A combined mathematical morphology and extreme learning machine techniques based approach to micro-grid protection

This work introduces a smart differential protection scheme for a microgrid system using a nonlinear signal transformation named as ‘mathematical morphology (MM)’. Here, the mathematical morphology is used as a feature extraction technique. Thus, the three elementary MM filtering operators like eros...

Full description

Bibliographic Details
Main Authors: Manohar Mishra, Rasmi Ranjan Panigrahi, Pravat Kumar Rout
Format: Article
Language:English
Published: Elsevier 2019-06-01
Series:Ain Shams Engineering Journal
Online Access:http://www.sciencedirect.com/science/article/pii/S2090447919300656
Description
Summary:This work introduces a smart differential protection scheme for a microgrid system using a nonlinear signal transformation named as ‘mathematical morphology (MM)’. Here, the mathematical morphology is used as a feature extraction technique. Thus, the three elementary MM filtering operators like erosion, dilation, and opening-closing-difference-filter (OCDF) are used to operate on the extracted phasor current signals and its symmetrical components for the extraction of differential feature vector. Further, the extracted feature vector is fed as an input to train and test two distinct extreme machine learning (ELM) classifiers meant for primary and backup protection. To justify and verify the better performance of the proposed method numerous fault and no-fault conditions are simulated by considering several operating conditions, such as topology of microgrid (radial/mesh) and mode of microgrid operation (islanding/grid-connecting). The experimental outcomes confirm the efficiency and reliability of the offered microgrid protection scheme (MPS) in diverse operating condition. Keywords: Distributed generation, Extreme learning machine, Microgrid, Mathematical morphology
ISSN:2090-4479