Shape effects of TEG mounted ventilated cavities with alumina-water nanofluids on the performance features by using artificial neural networks

Abstract

Shape effects of TEG mounted vented cavities on the performance characteristic during alumina-water nanofluid convection are numerically assessed by using finite element method. Rectangular, triangular, l-shaped and Ushaped cavities are used. The interface temperatures of hot and cold side are varied by changing the cavity shape, opening ratio (OR) and Reynolds number (Re). The highest hot side temperature is obtained with l-shaped cavity followed by U, T and R-shaped cavities. When using T, L and U shaped cavities, the rise of the power are calculated as 38%, 78% and 76% atRe=1000 when compered to rectangular cavity. The power rises with higher OR while increment amounts are 83.9%, 63.5%, 42% and 87.8% for R, T, L and U cavities a Re=1000. When nanofluid is used at the highest loading, 11%, 10.3%, 9% and 8.5% rise of power are achieved for R, T, L and U shaped cavities while the amount is than 5% when different sizes of particles are considered. Neural network modeling with 10 neurons in the hidden layer provides accurate power outputs for all shaped cavities.