離岸風電工程作業是複雜且擁有動態特性的，為了降低其危害風險事故的發生，除了工程技術門檻，也需降低作業人員發生事故的比率，為此台灣目前對於離岸風電施工安全的相關準則仍未完善，國外在離岸風電產業有相當豐富的經驗，如何借鏡這些經驗與台灣的氣候、海象與水文環境作結合，對此建立應用於台灣離岸風電施工作業的安全評估技術，是一門相當重要的研究議題。
安全風險評估多半有賴於專家經驗、知識與相關事件資料，然而缺少經驗豐富的專家以及事件資料不足時，如何建立明確的風險因子結構，並量化與管控風險因子，實屬重要之研究課題。本研究將以離岸風電工程作業中的｢駁船運送的葉片、機艙與塔架轉移到自升式工作船｣為例，利用危害與可操作性分析的專家知識找出因子間的因果關係，並轉換成動態貝氏網路的初始結構，接著使用EM演算法做參數學習完成條件機率表，完成動態貝氏網路之建構。研究結果顯示出3種情境分析的作業風險機率都呈現出符合事故金字塔理論的概念，因此可提供決策者進行暫停施工的參考，並能讓決策者找出人員傷亡機率低且重新作業的時間點。

Operation of offshore wind engineering is complex and has dynamic characteristics. In order to reduce the risk of accidents, in addition to improving engineering technology, it is also necessary to reduce the rate of accidents among operators. Taiwan's safety guidelines for the offshore wind power construction is still premature. Foreign countries have considerable experiences in offshore wind industry. How to combine these experiences with Taiwan's climate, marine meteorology and hydrological environment to develop safety assessment process for the operations of offshore wind power construction in Taiwan is an important research topic.
Risk assessment relies mainly on expert's experience, knowledge and related event data. However, when there are lack of experienced experts and insufficient event data, how to develop a clear structure of risk factors, quantify and control risk factors? This study uses the expert knowledge of Hazard and Operability Study to find causal relationship between factors in the offshore wind engineering for the operations of transferring blades, nacelles and towers from barge to self-elevating vessel. They are converted into initial structure of the dynamic Bayesian network, and EM algorithm is used to conduct parameter learning to complete conditional probability table of the dynamic Bayesian network. The results show that operational risk probabilities of all three scenarios analyzed conform to the concept of Heinrich's Pyramid theory. Thus it can assist decision making on whether to suspend operations, and enable decision makers to find out the time to re-operate when probability of casualty is low.

論文審定書 i
誌謝 ii
摘要 iii
Abstract iv
目錄 v
圖目錄 vii
表目錄 ix
第一章、緒論 1
1.1 研究動機 1
1.2 研究目的 2
1.3 研究流程 3
1.4 研究範圍 5
第二章、文獻回顧 7
2.1風險評估 7
2.2離岸風電的安全管理 11
2.3危害與可操作性分析 13
2.4 (動態)貝氏網路 14
第三章、研究方法與模型建置 17
3.1危害與可操作性分析 17
3.2動態貝氏網路 21
3.2.1貝氏網路 21
3.2.2動態貝氏網路基礎 24
3.2.3動態貝氏網路推論 26
3.2.4動態貝氏網路學習 29
3.2.5 GeNIe介紹 31
3.3 HAZOP轉換成動態貝氏網路架構 31
3.4參數學習建構條件機率表 46
第四章、研究成果 55
4.1情境分析 55
4.2結果與討論 80
第五章、結論與建議 82
5.1結論 82
5.2建議 84
參考文獻 85
附錄：HAZOP分析表 92
附錄二：蒲福風級表 95
意見回覆表 96

中文文獻
1. 江峰政 (2005)。有關EM法於具韋伯之設限資料其參數之估計，碩士論文，國立成功大學數學系應用數學碩博士班。
2. 吳士珍、陳俊瑜 (2001)。危害與可操作分析（Hazop）在化工業之應用。化工本質安全設計與應用專刊第48 卷第四期。p90-99。
3. 周顯光、鍾承憲 (2015)。離岸風場作業安全評估技術開發計畫。經濟部能源局研究機構能源科技專案(編號：103-D0106)。
4. 周顯光、鍾承憲 (2017)。離岸風場作業安全評估技術開發計畫。經濟部能源局研究機構能源科技專案(編號：105-D0106)。
5. 林怡姍 (2012)。考量人為因素之貝葉氏網路風險評估模型之建構-以連續模糊集合為基礎，碩士論文，國立臺灣海洋大學商船學系。
6. 柯俊宏 (2008)。運用貝氏網路建構不合格品質偵測輔助系統，碩士論文，中興大學土木學系所。
7. 張聖坤、白勇、唐文勇 (2003)。船舶與海洋工程風險評估，國防工業，北京市。
8. 黃清賢 (1996)。危害分析與風險評估(初版)。三民出版社。
9. 黃清賢 (2003)。危害分析與風險評估操作手冊(初版)。臺北縣：新文京。
10. 葉宇光 (2009)。事件樹於職業安全風險評估應用研究，碩士論文，國立中央大學環境工程研究所。
11. 潘家琳 (2006)。利用動態貝氏網路分析基因網路的調控，碩士論文，慈濟大學醫學資訊研究所。
12. 鄭祐昇 (2004)。運用資料探勘技術於知識圖之建立，碩士論文，東海大學工業工程與經營資訊研究所。
13. 賴瑞菊 (2011)。風險可接受度探討，碩士論文，國立中央大學環境工程研究所。

英文文獻
1. Ahlgren, E., & Grudic, E. (2017). Risk Management in Offshore Wind Farm Development.
2. Aspinall, P. (2006). HAZOPs and human factors, in: HAZARDS XIX – Process Safety and Environmental Protection Symposium. Institution of Chemical Engineers (IChemE), Manchester, UK, 28–30 March.
3. Baybutt, P. (2002). Layers of protection analysis for human factors (LOPA-HF). Process Safety Progress 21 (2), 119–129.
4. Bendixen, L., & O’Neill, J.K. (1984). Chemical plant risk assessment using HAZOP and fault tree methods. Plant/Operations Progress 3 (3), 179–184.
5. Boudali, H., & Dugan, J.B. (2005a). A new Bayesian network approach to solve dynamic fault trees. In: Proceedings of the IEEE Reliability and Maintainability Symposium. 451–456, January 24–27.
6. Boudali, H., & Dugan, J.B. (2005b). A discrete-time Bayesian network reliability modeling and analysis framework. Reliability Engineering and System Safety 87 (3), 337–349.
7. Chaiwat, P. (2016). Probabilistic Risk Assessment of Offshore Production Platform by Bayesian Network Application to HAZOP and Bow-tie Studies. (Master), Texas A & M University.
8. Dai, L., Ehlers, S., Rausand, M., & Utne, I. B. (2013). Risk of collision between service vessels and offshore wind turbines. Reliability Engineering & System Safety, 109, 18-31.
9. Demichela, M., Marmo, L., & Piccinini, N. (2002). Recursive operability analysis of a complex plant with multiple protection devices. Reliability Engineering and System Safety 77 (3), 301–308.
10. Dempster, A.P., Laird, N.M., & Rubin, D.B. (1977). Maximum likelihood from incomplete data via the EM algorithm (with discussion)
11. Dunjó, J., Fthenakis, V., Vílchez, J.A., & Arnaldos, A. (2010). Hazard and operability (HAZOP) analysis. A literature review. Journal of Hazardous Materials, 173, 19-32.
12. Embrey, D. (2002).Using influence diagrams to analyse and predict failures in safety critical systems. In: Proceedings of the 23rd ESReDA Seminar—Decision Analysis: Methodology and Applications for Safety of Transportation and Process Industries. Delft, The Netherlands, November 2002.
13. Enrique, C., Jose, M.G., & Ali, S.H. (1997), Expert Systems and Probabilistic Network Models, Springer-Verlag, New York.
14. EU-OSHA. (2013). Occupational safety and health in the wind energy sector. Luxembourg: Publications Office of the European Union: European Agency for Safety and Health at Work.
15. Furness, R. W., Wade, H. M., & Masden, E. A. (2013). Assessing vulnerability of marine bird populations to offshore wind farms. Journal of environmental management, 119, 56-66.
16. G+ Global offshore wind health and safety organisation. (2017). G+ Global offshore wind health and safety organisation 2016 incident data report: Energy Institute, London.
17. G9 Offshore wind health and safety association. (2015). G9 Offshore wind health and safety association 2014 incident data report: Energy Institute, London.
18. G9 Offshore wind health and safety association. (2016). G9 Offshore wind health and safety association 2015 incident data report: Energy Institute, London.
19. G9 Offshore Wind Health and Safety. (2014). Good Practice Guideline Working At Height In The Offshore Wind Industry: Energy Institute, London.
20. Greenberg, H.R., & Cramer, J.J. (1991). Risk Assessment and Risk Management for The Chemical Process Industry. Van Nostrand Reinhold, New York.
21. Gulvanessian, H., & Holicky, M. (2001). Determination of actions due to fire: recent developments in Bayesian risk assessment of structures under fire. Building Research Establishment, Garston, Watford, UK. Klokner Institute, Prague, Czech Republic.
22. Haluzan, N. (2011). Offshore wind power–Advantages and disadvantages. Renewable Energy Articles.
23. Heckerman, D. (2008). A tutorial on learning with Bayesian networks. In Innovations in Bayesian networks (pp. 33-82): Springer.
24. Hu, J., Zhang, L., Cai, Z., & Wang, Y. (2015). An intelligent fault diagnosis system for process plant using a functional HAZOP and DBN integrated methodology. Engineering Applications of Artificial Intelligence, 45, 119-135.
25. Hu, J., Zhang, L., Ma, L., & Liang, W. (2011). An integrated safety prognosis model for complex system based on dynamic Bayesian network and ant colony algorithm. Expert Systems with Applications, 38, 1431-1446.
26. Hulst, J. (2006). Modeling physiological processes with dynamic Bayesian networks. (Master), Delft University of Technology, Delft, Netherlands.
27. Jensen, F.V., & Nielsen, T.D. (2007). Bayesian networks and decision graphs. (2nd ed.) Springer, New York, NY.
28. Kennedy, R., & Kirwan, B. (1998). Development of a hazard and operability-based method for identifying safety management vulnerabilities in high risk systems, Safety Science 30 (3) 249–274.
29. Khakzad, N. (2015). Application of dynamic Bayesian network to risk analysis of domino effects in chemical infrastructures. Reliability Engineering and System Safety, 138, 263-272.
30. Khan, F. I., & Abbasi, S. A. (1998). Techniques and methodologies for risk analysis in chemical process industries. Journal of Loss Prevention in the Process Industries, 11(4), 261-277.
31. Kikuchi, R. (2010). Risk formulation for the sonic effects of offshore wind farms on fish in the EU region. Marine Pollution Bulletin, 60(2), 172-177.
32. Kim, M.C., & Seong, P.H. (2006). A computational method for probabilistic safety assessment of I&C systems and human operators in nuclear power plants. Reliability Engineering and System Safety, 91 (5), 580–593.
33. Kim, M.C., Seong, P.H., & Hollnagel, E. (2006). A probabilistic approach for determining the control mode in CREAM. Reliability Engineering and System Safety, 91 (2), 191–199.
34. Kletz, T., (1974). HAZOP and HAZAN-Notes on the Identification and Assessment of Hazards. Institute of Chemical Engineers, Rugby.
35. Kletz, T.A. (1998). Process Plants - A Handbook for Inherently Safer Design. 2nd Edition. Taylor & Francis, Philadelphia.
36. Lauritzen, S., Thiesson, B., & Spiegelhalter, D. (1994), Diagnostic systems created by model selection methods: A case study, In Cheeseman, P. and Oldford, R., editors, AI and Statistics IV, volume Lecture Notes in Statics, 89, pages 143-152. Springer Verlag. New York.
37. Lloyd, P. (2016). Health and safety of offshore wind farms. In Offshore Wind Farms (pp. 573-587): Elsevier.
38. McKelvey, T.C., (1988) "How to Improve the Effectiveness of Hazard and Operability Analysis", IEEE Transaction on Reliability, Vol. 37, No. 1, pp. 167-170.
39. Modarres, M., Kaminskiy, M. P., & Krivtsov, V. (2016). Reliability engineering and risk analysis: a practical guide: CRC press.
40. Montani, S., Portinale, L., Bobbio, A. Codetta-Raiteri, D. (2008). RADYBAN: a tool for reliability analysis of dynamic fault trees through conversion into dynamic Bayesian networks. Reliability Engineering and System Safety, 93,922–32.
41. Montani, S., Portinale, L., Bobbio, A., Varesio, M., & Codetta-Raiteri, D. (2006). A tool for automatically translating Dynamic Fault Trees into Dynamic Bayesian Networks. In: Reliability and Maintainability Symposium (RAMS 2006), 434–441.
42. Murphy, K. P., & Russell, S. (2002). Dynamic bayesian networks: representation, inference and learning.
43. Neapolitan, R. (2003). Learning Bayesian networks. Prentice Hall, Inc., Upper Saddle River, NJ, USA.
44. Ng, C., & Ran, L. (2016). Offshore wind farms: Technologies, design and operation: Woodhead Publishing.
45. Norrington, L., Quigley, J., Russel, A., & Van der Meer, R. (2007). Modeling the reliability of search and rescue operations with Bayesian Belief Networks. Reliability Engineering and System Safety, 93 (7), 940–949.
46. Øien, K. (2001). A framework for the establishment of organizational risk indicators. Reliability Engineering and System Safety, 74, 147–168.
47. Ozog, H. (1985). Hazard identification, analysis and control: a systematic way to assess potential hazards helps promote safer design and operation of new and existing plants. Chemical Engineering 92 (4), 161–170.
48. Ozog, H., & Bendixen, L.M. (1987). Hazard identification and quantification: the most effective way to identify, quantify, and control risks is to combine a hazard and operability study with fault tree analysis. Chemical Engineering Progress 83 (4), 55–64.
49. RenewableUK. (2014). Offshore Wind and Marine Energy Health and Safety Guidelines, RenewableUK.
50. Røed, W., Mosleh, A., Vinnem, J.E., & Aven, T. (2008). On the use of hybrid causal logic method in offshore risk analysis. Reliability Engineering and System Safety, 94 (2), 445–455.
51. Russell, S. J., & Norvig, P. (2010). Artificial Intelligence (A Modern Approach) (Third ed.): Prentice Hall.
52. Schurman, D.L., & Fleger, S.A. (1994). Human factors in HAZOPs: guide words and parameters. Professional Safety 39 (12), 32–34.
53. Shafiee, M. (2015). A fuzzy analytic network process model to mitigate the risks associated with offshore wind farms. Expert Systems with Applications, 42(4), 2143-2152.
54. Suk H-I, Sin B-K, & Lee S-W (2010) Hand gesture recognition based on dynamic Bayesian network framework. Pattern Recogn 43(9):3059–3072.
55. Trucco, P., Cagno, E., Ruggeri, F., & Grande, O. (2008). A Bayesian belief network modelling of organisational factors en risk analysis: a case study in maritime transportation. Reliability Engineering and System Safety, 93 (6), 845–856.
56. Tveiten, C. K., Albrechtsen, E., Heggset, J., Hofmann, M., Jersin, E., Leira, B., & Norddal, P. K. (2011). HSE challenges related to offshore renewable energy.
57. Wang, J., & Kieran, O., 2000. Offshore Safety Assessment and Safety-Based Decision-Making ─ The Current Status and Future Aspects. Journal of Offshore Mechanics and Arctic Engineering, 122(2), 93-99.
58. Wang, J., Sii, H. S., Yang, J. B., Pillay, A., Yu, D., Liu J., Maistralis E., & Saajedi, A., (2004). Use of Advances in Technology for Maritime Risk Assessment. Risk Analysis, 24(4), 1041-1063.
59. Weber, P., Medina-Oliva, G., Simon, C., & Iung, B. (2012). Overview on Bayesian networks applications for dependability, risk analysis and maintenance areas. Engineering Applications of Artificial Intelligence, 25, 671-682.
60. Wu, W., Yang, C., Chang, J., Château, P., & Chang, Y. (2015). Risk assessment by integrating interpretive structural modeling and Bayesian network, case of offshore pipeline project. Reliability Engineering and System Safety, 142, 515-524.
61. Yang, G., Lin, Y., & Bhattacharya, P. (2010). A driver fatigue recognition model based on information fusion and dynamic Bayesian network, Information Sciences, 180 (10), 1942–1954.