First International Conference on NDE in Relation to Structural Integrity for Nuclear and Pressurised Components, 20 - 22 October 1998, Amsterdam, Netherlands

TABLE OF CONTENTS

Abstract Introduction Advantages and Limitations of EMATS Evaluation of EMAT SH-Waves on cast stainless steel piping Evaluation of EMAT SH-Waves on austenitic pipe specimens Initial EMAT investigations Performance demonstration of EMAT raster scan procedure Expansion of EMAT procedure to include line/sector scan capabilities Results of EMAT investigations of field-removed IGSCC samples Potential for additional SH-EMAT evaluations Phased array Examination of seam-welded high energy piping Summary References

Abstract

Improvements are currently being made in nondestructive evaluation (NDE) equipment which have advanced the state of the art of piping examinations. Phased array ultrasonic examination technology has an increasing potential for utility applications with the availability of high-performance piezocomposite array probes and more affordable industrial array control systems. The demand for accuracy and efficiency of inspection techniques can be met by inspection techniques that use ultrasonic phased arrays where a single array probe can be controlled electronically to generate ultrasonic beams of many different angles and to focus at many different points in the inspected component. Likewise, electromagnetic acoustic transducer (EMAT) systems are becoming less expensive and have the potential to save examination costs while improving flaw discrimination capabilities. EMAT systems are being developed which produce horizontally polarized shear (SH) waves which are less affected by austenitic weld metal than are conventionally used vertically polarized shear waves. EMAT-produced SH waves have been used to successfully detect intergranular stress corrosion cracking (IGSCC) through austenitic piping welds and with fewer scan lines than are conventionally used, allowing single-sided examinations and reducing examination time. Lessons have also been learned concerning ultrasonic examination of dissimilar metal welds and welds in cast stainless steel. These difficult examination challenges require examiner familiarity with the materials and crack characteristics which have been observed in similar configurations.

Introduction

Fig 1: Automated Phased Array System Used in the Application of Shear Horizontal (SH) Waves to Austenitic Piping Welds

For many years, it has been well documented that difficulties exist with the examination of austenitic weldments using conventional ultrasonic techniques. Elastic anisotropy and the coarse grain structure of the weld tend to be the most likely reasons for these difficulties (1). To assist in improving the performance of ultrasonic examinations, much effort has been spent on developing and evaluating the application of horizontally polarized shear (SH) waves. Because SH-waves cannot be effectively applied using piezoelectric transducers, an electromagnetic acoustic transducer (EMAT) is required. Using an EMAT, SH-waves can be generated without the use of a couplant.

The new EMAT designs generate an acoustic wave directly in the component rather than in the transducer, as is the case for piezoelectric transducers. The phased array design of these probes allows the sound beam to be formed and steered electronically in a wide range of incidence angles of SH-waves. Figure 1 illustrates the application of the phased array EMAT SH-wave technology at the EPRI NDE Center for austenitic piping welds.

Advantages and Limitations of EMATS

When used for examination of austenitic piping components, SH-waves offer some fundamental advantages and disadvantages compared to longitudinal waves and conventional vertically polarized shear waves. The advantages include:

• SH-waves suffer no mode conversion,
• SH-waves are relatively unaffected by the grain structures of austenitic weldments,
• SH-waves at grazing incidence should provide good discrimination capabilities, and
• By taking advantage of phased array capabilities, SH-wave examinations might not require two dimensional scan patterns.

Limitations of SH-waves include:

• SH-waves are difficult to generate,
• High cost and limited availability of EMAT equipment and search units,
• EMAT SH-wave technique is not applicable for axial defects, and
• The relative inefficiency of the EMAT often yields lower signal-to-noise ratios than can be achieved with conventional probes.

Evaluation of EMAT SH-Waves on cast stainless steel piping

The reactor coolant loop piping in Westinghouse pressurized water reactors contain cast stainless steel (CSS) material. Regulatory requirements mandate that the welds be volumetrically examined to assess weld condition. The primary method of meeting this examination requirement is ultrasonic examination. To perform the ultrasonic examination, it typically requires that ultrasonic energy be propagated through the CSS material. Because the grain structure of CSS varies in orientation, size, and shape, it is difficult to achieve predictable ultrasonic examination results.

Since one of the fundamental advantages of SH-waves is their ability to be relatively unaffected by the grain structure of austenitic weldments, efforts were made at the NDE Center to evaluate the application of EMATs for examining CSS piping. Center staff used the phased array EMAT SH-wave probes to collect data on several Westinghouse Owners Group (WOG) wrought, statically cast, and centrifugally cast pipe samples. Figure 2 shows one of the WOG samples (FPE-02) that was examined using the EMAT SH-wave probe.

Fig 2: WOG Specimen No. FPE-02

This WOG specimen represents a 10-inch x 20.5-inch segment of a 27.5-inch diameter forged pipe-to-cast elbow joint mockup. The mockup contains a mechanical fatigue crack (approximately 2.2 inches long) on the elbow side. Using the phased array EMAT SH-wave probe, the mechanical fatigue crack was easily detectable through the weld joint and the crack length was measured to be approximately 1.8 inches. However, when the data was evaluated from the statically cast elbow side, the mechanical fatigue crack could not be discriminated from ultrasonic backscatter reflections.

Based on the results from the WOG samples, the phased array EMAT SH-wave inspection of coarse grained, anisotropic, statically cast samples offers no improvement over conventional longitudinal wave examinations (2). In fact, conventional ultrasonic techniques are preferable for these coarse grained components, due to the relative inefficiency of the phased array EMAT SH-wave.

Evaluation of EMAT SH-Waves on austenitic pipe specimens

The ultrasonic examination of austenitic welds is strongly complicated by a high degree of elastic anisotropy and the coarse grain structure of the weld material. Also, in some situations, the welds are accessible only from one side. For this limitation, it becomes extremely important that the chosen examination method be capable of inspecting the whole weld from a single side as illustrated in Figure 3. Since SH-waves have been shown to be less affected by the aforementioned effects than other wave modes, NDE Center staff have performed extensive investigations using a 750 kHz phased array EMAT SH-wave probe to examine a range of austenitic pipe specimens containing known flaws.

Fig 3: Definition of Single-Sided Access Pipe Weld Examination

Initial EMAT investigations
To test the feasibility of using phased array EMAT SH-wave probes for austenitic pipe welds, a selection of piping specimens containing a range of flaw sizes were identified. Using the specimens, an examination procedure was developed for single-sided raster scan examinations using phased array EMAT SH-wave probes. The primary objectives of the procedure were to detect and measure the length of known flaws while producing a minimum number of false calls (incorrect evaluation of unflawed regions). Initial EMAT data showed that the phased array EMAT SH-wave successfully eliminates confusing geometric and irrelevant signals. It is also apparent that the phased array EMAT SH-wave provides excellent length sizing accuracy of known flaws (3). Because the flaw end points are clearly identified with regards to accompanying geometry, flaw length is distinctly evident.

Performance demonstration of EMAT raster scan procedure
In order to test the phased array EMAT SH-wave raster scan procedure, a demonstration was performed on a set of austenitic piping samples without IGSCC, with single-sided access, using data that had been acquired earlier by someone other than the data analyst. Although the technical requirements of ASME Section XI, Appendix VIII and the Performance Demonstration Initiative (PDI) were met, the necessary administrative steps were not taken to permit the demonstration to be considered official through PDI. However, the test set did contain well over 20 flaws with more than 50% of them located on the far side of the weld to ensure that an adequate test was performed on the capability of the single-sided access procedure. With minimum (~ 1/2 day) exposure to EMAT SH-wave data, a technician completed the data analysis in less than two days. The results indicated that 100% of the flaws were detected and only one area was falsely called cracked (3). The results also satisfied the PDI length-sizing criterion with an RMS error of 0.5 inch. Until this demonstration, no one had ever met this criterion with single-sided access. In addition, the Performance Demonstration Administrator observed that "the EMAT SH-wave data provides an image which is considerably more definitive than results obtained from conventional techniques."

Expansion of EMAT procedure to include line/sector scan capabilities
After a successful single-sided performance demonstration of phased array EMAT SH-waves on austenitic piping samples without IGSCC with access using raster scans, NDE Center staff conducted additional investigations of EMATs using line/sector scans. As shown in Figure 4, this technique is accomplished by selecting a small number of fixed axial locations and scanning parallel to the weld. This technique was developed with the intent of performing double- or single-sided examinations. The objectives were to be able to reduce scan time by a factor of two while maintaining excellent flaw detection and length sizing capabilities.

Fig 4: Example of Phased Array EMAT SH-Wave Line/Sector Scan

Investigations were performed using a range of flawed specimens containing a selection of implanted and mechanical fatigue cracks. The resultant procedure utilized a number of line scans performed at fixed positions while the phased array EMAT investigated the weld region using a range of beam angles from approximately 40 to 90 degrees. Again, the resultant data showed that confusing geometric and irrelevant signals were virtually eliminated; reducing the chance of false calls or missed detections (3).

As part of this evaluation, length-sizing data was acquired from 28 flaws using the phased array EMAT SH-wave line/sector scan technique. The RMS error for the measured versus true flaw lengths for the case of the flaws on the same side of the weld as the probe was determined to be 0.19 inch. The RMS error for the measured versus true flaw lengths for the case of the flaws on the opposite side of the weld as the probe was determined to be 0.13 inch. As indicated by this data, the flaw length is very distinct using the phased array EMAT SH-wave line/sector scan data (3).

Results of EMAT investigations of field-removed IGSCC samples
Since the beginning of initial investigations into the application of phased array EMAT examinations to austenitic piping, NDE Center staff wanted to evaluate the performance of EMATs on field removed IGSCC samples. However, to support these investigations, Center staff was actively pursuing the acquisition of a higher frequency EMAT. The higher frequency EMAT was considered necessary since IGSCC has been shown to be frequency sensitive. When acquired, plans were to also evaluate the higher frequency EMAT for its through-wall sizing capabilities. However, despite extensive efforts, the NDE Center has not been able to acquire an acceptable higher frequency EMAT.

So, based on the promising performance of the phased array EMAT SH-wave line/sector scan procedure that uses the lower frequency (750 kHz) EMAT, a decision was made to evaluate the procedure's performance on field removed IGSCC samples. In 1998, a selection of PDI practice samples containing IGSCC were used to optimize the phased array EMAT SH-wave line/sector scan procedure. Although, a reduced signal-to-noise response was obvious, NDE Center staff were optimistic after initial investigations were performed. Based on this optimism, the investigators assembled a blind practice set of samples to more fairly assess the performance of EMATs for IGSCC.

The practice set of samples consisted of 6 field-removed IGSCC specimens that represented a total of approximately 111 inches of austenitic pipe weld. Available truth information for the practice samples showed that approximately 24% of the total weld length actually contained IGSCC. For data acquisition, the phased array EMAT SH-wave line/sector scan procedure was used to examine each sample from both sides of the weld. The quality of the acquired data was "typical" and considered adequate for use in producing the desired images necessary to assist with data analysis.

The phased array EMAT SH-wave data from the samples was evaluated by one of the investigators not involved with data collection. The results of this independent analysis showed that when using information from both sides of the weld, approximately 86% of the flawed area was correctly detected. However, when looking through the weld, only 38% of the flawed area was detected. Post test analyses showed that this reduced performance was primarily due to some significant mismatch conditions and a lower signal-to-noise ratio through the weld. In addition, the lower signal to noise response resulted in a higher incidence rate of "false calls" related to not being able to discriminate unflawed regions of the weld. More investigations are planned when a new EMAT can be acquired (4).

Potential for additional SH-EMAT evaluations
Based on the results obtained during the phased array EMAT SH-wave evaluations using actual IGSCC, efforts will continue to acquire a higher frequency EMAT or an EMAT with the ability to operate with higher efficiency on austenitic material. In addition, SH-waves may also be applicable and bring new approaches to weld overlay examinations, dissimilar metal weld examinations, erosion/corrosion examinations, and examination of weld-deposited cladding. With improvements in EMAT search units, future efforts should include investigations into the use of EMAT SH-wave examinations using time-of-flight diffraction techniques for through-wall sizing of cracks (4).

Phased array Examination of seam-welded high energy piping

Longitudinal seam welds in high energy components (piping and headers) of fossil-fueled power plants continue to present significant safety and reliability concerns due to initiation and propagation of creep damage. Although the damage mechanism is not fully understood, extensive studies of field-removed, hot reheat piping sections indicate a gradual-failure scenario in which damage begins almost exclusively near the midwall weld cusp and grows gradually as cracking occurs along the weld fusion line (Figure 5).

Fig 5: Typical Creep Damage Associated With Hot Reheat Piping

To ensure the integrity of the seam-welded piping systems, it is common practice to implement an inspection-based approach that relies on ultrasonic examination techniques to detect evidence of creep damage before major cracking, leaking, or rupture can occur. When conventional ultrasonic examination techniques are used, NDE procedures typically require that each weld be examined using multiple beam angles at increased levels of sensitivity. For each beam angle, the search unit is moved across the component surface in a time-consuming, two-dimensional scan pattern. In addition, supplementary flaw sizing techniques must also be used when indications are detected.

Phased array technology offers a means to reduce the scanning time by simplifying the scan pattern and improving the process used to analyze the resultant data. Because the array transducer is segmented into many individual, parallel elements; the timing or "phase" of each element may be controlled to simulate many different conventional probes. Without moving the probe, sound beams of many angles can be generated sequentially, inspecting a large portion of the component's cross-section (Figure 6).

Fig 6: Phased Array Steering of the Ultrasonic Beam

To evaluate the feasibility of using phased array techniques for examining seam-welded piping, a selection of seam-welded pipe specimens containing a range of flaw sizes and types were examined. Using the specimens, an examination procedure was developed and evaluated. The procedure involves selecting a small number of fixed axial locations and scanning parallel to the weld. By using this scanning technique with only one phased array probe, examination time is significantly reduced and the results indicate excellent flaw detection and sizing capabilities. In addition, because the data can be imaged, the data analysis process is simplified.

Summary

The capability of the phased array EMAT SH-wave technology has been evaluated for application to austenitic piping welds. Results of the evaluation were very promising and show that SH-waves are less affected by the grain structure of austenitic weldments. The capability of SH-wave examination procedures to meet the ASME Section XI Appendix VIII performance demonstration requirements has been demonstrated and showed that detection of flaws on the far side of the weld is possible with the EMAT technique. Flaw detection and length sizing results show considerable improvements over conventional ultrasonic techniques. Efforts to use SH-waves to perform reliable examinations of CSS components and examinations to detect IGSCC in austenitic pipe welds were less promising. However, efforts to obtain a higher frequency EMAT or an EMAT which generates acoustical energy more efficiently will continue at the NDE Center.

References

1. Salzburger, H.-J., G. Hübschen, M. Kroning, "Electromagnetic Ultrasonic (EMUS) Probes: State of the Art and Developments for Applications in Nuclear Power Plants." Proceedings of the 12th International Conference on NDE in the Nuclear and Pressure Vessel Industries, Philadelphia, PA (October 1993). Materials Park: ASM International, 1994, pp. 137-142.
2. MacDonald, D. E., "Ultrasonic Examination of the Westinghouse Owners Group Pipe Specimens using Horizontally Polarized Shear Waves." Letter Report, EPRI NDE Center. December 1995.
3. MacDonald, D. E. et al., "Demonstrating Feasibility of Electromagnetic Acoustic Transducers (EMATs) for Austenitic Weld Examination." EPRI Report MI-108240, EPRI NDE Center. December 1997.
4. Landrum, J. et al., "Evaluation of Electromagnetic Acoustic Transducers for Examination of Austenitic Piping Welds." Proceedings of the EPRI Phased Array Seminar, Portland ME. September 8-9, 1998.