Intelligent Control Based Estimation of Heat Transfer Coefficient from Four Flat Tubes with Different Attack Air Angles

Abstract

The inlet air flow over surface of heat exchanger in many cases for the direction is not orthogonal. To study of the heat transfer with change of air inlet angles, the paper presents how to predict the heat transfer coefficient for four flat tubes in a crossflow of air using an artificial neural networks (ANNs). The experimental setup with inclined the air incoming flow direction, the heat transfer coefficient of three air inlet angles (90°, 45° and 30°) are studied separately for five inlet air velocity 0.2, 0.5, 0.6, 0.8, and 1.2 m/s to corresponding the Reynolds number (Re), based on the transverse diameter are 1668, 2006, 2471, 2658, and 3782. The three cases of heat flux on the all tubes surface 13.2, 38.5, and 99.8 W/m2. The predicted results for heat transfer coefficient show a good agreement with experimental data. The accuracy between ANNs approach model and actual values (experimental) obtained with a mean relative error less than 2.2%, and the coefficient of determination (???? 2) around 99.8%.