الفهرس 
IntroductionOnly 14 pages are availabe for public view
 |  | from | 98 |  |  |
| | | from | 98 | | |
Abstract
PV systems became an important topic for research. Many scientists and engineers are searching for techniques and algorithms to increase the efficiency of photovoltaic systems and hence decrease its cost. Artificial intelligence (AI) algorithms and techniques play a critical role in increasing PV systems efficiency. This thesis first gives a literature study about the role of AI algorithms in PV systems design, control, and fault diagnosis. Then the thesis concentrates on the role of intelligent control in PV systems by introducing a Simulink model implementation for a fuzzy controller of MPPT. The thesis introduces a novel reconfigurable generic fuzzy MPPT controller on FPGA. The fuzzy controller is written using VHDL code and simulated using Xilinx ISE tool. The work investigates also applying the reinforcement learning AI algorithm for the MPPT problem. A complete Simulink model for the reinforcement learning MPPT algorithm is implemented. Finally, the thesis investigates the role of AI algorithm grid connected inverter. A 9.1 kW complete PV system is implemented in Simulink. The system is tested for two scenarios 2kw loads and 12kw loads. A hardware implementation of a complete MPPT and inverter control is provided and experimental results are presented.
The main contributions for the thesis are:
1- A complete study about the role of AI in PV research. The study shows the critical role that AI plays in the design, control and fault diagnosis of PV systems.
2- A novel reconfigurable generic fuzzy controller implemented in FPGA.
3- Using a novel reinforcement learning algorithm to solve the MPPT in PV systems.
4- A complete Simulink model for 9.1kw grid connected system.
The thesis is divided into seven chapters; the first chapter gives first a brief introduction to the thesis then it describes in detail the thesis contribution and finally introduces the thesis outline.
The second chapter presents a study on the role of AI techniques in the design, control and fault diagnose of PV systems. The chapter gives a study on parameter identification problem. The study shows that AI techniques achieve higher efficiency than conventional techniques. The chapter gives a study on the problem of sizing of PV systems. The study shows also that AI techniques achieve higher efficiency in this research area. The chapter gives a study on maximum power point tracking problem. The study shows the superior performance of AI techniques in MPPT problem. The chapter gives also a literature survey about inverter control in PV systems. The survey shows the critical role that AI plays in inverter control. The chapter also gives a study about the sun tracking problem. The study shows the critical role of AI algorithms in this problem. The chapter also gives a brief literature survey about solar irradiance forecasting. The survey shows that AI is essential in this research area. The chapter also gives a brief study about output power forecasting. The study shows the role of AI in this research area. Finally the chapter gives a literature survey about fault diagnosis problem. The survey shows the critical role that AI plays in solving this problem. Then the chapter gives a conclusion section and concludes the best AI algorithm for each application.
The third chapter provides in detail a mathematical model for PV panels. The chapter also gives a study on DC/DC boost converter main design equations. The chapter introduces perturb and observe algorithm and explains how to implement it on Simulink. The chapter then gives our proposed fuzzy controller of maximum power point tracking for A 60 watt PV system. The chapter provides detailed Simulink model representation and explained results. The chapter also gives a comparison between fuzzy and perturb and observe controllers. The comparison shows the superior performance of fuzzy controller in terms of efficiency, velocity and steady state oscillations.
The fourth chapter presents our novel reconfigurable generic fuzzy controller implemented in FPGA. The chapter first explains the importance of FPGA. Then the chapter gives a literature survey about the work done in this area. The chapter provides detailed description of the design including VHDL code and Xilinx tools simulations.
The fifth chapter describes our novel reinforcement learning algorithm and it is application for the maximum power point tracking point. The chapter first gives a brief introduction about reinforcement learning. Then the chapter gives a brief introduction about the four elements that represent the reinforcement learning problem. The chapter gives a comparison between fuzzy logic, perturb and observe and reinforcement learning maximum power point tracking algorithms. Then the chapter explains the future work in this point.
The sixth chapter introduces the grid connected PV system. We present A 9.1 Kw grid connected PV system is implemented in Simulink. The system is tested for two scenarios. The first scenario when the loads are lower than the power produced from the PV system and the other scenario when the loads need more power than the power produced by the PV system. Then the chapter introduces the hardware implementation of a complete grid connected PV inverter