Anticipation of power generated by photovoltaic panels based on artificial neural network and Matlab: article

This project is to design a feed forward multi­layer (MLP) neural network of PV panel by using artificial neural network (ANN) and Matlab application software. An approach based on artificial neural network (ANN) was developed in this project to anticipate the output power of photovoltaic (PV) panels. The Levenberg­Marquardt back propagation was used in the neural network as its learning algorithm. In this project, several sets of mono-crystalline historical data have been used to train and test the neural network in order to predict the performance mono-crystalline PV-panel. The inputs of the PV panels network are solar irradiance and ambient temperature of PV panels while the output is electrical power generated by the PV panels. Firstly, a dataset of 3x300 were used to train the network then another 3x165 data were used to test the trained network. The results of the trained and tested neural network based on Matlab were presented. The regression from data testing procedure shows that the use of AN for performance prediction of mono-crystalline PV panels is acceptable