Use this URL to cite or link to this record in EThOS:
Title: Robust fault diagnosis by GA optimisation with applications to wind turbine systems and induction motors
Author: Odofin, Sarah
ISNI:       0000 0004 7430 0305
Awarding Body: Northumbria University
Current Institution: Northumbria University
Date of Award: 2016
Availability of Full Text:
Access from EThOS:
Full text unavailable from EThOS. Please try the link below.
Access from Institution:
This investigation focuses and analyses the theoretical and practical performance of a dynamic system, which affords condition monitoring and robust fault diagnosis. The importance of robustness in fault diagnosis is becoming significant for controlled dynamic systems in order to improve operating reliability, critical-safety and reducing the cost often caused by interruption shut down and component repairing. There is a strong motivation to develop an effective real-time monitoring and fault diagnosis strategy so as to ensure a timely response by supervisory personnel to false alarms and damage control due to faults/malfunctions. Environmental disturbances/noises are unavoidable in practical engineering systems, the effects of which usually reduce the diagnostic ability of conventional fault diagnosis algorithms, and even cause false alarms. As a result, robust fault diagnosis is vital for practical application in control systems, which aims to maximize the fault detectability and minimize the effects of environment disturbances/noises. In this study, a genetic algorithm (GA) optimization model-based fault diagnosis algorithm is investigated for applications in wind turbine energy systems and induction motors through concerns for typical types of developing (incipient) and sudden (abrupt) faults. A robust fault detection approach is utilized by seeking an optimal observer gain when GA optimisation problems become solvable so that the residual is sensitive to the faults, but robust against environmental disturbances/noises. Also, robust fault estimation techniques are proposed by integrating augmented observer and GA optimisation techniques so that the estimation error dynamics have a good robustness against environmental disturbances/noises. The two case studies investigated in this project are: a 5MW wind turbine model where robust fault detection and robust fault estimation are discussed with details; and a 2kW induction motor experimental setup is investigated, where robust fault detection and robust fault estimation are both examined, and modelling errors are effectively attenuated by using the proposed algorithms. The simulations and experimental results have demonstrated the effectiveness of the proposed fault diagnosis methods.
Supervisor: Gao, Zhiwei ; Sun, Kai Sponsor: Not available
Qualification Name: Thesis (Ph.D.) Qualification Level: Doctoral
EThOS ID:  DOI: Not available
Keywords: H900 Others in Engineering