Optimisation studies on performance enhancement of spray cooling - Machine learning approach

Abstract

The performance optimisation of spray cooling heat transfer systems has been identified as a significant step in improving process efficiency, and the application of machine learning tools is a recent development that has enhanced this. This study aims to maximise the heat transfer coefficient for spray cooling at low heat flux levels. The effects of nozzle inclination angle, water pressure, and spray height on heat transfer coefficient were studied. Taguchi L₂₇ orthogonal array technique was used to perform the experiments. A maximum heat transfer coefficient of 181.4 kW/m²K was obtained at a nozzle inclination angle of 60°, spray height of 4 cm, and water pressure of 15 psi. Analysis of variance was performed to find the significance of each variable and its interactions. The results show that for the maximum heat transfer coefficient (181.4 kW/m²K), the optimum levels of the independent variables were A3H1P3, i.e., the highest level of nozzle inclination angle (60°), the lowest level of spray height (4 cm), and the highest level of water pressure (15 psi). The support vector machine outperformed the Random Forest algorithm and Multiple Regression analysis regarding prediction accuracy with a maximum error of 0.15 % and root mean squared error of 0.01.