基于天牛须搜索算法的电池模型参数辨识

Parameter identification of battery model based on beetle antennae search algorithm

为实现高精度的动力电池SOC估计,需要建立准确的电池模型。针对磷酸铁锂电池平台特性,采用二阶RC等效电路建立电池的等效电路模型;用双指数函数拟合OCV与SOC的函数关系;采用双线性变换法将电池的数学模型离散化,得到可辨识的电池数学模型。为保证电池模型的精度,要对电池模型的参数进行辨识。首次提出一种基于天牛须搜索算法的电池模型参数辨识方法。用该方法对建立的电池模型进行辨识后,用Simulink对模型进行了脉冲放电与恒流放电仿真测试,并与离线测量数据进行对比,结果表明模型的最大动态误差不超过30 mV,说明该辨识方法有较高的辨识精度,建立的模型能够准确模拟磷酸铁锂电池的动静态特性。

In order to achieve high precision SOC estimation of power battery,an accurate battery model needed to be established.According to the characteristics of lithium iron phosphate battery platform,the second-order RC equivalent circuit was used to establish the equivalent circuit model of the battery.The function relation between OCV and SOC was fitted by double exponential function.The mathematical model of the battery was discretized by bilinear transformation method and the identifiable mathematical model was obtained.In order to ensure the accuracy of the battery model,the parameters of the battery model should be identified.A battery model parameter identification method based on longicorn whisker search algorithm was proposed for the first time.The established battery model was identified by this method,and the simulation tests of pulse discharge and constant discharge were carried out with Simulink.Compared with the offline measurement data,the results show that the maximum dynamic model error is less than 30 mV.It shows that the identification method has higher recognition accuracy,the model can accurately simulate the dynamic and static properties of lithium iron phosphate batteries.