Copyright (c) 2011 IEEE. Personal use is permitted.

For any other purposes, Permission must be obtained from the IEEE by emailing [email protected].
This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication.
TPEL-Reg-2010-08-0743.R1

1

Abstract This paper presents a new technique to improve the
input power factor of a single-phase rectifier followed by an
inductive filter. It consists in the employment of a two-quadrant
active power filter, based on a conventional bidirectional DC-DC
converter, connected to the output side of the diode bridge. This
technique allows the extension of this rectifier range operation in
the continuous conduction mode. Circuit description, operation
principle, modeling, design and simulation results are presented
in this paper. A laboratory prototype was designed and built,
with the following specifications: input voltage: 220 V
RMS
; output
voltage: 200 V
DC
; rated power: 3 kW; switching frequency: 100
kHz. By the experimental results obtained, the system presented
a unity power factor and an efficiency of 95% at full load, while
the maximum efficiency value was 98% at 2 kW.

Index Terms Active power filter, power factor correction,
single-phase inductive filter rectifier, two-quadrant.
NOMENCLATURE
a Output voltage and input voltage peak ratio
f
C
Active filter capacitance
i
C Current compensator transfer function
o
C Load capacitance
v
C Voltage compensator transfer function
d Duty cyle
med
D Average duty cycle
in
f Input source frequency
N
F Band-stop filter transfer function
s
f Switching frequency
i
G Current sensor gain
Mi
G Current closed loop gain
mult
G Multiplier gain
pwm
G PWM modulator gain
v
G Voltage sensor gain
i
H Current model transfer function
v
H Voltage model transfer function
1p
I Input current peak
2
I Rectified input current
Lf
I Active filter inductor current
( )
Lo
I u Load inductor current
A
In UC3854B multiplier A input
B
In UC3854B multiplier B input
C
In UC3854B multiplier C input
o
I Load output current
o
I Parameterized load output current
oCR
I Critical parameterized load output current
f
L Active filter inductance
o
L Load inductance
oCR
L Critical load inductance
o
R Load resistance
5 pin
R UC3854B pin 5 resistance
1p
V Input voltage peak
o
V Load output voltage
oCCM
V
Load output voltage for continuous conduction mode
operation
out
V UC3854B pin 5 voltage
in
P Input average active power
o
P Load average active power
i
T Compensated current open loop transfer function
v
T Compensated voltage open loop transfer function
o Conduction angle
| Block angle
1MAX
i A
Parameterized maximum input current high frequency
ripple
I. INTRODUCTION
HE increasing of the electronic devices in industry and
residences has created a major concern on the electric
power quality issue, due to the fact that these equipments
draw a current with high harmonic content from the grid. The
devices which present this characteristic are known as
nonlinear loads.
Harmonic mitigation methods have been exhaustively
discussed in the literature [1]-[23], and the active power filters
have been considered a dynamic and adjustable solution to the
power quality issue, [6]- [23]. Shunt active power filters
(APF) consist of a power electronic converter which injects a
compensating current to the grid, with the harmonic contents
equal to the nonlinear load, but with opposite phase, resulting
in a sinusoidal grid current, in phase with the voltage,
A Single-Phase High Power Factor Rectifier,
Based on a Two-Quadrant Shunt Active Filter
Silvia Helena Pini, Ivo Barbi, Fellow Member, IEEE
T
Copyright (c) 2011 IEEE. Personal use is permitted. For any other purposes, Permission must be obtained from the IEEE by emailing [email protected].
This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication.
TPEL-Reg-2010-08-0743.R1

2
achieving an unity input power factor.
The single-phase diode rectifier is a usual nonlinear load,
which is widely used in industry for providing DC loads or
DC-DC converters and inverters. The conventional single-
phase rectifier topology consists of a full-bridge diode rectifier
followed by a capacitive filter, presents an input current with
impulsive characteristic, resulting in a high total harmonic
distortion (THD) and a poor input power factor. Many passive
[3]-[5] and active [11]- [23] methods for the input current
harmonic distortion reduction applied to this rectifier topology
have been already presented in literature.
Another example of usual nonlinear load is the single-phase
rectifier with output inductor, which is commonly used in
industry applications where high output current is demanded.
The addition of an inductor on the diode bridge output allows
the rectifier input current peak reduction and, therefore, its
total distortion harmonic reduction and the input power factor
improvement. Thus, the reactive power flow from this rectifier
is smaller than the conventional one. However, two drawbacks
of this rectifier topology must be emphasized: the inductor
large size and the possible resonance between the output
capacitor and inductor.
For the inductive filter rectifier, a passive method for input
current harmonic mitigation was proposed in [1]. The input
power factor of this topology was improved. Nevertheless, the
rectifier operation was in the discontinuous conduction mode
(DCM), not allowing the output voltage regulation. In [2], an
active power filter was used for input current harmonic
reduction applied to this rectifier topology operating as the
nonlinear load. However, as the APF was connected to the
alternating current (AC) side of the rectifier, for a load value
in which the original rectifier operated in the DCM, the APF
insertion to the system was not capable of changing the
rectifier operation mode, resulting in no output voltage
regulation.
The application of a shunt active power filter, for current
harmonics reduction, applied to an inductive filter rectifier
will process a smaller amount of reactive power than in the
case of a conventional rectifier, if the same active power flow
is taken into account. Then, the APF converter may be
designed for smaller power operation and control effort.
In this paper, a new technique to improve the input power
factor of a single-phase rectifier with output inductor, based
on a two-quadrant shunt active power filter, is proposed [24].
Still, the application of the proposed APF to this rectifier
topology enabled the extension of the rectifier operation range
in the continuous conduction mode (CCM), in other words,
with regulated output voltage.
II. PROPOSED SINGLE-PHASE SHUNT ACTIVE POWER FILTER
The proposed active power filter (APF) topology consists in
a bidirectional DC-DC converter with capacitive energy
storage. This converter presents the same operation stages of a
VSI converter used on conventional single-phase APF
application for the positive half-cycle of input voltage. The
proposed APF converter is connected on the DC side of a
rectifier followed by an inductive filter. The proposed circuit
is depicted on Fig. 1.
The active filter operation results in input side power factor
correction and input current harmonic contents reduction. In
addition, the extension of the operating range on continuous
conduction mode for this rectifier is another motivation for
this proposed APF application.
+
-
+
-
+
-
~
~
S1
S2
2
( )
S
I u
1
( )
S
I u
( )
Cf
I u
f
C
o
I
+
-
2
( ) I u
2
( ) V u
1
( ) I u
1
( ) V u
o
L
o
C
o
R
o
V
f
L
( )
Lo
I u
( )
Lf
I u
cf
V

Fig. 1 Proposed active power filter structure.
A. Load Characteristics Single-Phase Rectifier With
Inductive Filter
The non-unit power factor load employed in this work is
composed by a single-phase rectifier followed by an inductive
filter, as per Fig. 1. It is commonly used on applications that
demand high output current. Also, this rectifier presents a
higher input side power factor than the rectifier followed by a
capacitive filter, since the last one presents an input current
with impulsive characteristic, resulting in a high total
harmonic distortion (THD). Even though, the load introduced
on Fig. 1 contains some disadvantages, as the inductor volume
and the load inductor-capacitor resonance issue.
This rectifier can operate in the continuous conduction
mode (CCM), when the load inductor current never reaches
zero and, consequently, the diode bridge is always conducting.
When operating in this mode, the load output voltage, Vo,
depends only on the effective input voltage, as shown in
expression (1).
1
0, 9.
oCCM ef
V V = (1)
On the other hand, the operation in the discontinuous
conduction mode (DCM) occurs when the load inductor
current goes to zero for a period of time, resulting on the
diodes conduction block. When operating in the DCM, this
rectifier presents a varying load output voltage characteristic
which depends on two parameters: the effective input voltage
and the load current value. The DCM rectifier operation
analysis is detailed on the Appendix section. For very low
current load values, the output voltage can achieve the input
voltage peak.
The main theoretical waveforms for the inductive filter
rectifier on both operation modes are shown in Fig. 2.
However, in order to obtain a constant output load voltage for
several load current values, the operation in the CCM is
Copyright (c) 2011 IEEE. Personal use is permitted. For any other purposes, Permission must be obtained from the IEEE by emailing [email protected].
This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication.
TPEL-Reg-2010-08-0743.R1

3
desired.

Time
0
1
( ) V u
Time
1
( ) V u
0
Time
( )
Lo
I u
0
(a) (d)
(b) (e)
(c) (f)
CCM DCM
Time
1
( ) I u
0
Time
1
( ) I u
0
Time
( )
Lo
I u
0

Fig. 2 Theoretical waveforms for single-phase rectifier with inductive filter
structure CCM: (a) input voltage; (b) input current; (c) load inductor current;
DCM: (d) input voltage; (e) input current; (f) load inductor current.
The load inductor-resistance combination defines the
operation mode. The critical inductance value, that limits both
operation modes as a function of the load resistance value, is
given by (2).
360
o
oCR
R
L
t
= (2)
The parameterized load current is defined by expression (3)
.
120
1
2
o o Hz
o
p o
V L
I
V R
te
= (3)
Substituting the critical inductance value, described by (2),
in (3), the critical parameterized load current expression is
obtained, according to (4),
2
3 3
oCR
I a
t
= = (4)
where:
1
o
p
V
a
V
= (5)
With the CCM and the DCM load output voltage equations,
the load characteristic curve is plotted, as per Fig. 3. This
curve relates the output-input voltages ratio, (5), with the
parameterized load current, (3). Both operation modes regions
are limited by the vertical line, the critical mode operation.
The dotted line indicates the DCM operation, and the
continuous line is related to the CCM one.
o
1
p
V
=
V
a
0.667
oCR
I =

Fig. 3 Load characteristic curve for the single-phase rectifier with inductive
filter.
B. System Characteristics With The Proposed Active Power
Filter
By the exposed, the aims of the proposed APF application
are: input power factor correction, input current harmonic
contents reduction and the extension of the CCM range
operation for the inductive filter rectifier.
One particular point of this proposed structure is the APF
connection on the DC side of the load, unlike most active
power filters, which are connected on the AC side. This
connection leads to a converter operation in two quadrants
(bidirectional in current and unidirectional in voltage), being
pointed as a great advantage in comparison to the APF
converters connected on the AC side (four quadrants).
The two quadrants operation feature of the APF converter
results in its power topology simplification, as can be seen in
Fig. 1, and enables the UC3854B Integrated Circuit (IC)
utilization. This IC is commonly applied in power factor
correction control. Since the internal UC3854B multiplier
operates only in one quadrant, its usage for the four-quadrant
APF is not possible.
The APF operation is accomplished in a way that the
generated APF inductor current is the difference between the
output diode bridge current and the inductor load current,
according to (6). The output diode bridge current, I
2
(), is the
rectified input current. For power factor correction
achievement, the input current must be a sine wave in phase
with the input voltage, thus the output diode bridge current
ought to be a rectified sine. In Fig. 4, the main theoretical
currents waveforms, for the circuit presented in Fig. 1, are
shown along with the input voltage.
2
( ) ( ) ( )
Lf Lo
I I I u u u = (6)
From the proper APF operation, the input current is a sine
Copyright (c) 2011 IEEE. Personal use is permitted. For any other purposes, Permission must be obtained from the IEEE by emailing [email protected].
This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication.
TPEL-Reg-2010-08-0743.R1

4
wave, allowing input side high power factor and low input
current THD. Additionally, the diode bridge is operating in
the CCM, independent from the load current value, ensuring a
load output voltage, Vo, independent of the load power.
Therefore, even for load current values for which the original
inductive filter rectifier operates in the DCM, with the
proposed APF application the system operation in the CCM is
obtained, resulting in the extension of the CCM range
operation.
The duty cycle expression for the APF converter is given
by (7).
( ) 1 sin( ) d a u u = (7)

(a) (b)
( )
Lo
I u
Time
1
( ) V u
(c)
( )
Lf
I u
Time
1
( ) V u
(d)
0 0
2
( ) I u
Time
1
( ) V u
0
1
( ) V u
1
( ) I u
Time
0

Fig. 4 Main waveforms for the proposed active power filter structure (a)
input voltage and current; (b) input voltage and rectified input current; (c)
input voltage and inductor load current; (d) input voltage and inductor active
power filter current.
III. SYSTEM MODELING AND CONTROL STRATEGY
There are two main control strategies used in single-phase
active power filter control: the active power filter current
control and input current control. Both strategies require
voltage and current control. The APF DC-link voltage (V
cf
)
control is responsible for the system active power flow
control, since the APF active power flow must be zero,
besides the components losses. The current control is
responsible for the input current shape control, which must be
a sine waveform in phase with the input voltage.
A. Control Strategy
The control strategy applied here is based on the average
current mode control [25]-[26], using the rectified input
current as the control variable, which is similar to the input
current control strategy [11]-[14]. This strategy is effective
and simple, because it uses only one current sensor, and load
current harmonic contents calculation is not necessary. The
block diagram of the proposed control strategy applied to the
power circuit is presented in Fig. 5.

Copyright (c) 2011 IEEE. Personal use is permitted. For any other purposes, Permission must be obtained from the IEEE by emailing [email protected].
This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication.
TPEL-Reg-2010-08-0743.R1

5

( )
I
C s
PWM
G
f
C
o
I
sync
V ()
1
V()
1
I ()
2
I ()
Lo
I ()
osense
I
o
V
Lf
I ()
cfsense
V
cf
V
2 sense
I ()
f
L
o
L
o
C
o
R
vea
V
2 REF
I ()
I
c
2
() V
S2
CMD
S1
CMD
Io
K
()
a
V
( )
V
C s
( )
N
F s
cfREF
V
v
c

Fig. 5 Block diagram of the proposed control strategy
A few changes on this strategy were needed due to the
inductor-capacitor load resonance issue. For solving this
problem, a band-stop filter, F
N
(s), with the rejection center
frequency tuned in the load resonance frequency, was used in
the APF DC-link voltage monitoring. Therefore, the signal
applied on voltage loop control presented its resonance
frequency oscillations attenuated, not interfering on the
system control. The voltage controller also had to be
configured with a low cutoff frequency, resulting in an active
power flow control strongly slow. By that, the transient
voltage and current oscillations were increased, exceeding
design limits and compromising the system operation. For
solving the dynamic active power flow control problem, a
feedforward control of the load output current was necessary,
as represented in the block diagram shown in Fig. 5.
Due to the low components losses in the APF converter, it
is considered that the load average active power is the same as
the input one, (8) and (9). Then, using expression (1), the
average load output current is related to the input current peak
by expression (10). From this expression, it is seen that the
system active power flow can be controlled through the load
output current value. Expression (11) shows the constant gain
used in the feedforward output current loop for the system
active power flow control. In addition, the APF DC-link
voltage control is used for APF losses compensation.

o in
P P = (8)
1 1
.
.
2
p p
o o
V I
V I = (9)
1
4
o p
I I
t
= (10)
4
Io
K
t
= (11)
A. Current Transfer Function Modeling
After established the system control strategy, current and
voltage transfer functions models must be defined.
As the current loop control must be fast, since it is
responsible for the input current waveform control, the
obtained rectified input current model is high frequency
defined, being based on small-signal average current
modeling, for switching periods, [27]. The APF converter
equivalent circuit, in which the variables are the average
values in one switching period, is shown in Fig. 6, in which
the average voltage V
a
() is given by (12). From this
equivalent circuit, the rectified input current transfer function
model is obtained, according to (13).
( ) ( ) 1 ( )
a cf
V d V u u = (12)

Copyright (c) 2011 IEEE. Personal use is permitted. For any other purposes, Permission must be obtained from the IEEE by emailing [email protected].
This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication.
TPEL-Reg-2010-08-0743.R1

6

2
( ) V u
( )
Lf
I u
f
L
( )
a
V u

Fig. 6 Equivalent circuit for current transfer function modeling.
2
( )
( )
cf
f
V
I s
d s sL
A
=
A
(13)
B. Voltage Transfer Function Modeling
On the other hand, the APF DC-link voltage control must
be slow, even slower than the input source frequency. Its
transfer function model is obtained by the average values of
the APF variables accounted in one input source half-period.
Fig. 6 presents the equivalent circuit for voltage transfer
function modeling, resulting in expression (14), in which each
variable is defined for its average value in one input source
half-period.

2
( )
S
I u 2
( ) V u
( )
Lf
I u
1
( )
S
I u
( )
Cf
I u
f
C
cf
V

Fig. 7 Equivalent circuit for voltage transfer function modeling.

( )
1
( )
cf
med
Lf f
V s
D
I s sC
A

=
A
(14)
0
1
( )
med
D d d
t
u u
t
=
}
(15)
IV. PROTOTYPE DESIGN
For theoretical validation, a prototype implementation was
made. Table I shows the design specifications.












Table I Design specifications.
Specifications Value
Input Effective Voltage (V
1rms
) 220 V
APF DC-link Voltage (V
cf
) 400 V
Load Active Power (P
o
) 3 kW
Maximum Input Current Ripple
(i
1MAX
)
0.7A
Input Source Frequency (f
in
) 60 Hz
Switching Frequency (f
s
) 100 kHz

A. Power Components Determination
Load inductance, L
o
, was designed for original inductive
filter rectifier CCM operation from 30% to 100% of the
nominal load value, using expression (2). The APF
inductance, L
f
, was designed based on the maximum input
current high frequency ripple value, defines in Table I. The
expression (16) defines the APF inductance value, from the
specified maximum input current high frequency ripple value.
Load and APF capacitors, C
o
and C
f
, were designed for low
DC-link voltage ripple and their choices were limited by the
effective current value. The passive parameters components
values are shown in Table II.
1
1
0.32
1.4 mH
p
f
s MAX
V
L
f i

= =
A
(16)
Table II Passive parameters values
Parameter Value
Load Inductor (L
o
) 30 mH
Load Capacitor (C
o
) 4700 F
APF Inductor (L
f
) 1.4 mH
APF Capacitor (C
f
) 4700 F

The SKD30/08 diode bridge model from Semikron
manufacturer and the IRFP31N50L Mosfet model from
International Rectifier manufacturer were employed.
B. Control System Design
With the passive parameters values defined, the Bode
diagram of the current and voltage transfer functions models
are plotted. In Fig. 8, the current model Bode diagram is
Copyright (c) 2011 IEEE. Personal use is permitted. For any other purposes, Permission must be obtained from the IEEE by emailing [email protected].
This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication.
TPEL-Reg-2010-08-0743.R1

7
presented.

10 100 1 10
3
1 10
4
1 10
5

0
50
100
90.1
90.05
90
89.95
89.9
(Hz) frequency
( ) (dB)
i
H s ( ) (degrees)
i
H s

Fig. 8 Current model Bode diagram.
The current compensator chosen is a proportional integrator
controller with filter. Its cutoff frequency was established in
10kHz. Therefore, the designed current controller transfer
function is given by (17). In Fig. 9, the current controller
Bode diagram is shown.
( )
3
5
3
2 5 10
( ) 9.5 10
2 20 10
i
s
C s
s s
t
t
+
=
+
(17)

10 100 1 10
3
1 10
4
1 10
5

0
20
40
60
80
100
80
60
40
( ) (dB)
i
C s
(Hz) frequency
( ) (degrees)
i
C s

Fig. 9 Current controller Bode diagram.
Expression (18) represents the open loop compensated
current system model, having its Bode diagram depicted in
Fig. 10. The phase margin of this compensated system is
about 37, ensuring the system stability.
( ) ( ) ( )
i i pwm i i
T s H s G GC s = (18)

10 100 1 10
3
1 10
4
1 10
5

0
50
100
180
160
140
120
( ) (dB)
i
T s
(Hz) frequency
( ) (degrees)
i
T s

Fig. 10 Open loop compensated current system model.
Fig. 11 presents the Bode diagram of the voltage transfer
function model. For the voltage loop modeling, the band-stop
filter transfer function was needed, and its Bode diagram is
shown in Fig. 12.
1 10 100 1 10
3

40
20
0
20
90.1
90.05
90
89.95
89.9
(Hz) frequency
( ) (dB)
v
H s ( ) (degrees)
v
H s

Fig. 11 Voltage model Bode diagram.
0.1 1 10 100 1 10
3

30
20
10
0
100
50
0
50
100
(Hz) frequency
( ) (dB)
N
F s
( ) (degrees)
N
F s

Fig. 12 Band-stop filter Bode diagram.
The same controller type was chosen for voltage regulation,
and its transfer function is presented in (19), followed by its
respective Bode diagram, in Fig. 13.

( )
2 0.119
( ) 17.33
2 2.382
v
s
C s
s s
t
t
+
=
+
(19)
1 10 100 1 10
3

60
40
20
100
80
60
40
20
0
(Hz) frequency
( ) (dB)
v
C s ( ) (degrees)
v
C s

Fig. 13 Voltage controller Bode diagram.
Expression (20) represents the open loop compensated
voltage system model, which has its Bode diagram presented
in Fig. 14. The phase margin of this compensated system is
about 52, ensuring the system stability.
( ) ( ) ( ) ( )
v v mult Mi v N v
T s H s G G G F s C s = (20)
Copyright (c) 2011 IEEE. Personal use is permitted. For any other purposes, Permission must be obtained from the IEEE by emailing [email protected].
This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication.
TPEL-Reg-2010-08-0743.R1

8

(Hz) frequency
( ) (dB)
v
T s ( ) (degrees)
v
T s
1 10 100 1 10
3

100
50
0
200
100
0
100
200

Fig. 14 Open loop compensated voltage system model.

C. UC3854B Usage For Input Current Control
As mentioned previously, one of the main highlights of this
work is the UC3854B Integrated Circuit usage for the system
current control. Unlike most APF applications, this active
filter is connected on the DC side of the load, and then the
APF converter operates only in two quadrants, allowing this
IC application.
Among existing ICs that accomplish average current model
control for power factor correction, the UC3854B is quiet
required.
Some of this IC key features were used [28]: over current
protection, PWM modulator, two quadrants analog
multiplier/divider and the current amplifier. As the active
power flow control is made by the sum of the APF DC-link
voltage regulation signal and the load output current loop
signal. The resulting signal is generated externally from the
UC3854B and forced into the A input of its internal
multiplier. The multiplier B input is a current signal
responsible for the shape of the controlled current. It must be
an image of the rectified input voltage. The last multiplier
input, C, is for the effective input voltage feed-forward
control. The multiplier output is the current reference signal
for the rectified input current control.
The output multiplier voltage signal is defined by (21).
5 2
( ) ( 1, 5 )
( )
m B A
out pin
C
K In In V
V R
In
u
u

= (21)
For correct current reference configuration, the generated
multiplier A input must be given by expression (22).
2
5
4
1.5
C
A
pin B peak
In
In
R In t

= + (22)
V. EXPERIMENTAL RESULTS
The implemented system was tested for different load
values. The main waveforms obtained from this prototype are
presented in this section.
Fig. 15 shows the input voltage and current waveforms and,
also, the load output voltage for the inductive filter rectifier
operation without and with the proposed APF Fig. 15 (a)
and Fig. 15 (b), respectively in nominal load value. For this
load value, the original rectifier operates in CCM, therefore
the load output voltage remains the same in both cases.
However, with the APF operation, the input power factor
correction is achieved from 0,88 to 1 and the input current
harmonic contents are reduced from 43% to 4% as
expected.
1
( ) V u
1
( ) I u
o
V
1
( ) V u
1
( ) I u
o
V
(a) (b)
20ms
Time (s) Time (s)
20ms

Fig. 15 Steady state waveforms for nominal load (a) input voltage
(200V/div) and current (20A/div), output voltage (100V/div) for load
operation; (b) input voltage (200V/div) and current (20A/div), output voltage
(100V/div) for load operation with the proposed APF.
In Fig. 16 the main system current waveforms details, along
with the input voltage, are depicted, proving the theoretical
waveforms presented in Fig. 4.
1
( ) V u
1
( ) I u
1
( ) V u
2
( ) I u
1
( ) V u
( )
Lo
I u
1
( ) V u
( )
Lf
I u
(a) (b)
(c) (d)
Time (s)
4ms 4ms
4ms 4ms
Time (s)
Time (s)
Time (s)

Fig. 16 Steady state waveforms for the rectifier operation with the proposed
APF (Scales: 100V/div, 10A/div) (a) input voltage and current; (b) input
voltage and rectified input current; (c) input voltage and inductor load current;
(d) input voltage and inductor filter current.
The input current harmonic spectrum for the load operation
with and without the proposed APF, in nominal load value,
are presented in Fig. 17. It can be seen that the input current
harmonic contents were substantially reduced.

Copyright (c) 2011 IEEE. Personal use is permitted. For any other purposes, Permission must be obtained from the IEEE by emailing [email protected].
This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication.
TPEL-Reg-2010-08-0743.R1

9

Harmonic Component (n)

(b)
(a)
Harmonic Component (n)

Fig. 17 Input current harmonic spectrum, in nominal load value, for the
rectifier operation (a) without the APF; (b) with the proposed APF.
The input voltage and current waveforms, simultaneously
with the load output voltage, for 10% of nominal load value
operation, are presented in Fig. 18.

(a) (b)
1
( ) V u
1
( ) I u
o
V
1
( ) V u
1
( ) I u
o
V
Time (s)
20ms
Time (s)
20ms

Fig. 18 - Steady state waveforms for 10% of nominal load (a) input voltage
(200V/div) and current (10A/div), output voltage (100V/div) for load
operation; (b) input voltage (200V/div) and current (10A/div), output voltage
(100V/div) for load operation with the APF.
Again, waveforms for the original inductive filter rectifier
operation without and with the proposed APF Fig. 18 (a)
and Fig. 18 (b), respectively are presented. For this load
value, the original rectifier operated in DCM, and the load
output voltage was about 240V. With the APF operation,
besides the input power factor correction and the input current
harmonic contents reduction, the load output voltage was
regulated close to 198V, since the rectifier is operating in
CCM.
A load disturbance, from 65% to 100% of the nominal load
value, was applied to the system, with the APF connection, in
order to check the system dynamic response and stability. The
Fig. 19 shows the input voltage and current waveforms along
with APF and load output DC-link voltages. It can be seen
that the system continues operating as expected, the input
current remains a sine waveform. Both DC-link voltages
present small oscillations, and the system remains stable.
(a) (b)
1
( ) V u
1
( ) I u
cf
V
o
V
Time (s) Time (s)
20ms 20ms

Fig. 19 Main waveforms for load disturbance (a) input voltage (100V/div)
and current (10A/div); (b) output voltage (100V/div) and APF DC-link
voltage (100V/div) from 65% to 100% of the nominal load.
Moreover, the input current total harmonic distortion and
input power factor curves, obtained experimentally for load
values from 10% to 100% of the nominal load value, are
presented in Fig. 20 and Fig. 21, for the rectifier operation
with and without the proposed APF dotted and continuous
curves, respectively. With these curves, it is possible to
conclude that the active power filter had a satisfactory
operation for a wide load values range. For every tested case,
input power factor and input current THD were improved with
the APF application.
0
5
10
15
20
25
30
35
40
45
50
10 20 30 60 70 100
Load Load+APF
Load (% of Nominal)
T
o
t
a
l

H
a
r
m
o
n
i
c

D
i
s
t
o
r
t
i
o
n

(
%
)

Fig. 20 Experimentally obtained input current THD curves rectifier
operation with and without the APF for different load values.

Copyright (c) 2011 IEEE. Personal use is permitted. For any other purposes, Permission must be obtained from the IEEE by emailing [email protected].
This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication.
TPEL-Reg-2010-08-0743.R1

10

0
0,1
0,2
0,3
0,4
0,5
0,6
0,7
0,8
0,9
1
1,1
10 20 30 60 70 100
Load Load+APF
P
o
w
e
r

F
a
c
t
o
r
Load (% of Nominal)

Fig. 21 Experimentally obtained input power factor curves rectifier
operation with and without the APF for different load values.
Finally, the tendency curves for the efficiency of the
prototype obtained experimentally, for the rectifier operation
with and without the APF dotted and continuous curves,
respectively are presented in Fig. 22, for different load
values. The converter efficiency was not a focus of this work
and, yet, the efficiency results with the APF converter
operation are not away from the ones obtained for the original
load, operating from 30% to 100% of the nominal load value.
For lower loads, the switching effects on the system losses are
more substantial.

60
65
70
75
80
85
90
95
100
10 20 30 60 70 100
E
f
f
i
c
i
e
n
c
y

(
%
)
Load (% of Nominal)

Fig. 22 Tendency curves for the efficiency results obtained experimentally
rectifier operation with and without the APF for different load values.
VI. CONCLUSION
This paper has presented an active power filter topology,
based on a bidirectional DC-DC two-quadrant converter,
applied for the power factor correction of a single-phase
rectifier followed by an inductive filter. Theoretical analyses
were detailed, summarizing the main load characteristics and
the load operation with the APF. Moreover, the proposed
control strategy, system modeling and prototype design were
shown.
Finally, the experimental results, obtained by the
implemented prototype, validated the theoretical analysis.
Unity power factor and input current THD reduction were
achieved for nominal load value, and both were improved for
a large range of load values. In addition, the APF application
allowed the rectifier CCM operation extension, presenting an
output voltage load regulation for a wide load value range.
APPENDIX A
ANALYSIS OF THE SINGLE-PHASE RECTIFIER WITH INDUCTIVE
FILTER
The single-phase rectifier followed by an inductive filter,
for which the equivalent circuit is presented in Fig. 23, when
operating in the DCM, has its output characteristic guided by
a non-linear function of the load current and the input voltage.
The DCM operation analysis for this rectifier is detailed in
this section.
The rectifier stages operation described below are made for
the positive input voltage half-cycle, in which diodes D1 and
D4 conduct. For the negative input voltage half-cycle, the
obtained results are similar, but diodes D2 and D3 are
operating. The load output voltage ripple is disregarded for
this analysis.
1) First Stage (,): This stage is divided in two distinct
angular intervals. In the first one, (,), the input voltage
amplitude is higher than the load output voltage, and then
diodes D1 and D4 are conducting. The load inductor
voltage is positive and its current increases, as shown by
the waveforms in Fig. 24. Fig. 23 (a) presents the circuit
operation for this stage. In the second interval, (,), the
input voltage amplitude becomes smaller than the load
output voltage. Now, the inductor voltage is negative,
resulting in its current decreasing. But, as the inductor
current is still positive, the diodes D1 and D4 are still
conducting.
2) Second Stage (,+): The beginning of this stage occurs
when the inductor current reaches the zero value. Then,
the diodes D1 and D4 are blocked and this current
remains zero. The circuit presented on Fig. 23 (b) shows
the inductor filter rectifier operation on this stage.
1
( ) V u
2
( ) V u
( )
Lo
V u
o
L
( )
Lo
I u
1
D
2
D
4
D
3
D
o
V
o
I
1
( ) V u
2
( ) V u
( )
Lo
V u
o
L
( )
Lo
I u
1
D
2
D
4
D
3
D
o
V
o
I

Fig. 23 DCM inductive filter rectifier operation (a) first stage; (b) second
stage.
The main waveforms for this rectifier DMC operation
input voltage and current, output rectifier voltage and load
output voltage, load inductor current are shown in Fig. 24.
Copyright (c) 2011 IEEE. Personal use is permitted. For any other purposes, Permission must be obtained from the IEEE by emailing [email protected].
This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication.
TPEL-Reg-2010-08-0743.R1

11
1
( ) V u
1
( ) I u
0
2
( ) V u
o
V
0
0
o
|
t 2t
o t +
( )
Lo
I u
0

Fig. 24 DCM inductive filter rectifier operation main waveforms: input
voltage and current, rectifier output voltage and load output voltage, load
inductor current.
The featured angles and in Fig. 24 represent the
conduction and the block angles, which refer to the instant
when the inductor current starts to be conducted and the
instant when this current reaches the zero value, respectively.
By Fig. 24, it is possible to obtain an important relationship
between the input voltage and the load output voltage, given
by ( 23).
1
sin( )
o p
V V o = ( 23)
A. Load Inductor Current and Output Voltage Expressions
For this rectifier DCM operation, the inductor current
expression analysis depends on the block angle . When >
, the analysis must be split in two different systems, based on
the conduction interval [,]. If < , the analysis is restricted
only to the first system.
1) System 1: < .
The following expression, ( 24), is obtained from the
rectifier equivalent circuit, Fig. 23:
1
( )
sin( )
p
Lo o
o o
V
dI t V
t
dt L L
e
e = ( 24)
Integrating expression ( 24) results in ( 25).
1 1
1
( ) cos( )
p
o
Lo S i S
o o
V
V
I K
L L
u u u
e e

= + ( 25)
The constant Ki-Si can be found by the following system
initial condition, ( 26):
( ) 0
Lo
I o = ( 26)
Thus, the constant expression results in ( 27).
1
1
cos( )
p
o
i S
o o
V
V
K
L L
o o
e e

= + ( 27)
Then, the load inductor final expression for the rectifier
operation on System 1 region is given by ( 28).
( )
1
1
( ) cos( ) cos( )
( )
p
Lo S
o
o
o
V
I
L
V
L
u o u
e
o u
e

= +
+
( 28)
For < , the load output voltage expression is found,
knowing that the inductor current is zero for the block angle,
as given by ( 29). Then, the load output voltage expression is
represented by ( 30).
( ) 0
Lo
I | = ( 29)
1
1
cos( ) cos( )
o S p
V V
o |
| o

( 30)
2) System 2: < < +.
As the input voltage is negative for this system, the
following expression is obtained, ( 31):
1
( )
sin( )
p
Lo o
o o
V
dI t V
t
dt L L
e
e = ( 31)
Integrating expression ( 31) results in the inductor current
expression, ( 32).
2 2
1
( ) cos( )
p
o
Lo S i S
o o
V
V
I K
L L
u u u
e e

= + ( 32)
The following initial condition, ( 33), is used for the
constant K
i-S2
expression determination:
2 1
( ) ( )
Lo S Lo S
I I t t

= ( 33)
( )
1
1
( ) cos( ) 1 ( )
p
o
Lo S
o o
V
V
I
L L
t o o t
e e

= + + ( 34)
Therefore, the resulting inductor current expression for the
rectifier operation in System 2 region is given by ( 35).
( )
2
1
( ) cos( ) cos( ) 2
( )
p
Lo S
o
o
o
V
I
L
V
L
u u o
e
o u
e

= + + +
+
( 35)
Again, knowing that I
Lo-S2
() = 0, load output voltage
expression is obtained, ( 36).
Copyright (c) 2011 IEEE. Personal use is permitted. For any other purposes, Permission must be obtained from the IEEE by emailing [email protected].
This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication.
TPEL-Reg-2010-08-0743.R1

12
2 1
cos( ) cos( ) 2
o S p
V V
o |
| o

+ +
=

( 36)
B. Average Load Inductor Current Expression
By Fig. 23, it can be seen that load average current is given
by ( 37).
o Lo
I I = ( 37)
Again, load inductor average current expression is defined
for two different systems, which depend on the block angle .
1) System 1: < .
( )
1 1
1
2 2
1
( )
1
( ) cos( ) sin( ) sin( )
( )
2 2
o S Lo S
p
o
o
o
I I d
V
L
V
L
|
o
u u
t
| o o | o
t e
| o
| o o
e

= =

= + +

( | |
+ +
( |
\ .
}
( 38)
2) System 2: < < +.
( )
2 1 2
1
2 2
1
( ) ( )
1
sin( ) sin( ) ( ) cos( ) 2( )
( )
2 2
o S Lo S Lo S
p
o
o
o
I I d I d
V
L
V
L
t t o
o t
u u u u
t
o | | o o | o
t e
| o
| o o
e
+

(
= + =
(

= + + + +

( | |
+ +
( |
\ .
} }
( 39)
From the load average current expressions, ( 38) and ( 39),
the parameterized output current expression is defined, ( 40).
Thus, the parameterized load average current expressions for
System 1 and 2 regions are defined by ( 41) and ( 42),
respectively.
1
o o
o
p
I L
I
V
te
= ( 40)
1
2 2
cos( )( ) sin( ) sin( )
( )
2 2
o S
I
a
o | o | o
| o
o | o

= + +
| |
+ +
|
\ .
( 41)
2
2 2
cos( )( ) sin( ) sin( ) 2( )
( )
2 2
o S
I
a
o | o | o | o
| o
o | o

= + + + +
| |
+ +
|
\ .
( 42)
As shown before, the parameterized critical output current
expression is given by expression (4). Therefore, for
o oCR
I I <
the rectifier is operating in the discontinuous conduction
mode.
C. DCM Rectifier Output Characteristic
Gathering the expression ( 23) which is valid for both
system regions and relates the load output and the input
voltages , the output load voltage expressions ( 30) for
System 1 and ( 36) for System 2 , and the parameterized
output load current expressions ( 41) and ( 42), for System 1
and 2, respectively , two non-linear systems are obtained,
( 43) and ( 44). By these systems equations, the load
characteristic curve for the rectifier DCM operation is
generated, as shown in Fig. 3.
1) System 1: < .
( )
( )
2 2
sin( ) 0;
cos( ) cos( )
0;
cos( )( ) sin( ) sin( ) ( ) 0
2 2
o
a
a
I a
o
o |
| o
| o
o | o | o o | o

| |
+ + =
|

\ .
( 43)



2) System 2: < < +.
( )
( )
2 2
sin( ) 0;
cos( ) cos( ) 2
0;
cos( )( ) sin( ) sin( ) 2( ) ( ) 0
2 2
o
a
a
I a
o
o |
| o
| o
o | o | o | o o | o

| |
+ + + + =
|

\ .
( 44)
APPENDIX B
IMPLEMENTED CIRCUIT SCHEMATIC
The power and control stage schematic, for the
implemented prototype, is depicted in Fig. 25. The resistances
and capacitances values presented in the following schematic
consider the association of commercial components.

Copyright (c) 2011 IEEE. Personal use is permitted. For any other purposes, Permission must be obtained from the IEEE by emailing [email protected].
This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication.
TPEL-Reg-2010-08-0743.R1

13

100 nF
820 k
820 k
10 k
100 nF 22 k
22 k
10 k
10 k
470 k
470 k
560 k 2.2 F
100 nF
100 nF
2.2 F 560 k
4.5k
4.5 k 10 k 100 k 10 k
10 k
100 k
10 k
100 k
47 k
100 k
LF347
LF347
LF347
LF347 LF347
LF347
cfREF
V 1 V =
cfN
V
cv
V
+15 V
vea
V
5 4 3 11 7
16
1 12 14 13 10 15
8
6
9
2
100 k
100 k
47 k
3.9 k
10 k
22 k
1.5 M
820 k
18 k
12 k
8.2 k
47 k
22 k
8.2 k
10 k
470 pF
100 nF
680 nF
1 F
100 nF
1 nF 1 pF
1 nF
680 pF
220 pF
+15 V
220V:220V
820 k
1N4006
1
V
S
S2
V
HEX 4069UB
RMS
220 V
SKB30/08
30 mH
4700 F
4700 F
13 O
1.4 mH
LA 55-P
LA 55-P
osense
I
2sense
I
cfsense
V
IRFP31N50L
330 k
470
1 k
LF347
1
x
1
y
2
x
2
y
1
x
1
y
2
x
2
y
IRFP31N50L
1
V
gs S
2
V
gs S

Fig. 25 - Implemented prototype power and control stages schematic

REFERENCES
[1] S. B. Dewan, Optimum Input and Output Filters for a Single-Phase
Rectifier Power Supply, IEEE Transactions on Industry Applications,
vol. IA-17, no. 3, pp. 282-288, May/June 1981.
[2] J. Choi, G. Park, S. B. Dewan, Standby Power Supply with Active
Power Filter Ability Using Digital Controller, in Proc. IEEE Applied
Power Electronics Conf., 1995, vol. 2, pp. 783-789.
[3] A. R. Prasad, P. D. Ziogas, S. Manias, A Novel Passive Waveshaping
Method for Single-Phase Diode Rectifiers, IEEE Transactions on
Industrial Electronics, vol. 37, no. 6, pp. 521-530, December 1990.
[4] J. Yanchao, L. Xiaobing, L. Zhuo, J. Jisheng, L. Xinhua, An Improved
Passive Input Current Waveshaping Method for Single-Phase Diode
Rectifier, in Proc. IEEE International Conference on Industrial
Electronics, Control, and Instrumentation, 1996, vol. 2, pp. 695 699
[5] H. A. Kazem, An Improved Method of Passive Input Current
Waveshaping for Single-Phase Rectifier, IEEE Compatibility in Power
Electronics, pp. 1-4, 2007.
[6] H. Akagi, Active Harmonic Filters, Proceedings of the IEEE, vol. 93,
no. 12, pp. 2128-2141, December 2005.
[7] F. Z. Peng, Application Issues of Active Power Filters, IEEE Industry
Application Magazine, vol. 4, pp. 21-30, September/October 1998.
[8] B. Singh, K. Al-Haddad, A Review of Active Filters for Power Quality
Improvement, IEEE Transactions on Industrial Electronics, vol. 46,
no.5, pp. 960-971, October 1999.
[9] F. Z. Peng, H. Akagi, A. Nabae, A Study of Active Power Filters Using
Quad-Series Voltage-Source PWM Converters for Harmonic
Compensation, IEEE Transactions on Power Electronics, vol. 5, no. 1,
pp. 9-15, January 1990.
[10] F. Z. Peng, H. Akagi, A. Nabae, A New Approach to Harmonic
Compensation in Power System, in IEEE Industry Applications Society
Annual Meeting, 1988, vol. 1, pp. 874-880.
[11] F. Pttker, I. Barbi, Power Factor Correction on Non-Linear Loads
Employing a Single Phase Active Power Filter: Control Strategy, Design
Methodology and Experimentation, in IEEE Power Electronics
Specialists Conference, 1997, vol. 1, pp. 412-417.
[12] F. Pttker, I. Barbi, Power Factor Correction on Non-Linear Load
Employing a Single Phase Active Power Filter Based on Full-Bridge
Current Source Inverter Controlled Through the Sensor of the AC Mains
Current in IEEE Power Electronics Specialists Conference, 1999, vol.
1, pp. 387-392.
Copyright (c) 2011 IEEE. Personal use is permitted. For any other purposes, Permission must be obtained from the IEEE by emailing [email protected].
This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication.
TPEL-Reg-2010-08-0743.R1

14
[13] F. Pttker, I. Barbi, Single-Phase Active Power Filters for Distributed
Power Factor Correction in IEEE Power Electronics Specialists
Conference, 2000, vol. 1, pp. 500-505.
[14] D. Lindeke, S. A. Mussa, F. Pttker, and I. Barbi, Single Phase Active
Power Filter Controlled with a Digital Signal Processor DSP, in IEEE
Power Electronics Specialists Conference, 2004, vol. 2, pp. 1024-1027.
[15] Duke, R. M., Round, S. D., The Steady-State Performance of a
Controlled Current Active Filter, in IEEE Transactions on Power
Electronics, vol. 8, no. 3, pp 140-146, April 1993.
[16] Torrey, D. A., Al-Zamel, A. M. A. M., Single-Phase Active Power
Filters for Multiples Nonlinear Loads, in IEEE Transactions on Power
Electronics, vol. 10, no. 3, pp 263-272, May 1995.
[17] Akagi, H., Trends in Active Power Line Conditioners, in IEEE
Transactions on Power Electronics, vol. 9, no. 3, pp 263-269, May
1993.
[18] Kobayashi, Y., Funato, H., Current Control Method Based on
Hysteresis Control Suitable for Single Phase Active Filter with LC
Output Filter, in IEEE Power Electronics and Motion Control
Conference, 2008, pp 479-484.
[19] Hamzah, M. K., Ghafar, A. F. A., Hussain, M. N. M., Single-phase
Half-bridge Shunt Active Power Filter Employing Fuzzy Logic
Control, in IEEE Power Electronics Specialists Conference, 2008, pp
552-558.
[20] Rahmani, S., Al-Haddad, K., Kanaan, H. Y., Two PWM techniques for
single-phase shunt active power filters employing a direct current
control strategy, in IET Power Electronics, vol. 1, no. 3, pp 376-385,
September 2008.
[21] Matas, J., de Vicuna, L. G., Miret, J., Guerrero, J. M., Castilla, M.,
Feedback Linearization of a Single-Phase Active Power Filter via
Sliding Mode Control, in IEEE Transactions on Power Electronics,
vol. 23, no. 1, pp 116-125, January 2008.
[22] Fujita, H., A Single-Phase Active Filter Using an H-Bridge PWM
Converter With a Sampling Frequency Quadruple of the Switching
Frequency, in IEEE Transactions on Power Electronics, vol. 24, no. 4,
pp 934-941, April 2009.
[23] Zeng, F. P., Tan, G. H., Wang, J. Z., Ji, Y. C., Novel single-phase five-
level voltage-source inverter for the shunt active power filter, in IET
Power Electronics, vol. 3, no. 4, pp 480-489, July 2010.
[24] Pini, S. H., Study and Implementation of a Two-Quadrant Shunt Active
Power Filter Connected on the DC Side of a Rectifier with Inductor
Filter (in Portuguese), Florianopolis, 2010. Masters degree
dissertation, INEP, UFSC.
[25] L. H. Dixon, High power factor preregulators for off-line power
supplies, in Unitrode Seminar Proceedings, Sec. 12, pp.1-16, 1990.
[26] F. A. Huliehel, F. C. Lee, B. H. Cho, Small-signal modeling of the
single-phase boost high power factor converter with constant frequency
control, in IEEE Power Electronics Specialists Conference, 1992, vol.
1, pp. 475-482.
[27] R. W. Erickson, Fundamentals of Power Electronics. International
Thomson Publishing, 1997.
[28] P. C. Todd, UC3854 Controlled Power Factor Correction Circuit
Design, Unitrode Products & Application Handbook, 1995.