ML20113C635

Provides Justification for Extending Current Operating Cycle Interval to Complete Cycle for Unit 1 Cycle 18,960303 Through 971004.Ltr Updates Commitments Made in SG Insp Rept Ltr & Forwards RAIs & 0506

ML20113C635
Person / Time
Site:	Prairie Island
Issue date:	06/27/1996
From:	Wadley M NORTHERN STATES POWER CO.
To:	NRC OFFICE OF INFORMATION RESOURCES MANAGEMENT (IRM)
References
NUDOCS 9607020024
Download: ML20113C635 (43)

Text

Northern States Power Company Prairie Island Nuclear Generating Plant 1717 Wakonade Dr. East Welch, Minnesota 55089 June 27,1996 U S Nuclear Regulatory Commission Attn: Document Control Desk Washington, DC 20555 PRAIRIE ISLAND NUCLEAR GENERATING PLANT Docket Nos. 50-282 License Nos. DPR-42 50-306 DPR-60 January 1996 Steam Generator Sleevino issues Ninetv Dav Resoonse Letter This letter provides the justification for extending the current operating cycle interval to the complete cycle for Unit 1 Cycle 18, March 3,1996 through October 4,1997.

In our Steam Generator inspection Report letter dated February 28,1996', NSP committed to a mid-cycle shutdown for inspection of sleeves in 12 steam generator no later than eight operating months following the startup of Unit 1 due to circumferential and volumetric eddy current indications located in the upper sleeve weld region (Above the Tube Sheet or ATS Weld) at or above the sleeve welds. Further,5 sleeved tube samples were removed from 12 steam generator for detailed nondestructive and metallurgical examination. The results of that examination are contained in ABB Combustion Engineering Report CEN-628-P, Revision 01-P," Verification of the Structural Integrity of the ABB CENO Steam Generator Wolded Sleeve"2, which was forwarded to you by NSP letter dated May 3,1996. The Report documents that the sleeve weld indications are installation related and are not due to in-service induced degradation.

This letter:

updates the commitments made in the Steam Generator inspection Report letter dated February 28,1996, 9607020024 960627 PDR ADOCK 05000282 G

PDR sLvsoon.coe j.

f

{o

/

4 USNRC NORTHERN STATES POWER COMPANY June 27,1996 Page 2 summarizes the findings of ABB Combustion Engineering report CEN-628-P, e

Revision 01-P, provides justification for the deletion of the NSP Commitment to conduct a mid-cycle e

sleeve inspection and for continued operation of Prairie Island Unit 1 throughout the entire operating Cycle 18, and provides answers for the Request for Additional Information dated March 26,1996 on Prairie Island sleeving and inspection issues.

provides answers for the Request for Additional Information dated May 6,1996 on Prairie Island sleeving and inspection issues.

Background

During the Prairie Island Unit 19601 Refueling Outage Steam Generator Tube Inspection, eddy current indications were found in the upper sleeve weld region of 61 ABB Combustion Engineering welded tubesheet sleeves. These indications were characterized as single or multiple circumferential indications or volumetric indications.

All tubes with circumferential indications were removed from service by sample removal and/or plugging. The volumetric indications were evaluated for location by eddy current and those tubes with volumetric indications located below the centerline of the sleeve weld were plugged. 3 Five sleeve / tube samples were removed to determine the root cause of the eddy current indications. Four of the removed tubes contained circumferentialindications and one contained a volumetric indication.

The commitments made by NSP in the Steam Generator Inspection Report letter dated February 28,1996 ' were as follows:

We will begin a mid-cycle shutdown for inspection of sleeves in 12 steam generator no later than eight operating months following the startup of Unit 1.

[ Unit 1 startup was March 3,1996)... Subsequent to the receipt of the results of the metallurgical analysis, we will determine whether a Unit 1 mid-cycle shutdown is required to perform examinations of the sleeves in 12 steam generator. Mid-cycle shutdown will be determined unnecessary if the metallurgical analysis conclusively shows that the upper sleeve weld area indications are not service related or the analyses indicate that service related indications do not present flaw sizes of magnitudes great enough to present structural or leak hazards during the expected fuel cycle length.

SLV90 DAY. DOC

USNRC NORTHERN STATES POWER COMPANY June 27,1996 Page 3 i

STATUS: This letter justifies elimination of the mid-cycle inspection requirement j

During the next Unit 1 operating cycle, we will pursue formal qualification of e

eddy current examination for this application. [This applies to the CE welded tubesheet sleeve inspection, in particular, volumetric location.)

STATUS: Qualification of the elevation of volumetric indications by eddy current has been completed and will be submitted with the revised Combustion Engineering Sleeve Topical Report. An additional commitment is being considered for the

{

qualification of weld height determinations by ultrasonic examination.

Within 90 days of startup, we will submit the results of the metallurgical 1

analyses and our conclusions for NRC review and concurrence.

4 i

STATUS: The results of the metallurgical analysis are contained in ABB Combustion Ecgineering report CEN-628-P, Revision 01-P, " Verification of the Structural Integrity of the ABB CENO Steam Generator Welded Sleeve"2 which was forwarded i

to you by letter dated May 3,1996. Our conclusions are included later in this letter for your review and concurrence.

Although not a formal commitment, NSP stated that the EPRI Primary to Secondary e

Leak Rate Guidelines' would be incorporated into Pl Steam Generator Tube Leak Abnormal Operating Procedures.

STATUS: Operator training on and issuance of 1[2]C4AOP2, Steam Generator Tube Leak Abnormal Operating Procedure, incorporating the EPRI Guidelines, was completed April 9,1996.

Summarv of the Justification for No Unit 1 Mid-Cvele Sleeve Insoection Outage Sleeve weld indications are not service induced, but rather are the result of weld preparation cleanliness Sleeve weld indications show no growth in 3 years The Installation UT and the ET using + Point rotating coil probe provide an adequate examination of upper sleeve welds Avoid plant cycle to cold shutdown and resultant radiation exposure SLV90 DAY. DOC i

-. ~...- -. ~

l l

USNRC NORTHERN STATES POWER COMPANY j

June 27,1996 j

Page 4 i

j Five sleeved tubes were removed form Prairie is!and Steam Generator #12 during l

February 1996 and examined at the ABB CENO radioactive metallography laboratory i

for evaluation to determine the nature, extent and cause of the circumferential and j

volumetric eddy current indications identified by the + Point probe and also confirmed j

with 3 coil RPC.

}

The purposes of the sleeve examination were to 1) Identify the physical condition 1

which is the source of ET Indications,2) Evaluate the effect of the physical conditions on the structural integrity of the sleeve / tube joint,3) Identify the root cause and j

whether the indications are related to installation or service conditions, and 4) Evaluate l

leak tightness of the indications. Leak tightness was an issue since the plant i

developed a 23 GPD(maximum) leak on 11/17/96 and one sleeve, R7C63, leaked 1 j

drop per 6 minutes in the field under 750 psig secondary pressure.

l The NDE portion of the metallurgical examination consisted of visual observation, eddy l

current testing (ET), radiography, ultrasonic tes:ing (UT), silastic molds and outside j

diameter measurements. Each pulled sample sleeve weld was also subjected to a destructive examination in order to characterize the type of discontinuities as well as the i

size and circumferential extent of the discontinuities which were the source of the eddy j

current indications. Results of these examinations were presented to the NRC in i

j meetings on March 6, March 28, and May 9,1996.

1 i

i l

The five tubes / sleeves removed from 12 Steam Generator Hot Leg are:

)

Row Col.

Eddy Current indication at ATS Weld Year

1. of Welds -

5 48 Single CircumferentialIndication 9210 1

7 52 Multiple Circumferential Indications 9405 4

i 9

57 Single Circumferential Indication 9210 1

l 7

63 Volumetric Indication 9210 2

5 74 Multiple Circumferential Indications 9601 2

l Metalluroical Examination Results The following conclusions were made from the metallurgical examination of the 5 i

Prairie Island pulled sleeves. The bases for these conclusions are documented in ABB 2

Combustion Engineering report CEN-628-P, Revision 01-P I

The eddy current indications are a result of the radial component of either or both of j

two weld conditions termed Incomplete Fusion or Sleeve O.D. suckback. Both of i

these conditions are an anomaly associated with the sleeve installation process, i

SLv90 DAY Doc I

i b

f I

USNRC NORTHERN STATES POWER COMPANY June 27,1996 Page 5 with no evidence of in-service induced degradation.

Incomplete fusion manifests itself in two forms:

the formation of refractory oxides in the weld nugget creating laminar, non-linear, o

inclusions emanating from the intersection of the sleeve and tube faying l

surfaces.

l a lack of fusion flaw where the indigenous oxide layer on the tube surface o

l prevents wetting of the joint faying surfaces during welding and blocks coalescence.

l Both forms of incomplete fusion are a result of insufficient removal of the tube surface oxide from the inside diameter of the parent tube.

l; Sleeve O.D. Suckback is a term given to a rounded cavity formed on the edge of the weld. The rounded nature of this discontinuity indicates it is due to the evolution of a gas within the joint. The source of this gas could be the contamination of the tube i

surface or moisture behind the sleeve.

i l

Sleeve O.D. Suckback occurred predominately on the upper, non-pressure j

boundary portion of the weld. When it did occur below the weld it was localized and limited to less than 15% through the sleeve wall.

No evidence of service induced propagation, including environmental degradation, of any type of discontinuities was present. Most all inclusion termination points, i

many of which were rounded pores, were oxide filled. Furthermore, comparison of i

previous outage eddy current volumetric and circumferential signals exhibited no or i

very minor changes.

j Laboratorv Soecimen Test Proaram I

Combustion Engineering conducted a test program in order to produce quantities of l

sleeve to tube weld specimens that would duplicate the types of indications found i.

during the destructive analysis of the Prairie Island tube pull samples. By producing samples with known and varying severity of oxide inclusions, it was determined that pertinent test data could be obtained to verify leak rate and structural calculations, as well as, provide additional confirmation data for the future application of field NDE j

techniques. From these laboratory specimens the following conclusions can be stated:

i Samples with oxide inclusion in the weld, sleeve O.D. suckback inclusions and weld i

lack of fusion can be produced by welding sleeves to uncleaned tubes containing an i

SLV90 DAY DOC 4

m

=

._ - m USNRC NORTHERN STATES POWER COMPANY l

June 27,1996 1

Page 6 oxide coating and/or magnetite coating on the I.D. surface.

l Test samples with up to 360 degree weld lack of fusion were subjected to an axial push test. No sleeve axial movement was seen at loads up to normal operating and i

accident conditions which confirms that the hydraulic expansion contributes to the strength of the overalljoint.

j Fifty-four samples with varying amounts of lack of fusion up to 360 degrees, weld i

O.D. suckback, and oxide inclusions were subject to leak tests up to 4500 psig.

j Maximum observed leak rate was 0.016 gpm.

i i

Twenty samples were metallurgically examined to determine weld height. All i

samples which were found to be acceptable by UT had average weld heights greater than the original design value.

t Structural Intearity and Leakaa_e Evaluation j

Structural integrity and leakage evaluation were analyzed per the requirements of Section lli of the ASME code and draft Regulatory Guide 1.121. The following j

conclusions can be stated:

5 The sleeve licensing report, CEN -294-P demonstrated that the original design value for an average weld height of 0.080 inches met all the requirements of the 2

j ASME code with considerable margin. Structural Calculations in CEN-628-P show i

that the minimum required average load bearing ligament for the upper sleeve weld for Design conditions is actually 0.019 inches.

f Based on parametric studies of the leak rate analysis, the experimental leakage data is judged to be the appropriate leakage rate, as compared to the calculated l

values, for operating and steam line break condition evaluations of operating plants The ASME Code and Regulatory Guide 1.121 structural criteria were satisfied for the Prairie Island pulled tubes. However, four of the pulled tubes had average weld heights less than 0.080 inches, but more than 0.019 inches, and were documented in Table 5-7 of CEN-628-P and Prairie Island Unit 1 License Event Report 96-07.8 A probability of detection (POD) assessment has shown that the combined ET/UT examination program that was used at Prairie Island is capable of detecting 96.7%

of the sleeves with indications similar to those of the Prairie Island pulled tubes.7 SLV90 DAY. DOC

- _ _ _ - - ~

j USNRC NORTHERN STATES POWER COMPANY June 27,1996 Page 7 The metallurgical examination program and sample production program have resulted in improvements in the installation cleaning process and the post installation inspection process which will be committed to by submitting a revised topical report for Combustion Engineering welded sleeves using the License Amendment Process.

Root Cause of ET Indications in Sleeve Welds The root cause of the ET indications and the discontinuities in the sleeve welds was inadequate removal of the contaminants and oxide from inside of the parent tube prior to sleeve insertion resulting in oxide inclusions captured in the sleeve weld. It was

]

inadequate control and testing of changes to the sleeving installation cleaning process which led to inadequate cleaning. ' A Contributing Causal factor is Change Management l

- the effectiveness of the brushing step was never visually verified under field conditions following changes to the brushing system.

Summarv of Metallurgical Examination Program The ET indications are a result of weld oxide inclusions and sleeve O.D. weld suckback.

All of the indications are due to the welding process and were caused by inadequate removal of contaminants and the oxide layer of the parent tube prior to sleeve insertion.

There is no evidence of in-service induced propagation of the discontinuities. There is no evidence of in-service induced environmental degradation. The discontinuities are all in the weld and on the sleeve side of the weld. One sleeve which leaked during the secondary side pressure test, R7C63, also leaked in the lab. This sleeve weld was found to have an unacceptable lack of fusion and should not have been accepted using the existing criteria during its original installation i

Indeoendent Regulatorv Guide 1.121 Assessment APTECH Engineering Inc. was contracted by NSP to conduct an independent review of the significance of sleeve weld discontinuities found in the Prairie Island sleeve samples. Structural margins, leakage assessments, and fatigue considerations evaluated by Combustion Engineering were verified by APTECH Engineering. 8 Evaluation of Potential For Growth during the Cvele Oxide film inclusions in the weld metal are sometimes crack like circumferential discontinuities which terminate in the Alloy 690 sleeve material / weld material. The excellent corrosion resistance of Alloy 690 and the lack of growth in the pulled tube i

samples argue against growth of these discontinuities via a corrosion mechanism.

SLV90 DAY. DOC

i i

USNRC NORTHERN STATES POWER COMPANY June 27,1996 Page 8 Metallographic examinations of the ATS sleeve weld region of the pulled tubes did not reveal any evidence of growth of discontinuities by either corrosion or fatigue. Three of these sleeved tubes had been in service since 1992. The excellent corrosion performance of Alloy 690 steam generator tubing together with the destructive i

examination of the sleeved tubes with three years of exposure to service provides a i

good basis to argue against the need for a corrosion growth allowance in the evaluation of structurallimits.

2 8

Calculations were performed by both Combustion Engineering and APTECH to evaluate the possible growth of circumferential discontinuities by fatigue loading.

These discontinuities were conservatively treated as sharp cracks. A through wall crack of limited circumferential extent and a partial depth crack around the full tube j

circumference were evaluated. Flow induced vibration stresses are typically quite small in unsleeved tubes. Cyclic bending stresses in the sleeve of a sleeved tube will be reduced further. A generous allo vance of a stress range of 2 ksileads to a cyclic l

stress intensity range of about 1 ksiVin. Even at high mean loads, the applied cyclic stress intensity range will be comfortably below the threshold AKm for the onset of fatigue crack growth. Cyclic stress ranges for plant stop/ start cycles are 7 ksi and are only significant if one ignores the presence of the parent tube below the upper weld joint The applied stress intensity factor for either very long and deep partial through 1

j wall cracks or short (about 40 degrees) throughwall cracks are less than 10 ksiVin. The expected fatigue crack growth rate is less than one microinch per stop/ start cycle.

in summary, fatigue crack growth of circumferential discontinuities which may have

)

been undetected and remain in service is not an issue. Many cycles at the low cyclic stresses due to tube vibration will not result in the onset of growth; and fatigue cycles at j

higher cyclic axial tensile stresses, such as stop/ start cycles, will not lead to any reduction of structural integrity during the fuel cycie.

i Evaluation of Potential For Leakage during the Cvele Fifty four samples with lack of fusion from 0 to 360 degrees were tested for leakage at i

normal operation and steam line break pressures. The maximum measured leakage j

was 0.016 gpm. Each test showed the same leakage for both normal operation and steam line break pressures. Leak rate calculations were done using the PICEP code with parametric studies done on leakage path sizes.8 Leak rates varied from about 0.03 to 0.8 gpm for a complete lack of fusion as the gap opening varied from 0.0001 to 0.001 j-inches. No significant differences were found between the leak rates at normal j

operating conditions and at steam line break conditions. Experimental measurements i

indicate a maximum effective gap of less than 0.0001 inches due to the combination of l

actual gap size and the presence of an oxide film in actual weld joints with lack of fusion SLV90 DAY. DOC 3

a 1

1

-~ -

_ _. - - - _... - -. ~ ~

_ = -. -..

USNRC NORTHERN STATES POWER COMPANY June 27,1996 Page 9 paths. Measured leak rates (maximum of 0.016 gpm) are considered to be most representative of service performance.

Eddy current and ultrasonic inspections have led to the removal of any substantial ATS weld region discontinuities and most detectable lack of fusion leakage paths. The circumferential arc lengths of undetected leakage paths which may remain in service are much less than the bounding case of 360 degrees lack of fusion. Undetected arc lengths are less than 30 degrees. Thus, the limiting case of 0.016 gpm is conservative.

Conseauences of Structumi Failure The consequences of structural failure of the weld joint region from oxide inclusions is l

not severe. Leakage rather than burst is the failure mode. The bounding leakage rate i

is 0.016 GPM. A guillotine break of a steam generator sleeved tube is not credible due l

to the structural integrity of the parent tube not being degraded in the sleeve weld area.

Current Status of Sleeves at Prairie Island l

The number of Sleeve Weld Indications (WZls) in service is 34. These indications are l

classified as volumetric indications in the upper region of the sleeve weld. Of these 34,

(

three are now known to contain 25 degrees or less of lack of fusion and are discussed in Prairie Island Unit 1 LER 96-11.

The worst case scenario is an accident leakage value of.016 gpm for each of the 34 l

sleeves with volumetric indications in 12 Steam Generator corrected for the probability of detection of 0.967.

By multiplying the 34 sleeves by the worst case accident leakage rate of 0.016 gpm, a total leakage rate of 0.54 gpm could occur. The below equation accounts for a probability of detection of less than one and includes all of the initial l

identified sleeve weld indications:

' TotalNumberofVol. Indications - Vol.indicationsPlugged

• LeakRateperSleeve = TotalleakRate Pr obabilityofDetection or

' 61 - 27 *.016GPMperSleeve = 0.58GPM

(.967 This value is small compared to the allowable leak rate for the Main Steam Line Break off-site dose calculation of 5 gpm leak rate *' with an RCS activity level of l

1 microcurie / gram dose equivalent I-131 which produces off-site doses SLV90 DAY. DOC 1

l USNRC NORTHERN STATES POWER COMPANY June 27,1996 Page 10 equivalent to a small fraction of 10CFR100 dose limits. The current Dose Equivalent lodine in Unit 1 primary system is 2E-4 pci/ml.

i Justification for No Sleeve insoection Outaae The justification for elimination of the 8-month sleeve inspection is summarized as follows:

i The sleeve weld indications are not in-service induced, but rather are the result of weld preparation cleanliness. The sleeve weld indicatione have shown no growth in 3 years, both by eddy current evaluation and by metallurgical examination. The i

installation UT and the in-service ET using + Point rotating coil probe provided an adequate examination of ATS sleeve welds so that the number of possible sleeve indications left in service is limited and the main steam line break leakage is j

acceptable. Analysis has shown that the weld discontinuities will not grow I

significantly over the entire fuel cycle.

i Future Actions: Refueling Outage Insoection - October.1997 j

A 100% examination of all sleeves using current state-of-the-art Appendix H qualified eddy current techniques willbe done. A reexamination of all sleeve welds with resulting

)

volumetric and crack-like ETindications will be done by UT.

?

Future Actions: Sleeve License' Amendment i

We will submit, by December 1,1996, the new Combustion Engineering Sleeve Licensing Topical Report, which willinclude a variety of sleeve designs and the latest sleeve examination requirements, prior to the next installation of CE sleeves at Prairie l

Island for approval as a License Amendment.

I Conclusions i

l The Mid-Cycle Sleeve inspection is not required due to the nature of the weld discontinuities. We will be committing to the improved process control and acceptance l

testing via a Sleeve License Amendment Request. The existing Sleeve Weld Indications will be examined again at next refueling outage using improved UT and ET.

1, In this letter we have made new Nuclear Regulatory Commission commitments indicated as the italicized statements above.

1

)

SLV90 DAY. DOC a

i n

1 USNRC NORTHERN STATES POWER COMPANY l

June 27,1996 Page 11 i

Please contact Jack Leveille (612-388-1121, Ext. 4662) if you have any questions j

related to this letter.

l vf Michael D Wadley Plant Manager i

Prairie Island Nuclear Generating Plant i

i c: Regional Administrator - Region Ill, NRC j

Senior Resident inspector, NRC 4

NRR Project Manager, NRC J E Silberg i

2 Attachments:

1. Request for Additional Information, dated March 26,1996 on Steam Generator j

inspection Report, dated February 28,1996

2. Request for Additional Information, dated May 6,1996 l

NSP Prairie Island Nuclear Generating Plant Letter to NRC dated February 28,1996, i

" Steam Generator Inspection Reports" 2 ABB Combustion Engineering Report CEN-628-P Revision 01-P," Verification of the Structural Integrity of the ABB CENO Steam Generator Welded Sleeve", March 1996 Prairie Island Safety Evaluation Report 436: Eddy Current Indications in Welded Tubesheet Sleeve Upper Welds d

j EPRI Report TR-104788: PWR Primary-to-Secondary Leak Guidelines 5

Combustion Engineering Report CEN-294-P, Final Report, dated January 15,1985,

" Prairie Island Steam Generator Tube Repair Using Leak Tight Sleeves."

8 i

Prairie Island Nuclear Generating Plant Unit 1, Licensee Event Report 96-07, May i

28,1996.

l ABB CENO AMDATA (J P Lareau) letter 96-3-9038T Revision 1, June 14,1996, 7

{

" POD Assessment for NDE of Sleeves" 8

l Aptech Engineering Services, Inc. Report AES 96022668-1-1, June 1996, "A review I

of the Significance of Sleeve to Tube Weld Discontinuities at Prairie Island Unit 1"

)

Prairie Island DESIGN BASES DOCUMENT ACCIDENT ANALYSIS, DBD-TOP-01, Rev.1, Sections 3.4.4 and 5.2.1

]

Prairie Island FSAR Section 14.2.5," Rupture of a Steam Pipe", page 14.2-34d, amendment 33,4-9-73 l

1 SLV90 DAY. DOC m

+

a e

i i

Request for Additional Information, dated March 26,1936 on Steam Generator Inspection Report, dated February 28,1996 1.

Discuss the nature of the inside diameter single axial indication 2 inches above the first support plate on the cold leg side of steam generator No.11 (refer to page 1 of Attachment 1). Discuss your plans for addressing the root cause of this indication. Discuss industry experience, if known, with l

this form of degradation.

Answer:

l i

j The single axial indication (SAI) in tube R2C38 has a length of 0.29". The bobbin data j

i was reviewed from 8/88,1/90,6/91,10/92,5/94, and 1/96 and found the indication was j

present in all six outages with no appreciable change from outage to outage. The 1996 standard + Point, magnetic bias + Point", and 3-coil data on this indication was reviewed and it was concluded that the call made during the outage is correct..

We are not aware of any experience in the industry of cold leg, free span, inside 4

diameter, non-dented cracking phenomenon. However two plausible causes of such a signal are possible, both from manufacturing. One cause is a SLUG which is caused by metal buildup on the piercing mandrel with subsequent fusing of pieces of the metal to the inner wall of the tube and the other is GOUGING which is caused by friction between the sizing mandrel and the inside surface of the tube.

i 1

i i

i i

j

4 1

l June 27,1996 Page 2 2.

Discuss the nature of the " free span and top of tubesheet indications" i

cited on page 5 of Attachment 2. In addition, clarify the nature of the two free span crack-like indications reported on page 6 of Attachment 5.

Discuss the root cause of these indications.

i Answer:

The tubes with free span indications are listed in the table below. These tubes are plugged or sleeved.

i SG Row Col Indication Location Year Root Cause l

11 16 61 SAI TSH.1 to +.2 1996 OD IGA / SCC 11 2

38 sal 01C + 1.9 1996 Manufacture 11 8

4 40%

TSC +.5 1981 Unknown 11 9

3 40%

TSC +.5 1981 Unknown i

11 9

5 46%

TSC +.4 1981 Unknown l

11 16 68 89 %

07C + 2.5 1981 Unknown j

11 19 47 90%

07H + 2.5 1981 Unknown 1

12 14 38 SAI TSH +.5 to +.8 1996 OD IGA / SCC 21 44 60 45%

TSC + 6 1982 Unknown 21 6

80 48%

TSC + 1.6 1992 Unknown 22 35 78 46 %

TSC + 12 1980 Unknown The TSH indications are located in regions where hard sludge has been visually identified on the top of the tubesheet and where industry experience has found OD IGA / SCC due to the concentrating effects of the sludge.

The top of the tubesheet was visually inspected at the TSC indications locations in 11 steam generator in 1990. There were no loose parts visible at that location.

The indication in 21 steam generator at R6C80 was tracked by bobbin since 1985. It was classified as volumetric in nature by RPC when it was plugged in 1992 due to the bobbin call.

The two free span indications on Page 6 of Attachment 5 are the 1996 indicaticris at TSH in the above table. The correct length is 0.3 inches.

---

a

9 June 27,1996 Page 3 3.

Clarify the use of the UT-360 System in dispositioning volumetric indications in sleeve welds. That is, clarify if depth sizing of indications was performed only for indications which (1) were acceptable by visual i

examination; (2) were acceptable by installation UT examination; (3) were acceptable by bobbin coil inspection; (4) were physically located at the upper edge of the weld; (5) were located in a weld with an average weld height of greater than 0.080 inches; and (6) were volumetric in nature.

Answer The UT-360 system was used in a conservative manner as supplementary information only. No sleeve was left in service based on UT-360 data alone. Two sleeves were conservatively plugged because the UT-360 data gave an indication of a deep inside diameter cavity while the remaining inspection results supported leaving the sleeves in service.

To clarify the use of Framatome UT-360 System, depth sizing was performed on a few volumetric indications left in service which (1) were acceptable by the visual examination; (2) were acceptable by the installation UT examination; (3) were acceptable by bobbin coil inspection; (4) were physically located at the upper edge of the weld by ET; (5) were located in a weld with an average weld height of greater than 0.080 inches; and (6) were volumetric in nature.

If depth sizing was performed for other indications, provide the depth sizing qualification data for the UT-360 System.

Answer:

No other indications were depth sized with the Framatome UT-360 System.

June 27,1996 Page 4 4.

Clarify what is meant by the minimum average weld height (when referring to the height of the weld). Provide examples of how this quantity was l

calculated.

l Answer:

1 A hypothetical example of an neceptable weld is shown in the attached Hypothetical Acceptable Average We!d Height Example where blank spaces indicate lack of fusion.

An example of the determination of the average weld height is shown on the attached AMDATA ultrasonic examination results for the sleeve in R8C48. The upper edge of the weld is selected as 0.14 inches on the average. The lower edge of the weld is selected as 0.03 inches on the average. The difference between the twc elevations, 0.11 inches, is the average weld height.

A second example of the detarmination of the average weld height is shown on the attached AMDATA ultrasonic exemination results for the sleeve in R16C28. The upper edge of the weld is selected as 0.15 inches on the average. The lower edge of the weld is selected as 0.01 inches on the average. The difference between the two elevations,0.14 inches, is the average weld height.

I

1 1

I June 27,1996 Page 5 1

5.

Provide a summary of the procedure and your basis for the use of eddy l

current examination to determine the distance from the actual bottom of 1

the weld to the actual bottom of the volumetric indication.

i l

Answer:

i j

The basis for using the ET to locate the indications with respect to the weld centerline i

was due to the fact that it was the only examination method to detect all the indications i

and was therefore the logical method to build a foundation around. There was not an i

attempt made to determi% the distance from the actual bottom of the weld to the actual l

bottom of the volumetric indication, but rather whether the indication was above or j

below the centerline of the weld. The assessment was based on the preproduction weld O.D. suckback samples which were located either at the top or the bottom of the j

weld. The method to document the centerline of the weld and the centerline of the I

indication was co-developed between Dick Marlow of Rockridge and Scott Redner of

{

NSP and reviewed by the NRC on February 24,1996 during inspection activities at Rockridge, Inc. The method was further documented in a procedural form (enclosed) and distributed on March 8,1996 to Mike Sears of Com-Ed, Dick Marlow of Rockridge, l

and Jack Lareau of ABBCE for use in ABBCE sleeve data reviews of operating units j

and to be utilized on the 200 tube laboratory test program underway at ABBCE in j

support of an Appendix H qualification. The laboratory test program showed that the eddy current + Point" data could consistently assess the correct location of indications.

i l

t i

I I

.. - -. ~... - -

- - - - _ - - - - - _ ~ - -

1 f

i

'l L

l l

1 I

I i

l l

1 r

Hypothetical Acceptable Average Weld Height Example

__I j i l i i l l I I I F FTT I-l TliI I I I I I I

- ' ~ ~

l l

_._ _ _ l _ _.

l l

l l l l l l l l

~~ -

l I i i l l I l-1 I I I I I I I I I I I

(

j

} { { } } } } }

} } Average Weld Height = l.101 linch l l l l l l l l l

l 1

i l

l l

l 8

6 l

WELDHT.XLS i

a

(

I

9 x

e a

pucaAc( UElp RE ICST D E TCR r1 M M Iad j

j r

8 a

a

&w L/ul j1 STO e Cr( tJ Qgn A sleue RJ C W m

ett@ingst :;?g:w g

ChannelkG)ated=p&ysQffyqfQQf-x., : m-um u--dunme;;B-tSgancarih --hwnes mm.

Help 9ffY 2Q hQ

.:s

p.mac.b.1 F,ile

" :3.0;,. m. m %..

-4

..a n.v.w y w7p p am m v63 ise#y$f))!24 1g r8c48old W

.y$Ekamj;DNtis;f024 g9.6 Mii.tmP

..go 2

971.i0.JdB,iDY.L;%

Fifid'$WtifaxF d

NOM File:

e GaEef Sil69.W{WUVdNiSDi!

~ hdQ*

.Dh $ $"Sp?HP:

i31.

ip.

Channel:

c?pidF'F

!!P +

~ 2. O Ga'in:

' De P ul s eWi'd tb : h iMAlijif6W$s*3?d O(isiff5$3ldc80hPnis%Y:aT$

. Chime}n'20.0%

2 AdisidOVD0?u '

T TJ,

- h)"MM$%

Cordpr$isi'on!' 0FF"M4M d'dM'1'6%fnIslIN

• $:."SPNI"K71802 Transducer Ang:

0~.O'O.deg' Frect: 15.00?MlHNSIEE ' r$.

i 0"

Yf....

d O p e ra t o r :... t;i JD g.aNyW NSQ%p..fd@8}y@k$N@j%g

]a4 p

dtiSNNSfssF6 RGbbMS Coinidii's6t@Y ~ RO'ThTEliy$$T'0BU[!

F'rcWifd!f ~ 00' $'.e Stea Ged

'0DOFFset:

0 TB:q~pgp%...,

3 in

$l}.$@0RFiet:

SchfE 7 Asis R

~0 0.00 TO F

'ThWNfEl*dM:i$d718i$9W:kh[K*io Yd.Hy0i$

..__$radg*

InddxYAxis1InBEnfN?

l?y. "y "

R U$h~bMe a

1 e

ye v%.>.

e.n..

• 9 W;pecu M;eg W

,m:g:n m

~

... a..s.,.u. &.;.,. w.. %.;. m.

1 w

v. :..

__,) -

..:......-..c...,,.n,.,

mw.o 3yn. zyn m,

v.,

ggn,

_g.m ;

n. m.,,,.. -

b@i122 V

.df B

3

=.

3 P'eak

(%)

i 'm W & E GR%..a g ___7 y.nas j aary = =

. l mi 13==k c

.t n

w%.120.

..gg 0 112

• ~

@'/n8bi, -

t

" #naL __c:

a

. a_ annammu.Jput r

's====qt

^; ^ ; ;7i 40;

+

>3

~~

C e i J / g. v v e

__ =

u'u

= ---- n-n 2-- -- -

m,p.

01

' L "lTr qi

.-~

mmmearm,wr

..aum.w.- _.,.:: e:.. _ _ _ _ _ _

m M

n' x

)p NO h h h h i$i M i(( h {3f[(l%%f k g g) g g }.lgj;h R'OTATE PROBE 4 l Nnminid(d1996' AE$%71D*hTdNMM4 OS$%3N@MMEjf@@MfdiMM l'

9 E

AuCMGC La CL D p(l& [; y p(pfgh y/p y)yn

y. j s

A 3

a Pra ia < fs ka

12. STEM 1 G-cWE44tbR sLcCVE, A TR Il,C 2 8 Chanr[el;[NhkIb ffbhfff?

f ff,!{IffQMgs gelp File File: r16c28P22

. Exam Dste: A01@27M6MIllh&:?.13t[3s2413:39 O F F.1 '. 0 F F s'e t :..; 0T OWdlO'MId e$j; SW 1.* Mode:. Max.FrVfde6%

0PMVJidEAJFilt$r: 2 Channel: 1 Gate:

Phlier Gain: 74.0 dB Dac:

Sq.,ldaVe;.03Xffhst iCwdth':'g0l.15 rus:50&fijliiers LP: 5 HP: 2.0 dDampi;ng:

Pulser. Type:

Pulse Width: 80 Cdly:it;0 Cthrsh: 20.0%

Adly: 0.00 us Awdth: 3.00 us Compression

• OFF Thresh: 19%

I-Gate:'; ION; LAD! Rate:?1'O'0jMsps Transducer Ang: 0.00 deg Freq: 15.00-:MHz

' Skew Ang: 0.00 deg S/H: K71804 Operator: mJd Component: Steam Generator Sleeved Tube Ser Ho: N/at Scan Axis: ROTATE PROBE 4 From:

0.000:

To:-

362.000 Intv: 1,000 0FFset:

0 Index Axis: Index From:

0.000 To:-

0.300.Intv:.0,020 0FFset:

0.00 This Weld has been:

I Accept 6dE g.

_ ' Rejected;p

............................... MJ Dashuliesii ch, UT); Lev er* IIIlAlp, 2 f)

A

~

.,.~,..,c.-

ROT ATE PROB Eu 4 : ;. 149. 000, HIndek hip 01100),1.MMj 'h :

~

~

~~

AMP:~

1% -TOF:1:0.1400Lus; adBi@32V6MMPtM0/0'46 sin,Dpt 0._046 @ 92'Vf Peak (%) 5 i I 0.12 $v.wcW si si 120 0'06 i i n ,,,gThYM 80 d O 002 ,,,li,, -31 i i' % 5 = === '= i, ii A0, pe f.A 0 lil x 0 5O ib0 1b0 2bO' 2b^0' ' 3b0 3bO( <- ROTATE PROBE 4 Copyright (C)1996 ABB AMDATA s

Attichment 1 Question 5 Page 1 ABBCE SLEEVE WELD CENTERLINE DETERMINATION l 1.0 Observe upperjoint at a zoom setting of approximately 10 l displaying channel 3 '(75 KHZ axial sensitive) horizontal on the left chart and channe! P3 (75 KHZ circ. sensitive) vertical on the right chart. 2.0 Find the upper expansion transition (UET) and the lower expansion transition (LET) signals and select the center point j of the weld excursion signal between them. Label that point as i l " WELD-CL". Measure in both the positive and negative l-direction to the peak of the excursion to verify you have j assigned the centerline properly. Adjust as necessary. See Attachment #1. 3.0 Make sure your measurements read negative toward the tube i end and positive toward the top of the sleeve. These values will be different on a push verses a pull. You must change the i from hot too cold in the operator selectables if you are measunng incorrectly. ABBCE SLEEVE INDICATION CENTERLINE DETERMINATION 1.0 Observe the C-scan display of channel P3 with a Axial l Average Filter applied. Place the arrow in the circ. position of the peak of the signal by scrolling in the single circ. line box while holding down the right mouse button. See Attachment

1. 2.

2.0 Observe the lissajous display while scrolling axially trough the indication by holding down the right mouse button and dragging the mouse in the single axial line box. Count the number of hits you observe in the expanded strip chart display from top to bottom. On an indication with an odd number of hits select the center hit. On an indication with an even number of hits always select the center hit as the middle hit closest to the tube end. See Attachment #3.

r

ltIi;Jl!!

l 1J g$oTg3* q$=@oU 1 yTt o tl. l hh e s { C i o I O C i i s l I n o W i. C. 10 W i. j c 8 I TP 8 ( 1 5 + 1 a B h s sI e d n es s A eeO S C l 3 C I c i 2 f y tf e 9 I a 4 m ] B L 0 l a d O S 0 d e t e C C 1 se Ce 3 n m e et ( <s n 1 e I o V L a po W e B W K e e t tt e e lot a ut W'. d O N s n d a e o f = h a e m d P e ip C 1/ ll s y 2 e { l U te e s e 1 t e d s I 1 a s n 3 I S e P t h B e 0 D .w / l @ 3 e C L 4 d 8 .m 0 d 9 7 'P.. n. ie a a 8 A R n o s a 2 n ' m m o 1 o e t A "k e e< e t t N s A W t C M u e u A c s T> 9 o J C. re 1 y t i Z:~. t l L-4: 5O T V 9 a 0 Pi ( n w i n e e C r b O e w 1 d t C 'N* e. ( c y 1 7 C e d s 2 d d < ly l 2 1l s d De is e E3 e E e t e F 3 0 e a e f-F P C 0> n 0n A 9 e 1 d e 9t t

==' C j1 M n 2i e n 0 s n m a 0 u t re e = 0 t y s 1 r 6 d e e s A a y y y s t - d S-a' 1 H ?C h'* d C 2 6 6 s E g7 n Pk

• 1G ll 7

e s t e e ts n S 3 2 0 4 y 1 C m le e A-I a n p d a* U r' m e T i < a n a t _k D "' E r i E i a o _a e 9 o e o 1 = v L l L D' F-U e . # M N t r t n l F I J 1 l l j 1C o l mf ? 'l e" f L Y 0 l 3 t t i l f l 3t h s 1 pg l 5 l l l

C.

g = a a I 9 3 $^ m mf P. O 1 c l l J l" / Ji 5 t y1j]i s\\d l 1 a 3 lj2 U M_ M l <"9-c V n, t a r m [ u u t u n* t g i o l m f) ih i 2 i (G A l a ll M n 2 s t 11 il m 9 t t 4 l i C n 3 5 p 1 3 w n p V R t u b 21 l i. C l S E= t 9 ta 10 s 0 I DL e 0 s 0 A s s 1 C1 e 2 M u 1 iY t r 2 s t 1 tc l C i e S m a

s c

Y n a A s r t n L u V P fa a S t D S 0 t F 63 C [j a tf i l C s s N u _. t n P n a C e-k t - / A s ~ P si F m l G i n li re p t n p u v F i

e Question 5 Page 3 L _2 _E y lTV M 1 El y g L /s b e i i' s ,i h..I 3 8 y i e g I. 3 t 2-g I j 2 ) g r 5e 9 a u 2 3 g 3 3 8 4 = 4 s as l E [ }-8 8 =11 t e i s 3 5 4 qv 5 I k I N 3 = .A 2. 8

• 'o t,

J g 2 s ; i ng r ~ i L 7 4: X RG 2,5 3 "8 h h a if i i 1 In s 8 Es o- --7 g [ !

• z s l at s - 7 T

'] T EE $2 l2 l0 (( ~ ~T ?.y fil a e s a. 6Eh h Mi h- ~ 30 M t -- s -.m.- # D g c., ? a 3 D p g (8 g ~ a* s i O g f 1 i g : M s 2 su 5 3, 3 = n "Y a y 5 I 5 3 i o i s 2 I 4 h g 3 i /- a 4* z4 i hc e_ t a 5 x 4' E n > 1 e a ; g

. 2 c

-c. g w / $ 5 E. b bch 5 ,) d ~ ~ m4.( V 52 = a 5 i s i w a = s i t, u, \\ -= O \\ a m s \\j T ~E C OT \\ 5 ) ~ A-g

i, I 3-A I [ l S.h$ I I I 1 ?.A gf [s d f O e ( m / l g n 0 I /1f*] e .a, n wo C ) wM r, = lp M ( E gW e e t-C.

1. I' 7 l

a l i t P 5 e 1 a + 2 I n g.) t a 3 N i# e .,8 d 1 A c g e .( 8 p 3 C C M E Y-a 0 P + 4 c ) ? l n y y e L e h O e s r c g w a et ( C C e t 3 i e e e 0 D P i t. a c t 4 d t s. e e e N S d n le t A 3 . =. la 9 O a f e I :- b i g y T e c 0 0 A r E n d< e 1 t B =t kW e e A 0 u E

W m

M t t. c C. n N .mt o k i> 9 o F O 1 y R t v t a ] L-E P "I t C n t a,~ n C 8 i e O e is e y ( 5 5 o t 7 C d e d o0 te 3 l1 p s E t 3 s le T = M< la e A 3 o y s A O> n f-S P c n O0 A 8 n n, A ll w is c t s 21 3 s t 1 S = e n 0 U e m L P te y e e O L6 la e t t n s A a t a y y C s d S-o t 5 1 n ne a, d C t 7 C 2 n E 2 2 t r e e 1 li 5 5 n a S is e C e e s S 3 0 3 r e ly s 1 n G n a m o n } m k a y A-v }3 m. s 4 ee n a i a tc t 5 e E D e, e F s h I 5 L e t e u o e e D F-e u t e 5 t r t e e s r r I r i ,f M I m f j E u C N ' g Fl e E i f C m. e l 1I G o

C.

l VKi m l s ,D My O i g P p M f 3j - i , :1! js C o l L H I g [ d 7 g i 5 O 5 i o, g4 C 1 E y 2 l t A t p w4

m. T I

ll l ] u L e L 0 X C M A t s [ t f I wt g 8 I i D l 3 12 1 a t 0 l 2 I l 4 5 M C 0 4 3 1 I p M v R to 2 HZ 1 G 5 S 6 t 4 tM e 3 1 i 0O 0 L 0 A =s 0 1 C e ( 1 i m t r 2 1 T 1 C L S m l G E A C Vt es a. 4 A e I M W 0 L r 1 P fR 2 4 S L I 5 D C Y l C 9 1N j L i n so P e C t f c S t A /s l m t t i 5 l P e i l w l [.i t l FG I L C O V re E w e / - t F

i i l REQUEST FOR ADDITIONAL INFORMATION, dated May 6,1996 l

Reference:

CEN-628-P Revision 01-P l 1. VT l What are the process control feedback criteria that are used to determine the circumstances under which a visual inspection (VT-1), will be l performed? (Section 8.3.1.1 of CEN-628-P, Revision 0, indicates that a VT-1 inspection is not always performed.) i l ANSWER: l In the past post weld visual inspections were not required for all sleeve f installations. The field procedure required VT at the request of the process i engineer, or to resolve questions from the UT results. Also, all repair welds j required a VT. (Some utilities also implemented a program to VT all tubes with j either ET or UT indications, for both new and previously installed sleeves, for example, Zion 1, NCR #6, November,1995). In addition, ANO 2 (1995) and Prairie Island (1996) required post weld VT for all new installations. ) (Please note that the statement on VT in the second paragraph on page 8-8 of j the referenced report should read: All installed sleeves with WZI results from i the plus-point eddy current test were visually inspected at Zion 1 and 2; all l sleeves were visually inspected at ANO and for the 1996 Prairie Island j campaign....) There will be a VT-1 inspection done in accordance with the ASME Code Section XI on new welds. The VT-1 inspection program will be reduced to the previous procedure requirements following completion of additional qualification 3 work to establish the adequacy of the eddy current and ultrasonic tests. j j 2. UT FUSION, COMPLEMENTARY NDE If one ultrasonic testing (UT) pass over a weld indicates complete fusion for the width of the UT pickup, it is considered an acceptable weld. Based on this acceptance criterion, discuss the basis for leaving indications in service given that the continuous fusion could be above and/or below the weld centerlins. Given that the different techniques (visual inspection, ultrasonic testing, eddy current testing (ECT)) are complementary rather than supplementary, discuss the ability to align the VT, UT, and ECT results (i.e., if ECT detects an indication, how is it known whether this indication is above or below the continuous fusion path given that the continuous fusion path may be above or below the weld centerline). Provide v2o6 an st.voutomoc Page 1 of 9

i i supporting metallurgical data. ANSWER: The UT results are used to confirm a leak tight path around the tube / sleeve circumference. The stated concem is that, in the presence of a weld imperfection, the fusion could be above the imperfection. Either UT, ET, or VT is self sufficient to establish whether its identified imperfection is above or below the pressure boundary. VT is used for surface detectable conditions such as blow holes and incomplete welds and for these cases the image is self evident regarding relative location. UT is used for complete lack of fusion through the entire height of the weld, but can also be used to detect blow holes and can demonstrate whether there is a weld fusion path below the detected blow hole from the B and C scan images. Using the Appendix H qualification process, the ~ ET method has been used to correctly locate imperfections relative to the pressure boundary using laboratory samples. The Appendix H qualification report will be provided to the NRC as part of the revised topical reports to be submitted later this s,ummer. The results of this qualification are summarized in the attached letter,96-3-9038T Rev 1, dated June 14,1996. In this study, the correct location was assigned in 23/24 (95.8%) instances for all flaws and 17/17 (100%) instances for flaws exceeding 40% throughwall in the sleeve. These results are tabulated in Table 3 " Indication / Flaw Location Evaluation", in the attached letter. As described in Section 5 of the referenced document, both the suckback and inclusion imperfections have been shown to originate at either the upper or lower edge of the weld, which reduces ambiguity in the ET analysis. The interrelation among the NDE methods is depicted in the attached Sleeve NDE Process Flow Charts, Figures 2.1 and 2.2. The VT is used only for confirmation that a signal is associated with a surface detectable condition. UT is used to confirm the presence of a bond below the UT indication of a blow hole so no position relationship with VT is required. ET methods are used to locate flaws above or below the weld centerline in a stand alone manner. The enhanced UT analysis guidelines (attached) have been implemented for future installations to improve the ability to detect Lack of Fusion (LOF) throughout the weld height. 3. MISSED L ACK OF FUSION BY UT The UT field data analysis missed several unbonded regions (e.g., inclusions)in the Prairie Island pulled tubes. This observation required further modifications to the UT screening criteria (i.e., the voltage thresholds were adjusted and the types of scans analyzed were modified). (a) Given these modifications, discuss the basis for not repeating the review of all previously obtained UT data to ensure the tubes have adequate bonding. (b) Given that different types of scans may be needed to M4N RPP SLVRA110. DOC Page 2 of 9

j detect such unbonded regions (e.g., inclusions), discuss the basis for not reperforming the baseline UT examination for all inservice sleeves. l It was noted in Section 8.3.1.4 of CEN-628-P, Revision 0, that historical UT l and ET techniques were capable of detecting all rejectable welds (16 of 16). j If this is the case, (c) discuss why the unbonded regions of several tubes were missed at Prairie Island. (d) Discuss the possibility that the sensitivity of the analysts was heightened as a result of recent industry i experience therebJ resulting in the detection of these defects. (e) Given { that recent industry experience may have potentially heightened the l analysts' sensitivity, discuss the need for reanalyzing all historic ECT and l, UT data. l ANSWERS: ) (a) Given these modifications, discuss the basis for not repeating the j review of all previously obtained UT data to ensure the tubes have adequate bonding. To the extent possible, UT data was re-analyzed for all sleeves with WZI calls from the plus-point eddy current inspection. An evaluation of the NDE methods in use at the time of the inspections was evaluated and reported in the i referenced report and updated in ABB CE letter 96-3-9038T, which is attached. 1 This assessment indicates that there is only a small (3.3%) probability of missing a lack of fusion condition. From another perspective, the data set including i Prsirie Island removed tubes and intentionally faulted welds has only 5% of the welds with LOF conditions without associated radial inclusions segments that would leM to a WZI call (2/40 samples in the table in the attached letter). j Overall, the probability of producing these conditions is low and the probability of l nondetection is even lower. Of the LOF conditions not detected, even though a l WZl was reported, the affected area was less than 30 degrees of arc. For this type of condition, the measured leak rate in lab samples was less than 0.0027 l gpm as reported in Table 6.1 in the referenced report. Taking into consideration j the plant leak monitoring requirements in the technical specifications, our j conclusion is that there is sufficient defense in depth without re-analyzing UT l data. l l (b) Given that different types of scans may be needed to detect such unbonded regions (e.g., inclusions), discuss the basis for not reperforming i the baseline UT examination for all inservice sleeves. 1 The probability of detection (POD) was evaluated extensively in the i laboratory program reported in CEN-628-P, Rev.1, and updated in ABB CENO l letter 96-3-9038T, April 30,1996. In the former,16 metallographic results were available and in the latter, sixty four sections were completed. For this study, j both the historical NDE practices previously used and the enhanced NDE 6m6 RPP SLVRA110. DOC Page 3 of 9

i methods using detailed analysis guidelines were evaluated for POD. The results are summarized below. NDE POD STUDY HISTORICAL ENHANCED 4 j METHOD METHOD POD % POD % 4 UT ALONE, LOF 87.5 100 i ET ALONE, ALL FLAWS 87.5 90 ~ ET ALONE, FLAWS >40% 90 93.3 OR LOF COMBINED ET AND UT 96.7 100 FLAWS >40% OR LOF These POD results exceed the industry standard of Appendix H for both the historical and enhanced methods. The 96.7% POD for the historical methods was factored into the operability assessments to estimate the numt:er of effected tubes and these were handled conservatively, assigning the maximum observed possible leak rate to all effected tubes. All CE sleeves currently in service have been inspected by both UT and ET using the + Point coil. At the next inspection outage, a 100% inspection of all sleeves with the plus-point probe is planned with re-inspection or re-analysis of digital UT data for all WZI calls that are left in service based on a location determination above the weld centerline pressure boundary. (WZA results per Figure 2.2) Given the high POD and minimal consequences of a missed imperfection of the types observed from the pulled tubes, additional re-analysis of historic ET and UT data is not warranted. (c) discuss why the unbonded regions of several tubes were missed at Prairie Island. Two of the sleeves (R7C52 and R9C57) removed from Prairie Island had lack of fusion that could create a leak path that was not detected by the UT method. However, both of these tubes had ET indications that would have resulted in a decision to repair the tube. Metallographic analysis reported in CEN-628-P Rev. 1 indicated that these tubes had local, narrow fingers of unbond approximately can are st.vunomoc Page 4 of 9

i i l 5-20 degrees wide. The conclusion of the analysis was that an indigenous oxide i layer was left on the tube wall which remained in the weld. This oxide film could i either remain laminar or break up and " float" in the weld puddle with a resultant radial extent into the sleeve weld portion. If the oxide remains predominantly laminar, the UT method detects this condition. If the film breaks up into fingers, j some may remain laminar and others become radial. The radial fingers were j shown to be detected by the ET methods. Therefore, the condition that may j lead to a miss by UT increases the detectability by ET, and in combination the j techniques are capable of detecting these imperfections. i i (d) Discuss the possibility that the sensitivity of the analysts was i heightened as a result of recent industry experience thereby resulting in the detection of these defects. The detection of the previously undetected unbonds at Prairie Island was not due to heightened sensitivity, but rather by an additional analysis technique using amplitude based B scans. The earlier method of threshold crossing C scans does not provide any objective evidence of an unbond for these instances with narrow, oxide included LOF regions. (e) Given that recent industry experience may have potentially heightened the analysts' sensitivity, discuss the need for reanalyzing all historic ECT and UT' data. NSP was aware of the degradation of the parent tubing in the HEJ sleeves, recent experience with the + Point probe in CE sleeves (volumetric and geometric indications) and was set up to look for indications in the parent tube of the sleeve upper joint region, particularly at the expansion zone transitions. Analysts were steadily gaining experience with the new + Point probe which probably heightened their general awareness. Analysts were aware of the volumetric indications which were found at Prairie Island in 1994. The four circumferential indications found with the + Point probably gave a heightened awareness because it was a new phenomenon at Prairie Island. NSP is confident that the use of independent analysis teams for the ECT examination provides a high level of confidence that significant indications have been identified as provided in the Probability of Detection discussion above. A review of the available UT data for the Prairie Island tubes left in service with WZI reports (in this case, VOL indications above the weld centerline) was conducted. In this review, four additional tubes (one of which is plugged) with possible unbonded regions were detected using the enhanced B scan analysis techniques. These four are similar to the two LOF indications in sleeves removed from Prairie Island in so far as the LOF is in narrow fingers less than 30 degrees and there was a + Point coil detection of a WZl. This data has not been used in the determination of the POD for two reasons: 1)the true condition is not known through independent means and 2) there has been no assessment of the 6/24/96 RPP SLVRAll0 DOC Page 5 of 9

I i i number of correct detections of LOF with field data for tubes repaired and left in service. Only laboratory samples and pulled tubes have been used for the POD l analysis. Given the high POD as discussed in (b) above and the minimal consequences of a missed imperfection of the types observed from the pulled tubes, additional re-analysis of historic ET and UT data is not warranted. l 4. FLOW CHART { Clarify figure 8.3.1.3.A of CEN 528-P, Revision 0. If this flow chart had been j used, all volumetric indications should have been removed from service. l ANSWER: )- The error noted in the flow chart was corrected in CEN 628-P, Rev.1 which was j transmitted to the NRC by Northern States Power on May 3,1996. A copy of the l corrected flow chart "ECT Baseline New Practice" is attached. i i 5. MAGNETIC BIAS PROBES Discuss the qualification data supporting the use of the magnetic bias plus i point in dispositioning signals obtained with the non-magnetically biased l plus point. l ANSWER: } Th'e qualification study undertaken did not specifically address magnetically l biased probes as an essential variable. The magnet field strength to completely i saturate inconel materials is approximately 5 KiloGauss (per C.V. Dodd) which l would reduce the eddy current depth of penetration and suppress the eddy l current field to a point where it would be considered a no test with the current coil design. The Zetec manufactured magnetically biased plus point probes used } to date contain four (4) Super Neodynium post type permanent magnets which generate a saturation field between 0.1 and 0.15 Kilo Gauss as measured i, between the poles. With this weak field there is a slight reduction in the eddy i current depth of penetration and slight focusing of the eddy current field, but these variables are not considered as essential variables. The variations in the eddy current field have been accounted for during the analysis calibration setup by normalizing the span setting on the same calibration discontinuity and verifying detection of all calibration standard discontinuities. The magnetically biased probes in sleeve examinations have been only used in a supplemental or diagnostic capacity to date due to physical constraints on the manufacturers ability to both articulate the probe head and deploy permanent magnets. In the future, the same examination methodology will be utilized or, if these manufacturing constraints are overcome, examinations will be performed exclusively with magnetically biased probes. uww m st.vutovoc Page 6 of 9

6. ET AND UT CAPABILITY Discuss the nature of the two rejectable welds missed during UT examination (Section 8.3.1.4 of CEN-628-P, Revision 0). Discuss the nature of the two rejectable welds missed during ET examination (Section 8.3.1.4 of CEN-628-P, Revision 0). Discuss any conclusions which can be drawn i with respect to the capability and/or limitations of the two non-destructive examination methods based on these results (i.e., the types of defects that can be detected and/or missed). i l ANSWER: i The two welds with LOF missed during the field UT examinations at Prairie l Island were described in the response to Question 3. The two welds with a LOF that were not detected by the + Point eddy current coil were laboratory samples. In these samples, the oxide inclusion remained laminar rather than curling and { becoming predominantly radial, and, thus, as expected, were not detectable by { the ET method. The C scans using the amplitude threshold method (historical i method) are attached. The poor quality of the weld and the LOF are clearly i detectable for this situation (samples OS-15 and OS-17 in the table in the l attached letter). For comparison, a typical good weld C scan is also provided l (CT-36). The overall NDE capability was addressed in the response to Question ) 3. i 7. DATA COMPARISON AMONG PLANTS l In several plant assessments, a comparison is made to the indications at Prairie Island. This comparison is made with respect to both voltage and j are length. Discuss the sizing accuracy of the NDE techniques. If the { sizing accuracy is limited, discuss the limitations of this comparison. i Provide the supporting technical justification for the responses. } j lt was indicated that work is on-going with respect to using voltage to assess certain forms of weld degradation. In the plant's assessments, i discuss if voltage was used at all to assess the severity of degradation. If l It was, discuss how voltage is related to the severity of the degradation for all forms of degradation for which voltage was used to assess the severity of the degradation. I j ANSWER: j in the NDE assessment, a standardized method for reporting indications was i used for arc length and amplitude. The arc length was done using the EPRI l guidelines issued on January 15,1996 as part of the ODSCC programs. Various amplitude measurement techniques had been used, but a normalization i sww m st.vwmooc Page 7 of 9

-- _. _ -. ~. _ _ - -. _ - - - - - - - factor was applied to standardize these so that a relative comparison could be made among the signals from the various plants. In a plot of amplitudes of weld zone indications, tube R7C52 removed from l Prairie Island had the highest amplitude compared to both the lab samples l produced and tubes left in service at the other plants. A plot of amplitude versus maximum radial degradation depth is provided in Figure 7.1 for Prairie Island pulled tubes and the Appendix H data ut and this plot indicates that the laboratory samples were representative of the actual field responses. The comparisons were made solely on the basis of indication amplitude under the assumption that for the imperfection type encountered, more severe conditions would produce larger indications. There has not been an attempt to use the l' depth to voltage correlation as a direct assessment of degradation, only relative comparison among plant signals has been made. In the operability assessment, the conservative approach of assigning the highest potential leak rate to all tubes adds margin to the analysis. i 8. NDD REANALYSIS Some tubes that were originally characterized as having no detectable degradation (NDD) were reevaluated as having volumetric indications. Discuss whether the data for all the NDD tubes were reevaluated in this l analysis or just this one tube. If only a limited sample was reevaluated, discuss the basis for not expanding the reevaluation given the possibility that larger undetected defects could exist. ANSWER: Only one NDD tube was reanalyzed. The basis for not expanding the reanalysis was given in the answer to Question 3 and is based on the high POD, the estimated correction for potentially missed flaws and the bounding of the consequences of combined flaws based on the PI pulled tubes. At the writing of the technical report, the NDD tube at Zion 2 was compared to various signals from a suckback condition and an estimate of 10 to 15% depth was made in CEN-628-P Rev 01-P. Subsequently, this tube has been compared to various tubes with radial inclusions also and the signal is comparable to one sample with a 43% inclusion condition, which is still considerably less that the initial l allowable degradation calculation of 63.3% provided in CEN-628-P Rev 01_P l (page 7-9). 9. Some tubes had noisy C scan displays. Discuss the limits that are ( placed on the amount of noise in the data. f l 1 saw m st.vunomoc Page 8 of 9 l

ANSWER: Two noisy C Scans were noted in the ANO 2 review. This condition causes i i false positive indications of LOF that requires additional analysis using B scans to resolve. This condition is easily detected, and resolved by the analyst, and { does not pose any technical concern over the quality of the data. 1

10. Since the root cause of these indications has been attributed to the cleaning process, additional weld indications would not be expected. As a j

result, a reporting requirement to notify the staff when indications are detected in newly installed sleeves in the weld area would seem appropriate. Such indications may indicate a breakdown in the sleeving 3 l process. Discuss the appropriateness of such a reporting requirement. !~ l ANSWER: It is appropriate for the next sleeving campaign at each affected utility to provide the following report concerning the installation of Combustion Engineering welded sleeves withi.n 90 days following completion of the installation.

1. Number of sleeves successfully installed

2. Number of sleeves installed and then rejected and the reasons for rejection i

WM96 RPP SLVRAH0 DOC Page 9 of 9

Richard Pearson June 14,19% Northern States Power 96-3-9038T Praine Island Station Revision 01 1717 Wakondale Drive East Welch,MN 55089 FAX: 612-330-5743 cc: Mike Sears Com Ed (7084634299) DarolHamson ANO (501458-4685) SheriBernhoA WPS (414-433-5544) David Stepnick ABB

SUBJECT:

POD Assessment for NDE of Sleeves As a follow up to previous meetings and discussions on the assessment of the NDE methods to detect faulted sleeve welds, the attached write up is provided. Both the historical and enhanced methods of NDE have been evaluated relative to all the available metallography results. In the previous paesentation to the NRC only sixteen tubes had metallography information available at that time. We have now completed all planned metallography and a total of 40 faulted sleeve welds and 24 good welds have data for this compenson. The results of the cap r44d sample set are very cocststent with the original nue<= ment and these results do not change any of the earlier conclusions. This information is provided to support the individual plant operability evaluations. If we can be of any additional assistance, please feel free to call tric in Windsor. Sincerely, J. P. Lareau SeniorTechnical Advisor ABB CENO

4 i i 1 i l 4 NDE ASSESSMENT l SLEEVE WELD INSPECTION J Lareau ABB CE April 28,1996 i i { COMPARISON OF ET AND UT DATA TO AVAILABLE METALLOGRAPHY RESULTS HISTORICAL NDE EVALUATION The initial evaluation of the historical NDE process used for sleeve inspections reported that the j Probability Of Detection (POD) for the ET Plus Point inspection was 87.5% (14/16) and the IJr method

~

for detecting leak paths was also 87.5% (14/16). A statistically calculated combined POD would be { 98.4% and the actual observed POD was 100% for the sampic set of sixteen tubes. Additional metallographic work was performed on a total of 64 tubes to increase the data set. A total of forty (40) sleeve welds were found to have weld imperfections during metallographic sectioning, an additional twenty-four (24) sleeve welds were sectioned and had no weld imperfections. i 1 Of the forty sleeves with imperfections, ranging from 19%.100% through wall or with a leak path, the historical plus point method detected thirty-five (35/40), or a POD of 87.5% for ET. I Of these forty sleeves, a total of thirty (30) had repatrable conditions of either a !cak path or >40% i throughwall extent in the pressure bvouday. (He repair limit as a % of sleeve wall varies from plant to plant, 40% was selected as a conservative limit for this evaluation.) The historical plus point method l detected 27/30, or a POD of 90% for ET for repairable conditions. i Ultrasonic testing was also performed on these forty sleeve welds and of the five sleeves that were NDD, two had a detected leak path indication. The combination of ET and UT detected 37/40 welds with imperfections of all types and sizes, or a POD of 92.5% for ET and UT combined. Considering just the thirty repairable conditions, the combined detection was 29/30 or a combined POD j of 96.7% for repairable conditions. e ENHANCED NDE EVALUATION l t l A formal qualification program in accordance with EPRI Appendix H methods was performed for the plus point ET method for sleeve inspections. In this program, a more conservative requirement of using 40% flaws rather than the specified 60% was used. The primary changes that were developed came in the analysis guidelines, which were also formally qualified and published. All analysts involved in ET data 4 evaluation are receiving spectfic training and are required to pass a qualification exam using these j guidelines. I As part of the Appaa% H qualification, the method used to locate flaws relative to the weld centerline was evaluated. A total of twenty-four sleeve weld flaws were used for this study and the correct call was made in 23/24 cases for a correct call rate of 95.8%. The one error occurred when a large blow hole above j the weld obscured a smaller inclusion (35% through wall) below the weld. When only the repairable e i 1

( \\ i \\ conditions are considered, the correct call rate is 17/17, or 100%. These empirical results are valid for both the historical and enhanced ET methods for locaung flaws. l Overall, the results for the historical and enh=W NDE methods are tabulated below. HISTORICAL ENHANCED METHODS METHODS POD POD (%) (%) l ET ALONE, ALL FLAWS 87.5 90 ET ALONE, FLAWS >40% 90 93.3 COMBINED ET AND UT 96.7 100 FLAWS >40% The historical, combined NDE POD of %.7% is the best estimate of the actual field results to date. To determine the best estimate of the number of Pelly affected tubes in service, the number of reported indications from the plus point inspection should be divided by the POD value and then the number of tubes removed from service should be +A+ =1 As an example, assuming 43 plus point indications, the estimated number of affected tubes would be 43/0.967=44.5 tubes. If the original tubes with indications were plugged, then the remaining potennally affected tubes would be 44.5-43= 1.5 tubes. e l l i

1 POD ANALYSIS FOR PLUS POINT DATA I I APRIL 25,1996 J LAREAU l l 1 l l I TUBE ET CALL MET / LEAK DATA COMMENTS OS-2 W Z1 SB-61% OS-3 i W ZI LEAK /SB 18% OS-10 NDO SB+ 29% i OS-11 i WZI LEAK l 0 S-14 W ZI SB-31% j OS-15 I NDO LEAK UT LOF OS-17 i NDO LEAK UT LOF OS-20 l WZI LEAK OS-24 l WZ1 SB-32% OS-33 l WZ1 LEAK 0 S-34 i WZI LEAK OS-37 WZI LEAK OS-38 WZI LEAK /LOF OS-39 WZl LEAK l OS40 l W ZI LEAK /LOF/SB-13% OS-45 l W Z1 LOF l t I CS-5 i BH BH/ INCL-35% CS-13 W ZI SB+ 22% } CS-22 W Z1 SB/ INCL +/- 60% l OT-1 WZ1 INCL +/- 74% OT-2 W Z1 BH l OT-3 W ZI INCL-79% OT-5 W Z1 INCUSB+/- 38% OT-6 NDO INCL +/- 68% WZI WITH GUIDELINES, UT OK OT-11 NOD INCL +/- 31% NDO WITH GUIDELINES, UT OK OT-12 W Z1 INCUSB+/- 89% OT-16 BH BH l OT-17 W ZI INCL +/- 40% OT-20 WZI INCL +/- 65% OT-21 W Z1 INCUSB+/- 36% OT-23 WZI INCL +/- 19% OT-24 WZI INCL +/- 97% OT-31 W ZI LOF/SB+/- 25% l O-113 W ZI INCL-35% i CT-44 WZI UEX/ CON Pt 5/48 WZI SB/ INCL PI 9/57 W ZI SB/ INCL Pt 7/52 W ZI SB/ INCL Pt 7/63 W ZI SB/ INCL 4 PI 5/74 WZI SB/ INCL INCL

• INCLUSION BH: BLOW HOLE LOF: LACK OF FUSION

+/ ": ABOVE/BELOW WELD CENTERLINE l l WZl: WELD ZONE INDICATIONI UEX/ CON: UNEXPANDED, CONCAVE WELD TABLE 1 PLUS POINT PROBE POD ANALYSIS

POD ANALYSIS--NDD/ OK BY MET APRIL 25,1996 J LAREAU 4 TUBE I l CT-1 l CT-2 CT-3 i CT4 CT4 CT4 CT-30 CT-31 CT-35 CT-36 CT-38 CT-39 CT-41 CT-42 CT-45 CT-46 CT-47 CT-48 117/144A ~ 117/144C 38/44 H1 38/44 H2 i 23/43 H3 1 23/43 H4 4 4 4 N a 5 4 4 4 TABLE 2 NDD BY ET OK BY METALLOGRAPHY

s. ...es.=. -_x..a. .+ ~r.- w .t s-- s e LAB SAMPLES AND PI TUBES s h.' [.

b,

^. T ' fh. k,$.g

^~;[

t EU ?; h ' < :ii .k i i.bi. 5k. 5 p.M:;.fh g:.,h ".;.'i.:hf h h:his:. j-j. '.y g;., 'y ..o' g ,k, hf;F.:$: nt .) E 'h .s ',4+ < y

+g: m. s f'

.3%;. c e.> .A a x: s f.:. .I i: f,.h v 'i;;;. e "* t ; .. M... ,F ' s, $~p:{jp Ni . ' N :,,", J ;,3.s.., . k;;,[.zy.,f: Nf.i M

f

$CQ?;^jg[w.:'k.' > g.gi ;;.:@9:i.h.y ? > #. c. = , l.[: 4 m ;j;'O ,#%'M:.3, WF.x. .,%?i: s . ;x f M.e.g < ,: f.9f

• v; s
• g i:

n: 7 < >; ;

>.~. ;.:eN wy.;;# x ,; *W:;;,

yj 3 - e g.,

xx v

m. h, q 3 -

ti ,; i 'p: >%'s' v 5;', i,.pk,,[.Y,.3;. p:gs[ 5,g . ) ;t. 3,,; y < W.-

k. w:

.; [ m h ;h Y.'g:::f; [: t ii h.7,g.;.):h. i.4,,?.gif

f

.2

'N_, ):[ (.[.;^ h.Nflhhh f.) 4, s, p,, eih c / a 3, N,- ,.n.. :: 9.gj

py.

vo. g v ...:ae.;g. g;t kW'" m.m;w i M;tYNi! #5 4SB[ !M'p'gif m 6%.Mff.m4-ea W1 W a wip.re: qw N $Mb .@sg;334,:x f'j:NMM. [D1 b/ j$fj[$J ", 4 N$. ' d.4;<N '(,$f.h 9 j:f@Q ' t?'ff'b A /s<. 9gj0.(<. 45 j%gd:c6: h^ah.f $h <::T' !2 i g:p %A.T-? y?fJ p5h 4 5 h .St hh;, f Tf e

M:Q g y.:j:4;I My 5

v; '^ ?% m ^ 7 +' f x.::.g.i!. ' f::;p - m;:- pi/V:i:?qq Q 4.bfh'd 4 '.T' + 4 ,g.,- .j F.{. j' 0 k'$,.....;..;, ;. g. j$.,j.,..:.; s 6, ~ru.- .-.yy. 3 s -9s. v?w y g '4 - h.h -I [#[k -.v.$:.;:".h'f [F% hs. # h,pjt.:f;.y.M:h5(inh.[. e x < .k f: i 3 I. 4 kh[0 (h b , h.[h.

1. .j' h'l

.f+$! kg n:+:..:.,id>h'E [f: 3 .jd: Ev N'..,;,7:'jW# e..f,f e 3 .g>.f:d < h:;soijhbh.?. :x.y.I ' '.

q: +(M*i v N.;4. y^

g {gg ' fi [,s,fi ~ : !$: V

-l'

.c/m;:k:. v-:<v, v.' N:. ?:.s<..<.y?- J .g g'6: <N.ej' ff:{* ?Yw :p KW ( ' ', 7 g :g. ':. x '@i?. ' + f'n:jff d', ^? f".!.+;9( + ' :.^s,

-Ly M..s ys s

.^f^"'" C., 2

lt./.: r:

2n 4 v.x;;p%..,Jffj09 ,.s. p :^Q'fil o'ih; p ps.4.,.. y e v .,.:!r.q[;j V c.1$@ g f s. w. t.;Mh53.,j0 < i $ N

/p l

[ f p. f,w? ws

n. s n.oM

v!A ;s.i:dM w, ' y. u

.v. M> 8 :.i. w:i:e. b .>es~m n aye>: L-, 4,.,% M em~ y.: >w-n w.... 5% c e P.h..n.A>m.n an. t$.g.;. Mis,/.+gs.g m.x@s-$@,MeU Mg..;: s.*]e g:. c.n$@9 %+g.g= Av u w s ten M ~;4 ' ;p.7g ;y.}Y..s...f 7, Hey.+:: EM.y.. rm. ...c:. s 4: m gjg g5E p $y@$ ig v 9y e n a. s i NO* u> N !WM ,a su en z ?. , :ms.cm . jQgjj 8% L, n%ML e% m gn:t sn Q$q@J ( M.@mQ.[ w.M.,...,yh% M gw%$ MMS 'jil %9% g - @f. +%g@w w[;w( M Jh g5~-p1_i9/.. gg-- m +ym yA.i!x.yggs g~, .s y

u...

Q gggp n;..y,.;m& m n :+.g.%>p c >!pW' nu, -q MJ.' 4 :i.j a am si s i .w .e y:m x s a s%;#7:fjld G';

5Mb@p >f dFt ni

. m. fie%w!!Eif$ i %#.Rs[>Jhth j g:$.j g p:. 5:[.g b $y'f $ Q:M kih.:ihii$$; w>:s W ,"m. " s ^ 8 f/ g4i mysr4 yMi:h@ if:igQ$ Qgj, .ffs 43

.jfCi!!!!$:9?

hhiie "e:R ?$i'i:P!:4 $"h&g h. d$@.j % % M W Miy c M VM % '@h: n&O n o.M ~ m c P 2" k3hi< # d$ $ n~ MM5 Wiy 15 DO &# $?Nhb%h WN?s% ~ %[9% l .14$ MMfNE Qj$NM C} Q: ^ W Q Q($a%y@[@y e%sh e w&w %m&#hD wN Q gy%% W.Waid 4Mi+ W n ep 4 qw gg u =Wgyg p%ptegi. abp'> mgp x w y esq s NN ME b

h. N N

. +ik N - i 0 0 10 20 30 40 50 00 70 80 90 100 MAX % TiWtUWALL Figure 7.1 PLUS POINT ET RESPONSE TO LAB SAMPLES AND PRAIRIE ISLAND PULLED TUBES 4

INDICATION LOCATION EVALUATION APRIL 25,1996 J LAREAU i TUBE iET LOC MET i COMMENTS I 05 2 WZi - SB - 61% OS-3 BH/WZI - BH/SB - 18% OS-10 NDD SB + 29%. OS-14 WZI - SB - 31% CS-5 BH + BH +/ INCL - 35% BH OBSCURES INCL CS-13 WZI - SB-22% l CS-22 W Zl-SB/ INCL +/- 60% l OT-1 WZl-INCL +/- 74% OT-2 W Zl-BH-l OT-3 WZl-INCL-79% OT-5 W Zl-INCL /SB+/- 38% 0 T-6 NDD INCL +/- 68% WZl-W/ APP H PROC OT-11 lNDD INCL +/- 31% OT-12 W Zl-INCL /SB+/- 89% OT-16 BH-BH-l OT-17 W Zl-lNCL+/- 40% OT-20 W Zl-INCL +/- 65% OT-21 WZl-INCL /SB+/- 36% OT-23 WZl-INCL +/- 19% OT-24 W Zl-INCL +/- 97% OT-31 W Zl-LOF/SB+/- 25% I CT-44 UEX +/- CONCAVE >50% NO EXP, REJ WELD l Q-113 W Zl-INCL-35% l PI 5/48 W Zl. SB/ INCL-80% PI 9/57 W Zl-SB/ INCL-33%; +45% P17/52 WZl. SB/ INCL-100% PI 7/63 W Zl-SB/ INCL-50% PI 5/74 W Zl-SB/ INCL-100% TABLE 3 FLAW LOCATION EVALUATION

I u SLEEVE NDE PROCESS FLOW CHART 1 2 4 d d n "M POR RE CEAN l c'*** I accEr )~ i Y j' t M s f 1 i a nEwan y i f Jb 1 V i lNCOMMETE eLOW ^4.,, "m weto OE 3,cn,,, #*d!i, +.,7 wELo W >CLE ib e ,o 1 C' wao y m '3-(Ptuo) r r .4 5tm i stow eett i ASOVE WELD accEn i N i ur i i J I accEn 4 i l 1 i act d i d. t Y SEE CHART No. 2 1 1 'f ri are f

SLEEVE NDE PROCESS FLOW CHART 2 [. g l er r q; prii' ] i il l l Y I I I I l Y Y Y Y Y OmER ] N00 GEO W5t ga Y 1 1 Y Y acca m ou! NCR NCR NCR WOE WEE m l l l l Y Y Y Y vr vr vi vr REVIEW REVEW REVIEW REVEW ~ ~ i Y Y Accmasts accemste N Y I I I i Y Y Y Y ~ saa one was wm l l v i I v V Y <[m> accmans wza wza mew!w l i Y + V Aces m aut N BMA BHe @WI I n => fg vr4 S'

i ANALYSIS GUIDELINES FOR l SLEEVE WELD i ULTRASONIC INSPECTIONS ABB CENO l J Lareau May 1,1996

1. REVIEW THAT SLEEVE BACKWALL SIGNAL ABOVE AND BELOW WELD HAS A NOMINAL AMPLITUDE >80% FSH.

2. LOCATE B SCAN CURSOR NEAR MID PLANE OF WELD AS INDICATED BY C SCAN.

DISPLAY ZOOM SHOULD BE USED FOR ALL REVIEWS TO ASSURE ADEQUATE IMAGE SIZE.

3. USING WE MULTICOLOR B SCAN PRESENTATION, DETERMINE F A SIGNAL FROM THE TUBE OD IS PRESENT FOR 360 DEGREES AS A CONTINUOUS IMAGE ON THE B SCAN AT WE APPROPRIATE DEPTH (NOMINALLY 0.074'*). F A CONTINUOUS BACKWALL IS PRESENT ON.

A B SCAN, THEN 'ITIE WELD IS ACCEPTABLE WITHOUT FURTHER ANALYSIS.

4. IF SHORT SECTIONS OF THE TUBE BACKWALL ARE MISSING (<30 DEGREES IN ARC),

THEN EVALUATED THE ADJACENT B SCANS TO DETERMINE F THE BACKWALL IS PRESENT IN THE MISSING ARC LENGTHS, F SO, THE WELD IS ACCEPTABLE WITHOUT FURTEHR ANALYSIS. ~

5. E LARGER SECTIONS OF THE BACKWALL ARE MISSING, THEN THE SLEEVEm1BE INTERFACE HAS TO BE EVALUATED FOR CONTINUOUS REFLECTIONS ACROSS THE WELD HEIGHT. SUSPECT AREAS FROM BOTH THE C AND B SCAN HAVE TO BE EVALUATED FOR LACK OF FUSION (LOF) INDICATIONS AS FOLLOWS:

A. ON THE C AND B SCAN, DETERMINE F SIGNALS EXIST IN THE SLEEVEm1BE INTERFACE GATE REGION THAT EXCEED 20% OF THE NOMINAL SLEEVE BACKWALL SIGNAL. (EG, F SLEEVE BACKWALL IS 80%, THEN SIGNALS >l6% FSH SHOULD BE EVALUATED) B. IN REGIONS WHERE B SCAN PARTIAL LOF INDICATIONS EXIST, LOCA'IE A B' SCAN IN THAT REGION. DETERMINE F A LAMINAR SIGNAL EXISTS THAT BRANCHES FROM THE SLEEVE BACKWALL SIGNAL EDGE WROUGH THE WELD AND TO THE OWER EDGE OF THE SLEEVE BACKWALL. AMPLITUDES MAY VARY, BUT THE KEY SIGNAL FEATURE IS A CONTIUOUS LAMINAR SIGNAL SPANNING THE WELD HEIGHT AT THE SLEEVE WALL THICKNESS DEPTH OR A FEW MILS LESS. (F INCLUSIONS MOVE WITH THE WELD PUDDLE, WEY WILL MOVE CLOSER TO THE ID OF THE SLEEVE) F ONLY ONE OR TWO PIXELS ARE MISSINO IN A LAMINAR INDICATION THAT ALMOST SPANS THE WELD HEIGHT, F. VALUATE THE A SCAN FPR POSSIBLE SIGNAL DROPOtTT DUE TO LOSS OF GATE, AIR BULULE, ETC. IF ABSENCE OF A PORTION OF AN OTHERWISE NEARLY CONTINUOUS LAMINAR INDICATION IS ATTRIBIJTED TO SIGNAL DROPOUT, THEN PRESUME WE LOF IS COMPLETE OR RETEST.

6. ADDITIONAL DIAGNOSTIC UT MAY BE CONDUCTED BY111RNING OFF THE INTERFACE GATE TO DETERMINE BLOW HOLES AND COMPLETE CAPWRE OF THE INTERFACE. FOR EXAMPLE, IF WATER PATH VARIATIONS EXCEED 'IEE PROGRAMMED LIMIT, THE j

INTERFACE GATE WILL NOT LOCK ONTO THE PROPER INTERFACE SIGNAL AND COULD CAUSE ERRONEOUS RESULTS. ALSO, WELD SAG CAN BE IDENTEIED WITH A B' VIEW WITHOUT INTERFACE GATING. THE ET ANALYST MAY REQUEST ADDITIONAL DIAGNOSTIC INFORMATION TO SUPPORT GEO OR SAG INDICATIONS.

s a ECT BASELINE ~ NEW PRACTICE ABOVE OK AOCATIONWR ygg 3 PLUG m O BELOW BOTH 2 En PLUS POINT r 0 SCUMCI INDICATION No ggg NDD SADMAI PLUG O .}}