Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Electric vehicle power trains : high-performance control for constant power load stabilization Anun, Matias 2014

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata


24-ubc_2014_september_anun_matias.pdf [ 4.96MB ]
JSON: 24-1.0166006.json
JSON-LD: 24-1.0166006-ld.json
RDF/XML (Pretty): 24-1.0166006-rdf.xml
RDF/JSON: 24-1.0166006-rdf.json
Turtle: 24-1.0166006-turtle.txt
N-Triples: 24-1.0166006-rdf-ntriples.txt
Original Record: 24-1.0166006-source.json
Full Text

Full Text

Electric Vehicle Power Trains:High-Performance Control for Constant PowerLoad StabilizationbyMatias AnunIng., Universidad Nacional de Co´rdoba, 2010A THESIS SUBMITTED IN PARTIAL FULFILLMENT OFTHE REQUIREMENTS FOR THE DEGREE OFMASTER OF APPLIED SCIENCEinThe Faculty of Graduate and Postdoctoral Studies(Electrical & Computer Engineering)THE UNIVERSITY OF BRITISH COLUMBIA(Vancouver)August 2014Matias Anun 2014AbstractThe development of sustainable transport systems has experienced great improvements inthe last 15 years. As a result, electric vehicles, namely hybrid electric vehicles (HEVs) andall-electric or battery electric vehicles (BEVs), are slowly starting to coexist with regularinternal combustion vehicles around the world. The complex powering structure of auto-motive electric systems can be described as a distributed multi-converter architecture. Inpursuit of performance, constant-power behavior of tightly regulated downstream convertershas raised as an important challenge in terms of system stability and controllability.The first part of this work presents the theory and experimental validation of the unsta-ble behavior introduced by constant-power loads (CPLs) in power converters, more preciselyin a Buck+Boost cascade converter as the battery charge/discharge unit. Open-loop insta-bility with CPLs is studied in asynchronous and synchronous operating modes identifyingthis latter as a highly destructive scenario that creates undesirable unbounded oscillatorybehavior.The second part of this work presents the derivation of the Circular Switching Surfaces(CSS) and the implementation of the CSS-based control technique for CPL stabilization.The analysis shows that the constant-power load trajectories and the proposed CSS presenta wide, stable operating area and near-optimal transient response. Furthermore, impedanceanalysis of the converter in close-loop control shows advantageous reduced output sourceimpedance. This extremely high dynamic capability prevents the use of bulky DC capacitorsfor bus stabilization, and allows the implementation of metal-film capacitors, which havereliability advantages over commonly employed electrolytic capacitors, as well as reducedESR to improve system efficiency. Beyond the improved stabilization properties of theiiAbstractproposed CCS-based controller, a comparison with traditional compensated linear controllerand non-linear SMC highlights significant improvements in terms of dynamic response forsudden CPL changes. The analysis in this thesis is done in a normalized converter to providegenerality to the results and is valid for any power rating. Simulation and experimentalresults are provided to validate the work.The last part of this thesis work presents the design, construction, and testing of ahigh-power 3-phase converter. This platform is intended for electric motor driving and isable to manage 20kW of power flow and above, making it suitable for high power tractionsystem development. The platform features an Intelligent Power Module (IPM) to providewith flexibility allowing for changing the power module according to the requirements ofthe development. Testing of the platform was done in a 0.5HP AC induction motor drivecontrolled with Voltz-per-Hertz control technique. The integration of the BCDU and thehigh-power 3-phase motor drive platform conform a high-power bidirectional motor driveplatform for the development and testing of control techniques for energy management inEV.iiiPrefaceThis work is based on research performed at the Electrical and Computer Engineering De-partment of the University of British Columbia by Matias Anun, under the supervision ofDr. Martin Ordonez. Some experimental validation work was done in collaboration withIgnacio Galiano Zurbriggen.A version of Chapters 1 and 2 has been published at the IEEE Applied Power ElectronicsConference and Exposition (APEC) 2014 [1], and was submitted to a power electronicsjournal [2].Chapter 3 is based on the collaboration project between Future Vehicles TechnologiesCo. and Matias Anun under the supervision of Dr. Martin Ordonez and the technicalcollaboration of Peter Ksiazek. Hardware design, assembly and testing of the power platformwas done by Matias Anun.As first author of the above-mentioned publications and work, the author of this thesisdeveloped the theoretical concepts and wrote the manuscripts, receiving advice and technicalguidance from Dr. Martin Ordonez. The author developed simulations and experimentalplatforms, receiving contributions from Dr. Ordonez’s research team, in particular from thePhD. student Ignacio Galiano Zurbriggen, who assisted the author in the development ofsome experimental tasks.ivTable of ContentsAbstract . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . iiPreface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ivTable of Contents . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . vList of Tables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . viiList of Figures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . viiiAcknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xiDedication . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xii1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11.1 Motivation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11.2 Literature Review . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31.2.1 Battery Charge/Discharge Units . . . . . . . . . . . . . . . . . . . . 41.2.2 CPL in EV Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . 61.2.3 CPL Stabilization Methods . . . . . . . . . . . . . . . . . . . . . . . 81.3 Contribution of the Work . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81.4 Organization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 102 Constant-Power Load Instability in BCDU: Theory and Validation . . . 122.1 BCDU Implementation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12vTable of Contents2.2 Normalized Buck+Boost Cascade Converter . . . . . . . . . . . . . . . . . . 122.3 CPL Linear Analysis and Open-loop Instability . . . . . . . . . . . . . . . . 152.4 Experimental Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 162.5 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 193 High Performance CSS-Based Control for CPL Stabilization . . . . . . . 223.1 CSS-Based Control Strategy . . . . . . . . . . . . . . . . . . . . . . . . . . 223.1.1 CSS Derivation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 223.1.2 Closed-Loop Control Law Definition . . . . . . . . . . . . . . . . . . 243.2 Closed-Loop Operation and Performance Assessment . . . . . . . . . . . . . 283.3 Stability and Limits of Operation . . . . . . . . . . . . . . . . . . . . . . . . 313.4 Experimental Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 353.5 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 414 Development of High-Power Motor Drive Platform for Electric Vehicles 424.1 System Requirements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 424.2 Platform Description . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 434.2.1 Power Module . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 434.2.2 Interface Design . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 454.2.3 Open-Loop Test . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 464.3 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 495 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 505.1 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 505.2 Future Work . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51Bibliography . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52viList of Tables1.1 EVs energy storage system characteristics . . . . . . . . . . . . . . . . . . . 41.2 Electric motor implementation and main parameters . . . . . . . . . . . . . 62.1 Platform parameters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 173.1 Output capacitor comparison . . . . . . . . . . . . . . . . . . . . . . . . . . 364.1 AC motor parameters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46viiList of Figures1.1 Simplfied block diagram of the power structure of a BEV. . . . . . . . . . . . 21.2 Implementation of the BCDU. . . . . . . . . . . . . . . . . . . . . . . . . . . 51.3 Linearization of the CPL characteristic curve (a) allows to model the load asa negative incremental resistor (b) identifying the unstable nature of the load. 71.4 The CSS-based closed-loop control implemented in a Buck+Boost cascadeconverter as BCDU not only to provides high stabilization capabilities forCPL operation, but also remarkably improves performance during transientswith respect to recently-proposed linear and non-linear controls. . . . . . . . 92.1 BCDU implementation with the Buck+Boost cascade converter presents en-hanced properties such as increased performance and efficiency in comparisonto other bidirectional DC-DC converters. . . . . . . . . . . . . . . . . . . . . 132.2 Normalized Buck+Boost cascade converter structures (a) S2 = S3 = ON , (b)S1 = S3 = ON and (c) S1 = S4 = ON . . . . . . . . . . . . . . . . . . . . . . 142.3 The normalized Buck+Boost cascade converter loaded with the linear model ofthe CPL. Input-to-output transfer function analysis identifies potential right-half-plane poles, depending on the filter parameters. . . . . . . . . . . . . . . 152.4 Cascade converter experimental setup. . . . . . . . . . . . . . . . . . . . . . 172.5 Bounded oscillatory response is obtained for the open-loop cascade converterin asynchronous step-down operation when load is switched from CR to CP. 19viiiList of Figures2.6 Unbounded oscillatory response is obtained for the open-loop cascade con-verter in synchronous step-down operation when load is switched from CR toCP. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 203.1 Buck+Boost cascade converter trajectories in its three structures describecircumferences in the state plane. . . . . . . . . . . . . . . . . . . . . . . . . 233.2 For closed-loop operation, switching surfaces σ1 and σ2 are employed for step-down operation, while σ2 and σ3 are used for step-up operation. . . . . . . . 253.3 The CSS-based control strategy is conceptually analyzed in the state planewith the CPL trajectories. For the converter in its three structures, a wideoperating area is identified for stable operation and optimal transient response. 273.4 Closed-loop operation using the CSS for (a) step-down and (b) step-up op-eration. The controller addresses time-optimal start-up and stable operationwith near-optimal transient response for sudden load changes. . . . . . . . . 293.5 Simulation results for the proposed closed-loop control based on the CSS showsan important time-recovery improvement with reduced switching actions whencompared with linear compensated control and SMC. . . . . . . . . . . . . . 303.6 Impedance analysis results for the structure (a) shown in (b) reveal a highlyreduced Zsource for the cascade converter compared with the Zload of the CPLs,confirming the stabilization capabilities of the proposed control technique. . 333.7 Transient response for a family of CPL steps in (a) step-down and (b) step-upoperation shows transient response improvements for several loading conditions. 343.8 Experimental results for output stability obtained for the converter loadedwith 400W CPL when operation is switched from open loop to closed loopwith CCS. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 373.9 Experimental results for the cascade converter loaded with 1kW CPL inclosed-loop step-down operation. . . . . . . . . . . . . . . . . . . . . . . . . . 37ixList of Figures3.10 Experimental results for the 1kW platform in (a) step-down and (b) step-upoperation for a 500W CPL loading and unloading transients show the near-optimal operation with the CSS control strategy. . . . . . . . . . . . . . . . . 393.11 Experimental results for the 1kW platform in (a) step-down and (b) step-upoperation for several CPL steps show improved dynamic output response withminimum overshoot. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 404.1 Block diagram of the high-power motor drive platform. . . . . . . . . . . . . 434.2 Schematic diagram of the high-power IPM. . . . . . . . . . . . . . . . . . . . 444.3 3D design of the IPM interface board. . . . . . . . . . . . . . . . . . . . . . . 474.4 Platform test setup. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 484.5 Experimental results show the 3-phase stator currents for the 0.25HP ACinduction motor with no load running at the nominal speed of 1800RPM. . . 49xAcknowledgmentsI would like to thank my supervisor Dr. Martin Ordonez for recruiting me as part of hisresearch team. His support, extreme patience, and drive were a fundamental guiding lightduring the development of my Master’s program.I would also like to acknowledge my lab mates from the Alpha Technology Power Labo-ratory for sharing their experience and knowledge, making these years a fulfilling period. Inparticular, thanks go to my old friend Ignacio Galiano Zurbriggen. Without his input andbackup, this outcome would not have been possible.I want to express my deepest gratitude to my parents Monica and Simon and to mybrother Joaquin for their unconditional love and understanding. I could not have made itthrough this learning experience without their continuous support. To my girlfriend Vanesa,thank-you for enduring a long-distance relationship for these couple of years.xiFor my parentsxiiChapter 1Introduction1.1 MotivationIncreasing concern for the environment, as well as rise in the fossil-fuel prices have pushedsocieties to seek alternative means of transportation to reduce oil dependence and diminishair pollution. As a result, the development of a sustainable transport system has experiencedgreat improvements in the last 15 years. Electric vehicles (EVs), namely hybrid electricvehicles (HEVs) and all-electric or battery-electric vehicles (BEVs), are slowly starting tocoexist with regular internal combustion vehicles around the world [3–5].EVs present several alternatives for the implementation of the propulsion system. Besidesusing more than one energy source for propulsion, HEVs combine and Internal CombustionEngine (ICE) with an electric motor in a parallel or series architecture to improve fuelefficiency and lower emissions compared to regular ICE vehicles. In BEVs, the tractionsystem relies solely on electric motors, eliminating any emission generation. For this reason,BEVs are usually called zero-emission vehicles (ZEVs). A purely EV presents the followingcharacteristics:No exhaust pollutant emissions.High level of efficiency.Reduced maintenance for battery and electric motor systems.Virtually zero maintenance required for the associated electronics.No oil changes and cleaner operation.11.1. MotivationMain HVDC BusElectric MotorSpeed ControllerHigh- Power3-Phase InverterDC-DCConverterLV DC BusLow-PowerDC loadsLow-PowerAC LoadsDC/ACInverterHigh-powerDC LoadsBATTERY CHARGE/DISCHARGE UNITHV Battery PackBidirectionalDC-DCConverterBidirectionalDC/ACInverterTRACTION SYSTEMAUXILIARY SYSTEMSFigure 1.1: Simplfied block diagram of the power structure of a BEV.Longer brake lifetime with regenerative and dynamic braking.Minimized noise pollution.BEVs have reported efficiency levels as high as 90% while conventional vehicles (CVs)is 35% at best, making the former the ultimate solution to create a green and sustainabletransport system. The reason for the fast growth of HEVs is their profitability for the indus-try and affordability for consumers, allowing their introduction to the automotive market.HEVs can be thought of as an intermediate stage between CVs and BEVs (purely EVs).Driven by advances in power electronics, which have allowed the development of powerconverters with high levels of efficiency and reliability, the power structure of electric vehicleshas clearly evolved to a power electronics’s intensive solution, relying almost exclusively onthe electric implementation for automotive subsystems. In advanced automotive structures,besides the electric propulsion system, several electrical subsystems are used for control,safety and comfort in the vehicle. The relatively complex power structure of automotive21.2. Literature Reviewelectric systems can be described as a distributed multi-converter scheme including DC aswell as AC subsystems [6, 7]. A simplified block diagram of the distributed system of aBEV is illustrated in Fig. 1.1, where the main DC bus feeds the electric motor drive anddownstream converters, which power different automotive electric subsystems.To provide energy from the source to the electric system, a smart and efficient interfaceis required to achieve reliable and high-quality power flow. Depending on the characteristicsof the energy source and the requirements of the DC link, a particular DC-DC converteris required to perform the voltage conversion ratio. Moreover, in advanced systems theinterface, denominated Battery Charge/Discharge Unit (BCDU), provides the means forrecovering energy from the traction system to the energy storage (regenerative braking)through bidirectionality in the converter, highly improving energy usage.1.2 Literature ReviewAs an emerging trend currently in development, different technologies and approaches canbe found for the design, implementation and control of BEVs, all achieving particular ad-vantages and breakthroughs in performance.As was shown before, the power structure of EV systems has become complex in itsimplementation and it will continue to expand to include new features. The high complexityof the system combined with the latest advances in power electronics have created importantchallenges in terms of system design and controllability, where high-performance controllersare required not only to increase efficiency and competitiveness, but to provide reliableoperation with dynamic stability.The following literature review identifies the state-of-the-art technology for power flowmanagement in EV systems, focusing on the electric power train and battery interface. Inaddition, it presents a complete review on the research done on the existence and behavior ofconstant-power loads (CPLs), and on the current techniques developed for CPL stabilization31.2. Literature ReviewVehicle Battery type Energy [kWh] Nominal Voltage [V]Smart Li-ion 17.6 370Nissan Leaf Li-ion 24 360Testla Model S Li-ion 85 375Ford Focus Li-ion 23 325Toyota Prius Ni-MH 4.4 201.6Honda Insight Ni-MH 0.58 100.8BMW i3 Li-ion 22 380Table 1.1: EVs energy storage system characteristicsin automotive systems.1.2.1 Battery Charge/Discharge UnitsEVs are propelled using bidirectional DC-DC converters as BCDU to interface energy storagesystems, such as batteries, with the main DC bus [8]. These converters are suitable forcontrolling power flow in motoring and regenerative braking operation to improve the overallsystem efficiency and travel range. Two battery technologies predominate in the EV marketand are shown in Table 1.1. However, Ni-MH most likely be displaced by Li-Ion in theupcoming generation of EVs as, besides presenting higher energy and power density andreducing total weight (critical parameter for EV application) than Ni-MH, Li-Ion has becomemore reliable [9].The nominal DC-link voltage is in the range of 400V to adapt to the power requirementsof the propulsion system [10], and is typically higher than the energy storage voltage, thusstep-up operation is required to obtain the DC-link voltage level [11]. However, in caseswhere the nominal battery output voltage–which depends mainly on the number of cellsstacked and presents an approximate variation of (+15%,−30%)–overlaps with the DC-linkvoltage required, step-down operation is necessary as well [12]. The Buck+Boost cascadeconverter depicted in the lower part of Fig. 1.2 is capable of performing step-down and step-up operations and presents enhanced properties such as increased performance and efficiencyin comparison to other bidirectional DC-DC converters [13–16].41.2. Literature ReviewBatteryCharge/DischargeUnitBuck+Boost Cascade ConverterLnvccniLnS1S2S3S4vonC2n 1 2pi=C1n 1 2pi=1 2pi=BidirectionalDC/DCConverterHV Battery PackMain HVDC Busvccn von ...vxn =vxVccixn =ixIref= ixVccZoωn=ωωoNormalizing ParametersNormalizedVariablesZo= L CVccωo= L C1Figure 1.2: Implementation of the BCDU.51.2. Literature ReviewVehicle Type Motor type Power [kW]Honda Insight HEV (parallel hybrid) BLDC 10Toyota Prius HEV (series-parallel hybrid) BLDC 60Smart Fortwo BEV BLDC 55Nissan Leaf BEV BLDC 80Ford Focus BEV BLDC 107BMW i3 BEV AC induction 130Testla Model S BEV AC induction 310Table 1.2: Electric motor implementation and main parameters1.2.2 CPL in EV SystemsThe idea behind more-electric vehicles (MEVs) is to evolve from mechanic and/or hydraulicsubsystems to replace them with an electric implementation in order to improve systemefficiency and reliability. As a consequence, electric auxiliary systems in EVs have expandedfrom basic lighting and heating systems to more complex ones including electric steering,dynamic and regenerative braking, active suspension for ride-height control, to mention a few.For these subsystems, power electronics perform ON/OFF switching as well as transformingpower form and level. Overall power requirements for EVs auxiliary systems is estimated toincrease to more than 3kW in the next few years [12].The traction system in BEVs, on the other hand, consists of a high-power 3-phase DC/ACinverter that draws DC power from the batteries and delivers AC power to the the electricmotor, which converter the electrical energy into mechanical energy. For vehicle controllabil-ity, a speed/torque control strategy is implemented and controlled with the throttle pedal.Table 1.2 specifies the type of electric machine employed in the most popular EVs in themarket of the last 2 years. Power rating is also specified for the traction system, identifyingthe level of power required by the DC/AC inverter from the batteries. For purely electricEVs, power rating generally exceeds the 50kW . An HEV motor might require reduced power,depending on the topology of the propulsion system.As depicted in Fig. 1.1, the BCDU delivers power to the traction system and the auxiliarysystems. High-efficiency high-bandwidth power converters fed by the main bus create unique61.2. Literature Review(a) (b)V.I=PconstI∆ IV∆V XOPVOPIOP> 0∆V∆I < 0;ion = vonPon ≅ Reqn= VonPon2-Figure 1.3: Linearization of the CPL characteristic curve (a) allows to model the load as anegative incremental resistor (b) identifying the unstable nature of the load.upstream dynamic characteristics which have been the subject of study over the past fewyears. Under tight-speed regulation with constant torque-load relationship, the motor driveexhibits constant power (CP) behavior at the input of the DC bus [17, 18]. A similarsituation occurs in downstream converters under tight regulation [19–23]. As shown in Fig.1.3, the dynamic behavior of CPLs is equivalent to a dynamic negative resistance which,under certain conditions, can produce instability in the DC bus and consequently, in thesystem [19]. As it is deduced with the linearized model of the CPL, the instability effectis directly proportional to the CP level, hence major concern is focused on the tractionsystem CP behavior in relation to the power levels required for the vehicle’s motion. Onthe other hand, according to the assessment of the limitations of practical CPLs in realworld application which have been studied in [20], practical bandwidth limitations play afundamental role in CP behavior. In this work, an ideal (infinite bandwidth) CPL model isemployed in the analysis to account for the worst-case scenario.71.3. Contribution of the Work1.2.3 CPL Stabilization MethodsPassive stabilization methods have been developed to stabilize the open-loop converter. RCand RL networks were introduced at the resonance frequency of the filter in the converter,increasing the overall damping of the system [20, 21]. Although the method is simple, theaddition of elements to the filter reduces efficiency and negatively impacts size and cost.Active stabilization methods were presented in [22–24], which avoid the introduction ofadditional components. Nevertheless, while these techniques stabilize the open-loop system,a feedback controller is still required to achieve system regulation.Beyond traditional linear controllers, which rely on a linearization process that limits thevalidity of the controllers to a small area around the operating point, the analysis developedin [25–27] demonstrates the potential of non-linear techniques to address CPL instability.Boundary controllers have been widely studied and have proven to be robust with large-signal validity [28]. Furthermore, with the correct switching surface selection, remarkableimprovements can be made in terms of efficiency and dynamic response [29, 30].1.3 Contribution of the WorkCPL behavior of tightly-regulated converters is a common scenario in high-performanceautomotive developments and is still a major concern for the design and implementation ofstable and reliable power systems. An additional challenge is optimizing the packing of thecomplex electromechanical system with the associated power electronics.The simplest approach consists of adjusting the converter filtering parameters to achievea stable operating converter, but this penalizes size and cost. Although many techniqueshave been developed in the past few years to stabilize CP-loaded converters, there is still nounique solution that ensures safe operation under all operating conditions.This work proposes a simple and practical non-linear control, implementing CircularSwitching Surfaces (CSS) to address CPL instability in cascaded DC-DC converters used81.3. Contribution of the Worktn0.2 0.4 0.6 0.8 LnvonSliding mode controllerProposed controllerLinear controller0 1 2 3 4 5 600.3p onCPL00.50.75v on0.341.152.32Near-optimaltransient responseMinimumswitchingactionsBatteryCharge/DischargeUnitBuck+Boost Cascade ConverterLnvccniLnS1S2S3S4vonC2n 1 2pi=C1n 1 2pi=1 2pi=BCDU Close-loop ControlBidirectionalDC/DCConverterHV battery packCPLsMain HV DC busionσ30iLnvonvccnσ2σ1σ2von,targetStep-down op. Step-up op.point target op.pointvon,targetiLn,target {STRUCTURE II STRUCTURE IIISTRUCTURE IISTRUCTURE I(a) (b)Figure 1.4: The CSS-based closed-loop control implemented in a Buck+Boost cascade con-verter as BCDU not only to provides high stabilization capabilities for CPL operation, butalso remarkably improves performance during transients with respect to recently-proposedlinear and non-linear electric vehicle power systems as depicted in Fig. 1.4(a). The analysis of switchingsurfaces goes beyond typical applications with resistive loading studied in [30, 31], as wellas undamped systems with constant-current loading studied in [32–34]. Moreover, as shownin Fig. 1.4(b), the control technique proposed provides a solution to constant power loadingconditions while achieving outstanding dynamic response in comparison to recently proposedcontrol strategies.Bulky DC-link capacitance employed to provide system stability can be reduced giventhe improved large-signal stability achieved by the switching surfaces proposed in this work,highly reducing size and, ultimately, the cost. Furthermore, the proposed switching sur-91.4. Organizationfaces would allow electrolytic-type capacitors to be replaced by metal-film type capacitors,increasing the reliability of the system [35–39].It is important to note that the work is developed in a normalized domain depicted in Fig.1.2, which allows for the extrapolation of the results obtained to any converter, regardless ofvoltage and current ratings and filter parameters, and ensuring high-performance operation.Successful development of automotive-oriented power systems requires extensive designand testing stages in order to accomplish reliable and high-performance operation. For thisreason, a test bench is fundamental to emulate different working conditions and to validatecomplete system stability and performance. Through this thesis work, a scaled high-powerBCDU and a flexible motor-drive platform were developed to conform an EV power platform.These were used for the experimental validation of the present work.1.4 OrganizationThis thesis is organized as follows:Chapter 2 develops the concept of CPL and the instability effect introduced in power con-verters. As was determined through the literature review, analysis and experimental resultsare carried out in a Buck+Boost cascade converter as the BCDU, topology that allows forconsideration of any possible configuration regarding energy storage characteristics and DClink requirements. Instability scenarios are analyzed with a linearized Buck+Boost converterloaded with CPL. A more detailed and accurate stability criteria is derived by accountingfor inductor resistive losses as well as capacitor ESR, which is normally neglected. Proofof concept is attained with experimental results on a 1kW Buck+Boost cascade converterloaded with a fast current load externally controlled to behave as a CPL. Open-loop insta-bility with CPL was studied in asynchronous and synchronous operating modes, identifyingthe latter as the worst operating condition.In Chapter 3, the CSS are initially derived and later employed to define the CSS-based101.4. Organizationclosed-loop control strategy. Stabilization characteristics are analyzed with clarity in thestate plane, addressing the operational characteristics for start-up and sudden CPL changes.Time domain simulations are also included. Previous work proposed for CPL stabilization arecompared with the proposed technique to establish a benchmark and asses the performanceimprovements obtained. For the validation of the control technique proposed, several testresults are presented. All results were obtained on a 1kW Buck+Boost cascade converter.Chapter 4 presents the design, building and testing of a flexible high-power motor driveplatform for EV applications. The platform operates with an Intelligent Power Module(IPM) from Powerex to manage power flow. The board designed to interface the powermodule integrates voltage, current and temperature measurements, and allows connectionwith several IPM modules giving the flexibility to adapt the power module according tothe requirements. Testing of the platform was done in a 0.5HP AC induction motor drivecontrolled with Voltz-per-Hertz control technique.Lastly, Chapter 5 presents a summary, remarking on the major outcomes of the workcompleted within the thesis and outlining future ideas for work.11Chapter 2Constant-Power Load Instability inBCDU: Theory and Validation2.1 BCDU ImplementationThe Buck+Boost cascade converter, shown in Fig. 2.1, is a bidirectional topology obtainedby cascading a Buck with a Boost converter, allowing independent step-down or step-upoperation. The different structures of the converter are depicted in Fig. 2.2. In dischargeoperation (power flow from energy storage to the main DC bus), step-down conversion isobtained by switching between structures I and II. For a step-up operation, structures IIand III are employed. The behavior for the converter with CPL is described through thefollowing set of differential equations:LdiLdt = u1 vcc − u2 voCdvodt = u2 iL − io(2.1)In (2.1), u represents the state of the switches: u1 = 1 for switches S1 and S2, ON andOFF, respectively, and u2 = 1 for switches S3 and S4, OFF and ON, respectively.2.2 Normalized Buck+Boost Cascade ConverterIn order to give generality to the development, the converter is normalized, eliminating de-pendence on converter voltages, currents and power ratings, as well as filter characteristics:122.2. Normalized Buck+Boost Cascade ConverterBatteryCharge/DischargeUnitBuck+Boost Cascade ConverterLnvccniLnS1S2S3S4vonC2n 1 2pi=C1n 1 2pi=1 2pi=BCDUControlBidirectionalDC/DCConverterHV Battery PackPWM GateSignalsAnalog.Measurem.CPLs(DownstreamConverters)Main HVDC BusionFigure 2.1: BCDU implementation with the Buck+Boost cascade converter presents en-hanced properties such as increased performance and efficiency in comparison to other bidi-rectional DC-DC converters.vxn =vxVcc; ixn =ixIref= ixVccZo ; ωn =ωω0The normalization procedure makes use of the following base parameters: the inputvoltage vcc, characteristic impedance of the combined L and C values Zo =√L/C, andthe natural resonance period ω0 = 1/√LC. The subindex n is used to indicate normalizedvariables. Normalizing (2.1) yields the following:132.2. Normalized Buck+Boost Cascade Converter(a) Structure ICPLLnvccniLnS1S2S3S4vonC1n 12pi= =12pi C2n 12pi=(b) Structure IICPLLnvccniLnS1S2S3S4vonC1n 12pi= =12pi C2n 12pi=(c) Structure IIICPLLnvccniLnS1S2S3S4vonC1n 12pi= =12pi C2n 12pi=ionionionFigure 2.2: Normalized Buck+Boost cascade converter structures (a) S2 = S3 = ON , (b)S1 = S3 = ON and (c) S1 = S4 = ON .12pidiLndtn= u1 vccn − u2 von12pidvondtn= u2 iLn − ion(2.2)142.3. CPL Linear Analysis and Open-loop InstabilityRont1vccniLnRLn 1S1S2S3S4Rcn2piDt1 0 1 0-Reqn1 2piFigure 2.3: The normalized Buck+Boost cascade converter loaded with the linear modelof the CPL. Input-to-output transfer function analysis identifies potential right-half-planepoles, depending on the filter parameters.2.3 CPL Linear Analysis and Open-loop InstabilityThe presence of CPLs in power converters introduces unique dynamics into the system thatare not present with typical resistive or constant current loads, and which generate open-loop instability in the converter. An ideal CPL can be modeled as (2.3) where, for a givenoperating point (Von, Ion), the product of the load voltage and current is kept constant(Pon,const = VonIon) and the instantaneous value of the load impedance is positive (Ron =Von/Ion). A linear approximation is obtained by deriving (2.3) in the small area around(Von, Ion), yielding the linear equivalent incremental impedance Reqn. As shown in (2.4),Reqn at the given operating point, or at any other, is negative. The linearized CPL model is(2.5).ion =Ponvon(2.3) diondvon= −Ponv2on= − 1Reqn(2.4)ion,lin = 2PonVon− PonV 2onvon = 2Ion −1Reqnvon (2.5)A cascade converter in step-down operation mode is illustrated in Fig. 2.3(a) with nor-152.4. Experimental Resultsmalized inductor and capacitor parasitic resistances included in the model. By consideringthe converter loaded with a CPL, and by using the linearized model (2.5), small-signal anal-ysis for the input-to-output transfer function yields the following denormalized result:vˆo(s)vˆcc(s)= ReqReq −Rc1s2LC + s[ LRc −Req+ C(RL +Req RcReq −Rc)]+ RL −ReqRc −Req(2.6)Analyzing the roots of the characteristic equation in (2.6), it can be seen that the negativeequivalent impedance of the CPL produces a shift in the system’s poles which may bedisplaced to the right-half plane, depending on the filter parameters, making the systemopen-loop unstable. For a high-efficiency converter, the losses RL and Rc should remainminimal. Thus, stability of the converter relies on oversizing the output capacitance tocomply with the following:L| Rc −Req |< C(RL +Req RcReq −Rc)(2.7)Previous criteria is overly conservative and, as a sufficient condition, it ensures stableoperation of the open-loop converter.2.4 Experimental ResultsA scaled 1kW Buck+Boost cascade converter was implemented in hardware with the param-eters indicated in Table I. A fast dynamic CPL with a bandwidth of approximately 25kHzwas implemented as part of the test bench. A picture of the experimental setup is shown inFig. 2.4. For a nominal output power of Po = 1kW and for the filter parameters indicated,the linear stability analysis (2.7) for traditional controllers indicates that the system is highlyunstable.162.4. Experimental ResultsFigure 2.4: Cascade converter experimental setup.PARAMETER VALUEL 920µFRL−ESR 0.29ΩC 20µFRC(ESR) 9mΩvo 90Vvcc 50V − 200VPo(max) 1kWTable 2.1: Platform parameters172.4. Experimental ResultsFig. 2.5(b) and 2.6(b) show the experimental results of the cascade converter operatingin open loop in asynchronous and synchronous mode, respectively. Switching frequencyis set to a fixed value of fsw = 20kHz and duty cycle D = 0.75. Operation starts withresistive load Ron, and at t = 3.2ms, the load switches to CP. As expected, the negativeincremental impedance has a negative impact on the stability of the system, producingoscillations at the resonance frequency at the output of the converter. With asynchronousoperation (Fig. 2.5(a)), the load dynamics produce an increase in the energy stored bythe reactive elements cycle after cycle, increasing the oscillations until the converter entersdiscontinuous conduction mode (DCM). At this point, the inductor current cannot invert thedirection, and remains discharged while the capacitor continues to discharge. This processlimits the energy stored in the reactive elements, and the operation quasi-stabilizes aftert ≈ 19.2ms in a limit cycling which can be observed in the geometric domain.On the other hand, when switched to synchronous operation (Fig. 2.6(a)), inductorcurrent is allowed to reverse direction. The same loading process is applied, and in this case,the energy in the reactive elements rises without boundaries, increasing the oscillations untilthe system is shut down using a protection at t ≈ 21ms. This unbounded and destructivebehavior is a major concern that will be addressed by the proposed switching surfaces in thefollowing chapter.Besides the criteria for stable operation derived previously, the combination of CP as wellas regular resistive loads in the power system mitigates CP instability. However, the overallsystem stability should not rely on the existence of such resistive loads. Practical CPL modelsdiffer from (2.4) mainly due to bandwidth limitations of real converters/inverters acting asloads. In current applications (2014), the bandwidth of the CPLs can range from a few Hzto up to few kHz (power supplies connected to the DC Link). This work adopts the worst-case operating scenario of very large CPL bandwidth. If stability and high performanceis achieved for the theoretical high bandwidth CPL, then in real world applications, safeoperation and rapid response is guaranteed.182.5. Summary(b)iLvovoiL Limit cyclingCPLBounded oscillatory behaviorsw(a)Ront1vccniLnRLn 1S1S2S3S4Rcn2piDt1 0 1 0-Reqn1 2piFigure 2.5: Bounded oscillatory response is obtained for the open-loop cascade converter inasynchronous step-down operation when load is switched from CR to CP.2.5 SummaryThis chapter presented the theory and validation of the unstable behavior introduced byCPL in power converters. The concept was analyzed in a Buck+Boost cascade converter asa BCDU that enables step-up and step-down operation, covering a wide range of possible192.5. Summary(a)(b)Unstable behaviorUnbounded oscillatory behaviorvoiLCPLSystem shutdownswvoiLt1-ReqnRonvccn RLnS1S2S3S4RcnDt1 0 1 01 2pi1 2piiLnFigure 2.6: Unbounded oscillatory response is obtained for the open-loop cascade converterin synchronous step-down operation when load is switched from CR to CP.scenarios, depending on the energy source voltage level and the requirements of the DClink. A normalizing procedure was applied to the converter which was used to generalize theoutcomes of the work. To determine instability behavior of the converter loaded with CPL,the load was subject to a linearization process, allowing for computation of the input-to-202.5. Summaryoutput transfer function. Filter parasitic resistances were included in the model, since theyplay a vital role in system efficiency as well as system stability. With pole-locus analysis ageneral criteria was obtained to determine the behavior of the converter in open-loop.The experimental platform was designed with reduced output capacitance, allowing forthe use of metal-film caps highly recommended for system reliability. Using the criteriaderived for stability, the converter loaded with a CPL of Po(max) = 1kW was clearly identifiedas an unstable system. Open-loop instability with CPL was studied in asynchronous andsynchronous operating modes, identifying different behaviors. As could be observed in theexperimental results, this last operating mode caused the oscillations of the state variablesto grow without boundaries, creating a highly destructive operating scenario that requiredshutdown action for protection.21Chapter 3High Performance CSS-Based Controlfor CPL Stabilization3.1 CSS-Based Control StrategyThis section presents the derivation of the Circular Switching Surfaces (CSS) which are lateremployed to define the closed-loop control law for both operating modes.3.1.1 CSS DerivationTo derive the switching surfaces, the normalized differential system (2.2) is solved to obtainthe converter’s circular trajectories for all three of its structures depicted in Fig. 2.2. Thedifferential equations of the system, repeated in (3.1), are combined and solved, obtainingthe time domain solutions (3.2) and (3.3).12pidiLndtn= u1 vccn − u2 von12pidvondtn= u2 iLn − ion(3.1)von = [von(0)−u1u2vccn] cos(2pitn) + [u2 iLn(0)− ion] sin(2pitn) +u1u2vccn (3.2)iLn = [iLn(0)−1u2ion] cos(2pitn) + [u1 von(0)− u2 vccn] sin(2pitn) +1u2ion (3.3)223.1. CSS-Based Control Strategyλ3vccnioniLn(0)0ω0nω0niLnvonr1[von(0); iLn(0)]vccnionλ1λ2ω0nr2von(0)Figure 3.1: Buck+Boost cascade converter trajectories in its three structures describe cir-cumferences in the state plane.Combining (3.2) and (3.3), time is eliminated, resulting in the generalized trajectoryequation (3.4):λ : (von −u1u2vccn)2 + (u2 iLn − ion)2 − [von(0)−u1u2vccn]2 − [u2 iLn(0)− ion]2 = 0 (3.4)Replacing the parameters u1 and u2, the trajectories corresponding to each of the struc-tures depicted in Fig. 2.2 are obtained and shown in (3.5) to (3.7).λ1 : v2on + (iLn − ion)2 − v2on(0)− [iLn(0)− ion]2 = 0 (3.5)λ2 : (von − vccn)2 + (iLn − ion)2 − [von(0)− vccn]2 − [iLn(0)− ion]2 = 0 (3.6)λ3 :(vccnion)von + iLn −(iLn(0) +von(0) vccnion)= 0 (3.7)233.1. CSS-Based Control StrategyThese trajectories can be rewritten in the following form:λ : (x− Cx)2 + (y − Cy)2 − r2o = 0 (3.8)where the center of the circumferences is displaced by vccn and ion, and the radius rodepends on the operating point values von(0) and iLn(0). To facilitate visualization, the tra-jectories λ1 to λ3 are illustrated in Fig. 3.1 for a particular set of parameters (von(0), iLn(0)).The direction of the rotational speed ω0n is indicated by a set of arrows on each tra-jectory. λ3 is considered a particular case of a circumference in which the center shifts(Cx, Cy) → (∞,∞) approaching a straight line in the state plane.3.1.2 Closed-Loop Control Law DefinitionThe trajectories λ1 to λ3, defined in the previous section, are employed to define the controllaw based on the operating point and the intersection of the converter’s switching surfaces.For this task, some parameters need to be defined. vccn = 1, given that the normalization isdone with vcc. von(0) taking values in an interval (a < von,target < b), depending on the batterySOC. The boundaries a and b are given by the maximum and minimum operating valuesof the energy storage system. The parameter iLn(0) is matched with the targeted inductorcurrent (iLn,target) for each operating mode. For step-down operation, since structure I and IIare used, iLn,target = ion. For step-up, structures II and III are used, hence iLn is equal to theinput current of the converter and related to the output current ion as iLn,target = ion/(1−D)with D = 1 − vccn/von,target. Having all the required parameters defined, the control law ispresented for each operating mode.For step-down operation, u2 = 1 and the control law derived is as follows:Case I: (iLn > ion)if (σ1 > 0) then u1 = 0, else u1 = 1Case II: (iLn < ion)243.1. CSS-Based Control Strategyσ30iLnvonvccnσ2σ1σ2von,targetStep-down op.von,target<vccnioniLn,target =Step-up op.von,target>vccnioniLn,target= (1-D)Target op.point Target op.pointvon,targetiLn,target {STRUCTURE II STRUCTURE IIISTRUCTURE IISTRUCTURE IFigure 3.2: For closed-loop operation, switching surfaces σ1 and σ2 are employed for step-down operation, while σ2 and σ3 are used for step-up operation.253.1. CSS-Based Control Strategyif (σ2 > 0) then u1 = 1, else u1 = 0whereσ1 : v2on + (iLn − ion)2 − V 2on,target (3.9)σ2 : (von − vccn)2 + (iLn − ion)2 − (Von,target − 1)2 (3.10)In step-up operation, u1 = 1 and the control law is defined as follows:Case I: (iLn > ion/(1−D))if (σ2 > 0) then u2 = 1, else u2 = 0Case II: (iLn < ion/(1−D))if (σ3 < 0) then u2 = 0, else u2 = 1whereσ2 : (von − vccn)2 + (iLn − ion)2 − (Von,target − 1)2 − (ion Von,target − ion)2 (3.11)σ3 :( 1ion)von + iLn −(Von,target ion +Von,targetion)(3.12)Fig. 3.2 illustrates the control laws with the CSS established, showing the two areas ofoperation which are determined based on the dynamic value von with respect to vccn.To understand the operation and gain insight into the behavior of the CSS-based controlstrategy proposed, Fig. 3.3 presents the analysis of the CSS with the converter trajectoriesloaded with CPL for each structure depicted in Fig. 2.2. The points A shown in the sub-figures correspond to different initial conditions (ICs) for which the CSS have a sufficientlysmall error regarding the CPL trajectory, ensuring that the target can be achieved by switch-ing to the corresponding structure indicated as λx target. This condition is satisfied for all ICsthat are in the region between the curve Pon and a certain limit, which is determined by themaximum output voltage ripple, e.g. 2%. Thus, when the CSS represented as λx target within263.1. CSS-Based Control StrategyStructure II(b)CPL traj.CSSTarget_AlimitvoniLnABTarget_BPonλ1’λ3’vccn=1ABA’A’Structure I(a)CPL traj.CSSiLnlimitTargetλ1_targetBAPonvonλ2’A’Structure III(c)LimitiLnTargetPonABλ3_targetCPL traj.CSSv o nλ2’A’Figure 3.3: The CSS-based control strategy is conceptually analyzed in the state plane withthe CPL trajectories. For the converter in its three structures, a wide operating area isidentified for stable operation and optimal transient response.273.2. Closed-Loop Operation and Performance Assessmentthe given region is selected, the trajectories evolve until the condition v0n ± 2%, achievingsteady state.For points B outside the region where the CSS trajectory matched the target CPL tra-jectory, the circles that describe the target CSS trajectory now match different CPL tra-jectories that intersect the pon line at a von value which can be either higher or lower thanvon = Von,target (Point A in Fig. 3.3). However, it can be demonstrated that the operatingpoint is always able to return to the convergence area by switching to the proper structureas is indicated conceptually in Fig. 3.3 for the different cases. Consequently, once point A isreached, the structure is switched according to the operating mode, reaching again the tar-get CSS trajectory that allows the operating point to arrive at the desired target. Althoughextra switching actions are required, the operating region is extended.3.2 Closed-Loop Operation and PerformanceAssessmentThe CSS were implemented in simulation for the Buck+Boost cascade converter for start-upand CP loading transients. Results are shown in Fig. 3.4(a) and (b) for step-down andstep-up operation, respectively.In the first case, the converter is initially driven from ❶ to the target point with noload ❸. With the switches configured as in structure II, the operating point moves alongλ2. When it reaches the switching surface σ1(Pon=0) ❷, the converter switches to structureI, achieving steady-state in two switching actions in minimum time. At instant tn = 1, aCPL of Pon = 0.15 is applied and the new target point moves to ❺. According to the CSS,the converter switches again to structure II until it reaches the switching surface σ1(Pon=0.15),where the control now switches back to structure I. The operating point is directed towardsthe target, where it achieves steady state. The same start-up and loading process is donefor step-up operation and shown in Fig. 3.4(b). In this case, the controller uses structures283.2. Closed-Loop Operation and Performance AssessmentConverter trajectorySwitching Surfacev on No-load start-uptransientCPL step-uptransientp onS 1tn00.7501100.15CPLSwitching actionsvoni LnTarget pointNo-load start-uptransientCP loadingtransientPconst=0.15σ1(no load) σ1(Pon=0.15)0 0.2 0.4 0.6 0.800.20.40.6 λ2λ2(a)σ2(no load)σ2(Pon=0.2)Pconst=0.2No-load start-uptransientCP loadingtransient0.8 0.9 1 1.1 1.2 1.3voni Ln00.λ3λ311.331100.2tnCPLv onp onS 4No-load start-uptransientCPL step-uptransientSwitching actions(b)Target pointFigure 3.4: Closed-loop operation using the CSS for (a) step-down and (b) step-up opera-tion. The controller addresses time-optimal start-up and stable operation with near-optimaltransient response for sudden load changes.II and III to drive the operating point towards the target with CPL of Pon = 0.2. Theseresults show the stabilizing capabilities and enhanced transient responses for the CSS withsudden load changes in both operating modes.Important advantages in terms of performance and switching actions are obtained whendriving the operating point towards the target point through near-optimal trajectories. Acomparison between the closed-loop control implementing the CSS with linear compensatedcontrol and sliding-mode control (SMC) is shown in Fig. 3.5. Time domain and state-plane293.2. Closed-Loop Operation and Performance Assessmenttn0 1 2 3 4 5 600.3p onCPL00.50.75v on0.341.152.32Theoretical minimumtransient responses0.2 0.4 0.6 0.8 1− LnvonSliding mode controllerProposed controllerLinear controllerMinimumswitching actionsFigure 3.5: Simulation results for the proposed closed-loop control based on the CSS showsan important time-recovery improvement with reduced switching actions when comparedwith linear compensated control and SMC.plots show start-up and sudden load change transient performance for the three controltechniques. The linear controller was implemented in a compensated converter with the303.3. Stability and Limits of Operationtraditional approach, by analyzing root locus and frequency response for the design of thefeedback loop. Analyzing the response of the converter to a sudden CPL change, it is ob-served that the converter achieves stability after 2.32 normalized time units. SMC is aninteresting alternative control for stabilizing and controlling CPL systems. For this con-troller, a transient of 1.15 normalized time units is obtained for the same load change. Withthe CSS-based control, the recovery time shown for the same CPL step is 0.34 normalizedtime units, representing an improvement of 3.4 times with respect to SMC and of 6.8 timeswith respect to the linear controller. Analyzing the switching actions (in the state-planeplot) required by each technique to arrive at the new target point after the sudden CPLchange, it is observed that the proposed controller implements only two switching actions,while for the other two techniques, the number is extremely large.3.3 Stability and Limits of OperationSeveral approaches can be found in literature to analyze the stabilization capabilities of theproposed control technique. Middlebrook’s criterion [40], derived from the impedance ratioanalysis in filtered switching power converters has been found quite conservative. Neverthe-less, combined with gain- and phase-margin criteria, this analysis can deliver good insightinto the stabilization characteristics of the controlled converter.Power converters applied to traction systems might have limited response capabilities toensure smooth operation, especially in vehicle applications. On the other hand, cascadedconverters such as step-down converters for powering low-power loads present increased band-width. Fig 3.6 shows the output impedance Zsource of the cascade converter in closed-loopoperation overlapped with the input impedance Zload of two CPL cases. One case considersa high-bandwidth CPL which was employed for the experimental section, while the otherone has limited dynamic bandwidth, approaching a more realistic case.Impedance ratio criterion requires Zsource << Zload, which in practice can be extended313.3. Stability and Limits of Operationto −6dB difference to ensure stable operation. In this condition, a phase margin of 50 isacceptable for stability. As shown in the Bode plot in Fig 3.6, CPLs exhibit a negativeimpedance characteristic Zload(highBW ) and Zload(lowBW ). In the case of the high-bandwidthCPL, this behavior is maintained up to a frequency of 25kHz, at which point the gain startsto increase, reducing the instability effects. Variation on the phase also indicates that thebehavior is no longer pure CP. For the limited bandwidth CPL, the bandwidth goes downto 1kHz. The cascade converter loaded with Po(max) = 1kW , on the other hand, presentsa desirable reduced output impedance for a wide bandwidth, enough to ensure steady-stateoperation in both CPL cases.In terms of transient response, the analysis and results presented in the previous sectionare extended to different operating points along the CP curve, as well as implemented withdifferent CPL steps to characterize the behavior of the proposed controller. Fig. 3.7 presentsthe simulation results for the Buck+Boost cascade converter in step-down and step-up op-erating mode for a family of CPL steps. Under the operating conditions indicated in thefigure, successive CPLs are applied in steps of ∆Pon = 0.05. Curves obtained for the sim-ulation indicate that the controller is able to stabilize CPL steps of up to Pon = 0.25 forboth operating modes, exhibiting less than 5% overshoot. These results define an improveddynamic performance over an important set of working conditions, making the techniquewidely applicable for CPL stabilization and loading transient control.The sensitivity of the control technique regarding filter parameter drifting and toleranceswas also examined. Different scenarios were established for the normalization procedure ap-plied in Chapter 2, Section 2.2, focusing on the filter characteristic impedance Zo =√L/C.The worst case occurs when the value of L or C increases while the remaining parameterequally decreases. The filter parameters were subject to a variation of 10%. The effect ofsuch variations is a transformation in the switching surfaces, which become elliptical ratherthan circular. Simulations were developed for the different cases so that the response vari-ations could be analyzed. The distortion of the switching surfaces affected the evolution of323.3. Stability and Limits of Operation(a)Cascade converterCSS based close-loop controlHighbandwidthCPLvccZload(High BW)ZsourceLowbandwidthCPLZload(Low BW)(b)f [Hz]−80−40040−260−220−180102101 103 104 105Ampl. [dB]Phase [Deg.]ZsourceZload(High BW)Zload(Low BW)Figure 3.6: Impedance analysis results for the structure (a) shown in (b) reveal a highlyreduced Zsource for the cascade converter compared with the Zload of the CPLs, confirmingthe stabilization capabilities of the proposed control technique.the operating point during a sudden load change and showed less than 5% overshoot com-pared with the ideal case under the same operating conditions. Regarding steady state, thevariations observed were negligible. Moreover, by applying precise calibration, the distortioneffect due to parameter variation can be minimized.333.3. Stability and Limits of Operation(b)1.2 1.4vonxxxxxCPLtargetpointsCPLtransients1.3 1.500.20.4i Ln0.10.30.5vccvo = 1.331.21.30.5 0.7500.10.20.3v onP on0.4tnCPL steps 1.4<5% overshoot(a)0.5 0.75 LnvonxxxxxCPLtargetpointsCPLtransientsvccvo = 0.750.50.75v on0.5 100.20.4P ontn0.30.1CPL steps <5% overshootFigure 3.7: Transient response for a family of CPL steps in (a) step-down and (b) step-upoperation shows transient response improvements for several loading conditions.343.4. Experimental Results3.4 Experimental ResultsExperimental tests were done in the platform described in Chapter 2, Section 2.4, which wasdesigned with a 20µF metal-film output capacitor. This reduced output capacitance leadsto open-loop unstable behavior.Open-loop instability produced by the CPL was shown in Figs. 2.5 and 2.6, where theconverter was operating in step-down mode at a fixed switching frequency and constant duty-cycle. The load is initially resistive, dissipating Po = 250W . The response at the outputof the converter after switching to CPL at t1 = 3.2ms exhibits oscillatory behavior at theresonance frequency fo = 1.15kHz. Synchronous operation (Fig. 2.6) was identified as theworst case operating condition, since oscillation grew without boundaries. For protection,output voltage and current thresholds are set to shut down the power stage to avoid acatastrophic failure. If asynchronous switching were to be implemented, operation would berestricted to the first quadrant, thus enabling operation in DCM. In this case, instabilityeffects are reduced by preventing inductor current iLn from reversing direction and forcingthe trajectory to follow the normalized output voltage axis as shown in Fig. 2.5.To stabilize the open-loop converter, a 1.1mF electrolytic capacitor would be required.Besides the increased capacitance, this type of technology presents an ESR approximately10 times larger than the ESR of the metal-film type employed, which has significant impacton the filter losses. Additional advantages are the extended lifetime for reliable operationand reduced volume and mass. A detailed comparison is provided in Table 3.1.The proposed CSS was implemented in order to control the unbounded behavior andobtain predictable, stable operation. Experimental results for the converter operating underthe CSS-based control strategy are shown in Figs. 3.8 to 3.11. In all cases, the experimentalresults closely resembled the simulation results.Stabilization of the system under CPL load is shown in Fig. 3.8, where the converter isloaded with a Po = 400W CPL. In the first half of the oscilloscope capture, the converter isoperating in open loop and exhibits bounded oscillation. The demonstration of the unstable353.4. Experimental ResultsFilm ElectrolyticPARAMETER (MKP1848620704P4) (CGS112T500V4L)Capacitance 20µF 1100µFRC(ESR) 9mΩ 92mΩDimensions 43.00mm x 21.50mm 50.80mm()V olume 36.6mm3 238.1mm3Mass 36g 250gLifetime (@ 85◦) 100, 000hs 2, 000hsPrice(qty 50) (2014) $11.20 $85.01Table 3.1: Output capacitor comparisonbehavior is done in asynchronous mode to limit the amplitude of these oscillations. In thesecond half of the oscilloscope capture, the proposed CSS strategy is enabled, resulting instable operation. Note that the controller only performs two switching actions (structures Iand II) to drive the operating point towards the target point where it remains stable.In Fig. 3.9, the converter’s stability is further tested by loading the converter withPo(max) = 1kW . The converter exhibits stable operation with a switching frequency offsw = 8KHz, which is accomplished by employing an hysteresis band in the CSS controllaws to obtain the desired ripple. As for frequency shifting, a narrow variation is observedin steady-state operation. If necessary, this variation can be further mitigated by adjustingthe hysteresis band control in the CSS or forcing PWM triggering with precision timers.Experimental results for transient response to sudden CPL change were shown in Fig.3.10 for step-down and step-up operating modes. State-plane plots on the right-hand side ofeach experimental capture expose the evolution of the operating point through near optimaltrajectories, from no-load steady state to a 500W CP loading condition. Unloading to ano-load condition for both cases yielded the same results in terms of transient recovery.Further experimental results to evaluate the predicted transient response are shown inFigs. 3.11(a) and (b), where a family sudden CPL step changes is applied to the converter.In Fig. 3.11(a), the cascade converter is operating in step-down mode with Vo/Vcc = 0.75. At90V , the CPL step-ups are applied. In each case the control strategy responds by changing363.4. Experimental ResultsOpen-loop op. Closed-loop op.voiLswLimitcyclingSteady-stateoperationSteady stateλ2λ1 iLvoFigure 3.8: Experimental results for output stability obtained for the converter loaded with400W CPL when operation is switched from open loop to closed loop with CCS.1kWvoiLpo~11A90Viosw fsw=8KHzσ1σ2TargetpointiLvoFigure 3.9: Experimental results for the cascade converter loaded with 1kW CPL in closed-loop step-down operation.the converter structures according to the CSS control law, achieving fast recovery to steady-state. The results exhibit an output voltage overshoot of less than 5% for CPLs up to373.4. Experimental ResultsPo(max) = 500W as was predicted in section VI. Fig. 3.11(b) shows near-optimal responsefor a sudden CPL load change with negligible overshoot for the cascade converter operatingat Vo/Vcc = 1.25 (step-up operation).383.4. Experimental Results(b)voiLpoioswCP loading transientNo-loadtarget500WtargetCP unloadingtransientCPloadingtransientCPunloadingtransientStableoperation(a)voiLpoioswCP loading transientNo-loadtarget500WtargetCP unloadingtransientCPloadingtransientCPunloadingtransientStableoperationFigure 3.10: Experimental results for the 1kW platform in (a) step-down and (b) step-upoperation for a 500W CPL loading and unloading transients show the near-optimal operationwith the CSS control strategy.393.4. Experimental Results(a)voiopo<5% overshootCPL steps100W500W(b)voiopo<5% overshootCPL steps100W300WFigure 3.11: Experimental results for the 1kW platform in (a) step-down and (b) step-upoperation for several CPL steps show improved dynamic output response with minimumovershoot.403.5. Summary3.5 SummaryThis chapter presented the derivation and implementation of the CSS-based control techniquefor CPL stabilization. Closed-loop operation was analyzed with a simulation and validatedthrough experimental results in the scaled 1kW Buck+Boost cascade converter. Operationof the proposed control technique was conceptually analyzed in the state plane, comparingthe CSS and the CPL trajectories of the converter in its three structures and identifying awide operating area for stable operation and optimal transient response.Impedance analysis of the converter showed reduced output source impedance, indicatingextremely high dynamic capabilities and avoiding the use of bulky DC capacitors for busstabilization. This allows for the usage of metal-film capacitors, which have reliability ad-vantages over commonly employed electrolytic capacitors as well as reduced ESR, improvingsystem efficiency.Besides the stabilization characteristics of the controller proposed, comparison with atraditional compensated linear controller and a non-linear SMC showed significant improve-ments in terms of dynamic response for sudden CPL changes.The analysis was done in a normalized converter, allowing the results to be extrapolateto any platform by denormalization according to the specifications and requirements of agiven converter.41Chapter 4Development of High-Power MotorDrive Platform for Electric VehiclesEffective motor drive control for vehicle propulsion requires implementation of elaboratedcontrolled power structures to manage power flow in an accurate and reliable manner. Formotor drive operation, DC energy is delivered to the electric motor through a 3-phase inverterstructure controlled by a speed/torque control. A precise control is required to ensure smoothoperation of the motor. Meanwhile, during regeneration process, mechanical energy storedin the electric motor is converted to electrical energy and returned to the electric system.Effective energy management is required to ensure safe operation and avoid high DC voltagelevels in the bus.In terms of efficiency, it would be desirable to recover and store the total mechanicalenergy available from the drive train in the batteries for later usage. However, given theenergy storage capability, if the battery state-of-charge (SOC) is high, energy has to bemanaged in another way. Dynamic Braking uses a high-power shunt regulator across theDC bus to dissipate the extra energy, resisting rotation, and decreasing motor speed.4.1 System RequirementsIn order to achieve comprehensive and accurate bidirectional power flow control, a high-powermotor drive platform is required. The following are the key characteristics and capabilitiesachieved in the power stage designed:424.2. Platform DescriptionAC motoruvwDC linkBCDUControlPositionSpeedSensor3-Phase inverterChopperBCDUHV battery packSpeedControllerPWM GatesignalsAnalog.measurem.Gate signalsFigure 4.1: Block diagram of the high-power motor drive platform.• Multipurpose and flexible platform, allowing for adaptation to different needs in termsof motor driving, measurements and sensor connectivity.• Hardware robustness able to withstand mechanical stress and EMI.• Capable of handling 20kW of peak power flow within a 400V DC system.• The power stage must be able to operate at a switching frequency of fsw = 10kHzwith an output frequency of fo = 400Hz.• 3-phase voltage and current measurements.• Temperature sensing.4.2 Platform DescriptionFigure 4.1 presents the block diagram of the Power Drive Platform.4.2.1 Power ModuleThe Power stage is designed to work with Intelligent Power Modules (IPM) from Powerex.The heart of the power platform is the Powerex IPM PM300RL1A060, which is an isolated-434.2. Platform Description3-phase inverter/rectifierChopperIPM structureDC bus BRPNUVWNTCFigure 4.2: Schematic diagram of the high-power IPM.base module designed for power switching applications operating at frequencies of up tofsw = 20kHz. Voltage and current rating of the chosen power module exceed power handlingrequirements, allowing for future application in higher power motor driving.The IPM 7-pack structure is shown in Fig. 4.2 and consists of a 3-phase inverter com-plemented with a chopper structure for dynamic braking.7-pack module features:• Embedded Gate Drive Circuit.• Protection Logic with Fault-Operation output signal.• Short-Circuit detection (SCD).• Over-Temperature Detection (OTD) using On-Chip Temperature Sensing.• Under-Voltage Detection (UVD).• Low losses level.The advantages presented by this module are the built-in control circuits which provideoptimal gate drive for the switches, improving overall performance in terms of efficiencyand protection for the IGBTs and free-wheeling diodes. The combined integrated structureinverter + chopper allows for full control in most motor drive applications.444.2. Platform DescriptionIPM parametersIGBT Inverter SectorCollector-Emitter Voltage (Vce) 600VCollector-Emitter Saturation Voltage (Vce(sat)) 1.75VCollector Current (Ic) 300APeak Collector Current (Icp) 600ADiode Forward Voltage (Vec) 1, 7VPWM Input Frequency (fsw) 20kHzCollector Dissipation (PC) 833WIGBT Brake SectorCollector-Emitter Voltage (Vce) 600VCollector Current (Ic) 150APeak Collector Current (Icp) 300ADiode Forward Voltage (If ) 150APWM Input Frequency (fsw) 20kHzCollector Dissipation (PC) 520WIn terms of flexibility, several 6-pack and 7-pack IPM modules with different power ratesand the same hardware connection interface are available which can be interchanged andconnected to the power board.Table 4.2.1 presents the main characteristics of the IPM:4.2.2 Interface DesignTo drive the IPM, isolated power sources need to be supplied to each of the gate drive cir-cuits. 3 independent 1.5W DC-DC converters (V LA106 24151) are used to supply the 3 topswitches in the inverter. The 4 bottom switches, including the brake chopper, share com-mon ground, which allows for the use of a single 4.5W DC-DC converter (V LA106 24154).VLA606-01R is an interface module for IPM driving which contains isolation optocouplers.A differential line receiver is used in the power driver to convert the differential signals sentfrom the digital control into single-end signals. An inverting buffer is placed to fix the logicof the circuits and to provide the level of current needed to correctly drive the optodrivers.– Current measurement: for current measurement the LEM HASS-100S was selected,which provides a primary current measurement range of 300A. In a 3-phase system with454.2. Platform DescriptionPARAMETER VALUEOper.volt. 230VAmps(NL) 0.9APo(max) 0.25HpFreq. 0− 60HzMax speed 5400RPMTable 4.1: AC motor parametersa floating neutral point only 2-phase currents would be required to be sensed. For safetyissues, 3-phase current measurements are set, providing redundancy to the system whichallows for the detection of unbalanced and/or faulty operating conditions. Current in thebrake switch is also measured to aid in the control of the dynamic braking.– Voltage measurement: resistive voltage divider for 3-phase and bus voltage was imple-mented on the power board. This is a cost effective solution which allows voltage measure-ment in the inverter, providing the proper analog stage in the control board for filtering andsignal scaling.– Temperature measurement: the system has two temperature measurement points. Thefirst is embedded in the power module and is used for the internal protection circuitry. Toenable temperature monitoring in the control board, a second sensor based on a LM35is placed in the heat-sink plate. Signal from the sensor is sent to the control board fortemperature tracing.The interface board was designed in Altium Designer and the 3D render is shown in Open-Loop TestTesting of the high-power platform was done by implementing a scalar Volts-per-Hertz con-trol [41]. This technique provides variable speed control of an AC machine with a simpleimplementation in open-loop mode. The essence of the technique, which is briefly describedbelow, lies in the regulation of the input phase voltage according to the operating rotatingfrequency so that the voltage-to-frequency ratio is kept constant. In this way flux linkage is464.2. Platform DescriptionFigure 4.3: 3D design of the IPM interface board.maintained, avoiding saturation.If the mechanical system is externally loaded, and does not exceed the nominal torque,the angular speed can be calculated from the supply voltage frequency:ωs =2pifsp (4.1)With p = num.ofpoles The RMS value of the induced voltage Ef,ph is given by thefollowing:474.2. Platform DescriptionFigure 4.4: Platform test setup.Ef,ph =√2pifsNskwφ (4.2)If the stator resistance is small enough to be neglected, then the stator phase voltage Vs,phis equal to the induced voltage Ef,ph, and the flux linkage can be approximated as follows:φ = Vs,ph√2pifsNskw= kVs,phfs(4.3)The control technique was implemented in C28x Delfino MCU to control the AC motorwhose characteristics are depicted in Table 4.1. Fig. 4.5 shows the 3-phase current fed to themotor with no load running at 60Hz 1204poles = 1800RPM during the platform test. The powerplatform operated satisfactorily over the tests developed, confirming the interface design.484.3. Summaryphase Aphase Bphase Cfe≈60Hz1800RPMFigure 4.5: Experimental results show the 3-phase stator currents for the 0.25HP AC induc-tion motor with no load running at the nominal speed of 1800RPM.4.3 SummaryThis chapter presented the design, construction and testing of the 3-phase high-power plat-form for electric motor driving. The platform operates with an Intelligent Power Module(IPM) from Powerex to manage power flow. The board designed to interface with the powermodule integrates voltage, current and temperature measurements and connects with severalIPMs modules, giving flexibility to adapt the power module according to the development.Testing of the platform was done in a 0.5HP AC induction motor drive controlled withVoltz-per-Hertz control technique and satisfactory results were obtained.To provide position/speed sensing a resolver interface was designed composed of a high-current driver to excite the rotor coil and two analog stages for the quadrature stator signals.49Chapter 5Conclusions5.1 SummaryThis thesis investigated the unstable effect of constant power loads in power converters andproposed the implementation of the Circular Switching Surfaces (CSS) for controlling theBCDU in electric vehicle systems in order to mitigate instability behavior.CPL dynamics were investigated in a Buck+Boost cascade converter further implementedas the BCDU, interfacing with the energy storage and the HV DC link. Asynchronous andsynchronous operating modes were investigated, identifying this last as the worst case oper-ating condition, where oscillations grow without boundaries. As was shown with simulationas well as experimental results, the simple and practical geometric technique presented asolution for the challenging combination of high-bandwidth CPL and this unbounded oscil-lation due to synchronous operation of the converter, and yielded large-signal stability andenhanced dynamic transient response over a wide range of operating conditions required bythe targeted application.Comparison with state-of-the-art linear and non-linear controllers showed a transient re-covery improvement of up to 3.4 times and a considerable reduction in switching actions,thereby improving efficiency. The control laws based on the CSS were derived from a normal-ized converter, generalizing the application. Furthermore, the analysis led to a theoreticalframework in which predictable transient performance can be achieved.The application of this control technique enables a reduction in the size of the bulkyDC-link capacitance usually employed to stabilize the main bus. This allows the usage of505.2. Future Workmetal-film capacitors, which have reliability advantages over commonly employed electrolyticcapacitors as well as reduced ESR, improving system efficiency. Reduction of the capacitanceis very important as well since it allows for reductions in size, weight and finally the cost ofthe final implementation.Experimental results were developed in a scaled 1kW Buck+Boost cascade converter tovalidate the concepts described and the application of the CCS for CPL transient rejectionand steady-state operation.The contributions of this thesis are supported by publication in IEEE APEC 2014 [1] andit is currently submitted under review with minor changes to IEEE Transactions on PowerElectronics [2].5.2 Future WorkBCDU are bidirectional power converters providing power flow to the traction system andrecovering energy to the batteries through regenerative braking. The work developed inthis thesis covered the motoring operation of the BCDU in which energy is delivered fromthe batteries to the traction systems and to the rest of the EV power system. To controlthe energy recovered from the electric motor to the battery set, a different control strategyis required that will shape the input current to the battery according to the SOC. Theregenerative operation needs to be combined with dynamic braking due to the limitation inthe amount of energy that can be stored in the batteries.During this thesis work a high-power BCDU and motor drive platform were designed andconstructed and requires integration to conform a high-power bidirectional motor drive plat-form for the development and testing of control techniques for energy management in EVs.This would allow to develop high performance control algorithms to achieve comprehensiveregenerative braking as well as precision control of electric torque of the drive train.51Bibliography[1] M. Anun, M. Ordonez, I. Galiano and G. Oggier, “Bidirectional Power Flow with Con-stant Power Load in Electric Vehicles: A Non-linear Strategy for Buck+Boost CascadeConverters,” in Proc. IEEE Appl. Power Electron. Conf. (APEC) 2014, Fort Worth, TX,2014.[2] M. Anun, M. Ordonez, I. Galiano Zurbriggen, G. Oggier, “Circular Switching SurfaceTechnique: High Performance Constant Power Load Stabilization for Electric VehicleSystems,” submitted under review with major changes to IEEE Transactions on PowerElectronics, March 2014.[3] T. Trigg and P. Telleen, “Global EV Outlook: Understanding the Electric Vehicle Land-scape to 2020,” IEA, Paris, France, 2013.[4] D. Anair and A. Mahmassani, “State of CHARGE: Electric Vehicles Global WarmingEmissions and Fuel-Cost Savings across the United States,” UCS, Cambridge, MA, Rep.02138-3780, 2012.[5] H. Van Essen and B. Kampman, “Impacts of Electric Vehicles Summary report,” CEDelft, Delft, Netherlands, Rep. 11.4058.26, 2011.[6] A. Emadi, M. Ehsani, and J. M. Miller, “Automotive Power Systems,” in VehicularElectric Power Systems: Land, Sea, Air, and Space Vehicles., New York: Marcel Dekker,Dec. 2003.[7] Z. J. Shen, “Optimal Power Management and Distribution in Automotive Systems,”52Bibliographyin Handbook of Automotive Power Electronics and Motor Drives, A. Emadi Ed., BocaRaton, FL: CRC Press, 2005.[8] O. Hegazy, R. Barrero, J. Van Mierlo, P. Lataire, N. Omar, T. Coosemans, “An AdvancedPower Electronics Interface for Electric Vehicles Applications,” IEEE Trans. Power Elec-tron., vol. 28, no. 12 pp. 5508-5521, Dec. 2013.[9] N. A. Chaturvedi, R. Klein, J. Christensen, J. Ahmed, A. Kojic, “Algorithms for Ad-vanced Battery-Management Systems,” IEEE Control. Syst. Mag., vol. 30, no. 3, pp.49-68, June 2010.[10] S. Waffler and J.W. Kolar, “A Novel Low-Loss Modulation Strategy for High-PowerBidirectional Buck+Boost Converters,” Power Electronics, IEEE Transactions on ,vol.24, no.6, pp. 1589-1599, June 2009.[11] A. Khaligh, “Stability Criteria for the Energy Storage Bi-directional Dc/Dc Converterin the Toyota Hybrid System II,” in Proc. IEEE Veh. Power and Prop. Conf. (VPPC)2007, pp. 348-352, Sep. 2007.[12] A. Emadi, S. S. Williamson, and A. Khaligh, “Power Electronics Intensive SolutionsFor Advanced Electric, Hybrid Electric, and Fuel Cell Vehicular Power Systems,” IEEETrans. Power Electron., vol. 21, no. 3, pp. 567-577, May. 2006.[13] R. M. Schupbach and J. C. Balda, “Comparing DC-DC Converters For Power Manage-ment in Hybrid Electric Vehicles,” in Proc. IEEE Elec. Mach. and Drives Conf. (IEMDC)2003, vol.3, pp. 1369-1374, June 2003.[14] F. Caricchi, F. Crescimbini, F. Giulii Capponi, and L. Solero, “Study of Bi-DirectionalBuck-Boost Converter Topologies for Application in Electrical Vehicle Motor Drives,” inProc. IEEE Appl. Power Electron. Conf. (APEC) 1998, vol.1, pp. 287-293, Feb. 1998.53Bibliography[15] F. Caricchi, F. Crescimbini, G. Noia, and D. Pirolo, “Experimental Study of a Bidirec-tional DC-DC Converter for the DC Link Voltage Control and Regenerative Braking inPM Motor Drives Devoted to Electrical Vehicles,” in Proc. IEEE Appl. Power Electron.Conf. (APEC) 1994, vol.1, pp. 381-386, Feb. 1994.[16] F. Caricchi, F. Crescimbini, A. Di Napoli, and M. Marcheggiani, “Prototype of ElectricVehicle Drive With Twin Water-Cooled Wheel Direct Drive Motors,” in Proc. IEEEPower Electron. Spec. Conf. (PESC) 1996, pp. 1926-1932 , June 1996.[17] Y. A-R I. Mohamed, A.A.A. Radwan, and T.K. Lee, “Decoupled Reference-Voltage-Based Active DC-Link Stabilization for PMSM Drives With Tight-Speed Regulation,”IEEE Trans. Ind. Electron., vol. 59, no. 12, pp. 4523-4536, Dec. 2012.[18] X. Y. Liu, A. J. Forsyth, “Comparative Study of Stabilizing Controllers for BrushlessDC Motor Drive Systems,” in Proc. IEEE Elect. Mach. and Drives Conf. (IEMDC) 2005,pp. 1725-1731, May 2005.[19] A. Emadi, A. Khaligh, C. H. Rivetta, and G. A. Williamson, “Constant Power Loadsand Negative Impedance Instability in Automotive Systems: Definition, Modeling, Sta-bility, and Control of Power Electronic Converters and Motor Drives,” IEEE Trans. Veh.Technol., vol. 55, no. 4, pp. 1112-1125, July 2006.[20] A. M. Rahimi and A. Emadi, “An Analytical Investigation of DC/DC Power ElectronicConverters With Constant Power Loads in Vehicular Power Systems,” IEEE Trans. Veh.Technol., vol. 58, no. 6, pp. 2689-2702, Jul. 2009.[21] M. Cespedes, Lei Xing, and Jian Sun, “Constant-Power Load System Stabilization byPassive Damping,” IEEE Trans. Power Electron., vol. 26, no. 7, pp. 1832-1836, Jul. 2011.[22] A. M. Rahimi and A. Emadi, “Active Damping in DC/DC Power Electronic Converters:A Novel Method to Overcome the Problems of Constant Power Loads,” IEEE Trans. Ind.Electron., vol. 56, no. 5, pp. 1428-1439, May 2009.54Bibliography[23] A. M. Rahimi, G. A. Williamson, and A. Emadi, “Loop-Cancellation Technique: ANovel Nonlinear Feedback to Overcome the Destabilizing Effect of Constant-PowerLoads,” IEEE Trans. Veh. Technol., vol. 59, no. 2, pp. 650-661, Feb. 2010.[24] P. Magne, B. Nahid-Mobarakeh, S. Pierfederici, “General Active Global Stabilizationof Multiloads DCPower Networks,” IEEE Trans. Power Electron., vol. 275, no. 4, pp.1788-1798, Apr. 2012.[25] C. N. Onwuchekwa and A. Kwasinski, “Analysis of Boundary Control for Buck Con-verters With Instantaneous Constant-Power Loads,” IEEE Trans. Power Electron., vol.25, no. 8, pp. 2018-2032, Aug. 2010.[26] C. N. Onwuchekwa and A. Kwasinski, “Analysis of Boundary Control for Boost andBuck-Boost Converters in Distributed Power Architectures with Constant-Power Loads,”in Proc. IEEE Appl. Power Electron. Conf. (APEC) 2011, pp. 6-11, March 2011.[27] Y. Zhao, W. Qiao, D.Ha, “A Sliding-Mode Duty-Ratio Controller for DC/DC BuckConverters with Constant Power Loads,” IEEE Trans. Ind. Appl., vol. 50, no. 2, pp.1448-1458, Mar. 2014.[28] R. Munzert, P. T. Krein, “Issues in Boundary Control [of Power Convertors],” in Proc.IEEE Power Electron. Spec. Conf. (PESC) 1996, pp. 810-916, June 1996.[29] M. Ordonez, M. T. Iqbal, and J. E. Quaicoe, “Selection of a Curved Switching Surfacefor Buck Converters,” IEEE Trans. Power Electron., vol. 21, no. 4, pp. 1148-1153, Jul.2006.[30] K. K. S. Leung and H.S.-H Chung, “Derivation of a Second-Order Switching Surface inthe Boundary Control of Buck Converters,” IEEE Power Electron. Lett., vol. 2, no. 2,pp. 63-67, Jun. 2004.55Bibliography[31] T. Siew-Chong, Y. M. Lai, ; C. K. Tse, M. K. H.Cheung, “A Fixed-FrequencyPulsewidth Modulation Based Quasi-Sliding-Mode Controller for Buck Converters,”IEEE Trans. Power Electron., vol. 20, no. 6, pp. 1379-1392, Nov. 2005.[32] J. M. Galvez, M. Ordonez, F. Luchino, and J. E. Quaicoe, “Improvements in BoundaryControl of Boost Converters Using the Natural Switching Surface,” IEEE Trans. PowerElectron., vol. 26, no. 11, pp. 3367-3376, Nov. 2011.[33] M. Ordonez, J. E. Quaicoe, and M. T. Iqbal, “Advanced Boundary Control of InvertersUsing the Natural Switching surface: normalized geometrical derivation,” IEEE Trans.Power Electron., vol. 23, no. 6, pp. 2915-2930, Nov. 2008.[34] G.G. Oggier, M. Ordonez, J. M. Galvez, F. Luchino, “Fast Transient Boundary Controland Steady-State Operation of the Dual Active Bridge Converter Using the NaturalSwitching Surface,” IEEE Trans. Power Electron., vol.29, no.2, pp. 946-957, Feb. 2014.[35] M. Salcone and J. Bond, “Selecting Film Bus Link Capacitors for High PerformanceInverter Applications,” in Proc. IEEE Elect. Mach. and Drives Conf. (IEMDC) 2009,pp. 1692-1699, May 2009.[36] L. Gu, X. Ruan, M. Xu, and K. Yao, “Means of Eliminating Electrolytic Capacitor inAC/DC Power Supplies for LED Lightings,” IEEE Trans. Power Electron., vol. 24, no.5, pp. 1399-1408, 2009.[37] K. Yao, X. Ruan, X. Mao, and Z. Ye, “Reducing Storage Capacitor of a DCM BoostPFC Converter,” IEEE Trans. Power Electron., vol. 27, no. 1, pp. 151-160, Jan. 2012.[38] T. Messo, J. Jokipii, J. Puukko, T. Suntio, “Determining the Value of DC-Link Capac-itance to Ensure Stable Operation of a Three-Phase Photovoltaic Inverter,”IEEE Trans.Power Electron., vol.29, no.2, pp. 665-673, Feb. 2014.56Bibliography[39] H. Wang, H. S. -H. Chung and W. Liu, “Use of a Series Voltage Compensator forReduction of the DC-Link Capacitance in a Capacitor-Supported System,” IEEE Trans.Power Electron., vol. 29, no. 3, pp. 1163-1175, Mar. 2014.[40] R. D. Middlebrook, “Input Filter Considerations in Design and Application of SwitchingRegulators,” in Proc. IEEE Ind. Appl. Soc. Annu. Meet., 1976, pp. 366382.[41] P. C. Krause, O. Wasynczuk, and S. D. Sudhoff, “Induction Motor Drives,” in Automo-tive Power Systems., 2nd Ed. Piscataway, NJ: Wiley-Interscience , 2002.57


Citation Scheme:


Citations by CSL (citeproc-js)

Usage Statistics



Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            async >
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:


Related Items