ML13220A274

Comparison of an Ultrasonic Phased Array Evaluation with Destructive Analysis of a Documented Leak Path in a Nozzle Removed from Service

ML13220A274
Person / Time
Site:	North Anna
Issue date:	04/30/2012
From:	Cinson A D, Crawford S L, Hanson B D, MacFarlan P J, Mathews R A Battelle Memorial Institute, Pacific Northwest National Laboratory
To:	Office of Nuclear Regulatory Research
References
FOIA/PA-2013-0139
Download: ML13220A274 (9)

Text

Comparison of an Ultrasonic Phased Array Evaluation with Destructive Analysis ofa Documented Leak Path in a Nozzle Removed from ServiceAD Cinson, SL Crawford, PJ MacFarlan, BD Hanson, and RA MathewsPacific Northwest National Laboratory',

Richland, WA, USAABSTRACTNon-destructive and destructive testing methods were employed to evaluate adocumented boric acid leakage path through an Alloy 600 , .penetration from the -k -, ,j.North Anna Unit 2 reactor pressure vessel head that was removed from service in 2002.A variety of focal laws were employed to evaluate the signal responses from theinterference fit region. These responses were compared to responses from aW- `N that was used to determine detection limits and ...characterization capabilities for wastage and boric acid presence in the interference fit 1*0region. Nozzle 63 was destructively examined after the completion of the ultrasonic nondestructive evaluation (NDE) to visually assess the leak paths. These destructive and =nondestructive results compared favorably.

INTRODUCTION The United States Nuclear Regulatory Commission (NRC) has funded the PacificNorthwest National Laboratory (PNNL) to assess the effectiveness and reliability ofadvanced NDE methods including the ultrasonic PA technique for the evaluation ofcontrol rod drive mechanism (CRDM) nozzles [1, 2]. Nickel-based alloy primarypressure boundary components 9 primary water stress_ --inmt ,P,5i corrosion cracking (PWSCC).

This degradation mechanism may pose a safety concerndue to the potential for reactor pressure boundary leaks, the associated potential for boricacid corrosion of low-alloy steels, and the development of flaws in the pipin and welds.Alloy 82 and .82 .Other --have since been developed, such as Alloy 690 and its associated weld metals, Alloy 52and 152, to improve the pressurized water reactor (PWR) components' resistance toPWSCC [3, 4].In the 2002 refueling outage at North Anna Unit 2, a significant number of J-groovewelds in the CRDMs had flaw indications and cracking.

It was estimated that in excess of40 of these welds would require repair; instead, the utility decided to replace the reactorhead. The Electric Power Research Institute (EPRI) sponsored the extraction of severalnozzles from the removed head to be used for research.

Previous ISI inspections ofNozzle 63 revealed a probable leak path and axial indications on the outside diameter ofthe nozzle tube, Surface examinations of the J-groove weld identified flaw indications.

[.was' ofdatp!restt. ol-s .' ,The work was sponsored by the U.S. Nuclear Regulatory Commission under U.S. Department ofEnergy Contract DE-AC05-76RL0 1830; NRC JCN Y6783; Dr. Greg Oberson, Program Monitor.

ULTRASONIC PHASED ARRAY PROBFq- ---ND-- -TUPA calibration mockup specimen and Nozzle 63 were examined with a pulse-echo (PE)longitudinal phased-array probe with a nominal center frequency of 5 MIHz operating inan immersion mode. The PA probe was designed in a i-D annular configuration usingeight elements.

The eight elements were arranged in a Fresnel radius pattern starting witha radius of 3 mm (0.12 in.) and ending with an element radius of 9.72 mm (0.38 in.).Thus, the total aperture was 296.81 mm2 (0.46 in.2). The probe exhibited a 72%bandwidth at -6 decibels (dB). This particular design was chosen for enhanced depthfocusing capabilities.

Its beam-forming potential was modeled to show excellent insonification of the interference fit region of interest.

Fjgur.e.

shows an example of ...simple ray tracing for the probe on the left and the sound field density mapping on theright for a 15-mm (0.59-in.)

focal depth in the nozzle material.

It should be noted that thissimulation is performed in isotropic material; the longitudinal acoustic velocity of soundis constant for any angle.Formatter',

Stblkathrough Deldwd Figurm IFigure I -Side views: ray trace on the left and sound field simulation on the right at adepth focus of 15 mm (0.59 in.).Data acquisition on the mockup specimen and Nozzle 63 was accomplished with acustom, slave-encoded scanner mounted on the nozzle. The scanner had two steppermotors that controlled circumferential and axial (vertical) motion. It was attached to thenozzle by sliding the scanner over the top of the specimen and equally tightening threeset screws to align the PA probe head in the center of the tube. The setup is shown inFigure 2 with the scanner on the calibration specimen.

The probe was mounted on theextension arm that was adjustable along the pipe axis allowing probe positioning in theregion of interest.

2 to nw~b top~E UntTS,5md Atcift Mo herense it orFigure 2 -Left: Scanner on calibration specimen.

Right: Inspection schematic.

The nozzle was filled with enough water to immerse the regions of interest.

Datawere acquired with the ZETEC Tomoscan IlI system running UltraVision 1.2R4software.

Phased-array data were acquired over a range of inspection depths from 1-15mm (0.039-0.59 in.) at a normal or 00 angle of inspection.

Raster data were acquired at0.25-degree increments in the scan axis (circumferential) and at 0.5-rmm (0.019-in.)

increments in the index axis (vertical or nozzle axial direction) on the calibration

specimen, and 0.5 degree in the scan and 0.5-mm (0.0 19-in.) increments in the index forNozzle 63.CRDM MOCKUP SPECIMENregion contained electro-discharging machined (EDM) notches separated by varioussurface distances and at varying depths. One set of notches was placed in the '8 bceand the other in the steel head material.

The other region contained boric acid regions in aknown pattern.

Data were acquired on the mockup specimen with the ultrasonic PA probe ,on both the notches and the boric acid regions.

Figure 3 is a top view or C-scan image ,that displays the resolution notches in the upper left portion of the image, circled in red.The variable depth notches on top and variable width notches on bottom are also seen andare circled in red. An image from the lower interference fit region containing boric aciddeposits is displayed in Figure 4( The boric acid regiojnrsare outlined in red. The full_.....

detailed analysis of the calibration specimen can be found in the proceedings of the 2011ASME Pressure Vessel and Piping Conference

[5). In summary, the presence of an airgap in the interference fit resulted in a strong ultrasonic reflection.

This can be seen inFigures 3 and 4 by the red and orange colors. The presence of a tight fit was shown in theregions containing boric acid and resulted in a weak ultrasonic reflection.

These twoareas are noted by the blue and white colors. The nominal interference fit regioncontained a mixture of mid-range reflections, noted by the green and yellow colors. Thusthe mockup interference fit specimens provided the expected range of conditions thoughtto be present in Nozzle 63, Furthermore, the ultrasonic evaluation validated thecapabilities of the system to distinguish between the different nozzle conditions.

.-~ ~7 KW~ I~li~m i~:~-'11ýk7 ý ý I ... ..M. 71Mt40SDeleted:

Fiagure 43 331 -180 to 360 deg. Circumferential Figure 3 -Top view, plan view or C-scan ultrasonic image of the interference fit regioncontaining calibration notches in the carbon steel head.I;7-039b 60 to 240 deg. Circumferential Figure 4 -C-scan plan view of the boric acid deposits in the interference fit region. Thehorizontal axis represents the circumferential range of 60-240 degrees.Pmeng.. 3 O1rJ~Confli you, bay.:'V,4-mn zutbtIhUrAn=dmak publbcaflh Standard ultrasonic evaluation techniques used by ISI vendors include time-of-flight diffraction (TOFD) for detecting cracks in both the circumferential and axial orientation and zero-degree PE for an interference fit examination.

An examination conducted by ISIvendor, WesDyne International (data supplied by John P. Lareau),

discovered a probableleak path in Nozzle 63 during the 2002 outage. The data acquired with an industry-standard 5.0-MHz probe are shown in Figure 5. The data set indicates a leak path at thedownhill

position, traveling upward through the interference fit and is marked by the twoblack arrows in the figure.4 3200t14010 * ;' "210 "90 ý. 0 * ""h:'Figure 5 -Ultrasonic Data from Nozzle 63 as Obtained by WesDyne International (courtesy of John P. Lareau).

The image was acquired with a 5-MHz probe. Thehorizontal axis represents the nozzle circumference in units of degrees.

The vertical axisrepresents the nozzle axis in units of millimeters.

The leak path is indicated by the blackarrows.NOZZLE 63 DATA AND ANALYSESUltrasonic data were acquired on Nozzle 63 in a radiologically controlled laboratory atPNNL. Destructive examinations (DE) were conducted kabcock and Wilcox --- 'Technical Services Group (B&W) .~ynqkb~urg Yrgn-4eI.olwntse__

two separate activities the data were compared and showed a high degree of correlation

"" -. M .between the NDE and DE findings, Ultrasonic dataScanning commenced on Nozzle 63 using setup and data acquisition methods similar tothose used for the calibration specimen.

High-resolution scans were conducted with0.5-degree resolution in the scan and 0.5-mm (0.020-in.)

resolution in the index axes overregions of-94 to 94 degrees and 86 to 274 degrees.

Scans extended in the axial direction a distance of 0 to 360 mm (0 to 14.2 in.). Each region was scanned twice to observerepeatability.

A composite 360-degree scan of the interference fit region is shown inFi ur _and shows the fit region of interest as well as the i-groove weld. The J:&groove

--.weld is identified as the white to light-blue oscillating band in the lower third of thefigure. High ultrasonic transmission (low reflection) is signified by the white and light-blue colors in the figure. The color-bar scale is seen in the left of the figure, Mary ---The investigation of the interference fit region on Nozzle 63 shows a large variety inthe reflected ultrasonic responses.

Some responses are strong in amplitude, indicating anair gap and noted as red-orange in color, while others are weak in reflected amplitude, indicating a tight interference fit and noted as white-blue.

A search in the data image for ahigh-amplitude river pattern connecting the weld to the top of the interference fit regionreveals the primary leak path. This leak path originates at approximately the 180-degree location (downhill point on the nozzle) and is marked in Figure 6. The lower amplitude tCDhoft d: Figwe 65 Figure 6 -A composite 360-degree ultrasonic data image from Nozzle 63. The verticalaxis represents 360 mm (14.17 in.)(blue) cluster regions near the leak path in the interference fit indicate the potential presence of boric acid as was observed in the calibration specimen.

Boric acid depositshave been noted in actively leaking nozzles [1, 2].Nozzle 63 destructive analysisConfirmatory destructive testing was performed on Nozzle 63, R-'A_&W,_

to validate the_.-_ý .a~jq

• k _1ultrasonic characterizations of the leak path(s) and other areas of interest as described above. This activity required dismantling the interference fit region with full separation of the Alloy 600 tube from the reactor vessel head material to reveal true-state information regarding the leak path(s),

boric acid deposits, and wastage regions.Ultrasonic data correlation to destructive analysisThe exposed reactor pressure vessel (RPV) head was photographed and the composite image is displayed in Figure 7. The main leak path (yellow arrows) and other featuresobserved in the ultrasonic images are clearly evident.

Boric acid deposits are visible inwhite and corrosion products in the rust color. The interference fit band is apparent in thephotograph and is marked with the red line. For comparison to the ultrasonic data, referto Figure 6. The ultrasonic features well match the features seen visually on the RPVhead annulus.

Clearly, the main leak path was precisely imaged and other partial leakpaths are evident.RPV boric acid and corrosion analysis with comparison to ultrasonic dataAttempts were made to measure the thickness of the combined boric acid and corrosion layer deposits in the annulus as well as the extent of corrosion of the RPV head. Thesemeasurements were then compared to the ultrasonic data. The layer thicknesses were first6 Figure 7 -RPV head surface.

The red line marks the interference fit region and the twoyellow arrows identify the main leak path.measured at specific points using an eddy current thickness gage. Next, the RPV headsurface in the region of the primary leak path was replicated with a Microset

material, and layer thickness measurements were made on cross-sectional slices of the replica tovalidate several of the eddy current measurements.

Finally, the replicated sections wereexamined with a stereomicroscope to provide an indication of the corrosion extent.The separation of the Alloy 600 tube from the low-alloy steel of the RPV and ensuingpoint measurements of the boric acid and corrosion layer thicknesses allowedconfirmation of the ultrasonic findings.

Particular interest was placed on the main leakpath area and the region immediately adjacent to it. Boric acid and corrosion layerthickness values were acquired along lines traversing the leak path inside of theinterference fit and outside (below) the interference fit. Within the leak path there was noevidence of boric acid deposits, only a thin corrosion layer. This is indicative of acontinual flushing of any potential deposits during plant operation.

The corrosion layerwas measured at 10 microns (0.79 mils) or less while the high ultrasonic responses werenominally greater than 60-70%. Outside of the leak path, to either side, the ultrasonic responses decreased while the boric acid deposit layer increased.

Ultrasonic responses were typically less than 50% in the interference fit and greater than 50% outside theinterference fit. Boric acid deposits were 15 to 75_microns (0.59 to 2.95 mils) thick in theinterference fit and 130 to 190 microns (5.12 to 7.48 mils) thick outside the interference fit. The interference fit region would reflect less energy and also contain a thinner boricacid layer as compared to the larger gap in the annulus above and below the interference fit in the presence of a counter bore. This counter bore could potentially fill with boricacid deposits in a leaking nozzle. Furthermore, the deposits appear less compacted asnoted by the larger amplitude of the ultrasonic signals.

In summary, an inverserelationship between the ultrasonic responses and boric acid and corrosion layer deposits7 was seen across a line traversing the primary leak path. The leak path is characterized bya high ultrasonic response and low layer values while the regions on either side of theleak path are characterized by a lower ultrasonic response and higher layer thickness.

Lastly, replicated surfaces were viewed with a stereomicroscope to better documentthe surface conditions and to attempt to quantify the corrosion or erosion of the low-alloy steel in the annulus region. The replica of the main leak path is displayed on the left inFigure 8. Machining marks were observed on the replicated surfaces indicating minimalcorrosion,

erosion, or wastage throughout the leak path region. Piece 4 obtained at thetransition from below the interference fit to the interference fit shows the surfacemachining marks still present as seen on the right in Figure 8. It was estimated that thefinish within the interference fit region was equivalent to a turned finish of 1.6 micro-meters (63 giin.) while the finish below the interference fit was equivalent to a milled 1.6micro-meter-(63 gin.-) finish. A noted corroded area was visible at the exit point of themain leak path at the top of the head. It covered an area approximately 12.7 b 1.6 mm(0.5 by 0.06 in. and had a depth of 0.25 mm (0.01 in. .Figure 8 -Replicated leak path on the left and piece 4 on the right with machining marksevident.

The red line on the right represents 2.0 mm (0.80 in.) in length.CONCLUSIONS Nozzle 63 from the North Anna Unit 2 Nuclear Power Plant had showed signs of leakingthrough the annulus region and a suspected leak path was detected during in-service inspections.

PNNL was able to ultrasonically evaluate the nozzle at PNNL after the head8 was replaced and several nozzles were made available for further study. To first evaluatethe ultrasonic phased array system detection capabilities, a calibration/resolution mockupspecimen was built. It was fabricated with one interference fit containing machinednotches to simulate wastage and another with boric acid deposits.

The ultrasonic systemwas able to detect and size the notches and to distinguish between areas with and withoutboric acid deposits.

Wastage or voids were identified by a high-amplitude

response, typically greater than 60%, indicating a strong reflection.

Boric acid deposits wereidentified by a low-amplitude

response, less than 30%/o, indicating good energy transferthrough the interference fit region.The evaluation of Nozzle 63 revealed the previously identified leak path as well asother partial leak paths as high ultrasonic responses.

Potential boric acid deposits werenoted in areas surrounding the leak paths as lower ultrasonic responses.

Destructive evaluation validated the ultrasonic findings and confirmed the relationship betweenultrasonic response and the two conditions of voids (or wastage) and boric acid deposits.

A layer thickness composed of boric acid and corrosion products was measured with aneddy current probe at points traversing the main leak path. The data showed a minimallayer thickness in the main leak path, where conversely the ultrasonic response was high.On either side of the leak path, the layer thickness increased while the ultrasonic responsediminished.

Inside the interference fit, the boric acid deposits were in the 20-40 micronrange where a tight fit was expected.

Outside the interference fit, the layer deposits werein the 180-200 micron range; thicker deposits were due to a looser fit in the counter boreregion. Machining marks on the low-alloy steel head and only one small area of notedwastage were indicative of a low flow through the annulus during plant operation.

REFERENCES

[I] Cumblidge, S E, Doctor, S R, Schuster, G 1, Harris Jr., R V, Crawford, S L,Seffens, R ., Toloczko, M B, and Bruemmer, S M, Nondestructive and Destructive Examination Studies on Removed-from-Service Control Rod Drive Mechanism Penetrations, NUREG/CR-6996, PNNL-18372, Washington, D.C.: U.S. NuclearRegulatory Commission; 2009.[21 Cumblidge, S E, Crawford, S L, Doctor, S R, Seffens, R J, Schuster, G J, Toloczko, M B, Harris Jr., R V, and Bruemmer, S M, Nondestructive and Destructive Examination Studies on Removed-from-Service Control Rod Drive Mechanism Penetrations, PNNL-16628,

Richland, Washington:

Pacific Northwest NationalLaboratory; 2007.[3] Gorman, J, Hunt, S, and Riccardella, P, 2006, "Chapter 44, PWR Reactor VesselAlloy 600 Issues [in Part 12: Current Issues of B&PV Codes and Standards],"

Companion Guide to the ASME Boiler and Pressure Vessel Code, Volume 3,Second Edition, ASME Press, New York.[4] Howell III, A T, Davis-Besse Reactor Vessel Head Degradation Lessons-Learned Task Force, ADAMS #ML022760172, Washington, D.C.: U.S. Nuclear Regulatory Commission; 2002.[5] Cinson, A D, Crawford, S L, Mathews, R A, MacFarlan, P J, Hanson, B D, andDiaz, A A, "Ultrasonic Phased Array Evaluation of Control Rod Drive Mechanism (CRDM) Nozzle Interference Fit and Weld Region,"

Proceedings of the ASMEPressure Vessels and Piping Conference, 2011 (PVP201 1), Baltimore,

Maryland, American Society of Mechanical Engineers, New York, 2011. data/pdfs/trk-l0/PVP2011-57977.pdf 9