This month's NDT Solution compares the response of various ultrasonic probes to cracklike indications in cladding and underclad cracks. One of the primary conclusions drawn would appear obvious, but is often overlooked. It is always important to evaluate performance of ultrasonic probes with reflectors as similar as possible to real indications, as performance characteristics of the probes can depend on the type of the reflectors. G.P. Singh Contributing Editor

 

Introduction
Stainless steel claddings in the interior of reactor pressure vessels and other primary circuit components exert a strong influence on the ultrasonic inspection of the entire component as well as on the cladded area. Since the introduction of ultrasonic inspection for inservice inspections during the 1970s, the cladding influence has been investigated many times (Kolb and Wölfel, 1975; Wüstenberg et al., 1976; Wüstenberg and Schulz, 1972). Interest has increased in ultrasonic inspection of the cladding and the area close to it for studying certain corrosion aspects and possible cracks or crack growth. (Benoist et al., 1990; Ciorâu et al., 1987; Gruber et al., 1986; Heckhäuser and Gischler, 1988; Hesselmann and Wüstenberg, 1990; Ortega and Funes, 1988). This paper concentrates on evaluating the performance of specific ultrasonic probes frequently used for typical crack type indications in the cladding, especially for underclad cracks. All probes of interest here are probes with inclined incidence of the excited beams. Measurement of the cladding layer thickness and the tests for lack of fusion between the base material and the cladding were not considered.

 

---

The intention was to simulate worst case conditions for the crack detection.

---

 

Probe Types
Figure 1 shows the most frequently used probe types for the inspection of claddings. The transmitter/receiver technique for inclined longitudinal waves (TRL) with angles of incidence between 60 and 73 degrees is actually the most important concept for the ultrasonic inspection of claddings (Wüstenberg and Schulz, 1972). However, there are other approaches with some remarkable features, such as the SLIC probes which excite longitudinal waves as well as shear waves and are mainly based on diffraction at the extremities of indications (Gruber et al., 1986). Since this probe operates in a transmit/receive mode with a proper acoustic insulation between transmission and reception, a fairly good near probe resolution is achieved. The focusing probes with inclined longitudinal waves in the range of about 45 to 60 degrees without separation between transmission and reception are more or less restricted to the inspection of ferritic base material (Saglio et al., 1982; Wüstenberg et al., 1976). These probes have difficulties with cracks inside the cladding because of their limited near probe resolution. Therefore, our comparisons and evaluations were restricted to the two transmitter/receiver probe types shown in Figure 1. In addition, a phased array TRL probe with the possibility of changing the angle of incidence between 36 and 73 degrees for longitudinal waves was considered. This type of probe should also be able to guarantee an optimal detectability for cracklike reflectors almost perpendicular to the interface and to detect diffraction effects from crack tips.

 Figure 1 - Detection of underclad cracks with ultrasonic methods.

 

The primary goal of this investigation is to check the probes' ability to detect all cracks lying close to the surface and close to the interface of claddings, and which have a depth extension greater than 5 mm (0.20 in.) within the ferritic base material.

The work was done for Imatran Voima Oy from the Loviisa power plant in Finland (Brekow et al., 1992). Since the investigated type of cladding corresponds to the cladding used in many East European reactors, there was also a strong interest in checking the capacity of the probes to detect typical manufacturing discontinuities from the welding procedure within the cladding, for example, slag inclusions or lack of fusion between different layers and especially hot tears emanating from slag inclusions. This diversity of inspection tasks forced the use of TRL probes especially designed for different inspection tasks, adapting them to the depth areas by choosing a typical focusing. This can usually be achieved by the choice of appropriate transducer dimensions in the plane of incidence and in the plane perpendicular to it, and by an adaptation of the angle of skewing between the transmitter and the receiver system. Figure 2 shows the parameters of the applied TRL probes and their dimensions. The other important probe concept investigated is the SLIC 50 probe from the Southwest Research Institute in San Antonio, Texas (Gruber et al., 1986).

Table 1 Probe data and amplitude calibration for the TRL Probes

Figure 2- Probe data and amplitude calibration for the TRL Probes

 

Test Blocks and Reflectors
The cladded test blocks were provided with 25 reflectors, enabling 33 different probe to reflector conditions. The intention was to simulate worst case conditions for the crack detection by inclined reflectors and by an intentionally introduced probe skewing during some of the mechanized scans. The thickness of the cladding of the Loviisa test block ranges from 10 to 12 mm (0.40 to 0.48 in.). Figure 3 shows a top view of the test block with an 11 mm (0.44 in.) cladding. All reflectors are characterized with a length times depth extension and, if inclined, the azimuthal orientation is indicated in degrees. Indications starting from the surface are represented with thick black lines. Indications below the surface, mostly starting below the cladding interface, are indicated by dotted white rectangular bars. There are five underclad crack type reflectors. Numbering of the indications from 1.1 to 6.2 is shown in Figure 4. For the test block with real underclad cracks from the welding procedure and with 2 and 3 mm (0.08 and 0.12 in.) deep flat bottom holes, the cladding thickness is about 5 mm (0.20 in.). The natural underclad cracks are assumed to have a depth of about 1.5 to 2 mm (0.06 to 0.08 in.) and a length of 3 to 4 mm (0.12 to 0.16 in.).

Figure 3 - Test block Loviia I.

 

 

 

Figure 4 - Signal to noise ratio at Loviisa test block.

All test reflectors used during the investigations have different shapes, sizes, and positions and can be divided into the following categories:

• planar reflectors with different azimutal orientations, starting at the surface and remaining within the cladding
• planar reflectors from the surface running through the cladding towards the base material with different inclinations
• volumic reflectors with extensions corresponding to slag inclusions, partly combined with a planar reflector representing a hot tear starting at a slag inclusion
• cracklike reflectors in the base material, starting at the interface
• planar reflectors with a depth extension restricted to about 2 mm (0.08 in.)
• Additionally, we have used a test block with flat bottom holes representing underclad cracks and a block with real underclad cracks.

 

Measurements
Manual scanning. Manual scanning was carried out to measure the maximum achievable signal to noise ratio for all mechanical scanning directions. Therefore, the noise level measured on the screen and the maximum amplitude produced with manual optimization were recorded for all reflectors of the program. Figure 4 shows the signal to noise ratio at different reflector types for the SLIC 50 and the TRL probes.

Mechanized scanning. Scanning was done in two ways: a total surface scan with a meandering movement across the surface of the test block and a linear scan perpendicular and parallel to the main orientation of the reflectors. All data were recorded digitally with a linear A-scan converter (8 bit digitization with a 42 dB dynamic range) with 256 pixels per A-scan. To check the influence of skewing angles between indication and probe, one of the linear scannings at each reflector was carried out with an intentional misorientation between the probe and the indication by choosing ±15 and ±20 degree skewing angles during the scanning. The resulting dependency is an important factor for indication detectability. A typical result of the mechanical scanning is presented in Figure 5, showing a time displacement picture with some of the reflectors. Echodynamic curves can be derived from the stored A-scan data as presented in Figure 6.

Figure 5 - TD scan of real UCCs with a TRL probe.

Figure 6 - Comparison of echodynamic waves measured by lateral scan of test reflectors No. 1.1, 1.2, 1.3, and 1.4.

 

 

Evaluation of Measurement Results
The manual scan delivers a comparison of the probes simply based on the signal to noise ratio. Figure 4 shows a typical distribution of the signal to noise ratio for different reflector types. The comparison between the SLIC probe and standard TRL probes for inclined longitudinal waves shows a significant difference: the detectability of flat bottom holes representing underclad cracks and the detectability of real underclad cracks is remarkably smaller for the SLIC probe compared to TRL probes. This is apparently due to the fact that the SLIC probe mainly uses a crack tip diffraction for the detection of cracks, whereas the TRL probe is based on a geometric surface reflection.

In cases where the crack tip diffraction is reduced due to stresses or due to the specific transition behavior at the end of cracks, a weak interaction between the ultrasonic waves and the indication is to be expected (Figure 7). This is the case for the real underclad crack and for a 2 mm (0.08 in.) flat bottom hole where the circular shaped diffraction lines scatter the energy into the whole half space, transmitting back to the probe only a small portion of the energy. That specific difference is illustrated in Figure 7. An idealized crack was typically assumed for theoretical calculations and also applied within most of the national and international performance demonstration programs. For comparison, a second crack with a reduced crack tip diffraction is presented. The reduction is due to the basic fact that between the stress free condition at the crack surface and the uncracked area a smooth transition must be assumed depending on the special stress situation and other factors. Figure 8 compares the time displacement scans with an SLIC 50 probe and a TRL probe on a real underclad crack together with the A-scans. The typical advantage of the SLIC 50 probe is given by its sizing potential on larger cracks based on the crack tip diffraction. But there is some discussion needed about the justification of such a probe approach in cases where for conservation reasons a crack tip interaction cannot be assumed.

Figure 7 - Crack tip diffraction in underclad cracks.

 

Figure 8 - Comparison of indications from underclad cracks produced by a TRL probe and by an SLIC probe.

Based on echodynamic plots and TD images (Figures 5 and 6) from the mechanized scans, an evaluation of the results was carried out by several Level II ultrasonic inspectors. They classified all records according to the criteria listed below into "representing an indication" and "not representing an indication." Based on their findings, a relative operation characteristic evaluation was completed to support the signal to noise ratio results.

We have introduced three sets of different criteria for the evaluation of important indications. For all sets, the recording threshold was found to be based on the 6 dB drop of the echo from a 3 mm (0.12 in.) deep side drilled hole within the cladding.

• All indications with an echo amplitude above the recording threshold and a half value extension of echo dynamic plots larger than 10 mm (0.4 in.) are to be regarded as defect indications.
• All indications with echo amplitudes 6 dB above the noise level and with half value extensions greater than 10 mm (0.4 in.) can be regarded as defect indications.
• All indications with echo amplitudes 6 dB above the noise level and with half value extensions larger than 5 mm (0.2 in.) can be regarded as defect indications.

Besides those three criteria, others with gradually increasing recording thresholds and decreasing half value extensions of the echodynamic plots can be chosen. With the help of a graphical presentation in a probability plot, this procedure delivers additional values for virtual detectabilities and false calls rates. The result of the statistical evaluation of all reflectors can be represented in the ROC diagram shown in Figure 9, where the probability of reflector detection is plotted versus the probability of false calls. The diagram shows that for the SLIC probe one must take into account larger false call rates for the same reflector detectability as for the TRL probe. The diagram is restricted to an optimal adaptation of the horizontal orientation between probe and defect (no skewing). The ROC curves change dramatically if a horizontal skewing must be considered.Even with an optimized inspection technique based on an optimized TRL probe, a false call rate of 8 percent for a detectability of 80 percent, and of 13 percent for a detectability of 90 percent, must be taken into account. The application of a phased array probe instead on a TRL probe does not significantly improve the detectability in the near cladding area and the zone 5 mm (0.2 in.) below the interface, but even in this area those probes offer a better possibility of sizing. And since for an inspection zone of about 30 mm (1.2 in.) close to the coupling surface the conventional TRL concept is very difficult to enlarge without losing sizing and localizing capacity, the TRL phased array probe, with its different focusing and angles of incidence, may be a solution which combines coverage, detectability and sizing potential for the whole area.

Figure 9 - ROC curves for the SLIC 50 and the TRL probe.

 

Conclusions
The comparison of probes for the inspection of cladding shows a remarkable fact: depending on the reflectors used, a different performance characterization of the probe must be expected. It seems important for performance demonstration trials - especially those including the underclad area - to use reflector types as close as possible to the real cracks to be detected. For different detection approaches, reflectors representing underclad cracks should have a reduced diffraction at the crack tip areas in order to check their potential to detect the indication based on surface reflections of the ultrasonic waves.

 

References
Benoist, P., M. Serre, F. Champigny, and D. Sanchez, "Ultrasonic Inspection of Defects Close to the Inner Surface," at the 10th International Conference on NDE in the Nuclear and Pressure Vessel Industries, 1990, Glasgow.

Brekow, G., E. Schulz, and H. Wüstenberg, "Bewertung der Ultraschallprüfung für den plattierungsnahen Bereich des Kernkraftwerkes Loviisa in Finnland," in Report of the BAM Berlin for Imatran Voima Oy, Jun. 1992.

Ciorâu, P., S. Mândrilâ, C. Zambiloiu, and N. Fratutescu, "Practical Contributions to the Ultrasonic Testing of Cladded Welded Joints," in Proceedings of the 4th European Conference on Nondestructive Testing, Vol. 4, Sep. 1987, pp 2350-2357.

Gruber, G.J., D.R. Hamlin, H.L. Grothues, and J.L. Jackson, "Imaging of Fatigue Cracks in Cladded Pressure Vessels with SLIC 50," NDT International, Vol. 19, 1986, pp 155-161.

Heckhäuser, H., and K.H. Gischler, "Das Zipscan-System bei der Ultraschallprüfung an plattierten Bauteilen und Rohrleitungen; Automatisierung in der Ultraschallprüfung, Stand der Technik, Entwicklungstendenzen bei mobilen Prüfanlagen," in Proceedings of the Seminar, Mechanized Ultrasonic Inspection, DGZfP, Jul.-Aug. 1988, Berlin, pp 122-132.

Hesselmann, W., and H. Wüstenberg, "Beitrag von theoretischen Modellen zur Lösung von Problemen der Ultraschallprüfung an plattierten Komponenten," in Procedings of the Seminar, Model and Theory for the Ultrasonic Inspection, DGZfP, May-Jun. 1990, pp 190-205.

Kolb, K., and M. Wölfel, "The Ultrasonic Inspection of Reactor Pressure Vessels," Materialprüfung 17, No. 10, 1975.

Ortega, J., and A. Funes, "Experiences Related with Ultrasonic Inspection of Austenitic Materials with Corrosion Resistant Cladding," at IAEA Specialists Meeting on Inspection of Austenitic Dissimilar Material and Welds, Jun. 27-29, 1988, Espoo, Finland.

Saglio, R., A.M. Birac, J.C. Frappier, J. Viard, and B. Berger, "État des Études faites en France sur la surveillance des défauts sous revêtement," Conference on Periodic Inspection of Pressure Vessels, Oct. 12-14, 1982, London, England.

Wüstenberg, H., and E. Schulz, "Versuche zur Feststellung plattierungsnaher Reflexionsstellen an Reaktorteilen mit Ultraschall," presented at the annual assembly of the German Society for NDT (DGZfP), May 1972, Saarbrücken.

Wüstenberg, H., J. Kutzner, and W. Möhrle, "Focussing Probes for an Improved Defect Sizing at the Ultrasonic Inspection of Thickwalled Reactor Components," Materialprüfung 18, No. 5, 1976.

 

* Bundesanstalt für Materialforschung undprüfung, Unter den Eichen 87, 1000 Berlin 45, Germany; phone 030 8104-6210; fax 030 8119-396.

Copyright © 1996 by the American Society for Nondestructive Testing, Inc. All rights reserved.  

[ Materials Evaluation ]

IRRSP, NDT Handbook, The NDT Technician and www.asnt.org are trademarks of the American Society for Nondestructive Testing, Inc. ACCP, ASNT, Level III Study Guide, Materials Evaluation, Nondestructive Testing Handbook, Research in Nondestructive Evaluation and RNDE are registered trademarks of the American Society for Nondestructive Testing, Inc. ASNT exists to create a safer world by promoting the profession and technologies of nondestructive testing.