针对不同工况下滚动轴承振动数据分布差异大,单一用户数据量少且多个用户间数据不共享的问题,提出一种二次聚合个性化联邦的滚动轴承寿命预测方法。该方法用不同深度的自编码器提取多尺度特征信息并压缩为散点图,实现特征增强;利用无监督二元回归模型确定第一预测时间,构建分段退化标签;提出二次聚合个性化联邦学习算法,各用户构建改进的CNN-LSTM模型,并将其参数上传至服务端,服务端采用多任务学习框架,一次聚合多用户同种工况模型参数;在此基础上,利用批量归一化层参数统计信息计算一次聚合模型间相似度,引入权重更新机制指导模型参数二次聚合,减少不同工况模型间的负迁移现象并学习有益的全局知识,最终形成针对各工况的个性化预测模型。经实验验证,所提方法在保障数据隐私的前提下,可实现不同工况下滚动轴承寿命预测,并且预测的平均得分与不考虑数据隐私的集中式学习方法相当、相较于联邦平均算法平均得分提高0.2197。
Abstract
Addressing the issues of significant differences in vibration data distribution under various operating conditions for rolling bearings, limited data for individual user and the data not shared among multiple users, a life prediction method of rolling bearings under different operating conditions is proposed based on secondary aggregation personalized federated learning. Auto-encoders with different depths are utilized to extract multi-scale feature information and compress it into nested-scatter plot to achieve feature enhancement. An unsupervised binary regression model is employed to determine the first prediction time and construct segmented degradation labels. A secondary aggregation personalized federated learning algorithm is proposed. Each user constructs an enhanced convolutional neural network-long short term memory model and uploads its parameters to the server. The server adopts a multi-task learning framework, aggregates model parameters under the same operating conditions from multiple users for the first time; furthermore, the server utilizes batch normalization layer parameter statistics to calculate the similarity among one-time-aggregation-models. A weight update mechanism is introduced to guide the secondary aggregation of model parameters. Negative transfer effects among models are reduced under different operating conditions, and the learning of beneficial global knowledge can be facilitated. Ultimately, personalized prediction models tailored to specific operating conditions are established. Experimental validation demonstrates that the proposed method can achieve rolling bearing life prediction under different operating conditions while ensuring data privacy. The prediction average score is comparable to the centralized learning method without considering data privacy, and increases by 0.2197 compared to the federated average algorithm.
关键词
滚动轴承 /
多尺度特征提取 /
联邦学习 /
个性化 /
剩余寿命预测
{{custom_keyword}} /
Key words
rolling bearings /
multi-scale feature extraction /
federated learning /
personalized /
remaining life prediction
{{custom_keyword}} /
{{custom_sec.title}}
{{custom_sec.title}}
{{custom_sec.content}}
参考文献
[1] 邵海东, 肖一鸣, 邓乾旺, 等. 基于不确定性感知网络的可信机械故障诊断[J/OL].机械工程学报, 2024 1-13.
Shao Haidong, Xiao Yiming, Deng Qianwang, et al. Trustworthy Mechanical Fault Diagnosis Using Uncertainty-aware Network[J/OL]. Journal of Mechanical Engineering, 2024 1-13.
[2] Kheli Khelif R,Chebel Morello B,Malinowski S,et al.Direct remaining useful life estimation based on support vector regression[J].IEEE Transactions on Industrial Electronic,2016,64(3):2276-2285.
[3] Li Xiang,Wei Zhang,Qian Ding,et al.Deep learning-based remaining useful life estimation of bearings using multi-scale feature extraction[J].Reliability Engineering & System Safety,2019,182:208-218.
[4] 杨宇,张娜,程军圣.全参数动态学习深度信念网络在滚动轴承寿命预测中的应用[J].振动与冲击,2019,38(10):199-205.
Yang Yv,Zhang Na,Cheng Jun Sheng.Application of full parameter dynamic learning deep belief network in rolling bearing life prediction[J].Journal of Vibration and Shock,2019,38(10):199-205(in Chinese).
[5] 王玉静,王诗达,康守强,等.基于改进深度森林的滚动轴承剩余寿命预测方法[J].中国电机工程学报,2020,40(15):5032-5043.
Wang Yujing,Wang Shida,Kang Shouqiang,et al.Rolling bearing remaining life prediction method based on improved deep forest[J].Proceedings of the CSEE,2020,40(15):5032-5043(in Chinese).
[6] Suh S, Lukowicz P, Lee Y O. Generalized multiscale feature extraction for remaining useful life prediction of bearings with generative adversarial networks[J]. Knowledge-Based Systems, 2022, 237: 107866.
[7] Goddard M.The EU general data protection regulation (GDPR):european regulation that has a global impact[J].International Journal of Market Research,2017,59(6):703-705.
[8] Yang Q,Liu Y,Chen T,et al.Federated machine learning:concept and applications[J].ACM Transactions on Intelligent Systems and Technology (TIST),2019,10(2):1-19.
[9] Geng D Q,He H W,Lan X C,et al.Bearing fault diagnosis based on improved federated learning algorithm[J].Computing,2021:1-19.
[10] Chen X,Wang H,Lu S,et al.Bearing remaining useful life prediction using federated learning with taylor-expansion network pruning[J].IEEE Transactions on Ins1trumentation and Measurement,2023,72:1-10.
[11] 于春雨,张文韬,张庆海,等.基于EMD-AR与改进宽度学习系统的滚动轴承故障诊断方法[J/OL].中国电机工程学报,2022:1-13.
Yu Chunyu,Zhang Wentao,Zhang Qinghai,et al.Fault diagnosis method for rolling bearings based on EMD-AR and improved width learning system[J/OL].Proceedings of the CSEE,2022:1-13(in Chinese).
[12] 倪宣明,沈鑫圆,张海.面向异构数据的自适应个性化联邦学习—一种基于参数分解和持续学习的方法[J].中国科学:信息科学,2022,52(12):2306-2320.
Ni Xuanming,Shen Xinyuan,Zhang Hai.Adaptive personalized federated learning for heterogeneous data:a method based on parameter decomposition and continuous learning[J].Scientia Sinica Informationis,2022,52(12):2306-2320(in Chinese).
[13] Tan A Z, Yu H, Cui L, et al. Towards personalized federated learning[J]. IEEE Transactions on Neural Networks and Learning Systems, 2022.
[14] 焦润海,褚佳杰,李俊良,等.基于数据分解的多区域个性化联邦负荷预测方法[J/OL].中国电机工程报,2024,1-13.
Jiao Runhai,Zhu Jiajie,Li Junliang,et al.A multi region personalized federated load forecasting method based on data decomposition[J/OL].Proceedings of the CSEE,2024:1-13(in Chinese).
[15] Yoon J,Jeong W,Lee G.Federated continual learning with weighted inter-client transfer[C]//International Conference on Machine Learning PMLR,2021:12073-12086.
[16] Lu W,Wang J,Chen Y,et al.Personalized federated learning with adaptive batchnorm for healthcare[J].IEEE Transactions on Big Data,2022.
[17] 雷付煜,王兴东,刘亚径,等.基于自编码器的耐火板砖表面缺陷特征提取与分类方法研究[J].制造业自动化,2022,44(12):28-31+67.
Lei Fuyu,Wang Xingdong,Liu Yajing,et al.Research on feature extraction and classification method for surface defects of refractory bricks based on autoencoder[J].Manufacturing Automation,2022,44(12):28-31+67(in Chinese).
[18] Ragab M,Chen Z,Wu M,et al.Adversarial transfer learning for machine remaining useful life prediction[C]//2020 IEEE international conference on prognostics and health management (ICPHM).IEEE,2020:1-7.
[19] Li X,Zhang W,Ma H,et al.Degradation alignment in remaining useful life prediction using deep cycle-consistent learning[J].IEEE Transactions on Neural Networks and Learning Systems,2021,33(10):5480-5491.
[20] Nectoux P,Gouriveau R,Medjaher K,et al.PRONOSTIA:An experimental platform for bearings accelerated degradation tests[C]//IEEE International Conference on Prognostics and Health Management,PHM'12.IEEE Catalog Number:CPF12PHM-CDR,2012:1-8.
[21] Lei Y,Li N,Guo L,et al.Machinery health prognostics:A systematic review from data acquisition to RUL prediction[J].Mechanical systems and signal processing,2018,104:799-834.
[22] Guo L,Li N,Jia F,et al.A recurrent neural network based health indicator for remaining useful life prediction of bearings[J].Neurocomputing,2017,240:98-109.
[23] 康守强,邢颖怡,王玉静,等.基于无监督深度模型迁移的滚动轴承寿命预测方法[J/OL].自动化学报,2021:1-11.
Kang Shouqiang,Xing Yingyi,Wang Yujing,et al.A life prediction method for rolling bearings based on unsupervised deep model transfer[J/OL].Acta Automatica Sinica,2021:1-11(in Chinese).
[24] Lei Y,Han T,Wang B,et al.XJTU-SY rolling element bearing accelerated life test datasets:a tutorial[J].Journal of Mechanical Engineering,2019,55(16):1-6.
{{custom_fnGroup.title_cn}}
脚注
{{custom_fn.content}}