Files in this item

FilesDescriptionFormat

application/pdf

application/pdfLam_Frank.pdf (1MB)
(no description provided)PDF

Description

Title:A tool for reliability analysis of electrical power systems
Author(s):Lam, Frank J.
Advisor(s):Domínguez-García, Alejandro D.
Department / Program:Electrical & Computer Eng
Discipline:Electrical & Computer Engr
Degree Granting Institution:University of Illinois at Urbana-Champaign
Degree:M.S.
Genre:Thesis
Subject(s):Markov reliability models
fault coverage
reliability tool
Abstract:In this thesis, a computer tool for reliability analysis of electrical energy systems is presented. The tool is implemented in MATLAB/Simulink/PLECS and incorporates the concept of fault coverage, which is the probability that, given a fault has occurred, the system remains operational within some acceptable performance requirements. The tool’s computational engine automatically builds a Markov reliability model of the system under analysis from a Simulink/PLECS description of the system augmented to include fault behavior in passive components of the model. The transitions among the model’s Markov states are governed by component failure rates (to be input) and by the fault coverage, which is automatically calculated for each unique fault sequence. With the Markov reliability model constructed and solved, the reliability of the system under analysis is computed. Such a computer tool enables a thorough reliability analysis of a particular design of an electrical system before it is implemented, allowing weak points in the system design to be identified, which helps in redesigning the system for a more robust implementation. The system dynamics is described by a state-space model, where inputs are unknown-but-bounded, which results in the states also being unknown-but-bounded. The set that bounds all possible trajectories is called the reach set. In order to compute the fault coverage for a particular Markov state, the ellipsoid bounding the reach set of the system dynamics associated to the Markov state needs to be computed first. Initial conditions are first selected and all possible maximum and minimum inputs combinations are simulated. Once simulated, an ellipsoid is found that bounds all the trajectories of the simulations. From this bounding ellipsoid, initial conditions are selected on its surface, and simulations are run again for all of the input combinations. Again, the ellipsoid bounding the reach set is found, and this process repeats until the volume of this bounding ellipsoid is no longer increasing. The result is the ellipsoid bounding the reach set of the continuous dynamics associated with the Markov state. During each of the simulations, the trajectories are tracked to ensure that they remain within predefined performance requirements. Trajectories that do not remain within the defined performance requirements are deemed as failed and are not used in computing the bounding ellipsoid. Once all the simulations are completed and the ellipsoid bounding the reach set is found, the coverage can be found by taking the number of simulations that fail, dividing it by the total number of simulations run, and subtracting this quantity from one. Using this method to compute the fault coverage, along with the Markov reliability model construction, a tool is created using these ideas. A case study illustrating the application of the tool to the reliability analysis of a dc distribution system network is presented.
Issue Date:2010-01-06
URI:http://hdl.handle.net/2142/14687
Rights Information:Copyright 2009 Frank J. Lam
Date Available in IDEALS:2010-01-06
Date Deposited:December 2


This item appears in the following Collection(s)

Item Statistics