基于离散滑模观测器的锂电池荷电状态估计  

Charge State Estimation of Li-ion Batteries Based on Discrete-time Sliding Mode Observers

作　　者：孙冬[1] 陈息坤[1] 

机构地区：[1]上海大学机电工程与自动化学院,上海市闸北区200072

出　　处：《中国电机工程学报》2015年第1期185-191,共7页Proceedings of the CSEE

基　　金：国家863高技术基金项目(2011AA11A247)~~

摘　　要：锂电池的荷电状态(state of charge,SOC)估计是电池管理系统的重要组成部分,针对锂电池非线性的特性,提出了采用离散滑模观测器估计锂电池荷电状态的方法,给出了离散滑模观测器的设计方法及其稳定性证明。基于锂电池的戴维南等效电路模型,给出了该方法的设计过程,在不同的充放电电流倍率和环境温度下,进行了锂电池模型的参数辨识,通过与常用的扩展卡尔曼滤波法相比较,分析了离散滑模观测器对锂电池SOC估计的精度、鲁棒性和算法复杂度等方面的性能。实验结果表明,采用该算法可实现锂电池SOC快速精确地估计,误差可控制在约3%,验证了该方法的可行性。Estimation of the state of charge （SOC） is the key technique in the power management system for a li-ion power battery. For the inherent nonlinear property of the li-ion power battery, a method of SOC estimation is proposed applied to batteries, and the design of the algorithm for battery SOC estimation based on discrete-time sliding mode observers （DSMO） is given and the stability proof of DSMO is proven. Based on the Thevenin equivalent model, the detailed procedures of this estimation method are exhibited,and the model parameters are identified at different current rate and ambient temperature. The accuracy, the robustness and the time complexity of the extended Kalman filter （EKF） and the proposed method are analyzed in this comparative study. Experiments show that the arithmetic of the discrete-time sliding mode observers can be used to compute the battery SOC quickly and accurately with the dynamic error of 3%, and that the feasibility of the proposed algorithm is verified.

关 键 词：锂电池 荷电状态 离散滑模观测器 扩展卡尔曼滤波器 

分 类 号：U463[机械工程—车辆工程]