Sie sind auf Seite 1von 7

2004 35th Annual IEEE Power Electronics Specialists Conference Aachen, Germany, 2004

A Pulse-Width-Modulation Based Sliding Mode Controller for Buck Converters

Siew-Chong Tan, Y. M. Lai, Martin K. H. Cheung, and Chi K. Tse


Department of Electronic & Information Engineering
The Hong Kong Polytechnic University
Hong Kong, China
email: ensctan@eie.polyu.edu.hk

Abstract— This paper presents the design and analysis Hence, in this paper, we present the design of a PWM based
of a pulse-width-modulation based sliding mode voltage sliding mode voltage controlled (SMVC) buck converter, with
controller for buck converters from a circuit design emphasis on its practical and implementation details, from a
perspective. A practical design approach which aims at circuit design perspective, using the theoretical groundwork
systematizing the procedure for the selection of the control established in [7] and [8]. In contrast to [7], the design of this
parameters is introduced. Additionally, a simple analog controller at circuit level involves a different set of engineering
form of the controller for practical realization is provided. difficulty and consideration. Additionally, we introduce a prac-
It is found that this controller adopts a structure similar tical approach to the design and selection of the sliding coeffi-
to the conventional pulse-width-modulated voltage mode cients of the controller. This approach systematizes the design
controller. Simulation and experimental results show that procedure. It should be noted that it can also be employed
the response of the converter agrees with the theoretical for the design of other PWM based SM power converters.
design. Finally, an analog form of the controller that is suitable for
practical realization is provided. This controller can be easily
I. I NTRODUCTION implemented from the derived mathematical expressions with
only a few operational amplifiers and analog ICs. Simulations
Sliding mode (SM) controllers are well known for their and experiments are performed on the proposed converter to
robustness and stability. Most of the previously proposed validate the theoretical design.
SM controllers for switching power converters are hysteresis-
modulation (HM) (or delta-modulation) based [1]–[6]. Natu- II. THEORETICAL DERIVATION
rally, they inherit the typical disadvantages of having variable- This section covers the theoretical aspects of the SMVC
switching-frequency operation and being highly control sensi- converter. A practical method of designing the sliding coeffi-
tive to noise. Possible solutions are to incorporate constant cients is also introduced.
timer circuits into the hysteretic SM controller to ensure A. Mathematical Model of an Ideal Sliding Mode PID Voltage
constant switching frequency operation [5], or to use adaptive Controlled Buck Converter
hysteresis band that varies with parameter changes to control
and fixate the switching frequency [6]. However, these solu- SW L L R
tions require additional components and are unattractive for
low cost voltage conversion applications. C R1
An alternative solution to this is to change the modulation +
Vi D C Vo RL
method of the SM controllers from HM to pulse-width-
modulation (PWM). To the authors’ knowledge, this concept
R2
was first published in [7]. The idea is based on the assumption
that at a high switching frequency, the control action of a /C Vo
sliding mode controller is equivalent to the duty cycle control Sliding 2
Mode
action of a PWM controller. Hence, the migration of a sliding Controller
1
Vref
mode controller from being HM based to PWM based is made
possible. This proof was rigourously shown in a companion Fig. 1. Basic structure of an SMVC buck converter system.
paper [8] by the same authors. However, the work presented
is theoretical, and falls short of a practical consideration on Fig. 1 shows an SMVC buck converter. Here, the voltage
its implementation. error x1 , the voltage error dynamics (or the rate of change of
In another paper [9], some possible practical methods of voltage error) x2 , and the integral of voltage error x3 , can be
implementing a SM controller on buck type converters are expressed as
described. Termed as the indirect implementations of sliding
mode control, it is basically a straightforward circuit represen- x1 = Vref − βVo
  
tation of the idea provided in [7]. Although the discussion in β Vo uVi − Vo
x2 = x˙1 = − dt
[9] gives a clear direction as to how such controllers may be C RL L

developed, it provides only an overview of the PWM based
SM controller. x3 = x1 dt (1)

0-7803-8399-0/04/$20.00 ©2004 IEEE. 3647


2004 35th Annual IEEE Power Electronics Specialists Conference Aachen, Germany, 2004

where C, L, and RL are the capacitance, inductance, and load where iC is the capacitor current.
resistance respectively; Vref , Vi , and βVo are the reference, The above inequalities give the conditions for stability and
input, and sensed output voltage respectively; and u = 1 or provide a range of employable sliding coefficients. However,
0 is the switching state of power switch SW . Then, the state other than stability, no information relating the sliding coeffi-
space model of the system can be derived as cients to the converter performance is provided.
ẋ = Ax + Bu + D (2) C. Stability Condition With Design Parameters Consideration
where To alleviate the above problem, we propose to first tighten
⎡ ⎤ ⎡ ⎤ the design constraints by absorbing the actual operating pa-
0 1 0 0
rameters into the inequality. This is done by decomposing
A = ⎣− LC 1
− RL1C 0⎦ , B = ⎣− βVi⎦
,
LC (9) into two sections of inequalities and considering them as
1 0 0 0
⎡ ⎤ ⎡ ⎤ individual cases with respect to the polarity of the capacitor
0 x1 current flow. Since in practice α1 1
α2 > RL C , the left inequality
ref ⎦
D = ⎣ VLC , and x = ⎣x2 ⎦ . of (9) is implied by
0 x3  
α3 α1 1 ˆ
The basic idea of SM control is to design a certain sliding 0 < LC (Vref − βVo ) − βL − iC + βVo ,
α2 α2 RL C
surface in its control law that will direct the trajectory of the (10)
state variables towards a desired origin when coincided. For which can be rearranged to give
our system’s model, it is appropriate to have a control law that βVo + LC α
α2 (Vref − βVo ) 1
3
α1
adopts a switching function such as < + , (11)

1
α2 βL iˆ C
RL C
u = (1 + sign(S)) (3)
2 and the right inequality of (9) is implied by
where S is the instantaneous state variable’s trajectory, and is  
α3 α1 1 ˆ
described as LC (Vref − βVo ) + βL − iC + βVo < βVi
α2 α2 RL C
S = α1 x1 + α2 x2 + α3 x3 = JT x, (4) (12)
which can be rearranged to give
with J = [α1 α2 α3 ] and α1 , α2 , and α3 representing the
T

control parameters termed as sliding coefficients. α1 βVi βVo + LC α


α2 (Vref − βVo )
3
1
< − + , (13)
α2 ˆ ˆ R C
B. Stability Analysis in Circuit Terms βL iC βL iC L

As in all other SM control schemes, the determination of where iˆC is the peak magnitude of the bidirectional capacitor
the ranges of employable sliding coefficients for the SMVC current flow. Next, equations (11) and (13) can be recombined
converter must go through the process of analyzing the sta- and further tightened by considering the range of input and
bility behavior of the controller/converter system using the loading conditions of the converter to give
Lyapunov’s Direct Method [11]. This is performed by firstly ⎧  
combining (2), (3), and the time derivative of (4) to give ⎪
⎪ α1 Vo + LC α 3
β − Vo
Vref

⎪ α2 1

⎪ < +
1 1 ⎪
Ṡ = JT Ax + JT B + JT B sign(S) + JT D. (5) ⎪


α2 L iˆC RL(max) C
2 2 ⎪
⎪   

⎪ ≥ 2 + α3 Vref

Multiplying (5) by (4) gives the Lyapunov function candidate ⎪
⎨ for V i(min) V o LC α2 β V o
 (14)
1 T 1 T ⎪  
S Ṡ = S J Ax + J B + J B sign(S) + J D
T T ⎪

2 2 ⎪
⎪ Vo + LC α2 β − Vo
α3 V ref
1
  ⎪
⎪ α1 Vi(min)

1 1 ⎪
⎪ α2 < − +
= S JT Ax + JT B + JT D + |S|JT B. ⎪
⎪ L iˆC L iˆC R L(max) C
2 2
(6) ⎪
⎪   


⎩ for V < 2 V + LC α3 Vref − V
i(min) o o
To achieve global reachability and asymptotic stability, the α2 β
Lyapunov function is evaluated as S Ṡ < 0, which can be where RL(max) is the maximum load resistance and Vi(min) is
written as the minimum input voltage, which the converter is designed


ṠS>0 = JT Ax + JT B + JT D < 0 for. Additionally, the peak capacitor current iˆC is the maxi-
(7)
ṠS<0 = JT Ax + JT D > 0, mum inductor current ripple during steady state operation.
Theoretically, one may assume that at steady state operation,
i.e., the actual output voltage Vo is ideally a pure DC waveform
JT Ax + JT B + JT D < 0 < JT Ax + JT D (8) whose magnitude is equal to the desired output voltage Vod ≡
Vref
β . However, this is not true in practice. Due to the limitation
or rearranged in scalar representation, of finite switching frequency, there will always be some steady
 
α3 α1 1 state DC error between Vo and Vod , even with the error-
0 < LC (Vref − βVo )−βL − iC +βVo < βVi . reducing integral controllers (i.e. PI, PID). It is important
α2 α2 RL C
(9) to take this error into consideration for the design of the

3648
2004 35th Annual IEEE Power Electronics Specialists Conference Aachen, Germany, 2004

 
controller since the factor LC α 3
α2
Vref
β − V o is relatively By rearranging (19) and substituting ζ = 1 into the damping
large in comparison to Vo . ratio, the following design equations are obtained:
Now, considering that α1 α3
= 4πfBW and = 4π 2 fBW 2 . (20)
(a) in controllers with integral control function, the differ- α2 α2
ence between Vo and Vod is small, and when optimally
Thus, the design of the sliding coefficients is now dependent
designed, is normally limited to a range of within ±5 %
on the bandwidth of the desired frequency response in con-
of Vod ;
junction with the stability condition (15). Additionally, it is
(b) for PWM based SMVC converters, the DC average of Vo
worth mentioning that the design equations in (20) for the
is always lower than Vod ; and
SMVC controller are applicable to all other types of second
(C) the term LC α 3
α2 is always positive, order converters.
we can rewrite the stability condition (14) for PWM based
SMVC converter as III. I MPLEMENTATION OF PWM BASED SMVC

⎪ α1 < 0.95 Vod
⎪ +
1 C ONTROLLER


⎪ α2
⎪ L iˆC RL(max) C This section details the implementation of PWM based

⎪  

⎪ SMVC buck converter.

⎪ V ≥ 1.95 + 0.05 LC α3


for i(min) α2 Vod

A. Derivation of PWM Control Law
  (15)



⎪ α1 Vi(min) 1 + 0.05 LC α3
V od 1 It was previously derived in [8] that the control method [10]

⎪ < − 2
α
+

⎪ in sliding mode control is effectively a duty cycle control.

⎪ α 2 L iˆC L iˆC R L(max) C

⎪   From [10], the equivalent control input ueq can be formulated

⎪ by setting the time differentiation of (4) as Ṡ = 0, i.e
⎩ for Vi(min) < 1.95 + 0.05 LC α 3
α2 Vod

by substituting Vo = 0.95 Vod or Vo = Vod into the appropriate JT Ax + JT Bueq + JT D = 0. (21)


parts. Thus, the control parameters α1 , α2 , and α3 are now Now, solving for u yields
bounded by inequalities that have more stringent stability
 −1 T
constraints than in (9). ueq = − JT B J [Ax + D]
 
D. Selection of Sliding Coefficients LC α1 1 Vo α3 LC
= − x2 + + x1 (22)
Clearly, inequality (15) provides only the general stabil- βVi α2 RL C Vi α2 βVi
ity information, but give no detail on the selection of the where ueq is continuous and 0 < ueq < 1. Substituting (22)
parameters. The equation relating sliding coefficients to the into the inequality and multiplying by βVi gives
characteristic response of the converter during sliding mode
operation can be easily found by substituting S = 0 into (4), 0 < ueq ∗ < βVi (23)
i.e.,
 where
dx1  
α1 x1 + α2 + α3 x1 dt = 0. (16) ∗ 1 α1 α3
dt ueq = βL − iC + βVo +
LC (Vref − βVo )
RL C α2 α2
Rearranging the time differentiation of (16) into a standard (24)
second-order system form, we have which will provide the ideal average sliding motion on the
d2 x1 dx1 manifold S = 0.
2
+ 2ζωn + ωn 2 x1 = 0 (17) In terms of PWM based controlled system, the duty cycle
dt dt
 D is expressed as
where ωn = α3
α2 is the undamped natural frequency and
Vc
ζ = 2 α2 α3 is the damping ratio. Recall that there are three
√α1
D= (25)
possible types of response in a linear second-order system: V̂ramp
under-damped (0 ≤ ζ < 1), critically-damped (ζ = 1), and where Vc is the control signal to the pulse-width modulator
over-damped (ζ > 1). For ease of discussion, we choose to and V̂ramp is the peak magnitude of the constant ramp signal.
design the controller for critically-damped response1 , i.e., Since D is also continuous and bounded by 0 < D < 1, it
x1 (t) = (A1 + A2 t)e−ωn t , for t ≥ 0 (18) may also be written in the form

where A1 and A2 are determined by the initial conditions of 0 < Vc < V̂ramp . (26)
the system. In a critically-damped system, the bandwidth of
Comparing the equivalent control and the duty ratio control
the controller’s response fBW is
 [8], the following relationships can be established
ωn 1 α3
fBW = = . (19) Vc = ueq ∗ and V̂ramp = βVi (27)
2π 2π α2
1 The design for an under-damped controller can be performed using a for the practical implementation of PWM based SMVC con-
similar procedure as discussed hereafter. troller.

3649
2004 35th Annual IEEE Power Electronics Specialists Conference Aachen, Germany, 2004

Sw
L L continuous conduction mode for Vi = 18 V to 30 V and
iR = 0.5 A to 4 A. The calculated critical inductance is Lcrit =
C R R1 36 µH. The minimum required capacitance is Cmin = 9 µF.
Vi D C RL Vo The maximum allowable peak-to-peak ripple voltage is 50 mV.
H
C(mea)
R2 To study the compliance of the design equations with the
performance and its relationship with the transient response,
the controller is designed for two different bandwidths: at one
RI1
RI2 twentieth and at one tenth of the switching frequency fS ,
Isolated
Drive RSUM i.e., fBW = 10 kHz (i.e. first order response time constant
RINV
RINV RSUM RSUM
UI τ10 kHz = 15.915 µs) and fBW = 20 kHz (i.e. τ20 kHz =
Vo 7.956 µs). The controller is designed for maximum load
INV RV2
PWM
Vramp
current (i.e. minimum load resistance RL(min) )
SUM RV1
RSUM
UV
VREF A. Steady State Performance
RV1
RV2 Figs. 3 and 4 show the simulated (left) and experimental
(right) waveforms during steady state operation, for the SMVC
Fig. 2. Schematic diagram of the proposed PWM based SMVC buck converter with the 20 kHz bandwidth controller operating at
converter. full load (i.e. RL = 3 Ω). It can be seen that except for
some ringing noise in the experimentally captured Vc and V o
waveforms, the simulated and experimental waveforms are in
B. Implementation of Controller good agreement. The main difference is that for the simulation,
Fig. 2 shows the schematic diagram of the proposed PWM output voltage ripple V o ≈ ±4 mV (i.e. < 0.035 % of
based SMVC buck converter. The controller design is based Vod ), and for the experiment, V o ≈ ±8 mV (i.e. < 0.07 %
on (24) and (27). It basically adopts the structure of a voltage of Vod ). This discrepancy is mainly due to the presence of
mode PWM controlled converter. Interestingly, this controller parasitic resistance and equivalent series inductance (ESL) of
also inherits the input feed-forward voltage control scheme the capacitor in the practical converter, which are not modeled
of conventional PWM voltage mode control in its operation in the simulation program.
since the input ramp signal Vramp for modulation has peak Fig. 5 shows the corresponding set of experimental wave-
magnitude that is proportional to the input voltage Vi (refer to forms for the SMVC converter with the 10 kHz bandwidth
(27)). controller operating at full load (i.e. RL = 3 Ω). Except for
Briefly, the design of this controller can be summarized as Vc , there is no major difference between these waveforms
follows: selection of the desired frequency response’s band- and the experimental waveforms in Figs. 3 and 4. Due to the
width, calculation of the corresponding sliding coefficients higher magnitude of the sliding coefficients, Vc of the 20 kHz
using (20); inspection of the sliding coefficients’s appropri- bandwidth controller has a higher peak-to-peak value than Vc
ateness using stability condition (15); and formulation of the of the 10 kHz bandwidth controller.
control equations by substituting the calculated parameters into
(24) and (27). B. Steady State DC Error at Different Loading Conditions
Fig. 6 shows a plot of the measured DC output voltage
IV. S IMULATION AND E XPERIMENTAL R ESULTS against the different operating load resistances. At full load
operation (i.e. RL = 3 Ω), the converter employing the 20 kHz
TABLE I
bandwidth controller has a steady state DC output voltage of
S PECIFICATION OF B UCK C ONVERTER
11.661 V, which corresponds to a -2.825 % deviation from
Description Parameter Nominal Value Vod . The plot also shows that even though Vo increases with
Input voltage Vi 24 V RL , Vo is always less than Vod . This agrees with the previous
Capacitance C 150 µF assumption that the output voltage of PWM based system is
Capacitor ESR rC 21 mΩ
Inductance L 100 µH
always below the desired voltage. Furthermore, it also shows
Inductor resistance rL 0.12 Ω that the converter has satisfactory load regulation, having only
Switching frequency fS 200 kHz a 0.151 V deviation in Vo for the entire load range of 3 Ω ≤
Minimum load resistance RL(min) 3Ω
Maximum load resistance RL(max) 24 Ω
RL ≤ 24 Ω, i.e. the load regulation dR dVo
L
is only 0.72 % from
Desired output voltage Vod 12 V full load to minimum load.
For the converter employing the 10 kHz bandwidth con-
troller, the steady state DC output voltage at full load operation
The proposed design approach and analog controller for
is 11.633 V, which corresponds to a -3.058 % deviation from
the PWM based SMVC buck converter are verified through
Vod . For the entire load range of 3 Ω ≤ RL ≤ 24 Ω, Vo has
simulation2 and experiment. The specification of the converter dVo
a deviation of 0.189 V, i.e. the load regulation dR is 0.90 %
is given in Table 1. The converter is designed to operate in L
from full load to minimum load. Thus, it can be concluded
2 The simulation is performed using Matlab/Simulink. The step size taken that the 20 kHz bandwidth controller has better load variation
for all simulations is 10 ns. property than the 10 kHz bandwidth controller.

3650
2004 35th Annual IEEE Power Electronics Specialists Conference Aachen, Germany, 2004

Gate Pulse
u (5 V /DIV)
es )IV
lu D
P /V
tea (5
G u
u

Control Signal
VC (2 V/DIV)
la
ng )V
iS ID Vramp
lo /V Vramp VC
rt 2(
no V C

V ra mp (2 V/DIV)
C VC

Ramp Input
tu )VI
pn D/
I V
p 2(
m
aR pamr
V

Fig. 3. Simulated (left) and experimental (right) waveforms of control signal Vc , input ramp Vramp , and generated gate pulse u for SMVC converter with
the 20 kHz bandwidth controller operating at constant load resistance RL = 3 Ω.

)
eg VI/D
i L (1 A/DIV) Ripple (20 mV /DIV)
Output Voltage

alt V ~
Vo ~
oV m Vo
tu 0(1
pt le
up
O pi iL
R
Inductor
Current

u
ro t )IV
tc ne D/ L
ud rru A
nI C 1(L
Gate Pulse
u (5V /DIV)

es )V
lu I u
P /DV
et 5
a (
G u

o for
Fig. 4. Simulated (left) and experimental (right) waveforms of gate pulse u, and the corresponding inductor current iL and output voltage ripple V
SMVC converter with the 20 kHz bandwidth controller operating at constant load resistance RL = 3 Ω.
u (5 V /DIV)
Gate Pulse

i L (1 A/DIV) Ripple (20 mV /DIV)


Output Voltage

~
Vo
u

iL
Control Signal
V C (2 V/DIV)

Inductor
Current

Vramp u

VC
Gate Pulse
u (5V /DIV)
Vr amp (2 V/DIV)
Ramp Input

Fig. 5. Experimental waveforms of control signal Vc , input ramp Vramp , and generated gate pulse u (left) and waveforms of gate pulse u, and the
o (right), for the SMVC converter with the 10 kHz bandwidth controller operating at load
corresponding inductor current iL and output voltage ripple V
resistance RL = 3 Ω.

3651
2004 35th Annual IEEE Power Electronics Specialists Conference Aachen, Germany, 2004

11.84

11.82

11.8

Measured Output Voltage (V)


11.78

11.76

11.74

11.72

11.7

11.68

11.66

11.64 frequency bandwidth = 10 kHz


frequency bandwidth = 20 kHz
11.62

11.6
0 2 4 6 8 10 12 14 16 18 20 22 24
Load Resistance (Ohms)

Fig. 6. Plot of measured DC output voltage Vo against load resistance RL for SMVC buck converter with both the 10 kHz and 20 kHz bandwidth controllers.

) )
eg IV estimated eg VI
alt D/ D
response tal /V 232 mV
oV V.2 220 mV oV .2
tu 0( tu 0(
pt el 120 s pt el 83 s
u pp u pp
O iR O iR

tn 1.0 A tn 1.72 A
er )V er )V
ru ID ru ID
Cr /A Cr /A
toc 2( toc 2(
ud L ud L
nI nI

Fig. 7. Simulated waveforms of output voltage ripple V o and inductor current iL of the SMVC converter with the 10 kHz bandwidth controller (left) and
the 20 kHz bandwidth controller (right), operating at 5 kHz step load change between RL = 3 Ω and RL = 12 Ω.

)V )V
eg I/D eg I/D
tal V 200 mV tal V 250 mV
oV m oV m
tu 002 tu 002
pt (e 104 s pt (e 73 s
u lp u lp
O ip O ip
R R

tn 1.36 A tn 2.33 A
er )V er )V
ru ID ru ID
Cr /A Cr /A
ot 2( ot (2
cu L
cu L
dn dn
I I

Fig. 8. Experimental waveforms of output voltage ripple V o and inductor current iL of the SMVC converter with the 10 kHz bandwidth controller (left)
and the 20 kHz bandwidth controller (right), operating at 5 kHz step load change between RL = 3 Ω and RL = 12 Ω.

3652
2004 35th Annual IEEE Power Electronics Specialists Conference Aachen, Germany, 2004

C. Dynamic Performance An analog form of the controller is also presented. It is found


The dynamic performance of the controllers is studied using that the PWM based SM controller adopts a similar structure
a load resistance that alternates between quarter load (12 Ω) to that of a conventional PWM voltage mode controller. The
and full load (3 Ω) at a constant frequency of 5 kHz. Figs. 7 simulation and experimental results show that the response of
and 8 show respectively the simulated and experimental output the converter agrees with the theoretical design.
voltage ripple (top) and inductor current (bottom) waveforms ACKNOWLEDGMENT
of the converter for both the 10 kHz bandwidth controller (left)
The work described in this paper was supported by a grant
and the 20 kHz bandwidth controller (right). As illustrated in
provided by The Hong Kong Polytechnic University (Project
Fig. 7, the simulated output voltage has an overshoot ripple
No. G-T379).
of 220 mV (1.83% of Vod ) and a steady state settling time of
120 µs for the 10 kHz bandwidth controller, and an overshoot R EFERENCES
ripple of 232 mV (1.93% of Vod ) and a steady state settling [1] R. Venkataramanan, A. Sabanoivc, and S. Ćuk, “Sliding mode control of
time of 83 µs for the 20 kHz bandwidth controller, during the DC-to-DC converters,” in Proceedings, IEEE Conference on Industrial
load changes. As shown in Fig. 8, the output voltage has an Electronics, Control and Instrumentations (IECON), pp. 251–258, 1985.
[2] M. Castilla, L.C. de Vicuna, M. Lopez, O. Lopez, and J. Matas, “On
overshoot ripple of 200 mV (1.67% of Vod ) and a steady state the design of sliding mode control schemes for quantum resonant
settling time of 104 µs for the 10 kHz bandwidth controller, converters,” IEEE Transactions on Power Electronics, vol. 15 no. 15,
and an overshoot ripple of 250 mV (2.08% of Vod ) and a pp. 960–973, Nov. 2000.
[3] L. Malesani, L. Rossetto, G. Spiazzi, and P. Tenti, “Performance opti-
steady state settling time of 73 µs for the 20 kHz bandwidth mization of Cuk converters by sliding-mode control,” IEEE Transactions
controller, during the load changes. on Power Electronics, vol. 10 no. 3, pp. 302–309, May 1995.
Furthermore, consistent with a critically-damped response, [4] P. Mattavelli, L. Rossetto, and G. Spiazzi, “Small-signal analysis of DC-
there is no ringing or oscillations in transience. However, DC converters with sliding mode control,” IEEE Transactions on Power
Electronics, vol. 12 no. 1, pp. 96–102, Jan. 1997.
it should also be mentioned that there are some slight dis- [5] B.J. Cardoso, A.F. Moreira, B.R. Menezes, and P.C. Cortizo, “Analysis
agreements between the experimental and simulation results of switching frequency reduction methods applied to sliding mode
in terms of the overshoot ripple magnitudes and the settling controlled DC-DC converters,” in Proceedings, IEEE Applied Power
Electronics Conference and Exposition (APEC), pp. 403–410, Feb 1992.
times. These are mainly due to the modeling imperfection of [6] V.M. Nguyen and C.Q. Lee, “Tracking control of buck converter using
the simulation program, and the parameters’ deviation of the sliding-mode with adaptive hysteresis,” in IEEE Power Electronics
actual experimental circuits from the simulation program due Specialists Conference Record (PESC), vol. 2, pp. 1086–1093, June
1995.
to the variation of the actual components used in the setup. [7] H. Sira-Ramirez and M. Ilic, “A geometric approach to the feedback
control of switch mode DC-to-DC power supplies,” IEEE Transactions
V. C ONCLUSION on Circuits and Systems, vol. 35 no. 10, pp. 1291–1298, Oct. 1988.
[8] H. Sira-Ramirez, “A geometric approach to pulse-width modulated
A PWM based SMVC buck converter is presented from control in nolinear dynamical systems,” IEEE Transactions on Automatic
a circuit design perspective. The description of the design Control, vol. 34 no. 3, pp. 184–187, Feb. 1989.
[9] V.M. Nguyen and C.Q. Lee, “Indirect implementations of sliding-mode
methodology takes into account the different aspects of con- control law in buck-type converters,” in Proceedings, IEEE Applied
verter’s operating conditions. A practical approach to the Power Electronics Conference and Exposition (APEC), vol. 1, pp. 111–
design of the sliding coefficients is also proposed in this paper. 115, March 1996.
[10] V. Utkin, J. Guldner, and J.X. Shi , Sliding Mode Control in Electrome-
This approach uses an equation that is derived from analyzing chanical Systems. London, U.K.: Taylor and Francis, 1999.
the dynamic behavior of the converter during sliding mode [11] J.J.E. Slotine and W. Li, “Sliding Control”, Ch. 7, Applied Nonlinear
operation, in addition to the stability conditions of the system. Control. Englewood Cliffs, N.J.: Prentice Hall, 1991.

3653

Das könnte Ihnen auch gefallen