M.G. Lozev, R.L. Spencer, P. Patel, T.C. Huang - Edison Welding Institute
Due to flexibility of phased-array (PA) automated ultrasonic testing, many deep offshore components with thick cross sections, complex geometries, and difficult-to-inspect materials that were traditionally considered non-inspectable by ultrasonic testing (UT) now can be inspected. This paper from the well construction/risk track compares PA inspection data to conventional UT inspection data, predicted results, and metallographic measurements.
Over the last 10 years, automated ultrasonic testing (AUT) has been used increasingly in offshore pipeline construction to improve defect detection and sizing reliability. AUT inspection offers advantages over conventional manual ultrasonic testing (UT) including:
- Improved reliability and performance (defect detection and sizing)
- Ability to obtain an electronic copy of inspection results
- Increased speed.
Current AUT techniques are based on combinations of amplitude-based pulse-echo (P/E) method or pitch-catch mode using single-element multi-probes (focused or non-focused) or phased-array (PA) transducers with beam-fixed angles and the time-based time-of-flight diffraction (TOFD) method.
The most extensively used practices and standards for AUT system requirements, AUT procedure development, AUT system qualification, and acceptance criteria are based on the zonal approach that was developed in the 1990s. However, even with advanced AUT instrumentation, there are still uncertainties in defect detection using the current zonal discrimination techniques applying multi-probe or linear PA search units with focused, fixed angle beam and an amplitude-based zonal approach for defect sizing.
The results from development of improved ultrasonic assessment methods using non-zonal approach for pipeline girth welds inspection are presented in this study.
Improved AUT techniques
Detection and sizing of service-induced fatigue cracks is often challenging because of the narrow crack-tip opening and smooth face of the fracture surface. When using an angle beam, single probe P/E ultrasonic technique, the smooth fracture surface acts as a relatively good acoustical mirror that tends to reflect the ultrasonic sound beam so that very little sound energy is reflected back to the transducer.
If the crack is connected to the outside diameter (OD) or inside diameter (ID) surface, a strong corner trap signal is usually very apparent to help in the detection process; however, cracks with significant through-wall extension can be dramatically undersized using the P/E technique. In addition, mode-converted signals may be present that can be mistaken for reflections from the crack. If measurements are made from mode-converted signals, this will lead to inaccurate sizing.
The tandem pitch-catch ultrasonic technique can be effective for detecting fatigue cracks, but it requires the use of several different angle combinations and precise transducer spacing to obtain full inspection coverage. The use of pitch-catch for measuring the through-wall extent of fatigue cracks again can require different beam angles and precise transducer placement and spacing.
To overcome these problems, new emerging detection and sizing techniques such as PA electronic (E-) and sectorial (S-) scanning using multiple angles for better detection, advanced imaging, and time-based diffraction techniques for sizing were recently introduced. These techniques are based on non-zonal approach.
AUT approach
During the project, pipe sections containing fatigue cracks were scanned using conventional and PA UT. To evaluate the ultrasonic response from fatigue cracks, pipe sections containing fatigue cracks in the weld region, as well as those having fatigue cracks in the base metal, were scanned. A variety of beam angles and transducer combinations were used with both the P/E and pitch-catch techniques.
Except for the TOFD scans, calibration sensitivity in each case was achieved by adjusting the signal from a 1.2-mm (0.047-in.)-diameter side-drilled hole, at a depth of 25.4 mm (1.0 in.) to 80% full-screen height. Depending on the noise level, an additional 6 to 10 dB was added to the calibration sensitivity level during scanning and data acquisition. The test sensitivity level for TOFD scans was set at the highest gain setting that could be achieved without having excessive signal noise.
Based on experimental results from surface-breaking fatigue cracks in material up to 25 mm (1.0 in.) in thickness and with cracks having different through-wall heights, the following observations were made:
- Good detection of the corner trap from the cracks was achieved using P/E techniques with shear wave beam angles in the range of 45 to 50 degrees and a transducer frequency in the range of 4 to 10 MHz. These angles, however, were not good for determining the through-wall height of the cracks
- The best detection of the crack face on through-wall cracks was achieved using tandem pitch-catch and dual-PA pitch-catch techniques. These techniques provided the best detection with the lowest noise level
- Good through-wall sizing was obtained using shear wave beam angles in the range of 50 to 70 degrees and a transducer frequency in the range of 4 to 10 MHz
- Single-element and PA techniques were both useful for sizing provided a crack tip signal could be detected
- UT P/E techniques having small beam sizes in the through-wall dimension had consistently better sizing accuracy
- Fatigue cracks less than 1.5 mm (0.06 in.) in height were difficult to size because the relatively weak crack tip signal was usually not fully resolved from the strong corner trap signal and root geometry echoes
- Even with relatively high gain settings during scanning, crack faces, and crack tips from the fatigue cracks were difficult to detect
- PA P/E techniques worked well for crack sizing when using the sector scan technique to electronically steer the sound beam through a range of angles. Even when a tip signal could not be detected, the sector scan display gave a good estimate of the through-wall extension of the cracks
- TOFD was only tried on ID connected cracks having through-wall extensions less than 50%. However, in these cases, TOFD worked very well for sizing fatigue cracks when a good diffracted signal was received. Because of the low signal/noise ratio, it was sometimes difficult to resolve the diffracted signal from the crack in these experiments
- The use of data-merging techniques prov- ed very useful for visualizing the through-wall extent of cracks, as well as the crack length.
