基于遗传算法的锂离子电池等效电路参数识别

Parameter Identification of Lithium-Ion Battery Equivalent-Circuit Model Based on Genetic Algorithm

摘要:

结合锂离子电池双极性等效电路模型提出了一种基于遗传算法的参数识别方法,该方法通过指数函数对电路模型中的电阻、电容、恒压源等元件进行有理逼近,根据电池在不同充放电速率下的输出电压特性数据,通过实数编码遗传算法得到最优的函数参数,从而得到最优的电阻、电容,开路电压等电路参数值,针对电池在不同的工作状态,不同的工作参数下的运行数据,系列仿真和实验结果表明该算法原理简明,收敛较快,辨识得到的最优模型其电压输出特性与电池的实际电压输出特性基本吻合,能较精确的反映电池的实际特性,具有较高的辨识精度。

Abstract:

On account of lithium-ion battery dual polarization (DP) equivalent circuit model a genetic algorithm is proposed for parameter identification. The method makes resistance, capacitance and open circuit voltage source in the circuit model as exponential function based rational approximant. According to the characteristic data of output voltage in different charge and discharge rate, A real code genetic Algorithm (GA) is employed to estimate the battery model parameters including the resistance, capacitance and open circuit voltage. The algorithm is applied to several cases in different working conditions and working parameters. The simulation and experimental results demonstrate that the genetic algorithm to identify the parameters has concise principle and quickly convergence. The equivalent circuit model with the optimally extracted parameters can accurately predict the performance of the lithium-ion battery and has high identification precision.