Expected scattering of elastic waves by rough defects for applications in ultrasonic NDT
S Haslinger, M Lowe, P Huthwaite and R Craster
Ultrasonic NDE techniques, such as time-of-flight diffraction (TOFD) and pulse-echo (PE), have been developed using modelling and experiments and are well understood when cracks are smooth and straight. However, in environments where there are extreme changes in temperature and pressure, for instance in the nuclear industry, the damage that occurs is often not uniform. Rough cracks develop and these are much more challenging to characterise.
The qualification of ultrasonic inspections for defects that are expected to be rough (typically thermal fatigue or stress corrosion cracking) is presently very conservative, owing to the uncertainty of the amplitudes of rough surface reflections. No two rough surfaces are the same due to their inherent randomness and so scattering signatures differ from one surface to the next. The pragmatic solution is to apply a large safety threshold on the expected reflection amplitude, but this can lead to unnecessary shutdowns and early retirement of components.
An alternative approach has been developed by the authors, whereby the expected reflection from a rough defect can be predicted using a statistical model. Although every rough defect is randomly determined, it is possible to anticipate statistical metrics of a rough surface using extensive industrial databases of in-service defects. Given only the angle of incidence and two statistical parameter values used to characterise the defects, the expected reflection amplitude is obtained instantaneously for any scattering angle and length of defect. The method has been investigated for the scattering of both incident longitudinal and incident shear waves and includes the subsequent mode conversion. Validations using numerical methods (finite element and ray models) and experiments are provided.
The qualification of ultrasonic inspections for defects that are expected to be rough (typically thermal fatigue or stress corrosion cracking) is presently very conservative, owing to the uncertainty of the amplitudes of rough surface reflections. No two rough surfaces are the same due to their inherent randomness and so scattering signatures differ from one surface to the next. The pragmatic solution is to apply a large safety threshold on the expected reflection amplitude, but this can lead to unnecessary shutdowns and early retirement of components.
An alternative approach has been developed by the authors, whereby the expected reflection from a rough defect can be predicted using a statistical model. Although every rough defect is randomly determined, it is possible to anticipate statistical metrics of a rough surface using extensive industrial databases of in-service defects. Given only the angle of incidence and two statistical parameter values used to characterise the defects, the expected reflection amplitude is obtained instantaneously for any scattering angle and length of defect. The method has been investigated for the scattering of both incident longitudinal and incident shear waves and includes the subsequent mode conversion. Validations using numerical methods (finite element and ray models) and experiments are provided.
