Improved method of maximum power point tracking of photovoltaic (PV) array using hybrid intelligent controller

Generally, the solar photovoltaic (PV) system, which provides power to stand alone or grid connected systems, consists of a PV panel, Dc-Dc converter and a load. Maximum power point tracker (MPPT) is usually incorporated between the PV panel and a Dc-Dc converter to track the maximum power under changing solar irradiation and cell temperature. A fast and dynamic MPPT technique is desirable to track environmental variations without losing too much energy gains. In order to track the maximum power, an intelligent controller based MPPT algorithm for a standalone PV system is presented in this paper. For that purpose, hybrid techniques based on perturb and observe (P&O) Artificial Neural Network (PO-ANN) and Incremental conductance (INC) Artificial Neural Network (INC-ANN) are proposed and comparative analyses are made. For this purpose, a stacked auto encoders (SAEs) is trained with deep learning network with building blocks as a auto encoder to extract the maximum power from the solar panel. It is first trained using a greedy layerwise pattern, and then it uses a back propagation with supervised learning for fine-tuning the deep neural network with INC and PO to reach the maximum power. In addition to that, mathematical modeling of PV array is analyzed using a single-diode model using MATLAB/Simulink environment. It is evident from the results that the control scheme based on the hybrid INC-ANN with SAEs MPPT method is promising in tracking the maximum power with less oscillations under variable climatic conditions and load variations compared to other available techniques.