This article has been accepted for inclusion in a future issue of this journal.
Content is final as presented, with the exception of pagination.
IEEE TRANSACTIONS ON SUSTAINABLE ENERGY
An Improved MPPT Controller for Photovoltaic
System Under Partial Shading Condition
Kai Chen, Shulin Tian, Yuhua Cheng, Senior Member, IEEE, and Libing Bai
Abstract—Maximum power point tracking (MPPT) is an integral
part of a system of energy conversion using photovoltaic (PV)
arrays. The power-voltage characteristic of PV arrays operating
under partial shading conditions exhibits multiple local maximum
power points (LMPPs). In this paper, a new method has been
presented to track the global maximum power point (GMPP) of
PV. Compared with the past proposed global MPPT techniques, the
method proposed in this paper has the advantages of determining
whether partial shading is present, calculating the number of peaks
on P–V curves, and predicting the locations of GMPP and LMPP.
The new method can quickly find GMPP, and avoid much energy
loss due to blind scan. The experimental results verify that the
proposed method guarantees convergence to the global MPP under
partial shading conditions.
Index Terms—Maximum power point tracking (MPPT), multiple
maxima, partial shading, photovoltaic (PV).
I. INTRODUCTION
A PHOTOVOLTAIC (PV) cell is an electrical device that
converts the energy of light directly into electricity
through PV effect. PV cells have a complex relationship between
solar irradiation, temperature, and total resistance, and exhibit a
nonlinear output efficiency characteristic known as the P–V
curve. Therefore, maximum power point tracking (MPPT) tech-
niques should be developed in PV systems in order to maximize
the output power of PV systems. Nowadays, there have been
many MPPT methods reported in the literature, such as hill
climbing [1], perturb and observe [2], incremental conductance
(INC) [3], and ripple correction [4].
However, when there are multiple local power maxima [5],
[6], from partially shading or from installation on a curved
surface [7], [8], conventional MPPT techniques do not perform
well. Multiple maxima may occur due to bypass diodes [9], [10],
which are used to avoid hot spots from forming when some cells
in a module or some modules in a string receive less irradiance
than others [11]–[14]. Without the remediation of power elec-
tronics, the lost energy due to partial shading can be significant.
Manuscript received August 01, 2013; revised December 19, 2013 and February
17, 2014; accepted March 30, 2014. This work was supported in part by “the
Fundamental Research Funds for the Central Universities” (ZYGX2012YB029), in
part by the “Project on the Integration of Industry, Education and Research of
Guangdong Province” (Grant 2011B090400383), and in part by “the National
Natural Science Foundation of China’ (Grant 61102141). (Corresponding author:
Y. Cheng.)
The authors are with the School of Automation Engineering, University of
Electronic Science and Technology of China, Chengdu 611731, China (e-mail:
kaichen.uestc@gmail.com; shulin@uestc.edu.cn; yhcheng@uestc.edu.cn;
bailb991@163.com).
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.2014.2315653
1949-3029 © 2014 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission.
See http://www.ieee.org/publications_standards/publications/rights/index.html for more information.
1
Thus, it is imperative to utilize MPPT techniques that reliably
track the unique global power maximum present in shaded
arrays.
Some researchers have proposed global maximum power
point tracking (GMPPT) algorithms to address the partial shad-
ing condition. In [15], it is observed that the peaks follow a
specific trend in which the power at a peak point continues to
increase until it reaches the GMPP, and afterward, it continu-
ously decreases. In [16], the proposed algorithm incorporates an
online current measurement and periodic interruptions to address
certain challenges associated with rapidly changing insolation
and partial shading. A modified rule for updating the duty-cycle
is presented, which is based on the first and second derivatives of
the power as a function of the duty-cycle. In [17], a smart P–I
curve scanning is devised to detect the presence of the global
maxima. In [18], a dc–dc converter controlled by a dc signal of
adjustable amplitude is used to track the global MPP. Then the
perturb and observe (P&O) MPPT algorithm is applied in order
to continuously track the short-term variations of the previously
detected global MPP. In [19], it presents a novel model-based
two-loop control scheme for a particular module-integrated PV
and converter system, where bidirectional Cuk dc–dc converters
are used as the bypass converters and a terminal Cuk boost
functioning as a whole system power conditioner. The system
can increase power generation compared to the conventional
bypass diode structure. In [20], it uses a periodic scan sequence of
the P–V curve to detect local MPPs due to the long time required
for the completion of this tracking process, and the PV energy
production is reduced. The global MPPT methods that are
presented in [21] and [22] are based on the measurements of
the PV array open-circuit voltage and solar irradiation or short-
circuit current, respectively.
The above papers can be categorized into two groups. 1) The
global (or partial) scan of obtained PV output curves [15]–[20].
Then, a complex algorithm is used to calculate the GMPP of the
curves. One disadvantage of this method is the inability to
determine whether the PV cell is operating under the shading
conditions. Because the GMPP is blindly tracked, and the
conversion system works at low-power points most of the time
in the tracking process, energy waste is significant. 2) The other
group uses specialized sensors (e.g., for the measurement of solar
irradiation) to calculate the working status of the PV cells and
simulate the output characteristic curves [21], [22]. This method
can be an effective solution to mitigate the effect of partial
shading. The simulation results, however, obtained by measuring
environmental parameters and the actual case will be drastically
different, because the actual characteristic of the solar panels
depends on many factors (e.g., light intensity, temperature,
 This article has been accepted for inclusion in a future issue of this journal. Content is final as presented, with the exception of pagination.
2 IEEE TRANSACTIONS ON SUSTAINABLE ENERGY
Fig. 1. PV array under different partial shading conditions. (a) Sample PV array
with insolation of on the unshaded modules. (b) One shaded module
with . (c) Two shaded modules with . (d) Three shaded
modules with . The sample PV module’s parameters is
, , at
. Point A is the GMPP, while point B the local maximum power
aging, dust, and partial shading). In addition, the method in-
creases the PV system cost in practical commercial applications.
In this paper, a new method to track the GMPP of PV arrays
operating under partial shading conditions is presented. The rest
of the paper is organized as follows. Section II describes certain
critical observations on the current–voltage (I–V) and power–
voltage (P–V) curves of partially shaded arrays. Section III
describes the proposed algorithm to track the GMPP. Section IV
presents the experimental results. The main conclusion is sum-
marized in Section V.
II. PARTIAL SHADING CONDITIONS
Fig. 1(a) shows a PV array which has four PV modules
connected in series under uniform insolation conditions.
Fig. 2(a) illustrates typical I–V and P–V curves for the PV array
under a uniform solar irradiance of on all the PV
modules. The traditional MPPT algorithm can reach this peak
and continue oscillating around the MPP. The P&O method, e.g.,
perturbs the solar array voltage in one direction in each sampling
period and tests the power change afterward. It is assumed that
initially PV array is operating at point A, as shown in Fig. 2(a).
An operating voltage of the PV array is perturbed in a given
direction (from A to B), and an increase in output power is
observed ( > ). This means that point B is closer to the MPP
than point A, and the operating voltage must be further perturbed
in the same direction (from B to C). On the other hand, if the
output power of the PV array decreases (from D to E), the
operating point has moved away from the MPP, and therefore,
the direction of the operating voltage perturbation must be
reversed (from D to C). Through constant perturbation, eventu-
ally the operating voltage will reach and continue oscillating
around the MPP level.
However, in some practical conditions, the series strings of
PV modules are not under the same solar irradiance condition.
The partial shading condition is a common situation due to
the shadows of buildings, trees, clouds, dirt, etc. Fig. 1(b)–(d)
shows several different partial shading situations. Under the
partial shading condition, if there is one module in a PV string
that is less illuminated, the shaded module will dissipate some of
the power generated by the rest of the modules. It means that the
current available in a series-connected PV array is limited by the
current of the shaded module. This can be avoided by using
bypass diodes which can be placed in parallel with the PV
module. The method of using bypass diodes allows the array
current to flow in the correct direction even if one of the strings is
completely shadowed. Bypass diodes are widely implemented in
commercial solar panels. Because of bypass diodes, multiple
maxima appear under the partial shading condition.
The P–V curve of PV array in Fig. 1(b) possesses multiple
maxima under the partial shading condition, as shown in Fig. 2
(b). The unshaded modules in the sample PV array are exposed to
of solar insolation and the shaded module is
exposed to of solar insolation. There are two ob-
served peaks in the P–V curve, because of the natural behavior of
,
the bypass diode and PV array connection inside the module.
and
point (LMPP). When the area covered by the shadow changes,
the P–V curve and the location of GMPP also changes, as shown
in Fig. 2(c) and (d). Under these conditions, traditional algo-
rithms can only track either of the two MPPs, and cannot
distinguish between GMPP and LMPP. Continuing with the
P&O method as an example, both points satisfy the conditions to
be the “MPP.” If the operating point obtained by the PV array
algorithm is LMPP, the output power is significantly lower.
Some researchers proposed a global scan method to obtain the
PV output curves. Then a complex algorithm is required to
calculate the GMPP of the curves. This method is able to obtain
the GMPP, but it cannot determine whether the PV cell is
operating under shading conditions, and blindly and constantly
scans for the MPP, wasting the output energy. For these reasons,
a new improved MPPT method for the PV system under the
partial shading condition is proposed in this paper. In Section III,
the new method will be described in detail.
III. IMPROVED MPPT CONTROLLER METHOD
A. Analysis of Characteristic Curves Under Partial Shading
Conditions
In order to avoid blind global scan, methods to determine the
presence of partial shading are essential. It is noted that when a
series of PV array is under the identical solar irradiance condition
[Fig. 1(a)], every PV model works as a source, and all modules
are identical in their voltage, current, and output power at any
time. But this state changes when there is shadow.
Fig. 1(c) is an example in the following analysis. The models
in the series array are exposed to two different solar irradiances,
and the solar irradiation levels are 1000 and , respec-
tively. The voltages of the modules that are exposed to different
irradiation levels are completely different. The two peaks on the
P–V curve are divided into two separate parts, as shown in
Fig. 2(c). Part A is the curve containing the left peak (curve D–A–
C), and part B is the curve containing the right peak
(curve C–B–E). In part A, the current of the PV array is
greater than the maximum current that the PV module can
 This article has been accepted for inclusion in a future issue of this journal. Content is final as presented, with the exception of pagination.
CHEN et al.: IMPROVED MPPT CONTROLLER FOR PV SYSTEM
Fig. 2. P–V and I–V characteristics curves of a PV array under different partial
shading conditions, as shown in Fig. 1(a)–(d). (a) Unshaded modules. (b) One
shaded module with 400 . (c) Two shaded modules with 400
(d) Three shaded modules with 400
Fig. 3. Every module output voltage with array output power. (a) Unshaded
module. (b) Shaded module.
produce under the shade (M3 and M4); therefore, the current will
flow through the bypass diode of each module. At this stage, only
PV M1 and M2 are supplying power, and PV M3 and M4 have
been bypassed by the diodes. The characteristic curves of the PV
module voltage with output power are shown in Fig. 3(a) and (b).
The voltages of PV M3 and M4 are approximately negative 0.7 V
(the diode’s forward voltage drop) in part A, as shown in Fig. 3(b).
Therefore, the module voltages being equal to the negative
of the diode’s forward voltage can be used as one effective way
to estimate partial shading condition. In part B, all PV modules
are supplying power, but the unshaded and shaded modules are
in different working conditions. Because the PV modules
receive different amounts of solar radiation, the voltages of
the PV modules are different. In part B (curve C–B–E), the
voltage of the unshaded modules is greater than that of the
shaded modules, as shown in Fig. 4. It is evident that this is
another indicator to efficiently identify partial shading.
Following the above analysis, some of the observations are
listed as follows.
1) I–V curves under partial shading conditions have multiple
steps, while the P–V curves are characterized by multiple
peaks.
2) The number of peaks is equal to the number of different
insolation levels irradiated on the PV array, and any peak
point may be the GMPP.
3
Fig. 4. Array output power with unshaded module output voltage and shaded
module output voltage.
3) The voltages of PV modules that receive different solar
radiations are different.
4) The voltage of the PV module that is bypassed by a diode is
. equal to the negative of the diode’s forward voltage drop.
B. The Improved MPPT Controller Method
Fig. 5 shows the flowchart of the improved MPPT algorithm to
track the GMPP under partial shading conditions. The execution
of the algorithm always starts with a reference voltage ( )
value set to 85% of (block 1) as shown in the “Main
Program.” In the meantime, it calculates the number of PV
modules . When under uniform insolation, there is only one
peak in the P–V curve. Traditional MPPT methods such as the
P&O tracking process can work well. Therefore, until the partial
shading occurs, it maintains the operation at the GMPP by
continuously implementing the P&O method (blocks 2 and 3).
When an MPP is found, it will store the point information, i.e.,
array power and the voltage (block 4). A timer interrupt program
is used to ensure regular checking of the shading condition
(block 5). When a PV module voltage is greater than another
one ( > ) in the same point, it means that partial shading
has occurred. When the absolute difference between and
is greater than a predetermined constant (used to eliminate
sample disturbance and minor differences due to slight changes
in insolation), the “Main Program” calls the “GMPP track
subroutine” (block 7). The “GMPP track subroutine” finds the
true GMPP and then passes the control onto the “Main Program,”
which maintains the operation at this new GMPP.
The “GMPP track subroutine” determines the location of the
last MPP on the P–V curve (block 8). When the voltage of any PV
module is less than zero ( < ), it means that the last MPP is the
left peak on the P–V curve [as shown in Fig. 2(d)], and reference
voltage of the right peak is set to about 80% of (block 9).
Then, a conventional MPP technique (such as P&O) is applied to
track this peak (blocks 10 and 11). When any PV module voltage
is not less than zero, the last MPP is the right peak on the P–V
curve. It divides the voltage of all PV modules into two groups
according to their value and calculates , the number of
modules in the group with the smaller voltages (block 12). These
modules will be bypassed and will not output any power [as
shown in Fig. 2(d)], so the reference voltage of the left peak is
equal to about 80% of (block 13). Afterward,
 This article has been accepted for inclusion in a future issue of this journal. Content is final as presented, with the exception of pagination.

4 IEEE TRANSACTIONS ON SUSTAINABLE ENERGY

Fig. 5. Flowchart of the improved MPPT control.

the same MPP technique is applied to track this peak (blocks 14
and 15). By comparing the powers of this peak (new MPP) and
the previous one (last MPP), the true GMPP is obtained (blocks
16 and 17). Finally, the reference voltage is set to the voltage of
the true GMPP, and the control is passed onto the “Main
Program,” which maintains the operation at this GMPP until
the timer interrupt occurs again.
To understand the algorithm with an example, assume that a
PV array in the case of Fig. 1(c). The local peak [point B, as
shown in Fig. 2(c)] is tracked by the P&O method. Whether MPP
is reached or not is checked by determining the sign of the power
in two subsequent perturbations. When the local peak (point B) is
tracked and stored the current information ( , after making some modifications in the GMPP subroutine, the
), the algorithm measures the module voltage
( , ). Obviously, the judg-
ment ( > ) is established, and the “GMPP track subroutine”
is called. In the GMPP track program, no module voltage is
smaller than zero, so the module voltages are divided into two
groups ( , , and , ), and is equal to 2. Then the new
reference voltage is set to 37.1 V, and the P&O method is applied
to track another peak (point A). When the peak (point A) is
tracked, the new information ( ,
) is stored. The output power of the new MPP is greater
than the old one, so point A is the GMPP.
C. More Levels of Insolation
In general, a PV array is irradiated by only two levels of
insolation under the partial shading condition. However, some
extreme situation may also occur. For example, Fig. 6 shows the
characteristics of a PV array in which four PV modules are
connected in series and irradiated by four levels of insolation
conditions. Obviously, the above algorithm will no longer apply
to this situation, but the observations are still valid. Therefore,
previous algorithms can be successfully used for extreme situa-
tions. The “Main Program” will not have any change and can
continue to function well. The new flowchart of “GMPP track
subroutine” is shown in Fig. 7.
In order to determine the number of different insolation levels
irradiated on the PV array, the new “GMPP track subroutine”
starts with tracking the rightmost peak on the P–V curves. If any
PV module voltage is not less than zero, the last MPP is the
 This article has been accepted for inclusion in a future issue of this journal. Content is final as presented, with the exception of pagination.

CHEN et al.: IMPROVED MPPT CONTROLLER FOR PV SYSTEM 5

Fig. 6. Characteristics of a PV array under partial shading condition (the PV

module’s parameters are , , ,
at and ). (a) PV array configura-
tion. (b) I–V and P–V characteristics.

rightmost peak (block 2). Otherwise, the rightmost peak must be

tracked first, and the MPP information is stored (blocks 3–6). All
PV modules are divided into groups with respect to their
voltages and arrangements. The groups in a way that the cells
of the lowest voltage is labeled , while the cells of the largest
voltage is ( < < < ). Next, the subroutine cal-
culates the number of modules ( ) in every group
(block 7). Evidently, is the number of different insolation levels
irradiated on the PV array. There are MPPs information
that needs to be found, and the tracking will be carried out from
right peak to the left peak. When a PV array is operating at the
second rightmost peak on the P–V curves, modules, which
are at the lowest level of radiation, will be bypassed and will not
output any power, so the reference voltage of the peak is equal to
about 85% of (block 9). The same
method is used to find all other MPPs (blocks 8–13). After that,
by comparing the powers of all peaks, the true GMPP is obtained
(block 14). Finally, the reference voltage is set to the voltage of
the true GMPP, and the control is passed onto the “Main
Program,” which maintains the operation at this GMPP until
the timer interrupt arrives again.
IV. SIMULATION STUDIES
In order to evaluate the performance of the proposed MPPT
and compare with past scanning techniques, some MATLAB
simulation studies for a partial shading variation are carried out.
This process begins with an array from under no shadow [as
Fig. 1(a)] to partial shading [as Fig. 1(c)], then finally to complete
shade coverage (array with insolation of ). Mean-
while, two kinds of scanning techniques are chosen for compar-
ing: 1) a P–V curve periodic scan sequence; and 2) a rapid global
scanning algorithm [15]. In order to reasonably compare, all
methods are implemented through a boost circuit, the sampling
time is 0.01 s, and the voltage disturbance in the P&O method is
0.1 V. The disturbance of the rapid global scanning
algorithm is 13.3 V (60% of ) scanning algorithm nor the rapid global scanning algorithm can
The performance of the proposed algorithm under gradually
varying shading conditions is shown in Fig. 8(a). The insolation
level in the time range 0–3 s is , and the entire array
Fig. 7. Flowchart of the new “GMPP track subroutine.”
receives uniform insolation [Fig. 1(a)]. During this period, the
P&O method can work well and converge to the MPP. At
, a sudden change in the ambient conditions occurs,
which causes partial shading of the array. The insolation level
on the shaded module becomes [Fig. 1(c)]. At this
time, “GMPP track subroutine” begins to look for GMPP. First,
the P&O method traces the LMPP, and detects the presence of
partial shading. Then, the reference voltage of the left peak is set
to 35.5 V (80% of , , ,
). Afterward, the new GMPP is tracked down and replaces
the original LMPP at . The whole process takes 1.1 s.
When the whole array is covered in shadow at , the
“GMPP track subroutine” stops, and the P&O method finds a
new MPP. Compared with the proposed algorithm, the P–V
curve periodic scanning algorithm takes more time (8.6 s)
and loses more energy in the process of scanning, as shown in
Fig. 8(b). The rapid global scanning algorithm has good perfor-
mance on tracking speed and can fast track the GMPP (only
1.2 s), but some low-power work points still exist in the scanning
process, as shown in Fig. 8(c). Further, when the whole array is
covered in shadow (at ), neither the P–V curve periodic
determine the presence of partial shading. So, periodic scan
sequence or “GMPP track subroutine” cannot stop operating, and
cause more energy loss.
 This article has been accepted for inclusion in a future issue of this journal. Content is final as presented, with the exception of pagination.
6 IEEE TRANSACTIONS ON SUSTAINABLE ENERGY
Fig. 8. Performance of the three algorithms under varying shading conditions.
(a) The proposed algorithm. (b) P–V curve periodic scanning. (c) Rapid global
scanning algorithm.
V. EXPERIMENTAL RESULTS
A. Experimental Setup
To test the performance of the proposed MPPT technique, a
series of experimental tests was performed. Fig. 9 shows a boost
converter system adopted for the verified experiment. The PV
Fig. 9. Circuit schematic of the boost converter used to track GMPP.
TABLE I
DATA SHEET OF PV MODULE
array consisted of two modules of Solar Dongfang Electric
Corporation (DEC) 280 W, as shown in Fig. 10, and a bypass
diode is in parallel with each module. Table I shows the
characteristic values at standard test conditions (STCs; solar
irradiation , and cell temperature ).
The control unit of the proposed global MPPT system, which is
illustrated in Fig. 9, has been built with a Texas Instruments (TI)
Digital Signal Processor (DSP) TMS320F2812 microcontroller,
which features a 12-bit, 16-channel, A/D converter, and on-chip
8-bit pulsewidth modulation (PWM) outputs. The switching
frequency of the “PWM” control signal has been set to 20 kHz,
and the converter output has been connected to a 30- 600-W
power resistance. In addition, the PV arrays were installed at
different tilt angles to receive unequal solar irradiation. Thus, the
resulting P–V characteristic of the PV array indicates both local
and global MPPs.
B. Experimental Results
Initially, the P–V characteristic, I–V characteristic, and every
PV module voltage of the PV array were experimentally mea-
sured using a scan process. Then, the proposed global MPPT
method was applied. The experimentally measured operating
points of the PV array during the execution of the proposed
global MPPT process (i.e., points P1–P13) are also indicated in
Fig. 11. The time interval between two adjacent operating points
is 0.1 s, and the voltage disturbance in the P&O method is set to
2 V.
There are two observed peaks in the P–V curve, as shown in
Fig. 12. It proves that the PV array is exposed to two levels of
insolation. As seen in Fig. 12, the proposed algorithm can
quickly locate the GMPP. Fig. 13 shows the Please array output
voltage with every module output voltage.
The obtained maximum power, corresponding to the GMPP,
is 181.742 W, while the power corresponding to local peak is
120.192 W. The current and voltage at the GMPP are 2.822 A and
64.402 V, respectively. The values recorded for and are
 This article has been accepted for inclusion in a future issue of this journal. Content is final as presented, with the exception of pagination.

CHEN et al.: IMPROVED MPPT CONTROLLER FOR PV SYSTEM 7

87.034 V and 7.920 A, respectively. As two modules are

connected in series, voltages of the two modules connected in
series, V1 and V2, recorded at the GMPP are 34.530 and
29.872 V, respectively.

VI. CONCLUSION
The detection of global MPP is indispensable in maximizing
the PV system energy production in the case of PV array partial
shading. In this paper, a method was presented to track the global
MPP of a PV array. The method has the following advantages.
1) With comparison of the voltages of every PV module, the
method can detect the presence of partial shading and avoid
blind scanning.
2) By separating the voltages of all PV modules into different
Fig. 10. PV arrays used in the experimental setup.
groups, the method can calculate the number of peaks on
P–V curves and quickly locate the GMPP and LMPPs.
3) By calculating the number of PV modules in every group,
the method can predict the locations of GMPP or LMPPs.
This cuts down the tracking time.
The simulation and experimental results verify that the pro-
posed method determine the presence of partial shading and
guarantees convergence to the global MPP under partial shading
conditions.

Fig. 11. Experimentally measured P–V characteristic of the PV array and the PV
array operating points during the operation of the proposed GMPPT method.
Fig. 12. Experimental power response.
[10] E. Díaz-Dorado, A. Suárez-García, C. Carrillo, and J. Cidrás, “Influence of
[11] W. Herrmann, W. Wiesner, and W. Vaassen, “Hot spot investigations on PV
[12] M. C. A. García, W. Herrmann, W. Böhmer, and B. Proisy, “Thermal and
Fig. 13. Experimentally measured array output voltage with every module
output voltage.
REFERENCES
[1] W. Xiao and W. G. Dunford, “A modified adaptive hill climbing MPPT
method for photovoltaic power systems,” in Proc. Power Electron. Spec.
Conf. (PESC’04), vol. 3, Jun. 2004, pp. 1957–1963.
[2] N. Femia, G. Petrone, G. Spagnuolo, and M. Vitelli, “Optimization of
perturb and observe maximum power point tracking method,” IEEE Trans.
Power Electron., vol. 20, no. 4, pp. 963–973, Jul. 2005.
[3] F. Liu, S. Duan, and F. Liu, “A variable step size INC MPPT method for
PV systems,” IEEE Trans. Ind. Electron., vol. 55, no. 7, pp. 2622–2628,
Jul. 2008.
[4] J. W. Kimball and P. T. Krein, “Discrete-time ripple correlation control
for maximum power point tracking,” IEEE Trans. Power Electron., vol. 23,
no. 5, pp. 2353–2362, Sep. 2008.
[5] H. Patel and V. Agarwal, “MATLAB-based modeling to study the effects of
partial shading on PV array characteristics,” IEEE Trans. Energy Convers.,
vol. 23, no. 1, pp. 302–310, Mar. 2008.
[6] N. Thakkar, D. Cormode, V. P. A. Lonij, S. Pulver, and A. D. Cronin,
“A simple non-linear model for the effect of partial shade on PV systems,” in
Proc. IEEE Photovoltaic Spec. Conf. (PVSC), 2010, pp. 2321–2326.
[7] A. Karavadi and R. S. Balog, “Novel non-flat photovoltaic module geome-
tries and implications to power conversion,” in Proc. Energy Convers.
Congr. Expo., 2011, pp. 7–13.
[8] A. Karavadi, “Power electronics design implications of novel photovoltaic
collector geometries and their application for increased energy harvest,”
M.S. thesis, Dept. Electr. Comput. Eng., Texas A&M Univ., College
Station, TX, USA, 2011.
[9] S. Vemuru, P. Singh, and M. Niamat, “Modeling impact of bypass diodes on
photovoltaic cell performance under partial shading,” in Proc. IEEE Int.
Conf. Electro/Inf. Technol. (EIT), May 2012, pp. 1–5.
the shadows in photovoltaic systems with different configurations of bypass
diodes,” in Proc. Int. Symp. Power Electron. Elect. Drives Autom. Motion
(SPEEDAM), Jun. 2010, pp. 134–139.
modules—New concepts for a test standard and consequences for module
design with respect to bypass diodes,” in Proc. Photovoltaics Spec. Conf.,
1997, pp. 1129–1132.
electrical effects caused by outdoor hot-spot testing in associations of
photovoltaic cells,” Progr. Photovoltaics: Res. Appl., vol. 11, pp. 293–307,
2003.
 This article has been accepted for inclusion in a future issue of this journal. Content is final as presented, with the exception of pagination.
8 IEEE TRANSACTIONS ON SUSTAINABLE ENERGY
[13] E. Karatepe, M. Boztepe, and M. Çolak, “Development of a suitable model
for characterizing photovoltaic arrays with shaded solar cells,” Solar
Energy, vol. 81, pp. 977–992, Jan. 2007.
[14] A. Bidram, A. Davoudi, and R. S. Balog, “Control and circuit techniques to
mitigate partial shading effects in photovoltaic arrays,” IEEE J. Photo-
voltaics, vol. 2, no. 4, pp. 532–546, Oct. 2012.
[15] H. Patel and V. Agarwal, “Maximum power point tracking scheme for PV
systems operating under partially shaded conditions,” IEEE Trans. Ind.
Electron., vol. 55, no. 4, pp. 1689–1698, Apr. 2008.
[16] A. Zbeeb, V. Devabhaktuni, and A. Sebak, “Improved photovoltaic MPPT
algorithm adapted for unstable atmospheric conditions and partial shading,”
in Proc. Int. Conf. Clean Elect. Power, Jun. 2009, pp. 320–323.
[17] S. Kazmi, H. Goto, O. Ichinokura, and H.-J. Guo, “An improved and very
efficient MPPT controller for PV systems subjected to rapidly varying
atmospheric conditions and partial shading,” in Proc. Australas. Univ.
Power Eng. Conf., Sep. 2009, pp. 1–6.
[18] E. Koutroulis and F. Blaabjerg, “A new technique for tracking the global
maximum power point of PV arrays operating under partial-shading con-
ditions,” IEEE J. Photovoltaics, vol. 2, no. 2, pp. 184–190, Apr. 2012.
[19] B. V. P. Chong and L. Zhang, “Controller design for integrated PV converter
modules under partial shading conditions,” Solar Energy, vol. 92,
pp. 123–138, 2013.
[20] M. Trova, “Top 5 performers in PHOTON inverter tests,” in Proc. Photon’s
1st PV Inverter Conf., Stuttgart, Germany, Apr. 2010.
[21] M. Lei, S. Yaojie, L. Yandan, B. Zhifeng, T. Liqin, and S. Jieqiong, “A high
performance MPPT control method,” in Proc. Int. Conf. Mater. Renew.
Energy Environ., 2011, vol. 1, pp. 195–199.
[22] S. Kazmi, H. Goto, O. Ichinokura, and H.-J. Guo, “An improved and very
efficient MPPT controller for PV systems subjected to rapidly varying
atmospheric conditions and partial shading,” in Proc. Australas. Univ.
Power Eng. Conf., 2009, pp. 1–6.
Kai Chen received the B.S. and M.S. degrees from the
University of Electronic Science and Technology of
China, Chengdu, China, in 2008 and 2011. Currently,
he is pursuing the Ph.D. degree at the School of
Automation Engineering, University of Electronic
Science and Technology of China.
His research interests include testing and measure-
ment instrument design, nondestructive evaluation
(NDE), and renewable energy systems.
Shulin Tian received the B.S. and M.S. degrees in
automation engineering science from the University
of Electronic Science and Technology of China,
Chengdu, China, in 1989 and 1991, respectively.
Currently, he is a Professor and the Dean of the
School of Automation Engineering, University of
Electronic Science and Technology of China. Majored
in measuring and testing technology and instruments,
he has conducted a number of projects in high speed,
high-precision data acquisition and processing, high-
speed wave form generation, network/communication
testing, testing bus technology, and testing system integration.
Dr. Tian is a Member of the National Electronic Measuring Instrument
Standardization Technology Committee, and Chief Committeeman of the
Sichuan Automation Control Committee of Instrument, Institute of Sichuan
Province, China.
Yuhua Cheng (M’10–SM’13) received the Doctor
degree in measurement techniques and automation
devices from the Sichuan University, Chengdu,
China, in 2007.
He is an Associate Professor of Automation En-
gineering with the University of Electronic Science
and Technology of China, Chengdu, China. His re-
search interests include nondestructive evaluation
(NDE), structural health monitoring (SHM), and the
application of photonics in sensing and communica-
tions systems.
Libing Bai received the B.S. and Ph.D. degrees from
the University of Electronic Science and Technology
of China, Chengdu, China, in 2008 and 2013.
Currently, he is an Associate Professor of Auto-
mation Engineering with the University of Electronic
Science and Technology of China. His research inter-
ests include measurement and control technology
and instrument, sensor signal processing, structural
health monitoring, and nondestructive testing and
evaluation.