Correlating heat transfer and friction in helically-finned tubes using artificial neural networks

An artificial neural network (ANN) approach was used to correlate experimentally determined Colburn j-factors and Fanning friction factors for flow of liquid water in straight tubes with internal helical fins. Experimental data came from eight enhanced tubes with helix angles between 25° and 48°, number of fin starts between 10 and 45, fin height-to-diameter ratios between 0.0199 and 0.0327, and Reynolds numbers ranging from 12,000 to 60,000. The performance of the neural networks was found to be superior compared to the corresponding power-law regressions. The ANNs were subsequently used to predict data of other researchers but the results were less accurate. The ANN training database was therefore expanded to include experimental data from two independent investigations. The ANNs trained with the combined database showed satisfactory results, and were superior to algebraic power-law correlations developed with the combined database.