Application of feedforward controller based on fuzzy logic control with P&O algorithm for improving maximum power point tracking

This paper proposed a scheme combining both conventional P&O algorithm and a Feedforward compensator (P&O_FFC) adjusted by using FLC mechanism for improving the accuracy, and reducing oscillation in comparision with P&O or FLC applied in seperately.

APPLICATION OF FEEDFORWARD CONTROLLER BASED ON FUZZY LOGIC CONTROL WITH P&O ALGORITHM FOR IMPROVING MAXIMUM POWER POINT TRACKING

Dang Van Huyen*, Nguyen Van Chi

Abstract: The maximum power point tracking technique plays an important role

to improve the working efficiency of photovoltaic (PV) systems There are many algorithms introduced and applied on a maximum power point tracking task such as: Perturb and Observe (P&O), Increment Conductance (INC), Fuzzy Logic Control (FLC), Neural Network (NN), etc, in which P&O is the most widely used one because of its simplicity and easy implementation However, this algorithm draws back a low accuracy of the maximum power point, and causes a high oscillation In order to overcome these limitations, this paper proposed a scheme combining both conventional P&O algorithm and a Feedforward compensator (P&O_FFC) adjusted by using FLC mechanism for improving the accuracy, and reducing oscillation in comparision with P&O or FLC applied in seperately In order for testing the control algorithm and structure, the PV dynamic model and boost converter model are first established using Matlab/Simulink Next, the experimental setup will be used for validating the results

Keywords: Photovoltaic (PV); DC-DC Boost converter; PWM; P&O; INC; FLC; NN; Feedforward

I INTRODUCTION

Energies generated from fossil are limitted and cause harmful effects to natural environment, therefore using renewable, fresh, and unpoluted energies such as wind or solar resources gets more interested in engineers and researchers, especially solar energy Solar energy can be directly converted to electric power via PV module [1], [2], [3] Each

PV module has its own characteristics depended on the whether conditions: solar irradiation, temperature These factors are not maintained, but they change due to the daily time Hence, we need to figure out working points where the PV modules produce the highest power corresponding to these changes in order to get maximum power amount This method is generally called maximum power point tracking [4] (MPPT)

It can be understood MPPT such the way that matchs resistance between the PV module and system loads during the changes of the irradiation and temperature for making the PV module works at effective working points A boost converter [4], [5] is connected

in between the PV module and load for adjusting the resistance It means that both loads and boost converter can be seen as the overal system load There are several algorithms applied for tracking MPP, and basic ones can be considerd as: Perturb and Observe Algorithm [6] (P&O), Incremental Conductance Algorithm [6] (INC), Constant Voltage Algorithm [6] (CV), and some other popular development solutions are Fuzzy logic [7], [8], Neural network [8], [9], etc In which, the P&O is the most widely used one, because

of its simple structure for implementing easily However this method draws back strong fluctuations and inaccuracy of MPP More complex constructions such as Fuzzy logic or neural network techniques help to reduce fluctuated and inaccurate phenomena, but the more complexity means harder implementation

In order to overcome the obstructions faced while using these mentioned algorithms, this paper proposes a new deisgn, in which P&O algorithm is used for approaching rapidly MPP region, and Feedforward loop is added to compensate the fluctuation under the variation of the whether conditions The feedforward loop will be determined by Fuzzy

logic mechanism The paper is organized as following: Section II shows the mathematical model of PV system, and boost converter, Section III explains the new control design in which the P&O algorithm and Feedforward control loop will be demonstrated, Section IV gives Matlab/Simulink model and Simulation results, Finally, some conclusions are disscused in Section V

II MATHEMATICAL MODEL

1 Mathematical model of the PV system

Figure 1 a) The equivalent circuit of PV cell; b) The equivalent circuit of PV module

The equivalent circuit of PV cell is shown in Figure 1.a, and the I-V characteristic of the PV cell [1] can be described as Equation (1)

 

q V+R Is

pv s AkT

sh

V +R I I=I - I - I = I -I e 1

-R

(1)

Where, I is the panel current; Iph and ID are photogenerated and saturation current respectively; V is the panel voltage (V); Rs is the equivalent series resistance; Rsh is the equivalent shunt resistance, A is ideality factor; k is Boltzman constant; q is the electron charge; T is the temperature in Kelvin In case of Np, and Ns cells connected in parallel and series respectively, we can have the equivalent circuit of PV module as shown in Figure 1.b The I-V characteristic of the PV module [1]can be described as Equation (2)

p ph p D sh

The current generated from PV module is:

λ

I = I +K T-298

1000

The current through diode ID is:

s

I.R

q AkT

D s

(4)

The saturation current is:

g r

qE 1 1

-A.k T T

s rs r

T

T

(5)

sc

rs q.V

N A.k.T

I

I =

(6)

In which: Isc is short circuit current at ideal condition [A]; Ki is a ratio of current/temperature [A.oC];  is irradiation [W.m2]; Eg is energy to activate electron of silican; Tr is reference temperature; Irs is inversed saturation current at Tr; Voc is the open circuit voltage at reference temperature

The current passing through shunt resistance is:

s

sh p

sh

I.R

I =N

2 Boost converter

A boost converter (see Figure 2) is one type of several DC-DC converters which step

up output voltage magnitude from the input voltage magnitude

Figure 2 Boost converter circuit

It is established by a voltage source connected in series to an inductor, a biased diode, a capacitor, and a load of resistance R at the output terminal The regulation is normally obtained by pulse width modulation (PWM) and the switching device is normally MOSFET or IGBT The model of Boost converter is established by using Simcape/Simulink

3 Photovoltaic maximum power point using boost converter

Assume that the output of PV module is directly connected to the load It can be seen from the I-V characteristic curve (Figure 3) that the maximum power point (MPP) is depended on the load resistance

Figure 3 Effect of the load due to MPPT Figure 4 MPPT Boost configuration

Otherwise the I-V and P-V curves [1] vary due to the variation of solar irradiation and environment temperature Thus the load resistance needs to be properly adjusted, therefore the system will work at the maximum power point, unfortunately the load resistance should not manually adjusted, because there are a lot of MPPs due to the changing process

of the irradiation and temperature

The idea here is use of a DC-DC converter [4] which is a Boost converter in this case, and its connection is shown in Figure 4 Both Boost converter and load resistance are considered as equivalent impedance, and the equivalent resistance can be varied

by adjusting the duty cycle D of the Boost converter The formula indicating the relationship between the equivalent impedance and the duty cycle D is described as following:

pv load

load pv

V

pv

I

=1-D I = 1-D I

(8)

load load load

V

I

(9)

Replaces equation (8) into equation (9), the equivalent impedance is:

equ load

The other work finding the proper duy cycle D will be done by MPPT technique (see Figure 5) The MPPT technique can be P&O, INC or others [6]

III CONTROL DESIGN FOR MAXIMUM POWER POINT TRACKING

1 Conventional P&O algorithm

I sc

I MMP

V MMP V oc

0

V[V]

I[A]

MMP

Case1

Case2

Case3

Case4

P[W]

s D

Figure 6 P&O algorithm

explanation

If both power and voltage of PV

module vary in the same sign,

positive and negative, the duty cycle

D should be increased, vice versa Figure 7 Flowchart of P&O algorithm The P&O algorithm can be described in detail by the above flowchart (Figure 7) P&O

is the most widely used method for finding maximum power point, because it requires fewer number of sensors, and easy to implement in real equipments Due to the data reading from voltage and current sensors, it figures out the variation of the PV power and voltage at the current time and previous state If the PV power and voltage change in opposite sign, the duty cycle D should be reduced in order to generate the future cycle of pulse generator to force the working point of the PV module approaching to the MPP This

Figure 5 Basic structrure for MPPT

work will act automatically until the PV system works at the MPP region MPPT region is mentioned here, because the P&O can not help the voltage reach to precise value in steady state where the PV system can archieve the maximum power

2 Proposed Feedforward compensator added to improve P&O algorithm

In order to get rid of the limitations from P&O algorithm such as: slow response speed, oscillations in steady state, and wrong way tracking under rapidly changing conditions[6] [7] This paper introduces a new control design in which Feedforward control loop is added together with P&O algorithm The Feedforward control loop has two main functions In the transient time interval, the feedforward control will generate a large mount signal to add with the signal from the P&O control loop that helps reduce the time response, so this proposed design can adapt with fast changing whether conditions In the steady state, the feedforward control loop generates a small amount of control signal to modify the fixed step size signal of the P&O control loop, therefore it results of eliminating the oscillation The working principle of the feedforward control loop is operated based on the fuzzy logic mechanism As we all know that the fuzzy logic control acts as the capability to imitate human thinking There is no need to have mathematical model

The inputs of feedforward are selected as the variations of power and voltage for reducing a number of sensors instead of temperature and irradiation data The

configuration of the Feedforward control loop will be displayed in the Figure 8 Fuzzy Logic mechanism [7] includes four parts: Fuzzification block maps the inputs from a set of sensors (or features of those sensors such as amplitude or spectrum) to values from 0

to 1 using a set of input membership functions (MF) Inference block associates input variables with fuzzy rules and determines the output of the fuzzy controller Mamdani's inference method is used in this case with the max-min composition method Defuzzification block converts the fuzzy output into a crisp value that represents the control action

In this case, two input variables are power variation (ΔVpv) and voltage variation (ΔPpv) The membership functions of the utilized input and output variables for the proposed controller are illustrated in Figure 9, and 10 Figure 11 shows the MFs of the input variables Figure 12 is the MFs of the output (duty cycle compensation ΔDFF).For linguistic variables, P represents big positive, N represents negative, B, and Z are defined as big, and zero, respectively

Figure 8 The proposed scheme for MPPT

Figure 9 The Feedforward based on

Fuzzy logic mechanism

Table 1 Designed rule database

Figure 10 Feedforward explanation

For last index, P, V, and D stand for the power variation, voltage variation, and duty cycle compensation respectively

Each of the input variables ΔPpv and ΔVpv is mapped into five different linguistic values Therefore, there are 25 different rules

Figure 11 Input MFs for power variation (“delP”) and voltage variation (“delV”)

Figure 12 Output MFs for duty cycle “delD” (unit in percent)

The membership functions are designed based on the system understanding of the designer Otherwise some optimization methods can be used such as PSO, GA, etc [10]

IV SIMULATION RESULTS

The PV module used in this project is Ks80m-36 has parameters shown in Table 2 The system is implemented by using Matlab/Simulink for testing the proposed solution before exprementing in experimental setup

Table 2 Utilized PV module parameters at irradiation 1000W/m2, 25oC

Parameters Symbols Values

Voltage at Maximum power point VMPP 18V Current at Maximum power point IMPP 4.45A

Ratio of current/temperature at Isc Ki 0.0017A/oC The proposed scheme and conventional P&O solution will be tested in the same plant under the variation of the whether (irradiation) In order to make the simulation to be like the real conditions, so the system is set to operate at the temperature 10oC, and the irradiation changes like square waveform: 600 W/m2; 400 W/m2; 500 W/m2; 600 W/m2 every 0.5 seconds, and the load is assumed to be 50Ω If the PV module is directly connected to the load, the maximum power that the system can obtain is 6.5W due to these above condition

Figure 13 The output power response due to P&O and proposed P&O+FFC

Both P&O and proposed controller can reach the MPP point, but Figure 13 shows that the output power generated from PV module due to the proposed P&O combining with Feedforward controller (P&O_FFC) is higher and less fluctuating than the conventional P&O controller at the steady state time, furthermore P&O_FFC time response is faster P&O_FFC controller needs 0.066s, while only P&O controller takes 0.097s to seek to MPP and keep in steady state Otherwise the proposed P&O_FFC controller remains some drawbacks such that it causes more flipped states and larger fluctuating amplitude at the transient time

Figure 14.Output voltage of PV module due to P&O, and proposed P&O+FFC

Figure 15 indicates that changing irradiation leads duty cylce to be automatically adjusted to proper values In more detail the duty cyle gets 68% to get the highest power amount corresponding to 600W/m2 irradiation The duty cycle changes to 53% , when the irradiation steps down to 400W/m2 The output voltage of PV module when P&O_FFC acted on the system varies in lower range in comparision to P&O only (see in Figure 14) The larger range of the output voltage of the PV module is a result of the adjustment of duty cycle sending to the Boost converter

Figure 15 Duty cycle sending to Boost converter generated by P&O, and P&O+FFC

V CONCLUSION

In this paper, the P&O_FFC scheme is introduced for limitting the drawbacks from conventional P&O controller Feedforward element based on FFC mechanism is applied to compensate the fluctuating phenomena at the steady state, and to reduce the time response

at the transient time All results indicate that the proposed solution works better than conventional P&O controller Otherwise, the uneffectiveness of this method is the larger fluctuating amplitude at the transient time, and it needs to be demonstrated via the experimental device This work will be soon done in the future

REFERENCES

[1] L V Hartmann, A M N Lima et al., "Using the model of the solar cell for

determining the maximum power point of photovoltaic systems," Proc European

Conference on Power Electronics and Applications pp.1-10, 2007

[2] Edward E Anderson, "Fundamentals for Solar Energy Conversion", Addison Wesley

Pub Co., 1983

[3] D Yogi Goswami, Frank Kreith, Jan F Kreider, "Principles of Solar Engineering",

2nd Edition, Taylor & Francis, 2000, India Reprint, 2003, Chapter 9, Photovoltaics,

pp 411-446

[4] Bennett, T., Zilouchian, A., Messenger, R.: “Photovoltaic model and converter

topology considerations for MPPT purposes”, Sol Energy, 2012, 86, pp 2029–2040

[5] Katherine A Kim and Philip T Krein, “Photovoltaic Converter Module

Configurations for Maximum Power Point Operation”, University of Illinois

Urbana-Champaign Urbana, IL 61801 USA

[6] Femia, N.; Petrone, G.; Spagnuolo, G.; Vitelli, M “Optimization of perturb and

observe maximum power point tracking method” IEEE Trans Power Electron 2005,

20, 963–973

[7] Algazar, M.M.; AL-monier, H.; EL-halim, H.A.; Salem, M.E.E.K "Maximum power

point tracking using fuzzy logic control" Int J Electr Power Energy Syst 2012, 39, 21

[8] Chiu, C.S.; Ouyang, Y.L "Robust maximum power tracking control of uncertain

photovoltaic systems: A unified T-S fuzzy model-based approach" IEEE Trans Control

Syst Technol 2011,19, 1516–1526

[9] Mahmoud A Younis (University Tenaga National), Tamer Khatib (National University of Malaysia), Mushtaq Najeeb (Universiti Tenaga National), A Mohd

Ariffin (University Tenaga National), “An Improved Maximum Power Point Tracking

Controller for PV Systems Using Artificial Neural Network”, ISSN 0033-2097, R 88

NR 3b/2012, Pg 116-121

[10] Goldberg, E (1989) “Genetic Algorithms in Search Optimization & Machine

Learning” Reading, MA: Addison-Wesley

TÓM TẮT

ỨNG DỤNG BỘ ĐIỀU KHIỂN TRUYỀN THẲNG KẾT HỢP VỚI

THUẬT GIẢI NHIỄU LOẠN VÀ QUAN SÁT ĐỂ NÂNG CAO CHẤT LƯỢNG

CHO HỆ THỐNG DÒ ĐIỂM CÔNG SUẤT CỰC ĐẠI

Phương pháp dò điểm công suất cực đại (MPPT) đóng vai trò rất quan trọng trong các hệ thống pin năng lượng mặt trời giúp hệ thống làm việc đạt hiệu suất cao nhất Hiện nay, có rất nhiều thuật toán ứng dụng cho việc dò điểm công suất cực đại

đã được đề xuất và áp dụng, có thể kể đến các thuật toán như là: P&O, INC, Fuzzy, Neural network, Trong đó, phương pháp dò điểm công suất cực đại sử dụng thuật toán nhiễu loạn và quan sát P&O được ứng dụng rộng rãi nhất do tính đơn giản và

dễ dàng thực hiện Tuy nhiên, việc tìm điểm công suất cực đại sử dụng phương pháp này có độ chính xác chưa cao, độ dao động lớn Để hạn chế nhược điểm nêu trên, trong bài báo này nhóm tác giả đề xuất cấu trúc điều khiển kết hợp P&O với điều khiển truyền thẳng (Feedforward) được xác định bởi cơ cấu chỉnh định mờ để nâng cao độ chính xác điểm công suất cực đại, giảm độ dao động so với việc chỉ dùng P&O hoặc FLC riêng rẽ Thuật toán và cấu trúc điều khiển trước tiên sẽ được kiểm chứng thông qua mô hình động học của pin mặt trời và bộ biến đổi DC-DC thực hiện trên Matlab/Simulink, tiếp đến là kiểm chứng thông qua hệ thống thực nghiệm

Từ khóa: Pin mặt trời; Bộ biến đổi DC-DC; PWM; P&O; INC; FLC; NN; Điều khiển truyền thẳng

(Feedforward)

Received date, 02 nd May, 2017

Revised manuscript, 10 th June, 2017

Published, 20 th July, 2017

Author affiliations:

Thai Nguyen University of Technology;

*

Email: dangvanhuyen@tnut.edu.vn