Open AccessArtificial neural network and phasor data‐based islanding detection in smart grid
Author(s) -
Kumar Dhruba,
Bhowmik Partha Sarathee
Publication year - 2018
Publication title -
iet generation, transmission and distribution
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.92
H-Index - 110
eISSN - 1751-8695
pISSN - 1751-8687
DOI - 10.1049/iet-gtd.2018.6299
Subject(s) - phasor , islanding , artificial neural network , computer science , pattern recognition (psychology) , classifier (uml) , artificial intelligence , dimensionality reduction , electric power system , real time computing , data mining , power (physics) , physics , quantum mechanics
Islanding is an unusual condition in a power system where the generating station continues to supply the local load after one or multiple transmission line outage. This study develops a new islanding detection technique using the artificial neural network (ANN) classifier, which is provided with synchronised phasor measurements from a nine‐bus Western Electricity Coordinating Council power system. An excessive number of data frames are generated in the phasor data concentrator. Before sending these data to the classifier, multiplier‐based method (MBM) and Andrews plot‐based method (APBM) are applied for dimension reduction and feature extraction. Comparisons are prepared with other dimension reduction algorithms. The accuracy of the classifier has been increased by increasing the number of hidden layers, the best accuracy is observed at a certain level for APBM. Non‐detection zone (NDZ) for APBM is also evaluated. It is observed that the classification accuracy, and the detection time change when the neural network is retrained. All the results are compared and analysed statistically. This method can perform faster compared to other existing algorithms with an excellent accuracy and smaller NDZ.