IEEE TRANSACTIONS ON POWER ELECTRONICS. VOL. 9, NO.

2, MARCH 1994 153

A Novel Load Current Control Method for a Leading

Power Factor Voltage Source PWM Rectifier
Diego R. Veas, Juan W. Dixon, Member, IEEE, and Boon-Teck Ooi, Senior Member, IEEE

Abstract- A novel PWM voltage source rectifier, controlled

by the load dc current instead of the dc voltage, has been
developed. Its main characteristics are: a) there is neither input
current sensors nor dc voltage sensor; b) it works with an
unchangeable and predefined PWM pattern; c) it presents a very
strong stability; d) its stability does not depend on the size of
the dc capacitor; e) it can work at leading power factor for all
load conditions, and f) it can also work with zero regulation for
all load conditions. Digital simulations, analyses, and experiments
confirm all these characteristics of the control method.

I. INTRODUCTION TEMPLATE

T HE new generation of voltage source, pulse width mod-

ulated rectifiers has reached a good performance level in
terms of low harmonic distortion [1]-[3] and unity, or even
Fig. 1. Direct current control rectifier.

leading power factor operation [4]. Power reversal by current

reversal and good dynamic response have also been some

-
a
of its merits. The common factor of this kind of rectifier is
b
the method used in controlling the dc voltage, which is kept
close to a reference by using the error signal between that C

voltage and the reference. This error either controls directly I l l I-

the ac input currents (direct control) [ 5 ] , 161 or the input
voltage modulation (indirect control) [7], [8]. In both cases,
the pattem of the pulse width modulation (PWM) is being
adjusted permanently to satisfy the load requirements. For the
Ill
direct control method, precise input current sensors and one
output dc voltage sensor are required. In the indirect control
Fig. 2. Indirect current control rectifier.
method, the current sensors are eliminated, but then the system
becomes more unstable [8]. These two methods of control
require a dc capacitor large enough to ensure a good stability trolled in a similar form, controlling the dc voltage indirectly
margin, making the system transient response slower. Fig. 1 by adjusting the power angle through the dc load current.
shows a schematic circuit of the direct control rectifier, and This load-current-controlledrectifier can keep the dc voltage
Fig. 2 shows the indirect control without sensing it, and can operate at leading power factor for
A different approach is proposed in this w o r k 4 power all load conditions. The stability, unlike the other rectifiers
rectifier without dc voltage sensor and with an unchangeable mentioned here, neither depends on the input inductance nor
and unique PWMpattern. The idea is based on the principle on the size of the dc capacitor.
of operation for series connected, type B, PWM rectifiers The work analyzes two methods of operation based on
proposed in [9], in which a “master” unit controls the voltages this principle: a “zero regulation control” and a much sim-
of the other rectifiers called “slaves” without measuring their pler “linear control,” from which the analytical tools have
dc voltages. In the method proposed here, the rectifier is con- been developed. Also, digital computer simulations have been
used to understand the behavior of the rectifier under static
Manuscript received December 18, 1991; revised November 24, 1993. This
and dynamic operations. Finally, a 2-kW prototype has been
work was supported by the Fondo de Desarrollo Cienti fico y Technoldgico implemented using “linear control,” which has permitted the
(FONDECYT) under Project 600-89. verification of the analyses and the digital simulations.
D. R. Veas and J. W. Dixon are with the Department of Electrical
Engineering, Universidad Catdlica de Chile, Casilla 306, Correo 22, Santiago,
Chile. 11. PRINCIPLE OF OPERATION
B.-T. Ooi is with the Department of Electrical Engineering, McGill Uni-
versity, Montreal, P.Q., H3A-2A7, Canada. Fig. 3 shows the schematic of the proposed control system.
IEEE Log Number 9400087. The method used for controlling the rectifier is as follows:
0885-8993/94$04.00 0 1994 IEEE
I54 IEEE TRANSACTIONS ON POWER ELECTRONICS, VOL. 9, NO. 2, MARCH 1994

'1
Td

Fig. 5. Phasor diagram.

Fig. 3. Load current control rectifier.

V L R I i I. I^

I I

INPUT OUTPUT
Fig. 4. Equivalent diagram. Fig. 6 . I-c versus i z for different 0 angles ( S I R = 5).

the voltage Vmod modulated by the rectifier, is produced by in the d-q frame as follows:
an unique PWM pattem, which is shifted with respect to the
mains voltage V, to change the power angle 0 and hence the U, . i l = w modd . i d + w mod, .i,. (1)
amount of power flow transferred from the ac to the dc side.
When the power angle is negative (V mod lags V), the power On the other hand, at the ac side, the phasor diagram es-
flow goes from the ac to the dc side. When the power angle tablishes operating conditions between the mains and Vmod
is positive, then the power flows in the opposite direction. (the modulated voltage) as shown in Fig. 5.
The main idea is to control this power angle through the dc This phasor diagram can also be expressed in the d-q frame
load current. One important thing that must be mentioned here as
is that this rectifier establishes a stable dc voltage operation
for each dc current and power angle. With this characteristic,
it is possible to find a relation between the dc current and
the input power angle, to have zero dc voltage regulation for
where
all load conditions. Then, without the need of measuring the
dc voltage and with only one PWM pattem, the dc voltage
can be kept constant. This situation may include both rectifier
operation and inverter operation. As was already mentioned,
a simpler control version can be implemented by making the The rectifier works with a fixed PWM pattem, which means
power angle proportional to the dc current. This last version that there exists a linear relation between the amplitude of the
has been called "linear control." modulated voltage Vmod and the dc voltage Vc:

111. ANALYSISOF THE RECTIFIER IV mod I = Kw . Vc. (6)

The equivalent rectifier circuit for the fundamental (first When the system is operating under steady-state, the dc ca-
harmonic and dc component at input and output terminals of pacitor reaches a particular equilibrium voltage, which means
the rectifier, respectively) is shown in Fig. 4. In this figure, V is that the dc load current i 2 becomes equal to 21. If the power
the fundamental phase-to-neutral mains voltage. V mod is the angle 0 (angle between V and Vmod) is kept constant, the
fundamental of the PWM pattem. R and L are the resistance dc voltage will take a particular and well-defined value for
and inductance per phase located between the mains and the each load current. By using (1)-(6), it is possible to derive an
PWM pattem. The dc capacitor allows us to have a defined expression to see how the dc voltage V c depends on the load
dc voltage at the output terminals. current i 2 and the power angle 0.
The current source i l is controlled by the magnitude and
angle of Vmod and satisfies the power balance equation V c = (V/Kw) . [cos 0 - ( X / R )sin 0)
between the ac inputs and the dc output. It can be written - 22 (R2 + X 2 ) / ( 3 .KV2 R )
' ' (7)
VEAS PI ol.: A NOVEL LOAD CURRENT CONTROL METHOD FOR A LEADING POWER FACTOR VOLTAGE 15.5

60” I =U
R R

- 0.25 - 0.15

- aoo---- - - - -- --

Fig. 7. 0 - zz relation for I.-c = constant.

where
0 = arctg (w modd/w mod,). (8)
Equation (7) can be plotted to show how V c is affected by
the load current 22, keeping the power angle 0 constant. Fig.
6 shows V c versus 22 curves, for different 0 angles.

IV. “ZEROREGULATION” CONTROL

From Fig. 6, it can be seen that it is possible to link the load FIXED
PWM
dc current 22 with the angle 0 to get zero regulation operation.
To do that, 0 have to be modified each time the load current
i 2 changes to ensure that V c will be kept at the same value.
Fig. 8. “Zero regulation” PWM rectifier.
First, it is required to choose an arbitrary value of voltage V c .
For example, 0 = 0 for 22 = 0 can be selected and, then,
voltage for all load conditions. Fig. 7 shows these curves for
because of (7), the voltage V c takes the value
two different values of X / R and Kw = 1/3.
The relation plotted on Fig. 7 can be stored in a ROM
memory, where the input is the digital value of i ~and , the
output is the required power angle 0 to get zero regulation. A
This arbitrary value of V c form (9) can be replaced in (7) to rectifier with this kind of control is shown in Fig. 8.
get a 0 - i 2 relation for zero regulation operation. Under these It is important to realize that the rectifier cannot maintain
conditions, the voltage V c will always be equal to V/Kw and zero regulation beyond certain limits. From Fig. 7, it can
also, because of (6), [VI will be equal to IV mod I. be observed that there exists a limit for 22, I~CRIT. Beyond
this value, the dc voltage begins to fall. 1 2 ~ is~ evaluated
1 ~

cos 0 - ( X / R ). sin 0 - 1 through (dia/dO) and depends on the X / R relation. The

optimum value for X i R , in terms of higher I ~ C Ris IT .
J3.
+
= i 2 . (R2 X 2 ) / ( 3 V .KW. R). (10) IZCRIT also depends on the V I R rate, which is normally
greater than 100. For example, a rectifier working with V =
Equation (10) allows us to plot the required 0 - 22 curves 110 V and R = 0.5 R is able to keep zero regulation condition
to get zero regulation operation, which means the same dc until 12CRIT = 55 amp for X / R = J3.
156 [EEE TRANSACTIONS ON POWER ELECTRONICS, VOL. 9, NO. 2, MARCH 1994

\
d \

Fig. 9. I d magnitude under inverter and rectifier operation.

Fig. 11. The circle diagram for i2 = 0.

Fig. 10. Control block of the rectifier with @OFFSET.

V. LEADINGPOWERFACTOROPERATION - %
\

One important characteristic of the PWM rectifiers is that LOCUS OF Vmod FOR
“ZERO REGULATION”
they can operate at leading power factor. In other words, they LAGGING
can generate reactive power. The amount of reactive power is CURRENT
>-
__-cos ’ p = l
given by Q = 3 . V . I . sin 4. This equation, transformed to
the d-q frame, can be written as LEADING
CURRENT
Q J 3 . V . Id. (1 1)
INVERTER
When Id < 0, the converter absorbs reactive power. To avoid OPERATION
that, it would be required to have IV mod I greater than IVI; -- - - _--- ._
otherwise, the converter will work at lagging power factor
under inverter operation. This is because when 22 becomes
LOCUS FOR
WHEN
@OF< i 2 = 0 vAdd)--- OPERATION
RECTIFIER
- _--.

negative, the 0 - 22 relation (given in Fig. 7) makes 0 positive

and, then, if IV mod I remains equal to [VI because of (9), Id
\
‘ ‘ \ r - 7

becomes negative. This is shown in the phasor diagram of ........ -‘ d

Fig. 9. \
’$ !,EADMG
A solution for this problem is to increase the magnitude of LAGGING 6 CUZRENT
V mod . This is reached by adding a negative value of 0 in CURRENT
cos ’p= 1
the control block, defined as @OFFSET, which is shown in
Fig. 12. “Zero regulation” circle diagram.
Fig. 10.
@OFFSET shifts the power angle, allowing leading power
factor operation also during power inversion. Let us first with respect to the d-q axis. In this figure, the leading current
analyze the behavior of @OFFSET when 22 = 0. When limits for rectifier and inverter operation are clearly shown,
@OFFSET is modified, Vmod moves describing a circle, as and correspond to the dotted lines marked “COS 4 = 1” (unity
shown in Fig. 11. Under these conditions, the rectifier works power factor operation). From the geometry of the circles in
as a var compensator, which was described in detail in [lo]. Fig. 12, it can be seen that the power factor inside these limits
Now, adjusting @OFFSET to a particular value and adding is always leading but not constant.
the required “zero regulation” linking curve shown in Fig. 7, a
second circle diagram, which depends on 22, is described. The
radius of this circle is IV mod 1, which is constant because VI. LINEARCONTROL
the “zero regulation” control forces Vc to be constant. As Looking at Fig. 7 once again, it can be observed that the
Vmod and Vc are related through (6), it yields Vmod = relation between 0 and 22 is almost linear for normal currents
constant. Fig. 12 shows this second circle diagram drawn (22 < 22CRIT). This behavior allows a simplification in the
VEAS et al.; A NOVEL LOAD CURRENT CONTROL METHOD FOR A LEADING POWER FACTOR VOLTAGE 157

-
- 0.4Vc0
V ~ = V , ~ ( C O S ( 12
- -)R . ( TIz
+ - X- s m y F) -
R ) - I2

I o X 0 0 - 0.2vc0 V
[ 1 + (X)']
lo= R
-
I I 1 1

I I I I I I I I I I I I 1 I l l I I I I I I ~
-0.5 ,I 0 0.5 lo
'2
-I0 IO

Fig. 13. I-c - 72 characteristic using linear control.

control block by changing (10) for a very simple expression where

and
Kc can be chosen by taking the tangent to the curve of Fig. 7
[cf. (lo)] at the origin. This ensures adequate regulation near
the origin for both rectifier and inverter operation. With this V mod, = - J3(V mod I cos 0 . (19)
criterion,
These two differential equations, (16) and (17), represent the
+
K C = ( R 2 X 2 ) / ( 3 KV
. . V .X ) . (13) ac side of the rectifier. The differential equation for the dc
side of the converter is
By replacing (12) and (13) into (7), we get
C . dvc/dt = i l - 12 (20)
Vc =(V/Kv)[cos{(R/X)c~}
where il represents the output rectifier dc current, and 22
+
( X / R )sin { ( R / X ) a }- a] (14) is the forcing function. i2 is also the load dc current. Input
and output equations are related through the power balance
where equation given in (1). Replacing (18) and (19) into (1) and
= 22 . R . [I + ( X / R ) 2 ] / ( 3 K. V~) . because of (6), 21 can be evaluated in terms of id and 2,
CY (15)

Fig. 13 shows the regulation characteristic of the rectifier,

il = - J 3 . Kv(zd . sin 0 + z, . cos 0 ) . (21)
obtained for the linear control through (14). It can be noted The 0 angle depends on the forcing function 22 [cf. (lo)] and
that the regulation may be good enough for some applications then can be considered constant for the analysis. Replacing
in order to justify the simplification of the circuits with respect (4)-(6), (18), and (19) into (16) and (17),
to the zero regulation control. In the linear control, the ROM
look-up table which relates 22 with 0 is replaced by the simple L . did/dt = - R . id +
w . L .iq - Kd . vc (22)
proportional gain Kc given by (12). L . di,/dt = - W . L . i d - R . i, - K, . U, - J3V (23)
where
VII. STABILITY
Kd = - J 3 . K v . sin 0 (24)
Assuming that the frequency of the PWM pattem is high
enough, the harmonics can be neglected. Using the equivalent and
circuit given in Fig. 4 and analyzing the system through the
d-q frame, K , = - J 3 . KV . COS 0 . (25)
L . did/dt + R . id - X .iq = vd - modd (16) There are three unknown values in (22) and (23): id, i,, and
L . di,/dt +X . zd + R . z, = vq - v mod, (17) U,. The third equation is obtained by replacing (21), (24), and
158 IEEE TRANSACTIONS ON POWER ELECTRONICS, VOL. 9, NO. 2, MARCH 1994

9 *SO"

200
190
- 8 -6 - 4 - 2 180

2 4 6 8 10 12
---40" -8 -6 - 4 - 2 DC LOAD CURRENT i 2 [AI
T -50"
Fig. 15. Regulation for linear control.
Fig. 14. 0 - i a characteristic using linear control.

(25) into (20)

c . dv,/dt = Kd . id + K 4 . i, -iz.
I
Equations (22), (23), and (26) can be written in the form:

Z= [A].z+B (27)

where
-RIL w -Kd/L
-W -RIL -Kq/L
KdIC KqIC 0

The stability of the system can be found through the eigen-

values of the [A]-matrix
Fig. 16. Step response from 0 to 6 amps dc: (a) simulation and (b)
det { S [ I ]- [ A ] }= 0 (29) experiment.

or
under inverter operation, the load dc voltage increases. This
characteristic can be improved choosing a better value of Kc.
Fig. 16 shows (a) a simulation and (b) an experiment for a
Applying Routh's criterion ( a j > 0 for j = 0 , 1 , 2 , 3 and step response from 0 to 6 amp dc. Input voltages and currents
ala2 - aOa3 > 0), it is found that the system is stable if are displayed for both cases. The dc voltage was adjusted to
and only if R > 0 and L > 0. These two conditions are 220 V for i 2 = 0, and the ac voltage to 65 V rms (90 V peak).
automatically satisfied because L is necessary for commutation Fig. 17 shows similar results for a power reversal from
and R is always positive. The analysis also shows, unlike rectifier to inverter operation, with the same values of voltages.
other control methods, that the stability of this rectifier is The dc current changes from 5 to -6 amps dc. It can be
independent of the size of the dc capacitor. observed that the converter has a very fast dynamic response,
reaching steady-state during the first cycle.
VIII. SIMULATIONS AND EXPERIMENTAL
RESULTS
A 2-kW 120-250-V dc prototype was implemented and IX. CONCLUSIONS
tested to verify the analyses. The PWM pattem has a carrier A 2-kW PWM voltage source rectifier, controlled by the
frequency of 15 times the fundamental with a modulation load dc current instead of the dc voltage, has been developed
index m = 1. Some of the simulations and experimental and implemented. The rectifier has shown very fast dynamic
results are displayed in the figures cited below. response and stability characteristics which are independent of
Fig. 14 shows the 0 - i 2 curve obtained experimentally to the size of the dc capacitor. The simplifications introduced in
get "zero regulation" with the prototype implemented. It can this method of control with respect to the more conventional
be observed that this curve has a discontinuity in the origin "direct current control" are: a) no need of input current sensors;
due to the change of the value of "R" when the converter b) no need of output voltage sensor; c) an unchangeable and
changes from inverter to rectifier operation and vice versa. unique PWM pattem; and d) power angle directly controlled
Fig. 15 shows the regulation characteristic of the prototype through the dc current load. Other interesting characteristics
when linear control is used. K c from (13) was chosen to take of this rectifier are: 1) zero regulation operation; 2) very
the value of "R" under rectifier operation. As "R" changes simple control blocks; and 3) leading power factor operation
VEAS et al.: A NOVEL LOAD CURRENT CONTROL METHOD FOR A LEADING POWER FACTOR VOLTAGE 159

optimized pulse pattern with PROMC,” IEEE Trans. Ind. Appl., vol.
IA-23, pp. 1001-1009, Nov./Dec. 1987.
[6] J. W. Dixon, A. B. Kulkami, M. Nishimoto, and B. T. Ooi, “Character-
istics of a controlled-current PWM rectifier-inverter link,” IEEE Trans.
Ind. Appl., vol. IA-23, pp. 1022-1028, Nov./Dec. 1987.
[7] J. W. Dixon and B. T. Ooi, “Dynamically stabilized current controlled
SPWM boost type three-phase rectifier,” in IEEE-IAS Annual Meet.,
Oct. 1988.
[8] -, “Indirect current control of a unity power factor simusoidal current
boost type three-phase rectifier,” IEEE Trans. Ind. Electron., vol. IE-35,
pp. 508-515, NOV. 1988.
[9] -, “Series and parallel operation of hysteresis current controlled
PWM rectifiers,” IEEE Trans. Ind. Appl., vol. 25, pp. 644-65 1, July/Aug.
1989.
[IO] J. W. Dixon and D. R. Veas, “Stability analysis and performance
characteristics of an open loop PWM var compensator,” in Proc. IEEE-
IAS Annu. Meet., Seattle, WA, Oct. 1990, pp. 108&1091.

Diego R. Veas was born in Santiago, Chile. He

Fig. 17. Reversal of power from 5 to -6 amps: (a) simulation and (b) received the electrical engineering degree from the
experiment. Catholic University of Chile in 1990.
Since 1990, he has been working at the Catholic
University of Chile, where he is in charge of the Ex-
for all load conditions, providing the adequate adjustment of ternal Service Division of the Electnc Department.
the @OFFSET angle. He also works as an Investigator in the areas of
Some of the drawbacks of this system are: i) the power power electronics and electric traction.
factor is load dependent and, hence, despite the fact that it
can be adjusted to be leading for all load conditions, is not
constant; ii) under rectifier operation, there is a dc current
value, ~ ~ C R I T
beyond
, which it is not possible to get “zero
regulation control;” and iii) in practice, the resistance “R’ is
Juan W. Dixon (M’90) was born in Santiago, Chile.
not constant and, hence, the curves 0 - 22 have to be obtained He received the electncal engineenng degree from
experimentally to get zero regulation. the University of Chile in 1977, and the M.Eng. and
A simpler circuit configuration was implemented for labo- Ph.D. degrees in electrical engineenng, both from
McGill University, Montreal, P.Q., Canada, in 1986
ratory tests using “linear control.” The linear control does not and 1988, respectively.
work with zero regulation but needs only a proportional gain From 1977 to 1979, he worked at Ferrocamles
instead of a ROM look-up table. del Estado, the Chilean national railways company,
as a Chief of the Electncal Locomotives Section.
Finally, all the characteristics of the rectifier were also Since 1979, he has been working at the Catholic
studied with the help of digital computer simulations. University of Chile, where he IS an Associate Pro-
fessor in the Department of Elechcal Engineering, in the areas of power
electronics and electrical machines. His research interests have included
ACKNOWLEDGMENT electric traction, machine drives, frequency changers, high-power rectifiers,
The authors are indebted to M. Navarro for laboratory static var compensators, and active power filters.
assistance.

REFERENCES
Boon-Teck Ooi (S’69-M’71-SM’85) was bom in
[I] H. S. Patel and R. G. Hoft, “Generalized techniques of harmonic Kuala Lumpur, Malaysia. He received the B.Eng
elimination,” IEEE Trans. Ind. Appl., vol. IA-9, pp. 31CL313, 1973. (degree from the University of Adelaide, Australia;
[2] T. Kato, “Precise P W M waveform analysis of inverter for selected the S.M. degree from the Massachusetts Institute
harmonic elimination,” in Proc. IEEE-IAS Annual Meer., Oct. 1986, of Technology, Cambndge; and the Ph.D. degree
pp. 611-616. from McGill University, Montreal, PQ, Canada, all
[3] M. Boost and P. D. Ziogas, “State-of-the-art PWM techniques: A critical in electrical engineenng.
evaluation,” IEEE Trans. Ind. Appl., vol. 24, pp. 271-280, Mar./Apr. He is currently a Professor in the Department
1988. of Electncal Engineenng, McGill University. His
[4] B. T. Ooi, J. C. Salmon, J. W. Dixon, and A. B. Kulkami, “A three- research interests are in the areas of linear motors,
phase controlled current PWM converter with leading power factor,” repulsive magnetic levitation for high-speed ground
IEEE Trans. Ind. Appl., vol. IA-23, pp. 77-84, Jan./Feb. 1987. transportation, HVDC, static var controllers, power electronics, and subsyn-
[5] H. Kohlmeier, 0. Niermeyer, and D. Schroder, “High dynamic four chronous resonance instability in turbogenerators.
quadrant ac motor drive with improved power factor and on-line Dr. 001 is a Registered Engineer in the Province of Quebec.