Hyperband Optimization of Stacked Bidirectional Long Short-Term Memory Neural Network for the State of Charge Estimation

To attain net-zero carbon emissions, there are two significant challenges that need to be addressed: electrification of vehicles and increasing the utilization of renewable energy sources. Both these challenges require the use of battery energy storage systems, which not only entails efficient and optimal management but also safe usage. Thus, having accurate estimations of the State Of Charge (SOC) is crucial. Machine learning models for the estimation of SOC are becoming popular as they can effectively capture trends and learn complex relationships between different variables from large battery datasets. The paper proposes the use of hyperband optimization on a stacked, bidirectional LSTM model to achieve high accuracies for estimating SOC. The approach is validated using two public battery datasets.