基于DEPSO-RVM的B787电池剩余寿命预测  

Remaining useful life prediction for B787 plane battery based on DEPSO-RVM

作　　者：刘贵行[1] 穆东旭 LIU Guihang;MU Dongxu(Engineering Technology Training Center,Civil Aviation University of China,Tianjin 300300,China;School of Electronic Information and Automation,Civil Aviation University of China,Tianjin 300300,China)

机构地区：[1]中国民航大学工程训练中心,天津300300 [2]中国民航大学电子信息与自动化学院,天津300300

出　　处：《现代电子技术》2019年第20期94-98,102,共6页Modern Electronics Technique

基　　金：国家自然科学基金项目(U1733119);国家自然科学基金项目(U1333111);中央高校基本科研业务费：机载气象雷达冷却系统性能评估研究(3122016D008)

摘　　要：锂离子电池已经被应用于B787客机,为进一步提高B787锂离子电池的可靠性,针对传统基于相关向量机的电池剩余使用寿命预测方法的不足,提出一种把相关向量机、差分进化算法和粒子群优化算法融合的的方法。通过差分进化算法和粒子群优化算法对相关向量机的参数进行优化,增强其对电池历史监测数据退化趋势的预测能力。应用卡尔曼滤波器对融合算法实施优化,将优化后的预测结果作为在线样本添加到训练集中,对提出的模型重新训练,以此来动态调整系数矩阵和相关向量以执行下一次迭代预测。基于B787锂离子电池测量数据,对所提方法的有效性和鲁棒性进行了验证。In allusion to the shortcomings of the traditional battery residual useful life prediction method based on correla?tion vector machine,a new method integrating RVM(relevance vector machine),DE(differential evolution)algorithm and PSO(particle swam optimization)algorithm is proposed.The parameters of the RVM are optimized by the mean of DE algorithm and PSO algorithm to enhance its ability to predict the degradation trend of battery historical monitoring data.The Kalman filter is ap?plied to optimize the fusion algorithm.The optimized prediction results is added to the training set as an online sample,and the proposed model is retrained to dynamically adjust the coefficient matrix and the correlation vector to execute the next iteration prediction.On the basis of the measured data of lithium?ion batteryused in B787 plane,the effectiveness and robustness of the proposed method are verified.

关 键 词：剩余寿命预测 相关向量机 B787锂离子电池 差分进化算法 粒子群优化算法 卡尔曼滤波 

分 类 号：TN36?34[电子电信—物理电子学]