ML20070G121
| ML20070G121 | |
| Person / Time | |
|---|---|
| Site: | Monticello  |
| Issue date: | 12/10/1982 |
| From: | Office of Nuclear Reactor Regulation |
| To: | |
| Shared Package | |
| ML20070G118 | List: |
| References | |
| NUDOCS 8212210548 | |
| Download: ML20070G121 (19) | |
Text
.
UNITED STATES
~q,A NUCLEAR REGULATORY COMMISSION s
%, * * "
- j)
WASHING TON, D. C. 20555 SAFETY EVALUATION BY THE OFFICE OF NUCLEAR REACTOR REGULATION SUPPORTING AMENDMENT NO.14 TO FACILITY OPERATING LICENSE NO. DPR-22 NORTHERN STATES POWER COMPANY MONTICELLO NUCLEAR GENERATING PLANT DOCKET NO. 50-263 1.0 Intro' duction During this current outage, a planned replacement of piping insulation was carried out, that permitted inspection of all welds in the Recirculation System at the Monticello Nuclear Generating Plant.
The primary inspection was performed by ultrasonic (UT) methods, augmented by radiography to assist in evaluating suspect locations. This inspection resulted in the detection of cracks in 3 safe end to pipe welds and one pipe to elbow weld in the 12" Riser piping, and one crack in the 22" diameter Manifold End Cap to pipe weld.
On the basis of these results, the licensee decided to reinforce the End Cap weld and one Riser pipe to safe end seld with a weld overlay similar to that performed earlier on Quad Cities 1 Reactor Water Cleanup System piping, even though the cracks were considered to be very shallow.
During touch-up grinding preparatory to the weld overlay, a leak was N
noted on the Riser pipe to safe end weld.
The leak occurred at a different location than the indications identified by the ultrasonic examination. After the leak occurred, additional ultrasonic inspection was barely able to identify the leaking crack. After sealing the leak, the weld overlay was successfully accomplished.
Because the leak was not identified by ultrasonic testing (UT) and concern s
8212210548 821210 PDR ADOCK 05000263 P
2 t
~
that the cracks were deeper than originally determined, the licensee decided to overlay all the Riser welds showing indications.
During' this overlay process., two more pipe to safe end welds developed
's 3
small leaks, were sealed, and the weld overlay applied. The weld 1
overlay was accomplished without incident on the Manifold End Cap weld.
After all repairs, a hydrostatic test at 110% of operating pressure was performed.
This resulted in the detection of another very small leak in another Riser elbow to pipe weld.
This was successfully sealed and overlay welded.
In all, 3 Riser safe end to pipe welds and two Riser elbow to pipe welds were found to be cracked, and were reinforced by a weld overlay.
In addition, one cracked Manifold End Cap to pipe weld was reinforced by overlay welding.
In addition, this Safety Evaluation addresses an application dated July 6, 1981, in which the licensee proposed changes to the Technical Specifications to revise the requirements for the coolant leak dectection system.
2.0 Discussion On October 19, 1982, the Commission issued a Confimatory Action Letter which requested information on the cracks found on the recirculation system piping at Monticello.
Specifically, the Confimatory Action Letter requested that the licensee submit to the Commission, the results of the licensee's inspection, corrective actions, justification to return to power, and receive NRC concurrence 1
3 t
k-before returning the unit to power.
In a November 22, 1982 letter supplemented December 3,1982, the licensee submitted additional infonnation.
~
2.1 Description of Cracks Table 1, from the licensee's submittal of November 22, 1932, describes the details of the results of the inspections prior to the detection of the leaking elbow to pipe weld in Riser G during the hydrostatic test.
Note that all except two kere determined to be very short axial (90' to the weld) cracks, tiecause they are very short in A
comparison to the wall thickness, they are reported by the licensee as
" radial".
Short axial cracks have been noted previously, and leaks emanating from them were noted and reported at Quad Cities 1 in i980.
~
They probably 6ccur Tn'lo'ca'tions w7t3 high residual welding stresses in the circumferential direction.
They are typically short, because the sensitized heat affected zone extends less than 1/2" on either side of the weld, and intergranular stress corrosion cracking (IGSCC) requires sensitization to be present. As noted at Quad Cities 1, however, 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 e
i circumferential cracks that can grow through the wall and around the circumference of the pipe, for two reasons.
First,thestresson the axial crack is almost all caused by pressure, and typically the pressure stress is low compared to the totalltress acting on a circumferential crack, where bending stresses can be significant.
4 I
l
~
Second, because IGSCC is confined to sensitized material, they cannot grow to significant lengths.
This point will be covered more f'ully later un' der Fracture Analysis.
~
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 often can only be detected ht very limited transducer locations.
f This appears to have been the case on the Monticello riser welds.
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, even when they are through wall, they will cause very litte actual leakage, perhaps not enough to be detected with normal procedures.
s In summary, although axial or radial IGSCC cracks are hard to find by UT, they will cause only small leaks and will not grow long enough to initiate a pipe bufst unless the piping itself is completely sensitized.
2.2 Description of the Overlay Reinforcement The weld overlay on the Riser piping welds consisted of a complete circumferential reinforcement nominally.545 thick.
The nominal thickness of the piping is.75".
The axial length of the overlay
5 I
.y is 6", tapering at a nominal 18' angle. ' Tha overig on the Manifold t
End Cap to pipe weld is similar, with a thickness of.5" on a I
wall thickness of L1", and is 5 inches in minimum length.
The effect of the overlay is to provide a reinforcemen,t of IGSCC re.tistant material. The welding process also induces beneficial compressive residual stresses in'the underlying cracked pipe, f
in both the hoop and axial direc,tions.
2.3 Code Stress Analysis i
The repaired piping was evaluated according to Section III, and was found to meet all requirements including seismic and fatigure requirements. 'This was done by conservatively taking no credit '
Is for the entire circumferential volume where the cracks were detected.
Adoughnutshapedgroovewasassumedtoberemov$dinamannerto r' '
remove all of the cracked area.
Although the geometrical configura-tion is not typical of Code design, the stress analysis was performed
~
Osing the Code rules.
The fa'tigue analysis' used the stcidard set of i
f transient conditions, and included a strength reduct$on factor of 5 in~the calculation.
The calculations show that the repair of joints in the Riser piping and the fianifold End Cap met all Code requirements.
for at least 5 years.
t f
[
s Y
t e------,-
.--.-------,,..-w,--w--
.-.--+,--py,,
,A.,
, -o, e
...-r.
,,,-,------,-,,-----,-w--
-.,v-w.---.
,c
-s
./s t
6 a/
j 3.0 Evaluation i
3.1 Effect of the Overlay Repair on the Recirculation System
/,
The weld overlay shripkage induces beneficial residual compressive stresses in khe cracked pipe, but also causes other effects. These have been evaluated by NuTech for the licensee, and the results of their evaluation are summarized here. The overlay repair to the Riser safe end to pipe weld causes both an axial and radial shrinkage.
one.effect, caused by the 5/16" ra tal shrinkage is to compress the
.F
, area of the safe end where the internal secondary thermal sleeve l
is attached. The only, deleterious effect anticipated is that the ring nut holding the plate sprir.g and secondary thermal sleeve in place will be compresse1, m'aking removal by unthreading difficult
/
if not impossible. When the safe ends are replaced the parts can be cut apart. ~
Another minor effect is caused by the axial shrinkage pduced by the weld overlay in the horizontal runs of the riser piping.
This includes the three safe end to pipe welds, and the elbow to hort-zontal pipe weld.
In each case, a bending stress of approximately
^
7 ksi is induced in the Sweepolet to pipe weld.
As this is a displacement _ controlled stress, similar to a thermal stress, it i
represents a small additional secondary stress, and is acceptable.
a, A more significant effect is caused in riser D; where the vertical 7
p,ipe to albow weld was overlayed., Because the horizontal run to 4
t
~
7 t
~'
the safe end is fairly short, the bending stress induced at the pipe to safe end weld is large enough to require careful consideration.
Although the anal is performed treating this displacement controlled stress as an additional thermal stress showed that the limits of Section III of the Code are not exceeded, the propensity for IGSCC at the safe end to pipe may be increased.
We have concluded that this does not represent a serious safety concern, for the following reasons:
.L 1.
The safe end to' pipe itelds at Montic'eT1'o do not appear to be
~
partic'larly subject to circumferential cracking from IGSCC, u
which would 'se the type caused by high bending stress.
All safe end to pipe welds were inspected, and indications of circumferential1y oriented IGSCC was only found on one joint, (E Riser) where it appeared to be associated with short axial cracks, and was very short (1.06")'.
2.
The bending stress induced by the weld overlay is displacement controlled (self equilibrating loads) and would tend to,be relieved by initiation of cracking.
3.
If cracking did occur from this bending stress, ~it would tend to be asymmetrical, thus propagating through the wall in a local area. Thus it would be expected to Igak, and thereby be detected long before it could propagate circumferential1y to an extent that would jeopardize the overal.1 integrity of the pipe.
8 t
Therefore, we conclude that the possible increase in propensity for IGSCC in D Riser Safe End to pipe weld does not constitute,a significant,s3fety concern even if cracking should develop.
~ ~
'BecauseM.h4'MinTfdid'fhd-Cap overlay is litMe end of a piping run, the shrinkage induced has no effect on other parts of the system.
3.2 Fracture Analysis 3.2.1
Background
NuTech performed two types of fracture analyses to show that the safety margins against failure are at least equivalent to the margins inherent in the ASME Code.
One analysis method used is based on a new proposed flaw evaluation methodology for Section XI of the Code.
This includes IWB 3640,
" Acceptance Criteria for Flaws in Aust.enitic Stainless Steel Piping," and the associated Appendix C, " Evaluation of Flaws in Austenitic Stainless Steel Piping." Although thest. new sections,
have not yet been approved through the. Main Committee, they'have been approved through the first two levels, and, full approval is expected at the next Code meeting. The NRC will. review these modifica-tions to the Code, for concurrence.
y 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
..,, _. - - -. - l
9i
~ h'as been done under EPRI' contracts, and has been described in several I
reports, including References 1 and 2.
For Code use, this calcula-tional method has been used to develop simple tables, from which l
acceptable flaw sizes and shapes as a function of applied stresses can be read directly. These are Tables IWB 3642-1 and -2 for axial cracks, and Table 3641-1 and -2 for circumferential cracks. There are separate tables for Normal Conditions and Emergency and Faulted Conditions, with different rafety margins.
The tables pravide 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.
A comparison of the criterion with results of actual burst tests on stainless steel piping will be made later in this review, when the repair to the Manifold End Cap is discussed.
~
Note that the presence of more than one crack does not change the calcu-lations.
Multiple axial cracks do not interact, and are trected separately.
Because safety evaluations of flaws must include considera-tions of future growth, proposed Appendix C also includes rules for
~~
calculating growth by fatigue and stress corrosion.
The methodology for evaluating fatigue propagation appears acceptable, but we still have some reservations about the crac'k growth rates for IGSCC given in the. code. This is of no. concern for the repaired cracks at Monticello, l
(as will be described later) but it could affect our evaluation of other Cases.
O e
e a
,.,._,.,,,_a.
,n
l 10 f
3.2.2 Riser Flaw Repair Evaluation NuTech perfonned an Appendix C evaluation in accordance with the proposed Appendix C of the Code of the most limiting flaw and repair on the riser
- piping. The short axial (or radial) flaws were approximated by using an assumed one inch long thru-wall flaw in the pipe, and using minimum
' pipe and overlay thicknesses. Table IWB 3642-1 wn used to determine the allowable depth into the combined pipe-overlay wall. This gives a value of
.75 a/t, or.89".
The total thickness is a minimum of 1.19", comprised of a minimum wall thickness (allowing.for counterbore) of.687", plus a minimum b
overlay of.50".
Thus, the Cod 2 would pennit crack growth by fatigue and stress corrosion of (.89
.687") or.203".
NuTech calculated the crack growth due to fatigue to be only.005" during the next 5 years of operation.
NuTech also calculated growth by IGSCC in an axial direction.
(IGSCC is r,-'
expected to occur in the type 308L high ferrite weld overlay) and concluded that the maximum expected growth would add only.009" to the length of the existing crack.
Both of these values are ins,ignificant.
The calculations for the allowable depth of the crack are overly conservative in this case, because the Code arbitrarily cuts off the. allowable depths given in th'e tables for axial cracks to
.75 a/t.
Extrapolation of the values in the table would show that thru-wall cracks would be acceptable at the stress levels existing at these joints, if it weren't for the leakage problem.
j
= r,
. =
1.
4 11 f
We have checked NuTech's calculations and agree with its conclusions regarding the acceptability of the repaired riser welds according to Appendix C.
We also conclude that, because of the truncation of the tables at.75 a/t, even more margin than. required by the Code actually exists in these repaired welds.
For example, at the design pressure, the Code would permit a crack this deep,to be almost 6" long.
Further, because the Type 308L overlay is not subject to IGSCC, essentially no growth of the existing cracks is to be expected, making the repaired welds less likely to cause future problems than the unrepaired welds in the system.
' ome of the cracks in the riser welds were missed by the ult'rasonic S
inspection, and were only discovered,'by leakage. We have no assur,ance i
that other rispr weln do not also have short axial cracks essentially thru-wall. We have performed calculations in accordance with, Appendix C to evaluate the safety margin that would be' expected, should such undiscovered cracks be present.
Because the tables only include crack depths up to.75 a/t,' graphic extrapolation was used to estimate the length of a thru-wall flaw that would be acceptable from a safety stan'dpoint. This approach yielded a value of about 1.8 inches for a thru-wall axial crack in an unrepaired riser pipe joint. The maximum i
1 expected length of an axial IGSCC crack would not be more than about L70 inches.
This value assumes that a crack could grow completely across the weld at the OD, and 1/2 inch into b,ase metal on both sides of the weld.
As the sensitized zone probably does not extend' even 1/2 inch from the weld, this is a very conservative estimate.
~
12 i
We conclude, therefore, that the maximum length of a possible thre-wall axial crack would be acceptable using the limit load analysis, with a safety margin on pressure of at least 2.5.
3.2.3 Manifold End Cap Flaw Repair Evaluation NuTech performed similar calculations to show the acceptability of the End Cap weld repair.
In this case, it used the pipe thickness without overlay for its evaluation, even though the overlay was actually done. These calculations also showed complete Code acceptability using the proposed Appendix C and IWB 3640.
i Thesecalculatkonsaredependentontheaccuracyofthesizingof the shallow axial crack that was reported, and could be invalidated by either incorrect sizing or rapid IGSCC growth of the detected crack.
We therefore performed additional calculations using other assumptions.
The first calculation assumed that the detected crack was completely through the.95" pipe wall, and that the overlay was the specified minimum of.50".
This approach then, assumes a crack of.63 a/t in a wall 1.45" thick.
Using the table for Normal Conditions gives s
an acceptable flaw length of over 8", at the stress levels caused by <
design pressure (1248 psi vs. 1000 psi for operating).
We consider l
it very improbable that a crack of such length could form considering the compressive residual stresses induced in7the' pipe by the weld overlay.
I
y, -, - -,, _ - - -,,
_-,,_-,_,._-m,..,,,-.
_,.., _-,-,w,m-.,-m-
13 This particular geometry, 23"~0D and 1.45" wall, is,very similar to that of stainle,ss steel pipes used by Battelle for burst tests.
We compared the burst test results with the acceptaMe values in IWB 3642-1. We used the conditions imposed on the specimen used for test Number 25 described in Reference 3.
In this test, a section of 24" diameter pipe with a 1.5" wall was used.
A f7aw was machined in the pipe in the axial direction that was 6" long und.9 inch deep.
This almost duplicates the geometry used in the aboes t alculations for p.
the End Cap weld repair. The burst test.did not m in an actual.
burst, but was ter\\pinatedbecauseofexcessivelesay d the inability of the test specimen to hold pressure. The maximum pros ere attained was 4050 psi. With this crack geometry, the proposed Anr
'd C would permit a maximum stress ratio (stress /S,, where S, is the D;ie spo 54ad allowable stress intensity) of.75; whereas failure occurred tt z stress ratio of 1.92, demonstrating a safety, margin on pressure v at least 2.55 against burst.
We also performed Appendix C calculations in at:cordance with the proposed Appendix C of the Code for the hypothetica' case of an l
undetected thru-wall axial crack in another end cap. Under i
operating pressure, and again extrapolating the Code tables to thru-wall geometry, the results show that a thru-wall crack 3.2" long would meet the Code criterion. The maximos length expected by IGSCC would be less than 2".
We conclude, therefore, that even thru-wall undetected cracks in other end caps are not likely to grow to a size that would decrease the Code intended margin.
14 1
3.3 Tearing Modulus Fracture Evaluation NuTech also performed fracture mechanics calculations using the more sophisticated Tearing Modulus approach. This. type of elastic plastic fracture mechanics was initially developed on an NRC contract, and has been widely accepted and used during the past 5 years.
It is recognize'd that the limit load approach is conservative, and' that much larger margins are actually present in many cases.
r Tearing Modulus calculations were performed for both the repaired Riser welds and the Manifold End Cap weld.
As expected, the cal-culations show that very large margins against failure are present.
Although the material properties in the actual pipes and oyerlays may be somewhat lower than those used for the calculations, it is apparent that margins well over those intended by the Cc,de are shown to be present by this approach.
3.4 Conclusion of the Fracture Analysis Review The safety margins provided by the overlay repair to the cracked Riser and Manifold End Cap welds are shown by the proposed Code rules t.ited above to be acceptable.
Crack propagation to the extent of leakage is
~
considered very unlikely.
Staff calculations using the same Code rules also show. acceptable safety margins for postulated undetected and unrepai. red thru-wall cracks in Riser and End Cap welds, although sm'a'11 amounts of leakage would occur.
l
15 t
Tearing Modulus analyses of cracked welds show that even larger safety margins exist than are inherent in the Code approach.
l 3.5 Augmented Leak Detection By letter dated July 6, 1981, Northern States Power responded to Generic Letter 81-b4, " Implementation of NUREG-0313, Rev 1, Technical Report on Material Selection and Processing Guidelines for BWR Coolant Pressure Boundary Piping.* NUREG-0313, Rev. 1 recommends that leak detection should be augmented for plants that have piping susceptable to IGSCC.
In this letter, the licensee requested a change in the Technical Specifications to revise the Limitina Conditions for Operation and Surveillance Requirements in the leak detection systerr.
The improved leak detection proposed by the licensee consists of the following:
a.
In addition to the existing Technical Specification limit of 5 gpm Unidentified leakage, the licensee propo.ied to revise the Technical Specifications to add a condition, that in the event of an increase in unidentified leakage of two gallons / minute or more within any 4-hour
~ period, or 20 gallons / minute total leakage (averaged over a 24-hour period), the reactor will be placed in a cold shutdown condition within 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br /> for inspection.
t.
Drywell leakage will be seasured and recorded every four hours.
?,
c.
At least one of the leakage measurement instruments associated l
with each sump will be operable.
l l
16 f
d.
The drywell atmospheric particulate radioactivity monitoring
' system will be operable or a sample shall be taken and analyzed
~
every four hours.
We conclude that iklementation of these measures provide the aug-mentation irecomended in NUREG 0313. Rev.1, and wil.1. provide additional assurance that possible cracks in. pipes will be detected before growing to a size that will compromise the safety of the plant.
Therefore, we
).
find that the proposed changes to the Technical Specifications are acceptable.
3.6 Sumary and Safety Conclusions We have reviewed Northern States Power's submittals regarding actions taken during this refueling outage on the Recirculation Piping System in the Monticello Plant.
This includes location and descriptions of the defect found, description of repairs performed, stress and fracture analyses of'the present configuration of the system, and plans for augmented leak detection.
We 'one'lude that the Monticello plant can safely re urn to power c
and operate in ^.:: present configuration at least until the next l
refueling outage.
Nevertheless, we still have concerns regarding the long term growth
, of small IGSCC cracks that may be present ;.4 not detected during this" inspection.
Further,wefeelthattheeffectoftheadditkonal~
17 i
stress imposed on the Safe End to pipe we'1d in Riser D by the overlay i
may increase the probability of initiation of IGSCC at this location.
For these reasons, we require that plans for inspection and/or modifica-tion of the Recirculation Piping System during the next refueling, outage be submitted for our review and comment before the start of the outage.
4.0 Environmental Considerations i
We have determined that the amendment does not authorize a change in effluent l
types or total amounts nor an increase in power level and will not result in any significant environmental impact.
Having made this dete,rmination, we have furthe'r c'onc1uded=6 fit tReTm7ndmen't"idvIGlis 'aYacti7n~which is insignificant
~
from the standpoint of environmental impact, and pursuant to 10 CFR Sec-tion 51.5(d)(4) that an environmental impact statement,-or negative declaration and environmental impact appraisal need not be prepared in connection with the issuance of the amendment.
l 5.0 Conclusion i
~
We have concluded, based on-the considerations discussed above, that: (1) because the amendment does not involve a significant increase in the probability or' consequences of an accident previously evaluated..does not create the possibility of an accident of a type diffierent from any evaluated
- previously,,,
~
and does not involve a significant. reduction in,a, margin of safety, the
. cr..
amendment does not involve a significant hazards consideration, (2) there l
is reasonable assurance that the health and s'afety of th' public will not be e
endang-ed by operation in {'he proposed manner,'and (3) such activities will be 1
,,,_-.---,-~-..~-.-_.,-v-----------,..-.-,m..--
..~-..=w-w-.-, -
18 f
conducted in compliance wit'h the Commission's regulations and the is'uance of this s
amendment will not be inimical to the comon defense and security or to the health and safety of the public.
Dated:
December 10, 1982 Principal Contributors: Warren Hazelton Helen Nicolaras O
I
+
' References f
Reference 1.
EPRI NP-2472-SY '"The ' Growth and 3t' ability 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.
BMI-1866 " Investigation of the Initiation and Extent of Ductile Pipe Rupture," July,1969.
5
- e
.g 4
,