No abstract

The structural damage in the main shaft threatens the service safety of the wind turbine directly. The mechanical properties of the in-service main shaft tend to degrade severely due to the transverse cracks extending from the outer circumference. In the health monitoring of the shaft, it is of great importance not only to identify the cracks, but also to accurately evaluate the crack size. In this research, a novel quantitative method is proposed for the transverse crack characterization. The diffracted waves from the crack tip received in the central hole of the shaft are used to construct elliptic trajectories of beam path. The crack position and its depth can be determined by the intersection of the trajectories. Furthermore, a crack size quantification model is established, and the numerical demonstration is given for evaluation of crack position and depth. Meanwhile, an acoustic finite element model is developed, in which the path of the acoustic waves radiated from the end face of the shaft is analyzed in detail. The quantitative method is well confirmed by the simulation of cracks at different locations with different depths. Considering the accuracy of crack evaluation from a large size zone, the time of diffracted echo is calibrated by measuring the time shift of the transducers. A shaft sample with a transverse crack was used to implement the experimental verification. Totally, the sizing error is less than 5 mm, which indicates that this proposed method is effective for the evaluation of surface cracks in the main shaft.

For an operating wind turbine, the main shaft suffers severe structural damage. The surface breaking transverse crack threatens the safety of drive-train system directly. The crack examination method is available by diffracted waves in the central hole. In this research, an electromagnetic acoustic transducer (EMAT) for receiving diffracted longitudinal waves is proposed. Firstly, the converse piezomagnetic stress coefficients and magnetostriction coefficient were derived from the measured magnetic characteristic curve. The measured parameters were provided for sensor simulation. Considering the two coupling mechanisms, a receiving EMAT model was established for coil optimization. Secondly, the influence of materials, lift-off distance and cable type on sensor impedance would be analyzed in detail. Multiple factors were taken into account for impedance matching, and the signal amplitude was improved significantly. The angle divergence of the EMAT was also measured, showing the main lobe with 24 • . The crack detection experiments were carried out on a shaft sample. The results show that the developed EMAT can achieve the reception of diffracted longitudinal waves, and the signals are distinguishable. The crack was evaluated simultaneously by the received diffracted waves, and the errors showed less than 3 mm.