This thesis examines the problem of load flow feasibility, in other words, the conditions under which a power network characterized by the load flow equations has a steady-state solution. In this thesis, we are particularly interested in load flow feasibility in the presence of extreme contingencies such as the outage of several transmission lines. / Denoting the load flow equations by z = f(x) where z is the vector of specified injections (the real and reactive bus demands, the specified real power bus generations and the specified bus voltage levels), the question addressed is whether there exists a real solution x to z = f( x) where x is the vector of unknown bus voltage magnitudes at load buses and unknown bus voltage phase angles at all buses but the reference bus. Attacking this problem via conventional load flow algorithms has a major drawback, principally the fact that such algorithms do not converge when the load flow injections z define or are close to defining an infeasible load flow. In such cases, lack of convergence may be due to load flow infeasibility or simply to the ill-conditioning of the load flow Jacobian matrix. / This thesis therefore makes use of the method of supporting hyperplanes to characterize the load flow feasibility region, defined as the set the injections z for which there exists a real solution x to the load flow equations. Supporting hyperplanes allow us to calculate the so-called load flow feasibility margin, which determines whether a given injection is feasible or not as well as measuring how close the injection is to the feasibility boundary. This requires solving a generalized eigenvalue problem and a corresponding optimization for the closest feasible boundary point to the given injection. / The effect of extreme network contingencies on the feasibility of a given injection is examined for two main cases: those contingencies that affect the feasibility region such as line outages and those that change the given injection itself such as an increase in VAR demand or the loss of a generator. The results show that the hyperplane method is a powerful tool for analyzing the effect of extreme contingencies on the feasibility of a power network.
| Identifer | oai:union.ndltd.org:LACETR/oai:collectionscanada.gc.ca:QMM.100252 |
| Date | January 2007 |
| Creators | Khosravi-Dehkordi, Iman. |
| Publisher | McGill University |
| Source Sets | Library and Archives Canada ETDs Repository / Centre d'archives des thèses électroniques de Bibliothèque et Archives Canada |
| Language | English |
| Detected Language | English |
| Type | Electronic Thesis or Dissertation |
| Format | application/pdf |
| Coverage | Master of Engineering (Department of Electrical and Computer Engineering.) |
| Rights | © Iman Khosravi-Dehkordi, 2007 |
| Relation | alephsysno: 002769824, proquestno: AAIMR51463, Theses scanned by UMI/ProQuest. |
Page generated in 0.0018 seconds