汽车锂离子动力电池剩余寿命粒子滤波预测研究

Study on Particle Filtering Prediction of Remaining Life of Automotive Lithium-Ion Power Battery

针对汽车锂离子动力电池寿命存在明显不确定性,设计了一种基于标准粒子滤波粒子滤波的锂离子动力电池剩余寿命(RUL)预测方法。为获得更高精度的RUL预测结果,选择粒子滤波方法建立重要性密度函数,完成粒子滤波优化效果。研究结果表明:RUL预测结果获得33个循环预测误差,为8.56%,以粒子滤波进行预测的误差共26个循环,为6.81%。粒子滤波相对粒子群预测性能更优,增加训练数据样本量有助于电池RUL达到更低预测误差,获得更高预测精度。该研究易于实现,具有很好的推广价值。

Due to the obvious uncertainty of automotive lithium-ion power battery,this paper designs a residual life(RUL)prediction method of lithium-ion power battery based on standard particle filtering particle filtering.In order to obtain more accurate RUL prediction results,the particle filtering method is selected to establish the importance density function and complete the particle filtering optimisation effect.The research results show that the RUL prediction result obtains 33 cycles of prediction error,which is 8.56%,and the prediction error with particle filtering is a total of 26 cycles,which is 6.81%.Particle filtering has better prediction performance relative to particle swarm,and increasing the sample size of training data helps battery RUL to achieve lower prediction error and higher prediction accuracy.The study is easy to implement and has good value for promotion.