Sigma J Eng & Nat Sci 8 (2), 2017, 91-97

Publications Prepared for the Innovations on Intelligents Systems and

Applications Symposium ASYU 2016
Akıllı Sistemler ve Uygulamalardaki Yenilikler Sempozyumu
ASYU 2016 için Hazırlanan Yayınlar

Research Article / Araştırma Makalesi

MODEL PREDICTIVE CONTROL OF A DC-DC BUCK CONVERTER

Yahya DANAYİYEN*1, İsmail Hakkı ALTAŞ1, Erdinç ŞAHİN2

1
Karadeniz Technical University, Department of Electrical and Electronics Engineering, TRABZON
2
Karadeniz Technical University, Surmene Abdullah Kanca VHO, TRABZON

Received/Geliş: 02.12.2016 Accepted/Kabul: 11.12.2016

ABSTRACT

DC-DC converters are highly considered in recent years because of the increased use of the renewable energy
sources (RES). Since the energy output of the RES is depended on the weather conditions, DC-DC converters
are used to regulate variable RES output. In this study, simulation based model predictive control (MPC) of a
DC-DC buck converter is carried out. The performance of the designed controller is investigated under
variable load condition and results are compared with PID controller. The transfer function of the converter
model is used to design MPC controller in MATLAB/Simulink. MPC gives the better solution when it
compared to PID controller under condition of constant load. The overshoot is not occurred when the
predictive controller is used. PID provides better solution under the condition of load change whereas the
overshoot is increased.
Keywords: Model predictive control, buck converter, feedback control.

1. INTRODUCTION

DC-DC buck converter circuits include both linear and nonlinear parts. Switching devices
make its structure nonlinear. A digital signal which is triggering signal of the semiconductor
switches is used as an input signal of the system. The input and output voltage and current of buck
converter should be in a specific interval as such in the other power electronics converters [1].
These constraints should be taken into account in the controller design. The system constraints
and nonlinear parts are easily introduced the design in MPC method. This control method gives
desired system response when it is used in a buck converter [2]. MPC is a control method which
calculates the future values of the system behaviour using an accurate system model with
predetermined criteria in the objective function. The system model should be as close as the real
system and system mathematics should be well known to obtain a better MPC. A well-known
system model makes the parameter estimation easy [3]. There are some studies in which different
MPC approaches are used to control of a buck converter in the literature A nonlinear predictive
control method approach has been applied in [4] for a buck converter. Robust predictive
controller has been investigated in [5]. MPC method is applied to a buck converter including

*
Corresponding Author/Sorumlu Yazar: e-mail/e-ileti: [email protected], tel: (462) 377 20 86

91
Y. Danayiyen, İ.H. Altaş, E. Şahin / Sigma J Eng & Nat Sci 8 (2), 91-97, 2017

nonlinear inductance and nonlinear capacitance in [6] and [7], respectively and fourth order buck
topologies in [8].
In this study, the output voltage of a DC-DC buck converter is controlled by MPC method
which is designed in MATLAB/Simulink environment. The system constraints are taken into
account in the controller design process. The results are compared with PID controller. In the
second section, brief descriptions of MPC and buck converter are given. Finally the simulation
results are given in the third section.

2. MPC AND DC-DC BUCK CONVERTER

The general representation of a buck converter is shown in figure 1. MPC uses the output
voltage and reference voltage as two inputs. The optimization procedure is carried out using a
predefined criterion to calculate the optimal input sequence. The first term of this sequence is
applied to PWM generator as an input. PWM generator produces optimal duty cycle ratio and
apply it to switching element.

Q L

iL io +
V
Vin Vc C R Vo
-

control signal

PWM measured voltage

MPC
d generator reference voltage

Constraints
Vin, io

Figure 1. General block diagram of the system

2.1. Model Predictive Control

MPC method depends on an accurate system model. The future behaviour of the system is
calculated with an optimization cost function based on receding horizon principle using current
system values. The first element of the calculated input vector which is obtained every time step
is applied to the system as a control input within a predefined optimization window. The basic
principle of MPC is shown in figure 2. At the current time instant 𝑘, the calculated control
sequence ∆𝑈 is given below [9]. The first element of ∆𝑈 is chosen as an input at time instant 𝑘.
∆𝑈 = [∆𝑢(𝑘) ∆𝑢(𝑘 + 1) ∆𝑢(𝑘 + 2) . . . ∆𝑢(𝑘 + 𝑁𝑐 − 1) ]
𝑢(𝑘) = ∆𝑢(𝑘) + 𝑢(𝑘 − 1)
The discrete time model of a linear time invariant system is given in (1) and (2), respectively.
𝑥(𝑘 + 1) = 𝐴𝑥(𝑘) + 𝐵𝑢(𝑘 − 𝑟) (1)
𝑦(𝑘) = 𝐶𝑥(𝑘) (2)
The predicted state and output equations for the discrete time LTI system is given in (3) and
(4), respectively.

92
 Model Predictive Control of a Dc-Dc Buck Converter / Sigma J Eng & Nat Sci 8 (2), 91-97, 2017

𝑥̂((𝑘 + 𝑗|𝑘)) = 𝐴̂𝑥̂((𝑘 + 𝑗 − 1|𝑘)) + 𝐵̂𝑢̂((𝑘 + 𝑗 − 𝑟̂ |𝑘)) (3)

𝑦̂((𝑘 + 𝑗|𝑘)) = 𝐶𝑥̂((𝑘 + 𝑗|𝑘)) (4)
The cost function is given in (5). It is defined as a quadratic programming. 𝑄𝑤 and 𝑅𝑤 are the
weigth matrixes that should be chosen properly taken into account the system model. The
objective function 𝐽 is calculated in every time step within the optimization window [9].
𝐽 = (𝑅𝑠 − 𝑌)𝑇 𝑄𝑤 (𝑅𝑠 − 𝑌) + ∆𝑈 𝑇 𝑅𝑤 ∆𝑈 (5)

Figure 2. Basic principle of Model Predictive Control

2.2. DC-DC Buck Converter

A buck converter is a switch mode power electronic circuit which steps down input DC
voltage to a lower value output DC voltage. Schematic of the proposed converter is given in Fig.
3. The converter consists of an inductor (L), output capacitor filter (C), switching element (Q),
diode (D) and resistor (R) as a load.

Q L

iL io +
Vin d Vc C R Vo
-

Figure 3. The circuit diagram of buck converter

Buck converter operates in two-state. The first state is the switch On-state and the other state
is the switch Off-state. These states are briefly described in the following sections, respectively.

93
Y. Danayiyen, İ.H. Altaş, E. Şahin / Sigma J Eng & Nat Sci 8 (2), 91-97, 2017

2.2.1. Switch On-state

In this state, the diode is reverse biased and the inductor is charged by input voltage as seen in
Fig. 4. The equations of this state is given in (6-10).
𝑉𝑖𝑛 = 𝑉𝐿 + 𝑉𝑜 (6)
𝑑𝑖𝐿
𝑉𝑖𝑛 = 𝐿 + 𝑉𝑜 (7)
𝑑𝑡
𝑑𝑖𝐿 𝑉𝑖𝑛 −𝑉𝑜
= (8)
𝑑𝑡 𝐿
𝑑𝑉𝑜
𝑉𝑜 = 𝑉𝑐 𝑖𝑐 = 𝐶 (9)
𝑑𝑡
𝑑𝑉𝑜 1 𝑉
= (𝑖𝐿 − 𝑜 ) (10)
𝑑𝑡 𝐶 𝑅

iL ic io +
Vin Vc C R Vo
-

Figure 4. Switch On-state of a buck converter

2.2.2. Switch OFF-state

In this state, the diode is forward biased. The charged inductor behaves as a current source
and linearly discharged. The differential equations of this state is given in (11-12). The schematic
of this state is also shown in Fig. 5.
𝑑𝑖𝐿 𝑉𝑜
=− (11)
𝑑𝑡 𝐿
𝑑𝑉𝑜 1 𝑖 −𝑉𝑜
= (𝐿 ) (12)
𝑑𝑡 𝐶 𝑅

iL io +
Vc C R Vo
-

Figure 5. Switch OFF-state of a buck converter

The equations belong to aforementioned states are used to obtain average model of the
converter. The average model equations are given below.
1 𝑑
𝑥̇1 = − 𝑥2 + 𝑉 (13)
𝐿 𝐿 𝑖𝑛

94
 Model Predictive Control of a Dc-Dc Buck Converter / Sigma J Eng & Nat Sci 8 (2), 91-97, 2017

1 1
𝑥̇ 2 = − 𝑥1 + 𝑥2 (14)
𝐶 𝑅𝐶

Table 1. The parameters of the converter

Parameters Symbol Value
Load 𝑅 20 Ω
Inductance 𝐿 3.23mH
Capacitance 𝐶 6300µF
Input voltage 𝑉𝑖 10 V
Reference voltage 𝑉𝑟 4V

3. SIMULATION RESULTS

MPC method is used to control output voltage of the buck converter in order to obtain desired
output voltage without oscillations and fluctuations. The designed controllers are tested under
variable load. The system constraints are taken into account. In Fig. 6, the results are obtained
with the prediction horizon 𝑁𝑃 = 10 and the control horizon is 𝑁𝐶 = 6. For the tuning of PID
parameters, trial-error method is used. There is no overshoot when the MPC is used and
oscillation in the output signal is observed when PID is used as seen in Fig. 6. With the new
parameters 𝑁𝑃 = 20 and 𝑁𝐶 = 19, the system response is showed in Fig. 7. The result is
examined under condition of variable load and constant reference. PID gives better performance
compared to MPC as seen in Fig. 8. Because MPC controller depends on accurate system model.
The result is shown in Fig. 8.

Figure 6. System results with MPC and PID controller: Np=10, Nc=6, Kp=0,Ki=20,Kd=0

95
Y. Danayiyen, İ.H. Altaş, E. Şahin / Sigma J Eng & Nat Sci 8 (2), 91-97, 2017

Figure 7. System results with MPC and PID controller: Np=20, Nc=19, Kp=10,Ki=20,Kd=0.1

Figure 8. System results with MPC and PID controller under condition of variable load

96
 Model Predictive Control of a Dc-Dc Buck Converter / Sigma J Eng & Nat Sci 8 (2), 91-97, 2017

3. CONCLUSION

The output voltage of a DC-DC buck converter is controlled by using MPC in this study. The
controller is designed in MATLAB/Simulink and the constraints in the system such as voltage
limitations are taken into account in the controller design process. The reference tracking of
output voltage is aimed in the designed controller. The results are compared with PID controller
and effectiveness of MPC is evaluated.

REFERENCES / KAYNAKLAR

[1] M. P. Kumar, P. Ponnambalam, S. Sreejith, J. Belwinedward, and K. Krishnamurthy,

(2014) Comparison of fuzzy and MPC based buck converter, 2014 IEEE Int. Conf. Power
Electron. Drives Energy Syst. PEDES 2014,16-19 December 2014, Mumbai, India
[2] J. Rodriguez and P. Cortes, (2012) Predictive Control Of Power Converters And
Electrical Drives, Wiley-IEEE.
[3] S. Di Cairano and A. Bemporad, (2010) Model predictive control tuning by controller
matching, IEEE Trans. Automat. Contr., vol. 55, no. 1, pp. 185–190, 2010.
[4] A. Beccuti, M. Lazar, W. Heemels, G. Papafotiou, and M. Morari, (2007) Assessment of
Nonlinear Predictive Control Techniques for Dc-Dc Converters, IFAC Nonlinear Control
Design Symposium, pp. 395-401, 22-24 August 2007, Pretoria, South Africa
[5] J. Weigold and M. Braun, (2006) Robust Predictive Dead-Beat Controller for Buck
Converters, EPE-PEMC 2006: 12th International Power Electronics and Motion Control
Conference, Proceedings, pp. 951–956, 30 August-1 September 2006, Portoroz, Slovenia
[6] S. Mariethoz, M. Hercegy, and M. Kvasnica,(2008) Model Predictive Control of Buck
DC-DC Converter with Nonlinear Inductor, 11th IEEE Work. Control Model. Power
Electron. COMPEL 2008, pp. 1–8, 17-20 August 2008, Zurih, Switzerland.
[7] V. S. P. Cheung, H. S. H. Chung, and H. Wang, “Predictive Control of Buck Converter
Using Nonlinear Output Capacitor Current Programming,” Conf. Proc. - IEEE Appl.
Power Electron. Conf. Expo. - APEC, pp. 491–498, 21-25 Feb. 2010, CA, USA
[8] C. S. babu and M. Veerachary, (2005) Predictive Controller for Fourth Order Buck
Topologies, INTELEC 05 - Twenty-Seventh International Telecommunications
Conference, pp. 411-414., 18-22 Sept. 2005, Berlin, Germany
[9] Liuping Wand, (2009) Model Predicitive Control System Design And Implementation
Using MATLAB, Springer London.

97