Real-time estimation of break sizes during LOCA in nuclear power plants using NARX neural network

Abstract This paper deals with break size estimation of loss of coolant accidents (LOCA) using a nonlinear autoregressive with exogenous inputs (NARX) neural network. Previous studies used static approaches, requiring time-integrated parameters and independent firing algorithms. NARX neural network is able to directly deal with time-dependent signals for dynamic estimation of break sizes in real-time. The case studied is a LOCA in the primary system of Bushehr nuclear power plant (NPP). In this study, number of hidden layers, neurons, feedbacks, inputs, and training duration of transients are selected by performing parametric studies to determine the network architecture with minimum error. The developed NARX neural network is trained by error back propagation algorithm with different break sizes, covering 5%–100% of main coolant pipeline area. This database of LOCA scenarios is developed using RELAP5 thermal-hydraulic code. The results are satisfactory and indicate feasibility of implementing NARX neural network for break size estimation in NPPs. It is able to find a general solution for break size estimation problem in real-time, using a limited number of training data sets. This study has been performed in the framework of a research project, aiming to develop an appropriate accident management support tool for Bushehr NPP.

[1]  Enrico Zio,et al.  Identification of a line break by a neural network methodology , 1994 .

[2]  Mohammad B. Ghofrani,et al.  Accident management support tools in nuclear power plants: A post-Fukushima review , 2016 .

[3]  Man Gyun Na,et al.  Detection and Diagnostics of Loss of Coolant Accidents Using Support Vector Machines , 2008, IEEE Transactions on Nuclear Science.

[4]  Poong Hyun Seong,et al.  A dynamic neural network based accident diagnosis advisory system for nuclear power plants , 2005 .

[5]  Enrico Zio,et al.  Remaining useful life prediction of degrading systems subjected to imperfect maintenance: Application to draught fans , 2018 .

[6]  Man Gyun Na,et al.  Prediction of Leak Flow Rate Using Fuzzy Neural Networks in Severe Post-LOCA Circumstances , 2014, IEEE Transactions on Nuclear Science.

[7]  César Queral,et al.  Current Status and Applications of Integrated Safety Assessment and Simulation Code System for ISA , 2017 .

[8]  Man Gyun Na,et al.  ESTIMATION OF BREAK LOCATION AND SIZE FOR LOSS OF COOLANT ACCIDENTS USING NEURAL NETWORKS , 2004 .

[9]  M. Saghafi,et al.  Determination of PAR configuration for PWR containment design: A hydrogen mitigation strategy , 2017 .

[10]  Caro Lucas,et al.  Identification of a nuclear reactor core (VVER) using recurrent neural networks , 2002 .

[11]  T. Takeda.  ROSA/LSTF test and RELAP5 code analyses on PWR 1% vessel upper head small-break LOCA with accident management measure based on core exit temperature , 2018, Nuclear Engineering and Technology.

[12]  Francesco Saverio D'Auria,et al.  Application of FFTBM with signal mirroring to improve accuracy assessment of MELCOR code , 2016 .

[13]  Tsungnan Lin,et al.  What to remember: how memory order affects the performance of NARX neural networks , 1998, 1998 IEEE International Joint Conference on Neural Networks Proceedings. IEEE World Congress on Computational Intelligence (Cat. No.98CH36227).

[14]  A. Srivastava,et al.  Diagnostic system for identification of accident scenarios in nuclear power plants using artificial neural networks , 2009, Reliab. Eng. Syst. Saf..

[15]  Davide Roverso,et al.  Plant diagnostics by transient classification: The ALADDIN approach , 2002, Int. J. Intell. Syst..

[16]  Man Gyun Na,et al.  Diagnostics of Loss of Coolant Accidents Using SVC and GMDH Models , 2011, IEEE Transactions on Nuclear Science.

[17]  Takeshi Takeda,et al.  Uncertainty analysis of ROSA/LSTF test by RELAP5 code and PKL counterpart test concerning PWR hot leg break LOCAs , 2018, Nuclear Engineering and Technology.

[18]  Man Gyun Na,et al.  Smart Sensing of the RPV Water Level in NPP Severe Accidents Using a GMDH Algorithm , 2014, IEEE Transactions on Nuclear Science.

[19]  Sung Won Bae,et al.  ANALYSIS OF UNCERTAINTY QUANTIFICATION METHOD BY COMPARING MONTE-CARLO METHOD AND WILKS’ FORMULA , 2014 .

[20]  Poong Hyun Seong,et al.  A dynamic neural network aggregation model for transient diagnosis in nuclear power plants , 2007 .

[21]  Enrico Zio,et al.  Clustering for unsupervised fault diagnosis in nuclear turbine shut-down transients , 2015 .

[22]  Peter Tiño,et al.  Learning long-term dependencies in NARX recurrent neural networks , 1996, IEEE Trans. Neural Networks.

[23]  Alessandro Petruzzi,et al.  Thirty Years’ Experience in RELAP5 Applications at GRNSPG & NINE , 2016 .

[24]  Francesco Saverio D'Auria,et al.  Development and qualification of a thermal-hydraulic nodalization for modeling station blackout accident in PSB-VVER test facility , 2016 .