Heat Transfer and Friction in Helically-Finned Tubes Using Artificial Neural Networks

The last few decades have seen a significant development of complex heat transfer enhancement geometries such as a helically-finned tube. The arising problem is that as the fins become more complex, so does the prediction of their performance. Presently, to predict heat transfer and pressure drop in helically-finned tubes, engineers rely on empirical correlations. Tubes with axial and transverse fins have been studied extensively and techniques for predicting the friction factor and heat transfer coefficient exist. However, fluid flow in helically-finned tubes is more difficult to model and few attempts have been made to obtain non-empirical solutions. Friction and heat transfer in helically-finned tubes are governed by an intricate set of coupled and non-linear physical interactions. Therefore, obtaining a single prediction formula seems to be an unattainable goal with the knowledge engineers currently possess. Regression techniques performed on experimental data require mathematical functional form assumptions, which limit their accuracy. To achieve accuracy, techniques that can effectively overcome the complexity of the problem without dubious assumptions are needed. One of these techniques is the artificial neural network (ANN), inspired by the biological network of neurons in the brain. This chapter presents an introduction to artificial neural networks (ANNs), and a literature review of the use of ANNs in heat transfer and fluid flow is also discussed. In addition, this chapter demonstrates the successful use of artificial neural networks as a correlating method for experimentally- measured heat transfer and friction data of helically-finned tubes.