Optimization of effective parameters on heat transfer in a wavy sinusoidal channel flow containing water-aluminum oxide nano fluid using an artificial neural network

In the present study, the optimization of effective parameters on heat transfer in a wavy sinusoidal channel flow containing water-aluminum oxide nanofluid was conducted using an artificial neural network. Due to the wavy shape of their walls, wavy channels can significantly the enhance heat transfer. This research aims to design and implement a neural network using MATLAB software based on the data extracted from the modeling and analysis of the mentioned channel. To achieve the research objective, a backpropagation neural network was selected as the optimum model. The Nusselt number, Darcy–Weisbach coefficient, and performance efficiency of the sinusoidal channel compared to a straight channel were considered as the neural network’s output parameters. The Reynolds number ranged between 6000 and 22,000, while the nanoparticle volume fraction varied between 0% up to 4%. Additionally, the sinusoidal channel height (5 –15 mm) and the channel length (50–100 mm) were defined as variable input parameters in the designed neural network. The results of this study indicate that the channel’s performance efficiency reaches its optimum state at the lowest height and the highest length, including the maximum nanoparticle volume fraction (4%).