Abstract:
Aiming at the problems of slow charging, low charging efficiency and great damage to the battery, a lithium battery step–optimized charging strategy based on ELM–Takagi Sugeno (T–S) model is proposed. Firstly, lithium is obtained by the extreme learning machine (ELM). The mathematical model between the optimal charging current of the battery and the characteristic parameters such as temperature, internal resistance and polarization voltage, and the offline training parameters are stored in the STM32 processor. Each time the system collects data, the processor calls the ELM program to calculate the optimal charging current at the current moment. Secondly, the T–S fuzzy model is used to optimize the three different charging modes of constant voltage, constant current and pulse to achieve dynamic optimal charging. The simulation results show that the actual charging current of the lithium battery can track the optimal charging current in real time. The charging time is 20% shorter than the three-stage charging mode, and the charging efficiency is increased by about 25% compared with the CC–CV charging mode.