Download the full-sized PDF of Adaptive Approaches for Fault Detection and Diagnosis With ApplicationsDownload the full-sized PDF



Permanent link (DOI):


Export to: EndNote  |  Zotero  |  Mendeley


This file is in the following communities:

Graduate Studies and Research, Faculty of


This file is in the following collections:

Theses and Dissertations

Adaptive Approaches for Fault Detection and Diagnosis With Applications Open Access


Other title
fault-tolerant control
fault estimation
fault detection
Type of item
Degree grantor
University of Alberta
Author or creator
Li, Gang
Supervisor and department
Zhao, Qing (Electrical and Computer Engineering)
Examining committee member and department
Chen, Tongwen (Electrical and Computer Engineering)
Zhao, Qing (Electrical and Computer Engineering)
Tavakoli, Mahdi (Electrical and Computer Engineering)
Shi, Yang (Mechanical Engineering, University of Victoria)
Lipsett, Michael (Mechanical Engineering)
Department of Electrical and Computer Engineering
Control systems
Date accepted
Graduation date
2017-06:Spring 2017
Doctor of Philosophy
Degree level
Adaptive approaches for model-based fault detection, diagnosis and control are the focus of this research. Both analytical and data-driven model-based techniques are studied along with the investigation of their practical applications. This thesis is organized in two parts. In Part I, adaptive fault estimation (FE) and fault tolerant shape control (FTSC) designs are proposed for stochastic distribution systems. First, an adaptive FE and dynamic output feedback FTSC scheme is designed for a class of nonlinear non-Gaussian stochastic systems subject to time-varying loss of control effectiveness faults, where time-varying faults, Lipschitz nonlinear property and general stochastic characteristics are taken into consideration simultaneously. Then, the FE and FTSC schemes are extended to particle size distribution (PSD) processes subject to simultaneous time-varying actuator and sensor faults. The proposed adaptive FE and FTSC schemes are validated in both simulation and application examples. In Part II, adaptive data-driven model-based fault detection and diagnosis (FDD) methods are proposed for rotating machinery. First, a novel sinusoidal synthesis based adaptive tracking (SSBAT) technique is proposed based on a data-driven rotating machinery model identified using time-series vibration data and incorporating physical constraints of rotating equipment; the SSBAT model is then used as an adaptive predictor to generate residual for FDD. Then, a minimum entropy deconvolution based sinusoidal synthesis (MEDSS) model is proposed to improve the fault diagnosis performance of the SSBAT scheme; a time-weighted-error Kalman filter is designed to estimate the MEDSS model parameters adaptively. Both methods are tested and proved to be effective through simulation examples and a practical case study for rubbing fault diagnosis in an industrial steam turbine.
This thesis is made available by the University of Alberta Libraries with permission of the copyright owner solely for the purpose of private, scholarly or scientific research. This thesis, or any portion thereof, may not otherwise be copied or reproduced without the written consent of the copyright owner, except to the extent permitted by Canadian copyright law.
Citation for previous publication
G. Li, G. L. McDonald, and Q. Zhao, “Sinusoidal synthesis based adaptive tracking for rotating machinery fault detection,” Mechanical Systems and Signal Processing, vol. 83, pp. 356–370, 2017.G. Li and Q. Zhao, “Minimum entropy deconvolution optimized sinusoidal synthesis and its application to vibration based fault detection,” Journal of Sound and Vibration, vol. 390, pp. 218–231, 2017.

File Details

Date Uploaded
Date Modified
Audit Status
Audits have not yet been run on this file.
File format: pdf (PDF/A)
Mime type: application/pdf
File size: 1515219
Last modified: 2017:06:13 12:15:17-06:00
Filename: Li_Gang_201701_PhD.pdf
Original checksum: 92e15dff63ff7c14ecaa9e2ab25a8556
Activity of users you follow
User Activity Date