XML Print


Abstract:   (121 Views)
This paper presents a pattern recognition-based scheme for detection of islanding conditions in synchronous- based distributed generation (DG) systems. The main idea behind the proposed scheme is the use of spatial features of system parameters such as the frequency, magnitude of positive sequence voltage, etc. In this study, the system parameters sampled at the point of common coupling (PCC) were analyzed using reduced-noise morphological gradient (RNMG) tool, first. Then, the spatial features of the RNMG magnitudes were calculated. Next, to optimize and increase the ability of the proposed scheme for islanding detection, the best features with a much discriminating power were selected based on separability index (SI) calculation. Finally, to distinguish the islanding conditions from the other normal operation conditions, a support vector machine (SVM) classifier was trained based on the selected features.  To investigate the power of the proposed scheme for islanding detection, the results of examinations on the various islanding conditions including system loading and grid operating state were presented.  These results show that the proposed algorithm reliably detect the islanding condition within 32.7 ms. 
Full-Text [PDF 3479 kb]   (66 Downloads)    
  • A novel non-communication-based intelligent islanding detection scheme is proposed which is based on an effective simple digital filter being robust to the environmental noises.
  • In the proposed algorithm, the points of the signal (such as voltage, active power, etc.) with an irregularity are highlighted and, then be used for islanding detection.
  • To improve the performance of the proposed scheme, the separability index has been used to select the most valuable features.
  • The proposed scheme is capable of islanding detection in synchronous-based distributed generation systems in situations with the least exchanged power between the DG resources and the distribution system.

Type of Study: Research Paper |
Received: 2018/11/08 | Accepted: 2019/10/02 | Published: 2019/10/04

Creative Commons License
© 2019 by the authors. Licensee IUST, Tehran, Iran. This is an open access journal distributed under the terms and conditions of the Creative Commons Attribution-NonCommercial 4.0 International (CC BY-NC 4.0) license.