Herein, the heat-transfer characteristics of the subcooled flow boiling of swirl tubes are evaluated experimentally under one-side high heat load conditions. As the twist ratio of the tape decreased and the flow rate increased, the forced convective heat transfer was enhanced; thus, the single-phase heat transfer was enhanced, and the onset of nucleate boiling (ONB) heat flux increased. As the degree of subcooling increased, the ONB heat flux increased because of the faster condensation rate. As the pressure increased, the ONB heat flux increased, and the critical heat flux increased, due to the suppression of boiling and the decrease in the bubble size. The authors analyzed the heat transfer characteristics of the one-side heated swirl tube in the single-phase regime, the partially developed nucleate boiling (PDB) regime, and the fully developed nucleate (FDB) boiling regime. And the prediction performance of the existing heat-transfer correlations developed for application to each regime was evaluated. The PDB correlations showed a mean absolute error rate (MAE) of approximately 30%. However, the existing FDB correlations a high error rate (MAE higher than 48%) at a high heat flux of several MW/M ² outside the developed experimental range. Therefore, we develop a new swirl tube FDB heat-transfer correlation that can also be used in the development of a swirl tube flow boiling map using the Python code grafted with the artificial intelligence (AI) regression method.

In this study, the heat transfer characteristics of one-side heated screw tubes were analyzed using subcooled flow boiling experiments. The experiments were performed until the critical heat flux was reached, while the applied heat flux was gradually increased. Subsequently, the boiling curves derived from the experimental results were divided into partial nucleate boiling (PDB) and fully developed nucleate boiling (FDB) regimes. As the flow rate increases, the forced convective heat transfer is enhanced, improving the single-phase heat transfer performance. However, as the FDB region is approached, most of the heat transfer proceeds via nucleate boiling heat transfer, so the heat transfer coefficient curve converges into the same straight line. The same trend was observed when analyzing the effect of the degree of subcooling. However, as the system pressure increased, and the liquid surface tension and latent heat decreased, a faster transition was observed from the single-phase regime to the PDB regime. From evaluating the prediction performance of the existing subcooled flow boiling heat transfer correlations, among the PDB regime; Liu and Winterton correlation, a kind of Chen-type correlation, showed the highest accuracy with mean absolute error (MAE) of 13.62% and root mean square error (RMSE) of 16.13%. However, most of the FDB correlations did not accurately predict the heat transfer performance of the one-side heated screw tube under a high heat flux load of several MW/m². Therefore, a new screw tube FDB correlation was developed using Python code combined with artificial intelligence technology. The authors believe that the proposed correlation will be useful for not only predicting the heat transfer performance in the FDB region, but also establishing an operating map of the screw tube in the future.

In this study, the two-phase pressure drop of a one-side heated swirl tube (twist ratio = 3, 5) was experimentally studied. If the a heat flux is gradually increased, the vapor blocks the flow path, thus leading to an increase in the differential pressure. As the inset sub-cooling and mass flow rate increase, the vapor is removed by condensation more rapidly, and thus, the rate of increase of differential pressure is reduced. From the evaluation of the prediction performance of the existing two-phase pressure drop correlations, it was found that the mean error rates of all correlations exceeded 50%. This is because most of the existing correlations were not developed based on the microchannel, but they were not focused on the swirl tube. Therefore, the authors developed a new two-phase pressure drop multiplier correlation for a one-side heated swirl tube that reflects the effect of the twist tape by modifying the Manglick single-phase swirl tube pressure drop correlation using the AI regression method.

As the hypervapotron (HV) heat sink is used to cool many areas inside the fusion tokamak, it is essential to understand its heat transfer performance to calculate the thermal efficiency of the power generation system. Therefore, in this study, the single-phase (SP) heat transfer performance of HV heat sink was evaluated through sub-cooled flow boiling experiments under one-side high-heat load conditions. When vapor is generated inside the heat sink, flow instability and a potential risk of reaching the critical heat flux are created. Therefore, in commercial power plants, cooling systems tend to operate in the SP regime. System parameters that can be adjusted in the power generation system include the system pressure, mass flow rate, and subcooling, and the effect of these three parameters on the heat transfer performance in the SP regime was analyzed. It was experimentally observed that the mass flow rate was the most influential variable. The prediction performance of the SP forced convection heat transfer correlations of the existing conventiaonl channel were evaluated. The results revealed that they tended to under-predict the heat transfer performance of the HV heat sink. In addition, the same trends were found when the forced convection heat transfer correlation of the curved channel was evaluated. The reasons for the former and the latter are that the heat transfer enhancement effect by the vortex flow occurring between the fins of the HV heat sink is not reflected in the correlations, and the vortex effect of the HV heat sink is not expressed as a variable. Therefore, a new vortex forced convection heat transfer correlation was developed through the newly defined Dean number of the HV heat sink. The developed correlation recorded an average error rate of 0.48%.

In this study, the two-phase pressure drop in a one-sided high-heat-loaded screw tube was analyzed under subcooled flow boiling conditions. When the loaded heat flux is gradually increased under one-sided heating conditions, although the pressure drop changes slightly in the single phase, it starts to increase when loading above the onset of significant void (OSV) heat flux. The effect of system parameters on the two-phase pressure drop in relation to the change of heat transfer mechanism or fluid properties was analyzed according to each variable change. In addition, this study evaluated the prediction performance of existing two-phase pressure drop multiplier correlations developed under subcooled flow boiling conditions. Because most existing correlations are based on a smooth channel with a small diameter, they tend to overpredict the experimental values. The correlation proposed by Ramesh showed the lowest average error rate (31.94%); however, the results were underpredicted for high-heat-flux conditions. To obtain more accurate results, we developed a Python code using an artificial intelligence regression method to proceed with the optimization process based on the experimental values. Using this code, we developed a novel two-phase pressure drop multiplier correlation for a one-side-heated screw tube that can be used in fusion divertors under high-heat-load conditions.

In this study, the onset of flow instability (OFI) of the hypervapotron channel was experimentally explored under one-side high heat load conditions. The OFI was detected based on the heat flux ( q ″ ) where the standard deviation of the inlet pressure rapidly increased. Moreover, the effect of the system parameters was analyzed, showing that the OFI q ″ increased as the inset subcooling and mass flow rate increased. The reason is the high subcooling and flow rate that rapidly condense the vapor, which is the leading cause of OFI. Unfortunately, because most of the OFI correlations in previous studies were developed based on micro-channels, these correlations could not predict the hypervapotron OFI q ″ properly under one-side high heat load conditions. Finally, the authors found that the Reynolds number and Jakob number were inversely proportional to the ratio of the OFI heat flux ( q O F I ″ ) and saturation heat flux ( q s a t ″ ). Furthermore, a new correlation was developed using a PYTHON code and an artificial intelligence technique (mean absolute error rate = 12.89%, root mean square error rate = 14.81%).