基于在线LS-SVM的动力电池SOC估算方法研究

摘要

荷电状态(SOC)是描述动力电池状态的重要参数之一,提高SOC估计的准确性对电动汽车电池管理系统的研究至关重要.提出一种改进的最小二乘支持向量机(LS-SVM),动态地调整模型参数,对电池的开路电压(OCV)进行在线实时估计;通过SOC与OCV的关系确定初值,采用安时积分法估算SOC;并利用OCV的偏差信息对电池SOC进行修正,有效地补偿拟合误差和安时积分法产生的累计误差.仿真实验结果表明,在线LS-SVM算法能准确地逼近实际SOC值,平均绝对误差为1.279 3％.%State of charge (SOC) is one of the important parameters to describe the state of power battery and the accuracy of SOC estimation is crucial for the electric vehicle battery management system.An improved least squares support vector machine (LS-SVM) model was proposed;the model parameters were dynamically adjusted;the open-circuit voltage (OCV) for the battery was online estimated.Ah integration method was applied to estimate SOC based on the initial SOC value determined by the relationship between SOC and OCV.The OCV deviation was used to correct SOC to effectively compensate the fitting errors and the cumulative errors caused by Ah integration method.Simulation results show that the presented algorithm can accurately approximate the actual SOC value with the average absolute error of 1.279 3％.