Reconfiguration of power networks based on graph-theoretic algorithms

Abstract

My foremost and tremendous appreciation goes to my major professor, Dr. Chen-Ching Liu, who provided me with his continuous support, inspirations, challenges and encouragement during my research work. This dissertation work could not have been completed without his insightful and invaluable technical advice. The precious experience of working with him as a teaching assistant and as a research assistant on a variety of projects will have a profound influence on my future career development. I would also like to express my sincere thanks to Prof. Venkataramana Ajjarapu, Prof. Manimaran Govindalasu, Prof. Dionysios Aliprantis and Prof. Dan Nordman for serving on my committee. Their valuable feedback and suggestions were essential in improving the dissertation. I am very grateful to my parents Zhen-Ju Zheng and Quan-Xi Li, my beloved husband, Nanpeng Yu, for their endless support and love as always.

Optimisation of electrical network configuration: Complexity and algorithms for ring topologies

“…Firstly, given a required load scenario, dimensioning the network infrastructure to reach flow and reliability constraints [3,6,18], with objective to minimize the cost of investment of electric lines and switches. Secondly, the reconfiguration of networks in case of failures, where two objectives are usually to be optimised [8,12,15], namely to maximize the restore load and to minimize the size of the sequence of operations to be done from the configuration before fails, mainly the switches activation, to reach the decided re-configuration ( these two objectives can be considered simultaneously [4,7]). Finally, planning schedules [2,9,14] consisting in a schedule of consecutive configurations for different consecutive scenarios of load requirements.…”

“…In such planning schedules, vulnerability of configurations can be taken into account by considering various graph theory metrics [1]. Note that in this context, balanced planning schedules can also be provided by configuring the network in balanced subnetworks [21,9,12] or by considering balanced configurations in terms of power of loads [20].…”

New Multi Objective Approach for Optimal Network Reconfiguration in Electrical Distribution Systems Using Modified Ant Colony Algorithm

“…The reconfiguration of distribution networks for loss minimization has been proposed for the first time by Merlin and Back in [5] where a heuristic optimization approach has been used to determine an optimal configuration minimizing losses in a tree network with a specific charge. Since then, several other methods have been developed and oriented towards programming based on the C-language, genetic algorithms [6], evolutionary programming [7] which proposed a reconfiguration of distribution networks by the heuristic permutation method compared to the TABOU method. They found that the method of branch switching makes satisfactory results over the other.…”

Extended fast decoupled power flow for reconfiguration networks in distribution systems

“…Automated reconfiguration is a very relevant research theme since it deals with the most significant issues regarding DS: emergency service restoration after faults, power loss optimisation, and renewable sources integration. Since this is a complex problem that depends on the switching combinations of the control area, most of the methodologies applied to deal with network reconfiguration are based on the application of heuristic search, optimisation, or rule‐based algorithms [2–8]. In addition, since causes and consequences of reconfiguration manoeuvres are inevitably linked with steady‐state aspects of the network, such as voltage levels and power losses, power flow calculation is usually an imperative requisite in the reconfiguration process.…”