Real-time Battery Monitoring System for Solar

Infrastructure
Saurav Malve Akash Nayak Rohan Thape
Dept. of Electronics Dept. of Electronics Dept. of Electronics
Sardar Patel Institute of Technology Sardar Patel Institute of Technology Sardar Patel Institute of Technology
Mumbai, India Mumbai, India Mumbai, India
[email protected] [email protected] [email protected]

Govind T. Haldankar
Dept. of Electronics
Sardar Patel Institute of Technology
Mumbai, India
g [email protected]

Abstract—This paper presents a straightforward approach These initiatives have increased the investments in renewable
towards existing battery monitoring systems with solar input power, particularly solar. This has had a direct effect on the
which use a series of batteries. The presented system strips down tariffs of solar power generation, which have reduced to Rs.
the monitoring to systems requiring a single battery. The two
parameters being measured are battery voltage and charging 2.44 per unit from the previous Rs. 4.43 per unit. India has
current, from which many other characteristics like depth of recorded a 138% growth in renewable power capacity from
discharge, state of charge and state of health of battery can 34,000 MW to 82,580 MW, hence standing fifth globally in
be computed. Employing a simple mechanism for design where renewable power installed capacity. The country, being located
an Arduino UNO board is interfaced with easily available and in tropical suburbs, can extensively exploit the abundant sun-
economical voltage and current sensors for reading data, this
system further displays the readings and plots the characteristics light for meeting the energy demands of the nation [3].
on an easy to use Graphical User Interface (GUI) made using Nevertheless, solar power has its fair share of problems.
Python. The system presented has a mechanism for voltage cut-
One of them is its dependency on the weather conditions, i.e.
off to prevent over-charging of battery. The data acquisition
from the battery is done in real-time and displayed with all the Sun. Solar output depends on the nature and positioning of
the minute fluctuations in the output characteristics which occur the Sun and is prone to weather fluctuations. Power generation
due to weather changes from the solar panel input. Moreover, from Solar Photovoltaic plants is variable in nature due to
this data is sent to an Excel spreadsheet for recording purposes changes in solar irradiance, temperature and other factors
and further technical analysis.
[4]. These weather fluctuations result in the fluctuation of
Index Terms—battery monitoring, solar, Arduino, Python,
Excel output which is undesirable. Another drawback is solar energy
storage, which is done in expensive lithium-ion batteries and
I. I NTRODUCTION undergoes energy loss during transmission. An application
India being one of the fastest growing economies in the where inexpensive lead-acid batteries are used instead of
world, the need for energy demand in the country is enormous. efficient lithium-ion batteries are electric rickshaws. These
To meet such a demand, apart from the conventional sources commercially available and widely-used lead-acid batteries
of energy generation, India is focusing heavily on using that can also be connected to solar panels for energy storage do
renewable energy. Most of the people rely on the conven- not come with battery monitoring systems. Hence a monitoring
tional energy based electricity production for meeting their setup for applications having a single lead-acid battery which
daily needs [1]. With increasing demand, ensuring widespread lets the user know what the battery is going through and how it
availability of energy to every part of the country plays an affects their usage is essential. With the aid of such a system,
important role for growth. One of the solutions to achieve this various parameters of a battery, their behavior, fluctuations,
targeted growth is boosting renewable energy infrastructure in responses, characteristics can be obtained and monitored.
the country. Presently, various battery monitoring systems are opera-
According to the Ministry of New and Renewable Energy tional. However, a high cost of maintenance and operation
(MNRE) [2], during the year 2019-20, India has completed impede the exercise of systematic real-time monitoring. Other
a total of 75,919.99 MW of renewable energy capacity till monitoring systems have also been proposed, for instance
December 2019. With various government schemes in place, in [5], a virtual PV panel has been made in PSIM and is
India aims to achieve 175 GW of renewable power by 2022. then monitored. In [6], a battery management system with
978-1-7281-8873-7/20/$31.00 ©2020 IEEE

Authorized licensed use limited to: University of Cape Town. Downloaded on May 15,2021 at 20:11:33 UTC from IEEE Xplore. Restrictions apply.
three modes has been proposed, but without a user interface This data is acquired by the Arduino UNO board. The Arduino
and a record for data for easy end-user operation. A battery board and Python are then serially communicated and Python
monitoring system for lead-acid batteries has been proposed gets access to this data. A GUI made using Python, displays
in [7], which uses Arduino and sensors to monitor and display this data and plots the characteristics, namely current-time
battery parameters on an LCD screen. All these solutions are and voltage-time accordingly. The data from Python is sent
efficient and economical and leave a space for improvement to an Excel spreadsheet, where it is stored. Thus the task of
and advancement in further studies on battery monitoring. acquiring instantaneous real-time readings and recording them
The presented work takes advantage of this space open for is being achieved.
improvement and introduces a monitoring infrastructure which
consists of a parametric module. This module contains an Ar- III. S YSTEM D ESIGN
duino UNO board with current and voltage sensor modules. In This system focuses on the basic aspects of a traditional
addition a Graphical User Interface (GUI) has been made using battery monitoring system, hence all the components used, are
Python to display the battery readings in real-time. Microsoft readily available. Fig. 2 shows the actual setup of the proposed
Excel is used to store the data for records. The amalgamation system. The components and their uses are as follows:
of all these components is described subsequently.
• Solar panel: 38V output with 315W power rating.
The paper is organized as follows: Section II describes
• Lead-acid battery: 12V DC, 42Ah to store energy from
the proposed system. Section III describes the system design.
the solar panel.
Section IV describes the working of the system. Section V
• Inverter: Acts as an uninterrupted power supply in the
describes the observations and results and Section VI provides
system to convert DC supply to AC supply for further
the conclusion.
widespread use.
II. P ROPOSED S YSTEM
Fig. 1 shows the illustrative block diagram of the proposed
system. Solar infrastructure fundamentally consists of solar
panels, a battery and an inverter. The solar panel output cannot
be measured directly because it has a variable range, in some
cases hundreds of volts and becomes difficult to handle. Hence
a DC to DC buck converter module is used to scale down the
voltage as per the battery requirement. A battery is required
for storing energy from the panel output in DC form. The
battery can then provide input to the inverter. Since solar panel
output is contained in the direct current (DC) form, an inverter
is a required for the conversion of direct current to alternating
current (AC), on which most appliances work.

Fig. 2. Actual Setup

• DC-DC buck converter: Designed to reduce panel output

voltage from 38V down to 12V. A voltage regulator
cannot be used directly because of the high current and
voltages involved. If a commonly used voltage regulator
(LM317) is considered, its maximum voltage difference is
40V, hence it cannot be used. Furthermore, this regulator
provides an output current of 1.5A at best, resulting in
extremely slow charging of the battery, given that the
Fig. 1. Block Diagram of proposed system battery can be provided with 10.5A safely. We designed a
circuit of power transistors (2N3005) and power resistors.
This setup of battery and inverter is realized for their voltage 2N3005 is a common transistor used in power circuits,
and current. This is done via voltage and current sensors avail- having a rating of 60V, 15A. These transistors are used
able for Arduino development boards. The potential difference with heat-sinks so that the heat generated is efficiently
across battery and charging current of battery are obtained via dissipated. Fig. 3 shows the DC-DC buck converter circuit
these sensors. As solar infrastructure is critically dependent diagram simulated on Proteus software which was further
on battery, it is very essential to monitor battery parameters. used for hardware design.

Authorized licensed use limited to: University of Cape Town. Downloaded on May 15,2021 at 20:11:33 UTC from IEEE Xplore. Restrictions apply.
 function in Python requires the port number on which the
Arduino board is connected along with the baud rate on which
it operates. This ensures that Python reads the data provided by
Arduino board. Now that Python has the current and voltage
readings, a GUI is created to display these readings along
with the characteristic graphs of current-time and voltage-
time. These graphs can be interchangeable for parameters,
i.e. a current-voltage graph can also be plotted. A Python
library called tkinter is imported to create the GUI. This
library enables the GUI to live stream the data, i.e. display
the readings in real-time. A threshold value of 13.5V is set
in the GUI and whenever battery voltage exceeds 13.5V, the
GUI displays an alert message.
The data in Python can now be used to store in a spread-
Fig. 3. Schematic diagram of DC-DC Buck converter. sheet. This is made possible by using the xlwt and xlrt
libraries in Python from which many operations like writing
and reading on spreadsheet can be performed. A data table is
• Arduino UNO board: Contains the ATmega328P micro- created in Python and this data is sent to an Excel spreadsheet.
controller to interface current and voltage sensors and This spreadsheet is kept for record-keeping purposes and
acquire current and voltage readings from them, test them historical data can also be used for further analysis and to
on Arduino IDE and serially send the acquired data to gather insights.
Python.
• Voltage detection sensor module: 25V input capacity, V. OBSERVATIONS AND RESULTS
converts the input to 5V output required for the Arduino The system presented here gathers insights about the solar
board’s working. It is a 5:1 voltage divider that uses a panel behavior by monitoring the battery. Fig. 4 shows the GUI
30K and 5K resistors. displaying voltage and current readings along with the current-
• Current Sensor Module: 5A range based on the ACS712 time and voltage-time graphs. These readings are taken during
sensor which detects current accurately based on Hall the monsoon season of August 2019 in Mumbai. The season
effect. Used to detect current from the battery. was chosen as energy storage in batteries is most critical
IV. W ORKING OF THE SYSTEM during monsoon, when an increased use of stored energy takes
place. These weather conditions also allow to the GUI to
This system’s working starts off with the accurate con- showcase the fluctuations that occur in the battery parameters
nections of the solar panel output with the battery and the due to continuous rain in real-time.
inverter. The panel output is connected to a DC-DC buck
converter which steps down the voltage of the panel from
38V to 12V which is the voltage up to which the battery can
be charged. Hence, the battery must be chosen conveniently
according to the panel and converter output ranges. The
ACS712 current sensor is connected in series to the battery
to sense the charging current of the battery. The voltage
sensor is connected across the battery, to measure the potential
difference between the two terminals. Only current and voltage
are being monitored because many other parameters can be
effectively calculated from these two parameters.
These sensors are connected to the analog input pins A0
and A1 of the Arduino UNO board. These pins read the
data sent by the sensors, because the Arduino UNO board Fig. 4. Graphical User Interface displaying readings and characteristics.
is programmed to take input from these sensors by using
the AnalogRead function in Arduino IDE specified libraries. Another insight or proof of concept demonstration con-
These analog values of voltage and current are converted to ducted during this session is the change that occurs in charging
digital using the 10-bit analog to digital converter (ADC) in current and voltage when a load is applied to the battery.
the board. The compiled code is uploaded on the board and This test has been conducted in order to check the accuracy
the instantaneous analog values are displayed and tallied with of the proposed monitoring system. The load connected in
actual readings using multi-meter. this case is a 66W ceiling fan. The tables below show the
Serial communication between Arduino board and Python is readings gathered during two load conditions. Table I shows
established by using the pySerial library in Python. The serial the charging current and voltage when the fan does not work

Authorized licensed use limited to: University of Cape Town. Downloaded on May 15,2021 at 20:11:33 UTC from IEEE Xplore. Restrictions apply.
on the battery. Table II shows the same parameters when the if a battery has an open circuit voltage of 14V, and a reference
fan works on the battery. The values of the charging current level of 12V, its DoD is 15% approximately. When a battery
and voltage are in amperes and volts respectively. discharges completely, its DoD is 100%. State of Charge (SoC)
refers to the available capacity of the battery, relative to a
TABLE I reference level. If the DoD of a battery is 20%, then the SoC
READINGS WITH NO LOAD CONNECTED of the battery is 80%. State of Health (SoH) is estimated as
Charging Current (A) Voltage (V) the ratio of the current capacity of the battery to the initial
0.37 12.01 capacity. From the voltage readings of the battery which have
0.34 12.07 been measured here, these battery parameters can be easily
0.32 12.07
0.29 12.19 calculated, which are useful to determine the state and health
of the battery.

VI. C ONCLUSION
TABLE II
READINGS WITH LOAD CONNECTED This system’s objective is to provide a straightforward
digital solution for the increasing need for battery monitoring
Charging Current (A) Voltage (V)
0.33 11.96 in solar infrastructure, single battery systems and for the
0.41 11.85 inverters which do not have built-in monitoring software. One
0.52 11.73 factor which differentiates it from other systems is the use of
0.65 11.56
an interactive GUI and its clarity on the fluctuations that take
place in the battery. Furthermore, without needing experienced
As there is no load connected, the charging current de- technicians, it is easy to install and operate in a wide variety
creases and the voltage of the battery remains same. When the of settings where solar energy can be used, be it domestic or
load is connected, the charging current increases and voltage industrial. Additionally, this system has been used to verify
decreases. This proves the theoretical concept that the voltage the behavior of the battery during different load conditions,
of the battery is inversely proportional to the charging current. hence it can be used to identify and confirm well-known facts
and can act as an experimental apparatus.
This work leaves open a variety of modifications that can be
made to make the system better, like adding a human-machine
interface (HMI) that displays the GUI on a touch screen. Many
more parameters like the temperature of the battery and panel
can be added to further increase the scope. A separate product
or service in the scope of this paper can also be made to
enhance the customer superiority and control over their battery
and their solar infrastructure.

REFERENCES
[1] M.Suresh and R.Meenakumari, “Optimal Planning of Solar
PV/WTG/DG/Battery Connected Integrated Renewable Energy
Systems for Residential Applications using Hybrid Optimization,” Int.
J. Ind. Eng., vol.2, no.1, pp.15-20, Jan. 2018.
[2] Ministry of New and Renewable Energy, ”Annual Report 2019-
20”, Government of India, Jan. 2020. [Online]. Available:
https://mnre.gov.in/img/documents/uploads/file f-1597797108502.pdf
[3] Ambuj Gupta, Rishabh Jain, Rakshita Joshi and Dr. Ravi Saxena,
”Real Time Remote Solar Monitoring System,” in 3rd Int. Conf. Ad-
vances Comput., Commun. & Automat. (ICACCA) (Fall), 2017, doi:
10.1109/ICACCAF.2017.8344723.
[4] Soham Adhya, Dipak Saha, Abhijit Das, Joydip Jana, and Hiranmay
Saha, “An IoT based smart solar photovoltaic remote monitoring and
control unit,” in 2nd Int. Conf. Control, Instrum., Energy & Commun.
(CIEC), 2016, pp.432-436, doi: 10.1109/CIEC.2016.7513793
Fig. 5. Excel spreadsheet for data storage. [5] Aboubakr El Hammoumi, Saad Motahhir, Abdelilah Chalh, Abdelaziz
El Ghzizal and Aziz Derouich, ”Low-cost virtual instrumentation of
Further, Fig. 5 shows the data stored on an Excel spread- PV panel characteristics using Excel and Arduino in comparison with
traditional instrumentation,” Renewables: Wind, Water and Sol., vol.5,
sheet for further analysis. The readings are taken on a half no.1, Dec. 2018.
hourly basis. This data signifies the performance of the battery [6] William Kevin Siagian, Daniel Fernando Purba, Asrilani Sipahutar,
and the changes it undergoes. and Indra Hartarto Tambunan, “Design and implementation of battery
management system for on-grid system,” in Int. Conf. Control, Electron.,
Depth of Discharge (DoD) refers to the degree to which a Renewable Energy and Commun. (ICCEREC), 2018, doi: 10.1109/IC-
battery is discharged, relative to a reference level. For example, CEREC.2018.8712088.

Authorized licensed use limited to: University of Cape Town. Downloaded on May 15,2021 at 20:11:33 UTC from IEEE Xplore. Restrictions apply.
[7] Niraj Agarwal, Phulchand Saraswati, Ashish Malik, and Yogesh Krishan
Bhateshvar, “Design a battery monitoring system for lead-acid battery,”
in Int. Conf. Commun. & Comput. Technol., Dec. 2017, pp.306-311.

Authorized licensed use limited to: University of Cape Town. Downloaded on May 15,2021 at 20:11:33 UTC from IEEE Xplore. Restrictions apply.