Imaging of defects in girth welds using inverse wave field extrapolation of ultrasonic data

Ultrasonic non-destructive testing is a renowned method for the inspection of girth welds. However, defect sizing and characterization remains challenging with the current inspection philosophy. In addition, data display and interpretation is not straightforward and requires skill and experience from the inspector. A better and more reliable inspection result would contribute to safer pipeline construction and economic benefits (like low false call rates and the possibility to use smaller wall thickness). In seismic exploration, images of the subsurface have been obtained using acoustic data, for many years. The main objective of this thesis is to apply the imaging concepts developed in seismic exploration to the application field of ultrasonic NDI in 2D and to give a proof of concept in 3D. The 2D inverse wave field extrapolation (IWEX) imaging approach is presented, which is based on the Rayleigh II integral for back propagation. A 3D imaging approach is presented called 'two pass imaging'. This approach consist of two 2D imaging steps performed in two orthogonal directions. For homogeneous media, this two-pass process is known to be exact. With a linear array, a surface is scanned. In the direction parallel to linear array (the in-line direction) full offset imaging can be performed. In the direction perpendicular to the linear array (the cross-line direction) zero-offset imaging can be performed. The current traditional inspection philosophy is based on zonal discrimination. This philosophy is well regulated in codes and standards. From these codes and standards, a rough requirement was derived for the resolution of the IWEX imaging approach. The resolution that can be obtained with IWEX imaging is investigated on the basis of point spread functions. Aperture limitations and directivity effects of the ultrasonic elements are taken into account. Result of 2D images from different insonification and scatter paths are presented. The data was obtained from measurements on various test blocks with machined reflections using a linear array. The test blocks also contain reflectors that are representative for actual weld defects. To demonstrate the proof of principle of 3D IWEX imaging, test blocks were manufactured with suitable reflectors to study the characteristics of 3D imaging. Also, a final experiment was performed on a test block with an intentional defect in a real weld. The results are presented and discussed. The IWEX imaging approach basically removes propagation effects from source to scatterer and from scatterer to receiver. In ultrasonic data from carbon steel, two wave modes exist with different sound velocities (longitudinal and transversal). To construct an image, one of these modes must be chosen for the back propagation. As a consequence, the energy in the data from the other mode is not treated correctly and will produce leakage artifacts in the image. Artifacts in L-T images caused by L-L arrivals are discussed and illustrated with an analytical example. In addition, a procedure is presented and demonstrated to suppress these artifacts. It is concluded that the IWEX imaging approach is suitable to detect, size and characterize defects that are common in girth welds. The L-L image (direct longitudinal insonification path and direct longitudinal scatter path) will usually contain the most information. However, for the area just below the surface, the LL-LL image (insonification and scatter paths via the back wall) must be used. To confirm the L-L and the LL-LL images, the L-T image (direct longitudinal insonificaton path and direct transversal scatter path) can be used, after the artifacts are suppressed with the proposed procedure. It is also concluded that 3D IWEX imaging is feasible and that it contributes to an improved resolution in the lateral direction. Finally, recommendations are given for further research. Although ultrasonic imaging based on inverse wave field extrapolations offers a huge potential for ultrasonic NDI, still work must be done to apply IWEX in practice. Evaluation programs are necessary for the acceptance of the imaging technique. In addition, new codes and standards must be written for industrial use.