摘要
锂离子电池容量退化过程具有非线性强,参数影响多等特点,从而难以被预测。针对该问题,开展评估方法研究,提出融合神经网络与粒子滤波的锂离子电池容量退化估计算法。利用神经网络的非线性优势,采用三层神经网络模型,建立电池容量退化模型,并通过比较数据间相似性引入归一化参数,优化数据结构,然后采用粒子滤波追踪模型内部参数,进而预测电池容量状态。最后利用CALCE实验室电池数据进行验证,并与电池双指数经验模型的粒子滤波算法进行了对比,结果表明论文算法具有更好的准确性和稳定性。
Lithium-ion battery capacity degradation process has the characteristics of strong nonlinearity and many parame⁃ters,which makes it difficult to predict.Aiming at this problem,the evaluation method research is carried out,and the capacity deg⁃radation estimation algorithm of lithium ion battery with neural network and particle filter is proposed.Using the nonlinear advantage of neural network,a three-layer neural network model is used to establish a battery capacity degradation model,and the normalized parameters are obtained by comparing the similarity between the data to optimize the data structure.Then the particle filter is used to track the internal parameters of the model to predict the battery.Finally,the CALCE laboratory battery data is used for verification,and compared with the particle filter algorithm of the battery double exponential empirical model.The results show that the proposed algorithm has better accuracy and stability.
作者
彭发豫
张宁
蒋涛
PENG Fayu;ZHANG Ning;JIANG Tao(College of Weaponry Engineering,Naval University of Engineering,Wuhan 430033)
出处
《舰船电子工程》
2020年第9期44-47,66,共5页
Ship Electronic Engineering
关键词
电池容量估计
神经网络
粒子滤波
剩余寿命
battery capacity estimation
neural networks
particle filter
remaining useful life
