應用數位分身於馬達軸承之異常偵測

以作者查詢圖書館館藏

、以作者查詢臺灣博碩士

、以作者查詢全國書目

、勘誤回報

、線上人數：14

、訪客IP：3.149.240.185

姓名

吳孟勳(Meng-Hsun Wu) 查詢紙本館藏

畢業系所

資訊工程學系

論文名稱

應用數位分身於馬達軸承之異常偵測
(Applying Digital Twin for Anomaly Detection of Motor Bearings)

相關論文

★ 基於質譜儀資料使用機器學習辨識克雷伯氏肺炎桿菌之多重抗藥性	★ 結合多種訊號預處理方法於質譜儀資料以辨識細菌對抗生素之抗藥性
★ 利用機器學習預測濁水溪沖積扇區域之地下水位	★ 使用表徵學習和機器學習方法於晶圓線切割機台之異常偵測
★ 基於質譜儀資料利用人工智慧方法辨識革蘭氏陰性菌對環丙沙星抗藥性之特徵峰值	★ 基於光誘導介電泳影像處理檢測流體抗藥性
★ 利用機器學習方法基於多類型地層監測資料預測濁水溪沖積扇地區之地層下陷	★ 基於人工智慧模型預測抗菌肽的最小抑菌濃度於特定菌株上
★ 使用語言模型嵌入和不平衡調整之深度學習方法識別多功能抗菌肽	★ 使用權重組合模型預測雲林縣地層下陷
★ 基於深度學習從核醣核酸定序表達譜推斷外周血單核細胞之細胞組成

檔案

[Endnote RIS 格式]

[Bibtex 格式]

[相關文章]

[文章引用]

[完整記錄]

[館藏目錄]

至系統瀏覽論文 ( 永不開放)

摘要(中)

馬達系統被廣泛運用在現今工廠許多精密且複雜的機器中，一旦馬達發生異常或故障導致生產機台無預警的停機，將造成重大的損失，而其中的軸承扮演著重要的角色，許多故障原因與軸承均具有高度相關。近年來，數位分身越來越被廣泛應用於工廠機台之運作，藉由建置之虛擬模型，可以讓使用者能透過感測器或軟體收集的數據，即時了解目標對象或系統的運作情況。儘管許多研究表明數位分身的諸多優點且有針對馬達軸承來建構數位分身，目前仍缺乏以虛擬模型的模擬資料結合真實資料以及預警系統來提升對馬達軸承之異常偵測結果。本研究目標為針對馬達中的軸承，透過感測器收集到的資料，先使用線性回歸(linear regression)和決策樹回歸(decision tree regression)來建置其數位分身後，使用樣本合成技術(synthetic minority oversampling technique)平衡資料集並利用k鄰近演算法(k-nearest neighbors)、隨機森林(random forest)和單純貝氏分類(naïve bayes)演算法以及警報機制，來分析軸承的異常偵測。結果顯示與未採用數位分身相比，使用數位分身後可以明顯提升模型準確率。

摘要(英)

Motor systems are widely used in many sophisticated and complex machines in factories nowadays. Once the anomaly or failure occurs on motor and causes unexpected shutdown, it will lead to significant losses. The bearing plays an important role in motor, and many failures and anomalies in motor are highly related to bearings. In recent years, digital twin (DT) has become more and more widely used in factory machines. By building virtual model, it allows users to understand the real situation of the target object or system in real time through the data collected by sensors or software. Although many studies have shown advantages of DT and the construction of DT for motor bearings, there is still a lack of using simulation data of DT models combined with real data and early warning systems to improve anomaly detection results of motor bearings. The goal of this study is to use the data collected by sensors, first apply linear regression and decision tree regression to build DT for the bearings in motor. Then, use synthetic minority oversampling technique to balance dataset and adopt k-nearest neighbors, random forest, naïve bayes algorithm and alarm rules to analyze bearings anomaly detection. The results show that compared with not using DT, the accuracy of the model can be obviously improved by digital twin.

關鍵字(中)

★ 數位分身
★ 異常偵測
★ 不平衡資料集
★ 機器學習

關鍵字(英)

★ digital twin
★ anomaly detection
★ imbalance dataset
★ machine learning

論文目次

中文摘要 i
Abstract ii
致謝 iii
Table of Contents iv
List of Figures vi
List of Tables vii
Chapter 1. Introduction 1
1-1 Background 1
1-2 Related Works 3
1-3 Motivation and Goal 4
Chapter 2. Materials and Methods 5
2-1 Data Description and Preprocessing 6
2-2 Construction of Digital Twin Models 9
2-3 Classifiers for Time Windows 11
2-4 Alarm Rules 14
2-5 Evaluations 15
Chapter 3. Results 17
3-1 Results of Digital Twin Construction 17
3-2 Results of Classifiers 21
3-3 Results of Anomaly Detection 23
Chapter 4. Discussions and Conclusions 27
4-1 Discussions 27
4-2 Conclusions 29
References 30
Appendix 32

參考文獻

1. Li, B., et al., Neural-network-based motor rolling bearing fault diagnosis. 2000. 47(5): p. 1060-1069.
2. Sun, W., et al., Convolutional discriminative feature learning for induction motor fault diagnosis. 2017. 13(3): p. 1350-1359.
3. Ince, T., et al., Real-time motor fault detection by 1-D convolutional neural networks. 2016. 63(11): p. 7067-7075.
4. Haag, S. and R.J.M.L. Anderl, Digital twin–proof of concept. 2018. 15: p. 64-66.
5. Tao, F., et al., Digital twin in industry: state-of-the-art. 2018. 15(4): p. 2405-2415.
6. He, B. and K.-J.J.A.i.M. Bai, Digital twin-based sustainable intelligent manufacturing: a review. 2021. 9(1): p. 1-21.
7. Zheng, Y., et al., An application framework of digital twin and its case study. 2019. 10(3): p. 1141-1153.
8. Guo, J., et al., Modular based flexible digital twin for factory design. 2019. 10(3): p. 1189-1200.
9. Boschert, S. and R. Rosen, Digital twin—the simulation aspect, in Mechatronic futures. 2016, Springer. p. 59-74.
10. Tao, F., et al., Digital twin-driven product design framework. 2019. 57(12): p. 3935-3953.
11. Botkina, D., et al., Digital twin of a cutting tool. 2018. 72: p. 215-218.
12. Aivaliotis, P., K. Georgoulias, and G.J.I.J.o.C.I.M. Chryssolouris, The use of digital twin for predictive maintenance in manufacturing. 2019. 32(11): p. 1067-1080.
13. Zhang, S., et al., A product quality monitor model with the digital twin model and the stacked auto encoder. 2020. 8: p. 113826-113836.
14. Negri, E., L. Fumagalli, and M.J.P.M. Macchi, A review of the roles of digital twin in CPS-based production systems. 2017. 11: p. 939-948.
15. Tao, F., et al., Digital twin driven prognostics and health management for complex equipment. 2018. 67(1): p. 169-172.
16. Wang, J., et al., Digital twin for rotating machinery fault diagnosis in smart manufacturing. 2019. 57(12): p. 3920-3934.
17. Guivarch, D., et al., Creation of helicopter dynamic systems digital twin using multibody simulations. 2019. 68(1): p. 133-136.
18. Venkatesan, S., et al., Health monitoring and prognosis of electric vehicle motor using intelligent-digital twin. 2019. 13(9): p. 1328-1335.
19. Anis, M.D., S. Taghipour, and C.-G. Lee. Optimal RUL estimation: a state-of-art digital twin application. in 2020 Annual Reliability and Maintainability Symposium (RAMS). 2020. IEEE.
20. Cattaneo, L. and M.J.I.-P. Macchi, A digital twin proof of concept to support machine prognostics with low availability of run-to-failure data. 2019. 52(10): p. 37-42.
21. Chandola, V., A. Banerjee, and V.J.A.c.s. Kumar, Anomaly detection: a survey. 2009. 41(3): p. 1-58.
22. Xu, Y., et al., A digital-twin-assisted fault diagnosis using deep transfer learning. 2019. 7: p. 19990-19999.
23. Castellani, A., S. Schmitt, and S.J.a.p.a. Squartini, Real-world anomaly detection by using digital twin systems and weakly-supervised learning. 2020.
24. Pedregosa, F., et al., Scikit-learn: machine learning in python. 2011. 12: p. 2825-2830.
25. Hunter, J.D.J.I.A.o.t.H.o.C., Matplotlib: a 2D graphics environment. 2007. 9(03): p. 90-95.
26. Lim, P., C.K. Goh, and K.C. Tan. A time window neural network based framework for remaining useful life estimation. in 2016 international joint conference on neural networks (IJCNN). 2016. IEEE.
27. Li, X., et al., Remaining useful life estimation in prognostics using deep convolution neural networks. 2018. 172: p. 1-11.
28. Luo, M., et al., Using imbalanced triangle synthetic data for machine learning anomaly detection. 2019. 58(1): p. 15-26.
29. Chawla, N.V., et al., SMOTE: synthetic minority over-sampling technique. 2002. 16: p. 321-357.

指導教授

洪炯宗吳立青(Jorng-Tzong Horng Li-Ching Wu)

審核日期

2021-7-23

推文