Prace Naukowe Instytutu Maszyn, Napędów i Pomiarów Elektrycznych

Nr 63 Politechniki Wrocławskiej Nr 63

Studia i Materiały Nr 29 2009
PWM rectifier, AC/DC converter,
sliding-mode control, analysis
Michał KNAPCZYK
*
, Krzysztof PIEŃKOWSKI
**

VARIABLE STRUCTURE CONTROL OF PWM RECTIFIER
WITH SLIDING-MODE-BASED MODULATION
AND MINIMIZED NUMBER OF SENSORS
The paper presents the novel approach to the sliding-mode-based modulation
technique in application to the PWM rectifier with the minimized number of the
necessary voltage and current transducers. The crucial and technically most inconvenient
part of this sensorless technique is the preliminary reconstruction of the line three-phase
currents based on the proper sampling of the DC-link current within the specified PWM
periods. The reconstructed grid currents are then used to estimate their smooth
counterparts as well as the grid three-phase voltages. This can be realized through the
application of the sliding-mode current observer for the grid voltage mixed with the
conventional nonlinear asymptotic observer for the grid voltage. The novel approach to
the design of the sliding-mode control for the PWM rectifier is based on the application
of the nonlinear current controllers in the d-q coordinate frame and the direct look-up
table-based modulation. This solution allows eliminating the back transformation into the
three-phase coordinate system. The selected computational results based on the multi-
rate simulations have been presented and discussed.
1. INTRODUCTION
The three-phase two-level PWM rectifiers are greedily applied in the modern
frequency converters since they provide the four-quadrant operation with the
__________
*
3CapTechnologies GmbH, Germany, michal.knapczyk@3cap.de;
**
Politechnika Wrocławska, Instytut Maszyn, Napędów i Pomiarów Elektrycznych, 50-370
Wrocław, ul. Smoluchowskiego 19, krzysztof.pienkowski@pwr.wroc.pl.

`

possibility of the regenerative motor braking. Since the PWM rectifiers are mostly
subjected to the varying line and load conditions and disturbances it is convenient to
apply the nonlinear control strategies that provide robustness and better effectiveness of
the rectifier’s control system [1,2]. The PWM rectifiers represent a class of the
variable structure systems since the three-phase current is conducted along with the
six different electrical routes determined by the six active states of the power switches
under control.
The sliding-mode control stems from the nonlinear control strategies designed for
the variable structure systems. During the closed-loop operation this kind of the
nonlinear control method provides the control system insensitiveness to the particular
extent of the uncertainties and disturbances. Moreover the sliding-mode control turns
out to be a most effective tool in approach to the design of the state-variable observers
for sensorless control [7,8].
2. MODEL OF THE BOOST-TYPE PWM RECTIFIER
The boost-type PWM rectifier called also the AC/DC line-side converter has the
topology of the three-phase PWM voltage inverter commonly used to feed the
squirrel-cage induction motors. The converter topology presented in Fig.1 provides
the synchronous rectification and the power regeneration.


Fig.1. Boost-type PWM rectifier

The three-phase two-level PWM rectifier consists of the six fully-controlled IGBT
transistors connected to the grid through the three symmetrical line inductors. The
voltage-source output of the rectifier requires the DC-link capacitor to provide the
step-up operation. In order to make the implementation of the sliding-mode control
more intelligible the DC-link has been divided into two equal parts with the ground

3
connection between them. The equations (1) describe the dynamic model of the PWM
rectifier in the three-phase coordinate frame:


( )
( )
( )
( )
gC c gB b gA a
d d
load
dc
c b a
dc
gC g gC
g
gC
c b a
dc
gB g gB
g
gB
c b a
dc
gA g gA
g
gA
i K i K i K
C C
i
u
dt
d
K K K
u
i R e
L
i
dt
d
K K K
u
i R e
L
i
dt
d
K K K
u
i R e
L
i
dt
d
⋅ + ⋅ + ⋅ + − =
(
¸
(

¸

+ − − − − =
(
¸
(

¸

− + − − − =
(
¸
(

¸

− − − − =
2
1
2
6
1
2
6
1
2
6
1
(1)

where R
g
, L
g
– grid resistance [Ω] and inductance [H]; C
d
– DC-link capacitance
[F]; i
gA
, i
gB
, i
gC
– grid currents in three-phase coordinates frame [A]; e
gA
, e
gB
, e
gC
– grid
phase voltages in the three-phase coordinates frame [V]; u
dc
– DC-link voltage [V];
i
load
– load DC current [A]; K
a
, K
b
, K
c
– converter’s conduction states [-1,1].
The vector description of the PWM rectifier is presented by the equations (2):


g
T
g g g
i K
K A i e i
⋅ + − =
(
¸
(

¸

⋅ − − =
d d
load
dc
dc
g
g
C C
i
u
dt
d
u
R
L dt
d
2
1
6
1
(2)

where A is a gain matrix with the following elements that provide its singularity:

0 ) det( ,
2 1 1
1 2 1
1 1 2
=
(
(
(
¸
(

¸

− −
− −
− −
= A A (3)

and K is a vector of control states:


(
(
(
¸
(

¸

=
c
b
a
K
K
K
K (4)

3. SLIDING-MODE CONTROL DESIGN FOR PWM RECTIFIER
In case of the design of any control system the deviations between the real plant
and its simplified model are unavoidable due to the unknown variations of the
parameters and the assumption of the meaningful constraints.
According to the cascade operation the inner sliding-mode loop controls the
dynamics of the converter input currents, while the outer loop provides the desired
value of the converter output voltage through a linear PI control.
The crucial step is the choice of the switching functions for the current control:


0
0
= − =
= − =
gq gqref q
gd gdref d
i i s
i i s
(5)

The sliding-mode current controllers are applied directly in a (d−q) rotating frame
and return the sign of the current errors as follows:


) (
) (
q gq
d gd
s sign i
s sign i
− = ∆
− = ∆
(6)

Similarly to Direct Power Control for the PWM rectifiers, the control signals are
selected out of a pre-defined look-up table according to the instantaneous converter's
reference voltage rotating in the (α−β) coordinate frame [5,6]. Fig.2a presents the six
active voltage vectors on the complex plane divided into twelve sectors.


Fig.2. Six active positions of the converter's reference voltage: a) voltage plane divided into twelve equal
sectors; b) constraints of the line current reconstruction due to dead-time of power transistors

5
The values of the control signals as a function of the signs of the current errors and
the grid voltage position in the (α−β) coordinates have been presented in Table 1. The
sliding-mode occurs in the system when the system trajectory (current errors)
converges the sliding line (5) and remains on it. This can be achieved only for the
active converter states due to the singularity of the A matrix (3), otherwise the control
system may lose controllability. Hence the look-up table does not contain the zero
states.
Table 1. Twelve-sector look-up table for sliding-mode control of PWM rectifier

N:
1 2 3 4 5 6
∆i
d
>0 ∆i
q
>0 1 1-1 -1 1-1 -1 1-1 -1 1 1 -1 1 1 -1-1 1
∆i
d
>0 ∆i
q
<0 1-1 1 1-1 1 1-1-1 1-1-1 1 1-1 1 1-1
∆i
d
<0 ∆i
q
>0 1-1-1 1 1-1 1 1-1 -1 1-1 -1 1-1 -1 1 1
∆i
d
<0 ∆i
q
<0 1-1 1 1-1-1 1-1-1 1 1-1 1 1-1 -1 1-1

N: 7 8 9 10 11 12
∆i
d
>0 ∆i
q
>0 -1-1 1 1-1 1 1-1 1 1-1-1 1-1-1 1 1-1
∆i
d
>0 ∆i
q
<0 -1 1-1 -1 1-1 -1 1 1 -1 1 1 -1-1 1 -1-1 1
∆i
d
<0 ∆i
q
>0 -1 1 1 -1-1 1 -1-1 1 1-1 1 1-1 1 1-1-1
∆i
d
<0 ∆i
q
<0 -1 1-1 -1 1 1 -1 1 1 -1-1 1 -1-1 1 1-1 1

The fundamental requirement for the proper operation of the PWM rectifier is the
power balance condition. The instantaneous values of the electrical power at the
converter input must be equal to their corresponding quantities at the converter output
assuming that the power conversion process is lossless and the line voltages and
currents are of sinusoidal form (7). The coefficient 3/2 is a consequence of the applied
orthogonal Clarke vector transformation.

( )
load dcref gqref gq gdref gd
i u i e i e = +
2
3
(7)

The linear dynamics of the DC-link voltage control loop is presented in the
differential equation (8) and its transient depends only on the time constant that
results from the load resistance and the capacitance values:



( )
d load
dc dcref
d dc
load dc dcref
dc
C R
u u
C u
i u u
u
dt
d

=
⋅ −
= (8)
4. SLIDING-MODE-BASED OBSERVERS OF AC-SIDE VARIABLES
The strict knowledge of the AC voltage and AC current is necessary to determine
the line voltage vector position and the line current errors, which is the crucial
information for the design of all vector-oriented control methods for the PWM
rectifiers. It has already been proved in [3,4] that the grid voltages and grid currents
may be successfully estimated based only on the information from the DC-link
voltage sensor, DC-link current sensor and the actual PWM pattern. The following
section presents in detail the proposed methodology of the observer design.
4.1. LINE CURRENT PRELIMINARY RECONSTRUCTION
The mathematical relationship between the DC-link current and the line currents is
based on the following equation involving the PWM pattern:


gC c gB b gA a dc
i K i K i K i ⋅ + ⋅ + ⋅ = (9)

According to the equation (9) the partial information about each line phase current
can be achieved using the conditions presented in Table 2.
Table 2. Relationship between DC-link current and phase currents due to PWM pattern
PWM pattern i
dc
=
1-1-1 i
gA

1 1-1 -i
gC

-1 1-1 i
gB

-1 1 1 -i
gA

-1-1 1 i
gC

1-1 1 -i
gB


Since no AC current transducers exist physically in the rectifier application there
is no hardware-based protection in the control system. The most crucial problem in
the proposed line current reconstruction is the high rate of the sampling of the DC-
link current, which from a technical point of view is inconvenient. Fig.2b presents the

7
ranges where the DC-link current can be properly sampled. The problem occurs when
the converter's reference voltage vector crosses the boundary between each of the six
sectors. These blank regions depend on the system delays and the IGBT dead-time.
4.2. SLIDING-MODE OBSERVER FOR LINE VOLTAGE AND LINE CURRENT
The starting point of designing a sliding-mode current observer for source voltage
is the following discontinuous formula:

( ) ( )
(
¸
(

¸

− − − − − ′ ⋅ =
c b a
dc
gA g gA gA
g
gA
K K K
u
i R i i sign m
L
i 2
6
ˆ ˆ
1
ˆ
&
(10)

where m is an observer gain,
gA
i′ is the observed line current described as:


0 =
¹
´
¦
= ′
gAreconstr gAest
gAreconstr
gA
i when i
i
i (11)

The switching component ( )
gA gA
i i sign m
ˆ
− ′ ⋅ refers to the line voltage estimate and
must be low-pass filtered (LPF) in order to obtain a useful line voltage signal:

( ) ( )
4 1 − =
− ′ ⋅ =
e T
gA gA gA
filter
i
ˆ
i sign m LPF eˆ (12)

The total line voltage signal can be obtain using the following formula:

( )
gA gAest gA g gA
e i i R e ˆ + − ′ − = ′ (13)

The voltage signal
gA
e′ is next submitted to the nonlinear asymptotic observer that
filters out the input signal giving the resulting smooth form of the line voltage
gA
e ′ ′ :


1 1 2 2 1
2
2
2
2
1 1 1 2 1
3 3
x W L xˆ x ˆ xˆ
x W L xˆ xˆ
g
g
g
− +
|
|
¹
|

\
|
+ − =
− =
ω
α
ω
α
ω
&
&
(14)

where gA
1 1
ˆ e x x ′ − = , ω
g
is the grid pulsation [rad/s], L
1
, L
2
, α are observer gains.

W
1
is the "window signal" for the phase A that is defined by the following
formula:


0
0
1
1

¹
´
¦
=
gAreconstr
i
otherwise
if
W (15)
The estimated line voltage signal
gA
e ′ ′ is next the input signal of the line current
observer that is described by the following equation:

( )
(
¸
(

¸

− − − − ′ ′ =
c b a
dc
gA g gA
g
gAest
K K K
u
i R e
L
i 2
6
ˆ
1
&
(16)

The line voltage and line current for the phase B and C can be obtained in the
similar way. The block diagram of the proposed sliding-mode control system for the
PWM rectifier with the observers for AC-side variables is presented in Fig.3.


Fig.3. AC-sensorless sliding-mode control system for the PWM rectifier
5. INVESTIGATIONS OF PROPOSED TECHNIQUE
AND SELECTED SIMULATION RESULTS

9
The computer multi-rate simulations have been carried out for the AC-sensorless
sliding-mode control system of the PWM rectifier shown in Fig.3. The proposed
control method has been tested using Matlab/Simulink. The simulation model has
been divided into three subsystems which have been computed with the different
rates. The parameters of the simulation model are presented in Tab.3.
Table 3. Simulation parameters
Line phase voltage e
g
: 230 V
Line voltage frequency f
g
: 50 Hz
Line resistance R
g
: 100 mΩ
Line inductance L
g
: 10 mH
DC-link capacitance C
d
: 1000 µF
DC-link nominal voltage U
dc
: 600 V
Load resistance R
load
: 100 Ω
Observer gain m 500
Observer gain L
1
4000
Observer gain L
2
400000
Observer gain α 10
PWM rectifier’s sample time T
p1
: 1 µs
Line current reconstruction rate T
p2
: 10 µs
Control system’s sample time T
p3
: 30 µs

Fig.4a shows the transients of the estimated phase voltage e
"
gA
and the line
currents by applying the full load resistance R
load
. Fig.4b presents the transients of the
line current in the (d-q) coordinate frame under the step change of the converter load.
The converter fulfills the unity power factor condition, since i
gq
=0.

0.49 0.495 0.5 0.505 0.51 0.515 0.52 0.525
-40
-30
-20
-10
0
10
20
30
40
Time[s]
(1)0.1*egAest[V] (2)igA[A] (3)igB[A] (4)igC[A]
(1)
(2)
(4)
(3)

0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
-5
0
5
10
15
20
25
30
Time[s]
(1)igd[A] (2)igq[A]
(1)
(2)

Fig.4. a) estimated line phase voltage and three-phase line currents;
b) (d−q) components of line current
a) b)

Fig.5a presents the transient of the DC-link voltage under the step change of the
rectifier load. In sliding-mode due to the equation (8) the DC-link voltage tends to its
reference value u
dcref
with the time constant τ =R
load
⋅C
d
irrespective of the values of
the parameters of the control system and the AC grid. The DC-link current is
presented in Fig.5b and corresponds to the active component i
gd
of the line current.
0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
570
580
590
600
610
620
630
Time[s]
udc[V]

0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
-20
-10
0
10
20
30
40
Time[s]
idc[A]

Fig.5. a) transient of DC-link voltage; b) DC-link current at step change of load
Fig.6a demonstrates the pre-estimated low-pass-filtered signal of the line voltage
e'
gA
. A compromise between the signal distortion of the line voltage e'
gA
and the
associated phase lag has been made by adjusting an appropriate value of the low-pass
filter time constant T
f
=1e-4. Fig.6b presents the estimated line voltage e"
gA
smoothed
by the nonlinear asymptotic observer with the "window" signal (14).

0.7 0.71 0.72 0.73 0.74
-400
-300
-200
-100
0
100
200
300
400
Time[s]
egAestSMO[V]

0.7 0.71 0.72 0.73 0.74
-400
-300
-200
-100
0
100
200
300
400
Time[s]
egAestNAO[V]

Fig.6. a) estimated line voltage e'
gA
; b) smoothed line voltage e"
gA

The information about the line currents after their reconstruction according to
Table 2 is partial and insufficient for the direct application into the control system of
the PWM rectifier. Hence the line current estimation has to be involved
simultaneously with the line current reconstruction. According to (11) the resulting
a) b)
a) b)

11
estimate of the line current i'
gA
is made of the reconstructed line current and the
estimated line current in points where the reconstructed line current equals zero. This
procedure is presented in Fig.7a and allows using the resulting signal for the control
purposes. Fig.7b presents the switching line during the sliding-mode. The state
trajectory remains on the switching line and moves on it depending on the load
conditions.
0.495 0.5 0.505 0.51
-30
-20
-10
0
10
20
30
Time[s]
(1)igAreconstr[A] (2)igAest[A]
(1)
(2)

-10 -5 0 5 10 15 20 25 30
-10
-5
0
5
10
15
20
25
30
igdref[A]
i
g
d
[
A
]
sd=igdref-igd
idle state
loaded state

Fig.7. a) line current reconstruction; b) switching line in sliding-mode
5. CONCLUSIONS
The paper presents the design of the novel sliding-mode current control technique
for the PWM rectifier. The special stress has been put on the designing of the sliding-
mode line voltage and line current observer for the sensorless operation of the PWM
rectifier without the AC-side transducers. The line currents have been reconstructed
throughout the sampling of the DC-link current during the PWM period. The missing
AC-side variables have been then estimated based on the reconstructed line currents.
In the proposed control technique the sliding-mode current controllers have been
located in the rotating (d−q) coordinate frame and the PWM pattern for the converter
has been computed using the three-dimensional look-up table. This modification has
reduced the necessary number of the transformations and has led to the significant
simplification in the proposed control system.
The cascade control structure used in the multi-rate simulations has involved the
line current vector transformation into the (d−q) components oriented with the line
voltage vector. Hence the power factor of the PWM rectifier can be independently
controlled by the reactive component of the line current. The simulation results
demonstrate the excellent performance of the proposed sliding-mode current control
system of the PWM rectifier with the sliding-mode observers for the AC-side
sensorless operation.
a) b)

REFERENCES
[1] KNAPCZYK M., PIEŃKOWSKI K., Robust Current Control Technique for Boos-Type AC-DC line-
side Converter with Sliding-Mode Voltage Observer, Scientific Papers of the Institute of Electrical
Machines, Drives and Metrology of the Wroclaw University of Technology, No. 57, Studies and
Research, No. 25, Wroclaw, 2005, pp. 381-392.
[2] KNAPCZYK M., PIEŃKOWSKI K., Sliding-mode Virtual Flux Oriented Control of PWM rectifiers
with fixed switching frequency, Power electronics and electrical drives - selected problems, ed. by
Teresa Orłowska-Kowalska, Wroclaw University of Technology, 2007, pp. 122-136.
[3] LEE D.-CH., LIM D.-S., AC Voltage and Current Sensorless Control of Three-Phase PWM
Rectifiers, IEEE Transactions on Power Electronics, vol.17, no.6, November 2002, pp. 883-890.
[4] LEE W.-CH., HYUN D.-S., LEE T.-K., A Novel Control Method for Three-Phase PWM Rectifiers
Using a Single Current Sensor, IEEE Transactions on Power Electronics, vol.15, no.5, September
2000, pp. 861-870.
[5] MALINOWSKI M., Sensorless Control Strategies for Three-Phase PWM Rectifiers, PhD Thesis,
Warsaw University of Technology, Poland, 2001.
[6] RADULOVIC Z., SABANOVIC A., Active Filter Control Using a Sliding Mode Approach, 25
th

Annual IEEE Power Electronics Specialist Conference PESC’94, June, 1994, pp.177-182.
[7] UTKIN V. I., CHEN D.-S., ZAREI SH., MILLER J. M., Synchronous Rectification of the Automotive
Alternator using Sliding Mode Observer, Proceedings on the American Control Conference,
Albuquerque, 1997, pp.1141-1145.
[8] UTKIN V., GULDNER J., SHI J., Sliding Mode Control in Electromechanical Systems, Taylor &
Francis, 1999.
STEROWANIE ZE ZMIENNĄ STRUKTURĄ PROSTOWNIKIEM PWM Z MODULACJĄ
ŚLIZGOWĄ I MINIMALNĄ LICZBĄ CZUJNIKÓW POMIAROWYCH
Artykuł prezentuje metodę modulacji ślizgowej w zastosowaniu do układów sterowania
prostowników PWM bez czujników pomiarowych napięcia i prądu sieci zasilającej. Proces
rekonstrukcji prądów przewodowych trójfazowej sieci zasilającej jest oparty na odpowiednim
próbkowaniu sygnału prądu stałego w obwodzie pośredniczącym w czasie trwania impulsów
sterujących kluczami przekształtnika według zdefiniowanego schematu. Następnie
zrekonstruowane przebiegi prądów wejściowych prostownika PWM wraz z sygnałem napięcia
stałego obwodu pośredniczącego przekształtnika są poddane dalszej obróbce w obserwatorach
ślizgowych. Przebiegi napięcia sieci zasilającej uzyskano przy zastosowaniu połączenia
obserwatora ślizgowego z klasycznym obserwatorem nieliniowym. Zaprezentowano metodę
modulacji ślizgowej PWM opartą na zastosowaniu ślizgowych regulatorów prądu sieci w
wirującym synchronicznie układzie odniesienia (d−q) i trójwymiarowej tablicy wyboru wektora
napięcia przekształtnika. Przedstawiona metoda modulacji ślizgowej pozwala na
wyeliminowanie transformacji prostych przełączających do stacjonarnego, trójfazowego układu
odniesienia, co ma miejsce w klasycznym układzie sterowania ślizgowego prostownikiem
PWM. W artykule przedstawiono i omówiono wybrane wyniki badań symulacyjnych.

MODEL OF THE BOOST-TYPE PWM RECTIFIER The boost-type PWM rectifier called also the AC/DC line-side converter has the topology of the three-phase PWM voltage inverter commonly used to feed the squirrel-cage induction motors. Fig. The sliding-mode control stems from the nonlinear control strategies designed for the variable structure systems. Since the PWM rectifiers are mostly subjected to the varying line and load conditions and disturbances it is convenient to apply the nonlinear control strategies that provide robustness and better effectiveness of the rectifier’s control system [1. The voltage-source output of the rectifier requires the DC-link capacitor to provide the step-up operation. In order to make the implementation of the sliding-mode control more intelligible the DC-link has been divided into two equal parts with the ground .8]. Boost-type PWM rectifier The three-phase two-level PWM rectifier consists of the six fully-controlled IGBT transistors connected to the grid through the three symmetrical line inductors.1. 2. Moreover the sliding-mode control turns out to be a most effective tool in approach to the design of the state-variable observers for sensorless control [7.2].1 provides the synchronous rectification and the power regeneration. The converter topology presented in Fig. During the closed-loop operation this kind of the nonlinear control method provides the control system insensitiveness to the particular extent of the uncertainties and disturbances.possibility of the regenerative motor braking. The PWM rectifiers represent a class of the variable structure systems since the three-phase current is conducted along with the six different electrical routes determined by the six active states of the power switches under control.

Cd – DC-link capacitance [F]. udc – DC-link voltage [V]. egB.   − 1 − 1 2    and K is a vector of control states: K a  K = Kb    Kc    det( A ) = 0 (3) (4) . The vector description of the PWM rectifier is presented by the equations (2): d 1 ig = dt Lg u dc   e g − R g i g − 6 A ⋅ K    i d 1 u dc = − load + K T ⋅ ig dt C d 2C d (2) where A is a gain matrix with the following elements that provide its singularity:  2 − 1 − 1 A = − 1 2 − 1.3 connection between them. egA. iload – load DC current [A]. egC – grid phase voltages in the three-phase coordinates frame [V].1]. Ka. igC – grid currents in three-phase coordinates frame [A]. igA. Kb. The equations (1) describe the dynamic model of the PWM rectifier in the three-phase coordinate frame: u dc d 1   i gA = e gA − R g i gA − 6 (2 K a − K b − K c ) dt Lg   d 1 i gB = dt Lg u dc   e gB − R g i gB − 6 (− K a + 2 K b − K c )   u dc d 1   i gC = e gC − R g i gC − 6 (− K a − K b + 2 K c ) dt Lg   i d 1 u dc = − load + K a ⋅ i gA + K b ⋅ i gB + K c ⋅ i gC dt C d 2C d (1) ( ) where Rg. Lg – grid resistance [Ω] and inductance [H]. Kc – converter’s conduction states [-1. igB.

b) constraints of the line current reconstruction due to dead-time of power transistors . while the outer loop provides the desired value of the converter output voltage through a linear PI control. the control signals are selected out of a pre-defined look-up table according to the instantaneous converter's reference voltage rotating in the (α−β) coordinate frame [5. Fig.2.6]. Six active positions of the converter's reference voltage: a) voltage plane divided into twelve equal sectors.2a presents the six active voltage vectors on the complex plane divided into twelve sectors. SLIDING-MODE CONTROL DESIGN FOR PWM RECTIFIER In case of the design of any control system the deviations between the real plant and its simplified model are unavoidable due to the unknown variations of the parameters and the assumption of the meaningful constraints. According to the cascade operation the inner sliding-mode loop controls the dynamics of the converter input currents.3. Fig. The crucial step is the choice of the switching functions for the current control: s d = i gdref − i gd = 0 s q = i gqref − i gq = 0 (5) The sliding-mode current controllers are applied directly in a (d−q) rotating frame and return the sign of the current errors as follows: ∆i gd = − sign ( s d ) ∆i gq = − sign( sq ) (6) Similarly to Direct Power Control for the PWM rectifiers.

This can be achieved only for the active converter states due to the singularity of the A matrix (3). Table 1. The instantaneous values of the electrical power at the converter input must be equal to their corresponding quantities at the converter output assuming that the power conversion process is lossless and the line voltages and currents are of sinusoidal form (7). otherwise the control system may lose controllability. The sliding-mode occurs in the system when the system trajectory (current errors) converges the sliding line (5) and remains on it. Hence the look-up table does not contain the zero states. Twelve-sector look-up table for sliding-mode control of PWM rectifier N: ∆id>0 ∆id>0 ∆id<0 ∆id<0 ∆iq>0 ∆iq<0 ∆iq>0 ∆iq<0 N: ∆id>0 ∆id>0 ∆id<0 ∆id<0 ∆iq>0 ∆iq<0 ∆iq>0 ∆iq<0 1 1 1-1 1-1 1 1-1-1 1-1 1 2 -1 1-1 1-1 1 1 1-1 1-1-1 3 -1 1-1 1-1-1 1 1-1 1-1-1 4 -1 1 1 1-1-1 -1 1-1 1 1-1 5 -1 1 1 1 1-1 -1 1-1 1 1-1 6 -1-1 1 1 1-1 -1 1 1 -1 1-1 7 -1-1 1 -1 1-1 -1 1 1 -1 1-1 8 1-1 1 -1 1-1 -1-1 1 -1 1 1 9 1-1 1 -1 1 1 -1-1 1 -1 1 1 10 1-1-1 -1 1 1 1-1 1 -1-1 1 11 1-1-1 -1-1 1 1-1 1 -1-1 1 12 1 1-1 -1-1 1 1-1-1 1-1 1 The fundamental requirement for the proper operation of the PWM rectifier is the power balance condition.5 The values of the control signals as a function of the signs of the current errors and the grid voltage position in the (α−β) coordinates have been presented in Table 1. 3 e gd i gdref + e gq i gqref = u dcref iload 2 ( ) (7) The linear dynamics of the DC-link voltage control loop is presented in the differential equation (8) and its transient depends only on the time constant that results from the load resistance and the capacitance values: . The coefficient 3/2 is a consequence of the applied orthogonal Clarke vector transformation.

Relationship between DC-link current and phase currents due to PWM pattern PWM pattern 1-1-1 1 1-1 -1 1-1 -1 1 1 -1-1 1 1-1 1 idc= igA -igC igB -igA igC -igB Since no AC current transducers exist physically in the rectifier application there is no hardware-based protection in the control system. which from a technical point of view is inconvenient. which is the crucial information for the design of all vector-oriented control methods for the PWM rectifiers. LINE CURRENT PRELIMINARY RECONSTRUCTION The mathematical relationship between the DC-link current and the line currents is based on the following equation involving the PWM pattern: idc = K a ⋅ i gA + K b ⋅ i gB + K c ⋅ i gC (9) According to the equation (9) the partial information about each line phase current can be achieved using the conditions presented in Table 2. The following section presents in detail the proposed methodology of the observer design. DC-link current sensor and the actual PWM pattern.4] that the grid voltages and grid currents may be successfully estimated based only on the information from the DC-link voltage sensor.2b presents the .u dcref − u dc ⋅ iload u dcref − u dc d u dc = = dt u dc C d Rload C d 4.1. It has already been proved in [3. SLIDING-MODE-BASED OBSERVERS OF AC-SIDE VARIABLES ( ) (8) The strict knowledge of the AC voltage and AC current is necessary to determine the line voltage vector position and the line current errors. 4. Table 2. The most crucial problem in the proposed line current reconstruction is the high rate of the sampling of the DClink current. Fig.

L1. 4. These blank regions depend on the system delays and the IGBT dead-time.2. α are observer gains. L2. ωg is the grid pulsation [rad/s].7 ranges where the DC-link current can be properly sampled. The problem occurs when the converter's reference voltage vector crosses the boundary between each of the six sectors. . SLIDING-MODE OBSERVER FOR LINE VOLTAGE AND LINE CURRENT The starting point of designing a sliding-mode current observer for source voltage is the following discontinuous formula: u 1   & ˆ ˆ ˆ ′ i gA = m ⋅ sign i gA − i gA − Rg i gA − dc (2 K a − K b − K c ) Lg  6   ( ) (10) where m is an observer gain. i′ is the observed line current described as: gA i gAreconstr ′ i gA =   i gAest when (11) i gAreconstr = 0 ˆ The switching component m ⋅ sign i′ − i gA refers to the line voltage estimate and gA ( ) must be low-pass filtered (LPF) in order to obtain a useful line voltage signal: ˆ e gA = LPF m ⋅ sign i ′ − ˆ gA gA i ( ( )) T filter =1e − 4 (12) The total line voltage signal can be obtain using the following formula: ˆ ′ e′ = − Rg i gA − i gAest + e gA gA ( ) (13) The voltage signal e′ is next submitted to the nonlinear asymptotic observer that gA filters out the input signal giving the resulting smooth form of the line voltage e′′ : gA & ˆ ˆ x1 = x 2 − L1W1 x1  3α 2 2 & ˆ x 2 = − ω g + 2  ωg    ˆ + 3α x − L W x ˆ x 2 1 1  1 ωg 2  (14) ˆ where x1 = x1 − e′gA .

The block diagram of the proposed sliding-mode control system for the PWM rectifier with the observers for AC-side variables is presented in Fig. Fig.3.3.W1 is the "window signal" for the phase A that is defined by the following formula: if i gAreconstr ≠ 0 1 (15) W1 =  0 otherwise The estimated line voltage signal e′′ is next the input signal of the line current gA observer that is described by the following equation: u dc 1   & ˆ i gAest = gA e′′ − Rg i gA − 6 (2 K a − K b − K c ) Lg   (16) The line voltage and line current for the phase B and C can be obtained in the similar way. INVESTIGATIONS OF PROPOSED TECHNIQUE AND SELECTED SIMULATION RESULTS . AC-sensorless sliding-mode control system for the PWM rectifier 5.

3.515 0. The proposed control method has been tested using Matlab/Simulink. The simulation model has been divided into three subsystems which have been computed with the different rates.4b presents the transients of the line current in the (d-q) coordinate frame under the step change of the converter load.4.5 0.3. b) (d−q) components of line current .49 0.4a shows the transients of the estimated phase voltage e"gA and the line currents by applying the full load resistance Rload.52 0. Simulation parameters Line phase voltage eg: Line voltage frequency fg: Line resistance Rg: Line inductance Lg: DC-link capacitance Cd: DC-link nominal voltage Udc: Load resistance Rload: Observer gain m Observer gain L1 Observer gain L2 Observer gain α PWM rectifier’s sample time Tp1: Line current reconstruction rate Tp2: Control system’s sample time Tp3: 230 V 50 Hz 100 mΩ 10 mH 1000 µF 600 V 100 Ω 500 4000 400000 10 1 µs 10 µs 30 µs Fig.4 0.3 0. since igq=0.7 Time[s] 0.6 0. Table 3.51 0.8 0.9 The computer multi-rate simulations have been carried out for the AC-sensorless sliding-mode control system of the PWM rectifier shown in Fig.1*egAest[V] (2)igA[A] (3)igB[A] (4)igC[A] 40 30 20 20 b) (1)igd[A] 30 25 (2)igq[A] (1) (3) 10 (4) 0 10 15 -10 5 -20 -30 -40 0.495 (2) 0 (2) (1) 0. a) (1)0. The converter fulfills the unity power factor condition. a) estimated line phase voltage and three-phase line currents. Fig. The parameters of the simulation model are presented in Tab.5 0.505 0.9 1 Fig.525 Time[s] -5 0.

4 0.5.72 Time[s] 0. A compromise between the signal distortion of the line voltage e'gA and the associated phase lag has been made by adjusting an appropriate value of the low-pass filter time constant Tf=1e-4.6.3 0.8 0. Hence the line current estimation has to be involved simultaneously with the line current reconstruction. According to (11) the resulting .4 0.72 Time[s] 0.6 0.6b presents the estimated line voltage e"gA smoothed by the nonlinear asymptotic observer with the "window" signal (14).9 1 -20 0.73 0.71 0.74 0.5a presents the transient of the DC-link voltage under the step change of the rectifier load.5b and corresponds to the active component igd of the line current. a) estimated line voltage e'gA .6a demonstrates the pre-estimated low-pass-filtered signal of the line voltage e'gA. a) transient of DC-link voltage.74 Fig. In sliding-mode due to the equation (8) the DC-link voltage tends to its reference value udcref with the time constant τ =Rload⋅Cd irrespective of the values of the parameters of the control system and the AC grid. b) smoothed line voltage e"gA The information about the line currents after their reconstruction according to Table 2 is partial and insufficient for the direct application into the control system of the PWM rectifier.6 0.3 0.71 0.9 1 Fig.7 egAestSMO[V] b) 400 300 200 100 0 -100 -200 -300 -400 0.Fig.7 egAestNAO[V] 0.7 Time[s] 0. b) DC-link current at step change of load Fig.73 0.8 0. a) 400 300 200 100 0 -100 -200 -300 -400 0.7 Time[s] 0. The DC-link current is presented in Fig. udc[V] idc[A] a) 630 b) 40 620 30 610 20 600 10 590 0 580 -10 570 0.5 0.5 0. Fig.

a) line current reconstruction. This procedure is presented in Fig.7a and allows using the resulting signal for the control purposes.495 0. (2)igAest[A] sd=igdref-igd a) 30 (1)igAreconstr[A] b) 30 25 20 20 10 15 0 (1) -10 0 -20 (2) -5 -10 -10 igd[A] loaded state 10 idle state 5 -30 0. In the proposed control technique the sliding-mode current controllers have been located in the rotating (d−q) coordinate frame and the PWM pattern for the converter has been computed using the three-dimensional look-up table. The special stress has been put on the designing of the slidingmode line voltage and line current observer for the sensorless operation of the PWM rectifier without the AC-side transducers. The state trajectory remains on the switching line and moves on it depending on the load conditions. This modification has reduced the necessary number of the transformations and has led to the significant simplification in the proposed control system. Hence the power factor of the PWM rectifier can be independently controlled by the reactive component of the line current.7.5 Time[s] 0. The cascade control structure used in the multi-rate simulations has involved the line current vector transformation into the (d−q) components oriented with the line voltage vector. The simulation results demonstrate the excellent performance of the proposed sliding-mode current control system of the PWM rectifier with the sliding-mode observers for the AC-side sensorless operation.51 -5 0 5 10 15 igdref[A] 20 25 30 Fig. . The missing AC-side variables have been then estimated based on the reconstructed line currents. The line currents have been reconstructed throughout the sampling of the DC-link current during the PWM period.7b presents the switching line during the sliding-mode. b) switching line in sliding-mode 5. Fig.11 estimate of the line current i'gA is made of the reconstructed line current and the estimated line current in points where the reconstructed line current equals zero. CONCLUSIONS The paper presents the design of the novel sliding-mode current control technique for the PWM rectifier.505 0.

. Wroclaw University of Technology. Taylor & Francis. LIM D. Sensorless Control Strategies for Three-Phase PWM Rectifiers. [4] LEE W.-CH. Sliding-mode Virtual Flux Oriented Control of PWM rectifiers with fixed switching frequency. June. [6] RADULOVIC Z.. Proceedings on the American Control Conference. Sliding Mode Control in Electromechanical Systems.-S. pp. November 2002.177-182. pp. pp. A Novel Control Method for Three-Phase PWM Rectifiers Using a Single Current Sensor. pp. W artykule przedstawiono i omówiono wybrane wyniki badań symulacyjnych. 2001. co ma miejsce w klasycznym układzie sterowania ślizgowego prostownikiem PWM.-S. . [7] UTKIN V. Proces rekonstrukcji prądów przewodowych trójfazowej sieci zasilającej jest oparty na odpowiednim próbkowaniu sygnału prądu stałego w obwodzie pośredniczącym w czasie trwania impulsów sterujących kluczami przekształtnika według zdefiniowanego schematu.selected problems. LEE T. pp. 1999.-K. pp. vol. ed.5.. Robust Current Control Technique for Boos-Type AC-DC lineside Converter with Sliding-Mode Voltage Observer. Studies and Research. Active Filter Control Using a Sliding Mode Approach. Albuquerque.. No. STEROWANIE ZE ZMIENNĄ STRUKTURĄ PROSTOWNIKIEM PWM Z MODULACJĄ ŚLIZGOWĄ I MINIMALNĄ LICZBĄ CZUJNIKÓW POMIAROWYCH Artykuł prezentuje metodę modulacji ślizgowej w zastosowaniu do układów sterowania prostowników PWM bez czujników pomiarowych napięcia i prądu sieci zasilającej. IEEE Transactions on Power Electronics.-S. [8] UTKIN V. SABANOVIC A.17. 1994. Przebiegi napięcia sieci zasilającej uzyskano przy zastosowaniu połączenia obserwatora ślizgowego z klasycznym obserwatorem nieliniowym. No..6. IEEE Transactions on Power Electronics. 381-392. PhD Thesis. 25th Annual IEEE Power Electronics Specialist Conference PESC’94. HYUN D. [2] KNAPCZYK M.. I. 1997.-CH. SHI J. ZAREI SH. by Teresa Orłowska-Kowalska. vol.. Drives and Metrology of the Wroclaw University of Technology. Wroclaw. 883-890.. 122-136. 2005. Poland. no.. GULDNER J. 2007. [5] MALINOWSKI M. Warsaw University of Technology.. 861-870.REFERENCES [1] KNAPCZYK M. PIEŃKOWSKI K.. Scientific Papers of the Institute of Electrical Machines. AC Voltage and Current Sensorless Control of Three-Phase PWM Rectifiers. Przedstawiona metoda modulacji ślizgowej pozwala na wyeliminowanie transformacji prostych przełączających do stacjonarnego. MILLER J. Synchronous Rectification of the Automotive Alternator using Sliding Mode Observer. 25.. PIEŃKOWSKI K. no. Zaprezentowano metodę modulacji ślizgowej PWM opartą na zastosowaniu ślizgowych regulatorów prądu sieci w wirującym synchronicznie układzie odniesienia (d−q) i trójwymiarowej tablicy wyboru wektora napięcia przekształtnika.1141-1145. September 2000. CHEN D. [3] LEE D.. 57. Następnie zrekonstruowane przebiegi prądów wejściowych prostownika PWM wraz z sygnałem napięcia stałego obwodu pośredniczącego przekształtnika są poddane dalszej obróbce w obserwatorach ślizgowych.. M. trójfazowego układu odniesienia.. Power electronics and electrical drives ..15....