A new islanding detection approach using wavelet packet transform for wind-based distributed generation

Methods for detecting island in an electric power system are classified as active, passive or hybrid. For certain applications, passive methods are sometimes favored over other detection methods because they are cheap, simple and do not affect the power quality of the electric power system. This paper introduces an islanding detection approach based on wavelet packet transform using only local measurements of voltage and current at the point of common coupling. A new index named node rate of change of power defined in the time-frequency domain is introduced to quantify the rate of change of output power of the distributed generation (DG). The new index has been evaluated in seven different case studies including island case. The presented results have shown that the new index is effectively capable of determining the island case from other types of disturbances.