Sistem Cerdas untuk Menganalisis Parameter Pengisian Cepat Baterai Kendaraan Listrik menggunakan Raspberry PI 4

Advances in electric vehicle (EV) technology drive the need for fast and efficient battery charging. However, fast charging can cause problems such as overheating, cell degradation, and decreased battery performance. This research develops a Long Short-Term Memory Recurrent Neural Network (LSTM-RNN)-based intelligent system running on Raspberry Pi to monitor and predict battery charging parameters in real-time. The system processes data from voltage, current, power, temperature, and State of Charge (SOC) sensors to detect critical conditions such as overcharging and overheating. Equipped with a Human-Machine Interface (HMI) for live data visualization, the system is able to predict SOC with high accuracy (MAE 1.97%, RMSE 2.84%) and respond to automatic control in less than 2 seconds. This integration improves the efficiency and safety of EV battery fast charging.