: In recent years, the use of renewable energy instead of fossil fuels has seen significant growth in developed and developing countries. One of the sources of renewable energy that has been widely used is wind energy. Wind turbines are complex electromechanical systems that convert kinetic energy of the wind into electrical energy. Various parts of wind turbines are subjected to different types of faults. Subsystems of wind turbines are extensively interconnected. As a result, occurrence of a fault in one subsystem can be spread into other subsystems. Therefore, to prevent and reduce the costs caused by a fault in the system, we need a mechanism that detects the fault in a timely manner and takes corrective actions to reduce their effects. In this thesis, first a model of wind turbines, wind turbine common faults and fault detection and identification techniques are reviewed. Then an online method to detect faults of rotor speed, generator speed and torque sensors is proposed. The proposed method is based on a bank of unknown input observers. Unknown input observer is a method that produces residuals with emphasis on rejecting disturbances and other unknown inputs. Also, a fault tolerant mechanism is also proposed that eliminates the effects caused by the considered faults. In this technique, after identifying the time and type of a fault, the controller parameters are modified to compensate for the fault impacts. Efficiency of the proposed fault identification and tolerance methods are demonstrated through various simulations. Keywords: Fault Detection, Wind Turbine Control, Fault Accomodation, unknown Input Observers
