energies

Article
Design and Implementation of a Dual-Axis Solar
Tracking System
Huilin Shang * and Wei Shen

School of Mechanical Engineering, Shanghai Institute of Technology, Shanghai 201418, China

* Correspondence: shanghuilin@sit.edu.cn; Tel.: +86-216-087-3024

Abstract: A dual-axis solar tracking system with a novel and simple structure was designed and
constructed, as documented in this paper. The photoelectric method was utilized to perform the
tracking. The solar radiation values of the designed system and a fixed panel system were theoretically
estimated and compared, showing that the proposed system is more efficient in collecting solar energy
than a fixed solar panel with a 30◦ tilted fixed surface facing south. The experimental results verified
the validity of the prediction as well as the efficiency of the proposed solar tracking system. In a
comparison of the data obtained from the measurements, 24.6% more energy was seen to have been
obtained in the dual-axis solar tracking system compared to the fixed system. This study possesses
potential value in small- and medium-sized photovoltaic applications.

Keywords: solar tracking system; solar panel; dual-axis; solar angles; photoelectric tracking method

1. Introduction
Due to the already high and constantly increasing demand for energy worldwide, the
exploitation of various conventional energy sources has been induced [1]. Considering
the environmental issues that non-renewable energy sources cause—for instance, the
emission of greenhouse gases—renewable sources of energy such as solar, wind, biomass
and hydro have been adopted to limit and reduce this degradation [2]. Among these,
solar energy holds significant promise [3], as it is one of the most abundant, clean, cost-
free and inexhaustible resources of energy. To convert solar energy into electrical energy,
Citation: Shang, H.; Shen, W. Design
and Implementation of a Dual-Axis
Photovoltaic Voltage (PV) solar panels have been widely used as important components
Solar Tracking System. Energies 2023,
of solar tracking systems [4]. A PV solar panel comprises a large number of solar cells
16, 6330. https://doi.org/10.3390/ composed of silicon-like semiconductors. It works by transferring the energy of photons
en16176330 from sunlight that strikes the solar cells to silicon electrons, which then flow through
electrodes and the external circuit and generate an electric current [5].
Academic Editor: Manolis Souliotis
Fixed solar panels have been utilized over the past few decades since fixed track-
Received: 26 July 2023 ing can easily accommodate harsh environmental conditions [6]. There are three typical
Revised: 22 August 2023 structural installations for fixed solar panels: namely, fixed-angle, vertical and season-
Accepted: 29 August 2023 adjusted fixed-angle installations. Fixed-angle solar panels have been extensively applied
Published: 31 August 2023 in large solar power stations [7] and rooftop solar tracking systems [8], while vertical panels
are available for vertical structures with limited space, such as the façades of high-rise
buildings [9,10]. With season-adjusted fixed-angle systems, the solar panels’ tilt angles in
the horizontal direction can be modified for seasonal variations in order to maximize the
Copyright: © 2023 by the authors. collected amount of solar radiation [11,12]. Nevertheless, angle adjustment is not that con-
Licensee MDPI, Basel, Switzerland.
venient due to many site attributes, such as security measures for working on rooftops or
This article is an open access article
other high buildings and the human cost of adjusting a large number of solar panels in solar
distributed under the terms and
power stations.
conditions of the Creative Commons
Compared to fixed solar panels, solar tracking systems [13] that can track the position
Attribution (CC BY) license (https://
of the sun based on both the season and the moment of each day have higher solar energy
creativecommons.org/licenses/by/
4.0/).
collection efficiency [14], thus possessing broader applications and higher research value.

Energies 2023, 16, 6330. https://doi.org/10.3390/en16176330 https://www.mdpi.com/journal/energies

Energies 2023, 16, 6330 2 of 13

Based on the different degrees of freedom of structures, there are two different types of
solar tracking systems: single-axis and dual-axis [15,16]. The former is designed to track
the sun on a single axis according to the azimuth angle, while the latter is designed to track
it via dual axes corresponding to the azimuth and solar altitude angles. In recent years,
novel technologies and designs for both types have been developed to achieve high control
accuracy and structural reliability for these systems’ practical performance.
For the design of single-axis solar tracking systems, Poulek and Libra [17] have pro-
posed a simple solar tracker based on a new auxiliary bifacial arrangement connected
directly to a Direct Current (DC) motor; Mavromatakis and Franghiadakis [18] have pre-
sented an azimuthal tracker with the capability of moving the collector’s plane in two
directions via a special support structure so as to be efficient in smaller photovoltaic solar
applications; Kim et al. [19] have designed a single-axis Compound Parabolic Concentrator
(CPC) tracking solar collector and found it more stable and efficient than stationary CPC
solar collectors through a numerical evaluation of its performance; Kiyak and Gol [20] have
developed a single-axis solar tracking system based on both fuzzy logic and a Proportional
Integral Derivative (PID) controller and found via experimental results that the efficiency
of the fuzzy logic was much higher than that of the PID controller; and Li et al. [21]
have proposed a new solar tracking approach based on kirigami by taking the advantage
of paper-like Organic Photovoltaics (OPVs) made on flexible substrates, thus providing
stretchability and robustness.
Even though it is easier and more economical to construct single-axis solar tracking
systems, considering the higher possibility of the solar panels of dual-axis tracking systems
facing the sun perpendicularly and thus resulting in higher energy collection efficiency, an
increasing number of investigations of dual-axis tracking systems have been conducted re-
cently. Abdallah and Nijmeh [22] designed and constructed an electromechanical dual-axis
solar tracking system with an open-loop Programmable Logic Controller (PLC). Their ex-
perimental study showed that the measured solar energy collected on the system’s moving
surface was significantly larger than that on a fixed surface. Sungur [23] focused on the de-
sign of programmable logic control for a dual-axis solar tracking system and experimentally
verified that 42.6% more energy could be obtained from the system than from PV panels at
fixed positions. Based on the derivation and calculation of the mathematical formulation of
a dual-axis solar tracking system using its mechanical components and photoresistances,
Kentli and Yilmaz [24] proposed a system that could produce 30% more energy than a
fixed one. Syafii et al. [25] have presented a sensorless dual-axis solar tracker based on a
database of sun positions that uses sunrises and sunsets, created from exact calculations of
solar azimuths, elevations/geographic locations, altitudes and time zones. According to
theoretical studies of solar angles, which used time and geographic parameters, in Tunisia,
Skouri et al. [26] constructed three accurate dual-axis solar tracking systems. Laseinde and
Remere [27] have developed a maximum power point tracking algorithm for a dual-axis
servo motor feedback tracking system using an Arduino board, showing the advantages of
energy and space savings. Al-Rousan et al. [28] have proposed a dual-axis solar tracking
system by integrating supervised logistic regression and a supervised multilayer percep-
tron in order to increase the accuracy of tracking prediction. It is not hard to observe that,
compared to the structural design of dual-axis solar tracking systems, more attention has
been paid to the design of their control systems [29].
In the present study, a simple dual-axis solar tracking system has been designed
and implemented. The outline of this paper is as follows: The designed structure and
the operational principle of the system are presented in the next section. The collection
efficiency of solar energy is discussed analytically in Section 3. Experimental results are
provided to illustrate the efficiency of the solar energy collection system in Section 4. Finally,
Section 5 contains a discussion and conclusions.
Energies 2023, 16, x FOR PEER REVIEW 3 of 13

Energies 2023, 16, 6330 illustrate the efficiency of the solar energy collection system in Section 4. Finally, Section
3 of 13
5 contains a discussion and conclusions.

2. Structural Design and Operational Principles of the Solar Tracking System

2. Structural Design and Operational Principles of the Solar Tracking System
The mechanical structure of the system, shown in Figure 1, includes solar panel com-
The mechanical structure of the system, shown in Figure 1, includes solar panel
ponents, gears, support members, power storage units, photodiodes and control modules.
components, gears, support members, power storage units, photodiodes and control
In this solar tracking device, a Microcontroller Unit (MCU) is the core controller that ana-
modules. In this solar tracking device, a Microcontroller Unit (MCU) is the core controller
lyzes the signals transmitted from each component and controls the motor to rotate the
that analyzes the signals transmitted from each component and controls the motor to rotate
solar panel to the appropriate angle. Photoresistors on the boundaries of the solar panel
the solar panel to the appropriate angle. Photoresistors on the boundaries of the solar panel
are used to adjust the device. The battery stores power transferred from solar energy and
are used to adjust the device. The battery stores power transferred from solar energy and
supplies
supplies it
it to
to the control modules
the control modules and
and the
the dynamic
dynamic devices.
devices.

Figure 1.
Figure Structure diagram
1. Structure diagram of
of the
the solar
solar tracking
tracking device.
device.

For dual-axis
For dual-axis solar
solar tracking
tracking performance,
performance, the the structure
structure inin Figure
Figure 11 should
should have the
have the
ability to
ability to rotate
rotate inin the
the east–west
east–west andand south–north
south–north directions. The components
directions. The components that that fulfill
fulfill
this rotation can be seen in Figure 2. For the rotation of the solar panel in the east–west
this rotation can be seen in Figure 2. For the rotation of the solar panel in the east–west
direction, the
direction, thesupporting
supportingbasebaseisisdriven
drivenbybya motor
a motor gear. A diagram
gear. of itsofbottom
A diagram is depicted
its bottom is de-
in Figure 2a, illustrating that the bottommost fixed device is placed
picted in Figure 2a, illustrating that the bottommost fixed device is placed on the ground on the ground and
linked by a supporting base and pins. In the center of the supporting base, there is a base
and linked by a supporting base and pins. In the center of the supporting base, there is a
bearing. The stepping motor can lead to the rotation of the base via driving the motor gear,
base bearing. The stepping motor can lead to the rotation of the base via driving the motor
thus inducing the east–west rotation of the solar panel on the base, namely, the rotation in
gear, thus inducing the east–west rotation of the solar panel on the base, namely, the ro-
the plane xOy. For the south–north rotation of the solar panel, i.e., the rotation in the plane
tation in the plane xOy. For the south–north rotation of the solar panel, i.e., the rotation in
xOz, the stepping motor can trigger the rotation of the solar panel’s supporting frame by
the plane xOz, the stepping motor can trigger the rotation of the solar panel’s supporting
driving the rotations of the motor gear, the intermediate drive shaft gear, the transmission
frame by driving the rotations of the motor gear, the intermediate drive shaft gear, the
shaft and the double-sided drive gears (see Figure 2b). Accordingly, the rotations of the
transmission shaft and the double-sided drive gears (see Figure 2b). Accordingly, the ro-
panel in both directions are independent of each other.
tations of the panel in both directions are independent of each other.
As shown in Figure 2, for the spur gears to perform the rotation, a 20◦ pressure angle
As shown in Figure 2, for the spur gears to perform the rotation, a 20° pressure angle
was set, as it is the preference of most designers [30]. The solar panel was set to be a
was set, as itshape
rectangular is thewith
preference
a lengthofofmost
840 designers
mm, a width [30].ofThe
600solar
mm andpanel was set toofbe20a mm.
a thickness rec-
tangular
Considering shapethewith a length
stability of theofbase
840 of
mm,the asolar
width of 600
panel, themm
pitchand a thickness
diameter of theofbase
20 mm.
gear
Considering the stability of the base of the solar panel, the pitch diameter
was set to nearly the diagonal length of the solar panel, namely, d p b = 996 mm. Since of the base gear
was set tocommon
the most nearly the diagonal
module forlength of theissolar
spur gears 6 mm panel,
[30],namely,
we set the module 996mm. Since
for the the
gears
most
of thiscommon
device at module
M = 6formm.spurThegears is 6 mm
number [30], we
of teeth on set
thethe
basemodule
gear canfor be
thecalculated
gears of this
as
device
N =
d pbat M = 6 mm. The number of teeth on the base gear can be calculated as
= 166. For the motor gear meshed with the base gear, we used the dimension of
1 M
d pb
the
N1 =pitch =diameter, pm = 90 mm; thus, the number of teeth is N2 = 15. On this basis, in
166 . For dthe motor gear meshed with the base gear, we used the dimension of
FigureM2b, the gear ratio is mG1 = N 1
N2 = 11.07, and the dimensions of the gears are used
as follows:
the pitch diameter, pm = 90mm
the pitch ddiameter of; the driving
thus, of teeth is N
shaft intermediate
the number 2 = 15
gear, d ps = 120
. On thismm,
basis,and
in
the number of teeth, N3 = 20; the pitch diameter of the double-sided drive gears being the
same as that of the motor gear in Figure 2a, i.e., d pm = 90 mm, and the number of teeth
of each gear, N2 = 15; and the pitch diameter of the transmission shaft on the solar panel
 follows: the pitch diameter of the driving shaft intermediate gear, d ps = 120mm , and the
number of teeth, N3 = 20 ; the pitch diameter of the double-sided drive gears being the
same as that of the motor gear in Figure 2a, i.e., d pm = 90mm , and the number of teeth of

Energies 2023, 16, 6330 each gear, N2 = 15 ; and the pitch diameter of the transmission shaft on the solar panel
4 of 13
supporting frame, d pt = 462mm , and its number of teeth, N4 = 77 . From the gear ratio
in Figure 2b, the following can be concluded:
supporting frame, d pt = 462 mm, and d ps dits N N of teeth, N4 = 77. From the gear ratio in
number
pt
Figure 2b, the following can m = 2 = 3 2 4 = 6.84.
beG 2concluded: (1)
d pm N2
d ps d pt N N
mG2 = 2 = 3 2 4 = 6.84. (1)
d pm N2

(a)

(b)
Figure
Figure 2.
2. Components
Components to
to perform
perform rotation
rotation in
in two
two directions:
directions: (a)
(a) east–west
east–west and
and (b)
(b) north–south.
north–south.

We then applied the photoelectric method [31] of solar tracking. This method depends
on the outputs of the photoresistors on the edges of the solar panel, as shown in Figure 1.
Here, we note the signals of the photoresistors on the upper, lower, right and left edges
with the coefficients m1 , m2 , m3 and m4 , respectively. The route of the photoelectric tracking
method of the designed structure is displayed in Figure 3. m1 and m2 were converted into
digital signals with an Analog/Digital (A/D) conversion circuit and transferred to the
MCU, which then induced the rotation of the stepping motor so as to lead to the rotation
of the solar panel’s supporting frame, as shown in Figure 2b. This process is for the panel
adjustment of the rotation in the south–west direction, which would not stop until the
values of m1 and m2 were equal. Considering that this solar tracking system is sensitive
Energies 2023, 16, 6330 5 of 13

to practical weather disturbances such as clouds and shadows, we set the equality of m1
and m2 to mean that m1 was within the range of 97.5% to 102.5% of m2 . A tolerance range
Energies 2023, 16, x FOR PEER REVIEW ±2.5% was adequate for this application. For the adjustment of the rotation in
of 6 ofthe
13
east–west direction, the process was similar; the deterministic parameters were m3 and m4 .
The solar tracking was completed when the adjustment in the two directions was finished.

Start

Photoelectric
tracking method

Horizontal
m1 = m2 ? N m1 < m2 ? N motor is
reversing

Y Y

Horizontal motor Horizontal

stops running motor is turning

Base motor is
m3 = m4 ? N m3 < m4 ? N
reversing

Y Y

Base motor stops Base motor is

running turning

End

Figure 3.
Figure 3. Control
Control algorithm
algorithm of
of the
the solar
solar tracking
tracking system.
system.

The PID
3. Energy control algorithm
Harvesting Efficiency [31] was utilized to carry out the control of the intermittent
Analysis
position adjustments made by the motors.
On the basis of the photoelectric trackingThe control logic we
method, wasanalytically
simple, withevaluated
a fast reaction
and
time to meet the requirements of the device. The process included several
compared the total radiation of the solar panels of the designed and fixed systems steps. First,
in the
the
tracking error, m1 − m2 or m3 − m4 and the output of the control term in direct proportion
same location and with the same weather. To this end, formulas to determine the relevant
to this error were calculated. Next, the proportional effect provided a quick initial response
parameters, such as time, declination angle, day length, hour angle and ratio of hourly
by enabling the tracking device to rapidly move toward the target position. Then, based
radiation, were expressed at first.
on the error signal, the integral control term constantly adjusted the position of the device,
We adopted the time parameter reported in Ref. [32]. The Equation of Time (EoT) for
thus eliminating steady-state errors. When the target position was nearly reached, the error
each day of the year is expressed as:
would decrease, and the differential control term would respond to the change in the error.
EoT ( N ) = 9.87 sin(2 B ) − 7.53cos( B ) − 1.5sin( B ) , (2)
where
360( N − 81)
B= , (3)
364
Energies 2023, 16, 6330 6 of 13

If the change were too rapid, the differential effect would generate a reverse control output
to slow down the rate of the change, thus suppressing the overshoot and oscillation of the
system. Since each adjustment was independent, there was no cumulative error over time
in the system.
In the structure of the proposed solar tracking system, a few gears driven by step
motors could make the solar panel rotate in two directions to perform the tracking. Addi-
tionally, the solar supporting frame and columns, designed to distribute vertical pressure,
could avoid the bulking instability of the components. The structure is simple, with not
many components, and it is easy to follow the working principle of the control of this
system. Hence, the designed dual-axis solar tracking system can be easily installed and
assembled, which may also reduce the maintenance and possibility of failure of the system.

3. Energy Harvesting Efficiency Analysis

On the basis of the photoelectric tracking method, we analytically evaluated and
compared the total radiation of the solar panels of the designed and fixed systems in the
same location and with the same weather. To this end, formulas to determine the relevant
parameters, such as time, declination angle, day length, hour angle and ratio of hourly
radiation, were expressed at first.
We adopted the time parameter reported in Ref. [32]. The Equation of Time (EoT) for
each day of the year is expressed as:

EoT ( N ) = 9.87 sin(2B) − 7.53 cos( B) − 1.5 sin( B), (2)

where
360( N − 81)
B= , (3)
364
and N is the day number. For instance, on 1 January, N = 1.
The corresponding declination angle in degrees is expressed as the following formula:
 
360(284 + N )
δ = 23.45 sin , (4)
365

based upon this, the hour angle at sunset and the length of the day in hours can be expressed
as follows:
2
hss = arccos(− tan( L) tan(δ)), Dl = − tan( L) tan(δ), (5)
15
respectively. Here, the parameter L represents the local latitude. Since the time intervals
between noon and sunrise/sunset are equal, the following number, n, was introduced to
note the time interval in hours:
D
n = [ l ], (6)
2
within each time interval, namely, the number of the exact ranges from 1 to n, the angle for
each hour can be expressed as:

hi± = 15◦ ( LST + EoT ( N ) ± 4(SL − LL) − 12) i = 1 . . . n, (7)

where LST means Local Standard Time, SL means Standard Longitude and LL means Local
Longitude. The plus and minus signs in Equation (7) correspond to cases of afternoon hours
and morning hours, respectively. The solar incidence angle, θi , in each hour, depending on
the location of the solar panel, is expressed as:

θi = arccos(sin( L ± β 0 ) sin(δ) + cos( L ± β 0 ) cos(δ) cos(hi )) i = 1 . . . n, (8)

where β 0 is the tilt angle of the fixed solar surface from the horizontal plane, and the plus
and minus signs apply to the locations of the solar panel in the northern and southern
hemispheres, respectively.
Energies 2023, 16, 6330 7 of 13

Thus, the ratio of hourly to daily total radiation can be written as:

15π ( a + b cos(hi± ))(cos(hi± ) − cos(hss ))

ri± = i = 1 . . . n, (9)
360 sin(hss ) − 2πhss cos(hss )
where
a = 0.409 + 0.5016 sin(hss − 60◦ ), b = 0.6609 − 0.4767 sin(hss − 60◦ ), (10)

on this basis, the energy collected by the stationary solar tracking system for the whole day
can be obtained as follows:
n
W f ixed = HSη ∑ ri± cos(θi ). (11)
i =1

In the above equation, the parameters H, S and η represent the daily solar irradiation
per unit area, the solar panel area and the energy conversion efficiency of the solar panel,
respectively. Here, it is worth mentioning that the case for cos θi < 0 does not mean the
dissipation of solar energy via the solar panel; instead, it indicates that the sun is behind
the solar panel and, thus, no solar energy can be harvested. Therefore, when cos θi < 0, we
set cos θi = 0 in Equation (11). In most of the daylight hours, for the fixed panel, θi 6= 0,
namely, cos θi < 1.
For differences in the solar tracking system with the same solar panel and under the
same weather conditions, since the solar tracker keeps the solar panel perpendicular to
the solar radiation during daylight hours, cos(θi ) = 1. As the sum of the ratio, ri± , for a
n
whole day, ∑ ri± = 1. The collected energy of the solar tracking system for the same day
i =1
would be:
Wtracked = HSη, (12)
obviously, Wtracked > W f ixed .
Taking the location at the 30.8466◦ latitude and 121.5164◦ longitude in Shanghai, China,
as an example, the exact period we considered was five days long, from 27 March 2023, to
31 March 2023. Assuming sunny weather, we calculated the approximate collected energy
of the solar panels of the fixed solar panel and the proposed solar tracking system. The
values of some related parameters are provided in Table 1.

Table 1. Parameter table of values for calculation.

Symbol Meaning Value

LST Local Standard Time China Standard Time
N Annual Solar Altitude Days From 86 to 90
SL Standard Longitude 120◦ East Longitude
LL Local Longitude 121.5164◦ East Longitude
L Latitude 30.8466◦ North Latitude
β0 Tilt Angle of the Fixed Surface from Horizontal 30◦

In substituting the values of the parameters in the above table in Equations (11) and (12)
and noting the average collected energy of the five days as:

90 90
e f ixed = 1 ∑ W N , W
W e tracked = 1 ∑ W N , (13)
5 N =86 f ixed 5 N =86 tracked

we obtained
e tracked − W
W e f ixed
≈ 26.7%, (14)
We f ixed
Energies 2023, 16, 6330 8 of 13

implying that ideally, the energy collected via the designed solar tracking system would be
much higher than that of the fixed solar panel on a daily basis.
Whether the designed solar tracking system can be more efficient in energy harvesting
than the fixed system depends on not only the extra energy collected but also the energy
consumption required to control and drive the solar panel rotation in each system. Express-
ing W
e tracked as the average power of the solar panel, Psp , and the length of the day, Dl , and
substituting their form into Equation (14) yielded:

e tracked = Dl Psp ,
W (15)
and
e tracked − W
Wextra = W e f ixed = 0.2111Dl Psp . (16)
Next, we applied two stepping motors to the designed system. Every half hour,
the selected motors could work for 30 s, hence operating up to 4n times per day. Their
1
working period for each time is tm = 120 . Considering that the stepping motor operates at
full power and the total power consumption of the sensors and other control modules is
negligible relative to the energy consumption of the motor [33], the energy consumption of
the stepping motor can be written as:
[ Dl ]
Wconsume = 8ntm Pm = Pm (17)
30
since the designed structure is lightweight with typical materials [34], the power of the
motor to drive the rotation, Pm , can be less than Psp . Comparing the extra collected energy,
Wextra , and the consumed energy of the device, Wconsume , yields:

Wextra 6.333Psp
= >> 1. (18)
Wconsume Pm

It follows that the increase in collected energy when the proposed solar tracking device
is used to replace the fixed one can be far greater than the energy consumed in driving the
designed device, demonstrating the energy-harvesting efficiency of the designed system.

4. Experimentation and Results

In this section, experimental measurements are documented to verify the validity
of our analysis as well as the efficiency increase for solar energy with the designed solar
tracking system. A continuous test was carried out for 5 days, from 27 March to 31 March,
at the Shanghai Institute of Technology, Shanghai, China. The location of this test was
in Southeast China, at the latitude and longitude illustrated in Table 1. The outdoor
temperature ranged from 9 ◦ C to 20 ◦ C. Pictures of the prototype of the designed solar
tracker and the stationary solar panel are provided in Figure 4. Here, the dimensions of
the prototype depicted in Figure 4a were one-tenth of the dimensions of the numerical
model shown in Figure 1. The two solar panels, made of polysilicon, had the same-sized
rectangular shape, with a length of 84 mm, a width of 60 mm and a thickness of 2 mm.
Their maximum open-loop voltage was 6.3 V. The supporting base of the structure was
made of polymethylmethacrylate (PMMA) plastics, and the double-sided drive gears, the
motor gear, the support columns and the support frame of the solar panel were all made of
polyoxymethylene (POM) plastics. We selected a 28BYJ-48 stepping motor with a working
1 ◦
voltage range of 5 V–12 V and a step angle of 64 , as well as an STC89C 52 MCU with a
working voltage of 3.3 V and photoresistors with a light resistance range of 2 kΩ–5 kΩ and
a dark resistance of 200 kΩ. The fixed-surface solar panel, tilted at 30◦ toward the south
with the same-sized solar panel, is presented in Figure 4b.
 The experimental apparatus is illustrated in Figure 5. The multimeter was mounted be-
tween the dual-axis tracking device and the fixed solar collector, connecting to the data
logger, which was in turn connected to a computer. Multimeter readings were recorded
every hour. The collected and stored data were processed utilizing Microsoft Excel, and
the measured solar radiation values were averaged to obtain the average hourly solar ra-
Energies 2023, 16, 6330 9 of 13
diation power in Watts. The mobile power supply provided energy for the mechanical
system during the solar tracking process.

(a) (b)
Figure
Figure4.
4.Installation
Installationof
oftwo
twosolar
solarpanels:
panels:(a)
(a)tracking
trackingprototype
prototypeand
and(b)
(b)fixed
fixedpanel.
panel.

Both solar panels were installed facing the south and tested in outdoor field work. The
experimental apparatus is illustrated in Figure 5. The multimeter was mounted between
the dual-axis tracking device and the fixed solar collector, connecting to the data logger,
which was in turn connected to a computer. Multimeter readings were recorded every hour.
The collected and stored data were processed utilizing Microsoft Excel, and the measured
Energies 2023, 16, x FOR PEER solar
REVIEWradiation values were averaged to obtain the average hourly solar radiation power in 10 of 13
Watts. The mobile power supply provided energy for the mechanical system during the
solar tracking process.

(a) (b)
FigureFigure 5. Experimental
5. Experimental apparatus:
apparatus: (a) experimental
(a) the the experimental
setupsetup
and and (b) the
(b) the solar
solar tracking
tracking system in
system
operation.
in operation.

BasedBased on Equation
on Equation (6),calculated
(6), we we calculated that from
that from 27 March
27 March to 31 to 31 March
March 2023, 2023, the day
the day
lengths
lengths in hoursin hours were 12.16,
were 12.16, 12.19, 12.26
12.19, 12.22, 12.22,and
12.26 andrespectively.
12.29, 12.29, respectively. The average
The average numbernum-
ber of hours
of daylight daylight
washours
12.22.was
For12.22. For the convenience
the convenience of measurement
of measurement and calculation,
and calculation, we set we
set the for
the daytime daytime for each
each day in theday in the experimental
experimental period at period at 12 h:from
12 h: namely, namely,
6 a.m.from
to 66p.m.
a.m. to 6
Hence,p.m.
weHence, wethe
collected collected
average the average
solar solarpower
radiation radiation power
values values
of the two of the two
panels for panels
each for
hour, each
measured
hour, from 6 a.m.from
measured to 6 p.m.,
6 a.m.astoshown
6 p.m.,inasFigure
shown6.inThe energy
Figure power
6. The of the
energy solarof the
power
tracking device
solar was device
tracking observed wasto observed
be higher tothan
be that
higherof the fixed
than thatpanel
of the during
fixed the
paneldaytime.
during the
Additionally,
daytime.the solar radiation
Additionally, power
the solar of the two
radiation power panels both
of the twoincreased
panels both from 6 a.m. to
increased from 6
noon,a.m.
i.e., 12 p.m., but
to noon, i.e., decreased
12 p.m., but from 12 p.m. from
decreased to 6 p.m.
12 p.m. to 6 p.m.
 set the daytime for each day in the experimental period at 12 h: namely, from 6 a.m. to 6
p.m. Hence, we collected the average solar radiation power values of the two panels for
each hour, measured from 6 a.m. to 6 p.m., as shown in Figure 6. The energy power of the
solar tracking device was observed to be higher than that of the fixed panel during the
Energies 2023, 16, 6330 daytime. Additionally, the solar radiation power of the two panels both increased
10 of from
13 6
a.m. to noon, i.e., 12 p.m., but decreased from 12 p.m. to 6 p.m.

Figure 6. Output
Figure power
6. Output comparison
power comparison of
of the two
twosolar
solarpanels.
panels.

ViaVia
i
multiplicationby
multiplication
i
bythe
the time interval,∆t,
time interval, ∆t,the
theexperimental solar
experimental radiation
solar radiation values
values
P and P can be used to express the collected energy, Wtracked and W f ixed , respec-
Ptracked and P fixed can be used to express the collected energy, Wtracked and W fixed , respec-
i tracked i f ixed
tively, as follows:
tively, as follows:
2n i
Ptracked i +1
+ iP+1tracked 2n Pi
f ixedi ++ Pif+ 1
ixed
Wtracked = ∑ Ptracked + Ptracked ∆t, W f ixed = ∑
2n i 2 n i
Pfixed + Pfixed 1
∆t, (19)
Wtracked = 
i =0
2 t , W fixed =  i=0 2 t , (19)
i=0 2 i=0 2
herehere
thethe time
time interval,∆∆t,
interval, t, isisone
one hour,
hour, i.e.,
i.e.,3600
3600s, s,
and Ptracked
and
i
i and P
Ptracked and
i
represent
f ixed P
i the exper-
represent the ex-
fixed
imental values of the designed solar tracking system and the fixed solar panel, respectively,
perimental values of the designed solar tracking 0 system and the fixed solar panel, respec-
at the i-th hour after 6 a.m. For instance, Ptracked and 0P0f ixed are the data for 6 a.m. Since the
tively, at the i-th hour after 6 a.m. For
2n+1
instance,
2n+1
Ptracked and P fixed0
are the data for 6 a.m.
daytime is, at most, 2n hours, Ptracked and P are both zero. According to the data in
2 n +f 1ixed
Figure 6, we could calculate that Wtracked P=tracked
2 n +1
Since the daytime is, at most, 2n hours, and Pfixed
20768.4 J, W
are both zero. According to the
f ixed = 16664.4 J, and
data in Figure 6, we could calculate that Wtracked = 20768.4 J , W fixed = 16664.4 J , and
Wtracked − W f ixed
W tracked − W fixed ≈ 24.6%. (20)
W f ixed ≈ 24.6%. (20)
W fixed
Comparing the experimental values above with the ideal value predicted in Equation (14),
Comparing the experimental values above with the ideal value predicted in Equation
i.e., 26.7%, we could see the agreement of the experimental and theoretical results, which
(14),indicates
i.e., 26.7%, weproposed
that the could see the agreement
device of the to
is accurate enough experimental and
collect most of thetheoretical
solar energyresults,
of
an ideal tracker. Considering that the energy consumed by its mechanical system during
the tracking of the sun has such a negligible value that it can be omitted, the solar tracking
system that we designed and constructed is more efficient in collecting solar energy than
the fixed panel.
Compared to the fixed panel, the rough estimations of the extra cost for materials and
labor and the extra collected energy of the prototype are CNY 3.98 and 0.4161 kilowatt–hours
per year, respectively. Considering the local electricity price, it will take us nearly 14 years
to recover the cost. The extra collected energy of the full-scale solar tracker is nearly
100 times that of the prototype, while the extra cost did not increase that much since some
components chosen for the prototype, such as the MCU, photoresistors, crystal oscillator,
switch, D/A converter, connectors and circuit board, are still available. Additionally, in
the case of an industrial product, the price of the materials will be much lower than the
estimation above. Taking more materials and motors with higher power into account, it
is expected that the extra cost of the full-scale device with the same-sized fixed panel will
be 10 times that of the prototype. Therefore, the period for paying back the extra cost of
the full-scale device will be significantly reduced to 1.4 years, implying that after that time,
the full-sized solar tracker can be more economical in collecting energy than the full-scale
fixed panel.

5. Discussion and Conclusions

In this study, a novel dual-axis solar tracking system was designed and constructed to
enhance solar radiation yield. The proposed structure is simple, as it consists of a small
Energies 2023, 16, 6330 11 of 13

number of components, among which a few gears driven by step motors will make the solar
panel rotate in two directions for solar tracking. The working principles of the structure
and the control algorithm are easy to follow. The photoelectric method was utilized for
solar tracking. An extensive analysis of the total daily energy collection of the system was
performed. According to the results of these measurements, the prototype solar tracker
functioned as expected, specifically for small-sized solar panels. In Shanghai, China, where
the experiment was conducted, 24.6% more energy was obtained from the solar panel that
tracked the sun on two axes when compared with that of the 30◦ tilted fixed-surface panel.
The proposed dual-axis solar tracking system is characterized by a fairly simple and
economical electromechanical setup and ease of installation and operation. Since the base
is designed to rotate in the horizontal direction, thus determining the movement of the
solar panel in 1 degree of freedom, its dimensions should be a bit bigger than those of the
panel to ensure the dynamical stability of the device. With the fabrication cost and the
consumed energy of the base taken into account, the designed solar tracking system is not
applicable for large-sized photovoltaic applications. Instead, it should be suitable for small-
and medium-sized applications, such as individual rooftop solar tracking systems.
Due to the benefits that would be obtained via the utilization of the designed solar
tracking system, certainly fewer PV panels will be applied to it in proportion to the merits
obtained using fixed panels. Hence, the designed system will be more economical regarding
the number of PV panels used, which will decrease the investment costs. The use of this type
of system in Southern China, with significant value specifically in terms of sun exposure
time, has emerged as an important potential regarding energy savings and efficient use.
Our future research will focus on the fabrication of a full-scale model of the device and
its field tests. After obtaining a sufficient amount of field experience, we will calculate the
extra cost and the payback period of the developed system more precisely. As is known, a
strong wind—for instance, a Beaufort 10 wind—can endanger the safety of machinery [35].
For this design, we did not consider the effect of the wind load, as the wind was not
that strong in the area where these experiments were conducted. Still, for the overall
consideration of the proposed system, it will be necessary for us to apply an air velocity
transducer to the device to detect the wind speed, investigate the effect of the wind load
on the drive mechanism and the energy efficiency and add the corresponding wind speed
judgment to the control logic so as to minimize the damage to the device from strong winds.
On this basis, the optimization of the device geometry, epitaxial structure and control
algorithm will also be included in our future research.

Author Contributions: Conceptualization, H.S.; methodology, H.S. and W.S.; software, W.S.; vali-
dation, H.S. and W.S.; formal analysis, H.S. and W.S.; investigation, H.S. and W.S.; resources, H.S.;
data curation, H.S.; writing—original draft preparation, H.S. and W.S.; writing—review and editing,
H.S.; visualization, W.S.; supervision, H.S.; project administration, H.S.; funding acquisition, H.S. All
authors have read and agreed to the published version of the manuscript.
Funding: This research was funded by the National Natural Science Foundation of China, grant
number 11472176.
Data Availability Statement: The data presented in this study are available on request from the
corresponding authors.
Acknowledgments: The authors acknowledge support from the National Natural Science Foundation
of China under grant number 11472176.
Conflicts of Interest: The authors declare no conflict of interest.

References
1. Moghaddam, H.A.; Shorabeh, S.N. Designing and Implementing a Location-Based Model to Identify Areas Suitable for Multi-
renewable Energy Development for Supplying Electricity to Agricultural Wells. Renew. Energy 2022, 200, 1251–1264. [CrossRef]
2. Hafez, A.Z.; Yousef, A.M.; Harag, N.M. Solar Tracking Systems: Technologies and Trackers Drive Types—A Review. Renew.
Sustain. Energy Rev. 2018, 91, 754–782. [CrossRef]
Energies 2023, 16, 6330 12 of 13

3. Das, M.; Akpinar, E.K. Investigation of the Effects of Solar Tracking System on Performance of the Solar Air Dryer. Renew. Energy
2021, 167, 907–916. [CrossRef]
4. Guney, M.S. Solar Power and Application Methods. Renew. Sustain. Energy Rev. 2016, 57, 776–785. [CrossRef]
5. Gonzalo, A.P.; Marugán, A.P.; Márquez, F.P.G. A Review of the Application Performances of Concentrated Solar Power Systems.
Appl. Energy 2019, 255, 113893. [CrossRef]
6. Vieira, R.G.; Guerra, F.K.O.M.V.; Vale, M.R.B.G.; Araújo, M.M. Comparative Performance Analysis between Static Solar Panels
and Single-Axis Tracking System on a Hot Climate Region Near to the Equator. Renew. Sustain. Energy Rev. 2016, 64, 672–681.
[CrossRef]
7. Assadeg, J.; Sopian, K.; Fudholi, A. Performance of Grid-Connected Solar Photovoltaic Power Plants in the Middle East and
North Africa. Int. J. Electr. Comput. Eng. 2019, 9, 3375–3383. [CrossRef]
8. Uzum, B.; Onen, A.; Hasanien, H.M.; Muyeen, S.M. Rooftop Solar PV Penetration Impacts on Distribution Network and Further
Growth Factors—A Comprehensive Review. Electronics 2021, 10, 55. [CrossRef]
9. Quan, S.J.; Li, Q.; Augenbroe, G.; Brown, J.; Yang, P.P. A GIS-based Energy Balance Modeling System for Urban Solar Buildings.
Energy Procedia 2015, 75, 2946–2952. [CrossRef]
10. Baljit, S.S.S.; Chan, H.; Sopian, K. Review of Building Integrated Applications of Photovoltaic and Solar Thermal Systems. J. Clean.
Prod. 2016, 137, 677–689. [CrossRef]
11. Bailek, N.; Bouchouicha, K.; Aoun, N.; EL-Shimy, M.; Jamil, B.; Mostafaeipour, A. Optimized Fixed Tilt for Incident Solar Energy
Maximization on Flat Surfaces Located in the Algerian Big South. Sustain. Energy Technol. Assess. 2018, 28, 96–102. [CrossRef]
12. Kaddoura, T.O.; Ramli, M.A.M.; AL-Turki, Y.A. On the Estimation of the Optimum Tilt Angle of PV Panel in Saudi Arabia. Renew.
Sustain. Energy Rev. 2016, 65, 626–634. [CrossRef]
13. Diponkor, B.; Waliullah, G.M.; Hossain, M.A. Design and Implementation of Low Cost Dual Axis Solar Tracking System Using
Microcontroller. IITM J. Manag. IT 2021, 12, 88–92.
14. Abdollahpour, M.; Golzarian, M.R.; Rohani, A.; Zarchi, H.A. Development of a Machine Vision Dual-Axis Solar Tracking System.
Sol. Energy 2018, 169, 136–143. [CrossRef]
15. Awasthi, A.; Shukla, A.K.; Manohar, M.S.R.; Dondariya, C.; Shukla, K.N.; Porwal, D.; Richhariya, G. Review on Sun Tracking
Technology in Solar PV System. Energy Rep. 2020, 6, 392–405. [CrossRef]
16. Lamoureux, A.; Lee, K.; Shlian, M.; Forrest, S.R.; Shtein, M. Dynamic Kirigami Structures for Integrated Solar Tracking. Nat.
Commun. 2015, 6, 8092. [CrossRef] [PubMed]
17. Poulek, V.; Libra, M. A Very Simple Solar Tracker for Space and Terrestrial Applications. Sol. Energy Mater. Sol. Cells 2000, 60,
99–103. [CrossRef]
18. Mavromatakis, F.; Franghiadakis, Y. A Highly Efficient Novel Azimuthal Heliotrope. Sol. Energy 2008, 82, 336–342. [CrossRef]
19. Kim, Y.; Han, G.Y.; Seo, T. An Evaluation on Thermal Performance of CPC Solar Collector. Int. Commun. Heat Mass Transf. 2008,
35, 446–457. [CrossRef]
20. Kiyak, E.; Gol, G. A Comparison of Fuzzy Logic and PID Controller for a Single-Axis Solar Tracking System. Renew. Wind Water
Sol. 2016, 3, 7. [CrossRef]
21. Li, Y.; Chang, K.; Chang, J.; Yu, B.; Liu, L.; Liu, B.; Zhao, X.; Deng, W. Printed Kirigami Organic Photovoltaics for Efficient Solar
Tracking. Adv. Funct. Mater. 2022, 32, 2204004. [CrossRef]
22. Abdallah, S.; Nijmeh, S. Two Axes Sun Tracking System with PLC Control. Energy Convers. Manag. 2004, 45, 1931–1939. [CrossRef]
23. Sungur, C. Multi-Axes Sun-Tracking System with PLC Control for Photovoltaic Panels in Turkey. Renew. Energy 2009, 34,
1119–1125. [CrossRef]
24. Kentli, F.; Yilmaz, M. Mathematical Modelling of Two-Axis Photovoltaic System with Improved Efficiency. Elektron. Ir Elektrotech-
nika 2015, 21, 40–43. [CrossRef]
25. Syafii; Nazir, R.; Kamshory; Hadi, M. Improve Dual Axis Solar Tracker Algorithm based on Sunrise and Sunset Position. J. Electr.
Syst. 2015, 11, 397–406.
26. Skouri, S.; Ali, A.B.H.; Bouadila, S.; Salah, M.B.; Nasrallah, S.B. Design and Construction of Sun Tracking Systems for Solar
Parabolic Concentrator Displacement. Renew. Sustain. Energy Rev. 2016, 60, 1419–1429. [CrossRef]
27. Laseinde, T.; Ramere, D. Low-Cost Automatic Multi-Axis Solar Tracking System for Performance Improvement in Vertical
Support Solar Panels Using Arduino Board. Int. J. Low-Carbon Technol. 2019, 14, 76–82. [CrossRef]
28. AL-Rousan, N.; Isa, N.A.M.; Desa, M.K.M.; AL-Najjar, H. Integration of Logistic Regression and Multilayer Perceptron for
Intelligent Single and Dual Axis Solar Tracking Systems. Int. J. Intell. Syst. 2021, 36, 5605–5669. [CrossRef]
29. Mousazadeh, H.; Keyhani, A.; Javadi, A.; Mobli, H.; Abrinia, K.; Sharif, A. A Review of Principle and Sun-Tracking Methods for
Maximizing Solar Systems Output. Renew. Sustain. Energy Rev. 2009, 13, 1800–1818. [CrossRef]
30. Radzevich, S.P. Dudley’s Handbook of Practical Gear Design and Manufacture, 3rd ed.; CRC Press, Taylor & Francis Group: Boca
Raton, FL, USA, 2016; pp. 13–115.
31. Johnson, M.A.; Moradi, M.H. PID Control: New Identification and Design Methods, 1st ed.; Ian Kingston Publishing Services:
Nottingham, UK, 2005; pp. 1–106.
32. Soteris, K. Solar Energy Engineering: Processes and Systems, 1st ed.; Academic Press: San Diego, CA, USA, 2009; pp. 49–120.
33. Beling, A.; Xie, X.J.; Campbell, J.C. High-Power, High-Linearity Photodiodes. Optica 2016, 3, 328–338. [CrossRef]
Energies 2023, 16, 6330 13 of 13

34. Tisza, M.; Czinege, I. Comparative Study of the Application of Steels and Aluminium in Lightweight Production of Automotive
Parts. Int. J. Lightweight Mater. Manuf. 2018, 1, 229–238. [CrossRef]
35. Guo, S.J.; Chen, J.H.; Chiu, C.H. Fuzzy duration forecast model for wind turbine construction project subject to the impact of
wind uncertainty. Autom. Constr. 2017, 81, 401–410. [CrossRef]

Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual
author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to
people or property resulting from any ideas, methods, instructions or products referred to in the content.