基于粒子群算法的电池模型参数辨识

Parameters Identification of Battery Model Based on Particle Swarm Optimization

为实现动力电池精准的SOC估计,对磷酸铁锂动力电池建立二阶等效电路模型,考虑到汽车运行工况的复杂性,针对电池运行过程中的动静态特性,利用Arbin测试柜获取电池HPPC测试实验数据,采取粒子群算法计算电池模型参数全局最优解。另外,文章将HPPC离线电压数据与电池模型输出电压数据对比,最大误差为0.067V,能有效地反映出电池的动静态特性,为电池状态精确估计奠定了基础。

In order to achieve accurate estimation of power ithium iron phosphate battery state of charge （SOC）, a second-order equivalent circuit model for the traction lithium-ion battery of electric vehicle is established. Considering the complexity of vehicle operating conditions and dynamic and static characteristics of batteries during run time, experimental results of HPPC test for batteries are obtained by using Arbin testing cabinet and global optimum of model parameters for batteries has been achieved. Besides, we compare HPPC off-line voltage data with output voltage of battery model and find that maximal error is 0.067V, which means it could represent dynamic and static characteristics of batteries, thus laying the foundations for accurate estimation of state of lithium-ion.