Intelligent-based approach to islanding detection in distributed generation

This paper introduces a new intelligent-based approach for detecting islanding in distributed generation (DG). This approach utilizes and combines various system parameter indices in order to secure the detection of islanding for any possible network topology, penetration level and operating condition of the DG under study. Hence, every parameter index displays characteristics for a given set of events. The proposed technique uses the data-mining technology to extract information from the large data sets of these indices after they are screened off-line via massive event analyses using network simulations. The technique is tested on a typical DG with multiple distributed resources and the results indicate that this technique can successfully detect islanding operations. In addition, this technique can also overcome the problem of setting the detection thresholds inherent in the existing techniques by optimizing their settings.