Nuclear reactor vessel water level prediction during severe accidents using deep neural networks

  • Koo, Young Do (Department of Nuclear Engineering, Chosun University) ;
  • An, Ye Ji (Department of Nuclear Engineering, Chosun University) ;
  • Kim, Chang-Hwoi (Nuclear ICT Research Division, Korea Atomic Energy Research Institute) ;
  • Na, Man Gyun (Department of Nuclear Engineering, Chosun University)
  • Received : 2018.10.28
  • Accepted : 2018.12.24
  • Published : 2019.04.25


Acquiring instrumentation signals generated from nuclear power plants (NPPs) is essential to maintain nuclear reactor integrity or to mitigate an abnormal state under normal operating conditions or severe accident circumstances. However, various safety-critical instrumentation signals from NPPs cannot be accurately measured on account of instrument degradation or failure under severe accident circumstances. Reactor vessel (RV) water level, which is an accident monitoring variable directly related to reactor cooling and prevention of core exposure, was predicted by applying a few signals to deep neural networks (DNNs) during severe accidents in NPPs. Signal data were obtained by simulating the postulated loss-of-coolant accidents at hot- and cold-legs, and steam generator tube rupture using modular accident analysis program code as actual NPP accidents rarely happen. To optimize the DNN model for RV water level prediction, a genetic algorithm was used to select the numbers of hidden layers and nodes. The proposed DNN model had a small root mean square error for RV water level prediction, and performed better than the cascaded fuzzy neural network model of the previous study. Consequently, the DNN model is considered to perform well enough to provide supporting information on the RV water level to operators.


Supported by : National Research Foundation of Korea, Korea Institute of Energy Technology Evaluation and Planning (KETEP)


  1. Y. LeCun, Y. Bengio, G. Hinton, Deep Learning, Nature 521 (2015) 436-444.
  2. D. Goldberg, Genetic Algorithms in Search, Optimization, and Machine Learning, Addison-Wesley, Massachusetts, 1989.
  3. M. Mitchell, An Introduction to Genetic Algorithms, MIT Press, Massachusetts, 1996.
  4. D. Rumelhart, G. Hinton, R. Williams, Learning representations by backpropagating errors, Nature 323 (1986) 533-536.
  5. R. Henry, et al., MAAP4: Modular Accident Analysis Program for the LWR Power Plants, User's Manual, Fauske and Associates Inc., 1994-2005.
  6. D.Y. Kim, K.H. Yoo, G.P. Choi, J.H. Back, M.G. Na, Reactor vessel water level estimation during severe accidents using cascaded fuzzy neural networks, Nucl. Eng. Tech. 48 (3) (2016) 702-710.
  7. M.G. Na, On-line estimation of DNB protection limit via a fuzzy neural network, Nucl. Eng. Tech. 30 (3) (1998) 222-234.
  8. M.G. Na, Y.J. Lee, I.J. Hwang, A smart software sensor for feedwater flow measurement monitoring, IEEE Trans. Nucl. Sci. 52 (6) (2005) 3026-3034.
  9. D. Lewy, ANN Activation Function - Comparison, PyData Warsaw, Warsaw, Poland, 2017. October 18-20, 2017.
  10. A. Ng, K. Katanforoosh, Y.B. Mourri, Neural Networks and Deep Learning, Coursera, 2017.
  11. M.G. Na, W.S. Park, D.H. Lim, Detection and diagnostics of loss of coolant accidents using support vector machines, IEEE Trans. Nucl. Sci. 55 (1) (2008) 628-636.
  12. S.H. Lee, Y.G. No, M.G. Na, K.I. Ahn, S.Y. Park, Diagnostics of loss of coolant accidents using SVC and GMDH models, IEEE Trans. Nucl. Sci. 58 (1) (2011) 267-276.
  13. K.H. Yoo, Y.D. Koo, M.G. Na, Identification of LOCA and estimation of its break size by multiconnected support vector machines, IEEE Trans. Nucl. Sci. 64 (10) (2017) 2610-2617.