剩余寿命预测作为故障预测与健康管理技术（PHM）的重要组成部分，已成为当前工业领域中的关键研究方向之一。随着机械设备自动化程度的不断提高和智能化技术的快速发展，对设备运行状态的实时监测和剩余寿命的精准预测需求日益迫切。滚动轴承作为旋转机械设备的核心部件，其性能状态直接关系到设备的安全运行和生产效率。然而，尽管在该领域已经取得了一定的研究成果，但在滚动轴承剩余寿命预测领域仍然存在挑战，例如数据采集与处理、特征提取与选择、模型建立与验证等方面的问题仍有待进一步解决。因此，深入研究滚动轴承剩余寿命预测方法，提高预测准确性和可靠性，对于优化设备维护策略、降低生产成本和提高生产效率具有重要意义。本研究采用基于数据驱动的方法对滚动轴承的多域退化特征进行分析和建模，以实现滚动轴承剩余寿命的预测。本文的主要研究内容包括：

（1）研究振动信号降噪与高维多域退化特征提取方法。设备在实际工况环境中振动信号的特征信息通常较微弱且容易受到外部强噪声等因素的干扰，难以真实反映设备退化状态信息。此外，将振动信号从噪声中有效分离对滚动轴承的状态监测及剩余寿命预测等研究也具有重要影响。针对上述问题，提出了基于改进蜣螂算法优化变分模态分解的信号降噪方法，并结合多域退化特征提取方法对降噪后的振动信号进行特征提取。首先，利用信号模态分解降噪原理建立IDBO-VMD振动信号降噪模型，通过将原始信号分解为多组内涵模态分量和残差项，采用约束条件控制模态函数的光滑性和稀疏性，然后通过迭代优化更新和调整模态函数，有效地逐步减小信号中噪声成分的影响。通过振动仿真信号验证本文提出的降噪方法。然后对降噪后信号提取有量纲时域特征、无量纲时域特征、频域特征、熵特征以及谱峭度相关特征等构成多域混合特征集，从多维度描述设备退化过程进行滚动轴承的剩余寿命预测研究。

（2）研究基于多域退化特征的滚动轴承健康指标构建方法。首先，针对单一特征难以全面有效描述滚动轴承性能退化状态，本文通过分析多域特征指标的退化特性，构建了退化特征筛选准则，从多域混合特征集中获取对设备退化过程敏感的特征实现了高维多域特征的有效选取。接着，通过具有非线性特性的等距特征映射算法（ISOMAP）实现了多域退化特征融合，减少了冗余特征信息的影响，构建得到滚动轴承健康指标。然后对降维约简后的特征通过模糊核C均值聚类方法确定出滚动轴承运行过程的初始退化时间点。

（3）研究基于改进时间卷积网络的滚动轴承剩余寿命预测模型。首先，针对现有模型在机械设备退化过程中对关键信息的特征提取能力不足，缺乏对关键信息的重点关注等问题，提出了一种TCNMABG剩余寿命预测模型。该模型结合了时间卷积神经网络，双向门控循环网络以及多头注意力机制的优势，有效拓宽了模型的感受野，增强了模型的长时间信息学习能力。然后，利用XJTU-SY数据集对本章所提剩余寿命预测方法进行实验验证，实验结果表明，该方法能够更深层次地学习输入数据中的长期时序信息以及有效捕捉关键特征，增强了模型的预测性能。

（4）基于全生命周期数据采集实验平台对所提出的方法进行实验验证。通过在实验室模拟实际工况环境搭建减速器全生命周期退化实验硬件平台，并建立滚动轴承剩余寿命预测系统，实现研究算法的应用。重点采集减速器输入端滚动轴承水平和垂直方向振动信号退化数据集，应用本文所采用的多域退化特征提取与选择方法、健康指标构建方法和剩余寿命预测方法，对所提剩余寿命预测模型进行训练、测试、参数优化、性能对比，分析模型在实验室减速器平台上的预测误差与实际性能，进一步验证了本文所提方法的有效性。

As an important component of fault prediction and health management technology (PHM), the prediction of remaining useful life has become one of the key research directions in the current industrial field. With the continuous improvement of the automation level of mechanical equipment and the rapid development of intelligent technology, the demand for real-time monitoring of equipment operating status and accurate prediction of remaining useful life is becoming increasingly urgent. As the core component of rotating machinery equipment, the performance status of rolling bearings is directly related to the safe operation and production efficiency of the equipment. However, despite the research achievements in this field, there are still challenges in the field of rolling bearing remaining useful life prediction, such as data acquisition and processing, feature extraction and selection, model establishment and verification, which need to be further addressed. Therefore, in-depth research on rolling bearing remaining useful life prediction methods to improve prediction accuracy and reliability is of great significance for optimizing equipment maintenance strategies, reducing production costs, and improving production efficiency. This study uses a data-driven approach to analyze and model the multi-domain degradation characteristics of rolling bearings to achieve the prediction of rolling bearing remaining useful life. The main research contents of this article include:

(1) Research on methods for noise reduction and high-dimensional multi-domain degradation feature extraction of vibration signals. The characteristic information of vibration signals in actual working conditions is usually weak and easily affected by external strong noise and other factors, making it difficult to truly reflect the degradation state information of equipment. In addition, effectively separating vibration signals from noise has an important impact on the research of rolling bearing condition monitoring and remaining useful life prediction. In response to the above problems, a signal denoising method based on improved dung beetle algorithm optimization variational modal decomposition is proposed, and combined with multi-domain degradation feature extraction method to extract features from the denoised vibration signals. Firstly, using the principle of signal modal decomposition denoising to establish an IDBO-VMD vibration signal denoising model, by decomposing the original signal into multiple sets of intrinsic modal components and residual terms, using constraints to control the smoothness and sparsity of the modal function, and then iteratively optimizing and adjusting the modal function to effectively reduce the influence of noise components in the signal step by step. The denoising method proposed in this paper is verified through vibration simulation signals. Then, the denoised signals are extracted to form a multi-domain hybrid feature set consisting of dimensional time-domain features, dimensionless time-domain features, frequency-domain features, entropy features, and spectral kurtosis correlation features, which describes the degradation process of equipment from multiple dimensions for rolling bearing remaining useful life prediction research.

(2) Research on the construction method of health indicators for rolling bearings based on multi domain degradation features. Firstly, in response to the difficulty of comprehensively and effectively describing the performance degradation state of rolling bearings with a single feature, this paper analyzes the degradation characteristics of multi domain feature indicators and constructs a degradation feature selection criterion. By obtaining sensitive features to the equipment degradation process from the multi domain mixed feature set, effective selection of high-dimensional and multi domain features is achieved. Then, the non-linear isometric feature mapping algorithm (ISOMAP) was used to achieve multi domain degradation feature fusion, reducing the impact of redundant feature information and constructing the health indicators of rolling bearings. Finally, the first prediction time point of the rolling bearing operation process is determined using the fuzzy kernel C-means clustering method for the reduced features.

(3) Research on a remaining useful life prediction model for rolling bearings based on improved time convolutional networks. Firstly, a TCNMABG remaining useful life model is proposed to address the issues of insufficient feature extraction ability and lack of focus on key information in existing models during the degradation process of mechanical equipment. This model combines the advantages of time convolutional neural networks, bidirectional gated recurrent networks, and multi head attention mechanisms, effectively expanding the receptive field of the model and enhancing its long-term information learning ability. Then, the remaining useful life prediction method proposed in this chapter was experimentally validated using the XJTU-SY dataset. The experimental results showed that the method can learn long-term temporal information in the input data at a deeper level and effectively capture key features, enhancing the predictive performance of the model.

(4) Based on a full life cycle data collection experimental platform, the proposed method was experimentally validated. By simulating the actual working environment in the laboratory, a hardware platform for the full life cycle degradation experiment of reducers is built, and a remaining useful life system for rolling bearings is established to achieve the application of research algorithms. The focus is on collecting degradation datasets of horizontal and vertical vibration signals of the rolling bearings at the input end of the reducer. The multi domain degradation feature extraction and selection method, health indicator construction method, and remaining useful life method used in this paper are applied to train, test, optimize parameters, and compare performance of the proposed remaining useful life model. The prediction error and actual performance of the model on the laboratory reducer platform are analyzed, further verifying the effectiveness of the method proposed in this paper.

[1]金晓航,王宇,ZHANG Bin. 工业大数据驱动的故障预测与健康管理[J].计算机集成制造系统, 2022, 28(05):1314-1336.

[2]刘洋. 基于深度学习的滚动轴承剩余使用寿命预测方法研究[D].西安理工大学, 2024.

[3]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, 2021.

[4]李彦夫,韩特. 基于深度学习的工业装备PHM研究综述[J].振动.测试与诊断, 2022, 42(07):837-847+1029.

[5]杨磊磊. 基于振动信号与深度学习的滚动轴承剩余寿命预测研究[D].兰州理工大学, 2023.

[6]孙恒. 旋转机械剩余使用寿命预测研究[D].安徽大学, 2022.

[7]Chen Y, Peng G, Zhu Z, et al. A novel deep learning method based on attention mechanism for bearing remaining useful life prediction[J]. Applied Soft Computing, 2020, 86: 105919.

[8]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.

[9]卞文彬,邓艾东,刘东川等. 基于改进深度残差收缩网络的风电机组滚动轴承故障诊断方法[J].机械工程学报, 2023, 59(12):202-214.

[10]刘惠,刘振宇,郏维强等. 深度学习在装备剩余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):34-72.

[11]钟健飞. 旋转机械零部件剩余寿命预测方法研究[D].西安电子科技大学,2020.

[12]张倩,王涛,赵婕茹等. 基于局部均值分解的布里渊光时域分析传感系统降噪算法[J].光学学报, 2021, 41(13):78-86.

[13]丁锋,秦峰伟. 小波降噪及Hilbert变换在电机轴承故障诊断中的应用[J].电机与控制学报, 2017, 21(06):89-95.

[14]王应晨,段修生. 深度学习融合模型在机械故障诊断中的应用[J].振动.测试与诊断, 2019, 39(06):1271-1276+1363-1364.

[15]曹玲玲,李晶,彭镇等. 基于改进小波阈值降噪的滚动轴承故障诊断方法[J].振动工程学报, 2022, 35(02):454-463.

[16]吴龙文,聂雨亭,张宇鹏等. 基于变分模态分解的自适应滤波降噪方法[J].电子学报, 2021, 49(08):1457-1465.

[17]宋雷,游东东,郑振兴等. 基于经验模态分解的RV减速器运动参数降噪研究[J].振动与冲击, 2022, 41(18):266-272+290.

[18]许同乐,王营博,郑店坤等. 基于LMD-ICA降噪的滚动轴承故障特征提取方法研究[J].北京邮电大学学报, 2017, 40(01):111-116.

[19]孟宗,刘子涵,吕蒙. 基于改进奇异值分解滤波和谱峭度的滚动轴承故障诊断[J].中国机械工程, 2020, 31(20):2420-2428.

[20]孙鑫威,纪爱敏,陈曦晖等. 强噪声背景下动车组轴承微弱故障信号检测[J].电子学报, 2021, 49(11):2217-2224.

[21]孔运,王天杨,褚福磊. 自适应TQWT滤波器算法及其在冲击特征提取中的应用[J].振动与冲击, 2019, 38(11):9-16+23.

[22]张亢,麻云娇,袁志文等. MPA-MMD方法在变转速齿轮箱振动信号特征提取中的应用[J].振动与冲击, 2023, 42(24):127-135.

[23]刘湘楠,赵学智,上官文斌. 强背景噪声振动信号中滚动轴承故障冲击特征提取[J].振动工程学报, 2021, 34(01):202-210.

[24]任海军,韦冲,谭志强等. 基于CEEMDAN-IAWT方法的滚动轴承振动信号降噪[J].振动与冲击, 2023, 42(13):199-207+268.

[25]Tse P W, Wang D. Enhancing the abilities in assessing slurry pumps' performance degradation and estimating their remaining useful lives by using captured vibration signals[J]. Journal of Vibration and Control, 2017, 23(12): 1925-1937.

[26]Chen C, Xu T, Wang G, et al. Railway turnout system RUL prediction based on feature fusion and genetic programming[J]. Measurement, 2020, 151: 107162.

[27]Wu D, Wang H, Liu H, et al. Health monitoring on the spacecraft bearings in high-speed rotating systems by using the clustering fusion of normal acoustic parameters[J]. Applied Sciences, 2019, 9(16): 3246.

[28]陈冬梅,赵思恒,魏承印等. 船舶柴油机状态监测及预测性维护研究及应用[J].中国机械工程, 2022, 33(10):1162-1168.

[29]李天梅,司小胜,刘翔等. 大数据下数模联动的随机退化设备剩余寿命预测技术[J].自动化学报, 2022, 48(09):2119-2141.

[30]王彪,秦勇,贾利民等. 监测数据驱动的城轨列车轴箱轴承剩余寿命预测[J].西南交通大学学报, 2024, 59(01):229-238.

[31]景博,崔展博,孙宏达等. 失效物理与数据驱动融合的燃油泵在线寿命预测[J].仪器仪表学报, 2022, 43(03):68-76.

[32]Wang L, Zhang C, Tao C, et al. Prediction of multiple fatigue crack growth based on modified Paris model with particle filtering framework[J]. Mechanical Systems and Signal Processing, 2023, 190: 110124.

[33]Liang S Y, Li Y, Billington S A, et al. Adaptive prognostics for rotary machineries[J]. Procedia Engineering, 2014, 86: 852-857.

[34]El-Thalji I, Jantunen E. Dynamic modelling of wear evolution in rolling bearings[J]. Tribology International, 2015, 84: 90-99.

[35]李奎,高志成,武一等. 基于统计回归和非线性Wiener过程的交流接触器剩余寿命预测[J].电工技术学报, 2019, 34(19):4058-4070.

[36]徐洲常,王林军,刘洋等. 采用改进回归型支持向量机的滚动轴承剩余寿命预测方法[J].西安交通大学学报, 2022, 56(03):197-205.

[37]Tong S, Yang J, Zong H. A prediction model for complex equipment remaining useful life using gated recurrent unit complex networks[J]. Enterprise Information Systems, 2023, 17(5): 2008515.

[38]Kong Z, Jin X, Xu Z, et al. Spatio-temporal fusion attention: a novel approach for remaining useful life prediction based on graph neural network[J]. IEEE Transactions on Instrumentation and Measurement, 2022, 71: 1-12.

[39]裴洪,胡昌华,司小胜等. 基于机器学习的设备剩余寿命预测方法综述[J].机械工程学报, 2019, 55(08):1-13.

[40]于倩影,李娟,戴洪德等. 基于LASSO变量选择的航空发动机相似性剩余寿命预测[J].航空动力学报, 2023, 38(04):931-938.

[41]Wang H, Zhang X, Guo X, et al. Remaining useful life prediction of bearings based on multiple-feature fusion health indicator and weighted temporal convolution network[J]. Measurement Science and Technology, 2022, 33(10): 104003.

[42]梁伟阁,闫啸家,佘博等. 基于FA-LN-BiGRU的机械设备剩余寿命区间预测方法[J].振动.测试与诊断, 2023, 43(03):513-519+620-621.

[43]Zhang Z, Song W, Li Q. Dual-aspect self-attention based on transformer for remaining useful life prediction[J].IEEE Transactions on Instrumentation and Measurement, 2022, 71: 1-11.

[44]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.

[45]李天梅,司小胜,刘翔等. 大数据下数模联动的随机退化设备剩余寿命预测技术[J].自动化学报, 2022, 48(09):2119-2141.

[46]Liao L, Köttig F. A hybrid framework combining data-driven and model-based methods for system remaining useful life prediction[J]. Applied Soft Computing, 2016, 44: 191-199.

[47]刘建伟,宋志妍. 循环神经网络研究综述[J].控制与决策, 2022, 37(11):2753-2768.

[48]赵志宏,李晴,李春秀. 基于卷积GRU注意力的设备剩余寿命预测[J].振动.测试与诊断, 2022, 42(03):572-579+622.

[49]林亮行,马国政,孙建芳等. 基于VMD-EEMD-LSTM的涂层型关节轴承剩余使用寿命预测方法[J].机械工程学报, 2023, 59(09):125-136.

[50]周哲韬,刘路,宋晓等. 基于Transformer模型的滚动轴承剩余使用寿命预测方法[J].北京航空航天大学学报, 2023, 49(02):430-443.

[51]焦玲玲,陈捷,刘连华. 基于CAE和AGRU的滚动轴承退化趋势预测[J].振动与冲击, 2023, 42(12):109-117.

[52]沈天浩,丁康,黎杰等. 图结构联合时序数据驱动的装备剩余使用寿命预测方法[J].机械工程学报, 2023, 59(12):183-194.

[53]Wang Y, Deng L, Zheng L, et al. Temporal convolutional network with soft thresholding and attention mechanism for machinery prognostics[J]. Journal of Manufacturing Systems, 2021, 60: 512-526.

[54]于广滨,卓识,于军等. 基于InfoLSGAN和AC算法的滚动轴承剩余寿命预测[J].航空动力学报, 2020, 35(06):1212-1221.

[55]郭锐,赵之谦,贾鑫龙等. 基于ANFIS的外啮合齿轮泵寿命预测研究[J].仪器仪表学报, 2020, 41(01):223-232.

[56]任海军,韦冲,谭志强等. 基于CEEMDAN-IAWT方法的滚动轴承振动信号降噪[J].振动与冲击, 2023, 42(13):199-207+268.

[57]赵志宏,李乐豪,杨绍普等.一种振动信号降噪的堆叠降噪自编码器方法[J].振动.测试与诊断, 2022, 42(02):315-321+409.

[58]周涛,汪永超,张栩静等. 基于深度特征融合网络的数模联动随机退化设备剩余寿命预测[J].计算机集成制造系统, 2022, 28(12):3937-3945.

[59]She D, Jia M, Pecht M G. Sparse auto-encoder with regularization method for health indicator construction and remaining useful life prediction of rolling bearing[J]. Measurement Science and Technology, 2020, 31(10): 105005.

[60]余萍,曹洁. 基于图形特征的双输入卷积神经网络风力机轴承剩余寿命预测[J].太阳能学报, 2022, 43(05):343-350.

[61]徐洲常,王林军,刘洋等. 采用改进回归型支持向量机的滚动轴承剩余寿命预测方法[J].西安交通大学学报, 2022, 56(03):197-205.

[62]李天梅,司小胜,张建勋. 多源传感监测线性退化设备数模联动的剩余寿命预测方法[J].航空学报, 2023, 44(08):94-112.

[63]张东东,艾小川,刘畅. 基于相似样本特征提取的装备性能退化研究[J].系统工程与电子技术, 2022, 44(07):2374-2380.

[64]吕明珠,苏晓明,陈长征等. 基于PCA-UPF的风力机轴承剩余寿命预测方法[J].太阳能学报, 2021, 42(02):218-224.

[65]李红贤,韩延,吴敬涛等. 基于ICA包络增强MEMD的滚动轴承故障诊断[J].航空动力学报, 2021, 36(02):405-412.

[66]黄如意,李霁蒲,王震等. 基于多任务学习的装备智能诊断与寿命预测方法[J].中国科学:技术科学, 2022, 52(01):123-137.

[67]王玉静,李少鹏,康守强等. 结合CNN和LSTM的滚动轴承剩余使用寿命预测方法[J].振动.测试与诊断, 2021, 41(03):439-446+617.

[68]Peng C, Wu J, Wang Q, et al. Remaining useful life prediction using dual-channel LSTM with time feature and its difference[J]. Entropy, 2022, 24(12): 1818.

[69]Liu L, Song X, Chen K, et al. An enhanced encoder–decoder framework for bearing remaining useful life prediction[J]. Measurement, 2021, 170: 108753.

[70]Cao X, Lei Z, Li Y, et al. Prediction Method of Equipment Remaining Life Based on Self-Attention Long Short-Term Memory Neural Network[J]. Journal of Shanghai Jiaotong University (Science), 2023, 28(5): 652-664.

[71]Bai S, Kolter J Z, Koltun V. An empirical evaluation of generic convolutional and recurrent networks for sequence modeling[J]. arXiv preprint arXiv:1803.01271, 2018.

[72]张天缘. 基于深度学习的滚动轴承故障诊断和RUL预测方法研究[D].中北大学, 2024.

[73]薛正堂. 滚动轴承失效形式与性能特性分析及滚道表面形貌表征研究[D].合肥工业大学, 2023.

[74]陈彬强,曾念寅,曹新城等. 无监督学习驱动的高端轴承故障智能诊断算法[J].中国科学:技术科学, 2022, 52(01):165-179.

[75]刘靖纳. 数据驱动的滚动轴承故障诊断与性能退化评估方法研究[D].石家庄铁道大学, 2024.

[76]谢雨洁,肖友刚,王田天等. 基于异常检测的轴承退化阶段识别方法[J].中南大学学报(自然科学版), 2022, 53(05):1740-1749.

[77]杨晓晖,张圣昌. 基于多粒度级联孤立森林算法的异常检测模型[J].通信学报, 2019, 40(08):133-142.

[78]彭开香,皮彦婷,焦瑞华等. 航空发动机的健康指标构建与剩余寿命预测[J].控制理论与应用, 2020, 37(04):713-720.

[79]Hu L, Wang W, Ding G. RUL prediction for lithium-ion batteries based on variational mode decomposition and hybrid network model[J]. Signal, Image and Video Processing, 2023, 17(6): 3109-3117.

[80]Xue J, Shen B. Dung beetle optimizer: A new meta-heuristic algorithm for global optimization[J]. The Journal of Supercomputing, 2023, 79(7): 7305-7336.

[81]何成兵,车其祥,徐振华等. 基于参数自寻优变分模态分解的信号降噪方法[J].振动与冲击, 2023, 42(19):283-293.

[82]李得恺,张长胜,杨雪松. 融合多策略改进的黏菌优化算法[J].模式识别与人工智能, 2023, 36(07):647-660.

[83]张艺瀛,金志刚. 求解多峰优化问题的改进布谷鸟算法[J].哈尔滨工业大学学报, 2019, 51(11):89-99.

[84]刘渊博,陈相,刘妤. 基于卷积自编码器和时间卷积网络的轴承性能退化趋势预测[J].振动与冲击, 2023, 42(13):214-225.

[85]顾煜炯,贾子文,尹传涛等. 无量纲指标趋势分析法在风电机组齿轮箱故障预警诊断中的应用[J].振动与冲击, 2017, 36(19):213-220.

[86]莫仁鹏,司小胜,李天梅等. 基于多尺度特征与注意力机制的轴承寿命预测[J].浙江大学学报(工学版), 2022, 56(07):1447-1456.

[87]曾大懿,蒋雨良,邹益胜等. 一种新的轴承寿命预测特征评价指标构建与验证[J].振动与冲击, 2021, 40(22):18-27.

[88]孙丽,赵俊杰,袁春元等. 基于时域和谱峭度特征融合及指数模型的滚动轴承RUL预测[J].机床与液压, 2023, 51(10):203-209.

[89]Cheng Y, Hu K, Wu J, et al. A convolutional neural network based degradation indicator construction and health prognosis using bidirectional long short-term memory network for rolling bearings[J]. Advanced Engineering Informatics, 2021, 48: 101247.

[90]黄宇,冯坤,高俊峰等. 结合LSTM和Self-Attention的滚动轴承剩余使用寿命预测方法[J].振动工程学报, 2023,3 6(06):1744-1753.

[91]李海浪,邹益胜,曾大懿等. 一种基于特征聚类和评价的轴承寿命预测新方法[J].振动与冲击, 2022, 41(05):141-150.

[92]Zhang M, Yin J, Feng T. Lithium Battery SOH Estimation Based on Manifold Learning and LightGBM[J]. Applied Sciences, 2023, 13(11): 6540.

[93]戚晓利,王振亚,吴保林等. 基于ACMPE、ISSL-Isomap和GWO-SVM的行星齿轮箱故障诊断[J].航空动力学报, 2019, 34(04):744-755.

[94]刘璐,邱明,李军星等. 基于LTSA融合降维法的滚动轴承可靠性评估方法[J].航空动力学报, 2021, 36(02):413-420.

[95]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.

[96]Rai A，Upadhyay S H. Intelligent Bearing Performance Degradation Assessment and Remaining Useful Life Prediction Based on Self-organising Map and Support Vector Regression[J]. Joumnal of Mechanical Engineering Science, 2018, 232(6): 1118-1132.

[97]Fawaz H I, Forestier G, Weber J, et al. Deep learning for time series classification: a review[J]. Data Mining and Knowledge Discovery, 2019, 33(4): 917-963.

[98]Zhao B, Lu H, Chen S, et al. Convolutional neural networks for time series classification[J]. Journal of Systems Engineering and Electronics, 2017, 28(1): 162-169.

[99]高辉. 机械设备剩余寿命预测方法的研究[D].山东大学, 2020.

[100]雷亚国,韩天宇,王彪等. XJTU-SY滚动轴承加速寿命试验数据集解读[J].机械工程学报, 2019, 55(16):1-6.

[101]Wang B, Lei Y, Li N, et al. A hybrid prognostics approach for estimating remaining useful life of rolling element bearings[J]. IEEE Transactions on Reliability, 2018, 69(1): 401-412.

[102]叶壮,余建波. 基于多通道一维卷积神经网络特征学习的齿轮箱故障诊断方法[J].振动与冲击, 2020, 39(20):55-66.

[103]吕卫民,孙晨峰,任立坤等. 一种基于TCN-LGBM的航空发动机气路故障诊断方法[J].兵工学报, 2024, 45(01):253-263.

[104]张瑞聪. 基于时间卷积网络的设备运行状态预测方法与实现[D].中国石油大学(华东), 2024.

[105]莫仁鹏,李天梅,司小胜等. 采用残差网络与卷积注意力机制的设备剩余使用寿命预测方法[J].西安交通大学学报, 2022, 56(04):194-202.

[106]陈倩倩. 基于残差网络的滚动轴承故障诊断与剩余寿命预测方法研究[D].青岛理工大学, 2023.

[107]Zhang J, Huang C, Chow M Y, et al. A data-model interactive remaining useful life prediction approach of lithium-ion batteries based on PF-BiGRU-TSAM[J]. IEEE Transactions on Industrial Informatics, 2023.

[108]张哲. 基于时频Transformer的轴承剩余使用寿命预测与故障诊断方法研究[D].燕山大学, 2023.

[109]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.