https://doi.org/10.1140/epjp/s13360-026-07295-3
Regular Article
Remaining useful life prediction of Li-ion batteries using fusion-based data-driven GRU-CNN hybrid model
Department of Instrumentation & Control Engineering, Dr B R Ambedkar National Institute of Technology Jalandhar, 144008, Jalandhar, Punjab, India
a
This email address is being protected from spambots. You need JavaScript enabled to view it.
Received:
3
July
2025
Accepted:
3
January
2026
Published online:
13
January
2026
Abstract
The accurate capacity and fast life cycle predictions are important concerns for the safe and reliable operation of power sources. Recently, machine learning (ML) and deep learning algorithms have been employed to predict the remaining useful life (RUL) of batteries. This research work proposes a fusion-based hybrid model, which comprises a gated recurrent unit (GRU) with a convolutional neural network (CNN), i.e., GRU-CNN, to enhance the accuracy of RUL prediction of Li-ion batteries. The key feature of a Li-ion battery, i.e., capacity, is identified using linear, ridge, lasso, gradient boosting, and random forest regression ML models. Therefore, the capacity fading feature is used to predict the remaining life of the Li-ion battery. The moving average, exponential moving average (EMA), and Savitzky–Golay (Savgol) methods are applied for smoothing the data using a sliding window of ten data points. The proposed algorithm is trained on 80% of the entire dataset and validated on the remaining 20%. The effectiveness of the proposed hybrid algorithm is validated using two datasets: NASA and CALCE. The proposed model is compared with other existing algorithms, and statistical errors such as MSE, RMSE, and R2 are analyzed. It ensures the least prediction error, i.e., the highest accuracy of the proposed model among the existing algorithms for RUL prediction.
Copyright comment Springer Nature or its licensor (e.g. a society or other partner) holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law.
© The Author(s), under exclusive licence to Società Italiana di Fisica and Springer-Verlag GmbH Germany, part of Springer Nature 2026
Springer Nature or its licensor (e.g. a society or other partner) holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law.
