Dissertation/Thesis Abstract

A Fuzzy Logic-Based Fault Tolerant Control Approach for Wind Turbines
by Kanumalla, Neha, M.S., University of Louisiana at Lafayette, 2015, 99; 10002411
Abstract (Summary)

This thesis presents a FDI method using support vector machine (SVM) approach, which is preferred among other statistical methods like neural networks and principal component analysis, as in this approach the fault detection and isolation is done in less time. A controller is designed for fault tolerance using fuzzy logic. Support vector machine and Fuzzy controller are robust data based approach to process knowledge. The Fuzzy controller does not require exact model. The Fuzzy controller predicts the reference pitch angle and the power generation is lowered at high wind speeds to keep the system safe, after the fault has occurred pitch angle changes by which generator speed reaches to its nominal value by which the fault tolerance occurs. The fuzzy weight outputs are used to detect the faults at the earliest and accommodate the faults so the wind turbine operates in the specified optimal region. As the fuzzy controller and SVM is used the faults are detected in less time and are isolated by which the system stays safe. The variance corresponds to the matrix x is maintained in a way such that the false alarms are avoided. The novelty of this approach is that the faults are being detected earlier and false alarms are prevented by choosing a correct value of variance.

Indexing (document details)
Advisor: Afef, Fekih
Commitee: Mohammad, Madani, Zhongqi, Pan
School: University of Louisiana at Lafayette
Department: Electrical Engineering
School Location: United States -- Louisiana
Source: MAI 55/03M(E), Masters Abstracts International
Subjects: Electrical engineering
Keywords: Fault detection, Fault isolation, Fault tolerant control, Fuzzy logic, Support vector machines, Wind turbine
Publication Number: 10002411
ISBN: 9781339419213
Copyright © 2019 ProQuest LLC. All rights reserved. Terms and Conditions Privacy Policy Cookie Policy