• Nuclear Engineering and Technology
• /
• 제55권9호
• /
• pp.3409-3416
• /
• 2023

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.

• Nuclear Engineering and Technology
• /
• 제55권8호
• /
• pp.3088-3101
• /
• 2023

A pressurized vessel-pipe-safety valve (PVPSV) combination is a commonly used configuration in nuclear power plants, and a good numerical model is essential for the system design, sizing and performance optimization. However, owing to the large-scale and cross-scale features, it is still a challenge to build a system level numerical model with both high accuracy and efficiency. To overcome this, a novel system level modeling method which can synthesize the advantages of various models is proposed in this paper. For system modeling, the analytical approach, the method of characteristics (MOC) and the surrogate model approach are respectively adopted to predict the dynamics of the pressure vessel, the connecting pipe and the safety valve, and different models are connected through data interfaces. With this system model, dynamic simulations were carried out and both the stable and the unstable system responses were obtained. For the model verification purpose, the simulation results were compared with those obtained from experiments and full CFD simulations. A good agreement and a better efficiency were obtained, verifying the ability of the model and the feasibility of the modeling method proposed in this paper.

• International Journal of Fluid Machinery and Systems
• /
• 제9권3호
• /
• pp.265-276
• /
• 2016

Energy systems working coherently in different conditions may not have a specific design which can provide optimal performance. A system working for a longer period at lower efficiency implies higher energy consumption. In this effort, a methodology demonstrated by a jet pump design and optimization via numerical modeling for fluid dynamics and implementation of an evolutionary algorithm for the optimization shows a reduction in computational costs. The jet pump inherently has a low efficiency because of improper mixing of primary and secondary fluids, and multiple momentum and energy transfer phenomena associated with it. The high fidelity solutions were obtained through a validated numerical model to construct an approximate function through surrogate analysis. Pareto-optimal solutions for two objective functions, i.e., secondary fluid pressure head and primary fluid pressure-drop, were generated through a multi-objective genetic algorithm. For the jet pump geometry, a design space of several design variables was discretized using the Latin hypercube sampling method for the optimization. The performance analysis of the surrogate models shows that the combined surrogates perform better than a single surrogate and the optimized jet pump shows a higher performance. The approach can be implemented in other energy systems to find a better design.

• Nuclear Engineering and Technology
• /
• 제52권2호
• /
• pp.287-295
• /
• 2020

Group method of data handling (GMDH) is considered one of the earliest deep learning methods. Deep learning gained additional interest in today's applications due to its capability to handle complex and high dimensional problems. In this study, multi-layer GMDH networks are used to perform uncertainty quantification (UQ) and sensitivity analysis (SA) of nuclear reactor simulations. GMDH is utilized as a surrogate/metamodel to replace high fidelity computer models with cheap-to-evaluate surrogate models, which facilitate UQ and SA tasks (e.g. variance decomposition, uncertainty propagation, etc.). GMDH performance is validated through two UQ applications in reactor simulations: (1) low dimensional input space (two-phase flow in a reactor channel), and (2) high dimensional space (8-group homogenized cross-sections). In both applications, GMDH networks show very good performance with small mean absolute and squared errors as well as high accuracy in capturing the target variance. GMDH is utilized afterward to perform UQ tasks such as variance decomposition through Sobol indices, and GMDH-based uncertainty propagation with large number of samples. GMDH performance is also compared to other surrogates including Gaussian processes and polynomial chaos expansions. The comparison shows that GMDH has competitive performance with the other methods for the low dimensional problem, and reliable performance for the high dimensional problem.

• 한국항공우주학회지
• /
• 제48권1호
• /
• pp.13-21
• /
• 2020

유도무기의 설계 및 제어에서 6자유도 공력계수의 신속한 추정을 위해 공력계수 데이터에 기반한 예측 모형이 주로 이용된다. 고정확도의 공력계수 예측 모형은 다수의 풍동시험 데이터로 생성할 수 있지만, 이는 많은 시간과 자원을 요구한다. 따라서 본 연구에서는 소수의 풍동시험 데이터를 다수의 전산유체역학 데이터와 혼합한 코크리깅 기법을 활용해 고정확도의 공력계수를 신속하고 효율적으로 예측하고자 한다. 풍동시험과 전산유체역학 데이터를 혼용한 예측 모형의 우수성을 보기 위해, 전산유체역학 데이터 보조의 유무에 따라 두 가지 공력계수 예측 모형을 생성한 후 수치적 검증과 예측 경향성 점검으로 두 모형의 예측 정확도를 비교하였다. 그 결과, 전산유체역학 데이터의 도움 덕분에 코크리깅 모형으로 크리깅 모형보다 더 정확한 공력계수 산출이 가능한 것을 확인하였다.