Advanced SOC Estimation in EV Charging System with Sparrow Search Optimized LSTM Classifier

Improving accuracy of Battery State of Charge (SOC) estimation is essential for enhancing Electric Vehicles (EVs) performance, safety and energy efficiency. Accurate SoC determination helps manage battery power effectively, prevents overcharging, extends battery’s life span and more reliable range predictions, making EVs more practical and dependable for everyday use. Henceforth, this paper proposes a novel SOC estimation model using Sparrow Search Optimized Long Short-Term Memory (SSO-LSTM) classifier for real-time SOC predictions based on voltage and current data. In this work integrates a PV system with a boost converter and P&O MPPT controller for stepping up voltage level and extract maximum power from the PV source. Additionally, bidirectional converter is incorporate to facilitate both battery charging and discharging with a Proportional Integral (PI) controller ensuring optimal performance during these operation. These integrated system enhance energy efficiency and accurate SOC tracking for the battery. Furthermore, the use of the SSO-LSTM model in SOC estimation validated using Python software and the results demonstrates Root Mean Square Error (RMSE) of 0.2% and Mean Absolute Error (MAE) of 0.1.