Optimal Setting of ANN Weights using Resilient Backpropagation Approach in a Grid Connected Solar PV System

Optimal setting of weighted interconnections between the layers of an artificial neural network (ANN) is an essential task. In this paper, Resilient Backpropagation (RBP) machine learning approach has been used in order to set the weights optimally. As the fossil fuels are getting depleted, renewable energy sources (RESs) have become prominent in meeting the power demand globally. Solar photovoltaic (PV) systems on the other hand among all the RESs have grabbed massive attention of the globe in generating the power. However, extracting the peak power is again a challenging task for many as the nature of solar irradiation is highly intermittent. As conventional maximum power point tracking methods (MPPT) failed to address some issues like convergence speed, oscillations in the operating point, accuracy, ANN MPPT approach found to be the best alternative in addressing the above difficulties. As learning in an ANN is the key aspect, precise learning can be happened only through proper setting of ANN weights. The simulation is carried out through MATLAB/Simulink software.