BUCK TYPE PWM RECTIFIER.pdf

1412 IEEE TRANSACTIONS ON INDUSTRY APPLICATIONS, VOL. 38, NO. 5, SEPTEMBER/OCTOBER 2002

A Unity-Power-Factor Buck-Type PWM Rectifier forMedium/High-Power DC Motor Drive Applications

Hazım Faruk Bilgin, Member, IEEE, K. Nadir Köse, Gürkan Zenginobuz, Muammer Ermis¸, Member, IEEE,Erbil Nalçacı, Isık Çadırcı, Member, IEEE, and Hasan Köse

Abstract—This paper describes the application of a single-stageunity-power-factor buck-type pulsewidth modulation (PWM) rec-tifier to medium- and high-power variable-speed dc motor drives.The advantages of the developed system are low harmonic dis-tortion in ac supply currents (complying with IEEE Std. 519and IEC 555), nearly unity power factor over a wide oper-ating shaft speed range, and nearly level armature current andvoltage waveforms. These properties of output voltage and currentquantities of the converter eliminate entirely any failure risk incurrent commutation even for oldest motor designs, and furthermotor problems such as accelerated aging in motor insulation,and mechanical failure due to circulating bearing currents. Thedesign criteria and operating features and characteristics of thebuck-type PWM rectifier employed in a unity-power-factor dcmotor drive are discussed in the paper. The performance ofthe resulting system has been tested on 25–65-kW 100–150-A10–20-kHz dc motor drives, used in various processes of an ironand steel plant.

Index Terms—DC motor drives, power quality, pulsewidthmodulation, rectifiers.

I. INTRODUCTION

I N RECENT YEARS, increasing emphasis on power qualityhas directed researchers toward proposing and developing

inherently clean new power converter topologies. These con-verters operate at nearly unity power factor (PF), inject very lowharmonic content into the supply, and work at relatively highconverter efficiencies.

Several unity-PF rectifier topologies such as the buck, boost,and buck–boost-derived topologies with or without isolationhave been proposed to date. Their performances have beenreported in the literature [1]–[12]. Among these, the unity-PFbuck-type pulsewidth modulation (PWM) rectifier is oneof the most common topologies studied in detail over thepast ten years [2], [4]–[11]. These studies deal only with the

Paper IPCSD 02–037, presented at the 2001 Industry Applications SocietyAnnual Meeting, Chicago, IL, September 30–October 5, and approved for pub-lication in the IEEE TRANSACTIONS ONINDUSTRY APPLICATIONSby the Indus-trial Power Converter Committee of the IEEE Industry Applications Society.Manuscript submitted for review August 1, 2001 and released for publicationJune 7, 2002.

H. F. Bilgin, M. Ermis, E. Nalçacı, and I. Çadırcı are with theTÜBITAK-METU Information Technologies and Electronics ResearchInstitute, TR06531 Ankara, Turkey, and also with the Electrical and Elec-tronics Engineering Department, Middle East Technical University, TR06531Ankara, Turkey (e-mail: [email protected]; [email protected];[email protected]; [email protected]).

K. N. Köse and G. Zenginobuz are with the TÜBITAK-METU InformationTechnologies and Electronics Research Institute, TR06531 Ankara, Turkey.

H. Köse is with the ISKENDERUN Iron and Steel Plant, TR31319 Hatay,Turkey.

Publisher Item Identifier 10.1109/TIA.2002.803005.

application of unity-PF buck-type PWM rectifier to passiveloads [2], [4]–[6], [8]–[10] except [11], which investigatesthe converter–dc motor combination. Unfortunately, there arelimited publications on the use of this converter in practicalapplications [7].

Many high-frequency PWM techniques, such as the modifiedsinusoidal PWM [2], space-vector modulation [4], and deltamodulation [10] techniques have been proposed, and imple-mented for the single-stage three-phase buck-type PWM recti-fier. The operating principles and design criteria for buck-typePWM rectifier using modified sinusoidal PWM technique aregiven in [2]. A modified sinusoidal PWM-technique-basedalgorithm which is employing a separate control loop forcompensation of input current displacement factor in steadystate and output voltage regulation has been developed in [5].The PWM technique using a space-vector modulation has beendeveloped and implemented for buck-type PWM rectifiers in[4]. Isolated buck-type PWM rectifiers have been investigatedin [6] and [8], and zero-voltage switching for these types hasbeen realized in [6].

Buck-type PWM rectifier offers a good solution for directconversion of ac to dc at high power densities to meet the strictPF penalty limits imposed by electricity authorities and inputline current harmonic distortion limits dictated by various har-monic standards such as IEEE Std. 519, IEC 555, etc. One inter-esting application of this converter may be the upgrading workthat can be conducted on dc motor drives still working in in-dustry to comply with present power quality regulations. Thisconverter also offers superior output characteristics, especiallyfor old dc motor designs when combined with a simple and,hence, cheap high-frequency output filter.

This paper deals with the design and implementation of athree-phase buck-type PWM rectifier for medium/high-powerdc motor drive applications. Design criteria, important featuresof the resulting dc motor drives, and operating characteristicsof the buck-type PWM rectifier in variable-speed dc motor ap-plications are given in the paper. Successful performance fig-ures for the resulting dc motor drives have been obtained inthe field. This upgrading work has been applied to various pro-cesses of the Iskenderun Iron and Steel Plant, Hatay, Turkey,resulting in minimized downtimes, reduced periodic inspectionand maintenance work, and increased conversion efficienciessince mid-2000.

II. PROBLEM DEFINITION

In this paper, the three-phase unity-PF buck-type PWMrectifier has been applied to medium-power adjustable-speed

0093-9994/02$17.00 © 2002 IEEE

BILGIN et al.: UNITY-POWER-FACTOR BUCK-TYPE PWM RECTIFIER 1413

Fig. 1. Circuit diagram of the unity-PF buck-type PWM rectifier for separately excited dc motor application.

dc motor drives. The resulting new-generation single-quadrant(1-Q) and reversible two-quadrant (2-Q) dc motor drives havebeen applied to various processes of the Iskenderun Iron andSteel Plant within the scope of modernization work, such as thesintering process and the raw material charging process (seethe Appendix). The realization of this modernization work wasfound to be inevitable in 1998 owing to the following reasons:

• new penalty limits for PF imposed by electricity authori-ties (0.95 lag–0.98 lead on monthly basis);

• to comply with the present harmonic regulations in thecountry which are nearly the same with those of IEEE Std.519;

• the need for more efficient use of electricity in iron andsteel manufacturing in order to be able to compete in themarket.

At first sight, the use of variable-frequency ac motor drivesemploying squirrel-cage induction motors seems to be a feasiblesolution to this problem. However, a peer feasibility studyhas shown that preserving the present separately excited dcmotors, and upgrading their converters and control systemsis the most economical solution in comparison with the useof variable-frequency ac drive technology, for the followingreasons.

• The present dc motors are all in a good condition, andmany spare parts are available in stock, besides experi-enced personnel for maintenance and repair.

• Significant modifications in the mechanical parts andinstallation work are needed for the replacement ofpresent dc motors with squirrel-cage induction motors,even though relatively expensive 12-, eight-, and six-poleinduction machines for speed matching purposes wereused.

• The conventional dc drive continues to play a dominantrole when dynamic drives with constant load torque andstringent requirements for overload withstand capabilitythroughout a large speed setting range are involved. This isespecially important in the sintering band where the palletstend to be blocked frequently.

• The use of frequency-controlled variable-speed driveswould further necessitate special converter type ac mo-tors, or costly output filter systems in order to get rid ofhigh insulation stresses on the motor, causing acceleratedaging and insulation failures in the long term [15]–[17].

In order to meet the objectives of the project, the present dcmotor drives are upgraded by replacing primarily the magneticamplifiers, and Ward Leonard Systems, secondarily old versionsof six-pulse line-commutated bridge converters with unity-PFbuck-type PWM rectifiers (Fig. 1). The idea behind the choiceof the unity-PF buck-type PWM rectifier instead of modern ver-sions of thyristorized line-commutated ac–dc converters is asfollows. For the vast majority of operation time, these motors arerunning at speeds less than half of the rated value and, hence, atarmature voltages less than 250 Vfrom the available 0.4-kV


line-to-line 50-Hz ac supply. This means that a six-pulse thyris-torized ac–dc converter, if chosen, would operate at very highvalues of firing angle , thus consuming considerable amountsof reactive power from the supply, resulting in low input PF lessthan 0.5 for most of the time.

Furthermore, it injects significant amounts of low-frequencysuperharmonics such as fifth, seventh, and 11th current har-monic components into the supply. Since the variation in reac-tive power consumption is usually fast for some of the dc motordrives operating in iron and steel plants, this choice would makenecessary the use of costly reactive power compensation and fil-tering solutions such as, active power filters or dynamic var gen-erators designed in the form of passive shunt filters. However,the buck-type PWM rectifier chosen inherently operates at PFsnearly unity in a wide speed range, and does not produce anylow-order harmonics such as the fifth, seventh, 11, and 13 timesthe supply frequency. Since the chosen modulation technique isthe sinusoidal PWM, this converter produces harmonics as thesidebands of the switching frequency and its multiples. With thepresent semiconductor technology, switching frequency is in therange of 10–20 kHz in the moderate power range up to 100 kWfor hard switching. This makes necessary the use of a high fre-quency input filter which is smaller in size, has a fixed config-uration, and hence is more reliable, and cheaper in comparisonwith that of a six-pulse line-commutated ac–dc converter.

III. SYSTEM DESCRIPTION

The circuit diagram of the unity-PF buck-type PWM rectifierwhich supplies power to the armature circuit of a separately ex-cited dc motor is as shown in Fig. 1. The block diagram rep-resentations of the control system and speed feedback loopsare also shown in the same figure. Each power semiconductorswitch consists of an IGBT connected in series with an ultrafastrecovery diode, resulting in reverse voltage blocking capabilityand unidirectional current flow. A low-pass dampedLC filteris connected to the input side of the converter to filter out theswitching frequency harmonic components in the line currents.In order to avoid accelerated aging in motor insulation due tosteep voltage wavefronts, and mechanical failure due to circu-lating bearing currents, the switching frequency harmonics arefiltered out with a damped output filter. This problem has beenreported in the last decade only for variable-frequency inductionmotor drives [15]–[17].

In this paper, the modified sinusoidal PWM technique [2] ischosen to construct the switching signals. The switching signalsare generated by a 16-b microcontroller (Hitachi H8S2655) ata fixed frequency. Closed-loop speed control of the dc motor isachieved by digital implementation of a proportional plus inte-gral (PI) control strategy on the same microcontroller which isalso responsible for the unity-PF operation of the system. Twoadditional microcontrollers (NEC 78P058Y) are used for pro-tection, display, and communication purposes.

In the Iskenderun Iron and Steel Plant, since exhauster fansare constant-speed machines, the sintering process is controlledby varying the speed of the sintering strands in the moderateterm. Therefore, in these drives, and straightline coolers, 1-Q dcmotor drives are needed, resulting in speed control action only inthe first quadrant of the speed–torque plane, and a unidirectional

Fig. 2. Operating characteristic of the 2-Q reversible dc motor drives.

constant field excitation. However, boom-turning action in uni-versal machines needs shaft speeds in both directions. There-fore, for these machines, 2-Q reversible dc motor drives havebeen designed by the use of a unidirectional unity-PF buck-typePWM rectifier on the armature side (Fig. 2), and a four-quad-rant (4-Q) dc–dc converter on the field side (Fig. 3). The 4-Qdc–dc converter in Fig. 3 allows both reversing the field cur-rent rapidly, and keeping it constant at the preset value againstsupply voltage fluctuations.

IV. OPERATING PRINCIPLES OFBUCK-TYPE PWM RECTIFIER

IN THE DC MOTOR DRIVE

A. Comparison of PWM Techniques

PWM techniques have been commonly used in voltage-and current-source inverters (VSIs and CSIs, respectively) ofvariable-frequency ac motor drives. The control of unity-PFbuck-type rectifiers is also based on these techniques for lowdistortion in supply currents [2], [4]–[11]. In this paper, PWMtechniques are compared in view of the needs of the dc motordrive application. These are reliability and simplicity, goodtransient response, low modulation index (0–0.5) for mostof the operation time, high input power quality, optimizedswitching losses to make the semiconductor cooling systemimplementable, and optimum input- and output-filter sizeswith no interaction of input filter with the low order harmonicspresent in the supply.

Sinusoidal PWM technique (SPWM) has found a wide rangeof applications since the early development of PWM-VSI tech-nology. Although the control range of modulation index isrelatively narrow [3], [20], SPWM is a simple technique andhas a good transient response [19]. Wider control range ofmodulation index as compared to classical SPWM can beobtained by employing a modified SPWM technique, whichwas originally developed in [2], and can also be found in[3]. Space-vector PWM (SVPWM) has a wide control rangeof modulation index, simple implementation, good transientresponse, and optimum harmonic distortion factor over a widerange of modulation index at the expense of higher switchinglosses and acoustic annoyance [19]. In order to obtain overmod-ulation, Harmonic-injected PWM (HIPWM) can be appliedby injecting harmonics into the reference waveform [3], [20].The harmonic spectrum of the input currents is very similarto those of SPWM and SVPWM, except the contribution ofinjected harmonic component [3], [20]. Less acoustic annoy-ance can be obtained by the application of random PWM


Fig. 3. Operation modes of 4-Q field exciter.

techniques [19]. It distributes switching frequency harmonicsand its multiples to the overall spectrum. However, this resultsin calculation overhead, and higher current ripple in compar-ison with SPWM and SVPWM [19]. In order to reduce theswitching losses of continuous PWM techniques mentionedabove by an amount of 25%–50%, discontinuous PWM tech-niques (DPWM) can be employed [20]. Lower switching lossesin DPWM are at the expense of uncontrollability in the under-modulation region [20]. Although the number of switchings athigh frequency is reduced in DPWM, each switching still yieldshigh and , which are undesirable, especially inhigh-power applications, where gate-turn-off thyristors (GTOs)are commonly and insulated gate-commutated (IGCTs) arenewly used. For such applications, selective harmonic elimi-nation method (PWM-SHEM) [21] is the most suitable one.However, PWM-SHEM suffers from poor transient responseand high computational effort. Delta modulation technique [7],[10] (also known as hysteresis modulation) is very popularin voltage-source converter applications, but it needs onlinemonitoring of input current; moreover, it results in switchingfrequencies even higher than those of SPWM and SVPWM.

In view of the qualitative comparison given above, modifiedSPWM and SVPWM techniques meet largely the needs ofthe application. Simulation studies have shown that modifiedSPWM and SVPWM both yield satisfactory performance inthe operation, and control of a unity-PF buck-type rectifier–dcmotor combination. In a buck-type rectifier application, the twotechniques are very similar in shaping the input ac quantities,the only difference being in the power semiconductor switchingpatterns. The modified SPWM technique has been chosen forimplementation by considering the authors’ past experiencewith SPWM.

Fig. 4. Current-source buck-type rectifier.

B. Operating Principles

Operating principles of the buck-type PWM rectifier in thedc motor drive will be explained using the modified SPWMtechnique. A simplified circuit diagram of a three-phase cur-rent-source buck-type rectifier is shown in Fig. 4.

The switching function of a switch is defined as

ON denotes thesupply line to which theswitch connected

OFF where denotes theupper or lower part of thebridge to which the switchbelongs

(1)

and, at any time, must be satisfied.There are six time intervals, determined by the intersections

of the line-to-neutral voltages ( ) (or zero crosses of


Fig. 5. Six time intervals defined with respect to supply voltage waveforms.

the line-to-line voltages) shown in Fig. 5. In a certain time in-terval, the switch ( ) connected to the most posi-tive phase and the switch ( ) connected to the mostnegative phase during the interval are called “main switches.”The switches connected to the remaining third phase are called“subswitches.”

The most positive and the most negative line-to-neutralvoltage segments in each time interval in Fig. 5 are parts ofsinusoidal waveforms varying either from 30to 90 or from90 to 30 . The SPWM technique has been applied in the usualmanner by comparing a triangular carrier signal with a constantamplitude of (Fig. 6) with sinusoidal reference voltagesegments.

The sinusoidal reference voltage segments are being in phasewith six line-to-neutral voltage segments in Fig. 5. The meanvalue of the output dc voltage is adjusted by varying the ampli-tudes of reference voltage signals in Fig. 6. The switchingfrequency of the rectifier is obviously equal to the frequency ofcarrier signal and must be multiples of six times the supply fre-quency. The switching frequency is adjustable in the range from10 to 20 kHz in the implemented system. The switching func-tions of the main and subswitches in time interval III are givenin Fig. 6, as an example.

Power transfer from the ac supply to the dc motor will takeplace in time intervals where two main switches (one in theupper “u,” and one in the lower “l” parts of the converter), orone main switch and one subswitch are conducting simultane-ously. On the other hand, in the periods where two subswitchesare in conduction, freewheeling action takes place. In additionto freewheeling action of the converter switches, a freewheelingdiode (FWD) can be connected externally to the output of the

Fig. 6. Switching signals of the main- and subswitches in time interval III.

converter for protection purposes against possible overvoltages.This is permissible because in the mentioned application theunity-PF buck-type PWM rectifier is to be operated only inthe first and third quadrants, resulting in always positive outputvoltages. In the normal working range, the sequence of con-ducting semiconductor pairs is therefore main–main, main–sub,and sub–sub, e.g., from the starting point of Interval III,

, , until zero-crossing point of , andthen , , until the end of IntervalIII.

The modulation index is defined as

(2)

where and are the amplitudes of reference and car-rier voltage signals, respectively.

Since the output current is nearly constant within a switchingperiod due to large output inductance, the line currents of therectifier can be expressed as

(3)

where lowercase letters denote instantaneous values of associ-ated currents, and capitalis the mean output current, whichis equal to instantaneous current under the assumption of largeoutput inductance.


Output voltage of the rectifier can be expressed as

(4)

Rectifier input currents and output voltage can then beexpressed as in (5) and (6) [2], [5]

harmonics arisingfrom switchings



(5)

harmonics arising from switchings

(6)

where is the peak of the supply phase voltages.Mean output voltage of the rectifier is the direct compo-

nent of in (6). Mean value can be controlled by changing. It can also be controlled by shifting switching signals in ei-

ther direction with respect to supply voltage waveforms. Whenthe switching signals are shifted by anglewith respect tozero-crossing points of line-to-line voltages, this will reduceby a factor of [2], [5], as given in (7)

(7)

The rectifier input filter in Fig. 4 filters out almost all of thecurrent harmonics arising from switchings and, at the same time,introduces a phase lead to rectifier input currents with respect tosupply phase voltages by an angle of.

The supply line currents can, therefore, be expressed as

(8)

One of the major design objectives for the system in Fig. 4 isto have unity PF at the supply terminals. This makes necessarynot only shifting all switching signals by phase anglewithrespect to line-to-neutral voltage segments, but also adjustingold modulation index to . Then, rectifier line currentsbecome as follows:

harmonicsarising from switchings



(9)

This yields unity PF as can be understood from (10)

(10)

TABLE ITECHNICAL SPECIFICATIONS OFIMPLEMENTED DC DRIVES

where is the new modulation index, as given in (11)

(11)

The mean value of output voltage can then be expressed as

(12)

Equation (12) expresses the fact that, in order to bring theinput PF to unity while keeping constant at the desired value,not only should the switching signals be shifted by phase angle, but also the modulation index calculated from (6) should be

adjusted to . In case of a FWD at the output, shiftingswitching functions beyond 30results in a nonlinear relation-ship between the modulation index and the rectifier outputvoltage, .

V. DESIGN AND IMPLEMENTATION

A. Technical Specifications

The input and output specifications of the designed PWMbuck-type rectifier are as given in Table I.

B. Hardware

Fig. 7 shows the front panel and interior view of the im-plemented 25-kW 100-A 10–20-kHz dc motor drive for thesintering strand.

1) Circuit Layout: The layout of the power circuit hasbeen designed carefully in order to minimize stray inductancesarising from connections. This avoids generation of dangerousovervoltages in the circuit due to high . Custom-builtbusbars are used in the connections between power switchesand input/output filters [Fig. 7(b)].

2) Power Semiconductors:In the power circuit, third-gen-eration IGBTs are used. Ultrafast soft-recovery diodes are to bechosen in such a way that their turn-off time is less than or equalto that of IGBTs in order to avoid overvoltages across them.Both IGBTs and series diodes are subject to peak line-to-linesupply voltage, and carry peak of the rectifier output current.The ratings of all power semiconductors are chosen accordingto these criteria with a safety factor of two.


(a) (b)

Fig. 7. Implemented 25-kW 100-A dc motor drives based on buck-type PWM rectifier. (a) Front view. (b) Interior.

3) Input Filter: Rectifier input current harmonics arisingfrom semiconductor switches are filtered out by a low-passLCfilter tuned to 1 kHz (nearly one-tenth of switching frequency).As recommended in [13], it has been designed as a dampedLCfilter (Fig. 1) in order to eliminate the risk of resonance whichmay arise from coincidence of any supply side harmonic com-ponent with corner frequency. In the damped low-pass filter,maximizing capacitance value and minimizing inductancevalue would give us better filtering of switching frequency.However, a higher filter capacitance means generation ofmore capacitive vars to be compensated by delaying switchingsignals more, for unity-PF operation.

Since a FWD is present across the output terminals of therectifier for protection purposes, the maximum delay angle ofswitching signals has been limited to 30. This will restrict theoperating range in which the input PF can be kept at unity.However, as a design objective, it is intended to keep PF nearlyat unity from 100% to 10% of full load. Therefore, when se-lecting the capacitance value, increase in harmonic distortion inthe supply current must be balanced against reduction in usefulworking range for which PF could be kept at unity.

Optimum filter parameters are: H, F,and for these applications. Typical capacitor and in-ductor current waveforms are as given in Fig. 8.

4) Output Filter: A damped second-order filter is connectedto the output of the rectifier in order to filter out the switchingfrequency harmonics (10–20 kHz) which appear in the output

Fig. 8. Input filter current waveforms, Ch.3: capacitor current (10 mV/A), andCh.4: inductor current (10 mV/A).

voltage and, hence, across the armature terminals. These high-frequency harmonic voltage components cause stresses on thearmature insulation and induce currents through bearing due tohigh . The output filter is tuned to 1.5 kHz in order toobtain an attenuation of the switching-frequency ripple voltagenearly to its 10%. An amorphous metal core (1 mil Metglas) hasbeen chosen in both the input and output filter inductors to min-imize power dissipation and core volume. In view of availablestandard core sizes and costs, the output filter inductance valueis chosen as H, which is sufficiently large to keep


(a)

(b)

Fig. 9. Output filter inductor voltage and current waveforms, Ch.2: voltage,Ch.4: current (10 mV/A). (a) Time base: 2 ms/div. (b) Time base: 100�s.

the output current nearly level, and the corresponding capacitorvalue, as F.

Fig. 9 shows typical voltage and current waveforms of theoutput filter inductor.

5) Cooling System:Each system is placed into IP55 cab-inet furnished with a top-mounted industrial air conditioner asshown in Fig. 7(a) to provide internally closed air circulation.Since the most dissipative elements within the cabinet are semi-conductors, they are mounted on a fan-cooled heat sink. Theheat-sink arrangement is mounted on the rear surface of the cab-inet via sealing accessories in such a way that power semicon-ductors remain within the cabinet, while the fins and cooling fanare at the outside.

C. Controller Implementation

Control, protection, and monitoring tasks of the system areall performed digitally by microcontrollers. The circuit diagramof the controller hardware is given in Fig. 10. Microcontrollershaving most of the necessary peripheral devices embedded(analog-to-digital converter, PWM module, communicationblock, etc.) are preferred for higher noise immunity. PWMoperation is implemented by a 16-b microcontroller (HitachiH8S2655). Monitoring and supervision of the control system

and user interface are carried out by two 8-b microcontrollers(NEC78P058Y). One of them has been designed as the mastercontroller for the overall system including process controlactions. When the whole starting conditions are met, themaster controller initiates the 16-b microcontroller for PWMoperation. All signals except the supply voltages are taken byHall-effect devices. All analog signals (shown in Fig. 10) areconverted to digital form by the 10-b A/D converter embeddedin the microcontroller. Active power, reactive power, and PFangle are calculated by the microcontroller using line-to-linevoltage, and remaining line current .

The block diagram of PWM operation is as given in Fig. 11.Since switching strategy requires two different PWM patternsand their inverses for each of the time intervals in Fig. 5, twoPWM channels are sufficient. One of them compares triangularcarrier wave with a reference sine wave varying from 30to90 , whereas the other makes a comparison with a referencewave changing from 90to 30 . Sine values varying from 30to 90 are placed in a lookup table. Duty cycle register of PWMoperation is updated in 30-s steps in order to maintain the de-sired calculation accuracy for possible range of switching pe-riods changing from 100 to 50s. The frequency of carrierwaveform can be adjusted in the range from 10 to 20 kHz, de-pending upon the practical constraints imposed by loading andclimatic conditions. These PWM patterns P1 and P2 , and theirinversesP1 andP2 are directed to switches by the aid of a switchselector unit as shown in Fig. 10. The 3-b digital signal S0 S1 S2in Fig. 10 defines one of the six time intervals in Fig. 5 and thetruth table. The truth table for the switch selector unit is given inTable II. The hardware of the switch selector unit designed as acombinational logic circuit determines the main and subswitchpairs which should receive PWM patterns P1, P2, and their in-versesP1 andP2, for each time interval selected by S0 S1 S2signal. During operation, upon detection of a new line-to-linezero voltage crossing, PWM patterns are directed to the nextswitch group, in the sequence given in the truth table. The switchselector unit can be disabled by the system supervision and con-trol unit, and the microcontroller itself, in order to protect thedrive and the load.

Fig. 12 shows the flowchart of the main program which is re-sponsible for the operation of the system. The flowchart of themajor subroutines is also included in the same figure. There aretwo different PI controllers in the overall control system: speedand current PI controllers. The speed PI controller generates aset value of the armature current by comparing the shaft speedsignal with speed reference. The shaft speed signal can be takeneither directly from the tachogenerator or computed by usingarmature resistance voltage-drop compensation technique. Thecurrent controller generates modulation index by comparing theactual armature current signal with the reference signal gen-erated by the speed controller. A correction in the modulationindex computed in this way is needed by taking into account thePF at the supply side of converter according to (11). Deviation ofPF from the unity is compensated by shifting whole switchingpattern and switch selection instants in time. Phase shifting islimited to 30 in both directions owing to the presence of theFWD. PI-type speed and current controllers are implemented in


Fig. 10. Controller hardware.

Fig. 11. Block diagram of control system.

TABLE IITRUTH TABLE OF SWITCH SELECTOR

discrete time within the software. The current PI controller up-dates its output to change modulation index in each 250s. Itis much shorter than the electrical time constants of dc motors.

For flexible tuning of the parameters, the speed controller up-dates its output at an adjustable time. Since the speed controllerdoes not need to be as fast as the current controller, its minimumupdate time is chosen as 1.66 ms.

VI. EXPERIMENTAL RESULTS

Sixteen unity-PF buck-type PWM rectifier-based dc motordrives have been operating successfully in the plant since mid-2000. Test results obtained from a 25-kW 1Q dc motor driverunning in the sintering strand will be presented in the paper.The supply current and the corresponding line-to-neutral supplyvoltage waveforms during operation of the dc motor drive atnearly half-load and for a switching frequency of 9.6 kHz are asgiven in Fig. 13. The variations in input PF against load are asshown in Fig. 14. Experimental results collected on all installeddc motor drives show that input PF can be kept nearly constant at


Fig. 12. Flowchart of the dc drive.

Fig. 13. Ch.4: supply current (10 mV/A) and Ch.2: line-to-neutral voltage.

unity PF over the operating range from full load to partial loads.At light loads (lower than 10%–20% of full load depending upondrive size), input PF deviates from unity PF in the leading PFrange. This is an expected result because of the 30maximumimplementable phase shift for switching signals arising from theoperation of the externally connected FWD.

A sample simultaneous record of rectifier input currentbefore ( ) and after ( ) filtering is given in Fig. 15. Fig. 16shows their harmonic spectra. Harmonics at sidebands ofswitching frequency (9.6 kHz), and multiples of it are apparent

Fig. 14. Variations of input PF with load.

from Fig. 16(a). These harmonics are successfully filtered outas can be understood from the harmonic spectrum in Fig. 16(b).

Fig. 17 shows the output voltage of the rectifier, armaturecurrent, and applied armature voltage waveforms for the sameoperating and switching conditions. After filtering, armaturecurrent and applied armature voltage waveforms are nearlylevel. For old motor designs, if the motor were supplied froma line-commutated ac–dc converter, e.g., a six-pulse thyristorrectifier, sixth voltage harmonic and multiples of it woulddrive their own current harmonic components through thearmature winding, resulting in severe current commutationfailure problems in the commutator, especially at high speeds


Fig. 15. PWM rectifier input current before Ch.3:i (16 mV/A) and afterCh.4:i (10 mV/A) filtering.

(a)

(b)

Fig. 16. Harmonic spectra of supply currents. (a) Before input filter. (b) Afterinput filter.

under full load [18]. This would be eliminated by the use ofa sufficiently large dc choke at the output of the converter.

On the other hand, high ’s and transient peaks on theapplied motor voltage would cause high insulation stresses onthe armature, causing accelerated aging and insulation failuresin the long term. Furthermore, a long cable run between themotor and the converter worsens the situation mentioned above[16], [17]. Fig. 18 shows the rising edge of one applied motorvoltage pulse, even for a short motor cable run (50 m), afterdisconnecting the output filter from the circuit in the field.

Fig. 17. Output characteristics of the PWM rectifier, Ch.2: rectifier outputvoltage, Ch.3: armature terminal voltage (250 V/div), and Ch.4: armaturecurrent (10 mV/A).

Fig. 18. Rising edge of applied motor voltage pulse.

It is worth noting that, even for a 50-m cable run, transientpeak voltage is 750 V (50% higher than the normal value of550 V), and is 3.5 kV/ s. For a longer cable run (250 m),peak voltage reaches 1.2 kV.

VII. CONCLUSIONS

The field experience on the dc motor drives based on aunity-PF buck-type PWM rectifier has shown that the resultingsystem has superior operating characteristics over thyristorizedline-commutated dc motor drives at the expense of highercost and provides a good matching between standard supplyvoltage levels and standard dc motor voltage ratings, thuseliminating the need of a special rectifier transformer. Higherinitial cost is largely offset by eliminating the need of externallyconnected variable reactive power compensation, and relativelylarge harmonic filter sizes, especially in countries where strictharmonic and PF limits are applied.

Furthermore, by the use of a buck-type PWM rectifier in com-bination with a simple and cheap high-frequency output filter,output current and voltage can be approximated to nearly levelwaveforms. This is especially important for dc motor designs


Fig. 19. Application points of implemented dc motor drives based on buck-type PWM rectifier in sintering process of Iskenderun Iron and Steel Plant.

Fig. 20. (a) Universal machine, used for charging raw materials from the harbor to the sintering process. (b) DC motors, for turning the boom of the universalmachine. (c) Application point of the new-generation dc motor drive for boom-turning action of the universal machine.

in order to prevent these motors from the risk of commutationfailure at high speeds under load, insulation failure due to high

, and mechanical failure due to bearing currents.

The use of a unity-PF buck-type PWM rectifier in dc motorapplications introduces some opportunities such as the possi-bility of 4-Q operation, lower input current harmonic distortion,


etc., over the use of other known converter topologies. Withthe present IGBT technology, nearly 500-kW units seem to beimplementable.

APPENDIX

See Figs. 19 and 20.

REFERENCES

[1] R. Ridley, S. Kern, and B. Fuld, “Analysis and design of a wide inputrange power factor correction circuit for three-phase applications,” inProc. IEEE APEC’93, 1993, pp. 299–305.

[2] L. Malesani and P. Tenti, “Three-phase AC/DC PWM converter withsinusoidal AC currents and minimum filter requirements,”IEEE Trans.Ind. Applicat., vol. IA-23, pp. 71–77, Jan./Feb. 1987.

[3] M. A. Boost and P. D. Ziogas, “State-of-the-art carrier PWM techniques:A critical evaluation,”IEEE Trans. Ind. Applicat., vol. 24, pp. 271–280,Mar./Apr. 1988.

[4] B. H. Kwon and B. Min, “A fully software-controlled PWM rectifierwith current link,” IEEE Trans. Ind. Electron., vol. 40, pp. 355–363,June 1993.

[5] S. Hiti, V. Vlatkovic, D. Borojevic, and F. C. Lee, “A new control algo-rithm for three phase PWM buck rectifier with input displacement factorcompensation,”IEEE Trans. Power Electron., vol. 9, pp. 173–180, Mar.1994.

[6] V. Vlatkovic, D. Borojevic, and F. C. Lee, “A zero-voltage switched,three phase isolated PWM buck rectifier,”IEEE Trans. Power Electron.,vol. 10, pp. 148–157, Mar. 1995.

[7] D. Ciscato, L. Malesani, L. Rossetto, P. Tenti, G. L. Basile, M. Pasti,and F. Voelker, “PWM rectifier with low dc voltage ripple for magnetsupply,” IEEE Trans. Ind. Applicat., vol. 28, pp. 414–420, Mar./Apr.1992.

[8] J. W. Kolar, U. Drofenik, H. Ertl, and F. C. Zach, “VIENNA RectifierIII—A novel three phase single-stage buck derived unity power factorAC-to-DC converter system,” inConf. Rec. NORPIE/98, Aug. 1998, pp.14.1–16.

[9] Y. Sato and T. Kataoka, “A current-type PWM rectifier with activedamping function,”IEEE Trans. Ind. Applicat., vol. 32, pp. 533–541,May/June 1996.

[10] B. M. M. Rwinyiwiwa, P. M. Birks, and B. T. Ooi, “Delta-modulatedbuck-type PWM converter,”IEEE Trans. Ind. Applicat., vol. 28, pp.552–557, May/June 1992.

[11] A. T. Islier, “Three phase PWM rectifier driving a DC motor with mini-mized input harmonics,” M.S. thesis, Elect. Electron. Eng. Dept., MiddleEast Tech. Univ., Ankara, Turkey, July 1994.

[12] W. Kolar and H. Ertl, “Status of the techniques of three phase rectifiersystems with low effects on the mains,” inProc. 21st INTELEC, June1999, pp. 9–18.

[13] V. Vlatkovic, D. Borojevic, and F. C. Lee, “Input filter design for powerfactor correction circuits,”IEEE Trans. Power Electron., vol. 11, pp.199–205, Jan. 1996.

[14] IEEE Recommended Practices and Requirements for Harmonic Controlin Electrical Power Systems, IEEE Std. 519–1992, 1993.

[15] J. M. Erdman, R. J. Kerkman, D. W. Schlegel, and G. L. Skibinski, “Ef-fect of PWM inverters on AC motor bearing currents and shaft voltages,”IEEE Trans. Ind. Applicat., vol. 32, pp. 250–259, Mar./Apr. 1996.

[16] S. Bell and J. Sung, “Will your motor insulation survive a newadjustable-frequency drive?,”IEEE Trans. Ind. Applicat., vol. 33, pp.1307–1311, Sept./Oct. 1997.

[17] A. H. Bonnett, “A comparison between insulation systems availablefor PWM-inverter-fed motors,”IEEE Trans. Ind. Applicat., vol. 33, pp.1331–1341, Mar./Apr. 1997.

[18] C. E. Robinson, “Redesign of DC motors for applications with thyristorpower supplies,”IEEE Trans. Ind. Gen. Applicat., vol. IGA-4, pp.508–514, Nov. 1965.

[19] M. M. Bech, F. Blaabjerg, and J. K. Pedersen, “Random modulationtechniques with fixed switching frequency for three phase power con-verters,”IEEE Trans. Power. Electron., vol. 15, pp. 753–761, July 2000.

[20] A. M. Hava, R. J. Kerkman, and T. A. Lipo, “High-performance gener-alized discontinuous PWM algorithm,”IEEE Trans. Ind. Applicat., vol.34, pp. 1059–1071, Sept./Oct. 1998.

[21] H. R. Karshenas, H. A. Kojori, and S. B. Dewan, “Generalized tech-niques of selective harmonic elimination and current control in currentsource inverter/converter,”IEEE Trans. Power Electron., vol. 10, pp.566–573, Sept. 1995.

Hazım Faruk Bilgin (M’98) received the B.Sc.and M.Sc. degrees in electrical and electronicsengineering in 1998 and 2000, respectively, fromMiddle East Technical University, Ankara, Turkey,where he is currently working toward the Ph.D.degree.

He is also a Research and Development En-gineer with the TUBITAK-METU InformationTechnologies and Electronics Research Institute,Ankara, Turkey. His areas of research are powerconverters for power quality and electric motor drive

applications.

K. Nadir Köse received the B.S. degree in electricaland electronics engineering from Middle EastTechnical University, Ankara, Turkey, in 1993 andthe M.S. degree from Hacettepe University, Ankara,Turkey, in 2001.

He is a Senior Researcher with theTUBITAK-METU Information Technologies andElectronics Research Institute, Ankara, Turkey.His areas of research are active power filters,boost rectifiers, IGBT-based high-power dc motordrives, synchronous motor field exciters, static var

compensators, electrostatic precipitators, and microcontroller-based controlsystems.

Gürkan Zenginobuz received the B.Sc. and M.Sc.degrees in electrical and electronics engineering in1997 and 2000, respectively, from Middle East Tech-nical University, Ankara, Turkey, where he is cur-rently working toward the Ph.D. degree.

He is also Research and Development Engineerwith the TUBITAK-METU Information Technolo-gies and Electronics Research Institute, Ankara,Turkey. His areas of research are induction motorsoft starters at medium voltage and microprocessorcontrol of motor drives.

Muammer Ermis (M’98) received the B.Sc., M.Sc.,and Ph.D. degrees in electrical engineering fromMiddle East Technical University, Ankara, Turkey,in 1972, 1976, and 1982, respectively, and theM.B.A. degree in production management fromAnkara Academy of Commercial and EconomicSciences, Ankara, Turkey, in 1974.

He is currently a Professor of Electrical Engi-neering at Middle East Technical University andalso the Director of Power Electronics Group,TUBITAK-METU Information Technologies and

Electronics Research Institute, Ankara, Turkey. His current research interestsare high-power medium-voltage motor drives and static reactive powercompensation systems.

Dr. Ermis received “The Overseas Premium” paper award from the Institutionof Electrical Engineers, U.K., in 1992 and the IEEE Industry Applications So-ciety 2000 Committee Prize Paper Award from the Power Systems EngineeringCommittee.

Erbil Nalçacı received the B.Sc., M.Sc., and Ph.D.degrees in electrical and electronics engineering fromMiddle East Technical University, Ankara, Turkey, in1978, 1982, and 1996, respectively.

He has been an Instructor in the Electrical andElectronics Engineering Department, Middle EastTechnical University, for the past ten years. Heis also a part-time Senior Researcher with theTUBITAK-METU Information Technologies andElectronics Research Institute, Ankara, Turkey. Hismain areas of interest are electric arc furnaces and

power electronics for motor drives.


Isık Çadırcı (M’02) received the B.Sc., M.Sc.,and Ph.D. degrees in electrical and electronicsengineering from Middle East Technical University,Ankara, Turkey, in 1987, 1988, and 1994, respec-tively.

She is currently an Instructor in the Electrical andElectronics Engineering Department, Middle EastTechnical University, and also a Senior Researcherwith the TUBITAK-METU Information Technolo-gies and Electronics Research Institute, Ankara,Turkey. Her areas of interest include electric motor

drives and switch-mode power supplies.

Hasan Kösereceived the B.Sc. degree in electricalengineering from the Kocaeli Engineering Faculty,Yildiz Technical University, Istanbul, Turkey, in1982.

Since 1984, he has been a Senior Engineer in theSintering Department, Iskenderun Iron and SteelPlant, Hatay, Turkey.

BUCK TYPE PWM RECTIFIER.pdf

Documents