Modeling and Optimization of Nanofluid Based Heat Exchanger using ANN

Abstract

In the present study the heat transfer coefficient and friction factor is simulate by the artificial neural feed forward network for double tube heat exchanger using TiO2-water nanofluid with twisted tape. The back propagation learning algorithm is used by ANN model. The developed ANN model for heat transfer coefficient have regression coefficient of 0.9974. Heat transfer coefficient and friction factor are experimentally investigated in a turbulent flow regime. The volume concentration of TiO2-water nanofluid is 0.01%, 0.025% and 0.05% and twisted tape ratio 3.5, 4.6 and 6.5 respectively. The mass flow rate of hot side fluid ranges from 0.0163- 0.0816 Kg/s and cold side flow rate is fixed to 4LPM. The experiments were conducted both water and TiO2-water nanofluid. The result shows that the Dittus-Bolter predicted Nusselt number with good agreement with experiment for plain tube. The Nusselt number enhances with Reynold number for plain tube. The heat transfer coefficient for twisted tape increases with decrease in twisted tape that is due to the increase in residence time, swirl flow and better fluid mixing. The heat transfer coefficient enhances 53.4%, 47.9% and 34.4% with twisted tape ratio 3.5, 4.6 and 6.5 compared to plain tube. Furthermore with increase in concentration TiO2-water the maximum heat transfer coefficient enhance 3.17% with 0.05% TiO2-water the compared to water. The use of these enhancement techniques effects the friction factor too. The predicted friction by Blasius have good agreement with the experimental data. The developed ANN model for friction factor have regression coefficient of 0.9902. The friction factor decrease with increase in Reynold number. The use of twisted tape led to enhancement in friction factor as well. The results show that as twisted tape ratio decrease the friction factor enhances. The twisted tape ratio 3.5 provides highest friction factor 1.96 times compared to water in plain tube. The twisted tape insert with TiO2-water nanofluid provides the highest friction factor and heat transfer coefficient which is about 1.96 times and 53.4% compared to plain tube with water as fluid. The ANN predicts the result accurately compared to the correlation used to validate the tube and twisted tape. The predicted heat transfer coefficient and friction factor by ANN developed model are found to be in good agreement with experimental data.