Impact of distributed generation mix on the effectiveness of islanded operation detection

Tzelepis, Dimitrios and Dysko, Adam (2015) Impact of distributed generation mix on the effectiveness of islanded operation detection. In: Sixth Protection, Automation and Control World Conference (PAC World 2015), 2015-06-29 - 2015-07-03, Technology and Innovation Centre, University of Strathclyde.

[thumbnail of Tzelepis-Dysko-PAC-2015-distributed-generation-mix-on-the-effectiveness-of-islanded-operation-detection]
Preview
Text. Filename: Tzelepis_Dysko_PAC_2015_distributed_generation_mix_on_the_effectiveness_of_islanded_operation_detection.pdf
Accepted Author Manuscript

Download (487kB)| Preview

Abstract

Distributed generation can be understood as a process where large scale power generation is gradually replaced by smaller power generation facilities with reduced power yield, and mostly connected at the system distribution level. One of the most important requirements for interconnecting distributed generation to healthy power networks is the Loss of Mains (or Islanding) detection. During a Loss of Mains (LOM) event a part of the grid (including distributed generation) losses physical connection with rest of the grid. A condition like this should be detected and actions to disconnect distributed generation should be initiated, in order to protect life and property. A very common passive method used to detect an islanding event, is the Rate of Change of Frequency (ROCOF). Since distribution networks nowadays are accommodating a great amount converter-interfaced generation, there is a risk that such methods may fail to successfully operate or operate spuriously, putting system stability at risk. Most of the existing LOM protection performance studies, consider only a single generator within the islanded part of the network. While historically such approach was reasonable, rapidly increasing numbers of DG connections lead to high probability of islanding with more than one generator in the mix. Therefore, this paper, considers various mixes of generation to investigate how this impacts LOM detection performance. In particular studies are undertaken with a few identified most likely combinations of distributed generators.