Ultrasonic array design optimisation for defect characterisation

Speaker:
Safari, Ali; University of Bristol; United Kingdom

Authors:
Safari, A.; University of Bristol; United Kingdom
Bai, L.; University of Bristol; United Kingdom
Zhang, J.; University of Bristol; United Kingdom
Velichko, A.; University of Bristol; United Kingdom
Drinkwater, B.; University of Bristol; United Kingdom

ID: ECNDT-0230-2018
Download: PDF
Session: Ultrasonic Inspection 2
Room: G2
Date: 2018-06-13
Time: 11:50 - 12:10

Ultrasonic arrays are now increasingly used in many industrial inspections. Alongside this development, recent research has provided improved array image quality through Full Matrix Capture (FMC) and the use of a wide range of post-processing algorithms. Arrays have the capability to not only detect and locate defects, but also to provide information about the characteristics of the defect. For a crack, this information is size and orientation angle. But it is also important to be able to distinguish between cracks and other defects, such as volumetric voids and inclusions, which are often much less significant from a structural integrity perspective. An array insonifies a defect from a range of angles and thereby measures part of the matrix of defect scattering matrix. In this way, the scattering matrix encodes the defect characterisation information. Typically, this detail is lost when the FMC data is transformed into an image. However, here we extract this scattering information and use it to characterise small defects. Firstly, we show that under certain conditions, it is possible to determine the defect type, distinguishing between cracks and volumetric defects. Secondly, once the defect type is determined, we show that it is possible to accurately extract parameters such as size, orientation and aspect ratio. We show that this characterisation information is inherently probabilistic and introduce defect probability maps which reveal the most probable defect and the probability landscape. Finally, we show how such knowledge can lead to the design of new arrays, optimised specifically for characterisation. In this step, it is assumed that the defect has been detected and the requirement is now purely to determine its characteristics to the highest possible accuracy. We show that improved characterisation accuracy can be achieved with this optimised approach and suggest that this concept will have benefits for some safety critical applications.