Nuclear Engineering and Technology

Korean Nuclear Society (한국원자력학회)
권호 표지
DOI QR코드

DOI QR Code

Physics informed neural networks for surrogate modeling of accidental scenarios in nuclear power plants

  • Federico Antonello (Department of Nuclear Science and Engineering, Massachusetts Institute of Technology (MIT)) ;
  • Jacopo Buongiorno (Department of Nuclear Science and Engineering, Massachusetts Institute of Technology (MIT)) ;
  • Enrico Zio (Mines Paris-PSL University, Centre de Recherche sur les Risques et les Crises)
  • Received : 2023.01.10
  • Accepted : 2023.06.16
  • Published : 2023.09.25
Download PDF
⟨ Previous Next ⟩

Abstract

Licensing the next-generation of nuclear reactor designs requires extensive use of Modeling and Simulation (M&S) to investigate system response to many operational conditions, identify possible accidental scenarios and predict their evolution to undesirable consequences that are to be prevented or mitigated via the deployment of adequate safety barriers. Deep Learning (DL) and Artificial Intelligence (AI) can support M&S computationally by providing surrogates of the complex multi-physics high-fidelity models used for design. However, DL and AI are, generally, low-fidelity 'black-box' models that do not assure any structure based on physical laws and constraints, and may, thus, lack interpretability and accuracy of the results. This poses limitations on their credibility and doubts about their adoption for the safety assessment and licensing of novel reactor designs. In this regard, Physics Informed Neural Networks (PINNs) are receiving growing attention for their ability to integrate fundamental physics laws and domain knowledge in the neural networks, thus assuring credible generalization capabilities and credible predictions. This paper presents the use of PINNs as surrogate models for accidental scenarios simulation in Nuclear Power Plants (NPPs). A case study of a Loss of Heat Sink (LOHS) accidental scenario in a Nuclear Battery (NB), a unique class of transportable, plug-and-play microreactors, is considered. A PINN is developed and compared with a Deep Neural Network (DNN). The results show the advantages of PINNs in providing accurate solutions, avoiding overfitting, underfitting and intrinsically ensuring physics-consistent results.

Keywords

Acknowledgement

The research has been funded by the "Progetto Rocca PostDoctoral Fellowship".

References

  1. M.I. Radaideh, T. Kozlowski, Surrogate modeling of advanced computer simulations using deep Gaussian processes, Reliab. Eng. Syst. Saf. 195 (October 2019) (2020), 106731, https://doi.org/10.1016/j.ress.2019.106731.
  2. F. D'Auria, C. Camargo, O. Mazzantini, The Best Estimate Plus Uncertainty (BEPU) approach in licensing of current nuclear reactors, Nucl. Eng. Des. 248 (2012) 317-328, https://doi.org/10.1016/j.nucengdes.2012.04.002.
  3. X. Wu, T. Kozlowski, Inverse uncertainty quantification of reactor simulations under the Bayesian framework using surrogate models constructed by polynomial chaos expansion, Nucl. Eng. Des. 313 (2017) 29-52, https://doi.org/10.1016/j.nucengdes.2016.11.032.
  4. R.K. Tripathy, I. Bilionis, Deep UQ: learning deep neural network surrogate models for high dimensional uncertainty quantification, J. Comput. Phys. 375 (2018) 565-588, https://doi.org/10.1016/j.jcp.2018.08.036.
  5. M.I. Radaideh, T. Kozlowski, Analyzing nuclear reactor simulation data and uncertainty with the group method of data handling, Nucl. Eng. Technol. 52 (2) (2020) 287-295, https://doi.org/10.1016/j.net.2019.07.023.
  6. P. Benner, S. Gugercin, K. Willcox, A survey of projection-based model reduction methods for parametric dynamical systems, SIAM Rev. 57 (4) (2015) 483-531, https://doi.org/10.1137/130932715.
  7. S. Chaturantabut, D.C. Sorensen, Nonlinear model reduction via discrete empirical interpolation, SIAM J. Sci. Comput. 32 (5) (2010) 2737-2764. https://doi.org/10.1137/090766498
  8. C. Worrell, L. Luangkesorn, J. Haight, T. Congedo, Machine learning of fire hazard model simulations for use in probabilistic safety assessments at nuclear power plants, Reliab. Eng. Syst. Saf. 183 (February 2018) (2019) 128-142, https://doi.org/10.1016/j.ress.2018.11.014.
  9. J.E. Hurtado, Filtered importance sampling with support vector margin: a powerful method for structural reliability analysis, Struct. Saf. 29 (1) (2007) 2-15, https://doi.org/10.1016/j.strusafe.2005.12.002.
  10. G. Roma, F. Antonello, F. Di Maio, N. Pedroni, E. Zio, A. Bersano, C. Bertani, F. Mascari, Passive safety systems analysis: a novel approach for inverse uncertainty quantification based on Stacked Sparse Autoencoders and Kriging metamodeling, Prog. Nucl. Energy 148 (September 2021) (2022), 104209, https://doi.org/10.1016/j.pnucene.2022.104209.
  11. S. Yoon, M.J. Kim, S. Park, G.Y. Kim, Thermal conductivity prediction model for compacted bentonites considering temperature variations, Nucl. Eng. Technol. 53 (10) (2021) 3359-3366, https://doi.org/10.1016/j.net.2021.05.001.
  12. J.P. Yurko, J. Buongiorno, R. Youngblood, Demonstration of emulator-based bayesian calibration of safety analysis codes: theory and formulation, Sci. Technol. Nucl. Instal. 2015 (2015), https://doi.org/10.1155/2015/839249. Mcmc).
  13. B. Ebiwonjumi, C. Kong, P. Zhang, A. Cherezov, D. Lee, Uncertainty quantification of PWR spent fuel due to nuclear data and modeling parameters, Nucl. Eng. Technol. 53 (3) (2021) 715-731, https://doi.org/10.1016/j.net.2020.07.012.
  14. B. Sudret, C.V. Mai, Computing derivative-based global sensitivity measures using polynomial chaos expansions, Reliab. Eng. Syst. Saf. 134 (2015) 241-250, https://doi.org/10.1016/j.ress.2014.07.009.
  15. L. Puppo, N. Pedroni, F. Di Maio, A. Bersano, C. Bertani, E. Zio, A framework based on finite mixture models and adaptive kriging for characterizing non-smooth and multimodal failure regions in a nuclear passive safety system, Reliab. Eng. Syst. Saf. 216 (August) (2021), 107963, https://doi.org/10.1016/j.ress.2021.107963.
  16. N. Pedroni, E. Zio, An Adaptive Metamodel-Based Subset Importance Sampling approach for the assessment of the functional failure probability of a thermal-hydraulic passive system, Appl. Math. Model. 48 (2017) 269-288, https://doi.org/10.1016/j.apm.2017.04.003.
  17. X. Song, L. Lv, W. Sun, J. Zhang, A radial basis function-based multi-fidelity surrogate model: exploring correlation between high-fidelity and low-fidelity models, Struct. Multidiscip. Optim. 60 (3) (2019) 965-981, https://doi.org/10.1007/s00158-019-02248-0.
  18. L. Conner, C.L. Worrell, J.P. Spring, J. Liao, Machine learned metamodeling of a computationally intensive accident simulation code, Int. Conf. Nucl. Eng. Proc., ICONE 1 (2021) 1-6, https://doi.org/10.1115/ICONE28-66619.
  19. R. Abu Saleem, M.I. Radaideh, T. Kozlowski, Application of deep neural networks for high-dimensional large BWR core neutronics, Nucl. Eng. Technol. 52 (12) (2020) 2709-2716, https://doi.org/10.1016/j.net.2020.05.010.
  20. L. Sun, H. Gao, S. Pan, J.X. Wang, Surrogate modeling for fluid flows based on physics-constrained deep learning without simulation data, Comput. Methods Appl. Mech. Eng. 361 (2020), 112732, https://doi.org/10.1016/j.cma.2019.112732.
  21. A. Ayodeji, M.A. Amidu, S.A. Olatubosun, Y. Addad, H. Ahmed, Deep learning for safety assessment of nuclear power reactors: reliability, explainability, and research opportunities, Prog. Nucl. Energy 151 (August) (2022), 104339, https://doi.org/10.1016/j.pnucene.2022.104339.
  22. V. Ciriello, V. Di Federico, M. Riva, F. Cadini, J. De Sanctis, E. Zio, A. Guadagnini, Polynomial chaos expansion for global sensitivity analysis applied to a model of radionuclide migration in a randomly heterogeneous aquifer, Stoch. Environ. Res. Risk Assess. 27 (4) (2013) 945-954, https://doi.org/10.1007/s00477-012-0616-7.
  23. F. Shriver, C. Gentry, J. Watson, Prediction of neutronics parameters within a two-dimensional reflective PWR assembly using deep learning, Nucl. Sci. Eng. 195 (6) (2021) 626-647, https://doi.org/10.1080/00295639.2020.1852021.
  24. M.I. Radaideh, D. Price, T. Kozlowski, Modeling nuclear data uncertainties using deep neural networks, in: International Conference on Physics of Reactors: Transition to a Scalable Nuclear Future, PHYSOR 2020, 2020-March, 2020, pp. 2583-2590, https://doi.org/10.1051/epjconf/202124715016.
  25. F. Cadini, A. Gioletta, E. Zio, Improved metamodel-based importance sampling for the performance assessment of radioactive waste repositories, Reliab. Eng. Syst. Saf. 134 (2015) 188e197, https://doi.org/10.1016/j.ress.2014.10.018.
  26. D. Price, M.I. Radaideh, B. Kochunas, Multiobjective optimization of nuclear microreactor reactivity control system operation with swarm and evolutionary algorithms, Nucl. Eng. Des. 393 (January) (2022), 111776, https://doi.org/10.1016/j.nucengdes.2022.111776.
  27. H. Kim, J. Cho, J. Park, Application of a deep learning technique to the development of a fast accident scenario identifier, IEEE Access 8 (2020) 177363-177373, https://doi.org/10.1109/ACCESS.2020.3026104.
  28. S. Ryu, H. Kim, S.G. Kim, K. Jin, J. Cho, J. Park, Probabilistic deep learning model as a tool for supporting the fast simulation of a thermal-hydraulic code, Expert Syst. Appl. 200 (March) (2022), 116966, https://doi.org/10.1016/j.eswa.2022.116966.
  29. M. Ebad Sichani, J.E. Padgett, V. Bisadi, Probabilistic seismic analysis of concrete dry cask structures, Struct. Saf. 73 (June 2017) (2018) 87-98, https://doi.org/10.1016/j.strusafe.2018.03.001.
  30. R. Shrestha, T. Kozlowski, Inverse uncertainty quantification of input model parameters for thermal-hydraulics simulations using expectation-maximization under Bayesian framework, J. Appl. Stat. 43 (6) (2016) 1011-1026, https://doi.org/10.1080/02664763.2015.1089220.
  31. A. Barredo Arrieta, N. Diaz-Rodriguez, J. Del Ser, A. Bennetot, S. Tabik, A. Barbado, S. Garcia, S. Gil-Lopez, D. Molina, R. Benjamins, R. Chatila, F. Herrera, Explainable Artificial Intelligence (XAI): concepts, taxonomies, opportunities and challenges toward responsible AI, Inf. Fusion 58 (October 2019) (2020) 82-115, https://doi.org/10.1016/j.inffus.2019.12.012.
  32. G.E. Karniadakis, I.G. Kevrekidis, L. Lu, P. Perdikaris, S. Wang, L. Yang, Physics-informed machine learning, Nat. Rev. Phys. 3 (6) (2021) 422-440, https://doi.org/10.1038/s42254-021-00314-5.
  33. M. Raissi, P. Perdikaris, G.E. Karniadakis, Physics-informed neural networks: a deep learning framework for solving forward and inverse problems involving nonlinear partial differential equations, J. Comput. Phys. 378 (2019) 686-707, https://doi.org/10.1016/j.jcp.2018.10.045.
  34. S. Cai, Z. Wang, S. Wang, P. Perdikaris, G.E. Karniadakis, Physics-informed neural networks for heat transfer problems, J. Heat Tran. 143 (6) (2021), https://doi.org/10.1115/1.4050542.
  35. E. Schiassi, M. De Florio, B.D. Ganapol, P. Picca, R. Furfaro, Physics-informed neural networks for the point kinetics equations for nuclear reactor dynamics, Ann. Nucl. Energy 167 (2022), 108833, https://doi.org/10.1016/j.anucene.2021.108833.
  36. E. Haghighat, M. Raissi, A. Moure, H. Gomez, R. Juanes, A physics-informed deep learning framework for inversion and surrogate modeling in solid mechanics, Comput. Methods Appl. Mech. Eng. 379 (2021), 113741, https://doi.org/10.1016/j.cma.2021.113741.
  37. Y. Lu, B. Wang, Y. Zhao, X. Yang, L. Li, M. Dong, Q. Lv, F. Zhou, N. Gu, L. Shang, Physics-informed surrogate modeling for hydro-fracture geometry prediction based on deep learning, Energy 253 (2022), 124139, https://doi.org/10.1016/j.energy.2022.124139.
  38. J. Buongiorno, B. Carmichael, B. Dunkin, J. Parsons, D. Smit, Can nuclear batteries Be economically competitive in large markets? Energies 14 (14) (2021) 4385, https://doi.org/10.3390/en14144385.
  39. F. Antonello, J. Buongiorno, E. Zio, A methodology to perform dynamic risk assessment using system theory and modeling and simulation: application to nuclear batteries, Reliab. Eng. Syst. Saf. 228 (August) (2022), 108769, https://doi.org/10.1016/j.ress.2022.108769.
  40. F. Antonello, J. Buongiorno, E. Zio, Insights in the safety analysis of an early microreactor design, Nucl. Eng. Des. 404 (September 2022) (2023), 112203, https://doi.org/10.1016/j.nucengdes.2023.112203.
  41. E. Zio, Prognostics and Health Management (PHM): where are we and where do we (need to) go in theory and practice, Reliab. Eng. Syst. Saf. 218 (October 2021) (2022), https://doi.org/10.1016/j.ress.2021.108119.
  42. A.F. Henry, The application of reactor kinetics to the analysis of experiments, Nucl. Sci. Eng. 3 (1) (1958) 52-70, https://doi.org/10.13182/nse58-1.
  43. D.P. Kingma, J.L. Ba, Adam: a method for stochastic optimization, 3rd Int. Conf. Learn. Represent. ICLR 2015 - Conf.Track Proc. (2015) 1-15.
  44. D. Pedamonti, Comparison of Non-linear Activation Functions for Deep Neural Networks on MNIST Classification Task, vol. 3, 2018. http://arxiv.org/abs/1804.02763.
  45. C. Cortes, G. Research, N. York, L 2 regularization for learning kernels, in: Twenty-Fifth Conference on Uncertainty in Artificial Intelligence, 2004, pp. 109-116.