Liu 2014

This article was downloaded by: [Aston University]
On: 09 January 2014, At: 12:10

Publisher: Taylor & Francis
Informa Ltd Registered in England and Wales Registered Number: 1072954 Registered office: Mortimer
House, 37-41 Mortimer Street, London W1T 3JH, UK
International Journal of Control

Publication details, including instructions for authors and subscription information:
http://www.tandfonline.com/loi/tcon20
Observer-Based Higher Order Sliding Mode Control

of Power Factor in Three-Phase AC/DC Converter for
Hybrid Electric Vehicle Applications
a a a
Jianxing Liu , Salah Laghrouche & Maxime Wack
a
Laboratoire IRTES, Universit de Technologie de Belfort-Montbliard (UTBM), Belfort,
France
Accepted author version posted online: 26 Nov 2013.Published online: 26 Nov 2013.
To cite this article: International Journal of Control (2013): Observer-Based Higher Order Sliding Mode Control of Power
Factor in Three-Phase AC/DC Converter for Hybrid Electric Vehicle Applications, International Journal of Control, DOI:
10.1080/00207179.2013.868609
To link to this article: http://dx.doi.org/10.1080/00207179.2013.868609
Disclaimer: This is a version of an unedited manuscript that has been accepted for publication. As a service
to authors and researchers we are providing this version of the accepted manuscript (AM). Copyediting,
typesetting, and review of the resulting proof will be undertaken on this manuscript before final publication
of the Version of Record (VoR). During production and pre-press, errors may be discovered which could
affect the content, and all legal disclaimers that apply to the journal relate to this version also.
PLEASE SCROLL DOWN FOR ARTICLE
Taylor & Francis makes every effort to ensure the accuracy of all the information (the Content) contained
in the publications on our platform. However, Taylor & Francis, our agents, and our licensors make no
representations or warranties whatsoever as to the accuracy, completeness, or suitability for any purpose of
the Content. Any opinions and views expressed in this publication are the opinions and views of the authors,
and are not the views of or endorsed by Taylor & Francis. The accuracy of the Content should not be relied
upon and should be independently verified with primary sources of information. Taylor and Francis shall
not be liable for any losses, actions, claims, proceedings, demands, costs, expenses, damages, and other
liabilities whatsoever or howsoever caused arising directly or indirectly in connection with, in relation to or
arising out of the use of the Content.
This article may be used for research, teaching, and private study purposes. Any substantial or systematic
reproduction, redistribution, reselling, loan, sub-licensing, systematic supply, or distribution in any
form to anyone is expressly forbidden. Terms & Conditions of access and use can be found at http://
www.tandfonline.com/page/terms-and-conditions
November 21, 2013 12:57 International Journal of Control TCONA868609
International Journal of Control

Vol. 00, No. 00, Month 200x, 129
RESEARCH ARTICLE
Observer-Based Higher Order Sliding Mode Control of Power Factor in
Three-Phase AC/DC Converter for Hybrid Electric Vehicle Applications
t
Jianxing Liu, Salah Laghrouche and Maxime Wack
ip
Laboratoire IRTES, Universite de Technologie de Belfort-Montbeliard (UTBM), Belfort, France
(Received 00 Month 200x; final version received 00 Month 200x)
cr
In this paper, a full-bridge boost power converter topology is studied for power factor control, using output
Downloaded by [Aston University] at 12:10 09 January 2014
us
higher order sliding mode control. The AC/DC converters are used for charging the battery and super-capacitor
in hybrid electric vehicles from the utility. The proposed control forces the input currents to track the desired
values, which can control the output voltage while keeping the power factor close to one. Super-twisting sliding
mode observer is employed to estimate the input currents and load resistance only from the measurement
an
of output voltage. Lyapunov analysis shows the asymptotic convergence of the closed loop system to zero.
Multi-rate simulation illustrates the effectiveness and robustness of the proposed controller in the presence of
measurement noise.
M
Keywords: AC/DC Converter; Sliding Mode Control (SMC); Observer-Based Control; Super-Twisting
Observer; Unity-Power-Factor; Hybrid Electric Vehicle (HEV)
d
te
1 INTRODUCTION
With the advent of distributed DC power sources in the energy sector, the use of boost type
ep
three phase rectifiers has increased in industrial applications, especially, battery charger in hy-
brid electric vehicles (HEV) (Egan et al. 2007, Pahlevaninezhad et al. 2012a,b, Kuperman et al.
c
2012). Power-factor-corrected utility interfaces are of great importance in the HEV industry.
The complete energy conversion cycle of the HEV must convert electrical power from the util-
Ac
ity to mechanical power at the drive axle as efficiently and as economically as possible (Egan
et al. 2007, Guerrero et al. 2013, Liu et al. 2013). Different power conversion systems of plug-in
HEV power conditioning systems are presented in Cao and Emadi (2009), Lee et al. (2009),
Wirasingha and Emadi (2011), Camara et al. (2010), Amjadi and Williamson (2010).

Corresponding author. Email: salah.laghrouche@utbm.fr
ISSN: 0020-7179 print/ISSN 1366-5820 online

c 200x Taylor & Francis
DOI: 10.1080/0020717YYxxxxxxxx
http://www.informaworld.com
2 International Journal of Control
Fig. 1 shows the structure of the hybrid electric vehicle power conversion system, which consists
of an AC/DC converter, a three phase DC/AC inverter (Zhong and Hornik 2013, Hornik and
Zhong 2013), three DC/DC converters and various power storages i.e. source grid, battery, super-
capacitor and fuel cell. The AC/DC converter is used to charge the battery and super-capacitor
through its bidirectional DC/DC converters, while ensuring that the utility current is drawn at
unity power factor in order to minimize line distortion and maximize the real power available
from the utility outlet. The battery and super-capacitor supply power to the three-phase inverter
which feeds the three-phase motor. The AC/DC converter consists of two stages (Emadi et al.
2008, Hayes et al. 1999). The first stage is Power Factor Correction (PFC), which simultaneously
t
regulates the DC-link voltage level and the line current waveform. The second stage is a charger
ip
with different types of resonant or pulse-width-modulation (PWM) DC/DC converters (Egan
cr
et al. 2007). PFC is used to improve the quality of the input phase current that is sourced from
the utility by generating and tracking the desired current profile while the charger is used to
us
charge the battery and super-capacitor in a HEV.
Many schemes and solutions are proposed in the field of PFC. Linear control methods using
linear regulators for the output voltage control have been proposed in Pan and Chen (1993),
an
Dixon and Ooi (1988), which change the modulation index slowly, thus resulting in a slow
dynamical response. Consequently, the linear feedback control of the rectifier output voltage
becomes slow and difficult. Moreover, due to coupling between the duty-cycle and the state
M
variables in the AC/DC boost converter, linear controllers are not able to perform optimally for
the whole range of operating conditions. In contrast with linear control, nonlinear approaches
d
can optimize the performance of the AC/DC converter over a wide range of operating conditions.
Many nonlinear techniques have been proposed, such as input-output linearization (Lee 2003),
te
feedback linearization (Lee et al. 2000), fuzzy logic control (Cecati et al. 2005), passivity-based
control (Escobar et al. 2001), back-stepping technique control (Allag et al. 2007), Lyapunov-
ep
based control (Pahlevaninezhad et al. 2012a, Komurcugil and Kukrer 1998), differential flatness
based control (Pahlevaninezhad et al. 2012b, Houari et al. 2012, Thounthong 2012), and sliding
mode control (Shtessel et al. 2008, Silva 1999, Tan et al. 2007). However, most of the above works
c
need continuous measurements of AC voltages, AC currents and DC voltage. This requires a large
Ac
number of both voltage and current sensors, which increases system complexity, cost, space, and
reduces system reliability. Moreover, the sensors are susceptible to electrical noise, which cannot
be avoided during high-power switching. Reducing the number of sensors has a significant affect
upon the control systems performance.
A few results have been proposed to reduce the current sensors (Andersen et al. 1999, Pan and
Chen 1993, Lee and Lim 2002, Lee et al. 2001) where the input phase currents are reconstructed
International Journal of Control 3
from the switching states of the AC/DC rectifier and the measured DC-link currents, and then
used in feedback control. However, they require digital sampling of the DC-link current in ev-
ery switching cycle and numerical computations. The accuracy of measurement is inherently
controlled by the sampling rate. In Nollet et al. (2008), a flatness based second order sliding
mode (SOSM) control combined with an angular velocity second order sliding mode observer
was designed for the stepper motor. The angular displacement and the direct current are chosen
as flat outputs such that other states or input variables can be presented as a function of the
flat outputs and a finite number of their time derivatives. The practical stability of the closed-
loop system was obtained. For real time implementation of the designed control law, the online
t
parameter identification should be taken into consideration. Shtessel et al. (2008) designed a pa-
ip
rameter observer to estimate the load resistance and parasitic phase resistance. (Hamida et al.
cr
2013) proposed an adaptive interconnected observer with online parameter identification (the
stator inductance and the stator resistance). Both of (Shtessel et al. 2008, Hamida et al. 2013)
us
require the utilization of the current sensors.
The objective of this paper is to design an efficient AC/DC power converter that charges
the battery and super-capacitor in a HEV with unity power factor, by eliminating the using of
an
current sensors. Only voltage sensors are required for measuring the output voltage and source
voltage. A super-twisting (ST) sliding mode observer (SMO) and a ST parameter SMO are
designed to observe the phase currents and load resistance from the measured output voltage
M
respectively. The proposed ST SMO guarantees fast convergence rate of the current observation
error dynamics while the load resistance is estimated by so-called equivalent output injection,
d
facilitating the design of controllers. The observer-based controller design is based on the two
goals mentioned in (Vadim Utkin, Jurgen Guldner, Jingxin Shi 2009, Edwards and Spurgeon
te
1998), for the design of an efficient AC/DC power converter: 1) Unity power factor to maximize
the performance of the power conversion. 2) Ripple free output voltage.
ep
Sliding mode algorithm is known for the characteristics of robustness and effectiveness (Ed-
wards and Spurgeon 1998), making it an effective method to deal with the nonlinear behavior of
the boost rectifiers. The ST Sliding Mode Control (SMC) allows not only the achievement of the
c
high performance of the system but also the maintenance of the functionality under parametric
Ac
uncertainty and external disturbance.

The main contributions of this paper are as follows :
(1) Only the output DC voltage and source voltage are assumed to be measurable, without
requiring any current sensors. The proposed current observer ensures faster convergence
by employing the ST sliding mode algorithm (STA). The equivalent output injections
are obtained in a continuous way by the STA without the use of any low pass filters.
(2) The load resistance considered as a varying parameter is estimated via the ST SMO and
integrated in the proposed control design. The exponential stability of the closed loop
system is established via Lyapunov analysis.
The paper is organized as follows. In Section II, the mathematical model and control objectives
are presented. In Section III, the design of observer-based current controller based on STA is
presented. In Section IV, we show the design of the parameter observer for the system, and power
factor is also estimated. In Section V, simulations results of the performance of the obtained ST
SMC compared with the conventional PI controller are presented. Finally, some conclusions are
drawn in Section VI.
t
ip
cr
us
2 PROBLEM FORMULATION
2.1 System Modeling

an
The power circuit of the three phase voltage source AC/DC full-bridge boost converter under
consideration is shown in Fig. 2. It is assumed that an equivalent resistive load RL is connected
M
to the output of the AC/DC converter (Pahlevaninezhad et al. 2012a,b).
The control inputs, as they appear in the system, are defined as u = [u1 u2 u3 ]T , which take
values from the discrete set {1, +1}. The corresponding inverse control u
= [
u1 u 3 ]T takes
2 u
d
1 = 1 which corresponds to the

the opposite values at the same time, i.e. u1 = +1 means u
te
conducting state for the upper switching element T1 and nonconducting state for the bottom
switching element T4 (Shtessel et al. 2008).
ep
It is assumed that the input AC voltage is a balanced three-phase supply (i.e. Ug1 +Ug2 +Ug3 =
0), the mathematical model of the boost AC/DC converter in phase coordinate frame can be
obtained through analyzing the circuit (Pan and Chen 1993),
c
Ac

dia r U0 1

= ia (2u1 u2 u3 ) + Ug1 ,

dt L 6L L

dib r U0 1

= ib (2u2 u1 u3 ) + Ug2 ,

dt L 6L L
(1)

dic r U0 1

= ic (2u3 u1 u2 ) + Ug3 ,

dt L 6L L

dU U0 1

0 = + (ia u1 + ib u2 + ic u3 ).
dt RL C 2C
It can also be written as,

di r U0 1

= i Bu + Ug ,
dt L 6L L
(2)

dU U 1
0 = 0
+ uT i.
dt RL C 2C
where r is parasitic phase resistance (including voltage source internal resistance and impedance
of switching elements in open state); RL is the load resistance; L is phase inductor; C is output
h iT h iT
capacitor; U0 is output voltage; i = ia ib ic are the input phase currents; Ug = Ug1 Ug2 Ug3
are the source voltages which have different magnitudes but the same frequency and phase shift
h iT
t
of 2/3 electrical degrees (with respect to each other); and u = u1 u2 u3 are control signals.
ip
The gain matrix and source voltage are as follows,

cr
2 1 1 sin(t)
2

B = 1 2 1 , Ug = E sin(t ) . (3)
3
us
2
1 1 2 sin(t + )
3
where E is the magnitude of the source voltages (Komurcugil and Kukrer 1998).
an
For modeling and control design, it is convenient to transform three-phase variables into a
rotating (d, q) frame. The transformed variables is defined as,

M
ud id Ugd
udq = = T u, idq = = T i, T Ug = . (4)
uq iq Ugq
where
d

2 2
te
2 cos(t) cos(t 3 ) cos(t + 3 )

T = . (5)
3 sin(t) sin(t 2 ) sin(t + 2 )
3 3
ep
is the Parks transformation (Bose 2002, Lee 2003).

From Eqs.(3, 4, 5), it follows that Ugd = 0 and Ugq = E. The dynamical model of the AC/DC
c
converter in the rotating (d, q) frame can be expressed as (Komurcugil and Kukrer 1998, Lee
et al. 2000, Silva 1997)
Ac

did r U0

= id + iq ud ,

dt L 2L

di
q r E U0
= iq + id uq , (6)

dt L L 2L

dU0 U0 3(id ud + iq uq )

= + .
dt RL C 4C
where is the angular frequency of the source voltage. In the transformed state Eq. (6), the state
h iT h iT h iT
vector is defined as x = x1 x2 x3 = id iq U0 and the control input vector udq = ud uq
are the switching functions in synchronously rotating (d, q) coordinate. From the control point
of view, the model of AC/DC converter in (d, q) frame has the advantage of reducing the current
control task into a set-point tracking problem (Lee 2003).
2.2 Control Objectives
Assumption 2.1 The phase voltages Ug and output voltage U0 are measurable;
t
The control objectives are as follows,
ip
The input phase currents ia , ib , ic should be in phase with corresponding input source voltage
cr
Ug1 , Ug2 , Ug3 in order to obtain a unity power factor.
The DC component of the output voltage should be driven to some desired value U0 while its
AC component has to be attenuated to a given level.
us
3 OBSERVER-BASED SLIDING MODE CONTROLLER DESIGN
an
In observer-based sliding mode control, the real plant states are substituted by observer states,
reducing the number of measurements. It has been shown that the performance of an observer-
M
based sliding mode controller can be improved significantly by keeping the plant system and
the observer system operating closely (Sira-Ramirez et al. 1996). Fig. 3 shows the structure
d
of the observer-based control system for three phase AC/DC converters, which consists of two
important parts: sliding mode current observer and controller system. The desired current I is
te
calculated based on the Power Balance Condition. The observed current I is the output of the
observer system (10).
ep
3.1 Super-Twisting Sliding Mode Observer Design

c
The proposed observer is designed in the following two steps,

Ac
(1) Analyzing the observability of the nonlinear system;

(2) Construction of the ST SMO;
3.1.1 Observability Analysis of the System
In order to construct an observer for a system, it is necessary to verify the observability of

the system i.e. there exists the possibility of obtaining the states of a system only from the
knowledge of its inputs and outputs up to time t (Besancon 2007). Considering the following
nonlinear system,

x =fu (x, u),
h i (7)

y = h (x) h (x) T .
1 p
where x Rn are the state vectors, u Rm are the bounded inputs, y Rp are the outputs.
Assume that the vector field fu (, ) is a sufficiently smooth function.
Definition 3.1 (Hermann and Krener 1977) The system described by (7) is locally observable if
the matrix defined by (8) satisfies observability rank condition dim(O) = n at a point x0 ,
ip
dL0fu (h1 ) dL0fu (h2 ) dL0fu (hp )

1 1 1
dLfu (h1 ) dLfu (h2 ) dLfu (hp )

O = . (8)
.. .. ..
cr
..
. . . .

n1 n1 n1
dLfu (h1 ) dLfu (h2 ) dLfu (hp )
us
where Lfu (h) denotes the Lie derivative of h with respect to fu .
Taking the output as y = x3 = U0 , the application of Definition 3.1 leads to the following
observability matrix,
an

0, 0, 1

3ud 3uq 1
M

O= , , . (9)
4C 4C RC
3uq 3ud 3kudq k2
1 2
%ud , %uq , +( )
4C 4C 8LC RL C
d
3 r 1
where % = ( + ), and kudq k22 = u2d +u2q . Thus, it is possible to observe the currents id , iq
4C L RL C
te
from the measurement of the output voltage U0 when kudq k2 6= 0. In the case of singular inputs
ud = 0, uq = 0 which results the system (6) into a reduced system with detectability property.
ep
This property allows to construct an open-loop observer (Besancon 2007, Sarinana et al. 2000).
3.1.2 Construction of the Nonlinear Observer

c
Assume that the only measured variable is the output DC voltage, i.e. y = x3 = U0 . Denote
Ac
the observation errors e1 = id id , e2 = iq iq , e3 = U0 U0 . Then, a ST SMO for (6) is

constructed as follows

did r U0

= id + iq ud + k1 (e3 ),

dt L 2L

diq r U0 E
= iq id uq + + k2 (e3 ), (10)
dt
L 2L L

dU0 U0 3
= + (id ud + iq uq ) + (e3 ).
dt RL C 4C
where the STA is defined as

Z t
1
(e3 ) = |e3 | 2 sign(e3 ) + sign(e3 )d. (11)
0
with some positive constants , respectively.

Then the error dynamics is given by
r
e1 = e1 + e2 k1 (e3 ), (12)
L
r
e2 = e1 e2 k2 (e3 ), (13)
L
t
3
ip
e3 = (ud e1 + uq e2 ) (e3 ). (14)
4C
According to the Park transformation (4, 5) and the switching functions u1 , u2 , u3 which take
cr
values from the discrete set {1, +1}, the control variables ud and uq satisfy the following
saturation constraint
us
q r
2
(15)
kudq k kT kkuk = max (T T T )kuk 3 = 2.
3
an
where k k denotes the Euclidean norm of a vector or the spectral norm of a matrix.
The design problem is transformed into determining , and k1 , k2 which are the tuning
M
parameters to ensure the convergence of the error system (12, 13, 14).
Proposition 3.2: Consider the system (14) under the condition (15). Then, the trajectories of
d
the system (14) converge to zero in finite time, and the resulting reduced order dynamics (12,13)
are exponentially stable, if the gains , of the STA and tuning parameters k1 , k2 are chosen as
te
(Levant 1998),
ep
> F, 2 > . (16)
ud , if |e3 | = 0 uq , if |e3 | = 0
k1 = , k2 = (17)

Ac
0. otherwise 0. otherwise
where F and are some positive constants.
Proof The proof is divided into two steps. Firstly, the Eq. (14) is proven to be finite time stable.
Secondly, the resulting reduced order dynamics (12,13) are proven to be exponentially stable
with faster convergence rate than its open loop dynamics.
We start from |e3 | 6= 0, thus, the two correction gains k1 , k2 are zero according to Eq.(17).
h iT
Denoting e12 = e1 , e2 , the system (12,13) becomes its open loop dynamics
r

e12 = L r e12 . (18)

| {z L }
A
The system (18) is exponentially stable given that A is a Hurwitz matrix. Consequently, it follows
that
t
ke12 (t)k ke12 (t0 )k,
(19)
ip
ke12 (t)k kAkke12 (t)k.
cr
q
r2
where t0 is the initial value of time and kAk = max (AT A) = 2 + .
L2
In view of the Eqs. (18, 19), The Eq. (14) can be rewritten as
us

1
e3 = |e3 | 2 sign(e3 ) + ,
(20)

an
= sign(e3 ) + g(e1 , e2 , ud , uq ).
3 d
where g(e1 , e2 , ud , uq ) = (ud e1 (t) + uq e2 (t)). Due to the fact that the derivative of the
M
4C dt
control variable udq designed in Eqs.(41,42) is bounded, i.e., kudq k Udq . It follows that
3
|g(e1 , e2 , ud , uq )| (kudq kke12 (t)k + kudq kke12 (t)k)
d
4C
(21)
3
ke12 (t)k Udq + 2kAk .
te
4C
where Udq is a positive value and g is bounded by a decreasing function given that e12 (t) converges
ep
to zero exponentially. Thus, after certain time, there exists a positive constant value F such that
F > |g(e1 , e2 , ud , uq )|.
Given that the gains of the STA are chosen as (16), e3 , e3 converge to zero in finite time
c
(Levant 1998). Thereafter, the equivalent output-error injection (e3 ) in (14) can be obtained
Ac
directly without any low pass filters (Floquet and Barbot 2007),
3
(e3 ) = (ud e1 + uq e2 ). (22)
4C
In the second step of the proof, the convergence of the reduced order dynamics (12,13) is
performed. Once the reaching phase is arrived (e3 = 0, e3 = 0) and maintains on it thereafter,
the gains k1 , k2 will switch according to (17). Substitute the equivalent injection (22) into the
system (12,13), one can get the reduced order dynamics
12 .
e12 = Ae12 Ae (23)

3 u2d ud uq
where A = .
4C ud uq u2
q
Consider a candidate Lyapunov function for system (23)
V (e12 ) = eT12 P e12 . (24)

L
0
t

where P = 2r L which satisfies AT P + P A = I22 , and I22 is an identity matrix.
ip
0
2r
The time derivative of V along the trajectories of system (23) is presented
cr
V (e12 ) = eT12 e12 eT12 (AT P + P A)e
12
(25)
3L
us
= ke12 k2 (ud e1 + uq e2 )2 .
4rC
It can seen from Eq. (25) that V (e12 ) = ke12 k2 when kudq k = 0. For any positive and
kudq k 6= 0, Eq. (25) can be rewritten as
an
L T
V (e12 ) = ke12 k2 e Ae12 . (26)
r 12
M
which ensures faster convergence rate compared with the open loop dynamic (18) given that the
matrix A is positive semi-definite (p.s.d). The Proposition 3.2 is proven.
d
Remark 1 : It should be noted that e3 may cross zero during its reaching phase. Assume that
te
at time instant t = t1 that e3 crosses zero. Then, according to the condition in (17), Eq. (18)
becomes
ep
e12 = Ae12 + . (27)
where = (e3 )udq . Due to the fact that e3 does not exhibit finite time escape, and is a
c
bounded function, we can write the solution as (Hassan K. Khalil 2007)

Ac
Z t1 +t
e12 (t) = e(tt0 )A e12 (t0 ) + e(t )A d. (28)
t1 t
where t is arbitrary small. That is, the second term in the right side of Eq.(28) can be considered
as zero. Therefore, the condition (19) holds for all t t0 .
Remark 2 : Concerning the practical implementation of (17), the condition e3 = 0 in (17) can
not be satisfied in the presence of measurement noise and numerical approximations (Pisano and
Usai 2007). This obstacle can be overcome by modifying the condition e3 = 0 using dead-zone
technique (Slotine et al. 1991), as

ud ,
uq ,
if |e3 | < if |e3 | <
k1 = , k2 = (29)

0.
0.
otherwise otherwise
where in general is a small positive value.
In the next Subsection, an output feedback ST current control is designed in order to achieve
t
the objective of unity power factor and ripple free output voltage.
ip
cr
3.2 Output Feedback ST Sliding Mode Current Control
The STA is popular among the SOSM algorithms because it is an absolutely continuous sliding
us
mode algorithm, therefore it does not suffer from the problem of chattering (Levant 1993, 2007).
The main advantages of the ST SMC (Vadim Utkin, Jurgen Guldner, Jingxin Shi 2009) are as
follows:
an
(1) It does not need the evaluation of the time derivative of the sliding variable;
(2) Its continuous nature suppresses arbitrary disturbances with bounded time derivatives;
M
The control objectives are defined in the Subsection 2.2.
3.2.1 Desired current calculation with unity power factor

d
Normally, the value of the inductance L 1 in the system (6), and the right-hand sides of the
te
did diq dU0

equations in (6) have the values of the same order. Hence , , implying that the
dt dt dt
dynamics of id and iq are much faster than those of U0 (Vadim Utkin, Jurgen Guldner, Jingxin
ep
Shi 2009). Provided that the fast dynamics are stable, based on the singular perturbation theory
(Hassan K. Khalil 2007), let the first and second equations of (6) be zero formally, then ud , uq
can be obtained,
c
U r
Ac
0

ud = id + iq ,

2L L

U
0 r E
uq = iq id + , (30)

2L L L

dU0 = U0 + 3 (i ud + iq uq ).

dt RL C 4C d
Based on Eq. (30), the reference currents id , iq will be determined depending on the desired
system performance. Substitute the first and second equation into the third equation of (30)
yields,
dU0 U0 3 iq E Er
= + . (31)
dt RL C 2 U0 C
where Er = r(id 2 + iq 2 ) represents the power consumed by parasitic phase resistance r.

Considering the Power-Balance condition (Komurcugil and Kukrer 1998),
3 U 2
(iq E Er ) = 0 . (32)
2 RL
The reference current id is set to zero for guaranteeing Unity-Power-Factor which leads to the
t
ip
following calculation of the reference current iq ,
s
E2 8U02
cr
E 1
iq = . (33)
2r 2 r2 3RL r
us
under constraint for desired output voltage U0 ,
r
3RL
U0 E . (34)
8r
an
Finally, id and iq are obtained as follows (due to minimal energy consumption),

M

i =0,

d
s (35)

E 1 E2 8U02

iq = .
2r 2 r2 3RL r
d
As tracking error vector approaches zero, i.e., id id and iq iq , the zero dynamics have the
te
form (Lee 2003)

ep
dU0 U0 U02
= + . (36)
dt RL C RL CU0
Define a new variable Z = U02 , Eq. (36) can be rewritten as,

c
Ac
dZ 2
= (Z U02 ). (37)
dt RL C
For a positive initial value of the output voltage, the steady-state value of U0 will converge to
RL C
the desired level U0 with the time constant exponentially. Therefore, the tracking of the
2
reference current achieves the regulation of output voltage to the desired value U0 with a unity
power factor. In the following Subsection 3.2.2, the design of current control based on STA is
proposed.
3.2.2 Output sliding mode current control
We will now design ST SMC for the system (6) based on the proposed observer (10). The
sliding variables for the current control are defined as,

sd =id id = id id + id id = sd e1 ,
(38)

sq =i iq = i iq + iq iq = sq e2 .
q q
where id and iq are the desired values of the currents in the (d, q) coordinate frame and e1 , e2
are observation errors. The desired value is selected to provide the DC power balance between
t
the input power and the output power.
ip
h iT
Taking the first time derivative of sdq = sd sq yields,
cr
r
id iq u
sdq = L + U0 d . (39)
r E 2L uq
iq + iq + id
us
L L
The Eq. (39) can be rewritten as, an

r
sd sd + sq d u
= L + + U0 d . (40)
r 2L uq
sq sd sq q
L
M
r E
where d = iq , q = iq +i .
L q L
The control objective is to force the sliding variable sd , sq to zero. In order to satisfy the
d
saturation constraint (15), the controls ud , uq are designed as follows,

te

ud (md )
udq = = . (41)
uq (mq )
ep
where md , mq are designed as

c

2L r
sd sq (sd ) d , if ksdq k > Sdq
md = U0 L
Ac

2L r sd sq kd sd d . if ksdq k Sdq

U0 L
(42)

2L r
sd + sq (sq ) q , if ksdq k > Sdq
mq = U 0 L

2L r
sd + sq kq sq q . if ksdq k Sdq
U0 L
h iT
(sd ), (sq ) take the form of (11), sdq = sd sq , kd , kq , Sdq are some positive constants and
() is the standard saturation function
x
(x) = . (43)
max{1, |x|}
Remark 3 : For digital implementation of the controller (41), Digital Signal Processors (DSP)
is a good choice due to complicated calculations (Sabanovic et al. 2004). Since the control signals
(41) are continuous, a digital sawtooth signal is used to compare with the controller (41) in order
to produce the PWM pulses which can directly control the AC/DC converter in the real world.
t
3.3 Closed-loop stability analysis
ip
It is well known that the so-called separation principle is available for linear systems, may not
hold for nonlinear systems in general (Mazenc et al. 1994, Hong et al. 2000). This means that the
cr
observer and the controller can not be designed separately Nollet et al. (2008). Thus, Lyapunov
analysis of the closed-loop system is employed to prove the convergence of the system (40) by
us
taking the observation errors into account.
Theorem 3.3 : Consider the system (6) in closed loop with the saturated controller (41). The
an
state trajectories of the system (40) locally exponentially converge to the origin sd = 0, sq = 0,
moreover, the output voltage locally exponentially converge to the desired constant voltage U0 if
the gains of STA (sd ), (sq ) and U0 are chosen such that the following conditions are satisfied,
M
d >Fdq , 2d > d .
(44)
d
q >Fdq , 2q > q .
te
q 2 2
U0 > 2 Liq + E riq . (45)
ep
where Fdq is some arbitrary positive constant.
Proof The proof is divided into two parts, firstly, we consider the case when the control vector
h iT
c

kudq k 2, thus we choose udq = md mq . Secondly, we consider the case when the control

Ac
vector kudq k > 2, thus the control is chosen as kudq k = 2.

Case 1 : The controller is given by ud = md and uq = mq in Eq. (42). When ksdq k > Sdq , then
Eq. (40) can be rewritten as

sd (s ) 2
u u u e
= d 3 d d q 1 .
s q (sq ) 4C ud uq u 2
e2 (46)
q
| {z }
(ud ,uq ,e1 ,e2 )
where the function (ud , uq , e1 , e2 ) can be rewritten as
3 T
(ud , uq , e1 , e2 ) = u udq e12 (t). (47)
4C dq
It can be deduced from Eq.(42) that the derivative of the control variable udq is bounded, i.e.,
kudq k Udq . In view of Eqs. (15), one can write

d
= 3 2uT udq e12 (t) + uT udq e12 (t)
dt 4C dq dq
(48)
3
ke12 (t)k 2Udq + kAk .
t
2C
ip
It is shown from Eq.(48) that the derivative of the function (ud , uq , e1 , e2 ) is bounded by a
cr
decreasing function given that e12 (t) converges to zero exponentially.
Thus, after certain time,
d
there exists a positive constant value Fdq such that Fdq >
dt .
According to the result of Levant (1998), the trajectories of the system (46) will converge to
us
D = {sdq Rn | ksdq k Sdq } in finite time with the gains chosen as (44).
Inside the set ksdq k Sdq , according to the Eq. (42), Eq. (40) is presented as

an
s d kd 0 s
= d 3 uTdq udq e12 (t). (49)
s q 0 kq sq 4C
M
The solution of Eq. (49) is presented

Z t
3 T
d
(tt0 )K
sdq (t) = e sdq (t0 ) e(t )K u ( )udq ( )e12 ( )d, (50)
t0 4C dq
te

kd 0
where the matrix K = is Hurwitz.
ep
0 kq
Let us consider the bound ke(tt0 )K k k1 e1 (tt0 ) for some positive constants k1 and 1
(Hassan K. Khalil 2007). As Proposition 3.2 implies that e12 (t) converges to zero exponentially,
c
one can get ke12 (t)k k2 e2 (tt0 ) for some positive constants k2 and 2 . It follows that
Ac
Z t
1 (tt0 ) 3 T
ksdq (t)k k1 e ksdq (t0 )k + k1 e1 (t )
ku ( )udq ( )kke12 ( )kd
t0 4C dq
Z t
3
k1 e1 (tt0 ) ksdq (t0 )k + k1 k2 sup kuTdq ( )udq ( )k e1 (t ) e2 ( t0 ) d
4C t0 t t0 (51)
e2 (tt0 ) e1 (tt0 )
k1 e1 (tt0 ) ksdq (t0 )k + 3k1 k2
2C(1 2 )
= c1 e1 (tt0 ) + c2 e2 (tt0 ) ,
3k1 k2 3k1 k2
where c1 = k1 ksdq (t0 )k , c2 = are some constant values.
2C(1 2 ) 2C(1 2 )
It follows from (51) that sd and sq converge to zero exponentially. Given that e1 and e2
converge to zero exponentially, we can also conclude from Eq. (38), that sd and sq converge to
zero exponentially.
Case 2 : According to the result of Escobar et al. (1998), the sufficient condition for the control

vector udq to enter into the circle of radium 2, (i.e., kudq k 2) is that
q 2 2
U0 > 2 Liq + E riq . (52)
t
ip
Thus, Theorem 3.3 is proven.
Remark 4 : For practical implementation of the proposed controller, it is important to select
cr
the value of Fdq as the controllers gains depend on this bound. As can be seen from the Eq. (48)
that the value Fdq depends on e12 (t) which converges to zero exponentially ( lim ke12 (t)k = 0).
us
t
d

Thus, after certain time, a sufficient large value can be chosen for Fdq such that Fdq > .
dt
The larger value of kd and kq will ensure faster convergence of sd and sq to zero. Therefore, the
an
controllers gains can be simply chosen such that the condition (44) is fulfilled.
The proposed observer-based control law (41) requires real-time evaluation of sliding variables
M
sd = id id , sq = iq iq . However, the current reference iq in (35) requires the knowledge of
load resistance RL and parasitic phase resistance r. Due to this fact, a ST parameter observer
is employed to estimate the value of load resistance while phase resistance is assumed to have
d
nominal value (Shtessel et al. 2008).

te
ep
4 ST PARAMETER OBSERVER DESIGN AND POWER FACTOR ESTIMATION
4.1 Load Resistance Estimation

c
In this work, the load resistance RL in the system is assumed to vary around its nominal value
Ac
R0 . The last differential equation in (6) is used to construct the observer dynamics using ST
sliding mode technique
dU0 U0 3
= + (id ud + iq uq ) + (U0 ). (53)
dt R0 C 4C
where R0 is the nominal value of the load resistance.

Consider the observation error U0 = U0 U0 which has the following dynamics according to
(6) and (53),

U0 1 1
U 0 = (U0 )
C RL R0
(54)
3
+ (e1 ud + e2 uq ) = (U0 ) + RL .
4C
where () is the STA defined in (11). From Proposition 3.2, lim e1 = 0, lim e2 = 0 and
t t
U0 1 1
lim RL = . Sliding mode will be enforced with appropriate values of ,
t C RL R0
providing that the first time derivative of the term RL is bounded Levant (1993, 2007). It
follows that when a sliding motion takes place,
t
ip

U0 1 1 3
(U0 ) + (e1 ud + e2 uq ) = 0. (55)
C RL R0 4C
cr
The load resistance RL can therefore be estimated in terms of its nominal value and observers
output with appropriate parameters exponentially,
us
R0 U0
RL = . (56)
U0 R0 C(U0 )
an
since
lim kRL RL k = 0.
M
(57)
t
4.2 Power Factor Estimation

d
Almost all utility interfaced applications, especially in the electric vehicle industry concern about
te
power quality, i.e. higher power factor and less harmonic pollution. The unity-value power factor
converter does not introduce any distortion (phase current has only main harmonic) to the
ep
energy source, phase shift between input phase current and main source voltage and maximizes
the performance factor of the power conversion (Egan et al. 2007, Shtessel et al. 2008). Due to
the fact that one of our control objective is to obtain the unity power factor, thus, the estimation
c
of power factor value is of prime importance for analyzing the quality of the proposed observer-
Ac
based control law.

The definition of power factor is given as the following formula,
RMS(i1 (t))
PF = PFh PFd = cos(). (58)
RMS(i(t))
where PFh is the harmonic distortion and PFd is the displacement between input phase current
and source voltage.
The RMS( ) stands for the root-mean-square quantity which is calculated as follows
s
Z T
1
RMS(i(t)) = i2 ( )d . (59)
T 0
where T is the period of the phase current i(t), RMS(i1 (t)) characterizes the fundamental com-
ponent of the current (root-mean-square) and RMS(i(t)) corresponds to the total current (root-
mean-square).
The overall power factor for the three-phase AC/DC converter is calculated from the estimates
of phase current which will be a product of the three single phase power factor values,
t
ip
PFtotal =PFa PFb PFc
(60)
cr
RMS(ia1 (t)) RMS(ib1 (t)) RMS(ic1 (t))
= cos(a ) cos(b ) cos(c ).
RMS(ia (t)) RMS(ib (t)) RMS(ic (t))
us
where ia , ib , ic (t) and ia1 , ib1 , ic1 (t) are the estimates of the input phase currents, their funda-
mental components respectively. a , b , c are the phase shift between the input currents and
the source voltages.
an
The structure of single-phase power factor estimation is shown in Fig. 4, which includes two
important modules of Matlab/Simulink: Fourier analysis and Harmonic analysis. The first mod-
M
ule gives the phases of main-frequency input current and source phase voltage respectively. The
other module is used to measure the total harmonic distortion of input current.
d
5 SIMULATION RESULTS
te
The multi-rate simulation of the proposed three phase AC/DC boost power converter has been
ep
carried out, the parameters used in simulation are shown in Table 1. Multi-rate approach realizes
the achievement of realistic simulation results by taking into account some implementation issues:
c
(1) The control evaluation rate f2 is less than the simulation rate f1 (i.e., the integration
was carried out according to the Euler method).
Ac
(2) The power elements switch rate f3 is less than the control evaluation rate f2 due to
switching loss.
The simulation is made in order to test the robustness of the proposed controllers with respect
to the measurement noise in the system output. The function used in the simulation is
depicted in Fig. 5. Load resistance and frequency are varied to test controllers ability to handle
with varying conditions at time 1.0s and 1.5s respectively (see Table 1 for details).
The simulation results of the proposed observer-based ST SMC compared with a well tuned
linear PI regulator (Silva 1999) are shown in Figs. 6-9. Input phase current along with the
corresponding source voltage are shown in Fig. 6. It can be seen that both of the controller
make no phase shift between the input current and corresponding source voltage, however the
PI Control results in higher harmonics compared to the ST SMC. Fig. 7 shows the output
performance of the AC/DC converter. It proves that the observer-based ST SMC is rather robust
against disturbances (i.e. load varying and frequency varying), while the linear PI control results
in higher fluctuation around some DC level, and higher voltage overshoot compared with the
observer-based ST SMC. The proposed observer-based ST SMC is able to regulate the output
t
voltage to the desired level under the varying conditions. A good estimate for the load resistance
ip
is shown in Fig. 8. However, the PI Control results in higher fluctuation around the DC level
cr
and can not force the output voltage to its reference level compared with the observer-based ST
SMC. The PI control is not able to demonstrate its robustness with respect to load variation,
us
due to the fact that the gains kp , ki of the PI control depend on the load resistance RL (Silva
1999). It should be noted that the STA works as a robust differentiator and Levant has given
an estimate of the observation error with respect to the measurement noise in (Levant 2003).
an
Fig. 9 shows the separate power factor value of each phase and their product as a combined
characteristic of the AC/DC converter. It can be seen from the Figs. 9a and Fig. 9b that the
power factor value is greater than 97% in case of observer-based ST SMC, while the PI Control
M
results in less value (around 86%).
d
6 CONCLUSIONS
te
An observer-based ST SMC is proposed in this paper for the AC/DC boost converters. The use of
observer reduces the number of current sensors, decreases the system cost, volume and provides
ep
robustness to the change of operational condition (e.g. in load resistance RL and frequency of
the source voltage ). The input phase currents and load resistance are estimated only from
the measurement of output voltage. The proposed observer-based ST SMC maintains the power
c
factor close to unity. A strong Lyapunov function is introduced to prove the stability of the
Ac
observer and controller as a whole system. Simulation results show that the observer-based
controller performs better, compared to conventional PI control, with less overshoot and less
sensitivity to disturbance and parametric uncertainty.
20 REFERENCES
References
Allag, A., Hammoudi, M., Mimoune, S., Ayad, M., Becherif, M., and Miraoui, A. (2007), Track-
ing Control Via Adaptive Backstepping Approach for a Three Phase PWM AC-DC Con-
verter, in IEEE International Symposium on Industrial Electronics, ISIE, IEEE, pp. 371376.
Amjadi, Z., and Williamson, S. (2010), Power Electronics Based Solutions for Plug-In Hybrid
Electric Vehicle Energy Storage and Management Systems, IEEE Transactions on Industrial
Electronics, 57, 608616.
Andersen, B., Holmgaard, T., Nielsen, J., and Blasbjerg, F. (1999), Active Three-phase Rectifier
t
with only One Current Sensor in the DC-link, in Proceedings of the IEEE International
ip
Conference on Power Electronics and Drive Systems, Vol. 1, IEEE, pp. 6974.
Besancon, G., Nonlinear observers and applications, Vol. 363, Springer Verlag (2007).
cr
Bose, B.K., Modern Power Electronics and AC Drives, Prentice Hall (2002).
Camara, M., Gualous, H., Gustin, F., Berthon, A., and Dakyo, B. (2010), DC/DC Converter
us
Design for Supercapacitor and Battery Power Management in Hybrid Vehicle Applications-
Polynomial Control Strategy, IEEE Transactions on Industrial Electronics, 57, 587597.
an
Cao, J., and Emadi, A. (2009), A New Battery/Ultra-Capacitor Hybrid Energy Storage Sys-
tem for Electric, Hybrid and Plug-In Hybrid Electric Vehicles, in IEEE Vehicle Power and
Propulsion Conference, VPPC09, IEEE, pp. 941946.
M
Cecati, C., DellAquila, A., Lecci, A., and Liserre, M. (2005), Implementation Issues of a Fuzzy-
Logic-Based Three-Phase Active Rectifier Employing Only Voltage Sensors, IEEE Transac-
tions on Industrial Electronics, 52, 378385.
d
Dixon, J., and Ooi, B.T. (1988), Indirect Current Control of a Unity Power Factor Sinusoidal
te
Current Boost Type Three-Phase Rectifier, IEEE Transactions on Industrial Electronics, 35,
508515.
ep
Edwards, C., and Spurgeon, S., Sliding Mode Control: Theory and Applications, Taylor & Francis
(1998).
Egan, M., OSullivan, D., Hayes, J., Willers, M., and Henze, C. (2007), Power Factor Cor-
c
rected Single Stage Inductive Charger for Electric Vehicle Batteries, IEEE Transactions on
Ac
Industrial Electronics, 54, 12171226.

Emadi, A., Lee, Y.J., and Rajashekara, K. (2008), Power Electronics and Motor Drives in Elec-
tric, Hybrid Electric, and Plug-In Hybrid Electric Vehicles, IEEE Transactions on Industrial
Escobar, G., Chevreau, D., Ortega, R., and Mendes, E. (2001), An Adaptive Passivity-Based
Controller for a Unity Power Factor Rectifier, IEEE Transactions on Control Systems Tech-
REFERENCES 21
nology, 9, 637644.
Escobar, G., Ortega, R., and Van Der Schaft, A. (1998), A saturated output feedback controller
for the three phase voltage sourced reversible boost type rectifier, in Proceedings of the 24th
Annual Conference of the IEEE Industrial Electronics Society, IECON 98., Vol. 2, pp. 685
690.
Floquet, T., and Barbot, J. (2007), Super twisting algorithm-based step-by-step sliding mode
observers for nonlinear systems with unknown inputs, International Journal of Systems Sci-
ence, 38, 803815.
Guerrero, J., Loh, P.C., Lee, T.L., and Chandorkar, M. (2013), Advanced Control Architectures
t
for Intelligent Microgrids;Part II: Power Quality, Energy Storage, and AC/DC Microgrids,
ip
IEEE Transactions on Industrial Electronics, 60, 12631270.
cr
Hamida, M., De Leon, J., Glumineau, A., and Boisliveau, R. (2013), An Adaptive Intercon-
nected Observer for Sensorless Control of PM Synchronous Motors With Online Parameter
us
Identification, IEEE Transactions on Industrial Electronics, 60, 739748.
Hassan K. Khalil,, Nonlinear systems, Vol. 3, Prentice hall New Jersey (2007).
Hayes, J., Egan, M., Murphy, J., Schulz, S., and Hall, J. (1999), Wide-Load-Range Resonant
an
Converter Supplying the SAE J-1773 Electric Vehicle Inductive Charging Interface, IEEE
Transactions on Industry Applications, 35, 884895.
Hermann, R., and Krener, A. (1977), Nonlinear Controllability and Observability, IEEE Trans-
M
actions on Automatic Control, 22, 728740.
Hong, Y., Yang, G., Bushnell, L., and Wang, H.O. (2000), Global finite-time stabilization: from
d
state feedback to output feedback, in Proceedings of the 39th IEEE Conference on Decision
and Control, Vol. 3, IEEE, pp. 29082913.
te
Hornik, T., and Zhong, Q.C. (2013), Parallel PI VoltageH Current Controller for the Neutral
Point of a Three-Phase Inverter, IEEE Transactions on Industrial Electronics, 60, 13351343.
ep
Houari, A., Renaudineau, H., Martin, J., Pierfederici, S., and Meibody-Tabar, F. (2012),
Flatness-Based Control of Three-Phase Inverter With Output Filter, IEEE Transactions
on Industrial Electronics, 59, 28902897.
c
Komurcugil, H., and Kukrer, O. (1998), Lyapunov-Based Control for Three-Phase PWM
Ac
AC/DC Voltage-Source Converters, IEEE Transactions on Power Electronics, 13, 801813.

Kuperman, A., Levy, U., Goren, J., Zafransky, A., and Savernin, A. (2012), Battery Charger
for Electric Vehicle Traction Battery Switch Station, IEEE Transactions on Industrial Elec-
tronics, PP, 11.
Lee, D.C., Lee, G.M., and Lee, K.D. (2000), DC-bus Voltage Control of Three-Phase AC/DC
PWM Converters Using Feedback Linearization, IEEE Transactions on Industry Applica-
22 REFERENCES
tions, 36, 826833.

Lee, D.C., and Lim, D.S. (2002), AC Voltage and Current Sensorless Control of Three-Phase
PWM Rectifiers, IEEE Transactions on Power Electronics, 17, 883890.
Lee, T.S. (2003), Input-Output Linearization and Zero-Dynamics Control of Three-Phase
AC/DC Voltage-Source Converters, IEEE Transactions on Power Electronics, 18, 1122.
Lee, W.C., Lee, T.K., and Hyun, D.S. (2001), Comparison of Single-Sensor Current Control
in the DC Link for Three-Phase Voltage-Source PWM Converters, IEEE Transactions on
Industrial Electronics, 48, 491505.
Lee, Y.J., Khaligh, A., and Emadi, A. (2009), Advanced Integrated Bidirectional AC/DC and
t
DC/DC Converter for Plug-In Hybrid Electric Vehicles, IEEE Transactions on Vehicular
ip
Technology, 58, 39703980.
cr
Levant, A. (1993), Sliding Order and Sliding Accuracy in Sliding Mode Control, International
Journal of Control, 58, 12471263.
us
Levant, A. (1998), Robust Exact Differentiation via Sliding Mode Technique, Automatica, 34,
379384.
Levant, A. (2003), Higher-order sliding modes, differentiation and output-feedback control,
an
International Journal of Control, 76, 924941.
Levant, A. (2007), Principles of 2-Sliding Mode Design, Automatica, 43, 576586.
Liu, X., Loh, P.C., Wang, P., and Blaabjerg, F. (2013), A Direct Power Conversion Topology for
M
Grid Integrations of Hybrid AC/DC Resources, IEEE Transactions on Industrial Electronics,
PP, 110.
d
Mazenc, F., Praly, L., and Dayawansa, W. (1994), Global stabilization by output feedback:
examples and counterexamples, Systems & Control Letters, 23, 119125.
te
Nollet, F., Floquet, T., and Perruquetti, W. (2008), Observer-based second order sliding mode
control laws for stepper motors, Control Engineering Practice, 16, 429443.
ep
Pahlevaninezhad, M., Das, P., Drobnik, J., Jain, P., and Bakhshai, A. (2012b), A New Control
Approach Based on the Differential Flatness Theory for an AC/DC Converter Used in Electric
Vehicles, IEEE Transactions on Power Electronics, 27, 20852103.
c
Pahlevaninezhad, M., Das, P., Drobnik, J., Moschopoulos, G., Jain, P., and Bakhshai, A. (2012a),
Ac
A Nonlinear Optimal Control Approach Based on the Control-Lyapunov Function for an

AC/DC Converter Used in Electric Vehicles, IEEE Transactions on Industrial Informatics,
8, 596614.
Pan, C., and Chen, T. (1993), Modelling and Analysis of a Three Phase PWM AC-DC Convertor
Without Current Sensor, IEE Proceedings B, Electric Power Applications, 140, 201208.
Pisano, A., and Usai, E. (2007), Globally convergent real-time differentiation via second order
REFERENCES 23
sliding modes, International Journal of Systems Science, 38, 833844.

Sabanovic, A., Fridman, L., and Spurgeon, S.K., Variable Structure Systems: From Principles
to Implementation, IET (2004).
Sarinana, A., Bacha, S., and Bornard, G. (2000), On Nonlinear Observers Applied to Three-
Phase Voltage Source Converters, in 31st IEEE Annual Power Electronics Specialists Con-
ference, Vol. 3, IEEE, pp. 14191424.
Shtessel, Y., Baev, S., and Biglari, H. (2008), Unity Power Factor Control in Three-Phase
AC/DC Boost Converter Using Sliding Modes, IEEE Transactions on Industrial Electronics,
55, 38743882.
t
Silva, J. (1997), Sliding Mode Control of Voltage Sourced Boost-Type Reversible Rectifiers,
ip
in Proceedings of the IEEE International Symposium on Industrial Electronics, Vol. 2, IEEE,
cr
pp. 329334.
Silva, J. (1999), Sliding Mode Control of Boost-Type Unity-Power-Factor PWM Rectifiers,
us
IEEE Transactions on Industrial Electronics, 46, 594603.
Sira-Ramirez, H., Escobar, G., and Ortega, R. (1996), On Passivity-Based Sliding Mode Control
of Switched DC-to-DC Power Converters, in Proceedings of the 35th IEEE Decision and
Control, Vol. 3, IEEE, pp. 25252526.
an
Slotine, J.J.E., Li, W. et al., Applied nonlinear control, Vol. 1, Prentice Hall New Jersey (1991).
Tan, S.C., Lai, Y., Tse, C., Martinez-Salamero, L., and Wu, C.K. (2007), A Fast-Response
M
Sliding-Mode Controller for Boost-Type Converters With a Wide Range of Operating Condi-
tions, IEEE Transactions on Industrial Electronics, 54, 32763286.
d
Thounthong, P. (2012), Control of a Three-Level Boost Converter Based on a Differential Flat-

ness Approach for Fuel Cell Vehicle Applications, IEEE Transactions on Vehicular Technol-
te
ogy, 61, 14671472.

Vadim Utkin, Jurgen Guldner, Jingxin Shi,, Sliding Mode Control in Electro-Mechanical Systems,
ep
CRC Press, Taylor and Francis Group (2009).

Wirasingha, S., and Emadi, A. (2011), Classification and Review of Control Strategies for
Plug-In Hybrid Electric Vehicles, IEEE Transactions on Vehicular Technology, 60, 111122.
c
Zhong, Q.C., and Hornik, T. (2013), Cascaded Current-Voltage Control to Improve the Power
Ac
Quality for a Grid-Connected Inverter With a Local Load, IEEE Transactions on Industrial
24 REFERENCES
t
ip
cr
us
Figure 1. PEMFC powered hybrid system consists of Source Grid, Fuel Cell, Super-capacitor and battery
an
M
d
te
Figure 2. Electrical circuit of the three phase AC/DC boost converter

c ep
Ac
REFERENCES 25
Figure 3. Observer-Based Control Structure of a Three Phase AC/DC Converter
t
ip
cr
us
an
Figure 4. Structure of the single-phase power factor estimation
M
d
te
cep
Ac
26 REFERENCES
t
0.05
ip
0.04
0.03
Measurement Noise
0.02
cr
0.01
0
us
0.01
0.02
0.03
0.04
Noise:
0.05
0 0.1
an
0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
Time (s) 3
x 10
M
d
te
ep
Figure 5. Measurement noise affecting the system output

c
Ac
REFERENCES 27
(a) The case of observer-based ST SMC (b) The case of PI Control
Figure 6. Phase current and source voltage (0.2)
t
ip
cr
us
an
(a) The case of observer-based ST SMC (b) The case of PI Control
Figure 7. Output Voltage Performance

M
d
te
cep
Ac
Figure 8. Load Resistance Estimation Via Super-Twisting Observer

28 REFERENCES
(a) The case of super-twisting SMC (b) The case of PI Control
Figure 9. Power factor of the AC/DC converter
t
ip
cr
us
an
M
d
te
c ep
Ac
REFERENCES 29
Table 1. Parameters Used For Simulation
Parameter Value Description
f1 106 Simulation Rate, (Hz)

5
f2 10 Controller Evaluation Rate, (Hz)
f3 104 Pulse Width Modulator Rate, (Hz)
r 0.02 Parasitic Phase Resistance,
t=1.0s
RL 50 40 Load Resistance,
C 100 Output Capacitor, uF
L 2 Phase Inductor, mH
t=1.5s
w 150 300 Source Voltage Angular Speed, rad/s
t
E 150 Source Voltage, V
U0
ip
650 Desired Output Voltage, V
U0 (t0 ) 5 Initial Value of the Output Voltage, V
, , , 5, 2 103 , 5 102 , 0.01 Current Observer Parameters in (10)
cr
d = q , d = q , kd , kq , Sdq 200, 20, 5, 5, 0.05 ST SMC Parameters in (41)
kp , ki , 500, 20 PI Control Parameters
us
an
M
d
te
cep
Ac

Liu 2014

Загружено:

Сведения о документе

Авторское право

Доступные форматы

Поделиться этим документом

Поделиться или встроить документ

Параметры публикации

Этот документ был вам полезен?

Это неприемлемый материал?

Авторское право:

Доступные форматы

Liu 2014

Загружено:

Авторское право:

Доступные форматы

This article was downloaded by: [Aston University]

On: 09 January 2014, At: 12:10

International Journal of Control

Observer-Based Higher Order Sliding Mode Control

To link to this article: http://dx.doi.org/10.1080/00207179.2013.868609

PLEASE SCROLL DOWN FOR ARTICLE

International Journal of Control

Observer-Based Higher Order Sliding Mode Control of Power Factor in

Three-Phase AC/DC Converter for Hybrid Electric Vehicle Applications

ISSN: 0020-7179 print/ISSN 1366-5820 online

2 International Journal of Control

International Journal of Control 3

uncertainty and external disturbance.

4 International Journal of Control

2.1 System Modeling

1 = 1 which corresponds to the

International Journal of Control 5

It can also be written as,

2 cos(t) cos(t 3 ) cos(t + 3 )

is the Parks transformation (Bose 2002, Lee 2003).

6 International Journal of Control

2.2 Control Objectives

AC component has to be attenuated to a given level.

3.1 Super-Twisting Sliding Mode Observer Design

The proposed observer is designed in the following two steps,

(1) Analyzing the observability of the nonlinear system;

3.1.1 Observability Analysis of the System

In order to construct an observer for a system, it is necessary to verify the observability of

International Journal of Control 7

3.1.2 Construction of the Nonlinear Observer

the observation errors e1 = id id , e2 = iq iq , e3 = U0 U0 . Then, a ST SMO for (6) is

8 International Journal of Control

where the STA is defined as

with some positive constants , respectively.

> F, 2 > . (16)

where F and are some positive constants.

International Journal of Control 9

10 International Journal of Control

system (12,13), one can get the reduced order dynamics

V (e12 ) = eT12 P e12 . (24)

e12 = Ae12 + . (27)

bounded function, we can write the solution as (Hassan K. Khalil 2007)

International Journal of Control 11

where  in general is a small positive value.

3.2.1 Desired current calculation with unity power factor

did diq dU0

12 International Journal of Control

where Er = r(id 2 + iq 2 ) represents the power consumed by parasitic phase resistance r.

form (Lee 2003)

Define a new variable Z = U02 , Eq. (36) can be rewritten as,

International Journal of Control 13

3.2.2 Output sliding mode current control

The Eq. (39) can be rewritten as, an

saturation constraint (15), the controls ud , uq are designed as follows,

where md , mq are designed as

14 International Journal of Control

() is the standard saturation function

where Fdq is some arbitrary positive constant.

vector kudq k > 2, thus the control is chosen as kudq k = 2.

International Journal of Control 15

where the function (ud , uq , e1 , e2 ) can be rewritten as

The solution of Eq. (49) is presented

16 International Journal of Control

Remark 4 : For practical implementation of the proposed controller, it is important to select

where in general is a small positive value.

Figure 5. Measurement noise affecting the system output