Volume 6 Issue 2 December 2020 www.irjes.psyec.edu.

in

IOT BASED SOLAR PANEL FAULT MONITORING AND CONTROL

J. Subhashini
PG Scholar, Department of Electrical and Electronics Engineering, Pandian Saraswathi Yadav Engineering
College, Sivagangai,India.

M. Naveena
AssistantProfessor, Department of Electrical and Electronics Engineering, Pandian Saraswathi Yadav
Engineering College, Sivagangai,India.

Abstract
A novel datalogger based on free software and hardware has been designed, built, programmed and installed
as an experimental prototype in multiple sites. Remote monitoring extends the effectiveness of the datalogger
in areas deprived of electrical grid and traditional wired telecommunication networks. The integration of
Internet of Things (IoT) in solar measurement systems allows the remote monitoring of small stand-alone
photovoltaic (PV) systems, enhancing the performance and the maintenance of the system. The datalogger
measured electric and climatic parameters (up to 14 parameters, expandable) with the required accuracy
established by the IEC61724 standard; to include 3G technology in it allowed the stand-alone PV systems
system monitoring remotely via web or via mobile application, all at low cost. An outdoor campaign of over
12 months under the harsh environmental conditions at multiple locations was performed to test the new
datalogger under real and different conditions demonstrating the robustness and the reliability of the system.
Index Terms—3G mobile communication, Internet of Things, photovoltaic systems, remote monitoring

Introduction
An estimated 1.1 billion people currently lack electricity access [1], mainly residing in rural
areas of sub-Saharan Africa and Southeast Asia, and, to a lesser extend in the Middle East, Central
Asia and Latin America [2]. Due to the emergence of rural electrification programmes, thousands of
Solar Home Systems (SHSs) [3] are installed in remote regions, commonly in locations with no
electrical grid, no traditional wired telecommunication networks and often with difficult
accessibility by common transport. SHSs are stand- alone PV (SAPV) systems. Typically, these
installations use crystalline-silicon PV modules. Regarding the battery backup unit, the most
commonly battery type installed is lead-acid and most small SHS employs charge controllers using
PWM to regulate the charge current to the battery [4].The absence of adequate monitoring of SHS
and, therefore, the impossibility of detecting operation and maintenance problems, can lead to a
dramatic shortening of the useful life of the PV systems or even to the withdrawal of their use. [5].In
the case of grid-connected PV systems generally have high budgets, and the associated data
acquisition systems allow to monitor the main parameters to conduct the necessary maintenance
actions without impacting significantly in the total installation cost. However, it is very difficult to
supervise the operation of SHSs due mainly to the fact that the required commercial data loggers
available in the market are very expensive compared to the total system cost, they require special
software and they also generally demand an external power supply or a PC to be connected all of the
time .As a result, it is necessary to further develop data acquisition systems very accurate and

International Research Journal of Engineering Sciences Page 88

 Volume 6 Issue 2 December 2020 www.irjes.psyec.edu.in

independent of external sources, all at low cost. In recent years, analytical monitoring has been
progressively applied to small PV systems.

Related Works on Dataloggers for PV Monitoring

The literature includes numerous reports of PV monitoring systems during the last two decades.
Mukaro et al. [6] in 1998 developed one of the first low-cost systems designed for solar radiation
monitoring. With 4 analogue inputs, this monitoring system was based on an 8-bit microcontroller.
This datalogger was suitable for monitoring of meteorological parameters at remote stations,
especially in developing countries. The main disadvantage of this work was the connection of the
datalogger to PC. Koutroulis and Kalaitzakis [7] in 2003 reported another monitoring system for
renewable energy sources. With 16 single ended (eight differential analog) input channels, the
datalogger used a DAQ card installed in a PC and a 12-bit accuracy ADC converter. As a
disadvantage, the dependence on a PC and the use of commercial software increased the final cost of
the system. Forero et al. [8] in 2006 proposed a system for monitoring a stand-alone PV plant. The
system was designed for measuring environmental variables, as well as for monitoring the
performance of a stand-alone PV solar plant. The datalogger was based on precision electronic
modular field point I/O devices and a high speed data acquisition card. The main inconveniences
were the use of a graphical environment based on the Lab view program and the requirement of a
PC connection. Tina and Grasso [9] in 2014 proposed a dedicated data acquisition system for remote
monitoring the operation of a stand-alone PV appliance. With20 analog inputs, the monitoring
system measured electrical and climatic parameters using low-cost commercial sensors. However,
measurements were not compared with another calibrated datalogger and the errors were not
weighted. After reviewing the published literature, it was identified that it is difficult to track the
functioning of these SAPV systems with conventional data loggers mainly due to various
limitations:: external source dependency [10-13], requirement of commercial software [14, 15] and
in numerous designs the measurement of errors were not bounded following international standards
of accuracy [16, 17].
To solve the former technical and economic issues, a low- cost prototype based on open source
and free hardware technologies (ArduinoTM) for monitoring PV systems was proposed by M.
Fuentes et al. [18] in 2014. With 8 analogue inputs for measuring up to 8 electrical/meteorological
parameters and 3 inputs for low-cost analogue temperature sensors, the resolution of the datalogger
was 18-bits. Virtually unlimited inputs for digital temperature sensors were available. The
datalogger was especially designed for installing in remote areas or regions in developing countries.
The results indicated that the datalogger provided high accuracy, and it was autonomous, low-cost
and robust in harsh environments. Data was stored in SD card, allowing the installation of the
system in isolated areas, requiring minimal maintenance. But this minimal maintenance of the novel
datalogger was based on a manual procedure requiring human operator intervention for collecting
data, which can be an inconvenient at locations that are difficult to access as well as it increases the
maintenance costs and operation. The objective of this new work has been to develop a novel and

International Research Journal of Engineering Sciences Page 89

 Volume 6 Issue 2 December 2020 www.irjes.psyec.edu.in

improved version of this solar ArduinoTM low-cost datalogger based on the application of IoT
(Internet of the Things), adding internet connectivity and allowing the SAPV system to be
monitored remotely via web or via mobile application, which means that not only data of the long-
term system performance is retrieved but also that instantaneous information is received and so the
problems related to the operation and maintenance of the stand-alone PV system can be rapidly
detected and solved. Additionally, another novelty of this work lies in designing, manufacturing and
testing the prototype under real (electrical and environmental) working conditions.

Requirements for the SAPV Electrical Performance Monitoring System

The IEC61724 standard [20] entitled “Photovoltaic system performance” describes the general
guidelines for the monitoring and analysis of the electrical performance of photovoltaic systems
(stand-alone and grid connected). In Table I, a summary of parameters to be monitored in real time
PV systems is presented.

Table 1 Parameters to Be Monitored In Real Time PV Systems

A single current or power sensor can be used for the measurement of current or power for
directions or both input and output.
The standard defines three classes of monitoring systems, corresponding to different levels of
accuracy and different intended applications: Class A (high accuracy), Class B (medium accuracy)

International Research Journal of Engineering Sciences Page 90

 Volume 6 Issue 2 December 2020 www.irjes.psyec.edu.in

and Class C (basic accuracy). The accuracy classification and the suggested applications are
presented in TableII.

Table 2 Monitoring System Classification and Suggested Applications

Figure 1 Example of Application of the IEC61724 Standard: SAPV System

According to IEC61724 standard, Class B or Class C would be most appropriate for small
systems, such as smaller commercial and residential installations. An example of the IEC61724
standard appliance for SAPV systems is shown in Figure1.

Electrical and Climatic Parameters

The measurements of electrical parameters provide information on the consumption and the
efficiency of SAPV systems. According to IEC61724 standard, the accuracy (the most restrictive
International Research Journal of Engineering Sciences Page 91
 Volume 6 Issue 2 December 2020 www.irjes.psyec.edu.in

case, high accuracy class) of the voltage, current and power measurements, including signal
conditioning, must be better than 2% of the reading. IEC61724 standard indicates that climatic
parameters must be measured at a location that is representative of the SAPV system conditions. The
accuracy of temperature measures must be better than ±1 ºC and the accuracy of irradiance measures
must be better than 8% of the reading from 100 W∙m-2 to 1,500 W∙m-2 (including signal
conditioning). The measurement of optional parameters such as wind speed, also must comply the
standard requirements of accuracy: ≤ 0. 5 m⋅s-1 for wind speeds ≤ 5 m⋅s-1, and ≤ 10 % of the reading
for wind speeds greater than 5 m⋅s-1. According to the results obtained by López-Vargas et al. in
2018 [19], the previous version of the datalogger [19] met all the precision requirements established
by the IEC61724 standard mentioned above.

Monitoring Requirements
According to IEC61724 standard, for medium and basic accuracy systems (Class B and Class C,
as indicated earlier), the maximum sampling interval should be 1 min and maximum recording
interval should be 15 min (medium accuracy) and 60 min (basic accuracy). The recording interval
should be an integer multiple of the sampling interval, and an integer number of recording intervals
should fit within 1 h. Each record shall include a timestamp, and the time should refer either to local
time (not daylight saving time) or universal time, to avoid winter/summer time changes. When
multiple data acquisition units are involved that each independently apply time stamps, the clocks of
the units must be synchronized, preferably by an automated mechanism such as global positioning
system (GPS) or network time protocol (NTP).
To be able to apply the automated synchronization mechanisms advised by the IEC61724
standard, the connection to the Internet is required. To access the Internet, the standard does not set
any requirements concerning the type of communication method. However, in the case of
decentralized systems, wireless technologies are most appropriate: commonly SAPV systems are
located in remote and inaccessible regions, in locations without access to traditional wired
telecommunication networks.
The datalogger should be able to transmit the measured data anywhere; the location of the
SAPV system should not act to the detriment of the correct functioning of the monitoring system.
The communication technology should be capable of operating autonomously, without requiring
additional wired communications networks. To obtain real-time monitoring, the delivery speed of
the data transmission system should allow compliance with the sampling intervals established by the
IEC61724standard.

SAPV Monitoring System Connectivity

Among all the different standards deployed in the market [21], usually distributed in different
frequency bands and using multiple communication protocols, selecting the most suitable
connectivity technology for an IoT application can be a challenge. Wireless technologies have been
widely used in the case of decentralized systems, although Ethernet cable is also an alternative for
centralized systems [22].
International Research Journal of Engineering Sciences Page 92
 Volume 6 Issue 2 December 2020 www.irjes.psyec.edu.in

As there are different options to add wireless connectivity to the ArduinoTM solar data logger,
previous PV monitoring wireless systems according to their data transmission techniques have been
reviewed. The connectivity requirements of different types of IoT networks vary widely, depending
on their purpose and resource constraints. Relevant wireless communications techniques have been
studied: Table III [23] shows a wireless connectivity technologies (non-exhaustive) comparison
focused on range, bandwidth and cost.
A remote PV-system monitored via satellite was developed by Krauter and Depping [24] in
2004, where irradiance and electrical parameters were collected and sent to the ARGOS- SCD
satellite. The satellite data was transferred from the receiving ground station to the Internet,
providing worldwide access via WWW. Data transmission via satellite is very useful in regions
without access to telephone communications. But this data transmission was reported to be slow,
taking around 8 to 12 min, and the initial cost for the transmitter was very high.
Papageorgas et al. [25] in 2013 reported a solar panel monitoring system based on Zigbee. The
system measured data for each solar panel of a solar park installation and allowed the optimization
of electric power production through a web-based application. The client-server-web-publishing
software was based on an open-source platform. The integrated Zigbee communication technology
was based on a Zigbee RF modem available from ATMEL.

However, the system required a wired network due to this technology only covers short
distances. The same applies to the Bluetooth technologies that support simple wireless networking
but only cover short distances (100 m is the maximum range without repeaters [26]) as reported by
Hua et al. [27] in 2009: typically up to 10 m for Bluetooth class 2 and 100 m for Bluetooth class

International Research Journal of Engineering Sciences Page 93

 Volume 6 Issue 2 December 2020 www.irjes.psyec.edu.in

A series of novel wireless portable systems including calibration, monitoring and ISP systems
were implemented for fuel cell city buses, using Bluetooth for transmitting data.
Rosiek and Batlles [28] in 2008 developed a system for the data acquisition from remote
meteorological stations located in the north side of the Natural Park of Sierra Nevada (Granada,
Spain). The system was based on an ATmega16 microcontroller and they compared this system
against two other commercial acquisition systems obtaining an error in the order of 1 %. They used
mobile communications due to the considerable advantages in transmitting the information at big
distances in isolated areas. The experimental data was received from the remote stations network,
located about 100 km of distance.
Due to the location of PV systems in isolated regions, the study of communication techniques
focused on technologies that provide greater ranges (WWAN). The use of mobile communications
(2G, 3G) ensures an affordable and stable long-distance data transfer, but the effectiveness of this
technique is subject to the signal strength in the location of the stand-alone PV system. The
International Telecommunication Union (ITU) [29] in 2016 estimated that approximately seven
billion people, the 95 % of the global population, live in an area that is covered by a mobile network.
Regarding the mobile-broadband networks (3G or above) they reach 84 % of the global population,
67 % of the rural population [29].
On the other hand, a clear signal of the global mobile penetration rate is that in developing
countries, there are more households that own a mobile phone that have access to electricity or
drinking water: roughly 70 % of the bottom fifth of the population in developing countries own a
cell phone, as reported by the World Bank's World Development Report for 2016 [30]. According to
a Groupe Speciale Mobile Association (GSMA) report [31], in 2014, more than a half of all global
mobile money deployments were carried out in sub- Saharan Africa. The 63 % of Africans had
access to improved water supply and 32 % to electricity, compared to 82 % who had access to GSM
coverage. A United Nations research in India [32], in 2008, showed nearly 366 million people (31 %
of the population) had access to improved sanitation, meanwhile, 545 million mobile phones were
connected to service.
After studying the available wireless techniques, the use of mobile communications was selected
for adding wireless connectivity to the ArduinoTM datalogger, since this technology presents more
advantages than the other communication techniques studied and both, industrialized countries and
developing countries, are covered by mobile communications networks, even reaching areas where
the electric grid and the water supply network do not. In addition, other studies endorse the use of
mobile communications for this specific application. In 2014, R. Tejwani et al. [33] presented the
advantages and disadvantages of several monitoring systems for rural application based on different
techniques of communication. After discussing, they proposed remote monitoring using mobile
communications. In 2017, V. Villagrán et al. [34] presented the design and implementation of
monitoring system to study down-slope winds in the Laja River Valley, southern-central Chile. They
used mobile communications because of the continued growth of the mobile telecommunication

International Research Journal of Engineering Sciences Page 94

 Volume 6 Issue 2 December 2020 www.irjes.psyec.edu.in

industry; its coverage has also expanded in Chile, even in remote areas inhabited by small
communities.

Design of the 3G Datalogger for Solar Photovoltaic Monitoring

General Description
Figure 2 shows the diagram of the new wireless monitoring system as well as the basic
distribution of environmental and electrical sensors for monitoring the SAPV system. The new
datalogger measures meteorological and electrical parameters; data is sent via 3G and information is
stored in two different servers: a dedicated server (located in the University of Jaén) and a cloud
server (free storage platform). The connection to the internet allows to monitor the SAPV system
from any device or computer.

Hardware
The previous datalogger was designed around the ArduinoTM UNO board as it stands out in
comparison with the other open-source platforms due to its robustness, cost and developer
community [18]. However, it did not cover all the functionalities for PV monitoring itself, so
hardware enhancement was integrated by López-Vargas et al. [19] integrated in an ad-hoc PCB
(Printed Circuit Board):

(a) a bi- directional I2C™ bus, (b) sensors signal conditioning including the integration of
electronic elements for filter in gun desirable signals, (c) a RTC for tracking time and ensuring the
precise synchronization of the measurements, (d) two external ADCs of 18-bits resolution for

International Research Journal of Engineering Sciences Page 95

 Volume 6 Issue 2 December 2020 www.irjes.psyec.edu.in

accomplishing the IEC61724accuracyrequirementsand (e) a specific module to reduce the energy

consumption.
When designing the new datalogger with connectivity features, it was found that ArduinoTM
UNO would not allow to connect directly the board to the internet (it would require an extra shield
board). The 3G shields compatible with ArduinoTM found in the market that met the requirements of
transmission were extremely expensive so the use of these shields was discarded. As a low-cost
alternative, an Ethernet board with a nano router could perform the same connectivity functions

Therefore, ArduinoTM Ethernet (Rev 3) was selected as it has connectivity functions and similar
characteristics integrated all in a single board. Table IV shows the comparison between ArduinoTM
Ethernet and ArduinoTM UNO boards. The use of an ArduinoTM UNO with an Ethernet shield
stacked on top could be also considered as an alternative option as the price is comparable to that of
the ArduinoTM Ethernet board.
Regarding connectivity itself, studies based on PV system monitoring were reviewed (see
Section IV) to select a suitable data transmission technique. The use of mobile communications,
concretely 3G, offers the most significant advantages. Main advantages of the 3G prototype are
flexibility and unlimited reach, although the latter is restricted by the 3G mobile penetration rate.
The system based on 3G does not require the installation of additional networks and this
flexibility makes the 3G system more suitable for the monitoring of SAPV systems located in remote
areas, especially in developing countries. Among the numerous options to connect ArduinoTM to the
internet using 3G, the TL- MR3020 router from TP-LINK (Figure3a) stands out over the
restbecauseitoffersthepotentialtoconfigureWi-Fiand3G

International Research Journal of Engineering Sciences Page 96

 Volume 6 Issue 2 December 2020 www.irjes.psyec.edu.in

Figure 4 Development of PV Output Generation, Battery and Load Current and Voltage
Sensors

Transmission systems at a low price. This device requires an external 5VDC/1A power supply
and it operates over a temperature range from 10 ºC to 60 ºC and relative humidity conditions in the
humidity range of 10 % RH to 90 % RH (non-condensing).
As the router has an on-board interface USB 2.0 port, the 3G operating mode can be configured
providing internet connection through a 3G USB stick modem (Figure 3b). For this purpose, MA260
modem from TP-LINK was selected. This modem allows to acquire 3G mobile broadband access by
inserting a SIM card into the adapter. It supports 3G HSPA+ technology, which provides improved
data transfer rates. The USB modem operating temperature range is 0 °C to 40 °C. Thus, the
operating temperature of the entire system (router and modem) is limited by the most restricted

International Research Journal of Engineering Sciences Page 97

 Volume 6 Issue 2 December 2020 www.irjes.psyec.edu.in

values: 10 ºC - 40 ºC. This is particularly important since, depending on the location of the system,
temperature evacuation systems (fans, heat sinks…) may be required.
The novel datalogger was developed ad-hoc for the monitoring of SAPV systems that in most
cases are isolated systems, so an individual data transmission system was the most appropriate
option. However, in the case of finding several SAPV systems installed close enough, the
monitoring system was designed to provide scalability. The novel monitoring system stands out due
to its stackable hardware design. By adding boards with the hardware improvements, the number of
analog or digital inputs can be expanded, and various SAPV systems can be monitoring using a
single datalogger. To do this, only it would be necessary to extend the number of parameters
transmitted in each cycle by SW, achieving scalability at a reduced cost.

Meteorological and SAPV Sensors

Sensor characteristics for measuring both electrical and meteorological system parameters were
previously detailed by López-Vargas et al. [19] in their work on the ArduinoTM basic datalogger.
Employing exclusively low-cost sensors, the prototype for measuring SAPV systems includes a
wide range of climatic measures: irradiance (Gi), ambient temperature (Tamb), humidity (H), wind
velocity (W) and rainfall (R). DS18B20 temperature sensors from Maxim IntegratedTM were used to
acquire PV module temperature (Tmod) and PV battery temperature (Tbat). The electrical parameters
measured by the low-cost datalogger include the PV generator output voltage (VA), battery voltage
(VS), load voltage (VL), PV generator output current (IA), battery current (IS) and load current (IL).
Based on voltage divider circuits, low-cost voltage sensors that increase the ADC input range were
developed and three common-mode voltage amplifiers were also used to allow measuring
differential signals in presence of high common- mode voltage range accurately. Hall Effect sensors,
LTS 15- NP from LEMTM, were used for current measurements. These sensors presented a level of
accuracy from ± 0.2 % to ± 0.7% (operating with internal measuring resistance at 25 ºC). Figure 4
shows both voltage and current sensors developed.

Software
Figure 5 shows the flow diagram of the process followed by the microprocessor. At start, the
internet connection is established and the RTC is synchronized: these processes run every time the
routine is initialized or the reset button is pressed. In addition, once per day, at 00 AM, the RTC is
synchronized to avoid drifts. Every minute, in second 0 and second 30, parameters are measured;
data transmission and data storage processes are carried out. Software improvements have been
developed and new processes have been included:

International Research Journal of Engineering Sciences Page 98

 Volume 6 Issue 2 December 2020 www.irjes.psyec.edu.in

Figure 5 Diagram for the Routine Executed by Each Iteration: Sequence of Processes and
Communications

 The internet access provides the possibility of synchronizing the RTC using a NTP server. The
clock is synchronized at the start of the program or in case of reset. As the drift of the RTC is a
few seconds per week, to synchronize once a day (at 00 AM) is enough.
 The connection between the microcontroller and the server is established by a Hypertext
Transferr Protocol (HTTP) petition. After establishing the connection between datalogger and
server, the process of data transferring starts.
 The information is stored in a MySQLTM database hosted in a server, located in the University
of Jaén, with a specific order and format.
 Additionally, data is sent to Thing Speak™, an open source Internet of Thing analytics platform
service that allows inserting, visualizing and analyzing live data streams in the cloud.
Two types of data storage were implemented: a dedicated server (conventional method) and a
cloud service platform (novel application of IoT). Both alternatives of storage allow to monitor the
International Research Journal of Engineering Sciences Page 99
 Volume 6 Issue 2 December 2020 www.irjes.psyec.edu.in

system in a remote way. The use of free cloud service platforms allows to obtain a highly
accessibility and cost-effective monitoring system based entirely on free SW and HW.
In the first option (conventional), the measured data is stored in the main base PC (located in the
University of Jaén) which serves as the server for the website, such that users can retrieve data from
any previous date within the monitoring period. The web-based monitoring feature was included for
adding value to the global monitoring system. The website allows users to observe the monitored
parameters and statistical data, as well as other information: daily production, charge of battery or
consumptions. The website was developed in HTML language while the table and graph were
created by integrating PHP code and library from High charts. The parameters monitored are
processed and analyzed to produce information that can be understood easily by users; data is
protected by password and the information is represented in graphs. The website is accessible using
the following link: http://energiaagua.ujaen.es/FVMonitor/.
On the other hand, a second option was also developed, as the advantages of open cloud systems
are really interesting for this type of applications: they provide large data storage, high reliability
and low-cost. Cloud systems reduce the cost of data storage facilities and extra maintenance to keep
or maintain the data. Open cloud-based IoT platforms and services were reviewed [35]. Thing
SpeakTM, an open IoT data platform based on public cloud technology was selected mainly due to its
variety of third party platforms that include Arduino and Twitter amongst others. Thing Speak™
platform stores and retrieves data from things using HTTP over internet. Thing Speak™ allows to
analyze data using MATLAB without requiring the purchase of a license from Math works™. Each
monitoring system sends data to an individual channel. These channels can be configured as a public
or a private channel (protected by password). In addition, mobile applications allow monitoring
Thing Speak™ channels using a smart phone. Free users are limited to 10 million messages (3+
years at minimum update interval, 15s) and 8 fields per channel.
Irradiance (Gi), ambient temperature (Tamb), PV module temperature (Tmod), PV generator output
voltage (VA), battery voltage (VS), PV generator output current (IA), battery current (IS) and load
current (IL) parameters were selected for monitoring.

International Research Journal of Engineering Sciences Page 100

 Volume 6 Issue 2 December 2020 www.irjes.psyec.edu.in

Figure 6 SAPV System Installed at IMDEA Water Institute Facilities: PV Mono- Crystalline
Module (a), Charge Controller (b) and Lead-Acid Battery (c).

Experimental Set-Up
Studies on rural electrification in developing countries and renewable-based
renewable initiatives energy
programs were previously reviewed by López-VargasVargas et al. [19]. Two real SAPV systems, similar to
those typically employed in developing countries, were installed in different locations to verify the
t
proper functioning of the monitoring system under real conditions: countries, were installed in
different locations to verify the proper functioning of the monitoring system under real conditions:

Figure 7 SAPV System Installed at University of Linares Facilities: PV Mono- Crystalline

Module (a), Charge Controller (b) and Lead-Acid Battery (c)

International Research Journal of Engineering Sciences Page 101

 Volume 6 Issue 2 December 2020 www.irjes.psyec.edu.in

 SAPV system: The SAPV system (Figure 6) was located at the test facilities in IMDEA Water
Institute (Alcalá de Henares, Madrid, Spain, latitude 40.513° N, longitude 3.339° E). The PV
system included an 80W mono-crystalline module, a 12 V lead-acid battery (90Ah- C100) and a
PWM serial charge controller.
 Linares SAPV system: The SAPV system was composed by a 100W mono-crystalline module, a
12 V lead-acid battery (90Ah-C100) and a PWM serial charge controller. The real SAPV system
(Figure7) was installed at the test facilities in the University of Jaén, in the Linares School of
Engineering (Linares, Jaén, Spain, latitude 38,085ºN, longitude 3,646ºW).
Another version of the ArduinoTM solar datalogger was installed in the School of Science of the
University Benito Juarez of Oaxaca (Oaxaca, Mexico, latitude 17.047° N, longitude 96.712° W).
This datalogger was connected directly to the Internet via cable, using the Ethernet input of the
board. In this case, the datalogger measures meteorological parameters; data is sent to the Linares
server.
Typical domestic energy consumption of a household in developing countries will usually have
several lamps for lighting and it might also have a small household appliance. A real load profile
was designed by López-Vargas et al. [19] identifying all the commonly used appliances in terms of
power and hours of usage per day for estimating the daily energy consumption used in this work.
The load system was composed by two 12 V luminaries of 3 W and 50 W, respectively. The total
daily energy consumption of the designed load profile was 79 Wh.

Figure 8 Weather Stations Installed in Alcalá de Henares (a) And Linares (b)

Each SAPV system had also their own low-cost modular weather station (Figure 8), including a
calibrated solar cell from NousolTM to measure solar global irradiance in the same plane of the PV
mono-crystalline module (Gi), a DHT22 digital sensor with shield protector to monitor ambient
temperature (Tamb) and relativity humidity (H), a Small Wind Transmitter sensor from Thies Clima
to measure horizontal wind speed (W) and a rain gauge sensor from PronamicTM to measure rainfall
(R).

International Research Journal of Engineering Sciences Page 102

 Volume 6 Issue 2 December 2020 www.irjes.psyec.edu.in

Experimental Results
Figure 9 shows the final prototype of the new low-cost datalogger installed in Alcalá de Henares
(Figure 9a) and Linares (Figure 9b). At both sites, the results from the testing period indicated how
the new 3G datalogger monitored and sent the data correctly for the complete period of time.

Figure 9 Final Prototype Assembly. Datalogger Installed in Alcalá de Henares

a) datalogger installed in Linares

b) The datalogger uses 9 small PCBs for electrical sensing (current and voltage measurements).
The load control system includes an ArduinoTM UNO board and a relay system.

The measurements were compared with a commercial monitoring system that acted as a pattern.
This pattern commercial monitoring system was an AgilentTM 34972 datalogger of 22 bits with a
voltage accuracy resolution of 0.004% that uses LabviewTM. Every 30 s, this datalogger measured
and recorded the same parameters as the ArduinoTM datalogger.

Electrical and Meteorological Parameters Monitoring: Comparison with Commercial Data

Logger
Lopez-Vargas et al. demonstrated in 2018 that the ArduinoTM data logger works well under
adverse climatic conditions [19]. As expected, the new data logger continued to measure correctly
both the meteorological and electrical parameters, always complying with the IEC standard.
Figure10 shows an example of the daily variations of the solar irradiance and the ambient
temperature monitored by the low- cost ArduinoTM data logger and the commercial AgilentTM
datalogger in Alcalá de Henares (Figure10a) and Linares (Figure 10b), revealing the correct

International Research Journal of Engineering Sciences Page 103

 Volume 6 Issue 2 December 2020 www.irjes.psyec.edu.in

functioning of the ArduinoTM datalogger: the uncertainly of irradiance measures was below the 8 %
required by the standard and the mean absolute errors of measured ambient temperatures were better
than the ±1 °C required.

Figure 10 Irradiance and Ambient Temperatures Measured in Alcalá de Henares (27/07/2017)

and Linares (20/03/2017) Sites Under Clean Sky Conditions

Regarding electrical parameters, Figure 11 shows an example of the daily variations of the PV
generator output voltage, the battery voltage and the load voltage monitored by the low-cost
ArduinoTM data logger in Alcalá de Henares (Figure 11a) and Linares (Figure 11b) sites. Battery
voltage line and load voltage line overlap. The voltages measured by the ArduinoTM datalogger
using
low-cost sensors under clear sky conditions shown in the Figure 11 demonstrate the correct
functioning of the ArduinoTM datalogger. The uncertainly of electrical measures was below the 2%
required by the standard.

System Reliability
During the experimental period, it has been verified that all the installed data loggers send data
correctly and that the quality of the data is excellent (meteorological and electrical) as indicated
previously. All the data loggers have been intensively tested over a period of 12 months, identifying
the performance of the data logger when a failure occurred. The data loggers only failed in 4
occasions over the year, due to different reasons. One of these errors was an access failure (to the
server installed in Linares). The datalogger located in Oaxaca registered 1 error as a result of a

International Research Journal of Engineering Sciences Page 104

 Volume 6 Issue 2 December 2020 www.irjes.psyec.edu.in

network outage. The datalogger installed in Alcalá de Henares failed in 1 occasion due to a power
shortage. An equipment failure was detected in the datalogger installed in Linares. The experimental
results showedthatafter12monthsunderoperation,takinginto account the low number of failures and
that the most of the failures are due to external factors, the datalogger is highly reliable. Details on
the errors of different origin were registered:

Figure 11 Voltages Measured in Alcalá de Henares (27/07/2017) and Linares (20/03/2017) Sites
Under Clean Sky Conditions

Access Failure
On 16/11/2017, it was detected an access failure to the server installed in Linares. This failure
affected all three systems (data loggers sent data to the same server). In the experimental phase, one
of the problems encountered was the loss of data sent for brief periods of time. As previously stated,
each datalogger sends the collected data to a dedicated server, located in the Superior School of
Linares.

International Research Journal of Engineering Sciences Page 105

 Volume 6 Issue 2 December 2020 www.irjes.psyec.edu.in

Figure 12 Global Horizontal Irradiance (G0 ) and in-Plane Irradiance (G17 ) Data Registered
in the Network Failure Period in Oaxaca, Data Recorded on 17/6/2017 (a) and 19/6/2017 (b).

Eventually, because of performing the maintenance activities by the university, the server
installed in Linares was disconnected for short periods of time (it was beyond our scope of work:
these maintenance works were carried out by the general university technical unit). In these periods,
the datalogger could not establish the connection to the server and data was not received or stored.
The server downtimes lead to observe the behavior of the datalogger under this condition: it was
verified that as soon as the server functioned again, the monitoring system was reconnected and the
data was sent again and stored as normal.

International Research Journal of Engineering Sciences Page 106

 Volume 6 Issue 2 December 2020 www.irjes.psyec.edu.in

Figure 13 Temperatures Recorded by the Datalogger in Linares, a Previous Day Before the
Failure (24/06/2017)

Power Failure
Another data: in Alcalá de Henares, on 21/05/2017 a power failure was detected in the Alcalá de
Henares site. Problems with AgilentTM pattern datalogger occurred due to power shortages that
forced to initialize the software and perform maintenance. The ArduinoTM datalogger was designed
to start-up without external help after power shortages and it did not require any external
initialization. This result is of special interest since the SAPV systems studied in this work are
usually installed in isolated and inaccessible regions. The datalogger does not require external help
to recover from a power failure; the datalogger does not need an external operator. In addition to
providing the reliability to the system, it was empirically proved the proper functioning of the
system without the need for maintenance, reducing the operational costs of the monitoring system to
a minimum

International Research Journal of Engineering Sciences Page 107

 Volume 6 Issue 2 December 2020 www.irjes.psyec.edu.in

Figure 14 Online Web-Page View of Data Measured in the Linares Site: Graphs on 13/06/2017

Network Failure
The solar ArduinoTM datalogger installed in Oaxaca provided information on how the new
monitoring system responds to a network failure. Jointly with the Ardino datalogger, another
commercial data logger (LabJackTM U12 series) was installed. This commercial datalogger stored
the data in a folder shared over the Internet. On 17/06/2017, the datalogger stopped sending data to
the server as shown in Fig 12.a. It was observed that, the commercial datalogger, stopped sending
data too. On 19/06/2017, the solar datalogger (Fig 12b) and the commercial datalogger started
sending data again. As the commercial datalogger has to be initialized manually in case of power
failure, the bug was registered as a network failure.

International Research Journal of Engineering Sciences Page 108

 Volume 6 Issue 2 December 2020 www.irjes.psyec.edu.in

Figure 15 Thing Speak™ Visualization of Parameters Measured by the Monitoring System

Located in Alcalá de Henares

The ArduinoTM datalogger stopped sending data to the server due to a (WAN) network error.
This failure served to verify empirically that after a network failure, the datalogger can re-hook and
resend the data, as designed by the software. Subsequently, other networks failures were registered
with a certain frequency in the Oaxaca site, and these failures have led to the loss of data in the fault
period. It was found that, often in developing countries, due to the quality of the wired networks it is
common to suffer successive Internet drops; it was empirically proved that the monitoring system
based on transmission via mobile communications is more reliable than a data acquisition system
that is wired network dependent.

Equipment Failure
The datalogger installed in Linares recorded a fall on 29/06/2017. The system was completely
inoperative; it stopped sending data to the server and the cloud. In Linares, the summer period
usually contains some of the harshest environmental conditions for PV systems and electronic
equipment, including temperatures up to 45º, dust storms (from winds from the Sahara Desert), and
occasional electrical storms with high speeds and heavy rain. In the days before the failure,

International Research Journal of Engineering Sciences Page 109

 Volume 6 Issue 2 December 2020 www.irjes.psyec.edu.in

ambient temperatures (Tamb) were recorded above 40 degrees Celsius as shown in Figure 13. Table
V includes a summary of the daily environmental parameters recorded on 24/06/2017 by the
datalogger, including temperatures, irradiance, humidity and horizontal wind speed.
Although a cooling system consisting of two fans was installed, in some instances, the Tec,
temperature measured inside the electrical cabinet (where the datalogger was located), exceeded the
ambient temperature, reaching 47 degrees Celsius as shown in Table V. The operating temperature
of the entire connectivity system (router and modem) recommended by the manufacturer was 10 ºC -
40 ºC. As the maximum operative conditions of the system were reached, by all accounts, it was an
equipment failure in the connectivity system due to the high temperatures. After inspecting the
system it was found that, in effect, the data delivery system (consisted of the nano router and the
modem) stopped working. The connectivity system was not broken, but it was necessary to reset the
nano router and configure it again manually.

Data Visualizations
As the main novelty of this work, the remote visualization of the data in real time via web or
smart phone stands out, due to the incorporation of connectivity to the system via 3G. The
parameters monitored are processed and analyzed to produce information that can be understood
easily by users; the information is represented in graphs. Figure 14 shows a view of the dynamic
charts integrated in the website.
Figure 15 shows the visualization of data measured and sent to the cloud on 23/01/2017 in
Alcalá de Henares. The open cloud platform selected, ThingSpeak™, allows to monitor up to eight
parameters and to locate the monitoring system on the map via www.

Figure 16 Screenshots on an Android Smart Phone Using Thing viewTM App: Weekly (a) and
Daily (b) Ambient Temperature Monitored by the Datalogger Located in Alcalá de Henares.

International Research Journal of Engineering Sciences Page 110

 Volume 6 Issue 2 December 2020 www.irjes.psyec.edu.in

Table 6 Cost Estimation of the First Prototypes

The ultimate goal of this work was to develop an IoT-based prototype of a datalogger that uses a
mobile communication transmission system and is low-cost, accurate and autonomous. Table VI
describes the budget of the monitoring system. The final cost of the new prototype, including
sensing and connectivity, was approximately 141 €. Table VII shows the budget for the low cost
sensors. The budget of low-cost sensors (including the low-cost weather station) was approximately
268 €. Total costs could be reduced considerably when mass produced and introduced in the market.

This monitoring system has a maintenance cost due to the use of a 3G SIM card. This
maintenance cost depends on the fee of the contracted telephone service. As the information
transmitted is very low, with a cheap data plan should be sufficient, so this expense is not
significant.

International Research Journal of Engineering Sciences Page 111

 Volume 6 Issue 2 December 2020 www.irjes.psyec.edu.in

Conclusion
A novel datalogger for monitoring SAPV systems via website and mobile applications has been
designed, built and tested incorporating wireless technology in it, concretely 3G, along with IoT.
Two types of data storage were tested: a dedicated server (conventional method) and a cloud service
platform (novel application of IoT). Both alternatives of storage allow to monitor the system in a
remote way. The integration of an open cloud-based IoT platform allowed to monitor small stand-
alone PV systems remotely in real time via web or mobile app at low cost.
The novel datalogger based on free software and hardware has been installed as an experimental
prototype in multiple sites: Alcalá de Henares (Madrid, Spain), Linares (Jaén, Spain) and Oaxaca
(Oaxaca, Mexico). The datalogger was tested under real and different conditions demonstrating the
robustness of the system, working with 14 variables with low uncertainty. An outdoor campaign of
over ten months was performed. Considering the low number of failures and their origin (the most
of the failures were due to external factors), the experimental results showed that the datalogger is
highly reliable. The cost of the monitoring system is considerably lower than commercial devices
and allows high accurate remote monitoring.
The main novelty presented by this work is that the datalogger complies with the IEC standard
accuracy requirements and it is designed based on wireless communications taking into account the
limitations in rural areas of developing countries. The integrated solution has been tested and it
works, so the datalogger could be installed solving the problem of monitoring of PV systems located
in isolated regions, all at lowcost.
As future works, to avoid the loss of information in case of failures (network, access…), a
backup system based on a micro SD card would add value to the system; it would prevent the loss of
data, storing the measured parameters in the disconnection period. In addition, it was proven that the
connectivity system, working in extreme conditions (temperatures above 40 º C) fails. As a future
line, another type of cooling system more effective and another 3G connectivity system that
supports higher temperatures could be tested. To power the monitoring system using a solar cell
could be another future improvement. Due to the stackable hardware design (by adding boards)
several SAPV systems installed close enough can be monitoring using a single datalogger as a
central node. The development of these networks is another futurework.

References
1. IEA. Energy Access Outlook 2017report.
2. <http://www.iea.org/>(accessed19.09.2018)
3. R. Kempener, O. Lavagne, D. Saygin, J. Skeer, S. Vinci, and D. Gielen, “Off-Grid Renewable
Energy Systems: Status and Methodological Issues”, International Renewable Energy Agency,
2015.
4. N.J. Williams, E. Ernest van Dyk, F.J. Vorster, “Monitoring solar home systems with pulse
width modulation charge control”. J Sol Energy Eng2011;133:021006-1–021006-7.

International Research Journal of Engineering Sciences Page 112

 Volume 6 Issue 2 December 2020 www.irjes.psyec.edu.in

5. M.Fuentes, M. Vivar, H. Hosein, J. Aguilera, E. Muñoz-Cerón,„ Lessons learned from the field
analysis of PV installations in the Saharawi refugee camps after 10 years of operation‟,
Renewable and Sustainable Energy Reviews, Vol 93, pp.100-109, 2018..
6. R. Mukaro, X. Carelse, and L. Olumekor, “First performance analysis of a silicon-cell
microcontroller-based solar radiation monitoring system”, Sol. Energy, vol. 63, no. 5, pp. 313–
321, Nov.1998.
7. E. Koutroulis and K. Kalaitzakis, “Development of an integrated data- acquisition system for
renewable energy sources systems monitoring”, Renew. Energy, vol. 28, no. 1, pp. 139–
152,2003.
8. N. Forero, J. Hernández, and G. Gordillo, “Development of a monitoring system for a PV solar
plant”, Energy Convers. Manag., vol. 47, pp. 2329–2336,2006.
9. G. M. Tina and A. D. Grasso, “Remote monitoring system for stand- alone photovoltaic power
plants: The case study of a PV-powered outdoor refrigerator”, Energy Convers. Manag., vol. 78,
pp. 862–871, 2014.
10. M. Benghanem, “Low cost management for photovoltaic systems in isolated site with new IV
characterization model proposed”, Energy Convers. Manag., vol. 50, no. 3, pp. 748–755,
Mar.2009.
11. M. Demirtas, I. Sefa, E. Irmak, and I. Colak, “Low-cost and high sensitive microcontroller
based data acquisition system for renewable energy sources”, 2008 Int. Symp. Power Electron.
Electr. Drives, Autom. Motion, pp. 196–199,2008.
12. M. Benghanem and A. Maafi, “Performance of stand-alone photovoltaic systems using
measured meteorological data for Algiers”, Renew. Energy, vol. 13, no. 4, pp. 495–504,
Apr.1998.
13. M. Benghanem, A. H. Arab, and K. Mukadam, “Data acquisition system for photovoltaic water
pumps”, Renew. Energy, vol. 17, no. 3, pp. 385– 396, Jul.1999.
14. M. Benghanem, “Measurement of meteorological data based on wireless data acquisition system
monitoring”, Appl. Energy, vol. 86, no. 12, pp. 2651–2660, Dec.2009.
15. F. Touati, M. A. Al-Hitmi, N. A. Chowdhury, J. A. Hamad, and A. J. R. San Pedro Gonzales,
“Investigation of solar PV performance under Doha weather using a customized measurement
and monitoring system”, Renew. Energy, vol. 89, pp. 564–577,2016.
16. Purwadi, Y. Haroen, F. Y. Ali, N. Heryana, D. Nurafiat, and A. Assegaf, “Prototype
development of a Low Cost data logger for PV based LED Street Lighting System”, Proc. 2011
Int. Conf. Electr. Eng. Informatics, no. July, pp. 1–5, 2011.
17. M. Ikhsan, A. Purwadi, N. Hariyanto, N. Heryana, and Y. Haroen, “Study of Renewable Energy
Sources Capacity and Loading Using Data Logger for Sizing of Solar-wind Hybrid Power
System”, Procedia Technol., vol. 11, no. Iceei, pp. 1048–1053,2013.
18. M. Fuentes, M. Vivar, J. M. Burgos, J. Aguilera, and J. A. Vacas, “Design of an accurate, low-
cost autonomous datalogger for PV system monitoring using ArduinoTM that complies with IEC
standards”, Sol. Energy Mater. Sol. Cells, vol. 130, pp. 529–543, Nov.2014.

International Research Journal of Engineering Sciences Page 113

 Volume 6 Issue 2 December 2020 www.irjes.psyec.edu.in

19. López-Vargas, M. Fuentes, M. Vivar and F.J. Muñoz-Rodríguez, “Low-cost datalogger intended
for remote monitoring of solar photovoltaic stand-alone systems based on
ArduinoTM”,Submitted.
20. IEC61724-1. Photovoltaic System Performance- Part 1: Monitoring. 1st Ed., International
Electro technical Commission (IEC), Switzerland, Geneva, 2017.

International Research Journal of Engineering Sciences Page 114