You are on page 1of 10

IEEE TRANSACTIONS ON SUSTAINABLE ENERGY, VOL. 4, NO.

1, JANUARY 2013

11

Quasi-Z-Source Inverter-Based Photovoltaic


Generation System With Maximum Power
Tracking Control Using ANFIS
Haitham Abu-Rub, Senior Member, IEEE, Atif Iqbal, Senior Member, IEEE, Sk. Moin Ahmed, Member, IEEE,
Fang Z. Peng, Fellow, IEEE, Yuan Li, Member, IEEE, and Ge Baoming, Member, IEEE

AbstractThe paper proposes an artificial-intelligence-based


solution to interface and deliver maximum power from a photovoltaic (PV) power generating system in standalone operation.
The interface between the PV dc source and the load is accomplished by a quasi-Z-source inverter (qZSI). The maximum power
delivery to the load is ensured by an adaptive neuro-fuzzy inference system (ANFIS) based on maximum power point tracking
(MPPT). The proposed ANFIS-based MPPT offers an extremely
fast dynamic response with high accuracy. The closed-loop control of the qZSI regulates the shoot through duty ratio and the
modulation index to effectively control the injected power and
maintain the stringent voltage, current, and frequency conditions.
The proposed technique is tested for isolated load conditions.
Simulation and experimental approaches are used to validate the
proposed scheme.
Index TermsDCAC power conversion, nonshoot-through
state, pulsewidth modulated inverters, quasi-Z-source inverter
(qZSI), shoot-through state, solar power generation.

I. INTRODUCTION

HE rapidly increasing environmental degradation across


the globe is posing a major challenge to develop commercially feasible alternative sources of electrical energy generation. Thus, a huge research effort is being conducted worldwide to come up with a solution in developing an environmentally benign and long-term sustainable solution in electric power
generation. The major players in renewable energy generation
are photovoltaics (PV), wind farms, fuel cell, and biomass [1].

Manuscript received April 12, 2011; revised January 07, 2012; accepted
March 10, 2012. Date of publication May 25, 2012; date of current version
December 12, 2012. This work was supported by NPRP Grant [09-233-2-096]
from the Qatar National Research Fund (a member of Qatar Foundation). The
statements made herein are solely the responsibility of the authors.
H. Abu-Rub and Sk. Moin Ahmed are with the Department of Electrical and
Computer Engineering, Texas A&M University at Qatar, Doha, 23874, Qatar
(e-mail: haitham.abu-rub@qatar.tamu.edu; moin.sk@qatar.tamu.edu).
A. Iqbal is with the Department of Electrical Engineering, Qatar University,
Doha, 2713, Qatar, and also with the Department of Electrical Engineering, Aligarh Muslim University, Aligarh, 202002, India (e-mail: atif.iqbal@qu.edu.qa).
F. Z. Peng is with the Department of Electrical Engineering, Michigan State
University, East Lansing, MI 48824 USA (e-mail: fzpeng@egr.msu.edu).
Y. Li was with the Department of Electrical Engineering, Michigan State
University, East Lansing, MI 48824 USA. She is now with the Department of
Electrical and Information Engineering, Sichuan University, Chengdu 610064,
China (e-mail: yli@msu.edu).
G. Baoming is with the Department of Electrical Engineering, Beijing Jiaotong University, Beijing, 100044, China (e-mail: bm-ge@263.net).
Color versions of one or more of the figures in this paper are available online
at http://ieeexplore.ieee.org.
Digital Object Identifier 10.1109/TSTE.2012.2196059

These distributed power generation sources are widely accepted


for microgrid applications. However, the reliability of the microgrid relies upon the interfacing power converter [2]. Thus
the proper power regulation from the interfacing power converter will ensure a stable and reliable microgrid system [3].
Thus this paper focuses on the proposal of a new class of interfacing inverter, the quasi-Z-source inverter (qZSI) for off-grid
applications.
There are several power converter topologies employed in PV
systems; however, they differ by several characteristics: twostage or single-stage, with transformer or transformerless, and
with a two-level or multilevel inverter [4][14]. Single-stage inverters are becoming more attractive in comparison to two-stage
models due to their compactness, low cost, and reliability [15].
However, the conventional inverter has to be oversized to cope
with the wide PV array voltage changes because a PV panel
presents low output voltage with a wide range of variation based
on irradiation and temperature, usually at a range of 1 : 2. To interface the low voltage output of an inverter to the grid, a bulky
low-frequency transformer is necessary at the cost of a large
size, decrease in efficiency, loud acoustic noise, and high cost
[16]. The two-stage inverter applies a boost dc/dc converter instead of a transformer in order to minimize the required KVA
rating of the inverter and boost the wide range of voltage to a
constant desired value. Unfortunately, the switch in the dc/dc
converter becomes the cost and efficiency killer of the system.
For safety reasons, some PV systems have a galvanic isolation,
either in the dc/dc boost converter using a high-frequency transformer, or in the ac output side of a line frequency transformer.
Both of these added galvanic isolations increase the cost and
size of the whole system, and decrease the overall efficiency.
Transformerless topologies are especially deserving attention
because of their higher efficiency, smaller size and weight, and
a lower price for the PV system [16]. The Z-source inverter
(ZSI), as a single-stage power converter with a step-up/down
function, allows a wide range of PV voltages, and has been reported in applications in PV systems [17][19]. It can handle
the PV dc voltage variation in a wide range without overrating
the inverter, as well as implement voltage boost and inversion
simultaneously in a single power conversion stage, thus minimizing system cost and reducing component count and cost, and
improving the reliability. Recently proposed qZSIs have some
new attractive advantages that are more suitable for application
in PV systems. This will make the PV system much simpler
and lower its cost because the qZSI: 1) draws a constant current
from the PV panel, thus no need for extra filtering capacitors;

1949-3029/$31.00 2012 IEEE

12

IEEE TRANSACTIONS ON SUSTAINABLE ENERGY, VOL. 4, NO. 1, JANUARY 2013

2) features lower component (capacitor) rating; and 3) reduces


switching ripples to the PV panels [20][23]. This paper employed qZSI for interfacing the PV generation system for the
isolated load condition.
Artificial intelligence (AI)-based methods are increasingly
used in renewable energy systems [24][30] due to the flexible
nature of the control offered by such techniques. The AI techniques are highly successful in nonlinear systems due to the fact
that once properly trained they can interpolate and extrapolate
the random data with high accuracy. A review on the application of AI techniques in renewable energy generation system is
presented in [24]. Some applications of artificial neural network
in PV are presented in [25][28], and the use of fuzzy logic is
available in [29] and [30], while the adaptive neuro-fuzzy inference system (ANFIS) is utilized in [31]. In [31], voltage and
current are taken as the input to the ANFIS controller. In contrast, the presented technique utilizes the weather information
as the input to the ANFIS.
The neural network is a powerful technique for mapping
inputoutput nonlinear function; however, it lacks the heuristic
sense and it works as a black box [32]. On the other hand,
fuzzy logic has the capability of transforming heuristic and
linguistic terms into numerical values through fuzzy rules and
membership functions [33]. It also provides the heuristic output
by quantifying the actual numerical data into heuristic and
linguistic terms. However, the shortcoming of fuzzy computation is obtaining correct fuzzy rules and membership functions
which heavily rely on the prior knowledge of the system. The
ANFIS integrates the neural network and fuzzy logic, thus
this synergy offers the most powerful artificial intelligence
technique [34]. This paper thus uses ANFIS techniques to
determine the maximum power capability of a PV module
for variable solar irradiance and temperature conditions. The
other maximum power point tracking (MPPT) algorithms such
as perturb and observe, incremental conductance and their
improvements suffer from drawbacks such as oscillations at the
operating point and lack in fast dynamic response. The speed
of the algorithm in locating the correct operating point of PV
is a crucial factor especially when operating in grid interactive
mode. The proposed technique of using ANFIS-based MPPT
offers highly precise and fast control with robust operation and
is highly suitable for microgrid application in PV generation
systems.
This paper is organized in six different sections. Section II
discusses the basic working principle of a qZSI. The principle
and requirement of the proposed control system is laid down in
Section III. Section IV explains the application of ANFIS. The
simulated results are presented in Section V. The experimental
validation is illustrated in Section VI.

II. QUASI-Z-SOURCE INVERTER-CIRCUIT TOPOLOGY


AND CONTROL
Quasi-Z-source inverter (qZSI) topology has been introduced
recently to overcome some of the shortcomings of the Z-source
inverter. The qZSI offers several advantages over the Z-source
inverter such as continuous input current, reduced components
rating, higher reliability, and simple control strategy [20][23].

Fig. 1. The qZSI topology.

This topology of the inverter is identified as one of the most suitable power conditioning interface between the PV generation
system and the grid. This paper presents an improved power circuit topology of qZSI where one capacitor of the quasi-Z-source
network is replaced by storage batteries, as shown in Fig. 1, thus
flexible power conditioning functionalities can be achieved. In
the proposed structure, the load can be isolated or it can be a
micro grid or a full scale grid.
In order to capture the maximum solar energy, MPPT is necfrom the PV panels
essary to draw the maximum power
in PV applications, which is commonly implemented by regulating the PV voltage to follow a time-variant referencethe
. The value of
voltage of the maximum power point
is continuously tracked by certain MPPT algorithms, such as
the perturb and observe (P&O), incremental conductance (IncCond), or ANFIS presented in this paper. On the other hand, the
output power is another concern of the PV system. For an isolated load condition (standalone PV system), the output voltage
of the qZSI is regulated and the output power is determined by
corresponding load demands. While for a grid-interactive condition (grid-tied PV system), the output power of the qZSI can
be regulated by controlling the current injected to the grid accordingly. Notice that no matter the case, the input power from
PV panels, the output power to loads (isolated load or grid),
along with the power absorbed or released by energy storage
battery in the proposed topology, should be matched to maintain
a stable and sustainable operating approach, and it is the battery
that provides an energy buffer zone for both input and output
sides of the PV system. Given the condition that the battery is
within its valid charge or discharge status, MPPT can be accom
to
plished by sending the captured extra power
the battery, or the output power can be maintained for period of

from the battime by extracting absent power


tery. The second case is of essential importance for large-scale
PV systems interfacing the power grid, where stable and sustainable energy supply is always demanded, while contrastively
PV cells output power varies accordingly with temperature and
solar irradiation.
With the proposed topology in this paper, the state-of-charge
(SOC) of the battery is taken into consideration with the following concerns: 1) To choose a proper shoot-through duty
to
, and the battery is charged
ratio for regulating
or discharged through a desired current, which is determined
by the power difference of input and output together with the
battery voltage, and also is limited by the maximum acceptable
value according to the battery. 2) Once the battery has been
adequately charged, the voltage of the battery in the circuit
should be regulated as little current (theoretically no current)

ABU-RUB et al.: qZSI-BASED PV GENERATION SYSTEM WITH MAXIMUM POWER TRACKING CONTROL USING ANFIS

13

The average power relationship of the system during one


switching cycle is derived as follows:
The PV output power could be calculated by
(5)
The power of battery can be calculated by
(6)
The load power can be expressed by

(7a)
The power relationship in the system can be derived as
(7b)
Fig. 2. Equivalent circuit of the qZSI with battery. (a) Nonshoot-through state.
(b) Shoot-through state.

charging the battery any longer. Notice in this situation MPPT


may not be achieved because the energy buffer zone is unable
to absorb extra power. 3) If the battery is exhausted where the
battery voltage drops below one threshold and there is still a
shortage of power from PV panels, load shading needs to be
executed. 4) For all the above-mentioned statuses, a sufficient
should be kept for a valid output (as in
dc bus voltage
needs to be kept
another point of view, capacitor voltage
above a certain value). In order to satisfy this requirement,
MPPT may be sacrificed for the sake of an adequate voltage
boost.
The whole operation of a qZSI is divided into two distinct
modes called the shoot-through and nonshoot-through or active
mode and their equivalent representation is given in Fig. 2.
From Fig. 2(a), the inverter operated during the interval of the
and in Fig. 2(b), the inverter opernonshoot-through states
ated during the interval of the shoot-through states , one can
get

(1)
(2)
The dc-link peak voltage
two capacitor voltages,

could be derived from the sum of


and
, as
(3)

The relationship between two inductor currents


current flowing into the battery is shown as

and the

(4)

There are three operating states of the battery. When the battery
,
,
, and
is charging, thus
; when the battery is discharging,
,
,
, and
; no charging and discharging,
,
,
, and
.
Hence, the qZSI can buck or boost input dc voltage, it can
handle wide variation of the input voltage, particularly for the
PV system, and produce a desired voltage for the isolated load or
for the grid in a single stage. This feature results in the reduced
number of switches involved in the power electronics of the PV
system and, therefore, the reduced cost and the improved system
efficiency and reliability. When the solar irradiance is low and
the PV panel produces low voltage, the qZSI boosts the voltage,
which helps to avoid redundant PV panels for higher dc voltage
or unessential inverter overrating. As mentioned previously, it
is able to handle the shoot-through state; therefore, it is more
reliable than the conventional VSI. Additionally, for the qZSI,
there is a common dc rail between the source and inverter, which
is easier to assemble and causes less EMI problems.
III. PROPOSED CONTROL STRUCTURE
It is assumed that the PV power generating system is feeding
an isolated load. The objectives to be achieved by the proposed
control system are
1) Maximum power point tracking.
2) Desired stable output power to the isolated ac load. The
output power of the inverter should be controllable and
adjustable on the basis of users demand in case of the
isolated load conditions.
The MPPT is achieved using ANFIS. It is known from literature that the PV system needs to be operated at a specified
voltage for extracting maximum available power [35], [36]. The
ANFIS is trained for giving voltage output crisp value
corresponding to the maximum power delivery from the PV
panels. The inputs to the ANFIS are given as environmental conditions, i.e., the solar irradiance and temperature.
There are two control variables for this qZSI control system,
i.e., the shoot-through duty
and modulation index
.

14

IEEE TRANSACTIONS ON SUSTAINABLE ENERGY, VOL. 4, NO. 1, JANUARY 2013

Fig. 4. Training error versus epochs for the ANFIS.

for further processing or implementation of real-time control


system using the ANFIS controller. The schematic outline for
the ANFIS controller is shown in Fig. 3(b).
IV. APPLICATION OF ANFIS FOR MPPT TRACKING

Fig. 3. (a) Proposed control scheme with ANFIS-based MPPT. (b) Schematic
outline for ANFIS controller.

Both control variables should cooperate to achieve the above


set goals. In the proposed control strategy, a closed-loop control
of the input voltage
is combined with the trained ANFIS
to implement the MPPT control, as shown in Fig. 3. A proportional and integral regulator with feedforward will adjust the
shoot-through duty of the qZSI. In general, the battery voltage
depends on its SOC, instead of its current, and has a little change
in a suitable range of the SOC.
At constant temperature, the change of solar irradiation will
result in a great change of PV current at the maximum power
point (MPP), when compared to the resultant change of PV
voltage. The MPPT control could ensure a stable peak dc-link
voltage with little variation at a constant temperature. On the
other hand, the change of temperature will result in a great
change of PV voltage at the MPP, when compared to the resultant change of PV current, which will make the peak dc-link
voltage change greatly.
In the load side, the closed-loop control is employed to keep
the output voltage magnitude and frequency constant regardless
of the change in the input conditions, as shown in Fig. 3(a).
The modulation signals are adjusted to ensure a constant output
voltage, thus the change of load impedance will cause a change
of output power in the isolated loading conditions, where the
storage batteries are employed to meet the load demand.
The solar irradiance is measured using standard industrial
solar pyranometer and the temperature and other weather data
are collected using weather transmitters arrangements. The outputs of the solar irradiance and temperature transducers are current/voltage signals which are logged in real time using standard data loggers. These data are then be transferred to the PC

To validate the proposed control scheme, the simulated model


is developed in Matlab/Simulink for the whole system. Each
model is elaborated on next. The PV cell temperature varies
from 10 C to 70 C in a step of 6 C and the solar irradiance varies from 50 to 1000 W/m in a step of 50 W/m . By
varying these two environmental factors a set of data is generated in simulation. Two hundred sets of obtained data are then
used to train the ANFIS network for the purpose of MPPT. The
training is done offline using Matlab tool box. The network is
trained for 30 000 epochs. The target error is set to 1.1% and the
training waveform is depicted in Fig. 4. The overall neuro-fuzzy
structure shown in Fig. 5 is a five-layer network. The structure
shows two inputs of the solar irradiance and the cell temperature, which is translated into appropriate membership functions, three functions for the solar irradiance in Fig. 6 and three
functions for temperature in Fig. 7. These membership functions are generated by the ANFIS controller based on the prior
knowledge obtained from the training data set. The membership
functions shape varies during the training stage and the final
shape obtained after the completion of the training is shown
in Figs. 6 and 7. They are termed as low, medium, and
high. The common intersection areas between the low and
medium solar irradiance and the low and high irradiance are
nearly 75% and 50%, respectively. The common intersection
domain between the medium and high solar irradiance is nearly
80%. In the temperature membership function, the intersection
between low and medium temperature is nearly 70%, and almost no common area between low and high temperatures. The
fuzzy rules are depicted graphically in Fig. 8 for a specific case
of 40 C and 525-W/m irradiance. The corresponding MPP
voltages are shown.
The rule base depicts the relationship and mapping between
the input and output membership functions. One particular situation is shown in Fig. 8 when the temperature is at 40 C and
the solar irradiance is 525 W/m . By varying the slider on the
figure, all the conditions can be accessed. It can be seen that
the temperature varies from 10 C to 70 C, the solar irradiance
varies from 50 to 1000 W/m , and correspondingly, the maximum power point voltage varies as shown in the last column.

ABU-RUB et al.: qZSI-BASED PV GENERATION SYSTEM WITH MAXIMUM POWER TRACKING CONTROL USING ANFIS

15

Fig. 5. ANFIS-based MPPT structure.

Fig. 8. Rule base of ANFIS controller.

Fig. 6. Membership function of solar irradiance.

Fig. 9. Surface view created by ANFIS.

Fig. 7. Membership function of PV cell temperature.

There are nine rules that can follow, and more filled cells means
high values and the blank or less filled cells represents low
values; e.g., rule 8 can be read as if temperature input is low
(follow membership function low, Fig. 6) and the solar irradiance is medium (follow membership function medium, Fig. 7)
then the maximum power point voltage (output of ANFIS controller) is 15.5 V. The rulers (the vertical red line) shown in the
temperature and irradiance can be moved to check the rules for
other operating conditions.
The variation of the MPP voltage
with the changes
of the PV cell temperature and solar irradiance is shown in the
surface plot of Fig. 9. The surface depicts the typical behavior.
The proposed ANFIS-based MPPT is more stable and faster
than the traditional P&O-based MPPT method. This can be observed from Fig. 10. The simulation is done for an initial solar
radiation of 450 W/m . There is an irradiation step change at

0.4 s to 700 W/m and again at 0.7 s to 450 W/m . The temperature is randomly changed for a small amount for the whole
period.
V. SIMULATED RESULTS FOR ISOLATED LOAD CONDITION
Considering a real situation, the solar irradiance varies from
a certain minimum value to the maximum value and then goes
down to another minimum value. A similar pattern is also suitable for the PV cell temperature. To simulate a real time scenario, the solar irradiance and temperature is varied accordingly
as shown in Fig. 11. The solar irradiance is varied from nearly
300 to 400 W/m with a peak value of 1 kW/m and with a
10% ripple. The temperature is varied from 25 C to 40 C and
back to 25 C. The time scale is taken as 1.5 s due to the limited
memory of the digital computer.
Three different loads are connected at the output of the qZSI,
namely, the R-L load of 8 kW and 0.30 kVar applied from 0 to
0.4 s, followed by a resistive load of 13 kW from 0.4 to 1 s, and
finally another R-L load of 10.2 kW and 0.3 kVar from 1 to 1.5 s.

16

Fig. 10. PV output voltage


method.

IEEE TRANSACTIONS ON SUSTAINABLE ENERGY, VOL. 4, NO. 1, JANUARY 2013

with ANFIS and P&O-based MPPT

Fig. 11. Variation curves of solar irradiance and temperature.

A lithiumion battery pack with the nominal voltage of 100 V,


and 5 Ah with the discharge characteristic shown in Fig. 12, is
used as the energy storage. The small ampere-hour of the battery
is used due to the same reason of the limited computer memory.
The initial SOC of the battery pack is taken as 50%. The other
circuit parameters are
H,
F,
mH,
F, and carrier frequency is kept at
10 kHz. The line-to-line voltage across the load is kept at 208 V
rms with a frequency of 60 Hz as this is standard utility supply
voltage and frequency in the United States. The qZSI is modulated using a carrier-based scheme with upper and lower bound
voltages to control the shoot through states. The simulated results are shown in Fig. 13. It is evident from Fig. 13(a) that the
ANFIS outputs the MPP voltage of the PV array, and the terminal voltage of the PV array follows it very well for the entire duration to ensure the maximum output power from the PV
array, as shown in Fig. 13(b). The maximum power falls to a

Fig. 12. Discharge characteristic of battery pack.

minimum level and then rises again. This shows the effect of
variation in solar irradiance and cell temperature. The output
power of the PV array exceeds the load requirement from
to
s, causing the battery charge to maintain the power
balance. The battery current is negative and the battery terminal
voltage is high in Fig. 13(c). After 0.4 s, the battery floats as the
load increases and the PV array directly feeds the load. The situation changes again at
s, when the load requirement is
higher than the power produced by the PV array, as a result that
the battery also contributes the power to the load and hence it
discharges in Fig. 13(c).
The input current
changes slightly compared to the current of inductor
which varies greatly, depending on the state
of battery working. When the battery is charging,
;
when the battery is discharging,
; when there is no
charging and discharging,
, as shown in Fig. 13(d).
The load current is also shown in the same figure to present its
change in the magnitude due to the change of load conditions.
The output voltage of the qZSI is shown in Fig. 13(e), which
is evident that the voltage magnitude and the frequency remain
constant regardless of the loading conditions. Fig. 13(e) shows
its transient change in the voltage magnitude due to load change,
but it is corrected quickly by the PI controller. This shows the
successful implementation of the control strategy.
For the qZSI with the battery shown in Fig. 1, the dc-link
voltage depends on the value of input voltage due to the almost
constant voltage of the battery-based energy storage. The average duty only depends on the input voltage, regardless of
input power and output power. The battery will be charged if
the input power is larger than the output power, and discharged
when the input power is less than the output power. The peak
dc-link voltage changes if the input voltage varies. If a constant
peak dc-link voltage is desired, the designed control system in
this work is more suitable for application to the change of solar
irradiation, rather than the change of temperature. However, the
simulation result is shown for the extreme case of large variation in both solar irradiance and the cell temperature.

ABU-RUB et al.: qZSI-BASED PV GENERATION SYSTEM WITH MAXIMUM POWER TRACKING CONTROL USING ANFIS

17

Fig. 13. Simulated results for isolated load condition.

VI. EXPERIMENTAL INVESTIGATION


A prototype of the qZSI has been built on the basis of the
original ZSI prototype in the laboratory. Experiments of the

qZSI-based PV system have been done in isolated load condition to verify the concept and theoretical analysis presented
in the earlier sections. The controller of the prototype qZSI has
been built using a TMS320LF2407 DSP-based universal dig-

18

IEEE TRANSACTIONS ON SUSTAINABLE ENERGY, VOL. 4, NO. 1, JANUARY 2013

Fig. 16. Experimental results for load change (25 A/div, 140 ms/div).

experiment are as follows:

Fig. 14. Experimental setup for qZSI.

, system output frequency


Hz, and switching frequency
kHz, which is
also the sampling frequency of the controller.
Simple boost control was employed when
V,
where
. The qZSI works in boost conversion mode.
Voltage of capacitor C1 can be calculated as

The output line-to-line voltage was regulated to 104 V, with


output power 2.16 kW (resistive load, 5 ohms per phase), as
shown in Fig. 15(a). With a light load (resistive load, 10 ohms
per phase), the input voltage increased to 190 V, which is
adequate to provide 104-V output. Therefore, there is no
shoot-through involved, capacitor voltage
is the same
as the input voltage, and the output line-to-line voltage was
regulated to 104 V as well, with output power 1.08 kW, as
shown in Fig. 15(b).
The test results for the isolated load condition on the inductor
currents and load current with the change in the load condition
are depicted in Fig. 16. Initially the load is 1 KW, then increased
to 1.7 KW at 0.38 s, and finally increased to 2 KW at 0.8 s. The
input current
keeps constant as the shoot through duty ratio
remains constant during the whole operation. On the other side,
the inductor current
varies greatly because it depends on
the state of the battery working. When the battery is charging,
; when there is no charging and discharging,
; when the battery is discharging,
, as shown in
Fig. 16. It is evident that the inductor current changes with the
there is a change in the loading condition.
VII. CONCLUSION
Fig. 15. Experimental results of qZSI in load isolated condition, where
V,
V,
, output power 2.16 kW; and
(a)
V,
V,
, output power 1.08 kW.
(b)

ital-control board. The experimental setup for the qZSI is shown


in Fig. 14. The test results are provided for isolated load condition, where
V, output line-to-line voltage
V. The qZSI in the experiment is
operated with no dead time, and there are no snubbers on either
the individual switches or the dc link. System parameters in the

The paper proposes ANFIS-based PV power generation


system operating in standalone mode. The interface stage
between the generation source and the load is accomplished
by a qZSI. The shoot through duty ratio is controlled using
ANFIS to harness the maximum available power from the PV
system. The load side voltage and frequency is regulated by
controlling the modulation index of the interface of the qZSI.
Thus simultaneous control of shoot through duty ratio and
the modulation index ensure the control objectives achieved.
Simulation and experimental results are provided to verify
the proposed control approach. The grid tied operation for the
proposed scheme will be reported in the future.

ABU-RUB et al.: qZSI-BASED PV GENERATION SYSTEM WITH MAXIMUM POWER TRACKING CONTROL USING ANFIS

REFERENCES
[1] M. Boxwell, Solar Electricity Handbook, 3rd ed. Warwickshire,
U.K.: Greenstream Publishing, Apr. 2010.
[2] On-line resource Jun. 2010 [Online]. Available: http://www.
greenm3.com/2010/03/pike-research-forecasts-2010-to-2015-microgrid-growth-from-100-to-2000.html
[3] B. Azzopardi and J. Mutale, Smart integration of future grid-connected PV systems, in Proc. 34th IEEE Photovoltaic Specialist Conf.
(PVSC), 2009, pp. 23642369.
[4] M. M. Casaro and D. C. Martins, Grid-connected PV system: Introduction to behavior matching, in Proc. IEEE Power Electronics Specialist Conf., 2008, pp. 951956.
[5] S. Busquets-Monge, J. Rocabert, P. Rodrguez, S. Alepuz, and J. Bordonau, Multilevel diode-clamped converter for photovoltaic generators with independent voltage control of each solar array, IEEE Trans.
Ind. Electron., vol. 55, no. 7, pp. 27132723, Jul. 2008.
[6] Y. Huang, F. Z. Peng, J. Wang, and D. Yoo, Survey of the power conditioning system for PV power generation, in Proc. 37th IEEE Power
Electronics Specialists Conf. (PESC06), Jeju, Jun. 1822, 2006, pp.
16.
[7] Y. Huang, M. Shen, F. Z. Peng, and J. Wang, Z-source inverter for
residential photovoltaic systems, IEEE Trans. Power Electron., vol.
21, no. 6, pp. 17761782, Nov. 2006.
[8] H. Dehbonei, S. R. Lee, and H. Nehrir, Direct energy transfer for high
efficiency photovoltaic energy systems part I: Concepts and hypothesis, IEEE Trans. Aerosp. Electron. Syst., vol. 45, no. 1, pp. 3145,
Jan. 2009.
[9] H. Patel and V. Agarwal, A single-stage single-phase transformer-less
doubly grounded grid-connected PV interface, IEEE Trans. Energy
Convers., vol. 24, no. 1, pp. 93101, Mar. 2009.
[10] J. Selvaraj and N. A. Rahim, Multilevel inverter for grid-connected
PV system employing digital PI controllers, IEEE Trans. Ind. Electron., vol. 56, no. 1, pp. 149158, Jan. 2009.
[11] G. R. Walker and P. C. Senia, Cascaded DC-DC converter connection
of photovoltaic modules, IEEE Trans. Power Electron., vol. 19, no. 4,
pp. 11301139, Jul. 2004.
[12] L. Quan and P. Wolfs, A review of the single phase photovoltaic
module integrated converter topologies with three dc link configurations, IEEE Trans. Power Electron., vol. 23, no. 3, pp. 13201333,
May 2008.
[13] B. Yang, L. Wuhua, Y. Zhao, and H. Xianging, Design and analysis
of a grid-connected photovoltaic power system, IEEE Trans. Power
Electron., vol. 25, no. 4, pp. 9921000, Apr. 2010.
[14] B. M. T. Ho and S.-H. Cheng, An integrated inverter with maximum
power tracking for grid-connected PV systems, IEEE Trans. Power
Electron., vol. 20, no. 4, pp. 953962, Jul. 2005.
[15] H. Dehbonei, S. R. Lee, and H. Nehrir, Direct energy transfer for high
efficiency photovoltaic energy systems part I: Concepts and hypothesis, IEEE Trans. Aerosp. Electron. Syst., vol. 45, no. 1, pp. 3145,
Jan. 2009.
[16] H. Patel and V. Agarwal, A single-stage single-phase transformer-less
doubly grounded grid-connected PV interface, IEEE Trans. Energy
Convers., vol. 24, no. 1, pp. 93101, Mar. 2009.
[17] P. Xu, X. Zhang, C.-W. Zhang, R.-X. Cao, and L. Chang, Study of
Z-source inverter for grid-connected PV systems, in Proc. 37th IEEE
Power Electronics Specialist Conf. (PESC06), Jun. 2006, pp. 15.
[18] R. Badin, Y. Huang, F. Z. Peng, and H. G. Kim, Grid interconnected
Z-source PV system, in Proc. IEEE Power Elect. Specialist Conf.
(PESC), Oralndo, FL, Jun. 2007, pp. 23282333.
[19] F. Z. Peng, Z-Source inverter, IEEE Trans. Ind. Electron., vol. 39,
no. 2, pp. 504510, Mar./Apr. 2003.
[20] J. Anderson and F. Z. Peng, A class of quasi-Z-source inverters, in
Proc. IEEE Industry Applications Society Annual Meeting (IAS08),
Edmonton, AB, Canada, Oct. 59, 2008, pp. 17.
[21] Y. Li, J. Anderson, F. Z. Peng, and D. Liu, Quasi-Z-source inverter for
photovoltaic power generation systems, in Proc. 24th Ann. IEEE Applied Power Electronics Conf. Exposition (APEC2009), Washington,
DC, Feb. 1519, 2009, pp. 918924.
[22] J. Park, H. Kim, E. Nho, T. Chun, and J. Choi, Grid-connected PV
system using a quasi-Z-source inverter, in Proc. 24th Ann. IEEE Applied Power Electronics Conf. Exposition (APEC2009), Washington,
DC, Feb. 15-19, 2009, pp. 925929.

19

[23] J.-H. Park, H.-G. Kim, E.-C. Nho, and T.-W. Chun, Power conmditioning system for a grid connected PV power generation using a quasi
Z source inverter, J. Power Electron., vol. 10, no. 1, pp. 7984, Jan.
2010.
[24] A. Mellit and S. A. Kalogirou, Artificial intelligence techniques for
photovoltaic applications: A review, Prog. Energy Combustion Sci.,
vol. 34, pp. 574632, 2008.
[25] T. Hiyama and K. Kitabayashi, Neural network based estimation of
maximum power generation from PV module using environmental information, IEEE Trans. Energy Convers., vol. 12, no. 3, pp. 241247,
Sep. 1997.
[26] A. de Medeiros Torres, F. L. M. Antunes, and F. S. dos Reis, An artificial neural network-based real time maximum power tracking controller for connecting a PV system to the grid, in Proc. IEEE 24th Ann.
Conf. Industrial Electronics Society, 1998, vol. 1, pp. 554558.
[27] A. AI-Amoudi and L. Zhang, Application of radial basis function
networks for solar-array modelling and maximum power-point prediction, Proc. Inst. Elect. Eng., Generation, Transmission, Distribution,
vol. 147, no. 5, pp. 310316, Sep. 2000.
[28] S. Premrudeepreechacham and N. Patanapirom, Solar-array modelling and maximum power point tracking using neural networks, in
Proc. IEEE Bologna Power Tech. Conf., Bologna, Italy, Jun. 2326,
2003.
[29] T. L. Kottas, Y. S. Boutalis, and A. D. Karlis, New maximum power
point tracker for PV arrays using fuzzy controller in close cooperation
with fuzzy cognitive networks, IEEE Trans. Energy Convers., vol. 21,
no. 3, pp. 793803, Sep. 2006.
[30] F. Ansari, S. Chaterjee, and A. Iqbal, Fuzzy logic control scheme for
solar photo voltaic system for maximum power point tracker, Int. J.
Sustain. Energy, vol. 29, no. 4, pp. 245255, Apr. 2010.
[31] A. M. S. Aldobhani and R. John, Maximum power point tracking of
PV system using ANFIS prediction and fuzzy logic tracking, in Proc.
Int. Multiconf. Engineeris and Computer Scientists (IMECS), Hong
Kong, Mar. 1921, 2008, vol. II, CD-ROM.
[32] S. Haykin, Neural NetworkA Comprhensive Foundation, 2nd ed.
New York: Practice-Hall, 1999.
[33] G. Chen and T. T. Pham, Introduction to Fuzzy Sets, Fuzzy Logic, and
Fuzzy Control Systems. Boca Raton, FL: CRC Press, 2000.
[34] J. S. R. Jang, ANFIS: Adaptive network based fuzzy inference
system, IEEE Trans. Syst., Man., Cybern., vol. 23, no. 3, pp.
665685, May/Jun. 1993.
[35] T. Esram and P. L. Chapman, Comparison of photovoltaic array maximum power point tracking techniques, IEEE Trans. Energy Convers.,
vol. 22, no. 2, pp. 439449, Jun. 2007.
[36] R. A. Messenger and J. Ventre, Photovoltaic Systems Engineering, 2nd
ed. Hoboken, NJ: Wiley Interscience, 2003.

Haitham Abu-Rub (M99SM07) received the


M.Sc. degree in electrical engineering from Gdynia
Marine Academy in Gdansk, Poland, in 1990,
and the Ph.D. degree from Gdansk University of
Technology, in 1995.
He was assistant and associate professor at Birzeit
University in Birzeit, Palestine, for eight years. For
four years, he acted as the chairman of the Electrical
Engineering Department at the same university.
He is currently a full professor with Texas A&M
University at Qatar. His main research interests are
energy conversion.
Dr. Abu-Rub has earned many prestigious, international awards, such as
the American Fulbright Scholarship (at Texas A&M University), the German
Alexander von Humboldt Fellowship (at Wuppertal University, Wuppertal,
Germany), the German DAAD Scholarship (at Bochum University, Bochum,
Germany), the British Royal Society Scholarship (at Southampton University,
Southampton, U.K.), and Best Researcher of the Year 2012 award from
Texas A&M University at Qatar. He has published/accepted more than 130
journal and conference papers and supervised many renewable energy research
projects.

20

Atif Iqbal (M09SM10) received the B.Sc. and


M.Sc. degrees in electrical engineering, in 1991
and 1996, respectively, from the Aligarh Muslim
University, Aligarh, India, and the Ph.D. degree
in 2006 from Liverpool John Moores University,
Liverpool, U.K.
He has been employed as Lecturer in the Department of Electrical Engineering, Aligarh Muslim University, Aligarh, since 1991, and is currently working
as a Associate in the same university. He is on academic assignment and working at Qatar University,
Qatar. His principal area of research interest is power electronics and multiphase machine drives.
Dr. Iqbal is recipient of the Maulana Tufail Ahmad Gold Medal for standing
first at B.Sc. Engg. Exams in 1991, AMU and research fellowship from EPSRC,
U.K., for pursuing Ph.D. studies. He has published more than 150 peer reviewed
journal/conference papers. He is on the editorial board of several prestigious
journals.

Sk. Moin Ahmed (S10M12) was born in


Hooghly, West Bengal, India, in 1983. He received
the B.Tech. and M.Tech. degrees from Aligarh
Muslim University (AMU), Aligarh, India, in
2006 and 2008, respectively, where he is currently
working toward the Ph.D. degree.
He is also pursuing a research assignment at Texas
A&M University at Qatar. His principal areas of research are modeling, simulation, and control of multiphase power electronic converters and fault diagnosis
using artificial intelligence.
Mr. Ahmed was a Gold Medalist in earning the M.Tech. degree. He is a recipient of the Toronto Fellowship, funded by AMU. He also achieved the Best
Research Fellow Excellence Award from Texas A&M University, Qatar, for the
year 20102011.

Fang Z. Peng (M92SM96F05) received the


B.S. degree in electrical engineering from Wuhan
University, Wuhan, China, in 1983, and the M.S.
and Ph.D. degrees in electrical engineering from
Nagaoka University of Technology, Nagaoka, Japan,
in 1987 and 1990, respectively.
From 1990 to 1992, he was a Research Scientist
with Toyo Electric Manufacturing Company, Ltd.,
Toyo, Japan, where he was engaged in the research
and development of active power filters, flexible
ac transmission system (FACTS) applications, and
motor drives. From 1992 to 1994, he was with the Tokyo Institute of Technology, Tokyo, Japan, as a Research Assistant Professor, where he initiated a
multilevel inverter program for FACTS applications and a speed-sensorless
vector control project. From 1994 to 1997, he was a Research Assistant Professor with the University of Tennessee, Knoxville, where he was also a Staff
Member. From 1994 to 2000, he was with the Oak Ridge National Laboratory,
where, from 1997 to 2000, he was the Lead (Principal) Scientist with the Power
Electronics and Electric Machinery Research Center. Since 2000, he has been
with Michigan State University, East Lansing, where he is currently a Professor
with the Department of Electrical and Computer Engineering. He is the holder
of more than ten patents.
Dr. Peng was the recipient of the 1996 First Prize Paper Award and the 1995
Second Prize Paper Award of the Industrial Power Converter Committee of
the IEEE Industry Applications Society Annual Meeting, the 1996 Advanced
Technology Award of the Inventors Clubs of America, Inc., the International
Hall of Fame, the 1991 First Prize Paper Award of the IEEE TRANSACTIONS ON
INDUSTRY APPLICATIONS, and the 1990 Best Paper Award of the Transactions
of the Institute of Electrical Engineers of Japan. He was an Associate Editor for
the IEEE TRANSACTIONS ON POWER ELECTRONICS from 1997 to 2001 and has
been again since 2005. He was the Chair of the Technical Committee for Rectifiers and Inverters of the IEEE Power Electronics Society from 2001 to 2005.

IEEE TRANSACTIONS ON SUSTAINABLE ENERGY, VOL. 4, NO. 1, JANUARY 2013

Yuan Li (M09) received the B.S., M.S., and Ph.D.


degrees from Wuhan University, Wuhan, China, in
2003, 2006, and 2009, respectively, all in electrical
engineering.
From 2007 to 2009, she was a Visiting Student
with the Department of Electrical and Computer Engineering, Michigan State University, where she was
involved in research on utility interface and control of
renewable energy sources. Since 2009, she has been
with the Department of Electrical and Information
Engineering, Sichuan University, Chengdu, China, as
a Lecturer. Her research interests include acdc inverters, pulsewidth modulation rectifier, distributed generation, and digital control in power electronics.

Ge Baoming (M11) received the Ph.D. degree in


electrical engineering from Zhejiang University,
Hangzhou, China, in 2000.
He was a Postdoctoral Researcher in the Department of Electrical Engineering, Tsinghua University,
Beijing, China, from 2000 to 2002, a Visiting Scholar
in the Department of Electrical and Computer Engineering, University of Coimbra, Coimbra, Portugal,
from 2004 to 2005, and a Visiting Professor in the
Department of Electrical and Computer Engineering,
Michigan State University, East Lansing, MI, from
2007 to 2008. Currently, he is working with the Department of Electrical and
Computer Engineering, Michigan State University; he also is a Professor with
the School of Electrical Engineering, Beijing Jiaotong University, Beijing,
China, where he has been since 2002. His research interests include permanent
magnet synchronous, switched reluctance and induction motors, real-time
control of electrical machines, power electronics systems, nonlinear control
theories, and applications to electric drives.