旋转机械的剩余使用寿命（Remaining Useful Life，RUL）预测对于工业设备故障预测与健康管理（Prognostication and Health Management，PHM）具有重要意义，并且越来越多不同种类的传感器被用于旋转机械的监测和信息采集，通过多传感器信息进行旋转机械的剩余使用寿命研究与应用。研究旋转机械的退化规律，建立旋转机械剩余使用寿命预测模型，预测旋转机械的剩余使用寿命，合理地安排旋转机械维修计划，有效地保障旋转机械安全且稳定运行。本文以旋转机械为研究对象，利用旋转机械的多传感器信号，从数据驱动的角度出发，采用深度学习算法对旋转机械剩余使用寿命进行研究，                                                           主要研究内容如下：

（1）研究旋转机械的多传感器监测数据预处理。研究旋转机械的状态监测技术，保证旋转机械信息采集的正确性，并为旋转机械剩余使用寿命预测实验验证提供理论依据。

针对数据的异常影响模型特征训练和预测结果的准确性，研究基于3sigma准则剔除异常值方法和改进小波阈值降噪方法实现异常信号去除，提高数据质量，避免数据异常影响模型训练和剩余使用寿命预测的准确性。

（2）针对多传感器冗余数据导致旋转机械退化信息提取困难，并且使得信息融合结果不能准确表征旋转机械退化过程、剩余使用寿命预测准确性低的问题。本文提出一种基于核主成分分析（Kernel Principal Component Analysis，KPCA）的多维退化数据降维与信息融合方法。首先获取旋转机械多个传感器的退化曲线，选择退化趋势明显的传感器数据，采用KPCA算法进行数据融合；然后提取融合后数据的退化特征，构建综合评价指标选择出能够反映旋转机械退化的特征，基于KPCA将选择的退化特征进行融合，将满足KPCA贡献率阈值的主元数据作为预测模型的输入数据，实现旋转机械退化特征融合；最后引入旋转机械健康指标（Health indicators，HI）对旋转机械的健康状态进行表征，将满足KPCA贡献率阈值的主元数据归一化处理作为旋转机械的健康指标，通过多阶微分划分旋转机械健康状态，分析旋转机械的退化期，确定旋转机械真实的剩余使用寿命。

（3）深度学习方法中长短时记忆神经网络（Long-Short Term Memory，LSTM)有效解决了传统RNN模型的“记忆衰退”的问题，可以有效预测长时间序列数据，但LSTM模型最优参数难以确定。本文基于贝叶斯（Bayesian Optimization，BO）方法自动寻优LSTM模型的超参数。首先将满足KPCA累计贡献率的主元数据作为BO-LSTM模型的输入数据，从全局寻优确定LSTM模型的最优超参数；然后建立贝叶斯参数优化的LSTM模型；最后通过PHM2012数据集，将数据集划分为训练集与测试集，进行轴承剩余使用寿命预测。实验结果表明本文所提方法具有较好的预测效果，有效优化模型的预测过程，提高剩余使用寿命预测精度。

（4）搭建典型的旋转机械-减速器全寿命周期实验平台，并通过多种传感器采集设备运行数据并同步存储至云服务器，基于Labview实现上位机实时数据监测。并通过前端调用接口访问数据，建立旋转机械剩余使用寿命预测系统，实现研究算法的应用。最后通过减速器全寿命周期数据对本文提出的KPCA-LSTM剩余使用寿命预测模型进行验证。结果表明：KPCA-LSTM模型预测的绝对误差、均方根误差和Score 评分误差与得分均优于其它模型，并且KPCA-LSTM预测模型可以有效地实现信息融合，使得模型预测用时更短，收敛速度更快，可准确地预测旋转机械的剩余使用寿命。

综上所述本文所提的基于多传感器信息融合的旋转机械剩余使用寿命预测方法，能够准确地预测旋转机械的剩余使用寿命，对于旋转机械的预测性维护提供良好的理论指导。

The Remaining Useful Life (RUL) prediction of rotating machinery is essential for industrial equipment failure prediction and health management (PHM). Moreover, more and more different kinds of sensors are used to monitor and acquire rotating machinery information. Research and application of the remaining useful life of rotating machinery are carried out through multi-sensor information. Study the degradation law of rotating machinery, establish the remaining useful life prediction model of rotating machinery, predict the remaining useful life of rotating machinery, make reasonable arrangements for a rotating machinery maintenance plan, and effectively ensure safe and stable operation. This paper takes rotating machinery as the research object, utilizes multi-sensor signals of rotating machinery, and adopts a deep learning algorithm to study the remaining useful life of rotating machinery from the perspective of data drive. The main research contents are as follows:

The preprocessing of multi-sensor monitoring data for rotating machinery is studied.The condition monitoring technology of rotating machinery is studied to ensure the correctness of information collection of rotating machinery, and to provide a theoretical basis for the prediction of the remaining service life of rotating machinery.In view of the impact of data anomalies on the accuracy of model feature training and prediction results, the method of eliminating outliers based on the 3sigma criterion and the improved wavelet threshold denoising method are studied to remove abnormal signals, improve data quality and avoid data anomalies affecting the accuracy of model training and remaining service life prediction.

(2)The problem of difficulty in extracting the degradation information of rotating machinery due to redundant data from multiple sensors, the problem that the information fusion results cannot accurately characterize the degradation process of rotating machinery, and the low accuracy of remaining useful life prediction. In this paper, a method based on Kernel Principal Component Analysis (KPCA) for dimensionality reduction and information fusion of multi-dimensional degraded data.Finally, the health indicators (HI) are introduced to characterize the health status of rotating machines, and the primary metadata that meets the contribution threshold of KPCA is normalized as the health indicators of rotating machines. The health status of rotating machines is classified through multi-order differentiation. The health status of rotating machinery is organized, the degradation period of rotating machinery is analyzed, and the true remaining useful life of rotating machinery is determined.

(3)Long-Short Term Memory (LSTM) neural networks in deep learning methods can effectively solve the problem of "memory decline" of the traditional RNN model. It can effectively predict long-time series data. However, it is difficult to determine the optimal parameters of the LSTM model. This paper is based on the Bayesian Optimization (BO) method to automatically optimize the super parameters of the LSTM model.Firstly, the principal metadata that meets the KPCA contribution rate is taken as the input data of the BO-LSTM model, and the optimal hyperparameters of the LSTM model are determined from the global optimization. Then the LSTM model of Bayesian parameter optimization is established. Finally, The remaining useful life prediction method proposed in this paper is verified in PHM2012 data set.The experimental results show that the proposed method has a good prediction effect, effectively optimizes the process of model prediction, and improves the accuracy of remaining useful life prediction.

(4) A typical full-lifecycle experiment platform of rotary machine-reducer was built, and the operation data of the device was collected by various sensors and synchronously stored on the cloud server. Real-time data monitoring of the upper computer was realized based on Labview. By accessing the data through the front-end call interface, the prediction system of the remaining useful life of rotating machinery is established, and the application of the research algorithm is realized. Finally, the KPCA-LSTM remaining useful life prediction model is verified by the whole life cycle data of the reducer. The results show that the absolute error, root mean square error, and Score error and score of the KPCA-LSTM model are better than those of other models. Moreover, The KPCA-LSTM prediction model can effectively achieve information fusion, which makes the model prediction take less time and converge faster, and can accurately predict the remaining useful life of rotating machinery.

In summary, the prediction method of the remaining useful life of rotating machinery based on multi-sensor fusion proposed in this paper can accurately predict the remaining useful life of rotating machinery and provide good theoretical guidance for predictive maintenance of rotating machinery.

[1] 张龙,胡燕青,赵丽娟等.多通道信息融合与深度迁移学习的旋转机械故障诊断[J/OL].中国机械工程：1-12.

[2] 黄扣. 基于深度学习的旋转类机械预测性维护方法研究[D].江苏科技大学,2021.

[3] 葛阳,郭兰中,牛曙光,窦岩.基于t-SNE和LSTM的旋转机械剩余寿命预测[J].振动与冲击,2020,39(07):223-231.

[4] David Verstraete et al. A Deep Adversarial Approach Based on Multi-Sensor Fusion for Semi-Supervised Remaining Useful Life Prognostics[J]. Sensors, 2019, 20(1) : 176-176.

[5]孙博,康锐,谢劲松.故障预测与健康管理系统研究和应用现状综述[J].系统工程与电子技术,2007, (10): 1762-1767.

[6] 徐小力,韩秋实,高锦宏,等．面向旋转机械状态维护的智能化在线状态预测方法及系统[J].精密制造与自动化,2003(S1): 91-92.

[7] 程阳洋. 基于量子深度神经网络的旋转机械剩余寿命预测研究[D].四川大学,2021.

[8] H.E.Kim,A. C.C. Tan,J. Mathew, et al. Bearing fault prognosis based on health state probability estimation [J]. Expert Systems with Applications, 2012,39:5200-5213.

[9] Y.G. Lei,N. P.Li,L. Guo, et al. Machinery health prognostics: A systematic review from data acquisition to RUL prediction [J]. Mechanical Systems andSignal Processing, 2018,104: 799-834.

[10] 石慧,康辉,任谦力等.随机相关性影响的多部件系统剩余寿命预测[J].振动与冲击,2022,41(21):299-307+324.

[11] 程阳洋.基于量子深度神经网络的旋转机械剩余寿命预测研究[D].四川大学,2021.

[12] 钱志远.不完美维护活动干预下设备的剩余使用寿命预测与维护决策研究[D].江西理工大学,2022.

[13] 赵申坤,姜潮,龙湘云.一种基于数据驱动和贝叶斯理论的机械系统剩余寿命预测方法[J].机械工程学报,2018,54(12):115-124.

[14] 赵广社,吴思思,荣海军.多源统计数据驱动的航空发动机剩余寿命预测方法[J].西安交通大学学报,2017, 51(11): 150-155+172.

[15] 田瑶瑶. 基于机器学习的机电系统关键部件PHM技术研究[D].南京航空航天大学,2018.

[16] 李乃鹏,蔡潇,雷亚国,徐鹏程,王文廷,王彪.一种融合多传感器数据的数模联动机械剩余寿命预测方法[J].机械工程学报,2021,57(20):29-37.

[17] 张志壮. 考虑退化状态动态转移的剩余寿命预测研究[D].太原科技大学,2021.

[18] X. Fang, K. Paynabar, N. Gebraeel. Multistream sensor fusion-based prognostics model for systems with single failure modes[J]. Reliability Engineering & System Safety. 2017; 159: 322-31.

[19] Cao Yudong et al.Transfer learning for remaining useful life prediction of multi- conditions bearings based on bidirectional- GRU network[J].Measurement, 2021.

[20] Chen J, Jin H, Chang Y, Liu Q. Gated recurrent unit-based recurrent neural network for remaining useful life prediction of the nonlinear deterioration process.Reliability Engineering & System Safety.2019;185:372-82.

[21] Wang Yaping et al. Remaining useful life prediction of rolling bearings based on Pearson correlation-KPCA multi-feature fusion [J]. Measurement, 2022, 201.

[22] 马依琳,陶慧玲,董启文等.基于Transformer的多特征融合的航空发动机剩余使用寿命预[J].华东师范大学学报（自然科学版),2022,No.225(05):219-232.

[23] Liu Lilan et al. An Improved Entropy-Weighted Topsis Method for Decision-Level Fusion Evaluation System of Multi-Source Data[J]. Sensors, 2022, 22(17) : 6391-6391.

[24] 杨天琦.一种旋转机械的退化状态估计与剩余使用寿命预测方法研究[D].武汉科技大学,2020.

[25] 童清俊.基于数据驱动的滚动轴承性能退化评估与剩余寿命预测方法研究[D].东南大学,2020.

[26] 李广福,马萍,张宏立等.一种Informer模型的滚动轴承剩余寿命预测方法[J/OL].机械科学与技术：1-8[2023-03-27].

[27] 王拳.基于EMD和熵特征的滚动轴承故障诊断研究[D].沈阳化工大学,2021.

[28] 李乐豪. 基于自编码器的滚动轴承故障诊断与健康评估研究[D].石家庄铁道大学,2022.

[29] Lin D, Wiseman M,Banjevic D, et al. An approach to signal processing and condition- based maintenance for gearboxes subject to tooth failure[J]. Mechanical Systems & Signal Processing, 2004,18(5): 993-1007.

[30] Wang D, Tse P W.Prognostics of slurry pumps based on a moving-average wear degradation index and a general sequential Monte Carlo method[J]. Mechanical Systems & Signal Processing, 2015,56-57: 213-229.

[31] Widodo A,Yang B S. Application of relevance vector machine and survival probability to machine degradation assessment[J]. Expert Systems with Applications, 2011,38(3): 2592-2599.

[32] Jin x, Sun Y,Que Z, et al.Anomaly Detection and Fault Prognosis for Bearings[J]. IEEE Transactions on Instrumentation & Measurement, 2016,65(9): 2046-2054.

[33] 郭磊.基于核模式分析方法的旋转机械性能退化评估技术研究[D].上海交通大学,2009.

[34] Chen J, Jin H, Chang Y, Liu Q. Gated recurrent unit-based recurrent neural network for remaining useful life prediction of the nonlinear deterioration process.Reliability Engineering & System Safety.2019;185:372-82.

[35] Wang Yaping et al. Remaining useful life prediction of rolling bearings based on Pearson correlation-KPCA multi-feature fusion [J]. Measurement, 2022, 201.

[36] 马依琳,陶慧玲,董启文等.基于Transformer的多特征融合的航空发动机剩余使用寿命预[J].华东师范大学学报（自然科学版),2022,No.225(05):219-232.

[37] Liu Lilan et al. An Improved Entropy-Weighted Topsis Method for Decision-Level Fusion Evaluation System of Multi-Source Data[J]. Sensors, 2022, 22(17) : 6391-6391.

[38] 杨天琦.一种旋转机械的退化状态估计与剩余使用寿命预测方法研究[D].武汉科技大学,2020.

[39] 童清俊.基于数据驱动的滚动轴承性能退化评估与剩余寿命预测方法研究[D].东南大学,2020.

[40] 李广福,马萍,张宏立等.一种Informer模型的滚动轴承剩余寿命预测方法[J/OL].机械科学与技术：1-8[2023-03-27].

[41] 王拳.基于EMD和熵特征的滚动轴承故障诊断研究[D].沈阳化工大学,2021.

[42] 李乐豪. 基于自编码器的滚动轴承故障诊断与健康评估研究[D].石家庄铁道大学,2022.DOI:10.27334/d.cnki.gstdy.2022.000039.

[43] Lin D, Wiseman M,Banjevic D, et al. An approach to signal processing and condition-based maintenance for gearboxes subject to tooth failure[J]. Mechanical Systems & Signal Processing, 2004,18(5): 993-1007.

[44] Wang D, Tse P W.Prognostics of slurry pumps based on a moving-average wear degradation index and a general sequential Monte Carlo method[J]. Mechanical Systems & Signal Processing, 2015,56-57: 213-229.

[45] Widodo A,Yang B S. Application of relevance vector machine and survival probability to machine degradation assessment[J]. Expert Systems withApplications, 2011,38(3): 2592-2599.

[46] Jin x, Sun Y,Que Z, et al.Anomaly Detection and Fault Prognosis for Bearings[J]. IEEE Transactions on Instrumentation & Measurement, 2016,65(9): 2046-2054.

[47] 郭磊.基于核模式分析方法的旋转机械性能退化评估技术研究[D].上海交通大学,2009.

[48] Hu Y, Palmé T, Fink O. Deep Health Indicator Extraction: A Method based on Auto- encodersand Extreme Learning Machines[C]/Annual Conference of thePrognostics and Health Management Society. 2016,7∶7.

[49] 潘玉娜,陈进.小波包-支持向量数据描述在轴承性能退化评估中的应用研究[J.振动与冲击,2009,28(4): 164-167.

[50] Giantomassi A, Ferracuti F, Benini A, et al. Hidden Markov Model for Health Estimation and Prognosis of Turbofan Engines[C]// ASME 2011 International Design Engineering Technical Conferences & Computers and Information in Engineering Conference.2011: 1-6.

[51] Ocak H, Loparo K A, Discenzo F M.Online tracking ofbearing wear using wavelet packet decomposition and probabilistic modeling: A method for bearing prognostics[J]. Journal of Sound & Vibration, 2007,302(4): 951-961.

[52] Li Y, Xu M, Wei Y, et al. Health Condition Monitoring and Early Fault Diagnosis of Bearings Using SDF and Intrinsic Characteristic-Scale Decomposition [J]. IEEE Transactions on Instrumentation & Measurement, 2016,65(9): 2174-2189.

[53] ZIO E, MAIO F D.A data-driven fuzzy approach for predicting the remaining useful life in dynamic failure scenarios of a nuclear system[J]. Reliability Engineering & System Safety, 2013,95(1): 49-57.

[54] 裘冬平,孙建华.基于GA-BP神经网络的输油管道内腐蚀速率预测[J].青岛大学学报（工程技术版),2011,26(01):28-32.

[55] 钱志远.不完美维护活动干预下设备的剩余使用寿命预测与维护决策研究[D].江西理工大学,2022.

[56] 邓森,景博,周宏亮,朱海鹏,刘小平.一种改进的自适应增强支持向量回归机的故障预测方法[J].兵工学报,2012,33(08):991-996.

[57] Doksum k A,Hoyland A.Models for variable stress accelerated life testing experiments based on Wiener processrs and the inverse Gaussian distribution[J]. Technometrics, 1992, 34(1): 74-82.

[58] Huang Z Y, Xu Z G, Wang W H, Sun Y X.Remaininguseful life prediction for a nonlinear heterogeneous wienerprocess model with an adaptive drift. IEEE Transactions on Reliability, 2015, 64(2): 687−700.

[59] 朱磊,左洪福,蔡景.基于Wiener过程的民用航空发动机性能可靠性预测[J].航空动力学报,2013, 28(28): 1006-1012.

[60] 赵广社,吴思思,荣海军.多源统计数据驱动的航空发动机剩余寿命预测方法[J].西安交通大学学报,2017, 51(11): 150-155+172.

[61] Cong Chen et al. Railway turnout system RUL prediction based on feature fusion and genetic programming[J]. Measurement, 2020, 151(C) : 107162-107162.

[62] FAN J J, YUNG K C , PECHT M. Physics-of-failure-based prognostics and health management for high-power white light-emitting diode lighting[J]. IEEE Transactions on Device and Materials Reliability, 2011, 11(3): 407-416.

[63] 刘帮俊,陈荣刚,吴斌.火炮身管失效机理和寿命预测[J].兵器装备工程学报,2016, 37(12): 121-125+149.

[64] Maio ,Kwok Leung Tsui,Enrico Zio. Combining Relevance Vector Machines and exponential regression for bearing residual life estimation[J]. Mechanical Systems and Signal Processing, 2012, 31 : 405-427.

[65] ZHAO F,TIAN Z,ZENG Y.Uncertainty quantification in gear remaining useful life prediction through an integrated prognostics method[J].IEEE Transactions on Reliability, 2013,62(1): 146-159.

[66] Chiachio J , Chiachio M , Sankararaman S , et al. Condition-based prediction of time- dependent reliability in composites[J]. Reliability Engineering & System Safety, 2015, 142:134-147.

[67] LEI Y,LI N,GONTARZ S, et al.A model-based method for remaining useful life prediction of machinery[J].IEEE Transactions on Reliability, 2016,65(3): 1314-1326.

[68] 吴超勇. 大数据驱动的轴承寿命预测关键技术研究[D]. 华中科技大学,2017.

[69] Loutas T H, Georgoulas G. Remaining useful life estimation in rolling bearings utilizing data-driven probabilistic e-support vectorsre gression [J].IEEE Transactions on Reliability, 2013, 62 (4) :821-832.

[70] Bengio, Y,Courville, A.Vincent. Representation learning:a review and new perspectives. IEEE Transcations on Pattern Analysis Machine Intelligence, 2013, 35 (8) :1798-1828.

[71] 黎慧,张国文.基于灰色模型的滚动轴承剩余寿命预测[J].机械设计与研究,2018,34(1):113-116,120.

[72] 舒涛,张一弛,丁日显.基于灰色模型与LSTM网络的旋转机械寿命预测[J].系统工程与电子技术,2021,43(08):2355-2361.

[73] 张金豹.基于全寿命数据的滚动轴承运行状态评估和剩余寿命预测[D].哈尔滨工业大学,2020.

[74] 邹旺,吉畅.基于机器学习的轴承RUL预测方法综述[J].电脑知识与技术,2022,18(19):74-76+81.

[75] WANG L X,WU Z H,FU Y D,et al.Remaining life predictions of fan based on time series analysis and BP neural networks[C]//Information Technology, Networking, Electronic and Automation Control Conference, IEEE. 2016: 607-611.

[76] Bao Jinsong,Guangchao Yuan,Zheng Xiaohu,Zhang Jianguo,Ji Xia.A Data Driven Model forPredicting Tool Health Condition in High Speed Milling of Titanium Plates Using Real-Time SCADA[J].Procedia CIRP, 2017(59):61.

[77] Sbarufatti C, Corbetta M, Manes A, et al.Sequential Monte-Carlo Sampling Based on a Committee of Artificial Neural Networks for Posterior State Estimation and Residual Life time Prediction[J]. International Journal of Fatigue, 2015,83:10-23.

[78] Rai A,Upadhyay S H. The Use of MD-CUMSUM and NARX Neural Network for Anticipating the Remaining Useful Life of Bearings [J]. Measurement,2017,111:397-410.

[79] 禹鑫燚,施甜峰,唐权瑞,等. 面向预测性维护的工业设备管理系统[J]. 计算机科学,2020, 47(S2): 667-672+677.

[80] 程梦瑶. 大数据驱动下的PHM技术[J]. 软件和集成电路,2016 (05): 22-25.

[81] 李天梅,司小胜,刘翔,裴洪.大数据下数模联动的随机退化设备剩余寿命预测技术[J/OL].自动化学报：1-23.

[82] 刘惠,刘振宇,郏维强,等.深度学习在装备剩余使用寿命预测技术中的研究现状与挑战[J]. 计算机集成制造系统,2021, 27(01): 34-52.

[83] Deutsch J,He D. Using Deep Learning Based Approaches for Bearing Remaining Useful Life Prediction[C]. Annual Conference of the Prognostics and Health Management Society, Denver,USA,2016:292-298.

[84] Nectoux P, Gouriveau R, Medjaher K, et al. PRONOSTIA:An Experimental Platform for Accelerated Degradation Tests[C]. IEEE International Conference on Prognostics and Health Management(PHM), Denver,2012: 1-8.

[85] 罗金喜. 边缘计算中基于区块链的访问控制机制研究[D].南京邮电大学,2022.

[86] 房虹辰. 基于LSTM和CNN的频谱感知方法研究[D].宁夏大学,2022.

[87] Shen-Junxian and Xu-Feiyun. Method of fault feature selection and fusion based on poll mode and optimized weighted KPCA for bearings[J]. Measurement, 2022, 194.

[88] 周俊. 数据驱动的航空发动机剩余使用寿命预测方法研究[D].南京航空航天大学,2017.

[89] 尹桂宾. 基于皮尔逊-KPCA多特征融合的滚动轴承寿命预测研究[D].哈尔滨理工大学,2021.

[90] 魏熙朋.基于深度学习的滚动轴承健康状态评估及剩余寿命预测[D].西南交通大学,2021.

[91] 洪振麒.基于多特征融合的滚动轴承剩余寿命预测方法研究[D].沈阳大学,2021.

[92] Chen Zhaodong et al. Effective and Efficient Batch Normalization Using a Few Uncorrelated Data for Statistics Estimation.[J].IEEE transactions on neural networks and learning systems, 2020,PP : 1-15.

[93] 张其霄,董鹏,王科文等.基于贝叶斯优化LSTM的发动机剩余寿命预测[J].火力与指挥控制,2022,47(04):85-89.

[94] Ahmed Elsheikh et al. Bidirectional Handshaking LSTM for Remaining Useful Life Prediction[J]. Neuro computing, 2018, 323: 148-156.

[95] Zhao-HuaLiu,Xu-DongMeng,Hua-LiangWei,Bi-Liang.A Regularized LSTM Method for Predicting Remaining Useful Life of Rolling Bearings[J].International Journal of Automation and Computing,2021,18(04):581-593.

[96] 魏佳恒,郭惠勇.基于贝叶斯优化BiLSTM模型的输电塔损伤识别[J].振动与冲击,2023,42(01):238-248.

[97] Li Yuxiong et al. Wiener-based remaining useful life prediction of rolling bearings using improved Kalman filtering and adaptive modification[J]. Measurement, 2021, 182.

[98] Bi J X,Fan W Z,Wang S B.Remaining Life Prediction for Aircraft Turbine Engines Based on LSTM-RNN-HMM-APPROACH[J]. IOP Conference Series: Materials Science and Engineering, 2021, 1043(2):022033 (14pp).

[99] Chen Z,Wu M,Zhao R,et al. Machine Remaining Useful Life Prediction via an Attention Based Deep Learning Approach[J]. IEEE Transactions on Industrial Electronics, 2020, PP(99):1-1.

[100]Fengtao Wang et al. Remaining Life Prediction Method for Rolling Bearing Based on the Long Short-Term Memory Network[J]. Neural Processing Letters, 2019, 50(2): 2437- 2454.

[101]Venkateswarlu Gundu et al. PSO-LSTM for short term forecast of heterogeneous time series electricity price signals[J]. Journal of Ambient Intelligence and Humanized Computing, 2020, 12: 1-11.