隨著科技的演進，工業4.0的概念漸漸普及，而數據分析這項技術也在多年來對各項產業帶來重大影響，包括決策制定、提高生產率、降低成本等等。尤其工業大數據更在近年來受到各製造業的重視。由生產製程中所產生的數據十分大量，若要從龐大的數據中找出有用的資訊，製造業者勢必要具備處理複雜資料的能力。此外，在工業製程中，即使是微小的故障或錯誤都有可能對業者造成重大的損失，因此，機器、零件甚至是系統的失效診斷已成為學界與業界極為關注的議題。本研究針對軸承的失效偵測整合出一套基於時頻域的相似性分析流程。在此分析流程中，主成份分析被應用於降低資料維度，以轉換出最能夠代表原始資料的少數主成份。並結合霍特林T平方統計，計算資料群之間的相似性，以比較軸承於運作過程中，不同使用階段與健康狀態之間的相似性。而藉由此流程，能夠制定出一項足以準確偵測軸承失效時間的指標，以便讓機具的操作人員能在第一時間對異常的軸承進行檢查，降低異常軸承對產品、生產系統甚至操作人員安全所帶來的影響。

As technology keeps developing, the idea of Industry 4.0 has been widespread around the world. Data analysis, one of the concepts of Industry 4.0, has made changes to various fields of industries for years, including decision making, productivity improvement, lowering the cost, and so on. Specifically, in the field of manufacturing industry, industrial big data is a fairly popular technique nowadays. Since the amount of data generated from the machine and system is usually massive, it is important for manufacturers to be able to deal with the complex data, and further extract some useful information from the data. Moreover, a simple malfunction in a system may lead to huge loss in the manufacturing industry. Fault diagnosis, which aims to identify faults occurring on the components, machines, and even the whole system, has then become one of the issues that the manufacturers pay most attention to. In this work, we propose a system flow of conducting similarity analysis of bearing vibration data in the time-frequency domain. Principal component analysis, one of the important steps in this system flow, is applied to reduce the dimensionality of complex data. Simultaneously, a set of principal components, which can mostly explain the original data, are generated. In addition, Hotelling’s T2 Statistics is adopted to evaluate the similarity between groups of data, in order to identify the similarity between healthy bearings and used bearings. With this system flow, an indicator that is able to detect bearing fault is generated, which can offer the on-site operating personnel an exact time to check the abnormal bearings, and thus lower the influence caused by the abnormal bearings.

[1] J. Lee, H.-D. Ardakani, S. Yang, and B. Bagheri, “Industrial Big Data Analytics and Cyber-Physical Systems for Future Maintenance & Service Innovation,” Procedia CIRP, 38:3-7, 2015.
[2] N. Amruthnath, and T. Gupta, “Fault Class Prediction in Unsupervised Learning Using Model-Based Clustering Approach,” Proceedings of the 2018 International Conference on Information and Computer Technologies, May 2018.
[3] F. Al-Badour, M. Sunar, and L. Cheded, “Vibration Analysis of Rotating Machinery Using Time-Frequency Analysis and Wavelet Techniques,” Mechanical Systems and Signal Processing, 25(6): 2083-2101, August 2011.
[4] J. Lee, H.-A. Kao, and S. Yang, “Service Innovation and Smart Analytics for Industry 4.0 and Big Data Environment,” Procedia CIRP, 16:3-8, 2014.
[5] C. Gkerekos, I. Lazakis, and G. Theotokatos, “Ship Machinery Condition Monitoring Using Vibration Data Through Supervised Learning,” Proceedings of the International Conference on Maritime Safety and Operations, pp. 103-110, October 2016.
[6] A.-R. Mohanty, “Machinery Condition Monitoring: Principles and Practices,” CRC Press, 2015.
[7] R.-B. Randall, “Vibration-Based Condition Monitoring: Industrial, Aerospace and Automotive Applications,” John Wiley & Sons, 2011.
[8] Z. Gao, C. Cecati, and S.-X. Ding, “A Survey of Fault Diagnosis and Fault-Tolerant Techniques - Part I: Fault Diagnosis with Model-Based and Signal-Based Approaches,” IEEE Transactions on Industrial Electronics, 62(6): 3757-3767, June 2015.
[9] A. Nabhan, M. Nouby, A.-M. Sami, and M.-O. Mousa, “Bearing Fault Detection Techniques - A Review,” Turkish Journal of Engineering, Sciences and Technology, 3(2), 2015.
[10] B. Wang, Y. Lei, N. Li, and N. Li, “A Hybrid Prognostics Approach for Estimating Remaining Useful Life of Rolling Element Bearings,” IEEE Transactions on Reliability, 69(1):401-412, March 2020.
[11] C. Scheffer, and P. Girdhar, “Practical Machinery Vibration Analysis and Predictive Maintenance,” Elsevier, 2004.
[12] R.-J. Romero-Troncoso, “Multirate Signal Processing to Improve FFT-Based Analysis for Detecting Faults in Induction Motors,” IEEE Transactions on Industrial Informatics, 13(3):1291-1300, June 2017.
[13] F. Huang, A. Sava, K.-H. Adjallah, and Z. Wang, “Bearings Degradation Monitoring Indicator Based on Segmented Hotelling T Square and Piecewise Linear Representation,” Proceedings of the 2018 IEEE International Conference on Mechatronics and Automation, October 2018.
[14] C. Bianchini, F. Immovilli, M. Cocconcelli, R. Rubini, and A. Bellini, “Fault Detection of Linear Bearings in Brushless AC Linear Motors by Vibration Analysis,” IEEE Transactions on Industrial Electronics, 58(5):1684-1694, May 2011.
[15] H. Yu, H.-R. Li, Z.-K. Tian, and W.-G. Wang, “Rolling Bearing Degradation State Identification Based on LPP Optimized by GA,” International Journal of Rotating Machinery 2016, August 2016.
[16] E. Skubalska-Rafajłowicz, “Random Projections and Hotelling’s T2 Statistics for Change Detection in High-Dimensional Data Streams,” International Journal of Applied Mathematics and Computer Science, 23(2):447-461, June 2013.
[17] F. Tsung, and K. Wang, “Adaptive Charting Techniques: Literature Review and Extensions,” Frontiers in Statistical Quality Control 9, pp. 19-35, January 2010.
[18] J. Tian, C. Morillo, M.-H. Azarian, and M. Pecht, “Motor Bearing Fault Detection Using Spectral Kurtosis-Based Feature Extraction Coupled with K-Nearest Neighbor Distance Analysis,” IEEE Transactions on Industrial Electronics, 63(3):1793-1803, March 2016.
[19] J. Shlens, “A Tutorial on Principal Component Analysis,” arXiv:1404.1100, April 2014.
[20] S. Wold, K. Esbensen, and P. Geladi, “Principal Component Analysis,” Chemometrics and Intelligent Laboratory Systems, 2(1-3):37-52, August 1987.
[21] L.-J. Cao, K.-S. Chua, W.-K. Chong, H.-P. Lee, and Q.-M. Gu, “A Comparison of PCA, KPCA and ICA for Dimensionality Reduction in Support Vector Machine,” Neurocomputing, 55(1-2):321-336, September 2003.
[22] D. Lubo-Robles, “Development of Independent Component Analysis for Reservoir Geomorphology and Unsupervised Seismic Facies Classification in the Taranaki Basin, New Zealand,” University of Oklahoma Graduate College, May 2018.
[23] NIST/SEMATECH E-Handbook of Statistical Methods,
https://www.itl.nist.gov/div898/handbook/
[24] F. Aparisi, and C.-L. Haro, “Hotelling's T2 Control Chart with Variable Sampling Intervals,” International Journal of Production Research, 39(14):3127-3140, 2001.
[25] V. Chandola, A. Banerjee, and V. Kumar, “Anomaly Detection: A Survey,” ACM Computing Surveys, 41(3):1-58, July 2009.
[26] D. Shickel, “Reactive vs. Preventive vs. Predictive Maintenance,” June 2020, https://www.prometheusgroup.com/posts/reactive-vs-preventive-vs-predictive-maintenance
[27] Y.-G. Lei, T.-Y. Han, B. Wang, N.-P. Li, T. Yan, and J. Yang, “XJTU-SY Rolling Element Bearing Accelerated Life Test Datasets: A Tutorial,” Journal of Mechanical Engineering, 55(16):1, August 2019.
[28] P. Sharma, “The Ultimate Guide to 12 Dimensionality Reduction Techniques (with Python codes),” August 2018,
https://www.analyticsvidhya.com/blog/2018/08/dimensionality-reduction-techniques-python/
[29] S. Doğan, “Why Scree Plot is Important in PCA?” April 2020,
https://medium.com/@sercandogan/why-scree-plot-is-important-in-pca-a66cd7dcd624
[30] S. Trivedi, “Blind Source Separation in Magnetic Resonance Images,” January 2010
https://onionesquereality.wordpress.com/2010/01/30/blind-source-separation-in-magnetic-resonance-images/
[31] B. Brown, “Hotelling’s T2 Test,” https://www.slideserve.com/bette/hotelling-s-t-2-test