Machine Learning approach for Prediction of residual energy in batteries

The ability to predict battery residual life (RL) in advance is critical for ensuring a reliable supply of energy and the most efficient use of that energy. When it comes to precisely predicting the level of charge of batteries, battery management systems must be durable and trustworthy (SoC). Because of the non-linear nature of battery depreciation, it is extremely difficult to predict SoC estimation with considerably less degradation than is now possible. In this paper, we tend to reduce the data degradation for the prediction of RL using ensemble random forest model. The model enables collection of data, pre-processing and classification using random forest and ensemble random forest for the prediction of RL. The simulation is conducted in terms of R² and root mean square error (RMSE). The simulation shows that the ensemble random forest model achieves higher prediction accuracy.