製造業在科技的發展與輔助之下，智慧製造與智慧工廠的概念逐漸普及，製造業者積極啟動數位轉型並投入到智慧製造的行列之中，而數據分析在其中發揮了重要作用。對製造業來說設備正常運作是最重要的事情、其中最為關鍵的一點在於對設備故障進行預知與監控，若能從設備收集感測資料，藉由資料對設備零件退化有效診斷，能幫助工廠進行設備維護。在過去許多研究中，健康指標常用於預測剩餘壽命(Remaining Useful Life, RUL), 但由於方法的限制使得健康指標沒有明確邊界，難以從指標得知設備確切狀況而造成判斷失準。本研究針對設備健康檢測提出一套資料收集與分析的系統，能收集多個感測器資訊，並且能即時的分析並視覺化，我們也針對設備健康評估提出一套分析流程，是基於卷積神經網路建構健康指標，所提出的方法首先透過卷積神經網路學習頻譜中的重要特徵，接著將這些學習到的特徵轉化為健康指標。與傳統人為選擇特徵的方法不同，此方法可以彈性應用在不同的機具的感測資料上，不需針對不同應用場域選用不同特徵，且將健康指標進行平滑，使其更加穩定性與穩健，藉由此系統與流程，讓機具操作人員在設備發生問題之前提前預警。

With the development and assistance of technology, the concept of smart manufacturing and the smart factory had gradually become popular in the manufacturing industry. Manufacturers actively initiate digital transformation and engage in smart manufacturing, and data analysis plays an important role. Keeping the equipment running is the most important thing for the manufacturing industry. The most critical point is to predict and monitor the failures of equipment. If the sensing data can be collected from the equipment, and the components can be effectively diagnosed through the data analysis, it helps the factory maintain equipment. In previous studies, health indicators(HI) are often used to predict the remaining useful life (RUL). Due to limitations of methods, health indicators do not have clear boundaries, and it is difficult to evaluate the condition from the indicator to the equipment, resulting in inaccurate judgment. In this work, we propose a data collection and analysis flow for equipment health detection, which can collect multiple sensor information and analyze and visualize it in real-time. We also propose a set of analysis process based on a convolutional neural network to construct a health indicator. The proposed flow first extracts features through a convolutional neural network from the spectrum. Then these extracted features are constructed into a health indicator through a non-linear mapping operation. Unlike the traditional method of artificial feature selection, this method can be flexibly applied to different machine sensing data instead of choosing different features for different application fields. The technique of smoothing the indicator to make it more stable and robust. Through the system and flow, the on-site operating personnel can prevent problems on the equipment in advance.

[1] L. Chen, G. Xu, S. Zhang, W. Yan, and Q. Wu, "Health indicator construction of machinery based on end-to-end trainable convolution recurrent neural networks," Journal of Manufacturing Systems, 54, 1-11, 2020.
[2] R. Teti, K. Jemielniak, G. O'Donnell and Dornfeld, "Advanced monitoring of machining operations," Proceedings of CIRP annals, 59(2), 717-739, 2010.
[3] S. Arnautov, P. Felber, C. Fetzer, and B. Trach, "Ffq: A fast single-producer/multiple-consumer concurrent fifo queue," Proceedings of 2017 IEEE International Parallel and Distributed Processing Symposium, pp. 907-916, May, 2017.
[4] J. Wang, Y. Tian, and X. Fu, "EQueue: elastic lock-free FIFO queue for core-to-core communication on multi-core processors," IEEE Access, 2020.
[5] A. Barrington, S. Feldman, and D. Dechev, "A scalable multi-producer multi-consumer wait-free ring buffer," Proceedings of the 30th Annual ACM Symposium on Applied Computing (pp. 1321-1328), April, 2015.
[6] R. Hanlon, R. Dooley, S. Mock, M. Dahan, P. Nuthulapati, and P. Hurley, "A Case Study for NoSQL Applications and Performance Benefits: CouchDB vs. Postgres," Proceedings of the 2011 TeraGrid Conference: Extreme Digital Discovery. ACM, 2011.
[7] J. Lindström, V. Raatikka, J. Ruuth, P. Soini, and K. Vakkila, "IBM solidDB: In-Memory Database Optimized for Extreme Speed and Availability," IEEE Data Eng. Bull., 36(2), 14-20, 2013.
[8] J. Wang, Y. Tian, and X. Fu, "EQueue: elastic lock-free FIFO queue for core-to-core communication on multi-core processors," IEEE Access, 2020.
[9] Ş. G. Ögüdücü, M. Gayberi, E. Akpimar, and H. Kutluay, "A study for performance comparison of different in-memory databases," In 2013 7th International Conference on Application of Information and Communication Technologies (pp. 1-5). IEEE, October, 2013.
[10] J. Krueger, M. Grund, M. Boissier, A. Zeier, and H. Plattner, "Data structures for mixed workloads in in-memory databases," Proceedings of 5th International Conference on Computer Sciences and Convergence Information Technology (pp. 394-399). IEEE, 2010.
[11] R. Ramjee, J. Kurose, D. Towsley, and H. Schulzrinne, "Adaptive playout mechanisms for packetized audio applications in wide-area networks," Proceedings of INFOCOM'94 Conference on Computer Communications (pp. 680-688). IEEE, June, 1994.
[12] T. Y. Huang, R. Johari, and N. McKeown, "Downton abbey without the hiccups: Buffer-based rate adaptation for http video streaming," Proceedings of the 2013 ACM SIGCOMM workshop on Future human-centric multimedia networking (pp. 9-14), August, 2013.
[13] T. Y. Huang, R. Johari, N. McKeown, M. Trunnell, and M. Watson, "A buffer-based approach to rate adaptation: Evidence from a large video streaming service," Proceedings of the 2014 ACM conference on SIGCOMM (pp. 187-198), August, 2014.
[14] 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.
[15] A. Soualhi, K. Medjaher, and N. Zerhouni, "Bearing health monitoring based on Hilbert–Huang transform, support vector machine, and regression," IEEE Transactions on Instrumentation and Measurement, 64(1), 52-62, 2014.
[16] B. Zhang, S. Zhang, and W. Li, "Bearing performance degradation assessment using long short-term memory recurrent network," Computers in Industry, 106, 14-29, 2019.
[17] Y. Lei, N. Li, L. Guo, N. Li, T. Yan, and J. Lin, "Machinery health prognostics: A systematic review from data acquisition to RUL prediction," Mechanical Systems and Signal Processing, 104, 799-834, 2018.
[18] J. Zhu, T. Nostrand, C. Spiegel, and B. Morton, "Survey of condition indicators for condition monitoring systems," Proceedings of Annu. Conf. Progn. Heal. Manag. Soc (Vol. 5, pp. 1-13), September, 2014.
[19] W. Ahmad, S. A. Khan, and J. M. Kim, "A hybrid prognostics technique for rolling element bearings using adaptive predictive models," IEEE Transactions on Industrial Electronics, 65(2), 1577-1584, 2017.
[20] 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, 2018.
[21] Y. Lei, N. Li, and J. Lin, "A new method based on stochastic process models for machine remaining useful life prediction," Proceedings of IEEE Transactions on Instrumentation and Measurement, 65(12), 2671-2684, 2016.
[22] A. Widodo, and B. S. Yang, "Application of relevance vector machine and survival probability to machine degradation assessment," Expert Systems with Applications, 38(3), 2592-2599, 2011.
[23] L. Guo, N. Li, F. Jia, Y. Lei, and J. Lin, "A recurrent neural network-based health indicator for remaining useful life prediction of bearings," Neurocomputing, 240, 98-109, 2017.
[24] X. Li, W. Zhang, and Q. Ding, "Deep learning-based remaining useful life estimation of bearings using multi-scale feature extraction," Reliability Engineering & System Safety, 182, 208-218, 2019.
[25] L. Guo, N. Li, F. Jia, Y. Lei, and J. Lin, "Machinery health indicator construction based on convolutional neural networks considering trend burr," Neurocomputing, 292, 142-150, 2018.
[26] E. S. De Argandoña, A. Aztiria, C. García, N. Arana, A. Izaguirre, and P. Fillatreau, "Forming processes control by means of artificial intelligence techniques," Robotics and Computer-Integrated Manufacturing, 24(6), 773-779, 2008.
[27] W. Michael, "Involving Operators in Your Equipment Maintenance Plans," https://www.lce.com/Involving-Operators-in-Your-Equipment-Maintenance-Plans-1882.html