Design and implementation of an autonomous device with an app to monitor the performance of photovoltaic panels

Abstract

Photovoltaics (PV) utilize sunlight to generate electricity, thus playing a crucial role in generating clean energy and decreasing carbon emissions. Simultaneously, these systems encourage self-sufficiency in energy production. Consequently, it becomes imperative to monitor the performance of photovoltaic systems as an essential method for assessing and confirming these advantages. Precise measurement and analysis of performance data offer researchers and industry experts valuable insights into system effectiveness, power generation trends, as well as their overall ecological influence. The significance of PV monitoring in ensuring and enhancing system performance is emphasized by the research conducted. The ability to collect and analyze real-time data enables operators to identify inefficient modules, as well as shading or other obstacles that could hinder energy production. Furthermore, monitoring systems enable early identification of potential malfunctions, which allows for timely maintenance and repair actions. Ultimately, these practices enhance overall energy generation while extending the longevity of PV installations. This paper presents the design and implementation of a portable electronic device to measure the I-V and P-V curves of photovoltaic panels. This instrument acquires solar radiation, ambient temperature, electric current, and voltage signals from a PV panel via a cellphone through a mobile application. The device, capable of real-time characterization of PV panels up to 20 A and 500 V, features a 240 MHz Tensilica LX6 dual-core processor and 4 MB of storage memory. Experimental tests were carried out in two different geographical locations in Colombia: the city of Puerto Carreño and the city of Bogotá. Among the main results, an efficiency of 13.29 % was obtained for solar radiation of 755.47 W/m 2 and a temperature of 29.60 • C for a monocrystalline PV panel of 405 W.

Full text

Review Article
Design and implementation of an autonomous device with an app to
monitor the performance of photovoltaic panels
A. Ordo˜
nez
a
, J. Urbano
a
, F. Mesa
b,*
, M. Casta˜
neda
c
, S. Zapata
d
, B. Quesada
e
, O. García
a
,
A.J. Aristiz´
abal
a
a
Universidad Jorge Tadeo Lozano, Bogot´
a, Colombia
b
Fundaci´
on Universitaria Los Libertadores, Bogot´
a, Colombia
c
Universidad Cat´
olica de Valparaíso, Valparaiso, Chile
d
Universidad EIA, Envigado, Colombia
e
Earth System Sciences Program, Faculty of Natural Sciences, Universidad del Rosario, Bogot´
a, Colombia
ARTICLE INFO
Keywords:
Solar Photovoltaics
PV monitoring
Solar Energy
Renewables
PV performance
ABSTRACT
Photovoltaics (PV) utilize sunlight to generate electricity, thus playing a crucial role in generating clean energy
and decreasing carbon emissions. Simultaneously, these systems encourage self-sufciency in energy production.
Consequently, it becomes imperative to monitor the performance of photovoltaic systems as an essential method
for assessing and conrming these advantages. Precise measurement and analysis of performance data offer
researchers and industry experts valuable insights into system effectiveness, power generation trends, as well as
their overall ecological inuence. The signicance of PV monitoring in ensuring and enhancing system perfor-
mance is emphasized by the research conducted. The ability to collect and analyze real-time data enables op-
erators to identify inefcient modules, as well as shading or other obstacles that could hinder energy production.
Furthermore, monitoring systems enable early identication of potential malfunctions, which allows for timely
maintenance and repair actions. Ultimately, these practices enhance overall energy generation while extending
the longevity of PV installations. This paper presents the design and implementation of a portable electronic
device to measure the I-V and P-V curves of photovoltaic panels. This instrument acquires solar radiation,
ambient temperature, electric current, and voltage signals from a PV panel via a cellphone through a mobile
application. The device, capable of real-time characterization of PV panels up to 20 A and 500 V, features a
240 MHz Tensilica LX6 dual-core processor and 4 MB of storage memory. Experimental tests were carried out in
two different geographical locations in Colombia: the city of Puerto Carre˜
no and the city of Bogot´
a. Among the
main results, an efciency of 13.29 % was obtained for solar radiation of 755.47 W/m
2
and a temperature of
29.60 ◦C for a monocrystalline PV panel of 405 W.
1. Introduction
Photovoltaic energy is inherently sustainable, producing electricity
without direct greenhouse gas emissions or other harmful pollutants. It
plays a vital role in mitigating climate change by reducing carbon di-
oxide (CO
2
) emissions, thus contributing to global efforts to combat
global warming. As a result, PV energy stands as a promising solution to
achieve decentralized electricity, carbon neutrality and minimize the
environmental footprint of electricity generation (Hosseini et al., 2023).
As the adoption of PV technology continues to increase worldwide, it
is crucial to understand and address potential environmental concerns
throughout the entire lifecycle of these renewable energy systems. It´s
necessary to identify and evaluate key environmental factors affecting
the performance of photovoltaic panels (Jathar et al., 2023). There is a
crucial need for accurate and reliable testing and performance evalua-
tion of PV modules, providing valuable insights into their efciency and
behavior under varying environmental conditions. The PV industry en-
compasses the design and implementation of solar simulators, as well as
automatic I-V curve acquisition systems, which play a pivotal role in
streamlining the testing process and generating critical data for PV
module evaluation (Piccoli Junior et al., 2023). One of the main prob-
lems of PV technology pertains to the need for a fast, efcient, and
* Corresponding author.
E-mail address: fredy.mesa@libertadores.edu.co (F. Mesa).
Contents lists available at ScienceDirect
Energy Reports
journal homepage: www.elsevier.com/locate/egyr
https://doi.org/10.1016/j.egyr.2024.07.062
Received 22 April 2024; Received in revised form 18 July 2024; Accepted 29 July 2024
Energy Reports 12 (2024) 2498–2510
2352-4847/© 2024 The Author(s). Published by Elsevier Ltd. This is an open access article under the CC BY license ( http://creativecommons.org/licenses/by/4.0/ ).

accurate method to characterize the current-voltage (I-V) curves of
photovoltaic (PV) arrays. I-V curves are essential in evaluating the
performance, efciency, and health of PV arrays, as they provide crucial
information about the electrical behavior of the system under different
operating conditions. However, traditional methods of acquiring I-V
curves can be time-consuming and impractical, especially when dealing
with large-scale PV installations (Chen et al., 2020). When photovoltaic
panels are subject to defects or malfunctions, the traditional I-V curve
measurement methods prescribed by the IEC 60891 standard may lead
to inaccurate or unreliable results. These defects can manifest in various
forms, such as partial shading, hotspots, module degradation, or elec-
trical mismatches within the PV panel. When conventional I-V curve
measurement techniques are applied to defective panels, the resulting
data may not represent the true electrical characteristics of the PV
module (Li et al., 2021; Liu et al., 2021; Ma et al., 2020).
Various internal and external factors can lead to defects or mal-
functions in photovoltaic panels, compromising their functionality and
reducing their energy generation capacity. Full I-V characteristics pro-
vide a comprehensive understanding of the electrical behavior of the PV
panel under different operating conditions, allowing for a more detailed
fault analysis. PV researchers employ machine learning techniques to
process and interpret the full I-V characteristics for fault diagnosis.
Machine learning algorithms, such as support vector machines, decision
trees, or neural networks, are trained with data from healthy and faulty
photovoltaic panels to recognize patterns and correlations associated
with specic faults (Ma et al., 2020; Li et al., 2021; Liu et al., 2022; Ma
et al., 2021).
As the demand for clean and sustainable energy sources grows,
enhancing the efciency and output of PV systems becomes paramount
(Baghel et al., 2023). tackle the intricate task of evaluating and opti-
mizing the albedo (reectivity) and tilt angle of photovoltaic panels to
maximize their performance. By comprehensively examining how the
reective properties of surrounding surfaces and the orientation of
panels impact energy generation, the research seeks to identify the most
suitable albedo levels and tilt angles for different conditions. By
addressing these variables, the article contributes to the ongoing efforts
to boost the overall effectiveness and viability of solar PV systems,
leading to improved energy capture, system productivity, and environ-
mental benets. On the other hand, there is a key challenge related to
the integration of photovoltaic (PV) technology into building facades. As
the integration of renewable energy sources into architectural designs
gains momentum, there is a growing need to evaluate the performance
of large-sized PV modules specically designed for façade integration. In
(Assoa et al., 2023), the authors address the intricate issue of how such
modules perform under real-world conditions, considering factors like
solar exposure, shading, and architectural aesthetics. By analyzing the
performance of these specialized PV modules, the research seeks to shed
light on their energy generation capabilities and overall efciency when
integrated into building facades. The ndings of this study are crucial for
architects, engineers, and policymakers striving to create sustainable
and energy-efcient buildings by seamlessly incorporating solar tech-
nology into their designs, thus contributing to the ongoing progress in
renewable energy adoption and sustainable urban development.
The study (Yang et al., 2023) decided to optimize the performance of
PV modules and solar cells through the innovative application of a
hybrid and efcient chimp algorithm. By employing this algorithmic
approach, the research aims to ne-tune the parameters that signi-
cantly impact the energy conversion efciency of PV systems, thus
improving their overall power output. The study’s ndings hold signif-
icance for the renewable energy sector, offering insights into advanced
optimization techniques that can be applied to maximize the energy
yield of solar installations. This research contributes to the ongoing
pursuit of efcient and effective solar energy utilization, paving the way
for increased adoption of clean and sustainable power generation
technologies.
Other studies (Soler-Castillo et al., 2023; Padilla et al., 2022; Toledo
et al., 2023) address the intricate issue of modeling and simulating the
dynamics of I-V (current-voltage) curves, which are fundamental in
characterizing the behavior of solar cells under varying conditions. By
developing an approach to simulate the complex dynamics of these
curves, the research aims to enhance the predictability of PV system
performance. The insights gained from these studies are invaluable for
optimizing energy capture, forecasting power generation, and guiding
system design and maintenance strategies.
PV technology presents a necessary review of various circuit topol-
ogies employed for I-V curve tracing, a technique essential for assessing
the efciency and health of PV systems. By analyzing the strengths and
limitations of different tracer topologies, researchers aim to guide the
development of more effective and accurate measurement techniques.
Researchers contribute to the advancement of photovoltaic technology
by offering insights into the instrumentation and methodologies
required to better comprehend the behavior of solar cells, ultimately
leading to improved system design, monitoring, and maintenance
practices (Zhu and Xiao, 2020; Zhang et al., 2022; Olayiwola and Choi,
2023; Blakesley et al., 2020).
The challenge of developing I-V curve tracers is assumed by the
necessity of offering portable and modern devices to improve the ways
of evaluating PV performance. In (Casado et al., 2022) authors focused
on the development and implementation of a novel I-V (current-voltage)
curve tracer using Raspberry Pi, a versatile and affordable single-board
computer. The main problem tackled is the need for a cost-effective and
accessible method to trace I-V curves of PV modules, enabling a better
understanding of their behavior under varying conditions. By utilizing
Raspberry Pi as the core of the tracer, the research provides a practical
solution for researchers, engineers, and enthusiasts to measure and
analyze I-V curves in a user-friendly manner. This innovation has the
potential to democratize the process of characterizing PV modules,
contributing to the broader goals of advancing solar energy technology
and facilitating its widespread adoption. On the other hand, the study
(Jos´
e Mu˜
noz-Rodríguez et al., 2023) focuses on enhancing the perfor-
mance analysis of rooftop PV systems by introducing new parameters
derived from monitored data, in alignment with the guidelines provided
by IEC 61724 standard. The primary challenge addressed is the need for
more comprehensive and rened methods to assess the output and ef-
ciency of rooftop PV installations. By developing novel parameters and
metrics based on monitored operational data, the research aims to
provide a more accurate and comprehensive understanding of system
performance. This approach holds importance for stakeholders ranging
from solar energy industry professionals to policymakers, enabling
better decision-making, and systems optimization, and ultimately
fostering the integration of rooftop photovoltaics as a viable and ef-
cient renewable energy solution.
Several authors research about PV failures. In (Belhaouas et al.,
2024), a study on PV module failures reveals various challenges and
performance issues faced by photovoltaic systems. The analysis included
six types of PV modules with different outdoor exposure durations under
the Mediterranean climate. Most PV modules did not meet their ex-
pected yearly degradation rates (except for one type) with 17 years of
outdoor exposure showing a very low rate compared to the expected
rate. Most common failure modes include detached junction boxes due
to poor adhesion as well as degradation in electrical performance.
Another studies (Nieto-Morone et al., 2024; ¨
Ozkalay et al., 2024)
proposed several innovative solutions for improving PV module per-
formance and lifespan including, among others: utilizing advanced
diagnostic tools such as electroluminescence imaging and infrared
thermography for early detection of faults (e.g. cracks, hotspots,
moisture-induced degradation); implementing stringent quality control
measures throughout the manufacturing process; adopting international
standards (IEC, ISO) and adapting them to local conditions; establishing
national certication and training programs for PV installers and de-
signers can ensure high-quality installation practices; and, developing
circular supply chains for PV modules, including recycling,
A. Ordo˜
nez et al.
Energy Reports 12 (2024) 2498–2510
2499

refurbishment, and re-certication, can extend the lifespan of PV sys-
tems. This approach not only improves sustainability but also enhances
the overall economic viability of solar energy.
Multiple authors have focused on measuring and adopting solar
community methods. In (´
Angel-Antonio Bayod-Rújula, 2014), the in-
efciency and malfunction of traditional Maximum Power Point
Tracking (MPPT) techniques in photovoltaic systems under rapid irra-
diance changes and partial shading conditions were addressed. The
paper suggests a fresh MPPT method that incorporates photodiodes as
irradiance sensors, which enables accurate measurement of shadow
coverage, precise monitoring of light exposure levels, MPPT algorithm
adjustment to prevent malfunctions; all leading to more efcient energy
output optimization. This technique underwent theoretical examination
with conventional strategies while displaying improved efciency
feasibility for practical applications. The authors introduce an efcient
technique in (Kumar and Nayak, 2024) to identify and pinpoint faults
occurring within PV arrays, utilizing the cumulative sum (CUSUM) of
string current variations. This method relies solely on measuring indi-
vidual string currents and can promptly detect a range of fault types -
such as low mismatch or high resistance faults - with great precision
within 5 ms. Moreover, it excels at accurately distinguishing between
real faults versus non-fault situations like partial shading; thereby sur-
passing typical protective devices and other pre-existing techniques to
signal when damage occurs.
This article presents an innovative development of a portable and
fast characterizer of the performance of photovoltaic panels. The device
allows for the acquisition of solar radiation, ambient temperature,
electric current, and voltage generated by the solar panel to plot I-V and
P-V curves. Additionally, a mobile application has been developed that
enables the conguration and operation of the device over the internet
through any mobile phone. The paper is structured in a way that rst
introduces the theoretical model for the characterization of photovoltaic
panels, then describes the experimental device design along with the
methodology, and nally presents the obtained results and conclusions.
2. Device implementation
Accurate measurements and efcient system management are crucial
when developing our device to monitor photovoltaic panel performance.
Therefore, it is essential to understand and control ambient conditions
such as temperature. Temperature signicantly impacts the efciency of
photovoltaic panels, making it necessary to measure both the ambient
temperature and surface temperature of each panel with calibrated
sensors. The semiconductor’s thermal activity increases at higher tem-
peratures which can reduce a panel’s effectiveness in generating energy
output. To overcome this challenge, implementing cooling mechanisms
or selecting materials that have better thermal management properties
becomes imperative for mitigating these effects on solar cells’ overall
efcacy.
Monitoring solar irradiance is crucial for accurately measuring the
power output of photovoltaic panels. Pyranometers and photodiodes are
capable of capturing the intensity levels of solar radiation falling on a
panel, ensuring precise readings. To maintain accuracy in these mea-
surements, diligent upkeep must be followed to avoid obstruction or
interference with sensor function along with consistent calibration
against reference standards. Proper placement techniques that eliminate
turbulence should also be employed as they contribute to vital data
accuracy while recording wind conditions.
To shield the equipment from environmental factors, it is advisable
to incorporate weather-resistant sensors and protective housings.
Additionally, monitoring of soiling on the surface of PV panels caused by
dust and dirt should be done through visual inspections or soiling sen-
sors as this could cause signicant efciency reduction. The upkeep of
optimal panel performance necessitates periodic cleaning exercises
alongside assessments for any signs of accumulating dirt.
To guarantee precision, the sensors necessitate a series of
fundamental steps in their calibration process. The correct measure-
ments for temperature sensors are achieved by calibrating them against
an established standard thermometer amidst controlled circumstances.
Similar to this method is how irradiance sensors operate via calibration
using either STC-controlled conditions or other recognized standards as
they’re compared with reference pyranometers for accuracy verication
purposes respectively. Finally, electrical calibration requires state-of-
the-art multimeters and solid power sources functioning as references
while assessing current-voltage sensing devices and power meters dur-
ing testing processes conducted accordingly.
To maintain accuracy in the performance monitoring system, reca-
libration must be conducted regularly either as per manufacturer in-
structions or specied intervals. To account for variances in
temperature, wind conditions and irradiance; environmental compen-
sation algorithms can also be utilized to adjust measurements. More-
over, periodic validation of the entire system against a reference point or
under standard conditions should be carried out to ensure that long-term
precision is sustained without fail.
When deploying the monitoring device, it is crucial to merge all
sensors into a coherent data acquisition infrastructure that ensures
synchronized data logging and comprehensive performance analysis. To
swiftly detect any issues in system performance, real-time data pro-
cessing capacities are necessary; meanwhile, sturdy storage options for
information along with advanced analytical tools make evaluating long-
term changes simple while supporting predictive maintenance practices.
Additionally, having an easily navigable user interface helps users
interpret PV panel-related data without difculty by providing clear
graphical displays and remote access functionality through intuitive
dashboards.
Developing a PV panel performance monitoring device that observes
ambient conditions and follows a strict calibration process can yield
precise, dependable, and practical data to improve the effectiveness and
sustainability of photovoltaic systems.
If one tackles the surrounding circumstances and meticulously fol-
lows a calibration procedure, it is possible to create a monitoring gadget
for PV panels that offers precise, trustworthy, and practical information.
Such data can optimize both efciency and durability of photovoltaic
systems.
The performance measurements of photovoltaic panels can be
notably affected by environmental conditions present at testing sites,
where temperature, solar irradiance and dust levels have an essential
impact on the precision and dependability of collected data.
Fluctuations in ambient and panel surface temperatures can impact
the photovoltaic panels’ efciency. The effectiveness of these solar de-
vices is generally reduced by higher temperatures due to increased in-
ternal resistance, resulting in lower power output. Conversely, cooler
temperatures tend to boost their performance signicantly; hence tem-
perature sensors are vital tools for tracking uctuations and adjusting
relevant calculations correspondingly.
The performance of solar panels greatly depends on the amount and
strength of sunlight they receive - known as Solar Irradiance. Accurate
measurement and evaluation are crucial for gauging their efcacy under
varying conditions caused by cloud cover, time of day or seasonal
changes. Pyranometers and photodiodes measure irradiance levels but
must be calibrated periodically to ensure precision in readings.
The accumulation of dust and dirt on photovoltaic panels can block
sunlight, signicantly lowering their efciency. Maintaining optimal
performance requires regular monitoring and cleaning. One may use
visual inspections or soiling sensors to assess panel contamination levels.
To preserve the precision of sensors utilized for environmental
monitoring, regular calibration is essential. This entails subjecting
sensor readings to controlled conditions and comparing them against
established benchmarks. Moreover, maintaining both the monitoring
system and sensors on a routine basis guarantees reliable performance
measurements are obtained.
Accurate and reliable performance measurements of photovoltaic
A. Ordo˜
nez et al.
Energy Reports 12 (2024) 2498–2510
2500

panels can be achieved by meticulously examining and adjusting for
environmental factors. This enables an enhanced comprehension and
enhancement of the longevity as well as productivity of these panels
under real-life circumstances.
2.1. Theoretical model
Fig. 1shows the circuit model of a photovoltaic solar cell. Iph current
source represents the generated photocurrent, D diode represents the P/
N connection, Rs represents the serial number of the resistive device,
which is related to the strength of the material and electrical contact,
Rsh represents the current consumption of the resistive parallel device
and is related to the current consumption in the device volume
(“Practical Handbook of Photovoltaics, 2023).
Photovoltaic cells are often presented in the form of cells connected
in series to increase the output voltage to a desired value, and in parallel
to increase the current that a device can provide depending on the power
demand (Perez, 2023). Each solar cell behaves like a p/n rectifying
diode in the dark and generates a photocurrent when exposed to light.
A mathematical model of the equivalent circuit of a solar photovol-
taic cell is described below (Duru, 2006; Gow and Manning, 1999;
Seguel).
Applying current Kirchhoff’s laws, the current at the PV module
terminals is:
Ipv =Iph −ID −IRsh (1)
Where Ipv is the current generated by the system, Iph is the photocurrent
generated, ID is the current owing through the diode, and IRsh is the
current owing through the shunt resistor.
Eq. (2) is related to the photogenerated Iph current, which depends
on standard irradiance and temperature (Seguel). G is the radiation of
the existing station, T and Tr are the current station temperature and the
reference temperature. Δi is the temperature coefcient of current and
Isc is the short-circuit current of the battery at standard temperature.
Iph =G
1000 [Isc +Δi(T−Tr) ] (2)
The ID diode saturation current can be found by the Shockley
equation:
ID =Isat[exp(VD
Vt )]−1 (3)
where Isat is the saturation current of the diode, VD is the diode voltage,
Vt is the thermodynamic voltage of the diode, k is Boltzmann’s constant,
Tc is the temperature in the pn junction, and q is the electron charge
value corresponding to the diode ideality factor.
Applying Kirchhoff’s voltage law, we get:
VD =Vpv +Rs ∗Ipv (4)
We replace VD and clear IRsh, we have:
IRsh =Vpv +Rs ∗Ipv
Rsh (5)
And:
Vt =A∗K∗Tc
q(6)
The electrical characteristics of a PV module are given in terms of
current and voltage output (Ipv – Vpv). so:
Ipv =Iph −Isat[exp(VD
Vt )−1]−Vpv +Rs ∗Ipv
Rsh (7)
The Isat parameter must be evaluated as an open circuit, where:
Ipv =0 (8)
Vpv =Voc (9)
Under these conditions, it is rewritten as follows:
0=Iph −Isat[exp(Voc
Vt )−1]−Voc
Rsh (10)
By simplication, we get:
Isat =Iph −Voc
Rsh
exp(Voc
Vt )−1
(11)
The reverse saturation current also depends on temperature (Kumar
and Nayak, 2024):
Isat =Isatr ∗(Tc
Tr)3
∗exp[q∗Eq
K∗A][1
Tr −1
TC](12)
Tc =Ta + (0,2∗G)(13)
where Isatr is the reference saturation current, Tc is the battery tem-
perature, Tr is the reference temperature, Eq is the bandgap energy, q is
the electron charge, and K is the Boltzmann constant. A corresponds to
the ideality factor of the diode, Ta is the ambient temperature, and G is
the solar radiation.
In short, we have:
Ipv =Icc (14)
Vpv =0 (15)
Under these conditions, we have:
Icc =Iph −Iph −Voc
Rsh
exp(Voc
Vt )−1
−Icc ∗Rs
Rsh (16)
has been rewritten again:
Iph =
Icc(1−Rs
Rsh)−Voc⎛
⎜
⎜
⎝
exp(Icc∗Rs
Vt )−1
exp(Voc
Vt )−1
⎞
⎟
⎟
⎠
1−⎛
⎜
⎜
⎝
exp(Icc∗Rs
Vt )−1
exp(Voc
Vt )−1
⎞
⎟
⎟
⎠
(17)
Then the Rs and Rsh resistances are (“Practical Handbook of Pho-
tovoltaics, 2023):
Rs =[1−FF
FFo](Voc,STC
Isc,STC )(18)
Fig. 1. Solar cell electrical model.
A. Ordo˜
nez et al.
Energy Reports 12 (2024) 2498–2510
2501

Rsh =FFo(Vo +0,7)
(1−FF
FFo)Vo (Voc,STC
Isc,STC )(19)
where FF is the ll factor and FFo is the ideal ll factor (Green, 2023). If
Rs=0, then:
FF =Pm
Voc ∗Isc =Vm ∗Im
Voc ∗Isc (20)
FFo =Vo −Ln(Vo +0,72)
Vo ;with Vo =Voc
KTc (21)
Substituting Eqs. (11) and (17) into Eq. (7), the following Eq. (22) y
(23)are obtained:
Ipv =Iph −Iph −Voc
Rsh
exp(Voc
Vt )−1
⎛
⎝expVpv+Ipv∗Rs
VT −1⎞
⎠−Vpv +Ipv ∗Rs
Rsh (22)
Ipv =
Icc(1+Rs
Rsh)−Voc⎛
⎜
⎜
⎝
exp(Icc∗Rs
Vt )−1
exp(Voc
Vt )−1
⎞
⎟
⎟
⎠
1−⎛
⎜
⎜
⎝
exp(Icc∗Rs
Vt )−1
exp(Voc
Vt )−1
⎞
⎟
⎟
⎠
−
⎛
⎜
⎜
⎜
⎜
⎜
⎜
⎜
⎜
⎜
⎜
⎜
⎝
Icc(1+Rs
Rsh)−Voc⎛
⎜
⎜
⎝
exp(Icc∗Rs
Vt )−1
exp(Voc
Vt )−1
⎞
⎟
⎟
⎠
(exp(Voc
Vt )−1)−(exp(Icc∗Rs
Vt )−1)
−Voc
(exp(Voc
Vt )−1)∗Rsh
⎞
⎟
⎟
⎟
⎟
⎟
⎟
⎟
⎟
⎟
⎟
⎟
⎠
⎛
⎝expVpv+Ipv∗Rs
VT −1⎞
⎠−Vpv +Ipv ∗Rs
Rsh (23)
The battery temperature (Tc) is determined based on the operating
ambient temperature (Ta) and the NOCT (Battery Nominal Operating
Temperature) specied by the manufacturer.
Tsc =Ta+NOCT −20
800 G(W/m2)(24)
The following equations correspond to the parallel and series
impedance of PV modules, where Np is the number of cells connected in
parallel, Ns is the number of cells connected in series, Rp is the resistance
connected in parallel, and Rs is the series of resistors connected in
parallel.
Rsht =(Np
Ns)Rp (25)
Rst =(Ns
Np)Rs (26)
2.2. Experimental design
2.2.1. Device requirements
The device is aimed at acquiring and processing signals. Addition-
ally, the device will have a Platform as a Service architecture.
For the design of the device, the following considerations are taken
into account:
•The device must be capable of characterizing panels of up to 250
watts.
•It should be able to respond to requests via the HTTP protocol.
•The power stage must be isolated from the control stage.
•Operating voltages on the microcontroller must not exceed 3.3 V.
•Protection and regulation elements must be considered.
•Signals should be ltered if necessary.
•The device should create an Access Point for characterization
purposes.
•The device should be compact and have a protective casing.
For these purposes, an ESP32 microcontroller from Sparkfun has
been selected. It features a WiFi module capable of connecting to mobile
devices. It has 28 GPIO pins and also provides support for low-energy
Bluetooth connections. The versatility of the ESP32 shines in IoT proj-
ect execution. Its operating range is from 2.2 to 3.3 V, and it has low
power consumption (less than 100 mA). There are 18 analog-to-digital
converter (ADC) pins, 2 digital-to-analog converter (DAC) pins, and 3
SPI interfaces. The ESP32 microcontroller meets the industry-standard
SDIO card specication Version 2.0 and allows a host controller to ac-
cess the SoC, using de SDIO bus interface and protocol. Table 1 displays
the basic specications of the ESP32 microcontroller.
2.2.2. Sensors
- Current sensor
The ACS712 current sensor provides cost-effective and precise
solutions for detecting alternating current (AC) and direct current
(DC) in the industry. The device consists of a linear Hall effect sensor
with a copper conduction path located near the surface of the chip.
The supply voltage is 5 V, and the output voltage is proportional to
AC or DC currents. The ACS712 sensor has low noise in the analog
signal and a sensitivity range of 66–185 mV/A. It can measure up to
20 A of input current from a load.
- Voltage sensor
It’s essentially a voltage divider consisting of two resistors in se-
ries, powered by 7.4 V, designed to measure voltage from 0 V up to
50 V. This voltage is scaled from 0 V to 3.3 V so that it can be sent to
the ESP32 microcontroller.
- Temperature sensor
The NTC 10 K sensor is an ideal temperature sensor for measuring
ambient temperature. It comes with a stainless-steel housing and is
waterproof. It provides a linear output proportional to temperature,
starting at 0 volts at 0 degrees and a 10 mV output voltage change for
Table 1
Technical specications of the ESP32 microcontroller.
Parameter Description
Processor Dual-core Tensilica LX6
Clock Frequency Up to 250 MHz
SRAM 520 kB internal
GPIO 28
Operating Range From 2.2 V to 3.3 V
DAC 2
Flash Memory 4MB
ADC 17–18
Aditional features WiFi, Bluetooth, touch sensor, LI ion
A. Ordo˜
nez et al.
Energy Reports 12 (2024) 2498–2510
2502

every degree change, with a measurement range of −20–105 degrees
Celsius.
- Irradiance sensor
The SP Lite2 is a solar radiation sensor primarily used for measure-
ments in photovoltaic modules and can be used under various weather
conditions. Its operating principle relies on a photodiode that generates
an output voltage proportional to incoming radiation, with sensitivity
varying with the cosine of the incidence angle of the radiation. This
pyranometer meets the following international standards: EN
63000:2018 and EN 61326:2013. Table 2 shows the main technical
specications of the SPLite 2 sensor.
2.2.3. Signals conditioning
As a voltage regulator, we used a switching step-up module XL6009,
which allows for increasing voltages in the range of 5 V to 35 V with a
current load of 3 A. We also employed the XL4015 module, which is a
switching step-down voltage regulator, enabling voltage reduction in
the range of 1.25 V to 35 V with a current load of 5 A. For analog-to-
digital conversion, we utilized the ADS1015 module, providing a 12-
bit resolution at a rate of 3300 samples per second using the I2C bus.
Additionally, its programmable gain amplier allows conguration of
up to 16x for signals that are too small.
The AD620 module is an instrumentation amplier capable of
amplifying signals in the microvolt/millivolt range with gains ranging
from 1.5 to 1000 times the input signal. The AD620 module meets
JEDEC standards MS-001 and MS-012-AA. As an energy storage system,
we selected a 3500 mAh, 7.4 V LiPo battery with a module that allows
for monitoring and measuring its charge level.
MOSFET transistors are ideal for high-frequency switching since they
are not adversely affected at high speeds, which allowed us to determine
a frequency of 1.25 KHz for the PWM signal to be implemented. We
selected a P-type MOSFET transistor, specically the IRFZ44N, with a
maximum Drain-Source voltage of −100 V and a maximum drain cur-
rent of −13 A. To control the switching of the MOSFET referenced by the
panel and the 7.4 V battery, we use PWM generated by the ESP32, in
order to bring the MOSFET into the cut-off and saturation regions with
voltage variations. The output voltage of the microcontroller is refer-
enced from 0 V to 3.3 V; however, to drive the MOSFET, a higher voltage
is required. This is why it was necessary to implement an opto-coupler
stage to increase the signal coming from the microcontroller and thus
exceed the gate-source voltage (VGS) required for MOSFET switching.
Fig. 2 shows the initial switching circuit for characterizing the solar
panel.
This switching circuit features an IRFZ44N MOSFET that receives a
7.4 V PWM signal, allowing the MOSFET to enter the cut-off and satu-
ration states without any issues. The load in this case is the same
MOSFET, as it dissipates power when in saturation. Since it operates in
switching mode, heat dissipation is much lower compared to when it is
in the ohmic region. Nevertheless, a heatsink with forced ventilation is
added to reduce the impact of high currents.
Located between the drain and source is the panel to be character-
ized, which is connected in series with a 20 A ACS712 current sensor.
Since this sensor has a copper connector, it acts as a shunt resistor,
stabilizing the system. In parallel with the panel and the current sensor,
an RC branch is placed as a low-pass lter to reduce noise generated by
switching and obtain a more accurate measurement. Similarly, there is a
lter for the current signal.
Initially, this circuit does not yield very good results, so certain im-
provements are required. Fig. 3 shows the initial results of this circuit.
Fig. 3 displays the output variables (I and V) in their unconverted
form, meaning their analog values are shown. In this case, the output is
nearly linear, which could be attributed to the fact that the carrier signal
(PWM) is not being ltered. It is necessary to improve the ltering as the
carrier signal is at 1.25 KHz. Finally, a 4700µF capacitor is used to
smooth and lter the carrier signal, but it was observed that the
capacitor triggers a very high current peak when it enters a short circuit,
which could potentially cause permanent damage to the MOSFET. To
prevent this, current ltering is added, thus completing a second-order
lter. Fig. 4 shows the electronic model created in Proteus and KiCad
software for the second-order lter. Fig. 5a depicts the obtained I-V
curve.
With the assistance of the second-order lter, the MOSFET no longer
sustains damage, and the output signal closely matches the natural
behavior of the PV panel. As the nal outcome, the design of a PCB
circuit is initiated.
2.2.4. Printed circuit board design
The ESP32 serves as the central device in the PCB design, where all
input and output signals converge. Each signal has its respective ltering
stage. Since the ESP32 has a maximum output of 3.3 V, it is opto-coupled
to transmit the signal at 7.4 V, surpassing the required gate-source
voltage (VGS) for proper switching. Additionally, this helps ensure
that if the MOSFET tends to draw current due to parasitic capacitances,
the current is supplied by the 7.4 V battery rather than the ESP32,
preventing potential damage.
The ACS712 current module employs a resistor arrangement to
Table 2
Technical specications of the SPLite 2 sensor.
Parameter Description
Spectral range 400 nm – 1100 nm
Sensitivity (nominal) 72 uV/Wm
−2
Response time Lower than 1 s
Maximum radiation 2000 Wm
−2
Directional error ±5 % for angles above 80 degrees
Fig. 2. Initial circuit to PV panel characterization.
Fig. 3. PCB of the control circuit.
A. Ordo˜
nez et al.
Energy Reports 12 (2024) 2498–2510
2503

calculate the current passing through the sensor within a voltage range
of 0 V to 5 V. The voltage sensor utilizes a series circuit principle for
voltage measurement. The resistances were calculated to establish a
measurement window between 0 V to 5 V, which can be read by the
ADS1015. The instrumentation amplier module allows for the ampli-
cation of the radiation sensor’s voltage. It was congured to amplify
the signal within the 0 V to 5 V range to match the voltage levels of the
ADS1015. Fig. 5b shows the printed board circuit and, Fig. 5c depicts the
assembled sensors, connectors, and batteries.
2.2.5. Housing design
Since the device must be sturdy enough for regular use, it is neces-
sary for it to have a casing that protects it from impacts, and dust, and
isolates the circuit from human contact. This also ensures better
component security, preventing wear on the cables or loss of
components.
For the design, Autodesk Inventor 2022 software was chosen to
create a 3D model that ensures space for each component. The model is
intended to be user-friendly, so a small, easily transportable box-like
structure was chosen as a reference. This design allows for easy
manipulation of the system. Fig. 6 shows the implemented casing design.
2.2.6. Software design
The development of the software application for the device was
carried out using a service-oriented architecture, which is a communi-
cations technique that has grown due to the constant evolution of device
availability. A service is a piece of software that provides functionality to
other pieces within or outside the system. The software components that
consume or query these services are known as clients and can be any-
thing from a web page to a mobile application or even another service
that requires it to complete functionality.
A programming framework is a toolkit that assists the developer in
executing projects in an agile manner, utilizing libraries and native APIs
of each platform. Each framework manages its own structure with a
specic language (Java, C#, Javascript, etc.). The framework utilized in
this research was Cross-Platform, characterized by being a development
approach where multiple platforms can be targeted from a common
codebase through the use of different frameworks (Ionic, Flutter,
Xamarin.Forms, React Native, etc.). Xamarin.Forms was the framework
used for developing the application for the following platforms:
Android, Linux, iOS, Windows 10, and WPF.
3. Device calibration
3.1. Current sensor calibration
The calibration of the ACS712 current sensor was performed using a
20 A source connected in parallel to the sensor. Approximately 100
samples were taken, varying the current provided by the source, which
was reected in voltage variations in the sensor. Using the ADS1015, the
voltage value calculated by the module’s library was obtained for the
various points generated during the measurement, from which the
representative equation of the found straight line was generated. Fig. 7
Fig. 4. Electronic circuit using the second-order lter.
Fig. 5. (a) Initial results of the PV characterization, (b) 3D control circuit, and (c) Hardware of the assembled sensors, connectors, and batteries.
A. Ordo˜
nez et al.
Energy Reports 12 (2024) 2498–2510
2504

shows the calibration performed on the current sensor, using the 20 A
source, and the calibration equation for the current sensor.
3.2. Voltage sensor calibration
A voltage source with a range of 0 V to 50 V was used for calibrating
the voltage sensor. Approximately 116 samples were taken, varying the
voltage provided by the source. These variations were reected in
voltage changes in the sensor. Using the AD620, the analog value of the
various points generated during the measurement was obtained, and
from this, the representative equation of the found straight line was
generated.
3.3. Temperature sensor calibration
The temperature sensor used is a 10KOhm NTC thermistor. To
determine the temperature value, there is an array of resistors that
provides a reference, i.e., a xed value to compare with to detect the
voltage change. This change represents the temperature value in ohms.
However, it is necessary to arrive at a temperature unit, which is
described by the Steinhart-Hart equation.
The Steinhart-Hart equation describes a transfer function of the ideal
linear model, which is limited to a specic temperature range depending
on the manufacturing and model of the thermistor. The transfer function
helps eliminate this limitation.
3.4. Solar irradiance sensor calibration
The SP Lite2 pyranometer was calibrated using a certied reference
Eppley Serie: 35136f3. Fig. 8(a-b) shows the electronic calibration
process and depicts the certied pyranometer and the SP Lite2.
The pyranometer calibration process was carried out under real solar
radiation conditions. Fifty samples were taken over a span of 4 hours to
correlate solar radiation data at different times of the day. Using a
multimeter, the analog voltage value of various points generated during
the measurement was obtained, and from this, the representative
equation of the found straight line was generated. A measurement error
of 5.1 % was obtained between the standard pyranometer and the SP
Lite2 pyranometer.
3.5. Comparative cost analysis
1. Hardware Costs
•Microcontrollers/SoCs: The choice of microcontroller or System on
Chip (SoC) affects the overall cost. Popular choices include Arduino,
Raspberry Pi, and ESP32.
o Arduino Uno: $25
o Raspberry Pi 4: $35
o ESP32 (selected): $6
•Sensors: To monitor photovoltaic panels, various sensors are needed
for measuring voltage, current, temperature, and irradiance.
o Voltage and Current Sensors: $40 - $80 per unit
o Voltage and Current Sensors (selected): $10 - $30 per unit
o Temperature Sensors: $60 per unit
o Temperature Sensor (selected): $10 per unit
o Irradiance Sensors: $700 - $1000 per unit
o SP Lite 2 Irradiance sensor (selected): $500
•Communication Modules: Modules for Wi-Fi, Bluetooth, or cellular
connectivity.
o Wi-Fi Module (ESP8266 selected): $5
o Cellular Module (SIM800L): $15
2. Software Development Costs
•App Development: The cost of developing a mobile application varies
based on complexity and platform (iOS, Android).
o Simple App (selected): $200 - $500
o Advanced App: $20,000 - $50,000
•Firmware Development: Custom rmware for the microcontroller/
SoC.
o Basic Firmware (selected): Free
o Advanced Firmware: $5000 - $10,000
3. Cloud Services and Data Management
•Cloud Hosting: Costs for hosting data and running cloud-based
analytics.
Fig. 6. Implemented housing design.
Fig. 7. Process calibration for the current sensor.
A. Ordo˜
nez et al.
Energy Reports 12 (2024) 2498–2510
2505

o Basic Hosting (AWS Free Tier): $50 - $100 per month
o Advanced Hosting: $100 - $700 per month
•Data Analytics Tools: Tools and services for data processing and
analysis.
o Basic Tools (selected): Free
o Advanced Tools (AWS IoT Analytics): $5 - $50 per month
•No cloud services (Local data management and storage – selected):
Free
4. Testing and Quality Assurance
•Hardware Testing: Cost for testing hardware components and the
integrated system.
o Basic Testing (selected): $1000
o Advanced Testing: $4000
•Software Testing: Ensuring the app and rmware function correctly.
o Basic Testing (selected): $200
o Advanced Testing: $5000
5. Labor Costs
•Development Team: Cost for a team of developers, engineers, and
designers.
o Small Team (2 students in Master thesis - selected): Free
o Large Team (10–15 members): $200,000 - $500,000 per year
Total Estimated Costs
•Proposed device: $1000 - $1200
•Advanced System: $275,000 - $630,000
Table 3 shows a comparative analysis of the proposed device with
existing technologies.
•Device A (Solmetric PV Analyzer): This device offers comprehensive
measurement capabilities including IV curve tracing, irradiance,
temperature, voltage, and current. It is highly accurate and suitable
for professional use with detailed data logging and advanced di-
agnostics. However, it is expensive and requires professional instal-
lation and regular calibration.
•Device B (Fluke IRR1-SOL Solar Irradiance Meter): This is a moder-
ately priced, handheld device primarily focused on measuring solar
irradiance and temperature. It is easy to use and portable, making it
suitable for quick assessments. However, it lacks data logging ca-
pabilities and real-time monitoring, and offers moderate accuracy.
•Device C (SMA Sunny SensorBox): This device provides high accu-
racy measurements of solar irradiance, temperature, and wind speed
with integrated logging and remote data access. It is designed for
permanent installation with weather-resistant features, making it
highly durable. It is also expensive and requires professional setup
and regular calibration.
•Proposed Device: This proposed solution aims to provide essential
monitoring capabilities such as solar irradiance, temperature, cur-
rent, and voltage at a signicantly lower cost. It is designed for easy,
DIY installation and offers basic data logging with potential
Fig. 8. (a) Electronic calibration process, and (b) Certied pyranometer (up) and SP Lite2 (down).
Table 3
Comparative cost analysis between 3 devices and our proposed device.
Parameter Device A: Solmetric PV Analyzer Device B: Fluke IRR1-SOL Solar
Irradiance Meter
Device C: SMA Sunny
SensorBox
Proposed Device
Measurement
Capabilities
IV curve tracing, irradiance,
temperature, voltage, and current
Solar irradiance, temperature, tilt
angle
Solar irradiance, temperature,
wind speed
Solar irradiance, temperature,
current, voltage
Cost High ($4000+) Moderate ($2000-$3000) High ($4000+) Low ($1000)
Ease of Installation Moderate (requires professional
installation)
Easy (handheld device) Moderate (requires setup and
conguration)
Easy (DIY setup possible)
Accuracy High Moderate to High High Moderate
Data Logging Yes (detailed logging with software
support)
No (manual recording) Yes (integrated logging and
remote access)
Yes (basic logging, potential for
integration with software)
Real-Time
Monitoring
Yes No Yes Yes
Durability High (robust, designed for eld use) Moderate High (weather-resistant) Moderate (depends on component
quality)
Calibration
Required
Yes (regular calibration needed) Minimal Yes Yes (regular calibration for accuracy)
Additional Features Advanced diagnostics, remote access Simple to use, portable Weather monitoring, remote
data access
Customizable based on components
A. Ordo˜
nez et al.
Energy Reports 12 (2024) 2498–2510
2506

integration with software for enhanced features. While it may not
match the high accuracy and durability of the more expensive de-
vices, it provides an affordable and customizable option for wide-
spread adoption.
The proposed device, leveraging inexpensive components, strikes a
balance between cost-effectiveness and essential functionality, making it
a viable option for broader deployment in cost-sensitive markets.
4. Results and discussion
Experimental measurements of the device’s performance were con-
ducted at two different geographic locations and for two different types
of photovoltaic panels in order to evaluate the device’s performance
under real operating conditions.
4.1. Experimental measurements at Bogot´
a, Colombia
In Universidad Jorge Tadeo Lozano (located at 4.068496,
−74.04125) at Bogot´
a on which date and hour, Colombia was conducted
the rst experimental measurements. Fig. 9 shows the photovoltaic
system installed on the roof of Engineering Researching Building of
Universidad Jorge Tadeo Lozano, in which the portable device was
connected.
Table 4shows the technical data of the PV panel Trina Solar, ref. TSM
250PA05.08.
Samples were taken at different times of the day to guarantee the
plotting of curves in different radiations. The goal is to reach currents in
the panel that are measurable by the sensor. Fig. 10shows an example of
the graphical interface of the application on the cell phone with the
results of Power-Voltage (PV) and Current-Voltage (I-V) curves.
The climate classication for Bogot´
a is Cfb (Temperate Oceanic),
meaning cold to mild winters and cool summers. Precipitation is well
distributed throughout the year. The low temperatures in the Colombian
capital are due to the passage of some "eastern waves" through central
Colombia. These climatic events inuence atmospheric conditions and
create a colder and rainier conditions in the city. This affects Bogot´
a’s
solar radiation, registering values of around 3.8 HSS-4 HSS. Tempera-
tures in Bogot´
a can average 13.5◦C, with warm hours during the dry
season that can exceed 20◦C. During this same period, due to increased
surface outgoing longwave radiation enhanced by the absence of clouds,
frosts can occur (temperatures below 0◦C), especially in rural areas.
Results in Fig. 10a show an efciency of 7.418 % for the conversion
of solar irradiance to electricity by the PV panel with 503 W/m
2
of solar
irradiance and 16.2 ºC of temperature. Thanks to the app development,
the device can save the data for any measurement for export the infor-
mation for further analysis. Fig. 10b shows the I-V curve obtained with
data exported from the monitoring app. In this case, solar irradiance was
about 250 W/m
2
, affecting directly the produced power: short circuit
current registered about 1.56 A and open circuit voltage 12.6 V.
Dispersed data are visible in Fig. 10b as a consequence of the slow
irradiance level.
4.2. Experimental measurements at Puerto Carre˜
no, Colombia
In Universidad del Rosario experimental photovoltaic system
(located at 6.83486◦N; −67.483784E) at Puerto Carre˜
no, (Colombia)
was conducted the second experimental measurements. Fig. 11 shows
the photovoltaic system installed on the oor of the experimental area of
Universidad del Rosario, to which the portable device was connected.
Table 5 shows the technical data of the photovoltaic panel Talesun,
ref. TP6F72M-405.
The climate of Puerto Carre˜
no is tropical, with a rainy season from
April to November and a dry season from December to March. From
June to August, the rains are very abundant, at a monsoonal level,
although they decrease in El Ni˜
no years.
Temperatures are high throughout the year, but especially from
January to April, before the rains, when they can reach 38/39◦C. From
June to August, during the rainy season, temperatures decrease a bit,
especially during the day, but humidity reaches its peak values for the
year.
The city is located in the far east of Colombia, on the border with
Venezuela, at 6 degrees north latitude, at the conuence of the Orinoco
and Meta rivers. Fig. 12a shows the I-V curve obtained with data
exported from the monitoring app for a solar irradiance of 639.2 W/m
2
.
In Fig. 12a, results indicate a short circuit current of about 3.16 A
with an open circuit voltage of about 35 V. PV panel degradation and the
high environmental temperature could be responsible for the value of
registered open circuit voltage. Fig. 12b shows the I-V curve obtained
with data exported from the monitoring app for a solar irradiance of
168.2 W/m
2
. As it shows, for small solar irradiance levels, the PV panel
isn’t able to perform its nominal power. Little short currents data are
plotted for temperature levels since 15 ◦C. Also, with this small irradi-
ance level, the current sensor presents problems in acquiring data at the
beginning of the experimental test. The short circuit current reported for
168.2 W/m
2
was 1.1 A while the open circuit voltage dropped to 32.5 V.
4.3. Future enhancements
The integration of IoT capabilities enables the device to establish
communication with other intelligent devices within a solar power
system, leading to an enhancement in overall system management and
efciency. To enhance predictive maintenance further, implementation
of machine learning algorithms could identify patterns that predict po-
tential failures before they occur. In depth analytics on panel perfor-
mance includes assessing degradation rates, environmental impacts and
efciency trends resulting in more profound insights for optimizing PV
systems. Allowing users to customize their monitoring dashboards will
increase accessibility while integrating sensors measuring atmospheric
pressure, air quality and UV index results is essential in understanding
Fig. 9. Photovoltaic system at Universidad Jorge Tadeo Lozano,
Bogot´
a, Colombia.
Table 4
Technical specications of the PV panel Trina Solar.
Parameter Description
P
max
250 W
V
mp
30.3 V
I
mp
8.27 A
V
oc
37.6 V
I
sc
8.85 A
Max. system voltage 600 V DC
Bypass diode 15 A
Series breaker 15 A
Fire resistance C class
A. Ordo˜
nez et al.
Energy Reports 12 (2024) 2498–2510
2507

how conditions affect PV panel effectiveness.
To maintain optimal efciency of photovoltaic panels with minimal
manual intervention, it is benecial to develop automatic cleaning
mechanisms that are triggered by the monitoring system when high
levels of dust or soiling are detected. The accuracy and reliability of solar
radiation measurements can be improved by utilizing more precise
irradiance sensors with lower error margins. Temperature readings
critical for assessing thermal impacts on PV efciency will become even
more accurate through employing temperature sensors with higher
resolution and faster response times that withstand harsh environmental
conditions ensuring long-term reliability while reducing maintenance
obligations. By implementing self-calibrating technologies in sensors,
measurement accuracy over extended periods will not require constant
recalibration. Real-time monitoring processing speed up due to edge
computing capabilities which reduce latency along development loca-
tions recognized resistant algorithms facilitate efcient storage trans-
mission useful for continuous performance evaluation at impressive
rates.
5. Conclusions
With the development of new microcontrollers and the advent of
new technologies enabling electronics miniaturization, a device was
created with a convenient, user-friendly design capable of immediately
characterizing photovoltaic panels. This device competes with identical
technologies available in the market that are more expensive and chal-
lenging for users to handle. The innovative factor of the device is linked
to the development of the mobile application, allowing data visualiza-
tion from any operating system.
One of the signicant benets of using a smartphone as a datalogger
and data visualizer is the ability to work ofine. This means that if you
are working in an area with no coverage, you can still characterize
photovoltaic panels and later synchronize the data with a server. This
feature enables work in remote areas.
The collected data, when synchronized with a server, becomes
accessible from anywhere and allows analysis using data analytics al-
gorithms and Machine Learning. This enables the creation of predictive
models and the detection of deciencies in the system by identifying
individual components such as end-of-life, poor positioning, etc.
The compact nature of the device allows for expansion in future
versions with new functionalities, such as characterization of electrical
production systems or improving signals with electronics focused on
signal processing to achieve certication of quality according to estab-
lished standards. Additionally, the device’s protection, regarding its
casing, can be certied to withstand IP55 standards (NEMA Enclosures,
2024) or higher.
Assessing the degradation and performance of PV modules accu-
rately is crucial, which is where our developed device comes in handy.
By offering real-time data along with diagnostic capabilities, this
monitor helps detect potential issues like cracks or short circuits that
affect efciency and longevity of photovoltaic systems under uctuating
environmental conditions. As a result, it leads to better maintenance
strategies while optimizing reliability & efcacy for solar installations
over time.
By utilizing low-cost components like microcontrollers and DSPs, our
development of a photovoltaic performance monitor has revealed
valuable economic benets. This approach signicantly increases the
cost-effectiveness and accessibility of PV systems by reducing the overall
expense required for monitoring. As a result, this technology can be
more widely deployed - particularly in markets where costs are critical-
which may then reduce solar power’s levelized cost (LCOE) as well as
Fig. 10. (a) Mobile app interface showing I-V results and (b) I-V curve obtained for 650 W/m
2
irradiance.
Fig. 11. Photovoltaic system at Universidad del Rosario, Puerto
Carre˜
no Colombia.
Table 5
Technical data of the photovoltaic panel Talesun.
Parameter Description
Pmax 405 W
Power tolerance +3 %
Voc 49.3 V
Isc 10.5 A
Vmax 1000VDC
Temp. 45◦C
Dimensions 2008*1002*35 mm
Application class Class A
Weight 22.5 kg
A. Ordo˜
nez et al.
Energy Reports 12 (2024) 2498–2510
2508

improve the viability of solar energy projects economically. What’s
more, these inexpensive monitoring solutions could help optimize
module efciency while extending their life cycle; thus decreasing
maintenance expenses subsequently resulting in better return on in-
vestment for organizations that leverage them within their installations
powered by Solar Energy Technology.
CRediT authorship contribution statement
O. Garcia: Supervision, Resources, Methodology, Investigation,
Formal analysis. B Quesada: Writing – original draft, Validation, Su-
pervision, Project administration, Funding acquisition. A. Aristiz´
abal:
Writing – review & editing, Writing – original draft, Visualization,
Validation, Supervision, Methodology, Investigation. Fredy Mesa:
Writing – review & editing, Writing – original draft, Validation, Super-
vision, Project administration, Investigation. J. Urbano: Software,
Methodology, Investigation, Formal analysis, Data curation, Conceptu-
alization. S. Zapata: Validation, Methodology, Investigation, Funding
acquisition, Formal analysis, Data curation. M. Casta˜
neda: Validation,
Supervision, Software, Resources, Investigation. A. Ordo˜
nez: Software,
Methodology, Formal analysis, Data curation, Conceptualization.
Declaration of Competing Interest
The authors declare that they have no known competing nancial
interests or personal relationships that could have appeared to inuence
the work reported in this paper.
Data Availability
No data was used for the research described in the article.
Acknowledgments
The authors thank the project “Design and Implementation of a
Portable Device for Evaluating the Performance of Photovoltaic Panels”
between the University of Rosario, the Jorge Tadeo Lozano University,
and Fundaci´
on Universitaria Los Libertadores, funded by the Colom-
bian-French Researchers Association (COLIFRI) in the framework of the
FSPI project “Ecosystem of Renewable Energies in Puerto Carre˜
no”.
References
´
Angel-Antonio Bayod-Rújula, 2014. Jos´
e-Antonio Cebollero-Abi´
an, A novel MPPT
method for PV systems with irradiance measurement. ISSN 0038-092X Sol. Energy
Volume 109, 95–104. https://doi.org/10.1016/j.solener.2014.08.017.
Assoa, Y.B., Valencia-Caballero, D., Rico, E., Del Ca ˜
no, T., Furtado, J.V., 2023.
Performance of a large size photovoltaic module for façade integration. Renew.
Energy vol. 211, 903–917. https://doi.org/10.1016/j.renene.2023.04.087.
Baghel, N., Manjunath, K., Kumar, A., 2023. Performance evaluation and optimization of
albedo and tilt angle for solar photovoltaic system. Comput. Electr. Eng. vol. 110,
108849 https://doi.org/10.1016/j.compeleceng.2023.108849.
Belhaouas, N., Hafdaoui, H., Hadjrioua, F., Assem, H., Madjoudj, N., Chahtou, A.,
Mehareb, F., 2024. Failures and performance of different aged PV modules operated
under northern Algerian climate conditions: Analysis, assessment, and recommended
solutions. ISSN 1350-6307 Eng. Fail. Anal. Volume 163 (Part A). https://doi.org/
10.1016/j.engfailanal.2024.108504.
Blakesley, J.C., Castro, F.A., Koutsourakis, G., Laudani, A., Lozito, G.M., Riganti
Fulginei, F., 2020. Towards non-destructive individual cell I-V characteristic curve
extraction from photovoltaic module measurements. Sol. Energy vol. 202, 342–357.
https://doi.org/10.1016/j.solener.2020.03.082.
Casado, P., Blanes, J.M., Torres, C., Orts, C., Marroquí, D., Garrig´
os, A., 2022. Raspberry
Pi based photovoltaic I-V curve tracer. HardwareX vol. 11, e00262. https://doi.org/
10.1016/j.ohx.2022.e00262.
Chen, Z., Lin, Y., Wu, L., Cheng, S., Lin, P., 2020. Development of a capacitor charging
based quick I-V curve tracer with automatic parameter extraction for photovoltaic
arrays. Energy Convers. Manag. vol. 226, 113521 https://doi.org/10.1016/j.
enconman.2020.113521.
Duru, H.T., 2006. A maximum power tracking algorithm based on Impp=f(Pmax)
function for matching passive and active loads to a photovoltaic generator. Sol.
Energy vol. 80 (7), 812–822. https://doi.org/10.1016/j.solener.2005.05.016.
Gow, J.A., Manning, C.D., 1999. Development of a photovoltaic array model for use in
power-electronics simulation studies. IEE Proc. - Electr. Power Appl. vol. 146 (2),
193–200. https://doi.org/10.1049/ip-epa:19990116.
M.A. Green, “Solar cells: operating principles, technology, and system applications,” Jan.
1982, Accessed: Sep. 02, 2023. [Online]. Available: 〈https://www.osti.gov/biblio/
6051511〉.
Hosseini, A., Mirhosseini, M., Dashti, R., 2023. Analytical study of the effects of dust on
photovoltaic module performance in Tehran, capital of Iran. J. Taiwan Inst. Chem.
Eng. vol. 148, 104752 https://doi.org/10.1016/j.jtice.2023.104752.
Jathar, L.D., et al., 2023. Comprehensive review of environmental factors inuencing the
performance of photovoltaic panels: Concern over emissions at various phases
throughout the lifecycle. Environ. Pollut. vol. 326, 121474 https://doi.org/10.1016/
j.envpol.2023.121474.
Jos´
e Mu˜
noz-Rodríguez, F., Snytko, A., De La Casa Hern´
andez, J., Rus-Casas, C., Jim´
enez-
Castillo, G., 2023. Rooftop photovoltaic systems. New parameters for the
performance analysis from monitored data based on IEC 61724. Energy Build. vol.
295, 113280 https://doi.org/10.1016/j.enbuild.2023.113280.
Kumar, Shubham, Nayak, Paresh Kumar, 2024. An effective method for detection and
location estimation of faults in large-scale solar PV arrays. ISSN 0038-092X Sol.
Energy Volume 277. https://doi.org/10.1016/j.solener.2024.112727.
Li, B., Delpha, C., Migan-Dubois, A., Diallo, D., 2021. Fault diagnosis of photovoltaic
panels using full I–V characteristics and machine learning techniques. Energy
Convers. Manag. vol. 248, 114785 https://doi.org/10.1016/j.
enconman.2021.114785.
Li, B., Migan-Dubois, A., Delpha, C., Diallo, D., 2021. Evaluation and improvement of IEC
60891 correction methods for I-V curves of defective photovoltaic panels. Sol.
Energy vol. 216, 225–237. https://doi.org/10.1016/j.solener.2021.01.010.
Liu, Y., et al., 2021. Fault diagnosis approach for photovoltaic array based on the stacked
auto-encoder and clustering with I-V curves. Energy Convers. Manag. vol. 245,
114603 https://doi.org/10.1016/j.enconman.2021.114603.
Liu, Y., et al., 2022. Intelligent fault diagnosis of photovoltaic array based on variable
predictive models and I–V curves. Sol. Energy vol. 237, 340–351. https://doi.org/
10.1016/j.solener.2022.03.062.
Ma, M., Wang, H., Xiang, N., Yun, P., Wang, H., 2021. Fault diagnosis of PID in
crystalline silicon photovoltaic modules through I-V curve. Microelectron. Reliab.
vol. 126, 114236 https://doi.org/10.1016/j.microrel.2021.114236.
Fig. 12. I-V curve for Talesun PV panel at Puerto Carre˜
no, Colombia: (a) 639.2 W/m
2
and (b) 168.2 W/m
2
.
A. Ordo˜
nez et al.
Energy Reports 12 (2024) 2498–2510
2509

Ma, M., Zhang, Z., Xie, Z., Yun, P., Zhang, X., Li, F., 2020. Fault diagnosis of cracks in
crystalline silicon photovoltaic modules through I-V curve. Microelectron. Reliab.
vol. 114, 113848 https://doi.org/10.1016/j.microrel.2020.113848.
NEMA Enclosures, “IP55 Enclosures,” Accessed: Jul. 13, 2024. [Online]. Available:
〈https://www.nemaenclosures.com/enclosure-ratings/ip-enclosures/ip55-enclosur
es.html〉.
Nieto-Morone, M.B., Rosillo, F.G., Mu˜
noz-García, M.A., Alonso-García, M.C., 2024.
Enhancing photovoltaic module sustainability: Defect analysis on partially repaired
modules from Spanish PV plants. ISSN 0959-6526 J. Clean. Prod. Volume 461.
https://doi.org/10.1016/j.jclepro.2024.142575.
Olayiwola, T.N., Choi, S.-J., 2023. Superellipse model: An accurate and easy-to-t
empirical model for photovoltaic panels. Sol. Energy vol. 262, 111749. https://doi.
org/10.1016/j.solener.2023.05.026.
¨
Ozkalay, Ebrar, Virtuani, Alessandro, Eder, Gabriele, Voronko, Yuliya,
Bonomo, Pierluigi, Caccivio, Mauro, Ballif, Christophe, Friesen, Gabi, 2024.
Correlating long-term performance and aging behaviour of building integrated PV
modules. ISSN 0378-7788 Energy Build. Volume 316. https://doi.org/10.1016/j.
enbuild.2024.114252.
Padilla, A., Londo˜
no, C., Jaramillo, F., Tovar, I., Cano, J.B., Velilla, E., 2022.
Photovoltaic performance assess by correcting the I-V curves in outdoor tests. Sol.
Energy vol. 237, 11–18. https://doi.org/10.1016/j.solener.2022.03.064.
A.J.Q. Perez, “Prototipo fotovoltaico con seguimiento del Sol para procesos
electroquímicos”, Accessed: Sep. 02, 2023. [Online]. Available: 〈https://www.acade
mia.edu/29415500/Prototipo_fotovoltaico_con_seguimiento_del_Sol_para_procesos
_electroqu%C3%ADmicos〉.
Piccoli Junior, L.A., De Oliveira, F.S., Gasparin, F.P., Krenzinger, A., 2023. Design and
characterization of a continuous solar simulator for photovoltaic modules with
automatic I-V curve acquisition system. Sol. Energy vol. 256, 55–66. https://doi.org/
10.1016/j.solener.2023.03.057.
Practical Handbook of Photovoltaics - 1st Edition. 〈https://shop.elsevier.com/books/pra
ctical-handbook-of-photovoltaics/mcevoy/978-1-85617-390-2〉(Accessed Sep. 02,
2023).
Seguel., J.I.L., Aug. 2009. “Projeto de um sistema fotovoltaico autˆ
onomo de suprimento
de energia usando t´
ecnica MPPT e controle digital”.
Soler-Castillo, Y., Sahni, M., Leon-Castro, E., 2023. Performance predictability of
photovoltaic systems: An approach to simulate the I–V curve dynamics. Energy Rep.
vol. 9, 234–269. https://doi.org/10.1016/j.egyr.2023.03.075.
Toledo, F.J., Galiano, V., Blanes, J.M., Herranz, V., Batzelis, E., 2023. Photovoltaic
single-diode model parametrization. An application to the calculus of the Euclidean
distance to an I – V curve,”. Math. Comput. Simul., S0378475423000058 https://
doi.org/10.1016/j.matcom.2023.01.005.
Yang, C., Su, C., Hu, H., Habibi, M., Safarpour, H., Amine Khadimallah, M., 2023.
Performance optimization of photovoltaic and solar cells via a hybrid and efcient
chimp algorithm. Sol. Energy vol. 253, 343–359. https://doi.org/10.1016/j.
solener.2023.02.036.
Zhang, Z., Ma, M., Ma, W., Zhang, R., Wang, J., 2022. A data-driven photovoltaic string
current mismatch fault diagnosis method based on I-V curve. Microelectron. Reliab.
vol. 138, 114705 https://doi.org/10.1016/j.microrel.2022.114705.
Zhu, Y., Xiao, W., 2020. A comprehensive review of topologies for photovoltaic I–V curve
tracer. Sol. Energy vol. 196, 346–357. https://doi.org/10.1016/j.
solener.2019.12.020.
A. Ordo˜
nez et al.
Energy Reports 12 (2024) 2498–2510
2510