ML20081A604
| ML20081A604 | |
| Person / Time | |
|---|---|
| Site: | Quad Cities  |
| Issue date: | 02/15/1984 |
| From: | Office of Nuclear Reactor Regulation |
| To: | |
| Shared Package | |
| ML20081A603 | List: |
| References | |
| NUDOCS 8403050345 | |
| Download: ML20081A604 (11) | |
Text
_ _ - - _ _ _ _ _ _ - _ _ _
i p au
' [t ^ d UNITED STATES
~
NUCLEAR REGULATORY COMMISSION 3.
o
- $ p,ja E WASHINGTON. D. C. 20555 Ih4,/
++..
SAFETY EVALUATION BY THE OF'F CE OF NUCLEAR REACTOR REGULATION RELATED TO INTERGRANULAR STRESS CORROSION CRACKING COMMONWEALTH EDISON COMPANY AND IOWA-ILLINOIS GAS AND ELECTRIC COMPANY OUAD CITIES NUCLEAR POWER STATION, UNIT 2 DOCKET NO. 50-265 INTRODUCTION Quad Cities Unit 2 was shut down on September 4, 1983 in accordance with the confirming order issued on August 26, 1983 to inspect all ASME Class 1 austenitic stainless steel piping that are susceptible to intergranulhr stress corrosion cracking (IGSCC) in the Recirculation, Residual Heat Removal (RHR), Core Spray and Reactor Water Clean up (RWCU) systems.
During this shutdown period, ultrasonic examinations were performed on 225 nonconforming welds.
Of these,120 welds were in the Recirculation system,.64 welds were in the RHR system, three (3) welds were in the RWCU systems, 25 welds were in the Core Spray (CS) system, 12 welds were in the Control Rod Drive (CRD) system and one (1) was a Head Vent weld.
The licensee, Commonwealth Edison Company (CECO), indicated that except for 10 welds, all Class 1 welds susceptible to IGSCC in the above mentioned piping systems were ultrasonically examined.
The 10 welds (6 RHR welds, 2 CS welds and 2 RWCU welds) cannot be ultrasonically examined because of access limitations.
Personnel from Lambert, McGill, and Thomas (LMT) performed the ultrasonic testing (UT) for the licensee.
All LMT reported crack indications were confirmed by CEC 0's UT personnel.
Region III of the NRC has determined that their UT procedures, calibration standards, equipment and IGSCC detection capabilities were satisfactorily demonstrated in accordance with I&E Bulletin 83-02, and the same procedures and techniques were 8403050345 840215 PDR ADOCK 05000265 P
.c
' used in the UT examination.
Region III also indicated that all their UT personnel conducting these inspections have received appropriate training in IGSCC inspection using cracked thick-wall pipe specimens.
LMT used 45 and 60 shear waves for crack detection and 25% leading and trailing movement for crack depth measurement.
The results of the UT examinations indicated that a total of 23 welds showed reportable linear indications.
Of these, eleven are 28" Recirculation welds, two are 22" Recirculation Manifold welds, seven are 12" Recirculation Riser welds, two are 20" RHR welds, and one is 6" RWCU weld.
All reported UT indications were in the weld heat-affected-zone (HAZ).
Of the 23 defective welds, six (6) welds were reported to have a crack depth over 25% of the wall thickness.
The majority of the cracks in the
- defective welds were in the circumferential direction.
The deepest circumferential crack with a depth of 80% of the wall thickness.was reported in a 12" Riser weld (020-F6).
Short axial cracks were reported in two 28" Recirculation welds.
The reported crack length in all defective welds varied from half of an inch to 360 intermittently along the cir-cumference.
Five (5) defective welds were reported to have a crack length
- at-least 120* of the circumference.
As will be discussed later, Inter-national Test Laboratory (ITL) and Universal Testing Laboratory (UTL) performed supplementary UT examinations on those five (5) welds using advanced techniques.
Based on the results of the advanced UT techniques, UTL' determined that these five (5) welds were not cracked.
Subsequently, the " worst" of the five (5) welds (285-512) was confirmed to be not cracked by destructive metallography and radiography examinations.
General Electric (GE) performed Induction-Heat-Stress-Improvement (IHSI) for the licensee on a total of 87 welds (77 Recirculation welds and 10 RHR welds) including all the unrepaired defective welds.
The
_._--______________--_-_._-s_-
p i
h I licensee' indicated that all the large size (> 12 inches) stainless steel piping welds in the Recirculation and RHR systems up to the first isolation valves were treated with IHSI except the bimetallic welds, eight (8) sweepolet welds, three (3) overlay repaired welds and nine (9) welds requiring special heating coils because of geometry.
UT inspection was performed on each weld after IHSI.
38 IHSI treated welds were also UT inspected prior to IHSI.
There are no significant differences in UT results before and after IHSI.
Region III of the NRC indicated that IHSI was
-satisfactorily performed in accordance with the qualified GE procedures.
Based on a review of the UT data and the IHSI record, the licensea decided to weld overlay repair six (6) defective welds.
Of these, two (2) welds had axial cracks, three (3) welds had circumferential cracks with depth over 30% of the wall thickness and one (1) weld was improperly treated with IHSI.
In addition, one (1) RWCU weld (125-S27) was replaced with low carbon stainless steel piping material.
NUTECH performed flaw evaluations on the remaining 17 defective wclds for the licensee to determine whether overlay repair is needed or not.
The evaluations were based on the methodology provided in the new ASME Code Section XI IWB-3600.
The new Code IWB-3600 provides flaw acceptance criteria for the austenitic stainless steel piping based on a limit load approach which was approved by the ASME Main Committee in May 1983 and was published in Winter 1983 Addenda.
The results of NUTECH's flaw evaluations,' including crack growth calculations, indicated that dll the remaining 17 defeOtive Welds did not require weld overlay repair because the calculated flaw sizes of'those 17 welds at the end of an 18-month period did not exceed the staff's criteria of two-thirds of the new Code allowable limits.
In NUTECH's calculations, the cracks in each weld were essentially arrested by the surrounding compressive residual stresses induced by the IHSI treatment.
7
.~
.4-We had some concerns regarding five-(5) extensively cracked welds
('
1.2" Riser welds, 02F-F6.and 02J-F6, and 3 28" Recirculation welds, 02AS 102BS-F7 and 02BS-512) which were not recommended to be overlay repaired.
p
.The crack length in each of these five (5) welds is long and encompasses at.least 120*'of the circumference.
After several discussions with the licensee :regarding our concerns, the -licensee brought in'ITL and UTL to perform additional UT-examinations on the five (5)' welds.
Both ITL and
.UTL:used the'tachniques of 45* and 60* shear wave, crack tip diffraction and OD creeping ~~ wave to-size the crack. depth.
UTL also used an advanced Acchnique. based on ID creeping wave to detect the shallow cracks.
The creeping wave-is reported to be transparent to the metallurgical reflectors such as the' fusion ~1ine.
The indications ~ reported by ITL and UTL based on'the use of 45* and 60* shear wave were similar to-
- that. reported by-LMT.
However, base'd on the-results of using ID creeping wave,;UTL determined that these five (5) welds were not cracked and the indications were from metallurgica11 reflections.
To confirm 3
'UTL's determination, a trepan sample about one inch' in diameter was taken:from ona~of'the.five. welds (2BS-S12, 28" Recirculation weld)-
at1the' location; reported to have the strongest crack indication.
j.
No cracks-were~.found from the-trepan. sample both visually and destructively b
(metallographic examination).
Single wall radiography was also performed t.:
on weld 2BS-512 and no-crack indication was found.
In the meantime, the' licensee decided to weld overlay three (3) (2 12" Riser welds, 02F-F6 and 102J-F6, and 1 28" Recirculation weld, 02BS-F7) of the five (5) welds to ensure that-the required Code-safety margin-would be maintained during b
operation.
i EUTECH.also performed weld overlay design for the licensee.
Nine (9)'
i
. defective welds (4--12" Riser ~ welds, 4 28" Recirculation welds and 1 20" RHR
-weld)'were weld overlay repaired.
The. overlay thickness w3< designed to meet the1new IBW-3600 limits based on-an assumed crack depth which is two i
t l
P times the reported maximum crack depth.
The length of the overlay was selected to reinforce the weld structure and minimize the end effects.
The as-build overlay thickness was reported to vary from 0.27 inch to 0.462 inch and the as-built overlay length was reported to vary from 2.8 inches (only one side) to 6.5 inches.
Region III of the NRC has confirmed that the weld overlay repairs were performed in eccordance with the qualified and approved procedures consistent with ASME Code requirements.
The licensee reported that the as-measured axial shrinkages from the nine (9) overlay repaired welds were in the range of 0.019 inch to 0.23 inch.
The stresses caused by this shrinkage on all the unrepaired defective welds were calculated.
The largest value was reported to be 8,400 psi on a 12" Riser weld (020-53).
In NUTECH's crack growth calculation, the shrinkage stresses due to weld overlay were considered.
In summary, during the current Quad Cities Unit 2 confirming order outage, a total of 225 nonconforming austenitic stainless steel welds were UT examined.
This includes all the UT examinable stainless steel welds in the Recirculation, RHR, Core Spray, RWCU and CR0 piping systems.
In addition, 87 welds were treated with IHSI including all the unrepaired defective welds.
The UT results reported by LMT indicated that 23 welds showed reportable linear indications.
Of these, nine (9) welds were overlay repaired, one (1) weld was replaced and one (1) weld was shown not to be cracked by metallography and radiography examinations.
EVALUATION We reviewed the licensee's submittals, including NUTECH's analysis of the weld overlay designs, and the calculation of IGSCC crack growth
,p
.. nine (9) overlay repaired welds, the 12 IHSI mitigated defective welds (3 12" Riser welds, 6 28" Recirculation welds, 2 22" manifold welds and 1 20" RHR weld) and one (1) replaced weld (6" RWCU weld).
NUTECH used the favorable residual stress distribution resulting from IHSI treatment in their IGSCC crack growth calculations, because each unrepaired defective weld was mitigated by IHSI.
The results of NUTECH's calculations indicated that the cracks were essentially arrested in every IHSI treated weld.
This is caused by the presence of large compressive residual stresses at the front of the crack tip.
NUTECH indicated that the reported crack depth in each unrepaired defective weld is well within the staff's criterion of two-thirds of the new Code allowable limits.
We reviewed NUTECH's IGSCC' crack growth calculations and agree with their conclusion that.the continued operation for an 18-month fuel cycle with the !? defective welds mitigated by IHSI is justified because the Code design safety margin would be maintained.
Our con-
-clusion is based on the following considerations:
(1) Overcalls in UT Examination The UT indications reported by LMT were based on the technique of.
using 45 and 60* shear wave..The shear wave technique does not-have the capability of discriminating the indications reflected from a crack and that from a metallurgical reflection.
This was clearly demonstrated by UTL's test results of weld 285-12S using advanced UT techniques.
LMT's inspection results ii.dicated that weld 285-12S was extensively cracked; however, UTL determined that weld 28S-125 was not cracked and this was confirmed by the
' destructive metallography and radiography examinations.
Therefore, I
_______________-----a
9
. we expect additional overcalls might be present in the LMT's test results.
It is possible that some of the welds reported by LMT to be defeccive may not be cracked or, if it is cracked, it may not be cracked to the extent as reported.
(2) Short Cracks Of the 23 welds reported by LMT to be defective, nine (9) welds were overlay repaired, one (1) weld was replaced and two (2) welds were determined by UTL to be not cracked.
In the remaining 11 defective welds, six (6) welds have relatively short cracks with a crack length not exceeding 29 of the circumference and the other five (5) welds have crack length varied from 45 to 100 of the circumference.
Based on limit load analysis and the stress ratio reported by NUTECH, even if the cracks were assumed through-wall, it will not have a significant effect on the structural i,ntegrity of the welds.
(3)
IHSI Mitigation The licensee indicated that all the unrepaired defective welds were mitigated by the IHSI process.
IHSI is a process developed in Japan for treating weldments already fabricated or installed in a plant system.
It consists of heating the outside of the pipe by induction coils to controlled temperatures (~ 800 F) while cooling water is circulated inside the pipe.
The high thermal gradients produce the same effect as Heat Sink Welding (HSW).
The inside of the pipe is stretched during the process, causing high residual compressive stresses after the process is completed.
In the proposed NUREG 0313, Revision 2, the staff has accepted the IHSI as an effective IGSCC mitigation process when applied to uncracked welds because there is concern regarding the effectiveness l
- ?
. of IHSI process when applied to cracked welds especially for deep cracks (i: 50% of the wall thickness).
This concern is compounded f
by the uncertainties associated with the UT determination of the crack depth.
Therefore, at this time, we will consider the IHSI mitigation on welds with shallow cracks as a tecporary fix.
The Japanese have reported some limited field experience of applying IHSI to crecked weld in an EPRI sponsored seminar of counter-measures for BWR pipe cracking.
In Fukushima I, Unit 3, IHSI was applied to five (5) cracked sweepolet welds in the Recirculation piping system.
After one year's operation, the results of UT examination indicated that the cracks ir the five (5) sweepolet we'.ds did not propagate.
The licensee replaced one (1) defective RWCU weld (125-527) with 304L (SA 304, Grade WP, 304L) stainless steel material using HSW.
304L stainless steel material is considered resistant'to IGSCC, and HSW produces favorable compressive residual stresses at the inner surface further inhibiting IG',CC initiation and growth.
Therefore, we conclude that the replaceu 9 ective weld is acceptable for continued service.
NUTECH's overlay design for the nine (9) defective welds was based on assuming the crack depth to be two times the reported maximum crack depth.
The favorable residual stress distribution resulted from weld overlay repair is depended upon to inhibit further crack growth.
The residual stress distribution used in their analyses was calculated using a finite element model.
The required overlay thickness determined this way is much thinner than that determined by assuming all cracks
.are completely through-wall.
Although there are many important advan-tages in using thin overlay (lower fabrication time and cost less radiation exposures, increased UT inspectability of the overlaid weld, and-less distortion in the piping system), the thinner overlays do not I
.- provide complete Code compliance if the cracks were to continue to grow until it reaches the overlay.
The specific residual stress distribution used by NUTECH in this analysis appears to us to be somewhat more favorable than can be justified at this time; therefore, for our analysis, we relied on distributions derived from later work, and results from closely related work per-formed under EPRI contracts.
Even with our more conservative approach, we also concluded that all repairs will provide adequate assurance of safe operation during the next operating cycle.
There are 10 welds (6 RHR welds, 2 CS welds and 2 RWCU welds) not examinable by UT because of access limitations.
Of these, eight (8) are containment penetration welds, one (1) is a reinforced RWCU branch weld surrounded by a saddle weld, and nne (1) is a Head Spray weld located in a water barrier between the Drywell and the Reactor Cavity.
The licensee indicated that all eight (8) containment penetration welds are isolable; nowever, the 6" RWCU branch weld and the 4" Head Spray weld are not isolable.
During normal plant operation, there is no flow in the Head Spray piping and the temperature is less than 200 F.
Therefore, we do not exoect this Head Spray weld to be extensively cracked.
The non-isolable RWCU weld is made of small piping (6" diameter) and leakage in this size of piping is not expected to create any major LOCA, because the leakage is expected to be within the reactor coolant make-up capacity and can be detected by the leak detection system.
We conclude that the 10 uninspected welds will not create any major safety oroblein during the continuous operation of the unit for a normal operating cycle.
Leak Detection Although the conservative calculations discussed above indicate that the cracks in the unreinforced welds will not progress to the point of leakage during the operating fuel cycle, and very wide margins are expected
)
.1 to be maintained over crack growth to the extent of compromising safety, uncertainties in crack sizing and growth rate still remain.
- Further, not all welds were UT examinable, and cracks could be present in welds that were not examined.
Because of these uncertainties, it is prudent to tighten the requirements for the monitoring of unidentified leakage.
The licensee has agreed to additional monitoring and tighter limits on unidentified leakage, which are summarized below:
(1) The reactor coolant system leakage will be limited to a 2 gpm increase in unidentified leakage within any 24 hour2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br /> period.
Leakage shall be monitored and recorded once every 4 hours4.62963e-5 days <br />0.00111 hours <br />6.613757e-6 weeks <br />1.522e-6 months <br />.
Should this leakage limit be exceeded, the unit will immediately start an orderly shutdown.
The unit will be placed in at least hot shutdown within the next 12 hours1.388889e-4 days <br />0.00333 hours <br />1.984127e-5 weeks <br />4.566e-6 months <br /> and in cold shutdown within the following 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br />.
(2) At least one (1) primary containment sump collection and flow monitoring system will be operable.
With the primary containment sump collec-tion and flow monitoring system inoperable, restore the inoperable system to operable status within 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br /> or immediately initiate an orderly shutdown and be in at least hot shutdown within the next 12 hours1.388889e-4 days <br />0.00333 hours <br />1.984127e-5 weeks <br />4.566e-6 months <br /> and in cold shutdown within the following 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br />.
We conclude that implementation of the above measures will provide adequate assurance that possible cracks in pipes will be detected before growing to a size that could compromise the safety of the plant.
SUMMARY
AND CONCLUSIONS We have reviewed CECC's submittals regarding the actions taken or to be taken during this refueling outage on the inspection, analyses and repairs of Recirculation, RHR, RWCU, Core Spray, Head Spray and CRD piping systems in the Quad Cities Unit 2 plant.
This includes a description of ' ie defects found, description of repairs, stress and fracture mechanics analysis.
.. L We conclude that the Quad Cities Unit 2 plant can be safely returned to power and operated in its present configuration at least for one normal' operating cycle provided that the following items are satis-
~factorily completed:
(1) The Code required system pressure tests and nondestructive examina-tions on overlay repaired welds should be satisfactorily completed-
. prior to startup.
'(2) The additional leak detection requirements as listed in the section on Leak Detection should be in place prior to startup.
Nevertheless, we still have concern regarding the long term growth of small IGSCC cracks that may be present but not detected during this inspection. Therefore, plans for inspection and/or modification of the recirculation and other RCPB pipirig systems during the next refueling outage should be submitted for NRC review at least three (3) months before the start of the next refueling outage.
Principal Contributors: W..Koo and W. Hazelton Dated:
February 15,.1984 J