Prediction of LOCA break sizes using LSTM architecture for pressurized water reactors
-
Youssef Badr
,
Ibrahim Mohsen
Abstract
Neural networks (NNs) and deep learning have revolutionized several fields, and nuclear safety analysis is no exception. The proper operation of nuclear reactor safety systems is crucial and is designed to meet strict safety requirements. Such systems are expected to withstand certain postulated accidents known as design basis accidents (DBAs), which include the loss of coolant accident (LOCA). As the LOCA involves complex fluid mechanics and heat transfer, the pattern recognition abilities of NNs allow for excellent prediction capabilities and the bypassing of otherwise tedious conventional analysis methods. This work investigates the deep learning techniques through long short-term memory (LSTM) architecture, for its ability to deal with time-series problems which include LOCAs. The utilized model is taught to estimate the size of pipe breaks within the cooling system based off of the corresponding pressure drops. A range of 0.5 %–100 % break-size-time-variant parameters were collected using the WSC Inc. 1,400 MWe generic pressurized water reactor (GPWR) simulator, using two circulation loops. Neural networks were trained on parameters such as loop temperature, pressure, containment pressure and Boron concentration. The performance of the LSTM model showed a mean absolute error (MAE) of 5.185, mean squared error (MSE) of 76.50, root mean squared error (RMSE) of 7.953, R2 of 0.888 and Accuracy of 80.684 % within a tolerance of 15 % across 100 runs.
-
Research ethics: Not applicable.
-
Informed consent: Not applicable.
-
Author contributions: All authors have accepted responsibility for the entire content of this manuscript and approved its submission.
-
Use of Large Language Models, AI and Machine Learning Tools: None declared.
-
Conflict of interest: The authors state no conflict of interest.
-
Research funding: None declared.
-
Data availability: Not applicable.
References
Bays, S., Abou Jaoude, A., and Borlodan, G. (2019). Reactor fundamentals handbook. Technical Report INL/EXT-19-53301-Rev000, https://doi.org/10.2172/1615634.Search in Google Scholar
Choi, G.P., Yoo, K.H., Back, J.H., and Na, M.G. (2017). Estimation of LOCA break size using cascaded fuzzy neural networks. Nucl. Eng. Technol. 49: 495–503, https://doi.org/10.1016/j.net.2016.11.001.Search in Google Scholar
Desterro, F.S.M., Pinheiro, V.H.C., Pereira, C.M.N.A., and Schirru, R. (2023). Prediction of LOCA’s break size and location based on random forest and multi tasking deep neural network. Nucl. Eng. Des. 415: 112711, https://doi.org/10.1016/j.nucengdes.2023.112711.Search in Google Scholar
Elbordany, A.A., Kandil, M.M., Youness, H.A., and Abdelaal, H.M. (2025). An efficient AI algorithm for fault diagnosis in nuclear power plants based on machine deep learning techniques. Prog. Nucl. Energy 180: 105580, https://doi.org/10.1016/j.pnucene.2024.105580.Search in Google Scholar
Frepoli, C. (2008). An overview of Westinghouse realistic large break LOCA evaluation model. Sci. Technol. Nucl. Install. 2008: 498737, https://doi.org/10.1155/2008/498737.Search in Google Scholar
Fyza, N., Hossain, A., and Sarkar, R. (2019). Analysis of the thermal-hydraulic parameters of VVER-1200 due to loss of coolant accident concurrent with loss of offsite power. Energy Procedia 160: 155–161, https://doi.org/10.1016/j.egypro.2019.02.131.Search in Google Scholar
Glasstone, S. and Sesonske, A. (1994). Nuclear reactor engineering: reactor systems engineering, 4th ed. Springer Nature, Available: https://doi.org/10.1007/978-1-4615-2083-2 (Accessed 16 Sept. 2025).Search in Google Scholar
Gupta, D. (2025). Activation functions | fundamentals of deep learning. Analytics Vidhya, Available: https://www.analyticsvidhya.com/blog/2020/01/fundamentals-deep-learning-activation-functions-when-to-use-them/ (Accessed 16 Sept. 2025).Search in Google Scholar
Jo, H.S., Lee, S.H., and Na, M.G. (2024). Prediction of small-scale leak flow rate in LOCA situations using bidirectional GRU. Nucl. Eng. Technol. 56: 3594–3601, https://doi.org/10.1016/j.net.2024.04.009.Search in Google Scholar
Li, J., Huang, Y., Qiu, Y., Wang, S., Yang, Q., Wang, K., and Zhu, Y. (2025). Prediction of critical heat flux using different methods: a review from empirical correlations to the cutting-edge machine learning. Int. Commun. Heat Mass Transf. 160: 108362, https://doi.org/10.1016/j.icheatmasstransfer.2024.108362.Search in Google Scholar
Muhammad, W., Ullah, I. and Ashfaq, M., 2020. An introduction to deep convolutional neural networks with Keras. In Machine learning and deep learning in real-time applications (pp. 231–272). IGI Global.10.4018/978-1-7998-3095-5.ch011Search in Google Scholar
OECD/NEA (2009). Nuclear fuel behaviour in loss-of-coolant accident (LOCA) conditions. Nuclear Energy Agency (NEA), Available: https://www.oecd-nea.org/jcms/pl_14524/nuclear-fuel-behaviour-in-loss-of-coolant-accident-loca-conditions?details=true (Accessed 16 Sept. 2025).Search in Google Scholar
Omar, S. and Hasan, M.N. (2022). A PCTRAN based analysis on the effect of breaksize and comparative study between hot and cold leg loss of coolant accidents in VVER 1200 power reactor. Acta Mech. Malays. 5: 31–34, https://doi.org/10.26480/amm.02.2022.31.34.Search in Google Scholar
Racheal, S., Liu, Y., and Ayodeji, A. (2022). Evaluation of optimized machine learning models for nuclear reactor accident prediction. Prog. Nucl. Energy 149: 104263, https://doi.org/10.1016/j.pnucene.2022.104263.Search in Google Scholar
Reddi, S.J., Kale, S., and Kumar, S. (2019). On the convergence of Adam and Beyond, https://doi.org/10.48550/arXiv.1904.09237.Search in Google Scholar
Saghafi, M. and Ghofrani, M.B. (2019). Real-time estimation of break sizes during LOCA in nuclear power plants using NARX neural network. Nucl. Eng. Technol. 51: 702–708, https://doi.org/10.1016/j.net.2018.11.017.Search in Google Scholar
Shimeck, D.J. and Hartz, J.J. (2000). Description of the Westinghouse LOCTAJR 1-D heat conduction code for LOCA analysis of fuel rods. Technical Report WCAP-15578, Westinghouse Electric Company LLC, Nuclear Services Business Unit, Pittsburgh, PA, Available: https://www.nrc.gov/docs/ML0037/ML003759024.pdf.Search in Google Scholar
Xiao, X., Qi, B., Liang, J., Tong, J., Deng, Q., and Chen, P. (2024). Enhancing LOCA breach size diagnosis with fundamental deep learning models and optimized dataset construction. Energies 17: 159, https://doi.org/10.3390/en17010159.Search in Google Scholar
Zhang, A., Lipton, Z.C., Li, M., and Smola, A.J. (2023). Dive into deep learning, www.d2I.ai, https://doi.org/10.48550/arXiv.2106.11342.Search in Google Scholar
Zhou, G., Peng, M., and Wang, H. (2024). Enhancing prediction accuracy for LOCA break sizes in nuclear power plants: a hybrid deep learning method with data augmentation and hyperparameter optimization. Ann. Nucl. Energy 196: 110208, https://doi.org/10.1016/j.anucene.2023.110208.Search in Google Scholar
Zubair, M. and Akram, Y. (2024). Enhancement in the safety and reliability of pressurized water reactors using machine learning approach. Ann. Nucl. Energy 201: 110448, https://doi.org/10.1016/j.anucene.2024.110448.Search in Google Scholar
© 2025 Walter de Gruyter GmbH, Berlin/Boston
Articles in the same Issue
- Frontmatter
- Serpent 2 simulations of the historic Haigerloch B8 nuclear reactor of 1945
- Safety behaviour of a material testing reactor using U3Si2 fuel during reactivity insertion accident
- Augmentation of the neutronic safety aspect of high-density fuel research reactor using new control element design
- Analysis of neutronic performance of VVER 1200 reactor using accident tolerant fuel
- Reliability assessment of equal height difference passive containment cooling system based on the adaptive metamodel-based subset importance sampling method in inverse uncertainty quantification framework
- Numerical investigation of three-tube intertwined HX design
- Prediction of LOCA break sizes using LSTM architecture for pressurized water reactors
- Decoupling and controller design of multivariable systems for small modular reactors
- Adsorption behavior of Se(Ⅳ) and Se(Ⅵ) on sodium bentonite
- Calendar of events
Articles in the same Issue
- Frontmatter
- Serpent 2 simulations of the historic Haigerloch B8 nuclear reactor of 1945
- Safety behaviour of a material testing reactor using U3Si2 fuel during reactivity insertion accident
- Augmentation of the neutronic safety aspect of high-density fuel research reactor using new control element design
- Analysis of neutronic performance of VVER 1200 reactor using accident tolerant fuel
- Reliability assessment of equal height difference passive containment cooling system based on the adaptive metamodel-based subset importance sampling method in inverse uncertainty quantification framework
- Numerical investigation of three-tube intertwined HX design
- Prediction of LOCA break sizes using LSTM architecture for pressurized water reactors
- Decoupling and controller design of multivariable systems for small modular reactors
- Adsorption behavior of Se(Ⅳ) and Se(Ⅵ) on sodium bentonite
- Calendar of events
