Neural network methodology for heat transfer enhancement data

PurposeThe purpose of this paper is to study experimentally enhancement of heat transfer in a tube with axial swirling‐flow promoters. The geometric features of flow geometry to improve heat transfer can be selected in order to yield the maximum opposite reduction in heat exchange flow irreversibility by using exergy‐destruction method. The paper seeks to illustrate the use of neural network approach to analyze heat transfer enhancement data for further study in the scope of the experimental program.Design/methodology/approachFor this purpose, 402 experimental measurements are collected. About 225 of those are used as training data for neural networks, the rest is used for testing. Then, these testing results of artificial neural network (ANN) and experimental data are compared. A formula for presenting exergy loses in a tubular heat exchanger is derived first and then the thermodynamic optimum instead of economic optimum is found by minimizing the exergy losses in the system.FindingsResults from all configurations studied show that the heat transfer rate of the heated increases when the swirling‐flow promoter is inserted. From the heat transfer improvement number defined, it is observed that about 100 percent increase in heat transfer rate and five times increase in the pressure drop can be achieved under the condition of constant flow for the single promoter which has three blades, its blade angle is 30° and its location is in the middle of the tube length.Research limitations/implicationsThe back‐propagation (BP) algorithm was selected as the neural network algorithm, which uses the generalized delta learning rule. The training time of BP algorithm is considerably long. However, the testing of our neural network is real‐time.Practical implicationsThe experimental setup is established to collect the experimental data. It consists of an entrance region, test region (heat exchanger and steam generator), and, flow measurement and control. Also, a software program of neural networks trained BP is written by using Pascal high‐level languages.Originality/valueAn alternative and new approach is proposed in the paper to find optimum flow geometry for a pipe flow with an axial swirling‐flow promoter inserts. It is too difficult to predict the response of a complex physical system that cannot be easily modeled mathematically. The result thus obtained compare well with experimental results, but the computational effort of the ANN and time required in the analysis is much faster as compared. These results show that the ANN can be used efficiently for prediction.