Battery’s Cell Performance with Comparative Analysis of DeepTCN, TL, LSTM and GRU using Synthetic Dataset

This study deals with the application of various machine learning (ML) techniques for the estimation State of Charge (SoC) and State of Health (SoH) in battery management systems. The role of SoC and SoH is vital in the estimation of optimizing battery efficiency, the study evaluates methods such as Long Short-Term Memory (LSTM), Gated Recurrent Unit (GRU), Deep Temporal Convolutional Network (DeepTCN), and Transfer Learning (TL). Performance metrics like RMSE, MAE, and $R^{2}$ score are used which highlights the strengths and weaknesses of each model. The overall findings give conclusion on the effectiveness of these ML techniques in accurately predicting SoC and SoH of the battery system, offering betterment for enhancing battery management and its performance. Moreover, a comparative analysis is presented based on the results under certain parameters to give the most suitable and appropriate method for Soc and SoH prediction. Further, it has also been witnessed that under the taken circumstance DeepTCN is outperforming the other methods taken into consideration for comparative analysis.