Weighted Multi-Element Synthetic Aperture Focusing Technique Algorithm of Ultrasonic Non-Destructive Testing on Machinery

Abstract

key components of large industrial equipment for manufacturing need regular maintenance. The health state of these components are evaluated during maintenance, then future maintenance plan is continuously updated according to the latest health state. However, the health state of some components is difficult to evaluate since internal flaws are invisible and cannot be detected with passive inspection methods. Ultrasonic testing (UT) is an active non-destructive testing(NDT) method to inspect internal defects. But the use of UT in maintenance for machinery in manufacturing industry is limited due to the limited quality of result, i.e., the depth of inspection is limited, the resolution is not sufficient, and the complicated defects are difficult to interpret from 2D images. To overcome these problems, a multi-element synthetic aperture focusing technique (SAFT) algorithm with filtering and a novel weighting factor with linear array is studied. It can improve the quality of conventional ultrasonic non-destructive testing (UNDT) and prepare for future 3D model reconstruction. In the proposed algorithm, filters are implemented to denoise the acquired images, weighting factor can make the acquired image more balanced comparing to conventional algorithm, which simply sums received signal from all receiving elements at each pixel. The proposed algorithm is simulated using FIELD II in Matlab software and compared with single-element, multi-element SAFT and total focusing method (TFM) algorithms. An aluminium specimen is used as the simulation object. The results show that the proposed algorithm has clearer boundaries of defects, better contrast resolution (CR), contrast-to-noise ratio (CNR) and better overall result of the acquired image. The computational complexity of the proposed algorithm is higher than single element SAFT but lower than TFM. In the future study, 3D reconstruction of defective components can be implemented based on the results of this study.