Data Mining Approach to Threshold Settings of Islanding Relays in Distributed Generation

This paper introduces a new approach for determination of the threshold settings of islanding relays in distributed generation (DG) interconnections. This approach uses data-mining technology to extract the optimal relay settings information from a large data set of system parameters. This data is constructed from offline simulation analyses of events and consequences. The mining of these consequences defines the boundary limits of the threshold settings that could secure the detection of islanding operations under: minimum detectable zones, multiple distributed resources (DRs), diverse distributed resource technologies, various operating conditions, and different network topologies. The approach is tested on a typical DG with multiple distributed resources, and the results indicate that this approach can be used effectively to support the setting relay decision.