基于SVD-KDR算法的工业监测数据插补技术

doi:10.3901/JME.2021.02.030

机械工程学报 ›› 2021, Vol. 57 ›› Issue (2): 30-38.doi: 10.3901/JME.2021.02.030

扫码分享

基于SVD-KDR算法的工业监测数据插补技术

陈帅^1,2, 赵明¹, 郭栋², 林京³

1. 西安交通大学机械制造系统工程国家重点实验室西安 710049;
2. 汽车零部件先进制造技术教育部重点实验室(重庆理工大学) 重庆 400054;
3. 北京航空航天大学可靠性与系统工程学院北京 100083

收稿日期:2020-02-20 修回日期:2020-08-06 出版日期:2021-01-20 发布日期:2021-03-15
通讯作者: 赵明(通信作者),男,1984年出生,博士,副教授,博士研究生导师。主要研究方向为机械装备智能感知与健康监测。E-mail:zhaomingxjtu@mail.xjtu.edu.cn
作者简介:陈帅,男,1994年出生。主要研究方向为基于大数据的机械故障诊断。E-mail:chenshuai@stu.xjtu.edu.cn
基金资助:
国家自然科学基金（51875434）和汽车零部件先进制造技术教育部重点实验室开放课题基金（2019KLMT02）资助项目。

Missing Data Imputation Using SVD-KDR Algorithm in Industrial Monitoring Data

CHEN Shuai^1,2, ZHAO Ming¹, GUO Dong², LIN Jing³

1. State Key Laboratory for Manufacturing Systems Engineering, Xi'an Jiaotong University, Xi'an 710049;
2. Key Laboratory of Advance Manufacturing Technology for Automobile Parts(Chongqing University of Technology), Ministry of Education, Chongqing 400054;
3. School of Reliability and Systems Engineering, Beihang University, Beijing 100083

Received:2020-02-20 Revised:2020-08-06 Online:2021-01-20 Published:2021-03-15

摘要/Abstract

摘要： 监测数据的完整性和可用性是工业大数据时代信息提取与知识发掘的前提和基础。然而由于采集中断、传输干扰、存储不当等诸多原因，监测数据的缺失问题在工业场景中频繁发生，严重影响数据价值密度。提出一种基于SVD-KDR的高精度、高鲁棒性缺失数据插补算法。该方法将一维工业数据转换为高维矩阵，弥补了传统方法直接从低维空间插补工业监测数据的维度局限。通过发掘插补过程中非缺失数据的低秩特性，借助奇异值分解理论（Singular value decomposition，SVD）建立了鲁棒性更强的SVD-KDR算法模型，有效减弱了缺失数据对参数估计精度的不利影响。试验结果表明，相比于传统插补算法，所提出算法在高缺失率下仍具有较高插补精度和稳健性。此外，该方法不仅能够有效恢复缺失数据的波形，而且能充分还原原始数据所蕴含的波动信息。提出的SVD-KDR算法可有效解决数据缺失问题，为工业大数据分析提供了数据恢复与信息处理工具。

关键词: 缺失数据插补, 相空间重构, 奇异值分解, 低秩逼近

Abstract: Integrity and availability of data is the basis for information extraction and knowledge discovery in the era of industrial big data. However, data missing frequently occurs in industrial scenarios due to various reasons, such as collection interruption, transmission interference and improper storage, which seriously affects the data value density. In view of this, an interpolation algorithm with high precision and high robustness based on the SVD-KDR algorithm is proposed. The one-dimensional industrial data is converted into a high-dimensional matrix, which makes up for the limitation of the traditional method of interpolating industrial monitoring data directly from the low-dimensional space. By exploring the low-rank characteristics of non-missing data in the interpolation process, a more robust SVD-KDR algorithm model is established with the help of singular value decomposition theory, which effectively reduces the adverse effect of missing data on parameter estimation accuracy. Experimental results show that the proposed algorithm has higher accuracy and robustness at high missing rates compared with the traditional interpolation algorithm. In addition, the proposed method can not only effectively recover the waveform of missing data, but also restore the fluctuating information behind the original data. The SVD-KDR algorithm can effectively solve the problem of missing data, and provides data recovery and information processing tools for industrial big data analysis.

Key words: missing data imputation, phase space reconstruction, singular value decomposition, low-rank approximation

中图分类号:

TH17

陈帅, 赵明, 郭栋, 林京. 基于SVD-KDR算法的工业监测数据插补技术[J]. 机械工程学报, 2021, 57(2): 30-38.

CHEN Shuai, ZHAO Ming, GUO Dong, LIN Jing. Missing Data Imputation Using SVD-KDR Algorithm in Industrial Monitoring Data[J]. Journal of Mechanical Engineering, 2021, 57(2): 30-38.

参考文献

[1] GARCÍA-LAENCINA P J,SANCHO-GÓMEZ J L,FIGUEIRAS-VIDAL A R,et al. K nearest neighbours with mutual information for simultaneous classification and missing data imputation[J]. Neurocomputing,2009,72(7):1483-1493.
[2] SIDI Y,HAREL O. The treatment of incomplete data:Reporting,analysis,reproducibility,and replicability[J]. Soc. Sci. Med.,2018,209:169-173.
[3] ZHU J,GE Z,SONG Z,et al. Review and big data perspectives on robust data mining approaches for industrial process modeling with outliers and missing data[J]. Annual Reviews in Control,2018,46:107-133.
[4] CHENG C H,CHAN C P,SHEU Y J. A novel purity-based k nearest neighbors imputation method and its application in financial distress prediction[J]. Engineering Applications of Artificial Intelligence,2019,81:283-299.
[5] SAHA S,GHOSH A,BANDOPADHYAY S,et al. Missing value imputation in DNA microarray gene expression data:A comparative study of an improved collaborative filtering method with decision tree based approach[J]. International Journal of Computational Science and Engineering,2019,18(2):130-139.
[6] ANJOMSHOAA A,SALMANZADEH M. Filling missing meteorological data in heating and cooling seasons separately[J]. International Journal of Climatology,2019,39(2):701-710.
[7] CHEN X Y,HE Z C,CHEN Y X,et al. Missing traffic data imputation and pattern discovery with a Bayesian augmented tensor factorization model[J]. Transportation Research Part C-Emerging Technologies,2019,104:66-77.
[8] FARHANGFAR A,KURGAN L,DY J. Impact of imputation of missing values on classification error for discrete data[J]. Pattern Recognition,2008,41(12):3692-3705.
[9] 王强,郭广礼,査剑锋. 基于BP神经网络的岩移监测站缺失点数据插补[J]. 现代矿业,2010,26(2):73-75. WANG Qiang,GUO Guangli,ZHA Jianfeng. Interpolation of missing point data in rock movement monitoring station based on BP neural network[J]. Modern Mining,2010,26(2):73-75.
[10] RAHMAN M G,ISLAM M Z. Missing value imputation using decision trees and decision forests by splitting and merging records:Two novel techniques[J]. Knowledge-Based Systems,2013,53:51-65.
[11] 孟杰,李春林. 基于随机森林模型的分类数据缺失值插补[J]. 统计与信息论坛,2014,29(9):86-90. MENG Jie,LI Chunlin. Imputation of missing values for categorical data based on random forest model[J]. Statistics and Information Forum,2014,29(9):86-90.
[12] FOLGUERA L,ZUPAN J,CICERONE D,et al. Self-organizing maps for imputation of missing data in incomplete data matrices[J]. Chemometrics and Intelligent Laboratory Systems,2015,143:146-151.
[13] TSAI C F,LI M L,LIN W C. A class center based approach for missing value imputation[J]. Knowledge-Based Systems,2018,151:124-135.
[14] GUO Z J,WAN Y M,YE H. A data imputation method for multivariate time series based on generative adversarial network[J]. Neurocomputing,2019,360:185-197.
[15] ZHAO M,JIA X D. A novel strategy for signal denoising using reweighted SVD and its applications to weak fault feature enhancement of rotating machinery[J]. Mechanical Systems and Signal Processing,2017,94:129-147.
[16] PACKARD N H,CRUTCHFIELD J P,FARMER J D,et al. Geometry from a time-series[J]. Physical Review Letters,1980,45(9):712-716.
[17] XU B B,JACQUIR S,LAURENT G,et al. Analysis of an experimental model of in vitro cardiac tissue using phase space reconstruction[J]. Biomedical Signal Processing and Control,2014,13:313-326.
[18] LI H W,LIU J P,LI T,et al. Analysis of dynamic of two-phase flow in small channel based on phase space reconstruction combined with data reduction sub-frequency band wavelet[J]. Chinese Journal of Chemical Engineering,2015,23(6):1017-1026.
[19] CAO L Y. Practical method for determining the minimum embedding dimension of a scalar time series[J]. Physica D,1997,110(1-2):43-50.
[20] 鄢小安,贾民平. 基于改进奇异谱分解的形态学解调方法及其在滚动轴承故障诊断中的应用[J]. 机械工程学报,2017,53(7):104-112. YAN Xiaoan,JIA Minping. Morphological demodulation method based on improved singular spectrum decomposition and its application in rolling bearing fault diagnosis[J]. Journal of Mechanical Engineering,2017,53(7):104-112.
[21] FOLCH-FORTUNY A,ARTEAGA F,FERRER A. PCA model building with missing data:New proposals and a comparative study[J]. Chemometrics and Intelligent Laboratory Systems,2015,146:77-88.
[22] 雷亚国,贾峰,孔德同,等. 大数据下机械智能故障诊断的机遇与挑战[J]. 机械工程学报,2018,54(5):94-104. LEI Yaguo,JIA Feng,KONG Detong,et al. Opportunities and challenges of mechanical intelligent fault diagnosis under big data[J]. Journal of Mechanical Engineering,2018,54(5):94-104.
[23] ZHAO M,LIN J. Health assessment of rotating machinery using a rotary encoder[J]. IEEE Transactions on Industrial Electronics,2018,65(3):2548-2556.

基于SVD-KDR算法的工业监测数据插补技术

Missing Data Imputation Using SVD-KDR Algorithm in Industrial Monitoring Data

PDF

可视化

摘要/Abstract

引用本文

使用本文

参考文献

相关文章 15

编辑推荐

Metrics

本文评价

[1]	董绍江, 裴雪武, 吴文亮, 汤宝平, 赵兴新. 基于多层降噪技术及改进卷积神经网络的滚动轴承故障诊断方法[J]. 机械工程学报, 2021, 57(1): 148-156.
[2]	夏光, 张洋, 魏振亚, 唐希雯, 陈无畏. 驾驶员驾驶习惯识别与车辆自动变速换档修正[J]. 机械工程学报, 2020, 56(20): 120-133.
[3]	王华庆, 任帮月, 宋浏阳, 董方, 王梦阳. 基于终止准则改进K-SVD字典学习的稀疏表示特征增强方法[J]. 机械工程学报, 2019, 55(7): 35-43.
[4]	赵学智, 邵啟鹏, 叶邦彦, 陈统坚. 奇异值分解中考虑频率因素的矩阵维数[J]. 机械工程学报, 2019, 55(16): 7-16.
[5]	程礼, 杨武奎, 梁涛, 文璧, 姚东野. 基于声模态的压气机/风扇气路故障诊断[J]. 机械工程学报, 2019, 55(13): 38-44.
[6]	胡超凡, 王衍学. 基于张量分解的滚动轴承复合故障多通道信号降噪方法研究[J]. 机械工程学报, 2019, 55(12): 50-57.
[7]	赵宏晨,王刚,刘晓明. 多种群遗传算法与截断奇异值分解法在开关电弧反演中的应用[J]. 电气工程学报, 2018, 13(6): 14-19.
[8]	张铁, 叶景杨, 刘晓刚. 面向机器人砂带打磨的加权手眼标定算法[J]. 机械工程学报, 2018, 54(17): 142-148.
[9]	赵学智, 陈统坚, 叶邦彦. 变结构SVD算法及其在信号分离中的应用[J]. 机械工程学报, 2017, 53(22): 11-21.
[10]	黄涛, 杨开明, 杨进, 朱煜, 胡楚雄. 柔性结构的多输入多输出运动系统辨识方法[J]. 机械工程学报, 2016, 52(11): 42-49.
[11]	王建国, 李健, 万旭东. 基于奇异值分解和局域均值分解的滚动轴承故障特征提取方法[J]. 机械工程学报, 2015, 51(3): 104-110.
[12]	曾鸣, 杨宇, 郑近德, 程军圣. u-SVD降噪算法及其在齿轮故障诊断中的应用[J]. 机械工程学报, 2015, 51(3): 95-103.
[13]	赵学智, 陈统坚, 叶邦彦. 奇异值分解对连续Morlet小波变换的压缩和提纯[J]. 机械工程学报, 2015, 51(16): 57-70.
[14]	赵学智;叶邦彦;陈统坚. 基于小波—奇异值分解差分谱的弱故障特征提取方法[J]. , 2012, 48(7): 37-48.
[15]	林嘉睿;邾继贵;郭寅;郭庭航;叶声华. 现场大空间测量中精密三维坐标控制网的建立[J]. , 2012, 48(4): 6-11.