1) Phased array ultrasonic techniques were compared to other methods for sizing cracks in samples ranging from 6mm to 38mm thick.
2) PAUT provided crack shape images through S-scan displays and generally undersized cracks compared to optical and magnetic particle methods, with an average undersizing of 0.4mm.
3) Undersizing errors increased for cracks over 12mm deep and were reduced by using combinations of shear and longitudinal waves with higher angular resolution.
1) Phased array ultrasonic techniques were compared to other methods for sizing cracks in samples ranging from 6mm to 38mm thick.
2) PAUT provided crack shape images through S-scan displays and generally undersized cracks compared to optical and magnetic particle methods, with an average undersizing of 0.4mm.
3) Undersizing errors increased for cracks over 12mm deep and were reduced by using combinations of shear and longitudinal waves with higher angular resolution.
1) Phased array ultrasonic techniques were compared to other methods for sizing cracks in samples ranging from 6mm to 38mm thick.
2) PAUT provided crack shape images through S-scan displays and generally undersized cracks compared to optical and magnetic particle methods, with an average undersizing of 0.4mm.
3) Undersizing errors increased for cracks over 12mm deep and were reduced by using combinations of shear and longitudinal waves with higher angular resolution.
1) Phased array ultrasonic techniques were compared to other methods for sizing cracks in samples ranging from 6mm to 38mm thick.
2) PAUT provided crack shape images through S-scan displays and generally undersized cracks compared to optical and magnetic particle methods, with an average undersizing of 0.4mm.
3) Undersizing errors increased for cracks over 12mm deep and were reduced by using combinations of shear and longitudinal waves with higher angular resolution.
Contribution to Crack Sizing by Phased Array Ultrasonic Techniques.
Part 2: Comparison with Optical, Magnetic Particles, Fracture Mechanics and
Metallography for Last Significant Crack Tip. Ciorau, P. – Ontario Power Generation – Canada peter.ciorau@opg.com
Abstract: The paper presents phased array results for 1-D linear array probes of high frequency (7-10 MHz) in L-, and S-waves for detecting the crack shape and the last significant tip. Fatigue and stress-corrosion cracks with height ranging from 1.6 mm to 20.4 mm were detected in welded samples, piping welds and straight bars with thickness between 6 mm to 38 mm. The results of S-scan display are compared with different methods: optical, magnetic particles, fracture mechanics and metallography. The experimental results concluded the undersizing trend of PAUT in detecting the last crack tip or closure, in spite of using dynamic depth focusing, and/or focusing on crack tip. The average undersizing error is – 0.4 mm. This error increases for cracks with depth > 12 mm. The largest errors occur when the crack is sized from outer surface coupled with initiation from the outside surface with propagation towards the inside surface. These errors were reduced by a combination of shear and longitudinal waves and by increasing the angular resolution.
1.0 Introduction
Our previous published papers compared conventional and phased array ultrasonic technology [ 1- 4 ]. Two conclusions were PAUT provided more accurate sizing and produced an image of crack. These image patterns (see Figure 1) were more easily interpreted than the A-Scan from conventional instruments. Small cracks (h crack < 2 mm) could be sized due to redundancy in angles of S-scan.
Figure 1: Example of a stress-corrosion crack tip display by S-scan of 10-MHz longitudinal waves 1- D linear array probe (left) and comparison with crack shape (stereo microscope, 200X) (right).
Cracks could change the width, the angle and the closure aspect (see Figure 2).
1 Figure 2: Examples of thermal / corrosion fatigue crack morphology and crack closure of last significant tip. Generally, a transition of 0.2 to 0.4 mm is noticed from a “wider” crack (15- 25 µm) to a “narrower” crack (0.5-2 µm).
The next phase of our investigation focused on sizing accuracy and crack pattern display in S-scan. These results are discussed below.
2.0 Experimental Program
The samples with cracks are presented in Table 1.
Table 1: Samples with cracks used for sizing evaluation
The crack “actual” height was difficult to assess by optical or MP from the side of the sample. An example is given in Figure 3, for sample 9B.
2 Figure 3: Example of crack height measurement by three methods: optical (top left); MP (right) and after the crack was broken open (bottom left). Difference between 0.2 mm to 0.6 mm in height measurement was found.
The crack height was oversized by optical and MP. This may be explained by magnetic particle accumulation on the crack tip and by light reflection on metallic face of the sample. The crack height varied along the sample width, with the maximum height in the middle where PAUT was applied. Due to the importance of these samples for training and procedure validation, only one sample was broken open to measure the crack height via fractography. The remainder of the measurements were performed on the side of the samples by optical and MP methods. Welded samples were etched and a metallographic examination added for crack evaluation.
The probes used for evaluation are presented in Table 2.
Table 2: 1-D phased array probes used for crack height evaluation.
Frequency Pitch [mm ] /
Probe ID Remarks [ MHz ] nr. elements 2J 10 0.31 / 20 LW between -65° to 65°; variable focus depths 30 10 0.31 / 32 DDF / LW in the range 8 mm – 30 mm 52+70T 8 0.33 / 10 SW with 70° wedge for OD evaluation 43+60T 7 0.6 / 16 SW with 60° wedge; angles: 28° to 85° 2J + 60T 10 0.31 / 20 SW with 60° wedge for OD evaluation 18+52TW 10 0.5 / 64 SW with 52° and DDF for specific cracks (B18) 18 10 0.5 / 64 LW / DDF for thickness > 18 mm 42 10 0.5/ 32 LW / DDF for thickness > 18 mm
The probes were used in azimuthal and lateral scanning directions (Figure 4) and the maximum crack height along the sample width was recorded. All of the scans used positive and negative angles. Each scan was repeated five times. Screen shots and UT data were analyzed using S-B-A
3 layouts. Final crack height was measured in S-scan (true depth). Both OmniScan MX 32/32 and Focus LT 64 / 128 instruments were used to acquire data.
The sizing principle was based on back-scattering tip echo diffracted techniques and AATT (absolute arrival time technique) applied to a true-depth (volume-corrected) sectorial scan display (see Figure 5-as principle, and Figure 6-as real data related to specimen and probe).
Figure 4: Examples of crack height evaluation by: Azimuthal S-waves (left), by Azimuthal L-waves (middle) and by lateral L-waves (right).
Figure 5: Principle of crack sizing based on back-scattering diffraction for shear waves (left) and longitudinal waves (right).
Figure 6: Crack height sizing for shear waves (left) and L-waves (right). Note the SCC display with branched tips sized by L-waves (right). 4 3.0 Data Evaluation
Examples of crack height measurements are presented in Figure 7 to Figure 16.
Figure 7: Crack sizing comparison between MP (left) and PAUT (right) on sample 3E.
Figure 8: Crack height plotting for sample B12 (left – OD sizing; right - ID sizing)
Figure 9: Crack height sizing in sample C1 by OmniScan 32/32 and probe 30 (L-waves). A systematic undersizing of 0.3 mm was found by all three techniques compared to optical/MP.
5 (courtesy of OlympusNDT-Waltham-USA)
Figure 10: Stress Corrosion crack in sample OHR-20 at different refracted angles. Crack height has a minimum response normal incidence. See values from Table 3.
Table 3: Crack height measured at different angles for sample #B18 using a 10 MHz L-wave probe.
Angle [ ° ] 46 32 17 0 -17 -32 -46
Height [ mm ] 14.2 13.6 12 11.5 11.7 13.7 14
The best sizing angle range is between 30 – 35°, and symmetrical negative (-30 to -35°).
Figure 11: Comparison for sample 3B with three cracks. PAUT undersized by 0.5 mm.
6 Figure 12: Crack sizing using L-waves on sample C2. The crack was measured optically as 20.4 mm. Note the undersizing of OD technique.
Figure 13: Crack sizing using S-waves on sample C2. ID negative came very close to the optical value (20.3 mm vs. 20.4 mm-optical). Note the same trend of undersizing for OD technique.
7 Figure 14: Examples of maximum SCC evaluation on sample OHR-20.
Figure 15: Crack sizing by two systems on sample B18: left: OmniScan with 7MHz probe of 16- elements, sampling at 0.5 degree (data converted for Tomoview analysis); right: Focus LT, 10MHz with optimized wedge + Dynamic Depth Focussing (DDF) + sampling at 0.2 degrees.
Figure 16: Crack sizing and crack orientation performed from outer surface on sample C1.
Phased array results and the back-scattering from the last significant tip depend on crack opening, , load, oxide presence, probe access and optimizing the focus beam along the crack height [6-10]. An example of the influence of phased array set-up on crack display is presented in Figure 17.
8 Figure 17: Phased array set-up influence on SCC height measurement for 10-MHz probe, pitch 0.5 mm, 64 elements, L-waves mode.
When the full probe aperture was used with F=20 mm i.e. focused on the inside surface, the crack was not detected. The best PAUT response was with a 20-element aperture, DDF and a focal length of 15 mm.
The over-all performance of sizing for this experiment is presented in Figure 18.
Figure 18: Over-all sizing performance on cracks from Table 1. An undersizing trend was found. This degree of undersizing increased with crack height, especially for the OD technique.
4.0 Conclusions
1. The crack tip or closure i.e. the last 0.2-0.4 mm was difficult to detect and size. 2. Stress corrosion cracks were more accurately sized by L-waves 3. Fatigue cracks were more accurately sized by S-waves 4. A very narrow and energetic beam supplemented by DDF and a fine angular resolution resulted in the most accurate better sizing, especially on compressed cracks. 5. The cracks presented significant variations (between 0.2 to 0.6 mm) along the sample width 9 6. Only one sample was compared with fractography. The PAUT results presented in Figure 19, were very similar to the fractography measurements for the middle of the sample. 7. The results were consistent with our previous results i.e. the last significant tip with PAUT undersized by 0.2 to 0.6 mm. 8. Under field conditions, these techniques were expected to undersize by 0.5 mm for cracks with height < 8 mm and 0.8 - 1.0 mm for cracks with h > 8 mm. Recent field inspections of T- welds (ID)[4] and on repairs of outer surface-breaking cracks on outlet welds (Figure 20) confirmed these results.
Figure 19: PAUT results on sample 9B before it was broken open.
The material thickness at the crack location was 37.2 mm. The PAUT crack height = 8.7 mm. The actual crack height after breaking open the sample and fractography measurements (Figure 3), is between 8.8 – 8.9 mm. This measurement was from the centre of the sample.
Figure 20: Example of crack sizing in outlet weld using OmniScan and P52+60T. The crack was confirmed by MP (top left) and sized as h=7.3 mm (top right). Slag inclusions were also detected (see bottom). During repairs (excavation by grinding) the crack height was confirmed as h grinding = 7.6 mm.
The results of this experiment are used as a confidence boundary for ECA (see ref. 4).
10 Acknowledgements
The author wants to thank the following organizations and people:
• OPG-IMS Senior Management – for approving publication of this paper
• OlympusNDT – Waltham (USA) – for allowing publication of some figures from their book: ”Advances in Phased Array Ultrasonic Technology Applications”
• Wence Daks - CAD WIRE – Markham, Ontario, Canada - for plotting phased array data into 3-D specimens and for crack height measurements by optical and MP methods.
• Dick Gray - OPG-Thermal Production – Nanticoke for supporting this project and using PAUT as valuable ECA tool.
References
1 Ciorau, P.: “Contribution to Detection and Sizing Linear Defects by Conventional and Phased Array Ultrasonic Techniques” – Proceedings 16th WCNDT – Montreal, Sept. 2004, paper 233.
2 Ciorau, P.: “Contribution to Detection and Sizing Linear Defects by Phased Array Ultrasonic Techniques.”, ndt.net, v. 10, no. 11, Nov. 2005
3 Ciorau, P.: “Critical Comments on Detection and Sizing Linear Defects by Conventional Tip- echo Diffraction and Mode-converted Ultrasonic Techniques for Piping and Pressure Vessel Welds” – 4th PA seminar - EPRI-Miami – Dec 2005, ndt.net - vol. 11, no. 5, May 2006
4 Ciorau, P., Gray, D., Daks, W.: “Phased Array Ultrasonic Technology Contribution to Engineering Critical Assessment (ECA) of Economizer Piping Welds” - ndt.net – vol. 11, no. 5, May 2006
5 OlympusNDT: “Advances in Phased Array Ultrasonic Technology Applications”, March
2007, Waltham, USA.
6. Virkkunen, I Pitkänen, Kemppainen, M : “Effect of crack opening on UT response” –
Proceedings 9th ECNDT – Berlin - Oct. 2006, paper Th.4.4.2
7. Poidevin, C., Bredif, P., Dupond, O.: “A phased array technique for crack characterization”- Proceedings 9th ECNDT - Berlin - Oct. 2006, paper Th.1.1.2
8. Saka, M., Salam Akanda, M. A., 2004, “Ultrasonic measurement of the crack depth and crack opening stress intensity factor under a no load condition”, Journal of Nondestructive Evaluation, vol. 23 (2004), no. 2, pp. 49 – 63
9. Pitkänen, J., Kemppainen, M., Virkkunen, I., Laukkanen, A.: “Effect of stress on ultrasonic response in detection and sizing of cracks”, Proceedings 9th ECNDT – Berlin - Oct. 2006, paper Tu.3.3.4
10. Quimby, R.:” Practical limitations of TOFD on power station main steam pipework”, Insight, vol. 48, no. 9 (Sept. 2006), pp. 559-563.
