以燃氣渦輪(Gas Turbine)發動機做為基本動力機械裝置平台，已廣泛應用於輕工業、重工業、航空、航海等動力裝備。我國先進之國防武器中，許多戰機均使用這類機械裝置做為動力來源，機上也安裝多種的偵測器隨時監測發動機的運作狀態。此類發動機的維修方式多採用定期檢測與維修方式進行，但對於長時間處在高溫與高機械應力操作環境下的發動機而言，定期檢修的方式對於因組件傷害的累積所造成的發動機性能衰降，無法提供有效的性能預測藉以提高使用此類發動機器具使用的安全性。因此，本研究蒐集此類發動機相關的維修與飛行操作紀錄，嘗試運用資料探勘技術建立一個具有最低成本的發動機失效即時預測模式，提供維修人員發動機失效的預警，進而規劃適切之維修計畫，用以降低維修零組件的庫存與人力需求。本研究利用發動機性能衰降的歷史紀錄以及四種資料探勘技術建立預測模式，用以預測發動機性能衰降後之剩餘可用時間。本研究比較四種預測模式的良窳，提供建立發動機性能衰降預測模式的建立架構，期能讓傳統的定期維修升級為預兆式的維修管理，以提供良好的發動機使用安全。

Use of a gas turbine engine as the primary power source has been popular in light and heavy industries, aerospace engineering, marine engineering, etc. Gas turbine engine is also used in our modern national defense weapons in Taiwan. For instance, most of Air Force fighters use gas turbine engines as the source of power. Gas turbine engines are usually associated with various sensors for real-time condition surveillance and require periodical maintenance for providing proper functioning and safety guarantee. In contrast, real-time failure prediction of gas turbine engine components could be achieved by applying data mining or statistics techniques. However, such failure prediction will not be effective when applying to engines which are deteriorated by long-term running in high temperature and high stress environment. In this study, we collected maintenance and operating logs according to the engine deterioration history and established and empirically evaluated four different data-mining-based prediction models. The proposed data-mining-based prediction approach attempts to predict the time-to-deterioration for a gas turbine engine after its prior deterioration occurrence, to provide maintenance personnel accurate prediction for better making or revising maintenance schedules, and to achieve the “foreseeing maintenance and management policy.”

第一章 緒論
第一節 研究背景 1
第二節 研究動機與目的 4
第三節 論文結構 5
第二章 文獻探討
第一節 發動機性能衰降因素 6
第二節 飛機發動機組件更換預測 8
第三節 估計發動機組件失效時間的預測技術 11
第三章 蒐集及處理發動機性能衰降資料過程 17
第四章 實證評估
第一節 資料來源 21
第二節 模式效能評估準則 22
第三節 評估程序 24
第四節 預測結果與分析 26
第五節 討論 44
第五章 結論
第一節 研究結論及貢獻 46
第二節 未來研究方向 47
參考文獻
一、中文部份 48
二、英文部份 48

一、中文部份
1.方淳民、毛昭家，「基地之外場維修及品質管理探討」，空軍學術月刊，第573期，2004年。
2.侯光華，「長時間高溫使用之Inconel 718熱影響區裂紋成因與銲補製程研究」，行政院國家科學委員會專題研究計畫報告，NSC89-2216-E182-017，2001年。
3.柳耀華、張仁孚、蔡源成，「新世代武器系統自主式後勤支援管理架構概述與實例研究」，海軍學術月刊，第39卷，第2期，2005年。
4.郭祥之，「散熱風扇之性能曲線擬合分析」，中原大學機械工程學系碩士論文，2005年。
5.陳寬裕、何嘉惠、蕭宏誠，「應用支援向量回歸於國際旅遊需求之預測」，旅遊管理研究，第4卷，第1期，2004年，頁81- 97。
6.韓歆儀，「應用兩階段分類法提昇SVM法之分類準確率」，國立成功大學工業與資訊管理研究所碩士論文，2004年。

二、英文部份
1.Boyce, M. P., Gas Turbine Engineering Handbook, Gulf Professional Publishing, Houston, TX, 2001.
2.Carter, T. J., “Common Failures in Gas Turbine Blades,” Engineering Failure Analysis, Vol. 12, No. 2, 2005, pp.237-247.
3.Davies, P. C., “Design Issues in Neural Network Development,” Neurovest Journal, Vol. 5, No. 1, 1994, pp.21-25.
4.Denny, G., “F16 Jet Engine Trending and Diagnostics with Neural Networks,” Proceedings of the SPIE (International Society for Optical Engineering), 1993, pp.419-422.
5.Dietz, W. E., Kiech, E. L., and Ali, M., “Jet and Rocket Engine Fault Diagnosis in Real Time,” Journal of Neural Network Computing, Vol. 1, No. 1, 1989, pp.5-18.
6.Fish, K. E., Barnes, J. H., and Aiken, M. W., “Artificial Neural Networks - A New Methodology for Industrial Market Segmentation,” Industrial Marketing Management, Vol. 24, No. 5, 1995, pp.431-438.
7.Frank, E., Wang, Y., Inglis, S., Holmes, G., and Witten, I. H., “Using Model Trees for Classification,” Machine Learning, Vol. 32, No. 1, 1998, pp.63-76.
8.Freeman, J. A. and Skapura, D. M., Neural Networks Algorithms, Applications, and Programming Techniques, Addison-Wesley, Reading, MA, 1992.
9.GE Industrial Sensing, “Humidity Measurement in Gas Turbines,” Application Note 930-105A, April 2005 (available at: http://www.gesensing.com/products/ resources/application_notes_new/930-105A.pdf).
10.General Electric Company, “CT7-2D1 Turboshaft Engines Operating Instruction,” Revision 6, SEI-569, GE Aircraft Engines Customer Technical Training, June 1989.
11.Gunn, S. R., “Support Vector Machines for Classification and Regression,” Technical Report, Department of Electronics and Computer Science, University of Southampton, 1998.
12.Hatsipantelis, E., Murray, A., and Penman, J., “Comparing Hidden Markov Models with Artificial Neural Network Architectures for Condition Monitoring Applications,” Proceedings of the Fourth International Conference on Artificial Neural Networks, Cambridge, UK, 1995, pp.369-374.
13.Haykin, S., Neural Networks: A Comprehensive Foundation, Prentice Hall, Upper Saddle River, NJ, 1999.
14.Kulikov, G. G., Arkov, V. Y., and Breikin, T. V., “On Condition Monitoring of FADEC Information Channels (in Russian),” Izvestiya vuzov. Aviatsionnaya tehnika, N4, 1995.
15.Kulikov, G. G., Arkov, V. Y., and Breikin, T. V., “On Markov Model Applications in Aircraft Gas Turbine Engine Full Authority Digital Controller Test-beds,” Proceedings of the UKACC International Conference on Control '96, Vol. 1, 1996, pp.120-124.
16.Létourneau, S., Famili, F., and Matwin, S., “Data Mining for Prediction of Aircraft Component Replacement,” IEEE Intelligent Systems, Vol. 14, No. 6, 1999, pp.59-66.
17.Lippmann, R. P., “An Introduction to Computing with Neural Nets,” IEEE ASSP Magazine, Vol. 4, No. 2, 1987, pp.4-22.
18.Merrington, G., Kwon, O. K., Goodwin, G., and Carlsson, B.,“Fault Detection and Diagnosis in Gas Turbines,” Journal of Engineering for Gas Turbines and Power, Vol. 113, No. 4, 1991, pp.276-282.
19.Naeem, M., Singh, R., and Probert, D., “Implications of Engine Deterioration for a High-Pressure Turbine-Blade’s Low-Cycle Fatigue (LCF) Life-Consumption,” International Journal of Fatigue, Vol. 21, No. 8, 1999, pp.831-847.
20.Prescott, W. E., “CT7-2D1 Training Guide,” SEI-763, GE Aircraft Engines Customer Technical Training, June 1990.
21.Romeyn, A., “Analysis of Aircraft Propulsion System Failure,” Proceedings of the Australian Society of Air Safety Investigators Conference, Queensland, Australia, 2004.
22.Rumelhart, D. E., Hinton, D. E., and Williams, R. J., “Learning Internal Representation by Error Propagation in Parallel Distributed Processing,” In Parallel Distributed Processing: Explorations in the Microstructure of Cognition, Vol. 1, Rumelhart, D. E. and McClelland, J. L. (Eds.), MIT Press, Cambridge, MA, 1986, pp.318-362.
23.Smola, J. A. and Schölkopf, B., “A Tutorial on Support Vector Regression,” Statistics and Computing, Vol. 14, No. 3, 2004, pp.199-222.
24.Smyth, P., “Hidden Markov Models for Fault Detection in Dynamic Systems,” Pattern Recognition, Vol. 27, No. 1, 1994, pp.149-164.
25.Vapnik, V., Golowich, S., and Smola, A., “Support Vector Method for Function Approximation, Regression Estimation and Signal Processing,” In Advances in Neural Information Processing Systems, Mozer, M., Jordan, M., and Petsche. T. (Eds.), MIT Press, Cambridge, MA, 1997, pp.281-287.