首页|基于时空建模的锂离子电池温度预测

基于时空建模的锂离子电池温度预测

扫码查看
原文链接
  • 国家科技期刊平台
  • NETL
  • NSTL
  • 万方数据
锂离子电池温度具有时空耦合、强非线性和时变特性,建立准确的预测模型有困难.提出一种基于时空建模的锂离子电池温度分布预测方法.利用正交局部保持投影(OLPP)将电池温度分离为正交空间基函数和时间系数.以电流、电压为输入,时间系数为输出,建立基于带遗忘因子的在线顺序超限学习机(FFOS-ELM)的低阶时序模型.通过时空合成,重构出原始的温度分布.三元软包装锂离子电池温度预测结果表明,与基于拉普拉斯特征映射和在线顺序极限学习机的在线时空建模方法(LE-OS-ELM)相比,所提方法的预测精度更高,在恒流放电和城市动力测驾循环(UDDS)工况下,时间标准绝对误差分别在(0.030,0.155)和(0.095,0.110)区间内,均方根误差分别为0.0972及0.1084.
Temperature prediction of Li-ion battery based on spatiotemporal modeling
The temperature of Li-ion battery has the characteristics of spatiotemporal coupling,strong nonlinear and time-varying,it is difficult to establish an accurate prediction model.A temperature distribution prediction method for Li-ion battery based on spatiotemporal modeling is proposed.The battery temperature is separated into orthogonal space basis functions and time coefficients by orthogonal locality preserving projections (OLPP).With the current and voltage as input,time coefficient as output,a low-order time sequence model based on online sequential extreme learning machine with forgetting factors (FFOS-ELM) is established.The original temperature distribution is reconstructed according to the spatiotemporal synthesis.Results of the temperature prediction of a ternary pouch Li-ion battery show that compared with the online space-time modeling method based on Laplacian eigenmaps and online sequential extreme learning machine (LE-OS-ELM),the proposed method has higher prediction accuracy.Under galvanostatic discharge and urban dynamometer driving schedule (UDDS) conditions,the temporal normalized absolute errors are within the range of (0.030,0.155) and (0.095,0.110),the root-mean-square errors are 0.0972 and 0.1084,respectively.

Li-ion battery temperatureonline spatiotemporal modelingorthogonal locality preserving projection (OLPP)online sequential extreme learning machine with forgetting factor(FFOS-ELM)

吕洲、何波、宋连

展开 >

广州国科电技术有限公司,广东广州 511458

武汉理工大学重庆研究院,重庆 401120

锂离子电池温度 在线时空建模 正交局部保持投影(OLPP) 带遗忘因子的在线顺序超限学习机(FFOS-ELM)

国家自然科学基金面上项目重庆市自然科学基金

523733062023NSCQ-MSX2249

2024

10.19535/j.1001-1579.2024.04.012

电池
全国电池工业信息中心 湖南轻工研究院

电池

CSTPCD北大核心
影响因子：0.336
ISSN：1001-1579
年,卷(期)：2024.54(4)