Loading...

Passive Islanding Detection for Distributed Generation Using Decision Tree Algorithm

Madani, Sohail | 2012

691 Viewed
  1. Type of Document: M.Sc. Thesis
  2. Language: Farsi
  3. Document No: 43272 (05)
  4. University: Sharif University of Technology
  5. Department: Electrical Engineering
  6. Advisor(s): Abbaspour Tehrani Fard, Ali; Ranjbar, Ali Mohammad
  7. Abstract:
  8. The integration of renewable energy sources introduces several issues including islanding operation. Therefore, Islanding phenomenon should be detected by a fast and reliable islanding detection method. In this thesis an intelligent-based approach is proposed to detect islanding state for Photovoltaic (PV) and Doubly Fed Induction Generator (DFIG) units. Decision tree classification algorithm is chosen as weak classifier and by Adaptive Boosting (AdaBoost), detection accuracy is improved. 16 features are employed to construct feature vectors. Because of intermittency of renewable electricity generation, different states for PV and DFIG generation are assumed. Probable events are simulated under different system operating states to generate classification data set. The proposed method is tested on typical distribution system including PV, DFIG and synchronous generator. The studies show that this method succeeds in detection the islanding conditions with negligible detection delay. Detection speed is also increased significantly by using decision tree algorithm. Therefore, the proposed algorithm can be use for either “anti-islanding” or “microgrid” strategies. Results also show that Adaboost algorithm has improved the islanding detection accuracy compared to other data mining- based detection algorithms. This algorithm is capable of detecting islanding phenomenon under operating states with less than 1% power mismatch
  9. Keywords:
  10. Dispersed Generation ; Decision Making Tree ; Microgrid ; Islanding Operation

 Digital Object List

 Bookmark

No TOC