基于数据驱动的锂电池随机动态系统建模  

Modeling of Lithium-Ion Battery Stochastic Dynamic System based on Data-Driven

作　　者：刘晓程[1] 王建明[1] 王武[1] 

机构地区：[1]福州大学电气工程与自动化学院,福州350108

出　　处：《电气技术》2015年第5期17-21,共5页Electrical Engineering

基　　金：福建省自然科学基金资助项目(2012J01257);福州大学科技发展基金资助项目(2012-XY-3)

摘　　要：多数传统的锂电池的模型都是建立在理论简化的机理模型之上。但实际上对于锂电池,由于难于测量内部复杂的电化学反应过程并且易受外界环境影响,所以建立的理论模型存在一定的偏差,无法准确反应锂电池动态特性。针对这一问题,本文根据数据驱动的思想,采用一种基于EM算法的随机动态模型建模方式,提出锂电池放电过程的时间序列的随机动态模型。实验结果表明,利用本文所提算法建立的模型能够有效的契合实验数据,具有良好的稳定性和鲁棒性。Most of the conventional batteries model for lithium-ion batteries dependent on theoretical and simplified mechanism model. Actually for lithium-ion batteries, because of unable to measure the process of the internal complex electrochemical reaction and vulnerable to the impact of external environment, the error is exist by theoretical model, and can not accurately reflect the dynamic characteristics of lithium-ion batteries. To solve this problem, according to ideal of data-driven, this paper uses a stochastic dynamic modeling method based on the EM algorithm and a lithium-ion battery discharge time series of stochastic dynamic model is proposed. Experimental results show that the use of this model to establish the proposed algorithm can effectively fit the experimental data, with good stability and robustness.

关 键 词：数据驱动 锂电池 时间序列 EM算法 鲁棒性 

分 类 号：TM912[电气工程—电力电子与电力传动]