Estimação de parâmetros em ensaios não-destrutivos por ultrassom: diretividade e rugosidade

Abstract

Ultrasonic methods for non-destructive testing are widely used for object inspection due to advantages such as the ability to detect and characterize defects in the internal structure of solid objects, low cost, and safety of operation. However, the performance of those inspections depends on prior knowledge about the parameters of the ultrasound system, the test environment, and the inspected object. The acoustic beam directivity in ultrasonic transducers and the surface roughness of the objects under test are difficult parameters to be measured. Normally, the directivity function is obtained through a specific test that takes measurements of the sound pressure field around the transducer using a hydrophone. While roughness measurements, in most cases, are performed through contact and optical methods, which have limitations regarding the access to the surface to be inspected and the inspection environment. This study presents a straightforward method to determine the directivity of phased array transducers. The method consists of an immersion test using only the transducer under test and a small reflector. The Ascans signals, acquired in pulse-echo mode, are processed to obtain the parameters that generate the experimental directivity function. This study also proposes two methods to estimate the roughness. The first method is based on several theoretical attenuation models and proposes a joint nonlinear regression to estimate the roughness of external and internal surfaces through ultrasound tests. The second method is based only on Phase-Screen Approximation and aims to estimate the surface roughness in submerged tests, using a smooth surface as a reference. The proposed method for estimating directivity was applied to simulated and experimental data. The results were compared to theoretical models. Two observations were highlighted: the arising of undefined lateral lobes for both simulated and experimental functions and the narrowing of the main lobe of the experimental directivity function. Reconstructed images by Corrected Total Focusing Method algorithm resulted in a better performance when the experimental directivity function was used instead of the theoretical one. Specifically, artifacts were reduced and the amplitudes of identical reflectors in different positions were more uniform throughout the reconstructed image. The first method to estimate roughness was partially evaluated using simulated signals. However, the divergences found between the theoretical models and the simulated data made unfeasible both the tests of the complete model with simulated signals and the performance of experiments. The second method allowed the adequate estimation of roughness in simulated signals and the application of the method in experimental signals.