Extending the synthetic aperture focusing algorithm to deal with flat and curved features in NDT

Abstract

The synthetic aperture focusing (SAF) technique allows exploration of ultrasonic array imaging performance with minimal hardware and it can provide excellent imaging results for point defects, such as fatigue crack tips, commonly found in non-destructive testing (NDT). However, it performs less well when imaging defects which are flat or curved with radii large relative to the ultrasonic wavelength. Here, we report an enhanced time-domain algorithm which can deal with flat and curved defects and which we have termed "synthetic aperture correlation". It is based on processing signals from multiple overlapping subsets of elements from an entire array. The algorithm has roots in SAF, matched filtering, beam steering and biomedical ultrasonic imaging. The data collection process is the same as for SAF and the new algorithm operates entirely in reception, off-line. It has been tested using a 96-element linear array operating at 6 MHz on aluminum test objects. We present simulated and experimental results and compare these with B-scans showing timer-of-flight diffraction (TOFD) characteristics and with conventional SAF. We highlight the advantages of the new algorithm, such as its enhanced capability to deal with noise, and drawbacks such-as increased computational expense.