Implementation of the artificial neural network to predict the effectiveness of the solar system using Cu/water-ethylene nanofluid to save energy

Abstract

A solar heat panel with the nanofluids (NFs) flow is simulated. For this purpose, an 80 W Photovoltaic (PV) with its cooling system (CLS) is simulated. The CLS consists of a copper pipe, and a copper plate with 1 mm thickness is located under the panel. Cu/water-ethylene (70-30) NFs flows with different flow rates in the CLS under heat flux of different hours of the day. The neural network is employed to estimate the thermal behavior of a PV that provides heat flow. For better analysis, an optimization is performed on the variables, including pipe diameter (dPipe), heat flux, and NFs flow rate on the output parameters such as maximum temperature (T-Max) and pressure drop (ΔP). The simulation results show the extraordinary ability of the neural network to estimate the output variables so that MSE < 10−5, R2 = 0.999, and MOD < 0.5%. The results of optimization revealed that the greatest fluid flow should be conducted via the pipe with the smallest (dPipe) and lowest heat flux in order to decrease the T-Max. The maximum undesirable panel temperature occurs at the opposite point. The minimum ΔP occurs when the panel has the pipe with the largest diameter, and the fluid passes through it with the lowest velocity.