Inverse wave field extrapolation: a different NDI approach to imaging defects

Nondestructive inspection (NDI) based on ultrasound is widely used. A relatively recent development for industrial applications is the use of ultrasonic array technology. Here, ultrasonic beams generated by array transducers are controlled by a computer. This makes the use of arrays more flexible than conventional single-element transducers. However, the inspection techniques have principally remained unchanged. As a consequence, the properties of these techniques, as far as characterization and sizing are concerned, have not improved. For further improvement, in this paper we apply imaging theory developed for seismic exploration of oil and gas fields on the NDI application. Synthetic data obtained from finite difference simulations is used to illustrate the principle of imaging. Measured data is obtained with a 64-element linear array (4 MHz) on a 20-mm thick steel block with a bore hole to illustrate the imaging approach. Furthermore, three examples of real data are presented, representing a lack of fusion defect, a surface breaking crack, and porosity

I. Introduction

Ultrasonic techniques have been used successfully for many nondestructive inspection (NDI) applications. For girth weld inspection, ultrasonic techniques have proven to increase the probability of defect detection when used in combination with radiography. In some cases, ultrasonic techniques even can be used as a replacement for radiography [1]–[3]. Once a defect is detected, it must be evaluated in order to reject or accept the weld quality. The evaluation usually is carried out according to acceptance criteria as defined in codes. The evaluation requirements are different for most codes. Some codes require an evaluation based on the amplitude height (NEN 3650 [4], DNV [5]); other codes only prescribe rejection criteria based on maximum defect length and height (API 1104 [6], BS 4515 [7]). Evaluation based on amplitudes can be applied directly, because the result of an ultrasonic measurement contains the amplitude information. In case the criteria are based on defect height and length, the information must be extracted from the amplitudes and/or temporal information of the ultrasonic signals.