Numerical investigation and optimization of flow and thermal characteristics of nanofluid within a chaotic geometry

In this study, hydrothermal characteristics of the water-TiO₂ nanofluid are evaluated in the rectangular C-shaped chaotic channel by applying the two-phase Euler-Lagrange method. The results show that the fluid temperature distribution in the C-shaped channel is much more uniform than that in the straight rectangular channel due to the intense mixing in the chaotic geometry. By increasing the concentration and reducing the particle size, both convective heat transfer and pressure drop increase in the chaotic channel. For finer particles, due to more intense particle migration related to the Brownian motion, more uniform concentration distribution is obtained at the channel cross section. The model of mean convective heat transfer coefficient and pressure drop in terms of concentration and particle size was developed by neural network. In order to obtain some cases of the variables (i.e. size and concentration of nanoparticles) in which maximum heat transfer along with minimum pressure drop occurs, optimization on neural network model was implemented using methods of two-objective as well as single-objective (i.e. genetic algorithm coupled with compromise programming decision making method). Eventually, optimum cases were suggested for different importance levels of objective functions.