چكيده به لاتين
Since reaction wheel is one of the important actuators in satellite attitude control subsystem, the issues of fault diagnosis and fault tolerance are very essential in this element. Reaction wheel’s structure consists of two parts: electromechanical part and electronic board part. The purpose of this research is fault detection and isolation of electromechanical part including rotor, stator and bearing. Fault diagnosis is carried out using signal based methods. To reach this goal, at first, accurate model of reaction wheel is developed. Then by applying the fault and receiving the stator current of simulated model, fault diagnosis is done using signal based methods. For rotor fault diagnosis short time Fourier transform (STFT) has been used. This method distinguishes the fault frequencies that occur in stator current effectively. For detecting stator inter-turn fault, fast Fourier transform (FFT), continuous wavelet transform (CWT) and discrete wavelet transform (DWT) methods have been employed. Moreover, for bearing faults, vibration data due to these faults have been analysed using discrete wavelet transform method. The faults that introduced cause to effects on frequencies of current and vibration signals. By applying the signal based methods, frequency harmonics of faults in received signals is revealed. By comparing these frequency harmonics in faulty and healthy state, fault is diagnosed.
