Investigation of Nucleate Pool Boiling Heat Transfer by Computational Methods in Different Cylindrical Surfaces and Inclination Angles

Abstract

Boiling heat transfer, which allows a high amount of energy conversion with small temperature differences, is being investigated in many areas such as vapor boilers, heat exchangers, energy systems, and power plant. The present study investigated the boiling heat transfer occurring in the isolated bubble regime region on different cylindrical metal surfaces and at different inclination angles. Steel, aluminum, and copper were chosen as the surface material. In order to compare the experimental results carried out a numerical analysis of the boiling heat transfer in the material with smooth surfaces. The heat flux values resulting from the pool boiling heat transfer calculated with experimental data were modeled with pace regression (PR), multilayer perceptron (MLP), and decision tree (M5P) algorithms, which are machine learning algorithms. Among these algorithms, PR was used for the first time in surface pool boiling heat transfer modeling. The model results obtained were compared with the experimental results. 2.94 (MAE) error rate for steel material, 2.93 (MAE) error rate for aluminum material, and 2.66 (MAE) error rate for copper material and the heat flux values were modeled by the PR algorithm OLS method. With the help of the PR algorithm's boiling heat transfer equations were successfully modeled different materials' surface boiling heat flux values. Using these mathematical equations showed that boiling heat flux values could be determined for different surfaces under different experimental conditions.