SAFT – Is it a Tool for Improved Sizing in Ultrasonic Testing?

Jorma PITKÄNEN, Posiva Oy, Olkiluoto, Finland

Abstract. The classical tasks of ultrasonic NDT are detection, sizing and characterisation of structural discontinuities such as cracks, lack of fusion in welds, inclusions of foreign material or de-laminations. In ultrasonic testing the detection is important, nevertheless the characterisation and sizing are necessary for the acceptance of detected defects. The more exact characterisation of the defect gives information for discrimination of defects with regard to volume or planar defects and allows to interpret if the signal is a false call or not. SAFT is one tool for improved understanding of detected ultrasonic signals. The imaging system applied to improve the reliability of the assessment is the use of back propagation of elastic waves in synthetic aperture focussing technique, SAFT. SAFT technique is used for testing of pipes, turbines, plates, vessels or pump housings of austenitic and ferritic materials. At the moment there are many other techniques for sizing available starting from manual inspection with different types of probes and techniques such as diffraction technique, creeping wave technique, ultrasonic holography, ultrasonic tomography. Other techniques are also available for accurate visualisation and discrimination of defects in ultrasonic testing. In this study SAFT measurements were applied in sizing of cracks and other sizing techniques were evaluated as compared to SAFT.

Introduction

The typical tasks of ultrasonic NDT are detection, sizing and characterisation of structural discontinuities such as cracks or lack of fusion in welds, inclusions of foreign material or de-laminations. In ultrasonic testing the detection is important, nevertheless the characterisation and sizing for estimation of the defects are necessary. Through exact sizing the extension of the defects is well predicted and fracture mechanics is more precise as well the estimation of the lifetime. The more exact characterisation of the defects gives information for discrimination of defects to volume or planar defects and allow interpreting if signal is a false call or not. SAFT is one tool for improved understanding of detected ultrasonic signals.

The imaging system applied to improve the reliability of the assessment is the use of back propagation of elastic waves in synthetic aperture focussing technique, SAFT [1, 2, 3, 4, 5]. SAFT technique is used for testing of pipes, turbines, plates, vessels or pump housings of austenitic and ferritic materials. At the moment there are available many other techniques for sizing, e.g., manual inspection with several different types of probes and techniques such as diffraction technique [6, 7], creeping wave technique [8, 9], ultrasonic holography [10, 11, 12], ultrasonic tomography [13, 14, 15] and phased array [16, 17, 18, 19, 20, 21, 22]. There are also other techniques available for accurate visualisation and discrimination of defects in ultrasonic testing [23, 24, 25].

ECNDT 2006 - Poster 211

1

Features of SAFT inspection system are discussed and characterised and some results from measurements performed in field and in laboratory are described in the following: • The improvement of signal-to-noise ratio (SNR) is one of the main advantages, which

is considered in the theoretical background of SAFT. • SAFT reconstruction is valid in the far field (Fraunhofer Region). • This technique is mainly used for single crystal probes but applications of tandem

probes and TOFD have also been realised in various software. • The main advantage of SAFT is that the focusing possibility, which is independent on

the length of the sound path. This gives possibilities, which are not fully applied, yet. DDF (Dynamic Depth Focusing) in phased array technique is approaching this idea.

• SAFT reconstruction offers several tools for sizing using mainly crack tip detection, but not only necessarily. More exact ultrasonic field visualisation makes it easier to interpret the results.

• After SAFT reconstruction, dynamic 3D -sizing can improve crack detection, sizing and characterisation capabilities in ultrasonic testing.

Modelling gives information for understanding measured patterns of many ultrasonic inspection tasks also in SAFT. The measured signal has back propagated into the material. Even though the image is better than the raw data, the calculated image can still contain, for instance, some artefacts, which have to be recognised. In the SAFT measurements those artefacts can be seen, for instance, in turbine axel inspection near the surface. Similar artefacts can be seen for instance in the phased array field with strong focussing in the near area of the probe. The modelling gives possibility to improve the SAFT algorithm to take into consideration the changes in the inspection area such as the austenitic weld or changes in the grain size and anisotropy [26].

Inspection considerations In the SAFT inspection the probes should have the opening angle as large as possible. This is realised with small diameter piezo crystals. Echoes from several defects can be detected at the same position. The use of small piezo crystals affects also the resolution of the reconstruction as shown later. The measured ultrasonic data from the different depths of the tested object shown in the Figure 1 is used for reconstruction. The maximal distance between measured points should be half of the wavelength projected on surface, λS, taking in the consideration the angle of incidence, α, in the tested object.

λS = λ / (2 ∗ sin α) (1)

For example, for 2 and 4 MHz shear wave probe with 45° in austenitic steel (shear wave velocity 3110 m/s) the maximum distance between data acquisition points is correspondingly 1.1 mm and 0.55 mm.

2

Scanline x-direction

Defects inside the sound field

Time of flightRF-signal A(x,y,t) in the position (x,y)

Figure 1 The principle of SAFT

Important elements in practical SAFT measurements are the following:

For SAFT the point spread function calculated according to [27, 28] can be expressed as:

zRAz

RA

PSF

λπλ

π

2

)2sin(= , (2)

where probe sound field characteristics is not taken into consideration. First zero point of the spread function first zero point is when:

AzR2λ= . (3)

Factor A is equal to effective aperture length. For rectangular piezo crystal with width Ds the sound field width, WS, will be:

SS D

zW λ2= . (4)

By combining A = Ws it results in the SAFT resolution, which is:

4S

SAFTD

D = . (5)

When the sound field characteristic is taken into consideration, which in praxis means that the width of the sound field opening will increase more than twice. This confirms measurement that lateral resolution will be increased with increasing aperture

length, but it becomes nearer and saturates to the value of 2

SD [28].

The resolution of the reconstruction is D/2, the half of the crystal size of the probe. With the small element size, the near field is close to the probe and opening angle is large. Even though the resolution of small crystals is good, the power of ultrasonic response is low. To increase the power, especially with long sound paths, larger crystals are necessary. Thus, the resolution is decreased. It is possible to increase the power by using as wide

3

crystal as possible. Of course, the near field must be between the probe and defect. The axial resolution is equal to the pulse length and not related to the aperture. With shorter pulse width the amplitude variation in the near field area attenuates. By changing focus depth or delay lines (contact or immersion) SAFT algorithm can be used also for near surface area analysis. Step-to-step distance parameter influences directly the SNR and herewith how clearly the image of a defect is displayed above the noise. This parameter does not influence directly the resolution of the image [29, 30]. It is recommended to select a probe step distance of 0.3 mm, if the probe frequency lies between 1 MHz and 5 MHz (shear or longitudinal). In general, the maximum probe step distance should not exceed 1/5 of the effective crystal diameter of the probe. Sizing is independent from the selected probe distance. In conventional techniques using the DGS-diagram, amplitude variations influence immediately the equivalent flaw size: e.g., a drop of 6 dB would reduce the equivalent flaw size of 6 mm down to 4 mm. Therefore, possible variations in coupling directly affect the reliability of inspection. SAFT relies less on amplitude information but more on time-of-flight information. The image is formed summing up information at different probe positions. As long as there will be sufficient amount of probe positions the image intensity does not change much if there is loss of coupling during some probe positions. Loss of amplitude influences the SNR in the image

Additional items that can be considered are waviness of the surface and amplification of the measurements but one of the most important factor is the effect of stress state: residual stresses of the weld (1), thermal load from the outside or inside of the component (2), and mechanical load coming from the structure of component (3). The effect of the stress state on the crack itself has a strong effect on the detectability and sizing of the defect.

Figure 2 Effect of stress state on the ultrasonic response from a real thermal crack (all data from the same crack at different positions, crack height is about 10 mm) In Figure 2 it can be seen how the effect of stress state in the crack changes the ultrasonic response. This effect has also been partly described in [31]. The reconstructed

CR = Corner reflection, MCI = Mode conversion indication, CFI= Crack face indication, CTI = Crack tip indication, ISI = Inner surface indication, CFRI = Crack face reflection indication

4

image varies quite a lot depending on the stress state of the crack. This case exhibits a thermal crack, which is actually quite straightforward in thickness direction. Characterisation and sizing of defects

SAFT can be used to determine whether a defect is planar or volumetric. Strong evidence for a planar defect is the appearance of tip reflection echo(s). Especially in the case of large cracks, the crack face (fracture surface) can be seen in many cases in the reconstruction over the noise level (3-10 dB). Different sizing techniques can be used in general. A typical way is either to use tip diffraction or 6 dB drop technique. In tip diffraction method subsurface cracks are sized by measuring the vertical distance between the positions of tip echoes. Surface cracks are sized by measuring the position of the crack tip or by determining the vertical distance between the corner and tip echoes. In Figure 4 a corner reflection indication from a real IGSCC is presented. The size of this crack is about 5-7 mm, depending on the spot, where measurement is taking place. How well the crack tips are seen, depends on each measured crack. The lower pictures of Figure 3 show crack tip signals, and sizing is easy by measuring the tip position in depth. The upper pictures of Figure 3 show mainly reflection from the crack face. In the left upper picture of Figure 3 the crack is compressed together in the middle and reflection back to the probe is clearly lower. The sizing is made mainly by 6 dB drop method in these cases. Normally corner reflection from a defect is very strong and the defect will be detected easy, if it is on the outer or inner surface.

Figure 3 SAFT reconstruction from two fatigue cracks in ferritic pipe (upper figures), from thermal fatigue crack (left, lower figure) and from IGSCC (right, lower figure) in austenitic pipe Half-amplitude (-6 dB) method is appropriate, when the defect size equals or is greater than the beam diameter and the defect is approximately perpendicular to the beam axis. In the case of SAFT technique, the beam diameter is the focussed diameter, i.e., approximately half of the crystal size as shown previously. Measurement is carried out simply to determine vertical distance between points, where the amplitude has dropped to -

5

6 dB. This method is applicable for both planar and volumetric inner defects. Half-amplitude method is demonstrated in the upper pictures of Figure 3.

Figure 4 Demonstration of sizing capability from the crack face reflection (NESC-cylinder) Crack face gives also sometimes clear reflection, which is much more than normal noise level. Figure 4 shows that the crack is visualised in a very simple way by angle shear wave probe. The probe was moved on the surface directly to the crack. The ultrasonic beam hits the crack face during this movement. The reconstruction shows quite clearly the crack shape from the noise level and the sizing is easy to carry out. Volumetric defect gives generally strong main echo and weak creeping wave echo exactly on the beam central axis behind the main echo. Sometimes the main echo can be rounded in the reconstruction. Round volumetric defects can be sized by measuring the sound path S between reflected and creeping echo. Diameter D of the defect is calculated by using equation 6 [32].

π+=

24SD . (6)

Occasionally the use of shadow technique is a useful solution [33]. It is a rather inaccurate technique, when carried out by conventional manner, but it has certain advantages. For instance, it is not sensitive to the orientation of the defect. Interpretation of the acquired data can be improved significantly by SAFT and the height of a defect will be evaluated from the reconstruction. Dynamic sizing can be carried out either with phased array using angular scanning or with SAFT or Holography. After reconstruction the sizing is carried out dynamically going through the inspected material layer by layer to find the limits of the defects as in this case the depth and the length of the crack with help of the SAFT reconstructed data. Dynamic sizing can be carried out also in mechanised ultrasonic raw data form as made in [34]. Dynamic B-scan, C-scan or D-scan evaluation method will be in the future one of the main evaluation methods for defect characterisation. Crack tip indication is clearly seen in the depth of 9 mm and it starts to vanish in the depth of 10.4 mm. The dynamic evaluation is carried out visually looking from the indications above the noise level by going through different frames where the indication is detectable. The noise level can be changed during the evaluation, which gives valuable information to the analysis. This measured fatigue crack is in the ferritic specimen. The wall thickness of the material is about 40 mm. The measured crack is surface breaking. Dynamic sizing differs from the static because one can

6

go forward and backward within the reconstructed data and be sure about the indication. Of course, in this case it is also the main interest to choose the best possible probe configuration for the measurement. The same idea from dynamic sound field is in phased array inspections but in the opposite order. The ultrasonic field is dynamic and moving. Indications are moving due to variable sound field. This gives better understanding of the nature of the defects. By moving the probe, the crack front is starting to move according to the movement of the probe. The used software was developed by the IzfP [35]. Dynamic evaluation has been used in evaluation of the indication in Figure 5.

Figure 5 Dynamic 3D -sizing with SAFT reconstruction SAFT has not yet so often been used with phased array equipment. In 2004 some sizing results were presented by using phased array raw data for SAFT reconstruction [36]. According to measurements the tip signals were located and constructed according example shown in the Figure 6. In the experiments done by Erhard et al [36], it was shown that the

7

tip was better detected with longitudinal waves than with shear waves. It demonstrated also by Müller 1999 [37]. For the phased array application, the geometry of the PA-probe must be known quite exactly as well as also the movement on the surface. By changing the angles during the measurement the index point is changing in the PA-probe. For good SAFT reconstruction it is necessary to have exact knowledge of the probe sound field characteristics in different angles. The defect size of 1 mm was the minimum for sizing for the equipment. In Germany (IzfP FHG) was developed a new type of phased array equipment in 2005, which can calculate SAFT images much faster than conventional phased array equipment [22].

Figure 6 Phased array application for SAFT reconstruction [36]

Some SAFT applications

One example of application is the inspection of welds of the main gate valves. At first defects detected manually were analyzed with SAFT. Nowadays the valves are inspected with a mechanized system. A scanning manipulator is placed on the nozzle between the main cooling pipe with a diameter of 500 mm and the valve. The geometry did not allow direct insonification of the weld, but full skip (reflection from the back wall) had to be used. Therefore, the sound path was more than 300 mm and because of the strong attenuation of the ultrasonic pulse in austenitic forging, the SNR was not sufficient to make a good evaluation of the indications. This case has been handled more thoroughly in [32].

Figure 7 SAFT reconstruction for one indication in the NESC III blind test specimen SAFT measurements have been performed for the dissimilar metal weld of the emergency cooling nozzle in the Loviisa NPP. The aim of the SAFT measurement was to distinguish the geometrical indications from the real defect indications [32]. Several other

8

cases have been visualised in [32] such as primary circuit pump thread inspection, the analysis of under clad defect, radial support plate inspection. In Figure 7 one indication can be seen from dissimilar weld of the NESC III trial. The measurement of this indication in Figure 7 was carried out from the ferritic side. The indication is locating at the dissimilar metal interface between ferritic material and buttering weld. Some parts of that indication are extending to the 8 mm height. The thickness of the material was 52 mm. The ferritic material was clad from inside. This software cannot correct the anisotropy effect of the material, but small deviations on the sizing from ferritic side can be tolerated. Anisotropy can be taken into consideration according to Shlivinski er al. [26]. In practical inspections large defects are very seldom met and sized. In the round robin test of the international NESC I-project this was possible, Figure 8. In the following two examples are presented showing the results for defects B and R of the test cylinder. In normal testing it was difficult to detect the crack tip from the normal data of defect B. It means, that the crack tip was quite tight, see Figure 9. However, in one of the measurements performed by VTT the noise from the face of the crack was seen in the data. With the help of the SAFT reconstruction, the size of the defect B was determined. The comparison to the metallography gave a sizing error of about 2.7%. In the post test the tip of same defect was detected clearly with 0°L-probe, see Figure 9b. No SAFT measurements were performed in post-test phase for this defect. For the post-test phase one new fatigue crack (R) was manufactured. This defect was measured with normal multi channel ultrasonic equipment and with SAFT equipment. With normal ultrasonic measurement it is difficult to detect the crack tip profile, Figure 10a. In the SAFT reconstruction the crack tip profile can be detected more precisely, Figure 10b. The error in defect sizing measured ultrasonically with SAFT reconstruction was about 2.3% based on the metallography. In the NESC cylinder there was also one multi-branch defect, which was not included in the comparison of the results of the round robin test. This defect was a large defect of similar size as the two defects mentioned above.

Figure 8 NESC spinning cylinder containing different types of defects. The cylinder was tested in a round robin trial

9

a) b) Figure 9 Defect B measured in the pre-test phase with SAFT (a) and in the post-test phase with 0°L 5MHz probe (b). At the left (a) in Figure the crack characteristics and at the right (b) the crack tip profile are shown

Crack tip profile shadow

corner

Only highest spot of the crack can be seen

Shadow of a crack profile

Crack tip profile in thickness direction

Crack tip

Corner

Cladding

Figure 10 Defect R measured with normal multi-channel equipment and with one-channel SAFT equipment

The sizing capability of SAFT inspection

SAFT offers good sizing capabilities if the limitations are not exceeded. A large experimental study was carried out by Müller [28] and according to these results the 6dB method shows limitations especially in case of point reflectors, because the measured size is dependent on the reflector size and the used frequency. Figure 11 gives a good base for understanding the SAFT sizing in practice. Most of the sizing results have been obtained by the diffraction method. With SAFT it is often possible to size the defects more precisely as compared to the conventional ultrasonic testing, but it is not the only correct method. In general, the best result is based on the estimates achieved in combination by different methods. Figure 11 shows data from some inspections, data from Erhard [33] and data from the NESC I trial. In each case the metallurgical result has been compared to sizing result of SAFT measurements. It can be seen that the height sizing error of SAFT measurements is normally less than 2 mm from the real value. In three cases the error is equal to or more than 3 mm. Of course, the values tend to increase in percentages, when the sizes of defects are smaller, but the absolute value seems to be about the same. These values are only giving a tendency noticed in sizing performed with SAFT method. The defect sizing standards or nuclear standards describes sparsely or not at all the analysis of SAFT images [45, 46, 47 and 48]. Procedures should guide the inspector in order to do the analysis in proper manner. Sometimes the procedures are deficient [47] and people are missing to carry out SAFT measurements in practice internationally. One of the main reasons, why SAFT method is not popular is the low speed of the analysis. The high-speed inspections need high-speed analysis. The diffraction technique manually or with mechanised systems provides also sufficiently low sizing error compared to SAFT, but not as good and clear imaging as SAFT offers. TOFD is not as good for austenitic materials as for ferritic materials. Both methods have always problems when the defect detection suffers from the compressive stresses. Some teams are qualified with sizing capability of RMS

10

about 1 mm using TOFD technique. By measuring worst cases defects this level of competence is difficult to keep in this RMS limits. From Figure 11 we can see that the sizing capability of SAFT is about 2 mm using diffraction technique. ENIQ recommendations for sizing suggest that the RMS value should be about 3 mm for the blind test and similarly value is given also in the performance demonstration part of ASME XI. This can be achieved with sufficient knowledge and experience of the SAFT method. While the standards does not give proper procedure for the SAFT image analysis, the blind test according to the ENIQ and ASME principles gives simple way to test SAFT technique. SAFT has been used as a finer analysis method compared to conventional angle probe methods. One has to remember that sufficient training makes analysing easier. But the training defects as well easy blind test defects should include also difficult defects as can be expected in the practical inspection situations depending on the real defect types in practice. Dynamic analysis improves the defect detection, sizing and characterisation capability.

SAFT inaccuracy of depth sizing in various tests

-5,0

-4,0

-3,0

-2,0

-1,0

0,0

1,0

2,0

3,0

4,0

5,0

0,0 20,0 40,0 60,0 80,0 100,0

Defect depth [mm]

Sizi

ng E

rror

[mm

]

Figure 11 Defect heigth sizing error using SAFT measurement with different defect sizes

Summary and conclusion

SAFT reconstruction can be used in many applications in nuclear power plants to evaluate the defects detected in components. SAFT measurement can, especially, be applied to inspections where sound paths in ultrasonic testing are long. This includes especially pressure vessel, thick-walled piping, main gate valves, etc. SAFT method is mainly intended to be used for characterisation and sizing of defects.

The accuracy is depending on large amount of factors. The physical resolution of the method is decisive but also coupling, ultrasonic probe, sizing technique, human factors affects the result. The analysis of the SAFT image is not properly determined by the standards. Procedures should determine the way exactly how to do analysing. At the moment the analysis is based more or less on skills of the operator. Especially dynamic analysis of the SAFT images can provide better tools for sizing in the uttermost cases when the crack is under compressive stresses. Normally the analysis in these cases is carried out near noise level, sometimes as low as the noise level. The dynamic behaviour of the

11

reflected, diffracted and scattered sound waves can be distinguished better then with the static analysis methods.

References

[1] Doctor, S., Schuster, G. J., Reid, L. D. & Hall, T. E., 1995. Real-Time 3-D SAFT-UT System Evaluation and Validation. NUREG/CR-6344 PNNL-10571, 169 p + App. 78 p.

[2] Kröning, M. & Schmitz, V., 1997. Contributions of NDT, Risk Based Assessment of Industrial Components and Plants. 15th –17th December, Madras, Chennai, 31 p.

[3] Müller, W., Schmitz, V. & Schäfer, G., 1985. The Synthetic Aperture Focussing Technique and its Application in NDE - Principles – Signal Processing and Measurement. Philadelphia, USA September 25 p.

[4] Osetrov, A. V., 1995, Complex Flaws Reconstruction by Means of Second Order Synthetic Aperture Focusing Technique. Proceedings International Conference on Computer Methods and Inverse Problems in Nondestructive Testing and Diagnostics, CM NDT-95, 21st - 24th November, Minsk, Belarus, pp. 116 - 119.

[5] Schmitz, V., Müller, W. & Schäfer, G., 1988. Ultraschallprüfung mit synthetischen Aperture-verfahren (SAFT). Symposium für Werkstoffprüfung, Greifswald, Germany, Juni, 17 p.

[6] Gruber, G., 1980. Defect Identification and Sizing by the Ultrasonic Satellite-Pulse Technique. Deutsche Gesellschaft für zerstörungsfreie Prüfung, Jahrestagung, Göttingen, 12th -14th May 8 p.

[7] Silk M. G., 1987. Changes in Ultrasonic Defect Location and Sizing. NDT International, 20(1987)1 pp. 9-14.

[8] Erhard, A.1983. Untersuchungen zur Ausbreitung von Longitudinalwellen an Oberflächen bei der Materialprüfung mit Ultraschall. Forschungsbericht 88, BAM, Berlin, 32 p.

[9] Ermolov, I. N., Razygraev, N. P. & Scherbinski, V. G., 1978. The Use of the Head Type Waves for Ultrasonic Monitoring. Sov. Journal.of NDT, 14(1978)1 pp. 27-33.

[10] Wüstenberg, H., Haufe, U. & Erhard, A., 1984. Contribution to the Ultrasonic Detection of Inner Wall Surface Breaking Cracks in Pipes. Materials Evaluation, August, 42(1984)9 pp. 1142-1149.

[11] Mohr, F., 1991. KWU-HOLO. Technical report, Siemens, 17th May, 16 p. [12] Grebennikov, V., Badalyan, V., Grebennikov D. & Vopilkine, A., 2000. Comparative Analysis of

the Ways to Increase Signal to Noise Ratio at NDT Inspection of Austenitic Welds. 15th World Conference on Non-Destructive Testing, Rome, Italy, 15th - 21st October, 9 p.

[13] Wüstenberg, H., Erhard, A., Schenk, G. & Montag, H.-J., 1987. Anwendung der Ultraschall-tomographie an Turbinen und Generatorwellen. Materialprüfung, 29(1987)10 pp. 297-302.

[14] Rathore, S. K., Kishore, N. N., Munshi, P.& Singhal, S. K., 2000. Ultrasonic Tomography of Closed Spaced Defects. 15th World Conference on Non -Destructive Testing, Rome, Italy, 15th - 21st October 2000, 7 p.

[15] Munshi, P., 2003. X-ray and Ultrasonic Tomography. Insight 45(2003)1 pp. 47-50. [16] Komura, I., Hirasawa, T., Nagai, S., Takabayashi, J.-I. & Naruse, K., 2001. Crack Detection and

Sizing Technique by Ultrasonic and Electromagnetic Methods. Nuclear Engineering and Design, June, 206(2001)6 pp. 351-362.

[17] Erhard, A., Schenk G., Hauser, T., Völz, U., 2001. New Applications Using Phased Array Techniques. Nuclear engineering and Design, June, 206(2001)6 pp. 325-336.

[18] Matsuura, 2002. Study of Phased Array UT -Technique to Nuclear Power Plant. 6th EPRI Piping & Bolting Inspection Conference, Point Clear, Alabama, USA, 30th July- 1st August 21 p.

[19] Wüstenberg H., Erhard, A. & Schenk, G., 2000. Scanning Modes at the Application of Ultrasonic Phased Array Inspection Systems. 15th World Conference on Non-Destructive Testing, Rome, Italy, 15th - 21st October, 14 p.

[20] Selby, G., 2002. Phased Array Development Update. 12th EPRI NDE Issues Meeting, Kingston Plantation, Myrtle Beach, USA, 10th July, 78 p.

[21] Spies, M., Gebhardt, W. & Kröning, M., 1998. Neue Entwicklungen in Gruppenstrahler Technologie. Deutsche Gesellschaft für zerstörungsfreie Prüfung, Jahrestagung, Bamberg, Germany, 7th – 9th September, pp. 73 - 80.

[22] Bulaninov, A., 2005. Der getakte Gruppenstrahler. Dissertation des grades des doktors der Ingeniurwissenschaften der Naturwissenschaflich-Technischen Fakultät III Chemie, Pharmazie, Bio- und Werkstoffwissenschaften der Universität des Saarlandes, 11. August, 128 p.

[23] Kitahara M., Nakahata, K. & Hirose, H., 2000. Elastodynamic Inversion for Shape Reconstruction and Type Classification of Flaws. Wave Motion 36(2000)4 pp. 443 - 455.

12

[24] Stepiniski, T. & Lingvall, F., 2000. Automatic Defect Characterisation in Ultrasonic NDT. 15th World Conference on Non-Destructive Testing, Rome, Italy, 15th - 21st October, 9 p.

[25] Park, I. K., Park, U. S., Ahn, H. K., Kwun, S. I. & Byeon, J. W. 2000. Experimental Wavelet Analysis and Applications to Ultrasonic Evaluation, 15th World Conference on Non-Destructive Testing, Rome Italy 15th - 21st October, 9 p.

[26] Shlivinski, A., Langenberg, K. J., Marklein, R., Schmitz, V., Müller, W. & Mletzko, U., 2004. Modelling of Elastic Wave Propagation and Imaging of Defects in Dissimilar Welds. NDE in Relation to Structural Integrity for Nuclear and Pressurised Components, 6th - 8th December, London UK, pp. 900 – 909.

[27] Thompson, R. N., 1984. Transverse and Longitudinal Resolution of the Synthetic Aperture Focusing Technique. Ultrasonics, 22(1984)1 pp. 9-15.

[28] Müller, W. & Walte F., 1997. Auflösunfgsvermögen bei SAFT -Rekonstruktionen, ebene Aperturen, IzfP-FHG, Bericht Nr. 970104-TW, October, 115 p.

[29] Pitkänen, J., Kauppinen, P., Jeskanen, H. & Särkiniemi, P. 1998. Evaluation of Ultrasonic Indications by Using Synthetic Aperture Focusing Technique. RER/4/011 Regional Workshop on Flaw Discrimination and Sizing Methods and Techniques, IAEA Workshop, 20th - 24th April, Madrid, Spain, 7 p.

[30] Kauppinen, P., Jeskanen, H. & Schmitz, V. 1993. Analyse von Primärkreisventilen mittels PCSAFT. DGZFP-Jahrestagung, Garmisch Partenkirchen, 17th – 19th May, Berlin: Deutsche Gesellschaft für zerstörungsfreie Prüfung. pp. 659 – 663.

[31] Pitkänen, J., Särkiniemi, P., Kauppinen, P. & Jeskanen H. 1998. Ultrasonic Measurement of a Thermal Fatigue Crack Field Manually and with a Simple Hand Scanner. 7Th ECNDT European Conference on Non-Destructive Testing, Copenhagen, 26th - 29th May, pp. 746 - 750.

[32] Pitkänen, J. & Kauppinen P., 2001. Experiences from SAFT in Nuclear Applications. Proceedings of the Third International Conference on NDE in Relation for Nuclear and Pressurised Components, 14th-16th November, Seville, Spain, pp. 918-928.

[33] Särkiniemi, P. & Kauppinen P., 1994. The Use of Combined Ultrasonic Techniques for Examination of Under Cladding Flaws from Outside. Report STUK-YTO-TR, Helsinki, STUK. 24 p. (in Finnish)

[34] Kenefick, S. & Henry, G. 2001. Ultrasonic Imaging Technique for Characterizing Surface Breaking Flaws. Proceedings of the Third International Conference on NDE in Relation for Nuclear and Pressurised Components, 14th -16th, Seville, Spain, pp. 552-560.

[35] Schmitz, V. & Chakhlov, 2001. Fehleranalyse durch Ultraschallrekonstruktion 3D-CAD-SAFT-Handbuch v. 1.8. 92 p.

[36] Erhard, A., Tscharntke, D., Montag, H.-J., Tessaro, U., Voskamp R. & Schmidbauer, J.,2004. SAFT and TOFD Used for Defect Sizing at Steam Generator Feed Water Nozzle. NDE in Relation to Structural Integrity for Nuclear and Pressurised Components, 6th – 8th December, London, UK, pp. 251 – 260.

[37] Müller, W. & Bruche, D., 1999. Erkannbarkeit von Risspitzensignalen bei Vorwärts- und Rückstreuung – ein Vergleich. IZFP-FHG Bericht Nr. 990112-TW, March, 140 p.

[38] Pitkänen, J., Särkiniemi P., Kauppinen P. & Jeskanen, H., 1998. The Ultrasonic Detection and Evaluation of Defects in a Thick-Walled Cylindrical Component Having Cladding on Inside Surface. 7Th ECNDT European Conference on Non-Destructive Testing, Copenhagen, 26th – 29th May, pp. 1420 -1425.

[44] Erhard, A., Schulz, E., Brekow, G., Wüstenberg H. & Kreier. P., 1998. Critical Assessment to the TODF Approach for Ultrasonic Weld Inspection. 7th EC-NDT European Conference on Non-Destructive Testing, Copenhagen, 26th – 29th May, pp. 1236 - 1242.

[45] ASME V, 2004. Examination E470. Non-mandatory Appendix E, Edition 2004, pp. 69-73. [46] ASTM, 2003. E2192-02: Standard Guide for Planar Height Sizing by Ultrasonics. ASTM, 22 p. [47] Procedure IVO 511 M4, 1994. Defect Height Sizing Measurement with Ultrasonic Testing. 27 p. (in

Finnish) [48] KTA 3210-4, 1999. Components of Reactor Coolant Pressure Boundary of Light Water Reactors,

Part 4: In-Service Inspections and Operational Monitoring, pp. 19. [49] EN-583-5, 2000. Nondestructive testing – Ultrasonic examination – Part 5. Characterisation and

Sizing of the Discontinuities. CEN, 57 p.

13