Distribution System Reconfiguration for Annual Energy Loss Minimization Considering Variable Load and Distributed Generation Profiles

  • Ahmad M. Tahboub

Student thesis: Master's Thesis


Electrical distribution utilities strive to improve their network efficiency in an increasingly challenging deregulated setting dictated by stringent environmental requirements and increasing competition. An electrical power distribution utility operates in a radial network configuration with a number of tie-lines that provide alternate power feeds. These alternate paths have the potential to be utilized by dynamically reconfiguring the network to reduce the system losses in a combinatorial non-convex mixed-integer nonlinear programming (MINLP) problem called distribution system reconfiguration (DSR). The algorithm described in this thesis aims to minimize the annual energy losses as opposed to a snapshot of the system state minimizing power losses. This is achieved using a novel technique employing a fuzzy C-means (FCM) clustering algorithm to recognize actual annual patterns and variability of loads and distributed generation (DG) profiles applied to two wellstudied distribution systems, and employing a genetic algorithm (GA) to give an annually optimum configuration or a set of a predefined number of configurations. The main findings are highlighted and explained. It is shown that as the reconfiguration approach becomes more representative of the system actual states, further loss reduction can be achieved.
Date of AwardJun 2013
Original languageAmerican English
SupervisorHatem Zein El din (Supervisor)


  • Distributed generation (DG); Distribution System Reconfiguration; Energy Losses; Fuzzy C-means Clustering;Genetic Algorithm (GA); Load Variability.

Cite this