A Sparse Recovery Diagnosis Algorithm Applied on PMSG Wind Turbine

Many researchers have paid great attention to fault diagnosis technologies in order to reduce worst operations and maintenance costs. Accordingly, this work presents a new method to detect, isolate and estimate faults occurring in permanent magnet synchronous generator (PMSG). This method is based on a dynamical sparse recovery algorithm that is able to reconstruct online and with finite time convergence a sparse vector of numerous faults from few system measurements. The faulty system is modeled in case of permanent magnet demagnetization and voltage dip faults and its dynamical modelisation is adapted to the proposed algorithm under some theoretical conditions. The system model with the proposed algorithm is implemented under Matlab/Simulink environment and its effectiveness is verified upon simulation results for such fault scenarios.