基于高斯过程回归模型的锂电池数据处理
发布时间：2019-04-22 08:31
【摘要】：锂离子电池(简称锂电池)是一种绿色的高能充电电池,它具有放电电压稳定性高、工作时对外界温度限制宽、循环使用寿命长、所占空间体积小、电池质量轻、对环境无危害等优点,因此锂电池被广泛应用于电子产品、电动汽车、航空航天等领域。但是锂电池在工作过程中性能会逐渐衰减,有时候还可能发生意外的故障会导致锂电池失效而产生严重后果。因此对锂电池的健康状态监测和剩余循环寿命预测是至关重要的,这方面的研究对进一步指导锂电池的运行和维护,对于系统的安全具有重要意义。本文的主要工作有:(1)选择基于数据驱动的方法建立高斯过程回归模型,对锂电池的电压、电池容量的数据进行处理。对锂电池的健康状态和剩余使用寿命进行了预测和分析。同时,把预测结果和人工神经网络的方法进行对比,分析选择高斯过程回归的优越性。(2)研究了建立高斯过程回归模型中核函数选择的问题。对不同的核函数进行分析、比较和组合,将核函数分为局部核函数和全局核函数,针对电池数据的特点选择了最佳的核函数或组合核函数,提高了预测结果的精确度。(3)把稀疏高斯过程回归运用到锂电池的数据处理中。选取了稀疏伪输入法进行建模,在保证精确度的前提下,有效地减少了建模的计算量,提高了用高斯过程回归模型处理电池数据的实时性。
[Abstract]:Lithium-ion battery (Li-ion battery) is a kind of green high-energy rechargeable battery. It has high stability of discharge voltage, wide limit to external temperature, long cycle life, small volume of space and light quality of battery. Lithium battery is widely used in electronic products, electric vehicles, aerospace and other fields. However, the performance of lithium batteries will gradually decline in the process of operation, and sometimes unexpected failures may lead to the failure of lithium batteries, resulting in serious consequences. Therefore, it is very important to monitor the health status and predict the residual cycle life of lithium batteries. The research on this aspect is of great significance for further guiding the operation and maintenance of lithium batteries and for the safety of the system. The main work of this paper is as follows: (1) the Gao Si process regression model is established based on the data-driven method, and the data of lithium battery voltage and battery capacity are processed. The health status and residual service life of lithium battery were predicted and analyzed. At the same time, comparing the prediction results with the method of artificial neural network, the superiority of selecting Gao Si process regression is analyzed. (2) the problem of kernel function selection in establishing Gao Si process regression model is studied. Different kernel functions are analyzed, compared and combined. The kernel function is divided into local kernel function and global kernel function, and the best kernel function or combined kernel function is selected according to the characteristics of battery data. The accuracy of the prediction results is improved. (3) the sparse Gao Si process regression is applied to the data processing of lithium batteries. The sparse pseudo-input method is selected to model, which can effectively reduce the calculation amount of modeling and improve the real-time processing of battery data with Gao Si process regression model under the premise of ensuring the accuracy of modeling.
【学位授予单位】：北京交通大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TM912
本文编号：2462681
[Abstract]:Lithium-ion battery (Li-ion battery) is a kind of green high-energy rechargeable battery. It has high stability of discharge voltage, wide limit to external temperature, long cycle life, small volume of space and light quality of battery. Lithium battery is widely used in electronic products, electric vehicles, aerospace and other fields. However, the performance of lithium batteries will gradually decline in the process of operation, and sometimes unexpected failures may lead to the failure of lithium batteries, resulting in serious consequences. Therefore, it is very important to monitor the health status and predict the residual cycle life of lithium batteries. The research on this aspect is of great significance for further guiding the operation and maintenance of lithium batteries and for the safety of the system. The main work of this paper is as follows: (1) the Gao Si process regression model is established based on the data-driven method, and the data of lithium battery voltage and battery capacity are processed. The health status and residual service life of lithium battery were predicted and analyzed. At the same time, comparing the prediction results with the method of artificial neural network, the superiority of selecting Gao Si process regression is analyzed. (2) the problem of kernel function selection in establishing Gao Si process regression model is studied. Different kernel functions are analyzed, compared and combined. The kernel function is divided into local kernel function and global kernel function, and the best kernel function or combined kernel function is selected according to the characteristics of battery data. The accuracy of the prediction results is improved. (3) the sparse Gao Si process regression is applied to the data processing of lithium batteries. The sparse pseudo-input method is selected to model, which can effectively reduce the calculation amount of modeling and improve the real-time processing of battery data with Gao Si process regression model under the premise of ensuring the accuracy of modeling.
【学位授予单位】：北京交通大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TM912
【参考文献】
相关期刊论文 前9条
1 彭宇;刘大同;;数据驱动故障预测和健康管理综述[J];仪器仪表学报;2014年03期
2 魏克新;陈峭岩;;基于自适应无迹卡尔曼滤波算法的锂离子动力电池状态估计[J];中国电机工程学报;2014年03期
3 庞景月;马云彤;刘大同;彭宇;;锂离子电池剩余寿命间接预测方法[J];中国科技论文;2014年01期
4 何志昆;刘光斌;赵曦晶;王明昊;;高斯过程回归方法综述[J];控制与决策;2013年08期
5 罗广求;段艳丽;罗萍;;卫星用锂离子蓄电池性能分析与在轨管理方法[J];电源技术;2012年07期
6 申倩倩;孙宗海;;基于自适应自然梯度法的在线高斯过程建模[J];计算机应用研究;2011年01期
7 苏国韶;张研;燕柳斌;;隧道围岩变形预报的高斯过程机器学习模型[J];桂林理工大学学报;2010年04期
8 刘开云;刘保国;徐冲;;基于遗传 组合核函数高斯过程回归算法的边坡非线性变形时序分析智能模型[J];岩石力学与工程学报;2009年10期
9 苏国韶;燕柳斌;张小飞;江权;;基坑位移时间序列预测的高斯过程方法[J];广西大学学报(自然科学版);2007年02期
相关博士学位论文 前1条
1 徐冲;分岔隧道设计施工优化与稳定性评价[D];北京交通大学;2011年
，本文编号：2462681
本文链接：https://www.wllwen.com/kejilunwen/dianlidianqilunwen/2462681.html
