Conference Agenda

Due to its high latent heat, pool boiling exhibits excellent heat dissipation capability and is widely used in many industries such as electronic devices, steam generators, nuclear reactors, etc. This paper experimentally investigates the heat transfer enhancement effect of bi-conductive surfaces in pool boiling and reveals its underlying mechanism.Because of the enormous difference in conductivity, the heat transferred through the low-conductive epoxy can be ignored, and boiling only occurs on the high-conductive copper surface. This provides a method that can create specifically appointed spatial surface temperature variations and induce ordered liquid and vapor paths by changing the pattern of epoxy.This paper designs two types of epoxy patterns with a fixed copper area ratio of 60%, including reticular epoxy samples and more complex reticular epoxy with squares in the middle samples. At the same time, the number and width of stripes in the reticular epoxy are changed in each category. According to the result, almost all reticular epoxy samples enhance the CHF compared with bare copper. With the increase in the number of stripes in the reticular epoxy, the CHF displays the tendency to rise at first and decline in the end, reaching the peak of 77.2% CHF increase. Under the same number of stripes, reticular epoxy with squares in the middle samples have higher HTC and CHF compared with reticular epoxy samples because of superior wetting effect of epoxy. The results also suggest that overly narrow epoxy can lead to insufficient wetting effect and trigger CHF in advance.

During flow boiling, two different regimes are observed namely nucleate boiling and convective flow boiling. Nucleate boiling is dominant at high heat fluxes where bubbles produced at the wall are attributed the control of the heat transfer. Convective flow boiling is dominant at low heat fluxes and the heat transfer coefficient is observed to be directly dependent on the mass flux and the thermodynamic quality. The transition between to these two regimes has motivated vast research to determine if the transition is triggered sharply or there is a region where both mechanisms are presented. In this work, we study the nucleate flow boiling to convective flow boiling transition experimentally in a horizontal heated pipe.

Accurate prediction of heat flux is essential for various industrial and scientific applications, particularly in heat transfer and thermal management systems. Traditional methods for heat flux estimation often rely on complex physical modeling and intrusive sensor-based measurements, limiting their applicability in dynamic boiling conditions. Recent advancements in deep learning, particularly Convolutional Neural Networks (CNNs), have enabled non-intrusive heat flux prediction directly from boiling images. In this paper, we propose a multi-branch convolutional neural network architecture for heat flux prediction from boiling images. The key innovation lies in the introduction of multi-branch convolutional components (MB-Conv), which integrate multiple convolutional branches with varying kernel sizes to extract a comprehensive set of features.

The proposed model leverages a multi-branch architecture to enhance its representational power, allowing it to effectively learn from a diverse range of spatial features that are critical in predicting heat flux in boiling systems. Experimental results demonstrate that our model significantly improves prediction accuracy compared to the traditional single-path convolutional neural model. Moreover, the proposed multi-branch architecture outperforms several well-established CNN models, such as AlexNet, VGG, ResNet, DenseNet and EfficientNet in terms of predictive performance, highlighting the effectiveness of our approach in capturing the complex thermal phenomena present in boiling images. The ability to predict heat flux from boiling images opens up new possibilities for optimizing system performance and ensuring safety in thermal systems.

Enhancing the heat transfer is of interest for a wide spectrum of industries to achieve higher thermal efficiency. During subcooled flow boiling, liquid-to-vapour phase change causes high heat transfer rates, although the coolant bulk temperature is below its saturation temperature. When the required wall superheat is developed, vapour bubbles form over the heated surface marking the onset of nucleate boiling. When a bubble sufficiently grows in size, it departs from the heated surface, and its departure size and frequency dictate the enhancement of heat transfer rates. Formation of larger bubbles near the heated surface may result in their coalescence forming a local dry patch which may eventually lead to Critical Heat Flux (CHF). Although there are numerous correlations available in the literature, to estimate the bubble departure size, most of the correlations perform well at high pressure conditions. To this end, it is important to study the bubble departure size and its influence on the subcooled flow boiling characteristics at low pressure conditions. In the present study, Eulerian-Eulerian multiphase flow (EEMF) model is employed to simulate the subcooled flow boiling conditions. Instead of using an existing empirical correlation, a force balance model is developed to predict the departure size and validated against the experimental data at low pressure conditions. Based on this model, a correlation for departure diameter is developed which is coupled with the EEMF model framework. The developed correlation is found to be more accurately capturing the local vapour volume-fraction profiles compared to the existing correlations.

We employ three fast and synchronized optical techniques (white-light interferometry, infra-red thermography, shadowgraphy) to study the near-wall phenomena during the growth of a single bubble in saturated pool boiling of water at atmospheric pressure. Our focus is on the impact of applied heating on bubble growth dynamics, as well as the near-wall features: dry spot spreading, the liquid thin film (microlayer) that can form between the heater and the liquidvapor interface of the bubble and the interfacial thermal resistance. We found that varying the applied heating power does not significantly impact the bubble macroscopic and near-wall features. It is explained by large heat capacity of the heater. The only affected parameter is the waiting time, which decreases with the applied heating power. The interfacial thermal resistance shows no dependence with heat flux, and increases monotonously over time due to the progressive accumulation of impurities at the interface. We show that the triple contact line dynamics depends on the wall superheating at the contact line.