Article révisé par les pairs
Résumé : Surface crack sizing in metals by time-of-flight ultrasonic techniques is already well known. One of these techniques makes use of surface Rayleigh waves, and it correlates the time taken by this wave to contour the crack length. However, some questions concerned with the influence of some factors on the accuracy of this technique remain open. The present study focuses on one of them. It is the compressive stress state in the crack tip neighbourhood which gives origin to crack closure, hindering the Rayleigh waves contouring the crack, and causing undersizing. Moreover, the crack sizing suitability using not only the Rayleigh waves but also a shear wave generated at the crack tip by mode conversion are also studied. The material chosen was a C-Mn steel plate (classification ASTMA-283, gr.C) of 38 mm thickness. The ultrasonic waves were generated by commercial surface wave transducers of 4 MHz. Fracture mechanics test-pieces were machined having fatigue cracks normal to their surface of 2.5 to 11.0 mm length. Crack depth sizing was made using the Rayleigh wave that contours the crack path, and its shear wave mode conversion generated at the crack tip. This latter wave travels through the uncracked ligament of the test-piece, and comes back to the crack tip, where it reverts to a surface Rayleigh wave. The results showed that: □ a tensile loading is necessary to bring crack sizing accuracy to an acceptable level; □ nevertheless this loading does not introduce any risk of failure; □ crack sizing is suitably conducted in the elastic portion of the fracture mechanics test □ the Rayleigh wave time-of-flight technique shows a clear tendency to oversizing. The same cannot be said about the mode conversion shear wave.