Performance evaluation of hybrid photovoltaic double pass facade connected in series: An ANN approach

This paper presents an application of multi-layer perceptrons neural network to model performance of Nm semi transparent hybrid photovoltaic thermal double pass facades (HPVT-DPF) connected in series for New Delhi weather station of India. ANN has been developed using MATLAB 7.1 with four layers: input layer with 9 neurons having ambient air temperature, number of sunshine hours, number of clear days, temperature coefficient, cell efficiency, global radiation, diffuse radiation, number of rows and number of columns as input parameters and output layer with four neurons having electrical energy, thermal energy overall thermal energy and exergy as output parameters along with two hidden layers with 25 neurons. The feed-forward back propagation algorithm has been used for analysis. The transfer function, neural network configuration and learning parameters have been selected based on highest convergence during training and testing of network. Further the experimental validation of the proposed system for New Delhi weather station for one unit has also been performed.