Thermal aware fast EV charging and intelligent energy management using CNN-BiLSTM under highly variable renewable energy conditions

Extremely fast charging of electric vehicles (EVs) requires higher-capacity charging stations, which may pose an extra burden over the utility network if the charging station is supported by renewable energy sources having frequent power interruptions. Adaptive Sliding Mode Control (ASMC) offers robustness and consistency; its capability is restricted in the multi-objective functionalities such as battery state of charge estimation, thermal control, and power optimization during highly dynamic operating scenarios. To address these issues, this paper proposes a CNN-BiLSTM-assisted adaptive sliding mode control (ASMC), hereafter referred to as CBA, for renewable energy-supported EV charging stations connected to the grid. The CNN-BiLSTM model is used as a forecasting feedforward module to estimate the charging demand, uncertainty in renewable power, and battery thermal management control, while ASMC is employed for robust closed-loop trajectory tracking and constraint enforcement. This control framework executes rapid charging control, improved temperature management, and appropriate power sharing amongst renewable energy sources and the utility grid. Simulation and experimental results illustrate that the CBA controller achieves a battery state of charge (SoC) of 85% within 900 s compared to 80% in the standard ASMC controller, which is an increase of 6.25%. The experimental validation demonstrates the efficacy of the proposed technique, as 84% SoC is achieved in the same charging time, whereas ASMC achieves only 80%, which is an increase of 5.0%. Furthermore, the peak battery temperature is reduced by 8.33% in simulation and 10.2% in experimental validation using the suggested controller, with a reduction of the peak battery temperature from 48 °C to 44 °C and 49 °C to 44 °C, respectively. The proposed strategy also improves charging efficiency, stabilizes DC bus voltage, enhances renewable energy utilization, and reduces grid stress under highly dynamic operating conditions without sacrificing charging speed and battery health.