TS Fuzzy Output Control for a Boost Converter in

PV Systems

Abdellatif Belattar * Mohammed Chennanni Said Doubabi

Laboratory of Electric Systems and Laboratory of Electric Systems and Laboratory of Electric Systems and
Telecommunications Telecommunications Telecommunications
Faculty of Sciences and Technologies of Faculty of Sciences and Technologies of Faculty of Sciences and Technologies of
Marrakesh, Cadi Ayyad University Marrakesh, Cadi Ayyad University Marrakesh, Cadi Ayyad University
Marrakesh, Morocco Marrakesh, Morocco Marrakesh, Morocco
Belattar.abdellatif@gmail.com medchennani@gmail.com saiddouabi@gmail.com

Abstract—this paper presents fuzzy output tracking control Other controls deal with nonlinearity of the DC-DC converters.
of switching photovoltaic (PV) DC-DC converters. The Such as, state feedback-linearising control [4, 5], sliding-mode
proposed control ensures good performances and overall control [6], and fuzzy one. As very known, the feedback
large-signal stability of the system over the whole linearising control presents a very poor dynamic performance
operating space. To do this, we present a model of PV and is not easy to be implemented. The sliding mode technique
system and a state space model for DC-DC boost is widely used to control the nonlinear system. However, the
converter. Then, we introduce an added integral state of main problem with this approach is chattering phenomenon. In
the output tracking error. The resulting augmented system recent years, Takagi sugeno fuzzy logic modeling concept is
is converted into a Takagi sugeno fuzzy model (TSFM). used to improve dynamic behavior and zero tacking. This is
The Parallel distributed compensation (PDC) concept is due to their power for approximating a complex non linear
employed to design the state-feedback based control law system [7, 8]. Using this technique, many works are developed
whereby the control gains are fined using the linear to control the DC-DC converters [9, 10, 11 12]. In [9] shows
quadratic regulator (LQR) technique. Finally, the design the comparison between the PI controller and fuzzy one based
on a T-S fuzzy model .in [11] the others proposed an LMI-base
procedures are applied to a dynamics model of typical PV
fuzzy control of buck converter. In [12], a nonlinear design
power control system using a DC-DC boost converter to
methodology is proposed for fuzzy controllers that operate in
illustrate the effectiveness of the proposed control large-signal domain. Especially, output regulation of
techniques. photovoltaic DC-DC power converters system via a TS fuzzy
model-based approach is presented in [13] [14]. The study
Keywords; Pv system; DC-DC converter; Takagi sugeno; presented in [13] deals with the synthesis of a fuzzy controller
PDC, LQR. of DC-DC boost converter by using the integral action
performance. The work presented in [14] proposed TS fuzzy
I. INTRODUCTION controller to regulate the produced voltage and maintain the
Without pollutant emission and without noise, photovoltaic system stable during photovoltaic parameter uncertainties.
cells convert sunlight directly to electricity. However, the
major incontinent of PV panels is that the output voltage However, in these works the reference power is assumed to
depends strongly on sunlight. Therefore, the load cannot be constant. Nevertheless, the reference signal may be changed
directly connect to PV system. For that, the DC-DC converter according to the application case. In this paper, a fuzzy output
plays an interface between the load and PV system. tracking control based on TS fuzzy model with the duty ratio as
premise variable for switching photovoltaic DC-DC boost is
As the PV system is including in vehicle field where the presented.
sunlight can change considerably during a displacement and
the output voltage of the DC-DC converter should be in line The rest of this paper is organized as follows, the modeling
with voltage desired by load. For this subject, many approaches of the PV system and the state space model of the converter are
and studies are developed to control DC-DC converter. presented in the second section. The third section presents the
Control techniques that based on linear models which have design of fuzzy controller with integral action performance.
local validity. Therefore, it can only assure stability and The fourth section shows the simulation results obtained and
dynamic behavior around equilibrium point [1, 2, 3]. discussion. The conclusion is presented in the last section.

978-1-4673-7894-9/15/$31.00 ©2015 IEEE

Authorized licensed use limited to: Consortium - Algeria (CERIST). Downloaded on May 21,2021 at 11:10:19 UTC from IEEE Xplore. Restrictions apply.
 II. SYSTEM MODELING
A. Photovoltaic model
Photovoltaic technology consists to convert sunlight
directly to a current. The photovoltaic module is a nonlinear
device which can be modeled by a current source. The current
generated by a PV model neglecting the internal series
resistance is given by the equation (3) [15-16]:

(1)

where and are the output current and output voltage

of the PV module, respectively, is the generated current Fig. 2: DC-DC boost converter
under a given insolation, is the reverse saturation current, q
is the charge of an electron, k is the Boltzmann’s constant, A is
the ideality factor, is the temperature (K) of the PV module, And the second is during the "OFF" period, it can be
and is the current due to intrinsic shunt resistance. The expressed by these equations:
characteristic of the PV system used in this paper is presented
in Fig.1.
(3)
B. DC/DC boost converter
In this section, the modeling of the DC/DC converter based
According (2) and (3), the converter is described by a set of
on the state-space averaging method is presented as in [17].
state equations which is given by:
As seen in Fig.2, the DC/DC converter is composed of the
output capacitor C, the inductor L, the load R, the ESR resistor, i= {1, 2} (4)
diode D and active switch . In this study, the capacitor
resistance Rc is negligible with respect to the loading resistance With for the first configuration described by (2);
R. for the second configuration described by (3).
In addition, the Vin, iL, y represent the input voltage,
inductor current and output voltage respectively, The converter Where , , ,
switching frequency is denoted by and the switching period
is .
, , .
The capacitor voltage and the inductor current define the
state vector x=[x1 x2]T=[iL vc]T. During one Ts period, the system can be represented by its
state-space averaging model (SSAM) as it has been presented
In continuous conduction mode, two configurations are in several works [17, 18, and 19] and it can give by:
possible according to switch state [18].
The first is during the "ON" period; it can be expressed by (5)
following equations:
With ; ;
(2)
Where is the duty cycle, and represent the mean
over an interval of length Ts of and respectively. and we
can note , and .
Where , and are the DC steady-state of xm, ym and d,
respectively.
By neglecting the second order terms and given that C1=C2
the small signal model (SSM) is of the following form [17]

(6)
Where

In steady-state, , thus and

for duty cycle and with a fixed, the

Fig. 1: characteristic of the used PV module: (a) current-

voltage curve, (b) power-voltage curve.

Authorized licensed use limited to: Consortium - Algeria (CERIST). Downloaded on May 21,2021 at 11:10:19 UTC from IEEE Xplore. Restrictions apply.
conversion ratio M (d ) relating the output voltage to the input With and are local system matrices with appropriate
voltage is given by: dimensions.

(7) III.CONTROL STRATEGY

This work aims to control the output voltage of boost
From (6), it can be seen that M(d) dependent on R and RL converter where its input is PV generator. The control included
and it is plotted in Fig. 3 for RL/R=0.1 where it can be clearly in this document based on takagi-sugeno approach where the
seen the nonlinearity of the DC-DC converter. It is necessary Takagi sugeno fuzzy approximator (TFSA) provides the duty
to take into account the effect of this parasitic resistor in both ratio which is necessary to obtain and then fined
modeling and control. the corresponding (SSM). In the order hand the PDC approach
is used to calculate control law by state feedback for each sub
Therefore, we will adopt in the next section the TS fuzzy model of fuzzy model and through linear quadratic regulator
modeling approach for selecting local models so that can be (LQR) technique. In the following, the Takagi sugeno fuzzy
determined space local state over the entire extent of the cyclic controller (TSFC) infers the adequate state feedback gains by
ratio. using the grade of membership of .
C. TS model A. TAKAGI SUGENO APPROXIMATOR (TSFA)
The Takagi-Sugeno fuzzy model provides a general As mentioned above, the TFSA aims to predict the from
framework to represent a nonlinear system using a set of local a reference voltage and from the conversion ratio
linear models that are linked by nonlinear fuzzy membership characteristic (fig. 3).
functions.
For this, we first partition the conversion ratio interval into
"If-then" rules are used to represent the relationship between several sub-intervals. Thus the duty cycle value is precisely
different local models. Its form is as following: predicted over each sub-interval trough the related affine
If then function.

Where i=0…r, and is ith first fuzzy set, r is the rules (10)
th
number; and are i subsystem matrix. With i= 1....n and n is the number of the sub-intervals.
The global model of the system is achieved by fuzzy
combination of these linear models using nonlinear fuzzy The resulting TSFA combines all these functions in a
membership functions. It can be expressed as follow: nonlinear form in order to predict globally the duty cycle for
the whole conversion ratio range.
(8)
(11)
Where is membership degree with . With represents the degree of membership of the
D. Augmentated system
B. TS FUZZY CONTROLLER FOR THE BOOST
After linearising the system and obtaining the TS fuzzy
CONVERTER
model of the boost converter, And in order to achieve zero
steady-state error an added state variable as the integral of the After obtaining the TS fuzzy model of the boost converter,
difference between output and reference voltage is introduced we propose to design the TS fuzzy controller for the converter
and it is given by: with guaranteed stability and good dynamic. The PDC
technique is used to design fuzzy controller to stabilize fuzzy
system.
A four-rule fuzzy controller employing the same
Consequently, the TS-fuzzy plant model of the augment membership functions of the TS-fuzzy plant model of (9) will
system given by (8) is expressed as following [17]. be used to control the converter. The rules of the fuzzy
controller are as follow:
(9)

Where is the augmented system state, and

the new system matrices are given by:

The fuzzy system becomes the following form:

If is then .

Fig. 3: conversion ration M(d) for the averaged model and

its approximation with TSFA.

Authorized licensed use limited to: Consortium - Algeria (CERIST). Downloaded on May 21,2021 at 11:10:19 UTC from IEEE Xplore. Restrictions apply.
Controller Rule i: IF THEN IV. SIMULATION RESULTS
In this section, we well consider the problem of regulating
Where is linear feedback gain vectors associate to rule, the output voltage of a PV power control system using a DC-
the is determined by LQR technique which is embedded in DC Boost switching converter via fuzzy control. According to
Matlab software. the circuit diagram in Fig.4 this later has the following
The cost index to be minimized is expressed as parameters; L=1.08mH, RL=0.57Ω, C=470µF, and fs=40 KHz.
Matlab/Simulink software and "LMI Control Toolbox" are
(12) used to simulate converter MFTS and the corresponding fuzzy
Where and are positive semi-definite matrices of control law.
appropriate dimensions.
According to the modeling approach given in section II.C,
The state-feedback control law in the ith subspace (which the converter can be represented by TSFM with the following
minimizes the cost function ) is expressed as: rules:
(13) .
Where Pi is a positive definite and symmetric matrix
solution of the following Riccati equation:

The final model-based fuzzy controller is analytically

represented by:
(14)
Thus, the global system can be expressed as:
(15)

The feedback gains are determined by an LMI-based design

technique addressed in the following theorem.
Theorem
The equilibrium of fuzzy control system (14) is
asymptotically stable in the large if there exists a common
positive definite matrix P such that the following two ;
inequalities are satisfied:
For i= 1…r
, ;
For i<j<r
The input cost matrices used for minimization of the cost
With
index given by (11) are:

The proof of this theorem can be found in [20] and it can

be reformulated as a problem of LMI and interior-point And
algorithm which is included in Matlab/Simulink software to
resolve this problem.
By using the LMI method and LQR procedure in Matlab
toolbox the matrix P and feedback gains are:

Fig. 4: Bloc diagram of the TS fuzzy controller

Fig. 5. Membership functions of the premise variable

of: (a) TSFC. (b)TSFA

Authorized licensed use limited to: Consortium - Algeria (CERIST). Downloaded on May 21,2021 at 11:10:19 UTC from IEEE Xplore. Restrictions apply.
k1= [-3.1623 1,606 2.9599], k2= [-3.1623 1.5294 2.9673];
k3= [-3.1623 1.5549 2.9156], k4= [-3.1623 1.8236 2.428];

Different simulations are done to prove the effectiveness of

the proposed control. For this and in order To verify the
performance of the boost converter controlled by the integral
TS fuzzy controller, the variations of the sunlight, load and
reference output are considered.
Firstly, the scenario presented in Fig. 6, aims to examine
the dynamic performances of the controller through a step
variations in the reference voltage during the previously
attained steady states are shown in fig. 10. Indeed, the
reference voltage is respectively set
. As one can remark, the system
output tracks perfectly the reference voltage with no overshoot
in both output voltage and inductor current
Secondly, the figure Fig. 7 illustrates the output voltage
response controlled by the designed TS fuzzy controller when
the boost converter is subject to the sunlight variation. So, the
Sunlight changes from 1000 w/m2 to 800w/m2 and to 600w/m2,
1000 w/m2 and then 1200 w/m2. We assume that the cell
temperature is held constant at 25 ° C.
As seen, the output voltage converges perfectly to the
desired trajectory and did not get affected by the solar
irradiation with any overshoot.
Fig. 7: (a) The sunlight variation, (b) The output voltage
Thirdly, the Fig. 6 shows the response of output voltage as
response, (c) inductor current response
well as the inductor currents of the boost converter driven by
the TS fuzzy controller. Load R has been varied from
respectively. One can see that the
output voltage rejects quickly perturbations to the desired
voltage.
As seen in different results, the voltage response of the PI
controller presents an important oscillations and undulation in
transient regime, what prove that the fuzzy controller gives a
better performance than the PI controller.

Fig. 8: (a) Voltage response (b) inductor response for load

Fig. 6: Voltage response for yref= variation R=

Authorized licensed use limited to: Consortium - Algeria (CERIST). Downloaded on May 21,2021 at 11:10:19 UTC from IEEE Xplore. Restrictions apply.
 V. CONCLUSION Converters,” IEEE Trans. Ind. Electron.,vol. 56, no. 6,
june 2009.
In this paper, Takagi sugeno fuzzy controller for the control [9] H.K. Lam, F.H.F. Leung and P.K.S. Tam, “Fuzzy control
of the DC-DC boost converter connected in PV systems has of dc-dc switching converters based on TS-modeling
been presented. Based on the TSFM describing the switching approach,” in Proc. 24th Annual Conf. IEEE Ind.
converter working in large operating domain, the controller has Electron. Society IECON’98, pp. 1052-1054, Sept. 1998.
been proposed to handle the control problem. To track the [10] H.K. Lam, T.H. Lee, F.H.F. Leung , P.K.S. Tam, “Fuzzy
reference output voltage fuzzy state feedback controller was Control of DC-DC Switching Converters: Stability and
Robustness Analysis,” in Proc. 27th Annual Conf. IEEE
designed via PDC concept whereby the control gains are fined Ind. Electron. Society IECON’2001, vol. 2, pp. 899-902,
by LQR technique. Furthermore, Lyapunov theory is used to 2001.
guarantee the stability of the closed loop system operating in [11] K.Y.Lian, J.J.Liou. “ LMI based Integral Fuzzy Control
large signal. The regulated converter is subject to significant of DC-DC Converters, ” IEEE Trans. Fuzzy Systems, ,
changes of sunlight which affects the input voltage from PV pp. 71–80, Feb.2006.
and the load resistance; the simulation results prove that [12] M. Allouche, M. Chaabane, M. Souissi, D. Mehdi and F.
proposed takagi sugeno fuzzy controller can achieve good Tadeo." Takagi-Sugeno Fuzzy Control of DC-DC Boost
Converter in PV Systems" Proceedings of the 2nd
performance defined by a good transient responses and a less International Conference on Systems and Control,
sensibility to the disturbance effect than a PI controller Marrakech, Morocco, June 20-22, 2012.
[13] H. K. Lam, S. C. Tan ,”Stability analysis of fuzzy-model-
REFERENCES based control systems: application on regulation of
switching DC-DC converter,” IET Trans. Control Theory
[1] R. W. Erickson and D. Maksimovic, “Fundamental of & Applications, vol. 3, pp. 1093-1106, Aug 2009.
Power Electronics, ” 2nd ed. Norwell, MA: Kluwer, 2001. [14] E. Kamal, A. Aitouche, Member, IEEE, R. Ghorbani and
[2] E. William Z. Bustamante, J. Linares-Flores, E. G. M. Bayart." Output Regulation of PV Power System via a
Ramírez, and H. Sira-Ramírez, “A Comparison Between TS Fuzzy Model-based Approach" Proceedings of the 2nd
the GPI and PID Controllers for the Stabilization of a International Conference on Systems and Control,
DC–DC Buck Converter: A Field Programmable Gate Marrakech, Morocco, June 20-22, 2012.
Array Implementation,” IEEE Trans. Ind. Electron., vol. [15] Manoj, D. Tomonobu, S., Atsushi, Y., and Toshihisa,F."
58, no. 11 pp. 5251–5262, Nov. 2011. A fuzzy based method for leveling output power
[3] L. Guo, J. Y. Hung, R. M. Nelms. “Evaluation of DSP- fluctuations of photovoltaic-diesel hybrid power system".
Based PID and Fuzzy Controllers for DC–DC Renewable Energy, vol 36, 1693-1703, 201.
Converters,” IEEE Trans. Ind. Electron.,vol. 56, no. 6, [16] Nopporn, P., Suttichai, P.P., and Yosanai S. "Maximum
june 2009. power point tracking using adaptive fuzzy logic control
[4] Krein, P.T., Bentsman, J., Bass,R.M., and Lesieutre,B.L, for grid-connected photovoltaic system”, Renewable ”,
“On the use of averaging for the analysis of power Renewable Energy, vol 30, 1771–1788, 2005.
electronic systems,” IEEE Trans. Power Electron., , vol. [17] S. Elbeid and S. Doubabi,“DSP-Based Implementation of
5, no. 2, pp. 182–190, 1990. Fuzzy Output Tracking Control for a Boost Converter”
[5] Alvarez-Ramirez, J.,Cervantes, I., Espinosa-Perez, G., industrial Electronics, IEEE Transactions. Vol 61, pp. 196
Maya, P., andMorales, A,“A stable design of PI control – 209, jan. 2014.
for DC-DC converters with an RHS zero”, IEEE Trans. [18] R. D. Middlebrook and S. Cuk, “A general unified
Circuits Syst. I, Fundam. Theory Appl., vol. 48, no. 1, pp. approach to modeling switching converter power stages,”
103–106, 2001. in Proc. IEEE Power Electron. Spec. Conf., 1976, pp. 18–
[6] R-J. Wai, L-C. Shih “Design of Voltage Tracking Control 35.
for DC–DC Boost Converter Via Total Sliding-Mode [19] J. Sun and H. Grotstollen, “Averaged modeling of
Technique”. IEEE Trans. Ind. Electron., vol. 58, no. 6, pp. switching power converters: Reformulation and
2502–2511, Jun. 2011 theoretical basis,” in Proc. IEEE Power Electron. Spec.
[7] Zadeh LA. “Outline of a new approach to the analysis of Conf, vol. 2, pp. 1166–1172. 1992.
complex systems and decision processes,” IEEE Trans.
on Sys, Man, Cybern; vol.SMC-3, no. 1, pp. 28–44, 1973.
[8] L. Guo, J. Y. Hung, R. M. Nelms. “Evaluation of DSP-
Based PID and Fuzzy Controllers for DC–DC

Authorized licensed use limited to: Consortium - Algeria (CERIST). Downloaded on May 21,2021 at 11:10:19 UTC from IEEE Xplore. Restrictions apply.