Development of a New Model for a (Neuron-Fuzzy) Tracker for Photovoltaic Energy Systems based on Open Circuit Technique With DC-DC Boost Converter

Abstract

The research introduces a new methodology for developing a tracker based on the use of artificial intelligence technologies such as Artificial Neural Networks (ANN) and Fuzzy logic (FL) to track the Maximum Power Point Tracker (MPPT) for Photovoltaic Systems (PV). The proposed neuronal- Fuzzy tracker action is achieved in two successive phases. In the first stage, the optimum operating tension (Voltage of Maximum Power Point, VMPP) for the PV system is estimated using a developed model of neural tension estimator (VMPP ANN Estimator, VMPP-ANNE), using the use of open circuit technology to obtain the database necessary to train the problem neural network for the neuronal estimator model. In the second stage, the value of the appropriate duty cycle is determined, using a developed model for the Fuzzy logic Controller (DFLC). Where the error in voltage and its variability is the input variables of the Fuzzy controller DFLC, which is used to determine the appropriate the duty cycle, which is used to control the DC-DC Boost Converter switching cycle. Thus, the proposed MPPT-ANN-DFLC is characterized by high MPP point tracking performance, on the one hand because it is based on the use of neuron networks that are characterized by high accuracy and very fast VMPP determination. On the other hand, the proposed tracker, is based on the use of Fuzzy logic that improves the dynamic performance of the DFLC controller. Simulation results performed in Matlab/Simulink environment showed the best performance of the proposed MPPT-ANN-DFLC tracker, compared to using several other reference tracker models.