|NDT.net May 2006 Vol. 11 No.5|
Phased Array Ultrasonic Technology Contribution to Engineering Critical Assessment (ECA) of Economizer Piping WeldsCiorau,P.-Ontario Power Generation-Pickering - Canada - firstname.lastname@example.org
Gray, D.- Ontario Power Generation-Pickering - Canadaemail@example.com
Daks, W- CAD WIRE- Markham - Canada - firstname.lastname@example.org
1.0 BackgroundECA team was asked OPG-Inspection and Maintenance Services (IMS) to detect and size fatigue cracks in the counterbore area of carbon steel boiler feedwater piping welds (OD x t = 324 mm x 38 mm) of a thermal station. A sizing accuracy of 1 mm for the outer ligament was required. It was documented the fatigue cracks could be branched, change the morphology (from oxide-filled wide width to hair-tight aspect); they are tilt at different angles and the cracks have different shapes (see Figure 1).
Weld cap may not be totally removed in the field and access from component side is limited (Tee-welds). The technique must be validated on open and blind trials on retired-for-cause samples with service-induced cracks with known height. Some samples will undergo magnetic particle examination over the side faces for crack height confirmation. Other samples will be broken and fracture mechanics data will be compared with PAUT data. Some samples will be measured by high-accuracy optical methods. The field inspection will be witness by ECA engineer and by the customer. In order to get a reliable sizing, it was decided to report to ECA the crack ligament to outer surface. In this way, the errors due to crack morphology in reading the crack leg, inner surface irregularities and crack orientation in counterbore transition area (variable thickness) are eliminated (see Figure 2). As a conservative decision, ECA will subtract -2 mm on outer ligament value reported by PAUT.
It was also decided to use for calibration a retire-for-cause coupon with known crack height. In this way, the errors due to velocity change in the weld, pipe and component are taken into account in over-all assessment of ligament by PAUT.
2.0 Phased Array Procedure ValidationOMNISCAN 32/32 and FOCUS LT 32/128 were used in combination with 9 probes -hard face longitudinal waves-frequency range 3.5 - 10 MHz, and three probes in shear waves-frequency range: 5-8 MHz. They should be used for different possible field scenarios, such as: access on the top of weld, access on component side, un-even surface, piping roughness, a higher attenuation, rough inner surface, size the last significant tip (strong focusing at specific depth), different methods for sizing. Both systems produce repeatable results and their performance was documented.
Figure 3 presents a comparison between PAUT results before breaking the sample and the fracture mechanics of sample 9B. It could be noticed the crack height variation between 8.3 mm to 8.9 mm along a 30-mm pipe circumference length. Side measurements may not lead to the highest crack height. PAUT results cannot be related to the crack height at the end face of the sample.
Figure 4 presents an example of PAUT data comparison between FOCUS LT + Probe 23 and OMNISCAN 32/32 with probe 28. It may notice the crack height variation along the pipe circumference. In this case, crack height has a variation of 0.5 mm over 20 mm length.
Figure 5 presents data comparisons between magnetic particle testing and PAUT for sizing a tight crack in the 3E-header weld sample. Initially, this crack was reported by TOFD as 7-mm height.
Figure 6 presents data comparison for shallow crack sizing. An undersizing trend may be noticed.
Last significant tip (see Figure 7) is detected by phased array with high accuracy due to an improved signal-to-noise ratio on crack tip by the focus beam and to the diversity of scanning patterns.
The over-all performance of PAUT in sizing fatigue cracks in feedwater piping welds (counterbore zone) is presented in Figure 8. An undersizing trend of -0.5 mm may be noticed. However, the ECA approach took a conservative decision to add another 2 mm to PAUT reported data. The outside ligament value was corrected according to ECA criteria: L ECA = L PAUT - 2 mm.
3.0 Phased Array in the FieldExamples of crack detection, sizing and confirmation using lateral Scan L-waves and/or S-scan shear waves from the field are presented in Figure 9 to Figure 11.
Table 2 presents an example of PA reliability and the ECA decision for two cracks
The field inspection was performed on 10 welds for more than 40 locations within one day. The defect disposition was witnessed by ECA engineer. The straightforward S-scan display of the weld, HAZ and transition zones in the counterbore assured a reliable measurement of the crack profile around 360 pipe circumference. It was concluded the crack maximum height is located in 2-3 zones. This information was forwarded to ECA team for a better fine-tuning of finite-element analysis data input.
4.0 ConclusionsApplications of phased array to evaluate the crack ligament in the counterbore region of economizer piping welds was successful implemented in the field and the PA data were used by ECA for Tee disposition. Sizing of the last significant tip is a challenge for PA techniques. A combination of three to four techniques is required and a diversity of phased array probes with different frequencies and focus depth led to a reliable ligament evaluation. PA undersized the cracks by - 0.5 mm (average). Conservative ECA decision was to subtract -2 mm from the ligament reported by PA. Field inspection was reliable and very productive. ECA engineer witnessed the PAUT inspection. The sectorial scan data display was very friendly used by ECA specialist and the customer understood within minutes the data display, cursors readings and ligament evaluation. The disposition criterion for Tee with significant cracks was based on PAUT reliability (validation, diversity and redundancy) and ECA conservative decision. Calibration on economizer sample 9A with two known counterbore cracks contributed to a high-accuracy evaluation.
AcknowledgementsThe authors wish to thanks the following organizations and people for granting the publication of this paper and their support to the project: