Islanding Detection Method Of Micro-Grid Based On Multi-Point Connection Of Fuzzy Neural Network

In a Smart Grid(SG)configuration,islanding detection is an important aspect of interconnected micro-grids(MGs).In recent years,various island detection technologies have been developed.They are mainly divided into remote and local detection methods based on electrical signals measured on the SG and MG sides,respectively.Micro-grid can be divided into many smaller sub-grids,which include small MGs with multiple connection points.These small MGs are very stable grids,but these sub-networks and connections are not suitable for MG operation because of network protection and operational practices are not conducive to the operation of safe or reliable distributed generation.At present,the power system has changed,including advanced monitoring,control and communication technology,the SGs will be become more flexible and reliable.In this paper,a method of islanding detection for micro-grid based on multi-point connection of fuzzy neural network is proposed,which can solve the problem of islanding detection for multi-point connection between micro-grid and smart grid.In the simulation research,the signal measurement point is set on the side of the smart grid breaker,the acquisition signal are voltage and current,and 8 wavelet energy components are separated by discrete wavelet transform,which is used as the input of adaptive fuzzy neural network.probability of Islanding(PoI)microgrid is obtained by inference calculation of fuzzy neural network,then sent the value to the microgrid control center to determine whether islands occur in the microgrid.The method can not only realize active and passive island detection with communication,but also realize active and passive island detection in the case of communication failure.The simulation experiment shows that the island detection rate reaches 99.5%,which has a good application prospect.