Artificial neural networks/least squares fuzzy system methods to optimize the performance of a flat-plate solar collector according to the empirical data

Abstract

Reduced graphene oxide (rGO) has applications in Water purification, and Energy conversion. In this study, a water-based nanofluid containing rGO was formed at specific concentrations. The nanofluid Thermo-Rheology behavior was studied at room temperature to 50 °C. Viscosity was detected at specific RPMs from 10 to 100. The results showed that this nanofluid has excellent thermo-rheology properties. Flat plate solar collectors could heat the fluid inside using sunlight from a wide range of varied angles. Thus, the prepared nanofluid was used as the working fluid in the solar collector tubes. The results showed that this nanofluid can be used instead of water. The aims of this study are to optimize the process and lessen the examination costs, thus, Artificial Neural Networks algorithms of Orthogonal Distance Regression (ODR), Levenberg Marquardt (LM), and Fuzzy system of Recursive Least Squares were trained. Results proved that Artificial Neural Network and Fuzzy systems should be trained to predict the data of thermal conductivity and viscosity with acceptable coefficient of determination.