Heat Transfer and Pressure Drop Prediction in an In-Line Flat Tube Bundle by Radial Basis Function Network

Abstract

This paper aims to predict the heat transfer and pressure drop for an in-line flat tubes
configuration in a cross-flow using an artificial neural network. The numerical study of
a two-dimensional steady state and incompressible laminar flow for an in-line flat tube configuration in a cross-flow is also considered in this study. The Reynolds number varies from 10 to 320. Heat transfer coefficient and pressure drop results are presented for tube configurations at three transverse pitches of 2.5, 3.0, and 4.5 with two longitudinal pitches of 3.0 and 6.0. The predicted results for the average Nusselt number and dimensionless pressure show good agreement with previous work. The accuracy between the actual values and the neural network approach model results was obtained with a mean absolute relative error less than 4.1%, 4.8%, and 3.8% for the average Nusselt number, dimensionless pressure drop and average friction factor, respectively.