Research On Islanding Detection Technology For Distributed Generation Grid-connected Photovoltaic System

Energy is related to people's livelihood.In order to solve the increasingly prominent energy problem,People set their sights to photovoltaic,wind power represented by the new energy.photovoltaic power generation is the most widely used With the increase of PV generation system,it will bring a series of adverse effects to the safety and stability of the original distribution network,The main effect is the islanding effect.In order to minimize the harm of islanding,it is of great practical effect to study the islanding method with good performance.This paper described the background of detecting islanding effects,introduced the concept of islands and the the adverse effects,described the current research status of island detection methods.overview of current research on the status of islanding detection technology.And explored the mechanism of islands appear,introduced the methods used to detect islanding frequently and the standard.for evaluating the advantages and disadvantages of this method.The detection principle and the advantages and disadvantages of each island detection method are analyzed and summarized.The active phase shift method in active detection is taken as the object of study,the detection principle of this method is expounded.first building the system model to simulate in the Matlab /Simulink.the detection speed and the existence of the detection of blind areas as an indicator to evaluate its performance.and on the basis of the existing automatic phase shift method,analyze the reasons why the method detects need a long time,and an improved automatic phase shift method is put forward.The simulation results show that the proposed method can improve the detection efficiency and reduce the blind area.In order to overcome the larger problems of the traditional passive island detection method blind area,while without affecting the quality of the power.Based on wavelet packet energy entropy and back propagation neural network passive islanding detection method is researched.In this method,to achieve the purpose of islanding detection,wavelet packet energy entropy is adopted to analyze the thevoltage of point of common coupling and the inverter output current.The corresponding wavelet packet entropy is obtained as the characteristic variable and the characteristic value is used back propagation neural network to recognize classification results.the simulation results show that detection method which is based on wavelet packet energy entropy and back propagation neural network can achieve the purpose of islanding detection,and there isn't impact on power quality.