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

莫仁鹏; 李天梅; 司小胜; 朱旭

doi:10.7652/xjtuxb202204021

您当前的位置：

首页 >

文章列表页 >

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

更新时间：2025-07-09

- 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
- Remaining Useful Life Prediction for Equipment Using Residual Network and Convolutional Attention Mechanism
- 西安交通大学学报 2022年56卷第4期页码：194-202
- 作者机构：
  
  火箭军工程大学导弹工程学院,西安,710025
- 作者简介：
  
  : 2021-07-20。作者简介: 莫仁鹏(1997—),男,硕士生
  司小胜(通信作者),男,教授,博士生导师。基金项目: 国家自然科学基金优秀青年科学基金资助项目(61922089)
- 基金信息：
  
  : 2021-07-20。作者简介: 莫仁鹏(1997—),男,硕士生;司小胜(通信作者),男,教授,博士生导师。基金项目: 国家自然科学基金优秀青年科学基金资助项目(61922089);国家自然科学基金资助项目(61773386,62073336)
- DOI：10.7652/xjtuxb202204021
  中图分类号： TH17
- 网络首发：2022-04-10，
  
  纸质出版：2022
- 稿件说明：
移动端阅览
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(4):194-202.

MO Renpeng, LI Tianmei, SI Xiaosheng, et al. Remaining Useful Life Prediction for Equipment Using Residual Network and Convolutional Attention Mechanism[J]. 2022, 56(4): 194-202.
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(4):194-202. DOI： 10.7652/xjtuxb202204021.

MO Renpeng, LI Tianmei, SI Xiaosheng, et al. Remaining Useful Life Prediction for Equipment Using Residual Network and Convolutional Attention Mechanism[J]. 2022, 56(4): 194-202. DOI： 10.7652/xjtuxb202204021.

摘要

针对以往剩余使用寿命(RUL)预测方法均等对待深层特征中具有不同重要性的空间域和通道域特征

导致大量的计算资源浪费在不重要的特征上

进而造成RUL预测值偏差过大、不能对失效设备进行及时的维护以避免潜在的安全隐患的问题

提出一种采用残差网络与卷积注意力机制的端到端的RUL预测方法。该方法以卷积层和池化层对原始监测信号进行浅层特征提取与压缩; 利用堆叠残差模块在学习深层特征的同时

缓解梯度弥散以及网络退化现象的发生; 由卷积注意力模块对设备的深层退化特征进行加权赋值

分别在其空间维度上和通道维度上强化更重要的特征并抑制相对不重要的特征

使网络的注意力集中在对RUL预测任务更关键的信息上; 将加权后的特征输入到全连接网络中映射得到RUL预测值。通过PHM2012轴承数据集进行了实验验证

实验结果表明

卷积注意力和残差结构皆对改善模型的预测性能有着积极的作用

所提方法在测试轴承上的均方根误差和平均绝对误差分别为0.107 9和0.083 1

远低于其他对比方法。

Abstract

Regarding the previous remaining life prediction methods

the spatial domain and channel domain features of different importance in the deep features are often treated equally

causing a lot of computing resources to be wasted on unimportant features

which leads to excessive deviations in RUL predictions and failure to perform timely maintenance on failed equipment to avoid potential safety hazards

an end-to-end RUL prediction method using residual network and convolutional attention mechanism is proposed. This method first uses the convolutional layer and the pooling layer to extract and compress the shallow features of the original monitoring signal. Then uses the stacked residual module to learn deep features while mitigating the occurrence of gradient dispersion and network degradation Next

the convolutional attention module weights the deep degradation features of the device to strengthen the more important features and suppress the relatively unimportant features in its spatial dimension and channel dimension

so that the network's attention is focused on information that is more critical to RUL prediction task. Finally

the weighted features are input into the fully connected network and mapped to obtain the RUL prediction value. The experimental verification is carried out through the PHM2012 bearing data set. The experimental results show that both the convolutional attention and the residual structure have a positive effect on improving the prediction performance of the model. In addition

the root mean square error and average absolute error of the proposed method on the test bearing are 0.107 9 and 0.083 1

which are much lower than other comparison methods.

关键词

Keywords

references

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(8): 1-13.

PEI Hong, HU Changhua, SI Xiaosheng, et al. Review of machine learning based remaining useful life prediction methods for equipment [J]. Journal of Mechanical Engineering, 2019, 55(8): 1-13.

MOGHADDASS R, ZUO M J. An integrated framework for online diagnostic and prognostic health monitoring using a multistate deterioration process [J]. Reliability Engineering System Safety, 2014, 124: 92-104.

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/OL]. 自动化学报 [2021-06-25]. https:∥doi.org/10.16383/j.aas.c201 068.

LI Tianmei, SI Xiaosheng, LIU Xiang, et al. Data-model interactive remaining useful life prediction technologies for stochastic degrading devices with big data [J/OL]. Acta Automatica Sinica [2021-06-25]. https: ∥doi.org/10.16383/j.aas.c201068.

董永刚, 仪帅, 黄鑫磊, 等. 重载列车紧急制动过程车轮踏面疲劳裂纹萌生寿命预测 [J]. 中国铁道科学, 2021, 42(5): 123-131.

DONG Yonggang, YI Shuai, HUANG Xinlei, et al. Prediction of fatigue crack initiation life of wheel tread during emergency braking of heavy haul train [J]. China Railway Science, 2021, 42(5): 123-131.

HUANG Chenggeng, HUANG Hongzhong, LI Yanfeng. A bidirectional LSTM prognostics method under multiple operational conditions [J]. IEEE Transactions on Industrial Electronics, 2019, 66(11): 8792-8802.

孙秋云, 朱才朝, 樊志鑫, 等. 误差对风电齿轮箱轴承疲劳寿命影响 [J]. 太阳能学报, 2021, 42(7): 308-315.

SUN Qiuyun, ZHU Caichao, FAN Zhixin, et al. Effects of error on fatigue life of bearing for wind turbine gearbox [J]. Acta Energiae Solaris Sinica, 2021, 42(7): 308-315.

孙振铎, 侯东勃, 李伟, 等. 基于裂纹萌生和扩展的渗氮钢疲劳寿命预测 [J]. 兵器材料科学与工程, 2021, 44(5): 76-82.

SUN Zhenduo, HOU Dongbo, LI Wei, et al. Prediction of fatigue life of nitrided steel based on crack initiation and propagation [J]. Ordnance Material Science and Engineering, 2021, 44(5): 76-82.

张军, 张犁, 成瑜. IGBT模块寿命评估研究综述 [J]. 电工技术学报, 2021, 36(12): 2560-2575.

ZHANG Jun, ZHANG Li, CHENG Yu. Review of the lifetime evaluation for the IGBT module [J]. Transactions of China Electrotechnical Society, 2021, 36(12): 2560-2575.

苗强, 蒋京, 张恒, 等. 工业大数据背景下的航空智能发动机: 机遇与挑战 [J]. 仪器仪表学报, 2019, 40(7): 1-12.

MIAO Qiang, JIANG Jing, ZHANG Heng, et al. Development of aviation intelligent engine under industrial big data: chances and challenges [J]. Chinese Journal of Scientific Instrument, 2019, 40(7): 1-12.

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(4): 713-720.

PENG Kaixiang, PI Yanting, JIAO Ruihua, et al. Health indicator construction and remaining useful life prediction for aircraft engine [J]. Control Theory Applications, 2020, 37(4): 713-720.

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(1): 34-52.

LIU Hui, LIU Zhenyu, JIA Weiqiang, et al. Current research and challenges of deep learning for equipment remaining useful life prediction [J]. Computer Integrated Manufacturing Systems, 2021, 27(1): 34-52.

张钢, 田福庆, 梁伟阁, 等. 基于多尺度AlexNet网络的健康因子构建方法 [J]. 系统工程与电子技术, 2020, 42(1): 245-252.

ZHANG Gang, TIAN Fuqing, LIANG Weige, et al. Construction method of bearing health indicator based on multi-scale AlexNet network [J]. Systems Engineering and Electronics, 2020, 42(1): 245-252.

ZHANG Xiaoyong, XIAO Pengcheng, YANG Yingze, et al. Remaining useful life estimation using CNN-XGB with extended time window [J]. IEEE Access, 2019, 7: 154386-154397.

ZHU Jun, CHEN Nan, PENG Weiwen. Estimation of bearing remaining useful life based on multiscale convolutional neural network [J]. IEEE Transactions on Industrial Electronics, 2019, 66(4): 3208-3216.

HE Kaiming, ZHANG Xiangyu, REN Shaoqing, et al. Deep residual learning for image recognition [C]∥Proceedings of the 2016 IEEE Conference on Computer Vision and Pattern Recognition. Piscataway, NJ, USA: IEEE, 2016: 770-778.

HE Kaiming, ZHANG Xiangyu, REN Shaoqing, et al. Identity mappings in deep residual networks [C]∥Proceedings of the European Conference on Computer Vision. Berlin, Germany: Springer International Publishing, 2016: 630-645.

HOWARD A G, ZHU Menglong, CHEN Bo, et al. MobileNets: efficient convolutional neural networks for mobile vision applications [EB/OL]. [2021-06-01]. https: ∥arxiv.org/abs/1704. 04861.

YANG Zhibo, ZHANG Junpeng, ZHAO Zhibin, et al. Interpreting network knowledge with attention mechanism for bearing fault diagnosis [J]. Applied Soft Computing, 2020, 97: 106829.

方军, 管业鹏. 基于双编码器的会话型推荐模型 [J]. 西安交通大学学报, 2021, 55(8): 166-174.

FANG Jun, GUAN Yepeng. A session recommendation model based on dual encoders [J]. Journal of Xi'an Jiaotong University, 2021, 55(8): 166-174.

曹仰杰, 李昊, 段鹏松, 等. AHNNet: 融合注意力机制的行为识别混合神经网络模型 [J]. 西安交通大学学报, 2021, 55(5): 123-132.

CAO Yangjie, LI Hao, DUAN Pengsong, et al. AHNNet: human activity recognition based on hybrid neural network combining attention mechanism [J]. Journal of Xi'an Jiaotong University, 2021, 55(5): 123-132.

WOO S, PARK J, LEE J Y, et al. CBAM: convolutional block attention module [C]∥Proceedings of the European Conference on Computer Vision. Berlin, Germany: Springer International Publishing, 2018: 3-19.

王美华, 吴振鑫, 周祖光. 基于注意力改进CBAM的农作物病虫害细粒度识别研究 [J]. 农业机械学报, 2021, 52(4): 239-247.

WANG Meihua, WU Zhenxin, ZHOU Zuguang. Fine-grained identification research of crop pests and diseases based on improved CBAM via attention [J]. Transactions of the Chinese Society for Agricultural Machinery, 2021, 52(4): 239-247.

NECTOUX P, GOURIVEAU R, MEDJAHER K, et al. PRONOSTIA: an experimental platform for bearings accelerated degradation tests [C]∥Proceedings of the IEEE International Conference on Prognostics and Health Management. Piscataway, NJ, USA: IEEE, 2012: 1-8.

REN Lei, CUI Jin, SUN Yaqiang, et al. Multi-bearing remaining useful life collaborative prediction: a deep learning approach [J]. Journal of Manufacturing Systems, 2017, 43(2): 248-256.

GUO Liang, LEI Yaguo, LI Naipeng, et al. Deep convolution feature learning for health indicator construction of bearings [C]∥Proceedings of the 2017 Prognostics and System Health Management Conference. Piscataway, NJ, USA: IEEE, 2017: 1-6.

张继冬, 邹益胜, 邓佳林, 等. 基于全卷积层神经网络的轴承剩余寿命预测 [J]. 中国机械工程, 2019, 30(18): 2231-2235.

ZHANG Jidong, ZOU Yisheng, DENG Jialin, et al. Bearing remaining life prediction based on full convolutional layer neural networks [J]. China Mechanical Engineering, 2019, 30(18): 2231-2235.

浏览量

下载量

CSCD

文章被引用时，请邮件提醒。

提交

工具集

关联资源

频域注意力机制下的癫痫脑电信号分类

融合物理信息与深度算子网络的变压器绕组热-流场快速计算方法

语义区分和子图整合驱动的知识图谱表示学习方法

采用机器学习的大规模并行程序I/O性能建模与预测方法

隧道内燃料电池汽车氢气泄漏扩散特性及影响因素