基于KPCA与NRBO-Transformer的锂电池健康状态评估方法OA
Lithium battery health state assessment method based on KPCA and NRBO-Transformer
锂电池健康状态(state of health,SOH)可表征锂电池的老化状态.为准确评估SOH,首先,提取充电阶段的电流、电压、IC曲线中的6 个特征,为了提高输入特征的质量,采用核主成分分析(kernel principal component analysis,KPCA)结合Spearman相关性分析,消除多维特征的冗余性获取输入特征的关键信息.其次,为了降低模型复杂度,将全连接层代替Transformer解码器,并利用牛顿-拉夫逊优化算法(Newton-Raphson-based optimization algorithm,NR-BO)对模型的超参数寻优,提高预测精度.最后,利用公开数据集不同训练比例划分验证方法的有效性,并采用不同电池交叉验证与灰狼优化算法(gray wolf optimization algorithm,GWO)和鲸鱼优化算法(whale optimization algorithm,WOA)进行比较,结果表明:所提方法在精度和计算耗时方面均优于其他2 种算法.
The state of health(SOH)shows the aging state of lithium batteries.To accurately assess SOH,this paper first extracts six features in the current,voltage,and IC curves of the charging stage.To improve the quality of the input features,kernel principal component analysis(KPCA)combined with Spearman correlation analysis is employed to eliminate the redundancy of the multidimensional features and obtain the key information of the input features.Then,to reduce the model complexity,the fully connected layer is employed instead of the transformer decoder.The Newton-Raphson-based optimization algorithm(NRBO)is introduced to optimize the hyper-parameters of the model and improve the prediction accuracy.Finally,the effectiveness of the method is verified by using different training ratio divisions of the public dataset.Results show the proposed method excels in both accuracy and computational time compared with the gray wolf optimization algorithm(GWO)and the whale optimization algorithm(WOA).
刘富强;刘为国;朱洪波;胡凯;马旭东
安徽理工大学 电气与信息工程学院,安徽 淮南 232001安徽理工大学 电气与信息工程学院,安徽 淮南 232001安徽理工大学 电气与信息工程学院,安徽 淮南 232001安徽理工大学 电气与信息工程学院,安徽 淮南 232001安徽理工大学 电气与信息工程学院,安徽 淮南 232001
信息技术与安全科学
核主成分分析牛顿-拉夫逊Transformer模型锂电池健康状态
kernel principal component analysisNewton-Raphsontransformer modellithium batterystate of health
《重庆理工大学学报》 2026 (3)
35-44,10
国家自然科学基金项目(62003001)
评论