Investigation of antiislanding protection of power converter based distributed generators using frequency domain analysis

Islanding of a grid connected distributed generation (DG) unit occurs when a section of the utility system containing such generators is disconnected from the main utility, but the independent DGs continue to energize the utility lines in the isolated section (termed as an island). Unintended islanding is a concern, primarily because it poses a hazard to utility and customer equipment, maintenance personnel and the general public. The algorithm proposed by the Sandia National Laboratories is analyzed in this study because it is considered to be effective in detecting the formation of such islands. Previously, there has not been any quantitative analysis for tuning the control gains of the algorithm based on the rating and bandwidth of the DG power converter. The paper interprets the components of the algorithm that affect the voltage magnitude and frequency into block diagrams that can be linearized and studied using continuous time approximations. The model represents DG operation while connected or disconnected from the main utility. This paper uses frequency domain approach to analyze the range for the gains required by antiislanding algorithm to effectively determine the disconnection of the mains grid within an acceptable time duration. The effect of variation of DG´s operating power level is studied and demonstrated how the gains of the algorithm can be set to be effective for a wide power range. The results are validated using detailed time domain DG and power system simulations.