چكيده به لاتين
Fault diagnosis problem in wind turbine in the presence of disturbance and uncertainty is surveyed in this thesis in a way that for fault detection problem, disturbance and uncertainty are undesired factors which must be deleted from residual as much as possible for the sake of seeing effect of fault.
Description of general concepts in the fault detection and isolation problem such as residual signal and related issues, wind turbine structure and different existing fault detection methods are done. Focusing on a wind turbine, The types of faults that in a dynamic system are possible to happen are attended then their common methods and the conditions are specified in fault detection. Two methods are used and designed based on conditions and possibility of implement of for wind turbine system.
In this thesis, focusing on two optimal fault detection filter methods is firmed. After designing each of these filters, these methods are implemented on simulation of wind turbine system and threshold computing for fault detection of any type of faults are attended.
Finally, the above methods are compared with different thresholds, the advantages and disadvantages of each of them are surveyed based on 3 criterions: performance index, fault detection time and sensitivity to small faults. preference of each of this methods is specified as well.
Keywords: wind turbine, disturbance and uncertainty rejecting, sensitivity to fault, residual signal, optimal fault detection filter, threshold
