臺灣博碩士論文加值系統

基於數據驅動的故障診斷結合機器學習、資料探勘與大數據的演算法，以人工智慧技術取代診斷專家，更快速、有效地提供診斷結果，然而，欲透過智慧診斷得出準確的結果，僅基於數據驅動是不夠的，因過程不仰賴經驗知識，但設備中的問題常常隱含大量的知識及物理意義，因此完整的智慧診斷專家系統，應結合數據驅動、專家知識與物理模型三者，以確保分析結果不相違背。
本研究以往復式壓縮機為對象，建立失效與監控數據診斷關聯規則知識庫。期望將有備診斷的敘述性知識，轉換為可應用於專家系統中診斷的規則，以輔助演算方法的執行。由於執行正確的分析也是關鍵因素，因此本研究中以一往復式壓縮機振動參數，進行失效模式鑑別，並比較支援向量機、最近鄰居分類法及類神經網路等不同演算法，以探討往復式壓縮機失效模式鑑別的合適性。
診斷關聯規則知識庫的開發，可經由資料蒐集與彙整、知識庫欄位設計、資料擷取、資料轉換、關聯分析及結果輸出等步驟執行，最終則以量化的方式表示，當單個參數或雙參數同時發生異常時對應失效模式、失效原因、故障部位出現的機率做為診斷的基礎。針對失效模式鑑別之實驗結果顯示，支援向量機、類神經網路此兩種演算法，僅採用原始數據即可達到76%以上的準確度。未來若搭配合適訊號處理方法，可突顯訊號中的異常特徵(anomaly features)，則可進一步提升準確度。
未來可針對不同演算法進行開發，透過集成式演算法(Ensemble Learning Algorithms)，以發揮各自所長，並持續納入其他有關專家知識、物理模型等的輔助判斷基準，以提升診斷的準確度。

Data-driven fault diagnosis has become a trend that combines machine learning, data mining and big data algorithms to replace diagnostic experts with artificial intelligence to deliver diagnostic results more quickly and efficiently. However, it is not enough to rely on data-driven to obtain accurate results through intelligent diagnosis. Because it does not rely on empirical knowledge, the problems in the equipment often imply a large amount of knowledge and physical meaning. Therefore, a complete intelligent diagnostic expert system should combine data-driven, expert knowledge and physical models to ensure that the analysis results are not inconsistent.
In this study, a reciprocating compressor is used as a target to establish a knowledge base of failure and monitoring data diagnostic rules. It is desirable to explore a way to translate narrative knowledge about device diagnostics into rules that can be applied to diagnostics in expert systems to aid in the execution of algorithms. On the other hand, performing the right analysis is also a key point. In the study, the vibration mode of a reciprocating compressor was used to identify the failure mode. Compare the suitability of different algorithms such as Support vector machine, k-nearest neighbors, and neural network to the failure mode identification of reciprocating compressors.
The development of the diagnostic rule knowledge base, through data collection and collection, knowledge base field design, data acquisition, data conversion, correlation analysis and result output.Finally, it can be quantified to indicate that when a single parameter or a double parameter is abnormal at the same time, the failure mode, the cause of failure, and the probability of occurrence of the fault location are used as the basis for diagnosis.According to the experimental results of failure mode identification, the two algorithms of support vector machine and neural network are good for equipment fault diagnosis. The original data can achieve more than 76% accuracy. With the appropriate signal processing method, the anomaly features in the signal can be highlighted to further improve the accuracy.
For the blueprint of complete intelligent diagnosis, the future can be developed for different algorithms. By combining different algorithms, the equipment can be diagnosed in the most objective way. It should continue to include other auxiliary judgment criteria related to expert knowledge, physical models, etc., in order to achieve more effective diagnosis and avoid accidents.

摘要 I
ABSTRACT II
誌謝 IV
目錄 V
表目錄 VII
圖目錄 VIII
第一章 緒論 1
1.1. 研究背景 1
1.2. 研究動機 3
1.3. 研究目的 5
1.4. 論文架構 7
第二章 文獻回顧 9
2.1. 往復式壓縮機介紹 9
2.1.1. 往復式壓縮機的結構與構造 9
2.1.2. 往復式壓縮機的運作原理 16
2.1.3. 往復式壓縮機常見失效 20
2.2. 國際標準 22
2.2.1. API 618 22
2.2.2. ISO14224 22
2.2.3. ISO 17359 23
2.3. 基於數據驅動的智慧診斷 24
2.3.1. 類神經網路 24
2.3.2. 支援向量機 27
2.3.3. 最近鄰居分類法 30
第三章 診斷關聯規則知識庫開發 31
3.1. 建立流程 34
3.1.1. 文獻彙整 37
3.1.2. 欄位設計 41
3.1.3. 資料擷取 42
3.1.4. 資料轉換 45
3.1.5. 關聯分析 46
3.1.6. 資料輸出 47
3.2. 知識庫特性說明 48
第四章 設備故障診斷 50
4.1. 顯性失效模式鑑別 50
4.1.1. 案例數據說明 50
4.1.2. 實驗方法 53
4.2. 隱藏性失效維護策略 55
4.2.1. 隱藏性失效 55
4.2.2. 往復式壓縮機常見隱藏性失效 56
4.2.3. 隱藏性失效維護策略 57
第五章 結果與討論 59
5.1. 失效模式與參數之診斷關聯規則知識庫 59
5.2. 失效模式鑑別 69
5.2.1. 顯性失效模式鑑別 69
第六章 結論與建議 77
6.1. 結論 77
6.2. 建議 79
參考文獻 81
附錄一 84

[1]天下雜誌編輯群，2018，2018天下兩千大調查 智慧生活爭霸戰，台灣：天下雜誌。
[2]Crowl, Daniel A, and Joseph F Louvar. Chemical Process Safety: Fundamentals with Applications. Pearson Education, 2001.
[3]Shu, Yidan, Liang Ming, Feifan Cheng, Zhanpeng Zhang, and Jinsong Zhao. "Abnormal Situation Management: Challenges and Opportunities in the Big Data Era." Computers & Chemical Engineering 91 (2016/08/04/ 2016): 104-13.
[4]Lei, Yaguo. Intelligent Fault Diagnosis and Remaining Useful Life Prediction of Rotating Machinery. Butterworth-Heinemann, 2016.
[5]Hashemian, H. M. "State-of-the-Art Predictive Maintenance Techniques." [In English]. Ieee Transactions on Instrumentation and Measurement 60, no. 1 (Jan 2011): 226-36.
[6]曹常成，王振華，2013，設備維護管理結合設備完整性與風險基礎檢查技術研究，勞動部勞動及職業安全衛生研究所。
[7]Xu, Y., Y. M. Sun, J. F. Wan, X. L. Liu, and Z. T. Song. "Industrial Big Data for Fault Diagnosis: Taxonomy, Review, and Applications." [In English]. Ieee Access 5 (2017): 17368-80.
[8]Heng, A., S. Zhang, A. C. C. Tan, and J. Mathew. "Rotating Machinery Prognostics: State of the Art, Challenges and Opportunities." [In English]. Mechanical Systems and Signal Processing 23, no. 3 (Apr 2009): 724-39.
[9]Zhang, Zhenyou, Yi Wang, and Kesheng Wang. "Intelligent Fault Diagnosis and Prognosis Approach for Rotating Machinery Integrating Wavelet Transform, Principal Component Analysis, and Artificial Neural Networks." Advanced Manufacturing Technology 68, no. 1 (September 01 2013): 763-73.
[10]Ali, Yasir Hassan. "Artificial Intelligence Application in Machine Condition Monitoring and Fault Diagnosis." (2018).
[11]Tran, Van Tung, Faisal AlThobiani, and Andrew Ball. "An Approach to Fault Diagnosis of Reciprocating Compressor Valves Using Teager–Kaiser Energy Operator and Deep Belief Networks." Expert Systems with Applications 41, no. 9 (2014/07/01 2014): 4113-22.
[12]Lee, Jay, Fangji Wu, Wenyu Zhao, Masoud Ghaffari, Linxia Liao, and David Siegel. "Prognostics and Health Management Design for Rotary Machinery Systems—Reviews, Methodology and Applications." Mechanical Systems and Signal Processing 42, no. 1 (2014/01/01/ 2014): 314-34.
[13]Negnevitsky, Michael. Artificial Intelligence : A Guide to Intelligent Systems. 3rd ed. ed. Harlow, England ;: Addison Wesley/Pearson, 2011.
[14]Zhang, Wan, Min-Ping Jia, Lin Zhu, and Xiao-An Yan. "Comprehensive Overview on Computational Intelligence Techniques for Machinery Condition Monitoring and Fault Diagnosis." Chinese Journal of Mechanical Engineering 30, no. 4 (July 01 2017): 782-95.
[15]Dai, X. W., and Z. W. Gao. "From Model, Signal to Knowledge: A Data-Driven Perspective of Fault Detection and Diagnosis." [In English]. Ieee Transactions on Industrial Informatics 9, no. 4 (Nov 2013): 2226-38.
[16]Lei, Y. G., Z. Y. Liu, X. H. Wu, N. P. Li, W. Chen, and J. Lin. "Health Condition Identification of Multi-Stage Planetary Gearboxes Using a Mrvm-Based Method." [In English]. Mechanical Systems and Signal Processing 60-61 (Aug 2015): 289-300.
[17]李傑，2016，工業大數據：工業4.0時代的智慧轉型與價值創新，臺北市：天下雜誌出版，第一版。
[18]Cui, Houxi, Laibin Zhang, Rongyu Kang, and Xinyang Lan. "Research on Fault Diagnosis for Reciprocating Compressor Valve Using Information Entropy and Svm Method." loss prevention in the process industries 22, no. 6 (2009): 864-67.
[19]Griffith, William A, and Edward B Flanagan. "Online, Continuous Monitoring of Mechanical Condition and Performance for Critical Reciprocating Compressors." Paper presented at the Proceedings of the 30th Turbomachinery Symposium, 2001.
[20]鍾蒲生，2000，煉油設備操作實務，中國石油公司訓練所。
[21]API. "Reciprocating Compressors for Petroleum, Chemical, and Gas Industry Services." In API STANDARD 618: American Petroleum Institute, 2007.
[22]Bloch, Heinz P. A Practical Guide to Compressor Technology. John Wiley & Sons, 2006.
[23]Leonard, S.M. "Increase Reliability of Reciprocating Hydrogen Compressors." Hydrocarbon Processing (1996): Medium: X; Size: pp. 67-74.
[24]Goebel, DANIEL. "Reciprocating Compressor Suction and Discharge-Valve Monitoring." Compressor Tech Two (2014): 18-23.
[25]ISO. "Petroleum, Petrochemical and Natural Gas Industries -- Collection and Exchange of Reliability and Maintenance Data for Equipment." In ISO 14224: International Organization for Standardization, 2016.
[26]Kuemmlee, H., T. Gross, and J. Kolerus. "Machine Vibrations and Diagnostics the World of Iso." Paper presented at the Industry Applications Society 60th Annual Petroleum and Chemical Industry Conference, 23-25 Sept. 2013 2013.
[27]Si, Jia, Yibin Li, and Sile Ma. "Intelligent Fault Diagnosis for Industrial Big Data." Signal Processing Systems 90, no. 8-9 (2018): 1221-33.
[28]Liu, Ruonan, Boyuan Yang, Enrico Zio, and Xuefeng Chen. "Artificial Intelligence for Fault Diagnosis of Rotating Machinery: A Review." [In English]. Mechanical Systems and Signal Processing 108 (Aug 2018): 33-47.
[29]簡禎富，2014，許嘉裕譯，資料挖礦與大數據分析，新北市：前程文化，初版。
[30]王文傑，葉世偉，2004，人工智能原理及應用，北京：人民郵電出版社。
[31]Rossi, M., S. Benatti, E. Farella, and L. Benini. "Hybrid Emg Classifier Based on Hmm and Svm for Hand Gesture Recognition in Prosthetics." Paper presented at the 2015 IEEE International Conference on Industrial Technology (ICIT), 17-19 March 2015 2015.
[32]ISO. "Condition Monitoring and Diagnostics of Machines — General Guidelines." In ISO 17359: International Organization for Standardization, 2018.
[33]Loukopoulos, Panagiotis, George Zolkiewski, Ian Bennett, Suresh Sampath, Pericles Pilidis, Fang Duan, and David Mba. "Reciprocating Compressor Valve Leakage Detection under Varying Load Conditions." Cham, 2019.
[34]Giudici, Paolo, and Silvia Figini. Applied Data Mining for Business and Industry. Wiley Online Library, 2009.
[35]黃翊婷，2018，應用於設備智慧化診斷之數據處理與轉換研究，國立高雄第一科技大學，碩士論文。
[36]ABS. Guidance Notes on Equipment Condition Monitoring Techniques. American Bureau of Shipping, 2016.
[37]Wolpert, D. H., and W. G. Macready. "No Free Lunch Theorems for Optimization." IEEE Transactions on Evolutionary Computation 1, no. 1 (1997): 67-82.
[38]Wolpert, David H, and William G Macready. "No Free Lunch Theorems for Search." Technical Report SFI-TR-95-02-010, Santa Fe Institute, 1995.