Hybrid MPPT method based on Neural Network and Perturb & Observe for PV systems

To track accurately and fast the Maximum Power Point (MPP), a hybrid technique NN-P&O switched between Neural Network (NN) and Perturb and Observe method (P&O) according to the variation of irradiation was proposed. The considered methodology is based on voltage reference estimated by NN and achieved using proportional-integral controller (PI). The error between the actual power and the optimal power was minimized using a small duty cycle steps generated by P&O method, which initial duty cycle value was updated adaptively. To approve the efficiency of the proposed control algorithms, simulations have been performed considering different system responses as the current, voltage and essentially the power under changing weather conditions (irradiance or temperature values).