Optimizing electric bike battery management: Machine learning predictions of LiFePO4 temperature under varied conditions

The LiFePO4 battery temperature in electric bikes significantly affects their lifespan and performance. This study employed four machine learning models (Random Forest (RF), Long short-term memory (LSTM), Multilayer perceptron (MLP), and support vector machine (SVM)) to predict temperature at different speeds, with and without passengers, and during day/night charging. Random Forest outperformed the others, boasting the highest R-squared (0.9228) and lowest root mean squared error (RMSE) of 0.9023, signifying exceptional accuracy. The result shows that the temperature increases with passenger load and daytime charging, whereas the front side consistently showed higher temperatures due to external factors. These findings pave the way for optimized electric bike battery management and improved efficiency. Random Forest emerges as a valuable tool for accurate temperature prediction and optimal battery performance.