Doctoral Dissertations

Author

Bipul Luitel

Abstract

"Today, neural networks (NN) are used in several system identification and nonlinear control system applications. However, their implementation on large systems has been limited either because of their degraded performance or the amount of time and computation required for implementing them being impractical for real applications. The primary emphasis of this research is on the development of advanced NN - simultaneous recurrent neural networks (SRN) and cellular neural networks (CNN); and methods for learning the complexity of large systems.

US electricity infrastructure is one of the largest and the most critical infrastructure consisting of thousands of generators connected by transmission lines across the country. Complex interaction of generation, transmission and distribution components in a massively distributed network can result in system instabilities. Proper control action is needed at the right time, place and context in order keep it stable, lack of which can lead to cascaded failures and blackouts consequently causing loss of lives and property. Therefore, it is important to monitor the operation of these components, perform predictive state estimation, make intelligent decisions and carry out fast, automated control actions before a catastrophic event, thus avoiding it or taking measures to mitigate its effects. Development of such a monitoring, decision making and control system is, hence, crucial for smart grids. Development of intelligent monitoring systems using advanced NN for smart grids is addressed in this dissertation.

In order to reduce computational complexity and time required for implementation of advanced NN for smart grid applications, a parallel implementation on high performance computing platform is presented. The above concepts are illustrated for wide area monitoring and bus voltage prediction in smart grids. The results show that advanced NN with the new learning methods can handle the complexity of large systems and arc potentially scalable"--Abstract, page iv.

Advisor(s)

Venayagamoorthy, Ganesh K.

Committee Member(s)

Wunsch, Donald C.
Corzine, Keith, 1968-
Shi, Yiyu
Dagli, Cihan H., 1949-

Department(s)

Electrical and Computer Engineering

Degree Name

Ph. D. in Computer Engineering

Sponsor(s)

Real-Time Power and Intelligent Systems Laboratory
National Science Foundation (U.S.)

Publisher

Missouri University of Science and Technology

Publication Date

2012

Journal article titles appearing in thesis/dissertation

  • Quantum inspired PSO for the optimization of simultaneous recurrent neural networks as mimo learning systems
  • Decentralized asynchronous learning in cellular neural networks
  • A scalable situational awareness system for smart grids

Pagination

xii, 86 pages

Note about bibliography

Includes bibliographical references.

Rights

© 2012 Bipul Luitel, All rights reserved.

Document Type

Dissertation - Restricted Access

File Type

text

Language

English

Library of Congress Subject Headings

Smart power grids -- Design
Smart power grids -- Computer networks
Intelligent control systems
Neural networks (Computer science)
Swarm intelligence

Thesis Number

T 10269

Print OCLC #

870999173

Electronic OCLC #

919436653

Link to Catalog Record

Electronic access to the full-text of this document is restricted to Missouri S&T users. Otherwise, request this publication directly from Missouri S&T Library or contact your local library.

http://laurel.lso.missouri.edu:80/record=b10251258~S5

Comments

Financial support for research supplied by grants NSF EFRI #0836017 and NSF CAREER ECCS #0348221

Share

 
COinS