With the demand for energy, the marine current turbine (MCT) has been further developed. Since the blades are exposed to the sea water for a long time, they are prone to be affected by dirt and turbidity, and such imbalance fault features are difficult to be extracted owing to the frequent fluctuation of the current speed. Based on this, this paper proposes a method of Park's vector transformation and signal reconstruction. Firstly, the three-phase stator currents of a marine current generator are projected into the two-phase static coordinate system to obtain the square of modulus signal, which is filtered by wavelet transform and decomposed by empirical mode decomposition (EEMD) to obtain some IMF components, and then the signal reconstruction of some IMF components is carried out. By wavelet time-analysis, MCT blade state can be detected. The method proposed in this paper successfully detects the blade imbalanced fault.
|