Prediction of heat transfer enhancement of delta-wing tape inserts using artificial neural network

This work adopts the artificial neural network (ANN) for predicting the efficiency of a double-pipe heat exchanger which employs T-W tape inserts with different wing-width ratios (w/W) of 0.31, 0.47, and 0.63. The effects of friction factor (f), Nusselt number (Nu), and thermal performance (η) are predicted using the established multi-layer ANN. Different scenarios are examined using two parameters as inputs to the ANN: Reynolds number (Re) and w/W. The results prove that the developed ANN model is able to accurately predict the experimental data. The obtained mean square error has a value of less than 0.7 as compared to the experimental values. Furthermore, the proposed ANN-based approach is also effective to predict the thermal parameters, with the least variance and high precision. In addition, a multiple linear regression is employed to check the efficiency of the proposed model from which it is demonstrated that the suggested neural network method provides useful guidance for accurately predicting the thermal parameters with the least variance. The configuration of 2–10-1 is found to be the best for the current model, with a mean absolute error of 0.546896.