ML13113A233
| ML13113A233 | |
| Person / Time | |
|---|---|
| Site: | Calvert Cliffs  |
| Issue date: | 04/23/2013 |
| From: | Jay Collins Piping and NDE Branch |
| To: | |
| Collins J | |
| References | |
| TAC ME8871 | |
| Download: ML13113A233 (21) | |
Text
TECHNICAL LETTER REPORT EVALUATION OF LICENSEES ALTERNATIVE TO 10 CFR 50.55A(G)(6)(II)(F) FOR LIMITATIONS TO VOLUMETRIC EXAMINATIONS OF DISSIMILAR METAL WELDS CONSTELLATION ENERGY CALVERT CLIFFS NUCLEAR POWER PLANT - DOCKET NUMBER: 50-318 BACKGROUND By letter dated June 7, 2012, and with subsequent information in a letter dated January 10, 2013, the licensee, Constellation Energy, submitted an alternative to the examination requirements of Title 10 of the Code of Federal Regulations (10 CFR) 50.55a(g)(6)(ii)(F) which, in part, requires licensees implement American Society of Mechanical Engineers Boiler and Pressure Vessel (ASME) Code Case N-770-1, Alternative Examination Requirements and Acceptance Standards for Class 1 PWR Piping and Vessel Nozzle Butt Welds Fabricated With UNS N06082 or UNS W86182 Weld Filler Material With or Without Application of Listed Mitigation Activities. The CFR requirements include a baseline ultrasonic examination to be performed on each full penetration piping butt weld in the reactor coolant pressure boundary welded with Alloy 82/182 materials. Ultrasonic examination requirements are listed in ASME Code Section XI and ASME Code Case N-770-1, as modified by CFR.
The licensee submitted an alternative to volumetric examination coverage requirements for several dissimilar metal welds (DMWs) at Calvert Cliffs Nuclear Power Plant (Calvert Cliffs).
The alternative applies to limited coverage obtained on circumferential scans (for detection of axially-oriented cracking), as well as for limited coverage on axial scans (for detection of circumferential cracking) on select DMWs. The NRC requested that Pacific Northwest National Laboratory (PNNL) evaluate the licensees alternative with respect to claimed coverage and ultrasonic technique capabilities as applied to two reactor coolant pump (RCP)-to-primary loop piping welds identified as Weld 30-RC-21A-10 and Weld 30-RC-21B-10. The evaluation included theoretical modeling of the sound beams based on actual phased-array design parameters and component geometrical information provided by the licensee. Additionally, based on NRC request, PNNL performed limited backscattered acoustic energy response modeling from a simulated circumferential flaw. It should be noted that currently modeled sound field extents and densities, and reflected energy from the simulated flaw, represent only isotropic material, i.e., actual grain sizes and structures, velocity ranges, and other material variables that will affect sound beam attenuation, re-direction, specular reflections from flaws, and signal-to-noise values have not been applied, and in some cases, are presently unknown.
Another consideration is the potential for inconsistent transducer coupling. This variable is not addressed in the current modeling software, as CIVA1 tends to set perfect contact between the probe and the examined component. However, the actual component outside diameter (OD) weld surfaces, from where the manual phased-array examinations were performed, would typically possess varied waviness around the circumference of the pipe, as the weld crowns 1
CIVA is a trademarked acoustic modeling software package developed by CEA.
would have been manually ground or blended smooth prior to examination. Depending on the extent of surface irregularities that exist, intermittent and unpredictable losses of ultrasonic transducer coupling may affect transmitted sound beam coherence. This can be more pronounced with phased-array probes, as the array is generally required to be adequately coupled over the entire primary axis in order for wave-fronts emitted from individual elements to constructively interfere to produce proper beam steering and focusing. In addition, the licensee stated that probe wedges with flat contact surfaces were used for these examinations, which could also contribute to coupling inconsistencies in wavy regions, or if the probe were to rock during circumferential scans on the component.
Based on the physical limitations described above, the models should be viewed as a best-case representation only.
As requested, PNNL performed individual assessments for each of the subject welds. These are provided in the following sections of this report.
EVALUATION Weld 30-RC-21B-10 Weld 30-RC-21B-10 is a full penetration DMW on a discharge nozzle joining the reactor coolant pump outlet to a cast austenitic stainless steel (CASS) safe end. The carbon steel primary piping is clad with stainless steel on the inside diameter (ID), and the CASS safe end is welded directly to the RCP housing. An idealized cross-sectional depiction of Weld 30-RC-21B-10 is shown as Figure 1. Note that this scaled drawing shows the OD surface to have some irregularity. Because OD surface features of a ground weld crown can vary circumferentially, and no information was obtained from the licensee to depict actual field conditions, PNNL modeled Weld 30-RC-21B-10 with essentially flat surfaces and only accounted for OD dimensional changes such as the slight axial taper noted across the weld. Figure 1 also provides the ASME examination volume required as a red dashed area.
Figure 1. Idealized Cross Section of RCP Weld 30-RC-21B-10 The axial examination (for detecting circumferentially-oriented flaws) of this weld is physically limited by a welded structural steel insulation support that extends on a circumferential arc of approximately101.6 mm (4 in.) in length. Further limiting the axial examination is a spray nozzle branch connection that impacts axial scanning access for approximately 254 mm (10 in.) around the piping circumference, and is located just downstream of the insulation support. These physical limitations are illustrated in Figures 1 and 2. Due to these limitations, the licensee has postulated the maximum circumferential PWSCC flaw size that, due to scanning limitations, could go undetected in this weld. The semi-elliptical flaw is shown to have dimensions of 254 mm (10 in.) in length and a maximum of 30.5 mm (1.2 in.) in through-wall extent; highlighted as red in Figure 2.
Figure 2. End View of RCP Weld 30-RC-21B-10 Detailing Access Limitations and Potentially Circumferential Flaw in this Area The licensee also submitted a volumetric examination sketch, showing theoretical beam plots for axial scan coverage obtained; this sketch includes the postulated semi-elliptical planar flaw that could go undetected. The sketch is reproduced as Figure 3, and indicates that a flaw developing in the susceptible material region could grow to a maximum through-wall extent of 30.5 mm (1.2 in.) before detection (via flaw tip specular or diffracted responses) in the examination region containing the insulation support plate and spray nozzle. Without these limitations, axial scans can be initiated at a sufficient distance away from the weld on the carbon steel primary piping to ensure 100% volumetric coverage of the susceptible material, as full volumetric scans were completed for the remaining circumferential sections of the weld. Due to the scanning limitations described, the licensee estimated volumetric coverage for axial scans at approximately 93% of the required circumference of the weld, which includes the full inner one-third of the susceptible weld material, as examined from the ID-clad carbon steel piping. No coverage credit was taken for the CASS safe end material.
Figure 3. Licensee Calculated Volumetric Coverage and Largest Undetected Flaw for Weld 30-RC-21B-10 PNNL modeled theoretical ultrasonic beam intensities for the applied phased array examination on this weld in 10 degree increments from 0 to 80 degrees, using the licensees stated half path focusing method of 113 mm (4.4 in.). The focal law values, actual probe physical matrix parameters, and the idealized drawing submitted by the licensee were used as inputs to this model. Figure 4 shows examples of UT beams for 10 and 30 degrees, respectively, as simulated by CIVA. The models were generated using a 6 dB filter applied to the sound beam.
Although the licensee has not taken credit for angles below 30 degrees, beam coverage for 10 degrees is shown to demonstrate that ID coverage may potentially be improved by using lower refracted angles (between 10 and 30 degrees). As modeled, the 30 degree beam simulation correlates with the lower angle volumetric coverage provided by the licensees description of beams from 30 to about 70 degrees (as illustrated in Figure 3).
Figure 4. PNNL Mode eled -6 dB Beams B for 10 and 30 Degrrees Respectively The licennsees phase ed array probe was operrated with fo ocal laws deffined to prod duce steered d beams frrom 0 to 85 degrees, d at one-degree o increments, each focused at approxximately 113 3 mm (4.4 in.) of o metal path h after exitin ng the probe. This focal length is be eyond the ID surface for steered beams b less than t about 40 4 degrees, and only pro oduces 6 dB field denssities at the ID for beams frrom approxim mately 0 to 303 degrees. Similarly, steered beam ms above ab bout 65-70 degrees will not prod duce useful beam b profile es for detectiing flaws nea ar the ID because they are a focused ata too short a metal path h length.
The mod del predicts that t all soundd beams abo ove approxim mately 30 de egrees to ha ave less than n 6 dB beam m intensities near the ID of the weld. The 6 dB value repressents a poin nt where fieldd intensity is diminishe ed by 50% off the initial maximum; m evvery 6 dB iss an addition nal 50%
reduction n; for exampple, 12 dB iss 4 times low wer than initial sound en nergy. The model m repressents a best-ccase scenarrio; that is, no material attenuation or o sound bacckscattered from f the coaarse grain struucture of thee weld is included in the model. The ese factors, as a well as otther potentia al coupling issues desc cribed above e, will typically lower the amplitudes of signal ressponses from m real flawss, resulting in decreased d signal-to-noise ratio (S S/N), and ma aking flaw de etection significanntly more challenging ou utside the the eoretical 6d dB region.
Postulatted Flaw De etection in Weld W 30-RC-21B-10 PNNL evvaluated theo oretical deteection of the licensee-postulated ID-cconnected circumferent c ial flaw origiinating in the e susceptible e material re egion, using modeled de efect responsses. These simulatioons focused on assessin ng the detecttion of a plan nar type, semmi-elliptical flaw f located in the weld with a scan region that was physica ally limited by the insulattion support and spray
nozzle booss as previously depictted in Figure e 3. The sem mi-elliptical planar p defecct was defineed to have an arc length of 254 mm (10 in.) and a maximum th hrough-wall extent of 30 0.5 mm (1.2 in.)
own of the flaw, as illustrrated in Figu at the cro ure 5. The fllaw was placced in the bu utter/weld re egion of the component an nd was conne ected to the ID surface.
Figu ure 5. Top: Perspective P V View of Semii-elliptical Pla anar Flaw; Bottom:
B End View V of Flaw w
Theoreticcal flaw deteection was firrst assessedd by simulatiing a linear axial a pipe sccan over the weld, orig ginating withh the phased d array probee positioned d against thee insulation support s (as far f
back as possible p in relation r to the scan regioon), as show wn in Figure 6.
6 Insonifica ation angles set from 0 to o 60 degrees s in 5 degree e incrementss were seleccted for this simulation.
s
Fig gure 6. Persp pective Vieww of Flaw with h Linear Line e Scan Path Figure 7 shows results of the deffect responsse simulation n from the 30 0 degree sca an angle. The upper lefft image shows that the probep is in th he initial sca an position (aas far back as a possible from f
the weld in relation to o the insulattion support)). The ray trrace line exitting the centter of the proobe is the cenntral ray for the L-wave, 30 degree refracted r angle. The up pper right imaage shows that t
the results displayed correlate with w the initiall scan positio on indicatedd by the nummeric 0 labeling on the black cursor bar.
b The lowwer left image is an unprrocessed non-volume co orrected secttorial display showing s all angles a simulaated versus time at a pa articular scann position. Shown S are simulated d flaw respo onses for ang gles 0 to 60 degrees (0 degrees d on the left and 60 degrees on the right)), as generatted from the initial scan position. Th he black curssor in this im mage is centered on the shot numberr correspond ding to 30 de egrees. The lower right quadrant q displays the simulated d time-ampliitude A-Scan n image. Th he A-scan im mage shows that a 30 de egree angle does not result in a strong ID corner-trrap response e from the flaw, as angle es of 30 deggree and abo ove are projeected to be higher h than thhe ID conne ected region (a lower ang ed from this scan gle is require position to t insonify thhe ID connecction region)). A low amplitude top--of-flaw resp ponse is observed d in the figurre indicating that a postu ulated flaw ofo 30.5 mm (1.2 in.) throu ugh wall exteent could posssibly be detected at 30 degrees givven the axial scanning limitation.
Howeverr, the 30 deg gree angle beam would be b unable to o assess whe ether the flaw could be considere ed as ID-con nnected beccause this an ngle does no ot adequately insonify th he ID flaw coorner-trap regioon. The view w of the flaw from the inittial scan possition (lower left of the fig gure) indicattes that a low wer inspectio on angle, such as appro oximately 20 degrees, co ould possiblyy result in a modest amplitude a response from m the ID corn ner-trap of thhis flaw. As the scan pro ogressed froom the start position tow ward the weld d region, thee postulated flaw disappe eared from view v in the simulatioon. Thus, it iss imperative that the sca an begin at the simulated d starting po oint (as far back from weld d centerline as possible) for the grea atest amoun nt of volumettric coverage e.
Figure 7.
7 Defect Re esponse at 30 Degrees frrom the Initia al Scan Posittion At the req quest of NRC, a second d evaluation was conduccted using th he previouslyy defined semi-elliptical planar flaw to t determine e the effect of o through-w wall depth onn the top-of-fflaw responsse at a 30 degree inspection angle. The T ID corner-trap response at 20 de egrees from the initial sccan position was w selected d as a consttant referencce signal at 0 dB (maxim mum screen height) to normalize e the variable height top p-of-flaw responses. The corner trap p response is i unchangin ng at this angle e and positio on for all flaw ws evaluated d (regardless of through h wall depth) and so makkes a good com mmon respo onse for norm malizing amp plitudes of th he top-of-flaw w responsess. For refere ence, Figure 8 shows the A-Scan A view w of the corneer response at 0 dB give en by the 30.5 mm (1.2 in.)
i through wall w extent fllaw.
Fiigure 8. Corn ner Trap 0 dBB Referencee A-Scan at 202 Degrees Continuin ng to locate the probe aggainst the in nsulation suppport, the initial flaw dep pth of 30.5 mm m
(1.2 in.) was w made deeper in increments of 2.5 2 mm (0.1 in.). The theoretical top p-of-flaw response e was recordded at each increment to o define the response off the flaw witth increasingg flaw deptth; the peake ponse was allowed to subside as the ed flaw resp e flaw depth increased beyond the center off the 30-degrree beam. In this manner, the maximum flaw depth to prod duce the best top-of-flaw response r could be deterrmined. Figu ure 9 shows a series of A-Scans A
highlightiing response es from the 30.5, 3 35.6, 40.6, 4 and 433.2 mm (1.2, 1.4, 1.6, and 1.7 in.)
through wall w extent fllaws, respecctively, and insonified at the 30 degrree angle. Itt was observved that a flaww with a ma aximum throu ugh-wall exteent of 35.6 mm m (1.4 in.) results in th he highest theoretical top-of-flaw w response (lowest delta a amplitude)) when using g a 30 degre ee beam (see e Table 1)..
Figure 9. Variable Height H Defecct Signal Res sponses at 30 Degrees frrom the Initiaal Scan Posittion Table 1 - 30 Degree Top of Flaw w Response as a a Function n of Flaw Height Flaw Heigght Toop of Flaw Response at 30 mm (in..) Degre ees (dB)*
30.5 (1.22) -15.3 33.0 (1.33) -11.9 35.6 (1.44) -10.8 38.1 (1.55) -11.2 40.6 (1.66) -14.6 43.2 (1.77) -19.6
- 0
- dB set to ID D corner trapp response at 20 degrees on, flaw dete In additio ection of the original semmi-elliptical 30.5 mm (1.22 in.) through h-wall flaw was w
evaluatedd just outsid de of the region impacted d by the insuulation suppo ort, where only o the spray nozzle prrecluded a fuull scan pathh. In this reg gion, the proobe was ena abled to movve an additional
8.9 mm (0.35
( way from the weld, permitting greaterr inspection coverage off the susceptible in.) aw material. At this loca ation the flaw w has a nom minal through h-wall extent of 24.2 mm m (0.95 in.) and both the ID corner tra ap and top-oof-flaw respoonses are cle early seen in n the simulattion result fo or the 30 degre ee insonificattion angle, as a shown in Figure 10. Note N the nozzle boss is not n displayed d in this figuree.
Figure 10.
1 Top: Side e and End Viiews Respec ctively of Scaan without In nsulation Sup pport Limitattion; Bottom:
B Defe ect Signal Re esponse from m Flaw at 30 Degrees Finally, as a shown in Figure 11, th he far edge of o the flaw was w evaluate ed for detecttion under th he same axiial scan limittation as desscribed abovve (for the area immedia ately outsidee the insulation support),, where the nozzle n boss is the only physically p lim miting factor,, and the pro obe is thereffore allowed tot move the additional 8.9 mm (0.35 5 in.) away frrom the weld d. The simu ulation shown n in Figure 11 1 indicates that a strong ID corner-trrapped response from th he extremityy (near to thee end) of thhis flaw can also be see en with a 30 degree d insonification anngle, when th he additional scan leng gth is allowe ed. At this exxtremity, speecular respo onses from the ID cornerr and top-of--flaw are mostt likely superrimposed; th hus it would not n be expeccted that a shallow s throuugh-wall flaw w
could be accurately depth-sized d using standard tip diffraaction techniqques. Note the nozzle boss b
is not dissplayed in Figure 11.
Figure 11
- 1. Top: Side and End Vie ews Respecttively of Scan n without Ins sulation Support Limitation at Flaw Edge; Bottom: Defect D Signal Response from f Flaw Eddge at 30 Degrees
Weld 30-RC-22A-10 Weld 30-RC-22A-10 is a full penetration DMW on the RCP suction nozzle joining carbon steel, ID clad, primary piping to a CASS safe end. The safe end is welded directly to the RCP housing. An idealized cross-sectional depiction of Weld 30-RC-22A-10 is shown as Figure 12.
This weld varies from the outlet Weld 30-RC-21B-10 by an OD taper between the ferritic elbow to the CASS safe end, as depicted.
Figure 12. Idealized Cross Section for Weld 30-RC-22A-10 As with the previous weld, the licensee submitted sketches showing areas where no volumetric coverage was obtained for Weld 30-RC-22A-10. One such sketch is provided as Figure 13. The licensee estimated combined circumferential scan coverage (for axially-oriented flaws) to be approximately 91% of the required volume. The area of no coverage is shown to be an ID region of the weld nearest the carbon steel in the susceptible Alloy 82/182 weld/buttering material that has a potential to contain an undetected PWSCC flaw that is 5.1 mm (0.2 in.)
through-wall and 10.2 mm (0.4 in.) long (shown in red in the figure). Note that the licensee did not take credit for electronically skewing the UT beam a maximum of 10 degrees down into the weld ID region. As shown in the figure, the primary reason for limited coverage is the OD taper.
Figure 13. Licensee Calculated Volumetric Coverage for Weld 30-RC-22A-10 Optimized ID Impingement Flaw detection in austenitic weld materials is complicated, but is generally believed to require a corner-trapped, or crack-face, specular response to be back-scattered to the detecting probe with sufficient energy to yield a minimum 2:1 S/N. This amplitude is affected by several factors, including material acoustic properties, and impedance mismatching and orientation of the flaw, with respect to the impinging sound beam. When attempting to detect axially-oriented, ID surfacing-breaking planar flaws, there is a theoretical optimum range of ID impingement angles that should be designed into the transducer/wedge combination. The desired impingement angle should be within a critical range (above or below which sound would not optimally impact the ID surface to produce a corner-trapped flaw response) and can be calculated by the following:
OD sin ( ) = sin ( ) (1)
ID where: is the ID impingement angle, is the initial refracted angle from the probe on the OD surface, and OD/ID is the ratio of the outside-to-inside pipe diameters.
A graphical depiction of this relationship is shown as Figure 14. According to the industrys Performance Demonstration Initiative (PDI) generic DMW ultrasonic procedure 10 (PDI-UT-10),
the optimum ID impingement angle () for detecting PWSCC on the subject welds is in the range of 55-60 degrees, vis--vis, the transmitted refracted angle () should be in the range of 42-46 degrees.
Figure 14. Representation of Solution for ID Impingement Angle PNNL also modeled Weld 30-RC-22A-10 for beam coverage results for +/-10 degree lateral skews at steered angles of 42 and 46 degrees. Sound fields with a 6 dB filter were again used as shown in Figure 15 for the refracted L-wave beams. Here, due to the OD taper and slightly increased wall thickness as compared to Weld 30-RC-21B-10, an area of less than 6 dB coverage is shown to exist on the ID region of the weld; this appears to contradict the licensees stated coverage. Given the phased array parameters employed, the models show that full coverage (using 6 dB field intensities) is not obtained with beam steered angle ranges of between 42 to 46 degrees (optimized for ID impingement) even with electronic lateral skewing of +/-10 degrees. Figures 16 and 17 depict side views of the sound fields at 42 and 46 degrees positioned over carbon steel. The beam computation models show that a flaw height of approximately 12.69 mm (.49 in) could exist before the 6dB sound field at 42 degrees would be able to detect the tip of the flaw.
Figure 15 End View: PNNL Modeled 6 dB Coverage (42 Degrees) Positioned over Carbon & Weld (Weld 30-RC-22A-10)
Figure 16 Side View: Sound Field Intensity Profile at 42 Degrees on Weld 30-RC-22A-10
Figure 17 Side View: Sound Field Intensity Profile at 46 degrees on Weld 30-RC-22A-10 Shown below in Figures 18, 19, 20 and 21 are the 12dB (4 times less) theoretical sound intensity fields for the same +/-10 degree lateral skews at steered angles of 42 and 46 degrees, at locations adjacent to and over the weld. As can be seen, the lower intensity sound fields appear to insonify the entire ID region.
Figure 18 Skewed sound fields on Weld 30-RC-22A-10 for 12dB envelope at 42 degrees adjacent to weld over carbon steel; top left is top view, bottom left is pipe axial view, right is cross-section perspective Figure 19 Skewed sound fields on Weld 30-RC-22A-10 for 12dB envelope at 46 degrees adjacent to weld over carbon steel; top left is top view, bottom left is pipe axial view, right is cross-section perspective
Figure 20 Skewed sound fields on Weld 30-RC-22A-10 for 12dB envelope at 42 degrees over weld; top left is top view, bottom left is pipe axial view, right is cross-section perspective Figure 21 Skewed sound fields on Weld 30-RC-22A-10 for 12dB envelope at 46 degrees over weld; top left is top view, bottom left is pipe axial view, right is cross-section perspective
CONCLUSIONS Modeling of Weld 30-RC-21B-10 has shown 6 dB beam coverage at the ID region of the weld at a refracted angle of approximately 30 degrees, or less. The simulations are in agreement with the licensees prediction of coverage for the 30 degree angle only, as this is the lower value claimed to have been qualified. The simulated 6 dB coverage at 30 degrees results in a similar beam trajectory over the weld region as shown by the licensees sketch, and PNNL agrees that a flaw must grow to a minimum of 30.5 mm (1.2 in.) in through wall extent before a top-of-flaw response could theoretically be observed. Defect response simulations complement the beam coverage plots, indicating that a top-of-flaw response is observed from a flaw of this size. It should be noted that simulations were conducted without material noise or attenuation which could significantly reduce signal strength and quality.
Further defect response simulations indicate that as the flaw grows in through-wall depth, response(s) from the top of flaw increase with a maximum response from a flaw that is approximately 35.6 mm (1.4 in.) through wall, based on a 30 degree inspection angle. Thus, if volumetric coverage claims only begin using a 30 degree angle, it is more likely that a 35.6 mm (1.4 in.) through-wall flaw will be detected than the shallower flaw proposed by the licensee.
Defect response simulations, when scans are not limited by the insulation support, indicate that a 30 degree beam produces higher amplitude ID corner-trap responses, rather than significantly reduced top-of-flaw responses, which should result in improved flaw detection.
Ultrasonic beam computation models of Weld 30-RC-22A-10 partially agree with coverage maps provided by the licensee (see Figure 13), but potential flaws of the size provided by the licensee may go undetected during an inspection due to the lack of sufficient beam intensity at the ID surface region. Models using optimized impingement angles for detecting axially oriented flaws (L-wave refracted angles of 42-46 degrees) do not project 6 dB sound beams generally desired for flaw corner-trap detection at the ID region. In addition, based on the models, a roughly 17% through-wall, or approximately 12.69 mm (0.49 in.), deep flaw is the minimum depth needed to place the upper region of the flaw just into the 6dB sound field. It is important to note that, as in models of Weld 30-RC-21B-10, simulations on Weld 30-RC-22A-10 also do not account for material attenuation and/or noise.