Prediction of nanofluid flows’ optimum velocity in finned tube-in-tube heat exchangers using artificial neural network

Abstract

Abstract The average flow velocity in heat exchangers is considered less often and thus needs further and detailed investigation because of its crucial influence on the overall thermal performance of the application. The use of nanofluids has similar influences to finned tube designs. Considering the rise in heat transfer and pressure drop, uncertainties in cost analyses with the uses of fins and nanoparticles, evaluation of optimum operating velocity of the fluids is necessary. On the contrary, there aren’t enough experimental, parametric, or numerical investigations present on this subject. The use of machine learning techniques to heat transfer applications to make optimization becomes popular recently. In this work, important factors of the process as tube number, cleanliness factor, and overall cost as output factors have been estimated by an artificial intelligence method using 339 data points. The influence of input factors of Reynolds number, thermal conductivity, specific heat, viscosity, and total fin surface efficiency on the outputs have been studied. Total tube number, cleanliness factor, and total cost analysis have been determined with deviations of −0.66%, 0.001%, and 0.12% as a result of the solution with 6 inputs, correspondingly.