Estimation of heat transfer in oscillating annular flow using artifical neural networks Unal Akdag a, * , M. Aydin Komur b , A. Feridun Ozguc c a Department of Mechanical Engineering, Aksaray University, TR-68100 Aksaray, Turkey b Department of Civil Engineering, Aksaray University, TR-68100 Aksaray, Turkey c Department of Mechanical Engineering, Istanbul Technical University, TR-34437 Istanbul, Turkey article info Article history: Received 23 May 2008 Received in revised form 8 January 2009 Accepted 22 January 2009 Available online 26 February 2009 Keywords: Artificial neural network Oscillating flow Heat transfer Annular duct abstract In this study, the prediction of heat transfer from a surface having constant heat flux subjected to oscil- lating annular flow is investigated using artificial neural networks (ANNs). An experimental study is car- ried out to estimate the heat transfer characteristics as a function of some input parameters, namely frequency, amplitude, heat flux and filling heights. In the experiments, a piston cylinder mechanism is used to generate an oscillating flow in a liquid column at certain frequency and amplitude. The cycle- averaged values are considered in the calculation of heat transfer using the control volume approach. An experimentally evaluated data set is prepared to be processed with the use of neural networks. Back propagation algorithm, the most common learning method for ANNs, is used for training and testing the network. Results of the experiments and the ANN are in close agreements with errors less than 5%. The study showed that the ANNs could be used effectively for modeling oscillating flow heat transfer in a ver- tical annular duct. Ó 2009 Elsevier Ltd. All rights reserved. 1. Introduction Heat and mass transfer in oscillating flow have been a principal investigation field in the past decade. Oscillation-induced heat transport processes maintain an effective heat enhancement. They have many important applications in compact heat exchangers, cryocoolers, Stirling engines, internal combustion machines and other periodical thermal systems. Some experimental, numerical and analytical studies on oscillating flow heat transfer and fluid flow characteristics are given by Zhao and Cheng [1]. They reported a large literature survey in oscillating flow on heat transfer. Cha- twin [2] and Watson [3], among many others, discussed the mass transfer in oscillating flow, and they stated that the oscillating flow enhanced the mass transfer remarkably. Based on the analogy be- tween heat and mass transfer, Kurzweg and coworkers [4,5] inves- tigated the longitudinal heat transfer process in laminar oscillating flow inside a tube bundle which was jointed to a hot reservoir on the upper side and a cold reservoir at the bottom. They showed that the longitudinal heat transfer was increased by laminar oscil- lating flow to the levels similar to that of heat pipes. A number of numerical and experimental studies were also reported in the lit- erature concerning active modulated oscillatory flow and heat transfer characteristics [6–8]. Some of the investigators are focused on convective heat transfer in oscillating flow to obtain higher heat transfer rates for smaller duct heights, higher oscillation frequen- cies and large tidal displacement for cooling electronic compo- nents [9,10]. Thermal design engineers are investigating the parameters of convective heat transfer in a duct to develop more effective heat transfer devices. For complex heat transfer problems, there are some solutions in the literature, which are studied experimentally, but the cost of experimental studies is very high. Mathematical and numerical methods are alternative tools for further analysis be- cause they have a lower cost than others for the development of new devices. Another promising method is artificial neural net- works (ANNs). The available literature shows that this method pro- vides more reasonable results for the fields such as; analysis of energy systems [11], heat transfer data analysis [12], HVAC com- putations [13] and predictions of critical heat flux [14]. Recently, Islamoglu and Kurt [15] analysed the heat transfer phenomena in corrugated channels using neural networks approximation. Some investigators used the ANNs for the prediction of the performance of heat exchangers in a newly designed refrigerator [16]. Some investigators performed dynamical analysis and simulation of the time-dependent behavior of heat exchangers [17,18]. Furthermore, Sreekanth et al. [19] used ANNs for the evaluation of surface heat transfer coefficient at the liquid–solid interface. Kalogirou et al. [20] investigated the performance of a solar water heating system with ANNs. Hosoz et al. [21] predicted various performance param- eters of a cooling tower. Yigit and Ertunc [22], developed an ANN model for the prediction of air temperature and humidity at the outlet of a wire-on-tube type heat exchanger. Sozen et al. [23] used the ANN approach for thermodynamic analysis of ejector–absorp- 0965-9978/$ - see front matter Ó 2009 Elsevier Ltd. All rights reserved. doi:10.1016/j.advengsoft.2009.01.010 * Corresponding author. Tel.: +90 382 2150953; fax: +90 382 2150592. E-mail address: uakdag@aksaray.edu.tr (U. Akdag). Advances in Engineering Software 40 (2009) 864–870 Contents lists available at ScienceDirect Advances in Engineering Software journal homepage: www.elsevier.com/locate/advengsoft