面向实际工况的动力电池早期故障诊断研究
Early Fault Diagnosis of Power Battery for Real Working Conditions
金浩博 1王淑青 1汤璐 1袁晓辉2
作者信息
- 1. 湖北工业大学 太阳能高效利用及储能运行控制湖北省重点实验室,湖北 武汉 430068
- 2. 华中科技大学 土木与水利工程学院,湖北 武汉 430074
- 折叠
摘要
为了保证动力电池的使用安全性,提出了一种面向实际工况的动力电池早期故障诊断方法,可以准确地定位电池早期故障发生时刻.基于MATLAB平台对锂电池组进行仿真建模,研究电池故障演变特性.以实际工况电池模组中的电池状态信息作为模型的训练对象,来模拟电池在正常工作时的电池状态,当电池发生故障时即可定位发生状态变化的时刻.为进一步准确定位故障发生的时刻,对电池估计残差曲线进行了微分处理,能够以最短的时限定位故障发生时刻.该方法不仅能在仿真模拟的故障条件下准确定位故障,而且在数据条件较为严苛的实际工况数据中也能对电池故障进行识别.
Abstract
In order to ensure the safety of power battery use,an early fault diagnosis method for power battery is proposed for actual working conditions,which can accurately locate the moment of early battery fault occurrence.Simulation modeling of lithium battery pack based on MATLAB to study the battery failure evolution characteris-tics.The battery state information in the battery module is used as the training object of the model to simulate the battery state when the battery is in normal operation,and the moment of state change can be located when the bat-tery fails.In order to further locate the moment of failure accurately,the residual curve of battery estimation is dif-ferentiated to locate the moment of failure in the shortest time frame.This method can not only locate the fault ac-curately under the simulated fault conditions,but also identify the battery fault in the actual operating data with more severe data conditions.
关键词
实际工况/动力电池/电池早期故障/BP神经网络Key words
real working conditions/power battery/early battery failure/BP neural network引用本文复制引用
基金项目
国家重点研发计划(021YFC3200405)
出版年
2024
NETL
NSTL
维普
万方数据