Manoj K. Singh
School of Computer Science Engineering and Technology, Bennett University, Greater Noida, Uttar Pradesh, India.
Sangeeta Pant
Symbiosis Institute of Technology, Pune, Symbiosis International (Deemed University) (SIU), Pune, Maharashtra, India.
Shshank Chaube
Symbiosis Institute of Technology, Hyderabad, Symbiosis International (Deemed University), Pune, Maharashtra, India.
. Harsh
Department of Computer Science & Technology, Manav Rachna University, Faridabad, 121010, Haryana, India.
Anuj Kumar
School of Computer Science Engineering & Applications, D. Y. Patil International University (DYPIU), Pune, Maharashtra, India.
Jitendra Pal Singh
Department of Basic Sciences (Physics), Manav Rachna University, 121010, Faridabad, Haryana, India.
DOI https://doi.org/10.33889/PMSL.2026.5.2.018
Abstract
Machine learning is among the most popular techniques that use existing data to learn the mechanism that drives a particular phenomenon, in particular, state-of-health (SOH). Batteries are charged at a certain ambient temperature; however, depending on the battery degradation mechanism, the temperature of the battery could vary. In addition, voltage may also not be constant during a charging process. The features extracted for voltage and temperature in a charging cycle can then be useful in predicting the SOH of a battery in conjunction with a machine-learning model. We used complex networks to form a vector of features that can be useful in predicting SOH with these features as input to long short-term memory neural networks. The comparative analysis between measured SOH and predicted SOH shows that the features derived from complex networks used in this study can be useful in predicting SOH.
Keywords- Visibility graph, LSTM, State-of-health, Battery degradation.
Citation
Singh, M. K., Pant, S., Chaube, S., Harsh, .., Kumar, A., & Singh, J. P. (2026). Complex Network Analysis of Lithium-Ion Battery Life. Prabha Materials Science Letters, (2), 323-337. https://doi.org/10.33889/PMSL.2026.5.2.018.
