Neural network modelling and performance estimation of dual-fluid photovoltaic thermal solar collectors in tropical climate conditions

Abstract

A photovoltaic/thermal solar collector which uses air and water under the same collector area is known as a bi-fluid type or dual fluid photovoltaic thermal (PV/T) solar collector. A dual fluid PV/T system is more complex than a single fluid system, as two fluids are used simultaneously, making boundary conditions complicated. In addition, the influence of transient and variations in weather conditions leads to challenges in performance evaluation, both experimentally and theoretically. The aim of this study is to develop and test an Artificial Neural Network (ANN) model for a dual fluid PV/T collector which utilizes water and air as working fluids, to provide a consistent method for predicting dual fluid PV/T collector performance. 60%–80% of the data was randomly used to train the ANN, while 40%–20% was randomly selected for testing. New data was then used to validate the predictive model based on the best splitting percentage and the neural network architecture. Both the experimental and predictive models have been found to be in good agreement, with root mean square error RMSE values for total PV/T efficiency in air mode, water mode and simultaneous mode reaching 0.0069, 3.971 and 2.6338, respectively.