Prediction of Residual Energy in Batteries using CNN-BiGRU and Attention Mechanism Model

Maintaining a continuous energy supply and making the most of available energy necessitates accurate forecast of battery residual life (RL). The stability (SoC) of a battery management system is essential for leveraging reliable predictions of the battery’s charge state. Since battery deterioration is not linear, it is difficult to make accurate predictions about the So$\mathrm{C}'$s expected lifespan that are substantially longer than current estimates. The proposed method uses an F-divergence algorithm-based data quality evaluation methodology to investigate data unequal distribution, which improves the accuracy and consistency of the batteries. To avoid over-fitting and make more precise predictions, Principle Component Analysis (PCA) is often used to choose which features should be included in a model. CNN-BiGRU-AM based model is employed for training in the proposed method. When compared to two other models, such as CNN and GRU, the proposed method performs better.