[1]秦玉峰,孙敏奇,赵建印.基于最大均值差异的机械设备剩余寿命预测方法研究[J].机械与电子,2026,44(04):68-73.
　QIN Yufeng,SUN Minqi,ZHAO Jianyin.Remaining Useful Life Prediction Method for Mechanical Equipment Based on Maximum Mean Discrepancy[J].Machinery & Electronics,2026,44(04):68-73.
点击复制

基于最大均值差异的机械设备剩余寿命预测方法研究()
分享到：

《机械与电子》[ISSN:1001-2257/CN:52-1052/TH]

卷:
44
期数:
2026年04期
页码:
68-73
栏目:
智能检测
出版日期:
2026-04-27

文章信息/Info

Title:
Remaining Useful Life Prediction Method for Mechanical Equipment Based on Maximum Mean Discrepancy
文章编号:
1001-2257 ( 2026 ) 04-0068-06
作者:
秦玉峰; 孙敏奇; 赵建印
海军航空大学,山东 烟台 264001
Author(s):
QIN Yufeng ;  SUN Minqi ;  ZHAO Jianyin
( Naval Aviation University , Yantai 264001 , China )
关键词:
最大均值差异; 剩余寿命预测; 主成分分析; 相似性度量
Keywords:
maximum mean discrepancy ;  remaining useful life prediction ;  principal component analysis ;  similarity measurement
分类号:
TH17
文献标志码:
A
摘要:
针对现有深度学习网络模型存在的可解释性差等问题,提出了一种基于最大均值差异( MMD )的机械设备剩余寿命预测方法。首先,利用主成分分析( PCA )方法提取高维传感器数据的第一主成分作为设备性能退化健康指标;然后,基于 MMD 衡量退化健康指标之间的相似性,构建基于 MMD 的剩余寿命预测模型;最后,采用 C-MAPSS 数据集对所提方法进行验证,所得预测结果的 MAE 为 18.11 , RMSE 为24.21 ,与其他常见距离度量方法相比,所提方法得到的预测结果最佳。
Abstract:
To address issues such as poor interpretability in existing deep learning network models , a method for predicting the Remaining Useful Life ( RUL ) of mechanical equipment based on Maximum Mean Discrepancy ( MMD ) is proposed.Firstly , the first principal component of high-dimensional sensor data is extracted using Principal Component Analysis ( PCA ) to serve as a health indicator , reflecting equipment performance degradation.Subsequently , the similarity between degradation health indicators is measured based on MMD , and an MMD based RUL prediction model is constructed.Finally , the proposed method is validated using the C MAPSS dataset.The prediction results yield a Mean Absolute Error ( MAE ) of 18.11 and a Root Mean Square Error ( RMSE ) of 24.21.Compared with other common distance metric methods , the proposed method achieves superior prediction performance.

参考文献/References:

[ 1 ] NIE G C , ZHANG Z W , JIAO Z H , et al.A novel intelligent bearing fault diagnosis method based on image enhancement and improved convolutional neural network [ J ] .Measurement , 2025 , 242 ( part D ): 116148.

[ 2 ] HUANG D L , SU X D , YANG J H , et al.An improved dual-channelCNN-BILSTM fusion attention model for fault diagnosis of aero-engine bearings [ J ] . Measurement , 2025 , 253 ( C ): 117761.
[ 3 ] OSPINA-D?VILA Y M , OROZCO-ALZATE M. Dissimilarity-vector spaces based on dynamic time warpings of spectral / time frequency information for structural health monitoring [ J ] .Computers and structures , 2022 , 263 : 106754.
[ 4 ] GAO S Z , XU L T , ZHANG Y M , et al.Rolling bearing fault diagnosis based on SSA optimized self-adaptive DBN [ J ] .ISA Transactions , 2022 , 128 ( B ): 485-502.
[ 5 ] SIM Y S , LEE C K , HWANG J S , et al.AI-based remaining useful life prediction for transmission systems : integrating operating conditions with TimeGAN and CNN LSTM networks [ J ] .Electric power systems research , 2025 , 238 : 111151.
[ 6 ] ZHU T , CHEN Z , ZHOU D , et al.Adaptive staged remaining useful life prediction of roller in a hot strip mill based on multi-scale LSTM with multi-head attention [ J ] .Reliability engineering and system safety , 2024 , 248 : 110161.
[ 7 ] CHEN X D , LI K , WANG S F , et al.A hybrid prognostic approach combined with deep Bayesian transformer and enhanced particle filter for remaining useful life prediction of bearings [ J ] .Measurement , 2025 , 252 : 117184.
[ 8 ] LI X C , XU S Q , YANG Y J , et al.Spherical-dynamic time warping a new method for similarity-based remaining useful life prediction [ J ] .Expert systems with applications , 2024 , 238 ( B ): 121913.
[ 9 ] 秦玉峰,史贤俊 . 基于 MMD 的故障可诊断性定量评价方法[ J ] . 控制与决策, 2023 , 38 ( 10 ): 2529-2533.
[ 10 ] QIN Y F , SHI X J , ZHAO L.Quantitative evaluation of fault diagnosability based on maximum mean discrepancy [ C ] ∥2022 41st Chinese Control Conference ( CCC ),2022 : 3907-3911.
[ 11 ] SHI X J , QIN Y F , ZHAO L.Optimal test point placement based on fault diagnosability quantitative evaluation [ J ] .IEEE Access , 2022 , 10 : 74495-74507.
[ 12 ] QIAN Q , WANG Y , ZHANG T S , et al.Maximum mean square discrepancy : a new discrepancy representation metric for mechanical fault transfer diagnosis [ J ] .Knowledge based systems , 2023 , 276 : 110748.
[ 13 ] FENG H , WANG N , TANG J.Deep Weibull hashing with maximum mean discrepancy quantization for image retrieval [ J ] .Neurocomputing , 2021 , 464 : 95-106.
[ 14 ] WU J , WANG S P , SUN J.AMMD : attentive maximum mean discrepancy for few-shot image classification [ J ] .Pattern recognition , 2024 , 155 : 110680.
[ 15 ] LI J M , YE Z D , GAO J , et al.Fault transfer diagnosis of rolling bearings across different devices via multi domain information fusion and multi-kernel maximum mean discrepancy [ J ] .Applied soft computing , 2024 , 159 : 111620.
[ 16 ] BASHA S A H , KSHIRSAGAR P R , RAO P S , et al. PCA F SHCNNet : principal component analysis fused-shepard convolutional neural networks for lung cancer detection and severity level classification [ J ] .Bio- medical signal processing and control , 2025 , 107 : 107843.
[ 17 ] REN Z L , JIANG Y C , YANG X B.Learnable faster kernel-PCA for nonlinear fault detection : deep autoencoder-based-realization [ J ] .Journal of industrial information integration , 2024 , 40 : 100622.
[ 18 ] 侯佟泽 . 基于多元回归的过程监测研究及其工业应用[ D ] . 北京:北京化工大学,2024.
[ 19 ] 唐滔 . 基于非线性特征相域增强与交互分析的轴承故障诊断方法[ D ] . 重庆:重庆大学,2019.
[ 20 ] 赵洪利,张奔,张青 . 基于工况聚类和残差自注意力的发动机剩余使用寿命预测[ J ] . 航空科学技术 .2023 ,34 ( 4 ): 31-40.

相似文献/References:

[1]王 正１,２.基于机器学习的新能源汽车电池剩余寿命预测[J].机械与电子,2019,(12):9.
　WANG Zheng,Remaining UsefulLife Prediction of New Energy Vehicle Battery Based on Machine Learning[J].Machinery & Electronics,2019,(04):9.
[2]王 正１,２.基于DAE-HTPF的新能源汽车电池剩余寿命预测[J].机械与电子,2020,(03):3.
　WANG Zheng,RemainingLife Prediction of New Energy Vehicle Battery Based on DAE-HTPF[J].Machinery & Electronics,2020,(04):3.
[3]刘东林,秦玉焘,麻浩军,等.基于马尔科夫链的车桥齿轮剩余寿命预测研究[J].机械与电子,2023,41(03):76.
　LIU Donglin,QIN Yutao,MA Haojun,et al.Research on Residual Life Prediction of Axle Gear Based on Markov Chain[J].Machinery & Electronics,2023,41(04):76.

备注/Memo

备注/Memo:
收稿日期: 2025-11-24
作者简介:秦玉峰 ( 1995- ),男,黑龙江大庆人,博士,讲师,研究方向为装备故障预测与健康管理技术;孙敏奇 ( 1995- ),男,江苏溧阳人,硕士研究生,研究方向为装备故障预测与健康管理技术,通信作者, E-mai1 : js1y _ smq@163.com 。
更新日期/Last Update: 2026-08-20