Using artificial neural network models and particle swarm optimization for manner prediction of a photovoltaic thermal nanofluid based collector

The present study introduces a new approach to model a photovoltaic thermal nanofluid based collector system (PVT/N). Two artificial neural networks of radial-basis function artificial neural network (RBFANN) and multi-layer perception artificial neural network (MLPANN), as well as adaptive neuro fuzzy inference system (ANFIS) model are used to identify a complex non-linear relationship between input and output parameters of the PVT/N system. Fluid outlet temperature of the collector and the electrical efficiency of the photovoltaic unit (PV) are selected as two essential output parameters of the PVT/N system. In each model, the optimized structure is obtained through a Particle Swarm Optimization (PSO) technique. Zinc-oxide/water nanofluid is considered as the working fluid of the PVT/N setup. Experiments are repeated in ten days with thirteen data points in each day such that different environmental conditions are included in the measurements. Results of the three above-mentioned models are compared and validated with those of the measurements. All three models were found to be reasonably capable of estimating the performance of the PVT/N system. Moreover, the analysis of variance (ANOVA) results indicated that the ANFIS and RBFANN were more accurate in predicting the electrical efficiency and fluid outlet temperature, respectively. (C) 2016 Elsevier Ltd. All rights reserved.