摘要
针对极端学习机(ELM)方法在预测建模过程中因参数随机生成导致的结果不确定性,提出了一种加速鱼群算法(AAFSA)优化极端学习机的锂离子电池剩余寿命预测方法。针对基本人工鱼群算法(AFSA)存在的收敛速度慢、易陷入局部最优解,通过自适应确定感知距离和移动步长加快收敛速度,并在随机移动时增加混沌扰动提高种群多样性改善AFSA的寻优性能;利用AAFSA确定最优ELM隐含层参数,建立锂离子电池剩余寿命预测模型;利用NASA锂离子电池数据集验证,剩余寿命预测相对误差低于4%。与ELM相比预测结果更加稳定可靠,相比支持向量回归(SVR)及其改进方法预测精度有明显的提高,最高可提升近50%。
In this paper,aiming at the results uncertainty caused by the random generation of parameters in the prediction modeling process of extreme learning machine(ELM),an optimizing ELM method based on accelerated artificial fish swarm algorithm(AAFSA)was proposed to predict the remaining used life(RUL)of lithium-ion battery.Firstly,in view of the slow convergence speed and easy to fall into the local optimal solution of the AFSA,the adaptive visual distance and moving step were constructed to accelerate the convergence,and the chaotic disturbance was increased for improving the population diversity at the stage of random moving.Then,the optimal ELM hidden layer parameters were determined by AAFSA,and the RUL prediction model of lithium-ion battery was established.Finally,the effectiveness of the proposed method was verified by using the NASA lithium-ion battery data set,and the residual life prediction relative error was less than 4%.Compared with ELM,the prediction result is more stable and reliable.Compared with SVR and its improved method,the prediction results are significantly improved,maximum up to about 50%.
作者
何星
丁有军
宋丽君
殷春武
HE Xing;DING Youjun;SONG Lijun;YIN Chunwu(School of information and control engineering, Xi’an University of Architecture and Technology, Xi’an 710055, China;Laboratory of Science and technology on integrated Logistics Support, National University of Defense Technology, Changsha 410073, China)
出处
《兵器装备工程学报》
CSCD
北大核心
2022年第2期163-169,共7页
Journal of Ordnance Equipment Engineering
基金
陕西省教育厅专项科研计划项目(20JK0728)
国防科技重点实验室基金项目(6142003190204)
西安建筑科技大学自然科学专项项目(ZR20048)。
关键词
极端学习机
人工鱼群算法
数据驱动
剩余寿命预测
锂离子电池
extreme learning machine
artificial fish swarm algorithm
data-driven
remaining useful life
lithium-ion battery
