混合动力轿车用锂离子电池峰值功率预测模型研究

Research on Lithium Ion Battery Peak Power Prediction Model for Hybrid Electric Vehicles

混合动力汽车工况复杂,需对电池频繁充放电,锂电池抗滥用能力差,有必要研究电池功率特性,使其充分合理的使用。搭建电池测试系统,采用改进HPPC方法测试锰酸锂电池10 s脉冲峰值功率。基于ANFIS建立充放电峰值功率预测模型,输入量为可直接测量的温度、SOC、欧姆内阻和充放电状态,输出为10 s脉冲峰值功率。采用减法聚类分析算法产生模糊结构,用BP算法估计前提参数,用最小二乘法辨识结论参数。通过测试数据对模型进行验证,基于ANFIS的模型能够很好地估计电池的脉冲峰值功率。

The working conditions of hybrid electric vehicles are complicated, requiring frequent charging and discharging of batteries. Since the resistance to abuse of lithium batteries is poor, it is necessary to study the power characteristics of batteries so that they can be used rationally and fully. This paper has built the battery testing system and has adopted the improved HPPC method to test the pulse peak power of lithium manganate battery at 10 s. Based on ANFIS,the charging and discharging peak power prediction model has been built,and the input is the temperature,SOC,Ohm resistance and charging and discharging condition that can be measured directly,while the output is the pulse peak power at 10 s. This paper has adopted the subtraction clustering analysis algorithm to produce vague structure, to estimate premise parameter with BP algorithm and to identify conclusion parameters with the least square method. Through validating the model by testing data, the model based on ANFIS can be used to well estimate the pulse peak power of the battery.