ML18051A929
| ML18051A929 | |
| Person / Time | |
|---|---|
| Site: | Palisades  |
| Issue date: | 06/11/1984 |
| From: | Conrad H, Rajan J, Wu P Office of Nuclear Reactor Regulation |
| To: | |
| Shared Package | |
| ML18051A928 | List: |
| References | |
| NUDOCS 8406130163 | |
| Download: ML18051A929 (34) | |
Text
1.0 ENCLOSURE SAFETY EVALUATIONS BY THE OFFICE OF NUCLEAR REACTOR REGULATION 1983/84 STEAM GENERATOR EVALUATION AND REPAIR REPORT PALISADES NUCLEAR PLANT DOCKET NUMBER 50-255 INTRODUCTION By letter dated April 19, 1984, Consumers Power Company (CPC) submitted for our review, a report entitled, 11 1983/1984 Steam Gener*ator Evalua-ti on and Repair Program Summa ry 11
, and on April 25, 1984, they provided responses to our reque~t for additional information.
Infof~ation meetings were.held in Bethesda on October 25, 1983, February 21, 1984,
=-
~
~
and April 10, 1984.
Members of the Head.£._uarters and Regi on_.J 11 ~taff,
-* ~
~
-~
~ *.:.->~-:.:*_*....
along with a consultant from -the Oak _Ridg_g_ __ National Laboratory, ~et with CPC on December 7, 1983 at the Cer~.ate Research Center in_
Jatkson, Michigan in order to examine Eddy Current Test (ECT) data recbrds and to observe a demonstration of the-eddy current sy~tem equipment.
The 1983-1984 inspection was performed on 100% of the tubes with a 4 x 4 pancake probe.
The results of the final evaluation of the 1983-84 ECT data show a t_otal of 181 tubes with cracks and 29 tubes with inter-granular attack (IGA) greater than 50%.
With preventive plugging, a total of 277 tubes were.plugged in both steam generators.
r-94061301 03 040011... * :'. \\
I PDR ADOCK 05000~5~ - *'.
- )
P:
2.0 BACKGROUND
Pali~ades is a pressurized water reactor having two Combustion Engineering designed U-tube steam generators.
Each steam generator contains 8519 Inconel-600 tubes with a 0.750 inch outside diameter and a 0.048 inch minimum wall thickness.
On January 15, 1973 after approximately one year of t_ntermittent operation at less than full pow.er, the Palisades plant e,xperienced
,1
- its first steam gener~tof tube lea~.
ECT examinations of the.tubing detected general wastage attack in the U:bend area of tubes i*n_:t;[le
~-*-
first eleven rows from the divider plate.
The attack was attri~ted to the use of a coordinated phosphate. secondary water chemistry: *
~-
All tubes in these first eleven rows were plugged, and the plBot r~turned to service in March, 1973.
The p 1 ant operated at essentially full power unt i 1 August 11, 1973, when it was shut down because of ex~ess steam generator tube leakage.*
ECT examinations in September, 1973 showed measurable wastage on nearly half of the steam-generator tubes.
Evaluation of required tube strength showed that tubes with wastage less than 60% through-wall could remain in service.
All tubes with greater than 60% wastage~
defects were plugged.
, _ _. ~;:-.~::::.. :...
- In May 1974, preoperational hydrotests identified two leaking tubes.
ECT showed that wastage had increased and that IGA accompanied by shallo_w pitting was present between the. support plates in the upper hot leg.
All tubes with indications of IGA and those with greater than 50% wastage were plugged.
The IGA was attributed to a m~~ture of reduced forms of sulfur and sodium phosphates.
Beca~se the IGA appeared to be growing rapidly while the plant was in cold shutdown, the plant was permitted to return to powet at levels up to 60% in o_rder to flush or volatize any sulfur or solub.le.p_hos-
=
~
phorous compounds from the steam generat<>r's.
~
~-
Upon return to service in October 1974:, ~e
.. sec9ndary water chem1 stry-treatment was chan~e~ to an all-volatile treatment as recommended by.
the steam generator vendor in order to arrest wastage-type co.rr:_osion.
Subsequent ECT examinations in 1975 through 1981 showed that wastage
- tbrrosion of the steam generator tubing had essentially ceased although minor tube denting was occurring as a result of the switch-over to all-volatile treatment.
In March, 1982, a primary-to-secondary leak:in excess of the Technical
.,_.... ~-7:.:.: *.:.. -.-
Specification limit of 0._3 gpm occurred in steam generator 11A 11 Initial ECT examination (with the standard 11 bobbin 11 probe) of about 35 tubes which had been identifi~d as possible leakers by observation of
- moisture on the tubesheet, sho\\o.1ed no indications of tube degradation other than those seen during previous inspections.
Pressurization of the secondary side of the steam generator identified two leaking tubes.*
Subsequent ECT examinations of the faulted tubes with the 11 bobbi n 11 probe and an advanced 11 pancake 11 -style 4 x 4 ECT probe, which has better flaw detection capability for circumferential ~racks and IGA, i
showed the defects to be through-wa 11 with a ci rcumf e-i'ent i a 1 ori enta-ti on.
The~defect in one tube was found to be high in thi ~traight,
,1*
vertical hot-)eg section of the tube at tube support plate.No:*.9, --
whi 1 e the defect in the other tube was located in the hori zonta~*
.;::=---
~
section of a 11 Batwing.
11 Addi.ti ona l EC.T.~xami nations were perf orffied
,,-c:-_
with both the 11 bobbi n 11 and 114 x 4 11 prob~to pr.ovi de assurance that similar, pluggable defects did not exist in the remaining steam-generator tubes.
Also, in recognition of the fact that one *of~the defects was found near the 90-degree.bend plus the fact that the 4 x 4*ECT probe was incapable of transversing bends, CPC committed*
to develop a probe similar to the 4 x 4 probe but capable of trans-versing the:full length of the tube for the 1983 steam generator inspection.
In 1983-84, CPC performed a 100% inspection of the steam generators with the new flexible 114 x 4 F 11 pancake probe.
Initial ECT data, reported in October and November of 1983, indicated approximately 40,000 "non-quantifiable" indications and a large number of probable cracklike indications.
- 3.0 DISCUSSION An exten$ive effort was conducted by the Licensee to identify the extent and nature of the apparent indications and to identify remedial actions.
Portions of 56 steam generator tubes and approximately 70 tube-to-tube support plate intersections were removed* from both steam ge~erators for detailed metallurgical examination, chemical analysis and materials properties testing in order to correlate ECT indications with any ob5erved degradation.
In addition, the licensee conducted a pro.gram based on 81 l aborat-ory *
..::::=--
~
IGA samples in order to qualify the adeq~acy of the ECT techniqu~ to quantify the depth of IGA, to determi.Q_e..i,l~ sen_?itivity, *and to.'*
~*
~
establish its calibration. They confirmed that the IGA generated in the laboratory is metallurgical_ly similar to that found in the actual steam generator and that ECT measurements on laboratory samples corre-lated with IGA samples removed from the steam generator.
As a result of this program, they have shown that the technique could correlate the volumetric loss of tube wall material due to IGA a~d express it as an equivalent average depth of intergranular attack within a statistical error of+ 6% of wall penetration and to a threshold of detectability of 30% of wall.
Conclusions reached by the licensee from the ECT results and the metallurgical examination ~re summarized as follows:
(1)
Three types of OD initiated degradation are present on the tubes, these were general IGA, pits - shallow.and deep, and IGA spikes.
The deepest pit and the deepest IGA spike had 14% and 19% wall penetrations, respectively, A deep pit on one tube had 16% wall penetration.
\\,*
/'
(2)
The incipient IGA and pits were distributed over all ~he.tube sections, irrespective of their pos:ftion in relation to th~--
~-!"
antivibration bars (AVBs_).
The IGA_spikes appear to be lirM:ted
- ~
to only the AVB intersections, whe~JCT obtained.
indications were
~
(3)
The class of ECT indications which the ECT interpreters had
~riginally classed as 11 non-quantifiable 11 were shown not to be*
related to actual observable defects.
Twelve tube-to-tube support plate intersections with indications that were called either 11 non-quantifiable" or "no-detectable-defect" were metallurgically examined and no degradation greater than 25% through wall were found.
The eddy current signal associated with the non-quantifiable" i ndi cations was apparently caused by surface deposits.pr dents.
One intersection with a "no-detectable indication" was found to contain a 25% degradation.
-~*~.
(4)
Three intersections from the B steam generator cold leg which had been identified by ECT as containing deep quantifiable defects were metallurgically examined and.found to contain no degradation.
The ECT signal was apparently caused by dents or tube deformation.
(5} Five intersections containing actual degradation (three cracklike defects and two IGA) were metallurgically examined.. In all cases,
- ECT correctly called the degradation.
In the case of the two IGA
- defects~ ECT provided an accurate determination of th~maximum average depth.
- ,,.-=
- ::-*
.~
- ~ -*- -:.. -
(6)
The actual crack and IGA degradat i o~ was found to be less taan 75.
mils axial and 130° circumferenti-ar extent.
-~.
(7)
- The actual degradation was found to contain relatively l:arge amounts of sulfur and the tubing was found to be highly sensitized.
It is believed that the corrosion was caused by a reduced form of sulfur acting on sensitized Inconel-600 tubing and that it occurred prior to 1974, since similar pits and IGA spikes were observed in the tubes from the same steam generator in the 1974 examination.
(8)
Most def~cts reside adjacent to dents at the upper or lower edge*cif the tube support plate.
- .-.. (9)
Since sulfur in the form of sulfide is noncorrosive, the licensee chooses not to remove the remaining sulfur from the secondary side
.of the steam generators.
In summary, the results show a total of 181 tubes with cracks and 29 tubes with IGA defects greater than 50%.
The numbers for the 11 8 11 steam generator show 160 tubes with cracks and 22 tubes with IGA defects of greater than 50% depth.
Similar numbers for the 11 A 11 :sit.earn generator are 21 tubes with cracks and 7 tubes with IGA defects greater than 50%.
/*
The licensee concluded that the steam generator tube IGA aDd pitting degradation were caused by* reduced form -Of sulfur introduced i n;:I974.
- 4. 0 EVALUATION 4.1 D~termination of Causative Agent(s)
-:=-
~---;
--~
The lice~see and its consultants conducted extensive microstructural and fractographic examinations on Inconel-600*tubing specimens taken from the Palisades steam generators.
Degradation in the observed specimens exhibited a morphology characteristic of intergranular attack.
In some instances, in addition to -IGA, cracking was also observed.
Austen1tic stainless steels and certain nickel-base alloys, such as Inconel-600, under certain conditions, are known to be susceptible to IGA and intergranular stress corrosion cracking (IGSCC).
The occurrence of IGA requires the alloy to be in a sensitized condition or with a susceptible metallurgical condition and the presence of an aggressive environ~ent. However, for IGSCC to occur, three conditions must be present simultaneously:
(1) a high tensile stress, (2) a susceptible alloy microstru~ture, and (3) an aggressive environment.
Microstructural characterization of specimens taken from pulled tubes from the P~lisades steam generator confirmed that the tubing alloy is typical for Inconel-600.
In addition, the grain size (ASTM 5 and 7Y
~
is representative of mill annealed Alloy~OO microstructure wittl:_-heavy-carbide precipitation along grain bounda[ies. _Modified Hue; t~s~~~g and electrochemical potentiokinetic reactivation (EPR) testing con-
~:.-.
ducted by the licensee on removed tube sections confirmed that.the alloy microstructure has been heavil~ sensitized.
Based on this information, the staff agrees with the licensee's conclusion that the tubing in the Palisades S/Gs has been sensitized and is susceptible to IGA and IGSCC if containments are present in the S/G water.
Because the presence of IGA-causing corrosive species is a necessary condition for the observed tube degradation and cracking, the l*icense~
and its consuitants conducted numerous x-ray diffraction (XRD), scanning electron microscopy (SEM), and energy dispersive x-ray analysis *(EDAX) on speiimen taken from the defective tubes pulled from steam genera-tors.
The analytical results showed that high concentrations of sulfur existed.
Although the presence of chloride, phosphorus, silicon, and calcium was
~lso confirmed, sulfur, in the reduced form, is the onlY identified element that has been associated with IGA or IGSCC of Inconel-600(l, 2, 3)
Sulfur-induced IGA and IGSCC of sensitized stainless steels is a familiar phenomenon(l).
IGSCC of sensitized stainless steel by pqlythionic acid in defulfurizer hydrocrackers, and many other systems.,;.~of the petrochemical industry has been well documented( 2*3).
Furthermore, the'-*sulfur-induced.
/*
IGA and IGSCC have been previously observed in the nuc 1 ear.industry and reproduced in laboratory tests by the. l i ~nsee.
-.:-~-
Based on the above analysis, the staff aarees with the licensee's
~---
conclusion that the type of corrosion observed at Palisades steam generator tube was primarily sulfur-induced IGA and, in some -c~ses, accompanied by IGSCC.
4.2 Postulated Damaoe Scenario Two types of corrosion related tube degradation were observed in the metallography.
Intergrannular attack (IGA), and stress-assisted IGA or IGSCC.
The licensee stated that the reduced sulfur and acidic pH, which occurred in the 1974 dry layup following sodium sulfite (Na2so3) chemistry additions for oxygen control in secondary water, which is no longer used were responsible for the corrosion observed in 1974 and the currently detected steam generator tube degradation.
Given the -referenced chemistry of reduced sulfur aqueous solutions, tht postulated corrosion mechanism is supported directly by the foilowing facts and observations:
.:.~.
(l)
Removed tube degr~dation type, microstructure and sensitization, and reduced sulfur environment.
(2) Other similar commercial US PWR corrosion experience.
(3). CPC laboratory experience with generating IGA in sensiti.zed Alloy-600.
(4)
A definable plant event when sulfur corrosion is known to have occurred (1974 dry layup follo.,.,1ing Na2so3 chemistry (oxygen control).
This scenerio *is considered sufficient to account for local areas of IGA resulting from the dry layup event in 1973-1974.
However, some tubes were found, in certain instances, to contain cracks within the
- IGA.
We attribute these cracks to the stresses imposed (operational-heatup/cooldown axial stresses, bending/dent related local stresses).
(3) l n se.rv1 ce.
Based upon the chemical analysis, and the correlation between the defects and sensitization, we conclude that the observed corrosion can be attributed to reduced sulfur, acidic pH, and the resulting stress-assisted IGA.
The proposed damage scenario is consis..,tnt with known facts pertinent to the observed corrosion.
4.3 Determination of Interoranriular Attack Growth Rate To define an operating allowance, the_:,_J~osee _peveloped repair.
F.
~
criteria for the IGA detected during the 1983 steam generator
~
inspection.
This allowance consists of a minimum of two factors eddy current accuracy and flaw growth rate.
By comparing eddy current signals of pulled tube samples and laboratory IGA samples, and to a lessor degree on numerous historical data comparisons of in-generator suspected IGA indications, the licensee established a period in which IGA is believed to have occurred.
Detail analysis and determination on IGA growth rate are described in Section 3.3 of the licensee's April 16, 1984 submittal.
-.... ---~.-.. The licensee attempted to establish a zero growth rate thr6ugh the comparison of historical ECT data on IGA detec~ed in 1983 with the sensitive 4 x 4 F probe and bobbin probe data collected prior to 1982.
The licensee conceded-- that the ability of the bobbin probe to quantify IGA is poor.
Therefore; their evaluation was qualitative and their argument to establish a zero growth rate was not convincing.
ln additioq, the staff felt that the occurrence of tube leaks in 1982 and the metallographic examination finding of cracks and IGA
- =-
-~
depths of up to 100~~ throughwa l "1 in samp_tes removed in 1983-84..,;..
~
was not cons is.tent with the licensee' 5-contention of no defect growth since 1974.
The staff was uncqnvi~ced by the licensee*1 s*
~--
argument that since plant chemistry records show a small level of leakage that fluctuated between 0.01/gal/mtn and zero from June 1975, that the tube leak event that occurred in 1982 was
~ue to th~ openin~ up of a pre-existing (since 1974) throughwall-crack caused by tube lockup loads imposed by denting which held the tube rigid against the support plates.
For the above stated reason we could not accept the licensee's argument and rule out the possibility of the development of a*
throughwall defect during the next operational period.
We have
... therefore requested, and the licensee has agreed to implement a more stringent primary to secondary leakage rate limit than the 0.3gp_m limit imposed by the Technical Specifiction.
By letter dated*
May 15, 1984, the licensee agreed to reduce the allowable primary to
- secondary leakage limit to 0.1 gallon per minute.
These restrictions will be imposed either by a 11 Standing Order or a change to plant operating procedures.
The leakage limit for p~riods of startup and major load changes wi 11 remc. in as currently speci.ji ed in the Tethnical ~pecification.
We find this acceptable.
4.4 Actions to Prevent Recurrence
-- F-
=
--::-~--
~*/
Considering the nature of the observed steam generator tube degrada- --
~ ~-::..:.
tion, the probable cause, and the rate of progression, the lic~nsee h~s controlled the exposure of _the tube to air or oxygen as the main measure to prevent recurrence of the observed corrosion.
As such no speci~ic changes in current operating practice are necessary.
Con-.
tinued emphasis on reducing periods of air exposure of tubing and controlling dissolved oxygen contents of the operating or layup fluids should be sufficient, since these measures have been effec-tively employed prior to 1978, when the degradation was stiffled.*
Techniques for removal of the stable non-corrosive sulfur (presently as sulfide) has not been commercially demonstrated.
Oxidation of the sulfur to sulfate during desulfurization process would risk additional tube corrosion due to the formation of the polythionate
- anions which are most likely responsible for the observed corrosion.
Based on the above analysis, the staff agrees with the licensee's decision not to desulfurize the secondary side, and concludes that reducing tb.e exposure of the tubes to air or oxygen provides rea-sonable assurance that th*e observed tube degradation will not recur.
4.5 Conclusions We conclude that the licensee has (1) identified the agent which c~used the S/G tube degradation; (2} postulated a damage scenario which is consistent with the known facts pertinent to the observed corrosion; and (3) taken necessary actions to prevent recurrence of the observed SIG tube degradation.
The criteria of Regulat6ry Guide 1.121 provides for an operational allowance over the minimum steam generator tube wall required to preclude tube burst, consisting of two factors, an eddy current
-* -*:*::.. uncertainty and an allowance for possible flaw growth during the operational period.
The licensee has adopted-a plugging criteria of 51% ~hrough wall with the stipulation that all cracklike defects are plugged regardless of depth.
As discussed in Section 5.2 of this safety evaluation, the staff finds that a.defect which is limited to 0.380 inch~s axial, 360° circumferential and 82 percent through-wall is acceptable.
T~is provides an *operational allowance of 31%.
h'e conclude that this combined operational allowance
.for eddy. q.irrent uncertainty and degradation growth rate,.al.ong with the more stringent primary to. second2ry. leakage rate allowance of O.lgpm,.
provides reasonable assurance that the ~ali~ades Nuclear Plant~~!n be operated without undue risk to the publiir.health and safety.
~
5.0 STEAM GENERATOR TUBE INTEGRITY EVALUATION.
The requiremenis for steam generator tube pluggin~ have been calculated in accordance with the guidelines set forth in Regulatory Guide 1.121 entitled "Basis for Plugging Degraded PWR Steam Generator Tubes",
-*---~-Reference *s. The basic requirements are summarized below.
- 1.
for normal plant operation, primary tube stresses are limited such that a margin., of 3 is provided against exceeding the ultimate tensile strength of the tub~ material*, and the yield strengt~ of --
the tube material is not exceeded.
- 2.
Tor accident conditions, the requirements of paragraph NB~3225 ~of Section III of the ASME Code ar~ to be met.
In addition, it must be demonstrated that applied loads are less than the burst strength of the tubes at operating temperature as determined by testing.
- 3.
Fpr all design transients~ the cumulative fatigue usage factor must be less than unity.
In addition, leak before break must be demonstrated i.e., through wall cracks with a specified leakage limit during normal operation do not propagate and result in tube rupture during. postulated accident con-ditions.*
18 -
For the evaluation of the plugging limit for the Palisades tubes the following information was used:
Tubing Geometry:
Material Properties:
Operating Conditions:
0 350 11 Diameter 0.048 11 Thick Sy= 27.9 Ksi at 600°F Su = 80.8 Ksi at 600°F Normal: P. = 2150 psi 1
P 0 ~- _ 77D p s i A-- ~13so psi p -
Steam Line-Break (SLB) _
Ll p = 2150 psi Loss of Coolant Accident (LOCA)
Maximum external pressure, zero internal pressure The demonstration of.tube integrity for uniform thinning of up to 64 -
percent is contained in Reference 6.
For the purpose of establishing the tube-'. intearity in bending for other than uniform thinnina, a
- comparison of the structural capabilities relative to unifonn thinning, was used as the basis.
Since Regulatory Guide 1.121 constitutes an *operating rather than a design *requirement, the allowable stress limits are based on expected
__ l.ower boun_d actual material properties (as opposed to the Code specified minimum v&lues)~ Expected strength properties were obtained from analyses of ten~ile test data of actual production tubing.
These calculations were performed to support the uniform thinning plugging
- ~.,
margin evaluation reported by Reference 6i
. The degradation experienced in the Palisades-~t:eam-generators is* fimite"B~
~
in both axial and azimuthal extent.
The size of-actual defects removeO-
~----
-from the generator exhibits axial lengths of '81'Proximately 0.050" and -
,aiimuthal lengths of about 90 to 100 degrees.
This range of defects can
. be conservatively bounded by considering the axial extent to be 0.075 11 and the.~zimuthal extent to be 135° (which conservatively envelopes the Eddy current testing (ET) limits of detectability of 130° for azimuthal tube degradation) for evaluation purposes.
Defects limited to the extent specified c6uld be analyzed as cracks since Inconel 600 is a ductile material and the crack tip plastic zone could be on the order**of the axial extent dimension.
For a tight circumferential crack, crack tip blunting 'to the axial extent of Palisades defects would occur befor.e crack extension.
For defects longer than the 0.075" considered to be bounding, the burst behavior of the tubes may or may not be adequately described by treating the defects as crack-like.
In addition, since ET techniques are currently inadequate :o accurately describe defect axial lengths below about 0.200 11
, it is considered necessary to evaluate the Palisades degradation without relyin~-bn sp~cifica11y characterizing the defects as having crack-like behavior:
The usual *evaluation procedure e!'l".ployed to assess the effects of tube degradation* assumes that the degradation extent is around the full
.. circumference of the tube and is unlimited in the axial directipn,,(>2.0 in.).: Thus, the analys1s is based on evaluating a uniform tube with a thickness equal to the minimum remaining thick;~s-of the degraded tu~
.This has, in the. past"~ been the usual approach~ased on the expediency~
~- -
of performing the analysis and the lack of a _sig-nificant data base which could be used to quantify the beneficial* effeet of the degradation bei_ng of limited extent *
. Limiting-the ~xtent of degradation has the effect of also limiting the stresses in the material in the degraded area.
For degradation of li~ited axial extent, the magnitude of the hoop stress in the degraded area is restricted by the adjacent thicker material.
For degradation of 1 im1ted azimuthal extent, the adjacent material restricts the magnitude of the axial stress.
The reinforcing effects of adjacent undegraded
- tube material can be quantified using available testing informatioD on tubes with limited extent degradation, coupled with lower bound type collapse evaluation~.
~--,,---*
21 -
Effects of depth of thinning along with axial and circumferential extents.of thinning on the burst strength of Inconel 600 tubing have been represented on a plot of axial extent of thinning versus depth of thinning by a family of curves which present the loci of all geometries having a given burst pressure.
For tubes thinned completely around the ciHur;iferenc*e (360c) over a certain axial length, an empirical equation
- from Ref~rence 4 was used to calculate bursi pressure.
Results from limited axial extent, unifonn circumferential thinning
~.
burst tests are also reported in__ Reference 7.
In addition, an empirical equation was developed for relating the ratio ~ t~e burst pressure-~
. the degraded tube to.the burst pressure for: tJi~ndegraded tube, the
~=-
~-
remaining strength fraction (RSF), to the remaining wall fraction (RWF)=-
of the tube.
A comp-ariso.n of this empirical--.tion*\\*.rith test data indi~ates that for a RWF less than a specified value the bur~t pressure is overpredicted.
For normal tubes, and those in which the thinning is relatively long, the-mode of burst failure is characterized by the opening of an axially oriented split due to the hoop, stress being the maximum primary stress in the tube.
For degradation of limited extent, the hoop stress in the degraded area is restricted by the amount of hoop defonnation taking place in the undegraded region bf the tube.
The axial stress is howev-er, mainly dep~ndent on ~he remaining thickness.
For small axial extent some notch strengthening will take place, but this will be of mino~
significance.
For large depths and small axial extent the axial stress*
will become larger than the hoop stress and the mode_of failure will be a circumferential separation rather than axial. This, in effect, limits the amo~nt of ~trengthening which can-be realized to a factor of 2, the ratio of the magnitude of hoop to axial stress in a tube of unifonn thickness.
The licensee has developed relations to predict the strengthening due this reinforcement that can be achieved by the unde-grade& material.
In order to validate the use of t~ese relations for application to Palisades, the data base in Reference 7 was expanded to incllide:-Westinghouse data for a _variety of tube sizes and data speci*fic to the heats of material used for the PalisadeSktvb-es.
It was
- -~
=-
detennined that *the Reference 7 results are-: i.n;g'eneral conservativ-e -witfi-=-
respect to the added Westinghouse* an*d Palisa.des::--Oata.
On this basis i-r-
_,~
was judged that thes*e relations are appl.icab'le-=t.o Palisades and could be u~ed to quantify the effect of limited axial extent on the burst pressure.
In order to evaluate actual tube burst performance during SLB, a toler-an~e line with a 95 percent probability of forming a lower bound for 95 percent of the population underlying the Reference 7 data was developed.
The final fonn of a plugging ljmit relation for tubes with limited axial extent was developed to satisfy *the structura 1 requirements of Regula to-ry Guide 1.121_.
The requirement for demonstrating adequacy using a factor of safety of 3 relative to the ultimate tensile strength is more restrictive than the primary stress requirements relative to yield strength and accident condition loads.
For the yield strength requirement of*-Regulatory Guide 1.121 it is noted that:
so s
~ = 2.87 Sy Sy = Su/2.87.)-Su/3.0 Thus, limiting pri~ary membrane stress, Pm, based on a factor o~ 3" relative to ultimate is conservative as compared to limiting P to less
.m than the yield stress.
Similarly, the pressure differential. during.. s-te~m-lein break (SLB) is
--~--.
related to the normal operating pressure as:*--.
The stress limit during SLB is the lesser of 2.4 Sm or 0.7 Su where Sm is*found as the lesser of 2 Sy/3 or Su/3 at temperature.
Using the Palisades tube properties, P ts limited as given by the following rn equation:
P ~ 44.6 ksi = 1.67 (Su/3) rn Since the allowable stress is 67 percent larger than that during nonnal operation and the loading is only 56 percent larger, ~he requirement against the ultimate tensile strength is more restrictive.
For unl.imited thinning, it is shown in Reference E that the amount of un-ifonn* degradation which can be accommodated is 64 percent.
For limited axial extent thinning the wall thickness required to provide the same margin against circumferential burst is 0.0085 11
, for an all~wable degradat1on maximum of 82 percent.
In order to account for the rein-
/
forcfog.effect for flaws of limited azimuthal extent, information was*
provided considering the effect on burst press~ of circumferential ~-
~
cracks (representing zero axi a 1 extent)' the -e~c'f of burst pressure or ax i a 1 pa rt through wa 11 and through wa 11 cracis '*Trepresenti ng zero
-*~--.
azimuthal *extent), arid 90° and 180° upper bourrcFlimit solutions for re~t~ngular patth type degradation (finite but limited axial and aximuthal dimensions).
The evaluation for limited circumferential cracking relative to burst pressure is presented in the leak before break evaluation section of Reference 4.
Considering the pressure differential of 4140 psi, burst would be expected for a circumferential crack with an included angle of 145°.
Therefore, for cracks with less than* an included angle of 145°, a margin of 3 against burst at nonnal **.
operaing pressures exists relative to actual test data.
The plugging*
limit developed for 360° thinning can be considered equally applicable for thinning limited in -azimuthal extent to 145°.
__.~...:-.:..:::.:......
Analytical models have been developed to compute the burst stren~th of Inconel 600 tubes with thinned areas of limited axial-and circumferen-tial extent.
Effects of geometry on burst strength have been presented in terms of plots of axial extent of thinning and depth of thinning resulting in.a given burst pressure.
Effects of the extent. of cir-cumferential thinning on burst strength have been represented by a number of ~trength locus curves corresponding to different arc lengths of circumferential thinning. On the ba5is of these evaluations it has
-been demonstrated.;hat a tube with through wall degradation up to ~.075 inch~s in axial extent and up to~l45° in atimuthal extent is sufficient to withstand nonnal operating pressure with a f;ctor of safety of 3. -~
"C,.
An additional requirement from Regulatory Gui.de l.121 relative to burst
--~.-.
at steamline break, i.e., beyond the eva1uatiC>'r1of primary membrane
$tre~s, i~ that margin be provided against the ultim~te burst pressure
.. as determined by burst tests performed at operating temperature.
For this evaluation the lower.tolerance limit line from the Reference 4.data was used to determine axial burst values expected for steamline break conditions.
Since the lower tolerance limit is adequately below all of the ad~itional data plotted and the original data, it provides a suffi-cient basis for evaluating axial burst pressures.
In addition, a lower tolerance value for the undegr~ded burst pressure was used based on results reported in References 8 and 9. This evaluation demonstrates that axial burst pressur~s for the limited axial extent plugging
.* criteria for steamline break are above the criteria established using a safety factor of 3 based on ultimate tensile strength.
Beyond the evaluation requirements on-primary membrane stresses and the steamline break ultimate test pressures, Regulatory Guide 1.121 also
...... _requires that degraded tubes be evaluated in accordance with the re-quirements of the ASME Code Section III, Paragraph NB~3225.. This paragraph of the ASME Code invokes the rules of Appendix F of the Code for the evaluation of faulted condition limits.
In general, there are
,1 no primary bending stresses at degraded tube locations.
This is either due to:
(1) the ability to withstand bending fflc.ments is not*necessa~
- to satisfy requirements for equilibrium, 01 {-2fdefonnations under
~
accident conditions are limited by either the a:.:nTf-vibration devices or
--~.
the general defonnation of the overall tube 'buT'Tdle, thus defonnation of the degraded tube is displacement controlled.
However, a comparison was made for the limited axial extent plugging margin criteria with the uniform ~hinriing plugging margin criteria.
In this case the ultimate limit moments were calculated for each degradation condition.
As a limiting condition for the limited axial extent degradation, a 135° through wall crack was considered.
This crack represents the maximum ex~~cted azimuthal extent in the Palisades steam generators.
For thii condition the ultimate bending moment for the limited axial extent plugging criteria is 12 percent higher than the limiting ultimate moment for a unifonn.thinning criteria of 64 percent.
In addition, a compari-son of the section properties was made for degradation extending to 135°
.:. 27 -
and 82 percent throughwall.
In this case both the ~oment of inertia and the section modulus for the limited axial and azimutbal extent plugging margin criteria is greater than that for a tube degraded uniformly to 64 percent thro~gh wall.
Cpnsidera~ion of external collapse per the Code requirements can be made b~sed on ~ollapse data presented in Referen~e 1
7 The reference document demonstrates that for 0.875 inch diameter by 0.050 i~ch thick tubes, which have a collapse strength less than 84 percent of that for_ the Palisades 0.750 inch diameter: by_ 0.048 inch thick tubes, that the collapse pressure is *relatively unaffected for~niionn thinning 3/8 *~
. an inch long by 360° *azimuthally for degradat.i'3TI u-p to 60 percent -
F*-
i:..
through wall.
For degradation at*8o_percenf-th'roiigh wall, the collaps~
~--~,~
,pressure was demonstrated to be t to 1 times* k1fe undefected tube col- -
lapse pressure.
For all collapse tests performed, which included degradation up to 80 percent of the wall and 1.5 inches in axial l~ngth, the col}~pse ~res~ures found were well in excess of the external secon-dary pressure during LOCA.
In addition the limited degraded tube is ju~ged to have more resistance to collapse than a tube which has been uniformly degraded with unlimited aximuthal and axial extent up to 64 percent through wall.
In order to examine the fatigue *resistance for the limited extent plugging*.margin criteria, the evaluation was made of the tolerable stress concentration factor for a tube with no degradation, but
28 -
includjng the effects of denting (to consider the bounding effect of a tube locked-in at the tube support plates), vibration and thermal stresses. It was found that a maximum allowable stress concentration factor (SCF) of 5.9 could be tolerateJto result in a 40 year design life us~ge factor of 1.0.
For degradation of li~ited axial extent, excluding cracks, the maximum stress concentration factor required by the ASME Code is 5.0.
.For. the case of th.rough wall cracking, it is noted that there are,,Very few cycles of significant stress-levels. This results in low alternat-ing fracture stress intensity values.
The alteryating fractu.re stress;:
~
. intensity is directly related to the alternanle stress perpendicular ttr l!'-
~- -
~
the flanks of the crack multiplied by the sqµare-root of pi times i the crack* length, times a function of the crack length to thickness ratto._
. For the tase of the Palisades tubes, the alternating_stress is judge~ to be of the order of 7 ksi.
For crack lengths which were on the order of
. twice -t~e wall thickness; which would be the case for degradation with stress corrosion cracks, the alternating stress intensity factor would be less than the threshold required for crack growth.
In addition, very deep cracks would be needed to.plastically yield the remaining ligament of the tube.
Crack growth would be on the order of.the cyclic crack tip displacement, in which case several mils of growth would represent.a*
generous fatigue cratk growth allowance.*
-,.-:.:.=:.:~.....
In su~r~ the Regulatory Guide 1.121 structura*1 requirement for normal operating conditions (maintenance or a factor of safety of 3 against burst) and acci~ent conditions (compliance with pararaph NB-3225 of Section III of the ASME Code) have beeh met by the tube with 360° thinning of 82 percent with axial extent limited to 0.380 inches.
Such
.. - a-defect also satisfies the fatigue usage requirements for all appJicable* design transients.
~. l Leak Before Break Eva 1 ua ti ons
.t' Two separate approaches were considered applica"f?-e -to the Palisades
- defects to demonstrat~ leak before break.
Tire Tirst of these is b~sed ~
~
~---.
on the observed morphology that indicates that the defects could be
-* ~~-..
evalu*ated as cracks.
The second is based on considerations of finite
.axial extent in accordance with the range of the structura 1 criteria distussed earlier.
These evaluations demonstrate that through wall cracking would develop prior to the crack reaching the critical circumferential length for a postulated SLB event. The growth of part through wall cracks in tubes at Pali~ades have been observed to exhibit a limited aspect ratio which results* in extension through the wall prior to reaching the critical SLB bursting 1 ength.
- Furth~nnore, based on ge.ometri ca 1 considerations, circumferential crack extension. beyond 60° will lead to the axial bending stresses being a maximum at the crack front thus encouraging preferential _growth through the wall.
Evaluation 9f leak before break for degradation with finite axial extent was perfonned considering the limited extent upper bound of 82 percent to be unifotm thinning extend-ing 360° circumferentially and of unlimited axial extent.
Final infor-mation is provided int.he form of a pa-fr of curves, one for cracks leaking at the technical specification limit and the other for burst a
- SLB.
THe leak before break margin can be read as the distance between the curves* as a function of crack extent and degradation depth.
Based on these analytical evaluations the licensee has demonstrated that through wall crackjng would develop prior to the crack reaching.the length for a postulated SLB event.
5.2 Conclusions*
~:p Based* on a. review of the analyses and test data provided by the 1 icensee
, in Section 4 of Reference 4 the staff concludes:
(1) The structural
. limit of 82 percent for defects which are limited in axial extent to
. 0.075 and up to 135° in azimuthal extent is sufficient to meet the normal operating conditions, accident conditions and fatigue usage requirements outlined in Regulatory Guide 1.121. This includes the
~aintenance of a safety factor.of 3 relative to ultimate strength which is the most restrictive primary membrane stress limit.
(2) Leak before break has been demonstrated for degradation of extent limited such th~t it can be considered to behave as a crack.
(3) A tube with 360° thinning of 82-percent with axial extent limited to 0.380 inches, has a safety factor against burst comparable to the ~argin afforded pressure vessels designed *;n accordance with Section III, of the ASME B and PV Code.
(4) Leak before break of circumferential cracks has been
31 -
demonstrated for unifonn thinning degradation, extending 360° circumferentially and of unlimited axial extent, to a depth of 69 percent of the tube thickness.
( 5) The structural limit for uniformly thinned tubes with unlimited axial and circumferential extents is 64 percent thinning.
In accordance with these conclusions, a defect which is limited to 0.380 inches axial, 360°_, circumferential and 82 per~el\\-t through-wall,is.
acceptable.
Therefore, as discussed in Section 4.5 of this report, the threshold for corrective action was determined by reduci~ th.e 82 percent throu~all by (1) an allowance for eddy current uncertainty,~nd (2) an allowance fo~_ppssfble
~
~
~
-., __,_~-'.:"-.....
fl ow growth during th.e operational period.
l:.
- The staff finds the.repair criterion of 51% tti.r.._~'h-wall acceptable for-Palisades steam. genera tors.
6.0 SUPPORT PLATE DEGRADATION CONSIDERATIONS Evidence of den~ing rel~ted tube-to-tube support plate (TSP) interaction.
was observed during tube pulling and support plate removal operations at Palisades during the present outage.
In certain instances large forces were.required to remove tube specimens.
Steam generator tube denting*
results in tubes being locked into tube support plates.. The configura-tion of the TSP/tub~ bundle and c6rrosion mechanism leading to denting results in the development of compressive forces on the tubes and expansion of *the tube support plate. This interaction can eventually lead to significint distortion of the TSP and cracking at areas 6f high stress concentration (e.g., TSP flow holes and slots).
Secondary side visual inspe*ction of TSPs 14, 13, 12 and portions of 11 at the accessible periphery and in locations where TSP sections have been removed in the Palisades steam generators revea1ed no significant tube support plate degradation as a result of tube denting.
In addi-tion, ET dat~ indicates denting growt~_rate has decreased in recent years.
However, for conservatism, the structural analyses were performed to provide the basis for the determination that Palisades steam generator tube bundle integrity is maintained even if a single support plate is postulated to be missing, i.e. to lose its load carrying capabil_ity or fail to provide support for a tube.
A'primary load capability evaluation of the Pab:isa-Oes steam generate~~
tube bundle under an assumed genera 1 degrad~~j ~
w.as conducted using postulated SLB, LOCA and SSE loading*.
~
Tubing c'o:t l-apse potenti a 1 and tne resulting. loss of tube flow area during a pos.t.£a'ted [OCA and SSE event were analyzed. A detailed evaluati-0n of the Palisades steam generator tubes in terms of responses to such forcing functions as turbulence and
- vorte~ shedding mechanisms and fluid elastit stability criteria was completed (i.e., a normal operation fluid-structure interaction ev(I ua ti on).
Fatigue effects resulting from the fl ow induced motions were also evaluated:
Worst ca~e vertical and horizontal tube spans were analyzed with a single tube support plate missing.
Finally, tube wea~.
estimates were calculated assuming a postulated TSP fragment and partial iSP support.
...,_ - ~-* ;::...
l
6.1 CONCLUSION
S Based on review of the analyses presented by the license relative to the postulated. loss.of a single support plate, the potential tube flow area loss due to tube collapse during a LOCA and concurrent SSE, tube vi-bration and postulated loose plate pieces the staff concludes:
- ~- ***.*
The postulateq loss of a single support plate does not adversely affect the tube bundle integrity during LOCA, SSE or SLB accident condition loadings.
The potential tube flow area loss due to tub~ cq_llapse during LOCA.
and concurrent SSE is equivalent to less thaIL260 tubes (of whi~h 63 are currently plugged) with the postula.ted_Jps*s of load carrying -
capability of a ~ingle support plate.
~----
Tube vibration during normal operation should not ~e significant f~om flow-induced vibration mechanisms considering the postulated loss of a single support plate.
.Neither tube vibration nor postulated loose plate pieces lead to premature wear-through of a tube.
7.0 ACKNOWLEDGEMENT H. Conrad, P. Wu, and J. Rajan prepared this report.
Dated:.:)une 11, 1984
{ \\
.l:"'.
References
- 1.
Dravnieks A. and Sama:ns C. H., Corrosion Contr.ol in Ultra Forming,"
American Petroleum Institute, 37 (III), page 100, 1975.
- 2.
Samans, C.
H_. "Stress Corrosion Cracking Susceptibility of Stainless Steels and Nickel Base Alloys in Polythionic Acids and Acid Copper Sulfate Solution," Corrosion, 21, page 256, 1964.
'~-~---
- 3.
-*Ahmad, S. et al., "Stress Corrosion Cracking of Sensitized 304 Austenitic Stainless Steel in Petroleum Refinery Environment," Corrosion, Vol. 38, No.
6~ page 347~ 1982.
- 4.
Consumers Power Company Report ".1983/1984 Steam Generator Evaluation and Repair Program," April 1984. -_, =
- ~
-- ~
- 5.
USNRC_Regulatory Guide 1.121, "Bases for: Plugging Degraded PWR Steam Generator tubes," Issued for Corrment-;-August 1976.
- 6.
Combustion Engineering Report, "Analysis to Detennine Allowable Tube.Wa 11 *Degracia ti on for Pa 1 i sades Steam Generators," P. Anderson et al., Rev. 2, March 30, 1976.
- 7.
USNRC NUREG/CR-0718, "Steam.. Genera tor Tube Integrity Phase l R.eport," J. Alzheimer et al., September 1979.
- 8.
- CE Report CENC-1256, "Tube Burst and Leakage Test (Palisades),"
J. K.
Hayes~ February 1976.
- 9.
Sattel.le (Columbus) Topical Report, "Examination of Iconel 600 Tubi*ng from the A Steam Generator of the Palisades Nuclear Plant,"
\\.I. E. Berry et al., July 11; 1974.
..,. __ ;.:_:"*;...****