Optimized Machine Learning Framework for Battery State Prediction in Smart Charging Systems

Year : 2026 | Volume : 04 | Issue : 01 | Page : 11 23
    By

    Rajeshwari Mahantesh Thadi,

  • Sandhya Sharma,

  • Sarang M. Patil,

  • Mukesh Kumar Gupta,

  1. Research Scholar, Department of Electronics Communication & Engineering, Suresh Gyan Vihar University, Jaipur, Rajasthan, India
  2. Research Scholar, Department of Electronics Communication & Engineering, Suresh Gyan Vihar University, Jaipur, Rajasthan, India
  3. Professor, Department of Computer Science & Engineering, Amity School of Engineering & Technology, Mumbai, Maharashtra, India
  4. Professor, Department of Electrical Engineering, Suresh Gyan Vihar University, Jaipur, Rajasthan, India

Abstract

Good estimation of battery states, including State-of-Charge (SoC), State-of-Health (SoH), and Remaining Useful Life (RUL), are important in managing energy wisely and controlling the adaptive charging. This work introduces a streamlined machine learning model based on the ability to use multi-dimensional sensor measurements in terms of voltage, current, temperature, and cycle number to forecast battery conditions with high accuracy. Decent preprocessing, such as noise elimination, feature scaling, and calculated features, were done on the dataset to increase the reliability of the model. There are various models that were trained and compared through MAE, RMSE, and R2 which are XGBoost, LSTM, GRU, CNN+LSTM and Transformer architectures. Transformer-based model was observed to be better performing with less prediction error and high temporal stability. The findings prove that the suggested structure is able to achieve nonlinear trends in the degradation and this serves as the basis of real time adaptive charging and health-conscious decision-making in battery management systems.

Accurate estimation of battery conditions plays a vital role in improving the safety, dependability, and overall performance of contemporary energy storage technologies, especially in electric vehicles and renewable power systems. This research proposes a lightweight and computationally efficient machine learning framework designed to predict critical battery health parameters, namely State-of-Charge (SoC), State-of-Health (SoH), and Remaining Useful Life (RUL). The methodology utilizes multi-source sensor inputs, including voltage, current, temperature, and cycle count, to effectively model complex degradation behavior. To enhance prediction reliability, rigorous data preprocessing steps such as noise reduction, normalization, and advanced feature extraction are incorporated. Several machine learning and deep learning models—XGBoost, LSTM, GRU, CNN+LSTM, and Transformer—are developed and comparatively assessed using evaluation metrics such as MAE, RMSE, and R². Experimental findings indicate that the Transformer architecture outperforms the other models by achieving lower prediction errors and improved temporal stability. Its ability to learn nonlinear degradation characteristics and long-range dependencies makes it particularly effective for battery prognostics.

Keywords: Optimized Machine Learning, electric vehicles, Remaining Useful Life (RUL), deep learning (DL), State-of-Charge (SoC), State-of-Health (SoH).

[This article belongs to International Journal of Energy and Thermal Applications ]

How to cite this article:
Rajeshwari Mahantesh Thadi, Sandhya Sharma, Sarang M. Patil, Mukesh Kumar Gupta. Optimized Machine Learning Framework for Battery State Prediction in Smart Charging Systems. International Journal of Energy and Thermal Applications. 2026; 04(01):11-23.
How to cite this URL:
Rajeshwari Mahantesh Thadi, Sandhya Sharma, Sarang M. Patil, Mukesh Kumar Gupta. Optimized Machine Learning Framework for Battery State Prediction in Smart Charging Systems. International Journal of Energy and Thermal Applications. 2026; 04(01):11-23. Available from: https://journals.stmjournals.com/ijeta/article=2026/view=238785


References

[1]     M. T. Sarker, M. Al Qwaid, S. J. Shern, and G. Ramasamy, “AI-driven optimization framework for smart EV charging systems integrated with solar PV and BESS in high-density residential environments,” World Electric Vehicle Journal, vol. 16, no. 7, p. 385, 2025, doi: 10.3390/wevj16070385.

[2]     S. J. Shern, M. T. Sarker, M. H. S. M. Haram, G. Ramasamy, S. P. Thiagarajah, and F. Al Farid, “Artificial intelligence optimization for user prediction and efficient energy distribution in electric vehicle smart charging systems,” Energies, vol. 17, p. 5772, 2024.

[3]     X. Li, D. Yu, V. S. Byg, and S. D. Ioan, “The development of machine learning-based remaining useful life prediction for lithium-ion batteries,” Journal of Energy Chemistry, vol. 82, pp. 103–121, 2023, doi: 10.1016/j.jechem.2023.03.026.

[4]     L. Zhao et al., “Title not specified,” Journal of The Electrochemical Society, vol. 170, no. 9, p. 090537, 2023, doi: 10.1149/1945-7111/acf8ff.

[5]     X. Hu, Y. Che, X. Lin, and S. Onori, “Battery health prediction using fusion-based feature selection and machine learning,” IEEE Transactions on Transportation Electrification, vol. 7, pp. 382–393, 2021.

[6]     A. Fouka, A. Bousdekis, K. Lepenioti, and G. Mentzas, “A unified machine learning framework for Li-ion battery state estimation and prediction,” Applied Sciences, vol. 15, no. 15, p. 8164, 2025, doi: 10.3390/app15158164.

[7]     M. Martí-Florences, A. Cecilia, and R. Costa-Castelló, “Modelling and estimation in lithium-ion batteries: A literature review,” Energies, vol. 16, p. 6846, 2023.

[8]     B. Balagopal and M. Y. Chow, “The state-of-the-art approaches to estimate the state of health (SOH) and state of function (SOF) of lithium-ion batteries,” in Proc. IEEE 13th Int. Conf. Industrial Informatics (INDIN), Cambridge, UK, Jul. 2015, pp. 1302–1307.

[9]     D. Shi, J. Zhao, Z. Wang, H. Zhao, J. Wang, Y. Lian, and A. F. Burke, “Spatial-temporal self-attention transformer networks for battery state of charge estimation,” Electronics, vol. 12, p. 2598, 2023.

[10]   C. Vidal, P. Malysz, P. Kollmeyer, and A. Emadi, “Machine learning applied to electrified vehicle battery state of charge and state of health estimation: State-of-the-art,” IEEE Access, vol. 8, pp. 52796–52814, 2020.

[11]   Z. Shi and A. Chehade, “A dual-LSTM framework combining change point detection and remaining useful life prediction,” Reliability Engineering & System Safety, vol. 205, p. 107257, 2021.

[12]   Z. Fei, F. Yang, K. L. Tsui, L. Li, and Z. Zhang, “Early prediction of battery lifetime via a machine learning-based framework,” Energy, vol. 225, p. 120205, 2021.

[13]   V. Selvaraj and I. Vairavasundaram, “A comprehensive review of state of charge estimation in lithium-ion batteries used in electric vehicles,” Journal of Energy Storage, vol. 72, p. 108777, 2023.

[14]   V. S. Naresh, G. Ratnakara Rao, and D. Prabhakar, “Predictive machine learning in optimizing the performance of electric vehicle batteries: Techniques, challenges, and solutions,” Wiley Interdisciplinary Reviews: Data Mining and Knowledge Discovery, vol. 14, 2024.

[15]   K. Karthick, S. Ravivarman, and R. Priyanka, “Optimizing electric vehicle battery life: A machine learning approach for sustainable transportation,” World Electric Vehicle Journal, vol. 15, no. 2, p. 60, 2024, doi: 10.3390/wevj15020060.

Regular Issue Subscription Original Research
Volume 04
Issue 01
Received 10/02/2026
Accepted 28/02/2026
Published 10/04/2026
Publication Time 59 Days
8

Login


My IP

PlumX Metrics