Open AccessSynchrophasor based islanding detection for microgrids using moving window principal component analysis and extended mathematical morphology
Author(s) -
Radhakrishnan Rohikaa Micky,
Sankar Ashok,
Rajan Sunitha
Publication year - 2020
Publication title -
iet renewable power generation
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.005
H-Index - 76
eISSN - 1752-1424
pISSN - 1752-1416
DOI - 10.1049/iet-rpg.2019.1240
Subject(s) - islanding , phasor , principal component analysis , computer science , fault detection and isolation , grid , control theory (sociology) , filter (signal processing) , curse of dimensionality , real time computing , engineering , electric power system , distributed generation , renewable energy , mathematics , power (physics) , artificial intelligence , control (management) , physics , geometry , quantum mechanics , electrical engineering , actuator , computer vision
Extensive deployment of non‐dispatchable renewable energy sources in microgrids (MGs) has led to frequent fault events, DG and line trips and is arduous to discriminate islanding events (IEs) from other transient conditions. For enhanced grid security, time synchronised data from phasor measurement units is utilised. However, data handling at the control centre is tedious as system variables monitored often exhibit non‐linear characteristics that impose a problem to the operator in discriminating between islanding and non‐IEs. To cope with this, a zero non‐detection zone, reliable and precise islanding detection method for data‐intensive grid‐connected MG, is proposed. Voltage phasors, frequency and rate of change of frequency from various locations are processed through moving window principal component analysis (MWPCA) cascaded with extended mathematical morphological filter (EMMF). MWPCA reduces data dimensionality and Q statistics obtained are passed to EMMF, which acts as a non‐linear filter. Further, the islanding detecting factor identified all IEs within the prescribed time limit with minimum false alarms. Accuracy, precision and reliability demonstrated by using a case study model using DIgSILENT are encouraging. It can be adopted by operators to run MG securely. Further, testing of the proposed method on a typical utility feeder shows promising results.