See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/339268520

Feedback Linearization Control of Boost Converter

Conference Paper · July 2019

DOI: 10.1109/ICICICT46008.2019.8993341

CITATIONS READS

5 1,053

2 authors, including:

Punnya Priya
Sri Sairam Engineering college
4 PUBLICATIONS 7 CITATIONS

SEE PROFILE

All content following this page was uploaded by Punnya Priya on 18 February 2020.

The user has requested enhancement of the downloaded file.

 Feedback Linearization Control of Boost Converter
F. Punnya Priya Dr. K. Latha
Department of EEE Department of EEE
Anna University Anna University
Chennai, India Chennai, India
[email protected] [email protected]

Abstract—Boost converters are integrated with sources to input and output is one. And the other one is the internal
supply load at elevated voltages. Since Boost converter is a dynamics of the system which is hidden. In this paper, the
nonlinear system, it works good only at the operating point input-output feedback linearization techniques [18]–[21] are
when linear controllers are used. Whereas when nonlinear used for linearizing the DC-DC converter. This makes the
controllers are used, the operation becomes independent of its
system parameters independent of the operating point. The
operating point. In this paper, simplest cost effective feedback
linearization controller has been proposed for better dynamic proposed controller results are compared with PI controller
and steady state performance and widespread range of output results and the robustness of input-output feedback
voltage regulation. A 600W Boost converter is designed and linearization controller is justified.
simulation results are compared between PI controller and the In PI control algorithm used, both the inner current loop
feedback linearization controller. Simulation results shows the and outer voltage loop is controlled using Kp and Ki values
efficacy and robustness of the proposed controller. obtained by Ziegler-Nichols method. The designed boost
converter produced steady state error of one volt in open
Keywords— feedback linearization (FBKLN) controller, loop performance. Whereas in closed loop performance for
boost converter, state space averaging
both the designed controllers, steady state error was
I. INTRODUCTION removed.
The simulated results shows that the proposed method is the
DC/DC converters have found its use in several
versatile one when compared with the standard PI
industrial applications, aerospace and consumer electronic
controller.
devices. Hence dc/dc converters should be capable of
This paper is structured as follows. First, section II gives the
always providing the suitable output voltage required by
details of the mathematical model of Boost converter. Then,
these devices, in spite of changes in parameters and loads.
the feedback linearization technique approach is specified in
[1]– [8].
section III. The implementation of linearization techniques
Linearized small signal models are used for analyzing
in outer voltage loop and inner current loop and the
the stability of switching converters. But the relation
selection of feedback linearization coefficients are presented
between the converter dynamics and the initial conditions of
in section IV. The simulation results are compared and
the states and the post instability dynamics cannot be
justified in section V. Finally, the conclusions are given in
determined. In power electronic converters, the state space
section VI.
averaging method is the best modelling tool for designing
the stabilizing controllers.
Many advanced control algorithms such as adaptive II. TOPOLOGY OF BOOST CONVERTER
controllers [9]-[10], sliding mode controllers [11], predictive
controllers [12], deadbeat controllers [13], model predictive A. Mathematical Model of Boost Converter
controllers [14]–[17], and robust controllers [1] can be used
to improve the closed loop performance.
But these advanced controllers need information about
the converter parameters, to afford the desired closed-loop
performance and stability. To obtain a linearized model of a
nonlinear system, input state feedback linearization may be
used. But it requires availability of all states of the system,
which makes it not suitable for practical applications. A best
alternative for input state feedback linearization is input-
output linearization. It converts the system into a cascade Fig. 1. Circuit Diagram of Boost Converter
combination of two dynamics systems. The dynamics of
reduced order system obtained from the relation between
 a

Hi vin
y.LI
itli as YQ.si

1 we
solve for duty cycle
in equation
duty cycle
vanish
2 we make
in equation
set
solve
 Fig. 1 shows a DC–DC boost converter. It consists of a
DC input voltage source (Vi), a controlled switch (s), a 𝑧 𝐿 ℎ(𝑥)
diode (D), a filter inductor (L), a filter capacitor (C), and a 𝑧 ⎢𝐿 ℎ(𝑥)⎥
load resistor (R). ⎢. ⎥ ⎢ . ⎥
The equations describing the operation of the converter with 𝑧 = ⎢ . ⎥=⎢ ⎥ (6)
.
can be written for the switching conditions ON and OFF, ⎢. ⎥ ⎢ .
⎥
respectively as ⎢𝑧 ⎥ ⎢𝐿 ⎥
ℎ(𝑥)⎦
When switch‘s’ is ‘ON’: ⎣

𝑉𝑑 = 𝐿 (1) For a give scalar function ℎ(𝑥) of 𝑥 = [𝑥1 ⋯ 𝑥𝑛]T and the
vector field 𝑓 =[𝑓1 ⋯ 𝑓𝑛]𝑇 , the Lie derivative of ℎ(𝑥) along
When switch‘s’ is ‘OFF’: the vector field 𝑓(𝑥) can be written as follows:
𝑉 (1 − 𝑑) = 𝐿 + 𝑉 (1 − 𝑑) (2) ( ) ( )
𝐿 ℎ(𝑥) = 𝑓(𝑥) = ∑ 𝑓 (𝑥) (7)
For one full cycle, combining both the above equations; If the Lie derivative of the function 𝐿 ℎ(𝑥)along (𝑥) is not
𝐿 = 𝑉 − 𝑉 (1 − 𝑑) (3) equal to zero, i.e.,

Taking currents at output node; ( )

𝐿 𝐿 ℎ(𝑥) = 𝑔(𝑥) ≠ 0, (8)

1
𝑉

00
𝐶 = 𝐼 (1 − 𝑑) − (4) the system is said to have a relative degree 𝒓.
𝑅
The relative degree (𝒓) defines the feedback linearizability
𝑑=
0, 𝑤ℎ𝑒𝑛 𝑠 𝑖𝑠 𝑠𝑤𝑖𝑡𝑐ℎ𝑒𝑑 𝑜𝑓𝑓
1, 𝑤ℎ𝑒𝑛 𝑠 𝑖𝑠 𝑠𝑤𝑖𝑡𝑐ℎ𝑒𝑑 𝑜𝑛
Ed of a transformed linear system.
1. If 𝒓 = 𝒏, the feedback linearized system is called exactly
linearized
Thus the mathematical model given in (3) and (4) are 2. If 𝒓 < 𝒏, the feedback linearized system is called partially
used for simulation in Matlab/Simulink. linearized
III. FEEDBACK LINEARIZATION TECHNIQUE The feedback linearized system is obtained as
𝑧̇ 𝑧 𝐿 ℎ(𝑥)
Feedback Linearization technique is;
𝑧̇ 𝑧 ⎢ 𝐿 ℎ(𝑥) ⎥
 A nonlinear control technique which transforms a ⎢ ⎥ ⎢ ⎥
nonlinear system into a linear system or several 𝑧̇ = ⎢ .. ⎥ = ⎢ .. ⎥ = ⎢ .. ⎥
⎢ ⎥ ⎢ ⎥
linear subsystems using nonlinear coordinate ⎢ ⎥. . ⎢ . ⎥
transformation. ⎣𝑧 ̇ ⎦ ⎣𝑣⎦
⎣𝐿 ℎ(𝑥) + 𝐿 𝐿 ℎ(𝑥)𝑢⎦
 Linearization is independent of the operating point 𝐿 ℎ(𝑥)
 Order of the transformed system depends on the
⎢ 𝐿 ℎ(𝑥) ⎥
relative degree of the system
= ⎢ .. ⎥ = 𝐴𝑧 + 𝐵𝑣 (9)
 Any linear control technique can be applied for the ⎢ ⎥
transformed linear system ⎢ . ⎥
⎣𝑎(𝑥) + 𝑏(𝑥)𝑢⎦
Generalized form of a nonlinear system:
𝑥̇ = 𝑓(𝑥) + 𝑔(𝑥)𝑢 where 𝑎(𝑥) = 𝐿 ℎ(𝑥), 𝑏(𝑥) = 𝐿 𝐿 ℎ(𝑥) with
(5)
𝑦 = ℎ(𝑥) 0 1 ⋯ 0 0
where 0 0 ⋯ 0 0
𝐴= ,𝐵= and
 𝑥 is the state of the system which represents ⋮ ⋮ ⋮ ⋮ ⋮
different physical properties of the 0 0 0 0 1
system (e.g., voltages and currents) and if the order of the
system is 𝑛, there will be 𝑛 number of states, 𝑣 = 𝑎(𝑥) + 𝑏(𝑥)𝑢 = 𝐿 ℎ(𝑥) + 𝐿 𝐿 ℎ(𝑥)𝑢
 𝑓 and 𝑔 are the nonlinear functions with
𝑔 ≠ 0 which represent both physical properties and Any linear technique can be applied to design linear
parameters of the system, controller for the feedback linearized system.
 𝑢 is the control signal, Feedback linearized system can be written as follows:
 𝑦 is the output function.
If the original nonlinear system is transformed from 𝑥 states 𝑧̇ = 𝐴𝑧 + 𝐵𝑣
to 𝑧 states, the nonlinear coordinate transformation can be (10)
defined as follows: 𝑦=𝐶
𝑧 = 𝜑(𝑥)
 lo m
dkipf.fi did C
For this system, 𝑣 can be obtained by using any linear
IL
Based on the boost converter model in Eq. (4) and written in
control technique. For a state feedback control, 𝑣 can be the generalised form as in Eq. (5)
written as 𝑣 = −𝐾𝑧.
If the system is exactly linearized, 𝑟 = 𝑛, the control law (𝑢) 𝑉̇ = (𝑖 − 𝑖 − 𝑑𝑖 ) assume (18)
can be written as follows Substituting ‘d’ from Eq. (16) in Eq.(18) it becomes
𝑉̇ = 𝑖 − 𝑖

X
( ) ( )
𝑢= = (11)
( ) ( ) Usually its designed in such way that, the outer voltage loop
If the system is partially linearized, 𝑟 < 𝑛, the control responds slower than the inner current loop. Hence the value
law (𝑢) can be written as follows of iL is assumed as iL ref which gives the following voltage
model.
( ) 1 𝑉

11
( )
𝑢= = (12) 𝑉̇ = 𝑖 − 𝑖

of
( ) ( ) 𝐶 𝑉
(19)
𝑦 = 𝑉
IV. IMPLEMENTING FEEDBACK LINEARIZATION
TECHNIQUE TO BOOST CONVERTER
A. Linearization of inner current loop µ As given in section III, the relative degree of the system in
Eq.(19) is calculated first.
𝐿 ℎ(𝑥) =
( )
𝑔(𝑥) = ≠0 (20)
The dual loop control strategy can be employed to this
DC-DC converter. This dual loop consists of inner current
loop and outer voltage loop. The inner current loop is ( )
𝐿 ℎ(𝑥) = 𝑓(𝑥) = ≠0 (21)
considered first to implement the feedback linearization.
The inner current loop model is derived based on the boost According to the Eq. (8), ri = 1
According to the Eq. (12), the control input for the voltage
converter model in Eq. (3) and written in the generalised
form as in Eq. (5) _d loop can be found as

𝚤̇ =
ie[𝑉 − 𝑉 ] +VI[𝑉 ] 𝑑 𝑑= (22)

(13)
free Substituting this‘d’ in Eq. (19) where iL ref is
𝑦 = 𝑖
considered as control input ‘d’, the equation becomes
As given in section III, the relative degree of the system in 𝑉̇ =
Eq.(13) is calculated first.
+ +𝑚 =𝑚
I vitrack
𝑦 = 𝑉 (23)
( )
𝐿 ℎ(𝑥) = 𝑔(𝑥) = 𝑉 ≠0 (14) Taking the controllable variable ‘m’ as voltage error,
m=𝐾 𝑉 − 𝑉 the voltage loop is linearized. Thus
𝐿 ℎ(𝑥) =
( )
𝑓(𝑥) = ≠0 (15) the both control loops are linearized.
The linearized Boost converter system’s control block
diagram is shown in Fig. 2, where the L and C values are
According to the Eq. (8), ri = 1
included along with the feedback coefficients shown as Gv
According to the Eq. (12), the control input for current loop
and Gi and the resultant new system is a typical linear
i
can be found as
system model with no disturbance. The following section
𝑑=
Substituting this ‘d’ in Eq. (13),
iw (16)
o
gives the simulation comparison results to testify the control
performance of the linearized system.

𝚤̇ = [𝑉 − 𝑉 ] + [𝑉 ] =𝑣
𝑦 = 𝑖 (17)

𝑣=𝐾 𝑖
I
Taking the controllable variable ‘v’ as current error,
− 𝑖 the current loop is linearised. Fig. 2. Block Diagram of the Proposed FBKLN Controller

B. Linearization of outer voltage loop

C. Feedback Constants Selection
The outer voltage loop also can linearised using the same
From Fig. 1, ωo = 1/LC and ω1=1/RC are obtained
method.
corresponding to the natural vibration frequency and the

or E
constant RC circuit frequency respectively. Since the system
considered is second order sytem, based on the classic C. Tansienst Response to Load Variation
control theory, the constants are defined as K v = ωn2 and
Ki = 2 𝜁ωn. When the load is varied from R= 9.426Ω to R=17.24Ω at
time 2.5sec, the Fig. 5, shows that FBKLN controller takes
V. SIMULATION RESULTS
5ms to reach the final value at the load variation instant,
whereas the PI controller takes 0.9sec to reach the final
The designed nonlinear controller was effectively tested value at the load variation instant.
by MATLAB. The superiority was compared with PI
controller results. The circuit parameters for 600W Boost
D. Transienst Response to Reference Voltage Variation
converter are as follows: Vin = 30V, Reference Vo = 72V, RL
= 9.426Ω, fs = 50kHz, Inductor = 100mH, Capacitor =
110µF.
0 It can be seen from the Fig. 6, that Vref has been varied
Feedback Coefficient for this system is found as from 55V to 72V at 2.5sec. FBKLN controller makes the
ωn=3.97e3and 𝜁 = 1.14e-̵̶1 and substituted for the feedback system to reach final value within 5.5ms whereas the PI
linearization controller. controller takes 0.15sec to reach the final value.

A. Startup Response

Fig. 3, shows the dynamic behavior of output voltage for

both PI and FBKLN controller. The output reaches 72V
given as Vref at 6msec when FBKLN Controller is used and
at 0.22sec when PI controller is used. This shows the
robustness of the proposed controller.
B. Transienst Response to Input Voltage Variation

In Fig. 4, the input voltage is varied from 50V to 30V at

2.5sec. It can be seen that FBKLN controller is more robust
than PI. The former takes 5msec to reach the final value
whereas the latter takes 1.3sec to reach final value.

Fig. 5. Regulation of Output Voltage corresponding to change in Load

Current (Load Regulation)

Fig. 3. Startup Response of Feedback Linearization Controller and PI

Controller

Fig. 6. Tracking of Output Voltage corresponding to change in

Reference Voltage

Fig. 4. Regulation of Output Voltage corresponding to change in Input

Voltage (Line Regulation)
 TABLE I. COMPARISON OF SIMULATION RESULTS models,” Int. J. Adapt. Signal Process., vol. 12, no. 1, pp. 63–
80, Feb. 1998.
FBKLN [11] M. Salimi. “Sliding mode control of the DC-DC fly back
PI controller converter with zero steady-state error,” J. Basic Appl. Sci. Res.,
controller
Settling time 0.22sec 6msec vol. 2, no. 10, pp. 10 693–10 705, Oct. 2012.
At Startup
% Maximum
Response ̶ 2.05 [12] Q. Zhang et al., “Sensorless predictive current controlled DC-
Peak Overshoot
Change in Settling time 0.15sec 5.5msec DC converter with a self-correction differential current
Reference % Maximum observer,” IEEE. Trans. Ind. Electron., vol. 61, no. 12, pp.
̶ 1.388 6747–6757, Dec. 2014.
Voltage Peak Overshoot
Settling time 0.9sec 5msec [13] S. Bibian and H. Jin, “High performance predictive dead-beat
Change in digital controller for DC power supplies,” IEEE Trans. Power
% Maximum
Load Current ̶ 12.5 Electron., vol. 17, no. 3, pp. 420–427, May 2002.
Peak Overshoot
Settling time 1.3sec 5msec [14] P. Karamanakos, T. Geyer, and S. Manias, “Direct voltage
Change in control of DCDC boost converters using enumeration-based
% Maximum
Input Voltage ̶ 0.694 model predictive control,” IEEE Trans. Power Electron., vol.
Peak Overshoot
29, no. 2, pp. 968–978, Feb. 2014.
[15] S. -K. Kim, C. R. Park, and Y. I. Lee, “A stabilizing model
predictive controller for voltage regulation of a DC/DC boost
converter,” IEEE Trans. Control Syst. Technol., vol. 22, no. 5,
VI. CONCLUSION pp. 2106–2023, Sep. 2014.
[16] S.-K. Kim, J.-S. Kim, C. R. Park, and Y. I. Lee, “Output-
In this paper Feedback Linearization technique has been feedback model predictive controller for voltage regulation of a
demonstrated to have overcome the inadequacies of DC/DC converter,” IET Control Theory Appl., vol. 7, no. 16, pp.
traditional PI control strategy utilized in power converters. 1959–1968, Nov. 2013.
From Table I., the robustness of Input-Output feedback [17] A. G. Beccuti, S. Mariethoz, S. Cliquennois, S. Wang, and M.
Morari, “Explicit model predictive control of DC/DC switched-
linearization controller can be plainly observed. The mode power supplies with extended Kalman filtering,” IEEE
performance of both the controllers are compared based on Trans. Ind. Electron., vol. 56, no. 6, pp. 1864–1874, Jun. 2009.
robust, servo and regulatory responses. The simulation [18] Osvaldo J.Rojas, Graham C.Goodwin. “Preliminary analysis of
results ensures that the system remains stable in the a nonlinear control scheme related to feedback linearization”,
Proceedings of the 40 IEEE Conference on Decision and
th

presence of parameters and load uncertainties. Control, Oriando, Florida USA, December 2001: 1743-1748.
Thus the Input-Output Feedback Linearization controller [19] Carlos F.Pfeiffer, Thomas F.Edgar. “Robust feedback
provides the best solution for voltage regulation problems in linearization and fuzzy control”, Proceedings of American
DC-DC Boost converters. Control Conference, San Diego, California USA, June 1999:
1508-1514.
REFERENCES [20] Seok-Kyoon Kim and Kyo-Beum Lee, “Robust Feedback-
Linearizing Output Voltage Regulator for DC/DC Boost
[1] Y.-X. Wang, D.-H. Yu, and Y.-B. Kim, “Robust time-delay Converter” IEEE. Trans. on Ind Electronics, vol. 62, no. 11,
control for the DC/DC boost converter,” IEEE Trans. Ind. november 2015
Electron., vol. 61, no. 9, pp. 4829–4837, Sep. 2014. [21] Xiaohu Liu, Xinchun Lin, Feidong Xu, Yong Kang, “Research
[2] J.-H. Lee, J.-S. Lee, and K.-B. Lee, “Current sensorless MPPT on BOOST Converter Based on Nonlinear Control Strategy”
control method for dual-mode PV module-type interleaved Proceedings of IEEE 6th Internaitonal Power Electronics and
flyback inverters,” J. Power Electron., vol. 15, no. 1, pp. 54–64, Motion Control Conference, 2009.
Jan. 2015.
[3] J.-T. Su and C.-W. Liu, “Gain scheduling control scheme for
improved transient response of DC-DC converters,” IET Power
Electron., vol. 5, no. 6, pp. 678–692, Jul. 2012.
[4] P. Karamanakos, T. Geyer, and S. Manias, “Direct voltage
control of DC-DC boost converters using enumeration-based
model predictive control,” IEEE Trans. Power Electron., vol.
29, no. 2, pp. 968–978, Feb. 2014.
[5] X. Hu and C. Gong, “A high gain input-parallel output-series
DC/DC converter with dual coupled inductors,” IEEE Trans.
Power Electron., vol. 30, no. 3, pp. 1306–1317, Mar. 2015.
[6] R. Priewasser, M. Agostinelli, C. Unterrieder, S. Marsili, and M.
Huemer, “Modeling, control, and implementation of DC-DC
converters for variable frequency operation,” IEEE Trans.
Power Electron., vol. 29, no. 1, pp. 287–301, Jan. 2014.
[7] J. C. Lopez, M. Ortega, and F. Jurado, “New topology for
DC/DC bidirectional converter for hybrid systems in renewable
energy,” Int. J. Electron., vol. 102, no. 3, pp. 418–432, Mar.
2015.
[8] R. W. Erickson and D. Maksimovic, Fundamentals of Power
Electronics, 2nd ed. New York, NY, USA: Springer-Verlag,
2001.
[9] S. Oucheriah and L. Guo, “PWM-based adaptive sliding-mode
control for boost DC/DC converters,” IEEE. Trans. Ind.
Electron., vol. 60, no. 8, pp. 3291–3294, Aug. 2013.
[10] H. Sira-Ramirez, R. Ortega, and M. Garcia-Esteban, “Adaptive
passivity based control of average DC-to-DC power converter

View publication stats