本研究為龍門核電廠的自動升載系統與運轉員建置了一套輔助系統，其作用在於預測與判斷電廠狀態來避免不當操作。此輔助系統有兩大功能：預測與判斷。預測方面我們採用自適應式模糊類神經網路方法來實現，預測的目標為電廠5分鐘後的爐心壓力與蒸氣流量率；我們把神經網路模組化，爐心壓力預測採用階層的結構。判斷方面採用模糊邏輯方法來讓輔助系統對電廠的狀態進行判斷，其輸出是對功率要求的接受值，接受值高表示電廠狀態較穩定，輔助系統將接受此要求。輔助系統設計完成後，我們使用龍門模擬器來做為訓練資料來源與測試平台；龍門模擬器屬於動態全範圍模擬器，可以實時模擬龍門電廠各種轉情況。系統被進行自動升載測試與爐內幫浦跳脫一台測試；從自動升載測試的結果發現，加入輔助系統比起只用自動升載調節系統，可以有效避免汽機旁通閥開啟暫態；從爐內幫浦跳脫一台測試的結果發現，藉由模糊邏輯的輸入與輸出可以發現其電廠的異常狀態。從測試的結果可以展示，本研究所設計的輔助系統可以在極短的時間，得到合理的電廠參數動態預測，達到協助運轉員判斷與增進電廠安全、效率的目標。

This study presented an auxiliary system for the automatic power regulating system and operators of Lungmen nuclear power plant. The auxiliary system was design to predict and judge the plant state and to avoid improper operation. So the design system had two major functions: prediction and judgment. We used the adaptive neural fuzzy inference system to achieve the prediction function. Predicted targets are the reactor core pressure and steam flow rate of the next 5 minutes. We modularized the adaptive neural fuzzy inference system and proposed a hierarchical structures of the pressure prediction. To make the design system can judge the plant state, we used the fuzzy logic method. The output of judgment was the accept degree of the power demand. After the system was built, we used the Lungmen simulator, which is a full-scope dynamic simulator, as the test bench and the source of the training data. The auxiliary system was tested under the scenarios of the automatically increase power and the one reactor internal pump trip even. According to the result of the automatically increase power scenarios, the auxiliary system could avoid the turbine bypass value open even. Under the one reactor internal pump trip even, the input and output of the fuzzy logic could help operators to detect the abnormal state of the plant. The test result demonstrated that the design auxiliary system can make a reasonable prediction of the plant parameters in a very short time and can assist operators to improve plant safety and efficiency.

[ 1 ] 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
[ 2 ] Y. G. NO,J. H. KIM,M. G. NA,D. H. LIM, and K. I. AHN,“Monitoring Severe Accidents Using Ai Techniques”, Nuclear Engineering And Technology, VOL.44 NO.4 MAY 2012, P.P. 393-403
[ 3 ] A. de Callatay,“Active avoidance of threatening commands to integrate fuzzy systems in nuclear power plants＂, Fuzzy Sets and Systems, Vol. 74, Issue 1, 1995, P.P. 127–137
[ 4 ] 林聰明，「應用類神經網路於動態系統模擬」，國清華大學，碩士論文，1997
[ 5 ] 張世忠，「應用類神經網路建立核能電廠一般系統動態模式」，國立清華大學，碩士論文，1995
[6 ] L. Xinqing, L. H. Tsoukalas, R. E. Uhrig,“A Neurofuzzy Approach For The Anticipatory Control Of Complex Systems”, Proceedings of the Fifth IEEE International Conference on Fuzzy Systems, 1996, P.P. 587-593
[ 7 ] Keith E. Holbert and Kang Lin, “Nuclear Power Plant Instrumentation Fault Detection Using Fuzzy Logic＂, Science and Technology of Nuclear Installations, Vol. 2012, 2012
[ 8 ] R. H. Uhrigm, L. H. Tsoukalas,“Multi-Agent-Based Anticipatory Control For Enhancing The Safety And Performance Of Generation-Iv Nuclear Power Plants During Long-Term Semi-Autonomous Operation”, Nuclear Energy, Vol. 43, No. 1-4, 2003, p.p. 113-120
[ 9 ] M. MARSEGUERRA, E. ZIO, P. AVOGADRI,“Model Identification By Neuro-Fuzzy Techniques: Predicting The Water Level In A Steam Generator Of A PWR＂, Nuclear Energy, Vol. 44, No. 3, 2004, pp. 237-252
[ 10 ] M. G. Na, Member, IEEE, S. H.Shin, S. M. Lee, D. W. Jung, S. P. Kim, J. H. Jeong and B. C. Lee, “Prediction of Major Transient Scenarios for Severe Accidents of Nuclear Power Plants”, IEEE Transactions On Nuclear Science, Vol. 51, No. 2, 2004, p.p. 313-321
[ 11 ] T.C Hsia, system identification, 儒林圖書有限公司,台北市, 1983,chapter 4
[ 12 ] D. Ruan,“Fuzzy logic in the nuclear research world＂, Fuzzy Sets and Systems. Vol. 74, Issue 1, 1995, P.P. 5–13