Žukauskas A. Heat transfer from tubes in crossflow. Advanced Heat Transfer 1972; 8: 93–158.
 Benarji N, Balaji C, Venkateshan SP. Unsteady fluid flow and heat transfer over a bank of flat tubes. Heat Mass Transfer 2007; 44(4): 445–461.
 Kaptan Y, Buyruk E, Ecder A. Numerical investigation of fouling on cross-flow heat exchanger tubes with conjugated heat transfer approach. International Communication Heat and Mass Transfer 2008; 35(9): 1153–1158.
 Incropera FP, Dewitt DP, Bergman TL, Lavine AS. Fundamentals of heat and mass transfer. 7th ed. John Wiley & Sons, Inc: USA; 2011.
 Min JC, Webb RL. Numerical analyses of effects of tube shape on performance of a finned tube heat exchanger. Journal Enhanced Heat Transfer 2004; 11: 63–76.
 Webb RL, Kim N-H. Principle of enhanced heat transfer. 2nd ed. Taylor & Francis: New York; 2005.
 Atayılmaz ŞÖ, Demir H, Ağra Ö, Application of artificial neural networks for prediction of natural convection from a heated horizontal cylinder. International Communication Heat and Mass Transfer 2010; 37(1): 68–73.
Tikrit Journal of Engineering Sciences (2018) 25(2) 59- 67
Hybrid CFD-ANN Scheme for Air Flow and Heat Transfer Across In-Line Flat Tubes Array
|Ataalah Hussain Jassim||M.M. Rahman||Khalaf Ibrahim Hamada||M. Ishak||Tahseen Ahmad Tahseen
|Department of Mechanical Engineering, Tikrit University, Iraq||Faculty of Mechanical Engineering, Universiti Malaysia Pahang, Malaysia||Department of Mechanical Engineering, Tikrit University, Iraq||Faculty of Mechanical Engineering, Universiti Malaysia, Malaysia||Department of Mechanical Engineering, Tikrit University, Iraq|
Flat tubes are vital components of various technical applications including modern heat exchangers, thermal power plants, and automotive radiators. This paper presents the hybridization of computational fluid dynamic (CFD) and artiﬁcial neural network (ANN) approach to predict the thermal-hydraulic characteristics of in-line flat tubes heat exchangers. A 2D steady state and an incompressible laminar flow in a tube configuration are considered for numerical analysis. Finite volume technique and body-fitted coordinate system are used to solve the Navier–Stokes and energy equations. The Reynolds number based on outer hydraulic diameter varies between 10 and 320. Heat transfer coefficient and friction are analyzed for various tube configurations including transverse and longitudinal pitches. The numerical results from CFD analysis are used in the training and testing of the ANN for predicting thermal characteristics and friction factors. The predicted results revealed a satisfactory performance, with the mean relative error ranging from 0.39% to 5.57%, the root-mean-square error ranging from 0.00367 to 0.219, and the correlation coefficient (R2) ranging from 99.505% to 99.947%. Thus, this study verifies the effectiveness of using ANN in predicting the performance of thermal-hydraulic systems in engineering applications such as heat transfer modeling and fluid flow in tube bank heat exchangers.
Keywords: In-line flat tube, finite volume technique, CFD, modeling, friction factor, artiﬁcial neural networks