Enhancement of Charging Speed and Analysis of Battery

Configuration for Electric Vehicles Using Deep Learning

A.Subramaniya Siva 1, P.Vishal 2 and N.Vinothini 3

1
Assistant Professor, Department of EEE, K.Ramakrishnan College of Engineering
Trichy, India
[email protected]
2
Department of EEE, K.Ramakrishnan College of Engineering
Trichy, India
[email protected]
3
Research Scholar, Department of ECE, Thiagarajar College of Engineering, Madurai, India
[email protected]

Abstract. Coupled non-linear PDEs are first-principles models that include all the major
physics that govern the interior states of a lithium-ion battery. While these models can
produce accurate predictions, their high computational cost prevents them from being
employed for online control and monitoring. A reformed model, with a comprehensive
simulation running in milliseconds, may predict the internal states of the battery
without sacrificing precision. This study demonstrates the possibility of using this
updated prototype for real-time control-relevant applications. The rebuilt form is
utilized to calculate optimum rules for battery operation (MPC) to demonstrate that the
estimation price of each optimum controller calculation is very low to finish inside the
sampling period in analytical control. Then, observability experiments are provided to
demonstrate how the model will be applied to state estimation MPC. The moving
horizon estimator (MHE) method is employed because it can directly handle limitations
and nonlinear dynamics. The MHE uses the reformulated model to calculate reasonably
and practically. Because of the redesigned model's capacity to solve practically, a
sophisticated physics-based model could be linked to the battery management system
(BMS) to create and execute well monitoring and controller strategies.

Keywords: Lithium battery  Model predictive control Moving horizon estimator

Battery management systems

1 Introduction

Climate change is a prominent topic, and the world is confronted with enormous
challenges. Petroleum-based fuels dominate the transportation sector, which
contributes significantly to the release of CO2 and other pollutants into the
atmosphere. The electrification of vehicles and machines on land, water, and air is a
developing trend and a way to reduce our environmental effect. [1] Renewable
electric energy stored in batteries is a more environmentally friendly alternative to
petrol and diesel. Since its commercialization in 1991, the lithium-ion battery (LIB)
market has grown fast. [2] However, in order to perform effectively, battery packs in
vehicle applications require a sophisticated control system. The available energy of a
battery, known as the state of charge (SOC), is analogous to the fuel gauge of a car
driven by a combustion engine. Because SOC is not quantifiable, it must be
predicted using some method, most typically a mathematical model. [3] A problem
that affects battery longevity is the deterioration of battery cells, which results in
diminished performance. Electric vehicles (EVs) have rapidly increased in diversity
and use in recent years, and this has led to a boom in the market for this new mode
of transportation. The trend of rapid development is also projected to continue due to
tighter emission regulations and increased investment in the area. Due to their high
energy density & low self-discharge rate, lithium-ion power battery packs can
operate for longer periods and cover more miles per charge. [4]
 Due to that, lithium-ion batteries were frequently employed as energy storage &
power sources for electric cars. [5] The performance deterioration of lithium-ion
batteries, which is characterized by capacity fading or power attenuation, is one of
the issues they still face. Lithium-ion batteries must have a precise SOH. [6] The
emergence of big data technologies and the information age present potential for the
realization of unified battery pack health monitoring. Big data SOH monitoring
enhances public transportation's safety oversight and makes it easier to dispose of
EV batteries that have reached the end of their useful lives or have been sitting idle
for a while. The usage of minor use, vehicle-to-grid technologies, and the
assumption that every battery pack is accurately standardized for its SOH and
remaining value. (V2G) [7] Models of electrochemistry, analogous circuits, etc.
Many machine[8] learning techniques have been proposed and established among
the data-driven procedures. [9] For instance, combined infrared thermography,
ANN, and SVM approach to solving the problem of cycle life forecasting. To
investigate the worldwide degradation trend of [10] SOH, Qin et al. used SVM, and
subdivision swarm optimization was used to determine the kernel parameter (PSO).
Based on RF regression, Li et al. calculated the SOH of several batteries under
various cycle situations.

2 Existing Methodology

In the existing system Slope defined capacity trend analysis (SDCTA)

[11,12] has been implemented and the duration of the constant current charging phase
of the battery has a very direct connection to capacity and SOH since it stands for
most of the energy charged into the battery. Because the battery is rarely or never
completely discharged this duration is problematic to measure. [13,14] With
inspiration from the regional capacity method where the minimal slope of the voltage
curve is located, this novel method locates this location and measures the time (or
capacity) needed to reach cut-off voltage as shown below. The method works within
the mid ranges of SOC and can be implemented either through the polynomial fitting
or numerical differentiation. [15]
The decrease in capacity occurring through battery aging affects the shape
of the polarization during the CC-phase curve and increases with aging as the CC-
time period shortens. [16,17] The derivative of the voltage curve in the time interval
just before the cut-off voltage is reached, as seen in Figure 3.14, is easily
differentiated and can be used as an aging factor of the battery. [18] The slope can
also be calculated using the linear least square method on a population of data just
before the end of the CC phase. The key to this method being consistent is
measuring the slope at the same part of the voltage curve as the CC-phase shortens
with use and time. [19] The slope at a fixed SOC would be ideal but just like SOH,
SOC has to be estimated and therefore is somewhat inaccurate. By choosing the
interval just before the cut-off voltage is reached, where the voltage curve is close to
linear and easily found,[20] the accuracy of the method is increased. The SOC at the
cut-off voltage at room temperature is around 70%, a level that is commonly
circulated during operation.[21]

Fig. 1. Block Diagram of Existing Methodology

 Lithium Ion batteries play a vital in power systems & electric vehicles. They
can be used in mobile phones & portable equipment. The disadvantage is that it’s
having short lifespan compared to other batteries. [22] Therefore, it is crucial to
estimate SOH accurately. Three sections make up the x-Boost algorithm as it stands
today. Markov correction, feature selection, and XGBoost estimation. To characterize
the discharge process of the lithium-ion battery, we first employ voltage changes,
temperature deviation, and average voltage as XG Boost model inputs. XG Boost
estimation model is then developed to forecast the battery's SOH. To further increase
the accuracy of the SOH estimate, the Markov chain technique is to find the estimated
findings.[23,24,25]

3 Proposed Methodology
The battery SOH is calculated using long-term driving data. In the beginning, the
estimation model's input variables—the health features—are defined. The training data is
then chosen. The example dataset offers data for model development. The large data
platform's SOH estimation framework is then developed using the offline trained model.

Fig. 2. Block diagram of the Proposed Methodology

The Neural network is a model whose limitations don't directly correspond to

the novel model's physical or chemical structures. The two main types of NN architecture
design are feedforward and recurrent. They used a technique for predicting time series.
Yang et al. employed a back propagation neural network to determine the battery's SoH
using maximum accessible capacity. The limitations of the first-order ECM were found
using a straightforward parameter extraction method.
Then, a 3-coating back circulation neural network with the first-order ECM
parameters as inputs and the current value of SoH as output was suggested to estimate
SoH. According to the results of the research, SoH reduces as ohmic resistance rises, and
when SoC is between 20 and 90 percent, ohmic resistance increases and SoC decreases.
The simplicity of an artificial neural network (ANN) is well recognized. Since it can
handle nonlinear data, it is not required to include all of the battery's specifics in the
modeling process. Three distinct scenarios with various amounts of noise were used to
solve the estimating problem.
 Fig .3. Flow diagram of FFFN

4 Simulation Results & Characteristics

An adaptive system also called as an artificial neural network, learns by using
interconnected nodes or neurons in a layered structure that resembles a human brain. It
may be trained to analyze patterns classify the data, & predict the future occurrence from
the learned data.

Fig 4. Neural Network Training

Deep learning has gained popularity because it does away with the requirement
to extract features from images, a problem that machine learning applications for image
and signal processing had in the past. Even though feature removal can be skipped in
applications for image dispensation, it is frequently used in signal-processing activities to
increase model accuracy.
 Fig. 5. Voltage, Current, and Temperature Characteristics

Fig. 6. Battery Performance

The real capability is utilized to calculate the battery percent's SOH due to
the fact it's miles the maximum standard health indicator. because of the exquisite
efficiency of business lithium-ion batteries, the distinction between charging potential
and discharging capacity may be omitted even though the percent ability decided by
discharging capacity is not unusual. additionally, the usage of charging ability is fine for
the accuracy of the current integration technique for the reason that charging modern-
day profiles have discernible consistency in real operation. with the purpose of deciding
the capability requirements, this article uses fee statistics. in line with the records
accumulated, the impacts of using patterns and power management strategies bring
about the charging fragment never masking the entire SOC range. data from partial
charging can be used for SOH estimation to discover the reliable charging capacity.

We preprocess records with the aid of disposing of outliers and at ease accessible
facts earlier than the use of the battery information set as experimental facts. 4 units of
battery data with degradation characteristics each cycle is in the long run obtained. #5,
#6, #7, and #18 batteries.
 Fig. 7. Neural Network Training Fig 8. Neural Network Training Table

Fig 9. Battery Capacity Estimation Fig 10. Error of Neural Networking

The general battery specifications as well as the charging and discharging conditions are
outlined. However, due to the sensitivity of the data and the complexity of estimating, it
is not efficient to use all of the data even though there are numerous data points during
the charging process according to BMS settings

5 Conclusion

The SOH and battery aging are estimated in this work. The output of the proposed
GRNN-based SOH and the inputs are defined as the time of constant current during the
charging process, the output energy of the battery under a specified depth of discharge,
and the drop in voltage during the discharge. The estimated effect will be better in the
future when more battery charge/discharge characteristics are utilized in the model.
Users can find out from a battery's state of charge how long it will last before needing to
be recharged or replaced. Tracking the SoC is quite important, especially with EVs. A
novel estimation technique that combines coulomb counting, fuzzy logic, and neural
networks has been developed to combine the quality features included in them. The
precision of the State of Charge was improved by this MATLAB simulation approach.
 References

[1] Berecibar, M.; Gandiaga, I.; Villarreal, I.; Omar, N.; Mierlo, J.V.; Bossche, P. Critical review of
state of health estimation methods of Li-ion batteries for real applications. Renew. Sustain.
Energy Rev. 2016, 56, 572–587.
[2] A.Subramaniya Siva, P.Parthasarathy, K.Gijipriya, N.Vinothini, Albert Alexander Stonier,
“Anticipation of Battery Condition Using Internet of Things” IOP Conf. Series: Materials Science
and Engineering, doi:10.1088/1757-899X/1055/1/012157
[3] Andwari, A.M.; Pesiridis, A.; Rajoo, S.; Martinez-Botas, R.; Esfahanian, V. A review of battery
electric. vehicle technology and readiness levels. Renew. Sustain. Energy Rev. 2017
[4] Charging Optimization of Lithium-ion Batteries Based on Capacity Degradation Speed and
Energy Loss”, YanxiangLei, CaipingZhang, YangGao, 2018Charging Optimization of Lithium-
ion Batteries Based on Capacity Degradation Speed and Energy Loss”, YanxiangLei,
CaipingZhang, YangGao, 2018. 64
[5] A.Subramaniya Siva , N.Vinothini , S.Sathieshkumar, “Piezoelectric Based Electric Energy
Generation From Sound Energy” in International Journal of Scientific & Technology Research”
Volume 9, Issue 04, April 2020, ISSN2277-8616
[6] Optimal charging strategies in lithium-ion battery” Reinhardt Klein; Nalin A. Chaturvedi, 2016
[7] “Lithium-Ion Battery State of Charge and Critical Surface Charge Estimation Using an
Electrochemical Model-Based Extended Kalman Filter” , Domenico Di Domenico,Anna
Stephanopoulos, 2019.  Chen Z, Mi C C, Fu Y, et al. “Online Battery State of Health Estimation
Based on GeneticAlgorithm for Electric and Hybrid Vehicle Applications”. Journal of Power
Sources, 2020.
[8] A.Subramaniya Siva , S.Sathieshkumar, T.Santhosh Kumar, “Investigation of Harmonics &
optimal Power Flow in IEEE 14 Bus System Using ETAP Software”, in International Journal of
Scientific & Technology Research”, Volume 9, Issue 04, April 2020, ISSN2277-8616
[9] Yin Hua; Cordoba-Arenas, A.; Warner, N.etc. “A multi time-scale state-of- charge and state-of-
health estimation framework using nonlinear predictive filter for lithium-ion battery pack with
passive balance control”. Journal of PowerSources. 280(2019) : 293–312
[10] N. A. Chaturvedi, R. Klein, J. Christensen, J. Ahmed, and A. Kojic,”Algorithms for advanced
battery-management systems: modeling, estimation, and control challenges for Lithuim-Ion
batteries”, IEEE Control Systems Magazine, 30, pp. 49–68, 2010
[11] Wang, Yujie; Zhang, Chenbin Chen, Zonghai. “A method for state-of-charge estimation of Li-ion
batteries based on multi-model switching strategy”. APPLIED ENERGY . 137(2015): 427–434.
 M. Einhorn, F. V. Conte, C. Kral, and J. Fleig, “A method for online capacity estimation of
Lithium ion battery cells using the state.
[12] A. Subramaniya Siva, S.Sathieshkumar, T. Santhoshkumar, “Home Appliances Control by
Human Voice & Android Bluetooth using 89c51 Microcontroller” in International Journal of
Innovative Research in Electrical, Electronics Instrumentation and Control Engineering” Vol. 7,
Issue 5, May 2019, PP:72-76
[13] A. S. Siva, S. Sathieshkumar and T. Santhosh Kumar, "Analysis of Stability in IEEE 14 Bus
System using ETAP Software," 2020 Fourth International Conference on Inventive Systems and
Control (ICISC), Coimbatore, India, 2020, pp. 935-938, doi: 10.1109/ICISC47916.2020.9171115
[14] K. Dhayalini and A. S. Siva, "Flower pollination algorithm for economic dispatch problems with
non-smooth cost function," 2019 2nd International Conference on Power and Embedded Drive
Control (ICPEDC), Chennai, India, 2019, pp. 254-257, doi:
10.1109/ICPEDC47771.2019.9036499
[15] A. S. Siva, C. Balaji, V. Akileshwaran and N. Hemanathan, "Experimental Analysis on the
Combination of Jatropha Oil and Silk Cotton Oil for Transformer," 2019 Fifth International
Conference on Science Technology Engineering and Mathematics (ICONSTEM), Chennai India,
2019, pp.343-347,doi:10.1109/ICONSTEM.2019.8918903
[16] A. Subramaniya Siva, C.Balaji, V.Akileshwaran, N.Hemanathan ,” Experimental Analysis on the
Combination of Jatropha oil & Silk cotton oil for Transformer” in IEEE Xplore Digital Library,
DOI: 10.1109/ICONSTEM.2019.8918903 on 5 Dec 2019
[17] Archana R, A.Subramaniya Siva “A Perspective Analysis for the Impact of PV and Wind Hybrid
Distributed Generation Using ETAP”,International Journal of Advanced Research in Electrical,
Electronics and Instrumentation Engineering, Volume 7 issue 4-April 2018
[18] A.Subramaniya Siva, M. Bhavani “Mitigation of Harmonics by Shunt Active Power Filter using
Synchronous Detection Method”, International Journal of Engineering Trends & Technology
(IJETT), Volume 4 issue 6-2013
[19] Model-based unscented Kalman filter observer design for lithium-ion battery state of charge
estimation”, Taipeng Wang,Sizhong Chen, Hongbin Ren, 2018
[20] A.Sriram and P.Ramya “Automatic accident notification system using GSM and GPS modems
with 3g technology for video monitoring” International Journal of Emerging Trends in Electrical
and Electronics (IJETEE) Vol. 1, Is-sue. 2, March-2013, pp 11-13
[21] A.Subramaniya Siva, A.Prabhu, A.Arulwilfred “ Information Transmitter using Bluetooth Energy
Beacons” , International Journal of Research & Innovation in Applied Science, Volume 4, Issue
1, January 2019.
[22] A.Prabhu, A. Subramaniya Siva, “Short Circuit and Contingency Analysis Implementation for
IEEE – 14 Bus System using Mi Power Software” in International Journal of Advanced Research
in Computer and Communication Engineering” Vol. 8, Issue 5, May 2019,PP:156-1
[23] A.T. Sankara Subramanian, A. Subramaniya Siva, M. Ramesh, P. Sabarish,“Experimental
analysis on passive thermal management system for electronic gadgets using inorganic phase
changing material “(PCM), Materials Today: Proceedings,2022, ISSN
2214-7853,https://doi.org/10.1016/j.matpr.2022.08.101
[24] A. S. Siva and V. Pavithra, "Power Quality Enhancement in the Integration of Distributed
Generation using Particle Swarm Optimization," 2022 3rd International Conference on
Electronics and Sustainable Communication Systems (ICESC), 2022, pp. 155-160, doi:
10.1109/ICESC54411.2022.9885685.
1. Miguel, L.F.F., Miguel, L.F.F.: Shape and size optimization of truss structures considering
dynamic constraints through modern metaheuristic algorithms. Expert Systems with
Applications. 39, 9458-9467 (2012). doi: 10.1016/j.eswa.2012.02.113
2. Gomes, H.M.: Truss optimization with dynamic constraints using a particle swam
algorithm. Expert Systems with Applications. 38, 957-968 (2011). doi:
10.1016/j.eswa.2010.07.086
3. Kaveh, A., Zolghadr, A.: Democratic PSO for truss layout and size optimization with
frequency constraints. Computer and Structures. 130(3), 10-21 (2014). doi:
10.1016/j.compstruc.2013.09.002
4. Kaveh, A., Zolghadr, A.: A new PSRO algorithm for frequency constraint truss shape and
size optimization. Structural Engineering and Mechanics. 52(3), 445-468 (2014). doi:
10.12989/sem.2014.52.3.445
5. Kaveh, A., Javadi, S.M.: Shape and size optimization of trusses with multiple frequency
constraints using harmony search and ray optimizer for enhancing the particle swarm
optimization algorithm. Acta Mech. 225, 1595-1605 (2014).doi: 10.1007/s00707-013-1006-
z
6. Kaveh, A., Mahdavi, V.R.: Colling-bodies optimization for truss with multiple frequency
constraints. J.Comput. Civ. Eng. 29(5), 04014078-10 (2015). doi: 10.1061/(ASCE)CP.1943-
5487.0000402
7. Kaveh, A., Ghazaan, M.I.: Optimal design of dome truss structures with dynamic frequency
constraints. Struct. Multidisc. Optim. 53, 605-621 (2016). doi: 10.1007/s00158-015-1357-2
8. Tejani, G.G., Savsani, V.J., Patel, V.K.: Adaptive symbiotic organisms search (SOS)
algorithm for structural design optimization. Journal of Computational Design and
Engineering. 3, 226-249 (2016). doi: 10.1016/j.jcde.2016.02.003
9. Baghlani, A., Makiabadi, M.H.: Teaching-learning-based optimization algorithm for shape
and size optimization of truss structures with dynamic frequency constraints. Iran Journal of
Science and Technology. 37, 409-421 (2013). doi: 10.22099/IJSTC.2013.1796
10. Dede, T., Toğan, V.: A teaching learning based optimization for truss structures with
frequency constraints. Structural Engineering and Mechanics. 53, 833-845 (2015). doi:
10.12989/sem.2015.53.4.833
11. Farshchin, M., Camp, C.V., Maniat, M.: Multi-class teaching-learning-based optimization
for truss design with frequency constraints. Engineering Structures. 106, 356-369 (2016).
doi: 10.1016/j.engstruct.2015.10.039
12. Rao, R.V.: Jaya: A simple and new optimization algorithm for solving constrained and
unconstrained optimization problems. Int. J. Ind. Eng. Comput. 7, 19-34 (2016). doi:
10.5267/j.ijiec.2015.8.004
13. Rao, R.V.: Jaya: An Advanced Optimization Algorithm and its Engineering Applications.
Springer. (2019)
14. Dede, T.: Jaya algorithm to solve single objective size optimization problem for steel
grillage structures. Steel and Composite Structures. 26, 163-170 (2018). doi:
10.12989/scs.2018.26.2.163
15. Degertekin, S.O., Lamberti, L., Ugur, I.B.: Sizing, layout and topology design optimization
of truss structures using the Jaya algorithm. Applied Soft Computing. 70, 903-928 (2018).
doi: 10.1016/j.asoc.2017.10.001