ML20071A606
| ML20071A606 | |
| Person / Time | |
|---|---|
| Site: | Hatch  |
| Issue date: | 02/11/1983 |
| From: | Office of Nuclear Reactor Regulation |
| To: | |
| Shared Package | |
| ML20071A588 | List: |
| References | |
| TAC-49156, NUDOCS 8302240264 | |
| Download: ML20071A606 (24) | |
Text
+
UNITED STATES NUCLEAR REGULATORY COMMISSION WASHINGTON, D. C. 20656 SAFETY EVALUATION BY THE OFFICE OF NUCLEAR REACTOR REGULATION INSPECTI0tl. ANALYSES AND REPAIR OF RECIRCULATION AND BWR SYSTEMS PIPING AT EDWIN.I. HATCH NUCLEAR PLAtlT, UNIT NO. 1 DOCKET NO. 50-321 GEORGIAPOWE5 COMPANY OGLETHORPE POWER CORPOPATION I1VNICIPAL ELECTRIC AUTHORITY OF GEORGIA CITY OF DALTON, GEORGIA Introduction During the current 1982 maintenance /r'efuel outage, augmented inservice inspection was, performed on 23 austenitic stainless steel welds in the 4", 12", 22" and 28" recirculation (RECIRC) piping system and 11 austenitic stainless steel welds in the RHR piping system.. The results
/'of ultrasonic tests.(UT) indicated that three RECIRC welds- (two 22" end cap to p pe welds and one 22" sweepolet to man.' fold weld) and two RHR welds (one 20" elbow to pipe weld and one 24" pipe to pipe weld) showed reportable linear indications.
Because of the linear indications found on two 22" RECIRC end cap to manifold welds, the other two end cap to manifold welds were also. ultrasonically examined and were found to.show linear indications.
Subsequently, an additional 19 stainless steel welds in the 12", 22" and 28" RECIRC piping systems were" s6lected for ultrasonic inspection and no reportable linear indications were found.
Overall, a' total of 7 welds (5 welds in the RECIRC system and 2 welds in the RHR system) were found to show linear indications.,All indications were reported to be located in the general area of the base material heat affected zone (HAZ).
8302240264 830211 PDR ADOCK 05000321
=
During this outage, a total of 57 welds in the RECIRC and RHR piping systems were ultrasonically inspected. The selection of the welds for inspection was based on ASME Section XI requirements or the NUREG 0313 Revision 1 guidelines including the consideration of stress rule index, carbon content and the IGSCC experienced by other BWR plants.
Southern Company Services (SCS) and Southwest Research Institute (SWRI) per-formed the ultrasonic tests for Georgia Power Company.
Their UT procedures and calibration standards were satisfactorily evaluated on IGSCC cracked pipe samples at Battelle-Columbus in accordance with I&E Buligtin 82-03.
NuTech has evaluated the flaws found in the seven welds mentioned above.
Code stress analysis and various fracture mechanics analysis were performed.
The results of the evaluation indicate that six of the seven fla' ed welds required weld overlay repair to restore the original w
design safety margins.
The 22" RECIRC sweepolet to manifold weld was shown by analysis to continue to have the original design safety margin
.for at least one fuel cycle.
I Th'e six welds being reinforced'by a weld overlay of'IGSCC resistant, materials are four 22" RECIRC end cap to manifold welds, one 20" t
/-
RHR elbow to pipe we'.d, and one 24" RHR pipe to pipe weld. The design life of each everlay repair was calculated by NuTech to be at least 5 years.
Acoustic emission (AE) devices with automatic
3-strip chart recording will be installed on the unrepaired sweapolet to manifold weld to enhance the capability of the leak detection system to detectini, small leaks.
After all weld overlay repairs, the code required hydrostatic test and nondestructive examination will be performed.
Description of Cracks The reportable, linear indications found on 7 stainless steel welds a
in RECIRC and RHR piping systems are described in detail in the Attachment to the Georgia Power Company submittal of January 27, 1983. These indications are located generally in the base material of.the heat affected zone (HAZ) and are characterized predominantly as short axial or t.ransverse cracks.
Only two circumferential cracks were found; these were in the RHR system, at an elbow to pipe weld.
These were determined to be 1 1/2 inches long, not over 33% of the wall thickness in depth.
Numerous short, but deep axial cracks, were l
found in all 4 end cap to manifold welds with the lengths varying l
from 1/4 inch to 1/2 inch and with depths up to 72%.;of wall thickness..
l-Seven short transverse cra:ks (1/4 to 1/2 inch) were found in the.
The depth of these cracks is ' reported to be very shallow, not exceeding 12% of~the wall thickness.
The axial cracks found in the RHR pipe to pipe weld are also short e
h 4.
e
4-(1/4 to 1/2 inch long) with the depth not exceeding 47% of wall thickness.
The deepest axial crack reported was 94% of the wall thickness and about 3/8 inches long; it was found at the elbow to pipe weld.
Short, deep axial cracks have been noted previously, and leaks emanating from them were noted and reported at Quad Cities 1 in 1980, and Monticello in 1982.
They probably occur in locations with high residual welding stresses in the circumferential direction,.
They are typically short I
becausethesehsitizedheataffectedzone(NAZ)extendslessthan 1/2 inches on either side of the weld, and integranular stress cor-rosien cracking requires sensitization to be present.
As noted at Quad Cities 1, howeher, such cracks can propagate into and through the weld if it has high carbon and low ferrite.
Axial cracks are of much less concern from a safety standpoint than circumferential cracks, which can grow through the wall and around
.the circumference of the pipe, for two reasons.
First, the service stress on thd axial crack is almost all caused by pressure, and typically the presstre stress is low compared to the total str~esi' acting on a circumferential crack, where bending stresses can be significant.
Seconc, because IGSCC is c.onfined to the sensitized material of the narrow HAZ, axial cracks cannot grow to significant lengths.
g.
5-Axial or radial cracks, if short, are very difficult to detect and size by UT because they form under the crown of the weld, and it is usually difficult to direct the sound beam at the proper angle.
They of ten can only be detected at very limited transducer locations.
It should also be mentioned that because they can only be short in relation to the wall thickness, and the stresses tending to open them are low, they will cause very little actual leakage, perhaps not enough to be detected with normal procedures.
,f
/
g-In summary, altnough axial or radial IGSCC cracks are hard to find and size by UT, they will cause only small leaks and will not grow' long enough to initiate a pipe burst unless the piping itself is completely sensitized.
Description of the Overlay Rein'forcement The weld overlay was installed by depositing IGSCC resistant 308 L weld metal 360 degrees circumferential1y around the pipe.
Th'e weld deposited band over the cracks reinforces the pipe and introduces beneficial reduced stresses.
The minimum overlay thickness was e.
selected to restare the original design safety margins. The minimu'm overlay length used was equal to 4 rt, where r is the mean radius.
of the pipe and t is the wall thickness.' This is the minimum lengthrequiredtoensurethattherewobldbenoadverseendeffects.
O e
- -,,, ~,.. -,
s-,--
The ends of the overlay are tapered at a nominal angle of 45 degrees.
The overlays on the end cap to manifold welds (RECIRC), elbow to pipe weld (RHR) and the pipe to pipe weld (RHR) have a nominal thickness of 0.25 inch, 0.4 inch and 0.3 inch-respectively; and their respective lengths are 6.5 inches, 7 inches and 10 inches.
During overlay welding of the RHR pipe to pipe weld, toe cracks were observed adjacent to one end of the overlay. The toe cracks with a depth approximately 1/32 to 1/16 inch were formed due to portion of the stainless steel overlay being deposited on a neighboring Inconel weld.
The overlay was subse-
' quently extended 2 inches to a total length of 10 inches to cover completely the area where the toe cracks were removed by grinding.
Inconel weld material was used for the last portion of overlay which overlapped the neighooring Inconel weld.
Effect of' Overlay Reoair on the Recirculation or RHR Systems The weld overlay repair causes both an axial and radial shrinkage underneath the overlay.
The shrinkage induces beneficial residual
.conipressive stresses in the cracked pipe, but may adversely affect l
the weld joirits in cther locations of the systems if the shrinkage is of sigrificant magnitude.
Theseeffectshavebeen:evaluatedbyJuTsch r
for GPC and are jud;sd to have no deleterious effects on the recirculation or.RHR systems.
The results are summarized below.
e i
l
(
l
~
1
The effects of radial shrinkage are limited to the regions adjacent to and underneath the overlay.
NuTech indicated that based on their work performed for Monticello, the radial shrinkage stresses are less than yield stress at distances greater.than 4 inches from the end of the overlay.
~
NuTech has evaluated the effect of the weld overlay axial shrinkage on the Recirculation and RHR systems by using their computer program PISAR.
The 4 end cap weld overlays aie adjacent to the recirculation
?
' ' manifold free hnds and will not induce s' tress in any other section of the piping.
In the RHR system, the axial shrinkage of the elbow weld overlay and pipe to pipe weld overlay' was measured to be 0.25 inch and 0.19 inch,'respectively. The measured axial shrinkage is imposed in the model as boundary condition during evaluation. The maximum secondary stress calcul'ated from the model is less than 9 ksi and NuTech considers this to be acceptable.
We have noted that the flawed sweepolet to manifold weld is only 2 inches away from a nearby end cap to manifold weld overlay. There is a potential concern that the radial shrinkage of the end cap 'i
overlay may affect the flaw in the sweepolet weld.
We have concluded that this will not present a serious safety concern for th'e following reasons:
8-(1) Although the unrepaired'sweepolet weld is only about 2 inches away from the end cap to manifold repair overlay, the identified cracks in the sweepolet weld are at least 8 inches away. The magnitude of the radial shrinkage stresses at this distance are not expected to be of major significance.
4 (2) The cracks found in the sweepolet to manifold weld are all short cracks transverse to the weld.
As previously discussed, trans-verse cracks will not grow to any significant lenth as their l
1ength is, limited to the width of the sensitized heat affected Zone.
(3) The bending stress induced by the weld overlay is di.splacement controlled (self equilibrating) and would tend to be relieved by initiation of cracking.
(4) As will be discussed later, GPC is in the process of installing two acoustic emission (AE) devices on the isnrepaired sweepolet to manifold welds to detect the leakage.
The AE devices are very l
sensitiveandleakageassmallas0.1gpmcanreadilybe,,de.}ec,ted.
~
Therefore, any small leakage emanating from the unrepaired sweepo,let r, -
will immediately be det eted.
e
\\
s g
o.-e-..en,e.ge ese *
- e e e.=w e
- w n
ew ow,,
n
I Code Stress Analysis The repaired piping was evaluated according to Section III, and was found to meet all requirements including seismic and fatigue requirements. This was done by conservatively developing a finite element model with the use
[
of ANSYS computer program.
In the model, the material was removed to i
I represent the cracks.
Although the geometrical configuration is not typical of Code design, the stress analysis was performed using the l
Code rules.
The fatigue analysis used the standard set of transient l
conditions which consist of 38 startups,.25 small temperature change cycles and one' emergency cycle every five years, and included a
(
strength reduction factor of 5 in the calculation.
The calculations
~
show that the weld overlay repaired end cap to manifold welds, elbow to pipe weld and pipe to pipe weld will meet all Code requirements for at least 5 years.
Fracture Analysis
Background
NuTech performed the following three types of fracture analyses to show that the safety margins against failure are at lea.st equivalent to-the margins inherent in the ASME Code.
f Allowable Crack Depth Evaluation The calculation of allowable crack depth is based 'on a new proposed flaw evalaution methodology for Section XI of the Code.
This includes IWB 3640, o
" Acceptance Criteria for Flaws in Austenitic Stainless Steel Piping," and the associated Appendix C, " Evaluation of Flaws in Austenitic Stainless
Steel Piping." Although these new sections have not yet been approved through the Main Comnittee, they have been approved through the first levels, and full approval is expected shortly.
The basis for this criterion is the well known and accepted limit load for plastic collapse method of analysis.
Specific development of this method for the evaluation of flaws in stainless steel piping has been done under epi (J contracts, and has been d'escribed in several reports, including References 1 and 2.
For Code use, this calculational method has been used to develop simple tables, fr'om which acceptable flaw sizes and shap'es as a function of applied stresses can be read directly., These are Tables IWB 3642-1 and -2 for axial cracks, and Tables 3641-1 and -2 for circumferential cracks. There are sanarate
~
l l,
tables for Normal Conditions and Emergency and Faulted Conditions, with different safety margins.
The tables provide a safety margin
.of between 2.5 and 3 for Normal Conditions, and about 1.5 for Emergency and Faulted Conditions.
These are consistent with the overall basis of the Code.
It is noted that the presence of more than one crack does not change the calculations.
Multiple axial cracks do.not interact, and are treated separately.
1 l
4 6
=
._ n _
_ 11 Crack Growth Evaluation The crack growth due to fatigue and IGSCC is calculated based on rules provided in Appendix C to the proposed IWB-3640 for Section XI of the Code.
The methodology for evaluating fatigue propagation appears acceptable, but we still have some reservations about the IG5CC crack growth rate given in the Code.
This is of no concern for the repaired cracks and unrepaired crack at Hatch Unit 1 as will be described later.
,f Ultimate Failure Load The ultimate failure load is calculated with.a tearing modulus analysis.
This type of elastic plastic fracture mechanics was initia,11y developed on an NRC contract, and has been widely. accepted and used during the past 5 years.
It is recognized that the limit load approach is con-servative, and that much larger margins are actually present in many cases.
Tearing Modulus calculations were performed for both the repaired welds (end caps, elbow and pipe to pipe) and the unrepaired sweepolet weld.
~
l As expected, the calculations show that very large margins against '
l l
failure are present.
End Cap to Manifold Welds Repair Evalua' tion The flaws found in end cap to ' manifold welds are ali short axial cracks.
The largest axial crack has a length of 1/2 inch and a depth of 72% wall thickness.
NuTech performed an Appendix C egvaluation of the most limiting flaw in
12 -
the repaired end cap to manifold welds. The thickness of the overlay pipe wall is 1.24 inch with a minimum overlay thickness of 0.25 inch.
The allowable crack depth for the end cap to manifold welds is deter-mined to be 75 percent of the wall t'hickness from Table IWB-3642-1.
This corresponds to a crack depth of 0.93 inch in the weld repaired b/ overlay.
NuTech calculated the crack growth due to fatigue for 5 years of operation to be less than 0.05 inch.
a They also calculated the crack growth due to IGSCC for two cases. The -
firstcasewascalculatedbyconservatihelyassuminganinfinitelylong crack and by considering beneficial residual stess due to the weld overlay.
In this case, the crack will grow to a depth of approximately 0.E5 inch in five years, which is below the allowable crack depth of 0.93 inch, The second case considers the worst case for the end caps by assuming an axial crack completely through the original pipe wall.
The crack will not propagate into the overlay weld material due to its high resistance to IGSCC but will grow approximately 0.05
'ndh due to fatigue in five years of operation.
Thus, the total crack depth is about 1.05 inch which is about 82% of the overlay pi;e wall thickness.
This exceeds the allowable grack. depth by " ~ '
7 percent.
However, the calculations for the allowable depth of c
- e crack are overly conservative in this case, becau'se the Code e
~~
-y_
~ _
13 -
~
arbitrarily cuts off the allowable depths given in the tables for axial cracks to 75% of the wall thickness.
Extrapolation of the values in the table would show the allowable crack depth in excess of 82%.
Therefore, the overlay design is judged to be acceptable for 5 years.
We have reviewed NuTech's Appendix C calculation and agree with their conclusion regarding the acceptability of the overlay design based on the net section limit load analysis.
HuTe'chalsoperformedatearingmodulusev[aluationoftherepairedend cap to manifold welds based on the postulation that the l'argest size the existing crack could reasonably be expected to' grow to be a one inch radius flaw.
The predicted burst pressure (ultimate failure load) based on this flaw configuration is in excess of 5500 psig which is more than 4 times of the design ~ pressure.
This indicates that the design pressure, even in the presence of this worst case assumed
~
crack, has a safety factor well in excess of that inherent in the ASME Code.
RHR System 20 mil Elbow to Pipe Weld Repair Evaluation The largest cracks found in the elbow to pipe weld are an axial crack of-depth 94% of wall thickness and length of 3/3 inch,' and a circumferential crack of length 1-1/2 inches and depth of 33% of wall thickness., 'HuTegh performed similar calculations to show the acceptability of the elbow to pipe weld repair.
0 e
+
wm-g.
. O q
e--
-___..1._
~
9
. 1
~~
The overlay has a minimum thickness of 0.4 inch and the. repaired pipe section has a minimum thickness of 0.76 inch. The allowable crack depth determined from Tables IWB-3642-1 and IWB-3641-1 is 75%
of wall thickness for both the most limiting axial and circumferential cracks found in elbow to pipe weld.
The largest axial crack is essentially through the pipe wall and will propagate only by fatigue to a distance less than 0.05 mils. The axial crack depth after five years of operation would be 0.81 inch (70% of overlaid wall) which is less than the allowable crack depth of 75%. The circumferential. crack depth after 5 year's crack growth due to fatigue and IGSCC is?about 26% of the overlaid
~'
~
wall thickness which is substantially less than the allowable of 75L NuTech also performed a Tearing Modulus analysis based on a postulated
~
radius flaw of 0.8 inch. The predicted ultimate failure' load is in excess of 3 times the normal operating loads which provides a safety factor on normal operating loads larger than that inherent in the ASME Code, evt:n in the presence of this worst case assumed crack.
l
.We'have reviewed NuTech's Appendix C calculation and agree with their conclusion that the overlay design is acceptable based on the net section limit load analysis. We also note thaf there are 2.7 l
l circumferential cracks, each with a length of 1.5 inches in the elbow to pipe weld.
These two circumferential cracks'are located on different sides of the weld.
It is very unlikely that the two l
T
circumferential cracks are linked to each other.
We made a calculation based on Table IWB-3641-1 to show that with a crack depth of 40% wall thickness, the allowable length of circumferential crack can be half of the pipe circumference (~ 33 inches.). Therefore, even if the two circumferential cracks are linked together, it will still meet the '
Appendix C requirement.
Pipe to Pipe Weld Repair Evaluation The flaws found in pipe to pipe weld are all short axial cracks. The I
' largest axial crack has a length of 3/8 " inch and a depth of 47% wall.
The overlay applied has a minimum thickness of 0.3 inch and the repaired pipe section has a minimum wall thickness of'1.14 inches.
NuTech per-formed similar calculations to show the acceptability of.the pipe to pipe weld, repair.
The allowable crack depth' determined from Table IWB-3642-1 is 75% of wall thickness for the most limiting crack found in the pipe to pipe weld.
NuTech calculated the crack growth due to fatigue and IGSCC in the next five years.
The calculated crack growth is small and will not exceed the allowable crack depth.
NuTech also performed tearing modulus analysis based on a postulated 1.14 inch radius flaw.
The predicted ultimate failure loadsis in excess'of~ '
4 times the normal operating loads which provides a safety factor on
[.
normal operating load larger than that i,nherent in ths ASME Code, even in the presence of this worst case assumed crack.
~
16 -
l We have reviewed NuTech's Appendix C calculation and agree with their-j conclusion that the overlay design is acceptable based on the net section limit load analysis.
Sweepolet to Manifold Weld Evaluation
~
Seven'short transverse cracks (perpendicular to the weld) were found in the sweepolet to manifold weld.
The largest transverse crack is approximately 1/2 inch in length with a depth approximately 12% of the pipe wall.
Due to the difficulty in applying overlay in this pipebranchahea,theflawedsweepolettom'anifoldweldwasnot repaired.
NuTech performed similar Appendix C calculation to show that the subject sweepolet to manifold weld is acceptab.le for operation at least for one fuel cycle in the present.u'nrepaired condition.
The calculation showed that the allowable crack depth is 75*4 of pipe wall and the maximum crack depth after 5 years of operation
'is predicated to be 0.38 inch (38% of wakl thickness), which is well below the allowable crack depth.
NuTech also performed a tearing
. modulus analysis with a postulated 0.50 inch radius flaw in the weld.
l This flaw is' larger than any existing crack will grow to become in one fuel cycle based on the crack grc?th rate used for' Appendix CT Th'e' predicted ultimate failure load is in excess of 3.3 times the normal operating loads wr.ic, provides a safety factor on normal operating load larger than that inherent in the ASME Code even in the presence of tnis worst case assumed crack.
l 4
l O
We have received NuTech's evaluation and have concluded that the continuous operation of Hatch Unit 1' for at least one fuel cycle with the subject sweepolet to manifold weld in the as flawed condition does not represent a safety concern, and is acceptable provided that augmented leak detec-tion requirements which will be discussed later, are implemented. The bases for our conclusion are:
(1) The cracks found in the sweepolet to manifold weld are all short transverse cracks.
Any growth in the crack length by IGSCC is limited by the width of the sensitized.HAZ, which is generally not over 1/2 inch.
~
(2) Assuming the unlikely event that the cracks eventually grow through the wall, the leakage emanating from such short transverse cracks will be relatively small and will not cause any significant loss of reactor coolant.
~
(3) As will be discussed later, the licensee is in the process of installing 2 Acoustic Emission (AE) devices adjacent to the subject sweepolet to manifold weld.
AE is a very sensitive
.*s * -
device capable of detecting very tiny steam leaks.
In the unlikely event of a through wall crack, this device will provide an early warning to the operator to' initiate appropriate corrective action.
O i
18 -
(4) As will be discus' sed later, the' licensee _ sill [in61esent' augriedte'd' ~ " _
reactor coolant leakage detection requirements, which include
~
more frequent monitoring and more restrictive leakage limits.
Conclusion of the Fracture Analysis Review The safety margins provided by the overlay repair to the cracked end cap to manifold welds, elbow to pipe weld and pipe to pipe weld and the safe margins of the unrepaired sweepolet to manifold weld are shown by the proposed Code rules to be acceptable.
Crack propagation to the
' extent of leak' age is considered very un1'ikeiy.
~ '
Tearing modulus analyses of cracked welds show that even larger safety margins exist than are inherent in the Code approach.
Auamented Leak Detection The Hatch Unit 1 plant Technical Specificatiods reguire the op,erator
~ ~ ~ ~
~
h to initiate corrective action when 5 gpm unidentified leakage is l
. detected.
NuTech performed an evalaution based on the leak rate calculated iri reference (3) and concluded that there is considerable mar;in between the crack length to produce 5.0 gpm leakage and the
critical crack length.
We have reservations regarding these calcu]a
<l.
tions; therefore, we consider that tighter leak rate limits should be impcsed.
Specifically, we are concerned about the following:
l l
1
1 I
l (1) IGSCC cracks are known to be very tight and branched.
Preliminary l
experimental data provided by Argonne National Laboratory have i
shown that the leakage rate from IGSCC can be less than is usually assumed or calculated.
(2) All reactor water leaked from the pipe during normal operating conditions will not be collected by the sump monitoring system.
A large portion of the leakage will be flashed into
~.
steam or, evaporated before reaching.the sump.
Therefore, for s,_
the sump monitoring system to register 5 gpm leakage, more th'an 5 gpm has to be leaked out from the pipe.
In view of the above considerations, we have determined that the licensee should au'gment his leak detection procedures in accordance with the recommendations in HUREG-0313, Rev.1, by implementing the following items prior to the start-up of Hatch Unit 1:
a.
An additional operational limit on reactor coolant system leakage of an increase in unidentified leakage of two gallons / minute or, more within any 24-hour period. On exceeding this limit, or the existing limits of 5 gallons / minute unidentified leakage -
or 25 gallons / minute total leakage (averaged over a 24-hour period), the reactor will be placed in a cold shutdown condi-tion within 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br /> for inspection.
20 -
b, Drywell leakage will be measured and recorded every four hours.
At least one of the leakage measurement instruments associated c.
with each sump will be operable.
d.
The drywell atmospheric particulate radicactivity monitoring system will be operable or a sample shall be taken and analyzed every,four hours.
We conclude that implementation of the above measures will provide additional assurance that possible cracks in pipes will' be detected before growing to a size that will compromise thi safety of the
~
plant.
in addition, during the~1icensee's presentation to NRC on February 1,1983, the licensee has decided to install two. local acoustic emission devices l
on the flawed sweepolet to manifold weld to monitor the potential leakage.
.This system could provide even more assurance that small leaks will not go undetected.
Inservice Inspection Plan o
The licensee proposed the following future inspection; plans for the austenitic stainless steel reactor coolant pressure boundary piping in their submittal of January 27, 1983:
l-6
,. - ~ - -
.n
+ -.,
e 21 -
(1) The six overlay repaired welds and 50% of the recirculation sweepolet to manifold welds including the one found to have cracks will be examined during the next three successive scheduled refueling outages.
(2) All other, stainless; steel welds will be examined in accordance I
with the licensee's June 29; 1981 response to NRC Generic Letter' 81-04 dated February 26, 1981, regarding the implementation of NUREG 0313, Rev. 1.
C' We have reviewed the licensee's proposal'and have determined that'their proposed inspection plan is not adequate especially regarding the welds
^
in the recirculation piping system.
As NRC is in the process of revising NUREG 0313, Rev. 1, the licensee's June 29, 1981 response to NRC Generic Letter 81-04 will be addressed separately at a later date.
Due to the recent occurrences of IGSCC in recirculation lines in several BWR plants including Hatch Unit 1, such lines must be considered " service-sensitive," and augmented ISI.must be consistent with recommendations of NUREG 0313, Rev. 1.
Therefore, to increase the assurance of ihe integrity of recirculation piping in Hatch Unit 1, we have determined
that an augmented ISI program sir.ilar to that delineated bIslow should be carried out during the next refueling outage.
e 4
0 e
. = =. -.
ew..
.c._____ __
e.
(1) The six overlay repaired welds and the one unrepaired sweapolet to manifold weld should be ultrasonically examined.
(2) A minimum of ten welds in recirculation piping of 20 inches diameter, or larger should be ultrasonically ex' amined.
Those '
circumferential welds not ultrasonically inspected during l.
1982 refueling outage should be selected for inspection first.
(3) A minimum of ten welds of the jet pumps inlet riser piping and I
associated safe-ends should be ultrasonically examined.
Those circumferential welds not ultrasonically inspected during 1982 outage should be selected for inspection first.
(4) Stainless steel piping welds in other systems should be examined in accordance with the guidelines provided in the NUREG 0313, Rev.1 or its subsequent revision as appropriate.
Summary and Conclusion We have reviewed Georgia Power Company's submittal dated January 27, i
1983 regarding the a:ticas taken during this refueling. outage 'on t'he 0
analyses and repairs of recirculation and RHR piping system welds.
in the Hatch Unit 1 plant.
This include.s description of the defects found, description of repairs performed, stress and fracture analysis and future inspection plan.
I
\\
t.
g e m.-
_--m.-,-
-.- - - - - - +
23 -
We conclude that the Hatch Unit 1. plant can safely return to power and operate in its present configuration at least until the next refueling outage, provided the following items are satisfactorily completed prior to startup:
(1) The Code-required hydrostatic test and nondestructive examination on overlay repaired welds,should be satisfactorily completed.
(2) The additional leak detection requirements as listed in the section on Auamented Leak Detectiod, sh'ould be properly implemented.
~
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, we require that plans for inspection in accordance with the requirements provided herein and/or modification of the recirculation and RHR piping systems during the next refueling-outage be submitted for our review and comment before the start of the outage.
9 e
9 e
I f.
g a
24 -
References Reference 1.
EPRI NP-2472-SY "The Growth and Stability of Stress Corrosion Cracks in Large-Diameter BWR Piping",
July, 1982.
Reference 2.
EPRI NP-2705-SR " Structural Mechanics Program:
Progress in 1981, October, 1982.
Reference 3.
EPRI NP-2472, "The Growth and Stability of Stress Corrosion Cracks in Large-Diameter BWR _ Piping,"
e July, 1982.
e 8+
e e
e j
g
- I
~
e 9
s G
l' f
e e
=
f e
t
"