Fault Location Estimation Using Ensemble Averaging Decomposition and Hybrid Meta-Heuristic Optimized Kernel-Based ELM Technique for DG-Integrated Microgrid

Abstract

This paper presents a novel approach for fault location estimation in Distributed Generation (DG) based microgrid that combines a decomposition technique for feature extraction and a machine learning-based method for fault location computation. A hybrid meta-heuristic optimized-based KELM (HMOKELM) framework is implemented to improve the efficiency of the proposed scheme. Here, an energy-based strategy is used to extract the differential energy (DE) content of the three-phase signal through the ensemble averaging decomposition method. Then, a feature matrix is created using different features of the DE signal which is fed as input to the HMOKELM for fault location estimation. However, the proposed approach is tested in IEC-based standard microgrid model under MATLAB /SIMULINK environment. Further, different fault conditions, microgrid structures, and modes of operation are considered as a case study. Thereafter, the results are analyzed and extensively compared with other techniques to demonstrate the effectiveness and reliability of the proposed strategy.