This paper introduces a novel hyperparameter optimization framework for regression tasks called the Combined-Sampling Algorithm to Search the Optimized Hyperparameters (CASOH). Our approach enables hyperparameter tuning for deep learning models with two hidden layers and multiple types of hyperparameters, enhancing the model’s capacity to work with complex optimization problems. The primary goal is to improve hyperparameter tuning performance in deep learning models compared to conventional methods such as Bayesian Optimization and Random Search. Furthermore, CASOH is evaluated alongside the state-of-the-art hyperparameter reinforcement learning (Hyp-RL) framework to ensure a comprehensive assessment. The CASOH framework integrates the Metropolis-Hastings algorithm with a uniform random sampling approach, increasing the likelihood of identifying promising hyperparameter configurations. Specifically, we developed a correlation between the objective function and samples, allowing subsequent samples to be strongly correlated with the current sample by applying an acceptance probability in our sampling algorithm. The effectiveness of our proposed method was examined using regression datasets such as Boston Housing, Critical heat flux (CHF), Concrete compressive strength, Combined Cycle Power Plant, Gas Turbine CO, and NOx Emission, as well as an ‘in-house’ dataset of lattice-physics parameters generated from a Monte Carlo code for nuclear fuel assembly simulation. One of the primary goals of this study is to construct an optimized deep-learning model capable of accurately predicting lattice-physics parameters for future applications of machine learning in nuclear reactor analysis. Our results indicate that this framework achieves competitive accuracy compared to conventional random search and Bayesian optimization methods. The most significant enhancement was observed in the lattice-physics dataset, achieving a 56.6% improvement in prediction accuracy, compared to improvements of 53.2% by Hyp-RL, 44.9% by Bayesian optimization, and 38.8% by random search relative to the nominal prediction. While the results are promising, further empirical validation across a broader range of datasets would be helpful to better assess the framework’s suitability for optimizing hyperparameters in complex problems involving high-dimensional parameters, highly non-linear systems, and multi-objective optimization tasks.
The reflooding phase, a crucial recovery process after a loss of coolant accident (LOCA) in reactors, involves cooling overheated fuel rods with subcooled water. Its complex nature, notably in its flow regime and heat transfer, makes prediction challenging, resulting in high uncertainty and computation cost. In this study, we utilized the data assimilation (DA) technique to enhance the prediction of reflooding phenomena and subsequently deployed machine learning models to predict the accuracy of the safety and performance analysis code (SPACE) simulation. To generate the dataset for the machine learning model, we employed the sampling method for highly nonlinear system uncertainty analysis (STARU), providing a high-quality dataset for a complex problem such as a reflooding simulation. In this dataset, the physical models were assimilated under their selected uncertainty bands and utilized the effective sampling approach of STARU, generating the high-quality output and efficient enhancement of SPACE predictions. Consequently, the implemented machine learning model can be used to enhance model development and uncertainty quantification (UQ) analysis using the system code.
Symmetry-based domain reduction is widely applied in Monte Carlo reactor physics to reduce computational cost using reflective, white, and periodic boundary conditions. In lattice physics analyses, simulations based on minimum symmetric domains are often assumed to provide the best balance between accuracy and computational efficiency. However, this assumption depends on the validity of the imposed symmetry boundaries, which represent mathematical approximations rather than physical interfaces. This study investigates the effects of different boundary conditions on fuel assembly neutronics calculations using two representative geometries: a rectangular PWR fuel assembly and a hexagonal VVER fuel assembly. Infinity multiplication factors (k-inf) and pin-wise power peaking factors (PPPFs) are compared between full-assembly models and reduced-symmetry configurations employing reflective, white, and periodic boundary conditions. Results show that reflective and white boundary conditions provide strong agreement between full and reduced models, with R2 values up to 0.995 and PPPF differences below 1.8% for the PWR assembly, and R2 values up to 0.930 with maximum PPPF differences of 5.0% for the VVER assembly. In contrast, the symmetric 1/8 PWR domain with periodic boundary conditions exhibits lower agreement (R2 of 0.543 and PPPF differences of 12.5%), reflecting the stringent surface-matching requirements imposed by periodic boundaries in highly reduced computational domains. These results demonstrate that the predicted lattice physics parameters are sensitive to the choice of boundary condition and the adopted symmetry reduction level, even in highly regular lattices commonly idealized as infinitely repeated systems.