Random Forest Regression Based Temperature Estimation in Lithium-ion Batteries

Abstract

For lithium-ion batteries to operate safely and effectively, accurate temperature estimation is essential. The research presented here demonstrates the viability and efficacy of using machine learning to estimate the temperature in lithium-ion batteries and also improves the understanding of temperature-related dynamics while offering useful insights into battery behavior. The study presents a novel temperature estimation technique based on machine learning which employs the random forest regression algorithm. The state of health (SOH) which affects the thermal mechanism of the battery has been incorporated into the proposed approach. The model predicts the battery temperature using input features including voltage, current, time, and state of health (SOH). Through comprehensive experimentation and analysis, the model’s ability to precisely estimate the temperature of lithium-ion batteries has been demonstrated. The findings highlight the model’s remarkable accuracy and robustness and emphasize its potential for real-world applications in battery temperature control and safety.