Madubuike, K (2022) Fault diagnosis and fault tolerant control for wind turbine dynamic systems. Doctoral thesis, Liverpool John Moores University.

Preview

Text 2017kingsleymadubuikePhD.pdf - Published Version Available under License Creative Commons Attribution Non-commercial. Download (2MB) | Preview

Abstract

The aim of this research is to develop a fault tolerant control (FTC) and fault diagnosis (FD) methodology for nonlinear dynamic systems. This method is applied to the pitch system of a variable speed wind turbine system to verify the effectiveness of this method. The research is divided into three parts.
The first part proposes a robust fault detection approach using an unknown input observer method. This developed observer is sensitive to actuator faults in the benchmark model while it is robust to the system disturbance. A benchmark model consisting of a pitch system, a drive train, generator and converter and the state space model is proposed.
The second part proposes the use of neural network (NN) estimator to detect sensor faults in a wind turbine system. An independent radial basis function neural network (RBFNN) is developed for online diagnosis of the sensor faults. The RBF is trained using sample data during a fault free operating condition. The benchmark model of the wind turbine system proposed in the first part with three sensor faults simulated is developed on Simulink. This research will use two techniques to employ the RBF. One of the RBF will be used to model the wind turbine and generate residuals while the second RBF is developed as a classifier to isolate faults from the generated residuals.
In the final part the reliability of the wind turbine system is guaranteed by designing a variable speed wind turbine pitch angle control which can tolerate and detect faults. The pitch angle system consists of hydraulic pump, hydraulic and pitch gear system. The fault diagnosed here is the shaft rotary friction change which is caused by the break of shift or pitch gear set. The proposed fault tolerant control (FTC) method uses a disturbance observer to diagnose the fault. The FTC is implemented using the combination of a neural network (NN) estimator and a full order terminal sliding mode control. The post fault states can drive to the sliding surface and converge in finite time. The control law is derived using a Lyapunov stability method to ensure we guarantee stability for post-fault. Matlab and Simulink are used to simulate the electrohydraulic servo pitch system with the faults simulated. A third order nonlinear state space is derived, and physical parameters applied in the simulation. The simulation results are presented to validate the effectiveness of the proposed method.

Item Type:	Thesis (Doctoral)
Uncontrolled Keywords:	: fault detection, fault tolerant control, disturbance observer, hydraulic pitch system
Subjects:	T Technology > TA Engineering (General). Civil engineering (General) T Technology > TC Hydraulic engineering. Ocean engineering T Technology > TD Environmental technology. Sanitary engineering
Divisions:	Engineering
SWORD Depositor:	A Symplectic
Date Deposited:	24 Aug 2022 09:03
Last Modified:	24 Aug 2022 09:03
DOI or ID number:	10.24377/LJMU.t.00017240
Supervisors:	Yu, D, Zhang, Q and Gomm, B
URI:	https://researchonline.ljmu.ac.uk/id/eprint/17240

View Item