Neural network based model of a PV array for the optimum performance of PV system

This paper presents an application of Artificial Neural Network (ANN) for modeling a photovoltaic (PV) module. A feedforward neural network consisting of two hidden layers containing 6 and 12 neurons (first Linear and second logsig transfer function) has been used in the designed network, and a set of measured data, (acquired through a data acquisition system realized on purpose) and Levenberg-Marquard backpropagation optimization function has been used for the training. Efficiency of the realized Neural Network model is evaluated by comparing the predicted values with a set of experimental data different from those used in training the ANN. The results of this work are useful to estimate the site productivity and to proceed to a correct design and optimization of the entire PV system.