This study employs machine learning techniques â€“ random forest and extra trees – to predict the frictional pressure gradient during convective condensation in an inclined in-tube heat exchanger. The experimental data matrix (663) includes conditions for saturation temperatures of 30, 40, and 50oC, mass velocity 100-400 kgm-2s, quality 10-90%, and thirteen inclination angles between -90o and +90o for a smooth tube of an internal diameter of 8.38 mm. Based on statistical analysis, the extra trees outperforms the random forest. The average deviation (AD) and mean average deviation (MAD) are 2.88% and 6.72%, respectively, for random forest (RF) and 0.25% and 2.97%, respectively, for extra trees (ET).
Keywords rictional pressure gradient, inclination angle, inclined smooth tube, condensation, machine learning