functions corresponding to the three converter operational modes given in Figures 6. The switching functions of transistor and diode are S1 and S2. which take values of unity when the devices are turned on and zero when they are turned off. Figure 6.1(a) and Figure 6.16) where IL is the inductor current. The conjugate is presented by *. .14) (6. T is transistor’s switching period. respectively. Mode I: Transistor is on: 0 # t # d1T The voltage equations are given as Lo pI L = Vcd − Vco Ce pVco = I e Vco − Ro I L + Ro I e = 0 (6. the current through the output capacitor is Ie and the load resistance is Ro.3 (bd). respectively.
4(a) are Lo pI L = Vcd Ce pVco = I e − Vco − Ro I e = 0 . The voltage equations are derived below (6.23) (6.4 (bc).17) (6.25) (6.21) The equations of the three modes of operation are averaged using the switching functions and are given as Lo pI L ( S1 + S 2 ) = Vcd S1 − Vco ( S1 + S 2 ) Ce pVco = I e Vco − Ro I L ( S1 + S 2 ) − Ro I e = 0 .Mode II: Diode is on: d1T # t # (d1+ d2)T The voltage equations from figure 6.24) Mode I: Transistor is on: 0 # t # d1T The voltage equations from Figure 6.22) (6. Figure 6.27) . Mode II: Diode is on: d1T # t # (d1+ d2)T (6.3(d) are given as (6.18) (6. Boost DCDC Converter The model of the DCDC boost converter feeding a resistive load shown in Figure 6.1(b) is set forth in this subsection. (6.3 (c) are expressed as Lo pI L = −Vco Ce pVco = I e Vco − Ro I L − Ro I e = 0 .26) (6.19) Ce pVco = I e Vco − Ro I e = 0 .4(a) shows the inductor current of the boost DCDC converter with the switching functions and the three modes of operation of this converter are shown in Figure 6.20) (6. Mode III: Transistor and diode are off: (d1+ d2)T# t # (d1+ d2+d3)T The voltage equations from Figure 6.
4(c) are expressed as Lo pI L = Vcd − Vco Ce pVco = I e Vco − Ro I L + Ro I e = 0 .The voltage equations of this mode from Figure 6.28) (6. (a) Converter inductor current and switching functions. (6.3. Buck DCDC converter. (b) circuit when transistor is on.29) (6. . (d) circuit when both transistor and diode are off. (c) circuit when the diode is on.30) Figure 6.
(b) circuit when diode is on. (b) circuit when transistor is on. Boost DCDC converter. (a) Converter inductor current and switching functions. .Figure 6. (c) circuit when both diode and transistor are off.4.
(6.5.35) Cb[ pVbp = I b − Cb1 pVb1 = I b − Vbp Rbp Vb1 Rb1 .39) Ib Rb1 Rbs Cb1 Cbp Rbp Rbt + Vb Figure 6.4(d) are expressed as Ce pVco = I e Vco − Ro I e = 0 .31) (6.32) Lo pI L ( S1 + S 2 ) = −Vco S 2 + Vcd ( S1 + S 2 ) Ce pVco = I e Vco − Ro I e ( S1 + S 2 + S3 ) − Ro I L S 2 = 0 .36) (6. the dynamic equations describing the leadacid battery are given as (6. Finally. Leadacid Battery From [61] and Figure 6. 6.Mode III: Transistor and diode are off: (d1+ d2)T# t # (d1+ d2+d3)T The model equations from Figure 6.33) (6.34) (6.18)(6. the equations for three modes are combined and are given as (6.19).3 Steadystate Characteristics of the Generator Systems .5.37) Vc = Vbp + Vb1 + I b ( Rbs + Rbt ) (6. Equivalent circuit representation of a leadacid battery Appendix 6A contains the definitions of the different resistors and capacitors in Equations (6.
24) and (6.2(a) VqdpTp = rpTp I qdp + jω p λqdp 2 ˆ Np ' VsTs = rs'Ts I qds Ns 3 S qdI I d = I qdp Vd = Re al[Vqdp S qdv ] = Vcd ˆ Id = Io + I c Vco = Ro I o (6. Also.During steadystate operation.40) ˆ =I I L d ˆ (d + d ) = 0 Vcd d1 − V co 1 2 ˆ =R I ˆ V co o L ( d1 + d 2 ) VqdpTp = rpTp I qdp + jω p λqdp .22)(6. The Model of the generator system with boost converter in Figure 6. shunt capacitor and rectifier equations are constant.35) and setting the derivatives of the generator. The Model of the generator system in Figure 6. the models of the generator systems are obtained as following The Model of the generator system with buck converter in Figure 6. the average inductor voltages and capacitor currents in the rectifier output filter and the DCDC converters are zero. the qd state variables of the generator.1(b) Np 2 ' VdcTs = rs'Ts I qds Ns 3 S qdI I d = I qdp Vd = Re al[Vqdp S qdv ] = Vcd (6. Hence averaging the converter equations (6. rectifier variables and shunt capacitors to zero.1(a) VqdpTp = rpTp I qdp + jω p λqdp Np 2 ' VdcTs = rs'Ts I qds Ns 3 S qdI I d = I qdp Vd = Re al[Vqdp S qdv ] = Vcd (6.42) .33)(6.41) ˆ d =0 Vcd (d1 + d 2 ) − V co 2 ˆ −R I ˆ V co o Ld2 = 0 .
.44) ˆ are.ˆ =d I ˆ I c 1 L ˆ d1Vco = V s . it can be easily derived that the effective resistance seen at the input of the loaded DCDC buck converter and the phase resistance presented by the load at the output of the machine terminals are respectively given as [62.44) can be numerically solved given the duty ratio d1 and load to determine the characteristics of the generator system. The Model of the generator system in Figure 6. (6. Equations (6.2(b) VqdpTp = rpTp I qdp + jω p λqdp 2 ˆ Np ' VsTs = rs'Ts I qds Ns 3 S qdI I d = I qdp Vd = Re al[Vqdp S qdv ] = Vcd ˆ Id = Io + I c Vco = Vbp + Vb1 − I o ( Rbs + Rbt ) Io = − Io = − (6. ˆ V s is the averaged converter output voltage ˆ while V co is the capacitor voltage. S2.406. ˆ I L and (6. and d3. rectifier.43) Vbp Rbp Vb1 Rb1 ˆ =d I ˆ I c 1 L ˆ d1Vco = V s where the average of the switching functions S1. respectively and are related by d1 + d 2 + d 3 = 1 . Under steadystate operation. d3 is equal to zero.46) Hence.45) Rin = Ro (1 − d1 ) π2 12 . filter and the average of the states of the DCDC converter are constant and their time derivatives become zero. With these constraints enforced on Equations (6.43). When the converter operates in (CCM) mode. and S3 are d1.40)(6. respectively.63] Buck DCDC Converter Rin = Boost DCDC Converter Ro π 2 d12 12 (6. I c the averaged inductor current and input converter current. d2. the state variables of the generator. When the DCDC converters operate in the continuouscurrent conduction mode (CCM).
6(b) and 6. 2A and 4A.48 or 6. Then with known d1.51) These converter input resistances are further referred to the input side of the threephase diode rectifier using equation (6. it is determined whether the converter operates in the CCM or DCM mode by using Equation 6.42.43.However. duty ratio. d3 needs to be determined from Figure 6. Figures 6. The average inductor current expressed in terms of the load.48) d 3 = 1 − d1 − 2 Lo TRo d1 (6.43) and are expressed as Buck DCDC Converter Rin = Boost DCDC Converter Ro (1 − d 3 ) 2 π 2 d12 12 Ro (1 − d1 − d 3 ) 2 π 2 (1 − d 3 ) 2 12 (6. duty ratio.8(a) show the curves of load voltage vs. . when the operation is in the discontinuous condition mode (DCM).456. rotor speed.6 and 6. 6.7 show both steadystate calculated and measured characteristics of the generator system with a DCDC buck converter while corresponding results are displayed in Figures 6.43) are given as Buck DCDC Converter Rin = Boost DCDC Converter Ro (1 − d 3 ) 2 d12 Ro (1 − d1 − d 3 ) 2 (1 − d 3 ) 2 (6. Similar analysis for the boost converter using Figures 6. the equation below from which d3 can be determined is obtained: 2 Lo * 2 d 32 + d 3 (d1 − 2) + + +1 − d1 − 2d1 − TR ( (=0 o ) .53) It is observed that the converter load appears as a dutyratio dependent resistor.47) Using Equation 6.49) Using the same method of deriving Equations (6.41 or 6.46. Figures 6.406.5.45) and (6. Their effective input resistances are inverseproportional to the duty ratio. and decrease when the duty ratio deviates that value. Two control winding currents. the steadystate equations from any generator system model expressed by Equations (6. are numerically solved with the empirical equations of the generator parameters given the control winding excitation voltage. Load voltages of both have maximum values when the duty ratio is close to 0.40 or 6.50) Rin = .3(bd) for the buck DCDC converter. The rotor speed is set at 1200 rev/min.8(b) show the curves of effective input resistance of converter vs.49.52) Rin = .9 for the system with DCDC boost converter.6(a) and 6. the higher the duty ratio. d2 and d3.46). (6. (6. and the load resistances of the DCDC converter. (6. These two converters operate in the continuousconduction current mode of operation. the equation for d3 is given as (6.526. For a given load. are chosen in the experimental measurement and the calculation.8 and 6. the output load resistances referred to the input of the DCDC converter using Equations (6. the lower the effective input resistances. Figures 6.53).4(bd) and Equation 6.40)(6. input and output voltages is given as [64] ˆ = (V − V ˆ )d T ( d1 + d 2 ) I 1 L cd co 2 Lo . The load voltage of boost converter is much higher than that of the buck converter. converter switching frequency and other converter parameters.
which are displayed in Figure 6.12 that it is necessary to keep the duty ratio not less some values near 0. There is good agreement between measured and calculated steadystate performance curves.2 so that the generator system (shown in Figure 6.9(c) show the curves of perphase generator voltage vs.11. Since the rectifier output voltage also depends on the load impedance that in turn is influenced by the dutyratio and the generator load performance.7(b) and 6.7(c) and 6.6(c) and 6.12(c) shows the curves of perphase terminal voltage vs. generator perphase terminal voltage or DC load voltage. 900rev/min and 1350rev/min.12.12(d). (d) converter input voltage vs. Overall. The higher the rotor speed. The load powers of both have the maximum values when the duty ratio is close to 0. The steadystate performance of the generator system (shown in Figure 6. are used in the experimental measurement and the calculation.5. duty ratio. for the buck and boost converters give measured waveforms of the converters operating in the continuouscurrent mode. It is observed from the Figure 6. . Two rotor speeds.7 and Figure 6. (Rotor speed = 1200 rpm).2 can regulate these voltages.9. are similar each other. (c) load power vs.8(c) show the curves of load power vs.10 and 6. respectively.7(a) and 6.6. They both have the same maximum output powers.Figures 6.2(a)).14. Boost converter input voltage is higher than one of buck input voltage. duty ratio. duty ratio are shown in Figure 6. The converter input voltages of both are inverseproportional to the dutyratio. duty ratio. The curves of generator output power vs. generator output power while the curves of dutyratio vs. dutyratio.6(d) and 6. that is special because of its excitation depending on the generated output DC voltage from the power windings.12(a) and (b) show the curves of dutyratio vs.2(a)) avoids the collapse of the generator terminal voltage.9(b). duty ratio are shown in Figures 6. and then they decrease when the duty ratio leaves that value. we can see the output voltage profiles of these two converters differ remarkable from those of the converters fed with a constant voltage source. is shown in Figure 6. control winding DC current are shown in Figure 6.8(d) show the curves of converter input voltage vs. Figure 6. (a) Load voltage vs. the higher the attainable maximum control winding DC currents and generator output powers. Figures 6. Perphase generator voltage vs. Figure 6. duty ratio. Changing the dutyratio during the range from zero to near 0. Figure 6.9(a)). Figures 6.1(a). The little differences must be due to the presence of factors not accommodated in the derived models: significant harmonic components in the generator and input rectifier currents in addition to the nonnegligible overlap commutation of the rectifier diodes which can be seen in Figure 6. The steadystate characteristics of the generator system with either DCDC buck converter or DCDC boost converter. The steadystate characteristics of the generator system with DCDC buck converter in the Figure 6. Figures 6. there are good agreements between measurement and calculation results. perphase generator output power. (b) effective input resistance of converter vs. duty ratio.
(c) perphase generator voltage vs. (a) Generator output power vs. The steadystate characteristics of the generator system with DCDC buck converter in the Figure 6.Figure 6. duty ratio. (b) perphase generator voltage vs.1(a) (Rotor speed = 1200 rpm). duty ratio. generator output power.7. .
8. The steadystate characteristics of the generator system with DCDC boost converter in the Figure 6. (a) Load voltage vs.1(b) (Rotor speed = 1200 rpm). duty ratio. (c) load power vs. dutyratio. (d) converter input voltage vs. duty ratio.Figure 6. (b) effective input resistance of converter vs. . duty ratio.
(a) Generator output power vs.1(b) (Rotor speed = 1200 rpm). Figure 6. Bottom: Generator lineline voltage. (b) perphase generator voltage vs. duty ratio.10. (c) perphase generator voltage vs.1(a).Figure 6. Measured waveform of the generator with DCDC buck converter in the Figure 6. The steadystate characteristics of the generator system with DCDC boost converter in the Figure 6.9. (Rotor speed = 1200 rpm.6) (a) Top: Generator phase current. duty ratio. generator output power. dutyratio = 0. .
6 ( + : Wrm = 1350 RPM ) ( * : Wrm = 900 RPM ) 1 0. Bottom: Converter inductor current.1(a).8 0. Measured waveform of the generator with DCDC boost converter in the Figure 6.2 0 0 D 0. Bottom: Generator lineline voltage.4 0. 1 0.2 0 0 D 50 100 150 50 100 150 200 Vg ( W ) Vco ( V ) (a) (b) .(b) Top: Input converter current. Bottom: Converter inductor current.6) (a) (b) Top: Generator phase current. Top: Input converter current.11. dutyratio = 0. (Rotor speed = 1200 rpm.4 0.8 0. Figure 6.6 (+: Wrm = 1350 RPM ) (*: Wrm = 900 RPM ) 0.
6. Three starting stages are: (1) a battery is connected to the control windings while the converter is turned off. dc load voltage (d) Duty ratio vs. control winding dc current.(Figure 6.2.4 0. the starting transient of the generator system assuming that the root speed is constant is simulated and shown in Figure 6. the converter is turned on with the battery still connected for another 0. Measured and calculated steadystate characteristics of the doublyfed synchronous reluctance generator feeding an impedance load (RL=400Ω).140 120 (+: Wrm = 1350 RPM ) (*: Wrm = 900 RPM ) 1 0. (2) after 0.4 Simulation of Generator System For a complete understanding of the generator system in Figure 6. 13. generator perphase terminal voltage (b) Perphase terminal voltage vs.8 (*: Wrm = 900 RPM ) Vg ( V ) 100 80 60 40 20 0 0.4 seconds. generator output power (c) Dutyratio vs.6 (+: Wrm = 1350 RPM ) D 0.4 seconds and (3) finally.2(a)) (a) Dutyratio vs. the battery is .12.2 0 2 200 400 600 800 4 6 8 10 Pg ( W ) is ( A ) (c) (d) Figure 6.
2 0 0 . The battery initially provides an average current flowing into the converter and. With the generator shaft running at a constant speed and the converter turned on.14.6 0 .2 0 .4 0 . It is observed that the generator easily excites and the period taken to achieve steadystate operating condition depends on the battery voltage and the converter duty ratio. The electrical excitation process of the generator charging a leadacid battery shown in Figure 6.8 1 1 .8 1 1 .2 0 .6 0. In view of the low excitation current.15 gives the simulation results of the generator system when the converter duty ratio is reduced while operating at a stable steadystate point.4 0 .2 T ( sec. an effective current flows into the battery to charge it. a minimum converter duty ratio requirement for a given rotor speed and load must be met to sustain the generator operation. Figure 6.2(b) is simulated and the simulation results are shown in Figure 6. building up the generator DC output voltage. the generator excites. ) T ( sec. ) (a) (b) .16. the generator voltage collapses indicating that for a stable operation. The simulated steadystate waveforms and corresponding experimental results are now displayed in Figure 6. after the generator voltage has sufficiently risen. Figure 10 8 6 4 2 80 70 60 V co ( V ) Is(A) 50 40 30 20 10 0 0 0 0.disconnected with only the converter supplying the excitation current to the control windings.
) (a) (b) . load resistance = 400 ohms.2 5 T ( S ec .2 2 1 . (a) Control winding current.(c) (d) Figure 6.2 4 1 .13. Starting transient of the generator system shown in Figure 6.2(a). duty ratio = 0.2 1 1 .6. Rotor speed = 1350 rpm. (b) DC load voltage.2 3 1 .2 1 . (c) linetoline generator voltage. (d) generator phase ‘a’ current. 100 50 V ab ( V ) 0 5 0 1 0 0 1 .
(b) phase ‘a’ generator current.2 3 1 .6 4 Ipa ( A ) 2 0 2 4 6 8 1 .2 5 T ( Sec. load resistance = 400 ohms. 50V/div.2 4 1 .5 3 Is ( A ) 6 4 2 0 1 1 . Steadystate waveforms of the generator system feeding impedance load.2 1 .5 2 2 .5 3 T ( sec. Rotor speed = 1350 rpm. 2A/div.14.6.2 1 1 . 10 100 8 80 Vco ( V ) 60 40 20 0 1 1. ) (a) (b) . (a) Generator linetoline voltage.5 2 2. ) (c) (d) Figure 6. ) T ( sec.2 2 1 . duty ratio of converter = 0.
Generator deexcitation due to reduced DCDC converter duty ratio. (c) generator loinetoline voltage. (b) control winding current. (d) (a) DC load voltage.(c) Figure 6. (a) (b) . (d) generator phase current.15.
16. Electrical excitation transient for the doublyfed synchronous reluctance generator charging a leadacid battery.4 0. (d) control winding current.8 10 1 3 T ( Sec. (a) Current flowing into the battery. (c) phase ‘a’ generator current.6 0.2 0.(c) (d) Figure 6. ) (a) (b) . (b) generator DC voltage. 2 1 Io ( A ) 0 1 2 0 0.
(d) variation of charging curent with converter duty ratio for 900 rpm and 1350 rpm.2 0 0 ( + : Wrm = 900 RPM ) ( o : Wrm = 1350 RPM ) 0. 6. From this graph.1 0.4 0. Figure 6. Average charging current = 0.4 0. The average charging battery current is fixed at 0.5A/div. (c) generator phasse current. 0.17. Measured steadystate waveforms of the generator charging a 24V battery.7A.8 Io ( A ) 0. 10V/div.17(d) shows the measured variation of the average charging current as a function of the duty ratio of the converter for two rotor speeds. 6.6 0.17(a)(c) display measured steadystate waveforms of the generator charging a 24V battery when the rotor speed is 1350rpm and the duty ratio of the DCDC converter is 0.25.6 0. Finally.8 1 D (c) (d) Figure 6. (b) generator lineline voltage.2 0.5 Conclusion .7A. it is concluded that there is a maximum average charging current achievable below which there are two possible duty ratios that result in the same average charging current. (a) Battery current.
The generator systems successfully use buck or boost DCDC converter connected to the power windings as the DC source regulators. which can be an outside source or a self source from DC output of power windings. . The computer simulation results show that the excitation process is fast and reliable. The generator system characteristics are predicted based on the proposed models and the results of which compare favorably with measured results. These generator systems have a potential for use in standalone applications and in electric automotive applications. The excitation process of the generator systems feeding an impedance load or charging a battery are simulated and discussed.The modeling and analysis of the doublyfed synchronous reluctance generator with a DC excitation . are presented.
74]. which is used in the design of the control scheme I. Two control schemes are described in section 7. With controlled ac source connected to the control winding.10. The drive system operates exactly like a DC machine possessing the same ease control.6 and 7. the dynamical simulation results of two control systems are given and discussed. The feasibility of the doublyfed reluctance machine with controlled ac power and control winding excitations for accepted field orientationtype performance have been demonstrated [73.3. and the frequency in the power windings.5. It also gives the steadystate operation characteristics of the systems below and above the base speed. the fieldorientation principle is introduced. while the rotor circuit is connected to either an inverter controlled ac or converter controlled dc sources. the machine operates either in the synchronous. Section 7. A novel inputoutput linearization technique and Butterworth method are set forth in Section 7. . Sections 7. subsynchronous or supersynchronous modes permitting large speed operation range.7 give the detailed description of both voltagecontrolled and currentcontrolled VSI. Two control schemes are 161 investigated and the main focus are their operation characteristics in a wide speed range including the constant torque control below the base speed and constant output power above the base speed with maximum output torque. The power windings are connected to a voltage source inverter (VSI). In many low performance drive applications. This chapter proposes a novel highperformance control of the doublyfed synchronous reluctance in which the control winding is connected to a controlled current DC source. voltages. The theoretical derivation is also included in this section.2.CHAPTER 7 FIELD ORIENTATION CONTROL OF A DOUBLYFED SYNCHRONOUS RELUCTANCE MACHINE 7. which can be a currentcontrolled VSI or voltagecontrolled VSI. Finally. to regulate the axis currents. In section 7. we draw conclusions in Section 7.9.8 and 7.1 Introduction Doublyfed reluctance machines having two stator windings (power and control windings) have received renewed attention in the last few years in adjustablespeed drives where efficiency optimization and energy conservation are desirable.4 gives the design procedures of integral plus proportional (IP) controller. which presents mathematical algorithms to obtain the parameters of the IP controller with nonovershoot performance. In sections 7. the threephase power windings are connected to the utility supply.
The field should be maintained at a certain optimal level. The Field Orientation Principle (FOP) defines conditions for decoupling the magnetic field control from the torque control. The torque developed in the motor is a result of the interaction between current in the armature winding and the magnetic field produced in the field system of the motor. (2) Optimal conditions for torque production.7. resulting in the maximum torque per unit ampere. occur in the motor both in the steadystate and in transient conditions of operation. A field–oriented doublyfed synchronous reluctance motor should emulate a separatelyexcited DC motor in two aspects: (1) Both the magnetic field and the torque developed in the motor can be controlled independently. Independent control of the field and armature currents is feasible in separatelyexcited dc motors where the current in the stator field winding determines the magnetic field of the motor.. an electric motor can be thought of as a controlled source of torque. The physical disposition of the brushes with respect to the stator field ensures optimal conditions for torque production under all conditions. Accurate control of the instantaneous torque produced by a motor is required in highperformance drive systems. With fixed field.2 Field Orientation Principle In general.g. while the current in the rotor armature winding can be used as a direct means of torque control. sufficiently high to yield a high torque per unit ampere. e. but not too high to result in excessive saturation of the magnetic circuit of the motor. those used for position control. the torque is proportional to the armature current. The power winding and control winding qd equations of the doublyfed synchronous reluctance machine in the rotor reference frame are .
Pr (7. and Pr is equivalent pole numbers of the rotor and equal to p1+q.9) The relationship of the rotor speed ωr.11) . the control winding currents Ias.6) (7.1.1) (7.2) (7. Ics.3) (7.8) Torque equation is expressed as 3* ' ' Te = + (( p1 + q) Lm ( I dp I qs − I qp I ds ).7) (7. (7. Ibs .5) (7. .2) (7.4) where ' λqp = L p I qp + Lm I qs ' λdp = L p I dp + Lm I ds ' λ'qs = L's I qs + Lm I qp ' λ'ds = L's I ds + Lm I dp .Vqp = R p I qp + pλqp + ω p λdp Vdp = R p I dp + pλdp − ω p λqp ' ' Vqs = Rs' I qs + pλ'qs ' ' Vds = Rs' I ds + pλ'ds (7.10) J is the rotor inertia. load torque TL and electrical torque Te is J pω r = Te − TL . and current source Is have the relationships given in Equation (7 . If the control winding is connected as shown in Figure 7.
Control winding connection.12) After performing abcqd transformation as well as considering the turnsratio. the torque depends on the control winding qaxis current I qs and the power winding daxis current Idp.1.14) ' Hence.12): I as = I s . and Vdc are given in Equation (7. (7.Control Windings Ias Ibs Vas Vbs Vcs Current source Is +  Vdc Ics Figure 7. while the relationships among control winding voltages Vas.13) and ' I ds =0 ' I qs = I s' . It is .11) Vbs = Vcs . Vbs. Vcs. (7. Vdc = Vas − Vbs = Vas − Vcs Vas + Vbs + Vcs = 0 . 1 I bs = I cs = − I s 2 I as + I bs + I cs = 0 (7. control winding qd axis voltages and currents become 2 ' Vqs = Vdc 3 ' Vds =0 (7.
16) (7. with the daxis power winding current Idp set to the rated power winding current. hence it is called constantpower or fieldweakening region. The operating conditions for two modes are listed as follows [7980]: Mode I: current and voltage limited region: 2 2 2 V pm = Vdp + Vqp 2 2 2 I pm = I dp + I qp (7. From zero speed up to its base speed. We drive the motor in a wide speed range. 2 (7. the power winding voltage is kept at rated value while airgap flux linkage need to be decreased to realize the constant power output.15) ' ' is used to control magnet flux similar to Lmq I qs . the power winding voltage rises up to its maximum value at base speed and then limits to that value at higher speed. while Idp In field orientation control. Above base speed. There are two operating modes above the base speed.Te = 3Pr ' Lm I dp I qs .17) . To obtain maximum torque (and hence power) at any speed above base speed within the voltage and current rating. The outputpower is constant above base speed. Maximum ' is set at the rated constant torque is achieved when the control winding dc current I qs value. At the same time the qaxis power winding current is set equal to zero. This speed range is called constant torque region. I qs is used to control the torque. maximum torque fieldweakening control strategy is used. This goal is realized by regulating the power winding qaxis current Iqp and daxis currents Idp with the control winding current ' I qs set at the rated value.
Hence.22) when the machine is running below the base speed. the base speed ωbase can be calculated using Equation (7.18) (7. where Te is electrical torque. power winding qaxis current Iqp = 0 while the daxis Idp is set to be equal to the power winding peak rate current Ipm.19) Maximize Te at each value of speed. Vpm and Ipm are the voltage and current rated peak values. Idp and Iqp are the qd components of power winding currents. which is obtained from Equation (7.(7.21) The peak value of power winding voltage Vpm is constant.20) . respectively. Mode II: Voltagelimited region: 2 2 2 V pm = Vdp + Vqp 2 2 2 I pm ≥ I dp + I qp (7. Idp and Iqp are the qd components of power winding currents. It can be obtained from Equations (7.22): .23).where Vpm and Ipm are the voltage and current rated peak values.21) and expressed as 2 2 2 V pm = Vqp + Vdp ' = (R p I qp + ω p Lp I dp ) + (R p I dp − ω p L p I qp − Lm I qs ) 2 2 (7. Under the steadystate. control winding current I’qs is equal to the rated dc current I’s (considering turns ratio). Vdp and Vqp are the qd components of power winding voltage. (7.20) (7. respectively. Vdp and Vqp are the qd components of power winding voltage. the power winding qd voltage equations are Vqp = R p I qp + ω p L p I dp ' Vdp = R p I dp − ω p Lp I qp − Lm I qs . At the same time.
24) into Equation (7.17)) in mode I.24) Tbb = 2ω p Lm R p I s' 2 Tcc = 2ω p Lm L p I s' . the following equation is obtained: 2 2 2 2 2 2 (Tcc + Tbb ) ⋅ I qp − 2TaaTcc ⋅ I qp + (Taa − Tbb I pm ) = 0 . A simplified equation of ωbase is obtained by ignoring resistance Rp (Rp=0): ω base = V pm 2 '2 L2p I pm + L2 mIs . Mode I operation By extending Equation (7. Its solution is .23) 2 '2 S a = L2p I pm + L2 mIs S b = −2 Lm R p I s' I pm 2 2 2 Sc = R p I pm − V pm .22) and using the current constraint condition (Equation (7.17). we can obtain I dp = where 2 2 2 2 2 2 '2 Taa = V pm − (Rp +ω 2 p Lm ) I pm − ω p Lm I s Tcc I qp − Taa Tbb (7.ω base = − where Sb2 − 4 S a Sc Sb + 2Sa 2Sa (7. Substituting Equation (7.
207.18)) to yield 2 aa ⋅ I dp + bb ⋅ I dp + cc = 0 where 2 2 2 aa = R p +ωp Lp .21) into the voltage constraint condition (Equation (7.24) and (7.24) and (7.25) are obtained by ignoring the resistance Rp (Rp=0): I qp = − 2 2 2 2 −ω p + I s'2 ) V pm Lm ( I pm 2 2ω p Lm L p I s' I dp = 4 2 2 '2 2 2 2 2 2 −ω p + I s'2 )] 2 4ω p Lm L p I s I pm − [V pm Lm ( I pm ' 2ω 2 p Lm L p I s . Mode II operation Substituting Equations (7.25) where Tb = −2TaaTcc 2 2 2 Tc = Taa − Tbb I pm .I qp = − T 2 − 4TaTc Tb + b 2Ta 2Ta 2 2 Ta = Tcc + Tbb (7.25) are used to obtain the power winding qaxis and daxis current Iqp and Idp when the machine is operated in the Mode I region. The simplified equations of Equations (7. Equations (7. Torque is obtained by substituting the expression of Idp into torque Equation (7.15).
if Idp and Iqp are selected using Equations (7. in mode II. the torque equation can be expressed as Te = 3Pr Lm I s' f ( I qp ) .27) Hence. The simplified equations of Equations (7. not only the maximum torque is achieved. the torque equation is differentiated with respect to Iqp and then forced to zero.26) and (7. 2 To obtain the conditions of achieving maximum output torque.27). Idp is obtained and expressed as 2 ω p Lm R p I s' + V pm L2pω 2 p + Rp 2 2 Rp +ω2 p Lp I dp = .26) to replace variable Iqp in cc.bb = −2ω 2 p Lm R p I qs 2 '2 2 2 2 2 2 ' 2 cc = ω 2 p Lm I s + ( R p + ω p L p ) I qp + 2ω p Lm L p I s I qp − V pm . (7. Idp can be expressed as a function of variable Iqp : I dp = f ( I qp ) . dI qp dI qp The solution of this equation is I qp = − ' ω2 p Lm L p I s . 2 2 Rp +ω2 p Lp (7.26) and (7. that is dTe d = ( f ( I qp )) = 0. Hence. from Equation (715).27) are obtained by ignoring the resistance Rp (Rp=0) . but also the voltage constraint condition is satisfied.26) Using equation (7. Its solution is I dp = − bb + bb 2 − 4aa ⋅ cc 2aa where Idp is dependent on variable Iqp.
26) and (7.I qp = − Lm ' Is Lp . 2 4 2 2 C c = I pm Rp − Rp V pm .15).26) and (7.23) through (7. the expression of the boundary speed point is obtained from 2 8 6 2 4 2 Ca ω12 + 2CaCbω12 + (Cb2 − Cd + 2CaCc )ω12 + ( 2CbCc − Ce2 )ω12 + Cc2 = 0 (7. Three different controlwinding currents are used to . 2 2 2 2 '2 2 2 Cb = 2 R p L p I pm − L2 m R p I s − L pV pm . calculated from Equations (7. I dp = V pm ω p Lp By substituting the expression of Idp into Equation (7.27) into the current constraint Equation (7.28). the torque is given as Te = ' 3Pr Lm I sV pm . By substituting Equations (7. The performance curves for wide speed range (below the base speed and above the base speed including Mode I and Mode II). 2 ω p Lp Boundary speed ω12 At boundary speed point ω12 between Mode I and Mode II.27). are computed using Equations (7. also satisfy the current constraint condition in Mode I. C d = 2V pm Lm R p I s' L p . 2 ' C e = 2V pm Lm R p Is. shown in Figure (7.28) where 2 C a = I pm L4p − L2 L2p I s'2 .2). The simplified equation is obtained by ignoring the resistance Rp (Rp=0): ω 12 = V pm (L I 2 p 2 pm '2 − L2 mIs ) .17). the power winding qaxis and daxis current Iqp and Idp .
and so do the maximum speeds for this mode.4 0 . Unit speed is equal to 377 rad/s.2 0 4 3 2 1 0 1 s pe e d (p e r un it) 2 3 4 .6 pf Iqs = 8 A Iqs = 2 A 0 . below the base speed. We can also see from other figures that.obtain these curves. It is observed from Figure 7.2(c) that Mode I region with constant rated current and constant rated voltage is narrow.8 0 .2 (b) and 7. however it increases when the current winding current increases.m ) 15 2 00 Iqs=16A 1 50 Vp m (V) 2 3 4 Iqs=8A 10 5 0 4 1 00 50 Iqs=2A 3 2 1 0 1 s pe e d (p e r u nit) 0 4 3 2 1 0 1 s pe e d (p e r u nit) 2 3 4 (a) (b) 12 10 8 30 00 Iqs=16A 20 00 10 00 P o (W ) Ip (A) Iqs=8A 6 4 0 10 00 Iqs=2A Iqs=8A Iqs=16A Iqs=2A 2 4 3 2 1 0 1 s pe e d (p e r u nit) 2 3 4 20 00 30 00 4 1 3 2 (c) 1 0 1 s pe e d (p e r u nit) 2 3 4 (d) Iqs = 16 A 0 . Total power winding current and torque are held constant. the daxis power winding current is set to the constant rated values and the qaxis power winding current is equal to zero. but power 25 20 To rque (N .
15
Idp
10 C urre nt (A) 5 0 5
Iqs=2A Iqs=8A Iqs=16A
Iqp
10 4
3
2
1 0 1 s pe e d (p e r u nit)
2
3
4
(e)
(f)
Figure 7.2. Performance curves for extended speed range, maximum output torque
operation. (a) Torque, (b) power winding phase voltage, (c) power winding current, (d) output power, (e) power winding q and d axis currents, (f) power factor. winding voltage and output power increase with increasing value of rotor speed. The larger are the control winding currents, the larger the output torque; Above base speed, the daxis power winding current Idp decreases almost linearly with speed for a range, after which it tends to a constant value as the speed increases. The torque decreases almost linearly with speed, giving an almost constant output power characteristic. Figure 7.2(e) shows that there is a redistribution of the power winding axes currents with the
' winding current I qs set at its rated value above the base speed. The controllers need
accurately reflect this redistributing characteristic in the command axis current values. In Figure 7.2(f), we can see that increasing rotor speed or control winding currents will increase the power factor of the motor control system.
7.3 Description of Two Control Schemes
Two control schemes, with currentcontrolled or voltagecontrolled voltage source inverter, are investigated. Figure 7.3 gives the schematic diagram of the fieldorientation control scheme I. The two inputs are the reference rotor speed and the power winding qaxis current, which is set to zero for constant
torque operation and nonzero values for constant power operation. The threephase power winding currents, measured and transformed to the rotor reference frame using the rotor angle measured with an encoder, are fed to a nonlinear controller. The nonlinear controller generates reference qd voltage components that are transformed into threephase voltage reference for the VSI.
Iqp , Iqs, Idp, ωr
Nonlinear Controller linearization Vdp Vqp T ( θr)
Power Windings
IAp I
Control Windings Is
ωr* + I* qp
 ω r
+
Kdd
Vdd
+
+
1
 I qp
Kqq
+
Bp VSI PWM ICp
DFM ωr
Vqq
+
Iqp, Idp, ωr
T(θr) Idp
θr
!
Figure 7.3. Control Scheme I: Fieldorientation control of doubly fed synchronous
reluctance machine with voltagecontrolled voltage source inverter (VSI).
Power Windings
Control Windings Is ωr
VDC
VSI PWM
IAp
DFM
ωr* +
 ω r
IP
Current Control I* dp
* I* I* Cp Ap IBp
IBp ICp
ωb
ωb
I* qp
T (θr)
1
θr
!
FieldWeakening Control
Figure 7.4. Control Scheme II: Fieldorientation control of doublyfed synchronous
reluctance machine with currentcontrolled voltage source inverter (VSI).
The output of the VSI are used to drive the doublyfed synchronous reluctance motor that has control winding connection shown in Figure 7.1. Figure 7.4 gives the schematic diagram of the fieldorientation control of doubly fed synchronous reluctance machine with currentcontrolled voltage source inverter. The threephase power winding current are measured and used as a feedback signals for the current control of the VSI. The rotor angle is measured using an encoder and rotor speed is calculated using the integer. Below the base speed, fieldweakening controller is switched off, and the reference qaxis current is set to zero. However, the reference qaxis current is nonzero values above the base speed. The fieldweakening controller is switched on at this time. The speed error inputs to the IP controller that generates the reference daxis current, its values limited in some range set by the saturation controller. The reference qd currents are transformed to the reference threephase power winding
currents that compare with the threephase feedback currents to obtain switchingtime of the VSI’s transistors. The output voltage of the VSI drives the doubly fed synchronous reluctance motor its control winding fed by a current source.
7.4 Design of the IP Controller
The Integral plus proportional (IP) controller is applied here because of its novel features [65]: (1) IP controller can eliminates the current overshoot problem and has zero steadystate error. This is very useful for the protection of the electrical devices needing frequent stopping and starting.
(2)
IP controller can also eliminate the speed and position overshoots for step changes without sacrificing the load torque response due to a load torque change. The literature [66,67] presented a design procedure of IP controller that can be used for a threeorder
induction motor servo drive system. By using this design procedure, the parameters of the IP controller can be quantitatively decided and the control system can obtain no overshoot position response. We use the similar theoretical method to derive the design procedure for a secondorder speed control system with nonovershoot speed response. For a stable singleinput singleoutput (SISO) nth order system expressed by
n bo + b1S + b2 S 2 + ... + bn −1S n −1 h =. i . 2 n ao + a1S + a2 S + ... + S S + µi i =1
H (s) =
(7.29)
For a white noise input with variance
2 σx , the variance Eo of the impulse response, or what is called the
impulse energy of the output, can be determined to be [68 ]
Eo = ..
j =1 i =1
n
n
hi h j
µi + µ j
≅.
i =1
n
hi , S + µi
(7.30)
in which the impulse energy contribution corresponding to the eigenvalue µj is defined as [68]
ej ≅ .
i =1
n
hi h j
µi + µ j
(7.31)
For a stable linear system described by Equation (7.29), the variance Eo must be finite and positive, i.e., from Equation (7.30),
0 < (e1 + e2 + ... + en ) < ∞ .
(7.32)
IP Speed Controller TL
ω r∗
Drive System ωr
+

KI S
+
∗ I qp
Kt
+ Te

Kp
1 JS + B
Figure 7.5. Block diagram of IP speed control system.
The speed control loop of the control scheme II is described by the block diagram Figure 7.5, where Kt is torque coefficient, and a and b can be expressed by the system viscous damping coefficient B and the inertia constant J. they are expressed as
Kt =
3 ( p1 + q ) Lm I qs 2
(7.33)
a=
B J
1 . J
(7.34)
b=
(7.35)
The transfer function speed control loop can be found from Figure 7.5 as
ω r ( s) ω r* ( s ) T
=
L =0
K I Ktb . s 2 + ( B + K t K p )b ⋅ s + K t K I b
(7.36)
Secondorder system in Figure 7.5 can also be expressed as
ω r ( s) ω r* ( s ) T
=
L
( s ) =0
h1 h2 + . s + µ1 s + µ 2
(7.37)
The relationship of coefficients between Equations (7.36) and (7.37) can be found as following:
h1 + h2 = 0
(7.38)
µ1µ 2 = h1µ 2 + h2 µ1 = K I K t B µ1 + µ 2 = ( B + K t K p )b .
The unitstep response of transfer function is
(7.39)
(7.40)
y (t ) =
µ1
h1
h (1 − e ) + µ (1 − e ) .
− µ 1t
2
− µ 2t
(7.41)
2
For observing the overshoot, let d(y(t))/dt = 0 and 0<µ1<µ2, then
dy (t ) = h1e − µ1t + h2e − µ 2 t = 0 . dt
(7.42)
It is obvious that the overshoot of its step response will not occur if one of the following cases is satisfied: (1) h1 > 0, h2 > 0 ; (2) h1 (3) h1
< 0, h2 < 0 ; > 0, h2 < 0, h1 > −h2 ;
(4) h1 < 0, h2 > 0, − h1 > h2 . When any one of these conditions are satisfied, Equation (7.42) is equal to zero only at t = ∞, hence overshoot will not occur. For a secondorder system, the expressions of µ1 and µ2 are obtained from Equation (7.31) as follows
e1 =
h12 hh + 1 2 2 µ1 µ1 + µ 2
(7.43)
e2 =
2 h2 hh + 1 2 . 2µ 2 µ1 + µ 2
(7.44)
By using the nonovershoot conditions in Equations (7.43) and (7.44), the conditions (14) are transferred into corresponding conditions listed as follows (a) If h1 > 0 and h2 > 0 , or h1 < 0, h2 < 0 , then e1 > 0 and e2 > 0 ; (b) If h1 > 0, h2 < 0 and h1 > − h2 , or h1 < 0, h2 > 0 and − h1 > h2 , then e1>0, e2<0 and e1+e2>0. To match the general requirements for the speed drive, the following specifications are prescribed: (i) (ii) The tracking steadystate error of speed is zero. To avoid any overshoot in the command tracking response, the energy contributions of e1 and e2 corresponding to µ1 and µ2 are set according to the condition (a) and (b). For convenience of formulation, condition (b) is chosen. Accordingly, the relationships of e1 and e2 are set to e1 = K1e2 and K1<1. (iii) The response time tp is defined as the time for the unitstep response to increase from 0 to 90% of its final value. Based on these prescribed specifications, and using the related equations, the following nonlinear equations are constructed:
f1 ( µ1 , µ 2 , h1 ) =
µ1
h1
−
µ2
h1
−1 = 0
(7.45)
f 2 ( µ1 , µ 2 , h1 ) = e1 − K1e2 =
h12 h12 (1 − K1 ) K1h12 − − =0 µ1 + µ 2 2 µ1 2µ2
(7.46)
'h h −µ t −µ t $ f3 ( µ1 , µ 2 , h1 ) = 0.9 − % 1 (1 − e 1 p ) − 1 (1 − e 2 p )" = 0 . µ µ 2 & 1 #
(7.47)
To satisfy the condition of 0<µ1<µ2, let µ2 = µ1+ c2 and modified above nonlinear equations, we can obtain
f1 ( µ1 , c, h1 ) =
µ1
h1
−
h1 −1 = 0 µ1 + c 2
(7.48)
f 2 ( µ1 , c, h1 ) =
h12 h12 (1 − K1 ) K1h12 − − =0 µ1 + c 2 2 µ1 2µ1 + 2c 2
(7.49)
'h h1 −µ t −( µ + c 2 )t p $ (1 − e 1 )" = 0 f3 ( µ1 , c, h1 ) = 0.9 − % 1 (1 − e 1 p ) − 2 µ1 + c & µ1 #
(7.50)
where c is a finite positive constant. The unknown parameter h1, µ1, and c in Equations (7.48) through (7.50) can be solved using the Matlab program. Then the parameters of the controller Kp, KI can be found from following Equations (7.51) through (7.53):
µ 2 = µ1 + c 2
KI =
(7.51)
µ1µ 2
Ktb
(7.52)
' (µ1 + µ 2 ) $ − B )" % b # . Kp = & Kt
(7.53)
The Matlab program using to obtain these parameters for nonovershoot performance is listed at Appendix 7C.
7.5 Nonlinear Controller Design
In the last few years, nonlinear control theory has been used in some electrical machine control systems [7578]. The main reason of using nonlinear control techniques is to improve the existing control system and achieve high performance. Linear control methods rely on the key assumption of small range operation for the linear model to be valid. When the required operation range is larger, a linear controller is likely to perform very poorly or to be unstable, because the nonlinearities in the system cannot be properly
Consider a thirdorder system "1 = sin x 2 + ( x 2 + 1) x3 x " 2 = x15 + x3 x " 3 = x12 + u x y = x1 . The same control technique is used in control scheme I. The principles of nonlinear control inputoutput linearization with decoupling are used to design the controllers [71. .compensated for. The literature [73] has successfully used this control technique to control a doublyfed reluctance machine with axially laminated anisotropic rotor and singlephase control windings fed by a dc source. Finding a direct and simple relation between the system output and the control input constitutes the intuitive basis for the socalled inputoutput linearlization approach to nonlinear control design. However. The technological breakthroughs in digital signal processor (DSP) have made it possible to implement complex nonlinear control algorithms in the application of the electrical machine control.73]. Let us use an example to demonstrate this approach [72].72. In control scheme I. nonlinear controller is used to generate reference signals Vdp and Vqp for the voltage source inverter with space vector pulse width modulation technique. nonlinear controller may handle the nonlinearities in large range operation directly.
x 2 . let us define the error as e = y (t ) − y d (t ) where yd(t) is desired output. For instance. x3 ) y where f1 ( x1 . Now an explicit relationship between y and u has been found. x 2 . Choosing the new input v as "d − k1e − k 2 e " v=" y with k1 and k2 being positive constants. the nonlinearity above equation is canceled. and we obtain a simple linear doublyintegrator relationship between the output and the new input v. the error of the closed loop system is given by " " + k2e " + k1e = 0 e . let us differentiate again. " "=v y The design of a controller for this doublyintegrator relation is simple. let us differentiate the output y "=x "1 = sin x 2 + ( x 2 + 1) x3 . because of the availability of linear control techniques.To generate a direct relationship between the output y and the input u. x3 ) = ( x15 + x3 )( x3 + cos x 2 ) + ( x 2 + 1) x12 . We now obtain Since y " " = ( x 2 + 1)u + f1 ( x1 . y " is still not directly related to the input u. If we choose the control input to be in the form u= 1 (v − f 1 ) x2 + 1 where v is a new input to be determined.
54) (7. while the internal dynamics are comprised of nr states (n is total number of the system dynamic states).55) .10). (c) Study the stability of the internal dynamics.56): . the input variables are Vdp and Vqp are chosen as the input variables while ωr and Iqp are chosen as output variables. ) (7. (b) Choose the input to cancel the nonlinearities and guarantee tracking convergence. e(t) converges to zero exponentially.55) are combined as a matrix equation (7. perfect control is achieved. . and using the condition Ids = 0 because of the control windings connected with a DC current source."(0) = 0 .1) through (7. From Equations (7. The total number of differentiation for all the outputs is called the relative order. ) Equations (7. the following differential equations are obtained " = 1 V − Rp I − ω I I qp qp qp r dp Lp Lp ""r = ω ' .otherwise.1) and (7. the basic approach to design an inputoutput linearilized controller includes three steps: (a) Repeatedly differentiate an output function until it is explicitly related to the input. If initially e(0) = e e(t) ≡ 0.ZR p * Zω r Lm I qs Z ( I dp + Zω r I qp + Vdp + + − + Lp ( Lp Lp . Therefore.3Pr2 * where Z = + + 2J ( ( I qs Lm . then which represents an exponentially stable error dynamics. The inputoutput linearization and decoupling process ensures linear relationship between input and output variables with outputinput pairs decoupled each other. In the dynamic system described by Equations (7.10).54) and (7.
Hence. With expressions for Vdd and Vqq from Figure 7. "% " + 1 " &Vqp # % " Rp − I qp − ω r I dp % " Lp " L # % " p & # (7.1) and (7.56) This equation is linearized and decoupled and made to take the form of Equation (7.60) are yielded Vdd = L p pI dp = Lp Z K aYdp + Lp Z F1 (t ) (7.59) and (7.'Z ""r $ % L p 'ω % %I " "= & qp # % 0 % & ' ' ZR p $ Zω r Lm I qs $ 0" − I dp + Zω r I qp + % " 'Vdp $ % L p L p ". %V " = % Z Z " & qp # % L K Y + L ( F ( t ) − B ( t )) p 1 qp p 2 2 & # (7. Equation (7. Equations (7.58) are equated to the power winding qd voltage Equations (7.57) where Ydp and Yqp are new input variables.57) Equating Equations (7.60) are linear and decoupled from each other.60) Vqq = L p pI qp = L p K aYqp + Lp F1 (t ) .56).59) and (7.59) (7.58) If Equation (7. input variables Vqp and Vdp are determined and given as Lp Lp $ 'Vdp $ ' ( F1 (t ) − B1 (t )) " K aYdp + .58) is simplified as 'Vdp $ 'Vdd %V " = %V & qp # & qq − B1 (t ) $ .57) and (7.6 and substituting Equation . ""r $ ' K a 'ω %I " "=%0 & qp # & 0 $ 'Ydp $ ' F1 (t ) 0 $ %Y " + % " K a # & qp # & 0 F2 (t )" # (7.61) The quantities Vdd and Vqq are selected to be dependent on controlled and state variables to ensure that Equations (7.2). Cascaded IP (integralproportional) controller structure is adopted and shown in Figure 7.6. − B2 (t ) " # (7.
(7.6.62) ωr K d Kc Z . The Butterworth method locates the eigenvalues uniformly in the lefthaft Splane on a circle Iqp * +  Kq ! Vqq + Ka Iqp (a) Kqq Iqp * ω r + Idp  Kd ! + +  Kb Kc Vdd ωr Kdd ωr (b) Figure 7.63) ∗ The reference speed and reference qaxis power winding currents are ω r∗ and I qp .597. Integral proportinal linear and decouping controller structure.60). (b) rotor speed loop. Kd and KaKb are chosen to optimize the closedloop eigenvalue locations using the Butterworth polynomial [71]. respectively. . The parameters transfer functions Kq. the following transfer functions are obtained (with the disturbance load torque ignored): I qp ∗ I qp = Kq Lp S 2 + K a S + K q (7. = ∗ 3 2 ω r L p S + K c S + ZK b K c S + ZK c K d (7.10) into Equations (7. (a) Qaxis power winding current loop.
However. K c = 2ω o L p . also can be transferred as Kd Kc Z h h2 h3 = 1 + + Lp S + K c S + ZK b K c S + ZK c K d s + µ 1 s + µ 2 s + µ3 3 2 where h1 + h2 + h3 = 0 h1 ( µ 2 + µ3 ) + h2 ( µ1 + µ3 ) + h3 ( µ1 + µ 2 ) = 0 . the parameters of the speed control loop are given as Kb = ωo Z . K a = 2ω o Lp . K d = ω o2 2Z . (7. The Butterworth polynomials for a transfer function with a secondorder denominator is given as 2 S 2 + 2 Sω o + ω o =0 (7.58). with its center at the origin.63).66) and (7. (7.67) To realize nonovershoot current response.65) For a transfer function with a thirdorder denominator.of radius ωo. we need to use the design procedure described in the references [66.3.62) can be computed using the same procedure described in section 7.66) From Equations (7. (7.63).64) comparing Equations (7. the parameters of secondorder current control loop described by Equation (7.67] to obtain the parameters of a thirdorder speed control loop with nonovershoot speed response. the Butterworth polynomial is 2 3 S 3 + 2ω o S 2 + 2ω o S + ωo = 0. we have 2 K q = L pω o . The thirdorder speed control loop transfer function expressed in Equation (7.60) and (7.
2 2 . h2 . h3 ) = 0. Lp To obtain µ1. µ 2 . µ 2 . h3 ) = e1 − K 2 (e2 + e3 ) = 0 e2 = K1e3 K1 > 0 K 2 < −1 e1 = K 2 (e2 + e3 ) 2 µ 2 = µ1 + c2 2 2 µ3 = µ1 + c2 + c3 where c2 and c3 are finite positive constants. The detailed derivation and its physical meaning for this nonlinear equation system are described in the reference [66. h2 . µ 2 .67]. h3 ) = e2 − K1e3 = 0 f5 ( µ1 . µ 3 .µ1 h1 + µ2 h2 + µ3 h3 =1 µ1 + µ 2 + µ3 = Kc Lp Kb Kc Z Lp µ1µ 2 + µ 2 µ3 + µ1µ3 = µ1µ 2 µ3 = K d Kc Z .9 − { − (h2 + h3 ) µ1 [1 − e − µ1t re ] + µ2 h2 [1 − e − µ 2 t re ] + µ3 h3 [1 − e − µ 3t re ]} = 0 f 4 ( µ1 . the following nonlinear equations can be constructed: f1 ( µ1 . h3 ) = − (h2 + h3 ) + h2 + h3 −1 = 0 µ1 µ2 µ3 f3 ( µ1 . h2 . µ3 . h2 . µ 2 . µ3 . h3 ) = −( h2 + h3 )( µ 2 + µ3 ) + h2 ( µ1 + µ3 ) + h3 ( µ1 + µ 2 ) = 0 f 2 ( µ1 . µ3 . µ3 . h2 . µ 2 . µ2 and µ3 so that we can compute the parameters of nonovershoot speed response for this thirdorder speed control loop.
6 Voltage Source SPWMInverter A voltage source inverter using spacevector pulsewide modulation technique [47][69][70] is applied for the control scheme II. Supply Vpa C Vdc Ipa Vpb Ipb DFM Vpc Ipc A’ B’ C’ Rectifier () Inverter Figure 7. 7. A Matlab program used to obtain these parameters is listed at Appendix 7D.After solving this nonlinear equation system. .C. the parameters of the controllers can be computed using the expressions as follows: K c = L p ( µ1 + µ 2 + µ3 ) Kb = ( µ1µ 2 + µ 2 µ3 + µ1µ 3 ) Z ( µ1 + µ 2 + µ3 ) Kd = Z ( µ1 + µ 2 + µ 3 ) µ1µ 2 µ 3 . (+) Is L A B C A.7. The performance of this inverter is simulated using Matlab/Simulink. Circuit diagram of a threephase VSI.
On the other hand. then the potential of the corresponding output terminal is unknown to the control system of the inverter. the power switches in a given branch must never both be in the ONstate. only eight logic states are permitted for the whole power circuit. since this would constitute a short circuit. The circuit has a bridge topology with three branches (phases). if both the switches are in the OFFstate.69) (7.68) (7.69) .7. Defining the switching variables as 10 if A is OFF and A ' is ON S1 = 0 ' /1 if A is ON and A is OFF 10 if B is OFF and B ' is ON S2 = 0 ' /1 if B is ON and B is OFF 10 if C is OFF and C ' is ON S3 = 0 ' /1 if C is ON and C is OFF the instantaneous values of the linetoline output voltages of the inverter are given by Vab = Vdc ( S1 − S 2 ) Vbc = Vdc ( S 2 − S 3 ) Vca = Vdc ( S 3 − S1 ) where Vdc is the DC supply voltage of the inverter. diodebased rectifier.A diagram of the power circuit of the inverter is shown in Figure 7. Since only two combinations of states of the switches in each branch are allowed. In effect. (7. In the circuit. each consisting of two power switches and two freewheeling diodes. a switching logic variable can be assigned to each phase of the inverter. via a DC link that contains as LC filter in the inverted Γ configuration. The inverter is supplied from an uncontrolled.
75) 1 Vc = (2 S 3 − S1 − S 2 ) . we can obtain 1 1 3 Vqd = Vdc ( S1 − S 2 − S3 ) + j ( Vdc )( S 2 − S3 ) . Each vector corresponding to a given state of the inverter are listed in Table 7.73) hence.78) .70) in Equation (7. The space vector diagram of linetoneutral voltages of the VSI are shown in Figure 7.72) (7.73).4.717. the linetoneutral voltages of the inverter are given by Va = Vdc ( 2S1 − S 2 − S 3 ) 3 (7.76) (7.75).687.1. the linetoneutral voltages can be calculated from the linetoline voltages as 1 Va = (Vab − Vca ) 3 1 Vb = (Vbc − Vab ) 3 1 Vc = (Vca − Vbc ) 3 (7. 3 By performing abcqd transformation for Equation (7.In balanced threephase systems.747.77) The output voltages can be represented as space vectors in the stator reference frame.74) 1 Vb = ( 2S 2 − S 3 − S1 ) 3 (7. The relationship between the space vector and three phase voltages are expressed as # V = Va (t ) + a ⋅ Vb (t ) + a 2 ⋅ Vc (t ) (7.71) (7. after substituting Equation (7. 2 2 2 (7.
The magnitude of output voltage is adjustable if the modulation index M. is used and Vm is replaced by MVm .80) Vc (t ) = Vm cos(ωt + φ + (7. 1 1. 0. 1 7 # V7 = 0 .78). 0. Inverter states and Spave Vector S1. 1 1. Va(t). Vb(t). 0 0.where a = e −j 2π 3 . S3 i 0. 1. Substituting Equations (7. 0 1. 0 2 0. 3 2 2π 2π (7. 1. 0. 1.1. S2.81) into (7. we can obtain Table 7. 0#M#1. 0. 1. and Vc(t) are three phase output voltage: Va (t ) = Vm cos(ωt + φ ) = Vm j (ωt +φ ) [e + e − j (ω t + φ ) ] 2 2π 2π (7.79) Vb (t ) = Vm cos(ωt + φ − j (ωt +φ − ) − j (ωt +φ − ) V 2π 3 3 ) = m [e +e ] 3 2 j (ωt +φ + ) − j (ωt +φ + ) V 2π 3 3 ) = m [e +e ].81) φ is initial phase angle of the three phase voltages and Vm is the peak phase voltage.79) through (7. 0 3 4 5 6 # V3 = −Vdc # V4 = Vdc # V 3 V5 = dc − j Vdc 2 2 # V 3 V6 = dc + j Vdc 2 2 1. 1 0 1 # Vi # V0 = 0 # V 3 V1 = − dc − j Vdc 2 2 # V 3 V1 = − dc + j Vdc 2 2 0.
located in a given sector.8 and three phase output voltages finish one period.82) # When α changes from 0 to 2π. and Vo. which are framing the sector. Vj..8. can be synthesized as a linear combination of the two adjacent base vectors. # T T T V * = 1 Vi + 2 V j + o Vz . 2 (7. i. α = ωt + φ .# 3 V = MVme jα . we can see the nonzero base vectors divide the # cycle into six. (b) pulse patterns of symmetrical threephase modulation.83) T1 . V rotates one circle in Figure 7. and To are.8. and either one of the two zero vectors. 60o –wide sectors. we can obtain the following equation q V2 Ts Ts V6 V* V3 V7=V0 =0 α d V4 A B C To 2 To 2 To 2 To 2 T1 T2 T2 T1 V1 V5 (a) (b) Figure 7. Schematic of the space vector PWM. V * . . T2. Assuming that space vector V* locates in the sector as shown Figure 7. (a) Illustration of the space vector PWM strategy. respectively. Ts is the switching interval of the power switching components. the working time of state Vi .8. Vz is either Vo or V7. Ts Ts Ts (7. The desired voltage vector. In Figure 7.e. Vi and Vj .
2 2 This equation is equivalent to Equation (7. N = 0.6 .by substituting Vi = V4. which can be extended as follows: # V = 3 3 MVdc cosα + ( MVdc sin α ) ⋅ j . T2 and T0 are given as Equation (7. into Equation (7.1. 2 2 Hence. If 0 ≤ α ≤ 60o . the general expression to calculate the state working time T1.83): 3 1 V * = Vdc (T1 + T2 ) + ( VdcT2 ) ⋅ j . 2 2 Ts M T1 and T2 can be easily obtained by solving this equation system and are expressed as T1 = Ts M sin(60o − α ) T2 = Ts M sin α . Vj = V6.84): T1 = Ts M sin(60o − α ' ) T2 = Ts M sin α ' (7.84) To = Ts − Ta − Tb . whose values are looked up in Table 7.. and Vz=V7. we can construct the equation system as follows: Ts M 1 3 cosα = T1 + T2 2 2 3 3 sin α = T2 . If 60o ≤ α ≤ 360o . by equating their real parts and imagnary parts. When V* locates in any sector.82).... α ' = α − N ⋅ 60o . α ' = α .1.
VI. T1 + T2 (7. II.86) To' = 0. V.4(a).2.85) and (7. State Sequence 467764 620026 237732 310013 157751 540045 . (7. State Sequence of Each Sector State sector I.86) to generate the best approximate of the desired voltage vector. III. T1 + T2 Ts .2. Vdc If the overmodulation mode (T1 + T2 > Ts) occurs. The state sequence corresponding to each sector is listed in Table 7.where Vm is peak phase voltage and modulation index is expressed as M = 6Vm . Table 7. IV.87) The symmetrical pulse patterns of zerovector for one sampling period are illustrated in Figure 7. T1' = T1 ⋅ Ts . the time duration should be scaled as Equations (7.85) T2' = T2 ⋅ (7.
Vab = −(Vbc + Vca ). T2Vdc = −Vbc Ts . ∗ Vab =− T T1 ∗ ∗ ∗ ∗ Vdc . Ts Ts In Sector IV: ∗ ∗ T1Vdc = −Vab Ts . Vab = −(Vbc + Vca ). Vab = 2 Vdc . ∗ Vbc = T T1 ∗ ∗ ∗ ∗ Vdc .The time intervals T1. ∗ ∗ ∗ ∗ Vab = T1 T ∗ ∗ ∗ ∗ Vdc . Vbc = 2 Vdc . T2Vdc = −Vab Ts . Ts Ts In Sector II: ∗ ∗ T1Vdc = −Vca Ts . T2. Ts Ts In Sector V: ∗ ∗ T1Vdc = Vca Ts . Vbc = − 2 Vdc . T2Vdc = Vab Ts . Ts Ts In Sector III: ∗ ∗ T1Vdc = Vbc Ts . Vb and Vc of the VSI: In Sector I: ∗ ∗ T1Vdc = Vab Ts . Vbc = −(Vab + Vbc ). T2Vdc = Vca Ts . Vca = 2 Vdc . Ts Ts . and Ts are used to calculate the reference output voltages Va . T2Vdc = Vbc Ts . Vca = −(Vab + Vbc ). ∗ Vca = T1 T ∗ ∗ ∗ ∗ Vdc . ∗ Vca =− T1 T ∗ ∗ ∗ ∗ Vdc . Vca = −(Vab + Vbc ). Vab = − 2 Vdc .
88) Vc∗ = ∗ ∗ Vca − Vbc .3. ∗ ∗ ∗ 50 2 1 V pa (V) 0 Ip a(A) 0. then the interval times T1 and T2 are obtained from Equation (7.9 shows the simulation waveforms of power winding phase voltage and current whose frequency is equal to 60Hz with 5KHz switching frequency. T2Vdc = −Vca Ts .04 0 . Vca = − 2 Vdc . ) T(s e c . 3 In Figure 7.1 0 1 2 50 0 0 0 . ∗ Vbc =− T1 T ∗ ∗ ∗ ∗ Vdc .06 0 . 02 0. the control signals Vqp and Vdp are used to computed the modulation M and initial phase angle φ.88) are used to compute output phase voltages Va . Vb and Vc are given as Va∗ = ∗ ∗ Vab − Vca 3 Vb∗ = ∗ ∗ Vbc − Vab 3 (7. 04 0 . Figure 7.84).02 0 . Equation (7.In Sector VI: ∗ ∗ T1Vdc = −Vbc Ts . Vab = −(Vbc + Vca ). Vb . Vbc and Vca are linetoline voltages and phase voltages Va .08 0 .08 0 .1 T(s e c . Ts Ts ∗ ∗ ∗ ∗ ∗ ∗ Vab . Finally. and Vc of the VSI.06 0 . ) (a) (b) .
The width of the hysteretic loop.7 Current Control in Voltage Source Inverters A block diagram of a currentcontrolled VSI that is used in control scheme II is shown in Figure 7. are then applied to the hysteretic current controllers. must be increased. current Ia is unacceptably lower than the reference current. ∆Ia.10. is greater than h/2. denoted by h.10. The output currents.9. i. Ib. represents the tolerance bandwidth for the controlled current. 7. (b) power winding current waveform.Figure 7. the corresponding linetoneutral voltage. and ∆Ic. hence it is this variable that is regulated by the controller. Ia. I a .e.. Simulation of the voltagecontrolled PWMVSI. and Ic. which generate switching signals for the inverter switches. I a . and I c . Current error signals. ∆Ia. This voltage is most strongly affected by the switching variable a. If the current error. “on” state of a power switch component) in the Ipa + ∗ ∗ ∗ ∗ Vdc  Inverter  Ipb Ipc + DFM I* pa I* pb * + Ipc (a) ∆Ipa ∆Ipb + ∆Ipc a 1 (b) h/2 Hysteretic Current Controller h/2 0 ∆Ia Hysteretic Current Controller inputout characteristic . of the inverter are sensed and compared with the reference current signals. The inputoutput characteristic of the phaseA hysteretic current controller is shown in Figure 7. (a) Power winding voltage waveform. ∆Ib. and is set to a logic variable 1 (equivalently. I b . Va..
8 S W1.2 0 0 0 .2 0 0 0 .02 T(s e c ) 0. 04 (a) (b) 200 100 Vp a (V) 0 100 200 0 0.4 0 .005 0 .6 0 .03 0.01 0. (a) Phase current with 2% tolerance bandwidth. Simulation of the currentcontrolled VSI. Block diagram of a currentcontrolled VSI.015 0. off = ‘0’) with 10% tolerance bandwidth. (e) transistor switching signal (on = ‘1’. 02 T(s e c ) 0.02 T(s e c ) 0. (d) phase voltage with 10% tolerance.01 5 0 .4 0 . 0 .10. 1 0 . 02 T(s e c ) 0. 03 0.6 0 . . (f) transistor switching signal (on = ‘1’.04 Vp a (V) 2 00 1 00 0 1 00 2 00 0 0. 04 (c) (d) 1 0 . 01 0.01 T(s ec ) 0 . off = ‘0’) with 2% tolerance bandwidth.0 1 T(s e c ) 0 . (c) phase voltage with 2% tolerance. 03 0.01 0. (b) phase current with 10% tolerance bandwidth.8 0 .0 2 (e) (f) Figure 7.0 2 S W1. 01 0.Figure 7.03 0.00 5 0 .11.04 2 0 0 2 2 4 4 0 0. 4 4 2 Ip a(A ) Ip a(A) 0 0.
2 s.described situation. The parameters of IP controller in control scheme II and nonlinear controller in control scheme I are listed in Appendix 7A. The load torque is 4 N. The width. The other two controllers operate in a similar manner. of the tolerance band affects the switching frequency of the inverter.11.8 Simulation Results Below the Base Speed The simulations below illustrate the control performance of two control schemes presented in this chapter. Figure 7. “off” state of a power switch component) in order to decrease the voltage and current in question. .11(c) and (d) show the transistor switching signals with 10% tolerance bandwidth and 2% tolerance bandwidth. The typical examples of constant torque application below the base speed are actuators and servo systems.5 s to 2.m from 0 to 1. It is obvious that the narrower the bandwidth. Conversely. which require maximum torque availability at all speeds to ensure maximum dynamic response. the more frequent switching takes place and the higher quality of the currents. No action is taken by the controller when current Ia stays within the tolerance band. depicting simulation waveforms of power winding current and voltage for an inverter supplying the power windings of the machine at values of h equal to 10% in Figure 7. 7. The narrower the band. an error less than –h/2 results in a = 0 (equivalently.5 s and changes to 8 N.11(a) and 2% in Figure 7. the higher the switching frequency of transistor. The control winding current is chose as 8 A. This is illustrated in Figure 7.m from 1. To investigate the performance of doublyfed synchronous reluctance motor drive system below the base speed. of the amplitude of the reference current.11(b). the following simulation are used to demonstrate the dynamic response to step speed change and step torque change. h.
which are decided by q axis and daxis current componenets.12 illustrate the dynamic response of the motor drive for a step change of the reference speed from zero to the rated electrical base speed. there is no overshoot speed response in the simulation results.13(b) to (d) show this dynamic process.12(b) to (d).S e p 1 9 9 8 (a) (b) . 5 T(s ec ) 1 1. which can be seen in Figure 7.Figure 7. The qaxis current of the power winding aligns to the zero value so that the field orientation requirements are satisfied. The step change of the load torque causes a speed dip as shown in Figure 7. and at last the torque reaches an constant torque that is equal to 4 N. The response to change load torque from 4 N.12(e) to (f). 2 0 0 0. 6 0. When the speed reaches the base speed. Because we use nonovershoot design procedure to obtain the control parameters for twoorder speed control loop (control scheme II) and for threeorder speed control loop (control scheme I). power winding qaxis current has a little ripple around the zero value and aligns to zero value after a short transient. In Figures 7. The dynamic response of the power winding qaxis and daxis flux 1 W rm (p. the torque and power winding daxis current increase quickly to reach their maximum values. we can see that.12(a). 8 0.13. are illustrated in Figures 7.m to 8 N. but it recovers its value after a short transient because power winding daxis current sharply increases to generate a bigger electrical torque against the change of load torque.m load torque for this simulation example. 4 0.13(a).m at t = 1. Figures 7.5 s is displayed in Figure 7. The power winding qaxis and daxis flux linkages. The daxis power winding current generating the corresponding torque also reachs its steadystate value. u ) 0. During this process. then they begin to decrease when the speed closes on the base speed. 5 2 9.
1. 4 1. 1 1. 6 1. (f) power winding qaxis fluxlinkage. u) 1 0.0 5 Wrm (p. 2 1. (a) Reference and actual speed. (e) power winding daxis fluxlinkage. (b) torque (c) power winding daxis current. 9 1.(c) (d) (e) (f) Figure 7. 2 (a) (b) .9 5 0. 8 T(s ec ) 2 2. Dynamic response to a stepchange of speed. (d) power winding qaxis current.12.
(e) power winding daxis fluxlinkage. (f) power winding qaxis fluxlinkage. Response to change of load torque. (a) (b) . (a) reference and actual speed.(c) (d) (e) (f) Figure 7. (d) power winding qaxis current. (b) torque (c) power winding daxis current.13.
m torque load.14. To obtain constant power control above base speed.5 s to 4 s.m load torque is used from 0 to 1. daxis and qaxis currents of the power windings are illustrated in Figures 7. sharply reach their maximum values which cause the speed to quickly increases. which operate in a 2.m from 1.5 s and then changes for 8 N.15(a) shows the speed response to this twostage speed step change command.Figure 7. the voltage and current sharply increase and reach the steadystate values corresponding to a 4 N.15(c). is align to zero below the base speed and .5 s due to change load torque from 4 N. Idp.m to 8 N. Figure 7. 7. While the qaxis current.m torque load. They are affected by the change of the qaxis and daxis current components.14. The control winding current is chose as 8 A.5 s.13(e) to (f). They increase at 1. where we can see that the torque and its control current. Finally they reach new steadystate values corresponding to 8 N. (a) Phase current. The torque. The following three cases demonstrate the dynamic process of a doublyfed synchronous reluctance motor drive systems.m. A familiar example is the electric vehicle traction that requires high torque for lowspeed acceleration and constant power (reduced torque) for highspeed cruising.15(b) and 7.5 s and the base speed from 2.5 s to 4 s.5 times base speed range and rotate in two directions. linkages are illustrated in Figure 7. from 0 to 1. Case 1. A 4 N.5 times the base speed from 0 to 2. The profiles of the power winding phase current and voltage are displayed in Figure 7. fieldweakening techniques are used. Profiles of the phase current and voltage during the simulation process. We can see.9 Simulation Results Above the Base Speed Some industrial and commercial applications require motors to operate over a wide speed range and rotating in both directions. (b) phase voltage. Iqp. We simulate the dynamic process of the control system with twostage speed stepchange: 2.
5 s.16 gives the profile of phase current and voltage of power windings during the whole transient process.5 s to 2.m to 8 N. At t = 1. Phase current Ipa of the power windings goes through its second transient stage. Idp.shapely changes to a negative constant value above base speed because of the field weakening control. .5 times the base speed but load torque is changed from 4 N.15(e) and 7. which causes the rotor speed a little dip but it is recovered quickly because the daxis current Idp and its corresponding torque quickly increase to new values to resist the change of speed. In Figure 7. When the speed closes on the speed command.m. Finally the speed reach its command value.. sharply falling down to their minimum values.5 times the base speed and load torque is set to 4 N. Phase current Ipa of the power windings goes through its first transient stage. During this transient stage. the transient current quickly increases and reaches to a new steadystate value.5 times the base speed to the base speed. (2) From 1. we can clearly see that there are three transient stages in the whole transient process of phase current Ipa: (1) From 0 to 1. The qaxis is still kept in some negative values before the speed falls down to the base speed and then change to the zero again. Figure 7. During this transient stage.5 times the base frequency.m. which can generate a new electrical torque to balance the new load torque. transient current sharply increases to maximum value and then slowly decreases to a steadystate value.5 s. The current frequency is kept constant. speed command is equal to 2. At t = 2. torque and current also reach their steadystate values. which farther causes the rotor speed to quickly decrease. The daxis and qaxis flux linkages of power windings are shown in Figures 7.5 s. The frequency of the current increases step by step and then reaches 2.m. the load torque is changed from 4 N. the speed command proceeds to second stage by changing the speed command from 2. the torque and daxis current Idp begin increase in inversely proportional to speed.16(a). Then the speed increase is slowed down with the decrease in torque and qaxis current.m to 8 N. speed command is equal to 2. which are dependent on the dq current components of power windings and control windings. The speed change causes the torque and torque control current.5 s.15(f).
5 times base speed. Dynamic response to a wide speed range of 2.3 2 .5 4 (a) (b) (c) (d) (e) (f) Figure 7.5 Wrm (p . (f) power winding qaxis fluxlinkage. u ) 1 . (d) power winding qaxis current.5 0 . (a) Reference and actual speed.5 0 0 0 .5 1 . (b) torque (c) power winding daxis current.5 2 . (e) power winding daxis fluxlinkage. .15.) 3 .5 T(S ec .
As shown in Figure 7. Case 2. (4) The speed is kept constant at 0. a 4 N.m load torque is used and control winding current is chose as 8 A.5 times the base speed to 0. the command speed is seperated into six stages: (1) The speed slopes up to twice the base speed from zero in first second. Phase current Ipa of the power windings goes through its last stage. (3) The speed slopes down to 0. (3) From 2.5 s to 3.5 times base speed for 1 second.17(a). speed is changed from 2. (a)Phase current. (2) The speed is kept constant at twice the base speed for 1.m.5 seconds. . During this transient stage.5 s.5 times the base speed and torque is kept at 8 N. During the whole simulation process.(a) (b) Figure 7. Profiles of the power winding phase current and voltage during the simulation process.5 s to 4 s.16. (b)phase voltage. We simulate the dynamic process of the control system with trapezoid speed command. The current frequency decreases to the base current frequency. the transient current increases to some value and then decreases and finally reaches to a new steadystate value.5 times the base speed from 2.
(f) power winding qaxis fluxlinkage.) 4 .5 T(s ec . (a) Reference and actual speed.2 .5 (a) (b) (c) (d) (e) (f) Figure 7.5 1 0 . (b) torque (c) power winding daxis current. Dynamic response to a trapezoid speed command.5 3 .17. (e) power winding daxis fluxlinkage. .5 2 Wrm (p . (d) power winding qaxis current.5 0 0 1 2 . u ) 1 .5 5 .5 6 .
(a) Phase current. Figures 7. Figures 7.18(e) and (f) illustrate the dynamic process of daxis and qaxis flux linkages of power windings. The rotor speed follows the speed command very well and reaches the desired steadystate speeds: twice the base speed or 0. Figure 7.5 times the base speed.Figure 7.5 s to 5. we can clear see that there are six transient stages corresponding to the sixstage speed command.18. (6) The speed is kep constant at twice the base speed for 1. The dynamic process of the phase current and voltage of the power windings is shown in Figures 7.17(b) and (c) illustrate the dynamic process of daxis current of power windings. . (5) The speed slopes up to twice the base speed again in the time of 4. When the speed reaches its stadystate values: 0. which is align to zero when speed is below base speed or becomes negative value when the speed is above the base speed. (b) phase voltage.5 times the base speed and 2 times the base speed.17(d) displays the dynamic process of qaxis current of power windings.5 seconds.19(a) and (b). The upslope speed causes the torque control current Idp and its coresponding torque increament while the downslope speed causes them decreament. Idp and its corresponding torque also reach their teadystate values. Profiles of the power winding phase current and voltage during the simulation process.5 s. Their profiles look similar to those of the daxis and qaxis currents of the power windings because they are mainly dependent on theses currents.
The steadystate rotor speeds can also be obtained in both directions . We simulate the dynamic process of the control system with a twodirection trapezoid speed command. a 4 N. (3) The speed slopes down to negative twice the base speed from 2. u ) (a) (b) . Idp and torque increase quickly and then keep at some values until the rotor speed closes its desired speed.5 s.Case 3.m load torque is used and control winding current is choosen as 8 A. Figures 7. The rotor speed follows this speed command very well either in clockwise direction or counter clockwise direction. (4) The speed is kept constant at negative twice the base speed for 1.5 Wrm (p.5 s. (6) The speed is kep constant at twice the base speed for 1. When the speed command is on its downslope sides. the speed command is seperated into six stages: (1) The speed slopes up to twice the base speed from zero in first second.19(a). ) 8 9 .5 s.19(b) and (c) display the dynamic process of the torque control current Idp and electrical torque. Idp and 2 1 0 1 2 0 1 2 . Idp and torque quickly decrease and hold at some values which generate an electrical torque to balance the load torque.5 6 T(s e c. When the speed command is on its upslope sides. (5) The speed slopes up to twice the base speed in the time of 6 s to 8 s. During the whole simulation process.5 seconds. As shown in Figure 7.5 4 . At the monent of the rotor speed reaching its stedystate values. (2) The speed is kept constant at twice the base speed for 1.5 s to 4.
(e) power winding daxis fluxlinkage. (b) torque (c) power winding daxis current. (f) power winding qaxis fluxlinkage. (d) power winding qaxis current. (a) (b) . (a) Reference and actual speed. Dynamic response to a twodirection trapezoid speed command.19.(c) (d) (e) (f) Figure 7.
At the monent of the rotor speed reaches its steadystate values. the voltagecontrolled PWMVSI is simulated using average voltage of switching period to replace true pulsating voltage. Idp and torque quickly increase and reach their steadystate values for balancing the load torque. hence it is approximate simulation. Above simulation results have demonstrated the dynamic performance of control scheme II (using IP and hysteretic current control).19(d) shows that qaxis current Iqp is align to zero below the base speed or is align to the negative values above the base speed because of the field weakening control. if the trapezoid speed command is used.9 for voltage . The qaxis and daxis flux linkages of the power windings are illustrated in Figures 7. However. The simulation waveforms are shown in Figure 7.Figure 7. torque decrease quickly and then keep at some values so that the rotor speed quickly decreases and closes its desired speed. It is observed from Figure 7. Figures 7. this is noted that. they are dependent on the qaxis and daxis currents of both the power windings and control windings. the change of the torque control current Idp and torque are not as sharp as those of the step change speed command.20(a) that there are six transient stages corresponding to the speed command. A few factors cause transient difference between control scheme I and control scheme II in case 3: (1) Two control schemes use different controller. Profiles of the power winding phase current and voltage during the simulation process.19(e) and (f). Figure 7. (a)Phase current.21 and 7. Figure 7.20 illustrates the profiles of the phase current and voltage of the power windings during the whole dynamic process.20. The control scheme I (using linearization control) also has similar simulation results. By comparing three cases. using the trapezoid speed command is a good way to avoid the too big command current Idp*. the currentcontrolled PWMVSI in control scheme II is simulated directly using true pulsating voltage. (2) In control scheme I. (b)phase voltage. One uses IP controller. Hence.22 give the evidence that the dynamic performance is the same as one of control scheme II in case 3. the another uses linearization controller.
5 6 T (s e c . 2 0 .1 0 .5 6 T (s e c . 2 7 6 To rqu e (N.1 5 0.22(b) just shows the approximate phasevoltage waveform (average voltage of switching period).u ) 5 4 3 2 1 0 1 2 0 0 1 2 .2 5 0. 5 4.2 0 .5 4 .) 8 9.5 1 Wrm(p . 5 KHz switching frequency is used in currentcontrolled PWMVSI while 2 KHz switching frequency is used in voltagecontrolled PWMVSI.) 8 9.controlled PWMVSI and Figure 7.m ) 0 1 2 .5 4 .5 6 T (s e c .5 6 0 1 2 .11 for currentcontrolled PWMVSI.5 6 T (s e c . 5 6 T (s e c .5 (a) 7 6 5 Id p(A ) Iq p(A ) 4 3 2 1 0 0 1 2 . ) 8 9 .5 (e) (f) .1 5 0. 1 0 .0 5 0 .5 (c) 0 .0 . Figure 7.5 6 T (s e c .0 5 (d) 0 1 2.5 0 1 2. ) 8 9 .5 4 .5 4. (3) The higher switching frequency is used in control scheme II.5 0 2 (b) 2 4 4 .20(b) shows the true phasevoltage waveform (pulsating voltage) while Figure 7. ) 8 9 .0 5 0 LAMD Aqp (Wb) LAMD Adp (Wb) 0. ) 8 9 .
Figure 7. (b) torque (c) power winding daxis current. (d) power winding qaxis current. Dynamic response to a twodirection trapezoid speed command for control scheme I.22.21. they should much close to the true results. (b)phase voltage. Although the dynamic responses of the control scheme II have more harmonic components. Profiles of the power winding phase current and voltage during the simulation process for control scheme I. 7. (a)Phase current. (a) (b) Figure 7. (a) Reference and actual speed.10 Conclusion . (e) power winding daxis fluxlinkage. (f) power winding qaxis fluxlinkage.
. are described in details and their performance is successfully simulated using Matlab/Simulink. which are used as a controllable source to supply the power winding of the machine.This chapter demonstrates the application of field orientation principle in a doublyfed synchronous reluctance machine with dc control windings. The simulation results show that the prescribed control performance is achieved for IP controller and nonlinear controller. Control scheme I with voltagecontrolled PWMVSI and control scheme II with currentcontrolled PWMVSI are also simulated using Matlab/Simulink. They both demonstrate the similar performance. The expressions to determine maximum torque and obtain fieldweakening operation performance above base speed are given and satisfactorily used in the simulation. Their effectiveness shows that they are worth to be verified by the experiment. The novel design methodology using inputoutput linearlization technique and Butterworth method are used for designing a nonlinear controller for the machine. The simulation results prove that this machine can be controlled like a separately excited dc motor. The machine is controllable in a wide speed range including constant torque region and constant power region. which can be used to obtain the performance of the nonovershoot speed response. Currentcontrolled PWMVSI and voltagecontrolled PWMVSI. A design procedure and programs to quantitatively decide the IP parameters for a secondorder speed loop are given.
The machine inherent parameters are obtained by using its steadystate equivalent circuits and experimental measurement. The potential applications of these generator systems had been revealed from the results of dynamic simulation and steadystate calculation. (b) Doublyfed reluctance synchronous generator systems. Their dynamic performances were simulated using MATLAB/Simulink and ACSL program. simulation. A accurate mathematical model. Based on this mathematical model. and application of a doublyfed reluctance machine. the mathematical models of a few generator systems were described. Some generator systems investigated in this dissertation include: (a) Selfexcited doublyfed reluctance generator systems. 217 . (c) Regulated DC power generation systems using doublyfed synchronous reluctance machine.1 Conclusion The research work on this dissertation has focused on modeling. which considers the coreloss of the machine. The steadystate characteristics of these generator systems were obtained by simulation and experimental measurement. and its corresponding dynamic qd equivalent circuits with series coresloss resistance or shunt coreloss resistance are presented.CHAPTER 8 CONCLUSIONS AND SUGGESTION FOR FURTHER WORK 8.
It provided a more accurate model for the performance evaluation and the design of doubly fed reluctance machines. These works revealed their potential applications. (3) Two field orientation control schemes for a doublyfed synchronous reluctance machine were investigated for highperformance operation both in the constant torque and constant power regions. An approach to determine machine parameters from measurement was also provided. One had IP controller and currentcontrolled voltagesource inverter. The simulation results prove that this machine can be controlled like a separatedexcited dc motor in a wide speed range including constant torque control below the base speed and field weakening control above the base speed. and another had linearization controller and voltagecontrolled PWM voltagesource inverter.The application of fieldorientation principle in a doublyfed synchronous reluctance machine with dc control windings was investigated. Two control schemes were investigated in the field orientation control system of doublyfed synchronous reluctance machine. were investigated and analysis were presented. The design procedures to obtain the parameters of controller were given. The research work has made contribution in these areas: (1) A dynamic model of doublyfed reluctance machine considering the effects of coreloss and saturation was presented. (2) The performance characteristics of doublyfed selfexcited reluctance generator and doublyfed synchronous reluctance generator as well as an innovative DC power generator scheme using doubly fed reluctance machine. . The simulation results have provided valuable information for realization and application of the control systems.
The simulation results have demonstrated that two fieldorientation control schemes of doublyfed synchronous reluctance machine are effective and robust.8. (2) Its core losses. To verify the theory set forth and the effective of these two control schemes.2 Suggestion for Further Work The research work revealed some inherent disadvantage of this experimental machine with simple salient laminated pole rotor structure. The disadvantages include: (1) The inherent oscillatory instability of this class of reluctance machine when fed with variablevoltage. . An axially laminated rotor structure will drastically reduce the effects of these disadvantages. saturation effect. proper rotor structure and optimization designs are necessary. experimental work will be necessary. variable frequency source under openloop control. We hope that the proposed further research will improve the performance of this kind of machine and promote its applications in industry. To eliminate these inherent disadvantages. military and power system given its low cost. and leakage inductances are relatively high. high reliability and flexible control methods.
q = 3 .APPENDIX 228 Appendix 2A The parameters of the experimental machine are: Number of power winding pole pair. P = 1 Number of control winding pole pair.
2p r pr n=0 − cos 2 ( K 1 p θ rm − θ p +  . Rp = 1. Rs’ = 0.Power winding perphase resistance. cos 2 p r −1 3µ 0 r ! N 2 3µ 0 r ! N 2 2 K 1pπn 2 K 1pπn p p πα − sin 2 ( K 1 p θ rm − θ p ) . Appendix 3A L qqp = 3µ 0 r ! N 2 p πα 2g1 + 3µ 0 r ! N 2 p 4g1K 1p sin K 1 p πα {cos 2 ( K 1 p θ rm − θ p ) 2p r 2 p r −1 n=0 . Np/Ns = 0. sin + }+ pr p 2 g 8 g K p 2 1 2 n=0 r {sin 2 ( K 1 p θ rm − θ p − 2 p r −1 n=0 K 1 p πα 2 p r −1 2 K 1 p π ( n + 1) K p πα + cos 2 ( K 1 p θ rm − θ p − 1 ) .439 Ohms Refered control winding perphase resistance. sin 2 K 1 p π ( n + 1) K p πα 2 p r −1 2 K 1pπn − sin 2 ( K 1 p θ rm − θ p + 1 ) . sin }. cos pr 2p r pr n =0 K 1 p πα 2 p r −1 2 K 1pπn ) . cos ) 2p r p 2p r n =0 r .723 ohms The ratio of the effective power winding to control winding turns.5.
sin − }+ pr pr 2g 2 8g 2 K 1 p n=0 {sin 2 ( K 1 p θ rm − θ p − 2 p r −1 n=0 K 1 p πα 2 p r −1 2 K 1 p π ( n + 1) K p πα + cos 2 ( K 1 p θ rm − θ p − 1 ) .L ddp = 3µ 0 r ! N 2 p πα 2g1 − 3µ 0 r ! N 2 p 4g1K 1p sin K 1 p πα {cos 2 ( K 1 p θ rm − θ p ) 2p r 2 p r −1 n=0 . 2p r pr n=0 − cos 2 ( K 1 p θ rm − θ p +  L qdp = 3µ 0 r ! N 2 p 4g1K 1p sin 2 p r −1 K 1 p πα 2 K 1pπn {sin 2 ( θ p − K 1 p θ rm ) . cos 2 p r −1 3µ 0 r ! N 2 3µ 0 r ! N 2 2 K 1pπn 2 K 1pπn p πα p − sin 2 ( K 1 p θ rm − θ p ) . cos pr 2p r pr n =0 K 1 p πα 2 p r −1 2 K 1pπn ) . cos 2p r pr n =0 2 p r −1 n =0 − cos 2 ( θ p − K 1 p θ rm ) . sin 2 K 1 p π ( n + 1) K p πα 2 p r −1 2 K 1 pπn ) . cos ) + sin 2 ( θ p − K 1 p θ rm + 1 2p r pr 2p r n =0 . sin }.  . sin }. sin 3µ 0 r ! N 2 2 K 1pπn p }+ pr 8g 2 K 1p {cos 2 ( θ p − K 1 p θ rm + 2 p r −1 n=0 K 1 p πα 2 p r −1 2 K 1 p π ( n + 1) K p πα ) . cos ) 2p r p 2p r n =0 r . 2p r pr n =0 − sin 2 ( θ p − K 1 p θ rm − . cos − cos 2 ( θ p − K 1 p θ rm − 1 pr 2p r pr n =0 K 1 p πα 2 p r −1 2 K 1pπn ) . sin 2 K 1 p π ( n + 1) K p πα 2 p r −1 2 K 1pπn − sin 2 ( K 1 p θ rm − θ p + 1 ) .
sin }. cos − sin 2 ( K 2 q θ rm − θ s + 2 pr 2p r pr n=0 K 2 q πα 2 p r −1 2 K 2 qπn ) . cos ) + cos 2 ( K 2 q θ rm − θ s − 2 2p r p 2p r n =0 r . cos {sin 2 ( K 2 q θ rm − θ s − 2 p r −1 n=0 K 2 q πα 2 p r −1 2 K 2 q π ( n + 1) K q πα ) . sin 2 K 2 q π ( n + 1) K q πα 2 p r −1 2 K 2 qπn ) . cos {sin 2 ( K 2 q θ rm − θ s − 2 p r −1 n=0 K 2 q πα 2 p r −1 2 K 2 q π ( n + 1) K q πα ) . 2p r pr n=0 − cos 2 ( K 2 q θ rm − θ s +  . cos ) + cos 2 ( K 2 q θ rm − θ s − 2 2p r pr 2p r n =0 .L qqs = 2 2 3µ 0 r ! N s K q πα πα 3µ 0 r ! N s sin 2 {cos 2 ( K 2 q θ rm − θ s ) + 2p r 2g1 4g1K 2 q 2 2 2 p r −1 3µ r ! N s πα 3µ 0 r ! N s 2 K 2 qπn 2 K 2 qπn }+ 0 − sin 2 ( K 2 q θ rm − θ s ) . sin K q πα 2 p r −1 2 K 2 qπn 2 K 2 q π ( n + 1) ) . sin − pr pr 2g 2 8g 2 K 2 q n =0 2 p r −1 n=0 . 2p r pr n=0 − cos 2 ( K 2 q θ rm − θ s +  L dds = 2 2 K q πα 3µ 0 r ! N s πα 3µ 0 r ! N s sin 2 {cos 2 ( K 2 q θ rm − θ s ) − 2g1 4g1K 2 q 2p r 2 2 2 p r −1 3µ r ! N s πα 3µ 0 r ! N s 2 K 2 qπn 2 K 2 qπn }+ 0 − sin 2 ( K 2 q θ rm − θ s ) . cos − sin 2 ( K 2 q θ rm − θ s + 2 pr 2p r pr n=0 K 2 q πα 2 p r −1 2 K 2 qπn ) . sin + pr pr 2g 2 8g 2 K 2 q n =0 2 p r −1 n=0 . sin }.
sin 2 K 2 q π ( n + 1) K q πα 2 p r −1 2 K 2 qπn ) . 2p r pr n=0 − sin 2 ( θ s − K 2 q θ rm − . sin }.  . cos − cos 2 ( θ s − K 2 q θ rm − 2 pr 2p r pr n =0 K 2 q πα 2 p r −1 2 K 2 qπn ) . cos ) + sin 2 ( θ s − K 2 q θ rm + 2 2p r p 2p r n=0 r . cos 4g1K 2 q 2p r pr n=0 2 p r −1 n =0 2 2 K 2 qπn 3µ 0 r ! N s }+ pr 8g 2 K 2 q − cos 2 ( θ s − K 2 q θ rm ) . sin {cos 2 ( θ s − K 2 q θ rm + 2 p r −1 n=0 K 2 q πα 2 p r −1 2 K 2 q π ( n + 1) K q πα ) .L qds = L dqs = 2 2 p r −1 3µ 0 r ! N s 2 K 2 qπn K q πα sin 2 {sin 2 ( θ s − K 2 q θ rm ) .
sin . cos 2p r pr n =0 ( K 1 p − K 2 q ) πα 2 p r −1 ( K 1p − K 2 q )πn ] . sin ( K 1p + K 2 q )πn ( K 1 p + K 2 q ) πα + cos[( θ p + θ s ) − ( K 1 p + K 2 q ) θ rm − ] pr 2p r ( K 1 p + K 2 q ) πα ( K 1p + K 2 q )πn + sin[( θ p + θ s ) − ( K 1 p + K 2 q ) θ rm + ] 2p r pr ( K 1p + K 2 q )πn ( K 1 p + K 2 q ) πα − cos[( θ p + θ s ) − ( K 1 p + K 2 q ) θ rm + ] pr 2p r 3µ 0 r ! N p N s ( K 1p + K 2 q )πn }+ {− sin[( θ p − θ s ) − ( K 1 p − K 2 q ) θ rm pr 4g1 ( K 1p − K 2q ) 2 p r −1 n=0 2 p r −1 n=0 2 p r −1 n=0 − ( K 1 p − K 2 q ) πα 2 p r −1 ( K 1 p − K 2 q ) πn + cos[( θ p − θ s ) − ( K 1 p − K 2 q ) θ rm ] . sin 2p r pr n =0 ( K 1 p − K 2 q ) πα 2 p r −1 ( K 1 p − K 2 q ) πn − cos[( θ p − θ s ) − ( K 1 p − K 2 q ) θ rm ] . sin } 2p r pr n =0 − − − . cos . cos 2p r pr n =0 ( K 1 p − K 2 q ) πα 2 p r −1 ( K 1p − K 2 q )πn + sin[( θ p − θ s ) − ( K 1 p − K 2 q ) θ rm ] . cos .L mqq = 3µ 0 r ! N p N s 4g1 ( K 1p + K 2q ) {− sin[( θ p + θ s ) − ( K 1 p + K 2 q ) θ rm − ( K 1 p + K 2 q ) πα ] 2p r 2 p r −1 n=0 .
. sin + sin[(θ p + θ s ) − ( K 1 p + K 2 q ) θ rm 2p r pr n =0 ( K 1 p + K 2 q ) πα 2 p r −1 ( K p + K 2 q ) πn ] . .+ 3µ 0 r!N p N s 4g 2 ( K 1 p + K 2 q ) {− sin[(θ p + θ s ) − ( K 1 p + K 2 q )θ rm + ( K 1 p + K 2 q ) πα 2 pr −1 ( K p + K 2 q ) π( n + 1) ] . . cos ( K 1 p − K 2 q ) π( n + 1) ( K p − K 2 q ) πα + cos[(θ p − θ s ) − ( K 1 p − K 2 q )θ rm + 1 ] pr 2p r ( K 1 p − K 2 q ) π( n + 1) ( K p − K 2 q ) πα + sin[(θ p − θ s ) − ( K 1 p − K 2 q )θ rm − 1 ] pr 2p r ( K 1 p − K 2 q ) πn ( K p − K 2 q ) πα − cos[(θ p − θ s ) − ( K 1 p − K 2 q )θ rm − 1 pr 2p r ( K 1 p − K 2 q ) πn } pr sin cos sin  . cos 1 + cos[(θ p + θ s ) − ( K 1 p + K 2 q )θ rm 2p r pr n =0 ( K 1 p + K 2 q ) πα 2 pr −1 ( K 1 p + K 2 q ) π( n + 1) ] . cos 1 − cos[(θ p + θ s ) − ( K 1 p + K 2 q )θ rm 2p r pr n =0 3µ 0 r!N p N s ( K 1 p + K 2 q ) πα 2 pr −1 ( K 1 p + K 2 q ) πn sin }+ . 2p r pr 4 g 2 ( K 1 p − K 2q ) n =0 ( K 1 p − K 2 q ) πα ] 2p r + − − {− sin[(θ p − θ s ) − ( K 1 p − K 2 q )θ rm + 2 p r −1 n =0 2 p r −1 n =0 2 p r −1 n =0 2 p r −1 n =0 .
. . cos 2p r pr n =0 ( K 1 p − K 2 q ) πα 2 p r −1 ( K 1p − K 2 q )πn ] . . sin } 2p r pr n =0 − − − . cos ( K 1p + K 2 q )πn ( K 1 p + K 2 q ) πα + cos[( θ p + θ s ) − ( K 1 p + K 2 q ) θ rm − ] pr 2p r ( K 1 p + K 2 q ) πα ( K 1p + K 2 q )πn + sin[( θ p + θ s ) − ( K 1 p + K 2 q ) θ rm + ] 2p r pr ( K 1p + K 2 q )πn ( K 1 p + K 2 q ) πα − cos[( θ p + θ s ) − ( K 1 p + K 2 q ) θ rm + ] pr 2p r 3µ 0 r ! N p N s ( K 1p + K 2 q )πn }+ {− sin[( θ p − θ s ) − ( K 1 p − K 2 q ) θ rm pr 4g1 ( K 1p − K 2q ) sin cos sin − ( K 1 p − K 2 q ) πα 2 p r −1 ( K 1 p − K 2 q ) πn + cos[( θ p − θ s ) − ( K 1 p − K 2 q ) θ rm ] . cos 2p r pr n =0 ( K 1 p − K 2 q ) πα 2 p r −1 ( K 1p − K 2 q )πn + sin[( θ p − θ s ) − ( K 1 p − K 2 q ) θ rm ] . sin 2p r pr n =0 ( K 1 p − K 2 q ) πα 2 p r −1 ( K 1 p − K 2 q ) πn − cos[( θ p − θ s ) − ( K 1 p − K 2 q ) θ rm ] .L mdd = − 3µ 0 r ! N p N s 4g1 ( K 1p + K 2q ) {− sin[( θ p + θ s ) − ( K 1 p + K 2 q ) θ rm − ( K 1 p + K 2 q ) πα ] 2p r 2 p r −1 n=0 2 p r −1 n=0 2 p r −1 n=0 2 p r −1 n=0 .
cos . sin .− 3µ 0 r!N p N s 4g 2 ( K 1 p + K 2 q ) {− sin[(θ p + θ s ) − ( K 1 p + K 2 q )θ rm + ( K 1 p + K 2 q ) πα 2 pr −1 ( K p + K 2 q )π( n + 1) + cos[(θ p + θ s ) − ( K 1 p + K 2 q )θ rm ] . cos . sin 2p r pr 4 g 2 ( K 1 p − K 2q ) n =0 ( K 1 p − K 2 q ) πα ] 2p r + − − {− sin[(θ p − θ s ) − ( K 1 p − K 2 q )θ rm + 2 p r −1 n =0 . cos 1 2p r pr n =0 3µ 0 r!N p N s ( K 1 p + K 2 q ) πα 2 pr −1 ( K 1 p + K 2 q ) πn }+ . sin 2p r pr n =0 ( K 1 p + K 2 q ) πα 2 p r −1 ( K p + K 2 q ) πn − cos[(θ p + θ s ) − ( K 1 p + K 2 q )θ rm ] . cos 1 2p r pr n =0 ( K 1 p + K 2 q ) πα 2 pr −1 ( K 1 p + K 2 q ) π( n + 1) + sin[(θ p + θ s ) − ( K 1 p + K 2 q ) θ rm ] . sin ( K 1 p − K 2 q ) π( n + 1) ( K p − K 2 q ) πα + cos[(θ p − θ s ) − ( K 1 p − K 2 q )θ rm + 1 ] pr 2p r ( K 1 p − K 2 q ) π( n + 1) ( K p − K 2 q ) πα + sin[(θ p − θ s ) − ( K 1 p − K 2 q )θ rm − 1 ] pr 2p r ( K 1 p − K 2 q ) πn ( K p − K 2 q ) πα − cos[(θ p − θ s ) − ( K 1 p − K 2 q )θ rm − 1 pr 2p r ( K 1 p − K 2 q ) πn } pr 2 p r −1 n =0 2 p r −1 n =0 2 p r −1 n =0  .
Appendix 3B
3µ 0 r ! N 2 K p πα d p ( L ddp ) = − sin 1 {− sin 2 ( K 1 p θ rm − θ p ) d θ rm 2g1 2p r
2 p r −1 3µ 0 r ! N 2 2 K 1 pπn 2 K 1 pπn p − cos 2 ( K 1 p θ rm − θ p ) . sin }− 4 g pr p n =0 r 2
L'ddp =
2 p r −1 n =0
.
cos
{cos 2 ( K 1 p θ rm − θ p −
2 p r −1 n =0
K 1 p πα 2 p r −1 2 K 1 p π ( n + 1) K p πα − sin 2 ( K 1 p θ rm − θ p − 1 ) . cos ) 2p r pr 2p r n =0
. sin
2 K 1 p π ( n + 1) K p πα 2 p r −1 2 K 1 pπn − cos 2 ( K 1 p θ rm − θ p + 1 ) . cos pr 2p r pr n =0 K 1 p πα 2 p r −1 2 K 1 pπn ) . sin }. 2p r pr n =0
+ sin 2 ( K 1 p θ rm − θ p +

L'qqp =
3µ 0 r ! N 2 K p πα d p ( L dp ) = sin 1 { − sin 2 ( K 1 p θ rm − θ p ) 2g1 2pr d θ rm
2 p r −1 3µ 0 r ! N 2 2 K 1pπ n 2 K 1pπ n p }+ − cos 2 ( K 1 p θ rm − θ p ) . sin pr p 4 g n=0 2 r
2 p r −1 n=0
.
cos
{cos 2 ( K 1 p θ rm − θ p −
2 p r −1 n=0
K 1 p πα 2 p r − 1 2 K 1 p π ( n + 1) K p πα ) . cos ) − sin 2 ( K 1 p θ rm − θ p − 1 2p r p 2p r n=0 r
.
sin
2 K 1 p π ( n + 1) K p πα 2 p r − 1 2 K 1pπn ) . cos − cos 2 ( K 1 p θ rm − θ p + 1 pr 2p r pr n=0 K 1 p πα 2 p r − 1 2 K 1pπn ) . sin }. 2p r pr n=0
+ sin 2 ( K 1 p θ rm − θ p +

L'dds =
2 3µ r ! N s K q πα d ( L dds ) = − 0 sin 2 {− sin 2 ( K 2 q θ rm − θ s ) d θ rm 2g1 2p r 2 2 p r −1 3µ r ! N s 2 K 2 qπn 2 K 2 qπn − cos 2 ( K 2 q θ rm − θ q ) . sin }− 0 pr pr 4g 2 n =0
2 p r −1 n=0
. cos
{cos 2 ( K 2 q θ rm − θ s −
2 p r −1 n=0
K 2 q πα 2 p r −1 2 K 2 q π ( n + 1) K q πα − sin 2 ( K 2 q θ rm − θ s − 2 ) . cos ) 2p r p 2p r n=0 r
. sin
2 K 2 q π ( n + 1) K q πα 2 p r −1 2 K 2 qπn − cos 2 ( K 2 q θ rm − θ s + 2 ) . cos pr 2p r pr n=0 K 2 q πα 2 p r −1 2 K 2 qπn ) . sin }. 2p r pr n =0
+ sin 2 ( K 2 q θ rm − θ s +

L'qqs =
2 3µ 0 r ! N s K q πα d ( L qqs ) = sin 2 {− sin 2 ( K 2 q θ rm − θ s ) d θ rm 2g1 2p r 2 2 p r −1 3µ r ! N s 2 K 2 qπn 2 K 2 qπn − cos 2 ( K 2 q θ rm − θ q ) . sin }+ 0 pr pr 4g 2 n =0
2 p r −1 n=0
. cos
{cos 2 ( K 2 q θ rm − θ s −
2 p r −1 n=0
K 2 q πα 2 p r −1 2 K 2 q π ( n + 1) K q πα − sin 2 ( K 2 q θ rm − θ s − 2 ) . cos ) 2p r p 2p r n=0 r
. sin
2 K 2 q π ( n + 1) K q πα 2 p r −1 2 K 2 qπn − cos 2 ( K 2 q θ rm − θ s + 2 ) . cos pr 2p r pr n=0 K 2 q πα 2 p r −1 2 K 2 qπn ) . sin }. 2p r pr n =0
+ sin 2 ( K 2 q θ rm − θ s +

L'dqp =
3µ 0 r ! N p K p πα d sin 1 {− cos 2 ( K 1 p θ rm − θ p ) ( L dqp ) = − 2g1 2p r d θ rm
2 2 p r −1 3µ 0 r ! N p πα 2 K 1pπn 2 K 1 pπn − sin 2 ( K 1 p θ rm − θ p ) . sin }− 4g 2 pr pr n =0 2
2 p r −1 n=0
. cos
{− sin 2 ( K 1 p θ rm − θ p −
2 p r −1 n=0
K 1 p πα 2 p r −1 2 K 1 p π ( n + 1) K p πα + cos 2 ( K 1 p θ rm − θ p − 1 ) . cos ) 2p r p 2p r n=0 r
. sin
2 K 1 p π ( n + 1) K p πα 2 p r −1 2 K 1 pπn + sin 2 ( K 1 p θ rm − θ p + 1 ) . cos pr 2p r pr n=0 K 1 p πα 2 p r −1 2 K 1 pπn ) . sin }. 2p r pr n =0
− cos 2 ( K 1 p θ rm − θ p +

L'mqq =
2 p r −1 n=0
3µ 0 r ! N p N s 4g1
{cos[( θ p + θ s ) − ( K 1 p + K 2 q ) θ rm −
( K 1 p + K 2 q ) πα ] 2p r
. cos . sin . cos . sin
( K 1p + K 2 q )πn ( K 1 p + K 2 q ) πα + sin[( θ p + θ s ) − ( K 1 p + K 2 q ) θ rm − ] pr 2p r ( K 1p + K 2 q )πn ( K 1 p + K 2 q ) πα − cos[( θ p + θ s ) − ( K 1 p + K 2 q ) θ rm + ] pr 2p r ( K 1p + K 2 q )πn ( K 1 p + K 2 q ) πα − sin[( θ p + θ s ) − ( K 1 p + K 2 q ) θ rm + ] pr 2p r 3µ 0 r ! N p N s ( K 1p + K 2 q )πn {cos[( θ p − θ s ) − ( K 1 p − K 2 q ) θ rm }+ 4g1 pr
2 p r −1 n=0
2 p r −1 n=0
2 p r −1 n=0
−
( K 1 p − K 2 q ) πα 2 p r −1 ( K 1 p − K 2 q ) πn + sin[( θ p − θ s ) − ( K 1 p − K 2 q ) θ rm ] . cos 2p r pr n =0 ( K 1 p − K 2 q ) πα 2 p r −1 ( K 1p − K 2 q )πn − cos[( θ p − θ s ) − ( K 1 p − K 2 q ) θ rm ] . sin 2p r pr n =0 ( K 1 p − K 2 q ) πα 2 p r −1 ( K 1 p − K 2 q ) πn − sin[( θ p − θ s ) − ( K 1 p − K 2 q ) θ rm ] . cos 2p r pr n =0 ( K 1 p − K 2 q ) πα 2 p r −1 ( K 1p − K 2 q )πn ] . sin } 2p r pr n =0
−
−
−
sin .+ 3µ 0 r ! N p N s 4g 2 {cos[( θ p + θ s ) − ( K 1 p + K 2 q ) θ rm + ( K 1 p + K 2 q ) πα ] 2p r 2 p r −1 n=0 . sin 2p r pr n =0 ( K 1 p − K 2 q ) πα 2 p r −1 ( K 1p − K 2 q )πn − sin[( θ p − θ s ) − ( K 1 p − K 2 q ) θ rm ] . cos . cos . cos 2p r pr n =0 ( K 1 p − K 2 q ) πα 2 p r −1 ( K 1p − K 2 q )πn ] . sin } 2p r pr n =0 + − −  . sin ( K 1 p + K 2 q ) π ( n + 1) ( K 1 p + K 2 q ) πα − sin[( θ p + θ s ) − ( K 1 p + K 2 q ) θ rm + ] pr 2p r ( K 1 p + K 2 q ) π ( n + 1) ( K 1 p + K 2 q ) πα − cos[( θ p + θ s ) − ( K 1 p + K 2 q ) θ rm − ] pr 2p r ( K 1p + K 2 q )πn ( K 1 p + K 2 q ) πα − sin[( θ p + θ s ) − ( K 1 p + K 2 q ) θ rm − ] pr 2p r 3µ 0 r ! N p N s ( K 1p + K 2 q )πn {cos[( θ p − θ s ) − ( K 1 p − K 2 q ) θ rm }+ 4g 2 pr 2 p r −1 n=0 2 p r −1 n=0 2 p r −1 n=0 + ( K 1 p − K 2 q ) πα 2 p r −1 ( K 1 p − K 2 q ) π ( n + 1) + sin[( θ p − θ s ) − ( K 1 p − K 2 q ) θ rm ] . cos 2p r pr n =0 ( K 1 p − K 2 q ) πα 2 p r −1 ( K 1 p − K 2 q ) π ( n + 1) − cos[( θ p − θ s ) − ( K 1 p − K 2 q ) θ rm ] .
cos . sin . cos 2p r pr n =0 ( K 1p − K 2 q )πn ( K 1 p − K 2 q ) πα 2 p r −1 ] . cos 2p r pr n =0 ( K 1 p − K 2 q ) πα 2 p r −1 ( K 1p − K 2 q )πn − cos[( θ p − θ s ) − ( K 1 p − K 2 q ) θ rm ] . cos . sin } 2p r pr n =0 − − − .L'mdd = 2 p r −1 n =0 3µ 0 r ! N p N s 4g1 {− cos[( θ p + θ s ) − ( K 1 p + K 2 q ) θ rm − ( K 1 p + K 2 q ) πα ] 2p r . sin 2p r pr n =0 ( K 1 p − K 2 q ) πα 2 p r −1 ( K 1p − K 2 q )πn − sin[( θ p − θ s ) − ( K 1 p − K 2 q ) θ rm ] . sin ( K 1p + K 2 q )πn ( K 1 p + K 2 q ) πα − sin[( θ p + θ s ) − ( K 1 p + K 2 q ) θ rm − ] pr 2p r ( K 1p + K 2 q )πn ( K 1 p + K 2 q ) πα + cos[( θ p + θ s ) − ( K 1 p + K 2 q ) θ rm + ] pr 2p r ( K 1p + K 2 q )πn ( K 1 p + K 2 q ) πα + sin[( θ p + θ s ) − ( K 1 p + K 2 q ) θ rm + ] pr 2p r 3µ 0 r ! N p N s ( K 1p + K 2 q )πn {cos[( θ p − θ s ) − ( K 1 p − K 2 q ) θ rm }+ 4g1 pr 2 p r −1 n =0 2 p r −1 n =0 2 p r −1 n =0 − ( K 1 p − K 2 q ) πα 2 p r −1 ( K 1p − K 2 q )πn + sin[( θ p − θ s ) − ( K 1 p − K 2 q ) θ rm ] .
sin } 2p r pr n =0 + − −  . cos . sin ( K 1 p + K 2 q ) π ( n + 1) ( K 1 p + K 2 q ) πα − sin[( θ p + θ s ) − ( K 1 p + K 2 q ) θ rm + ] pr 2p r ( K 1 p + K 2 q ) π ( n + 1) ( K 1 p + K 2 q ) πα − cos[( θ p + θ s ) − ( K 1 p + K 2 q ) θ rm − ] pr 2p r ( K 1p + K 2 q )πn ( K 1 p + K 2 q ) πα − sin[( θ p + θ s ) − ( K 1 p + K 2 q ) θ rm − ] pr 2p r 3µ 0 r ! N p N s ( K 1p + K 2 q )πn {cos[( θ p − θ s ) − ( K 1 p − K 2 q ) θ rm }+ 4g 2 pr 2 p r −1 n =0 2 p r −1 n =0 2 p r −1 n =0 + ( K 1 p − K 2 q ) πα 2 p r −1 ( K 1 p − K 2 q ) π ( n + 1) + sin[( θ p − θ s ) − ( K 1 p − K 2 q ) θ rm ] . sin 2p r pr n =0 ( K 1 p − K 2 q ) πα 2 p r −1 ( K 1p − K 2 q )πn − sin[( θ p − θ s ) − ( K 1 p − K 2 q ) θ rm ] . cos .+ 3µ 0 r ! N p N s 4g 2 {− cos[( θ p + θ s ) − ( K 1 p + K 2 q ) θ rm + ( K 1 p + K 2 q ) πα ] 2p r 2 p r −1 n =0 . cos 2p r pr n =0 ( K 1p − K 2 q )πn ( K 1 p − K 2 q ) πα 2 p r −1 ] . cos 2p r pr n =0 ( K 1 p − K 2 q ) πα 2 p r −1 ( K 1 p − K 2 q ) π ( n + 1) − cos[( θ p − θ s ) − ( K 1 p − K 2 q ) θ rm ] . sin .
Appendix 4A ∆ R = Bs1 (−ω pω s Lo C q Bss 2 − Ro Bs 2 ) + Ts1 ( Ro (Ts 2 + ω s2 C q ) + ω pω s Lo C q Tss 2 ) 2 + (1 − ω 2 p C p Lo )( − B s 2 B ss1 + ω q C q Tss1 + Ts 2Tss1 − ω p ω s C q Tss 2 ) 2 + ω p C p Ro ( −ω s C q Bss 2 Bss1 + ω q C 2ω p + Ts 2ω p + ω q C q Tss1Tss 2 ) ∆ I = B1 (−ω 2Cq Bss 2 Ro + ω p Lo Bs 2 ) + Ts1 ( Roω s C2Tss 2 − ω p Lo (Ts 2 + ω s2Cq )) 2 + (1 − ω 2 p C p Lo )( −ω s Cq Bss 2 Bss1 + ω s Cqω p + Ts 2ω p + ω s C2Tss 2Tss1 ) − ω p C p Ro (− Bs 2 Bss1 + ω s2CqTss1 + Ts 2Tss1 − ω sω p CqTss 2 aa = −C pCq Lo (Tss1 + Tss 2 ) ba = C p Lo (2Tss1 + Tss 2 )Cqω r ca = LoCq ( Bs1Bss 2 − Ts1Tss 2 ) + Cq (Tss1 + Tss 2 ) − C p Lo ((Tss1Ts 2 − Bss1Bs 2 ) + Cqω r2Tss1 da = LoCqω r (Ts1Tss 2 − Bs1Bss 2 ) − (2Tss1 + Tss 2 )C2ω r ea = (Tss1Ts 2 − Bss1Bs 2 ) + Cqω r2Tss1 fa = C p Cq ga = −2C p Cqω r ha =C p Cq ( Bss 2 Bss1 − Tss1Tss 2 ) + (Ts1Cq + Ts 2C p ) + C p Cqω r2 ia = ω r C pCq ( (Tss1Tss 2 − Bss1 Bss 2 ) − 2Ts1Cqω r ja = (Ts1Ts 2 − Bs1Bs 2 ) + Ts1Cqω r2 la = −C p Cq (Tss1 + Tss 2 ) ma = 2C p CqTss1ω r + C p CqTss 2ω r .
na = Cq ( Bs1 Bss 2 − Ts1Tss 2 ) + C p ( Bss1Bs 2 − Tss1Ts 2 ) − C p CqTss1ω r2 qa = Cqω r (Ts1Tss 2 − Bs1 Bss 2 ) sa = −C p Cq Lo ta = 2C p Cq Loω r ua = − LoTs1Cq + Cq − C p Cq Lo ( Bss 2 Bss1 − Tss 2Tss1 ) − LoTs 2C p − C p Cq Loω r2 va = 2 LoTs1Cqω r − 2Cqω r + C p Cq Loω r ( Bss 2 Bss1 − Tss 2Tss1 ) xa = Lo ( Bs1Bs 2 − Ts1Ts 2 ) − LoTs1Cqω r2 + Cq ( Bss 2 Bss1 − Tss 2Tss1 ) + Cqω r2 + Ts 2 ya = −Cqω r ( Bss 2 Bss1 − Tss 2Tss1 ) C1 = − f a sa C 2 = −( f a t a + g a sa ) C3 = aa la − f a ua − g a ta − ha sa C4 = aa ma + ba la − f a va − g au a − hata − ia sa C5 = aa na + ba ma + ca la − f a xa − g a va − hau a − ia ta − ja sa C6 = aa oa + ba na + ca ma + d a la − f a ya − g a xa − ha va − ia ua − ja ta C7 = ba oa + ca na + d a ma + ea la − g a ya − ha xa − ia va − ja ua C8 = Ca Oa + d a na + ea ma − ha ya − ia xa − ja va C9 = d a oa + ea na − ia ya − ja xa C10 = ea oa − ja ya .
%ktt=0.c1=168. %ktt=0. ktt=0.e6.8933.rr1=1. end.00015. for num=1:1:3 if num==1.rr2=2.Appendix 4B Matlab Program %Doublefed generator steadystate analysis. if num==3.wr=377*2*55/60.lo=0. end. for x=1:1:41. c2=90.27.c1=168.c1=168.wr=377*2*45/60. lamx=(x1)*ktt/40.244. ktt=0. %Power winding denoted by “1” and control winding denoted by “2”. .185. ktt=0.2.e6. %ktt=0.4392. end.e6.207.e6.wr=377*2*50/60.215. if num==2.
f=1/((1/lm+1/lls1+1/lls2)*lls2).bss2=r2*lm/(ls1*ls2lm^2). r1=rr1+rm1.36*lamm^2+0. e=1/((1/lm+1/lls1+1/lls2)*lls1). ts1=ls2/(ls1*ls2lm^2).44*lamm^2+0.085*lamm+0.bss1=r1*lm/(ls1*ls2lm^2). ba=c1*lo*(2*tss1+tss2)*c2*wr. aa=c1*c2*lo*(tss1+tss2). ls2=0. rm1=4. ca=lo*c2*(bs1*bss2ts1*tss2)+c2*(tss1+tss2)c1*lo*((tss1*ts2bss1*bs2)+c2*wr^2*tss1).217.bs2=lm/(ls1*ls2lm^2). .09*lamm+0. ls1=0.55*lamm^2+0.bs1=lm/(ls1*ls2lm^2).x)=lamx.09*lamm+0. r2=rr2+rm2. rm2=7.47*lamm+1.0385.58. da=lo*c2*wr*(ts1*tss2bs1*bss2)(2*tss1+tss2)*c2*wr. tss2=r2*ls1/(ls1*ls2lm^2).24*lamm^24. lls1=ls1lm.0373.0385. lls2=ls2lm. ts2=ls1/(ls1*ls2lm^2).05*lamm^21.25*lamm+1. lm=0. lam(num. tss1=r1*ls2/(ls1*ls2lm^2).lamm=lamx.
xa=lo*(bs1*bs2ts1*ts2)lo*ts1*c2*wr^2+c2*(bss2*bss1tss2*tss1)+c2*wr^2+ts2. ha=c1*c2*(bss2*bss1tss2*tss1)+ts1*c2+ts2*c1+c1*c2*wr^2. . na=c2*(bs1*bss2ts1*tss2)+c1*(bss1*bs2tss1*ts2)c1*c2*tss1*wr^2. la=c1*c2*(tss1+tss2). ma=2*c1*c2*tss1*wr+c1*c2*tss2*wr. ua=lo*ts1*c2lo*ts2*c1+c2c1*c2*lo*(bss2*bss1tss2*tss1)c1*c2*lo*wr^2. cw1=aa*lafa*uaga*taha*sa. fa=c1*c2. bw1=fa*taga*sa. aw1=fa*sa. sa=c1*c2*lo. ya=c2*wr*(bss2*bss1tss2*tss1). va=2*lo*ts1*c2*wr2*c2*wr+c1*c2*lo*wr*(bss2*bss1tss2*tss1). dw1=aa*ma+ba*lafa*vaga*uaha*taia*sa. ia=wr*c1*c2*(tss1*tss2bss1*bss2)2*ts1*c2*wr. oa=c2*wr*(ts1*tss2bs1*bss2).ea=(tss1*ts2bss1*bs2)+c2*wr^2*tss1. ga=2*c1*c2*wr. ta=2*c1*c2*lo*wr. ja=(ts1*ts2bs1*bs2)+ts1*c2*wr^2.
gw1=(ba*oa+ca*na+da*ma+ea*laga*yaha*xaia*vaja*ua). if real(w1(mm))>1.x)^2+na*ww1(num.test1=1.x)=(sa*ww1(num.x)^2+xa*ww1 (num.x)+oa). hw1=(ca*oa+da*na+ea*maha*yaia*xaja*va). end.x)*j).x)^3+ma*ww1(num. w1=roots(cw). iw1=(da*oa+ea*naia*yaja*xa). kw=test1*test2*test3*mm. ro(num.end.test2=1. ww2(num.end. if kw>0.end.x)ww1(num.x)+ya)/(la*ww1(num.else test1=0.x)*c1*ro(num.x)^5+ta*ww1(num.ew1=(aa*na+ba*ma+ca*lafa*xaga*vaha*uaia*taja*sa). fw1=(aa*oa+ba*na+ca*ma+da*lafa*yaga*xaha*vaia*uaja*ta).x)*lo*j)/(1ww1(num.test3=1. jw1=(ea*oaja*ya). . ww1(num.km=kw.x)^3+va*ww1(num.x)=w1(km). cw=[aw1 bw1 cw1 dw1 ew1 fw1 gw1 hw1 iw1 jw1].else test3=0. ao=ts1*(ro(num. if real(w1(mm))<wr. for mm=1:1:9.else test2=0.x)=wrww1(num.x)^4+ua*ww1(num.x). end.x)^2*c1*loww1(num. if imag(w1(mm))==0.
x)^2*c1*loww1(num.0*pi).x)*c1*ro(num.x)/(2.0*pi).x)=1.k12=ww2(num.x))^2+imag(iqdo(num.x)*j). end. vs(num.x))^2).k22=bo+bss1.x)=sqrt(real(iqdo(num. iqds1(num. vp(num. k11=bss2bs2/(ww2(num. io(num.x)*j+tss1.x))^2+imag(vqds2(num. rro(num.x)ww1(num.x)/(2. lamqds2=(k11*lamx)/(e*k12f*k11).x)=ww1(num. vqds2(num.x)*c2*j).x))^2).x)ww1(num.x))^2). vqds1(num.x))^2+imag(vqds1(num. iqds2(num. k21=ao+ww1(num.x)^2*ro(num.x)/(ro(num. lamqds1=(k12*lamx)/(e*k12f*k11). f2(num.x)=vqds1(num.x)*lo*j)/(1ww1(num.x)=ao*lamqds1+bo*lamqds2.x)/2.x)=(ts2*lamqds2+bs2*lamqds1)/(ww2(num.bo=bs1*(ro(num.x))^2).x)=sqrt(real(iqds1(num.x)=ww2(num.x)=ts1*lamqds1bs1*lamqds2. is(num.x)=ts2*lamqds2bs2*lamqds1.x).x))^2).x)*lo*j). lamqdss1=(k22*lamx)/(e*k22f*k21).0.x)*c2*j).x)=sqrt(real(vqds1(num. ip(num. lamqdss2=(k21*lamx)/(e*k22f*k21).x)=sqrt(real(iqds2(num.x)*c2*j).x)=3*io(num.0/ro(num. po(num. f1(num.x)*j+tss2ts2/(ww2(num.x))^2+imag(iqds1(num. iqdo(num.x)=sqrt(real(vqds2(num. .x))^2+imag(iqds2(num.
LSD1=0.0.e4.VDIC2=2..4392..WRR=377.0.MQSIC2=5.....7233.R22=0.E4 INITIAL RQ1=RD1 RQ2=RD2 LSQ1=0...THIC=0..C2=360.8 CINTERVAL CINT=1..8.E6.e5.094395.0..NSTOP=0.e5. Appendix 4C ACSL PROGRAM DoubleFed Generator Starting Transient Dynamic Simulation .LSD2=0.NRIC=0.VQIC1=5..0 CONSTANT PI=3..R11=1...0373.VDIC1=5.0 CONSTANT VQIC2=5.WEIC=0..0373 .E6. RO2=16.LO=0.e5...e2.6568 CONSTANT RD2=2. end...e5.e5...IDIC=1.e5 CONSTANT MDSIC2=3....0164 CONSTANT MQSIC1=5.RO=16 CONSTANT C1=168. RO1=16..RD1=2.MDSIC1=5.00015..3022..CONSTANT PHASE=2.end.0385.0 CONSTANT TSTOP=0.LMD=0.E4 MAXTERVAL MAXT=5.1412.0.e5 CONSTANT IQIC=1.
0385 LMQ=0.MQSIC1) LAMDS1=INTEG((VDS1+TSSD1*LAMDS1+W1*LAMQS1+BSSD1*LAMDS 2).RO1.0/C1 GUNC2=1.0/(1.0164 LLSD1=(LSD1LMD) LLSD2=(LSD2LMD) LLSQ1=(LSQ1LMQ) LLSQ2=(LSQ2LMQ) LMMQ=1.0*55/60 THETAR=INTEG((W1WR).LSQ2=0.THIC) RO=RSW((T.NRIC) THETER=INTEG((W1).0).0/LMQ+1.MDSIC1) LAMQS2=INTEG((VQS2+TSSQ2*LAMQS2(W1WR)*LAMDS2+BSSQ2*LAMQS1).LT.0/C2 END DYNAMIC DERIVATIVE WR=WRR*2.1.0/LLSQ2) GUNC1=1.RO2) LAMQS1=INTEG((VQS1+TSSQ1*LAMQSW1*LAMDS1+BSSQ1*LAMQS2).0/LLSQ1+1.MQSIC2) .
VQIC1) VD=GUNC1*TSD1*LAMDS1+BSD1*GUNC1*LAMDS2+W1*VQS1GUNC1*IDO VDS1=INTEG(VD.VQIC2) VDS2=INTEG((GUNC2*TSD2*LAMDS2+BSD2*GUNC2*LAMDS1+(W1WR)*VQS2).VDIC2) IQO=INTEG((VQS1/LORO*IQO/LOW1*IDO).0/3.VDIC1) VQS2=INTEG((GUNC2*TSQ2*LAMQS2+BSQ2*GUNC2*LAMQS1(W1WR)*VDS2).0/3.IQIC) IDO=INTEG((VDS1/LORO*IDO/LO+W1*IQO).MDSIC2) VQ=GUNC1*TSQ1*LAMQS1+BSQ1*GUNC1*LAMQS2W1*VDS1GUNC1*IQO VQS1=INTEG(VQ.IDIC) IQS1=(TSQ1*LAMQS1+BSQ1*LAMQS2) IDS1=(TSD1*LAMDS1+BSD1*LAMDS2) IQS2=(TSQ2*LAMQS2+BSQ2*LAMQS1) IDS2=(TSD2*LAMDS2+BSD2*LAMDS1) VAS1=(VQS1*COS(THETER)+VDS1*SIN(THETER))*SQRT(2.LAMDS2=INTEG((VDS2+TSSD2*LAMDS2+(W1WR)*LAMQS2+BSSD2*LAMDS1).0) VBS1=(VQS1*COS(THETERPHASE)+VDS1*SIN(THETERPHASE))*SQRT(2.0) .0) VCS1=(VQS1*COS(THETER+PHASE)+VDS1*SIN(THETER+PHASE))*SQR T(2.0/3.
VAS2=(VQS2*COS(THETAR)+VDS2*SIN(THETAR))*SQRT(2.0)*2.0 VBS2=(VQS2*COS(THETAR+PHASE)+VDS2*SIN(THETAR+PHASE))*SQR T(2.0) WA=VDS1*LLSD2+VDS2*LLSD1LLSD1*WR*LAMQS2 WB=LAMDS1*(TSSD1*LLSD2+BSSD2*LLSD1LLSD2**2*BSSD1/LLSD1LLSD2*TSSD2) .0 IBS2=(IQS2*COS(THETAR+PHASE)+IDS2*SIN(THETAR+PHASE))*SQRT( 2.0/3.0/3.0) ICS2=(IQS2*COS(THETARPHASE)+IDS2*SIN(THETARPHASE))*SQRT(2.0/3.0/3.0/3.0) IBS1=(IQS1*COS(THETERPHASE)+IDS1*SIN(THETERPHASE))*SQRT(2.0/3.0/3.0) VS1=SQRT(VQS1**2+VDS1**2) VS2=SQRT(VQS2**2+VDS2**2) IAS1=(IQS1*COS(THETER)+IDS1*SIN(THETER))*SQRT(2.0) VCS2=(VQS2*COS(THETARPHASE)+VDS2*SIN(THETARPHASE))*SQRT(2.0/3.0) Ipa=ias1 Vpa=vas1 IAS2=(IQS2*COS(THETAR)+IDS2*SIN(THETAR))*SQRT(2.0)/2.0) ICS1=(IQS1*COS(THETER+PHASE)+IDS1*SIN(THETER+PHASE))*SQRT(2 .0/3.
0) BSD2=LMD/(LSD1*LSD2LMD**2.0*pi) TSSQ1=RQ1*LSQ2/(LSQ1*LSQ2LMQ**2.0) BSQ2=LMQ/(LSQ1*LSQ2LMQ**2.0*pi) f2=w2/(2.0) TSSQ2=RQ2*LSQ1/(LSQ1*LSQ2LMQ**2.0) TSSD1=RD1*LSD2/(LSD1*LSD2LMD**2.0) BSSD2=RD2*LMD/(LSD1*LSD2LMD**2.0) TSD1=LSD2/(LSD1*LSD2LMD**2.0) BSSD1=RD1*LMD/(LSD1*LSD2LMD**2.0) TSQ2=LSQ1/(LSQ1*LSQ2LMQ**2.W1=(WA+WB)/(LLSD2*LAMQS1+LLSD1*LAMQS2) w2=wrw1 f1=w1/(2.0) BSSQ2=RQ2*LMQ/(LSQ1*LSQ2LMQ**2.0) BSSQ1=RQ2*LMQ/(LSQ1*LSQ2LMQ**2.0) BSQ1=LMQ/(LSQ1*LSQ2LMQ**2.0) TSSD2=RD2*LSD1/(LSD1*LSD2LMD**2.0) BSD1=LMD/(LSD1*LSD2LMD**2.0) TSQ1=LSQ2/(LSQ1*LSQ2LMQ**2.0) TSD2=LSD1/(LSD1*LSD2LMD**2.0) PROCEDURAL LAMQM=LMMQ*(LAMQS1/LLSQ1+LAMQS2/LLSQ2) .
0+0.47*LAMM+1...0/(1...GE..2)..2)...24*LAMM**2.LAMM=ABS(LAMQM) RME1=4.0164 LSE1=0...0906*LAMM+0.0.0.....0385 RM1=RSW((LAMM.0+0..0373 .LSE2=0..0.GE.GE.0/LMQ+1..0.......58 LME=0.05*LAMM**2.LME) RQ1=R11+RM1 RQ2=R22+RM2 LLSQ1=(LSQ1LMQ) LLSQ2=(LSQ2LMQ) LMMQ=1..44*LAMM**2.2).0.0.55*LAMM**2...2).0851*LAMM+0..36*LAMM**2..01577...0/LLSQ1+1...2).GE...7608.7408.04..LSE2) LMQ=RSW((LAMM.25*LAMM+1..GE....0+0..0.0/LLSQ2) END END TERMT(T..041..GE.RME1) RM2=RSW((LAMM....0345.217 RME2=7.01...0...0.TSTOP) END END ...0...09*LAMM+0..RME2) LSQ1=RSW((LAMM.LSE1) LSQ2=RSW((LAMM.
Vas = Rs I as + pλas = Rs I as + p[ Laa I as + Lab I bs + Lac I cs + LaA I Ap + LaB I Bp + I aC I Cp ] 3 = Rs I as + p{( Lm12 + Lls ) I as + Lm12 cos[( p1 + q )θ rm ]I As 2 + cos[( p1 + q)θ rm − 2π 2π ]I Bs + cos[( p1 + q )θ rm + ]I Cs } 3 3 = Rs I as + p{Ls I as + Lm12 [ I Ap cosθ r + I Bp cos(θ r − β ) + I cp cos(θ r + β )} = Rs I as + p{Ls I as + Lm12 [Re(e jθ r ) R( I qdp e + Re(e j (θ r + β ) ) Re( I qdp e = Rs I as + p{Ls I as + + Re[ I qdp e j (θ r +θ p ) j (θ p + β ) jθ p ) + Re(e j (θ r − β ) ) Re( I qdp e j (θ p − β ) ) )]} 1 j (θ +θ ) j (θ +θ ) Lm12 (Re[ I qdp e r p ] + Re[ I p e r p e − j 2 β ] 2 j (θ r −θ p ) * e j 2 β ] + 3 Re[ I qdp e ]} = Rs I as + p{Ls I as + 3 j (θ −θ ) * Lm12 Re[ I qdp e r p ]} 2 Replacing the Vas. Ias with the equation (5.17) and (5.Appendix 5A Control Winding Mathematical Model of doublyfed synchronous reluctance machine with dc control winding excitation.18) gives .
Ias.11).Re[Vqds ] = Rs Re[ I qds ] + p{Ls Re[ I qds ] + 3 j (θ −θ ) * Lm12 Re[ I qdp e r p ]} 2 = Re[ Rs I qds + p{Ls I qds + 3 j (θ −θ ) * Lm12 I qdp e r p }] 2 Removing the ‘Re’ operator on both sides of the equation obtains the qd complexform equation of the power winding Vqds = Rs I qds + p{Ls I qds + 3 j (θ −θ ) * Lm12 I qdp e r p }. 2 Appendix 5B Power Winding Mathematical Model of doublyfed synchronous reluctance machine with dc control winding excitation.10). Iap.17) and (5. Ibs and Ics with the equations (5. (5. VAp = R p I Ap + pλ Ap = R p I Ap + p[ LAA I Ap + LAB I Bp + LAC I Cp + LAa I as + LAb I bs + LAc I cs ] = ( R p I Ap + Lp pI Ap ) + Lm12 p{cos[( p1 + q)θ rm ]I as + cos[( p1 + q )θ rm − β ]I bs + cos[( p1 + q )θ rm + β ]I cs } = ( R p I Ap + Lp pI Ap ) + Lm12 p{Re(e jθ x ) I as + Re(e jθ x e − jβ ) I bs + Re(e jθ x e jβ ) I cs } Replacing Vap. (5.18) gives .
I qdp = R p I qdp + jω p [ L p I qdp + ’ from both sides of the 3 3 j (θ −θ ) j (θ −θ ) * * Lm12 I qds e r p ] + p[ L p I qdp + Lm12 I qds e r p ].Re[Vqdp e jθ p ] = ( R p + Lp p ) Re[ I qdp e jθ p ] + Lm12 p{Re[e jθ r ] Re[ I qds ] + Re[e j (θ r − β ) ] Re[ I qds e − jβ ] + Re[e j (θ r + β ) ] Re[ I qds e jβ ]} = ( R p + L p p ) Re[ I qdp e jθ p ]+ 1 Lm12 p{Re[ I qds e jθ r ] + Re[ I qds e jθ r e − j 2 β ] 2 j (θ r −θ p ) * + Re[ I qds e jθ r e j 2 β ] + 3 Re[ I qds e e jθ p ]} = Re[( R p + Lp p ) I qdp e * + 3 Re[ I qds e j (θ r −θ p ) jθ p jθ p ]+ 1 Lm12 p{Re[ I qds e jθ r (1 + e − j 2 β + e j 2 β )] 2 e ]} jθ p = Re[ R p I qdp e + jθ p + ( jω p Lp + Lp p) I qdp e ] 3 j (θ −θ ) jθ * Lm12 Re[( jω p + p) I qds e r p e p] 2 Removing the ‘Re’ operator from two sides of the equation obtains I qdp e jθ p = R p I qdp e jθ p + ( jω p L p + L p p ) I qdpe jθ p + 3 j (θ −θ ) jθ * Lm12 ( jω p + p) I qds e r pe p 2 jθ p The qd complexform equation is obtained by dividing ‘ e equation. 2 2 .
Te = [ I Ap I Bp I Cp ] ' LAa d % LBa dθ rm % % & LCa LAb LBb LCb LAc $ ' I as $ % " LBc " " % I bs " LCc " # & I cs " #% ' LAa %L % Ba % & LCa LAb LBb LCb 2π 2π $ ' ] cos[( p1 + q )θ rm + ] cos[( p1 + q )θ rm ] cos[( p1 + q )θ rm − % 3 3 " LAc $ " % 2π 2π " %cos[( p1 + q)θ rm − θ θ LBc " L p q p q = ] cos[( + ) + ] cos[( + ) ] m 12 1 rm 1 rm " 3 3 " % LCc " 2π 2π " # % cos[( p1 + q)θ rm + ] cos[( p1 + q)θ rm ] cos[( p1 + q)θ rm − ] % 3 3 " # & ' Re[e jθ r ] Re[e j (θ r − β ) ] Re[e j (θ r + β ) ]$ % " = Lm12 %Re[e j (θ r − β ) ] Re[e j (θ r + β ) ] Re[e jθ r ] " %Re[e j (θ r + β ) ] Re[e jθ r ] Re[e j (θ r − β ) ]" & # ' LAa d % LBa dθ rm % % LCa & LAb LBb LCb LAc $ LBc " " " LCc # ' Re[ je jθ r ] Re[ je j (θ r − β ) ] Re[ je j (θ r + β ) ]$ % " = ( p1 + q) Lm12 %Re[ je j (θ r − β ) ] Re[ je j (θ r + β ) ] Re[ je jθ r ] " %Re[ je j (θ r + β ) ] Re[ je jθ r ] Re[ je j (θ r − β ) ]" & # [I Ap I Bp I Cp = Re[ I qdp e ] [ jθ p ] Re[ I qdp e j (θ p − β ) ] Re[ I qdp e j (θ p + β ) ] ] [I Ap I Bp I Cp d ] dθ rm ' LAa %L % Ba % & LCa LAb LBb LCb LAc $ ' AA $ % " = ( + ) LBc " p q L 1 m12 % BB " " % LCc " # # &CC " Where .Appendix 5C Torque equation of doublyfed synchronous reluctance machine with dc control winding excitation.
AA = Re[ I qdp e jθ p ] Re[ je jθ r ] + Re[ I qdp e j (θ p − β ) ] Re[ je j (θ r − β ) ] + Re[ I qdp e j (θ p + β ) ] Re[ je j (θ r + β ) ] 1 j (θ +θ ) j (θ +θ ) j (θ +θ ) j (θ −θ ) * e r p ]} = {Re[ jI qdp e p r ] + Re[ jI qdp e p r e − j 2 β ] + Re[ jI qdp e p r e j 2 β ] + 3 Re[ jI qdp 2 = 3 j (θ −θ ) * Re[ jI qdp e r p ] 2 jθ p BB = Re[ I qdp e ] Re[ je j (θ r − β ) ] + Re[ I qdp e j (θ p − β ) ] Re[ je j (θ r + β ) ] + Re[ I qdp e j (θ p + β ) ] Re[ je jθ r ] 1 j (θ +θ ) j (θ +θ ) j (θ +θ ) j (θ −θ ) * e r p e − jβ ]} = {Re[ jI qdp e p r e − jβ ] + Re[ jI qdp e p r ] + Re[ jI p e p r e jβ ] + 3 Re[ jI qdp 2 = 3 j (θ −θ ) * Re[ jI qdp e r p e − jβ ] 2 jθ p CC = Re[ I qdp e ] Re[ je j (θ r + β ) ] + Re[ I qdp e j (θ p − β ) ] Re[ je jθ r ] + Re[ I qdp e j (θ p + β ) ] Re[ je j (θ r − β ) ] 1 j (θ +θ ) j (θ +θ ) j (θ +θ ) j (θ −θ ) * e r p e jβ ]} = {Re[ jI qdp e p r e jβ ] + Re[ jI qdp e p r e − jβ ] + Re[ jI p e p r ] + 3 Re[ jI qdp 2 = 3 j (θ −θ ) * Re[ jI qdp e r p e jβ ] 2 [I Ap I Bp I Cp ] ' LAa d % LBa dθ rm % % & LCa LAb LBb LCb LAc $ ' I as $ % " LBc " " % I bs " LCc " # & I cs " #% T $ ' 3 j (θ r −θ p ) * ] " % 2 Re[ jI qdp e " %3 j (θ −θ ) * = ( p1 + q) Lm12 % Re[ jI qdp e r p e − jβ ]" " %2 % 3 Re[ jI * e j (θ r −θ p )e jβ ] " qdp " % # &2 ' Re[ I qds ] $ % jβ " % Re[ I qds e ] " %Re[ I qds e − jβ ]" # & .
000 F Discharging resistance Rb1 = 0.001Ω Discharging capacitance Cb1 = 1. Lm = 2 Lm12 .000 Ohms Open circuit battery voltage = 24 V Connecting resistance Rbt = 0.0F Selfcharging Resistance Rbp = 10. 3 Appendix 6A The parameters of the lead acid battery [61] Battery capacitance Cbp = 54.= 3 j (θ −θ ) j (θ −θ ) * * ( p1 + q ) Lm12{Re[ jI qdp e r p ] Re[ I qds ] + Re[ jI qdp e r p e − jβ ] Re[ I qds e jβ ] 2 j (θ r −θ p ) * + Re[ jI qdp e e jβ ] Re[ I qds e − jβ ]} = 3 1 j (θ −θ ) j (θ −θ ) * * * ( p1 + q ) Lm12 ( ){3 Re[ jI qdp e r p I qds ] + Re[ jI qdp e r p I qds (1 + e − j 2 β + e j 2 β )] 2 2 3 3 j (θ −θ ) * ( p1 + q )( Lm12 ) Re[ jI qdp e r p I qds ] 2 2 = The final form of torque equation is 3 j (θ −θ ) * Te = ( p1 + q ) Lm Re[ jI qdp I qds e r p ] 2 Where.05Ω Internal resistance Rbs = 0.002Ω .
wp=377.end if mm==3.wp=377. for mm=1:1:3 %Ojo's machine parameters %vpm=187.ssc=rp^2*ipm^2vpm^2.Appendix 7A %Program to obtain the electrical characteristics of DF machine %for a wide speed range including above base speed and below speed range.lp=0.end %calculate the peak rate current. %rate value calculation: clear all basetime=4.ipm=9.0385.4..3477.iqs=2..pr=8.lp=0.iqs=16.7107.lm=0.rp=0. %ipm=(wp*lm*rp*iqs)/(rp^2+wp^2*lp^2)+sqrt((wp*lm*iqs*rp)^2.pr=4..iqs=8.021. %(rp^2+wp^2*lp^2)*((wp*lm*iqs)^2vpm^2))/(rp^2+wp^2*lp^2) %calculate the base speed.043.ssb=2*lm*rp*iqs*ipm.lm=0.7. if mm==1. ssa=lp^2*ipm^2+lm^2*iqs^2.rp=1. %Wu's machine parameters vpm=187. %(rp^2+wp^2*lp^2)*((wp*lm*iqs)^2vpm^2))/(rp^2+wp^2*lp^2) %iss(mm)=ipm %vpm=sqrt((rp^2+wp^2*lp^2)*ipm^2+wp^2*lm^2*iqs^22*wp*lm*rp*iqs*idp) %ipm=(wp*lm*rp*iqs)/(rp^2+wp^2*lp^2)+sqrt((wp*lm*iqs*rp)^2.end if mm==2. .041.ipm=11..
kk3=lb^2ld^2+2*la*lc. wn=roots(mwm). aw12(mm)=vpm/sqrt(lp^2*ipm^2lm^2*iqs^2) . kk4=2*lb*lcle^2. kk5=lc^2. awbb(mm)=vpm/sqrt(lp^2*ipm^2+lm^2*iqs^2) %boundary point between Mode I and Mode II la=ipm^2*lp^4lm^2*lp^2*iqs^2. le=2*vpm*lm*rp^2*iqs. awnn=sqrt(wn). wee=basetime*377. kk2=2*la*lb. ld=2*vpm*lm*rp*iqs*lp. nwee=w12(mm). if wee>=basetime*377. end. mwm=[kk1 kk2 kk3 kk4 kk5]. %approximated equation to calculate the boundary speed by choosing rp=0. kk1=la^2. w12(mm)=max(awnn) wee=w12(mm). lb=2*rp^2*lp^2*ipm^2lm^2*rp^2*iqs^2lp^2*vpm^2.wbb(mm)=ssb/(2*ssa)+sqrt(ssb^24*ssa*ssc)/(2*ssa) %approximated equation to calculate the base speed by chossing rp=0. nwee=basetime*377. lc=ipm^2*rp^4rp^2*vpm^2.
nn)=wbb(mm)*(nn1)/20. niidp(mm.nn)=sqrt((rp^2+wp^2*lp^2)*iip(mm. end for nn=1:1:21. npo(mm.nn)=sqrt(iidp(mm.. iip(mm. ntor(mm.5*4*iqs*iidp(mm.nn)).nn)*nwbe(mm.nn)^2+iiqp(mm. po(mm.nn)=sqrt(niidp(mm.nn)*lm.5*4*iqs*niidp(mm. +2*wp^2*lm*lp*iqs*niiqp(mm.nn)=0..nn)).nn)=ipm.nn)^2+niiqp(mm. iiqp(mm.nn)=tor(mm.nn)=wbb(mm)*(nn1)/20. wbe(mm.nn)/4. +2*wp^2*lm*lp*iqs*iiqp(mm.nn)).nn). niip(mm.nn)2*wp*lm*rp*iqs*niidp(mm.nn)2*wp*lm*rp*iqs*iidp(mm. nvpo(mm.nn)=1.%below the base speed for nn=1:1:21.nn)*lm. vpo(mm.nn)=1.. wp=abs(nwbe(mm.nn)=ipm.nn)^2). nwbe(mm. end . tor(mm. wp=wbe(mm.nn)*wbe(mm.nn)=ntor(mm.nn)=sqrt((rp^2+wp^2*lp^2)*niip(mm.nn)^2+wp^2*lm^2*iqs^2.nn)^2). iidp(mm.nn)=0. niiqp(mm..nn)/4.nn)^2+wp^2*lm^2*iqs^2.
niidp(mm.s3=10e4*(2*wp^2*lm*lp*iqs).nn).nn)=tor(mm. iip(mm. iiqp(mm. s1=10e4*(vpm^2(rp^2+wp^2*lp^2)*ipm^2wp^2*lm^2*iqs^2).nn))/s3.%above the base speed %Mode I for nn=22:1:72.tb=2*s1*s2. vpo(mm.nn)=sqrt((rp^2+wp^2*lp^2)*iip(mm. s2=10e4*(2*wp*lm*rp*iqs).nn)^2). wp=wbe(mm.nn)).nn)=(s1+s2*iidp(mm. +2*wp^2*lm*lp*iqs*iiqp(mm. end for nn=22:1:72.nn)). nwbe(mm.nn)*lm.5*4*iqs*iidp(mm.nn)*wbe(mm.. .nn)=tb/(2*ta)+sqrt(tb^24*ta*tc)/(2*ta). tor(mm. po(mm. wbe(mm. s2=10e4*(2*wp*lm*rp*iqs).nn)2*wp*lm*rp*iqs*iidp(mm.nn)^2+iiqp(mm. iidp(mm.nn)/4.tb=2*s1*s2. ta=s2^2+s3^2. ta=s2^2+s3^2..nn)^2+wp^2*lm^2*iqs^2.nn)=sqrt(iidp(mm.tc=s1^2s3^2*ipm^2.nn)=tb/(2*ta)+sqrt(tb^24*ta*tc)/(2*ta). wp=abs(nwbe(mm.nn)=wbb(mm)+(weewbb(mm))*(nn21)/51.nn)=wbb(mm)+(nwee+wbb(mm))*(nn21)/51.tc=s1^2s3^2*ipm^2. s1=10e4*(vpm^2(rp^2+wp^2*lp^2)*ipm^2wp^2*lm^2*iqs^2).s3=10e4*(2*wp^2*lm*lp*iqs).nn)=1.
. +2*wp^2*lm*lp*iqs*iiqp(mm.nn)=1.nn))..nn)^2)..niiqp(mm. iip(mm.nn)^2+iiqp(mm.5*4*iqs*niidp(mm.nn)=(s1+s2*iidp(mm. wbe(mm.nn)/4. end.nn)*lm.nn)2*wp*lm*rp*iqs*niidp(mm.nn)).nn). npo(mm.nn))/s3. nwbe(mm. vpo(mm. ntor(mm. iidp(mm.nn)=sqrt(niidp(mm.nn)*wbe(mm.nn)=wee+(basetime*377wee)*(nn73)/50.nn)).nn)^2+wp^2*lm^2*iqs^2..nn)^2). wp=wbe(mm.nn)/4.nn)2*wp*lm*rp*iqs*iidp(mm.nn)=(wp*lm*rp*iqs+vpm*sqrt(lp^2*wp^2+rp^2))/(rp^2+wp^2*lp^2).nn)=tor(mm. wp=abs(nwbe(mm.nn)^2+niiqp(mm.nn)=sqrt(iidp(mm.nn)=wp^2*lm*lp*iqs/(rp^2+wp^2*lp^2).nn)^2+wp^2*lm^2*iqs^2. po(mm.nn)=wp^2*lm*lp*iqs/(rp^2+wp^2*lp^2). iiqp(mm. niiqp(mm.nn)=sqrt((rp^2+wp^2*lp^2)*niip(mm.nn)*nwbe(mm. tor(mm. for nn=73:1:123.nn)=sqrt((rp^2+wp^2*lp^2)*iip(mm. niip(mm..nn)=1.5*4*iqs*iidp(mm.nn)=ntor(mm.nn)*lm.nn)=nwee+(basetime*377+nwee)*(nn73)/50. nvpo(mm. end %mode II for nn=73:1:123. +2*wp^2*lm*lp*iqs*niiqp(mm.
nvpo(mm. +2*wp^2*lm*lp*iqs*niiqp(mm.nn)=ntor(mm. niip(mm.nn)*nwbe(mm.0.niidp(mm.nn)=1.1042 Kc = 2. Control Scheme II: . npo(mm.nn)^2+wp^2*lm^2*iqs^2.nn)=sqrt(niidp(mm.nn)).nn)/4..6955 J = 0.nn)2*wp*lm*rp*iqs*niidp(mm.nn)^2+niiqp(mm.nn)=sqrt((rp^2+wp^2*lp^2)*niip(mm. end.54 Kb = 0.nn)^2).nn)*lm.0025 Kgm2 B = 0.nn)=(wp*lm*rp*iqs+vpm*sqrt(lp^2*wp^2+rp^2))/(rp^2+wp^2*lp^2). ntor(mm.5*4*iqs*niidp(mm.. end Appendix 7B Control Scheme I: Ka =35.4019 Kq =15791 Kd = 0.
m clear all %parameters global u1 c1 h1 tre k1.b=1/J.0.4.[uo co ho]') u1=ss(1) c1=ss(2) h1=ss(3) u2=u1+c1^2 ki=u1*u2/(kt*b) kp=((u1+u2)/bB)/kt .0025.tre=0.041.KI = 2.58 J = 0. uo=5.ho=1 ss=fsolve('KIP'. Appendix 7C %Program used to calculate the IP parameters for twoorder speed loop.0. %aao/(s^2+aa1*s+aa2) %main programKIP2.iqs=8.lp=0.B=0.k1=2.co=1.m and subprogramKIP. kt=(3/2)*pr*lm*iqs.lm=0.02.0025 Kgm2 B = 0. pr=4. J=0.871 Kp = 0.
co3=2.k2=2. %Subprogram Kip2. %aao/(S^3+aa2*S^2+aa1*S+aao) %main programKIP3. q(3)=0. q=zeros(3. uo1=20. u1=p(1).h1=p(3).c1=p(2).9((h1/u1)*(1exp(u1*tre))(h1/(u1+c1^2))*(1exp((u1+c1^2)*tre))). q(2)=h1^2/(2*u1)h1^2*(1k1)/(2*u1+c1^2)k1*h1^2/(2*u1+2*c1^2). k1=1.1).3.ho3=3.4. end. .m and subprogramKIPP.m function q=KIP(p) global u1 c1 h1 tre k1.ho2=0. %coefficiences parameter q(1)=h1/u1h1/(u1+c1^2)1.end.2.tre=0.m clear all %parameters global u1 c2 c3 h2 h3 tre k1 k2. Appendix 7D %program used to calculate the IP parameters for threeorder speed loop.co2=2.
lp=0. %coefficiences parameter . aa1=u1*u2+u2*u3+u3*u1. u1=p(1).ss=fsolve('KIPP'.c2=p(2).0025. q=zeros(5. z=(3*pr^2*lm*iqs)/(4*j).m function q=KIPP(p) global u1 c2 c3 h2 h3 tre k1 k2. k3=lp*aa2 kd=lp*aao/(z*k3) k2=lp*aa1/(z*k3) end.[uo1 co2 co3 ho2 ho3]') u1=ss(1) c2=ss(2) c3=ss(3) h2=ss(4) h3=ss(5) u2=u1+c2^2.iqs=10.c3=p(3).lm=0.Ojo controller pr=4.04. aa2=u1+u2+u3.02. u3=u1+c2^2+c3^2. %Dr.h2=p(4). h3=p(5).j=0.1). aao=u1*u2*u3. %Subprogram of Kip3.
... (h2+h3)*h3/((u1+c2^2+c3^2)+u1)).. q(3)=0. k1*((h2+h3)*h3/(u1+(u1+c2^2+c3^2))... .. q(5)=((h2+h3)^2/(2*u1)(h2+h3)*h2/(u1+(u1+c2^2)). +h3^2/(2*(u1+c2^2+c3^2)))). +h3*(u1+(u1+c2^2)).9((h2+h3)*(1exp(u1*tre))/u1. k2*(((h2+h3)*h2/(u1+(u1+c2^2)).. +((h2+h3)*h3/(u1+(u1+c2^2+c3^2))+h2*h3/((u1+c2^2)+(u1+c2^2+c3^2)). +h2^2/(2*(u1+c2^2))+h2*h3/((u1+c2^2)+(u1+c2^2+c3^2))). +h2*(1exp((u1+c2^2)*tre))/(u1+c2^2)...q(1)=(h2+h3)*((u1+c2^2)+(u1+c2^2+c3^2))+h2*(u1+(u1+c2^2+c3^2)). q(4)=((h2+h3)*h2/(u1+(u1+c2^2))+h2^2/(2*(u1+c2^2))... q(2)=(h2+h3)/u1+h2/(u1+c2^2)+h3/(u1+c2^2+c3^2)1... end.... +h2*h3/((u1+c2^2)+(u1+c2^2+c3^2))).... +h3*(1exp((u1+c2^2+c3^2)*tre))/(u1+c2^2+c3^2)).. +h2*h3/((u1+c2^2)+(u1+c2^2+c3^2))+h3^2/(2*(u1+c2^2+c3^2)))..
REFERENCES 220 .
Seattle. A. 117. Spee. K. 11681174. 1983. August 1973. 1907. (2). Cook and B. IEE. "Stability and stabilization of doublyfed singleframe cascade induction machines." IEEE Industry Applications Society. L. K. Power Appl. on Energy Conversion." J. Sao Paulo. pp. 118. 1970. 648677 2.14. Wallace and R. (2). C." IEEE Trans. 511537. Roy. 120. "Brushless statorcontrolled synchronousinduction machine. D. "Induction motors: old and new. 15931600.246263. C. Broadway. "Speed control of a single frame cascade induction motor with slip power pump back. pp. pp. Shibata and K. "The potential of brushless doublyfed machines for adjustable speed. C. (11). R. 11. 15291535. "Some developments in multispeed cascade induction motors. 9. Brazil. 860866.). "Research development and application studies for brushless doublyfed machines. R. Kusko and B. K. 1991. (2). Creedy. W. "A new type of induction motor. pp. 4550. (7). "Selfcascaded machine: a lowspeed motor or highfrequency brushless alternator. pp. pp.2. Rawcliffe.'' Proc. 1990. Burbridge. W. on Industry Application." Proc. Electr. 1987. 121." Proc. A. Hunt." Proc. 1979. 1921. vol. 12771290 4. Vol. Broadway and G. R. R." Seminarios International De Motors Electricos E acionamentos Regulares. Lauw. Vol. A. 126." Proc. IEE. 5. 130. Juy 1970. H. Smith. Spee and H. 59." IEEE Trans. 38. Cook and B.117. "Effects of machine parameter values on dynamic response and stability regions of doublyfed cascade induction machines. G. 10. Broadway. 6. pp. R. IEE. vol. 1971. H. pp. 7. vol. A. Wallace. (Lond. 8. 1978.). A. Soc. IEE. H. pulp and paper industries annual conference.1974.). 97105. pp. pp. IEE (Lond. pp. 12771290. "Brushless cascade alternator. 12. D. F. vol." Proc. IEE.1. F. Somuah. L. IEE (Lond. W. pp. A. "Cageless induction motor. 13. IEE.” Proc. A. 3. 137142. Taka "speed control system for brushless cascade induction motors in control range of slips S1>1 and S2>1. J." J. Thomas. Smith. 4160. pp. Broadway. W. .
K.32 (1). pp. M. on Energy Conversion.19. 453460. A. 5459. Wallace.. Lauw. M. Spee and P. 738743. Wallace. A. "Generalised theory of the brushless doublyfed machine. pp.6. Electr. "Rotor modelling and development for brushless doublyfed machines. pp. 1989.” IEEE Trans. K. K. 189194. ElAntably. Wallace and R. Lipo. A. 22. R. Li. . "Performance simulation of bruless doublyfed adjustable speed drives. Electr. R. K. (3). A. R. 23. Rochelle. Williamson and A. Spee and A." IEEE Industry Applications Society Annual Meeting.. 19. R. 16. Report ORNL/SUB/8195013/1. S. Lauw. Ferreira. S. "Generalised theory of the brushless doublyfed machine. R. 125138. 1997. "Two axis model development of cagerotor brushless doublyfed machines. "Dynamic simulation of brushless doublyfed machines. 1990. Wallace. 25. pp.. pp. C. Xu and T. S. Vol. Ferreira. pp. Li." IEE Proc. pp. 111121." IEEE Trans. R. on Energy Conversion. "Iron loss and saturation in brushless doublyfed machines. K. K. Spee. Power Appl. Liang. pp. pp." IEEE Industry Applications Society Annual Meeting. on Industry Application. F. (2). "Dynamic modeling of brushless doublyfed mahines.14. C. "Investigation of appropriate pole number combinations for brushless doublyfed machines applied to pump drives. Power Appl. J. 1991. ICEM'90." Electric Machine and Power System. 1991. Boger. K. 17. A. Vol." IEEE Trans.. "The effect of stator winding configuration on the performance of brushless doublyfed machines in adjustable speed drives. Spee. pp. Spee. R. Spee. K. "dq analysis of a variable speed doublyAC excited reluctance motor. 1997. K. "Reluctance and doublyexcited reluctance motors. Part 2: Model verification and performance. C. 329334. 445452. Boston. 97103. 15. 18. 1996. L. 1991." IEEE Industry Applications Society Annual Meeting. Wallace and H. Oak Ridge National Laboratories. Vol. Wallace." Final report. Rochelle. A. pp. 144. Williamson and A. Wallace and R. 144. C." IEE Proc. R. P. 1997. 21. 24. Heyne and A. A. 123129. Li. (2). 20.6. Ferreira and S. 1989. 1990. Vol. A. 331337. (3)." IEEE Industry Applications Society Annual Meeting. New Orlean. Williamson. Spee and H. Part 1: Analysis." International Conference on Electrical Machine.
Wallace and R." IEE Proc. Xu and Y. L. vol. July/August. 1992. F. on Power Electronics. 29.408413. 1994.30. Liao and C. Vol. Sun. pp. 1996. 30. Alexander.437444.. Elsonbaty. August 912. on Energy Conversion. 1996. S.. January. on Industry Application.." IEEE Industry Application Society Annual Meeting. Park. 28. Wu. Zhen and E." IEEE Trans. K. 31. R." IEEE Industry Applications Society Annual Meeting.10. A. "Stator field oriented control of doublyexcited induction machine in wind power generating system. pp. 1994.4. Novotny. L. Jeong. on Industry Application. 1995. and N. 1996. Li." 35th Midwest Symposium on Circuits and System". Vol. O. 36.1. Kim. and T. L. Y. A. 1993. L. robust position sensorless control scheme for doublyfed reluctance motor drives. D.6. Li. pp. "Transient Model of a doublyexcited reluctance motor. Xu.30. A. Vol. "A novel windpower generating system using field orientation controlled doublyexcited brushless reluctance machine. Sep. 35." IEEE Trans. Lipo." IEEE Trans. pp. . pp.131. P. No5. R. 12101218." IEEE Industry Application Society Annual Meeting. Xu. 32. No. "Steady state analysis of a stand alone wound rotor induction generator excited by a PWM inverter. 33. C. Vector Control and Dynamics of AC drives. R. L. Electr. No. 1992. Liang and T. pp." IEEE Industry Application Society Annual Meeting. "A novel position sensorless control scheme for doubly fed reluctance motor drives. "Field orientation control of a doubly excited brushless reluctance machine. March 1984. March 1991. pp. Liao and C. Power Appl. H. Y. pp. and M.1. 27. G. "Cyclonvertorexcited dividedwinding doublyfed machine as windpower convertor. Spee. 790797. "A low cost. Holmes. Vol. A. G. Wallace and G. Lipo. 34./Oct. 126132. No. Y. Tang. K. Sun. H. Xu. Tang.:"Determination of converter control algorithms for brushless doublyfed induction motor drive using Floquet and Lyapunov techniques. 794800. DC. Oxford University Press. A. Ojo and Z.6169. Spee. 1987. 37.319325. "Modeling of the doublyfed reluctance machines using harmonic balance technique and ManleyRowe power frequency relationship. W. "Synchronous drive performance of brushless doublyfed motors." IEEE Industry Application Society Annual Meeting. Wachington." IEEE Trans.26.
Vol. Ojo and Z. Power Appl. Power Appl. "A doubly fed induction generator using backtoback PWM converters supplying an isolated load from a variable speed wind turbine. 1987. Williamson. Boger. pp. Xu and F. HC. 1988.. S. 49.5. L. on Energy Conversion. Electr. A. R. Laithwaite. 143. J. pp. 121129. No. 132. (Lecture Notes). S. No. Williamson and E. N. 142148. 48. Wu. 24. W. Lipo.. "Impact of interbar currents on the performance of brushless doublyfed motor. (book). R. Wang. T. .1. 44. O. Issue 6. 326331. Analysis of synchronous Machines. 1965." IEEE Industry Application Society Annual Meeting. 40. "Regulated DC power generation using mixedpole synchronous reluctance machines. "Synchronous operation of a dualwinding reluctance generator. Williamson and M. pp. "Performance characteristics of dualwinding reluctance generators. 188195. pp. March 1994. B Heinz and B. pp. "Comparative study of magnetic coupling for a doubly fed brushless machine with reluctance and cage rotors. 1997. Electr. 130. New York. "Analysis and realization of a pulsewidth modulator based on voltage space vectors.9." IEE Proc." IEE Proc. 461468. 41. 380387. Y. Vol. 165171. Power Appl." IEE Proceeding. (2). O. vol. 157163. S. 1997. Ye. (3). No. Xu. Tang and L. January/February. pp. Power Appl. Skudelny.. W. New Orlean." IEE Proc." IEEE Trans. New Orlean. Ronald Press. L. Pene. Asher. "Power factor improvement in cagerotor induction motors. Ojo and Z." IEEE Trans. Vol. Clare and G. D. 1997." to be published in the IEEE Trans. Schmitz and D. 39. Novotny. Wu. O.. 1. Wu. November 1997. 46. University of WisconsinMadison.38. M.. on Energy Conversion. 1985. 45. 42. pp. 47.Electr. S. Introductory Electromechanics. 1983." IEEE Industry Applications Society Annual Meeting. 144. Electr. pp. Ojo and Z. on Industry Application. "Comparison study of rotor structures of doubly excited brushless reluctance machine by finite element analysis. C. pp. V. 43." IEEE Industry Application Annual Meeting. 263270. September 1996. L. "Generalized harmonic analysis for the steadystate performance of sinusoidallyexcited cage induction motors.
vol.7. 57. 9398.19101913. Electric Machinery and Transformers. Manley and H. E. Chapman.8997. O. Analysis and Performance Test of a Doubly Fed Reluctance Machine. 58. Rowe. On Aerospace and electronics. vol. no. A. Balachandran. M. E. September 1982. R. 56. Gury and H." Ph. "Simplified analysis of PWM converters using model of PWM switch part II: Discontinuous conduction mode." (Master thesis). FrequencyPower Formulas. S. M. 1988. (book). The Ohio State University. B. Vatche Vorperian. on Energy Conversion. P. 61. C. J. no. 52. A.46.10. University of WisconsinMadison.12. pp.659665. Ojo." IEEE Trans. Salameh. On Energy Conversion. 1990. pp. E. No. no. Russell and I. "General power relationship for positive and negative nonlinear resistive elements. "PWMswitch modeling of DCDC converters. New York.. vol. Akademiai Kiads. 54." Proceedings of the IRE. L. J. M. Spooner and A. J. D. "An analysis of the induction motor. O'Sullivan and J. Hiziroglu. A. no. 1994. John Wiley and Sons." IEEE Trans. "Design.3. Retter. "Design and Evaluation of a reluctance machine drive system. 53. vol. 63. Klassens. B. (book). 1995. pp. H. G. 1985. 1960. (book). D. December 1997. Penfield. May 1990. McGrawHill.6.2. E." IEE Proceedings. (book). Harcourt Bruce Jovanovich.44. J. vol.5. part 1: general energy relations. Nayfeh and B. part 3power/frequency relationships. March 1992. Spruijt. Electric Machinery Fundamentals. H. no. Publishers. Lynch. vol. "A mathematical model of a leadacid batteries. vol. J. pp. Matrix and SpacePhasor Theory of Electrical Machines. Williamson. Cox and Z.D. 51. A. pp.1. "DC power generation using interior permanentmagnet machines. S. Budepest." IEEE Trans.904913. 1958.243247. MIT press. L. 59. 1956. Dijk. N.26. Ye. Xu.12. pp. P. pp. Pantell. no. Inc.4." IEE Proceedings. 60. 1998. (book). Z. Wu. on Power Electronics. Applied Nonlinear Dynamics." IEEE Trans. 62. pp. November 1995.50.129. . dissertation. "Mixed pole windings and some applications. Pickup. Casacca and W. 55. Part B. "Some properties of nonlinear elements. vol.497505. 64. 1987. R.137. Van.351356." Proceedings of the IRE.
A. J. July 1993. F. Friedland. pp. O. no. Park. Pan. 74. 1. no. Zhen and E. H. 5561. vol. L. on Automatic Control. (book). no. B. K.12761282. The Field Orientation Principle in Control of Induction Motors. "Field orientation control of a doublyexcited brushless reluctance machine. Control System Design. C. M. Oct. 148155. Park and K. Li. 1991. McGrawHill. 69. L. 71. 680686. vol. on Industrial Electronics.5." IEEE Trans. pp. January 1987. "High performance current regulator for a fieldorientation controlled induction motor drive. 76. pp. T. New York. Lin. Applied Nonlinear Control. Slotine and W. Lin and Y. 5. "Field orientation control of a doublyfed synchronous reluctance machine. pp. on Inductry Applications." IEEE ProceedingsB. L. 66. T. “Experimental nonlinear torque control of a permanentmagnet synchronous motor using saliency. and B. Lee. vol. "A digital signal processor based robust integralproportional controller for an induction motor servo drive.30.32. On Industry Electronics. Kim. M. 1986. vol." Electric Power Systems Research vol. 73. no. –S. "Model reduction by power decomposion and frequency response matching.” IEE Proceedings Control Theory Application." Conference Record of the 29th Annual IEEE PESC conference. 34. Akhrif. vol. 72.12471257. Liaw.65. 145. Kluwer Academic Publishers. P. 140. S. F. Nandam and P. O. (book). Grenier. May 1997. 1996. January/ February 1998. pp. pp." IEEE Trans.34. 1997. Liaw. "Analog and digital speed control of DC drives using proportionalintegral and integralproportional control techniques.2.241248. on Industry Applications. 1998. 4. January 1998. Xu. Ouyang.129136. pp. 1994. A. D. J. no. Pioufle. “SMCbased adaptive inputoutput linearising control of induction motors. D. pp. K. Prentice Hall. 68.37. 70. S. L." IEEE Trans. . Dessaint. C. (book). Sen. M. and C. no. 67.” IEEE Trans. no.1. "Design and implementation of a high performance induction motor servo drive. 1.G. Kung.. New York. 75. Ojo. vol.227233. vol.44. Bonnassieux. C. September/October. an Introduction to StateSpace Methods." IEEE Trans.5962. C. M. Trzynadlowski. J. Lee. Sal and M. pp. Y.
Chiasson. “Highperformance nonlinear feedback control of a permanent magnet stepper motor. M. pp. “Variablefrequency synchronous motor drives for electric vehicles.Electr. 1996. New York. On Industry Applications. 1996. Tarbouchi and H. On Control Systems Technology. 81. 141. 4. Rekowski. Power Appl. 1996. Miller. Gosden. March 1993. 896903. 80. Omozusi.77. Bodson. Soong and T. pp.L. 79. R. 32.” IEE Proc. June. B. Vol. pp. B. pp. N0. No. F. Tennessee Technological University. O. Musaba and D. no. Novotnak and R. John Cox. 82..November 1994.513. 6.J. 78. . July/August.E. L. M. W. 373378. Analysis of an Interior Permanent Magnet Machine Operating in Generator Mode. LeHuy.” IEEE Trans. Chalmers. 929935. “Highperformance control by inputoutput linearlization technique of an induction motor. T.” IEEE Trans. N. Master Thesis. 1. pp.” IEEE Power Electronics Specialists Conference Record. M. Vol. “Fieldweakening performance of brushless synchronous AC motor drives.” IEEE Industry Application Annual Meeting. 1. “Modeling and analysis of an interior perfornentmagnet DCDC converter generator system. J. J. Ojo and O. 1997. vol. 331340.