Experimental validation of two elastodynamic models for the wave field generated by ultrasonic transducers.

Abstract

Two different three-dimensional elastodynamic models are introduced to simulate the wave field generated in steel by two types of surface mounted ultrasonic transducers. By replacing the actual transducer by an equivalent surface source distribution, the models become amenable to an exact analytical analysis. The first model simulates the action of a contact transducer through a distribution of nonmoving line segment sources. The second model simulates the action of an angle beam transducer through a single moving line segment source. Almost any transducer aperture shape may be modeled, while the source may apply a nonuniform traction. To speed up the numerical space-time domain calculations, the Cagniard-De Hoop method is employed to analytically evaluate the wave field produced by a single nonmoving line segment source. This solution provides the integrand for both single-integral models. The models are experimentally validated for a contact transducer and three different angle beam transducers. The validation involves a comparison of the wave-field patterns, the directivity curves and some time-domain signals from the wave field. It is shown that the models reliably identify the wide variety of waves generated by ultrasonic transducers, such as focused waves, edge waves, Rayleigh waves and head waves.