Abstract-The world wide resource crisis led scientists and engineers to search for renewable energy sources. Photovoltaic systems are one of the most important renewable energy sources. In this paper we propose an intelligent solution for solving the maximum power point tracking problem in photovoltaic systems. The proposed controller is based on reinforcement learning techniques. The algorithm performance far exceeds the performance of traditional maximum power point tracking techniques. The algorithm not only reaches the optimum power it learns also from the environment without any prior knowledge or offline learning. The proposed control algorithm solves the problem of maximum power point tracking under different environment conditions and partial shading conditions. The simulations results show satisfactory dynamic and static response and superior performance over famous perturb and observe algorithm.Index Terms-Photovoltaic, maximum power point tracking, reinforcement learning.
A photovoltaic system is one of the major sources of renewable energy. The grid-connected inverter controllers play an important role in the conversion and transmission of solar energy. Therefore, they must be improved to meet the demands for grid interconnection. This article introduces the design and hardware implementation of the intelligent fuzzy-PI controller of the inverter part of the grid-connected photovoltaic system. First, the paper discusses the design of the three-phase grid-connected fuzzy-PI controller. Next, the paper describes the implementation of a Matlab graphical user interface (GUI) to design any grid-connected inverter and size the photovoltaic systems. The code generation of the fuzzy-PI controller of the system is accomplished by using Matlab Simulink simulation software. The hardware components of the PV system are implemented experimentally. In the hardware implementation, a 70 W prototype is realized to test the functionality of the controller, such that one can develop a realistic controller without taking risks or falling into security concerns in the case of performing experiments on high-power systems. The prototype proves that the controller model can be directly transformed from Simulink to the control device. It also shows that the fuzzy-PI controller is working properly in the 70-watt prototype. The achieved performance parameters of the proposed fuzzy-PI controller are satisfactory. The proposed method to design and implement the fuzzy-PI controller does not require complicated programming, where a Matlab coder is proposed to transform the Simulink controller into C code that can be directly utilized as a software control program loaded in the microcontrollers embedded in the hardware of the controller. The main result is that the fuzzy-PI controller for the three-phase grid-connected systems can be implemented using low-cost reconfigurable microcontrollers.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.