面向不平衡数据的云模型旋转机械故障识别方法

PDF(1305 KB)

振动与冲击 ›› 2022, Vol. 41 ›› Issue (22) : 70-78.

论文

赵楠，赵荣珍

作者信息 +

Rotating machinery fault identification method based on the cloud model confronting unbalanced data

ZHAO Nan，ZHAO Rongzhen

Author information +

文章历史 +

摘要

因正常数据丰富、故障数据匮乏而引起的数据不平衡，已经成为工业大数据智能决策技术面临的关键问题之一。因此针对机械设备故障诊断研究中经常遇到的不平衡数据集中少数样本类别识别精度偏低的问题，提出一种基于云模型的集成学习方法并将其用于旋转机械不平衡数据的模式识别。该方法首先将提取出的轴承故障特征数据集通过ReliefF算法计算各特征的权重，依据特征权重值降序排列的结果提取出权重趋大的特征构成低维特征集，并将低维特征集划分为不平衡训练集、测试集两部分；其次通过云模型理论中的正向云发生器、逆向云发生器对低维特征集中各个特征分别绘制云图，得到单一特征下各状态的训练数据与测试数据云图；然后通过距离公式判别与待测样本距离最近的训练数据云图，判断出单一特征下待测样本的类别；最后通过集成学习方法将各个特征下的识别结果进行整合，以相对多数投票法识别出待测样本的所属类别结果。与传统的BP神经网络、支持向量机两种分类器进行对比的实验表明，本方法不仅对不平衡数据的待测样本识别精度较高，而且具有一定的泛化性能。
关键词：故障识别；不平衡数据；云模型；ReliefF算法；集成学习

Abstract

The data imbalance caused by the abundant normal data and the lack of fault data has become one of the key problems faced by the industrial big data intelligent decision-making technology. Therefore, in order to solve the problem that the recognition accuracy of a small number of sample categories in imbalanced data sets is often encountered in the research of mechanical equipment fault diagnosis, an ensemble learning method based on cloud model is proposed and used for pattern recognition of unbalanced data of rotating machinery. First, The method uses the ReliefF algorithm to calculate the weight of each feature from the extracted bearing fault feature data set. According to the result of the feature weight value in descending order, features with higher weights are extracted to form a low-dimensional feature set, and the low-dimensional feature set is divided into unbalanced training set and test set. Secondly, through the forward cloud generator and reverse cloud generator in the cloud model theory, each feature in the low-dimensional feature set is drawn separately to obtain the training data and test data cloud diagrams of each state under a single feature. Then, the distance formula is used to discriminate the training data cloud image with the closest distance to the sample to be tested, and the category of the sample to be tested under a single feature is judged. Finally, the recognition results under each feature are integrated through the ensemble learning method, and the relative majority voting method is used to identify the sample to be tested. Method of this article compared with the traditional BP neural network and support vector machine, the experiment shows that this method not only has high recognition accuracy for the unbalanced data to be tested, but also has a certain generalization performance.
Key words: Fault identification; Unbalanced data; Cloud model; ReliefF algorithm; Ensemble learning

导出引用

赵楠，赵荣珍. 面向不平衡数据的云模型旋转机械故障识别方法[J]. 振动与冲击, 2022, 41(22): 70-78

ZHAO Nan，ZHAO Rongzhen. Rotating machinery fault identification method based on the cloud model confronting unbalanced data[J]. Journal of Vibration and Shock, 2022, 41(22): 70-78

参考文献

[1] Liu R, Yang B, Enrico Z, et al. Artificial intelligence for fault diagnosis of rotating machinery:A review[J]. Mechanical Systems and Signal Processing, 2018, 108: 33-47.
[2] 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(05): 94-104.
LIE Ya-guo, JIA Feng, KONG De-tong, et al. Opportunities and Challenges of Machinery Intelligent Fault Diagnosis in Big Data Era[J]. Journal of mechanical engineering, 2018,54(05): 94-104.
[3] 吴耀春, 赵荣珍, 靳伍银. EWT与加权多邻域粗糙集结合的旋转机械故障特征提取方法[J]. 振动与冲击, 2019, 38(24): 235-242.
     WU Yao-chun, ZHAO Rong-zhen, JIN Wu-yin. Fault feature extraction of rotating machinery based on EWT and a weighted multi neighborhood rough set[J]. Journal of vibration and shock, 2019, 38(24): 235-242.
[4] 石明宽, 赵荣珍. 基于局部质心均值最小距离鉴别投影的旋转机械故障数据降维分析研究[J]. 振动工程学报, 2021, 34(02): 421-430.
     SHI Ming-kuan, ZHAO Rong-zhen. Dimensional reduction analysis of rotating machinery fault data based on local centroid mean minimum-distance discriminant projection[J]. Journal of vibration engineering, 2021, 34(02): 421-430.
[5] 马森财. 旋转机械典型故障辨识及分类可视化问题研究[D]. 兰州: 兰州理工大学, 2020.
     MA Sen-cai. Research on Pattern Identification and Classification Visualization of Typical Faults from Rotating Machinery[D]. LanZhou: Lanzhou University of Technology, 2020.
[6] 李艳霞, 柴毅, 胡友强, 等. 不平衡数据分类方法综述[J]. 控制与决策, 2019, 34(04): 673-688.
     LI Yan-xia, CHAI Yi, HU You-qiang, et al. Review of imbalanced data classification methods[J]. Control and decision, 2019, 34(04): 673-688.
[7] Guo H, Li Y, Jennifer S, et al. Learning from class-imbalanced data: Review of methods and applications[J]. Expert Systems with Applications, 2017, 73: 220-239.
[8] 董勋, 郭亮, 高宏力, 等. 代价敏感卷积神经网络:一种机械故障数据不平衡分类方法[J]. 仪器仪表学报, 2019, 40(12): 205-213.
     DONG Xun, GUO Liang, GAO Hong-li, et al. Cost sensitive convolutional neural network: a classification method for imbalanced data of mechanical fault[J]. Chinese journal of scientific instrument, 2019, 40(12): 205-213.
[9] 阴艳超, 丁卫刚. 切削加工表面粗糙度的多维多规则云预测方法[J]. 机械工程学报, 2016, 52(15): 204-212.
     YIN Yan-chao, DING Wei-gang. A Novel Cloud Model Prediction for Surface Roughness Based on Multidimensional & Multi-rules Reasoning[J]. Journal of mechanical engineering, 2016, 52(15): 204-212.
[10] Wu Y, Chu H, Xu C, et al. Risk assessment of wind-photovoltaic-hydrogen storage projects using an improved fuzzy synthetic evaluation approach based on cloud model: A case study in China[J]. Journal of Energy Storage, 2021, 38.
[11] Wang M, Liu Q, Wang X, et al. Prediction of Rockburst Based on Multidimensional Connection Cloud Model and Set Pair Analysis[J]. International Journal of Geomechanics, 2020, 20(1).
[12] Peng T, Deng H, Lin Y, et al. Assessment on water resources carrying capacity in karst areas by using an innovative DPESBRM concept model and cloud model[J]. Science of the Total Environment, 2021, 767: 144353.
[13] Liu Q, Wang M, Zhou T, et al. A connection cloud model coupled with extenics for water eutrophication evaluation[J]. Earth Science Informatics, 2019, 12(4): 659-669.
[14] 王国胤. 云模型与粒计算[M]. 北京: 科学出版社, 2018.
WANG Guo-yin. Cloud model and granular computing[M]. Beijing: Science Press, 2018.
[15] 董兴辉, 张鑫淼, 张光, 等. 基于云模型的风电机组输出功率特性分析[J]. 机械工程学报, 2017, 53(22): 198-205.
     DONG Xing-hui, ZHANG Xin-miao, ZHANG Guang, et al. Analysis of Wind Turbine Output Power Characteristic Based on Cloud Model[J]. Journal of mechanical engineering, 2017, 53(22): 198-205.
[16] 田睿, 孟海东, 陈世江, 等. RF-AHP-云模型下岩爆烈度分级预测模型[J]. 中国安全科学学报, 2020, 30(07) : 166-172.
     TIAN Rui, MENG Dong-hai, CHEN Shi-jiang, et al. Prediction model of rockburst intensity classification based on ＲF-AHP-Cloud model[J]. China safety science journal, 2020, 30(07) : 166-172.
[17] 国强, 李明松, 周凯. 基于势距图与改进云模型的多模雷达分选[J/OL]. 吉林大学学报(工学版): 1-10 [2021-08-03].http://kns.cnki.net/kcms/detail/22.1341.T.20210716.1526.002.html.
     GUO Qiang, LI Ming-song, ZHOU Kai. Multi-mode Radar Signal Sorting Based on Potential Distance Graph and Improved Cloud Model[J/OL]. Journal of Jilin University (Engineering and Technology Edition): 1-10 [2021-08-03]. http://kns.cnki.net/kcms/detail/22.1341.T.20210716.1526.002.html.
[18] 李绍红, 王少阳, 朱建东, 等. 基于权重融合和云模型的岩爆倾向性预测研究[J]. 岩土工程学报, 2018, 40(06): 1075-1083.
     LI Shao-hong, WANG Shao-yang, ZHU Jian-dong, et al. Prediction of rock burst tendency based on weighted fusion and improved cloud model[J].Chinese journal of geotechnical engineering, 2018, 40(06): 1075-1083.
[19] 杨洁, 王国胤, 刘群, 等. 正态云模型研究回顾与展望[J]. 计算机学报, 2018, 41(03): 724-744.
YANG Jie, WANG Guo-yin, LIU Qun, et al. Retrospect and Prospect of Research of Normal Cloud Model[J]. Chinese journal of computers, 2018, 41(03): 724-744.
[20] 薛瑞, 赵荣珍. ReliefF与QPSO结合的故障特征选择算法[J]. 振动与冲击, 2020, 39(11): 171-176+208.
     XUE Rui, ZHAO Rong-zhen. The fault feature selection algorithm of combination of ReliefF and QPSO[J]. Journal of vibration and shock, 2020, 39(11): 171-176+208.
[21] 赵荣珍, 马森财, 吴耀春. 云模型和集成极限学习机相结合的滚动轴承故障诊断方法[J]. 兰州理工大学学报, 2021, 47(04): 33-39.
     ZHAO Rong-zhen, MA Sen-cai, WU Yao-chun. Fault diagnosis method for rolling bearing based on cloud model and ensemble extreme learning machine[J]. Journal of Lanzhou University of Technology, 2021, 47(04): 33-39.