Critical heat flux (CHF) marks the onset of boiling crisis in light-water reactors, defining safe thermal-hydraulic operating limits. To support Phase II of the OECD/NEA AI/ML CHF benchmark, which introduces spatially varying power profiles, this work compiles and digitizes a broad CHF dataset covering both uniform and non-uniform axial heating conditions. Heating profiles were extracted from technical reports, interpolated onto a consistent axial mesh, validated via energy-balance checks, and encoded in machine-readable formats for benchmark compatibility. Classical CHF correlations exhibit substantial errors under uniform heating and degrade markedly when applied to non-uniform profiles, while modern tabular methods offer improved but still imperfect predictions. A neural network trained solely on uniform data performs well in that regime but fails to generalize to spatially varying scenarios, underscoring the need for models that explicitly incorporate axial power distributions. By providing these curated datasets and baseline modeling results, this study lays the groundwork for advanced transfer-learning strategies, rigorous uncertainty quantification, and design-optimization efforts in the next phase of the CHF benchmark.

The use of ML in engineering has grown steadily to support a wide array of applications. Among these methods, deep neural networks have been widely adopted due to their performance and accessibility, but they require large, high-quality datasets. Experimental data are often sparse, noisy, or insufficient to build resilient data-driven models. Transfer learning, which leverages relevant data-abundant source domains to assist learning in data-scarce target domains, has shown efficacy. Parameter transfer, where pretrained weights are reused, is common but degrades under large domain shifts. Domain-adversarial neural networks (DANNs) help address this issue by learning domain-invariant representations, thereby improving transfer under greater domain shifts in a semi-supervised setting. However, DANNs can be unstable during training and lack a native means for uncertainty quantification. This study introduces a fully-supervised three-stage framework, the staged Bayesian domain-adversarial neural network (staged B-DANN), that combines parameter transfer and shared latent space adaptation. In Stage 1, a deterministic feature extractor is trained on the source domain. This feature extractor is then adversarially refined using a DANN in Stage 2. In Stage 3, a Bayesian neural network is built on the adapted feature extractor for fine-tuning on the target domain to handle conditional shifts and yield calibrated uncertainty estimates. This staged B-DANN approach was first validated on a synthetic benchmark, where it was shown to significantly outperform standard transfer techniques. It was then applied to the task of predicting critical heat flux in rectangular channels, leveraging data from tube experiments as the source domain. The results of this study show that the staged B-DANN method can improve predictive accuracy and generalization, potentially assisting other domains in nuclear engineering.

This study explores the development of a hybrid deep convolutional neural network (DCNN) model enhanced by autoencoders and data augmentation techniques to predict critical heat flux (CHF) with high accuracy. By augmenting the original input features using three different autoencoder configurations, the model's predictive capabilities were significantly improved. The hybrid models were trained and tested on a dataset of 7225 samples, with performance metrics including the coefficient of determination (R2), Nash-Sutcliffe efficiency (NSE), mean absolute error (MAE), and normalized root-mean-squared error (NRMSE) used for evaluation. Among the tested models, the DCNN_3F-A2 configuration demonstrated the highest accuracy, achieving an R2 of 0.9908 during training and 0.9826 during testing, outperforming the base model and other augmented versions. These results suggest that the proposed hybrid approach, combining deep learning with feature augmentation, offers a robust solution for CHF prediction, with the potential to generalize across a wider range of conditions.

The critical heat flux (CHF) represents the maximum heat flux in the nucleate boiling process, marking an abrupt increase in surface temperature. As a crucial factor in high heat-flux systems to ensure safe operation and prevent system damage, CHF diagnosis has been extensively researched, leading to the development of various mechanistic models explaining the triggering mechanisms of CHF [1][2][3][4]. Among these models -- such as the hydrodynamic instability model, macrolayer dryout model, and interfacial lift-off model -- the hot/dry spot model suggests that irreversible dry patch formation leads to increasing temperature, resulting in the postulation that the development of the irreversible dry spot's temperature hinders the wetting of the heated surface by the supplied liquid [5]. The dry patch is first generated at high heat flux, then coalesces and expands again under the remnant bubble to trigger CHF [6]. To validate and improve such models, visual observation methods have been developed [7][8]. Total reflection visualization and (TR) infrared thermometry (IR) are arguably the most important techniques for visualizing the formation of dry patches while measuring the coincidental temperature evolution of the liquid-vapor system [9][10][11]. Through the methods, the behavior of the bubble structure and dry patch under flow boiling has been observed, and the hydrodynamic mechanism of the irreversible dry patch have been analyzed. Also, there have been attempts to determine CHF based on the temperature of the dry patch periphery [6][12]. Besides, following recent advancements in Convolutional Neural Networks (CNNs), which excel in capturing visual information characteristics, neural networks are expected to have the potential to simplify infrared thermal imaging, as the process typically involves tedious experimental setups and extensive data reduction [13].