Abstract

The article introduces results of a program for validation of different ultrasonic test techniques for crack depth determination of surface breaking cracks. Two techniques, the TOFD technique and the angle beam testing with corner effect, both using time of flight measurement of the diffracted signal from the crack tip, are employed. Since it is not certain that all crack edges generate diffracted waves, the optimized ultrasonic pulse echo technique should be used. Applying TOFD as an additional analytic method for verification of Impulse Echo Technique or Radiography results seems to be more applicable.

Table of contents

• Introduction
• Ultrasonic testing for defect detection
• The TOFD-(Time of flight diffraction-)Technique
• Crack depth determination with Impulse-Echo-Technique
• Conclusion
• Literature

Introduction

The first intensive programs focusing on detection of cracks in austenitic piping of nuclear power plants began in 1992. Defects studied include cracks in the weld section; lying circumferentially and more or less perpendicular to the inner surface. Analysis showed stress corrosion cracking to be the cause. Nonuniformities in the performance of weld geometry, such as misalignment and weld root penetration, make clear ultrasonic or radiography testing difficult.

Ultrasonic diffraction technique was also employed as a test method. In the USA this method was first applied for crack detection and sizing (depth, length) in austenitic pipe welds. The tested welds were investigated destructively and the ultrasonic results were compared with those obtained via metalography. The inspection method and its parameters were validated.

Ultrasonic testing for defect detection

In Germany ultrasonic testing is usually applied for in-service inspection of plants by use of pulse echo techniques, such as 45° shear wave probes with a corner effect or 70° SEL (creeping waves) probes. All of these techniques use reflection which relies as much as possible on a perpendicular position of the crack surface to the probe index.

UT with the 70°SEL technique was qualified for testing of austenitic piping with 5mm to 15mm wall thicknesses. This uses the mode conversion effect on transversal waves at the far surface [1], [2]. By use of longitudinal waves, which penetrate well in austenitic materials, and applying with a large (70°) angle of incidence , a signal from the reflector surface (main echo) can be achieved . Rood defects of less depth are distinguished from defects with greater depth by the absence of neighboring echoes (triangle reflection in fig 1a.). That is performed as a detection technique. The height of the main echo cannot be used to determine the crack depth. Fig 1b. shows the so-called notch diagram of the 70° SEL small probe (KWU) (main and neighboring echoes). The different test notches and the edges of the reference block display an echo amplitude similar to the main echo. A clear echo resolution exists for determination of the registration level (echo amplitude of the 1.5 mm notch minus 6 dB). Those echo signals which exceed the registration level and the typical neighboring echoes are classified as defects. The echo amplitude of the neighboring echoes (triangular reflection) increases with crack depth. The evaluation using the displayed echo signal pattern is only qualitative by means of a yes/no decision, since the echo amplitude changes significantly with the weld root geometry.

The TOFD-(Time of flight diffraction-) Technique

In 1977 Silk developed the TOFD technique which use diffracted signals and also introduced the Zipscan equipment to the market [3,4,5]. The principle is shown in fig.2. Two probes in a transmitter-receiver arrangement are positioned in through transmission, each on both sides of the weld. Longitudinal probes are applied with an angle of incidence range of 45° to 70°. Fig. 2a shows the main sound beam propagation (back wall echo), the lateral wave on the surface and the diffracted waves from the upper and lower crack tip which appear with different times of flight (fig 2b). By comparing the time of flight values of these signals with known component geometry and probe distance, the depth of the crack can be determined. The exact location of the defect can not be calculated. The literature mentioned the best angle of incidence (longitudinal waves) of 65° [3]. Under easy geometry a longitudinal probe scan normal to the weld is efficient enough, coupled on a uniform surface. By 1980, the Zipscan equipment already had the capability for digitalization of the complete A-scan as well as post processing as B-scan images with linear scales by use of the SAFT algorithm.

Fig. 2c shows an example of a crack open to the far surface (artificial notch), which is detectable by the diffraction echoes and the absence of the back wall echo. The crack depth can be depicted as a linear B-scan image. The echo amplitude is displayed as gray scale, usually zero amplitude light gray (negative maximum amplitude black, positive maximum amplitude white). An accuracy of approx. +/- 1 mm can be achieved for defects 10 mm under the surface [3]. The lateral wave hinders the evaluation in a depth of approx. 5mm.

The diffracted signals have a low amplitude ( approx. 20 dB less than a 3mm Ø -side drilled hole). A problem is that the test has to be performed with high gain, increasing chances that small inhomogeneities or slake will generate significant signals. Our task was to qualify TOFD as a detection technique on gladding components . The results showed that a gladding nonuniform surface interfered with the backwall echo and lateral wave; these could be evaluated as surface cracks or loss of coupling. That necessitated verification by use of the impulse echo technique. It is very difficult to distinguish significant signals in the area of noise (rough surface, conclusions). When TOFD is employed as a detection technique the scan of the total surface must be performed with clear and sufficient sensitivity to produce a reliable test result.

When applied on austenitic welds an insufficient signal to noise ratio is another problem. TOFD employed as an additional analytic method for verification of Impulse Echo Technique or Radiography results seems to be more applicable.

Crack depth determination with Impulse-Echo-Technique

For the determination of crack depth this technique uses the same diffracted signals as the TOFD technique. But here one probe is used as both transmitter and receiver, and the defect location can be determined.

To identify a diffracted echo a corner echo from the crack is needed. The corner echo is generated by use of the corner effect between far surface and crack surface. There are two techniques:

• "Relative time of flight technique" (RLT, fig 3a) , which measures the time of flight between the corner echo and the diffracted echo.
  
• "Absolute time of flight technique" (ALT, fig 3b), which measures only the time of flight until the diffracted echo.

Twin probes with 60° or 45° generating shear waves are applied. To make the crack detection more reliable for finding cracks in deeper areas as well, the diffracted echo is also evaluated after travelled as V-transmission; that means after it has been reflected on the far surface. A test with 60° longitudinal waves is subsequently performed. When welds of piping that can be only reached from one side of the weld are tested and the echo must be transmitted through the weld, only longitudinal probes are applied.

The crack depth can be easily calculated, thus measuring the time of flight when the peak amplitude of the diffracted signal is detected. There are also methods that measure when the signal registers half value or just disappears within the noise. That should prevent the error on measuring diffracted signals from a lower part of the crack instead from the tip. All signals higher than noise and echo dynamic and time of flight are evaluated as possible cracks. If a diffracted signal does not exist, but an echo signal from the far surface appears, the result is classified as a crack less than 1 mm.

In the 80's, this technique was investigated in the USA by the EPRI Institute for detection and sizing of cracks [6]. An American standard describes the necessity of using ultrasonic testing for crack depth measurement. The performance of the method used and the personnel qualifications must be demonstrated [7].

Our task was the inspection audit of austenitic piping which was tested by the service company General Electric. The test was performed mechanically by the Smart 2000 system; all A-scans can be stored. The evaluation is performed with post processing of all data after the test is finished. Even though the company had a lot of experience in our specific area (other component geometry, weld geometry) further study was necessary in order to achieve a reliable result. The results of the ALT Method on notches and nature cracks are displayed in fig 4. Our comments for these test results:

• the regression curve and the regression coefficient matches the ASME-criteria [7] excellently;
  
• the best measurement results can be achieved for crack depths in the range of 30% of wall thickness;
  
• small crack depth is calculated bigger, large cracks depth is calculated smaller. The crack in fig 5 was also calculated significantly smaller than the real value by use of the 6 dB method ;
  
• the number of investigated cracks are not enough for a final result . Results of practical existing cracks are still missing (corroded crack tips, stress).
  
• the method can determine the crack depth with an accuracy of 20% of the wall thickness.

Conclusion

From our point of view, the methods suitable for defect analysis are those which use diffracted ultrasonic signals with their time of flight measurement for crack depth determination. Therefore the adjusted gain should be verified against a reference block to notice a higher attenuation on the work piece and also to yield reliable results for in-service inspection.

Since it is not certain that all crack edges generate diffracted waves, the optimized ultrasonic pulse echo technique, (described above) should be used. That method uses reflected signals from the crack surface or corner. In case of a suspected defect other tests, e.g., radiography or ultrasonic analysis, need to be employed to obtain a test result which is based on all test methods.

Literature

1. J. Wessels Aktuelle Erkenntnisse über die Leistungsgrenzen der ZfP DGZfP-Seminar"Leistungsnachweis bei ZfP-Methoden DGZfP-Berichtsband 38, Seite 42-57
   
2. G. Csapo, T. Just, H. Eggers, E. Hein, R. Nimtz Fehlernachweisvennögen der Zerstörungsfreien Prüfverfahren an dünnwandigen austenitischen Schweißnähten 6th European Conf. on NDT, 24. bis 28.10.1994, Nizza, Seite 995-999
   
3. P. Carter Experience with the Time-of-flight Diffraction Technique and an accompanying Portable and Versatile Ultrasonic Digital Recording System Brit. J. Nondestr. Testing 26(84), 6, 354-361
   
4. H. Heckhäuser, K.-H. Gischler Das Zipscan-System der Ultraschallprüfung an plattierten Bauteilen und Rohrleitungen DGZFP Berichtsband 17, (1988) S. 122-132
   
5. M. G. Silk Benefits of signal processing in ultrasonic inspection Insight, Vol. 36, No. 101 October 1994, p.776-781
   
6. Electric Power Research Institute UT Operator Training for Planar Flaw Sizing EPRI NDE CENTER, Charlotte, N.C. (1985)
   
7. ASME Boiler and Pressure Vessel-Code, Section XI -Subsection IWA, IWA 3300 "Flaw Characterisation", -Appendix VIII "Perfomance Demonstration for Ultrasonic Examination Systems"