Amélioration de l’algorithme MPPT pour un système PV par l’intégration des nouvelles techniques basées sur l’intelligence artificielle

Abstract

With the increasing integration of renewable energies into power grids, much research is being conducted to improve the efficiency of conversion systems. The global challenge is to reduce the cost of production and maximize the use of sustainable resources. In this regard, photovoltaics is considered a very promising source in terms of implementation costs and the number of applications using PV panels. Several MPPT techniques have been developed to track the maximum power point. The Perturb and Observe (P&O) technique is one of the conventional techniques used for MPP tracking, but it has several drawbacks in terms of response time and output signal quality. In our research, a method based on a fuzzy logic controller (FLC) that corrects the limitations of conventional algorithms has been developed. This technique combines the advantages of P&O-MPPT to account for slow and fast variations in solar irradiance, as well as the reduced processing time of FLC-MPPT, to solve complex technical problems with a reduced number of membership function rules. As a result, the proposed technique achieves average tracking efficiencies of about 99.6% in the standard EN50530 test, and this method offers better response time and less oscillations than conventional methods. As a second part of our study, a model based on the adaptive neuro-fuzzy inference system (ANFIS) was developed. This technique combines a fuzzy logic controller and artificial neural networks (ANN). Test results are simulated on Matlab/Simulink with a comparison of the proposed models with other MPPT techniques such as Perturb and Observe (P&O) and Incremental Conductance (INC).

Date	4 Jul 2022
Original language	French
Awarding Institution	École de technologie supérieure
Supervisor	Ambrish Chandra (Supervisor)