High fidelity ultrasonic measurements for defect detection and characterisation

Abstract

Repeatable measurements with detachable transducers such as phased arrays have much potential for enhancing the sensitivity of nondestructive evaluation (NDE), as baseline subtraction can be employed. This paper aims to develop post-processing methods to overcome the variation between inspections which leads to poor subtraction performance. In this paper, ultrasonic immersion inspections are performed on a brass sample and a steel sample with a corroded backwall. Reference measurements are taken, and then small defects are machined to simulate defect growth in industrial applications. Defect types include two 1 mm diameter flat bottom holes (FBH) extending from 15 mm to 17 mm in the brass sample and two 2 mm diameter holes drilled to 1 mm depth from the corroded surface of the steel sample to represent the growth of localised pits. Although FMC/TFM is used in this paper, the defects have similar response amplitudes to the structural noise, and they cannot be detected or characterised directly from the images by the conventional 6 dB drop method. This paper demonstrates the utilisation of ultrasonic measurements, together with cross-correlation, to obtain inspection variables which include array position and couplant wave velocity. By knowing the variations between inspection conditions, compensation is made in post-processing to improve image repeatability, and hence the performance of baseline subtraction. The SNR is improved by approximately 10–25 dB depending on how great the actual variations are. It is shown that these previously undetectable defects can be seen and accurately sized after conducting baseline subtraction with compensation.