Microsoft Word - 32-3121_s1_ETASR_V9_N6_pp5041-5046


Engineering, Technology & Applied Science Research Vol. 9, No. 6, 2019, 5041-5046 5041 
 

www.etasr.com Dhaked et al.: Battery Charging Optimization of Solar Energy based Telecom Sites in India 

 

Battery Charging Optimization of Solar Energy based 

Telecom Sites in India 
 

Dheeraj Kumar Dhaked 

Department of Electrical Engineering  
Rajasthan Technical University 

Kota, Rajasthan, India 

ddhakar9@gmail.com 

Yatindra Gopal 

Department of Electrical Engineering  
Rajasthan Technical University 

Kota, Rajasthan, India 

ygopal.phd@rtu.ac.in 

Dinesh Birla 

Department of Electrical Engineering  
Rajasthan Technical University 

Kota, Rajasthan, India 

birlartu2@gmail.com 
 

 

Abstract—Telecom sites get the power normally from the grid. At 

the occurrence of power outages power need to be supplied for 
telecom sites. The battery bank is a good option for telecom sites 

to fulfil power demand. This paper discusses a smooth battery 

bank charging and discharging system with solar power as the 

input supply source. The system requires a large capital 

investment, but it provides uninterrupted emergency power when 

needed. Maintaining battery banks is essential for getting 

optimum performance. This paper also discusses the power 
requirements for telecom sites backup and various parameter 

impacts on battery life. Methods are derived to optimize charging 

management of batteries in order to get maximum lifespan in 
addition to better battery performance throughout its useful life. 

Keywords-telecom site; battery bank; charging-discharging; 

maximum power point tracking (MPPT) 

I. INTRODUCTION  

The telecom industry of India is growing at a very fast 
pace. Telecom sites require uninterrupted power supply [1]. 
The main problem of the telecom sites is that rural areas have 
only 6 to 10 hours of power supply from the grid. To fulfil this 
demand, two ways are possible, either by grid or by some other 
renewable source, i.e. solar, biomass, wind, etc. [2]. Also, there 
are other issues which make this solution worthy: 

• The dismal state of rural electrification  

• Non-availability of grid power in rural areas  

• The grid might be away from the site 

• Poor grid power quality 

• Unreliable & erratic power behavior  

The use of renewable power sources also reduces 
environmental deteriorating agents, as carbon emission for 
power generation, thus rural sites will get reliable power for 
operation. Solar energy is one of the best options as it is totally 
free of cost and can be harvested easily with available means 
like solar plates, inverters, and converters. The power harvested 
from the sunlight during the day can also be used during the 
night hours if stored in battery banks [3-4]. The power demand 
for a single telecom site ranges from 1 to 4kW for 24×7 hour 
operation. For back-up, it requires a 15-20kVA diesel generator 
set (DG) which has 4l/h diesel consumption. The running time 

of the DG set is 6-10 hours per day in grid-connected areas but 
for about 14-16 hours in rural areas the grid is not available. 
The burning of diesel causes the emission of greenhouse gases 
(GHG) (carbon dioxide, methane, nitrous oxide etc.). Solar 
energy can fulfil telecom sites’ load in day-time, while during 
night hours charged batteries can supply the system. The 
battery banks set-up will be placed at the telecom site and as an 
economical solution for back-up [5]. Lead-acid battery model 
for simulation with solar system via a DC-DC converter with 
battery set-up was used in [6-7]. 

In this paper, particle swarm optimization (PSO) algorithm 
is used for battery charging. This paper also describes an 
equivalent circuit model for battery and solar cell, and a battery 
model is made to match the model from the datasheet of the 
manufacturer. The performance of the battery cell, while being 
charged by solar array cell is also simulated. Multiple modes of 
operation and scenarios which cause early battery failure are 
described in [8-11] and are taken into consideration while 
proceeding to develop the most suitable battery charging 
method. 

A. Study Objectives 

• To introduce the reader to current energy-efficient 
technologies which are environment supportive. 

• To study the solar fed battery bank system with model 
implementation in Matlab. 

• To study the behavior of the model for battery charging and 
discharging and analyze its characteristics. 

B. Problem Statement 

• Energy consumption from telecom network is an increasing 
contributor to GHG emission. 

• Limiting carbon dioxide emissions will be more difficult for 
most of the countries where electric generation comes 
primarily from coal such as India. 

• The carbon emission footprint of the telecom sector has 
risen significantly and will rise despite the development of 
energy-efficient technologies. 

Corresponding author: Dheeraj Kumar Dhaked 



Engineering, Technology & Applied Science Research Vol. 9, No. 6, 2019, 5041-5046 5042 
 

www.etasr.com Dhaked et al.: Battery Charging Optimization of Solar Energy based Telecom Sites in India 

 

II. METHODOLOGY OF BATTERY CHARGING OPTIMIZATION 

As the scope of this manuscript is to develop modified 
charging algorithms for batteries, one has to reach to the exact 
traditional charging profile before using the suggested 
algorithms, according to battery chemistries and strict Do’s and 
Don’ts which must be kept in the algorithms. While varying 
parameters of batteries and changing operating conditions, 
critical parameters, like thermal dissipation which causes 
gassing and ageing of batteries, have to be observed and 
improved [6]. Input varying conditions, like solar charging 
current, have to be simulated in most real-time scenarios to see 
the impact and sizing of solar panels to be optimized [4]. In 
most cases, the power input from solar panels is insufficient to 
keep the batteries full charged, especially during sunless days. 
In such conditions, the battery remains in partial state of charge 
(PSOC) and deep cycling. Solar systems are installed in 
conditions where battery temperatures remain usually high [7-
8]. Normal lead-acid batteries fail in such conditions due to 
sulphating, corrosion, and active material shedding. Water top-
up in remote sites is difficult also. Hence, the best option is gel 
batteries with VRLA technology. For solar applications, 
charging hours are limited by sunlight availability. Boost 
charge can be given every day since charging current and 
ambient temperature during the evening are already decreased 
to a certain level [9-11]. It would not lead to any heat 
generation or any other negative effect, but will help keep 
sulphation away because higher voltage kills sulphation and 
hence plates remain fresh and battery cycle life increases. In 
order to achieve the implementation of the above, a nominal 
charging algorithm has been modified to give everyday 
equalization charge keeping close monitoring of cell 
temperature avoiding any ill effect of heat to cycle life [12]. By 
removing sulphation, lead plates remain fresh and hence 
battery lasts longer as lead plate sulphation is a significant part 
in determining the battery life [13-15]. 

III. MPPT AND BATTERY CHARGING OPTIMIZATION 

Optimization is the mechanism which finds the maximum 
or minimum value of a function or process. It is used in fields 
such as physics, chemistry, economics, and engineering where 
the goal is to maximize efficiency, production, or some other 
measure. To gain the maximum power output, an MPPT 
technique is used in the form of an electronic system that 
operates with the PV system. The maximal power point (MPP) 
doesn’t reach an exact end or point but it gets moving around 
the PV curve that depends on light intensity of irradiation and 
the temperature. The PSO MPPT is used and the flowchart of 
this algorithm is given in Figure 1. The PSO algorithm finds 
the best possible resolution of swarm particles through the 
particle’s progress in the investigation space with its optimized 
velocity and position. It finds the optimum solution by 
obtaining the minimum value of the given objective function 
[16]. The outline of the PSO algorithm is presented below. 

Step I: Initialize the particles with random numbers having 
a consistent distribution: Y=Vrand (plowerlim, pupperlim). 
Assign this position to the best known position array: r=Y. 
Initialize particle velocity: V=Y. If the number of particles is 
Nump then, Y is the size of the array of particle positions. 

Similarly, r is an array of pbest positions, and V is a Nump-size 
array of particle velocities. 

Step II: Calculate the fitness function: Ey=F(Y), Er=F(r), 
and eg=f(gbest), where Ey and Er are the fitness evaluation 
arrays for y and r, respectively, and eg is the function 
evaluation at gbest. 

Step III: Update pbest value for each particle of the 
population: if Ey(i)<Er(i) then r(i)=Y(i), where i= 1, 2, . . . 
Nump. 

Step IV: Update gbest value for the entire population:  

if Er(i)<eg then gbest=r(i). 

Step V: Update the velocity and position of the particles: 

( ) ( ) ( ) ( ) ( )( ) ( )

( )( ) ( ) ( ) ( )
1 2

1   

,

           ,  

1 0,1 0  1

  1   

V i wV i C rrand r i Y i C rrand

gbest Y i Y i Y i V i

+ = + − +

− + = + +
 

where w is the initial weight, C1 is the cognitive parameter, and 
C2 is the social parameter. 

Step VI: Exit criterion is generally a fitness threshold or a 
highest number of iterations. If the criterion is met, then exit or 
else, go to step II. 

The PSO works by simultaneously maintaining several 
candidate solutions in the search space [6]. During each 
iteration, each candidate solution is evaluated by the objective 
function to be optimized, determining the fitness of that 
solution. Each candidate solution can be thought of as a particle 
“flying” through the fitness landscape finding the maximum or 
minimum of the objective function. 

IV. MODELING AND SIMULATION  

The equivalent circuit model of a lead-acid battery is a 
voltage source connecting with the internal resistance. State of 
Charge (SOC), temperature and other elements of the battery 
parameters of circuit are related by the relationship between the 
resistance and voltage of capacitor. Each cell of the lead-acid 
battery produces 2V. Total terminal voltage will vary according 
to the working conditions. The concentration of the acid will 
change during charge and discharge [10]. The experiment has 
the following steps. 

• Study of the battery used in the experiment and research 
data from various sources. 

• Design of the battery circuit and the block diagram of a 
mathematical model using Simulink. 

• Simulations at different temperatures from 25 to 60
o
C have 

to be conducted and recording for current and voltage graph 
has to be done. 

In this section, modeling of lead acid battery is done using 
the Simscape language in order to implement nonlinear 
equations of the component of equivalent circuit as shown in 
Figure 2. In this way, defining physical equations and the 
connection between model components described. The study in 
[13] is referred for defining equations and their validation. 



Engineering, Technology & Applied Science Research Vol. 9, No. 6, 2019, 5041-5046 5043 
 

www.etasr.com Dhaked et al.: Battery Charging Optimization of Solar Energy based Telecom Sites in India 

 

Start

Set default values of PWM

Initialize Controller drivers

Initialize ADC drivers

Get battery capacity from user

Set battery current limit threshold to capacity/5 i.e. C/5

Set battery LVD, LV, OVD, OVR thresholds

Read battery voltage, battery current & temperature

Calculate battery SOC

Calculate battery voltage and current threshold based on temperature

Set battery charging mode-boost or float

Is battery
Voltage <LVD

Is battery
Voltage >LVR

Is battery
Voltage>HVD

Is battery
Voltage <HVR

Is charging
Threshold reached

Load charging profile from memory, Raise the PWM 

Store the report in memory and regulate the current and voltage

Turn off Load port

Reconnect the load

Disable charging Port

Enable charging Port

Yes

Yes

Yes

Yes

Yes

No

No

No

No

No

 
Fig. 1.  PSO algorithm flowchart for digital control loop for battery 
charging 

Initially, the battery is discharged at a load current of 10A 
in the simulation and then further recharged at a current of 10A 
to return to the initial state of charge. A simple thermal model 
is used to model battery temperature. It is considered that 
cooling is done primarily via convection method and that 
heating is mainly from the internal battery resistance R2. A 
standard 12V lead-acid battery is modelled by connecting six 
2V battery cell blocks in series, or for telecom applications of 
48V, more 2V cells should be used in series [10-11]. Once the 
battery model is developed, then to simulate the application, a 
simulation profile is created where both charging and 
discharging parameters using controlled current sources and 
controlled load sources are programmed to simulate the actual 
conditions of solar panel and grid-based battery charging [14-
15]. 

 

 

Fig. 2.  Battery simscape model 

 
Fig. 3.  Simulink model for lead acid battery capacity=100Ah, battery nominal voltage=52V, response time=30s, initial SOC=100% 

V. DESCRIPTION OF SIMULATION CIRCUIT 

The simulation circuit components of Figure 3 are: 

• Energy source: A solar panel providing 35V voltage and 
10A current. 

• Battery: A battery model with 48V nominal voltage and 
respective internal resistance as per battery model. 

• Load: One controlled resistance is used in the simulation 
circuit which represents the telecom load for which the 
entire system has to be made. 

• Voltage conversion module: A DC-DC boost converter to 
change the solar panel voltage to battery charging voltage 
in a controlled manner. 



Engineering, Technology & Applied Science Research Vol. 9, No. 6, 2019, 5041-5046 5044 
 

www.etasr.com Dhaked et al.: Battery Charging Optimization of Solar Energy based Telecom Sites in India 

 

• Monitoring modules: The necessary voltage and current 
measurement devices are used in the simulation circuit. 

• Simulation: The DC-DC boost converter is used to boost 
the voltage to battery voltage and the whole circuit acts as a 
current limited source to charge the battery up to the desired 
voltage using the current source. Using the PWM controller 
in the boost converter, the desired charging current can be 
set when charging the battery in bulk charging mode. It 
depends on the profile which is based on the charging and 
discharging time status of the battery voltage, current, solar 
module voltage and power output w.r.t. sunlight. 

Once the battery voltage reaches full charge, it can be 
controlled to enter in constant voltage-based charging. In this 
case, the battery voltage reaches a constant value of 56V and 
the battery current starts decreasing slowly with time. With 
different charging current limits the battery voltage and SOC 
can be plotted with charging current. With different charging 
voltage settings in constant charging voltage mode, the 
charging performance can be optimized [17-19]. With this 
practical simulation model, any permutation and combination 
of charging voltage versus charging current and battery SOC 
can be developed and tested. The PSO algorithm is loaded in 
the PWM controller in the simulation model. Different 
charging and discharging curves are below simulating different 
battery sizes, charging voltages and charging current to confirm 
the system’s functionality. 

 

 

 
Fig. 4.  Discharge characteristics of lead-acid battery: Capacity=100Ah, 
nominal voltage=52V, response time=30s, initial SOC=100% 

The discharge characteristic of lead-acid battery for 100Ah 
capacity, 52V battery voltage and 100% SOC is shown in 
Figure 4. The discharge characteristic of lead-acid battery for 
50Ah capacity, 13.5V battery voltage, and 90% SOC, is shown 
in Figure 5. The discharge characteristic of lead-acid battery for 
capacity of 100Ah, 26.5V battery voltage and 80% SOC is 
shown in Figure 6. 

 

 
Fig. 5.  Discharge characteristics of lead-acid battery: Nominal 

voltage=13.5V, rated capacity=50Ah, initial SOC=90%, battery response 

time=30s 

 

 
Fig. 6.  Discharge characteristics of lead-acid battery: Initial SOC=80%, 
capacity=100Ah, nominal voltage=26.5V, response time=20s 

VI. SIMULATION RESULTS 

The results of the above mentioned simulations are: 

• Battery temperature rise is controlled by maximum battery 
charging current. Current control is done through measuring 
the battery internal temperature by mounting the 
temperature sensor on the negative terminal of the battery. 

• When battery temperature control is done, the cycle life of 
batteries can be determined directly by the exclusive data 
provided by the manufacturer. This is usually at 25

o
C as a 

function of depth of discharge in the cycle and the number 
of cycles. 



Engineering, Technology & Applied Science Research Vol. 9, No. 6, 2019, 5041-5046 5045 
 

www.etasr.com Dhaked et al.: Battery Charging Optimization of Solar Energy based Telecom Sites in India 

 

• The charging profile is matched to the ideal charging 
profile given by the battery manufacturer. Gassing can be 
controlled in simulations and the desired battery 
performance will be achievable when the system's 
operational needs are very close to the ideal conditions. 

• When the battery is charged and discharged using properly 
rated charging power, daily SOC are at the higher sides of 
the SOC scale. Minimum battery SOC after overnight 
discharge is 80% and 100% during charging time at around 
3 PM. This results in highest cycle life as given in the 
Depth of Discharge (DOD) vs cycle life chart provided by 
the battery manufacturer. 

• When the battery capacity selected for required load is as 
mentioned above, batteries are usually sized as per standard 
discharge current and capacity. Usually, all batteries are 
rated as C/20 or C/10. 

• In telecom conditions, more than 25 hours back-up is 
needed for the batteries to operate to C/40 or greater (up to 
C/100) rating. This provides a very high cycle life but it 
requires increasing the low voltage disconnect threshold 
level depending on discharging current rating. 

VII. PRACTICAL RESULTS  

After applying the PSO algorithm, it has been observed that 
the daily SOC battery level stayed from 80% to 100% and the 
voltage level ranged between 65% and 85%, hence the cycle 
life is certainly expected to increase. Figure 7 shows the SOC 
variation before applying PSO MPPT algorithm. The SOC 
stayed between 65% and 80% throughout the week.  

 

 
Fig. 7.  SOC variation before applying PSO algorithm. SOC stays between 
65% and 80% 

 
Fig. 8.  SOC variations after applying PSO. SOC stays above 80% 

Battery’s life cycle was around 4000 cycles, as shown in 
Figure 9. Figure 8 shows the SOC variation after applying the 
PSO MPPT algorithm. It can be seen from Figure 9 that SOC 
stayed above 80% throughout the week. The battery cycle life 
became around 5500 cycles (Figure 9). It is observed that 
battery life increased by 30% with the suggested optimization. 

 

 
Fig. 9.  Battery cycle life with respect to Depth of Discharge 

VIII. CONCLUSION 

In this paper, various parameters of charging battery and 
their respective impact on the battery cycle life were studied. 
While carefully considering all of them, an optimized charging 
method was driven which also takes proper sizing of battery 
bank and solar PV array into consideration. A lead-acid battery 
model was programmed in Matlab and its characteristics were 
studied. DC-DC boost converter-based current source and 
constant-voltage source equations were programmed in Matlab 
and their characteristics were studied. Advance battery 
charging method was driven and the performance impact on 
batteries was studied while operating with solar panels. Solar 
PV power system performance for 24x7 operating loads, such 
as telecom loads, is satisfactory when batteries perform for a 
long time and give better revenues to operators with less 
maintenance expenses. Optimum calculations of system sizing 
for solar PV array and battery banks capacity were also 
derived. With the help of the developed algorithm, 30% more 
cycle life was achieved. 

REFERENCES 

[1] Intelligent Energy, The true cost of providing energy to telecom towers 
in India, available at: https://www.gsma.com/membership/wp-

content/uploads/2013/01/true-cost-providing-energy-telecom-towers-
india.pdf, 2012 

[2] R. Dabkeya, D. K. Dhaked, M. Lalwani, “An overview of gasifier and 
its application in Indian context”, 3rd International Conference on 
Emerging Trends in Electrical, Communication and Information 

Technologies, Ananthapuramu, India, December 20-22, 2018 

[3] N. Bashir, H. S. Sardar, M. Nasir, N. U. Hassan, H. A. Khan, “Lifetime 
maximization of lead-acid batteries in small scale UPS and distributed 
generation systems”, 2017 IEEE Manchester PowerTech, Manchester, 

UK, November 5-8, 2017 

[4] F. Chermat, M. Khemliche, A. Badoud, S. Latreche, “Techno-economic 
feasibility study of investigation of renewable energy system for rural 

electrification in southern Algeria”, Engineering, Technology & Applied 
Science Research, Vol. 8, No. 5, pp. 3421–3426, 2018 



Engineering, Technology & Applied Science Research Vol. 9, No. 6, 2019, 5041-5046 5046 
 

www.etasr.com Dhaked et al.: Battery Charging Optimization of Solar Energy based Telecom Sites in India 

 

[5] M. Garba, M. A. Tankari, G. Lefebvre, “Using of distributed energy 
resources for microgrid resilience achieving”, IEEE 6th International 

Conference on Renewable Energy Research and Applications, San 
Diego, USA, December 5-8, 2017 

[6] S. Chen, C. Kang, Z. Zhang, H. Zhu, “A method for SOC estimation for 
lead-acid battery based on multi-model adaptive extended Kalman 
filtering estimation”, IECON 42nd Annual Conference of the IEEE 

Industrial Electronics Society, Florence, Italy, October 23-26, 2016 

[7] K. Dhaked, S. Saini, P. Sharma, “Analysis of different converters for 
reduced total harmonic distortion and improved power factor”, IEEE 

Fifth International Conference on Signal Processing and Integrated 
Networks, Noida, India, February 22-23,2018 

[8] C. C. Freitas, M. B. Ketzer, M. R. A. Morais, A. M. N. Lima, “Lifetime 
estimation technique for lead-acid batteries”, IECON 42nd Annual 

Conference of the IEEE Industrial Electronics Society, Florence, Italy, 
October 23-26, 2016 

[9] J. F. A. Leao, L. V. Hartmann, M. B. R. Correa, A. M. N. Lima, “Lead-
acid battery modeling and state of charge monitoring”, 2010 Twenty-
Fifth Annual IEEE Applied Power Electronics Conference and 

Exposition, Palm Springs, USA, February 21-25, 2010 

[10] Z. M. Salameh, M. A. Casacca, W. A. Lynch, “A mathematical model 
for lead-acid batteries”, IEEE Transactions on Energy Conversion, Vol. 

7, No. 1, pp. 93-98, 1992 

[11] M. Y. Chen, M. F. Tsai, C. S. Tseng, S. Y. Yen, “A novel control 
strategy for multiphase battery chargers without Hall-effect current 

sensors”, International Power Electronics Conference, Sapporo, Japan, 
June 21-24, 2010 

[12] P. K. Manoharan, R. Mohan Kumar, K. Karthick, R. Sowmya, “SoC 
estimation and Mmnitoring of Li-ion cell using Kalman-rilter 
Algorithm”, International Journal on Electrical Engineering and 

Informatics, Vol. 6, No. 4, pp. 418-427, 2018 

[13] R. A. Jackey, “A simple, effective lead-acid battery modelling process 
for electrical system component selection”, SAE World Congress & 

Exhibition, April 2007 

[14] J. Jana, H. Samanta, K. D. Bhattacharya, H. Saha, “Design and 
development of high efficiency five stage battery charge controller with 
improved MPPT performance for solar PV systems”, International 

Journal of Renewable Energy Research, Vol. 8, No. 2, pp. 941-953, 
2018 

[15] A. Hassoune, M. Khafallah, A. Mesbahi, T. Bouragba, “Power 
management procedures of electric vehicle charging station based grid 
tied PV-battery system”, International Journal of Renewable Energy 

Research, Vol. 8, No. 2, pp. 851-860, 2018 

[16] M. M. Hoque, M. A. Hannan, A. Mohamed, “Charging and discharging 
model of lithium-ion battery for charge equalization control using 

particle swarm optimization algorithm”, Journal of Renewable and 
Sustainable Energy, Vol. 8, No. 6, Article ID 065701, 2016 

[17] M. S. Hossain Lipu, M. A. Hannan, A. Hussain, M. M. Hoque, P. J. Ker, 
M. H. M. Saad, A. Ayob, “A review of state of health and remaining 
useful life estimation methods for lithium-ion battery in electric 

vehicles: Challenges and recommendations”, Journal of Cleaner 
Production, Vol. 205, No. 1, pp. 115-133, 2018 

[18] U. B. Tayab, M. A. Al Humayun, “Modeling and analysis of a cascaded 
battery-boost multilevel inverter using different switching angle 
arrangement techniques”, Engineering, Technology & Applied Science 

Research, Vol. 7, No. 2, pp. 1450-1454, 2017 

[19] Y. Gopal, K. Kumar, D. Birla, M. Lalwani, “Banes and boons of perturb 
& observe, incremental conductance and modified Regula Falsi methods 

for sustainable PV energy generation“, Journal of Power Technologies, 
Vol. 97, No. 1, pp. 35-43, 2017