摘要
在現今的數位世代，面對不斷增長的的消費性電子產品需求，造就了半導體行業的激烈競爭。半導體製造商為確保市場競爭力，除開發新技術外，提高生產效率以及降低製造成本也是競爭策略之一。其中，製造設備的可用性與性能將直接影響生產效率，若發生計畫外的設備故障將導致大量的產能損失，甚至影響產品品質。預防性保養是當前產業用以避免非預期性設備異常的常見做法，但此方法未能有效的考量設備狀況，存在設備利用率不足的可能，因此狀態性維修(CBM)的概念逐漸萌芽。
狀態性維修策略是藉由製造設備上的感應器，大量蒐集生產過程的生產數據，在考量設備狀況下進行維護策略的製訂，以確保設備利用率。本研究提出以數據驅動的狀態性維修架構，利用設備感應器所蒐集的資料，先以偏最小平方判別分析(PLS-DA)進行數據降維，再基於變量投影重要性(VIP)選取特徵子集。經提取後的特徵一方面可經由轉換，進一步作為設備健康度指標(HI)，用以監測設備的當前狀況；另一方面則做為基於人工神經網路(ANN)模型的輸入，透過模型的構建來檢測設備未來的健康狀況。倘若設備的健康狀況開始出現下降狀態，則建立一個長短期記憶(LSTM)網路來預測設備的剩餘使用壽命(RUL)。
本研究以台灣某半導體公司提供的設備生產數據進行實證研究，並驗證其模型效度與可行性。
關鍵字：狀態性維修、機台狀態監控、預測性維修、人工神經網絡、長短期記憶、偏最小平方法。

Abstract
The technology revolution has created an ever-increasing demand for consumer electronics, which has consequently resulted in an intense competition in the semiconductor industry. Manufacturers focus on enhancing production efficiency and lowering their manufacturing costs. This is heavily dependent on the availability and performance of the manufacturing equipment. Unscheduled breakdown of fabrication equipment may cause significant production losses, and affect the product quality. Current industry practices are predominantly preventive or reactive maintenance. Regular maintenance performed without considering the equipment condition, which causes disruption in the production process and may lead to the underutilization of the equipment. The advancement in manufacturing technology allows the collection of sensor data from the equipment during production process. Condition Based Maintenance (CBM) strategy effectively integrates this data for making maintenance decisions. The proposed framework is a data driven approach that uses features extracted from the sensor data to develop a method for the condition monitoring of the equipment health. Partial least square discriminant analysis (PLS-DA) is used for dimensionality reduction of the data, and a subset of feature is selected based on variable importance projection (VIP). These features are further transformed into a health indicator (HI) that provides information about the current state of the equipment. Then an artificial neural network (ANN) based model is constructed for detecting the future health state. Finally, a long short-term memory (LSTM) network is built to predict the remaining useful life (RUL) of the equipment, once it enters the degraded state. An empirical study is conducted at a semiconductor company in Taiwan to validate this research.
Keywords: condition based maintenance, predictive maintenance, remaining useful life, artificial neural networks, long short-term memory, partial least square.

Reference

Al-kharaz, M., Ananou, B., Ouladsine, M., Combal, M., and Pinaton, J. (2019), "Quality Prediction in Semiconductor Manufacturing processes Using Multilayer Perceptron Feedforward Artificial Neural Network," Proceedings of 2019 8th International Conference on Systems and Control (ICSC).
Alaswad, S. and Xiang, Y. (2017), "A review on condition-based maintenance optimization models for stochastically deteriorating system," Reliability Engineering & System Safety, Vol. 157, No., pp. 54-63.
Ali, J. B., Chebel-Morello, B., Saidi, L., Malinowski, S., and Fnaiech, F. (2015), "Accurate bearing remaining useful life prediction based on Weibull distribution and artificial neural network," Mechanical Systems and Signal Processing, Vol. 56, No., pp. 150-172.
Ayo-Imoru, R. and Cilliers, A. (2018), "A survey of the state of condition-based maintenance (CBM) in the nuclear power industry," Annals of Nuclear Energy, Vol. 112, No., pp. 177-188.
Barker, M. and Rayens, W. (2003), "Partial least squares for discrimination," Journal of Chemometrics: A Journal of the Chemometrics Society, Vol. 17, No. 3, pp. 166-173.
Bengio, Y., Simard, P., and Frasconi, P. (1994), "Learning long-term dependencies with gradient descent is difficult," IEEE transactions on neural networks, Vol. 5, No. 2, pp. 157-166.
Bousdekis, A., Magoutas, B., Apostolou, D., and Mentzas, G. (2018), "Review, analysis and synthesis of prognostic-based decision support methods for condition based maintenance," Journal of Intelligent Manufacturing, Vol. 29, No. 6, pp. 1303-1316.
Bousdekis, A., Mentzas, G., Hribernik, K., Lewandowski, M., von Stietencron, M., and Thoben, K.-D. (2019), "A Unified Architecture for Proactive Maintenance in Manufacturing Enterprises," in: (eds.), Enterprise Interoperability VIII, Springer, pp. 307-317.
Chien, C.-F., Hsu, C.-Y., and Chen, P.-N. (2013), "Semiconductor fault detection and classification for yield enhancement and manufacturing intelligence," Flexible Services and Manufacturing Journal, Vol. 25, No. 3, pp. 367-388.
Cholette, M. E., Celen, M., Djurdjanovic, D., and Rasberry, J. D. (2013), "Condition monitoring and operational decision making in semiconductor manufacturing," IEEE Transactions on Semiconductor Manufacturing, Vol. 26, No. 4, pp. 454-464.
Crespo Marquez, A., Gupta, J. N., and Ignizio, J. P. (2006), "Improving preventive maintenance scheduling in semiconductor fabrication facilities," Production Planning & Control, Vol. 17, No. 7, pp. 742-754.
Ding, S.-H. and Kamaruddin, S. (2015), "Maintenance policy optimization—literature review and directions," The International Journal of Advanced Manufacturing Technology, Vol. 76, No. 5, pp. 1263-1283.
Dragomir, O. E., Gouriveau, R., Dragomir, F., Minca, E., and Zerhouni, N. (2009), "Review of prognostic problem in condition-based maintenance," Proceedings of 2009 European Control Conference (ECC).
Fordellone, M., Bellincontro, A., and Mencarelli, F. (2018), "Partial least squares discriminant analysis: A dimensionality reduction method to classify hyperspectral data," arXiv preprint arXiv:1806.09347, Vol., No., pp.
Gers, F. A., Schmidhuber, J., and Cummins, F. (1999), "Learning to forget: Continual prediction with LSTM," Vol., No., pp.
Hochreiter, S. and Schmidhuber, J. (1997), "Long short-term memory," Neural computation, Vol. 9, No. 8, pp. 1735-1780.
Hsieh, Y.-S., Cheng, F.-T., Huang, H.-C., Wang, C.-R., Wang, S.-C., and Yang, H.-C. (2012), "VM-based baseline predictive maintenance scheme," IEEE Transactions on semiconductor Manufacturing, Vol. 26, No. 1, pp. 132-144.
Jardine, A. K., Lin, D., and Banjevic, D. (2006), "A review on machinery diagnostics and prognostics implementing condition-based maintenance," Mechanical systems and signal processing, Vol. 20, No. 7, pp. 1483-1510.
Jin, X., Siegel, D., Weiss, B. A., Gamel, E., Wang, W., Lee, J., and Ni, J. (2016), "The present status and future growth of maintenance in US manufacturing: results from a pilot survey," Manufacturing review, Vol. 3, No., pp.
Krenek, J., Kuca, K., Blazek, P., Krejcar, O., and Jun, D. (2016), "Application of artificial neural networks in condition based predictive maintenance," in: (eds.), Recent Developments in Intelligent Information and Database Systems, Springer, pp. 75-86.
Laayouj, N., Jamouli, H., and El Hail, M. (2016), "Prognosis of degradation through a dynamic estimation of remaining useful life," Proceedings of 2016 5th International Conference on Systems and Control (ICSC).
Lee, C.-Y. and Dong, Z.-H. (2019), "Hierarchical Equipment Health Index Framework," IEEE Transactions on Semiconductor Manufacturing, Vol. 32, No. 3, pp. 267-276.
Lin, K.-Y., Hsu, C.-Y., and Yu, H.-C. (2014), "A virtual metrology approach for maintenance compensation to improve yield in semiconductor manufacturing," International Journal of Computational Intelligence Systems, Vol. 7, No. sup2, pp. 66-73.
Liu, L., Liu, D., Zhang, Y., and Peng, Y. (2016), "Effective sensor selection and data anomaly detection for condition monitoring of aircraft engines," Sensors, Vol. 16, No. 5, pp. 623.
Luo, M., Yan, H.-C., Hu, B., Zhou, J.-H., and Pang, C. K. (2015), "A data-driven two-stage maintenance framework for degradation prediction in semiconductor manufacturing industries," Computers & Industrial Engineering, Vol. 85, No., pp. 414-422.
Monostori, L. (2014), "Cyber-physical production systems: Roots, expectations and R&D challenges," Procedia Cirp, Vol. 17, No., pp. 9-13.
Mrugalska, B. (2018), "Remaining useful life as prognostic approach: a review," Proceedings of International Conference on Human Systems Engineering and Design: Future Trends and Applications.
Munirathinam, S. and Ramadoss, B. (2014), "Big data predictive analtyics for proactive semiconductor equipment maintenance," Proceedings of 2014 IEEE International Conference on Big Data (Big Data).
Muthanandan, S. and Nor, K. A. B. (2019), "Condition Monitoring and Assessment for Rotating Machinery," in: (eds.), Rotating Machineries, Springer, pp. 1-22.
Nwadinobi, C. and Ezeaku, I. (2018), "Review of Maintenance Scheduling and Optimization Models," Vol., No., pp.
Qian, P., Tian, X., Kanfoud, J., Lee, J. L. Y., and Gan, T.-H. (2019), "A novel condition monitoring method of wind turbines based on long short-term memory neural network," Energies, Vol. 12, No. 18, pp. 3411.
Romain, D., Dickie, R., Robu, V., and Flynn, D. (2017), "A review of the role of prognostics in predicting the remaining useful life of assets," Proceedings of 27th European Safety and Reliability Conference.
Rostami, H., Blue, J., and Yugma, C. (2016), "Equipment condition diagnosis and fault fingerprint extraction in semiconductor manufacturing," Proceedings of 2016 15th IEEE International Conference on Machine Learning and Applications (ICMLA).
Ruschel, E., Santos, E. A. P., and Loures, E. d. F. R. (2017), "Industrial maintenance decision-making: A systematic literature review," Journal of Manufacturing Systems, Vol. 45, No., pp. 180-194.
Sikorska, J., Hodkiewicz, M., and Ma, L. (2011), "Prognostic modelling options for remaining useful life estimation by industry," Mechanical systems and signal processing, Vol. 25, No. 5, pp. 1803-1836.
Tsui, K. L., Chen, N., Zhou, Q., Hai, Y., and Wang, W. (2015), "Prognostics and health management: A review on data driven approaches," Mathematical Problems in Engineering, Vol. 2015, No., pp.
Widodo, A. and Yang, B.-S. (2007), "Support vector machine in machine condition monitoring and fault diagnosis," Mechanical systems and signal processing, Vol. 21, No. 6, pp. 2560-2574.
Wu, J., Hu, K., Cheng, Y., Zhu, H., Shao, X., and Wang, Y. (2019), "Data-driven remaining useful life prediction via multiple sensor signals and deep long short-term memory neural network," ISA transactions, Vol., No., pp.
Wu, Y., Yuan, M., Dong, S., Lin, L., and Liu, Y. (2018), "Remaining useful life estimation of engineered systems using vanilla LSTM neural networks," Neurocomputing, Vol. 275, No., pp. 167-179.
Yu, H.-C., Lin, K.-Y., and Chien, C.-F. (2014), "Hierarchical indices to detect equipment condition changes with high dimensional data for semiconductor manufacturing," Journal of Intelligent Manufacturing, Vol. 25, No. 5, pp. 933-943.
Yu, W., Kim, I. Y., and Mechefske, C. (2019), "Remaining useful life estimation using a bidirectional recurrent neural network based autoencoder scheme," Mechanical Systems and Signal Processing, Vol. 129, No., pp. 764-780.
Zhang, S., Zhang, C., and Yang, Q. (2003), "Data preparation for data mining," Applied artificial intelligence, Vol. 17, No. 5-6, pp. 375-381.
Zhang, Y., Hutchinson, P., Lieven, N. A., and Nunez-Yanez, J. (2020), "Remaining Useful Life Estimation Using Long Short-Term Memory Neural Networks and Deep Fusion," IEEE Access, Vol. 8, No., pp. 19033-19045.
Zhao, R., Yan, R., Chen, Z., Mao, K., Wang, P., and Gao, R. X. (2019), "Deep learning and its applications to machine health monitoring," Mechanical Systems and Signal Processing, Vol. 115, No., pp. 213-237.
Zhao, R., Yan, R., Wang, J., and Mao, K. (2017), "Learning to monitor machine health with convolutional bi-directional LSTM networks," Sensors, Vol. 17, No. 2, pp. 273.
Zheng, S., Ristovski, K., Farahat, A., and Gupta, C. (2017), "Long short-term memory network for remaining useful life estimation," Proceedings of 2017 IEEE International Conference on Prognostics and Health Management (ICPHM).
Zou, J., Chang, Q., Arinez, J., Xiao, G., and Lei, Y. (2017), "Dynamic production system diagnosis and prognosis using model-based data-driven method," Expert Systems with Applications, Vol. 80, No., pp. 200-209.