Application of radial basis function neuron networks in an identification and prediction of maximum power point of photovoltaic-array

Photovoltaic (PV)-array is difficult to model with analytical methods because it is highly non-linear characteristics, and the traditional model is not satisfied for the design of control systems. Because radial basis function neuron networks (RBFNN) have excellent ability to approach any complex non-linear function, this paper sets up the model of PV-array based on the radial basis function neuron network technique to avoid the complexity of the modeling. The solar radiation and ambient temperature impacting on PV-array are input variables of the RBFNN, the output voltage and current corresponding to the maximum power output of PV-array are output variables of the RBFNN. The training samples are a practice PV-array data given in predecessors' paper. The simulation results presented in this paper verify the RBFNN is an excellent modeling method for PV-array, the RBFNN model can be used to identify and predict the maximum power point of PV-array in its control system.