1979年三哩島核電廠事故(Three Mile Island, TMI)發生，由於運轉員並未察覺造成電廠異常的主要原因，加上未採取正確的舒緩措施，導致嚴重的爐心熔毀事故(core melt accident)。由此可知，對於核電廠事故發生後當下的判斷以及後續的處理是有相當程度的關聯性的。
近幾年來，有許多學者的研究都在關注核電廠的肇始事件，並有多人採用類神經網路去進行事件分析。由於類神經網路的計算量較高，且需要大量的時間去學習，因此有了他種分析方法之需求。並且探討多訊號分類演算法(Multiple Signal Classification, MUSIC)在核災事故方面的適用性。
本研究使用MAAP 5程式模擬核電廠在各種位置、面積下的破管事件數據，以建立事故鑑別系統所需的資料庫，利用感測器空間訊號的差異量為主要想法，取事件前60秒的數據，並以MUSIC演算法作為偵測方式，建立偵測系統，期望能在事件發生初期，便能準確判斷破口定位。
本研究將破口事件分為18類，事件由四種感測器組合而成，最後流率感測器的辨識率對於所有破管事件都達到了95 %以上，因此可證明MUSIC演算法是能適用於核電廠的事故偵測上的。

Three Mile Island nuclear power plant(NPP) accident occurred in 1979.Since the operator made a wrong decision, causing continuous human error, leading to the serious result, Core Melt. After the nuclear power plant accident occurred, the current judgment and subsequent treatment are quite relevant. However, during the occurrence of an event, sensor readings normally undergo high-level oscillations that make this manual examination challenging. Operator can’t analysis such a large amount of information in the short time. Their all decisions depend on years of analysis experience.
In recent years , many experts have used different kinds of neural networks to establish identification system and have great performance .Because neural network have high amount of calculation to learn how to detection accidents. It extends the use of the MUSIC algorithm to judge important accidents in nuclear power plants with fewer calculation. In this issue we discuss the problem of pipe break accident and the applicability of MUSIC algorithm in nuclear power plants.
Modular Accident Analysis Program5(MAAP 5) is used to simulate the different things, locations and area of nuclear power plant accident, to establish the data base. With these data base, this project develops an accident identification system based on MUSIC (MUltiple Signal Classification) algorithm. This system can identify the accidents during the early stage and help the operator remedy the accident.
All accident of nuclear power plant are divided into 18 categories .The result of flow rate sensor can recognize the accident of nuclear power plant, and the recognition rate already reach 95 %. According to the result, MUSIC can be used to detect the accident of nuclear power plant.

[1] CNCS, "Fukush task force report," INFO-0824, 2011.
[2] S.-C. Wu, K.-Y. Chen, T.-H. Lin, and H.-P. Chou, "Multivariate algorithms for initiating event detection and identification in nuclear power plants," Annals of Nuclear Energy, vol. 111, pp. 127-135, 2018.
[3] 邱永芳, "MUSIC演算法應用於高頻雷達線性陣列天線訊號解算之研究," 國立成功大學, 2012.
[4] B. M. Wise and N. B. Gallagher, "The process chemometrics approach to process monitoring and fault detection," Journal of Process Control, vol. 6, no. 6, pp. 329-348, 1996.
[5] R. Schmidt, "Multiple emitter location and signal parameter estimation," IEEE transactions on antennas and propagation, vol. 34, no. 3, pp. 276-280, 1986.
[6] 張安成 and 陳志忠, "雙基地多輸入多輸出雷達之聯合 DOD 和 DOA 估測," 資訊科技國際期刊, vol. 7, no. 2, pp. 70-76, 2013.
[7] S. C. Wu, A. L. Swindlehurst, and Y. C. Yao, "Direct interference suppression in EEG/MEG dipole source localization," in 2010 IEEE International Conference on Acoustics, Speech and Signal Processing, 2010, pp. 574-577: IEEE.
[8] M. G. Na, S. H. Shin, D. W. Jung, S. P. Kim, J. H. Jeong, and B. C. Lee, "Estimation of break location and size for loss of coolant accidents using neural networks," Nuclear Engineering and Design, vol. 232, no. 3, pp. 289-300, 2004.
[9] 張鴻仁, "使用圖樣比對法進行核能電廠暫態辨識," 碩士論文, 國立清華大學, 2010.
[10] R. E. Uhrig and J. Hines, "Computational intelligence in nuclear engineering," Nuclear Engineering and Technology, vol. 37, no. 2, pp. 127-138, 2005.
[11] 台灣電力公司第二核能發電廠, "沸水式核能電廠," 1979.
[12] D. Roverso, "Soft computing tools for transient classification," Information Sciences, vol. 127, no. 3-4, pp. 137-156, 2000.
[13] T.-H. Lin, S.-C. Wu, K.-Y. Chen, and H.-P. Chou, "Feature extraction and sensor selection for NPP initiating event identification," Annals of Nuclear Energy, vol. 103, pp. 384-392, 2017.
[14] 劉盈廷, "以MAAP程式驗證龍門電廠在預期暫態未急停事故情況下之緊急操作程序書," 清華大學, 2008.
[15] D. Roverso, "Soft computing tools for transient classification," Information Sciences, vol. 127, no. 3-4, pp. 137-156, 2000.
[16] D. Roverso, "Plant diagnostics by transient classification: The aladdin approach," International Journal of Intelligent Systems, vol. 17, no. 8, pp. 767-790, 2002.
[17] M. J. Embrechts and S. Benedek, "Hybrid identification of nuclear power plant transients with artificial neural networks," IEEE Transactions on Industrial Electronics, vol. 51, no. 3, pp. 686-693, 2004.
[18] T. Santosh, G. Vinod, R. Saraf, A. Ghosh, and H. Kushwaha, "Application of artificial neural networks to nuclear power plant transient diagnosis," Reliability Engineering & System Safety, vol. 92, no. 10, pp. 1468-1472, 2007.
[19] K. Mo, S. J. Lee, and P. H. Seong, "A dynamic neural network aggregation model for transient diagnosis in nuclear power plants," Progress in Nuclear Energy, vol. 49, no. 3, pp. 262-272, 2007.
[20] 陳光佑, "壓水式反應器肇始事件偵測與辨識之研究," 清華大學, 2007.
[21] K.-C. Kwon and J.-H. Kim, "Accident identification in nuclear power plants
using hidden Markov models," Engineering Applications of Artificial Intelligence, vol. 12, no. 4, pp. 491-501, 1999.
[22] G. P. Choi, K. H. Yoo, J. H. Back, and M. G. Na, "Estimation of LOCA break
size using cascaded fuzzy neural networks," Nuclear Engineering and Technology, vol. 49, no. 3, pp. 495-503, 2017.
[23] 楊昭義，核能發電工程學，水牛出版社，1997
[24] L. Li, Y. Zhang, W. Tian, G. Su, and S. Qiu, "MAAP5 simulation of the PWR
severe accident induced by pressurizer safety valve stuck-open accident," Progress in Nuclear Energy, vol. 77, pp. 141-151, 2014.