Reconfiguration of distribution system with distributed generation using an adaptive loop approach

Abstract

A new concept of switches selection in the meta-heuristic optimization process of optimal distribution network reconfiguration has been proposed. Based on the adaptive set of selectable candidates, the proposed concept determines the switch status. This approach prevents the creation of unfeasible solutions (non-radial and unconnected configurations), and significantly reducing the number of searches and accelerating the optimization process. Unfeasible solutions, created by meta-heuristic optimization rules, can be corrected by means of the proposed adaptive loop concept. The correct parts of the unfeasible solution are retained, while only the defective parts are replaced by the adaptively formed loops from the currently available conditions that respect the correct switching operations. In this way, the basic characteristics of the optimization process have been retained to the greatest possible extent. Tests were performed on a two different size standard distribution networks.

Fault detection in power grids based on improved supervised machine learning binary classification

“…Consequently, feature extraction methods are crucial to separate these signals and to accurately diagnosis fault signals [2]. Many approaches were proposed to handle feature extraction problem such as transformation [3,4], dimensionality reduction [5], and modal transformation (Clarke transformation) [6,7].…”

“…An approach based on artificial neural network (ANN) and support vector machine (SVM) relies on the gathered measurements at the substation, circuit breaker, and relay states for locating faults in radial distribution systems was presented in [5]. Many methods of fault detection and classification were utilized neural network [7,8]. Zin et al [7] proposed an algorithm that combines back-propagation neural network (BPNN) based on modal Clarke transformation and discrete wavelet transform (DWT) for fault detection in transmission lines.…”

“…Many methods of fault detection and classification were utilized neural network [7,8]. Zin et al [7] proposed an algorithm that combines back-propagation neural network (BPNN) based on modal Clarke transformation and discrete wavelet transform (DWT) for fault detection in transmission lines. A hybrid classification approach based on DWT and long short-term memory network (LSTMN) for detecting insulated overhead conductors fault according to partial discharge [9].…”

Implementation of Distributed Generation with Solar Plants in a 132 kV Grid Station at Layyah Using ETAP

“…In order to exploit the full potential of DG, versatile and competent work force is required to cope with the broad spectrum of technical challenges. Few of the associated issues hindering the robust operation are the requirements of decentralized control [9], optimal placement of plants [3,10], fault location [11], distribution system protection [12], reconfiguration [13], and integration [14].…”