Artificial Neural Network Assisted P &O-Based MPPT Controller for a Partially Shaded Grid-Connected Solar PV Panel

This paper presents an offline trained artificial neural network (ANN)-assisted perturb and observe (P &O) algorithm to track the global maximum power of a solar PV under partial shading conditions. The conventional hill climbing algorithms work very well to extract maximum power under uniform shading conditions, but fails under partial shading conditions. On the other hand, the offline trained ANN can track the power closer to global maximum power point (GMPP) with less oscillations. Hence, two methods are combined to design the MPPT controller for tracking the global maximum power accurately with less oscillations. Initially, ANN predicts the voltage near to the global maximum power point with solar PV voltage and current as the inputs. This predicted voltage from ANN is used as the initial values for the P &O algorithm, and the P &O algorithm is triggered for tracking the actual GMPP. The proposed MPPT controller is validated in the MATLAB/Simulink environment and the performance of the ANN model for the prediction of GMPP is validated with the experimental data sets.