Study On Islanding Detection Of Photovoltaic Grid-connected Inverters

In recent years, photovoltaic power generation is developed rapidly and thegrid-connected operation is achieved. With more and more photovoltaic power generationsystems are connected to the utility grid, the resulting islanding effect gets a great attention.The existence of islanding effect may bring damage to the safety of the system equipment andthe staffs, thus effective detection of islanding effect is a key problem which must be solvedfor the photovoltaic grid-connected system. The existing islanding detection methods includeremote detection methods and local detection methods. However the two kinds of methodshave their advantages and disadvantages respectively. Therefore, it has important practicalsignificance to research islanding effect and corresponding detection methods.Firstly, the dissertation introduces the background, development situation andsignificance of the islanding detection for the photovoltaic grid-connected inverters, and thenthe main research contents are presented. Secondly, the occurrence mechanism of islandingeffect is discussed through analyzing power flow of the photovoltaic grid-connected system.The power match of the photovoltaic grid-connected inverter and local loads is expatiated,and then the relationship between the power match and the variation of output voltage oroutput frequency of the photovoltaic grid-connected inverter is expounded. Thirdly, thedissertation analyzes remote islanding detection methods, and studies the basic detectionprinciples and the advantages and disadvantages of various remote detection methods.Particularly, the smart remote detection methods are presented, and the comparison in usingof equipment and the adaptability to system extension for these methods are carried out indetail. Then the dissertation elaborates the detection principle and detection performance oflocal detection methods, and deduces the criterion which describes the non-detection zone ofthe methods. Fourthly, aiming at the problem that the conventional AFD scheme reduces thequality of the power delivered to the grid, the improved frequency perturbation scheme isdesigned to detect the islanding effect. The performance of the improved frequencyperturbation scheme is derived analytically, and the scheme is compared with the AFDscheme. Finally, a single phase the photovoltaic grid-connected system is modeled, and theimproved frequency perturbation scheme is simulated using MATLAB.Through the analysis of the power match of the photovoltaic grid-connected system, thedissertation presents the reasons influencing the performance of islanding detection, and theimproved frequency perturbation scheme for the photovoltaic grid-connected system isdesigned. The theoretical analysis results show that the improved frequency perturbationscheme can achieve a better performance which results in faster islanding detection andgenerates less THD than the AFD scheme. The simulation results in MATLAB prove the validity and feasibility of the improved frequency perturbation scheme.