透過您的圖書館登入
IP:3.17.176.70
  • 學位論文

進步型沸水式電廠冷卻水流失事故分類系統設計

Classification System Design of Loss of Coolant Accident for Advanced Boiling Water Reactor

指導教授 : 周懷樸
若您是本文的作者,可授權文章由華藝線上圖書館中協助推廣。

摘要


在2011年福島核災事故後,對於運轉員面臨嚴重事故的處理便受到更大的關注,通常運轉員在控制室遇到暫態事故發生時,僅能從儀控盤面上的各種參數燈號,再搭配自己的知識與經驗來判斷是何種事故或暫態,進而依照程序書來做電廠異常狀態排除。因此,如果我們能精準且快速的分類事故初期的暫態事故類型,便可以有效的幫助運轉員操作系統回到安全狀態,降低嚴重事故發生的風險。 本研究將使用人工智慧網路系統,藉由系統一些主要參數來做事故與暫態的判斷,期望在系統的協助下,在暫態發生後短時間內,即能辨識出結果供運轉員參考。在福島事件後的經驗:電廠發生地震與海嘯這類複合式天然災變時,對於電廠的影響是整個系統的面衝擊,如電廠全黑事故與冷卻水流失事故同時發生,在此狀況下,儀控盤面也無法正常顯示電廠相關異常警示訊息,運轉員面對這類事故的控管更是一大挑戰,或是複合式災變發生時,許多燈號同時亮起,運轉員面對此混亂情境的誤動作風險也會相對上升。 研究的目標電廠為龍門電廠,主要分析的暫態為冷卻水流失事故,在事故引發跳機事件後,以條件式辨識系統來判斷該暫態是否為冷卻水流失事故,並在確立該事故為冷卻水流失事故後,由建置在儀控盤模擬器上的系統參數記錄器來記錄參數變化值,該記錄值將做為人工智慧網路:支持向量分類法的輸入值,根據系統參數的紀錄值來診斷破管位置為主蒸汽管路還是飼水管路。供運轉員於冷卻水流失事故發生後排除異常之參考訊息,進而幫助運轉員更快速操作使電廠回到安全狀態。

並列摘要


After Fukushima nuclear accident, there has been increasing concern regarding monitoring and management of severe accident. When transients or accidents happened in nuclear power plant, plant operator will try to identify transients by observing the trend of some important parameters. However, under the accident scenario, operator will face with hundreds of alarms and warning information, which might cause confusion and raise the risk of operational error. Therefore, accurately and fast classification of the initiating event is an important and valuable information to successfully manage the severe accident. With the result of classification, plant operators can follow the consequence to find out the sequence of management from emergency operating procedure (EOP). In order to classify loss of coolant accident (LOCA), present research employs the rule-based classification system and artificial intelligence (AI) techniques to diagnose accidents. Taipower Lungmen nuclear power station (LNPS), an advanced boiling water reactor (ABWR), is chosen as the target plant. The AI approach is to construct the database of operators’ knowledge and then make classification based on the value and trend of important operation parameters. Demonstration has shown that the present technique is a feasible approach for events classification.

參考文獻


[1] Santosh G. Vinod, A.K. Babar, H.S. Kushwaha, V. Venkat Raj, 2003, “Symptom based diagnostic system for nuclear power plant operations using artificial neural networks” , Reliability Engineering and System Safety, vol. 82, p.p.33–40.
[2] 蕭德勇,「壓水式反應器暫態辨識研究」,國立清華大學,博士論文,2011
[3] M. G. Na, S. M. Lee, S. H. Shin, and D. W. Jung, S. P. Kim, J. H. Jeong, and B. C. Lee, 2004, “Prediction of Major Transient Scenarios for Severe Accidents of Nuclear Power Plants” , IEEE Trans. Nucl. Sci., vol. 51, no. 2, p.p. 313-321.
[5] S. H. Lee, Y. G. No, M. G. Na, K.-I. Ahn and S.-Y. Park, 2011, “Diagnostics of Loss of Coolant Accidents Using SVC and GMDH models”, IEEE Trans. Nucl. Sci., vol. 58, no. 1, p.p. 267-276.
[6] Y. G. No, J. H. Kim, M. G. Na, D. H. Lin, and K.-I. Ahn, 2012, “Monitoring Severe Accidents Using AI Techniques”, Nucl. Eng. Tech., vol. 44, no.4, pp.44, pp.393-404.

延伸閱讀