Research On MPPT And Islanding Detection Technology In Photovoltaic Grid-connected Power Generation System

In the context of the global energy crisis and the increasingly serious ecological environment,renewable energy has been booming.Among them,solar energy has become the most important alternative form of fossil energy due to its many advantages.Solar photovoltaic power generation technology is an important way of using solar energy,so the research on photovoltaic power generation technology has great theoretical significance and practical value.The thesis takes photovoltaic grid-connected power generation system as the research object,and conducts in-depth research on the two most critical issues,namely maximum power point tracking technology and islanding detection technology.The thesis introduces the important components of photovoltaic grid-connected power generation system,and establishes corresponding mathematical model based on the topology.In the thesis,improved the most commonly used conductance increment method in MPPT control,and proposed the adaptive differential conductance increment method,built a simulation model of the two algorithms.The improved adaptive differential conductance increment method can effectively solve the problem of difficult step selection before the improvement,and can simultaneously satisfy the speed and accuracy of the maximum power point tracking,which can improve the dynamic performance of photovoltaic systems.In order to achieve global maximum power point tracking under partial shading condition,the thesis uses RBF neural network to perform global maximum power point tracking of photovoltaic power generation system.Then the deficiencies of the method are analyzed,and the RBF neural network by improved genetic algorithm is proposed,and optimizes the data center,weight and expansion constant of RBF neural network by genetic algorithm.In order to improve the convergence speed of the genetic algorithm,improved the genetic mechanism of the genetic algorithm.Finally,the GMPPT prediction of the photovoltaic power generation system is carried out.The prediction results verify that the improved algorithm can reduce the network prediction error and improve tracking of global maximum power point.Aiming at the problem of island detection in photovoltaic grid-connected power generation system,the thesis analyzes the traditional SMS island detection technology in detail,and proposes a new perturbation strategy to narrow the non-detection zone.Aiming at the selection of feedback coefficients,the thesis proposes an SMS islanding detection algorithm based on fuzzy control.By fuzzy optimization of the coefficients,shortened the island detection time and reduced the influence on the power quality of the grid.Built an improved island detection simulation model in MATLAB/Simulink,which verifies the reliability and superiority of the island detection method.