Développement d'une stratégie intelligente de contrôle et de gestion d'un système hybride connecté au réseau

Abstract

This thesis focusses on the development and implementation of advanced intelligent control and management strategies for a grid-connected hybrid renewable energy system, integrating photovoltaic (PV) arrays, proton exchange membrane fuel cells (PEMFCs), and wind energy conversion systems (WECS). Unfortunately, the mentioned sources face several challenges, including low conversion efficiency, particularly under fluctuating weather conditions, and the nonlinear characteristics of their output power and current. Therefore, the primary objective is to optimize power extraction, enhance grid integration, and ensure system stability under dynamic operating conditions. Novel maximum power point tracking (MPPT) algorithms are proposed for PV systems, including a high-order sliding mode Super-Twisting (STA) technique to mitigate chattering effects, a two-level artificial neural network-based model reference adaptive control (ANN-MRAC), and a double-stage ANN-finite control set model predictive control (ANN-FSC-MPC). Moreover, this research also focuses on the modeling and control design of variable-speed wind turbines. The aim is to optimize energy extraction from the wind below the rated power range, while controlling electrical power output above the rated power range, all while reducing mechanical transient disturbances. As matter of fact, the first control scheme is designed using classical regulators, specifically employing a PI controller. However, given the strong non-linearity and uncertainty of the wind turbine aerodynamics, a robust controller based high order slidi ng mode algorithms has been proposed. Additionally, a grid-side control strategy employing a finite control set model predictive controller (FCS-MPC) for a multilevel neutral-point clamped (NPC) inverter is introduced. This approach ensures DC-link voltage stability, minimizes switching frequency, reduces total harmonic distortion (THD), and maintains high-quality power injection into the grid, even during variable power generation from hybrid sources. To guarantee that the hybrid system can reliably meet power demands while respecting the operational limits of each energy source, an energy management strategy has been incorporated and critically discussed. The latter approach aims to balance energy distribution efficiently, ensuring uninterrupted power supply without exceeding the capabilities of individual components. Simulation results validate the superiority of the proposed methods, demonstrating enhanced efficiency, robustness, and dynamic response compared to traditional techniques. The developed strategies collectively address critical challenges in renewable energy integration, offering a scalable framework for reliable, sustainable, and intelligent hybrid energy systems