Artificial Intelligence Based Fast Charging Method for Battery Management Systems

Electric vehicle batteries are one of the most crucial components eco-friendly transportations. The main goals in their development include achieving high energy density, a long lifespan, and fast charging capabilities. Different charging methods are available for these batteries, with the CC-CV method being the most prominent. In this method, the charging procedure adheres to the safe current and voltage limits set by the manufacturer of the cell. However, in fast charging methods, the charging is conducted by exceeding the recommended safe current. Nevertheless, it is widely recognized that these fast-charging methods can be detrimental to battery health. This study aims to create a fast-charging method that reduces damage to li-ion cells in an electric vehicle battery pack during fast charging. The proposed method will dynamically determine the fast charging current in real time by obtaining current, voltage, and temperature data from the BMS. The determined charging current value will be shared with the vehicle’s charging device via the BMS. In the targeted method, the lithium-ion binding capacity, which is directly related with charging, will be estimated by determining the anode potential of the cell using an AI algorithm. Based on this estimation, a dynamic charging method will be developed, where the charging current is adjusted via a PID controller. In this study, fast and safe charging was aimed by providing at least 50% improvement compared to the standard charging method and the charging time, which was 4.5 h, was theoretically reduced to 1.5 h.