State of The Art PWM Techniques A Critical Evaluation

Download as pdf or txt
Download as pdf or txt
You are on page 1of 10

IEEE TRANSACTIONS ON INDUSTRY APPLICATIONS, VOL. 24, NO.

2, MARCHIAPRIL 1988

27 1

State-of-the-Art Carrier PWM Techniques: A Critical Evaluation


Abstract-With the introduction and wide acceptance of gate turn-off power devices (e.g., bipolars, power FET's, GTO's, etc.), the switching behavior of converters has reached the point where further improvements in firing and switching networks bring only marginal benefits. Consequently, the research interests in the area of static converters have been shifting toward improving the process of power conversion through a combination of new circuit topologies and improved voltage and harmonic control PWM techniques. As a result, several such techniques have been proposed lately. Although these techniques are clearly superior to the original sine PWM technique, little or conflicting data is available about their merits relative to each other. Consequently, selection of the best PWM technique for most applications is accompanied by uncertainty, which can lead to less than optimum results. A critical evaluation of the aforementioned PWM techniques on the basis of application is provided, thereby giving the framework and guidelines for the selection of the best technique for each area of application.

.: C I

I. INTRODUCTION

-I
-2

A . General N power electronics, pulsewidth modulation (PWM) is an operation performed on "raw" voltage and current waveforms to shape their spectra in a way beneficial to the application under consideration. Spectra shaping typically means the creation of a "dead band" between wanted and unwanted spectral components. For a given switching frequency it is desirable that the dead band be as wide as possible. To illustrate this point, Figs. l(a), l(b), l(d), and l(e) show a typical inverter line-to-line output voltage waveform and its respective spectra before and after it has been pulsewidth modulated (PWMed). Figs. l(c) and l(f) show the resulting line current waveforms obtained with a load PF = 0.8 lagging. In particular, Fig. l(f) shows that PWM allows static inverters to generate close to ideal output waveforms while providing variable-voltage and variable-frequency operation. Further investigation of the results shown in Fig. 1 also reveals that PWM has several disadvantages, which include

.5
E -.5
ut

(e)
2 -

-1

ut

i) attenuation of the wanted fundamental component of the PWMed waveform, in this case from 1.1-0.866 pu; ii) drastically increased switching frequencies (in this case from 1 pu to 21 pu)-this means greater stresses on associated switching devices and therefore derating of those devices:
Paper IPCSD 87-29, approved by the Static Power Converter Committee of the IEEE Industry Applications Society for presentation at the 1986 Power Electronics Specialists Conference, Vancouver, BC, Canada, June 23-27. Manuscript released for publication June IO, 1987. The authors are with Concordia University, 1455 de Maisonnueve Boulevard West, Montreal, PQ, Canada H35 1M8. IEEE Log Number 8717547.

Fig. 1. Inverter voltage and current waveforms with six-step and sinusoidal PWM (SPWM) operation. (a) Square-wave output line-to-line voltage. (b) Line-to-line voltage spectrum (square wave). (c) Output line current (square wave). (d) SPWM output line-to-line voltage. (e) Line-to-line spectrum (SPWM). (f) Output line current (SPWM).

iii) generation of high-frequency harmonic components previously not present. However, in addition to the importance of spectral shaping, the main advantage of pulsewidth modulation is that it allows linear amplitude control of the output voltages/currents from

0093-9994/88/03OO-O271$01 .OO 0 1988 IEEE

Authorized licensed use limited to: Universidad de Concepcion. Downloaded on October 13, 2009 at 10:18 from IEEE Xplore. Restrictions apply.

272

IEEE TRANSACTIONS ON INDUSTRY APPLICATIONS, VOL. 24, NO. 2 , MARCH/APRIL 1988

within the converter. Consequently, inverters can be supplied from diode (instead of thyristor) rectifiers, resulting in a simpler and cheaper power conversion system.

B. Specifics of PWM Techniques Treated in this Paper

Since the advantages of PWM clearly outweigh the respective disadvantages, a considerable research effort has gone into minimizing the PWM disadvantages mentioned earlier. As a result, several specific PWM techniques that focus on improving source utilization and generated input/output harmonic content have been proposed lately [2]-[4]. It is noted that the advantages of all these improved PWM techniques over the original sine PWM technique are obtained at the cost of generating third line-to-neutral voltage harmonics. However, under balanced and open neutral operating conditions, third-harmonic currents cannot flow and thus the thirdConsequently, it appears that even with projected technoorder voltage harmonics are neutralized. logical advances the programmed PWM techniques could only Previous work related to these improved techniques has partially replace the carrier PWM techniques discussed in this been focused on converter output spectra, neglecting other paper. important criteria such as input spectra, input/output distortion 11. CHARACTERIZATION OF IMPROVED PWM TECHNIQUES factor, switching frequencies and hardware implementation A . Preliminary considerations. Furthermore, another focus of previous work has been inverter applications neglecting in the process some The main features of the improved PWM techniques treated other equally important areas such as dc drives, rectifier in this paper are brought out in this section. These features power supplies, and rectifier-inverter type frequency chang- include the derivation of switching points, the resulting ac and ers. Consequently, selection of the best PWM technique for dc term gains, the resulting frequency spectra, the harmonic most applications is accompanied with uncertainty, which can distortion at the converter input and output terminals, and lead to less-than-optimum results. finally, the degree of difficulty of hardware implementation For these reasons this paper provides a critical evaluation of for each technique. Since there are four PWM techniques and all the aforementioned PWM techniques on the basis of three types of converters involved in this evaluation, care must application, thereby providing the framework and guidelines be taken to present relevant results with clarity. For this for the selection of the best technique for each area of purpose, a generalized bridge converter (Fig. 2 ) comprised of application. The applications considered include six ideal four-quadrant switches has been employed. The main advantage of this converter is that because of the 1) voltage and current source inverters for nature of its switches it can function either as an inverter a) variable-speed ac motor drives, (voltage or current source) or as a rectifier by simply applying b) UPS power supplies; the proper gating signals. The generalized converter thus 2) rectifiers for permits the ac terminal waveforms (ac term) to represent a) dc motor drives, [6] either b) power supplies; a) the line-to-line voltage of a voltage source inverter 3) high-performance rectifier-inverter type frequency changers. (VSI); b) the output line current of a current source inverter (CSI); C. Other Improved P WM Techniques or c) the input line current of a controlled rectifier (CR); Another class of improved PWM techniques that are not included in this paper are the so-called programmed PWM and the dc terminal (dc term) waveforms to represent techniques [7]-[9]. These techniques offer even better voltage a) the input current of a VSI; utilization and lower switching frequencies when employed b) the input voltage of a CSI; or with inverters supplied from independently regulated voltage c) the output voltage of a CR. or current sources. However, when these techniques are employed with variable-frequencyivariable-voltageinverters Also, in the evaluation that follows, the switching frequency and an unregulated dc bus, they have a number of disadvan- of the bridge is kept constant to provide a common ground for tages, which include the following. comparison of each technique under rectifier and inverter operation. i) Quite sophisticated control hardware is required to store and access the required large number of switching B. The Original Sine P WM Technique patterns. ii) For ac motor drives, and at low operating frequencies, This technique [ 2 ] has been included for the purpose of using it as a benchmark to evaluate respective improved

the number of switching points becomes too large to evaluate analytically, even with mainframe computers. iii) When a large number of harmonics needs to be eliminated, the respective switching pattern cannot be reproduced accurately by the inverter due to delays associated with the inverter switches and their respective base drives. iv) Associated hardware becomes practical and cost effective only through VLSI implementation. The expertise required to design such hardware is not available to most small to mid-size companies. Also the number of units produced must be large enough to justify initial development costs. v) Reliability of VLSI technology in high-current switching environments is still questionable.

Authorized licensed use limited to: Universidad de Concepcion. Downloaded on October 13, 2009 at 10:18 from IEEE Xplore. Restrictions apply.

BOOST AND ZIOGAS. STATE-OF-THE-ART CARNER PWM


I
I

273
I 7

RC

TERMINRL
( D C TERM1

TE R M I N R L ( R C TERM)

Fig. 2 .
1

Generalized bridge configuration.

-1

tV v v v vi v v v v
90 180

2 0

w t

270

360

wt

(b)

r
w t

90

180

270

360

(e)

.26

.26

.31

e
0

ie

24

42

360

FCpu)

(f)

90

1B 0

270

w t

(g)
Fig. 3 . Sinusoidal PWM (SPWM). (a) SPWM scheme. (b) VSI switch #1 gating signal (SWI). (c) AC term. (d) AC term spectrum (e) DC term. (f) DC term spectrum. (g) CSIKR switch #1 gating signal.

techniques. Its main intrinsic features are shown in Fig. 3. The main disadvantage with this technique is that the maximum possible ac term and dc term gain values are only Gac = 0.866 and Gdc = 0.75, respectively, where i) ac term gain is the ratio of maximum value (peak) of the fundamental component of the ac term to the amplitude

of the unfiltered pulses comprising the same term; and ii) dc term gain is the ratio of the maximum value of the dc component of the term to the maximum amplitude of the unfiltered pulses comprising the same term.

For many applications this low Gac value means the use of a voltage-matching transformer. The main advantage of this

Authorized licensed use limited to: Universidad de Concepcion. Downloaded on October 13, 2009 at 10:18 from IEEE Xplore. Restrictions apply.

274

IEEE TRANSACTIONS ON INDUSTRY APPLICATIONS, VOL. 24, NO. 2, MARCHiAPRIL 1988

technique is that it generates line-to-neutral ac spectra with no low-order third harmonics. This allows the use of neutral-toneutral connections (if required) and decoupled (individual) control of each one of the three inverter phases.

a) The analytical expression for the reference waveform is now y = 1.15 sin (wt) + 0.27 sin (3 wt) - 0.029 sin (9

wt).

C. The Modified Sine P WM Technique


This is the first of the improved PWM techniques [3] treated in this paper. Its main intrinsic features are shown in Fig. 4. For this technique spectra are shown for VSI and CSIKR operation, respectively, in order to maintain the same switching frequency. The particulars follow. a) This technique defines the ac term (Fig. 4(c)) on a lineto-line basis for VSIs and on a line basis for CSIs and CRs. b) As shbwn in Figs. 4(a) and 4(b), only the first and last 60" intervals (per half-cycle) of the ac term waveform are directly defined through intersections of respective sine reference) and triangular (carrier) waveforms. The 60" to 120" intervals are obtained by "folding" the first and last 60" intervals around the 60" and 120" points, respectively. c) As shown in Fig. 4(d), this technique provides a substantially higher ac term gain as compared with the original sine PWM technique (Fig. 3). However, from the aforementioned b) it follows that hardware implementation for this technique is rather complex. Also, this technique generates a substantial (2 1-percent) ac term third-harmonic component on a line-to-neutral basis.

b) The ac term again (Fig. 6(d)) is equal to the one obtained with the previous two improved PWM techniques while the harmonic spectra of ac and dc terms are clearly better. Again, as shown in Fig. 6(a), the hardware implementation of this technique is as simple as with the original sine and the third-harmonic injection techniques.

F. Waveform Quality Under Variable Modulation Index Conditions


Although the waveforms shown in Figs. 3, 4, 5, and 6 provide an accurate characterization of the evaluated PWM techniques, they also have the disadvantage of showing respective harmonic spectra at only one modulation index value (i-e., M = I). However, in most applications, output converter power is controlled by varying the respective modulation index value. Therefore it is also necessary to investigate the quality of the various converter (voltage and current) waveforms obtained with each technique for all modulation index (M index) values. In particular, these waveforms are the ac and dc terminal waveforms (Figs. 3(c), 3(e); 4(c), 4(e); 5(c), 5(e); and 6(c), 6(e)) obtained with the generalized converter shown in Fig. 2. They include the lineto-line voltages of VSIs output line currents of CSIs, and input line currents of CRs (all three represented by the ac terminal waveform); and the input (dc link) current of VSIs, the input voltage of CSIs and the output voltage of CRs (all three represented by the dc terminal waveform). Furthermore, the three quality indexes used to evaluate these waveforms are defined as follows:

D. The Third-Harmonic Injection P WM Technique


This improved technique 121, [4] (Fig. 5) has been derived from the original sine PWM technique (Fig. 3) through the addition of the 17-percent third-harmonic component to the original sine reference waveform. The resulting flat-topped waveform (Fig. 5(a)) allows overmodulation (with respect to the original sine PWM technique) while maintaining excellent ac term and dc term spectra. The particulars follow. a) The analytical expression for the reference waveform is y = 1.15 sin (wt) + 0.19 sin (3 wt). b) The ac term gain (Fig. 5(d)) is equal to the respective gain obtained with the modified sine PWM (Fig. 4(d)) and substantially higher than the one obtained with the original PWM technique (Fig. 3(d)). Furthermore, as shown in Fig. 5(a), the hardware implementation of this technique is quite simple. However, this technique also generates a substantial ac term third-harmonic component (17 percent) on line-to-neutral basis.

where

DF, ac terminal distortion factor for second-order ac-side


filtering;

DFZ ac terminal distortion factor for first-order ac-side


filtering;

E. The Harmonic Injection P WM Technique


This technique (Fig. 6) is a variation of the previously discussed third-harmonic injection technique. The variation is obtained by injecting additional harmonics in the respective reference waveform. The resulting flat-topped waveform (Fig. 6(a)) again allows overmodulation while improving even further the resulting frequency spectra of the ac term and dc term waveforms. The particulars follow.

OF3 dc terminal distortion factor for first-order dc-side


filtering;

H ( n ) amplitude of the nth harmonic; H(1) maximum ac gain Gac; H(0) maximum dc gain Gdc.
The variations of these three quality indices as a function of the M index are shown in Figs. 7, 8, and 9, respectively.

Authorized licensed use limited to: Universidad de Concepcion. Downloaded on October 13, 2009 at 10:18 from IEEE Xplore. Restrictions apply.

BOOST AND ZIOGAS: STATE-OF-THE-ART CARRIER PWM

275

0
-1

w t

90

180

270

360 w t

(b)
l - -

I I

0 -

Ig

98
-I

180

a;'E

360

'

wt

-_

90

270

360

w t

0 90 180

I IIIII I
270

360 I) wt

-AI

.I4

0 0

.I4
c

30

36

F(pu)

Authorized licensed use limited to: Universidad de Concepcion. Downloaded on October 13, 2009 at 10:18 from IEEE Xplore. Restrictions apply.

276
1

IEEE TRANSACTIONS ON INDUSTRY APPLICATIONS, VOL. 24, NO. 2 , MARCHIAPRIL 1988

-1

tV v v v r v v v v WNY v v v V !
2 0
90

wt

I80

270

360 w t

(b)
1

-1

lh

90

180

270

360

W t

(e)

90

I88

270

360 w t

Fig. 5.

(g) Harmonic injection sinusoidal PWM (HIPWM) (1st and 3rd). (a) HIPWM scheme. (b) VSI switch # I gating signal (SWI). (c) AC term. (d) AC term spectrum. (e) DC term. (f) DC term spectrum. (g) CSIlCR switch # I gating signal.

It is noted that DF,, DF2, and OF3 have been defined in ways that reflect actual levels of harmonic distortion experienced in practical applications. For example, practical static uninterruptible power supplies (UPS) employ a second-order LC filter between respective inverters and loads. Such filters provide harmonic attenuation, which is approximately inversely proportional to the square of the order (n) of the harmonic. Therefore the OF, data shown in Fig. 7 are relevant to UPS (i.e., load voltage) and any other static power supply that employs a second-order filter. Similarly, ac and dc motors supplied from PWM static converters utilize their respective leakage and armature inductances to produce quasi-sinusoidal

and quasi-dc input current waveforms. These inductances provide first-order attenuation to voltage harmonics, which is equivalent to dividing the amplitude of each harmonic by its respective order. Therefore the DF2 and OF3 data shown in Figs. 8 and 9 are applicable to ac and dc motor drives and any other application that uses an actual or equivalent first-order filter. Moreover, to stress emphasis on the input spectral content of each technique, a quality index is defined for the rms ripple value that the input filter must tolerate or support. The rms current rating Ici of the dc link capacitor for VSI operation and the rms voltage rating V,i of the dc link inductor for CSI operation is shown in Fig. 10 as a function of

Authorized licensed use limited to: Universidad de Concepcion. Downloaded on October 13, 2009 at 10:18 from IEEE Xplore. Restrictions apply.

BOOST AND ZIOGAS: STATE-OF-THE-ART CARRIER PWM

277

1719

2325

3537

4143

4749

F(pu)

IY
90

270

360

WT.

1
I

I)

Fig. 6 . Harmonic injection sinusoidal PWM (HIPWM) (lst, 3rd, and 9th). (a) HIPWM scheme. @) VSI switch #1 gating signal (SW1). (c) AC term. (d) AC term spectrum. (e) DC term. (f) DC term spectrum. (9) CSIKR switch #1 gating signal.

El

'

.2

'

'

.4

'

.6

'

.8

'

'

I
H index

Fig. 7 . Second-order filter distortion factor (ac term). (a) SPWM. (b) HIPWM (1st and 3rd). (c) MSPWM VSI operation. (d) HIPWM (lst, 3rd, and 9th). (e) MSPWM CSI/CR operation.

Authorized licensed use limited to: Universidad de Concepcion. Downloaded on October 13, 2009 at 10:18 from IEEE Xplore. Restrictions apply.

278

IEEE TRANSACTIONS ON INDUSTRY APPLICATIONS, VOL. 24, NO. 2, MARCHIAPRIL 1988


Irns pu
7-

.375

2 -

;
I -

B&D

"

"

"

"

'

~
Fig. 11. Rectifier input capacitor rms ripple current. (a) SPWM. (b) HIPWM (1st and 3rd). (c) MSPWM VSI operation. (d) HIPWM (lst, 3rd, and 9th). (e) MSPWM.

where VSI rms input ripple current, CSI rms input ripple voltage, Icr CR rms input ripple current, Vin(n)/Zin(n) rms magnitude of n th harmonic.
Zci

v/i

81

'

.2

'

.4

.6

'

'

.E

'

1 M index

Fig. 9. First-order filter distortion factor (dc term). (a) SPWM. (b) HIPWM (1st and 3rd). (c) MSPWM VSI operation. (d) HIPWM (lst, 3rd, and 9th). (e) MSPWM CSI/CR operation.

The variation of these performance indices as a function of modulation ( M index) and rated load current conditions is shown in Figs. 10 and 11 (for all four PWM techniques treated in this paper). It is finally noted that although the exact DF,, DF2, DF3, and IC;, V/i,ZCr values shown in Figs. 7-11 are valid for only one particular common-carrier frequency (shown in Figs. 3(a), 4(a), 5(a), and 6(a)), their shapes and their relative position are independent of carrier frequency. Consequently, these data can be used for the general evaluation of the subject PWM techniques.
1 1 1 . EVALUATION OF IMPROVED PWM TECHNIQUES

ICl/VII pu

.75 r

Fig. 10. Inverter dc terminal rms ripple. VSI operation-capacitor ripple current (Ic,); CSI operation-inductor ripple voltage (V,,).(a) SPWM. (b) HIPWM (1st and 3rd). (c) MSPWM VSI operation. (d) HIPWM (lst, 3rd, and 9th). (e) MSPWM CSI operation.

In this section the previously discussed and analyzed techniques are evaluated by using relevant data obtained in previous sections. For the meaningful interpretation of evaluation data, the switching frequency variable has been eliminated by using the same switching frequency for all PWM techniques. However, careful investigation of the modified sine PWM technique shows [3] that for the same ac term and dc term waveforms, CSIs, and CRs require lower switching frequencies than VSIs. Because of this asymmetry, this technique has been represented twice in Figs. 7-10.

A. A C and DC Term Gains


Respective ac term and dc term gains for all four PWM techniques are shown in Figs. 3(d), 3(f), 4(d), 4(f), 4(h), 4(i), 5(d), 5(f), 6(d), and 6(f) and are summarized in Table I. It is noted that these gain values are independent of switching frequency and are directly proportional to the respective modulation index values ( M index). From this table it is obvious that regarding ac/dc gains, all improved PWM techniques are a) better than the original sine technique; b) equivalent among themselves.

modulation index. The rms current of the input capacitor Zcr for controlled expressions for IC;, V,, and Zcr are

I m

B. Quality Factors DF,, DF,, DF, and Zcj, ZIi, IC,


The variations of DF,, DF,, and DF, as a function of M index are presented in Figs. 7-9. These figures show that

Authorized licensed use limited to: Universidad de Concepcion. Downloaded on October 13, 2009 at 10:18 from IEEE Xplore. Restrictions apply.

BOOST AND ZIOGAS: STATE-OF-THE-ART CARRIER PWM

279
TABLE 1

MAXIMUM A C AND DC GAIN VALUES Gac AND Gdc

(rn

= 1)

I I

TECHNIQUE ORIGINRL S I N E PWM ( f l g . 3 )

Gac

I
I
I

I
0.866

Gdc
8.758

I
I
I

MODIFIED SINE P W W (f I g . 4 ) T H I R D HARMONIC INJECTION


~ ~~

I
I

ctcg.5,
HRRWONIC INJECTION (flg.6)

8.866

0.866

T A B L E I1 OPTIMUM P W M TECHNIQUES ACCORDING TO APPLICATION


RPPLICRTION V S I BRSED RC MOTOR D R I V E S OPTIMUM P W W TECHNIQUE LOW SPEEDS: HRRWONIC I N J E C T I O N TECHNIQUE H I G H SPEEDS:PROGRRWMED HARMONIC E L I M I N R T I O N TECHNIQUES ( P H E T s ) C S I BRSED RC MOTOR D R I V E S H I G H PERFORMRNCE: M O D I I F I E D S I N E PWM TECHNIQUE TYPICRL: PHETt PWM R E C T I F I E R S FOR C S I INVERTER RND DC WOTOR DRIVES UPS ( W I T H UNREGULRTED DC B U S ) UPS ( W I T H REGULRTED BUS) HARMONIC I N J E C T I O N TECHNIQUE HRRMONIC I N J E C T I O N TECHNIQUE PHETs REFERENCES F I G S . 6 8.9.10
SEE R E F S . C 7 1 C83 [ 9 1

F I G S . 4.8,9.10

SEE REFS. C7 I C8 I C 9 I F I G S . 6 8.9,ll

F I G S . 6.7

3.18

DC

SEE R E F S . C 7 3

[ E l C91

i) for voltage source inverter (VSI) applications the harmonic injection technique offers the overall best quality ac term waveforms; ii) for current source inverter (CSI) applications the harmonic injection technique is best for M < 0.65 (approximately) while the modified sine technique becomes increasingly better for M > 0.065 (approximately). Next, the variation of IC;,V,i, and I , , as a function of the M index is shown in Figs. 10 and 11. It is noted that the maxima of the I,; and V,; curves occur around the M = 0.6 point instead of the M = 1 point. This fact should not be overlooked when considering the ratings of respective filter components. From the figures it can be seen that i) for VSI operation, the harmonic injection PWM (HIPWM) technique demands the lowest ripple current from the input filter capacitor; ii) for CSI operation, modified sinusoidal PWM (MSPWM) produces the lowest ripple voltage across the link reactor; iii) for CR operation, HIPWM and MSPWM offer the best results. IV. SELECTION OF OPTIMUM PWM TECHNIQUE ACCORDING TO APPLICATION As stated earlier, quality factors D F I ,DF2, and DF3 reflect actual levels of harmonic distortion experienced in practical applications. Therefore it is possible to use respective analytical data from Figs. 7-1 1 to select the optimum technique(s) for each of these applications. For example, from the DFI definition given earlier and the results presented in Fig. 7 it is justifiable to say that for UPS VSI applications (without a dc

Fig. 12. Experimental waveforms for VSI using H I P W M (1st. 3rd, and 9th). (a) Output line-to-line voltage. (b) Output line-to-line voltage spectrum. (c) Input current. (d) Input current spectrum.

voltage control stage) the harmonic injection technique is the best choice. By following the same approach, similar conclusions can be drawn for CSI-based UPS, VSI, and CSI-based motor drives, etc. Because of the apparent importance of this subject, detailed information is given in Table 11. V. EXPERIMENTAL RESULTS The validity of selected results predicted in previous sections has been verified experimentally on a 1-kVA laboratory prototype system. Samples of these results are shown in

Authorized licensed use limited to: Universidad de Concepcion. Downloaded on October 13, 2009 at 10:18 from IEEE Xplore. Restrictions apply.

280

IEEE TRANSACTIONS ON INDUSTRY APPLICATIONS, VOL. 24, NO. 2 . MARCHIAPRIL 1988

optimum PWM technique. Detailed information has been presented in Tables I and 11.

REFERENCES
[l] V. R. Stefanovic, Present trends in variable speed ac drives, presented at IPEC 1983, Tokyo, Japan. [2] R. Bonnert and R. S. Wu, Improved three phase pulse width modulation for overmodulation, in Conf. Rec. 1984 IEEE-IAS Annual Meeting, Chicago, IL. [3] T. Onishi and H. Okitsu, A novel PWM technique for three phase inverter/converter, in Conf. Rec. 1983 IPEC, Tokyo, Japan. 141 J. A. Houldsworth and D. A. Grant. The use of harmonic distortion to increase the output voltage of a three-phase PWM inverter, IEEE Trans. Ind. Appl., vol. IA-20, no. 5 , SeptJOct. 1984. [5] A. Schonung and H. Stemmler, Static frequency changers with subharmonic control in conjunction with reversible variable speed ac drives, Brown Boveri Rev., Aug./Sept. 1964. [6] E. P. Wiechmann, P. D. Ziogas, and V. R . Stephanovic. Time domain functional model for three phase PWM inverterirectifier converters, in IEEE-IAS I985 Conf. Record, Toronto, O N , Canada. [7] H. S. Patel and R . G. Hoft, Generalized techniques of harmonic elimination and voltage control in thyristor inverters: Part I , harmonic elimination, IEEE Trans. Ind. Appl., vol. LA-9, no. 3, MayiJune 1973. [8] H. S. Patel and R. G . Hoft, Generalized techniques of harmonic elimination and voltage control in thyristor inverters: Part 11, voltage control techniques, IEEE Trans. Ind. Appl., vol. 1A-10. vol. 5 , Sept. /Oct . 1974, [9] W. Lienau, A. Muller-Hellmann, and H. C. Skudelny, Power converters for feeding asynchronous traction motors of single phase ac vehicles, IEEE Trans. Ind. Appl., vol. IA-16, no. 1, Jan./Feb. 1980.
. 1

Fig. 13. Experimental waveforms for CSI using MSPWM. (a) Output line current. (b) Output line current spectrum. (c) Input voltage. (d) Input voltage spectrum.

Figs. 12 and 13. In particular, Fig. 12 shows the experimental terminal waveforms and the associated spectra of a VSI operating with HIPWM (lst, 3rd, and 9th) at modulation index = 1 and a carrier frequency of 21 pu. The theoretical results for the same operation are shown in Fig. 6. Fig. 13 shows the experimental terminal waveforms and the associated spectra of a CSI operating with MSPWM at modulation index = 1 and a carrier frequency of 33 pu (results in 22-pu switching frequency). The theoretical results for the same operating conditions are shown in Fig. 4. The close agreement between the analytical and experimental results prove the validity of the theoretical evaluations made. VI. CONCLUSION In this paper a critical evaluation of state-of-the-art carrier PWM techniques has been presented. Through the use of relevant theoretical and experimental results it has been shown that for each major area of application there is a specific

Micliael A . Boost (S82) obtained the B.Eng. degree in 1983 and is currently enrolled in the Ph.D. program at Concordia University, Montreal, PQ, Canada He has participated in several industrial research and development projects in the area of mediumpower UPS

Phoivos D. Ziogas (S75-M78) received the B S , M.S., and Ph D degree\ from the University of Toronto, Toronto, ON, Canada, in 1973, 1974, and 1978, respectively Since 1978 he has been with the Department of Electrical Engineering of Concordia University in Montreal, PQ, Canada, where he is engaged in teaching and research in the area of static power converters He also has participated as a Consultant in several industrial projects

Authorized licensed use limited to: Universidad de Concepcion. Downloaded on October 13, 2009 at 10:18 from IEEE Xplore. Restrictions apply.

You might also like