ML20084N540
| ML20084N540 | |
| Person / Time | |
|---|---|
| Site: | Crane  |
| Issue date: | 05/24/1983 |
| From: | Johnston W Office of Nuclear Reactor Regulation |
| To: | Lainas G Office of Nuclear Reactor Regulation |
| Shared Package | |
| ML20079G498 | List:
|
| References | |
| FOIA-83-243, FOIA-83-A-18 TAC-47484, NUDOCS 8306030387 | |
| Download: ML20084N540 (65) | |
Text
--
e Gr i
5 2 j-f e%'o UNITED STATES.
!E 7g NUCLEAR REGULATORY COMMISSION P #
n
,h E
WASHINGTON D. C. 20555 C. w./
MEMORANDUM FOR: Gus C. Lainas, Assistant Director for Operating Reactors Division of Engineering
~
William V. Johnston, Assistant Director FROM:
Materials, Chemical & Environmental Technology Division of Engineering
SUBJECT:
SAFETY EVALUATION OF THREE MILE ISLAND UNIT NO.'l FOR RETURN TO SERVICE AFTER REPAIR OF OTSG CORROSION (TAC #47484)
Plant Name: Three Mile Island Unit No. 1 Utility: General Public Utilities Nuclear Corporation Supplier: Babcock & Wilcox, Gilbert Licensing Stage: OR Docket No.: 50-289 Responsible Branch & Project Manager: ORB #4, J. Van Vliet DOE Reviewers:
C. McCracken (Lead), L. Frank, S. Kirslis, J. Rajan, D. Sellers, P. Wu Task
Description:
Safety Evaluation of the OTSG's and RCS for return to service Review Status:
SER Complete Attached is the evaluation by Chemical Engineering, Materials, Engineering, and Mechanical Engineering Branches. This review is based primarily on the licensee's Topical Report 008, Rev. 2 which was submitted by letter dated March 31, 1983. We have reviewed the cause(s), repairs and corrective actions associated with the OTSG corrosion problem.
Our consultants, Franklin Research Center (FRC), Brookhaven National Laboratory (BNL), Pacific Northwest Laboratory (PNL), Dr. Digby Mcdonald, Ohio State University (OSU) and Oak Ridge National Laboratories (ORNL) have provided assistance in the evaluation. The TER's from these con-sultants are attached to the SER. However, attachment No. 1 from FRC is a proprietary document and therefore is not enclosed.
In the SER we have not addressed the long-term potential for adverse results I.
during 0TSG leak / hydrostatic testing, peroxide flushing and pre-critical hot functonal testing.
However, we do anticipate that some tube leaks will be found and additional tube plugging or repairs implemented, using the procedures that have been evaluated in the SER.
In case of an unanticipated occurrence, the plant Technical Specifications will ensure that the licensee carries out these functions in a manner which will not adversely affect public health and safety and that the staff will be notified if additional problems arise.
(
T3cGo303879 XA Copy Has Been Sent to PDR
~i
Gus C. Lainas C On the basis of our evaluation, we conclude that:
the causative agent of the corrosion and its potential sources have been identified; the remainder of the RCS has been returned to a serviceable condition; the repair method established a new pressure boundary without significantly degrading the OTSG performance; the contaminants have been or are being removed and procedures incorporated to prevent re-introduction; and administrative procedures'will be incorporated to require plant shutdown well before reaching the current Technical Specifications leakage limit of 1.0 gpm per steam generator.
Therefore, we find that GDC 1, 14, 15 and 31 have been met as they apply to minimizing the possibility of a rapidly propagating failure of the RCS pressure boundary and that reasonable assurances exist that the public health and safety is protected.
To ensure timely verification of anticipated performance we recommend that the license be conditioned to require:
1.
A mid-cycle ECT (SER Section 3.3) 2.
Administrative limits for primary-to-secondary leakage (SER Section 3.3),
3.
Results of the extended life qualification program, by the first.
refueling (SER Section 3.4) and 4.
Timely notification of results from the lead corrosion test program (SER Section 3.6).
T6e suggested wording for the four proposed license conditions is ir.
(
Enclosure No. 1.
i g'
g,,..,,, E.. a.. ', w William V. Johnston, Assistant Director Materials, Chemical & Environmental Technology Division of Engineering
Enclosure:
As stated cc:
R. Vollmer L. Frank D. Eisenhut S. Kirslis j
T. Novak J. Rajan J. Knight D. Sellers V. Benaraya P. Wu J. Stolz P. Grant B. D. Liaw S. Young (Resident)
R. Bosnak H. Gray (Region)
C. Cheng J. Van Vliet S. Pawlicki T. Sullivan i
H. Conrad W. Hazelton j (
H. Brammer P. Grant l
t
[
~
.-__- - - = - -
m.
s Enclosure No. 1
('
Proposed License Conditions for TMI-1 OTSG Return to Service 1) the licensee will conduct eddy-current examinations, consistent with the inspection plan defined in Table 3.3-1, either 90 calendar days after reaching full power, or 120, calendar days after exceeding 50%
power operation whichever comes first (SER Section 3.3).
2) an administrative shutdown limit will be imposed if primary-to-secondary leakage increases significantly above baseline. The licensee will establish a new baseline leakage limit during the precritical hot functional test.
If primary-to-secondary leakage exceeds the new baseline by a significant amount the plant will be shutdown and leak tested. Current information indicates that increases in leakage of approximately 0.1 gpm are measurable. The
(
licensee should establish a limit for detecting increases in primary-to-secondary leakage upon restart. This limit should be either 0.1 gpm or an established measurable limit which does not significantly exceed 0.1 gpm.
If leakage is due to defects in the tube free span, the tube will be removed from service.
Subsequent to shutdown for primary-to-secondary leakage, the amount of leakage from within the tube /tubesheet crevices will be considered as background, and will be used to establish a new baseline at restart, provided that the present technical specification limit of 1.0,gpm is not exceeded (SER Section 3.3).
3)
The licensee will submit the extended life cyclic program qualification test results by the first refueling (SER Section 3.4).
4)
The licensee will provide' routine reporting of the long term corrosion " lead tests" test results on a quarterly basis as well l-as more timely notification if adverse corrosion test results are
(.
discovered (SER Section 3.6).
~
G
(.
Sc.fety Evaluation GPU Nuclear Corporation Three Mile Island Nuclear Station, Unit No. 1 Docket No. 50-289
- 1. 0 Background By letter dated April 30, 1982, the licensee indicated its intent to conduct steam generator repairs under the authority of 10 CFR 50.59.
By letter dated August 23, 1982, the staff responded to the licensee by stating that the steam generator problem contained several aspects that appear to involve unreviewed safety questions. The staff's major concerns were:
1.
The corrosion mechanism and extent of corrosion in the steam generators are unique. The Staff had not reviewed the potential consequences of additional plant operation. subsequent to the
(
repair of the defects.
In particular, the potential for this type of corrosion to reinitiate during operation and to rapidly progress, thus adversely affecting the steam generator primary pressure boundary, will need to be reviewed by the Staff..
2.
The potentiaT exists for this type of corrosion to attack other l
primary pressure boundary materials.
l The proposed tube repair technique has not previously been tpproved by the staff as an acceptable method for repairing defective steam generator tubes.
l l
l O
~
2-4.
Since portions of the tubes within the tubesheet contai.i defects
. greater than 40% through-wall and the repair method fo-the majority of these defects will not involve plugging, an amendment to the plant Technical Specification (TS) 4.19 will be needed prior to return to power operation.
On October 13, 1982 the staff issued a safety evaluation for the OTSG kinetic expansion repair technique.
In the safety evaluation, the staff found that the repair process did not involve an unreviewed safety question or a modification to the technical specif.ications, and hence' could be conducted without NRC approval. However, the staff stated that NRC review and approval of the overall program to return the OTSG to service was required prior to any subsequent power operation.
This SER is the staff's review of the overall program to evaluate the licensee's justification for return to service.
(
~
.Cp
M
. 2. 0 Introduction In late November 1981, while performing reactor coolant system hydrostatic testing with the reactor shut down, primary-to-secondary system leakage was
~
detected in both once-through steam ' generators (OTSG).
Subsequently, e detailed examinations revealed many of defective tubes. Metallo-graphic examination of portions of removed tubes confirmed that the tube failures were initiated from the primary side (ID) of the tubes in the form
,, of cirew.iferential stress-assisted intergranular cracks.
The active
chemical impurity causing the corrosion was sulfur in reouced forms, which had been inadvertently introduced into, the reactor coolant system. The vast majority (approximately 95%) of the defects occurred within the top 2-3 inches of the 24-inch thick upper tubesheet (UTS). The corrosion attacks most rapidly at,.the air / water interface and during layup. The air / water interface was located in the UTS'during a significant portion of the post-hot-functional shutdown period. To repair the defective OTSG tubes within the UTS, the licensee has applied a kinetic explosive expansion repair
(
- tecnnique. This will establish a new primary pressure boundary, by expanding and tightly sealing,the tubes within the tubesheet below the degraded region, thereby establi hing a new 'l'eak limiting / load carrying seal.
The licensee has
~
applied the kinetic expansion repair technique to all tubes within the UTS to ensure that all potentially degraded tubes are sealed. Tubes which have defects'that are not repairable by the kinetic expansion process have been removed from service by plugging.
\\
l The licensee's overall program to return the OTSG to service includes:
S r
determinatic,n of the causative agent (s)
~
l eximinations.of the~ remainder of the reactoricoolant systems (RCS)'
l I
OTS rp
[
0TSG' performance evaluation subsequent to; repairs e'
cleanup of the contaminant
[
procedures to prevent a re-introduction of contaminants crack arrest considerations
\\
)
g% <
4
[
j.s,
l
s
. 2.1 Description of the Repair Method i
The as-built OTSG has two twenty-four-inch thick tubesheets, one at the top (UTS) and one at the bottom. The once through straight tubes are nominally fifty-six feet and one inch in length, with fifty-two feet of
~
heat transfer surface between the tubesheets. The remaining four feet and one inch includes two feet in each tubesheet and one-half inch protruding into the primary head at each end of the OTSG.
To provide structural integrity and leak tightness during shop fabric-ation, all tubes are hard rolled to a nominal depth of one and one-quarter inches, and seal welded on the primary :ide of the tubesheet surfaces. The tubesheet has a nominal twenty-two and three quarter-inch long, eight-mil radial gap (crevice) between the outer tube surface and the drilled tubesheet hole. As stated in the introduction, the preponderance
- of defects are located within the top two inches of the UTS with a rapidly decreasing number of defects down through the depth of the upper tubesheet.
(
The repair method utilizes a kinetic expansion process to form the tube against the tubesheet; i.e., close the eight-mil radial gap.
The kinetic expansion process closes the eight-mil gap and produces an interference fit between the tube OD and tubesheet drilled holes ID to achieve a leak-tight, load-carrying joint. The tube repair procedure requires that all repaired tubes have a two-inch defect-free unexpanded section within the UTS above the secondary side tubesheet interface. This unexpanded section will prevent tube pullout in the event of a defect at the repair transition zone.
Developmental testing has been conducted to demonstrate that a kinetically expanded six-inch long defect-free section of tube (qualification zone) can provide the necessary leak tightness and load carrying capabilities required for operation. Therefore, all tubes which have defects down to a depth of sixteen inches into the tubesheet can be repaired. The sixteen-inch section plus the two-inch unexpanded zone and the six-inch qualification zone account for the full depth of the twenty-four-inch thick tubesheet. The forming technique
(
-c
-=n
. consists of inserting a polypropylene sheath into each tube. The poly-propylene sheath contains a prima-cord which, when ignited, forces the colypropylene sheath against the tube. The resultant force expands the tube.
Tae polypropylene sheath and prima-cord assembly is called a candle.
The candles are detonated by a blasting cap which is maintained outside the steam generator in a sealed container and ignited electrically by a licensed blaster. The two OTSG have a total of approximately 31,000 tubes, all of which have been expanded. After all tubes-were expanded, those tubes which contained non-repairable defects were plugged.
Prior to kinetic expansion the crevices between the tube and tubesheet were flushed with hydrazine-treated water and then dried out with electrical heaters to remove moisture.
i e
a 9
i I
I 1
t
- = -
=.
p..m
.e
,mdeom.
e.w6-ew=
6-3.0 Evaluation By letter dated' March 31, 1983, with attached Topical Report 008, Rev. 2, the licensee provided the Assessment of TMI-1 Plant Safety for Return to t
Service after Steam Generator Repair. Topical Report 008, Rev. 2. super-ceded Topical Report 008, Rev. 1 that was submitted by letter dated l
December 10, 1982.
i Several consultants helped NRC in the evaluation of the licensee's program for return to service. The staff consultants included representatives from the following organizations.
Brookhaven National Laboratory (BNL; Attachment No. 2)
Franklin Resea.cch Center (FRC; Attachment No.1)
Oak Ridge National Laboratory (ORNL; Attachment No. 5)
Ohio State University (OSU; Attachment No. 4)
Pacific Northwest National Laboratory (PNL; Attachment No. 3)
In addition to the evaluation conducted by the staff and staff consultants, the licensee established an independent third party review group (TPR) of industry experts who are not employed by the licensee or its affiliated companies.
By letters dated April 4, and May 20, 1983 the licensee submitted copies of the TPR ffnal reports. The TPR provided an independent operational and safety evaluation which is enclosed as Attachment No. 6 and consists of an original report dated February 18, 1983 and a supplemental report dated May 16, 1983.
In the cover memo for the May 16, 1983 supple-mental report, the TPR concludes that " comments and recommendations relating to safety of the steam generator repair have been satisfactory resolved by GPU nuclear." Some additional comments by the TPR which are related to plant operation rather than safety issues are being considered by the licensee. However, resolution of these operational comments is not necessary to ensure that the public health and safety is protected.
(
l
~"
'~
j('
Our evaluation is divided into eight areas associated with the cause, repair and recovery from the OTSG corrosion problems:
l
{
1.
. Determination, of causative agents 2.
Examination of the remainder of the RCS 3.
OTSG examinations to determine the extent of m radation 4.
OTSG repair I
5.
OTSG performance evaluation subsequent to repairs 6.
Cleanup of contaminants 7.
Procedures to prevent re-introduction of contaminants 8.
Crack arrest considerations 3.1 Determination of Causative Aaent(s)
The licens'ee and its consultants conducted extensive microstructural and fractographic examinations on Inconel 600 tubing specimens taken from the TMI-1 OTSG. Cracks in the observed specimens exhibited a morphology characteristic of stress-assisted intergranular attack.
Austenitic stainless steels and certain nickel-base alloys, such as the Inconel 600, under certain conditions, are known to be susceptible to intergranular stress corrosion cracking (IGSCC). However, the occurrence of IGSCC requires that three. conditions be present simultaneously: 1) a high tensile stress, 2) a susceptible alloy microstructure, and 3) an aggressive environment.
Stress analysis on the OTSG tubes conducted by the licensee indicates that: 1) axial tensile stresses in tubing are largest during cooldown when they may approach the yield strength, 2) axial tensile stresses also exist during cold shutdown, 3) locally high axial tensile stresses are possible in the seal weld heat affected zone and in the vicinity of the roll transition, 4) under heatup at full operating temperature, the hoop stress generally is larger than the axial stress, and 5) the axial stresses are generally larger at the periphery tubes than those in the center of the tube bundle. Based on this analysis, the licensee concluded l
M M
that the tube cracking occurred during cooldown or cold shutdown because tensile stresses were highest under these conditions.
By independent analysis (Attachment 1), the staff consultant confirmed the licensee's finding that high tensile stresses exist in the cooldown and cold shutdown conditions and their distribution is consistent with that reported.
Consequently, the staff agrees with the licensee's conclusion that IGSCC of the OTSG tubes occurred during the cooldown or the cold shutdown conditions. Furthermore, the stress level and its distribution are consistent with the observed failure pattern.
Sensitization is a term used to describe the precipitation of carbides at grain boundaries of alloys during high-temperature exposure. The essence of' sensitization is' that the rate of attack of Inconel 600 and austenitic stainless steels under intergranular attack of stress-corrosion-cracking conditions is sensitive to the carbides precipitated in the 800-1600*F (427-871*C) temperature range. The staff's review of the TMI-1 OTSG fabrication history indicates that the Inconel 600 tubes were present in the steam generators when they were given a post-weld stress-relief heat treatment for 18 hours2.083333e-4 days <br />0.005 hours <br />2.97619e-5 weeks <br />6.849e-6 months <br /> at about 1150*F (621*C).
In the staff's opinion, this treatment produced a sensitized microstructure in the Inconel 600 OTSG tubing. Therefore, the staff agrees with the licensee's conclusion that the sensitized microstructure of the alloy also contributed to the overall degradation and cracking of the Inconel 600 OTSG tubing in TMI-1.
The same conclusion was also reached by the staff consultants (Attach-ments 2-4).
Because the presence of crack-causing corrosive species in the reactor coolant is a necessary condition for the observed tube failure, the j
licensee analyzed numerous samples of the reactor coolant and
~
interconnecting systems to identify the corrodant(s).
The analytical i
results show high concentrations of sulfur in the reactor coolant.
In i
addition, the presence of sulfur was also observed and identified at the l'
cracking surface as well as the inside surface of the pulled steam k
i
{
i
~.
.g-generator tubes. The licensee concluded that the presence of sulfur, as discussed later in this section, and oxygen, because of venting the RCS through the Control Rod Drive Mechanism (CRDM) vent during shutdown period, has led to the OTSG tube degradation.
Sulfur-induced stress corrosion cracking of sensitized stainless steels has been known for a long time.(1) Intergranular stress corrosion cracking (IGSCC) of sensitized stainless steel by polythionic acid in desulfurizers, hydrocrackers, and many other systems of the petrochemical industry has been well documented.(2,3)
Extensive laboratory testing by the licensee and staff consultants (Attachment 2) has demonstrated that sensitized Inconel 600 is susceptible to IGSCC in polythionic acid environments at low temperatures just like sensitized stainless stesis.
Furthermore, th'e cracking morphology was the same as that observed on the damaged Inconel 600 tube sections pulled
(
from the TMI-1 steam generators. Based on the results on reactor coolant l
chemistry analyses and metallurgical examination, Scanning Electron l
Microscopy (SEM), and Energy Dispersive X-Ray Analysis (EDAX) characteri-zation of the pulled TMI-1 OTSG tube samples, the staff agrees with the licensee's conclusion that the failure mode of the steam generator tubes l
at TMI-1 was sulfur-induced IGSCC. The same conclusions are also stated I
by staff consultants (Attachments 2-4).
In an effort to identify the source (s) of sulfur contaminaton in the reactor l
coolant system, the licensee and its consultants examined operating history 1
and coolant chemistry control of TMI-1 before and after the hot functional tests conducted during. August-September, 1981. Generally, the reactor coolant system remained within specifications for those parameters for which an analysis requirement existed for the period extending from April 1979 through November 1981. Three instances of intrusion of ionic substances and oil not accounted for by specific analyses have been identified. The ionic species from the first contamination incident in July 1980 were e
. (
removed from the bulk liquid by demineralization in August 1980. The ionic species from the second contamination incident in May 1981 appear to have been only partly removed by processing through a resin precoat filter in August 1981. The licensee estimates that 1 to 2 ppm thiosulfate residual could have remained in the system at the start of September 1981.
Further review of the operational history shows that additional sodium thiosulfate in the RCS may have resulted from injections of the Borated Water Storage Tank (BWST)
~
contents during the September 1981 cooldown from hot functional testing.
In the Topical Report 008. Rev. 2, the licensee stated that sulfide (S~2) and possibly sulfur and other intermediate species were generated because of hydrogenating and heating the water to perform a hot functional test.
Subsequent cooling to room temperature and oxygenating following the hot
-2 functional tests rapidly oxidize S to S and possibly other sulfur species with higher oxidation states.
In addition, the licensee also indicated that oil or grease and sulfuric acid contamination of the RCS occurred in 1979.
The licensee further concluded that the sulfate-contaminated water and the temperature and oxidation potential transient associated with the hot func-tional test provided an aggressive environment which led to the IGSCC of OTSG tubes.
The licensee proposed.the following failure scenario for the TMI-1 OTSG tube degradation:
a.
During layup the primary system was contaminated with sulfur by the accidental introduction of sulfuric acid, sodium thiosulfate, and possibly a sulfur-containing oil. The amount of sulfur present may have reachad several ppm, but the contaminated water was not aggressive enough to crack mill annealed plus stress-relieved Inconel 600. The corrosion tests confirm that cracking would not have been expected to occur at this stage.
(
b.
The temperature and oxidation potantial transient associated with the hot functional test resulted in a change in the types and con-concentrations of sulfur species present in the primary water.
Further changes occurred when thiosulfate-contaminated oxygenated water was injected during the tests of the High Pressure Injection (HPI) and Low Pressure Injection (LPI) systems.
c.
When the water level in the OTSG's was lowered following the hot functional test, high concentrations of aggressive metastable sulfur species developed in the dry-out region at the top of the generators due to the.combinert effects of solution concentration by evaporation and the comparatively high availability of oxygen.
Changes in the sulfur species in the more dilute bulk solution
~
proceeded more slowly resulting in lower concentrations of aggressive sulfur species.
{
d.'
Sulfur-induced IGSCC of the Inconel 600 tubing occurred rapidly in N.
the dry-out zone with preferential attack at high stress locations in the most highly sensitized tubes. Cracking occurred to a lesser extent at the lower portion of tubes in the generator.
Statistically this would be expected because the bulk solution was less aggressive than the environment seen by tubes in the dry-out zone.
e.
Cracking terminated either because continued chemistry changes resulted in the formation of less aggressive sulfur species or because the environment.in the dry-out region was diluted by the slowly rising bulk solution. By the time the water level was dropped again, the chemical state of the sulfur in the primary water was sufficiently different from its state immediately after the hot functional tests to prevent a recurrence of above steps c and d in the new dry-out zone.
m w o m ae
'O c
p y
12 -
f.
Cracking wis discovered when the OTSG's were pressurized.
The staf" concludes that sodium thiosulfate at concentrations of 4-5 ppm is the contaminant which most likely caused the OTSG tube degradation.
The staff finds that although the introduction of oil or grease and the inadvertent contamination of sulfuric' acid would have caused intrusion of sulfur-bearing species into the reactor coolant,. based on plant operating history in the period between 1979 and 1931, the amount of sulfur present in the reactor coolant system, as estimated by the licensee, could not be accounted for from the oil and sulfuric acid intrusions alone.
Furthermore, staff consultant evaluation (Attachment 4) indicated that sulfur in oil exists in the form of organo-sulfur compounds in the lowest oxidation state, state, therefore, it is unlikely that they made any significant contribution to the polysulfur species inventory in the RCS.
In case of sulfuric acid, the sulfur exists in tNe highest oxidation state, and hence would not act as a source of elemental sulfur due to the kinetics that existed in the RCS.
C' Furthermore, sulfate itself has not been known to cause IGSCC of Inconel 600 or stainless steel under the cold shutdown conditions.
In the case of sodium thiosulfate, the sulfur is in the S*2 state.
It could be reduced during hot functional testing, when the dissolved oxygen concentration is low, and subsequently oxidized again during cooling down or the shutdown period while the top of the OTSG were vented thrcugh the CRDM vent. The dissociated or activated sulfur species generated during this period t
1 could have contributed to the OTSG tube cracking. On this basis, the staff I
agrees with the licensee's conclusion that the main source of polysulfur l
species was the thiosulfate contamination, and the cracking occured during the cooling period or the shutdown period after hot functional testing.
The licensee stated that it is not possible to deduce from available past data the identities or concent aations of transient sulfur species which were present in the reactor coolant system following the hot functional test.
The staff is of the opinion that cracking of the sensitized Inconel 600 OTSG tubing was most likely caused by absorption of elemental sulfur or some
(
13 -
highly. active form of sulfur species, formed by dissociation of the thiosulfate. One postulated mechanism is that these sulfur species depassivated the crack tip surfaces by forming various metal sulfides, which lead to rapid propagation of the crack. The staff consultants expressed the same opinion (Attachment 2-4).
The staff consultant (Attachment-3) expressed concern about an incon-
~
sistency in the licensee's Topical Report 008, Rev. 2.
In pages 13-14 of this report, the licensee stated that sulfur reduction might have occurred during the hot functional test, and that the subsequent OTSG tube degradation was as a consequence of reduced sulfur species.
In the Test Section of the same report, laboratory data indicate that cracking of sensitized type 304 Stiinless Steel (SS) and Inconel 600 specimens in low temperatura, cxygenated water, contaminated with thiosulfate proceeds without the presence of other reducing agents. Apparently, the consultant's concern is that in one case reduced sulfur species is suggested as the corrosion initiator, while in the other case it is shown that corrosion will occur in the absence of reduced species. We are of the opinion that irrespec-tive of the exact scenarics, the thiosulfate contaminant and its source, the thiosulfate tanks, have been removed. The intermediate states which may have contributed to the degradation of the components are not germane to the staff's final. conclusion that at TMI-1 thiosulfate contamination combined with the presence of ox; gen caused the OTSG tube degradation.
Conclusion Based on the above evaluation, the staff determined that the licensee has identified the presence of sulfur in the reactor coolant system and on the
~
l OTSG tube inside surface. The staff concludes that the main source of sulfur contamination was from the thiosulfate tank. The specific mechanistic steps involved in the sulfur-induced stress corrosion cracking phenomenon have not been clearly established; however, the fact that thiosulfate, like tetrathionate, can cause IGSCC of sensitized stainless steels has been
(
well recognized and investigated since the 1950's(1), and furthermore, l
i
~
~ : _E___.
. (
experimental results obtained by the licensee and the staff consultant indicate that the TMI-1 steam generatur tubing specimens cracked in borated aqueous solutions at room temperature with thiosulfate con-centration as little as one ppm. Therefore, we conclude that sulfur-induced SCC is the main cause of tha TMI-1 OTSG tube degradation and
,nat it occurred during the cooldown'or cold shutdown after the hot functional tests. The same conclusion was also stated by tha staff consultants through an independe'nt evaluation (Attachments 2-4).
(
o e 6
em>
~
v-ns-w g,
+
y
.~
3.2 Examination of the. Remainder of the RCS The staff evaluated the suitability of the existing TMI-1 reactor coolant system and primary side auxiliary system materials and components for continued safe operation. To accomplish this, at the staff's request, the licensee developed a test program to determine whether any corrosion damage of the RCS materials and components o'Jtside the OTSG occurred as a result of exposure to potentially aggressive environmental conditions known to be present in portions of the reactor coolant system.
As part of the inspection program were generic B&W problem areas, including core barrel bolting, fuel hold down springs and High Pressure Injection (HPI) thermal sleeves.
The inspection plan instituted by the licensee included material / stress combinations within three different environmental conditions described as follows:
(
The primary coolant liquid / vapor interface areas.
This is where the known attack occurred in the OTSG; Those areas that are wetted during hot, pressurized operations and dry when the plant is cold, depressurized and under shutdown conditions; and Areas fully covered by reactor coolant water.
Each inspection was evaluated by one or more of the following methods:
Ultrasonic inspection Visual inspection Eddy current inspection Dye penetrant inspection Destructive metallurgical examination
(
Radiographic examination Functional check examination Wipe (surface samples)
C To perform the inspection of the remainder of the reactor coolant system, the licensee removed the vessel head and the plenum. The plenum was installed in the deep end of the refueling canal for inspection. Once the plenum was removed, many of the inspections, such as ultrasonic inspection of plenum cover to cylinder bolts, and the video examination of upper core components were conducted in parallel. While these parallel inspections were being performed, the following components / parts were removed from the
~
plant and shipped to B&W Lynchburg Research Center (LRC) for destructive testing.
Three pieces of vessel inner 0-ring Regenerative Neutron Source Retainer Control Rod Drive Mechanism (CRDM) end closure To determine if contaminants had caused corrosion of other RCS materials, the licensee instituted an inspection and requalification program.(4)
The inspection program included visual, remote visual, ultrasonic, eddy current tests, dye penetrant, radiography and destructive examinations.
The materials selected for examination were picked by listing all materials in the RCS and making a determination as to their susceptibility to sulfur-induced corrosion. Components representing virtually all RCS materials were examined. Those materials which were determined.to be most susceptible and which were destructively examined included Inconel 600, Inconel X-750 and 304 SS.
Examination of the RCS included removal of the reactor vessel head, inspection of reactor vessel internals, CRDM's and two fuel l
bundles.
Extensive inspections were also coriducted in systems which interface with the reactor coolant system. An integral part of these inspections was a requirement to expand the scope of the inspection if corrosion was found at any location.
(
_ C During a supplemental examination of systems which interface with the reactor coolant system, evidence of sulfur-induced corrosion was found in a waste gas system stainless steel line. The extent of corrosion was quantified and all corrosion affected sections in the waste gas system have been replaced. Based on environmental similarities between the waste gas system and the pressurizer, the power operated relief valve (PORV) was removed for examination. Components of the PORV were
~
found to exhibit pitting corrosion attributable to sulfur which could have reduced the valve's capability to function but did not affect its structural integrity. By letter dated May 20, 1983 the licensee provided results of the pressurizer corrosion examination.
Examination of the PORV block valve, the connecting piping, safety relief valves, and the remainder of the pressurizer ~ revealed only shallow pitting on one of the two safety relief valves.
Based on this examination, the PORV was rebuilt with uncorroded parts and both safety relief valves were replaced.
-[
In addition.to the pitting corrosion found in the PORV, cracks were noted in the metallographic examinations of the Inconel 718 vessel head 0 rings. The cracks observed in the reactor vessel 0-ring have been characterized by the licensee as ductile ruptures as indicated by the dimpled feature of the fracture surfaces. The staff examined the surface topography of the existing defects and agrees with the 1.icensee's conclusion that the fractography is j
not indicative of stress corrosion cracking (SCC) morphology.
Further, the indications observed on the surface were very shallow (.001','.002") and when l
opened by bending, revealed only ductile rupture. This type of ductile rupture indicates that these defects were most likely due to deformation of the 0-ring near the seating surfaces.
l Normal sensitivity of the ultrasonic (UT) examination method is sufficient to detect indications having a depth of 20% of the diameter of the bolts. This sensitivity is based on examining bolts which are fully accessible. There was no in place standard for a UT examination performed from the head of the bolt, which is l
l
. ~.. _ _ _. - _
- ~
necessary for the examination of bolts in the installed components.
The licensee had to develop a test program which would detect a small indication without ambiguity. The sensitivity of ultrasonic testing of bolts is dependent, in part, on the size of the bolt and the depth of the threads.
In some cases, defect depths of 10% of the bolt diameter are detectable. Where threads are deeper, 20% is the minimum detectable defect depth. The " worst case" condition is the core barrel assembly bolts, which have a 20% detectability limit and are higMy 1caded. These Incens! X-750 bolts had c preload of 24 to 30 ksi at installation, and were exposed to sulfur in solution at the the same time as the remainder of the RCS. The licensee concluded and the staff agreed that at least one of the ninety-eight bolts inspected would be expected to exhibit detectable damage if the bolts were in an environment conducive to stress corrosion cracking. As a result of these tests, no damage was seen. This test was supplemented by an examination of.the holddown springs, which were also Inconel X-750 and were subject to
(
a similar environment at higher stress levels (100 ksi). Again, no corrosion indications were detected.
~
The licensee used a variety of nondestructive and destructive techniques in the RCS inspection. A variety of different materials were metallo-graphically examined by the licensee. The program, as a whole, was planned to generate a large data base to assess possible areas or types of damage in the remainder of the RCS. No evidence of stress corrosion cracking was found outside of the OTSG. The results of all of the examinations performed indicated that the items examined, with the exception of the PORV, were in an acceptable condition. The pitted portions of the POPN were replaced. A summary of reactor coolant system materials examined is shown in Table 3.2-1.
(
G
@H@
- e
~
.. c-The inspection plan, based on the premise that generic material groups will behave similarly, included items of sufficient diversity to reficct the different materials, stresses, and environments that are present in the RCS. The inspection developed also accounted for critical func-tions of the RCS items. The functions of the pressure boundary, core support, and fuel integrity, because they are safety-related, received the most emphasis in determining the general condition of the system.
~
The staff finds that this selection process took into consideration materials which had shown a tendency for this type of attack based on a literature search performed by B&W and EPRI.(5) Based on the above examinations, the staff has, reasonable assurance that the Nmainder
~
of the RCS has not been subjected to sulfur-induced corrosion and is therefore acceptable for continued power operation.
Conclusion
(
The staff finds that the PORV and safety relief valves, which exhibited pitting corrosion, were replaced and therefore are now acceptable. The remainder of the reactor coolant system which was inspected within the limitations of the inspection methods employed, disclosed no defects attributable to sulfur induced corrosion.
Therefore, the staff finds.that GDC 1, 14,'15 and 31 have been met as they apply to reducing the possibility of a' rapidly propagating failure of the RCS pressure bo'undary, and that reasonable assurance exists that the public health and safety is protected.
k
'20 Table 3.2-1 9
,(
~~ * '
Summary of Reactor Coolant Systec."aterials
~
KCS caterial Condition Representative item
- nconel 600, SB-166, As-o rdered RV CRDM nozzle, incore detector drive lines
-167, -168 As-welded RV CRDM nozzle to SS 304 flange PWHT/as-welded HPI safe-end to SS 304 pipe, spray line nozzle safe-end to SS 316 pipe, decay heat nozzle safe-end weld Welded /PWHT OTSC tubes, HP1 safe-end, spray line nozzle safe end, tubesheet cladding Incor.el71d As-welded, annealed RV O-ring j
Welded, age-Spacer grid, fuel assembly hardened Inconel X750 As-ordered Fuel assembly holddown spring, RNS retainer spring, core support bolts Inconel 660 As-ordered Vent valve retainer spring Stsinless Steel, As-ordere'd RV CRDM flange, decay heat nozzle safe-end, typs 304 (annealed)
CRD.M leadscrew As-welded RV CRDM flange (HA2), CRDM motor tube, LPI '
(
piping veld PWHT/as-welded HAZ locking cup welds on RV internals Welded /PKHT Stainless cladding, plenum cylinder welds Cold-worked Plenum lug bolts, plenum cover-to-cylinder bolts Cold-worked and Control rod assemblies welded Stainless steel, As-ordered Surge line pipe, spray line pipe typa 316 (annealed)
As-welded Surge line pipe, spray line pipe Welded /PWHT None i
Stainless steel, As-ordered, Pressurizer heater t7p2 316L as-velded Ca sting Fuel assembly end fittings Stainless steel, As-ordered, CRDM Retainer assembly Ypc 403 condition A l
As-ordered CRDM leadscrew coupling, makeup pump shaft
(
i (
-.,---s"e
--p_
e-vm.r,4---ygme q-
~,;,
e..m.e
-,:-r-a--
m m4
-s-e
~
=,w---
l o
- 21 Iable 3.2-1,
(.
_S..e:a ry of Reactor Coolant System Materials (Cont.)
RCS material Condi tion Recresentative item Misc. stainless, As-ordered CRDM internals cobalt, and Inconel alloys AMS 5737 C ( A286)
As ordered Vent valve jack screw Stainless steel, As ordered type 410 Decay heat punp shaf t Stellite No. 12 As-welded RC pump rotor assembly Zircaloy-4 As-o rdered Fuel rod Braze material, As-welded 83Ni-7Cr-3Fe, RV internals rod guide bra:ecent 3B-4.5Si.1 max C Stainless steel, As-beased type 304L Spacer grids, rod guide seg=ents Stellite No. 6 As-welded Vent valve bushing, CRDM internals 15-5-PE ( AMS 5658)
As-ordered Vent valve retaining ring Stainless steel.
As ordered Vent valve shaft type 431 9
4 9
S e
IIT-S
~ ~. -.,
e
, 3.3 OTSG Examinations to Determine Extent of Degradation Initial eddy-current examinati~ons of the steam ger.arators with a 0.510-inch diameter standard differential probe indicated that tube wall defects were on the inside surfaces of tubes in both steam generators.
These defects were distributed radially throughout the steam generators and predominantly in the upper portions axially. The vast majority of the defect indications were in the UTS region and particularly confined to the tube roll transition zone.
In order to better quantify and characterize the defects, the licensee established an extensive testing program whereby special eddy-current testing (ECT) techniques were developed so that a more accurate picture of the extent of damage could be developed.
Using machined notches and laboratory grown cracks as standards for
(
calibration, and comparing field data to laboratory ECT and metallurgical examinations of tubes removed from service, the licensee used a self-developed eddy-current system for 100% full length inservice inspection of the tubes in both steam generators.
~
The eddy-current system which evolved from the testing program was a standard differential probe of 0.540-inch diameter, with an effective gain of approximately 60.
Two base frequencies (400 KHz and 200 KHz) l and an "ID" inix to enhance detection of ID defects and minimize the effect of chatter and tube noise were used. The testing program compared absolute systems to the 0.540-inch diameter standard differential high gain probe and found the latter system as sensitive as the absolute systems for detecting circumferential cracks.
l
(
l
-s-,
m,
23 -
(-
The licensee's inspection results, with the 0.540-inch high gain probe eddy-current inspection system, revealed indications in addition to those previously identified by the 0.510-inch standard gain probe. However, no significant. pattern of crack growth was apparent in the six-month interval between the initial 0.510-inch standard gain inspections and the 0.540-inch high gain inspections.
The licensee interpreted the eddy-current test measurements of the through-wall depth of the indications in the UTS transition zone regions as greater than 40% through-wall, and hence were characterized as defective tubes.
Metallurgical examination conducted by the licensee on tubes removed from service confirmed that.the flaws in the transition zone region all exceeded 50% through-wall, with the' majority 100% through-wall.
Based on the eddy-current results indicating the depth of the defects and the extent of the observed degradation, the licensee decided that all
(
unplugged tubes would be kinetically expanded within the tubesheet to establish a new primary system pressure boundary below the defect area.
After kinetic expansion, the eddy current examinations, using an 8 x 1 absolute probe, revealed 9 of 151 tubes examined in steam generator A, and 6 of 284 tubes examined in steam generator B, with indications not previously seen by the 0.540-inch standard differential high gain probe.
Fiberscope examinations of the new eddy current indications revealeo small pits and scratches which were below the sensitivity of the 0.540-inch standard differential high gain probe. These indications, which are not considered to be of safety significance, will be reexamined at the mid-cycle ECT inspection and evaluated to confirm the decision that they are acceptable.
Any future eddy current examinations in the kinetic expansion region will be performed with the 8 x 1 absolute probe.
\\
e e
4 "b*
,4-r y
w
. Of the 31,062 tubes in both steam generators, 29,838 with no known defects within 16 inches below the primary surface of the UTS were repaired by kinetic expansion and returned to service. A total of 347 tubes had been removed from service by plugging prio.r to the start of the kinetic expansion repair program. An additional 811 tubes with greater than 40%
through-wall indications 16 inches below the primary surface of the UTS were removed from service by plugging after kinetic expansion. Approximately 475 of the plugged tubes were also stabilized with internal rods to prevent damage to adjacent tubes in the event the degradation continues and the tube severes.
The purpose of the stabilization of plugged tubes is to reduce the risk due to propagation of tube defects located in regions with high potential for flow induced vibration, resulting in circumferential tube severance, thus causing damage to adjacent tubes or creating loose parts. The affected tubes, which are in the area of high steam cross flow (16th
(
span, were stabilized to the 14th tube support plate.
Fracture mechanics analysis of circumferential tube defects conducted by the licensee shows that circumferential defects less than 40% through-wall are acceptable because they would not propagate during normal operation or accident conditions.
The staff consultant's analysis (Attachment 1) confirmed the licensee's conclusion and, therefore, the staff agrees with the licensee that circumferential defects less than 40% through-wall are acceptable, and they would not propagate during normal or accident conditions. Approximately 66 tubes with ECT indications between 20-40% through-wall, as verified by an 8 X 1 absolute probe and located 16 inches below the primary surface of the UTS, are considered degraded tubes. Approximately seventeen of the degraded tubes have ID indications which are attributable to the sulfur-induced corrosion problem. These 66 tubes will be left in service and monitored in the extended Inservice Inspection (ISI) program.
The extended ISI program will include 100% reinspection of tubes with 40% and less through-I wall indications as a separate subset during subsequent examinations.
If eddy-current examinations show no substantial growth in the cracks, they will be left in service.
Tubes showing signs of crack propagation based e
e
' on previously established acceptance criteria will be taken out of service.
Lack of defect propagation will give additional assurance that the cracking mechanism has been arrested for the long term.
A summary of the licensee's post repair eddy current inspection plan is shown in Table 3.2-1.-
Table 3.3-1 Summary of Post Repair Eddy Current Inspection TOTAL NUMBER OF TUBES DESCRIPTION SCOPE PROBE BASELINE AFTER 90 DAYS
- 15 Tubes with Previous 8x1 15 15 Kinetic Expansion Indications (6 Qual. Length)
- 3% Baseline /0TSG 8x1 930 930
(
- 10 P6ripheral/0TSG 0.540 SD 60 60 Wear
- 10 with Defect in (Inservice Tubes 15th, 10th or 1st Adjacent to Unsta-Span /0TSG 0.540 SD 120 120 bilized Plugged
.5 with.540 SD>3V 0.540 30 60 60 Tubes)
- Defect Location 8x1 66 66 Inservice
- Full Length 0.540 SD (F40% TW)
High Plugging
- 50 Full Length /0TSG 0.540 SD 100 100 Density Completed Standard
- 3% Full Length /0TSG 0.540 SD in 1982 930
(
Inspection l
TOTAL 1350 2300
_ =
4
(-
CONCLUSIONS Based on the above evaluation, the staff finds the ISI program and the administrative controls instituted by the licensee acceptable. The staff concludes that the eddy-current techniques developed and qualified for inspection of the OTSG tubing demonstrated the ability'to reliably detect and size, with a high degree of sensitivity, the defects that were present in the tubing. The 100% tube inspection, preventive tube plugging and staking of critical defective tubes give reasonable assurance that defective i
tubes have been identified for repair or removal from service, so that the return of the plant to service will not adversely affect safety.
However, the staff will require that the following license conditions be imposed to mitigate the potential for primary to secondary leakage:
- 1) the licensee will conduct eddy-current examinations, consistent with the inspection plan defined in Table 3.3-1, either 90 calendar days after reaching full power, or 120 calendar days after e.tceeding 50%
~
power operation whichever comes first and 2) an administrative shutdown limit will be imposed if primary-to-secondary leakage increases signi-ficantly above baseline. The licensee will establish a new baseline leakage limit during the precritical hot functional test.
If primary-to-secondary leakage exceeds the new baseline by a significant amount the plant will be shutdown and leak tested.
Current information indicates that increases in leakage of approximately 0.1 gpm are measurable.
The licensee should establish a limit for detecting increases in primary-to-secondary leakage upon restart. This limit should be either 0.1 gpm or an established measurable limit which does not significantly exceed 0.1 gpm.
If leakage is due to defacts in the tube free span, the tube will be removed from l
service.
Subsequent to shutdown for primary-to-secondary leakage, the l
amount of leakage from within the tube /tubesheet crevices will be considered as background and will be used to establish a new baseline at restart, provided that the technical specification limit of 1.0 gpm is not exceeded.
, l' 3.4 Once Through Steam Generator Repair 3.4.1 Desian Requirements of the Reoaired Joint To establish acceptability of the repaired OTSG for return to service, the licens'ee instituted a test program to demonstrate that the repaired joint would meet the original design basis. The following is a summary of the design requirements established by the' licensae which were used to qualify the repaired joint.
a.
Axial Load The joint should be able to sustain a tensile load of 3140 pounds with no slippage between the expanded area and the tubesheet at an axial strain corresponding to this load.'
b.
Thermal Air Pressure Cycle Loading The goal for the repaired joint is to maintain its ioad carrying and leak-tight capabilities for the remaining plant design life of 35 years, assuming design basis thermal cycling and transients. To demonstrate design life capabilities, the licensee instituted a multi phase test program. The first phase of the test program includes qualification testing to show capability to operate for the first 5 years, to justify restart. The second testing phase includes extended qualifications which will be completed subsequent to restart.
c.
Tube Preload The tensile preload on the tube should not change by more than a i
predetermined load as a result of expansion. This objective assures that the vibrational characteristics of the tube will remain unchanged.
d.
Residual Stresses A transition length between 1/8 inch and 1/4 inch is established as a goal.
A more abrupt transition would result in a higher residual stress.
The residual stresses resulting from the kinetic expansion process should
(
be limited to 45% of the 0.2% offset yield strength at room temperature.
. e.
Leak-Tight Integrity The kinetically expanded joints are designed to be leak-tight.- Leakage in the joints has been qualified by drip and bubble tests. Those joints determined to have leakage in excess of acceptable limits have been removed by plugging.
A design goal of one pound per hour total leakage from both generators has been set for the qualification program. The bubble test can detect leak rates as low as 0.83 pound per day or 0.1 gallon per day.
During normal plant operation, total primary-to-secondary steam generator leakage limits will continue to be set by the plant Technical Specifications limit of one gallon per minute. However, an administrative shutdown limit of approxi-
~
mately 0.1 gal 1on per minute above the baseline primary to secondary leakage limit will be imposed.
3.4.2 Mechanical Tests to Qualify the Repair Process
(
The licensee performed a series of mechanical tests to qualify the kinetic expansion process and qualify the joint to meet the design goals of load carrying capability, leak-tightness, residual stress and tube preload variations. The primary vehicle for these tests was a test block which was fabricated utilizing archive tubesheet sections and either archive tubes or actual tubes removed from TMI-1 OTSG. The test blocks were then assembled and the tubes kinetical y expanded using the same process as in the actual OTSG. A more detailed descrip-tion of the test blocks is in Attachment 1.
Similar tests were performed by the staff consultant (Attachment 1) to provide an independent confirmation of the test results.
In addition, the licensee conducted tests on a full size steam generator at B&W's Mt. Vernon facility.
A detailed description of the testing program by the licensee is provided in Reference 6.
a.
Axial Load Test The repaired tube is exp,ected to sustain a maximum design basis axial load of 3140 pounds with no slippage. This criterion envelopes
29 -
(*
the main steam line break accident loading that the joint is designed for. Seven test blocks were subjected to pullout tests at room temperature. One test block was tested at 330*F.
Pullout loads for high and low strength tubing with the ti-inch expansion on both corroded and uncorroded blocks showed pullout loads significantly in excess of 3140 pounds for all. qualification conditions. The effects of thermal and pressure cycling on pullout load were minimal over the initial qualffication period.. Thermal cye' ling tends to decrease pullcut load, however thermally cycled blocks pulled at 70*F were determined to have pullout loads in excess of'3140 pounds.' Most of the leak and slip-load qualification tests were performed at room temperature, with one block loadeo at'about 330*F,and one leak tested at 400*F. The licensee had determined, prior to qualification testing, that the effect of the different coefficients of thermal expansion of Inconel tubing and carbon, steel tubesheet made room temperature testing more conservative i
than tiesting at elevated temperatures. However, it was found during testing that slip load was more dependent on tube yield strength than interference pressure.
Hence slip load is reduced at elevated tempera-tures due to the reduction in yield strength of the tube material.
The ten-tube block'which was slip-load tested at 330*F was determined to have a 286 pound reduction in mean pullou+. load, but was still in excess of 3140 pounds. This reduction in mean pullout load is attributed to the reduction of tubing yield strength at elevated temperatures. The mean slip load for low yield-strength tubes in test blocks receiving identical conditioning dropped by 6.2% when pullout temperatures were increased from 70* to 330*F. The true criterion for. tube slippage is the strain in the tube at slippage. Tube strains measured at the initiation of slippage for the block tested at 330*F were equal to or greater than those at room temperature. The staff therefore finds the slip load qualification tests at room temperature to be adequate.
(
i x
w
-~
r w-
- r
-,e--,m y
,---,--n--~
,-=+r,,
~
- A pull test performed on a full scale OTSG with expanded tubes at Mt. Vernon also showed a load carrying capability in excess of 3140 pounds. The above test data indicate that adequate margin in the pull-out strength is available at a normal operating temperature of 600'F.
Pullout loads obtainea by the staff consultant (Attachment 1) through independent testing on ten test specimens all exceeded 3140 pounds and provide additional confirmation of the pullout strength.
b.
Leak Rate Tests The leak rate tests conducted by the licen m confirmed that the expanded joint will meet the design objective of limiting total primary-to-secondary leakage from the TMI-1 OTSG's to 1 pound per hour (lb/hr) per plant or less
-5 (or 3.2 x 10 lb/hr per tube) under plant operating conditions. The tests were conducted by the licensee in accordance with approved procedures.(6)
Seven test blocks consisting of 10 tubes each were subjected to a series of leak tests using demineralized water at 70*F.
The effects of thermal cycling and second expansion shots were examined.
Leak rates were determined at normal operating pressures as well as simulated accident conditions. The leak rate results indicated that the water leakage rate for the normally expanded
-5 tubes is less than 12.15 x 10 lb/hr per tube. While this rate exceeds the
-5 design objective of 3.2 x 10 lb/hr/ tube, it is considered a very low leak rate.
If every tube in both steam generatort leaked at this maximum rate; j
the cumulative leak rate would still be less than one hundredth of the plant i
Technical Specifications limit of 1.0 gallon per minute. One test block was cycled between 10 F and 400*F and showed a reduction in leakage of less
~
than 1 x 10 lb/hr/ tube, leading to the conclusion that the leak rate for a tube at operating temperatures would be the same or slightly less than it would be at room temperature. Therefore, based on these tests, leakage l
rates can be anticipated to be significantly less than Technical Specifi cations limits, and are therefore, acceptable.
t
(
1 l
l
7_;.
~
31 -
t c.
Residu'al Stresses in the Transition Length of the Expansion The qualification objective is to lower the residual stresses at the transition region between the expanded and unexpanded portions of the tubes on the downstream side of the expansion in order to reduce the possibility of stress corrosion cracking'. The qualification goal of limiting the transition length to between 1/8 and 1/4 inch minimizes
~
the residual stresses. The actual residual stresses were measured in the special test blocks using X-ray diffraction and strain gage techniques to determine the post-kinetic expansion tube stresses in the transition area at the bottom of the expansion, and at a second point near the middle of the expansion. - Beth hard rolled and kineti-cally expanded tubes were evaluated using'hi'gh and low strength materials. These tast results, obtained by the licensee's consultant, indicated that the residual stresses do not exceed the design goal of
~
0.45 yield strength.(7)
Sample Inconel 600 tubes were expanded by rolling and kinetic processes in order to compare the resulting hardness'. The hardening effect in the mechanically expanded tu'Je was determined to be more pronounced and less uniform than the kinetically expanded tube. A such, the
, latter, due.to lower residual stresses, may be expected to be less susceptible.to stress corrosion cracking; The licensee _and its consultants perf?rmed accelerated stress corrosion cracking tests on expanded tube /tubesheet mockups.
The mockups were tested in a standard 10% sodium hydroxide (NaOH) at constant potential ancien. mined for stress corrosion cracking. Examination of the specimens witt a fiberopticscope revealed no cracks after 5 days in r
an autoclave. Thisexposure"correspondstbapproximately8.5 years
+
in PWR secondary side water at 650*F.
If significant residual stresses had been introduced by the expansion repair process, t1is
(
T 3
+
e ee me 4
umar
(-- test would have initiated SCC. The test results indicate that there is no detectable increaseu SCC susceptibility of expanded tubing over unexpanded tubing.
s.
d.
Effects of the Kinetic Expansion on the Tube Pretension The coefficient of expansion for the Inconel 600 tubes is greater than for the carbon steel shell. Therefore, as the OTSG heats up, the tubes tend to elongate proportionally in relation to the shell. This increased proportional length of the Inconel 600 tubes has to be accounted for in the OTSG design to prevent excessive elongation and subsequent bucklir.g on heatup.
If it can be assured that.the tubes are under insignificant compression or even in tension under cold conditions, then buckling will not occur at design operating conditions. The problem of buckling is resolved by installing the tubes under tension; i.e., they are stretched and seal welded in the stretched or pretensioned position. As a result,
(
subsequent elongation of the tube is not sufficient to cause buckling.
The change in the pretension in the tubes due to the kinetic expansion process is directly attributable to a change in the tube length caused by the expansion process.
The licensee conducted induced strain tests on test blocks to determine l
l tube lengths before and after the expansion.
Results show that the l
l process has a minimal effect on the overall longitudinal tube strain and as-fabricated preload induced strain. Measurements taken before and after expansion indicated changes in maximum longitudinal strain l
values of less than 0.04%. This relates to a reduction in the tube pretension of about 16 pounds, which is less than 30% of the as-fabricated pretension. Therefore, reasonable assurance is provided that the repaired tubes are not in compression while cold and will not buckle during hot operations.
(
Nmme MAIO
~
. (-
In addition to the mechanical tests discussed above to qualify the
' expanded joint, the effects of the explosive expansion on the tube-sheet ligaments and welds were determined. The dimensions of adjacent holes in the tubesheet were measured before and after the expansion and compared.
Results show that there is practically no effect on the diameter of adjacent holes due to tube expansion.
Full scale testing conducted by the licensee and its consultant in a steam generator at B&W's Mt. Vernon facility using strain gages and profilometry showed no degradation of the tubesheet ligaments.N During preliminary and qualification testing, some candles have fractured, i.e. blow throughs have ociurred, creating a concern that parts of the polypropylene cartridges could be left in OTSG tubes. To ensure that fragments of cartridges do not remain trapped or wedged in tubes, free flow air tests were conducted for each tube after the expansion. The
(
final cleaning involved blowing felt plugs through aach individual tube.
The head and the tubesheet were manually wiped down and then the generator was flushed to remove any remnants of the repair process. Non-destructive
(
and visual examinations revealed no tube deformation as a consequence of the frtctured candles.
To determine the effect of the kinetic expansion process on the welded connections in the vicinity of the tubesheet, the licensee obtained strain and acceleration data during a kinetic expansion cn a full scale steam generator at Mt. Vernon. One measurement was made at the junction between the inlet header and the tubesheet and the other at the welded location underneath the tubasheet. The strain gage measurements were taken at the two ends of a diametral row of 132 tubes. On the basis of these data, the peak stresses and stress intensities were calculated for the fatigue evaluation. The cumulative usage factor was determined to be 0.12 at these locations.
Based on these data, the licensee concluded that the welded connections in the tubesheet/shell 1
1
34 -
section will not be affected by the expansion process.
Independent analysis by the staff consultant (Attachment 1) confirmed the licensee's calculation. Therefore, we agree with the licensee's conclusion.
A number of tube stub sections were cracked at the tubesheet primary face during the, expansion process. This was most pra5 ably caused by the normal pressure pulse generated during expansion. The portions of the tubes above the tubesheet, being unrestrained by the tubesheet were subjected to higher deformations during expansion.
Pre-existing defects in this region resulted in cracking at the welds at the top of the tubesheet.
Because these cracked stub tube pieces are not part of the primary pressure boundary, the structural integrity will not be t
affected. However, there was a concern that loose pieces cauld be introduced to the RCS.
For this reason, the licensee milled all stub i
tubes down to the tubesheet primary face to remove all stub tube sections which could potentially become loose.
The effect of high cycle flow-induced vibration loading on the repaired steam generator tubes has been evaluated by the licensee. The steady axial and high cycle bending loads define the tube loading. A linear elastic fracture mechanics (LEFM) Code "BIGIF" developed by the Electric Power Research Institute (0) has been used to determine when a crack of a given initial. size can be expected to propagate through wall.
A key parameter in this analysis is the stress intensity factor which quantifies the interaction of crack size, shape, boundary geometry and stress field. The stress intensity factor calculation includes the loading due to internal pressure, axial and bending loads which would tend to open up a crack during flow-induced vibration (FIV).
Ouring steady state operation the steam generator shell to tube temperature difference could cause an axial tension of up to 500 pounds on a single tube.
(
1
.-+--%..-,..m4 p..
g m
_...-,,.a
--m---*c-
'y
-rn e,N'r
"*P "P
35 -
In the analysis, the load cycle imposed on the tubes included mechanical and thermal factors.
Low cycle, long duration loads were combined with high cycle flow induced vibration loading. The vibrational load amplitude was selected to be the maximum tube displacement seen under steady-state loading. The maximum tension excursion, represented by the 100*F/ hour cooldown, which results in an axial load of 1107 pounds, was combined with high cycle loading.
A modified Paris equation (8) was incorporated in the Code "BIGIF" with the feature that if the stress intensity factor range did not exceed thres-hold, no growth would occur.
In the analysis, 1.0 ksi in. b' was used for the threshold stress intensity factor (WKth), the value below which a crack will not propagate. The result of this analysis indicates that ECT is capable of detecting crack sizes which are smaller than those that can propagate by mechanical cyclic stress. Therefore, cracks which are large enough to propagate to failure can be detected and removed from
(
service.
The licensee performed additional calculations to determine the maximum crack size that would remain stable under loads experienced during a main steam line break (MSLB) accident.
Results of the calculations indicate that cracks which would remain stable during a MSLB accident can be detected by ECT.
In order to bound different crack geometries and plant conditions, leakage rates for various tube axial loadings and crack arc lengths were determined.
Results shcw tuat leakage rate increases with tube axial loads and is detectable under conditions which might cause tube failures.
The minimum detectable leakage ranging from 0.1-2.5 gpm for various loadings is well below the threshold crack size that would fail during MSLB conditions. Based on a review of the licensee's analysis and independent staff calculations, the staff finds that:
k 9
Ob8
,,3<.w-
,,-a-r n
a
-, e g m
,r,-,,,,,nm.,-,,--m.,,,-
,-,---w.-+,
<-, ~ -
--f-
,--.,.e m.
. 1.
Cracks which are large enough 1.e. critical size, to propagate due to flow-induced vibration are readily detectable by ECT; 2.
Cracks which are below the threshold of ECT detectability will not propagate under combined cyclic, flow-induced and thermal loadings;
~
3.
The maximum crack size which will remain :: table during a MSLB has been determined; 4.
Through-wall defects which may propagate during operation can be detected well below the threshold size that could fail during a MSLB. Therefore, reasonable assurance exists that the potential for rapidly propagating failure of steam generator tubes due to flow-induced vibration is limited.
e.
Effect of Expansion on Existing Plugs The TMI-1 OTSG tubes have been taken out of service using four
(
different procedures:
1.
Explosively welded plugs - Plugs inserted into the tube and explosively welded in position within the tubesheet.
2.
Welded plugs - Plugs welded to the tube ends or tubesheet at the top of the upper tubesheet.
3.
Hydraulically expanded tubes sealed with a welded plug - Tubes that have been immobilized by expansion after a short section of the tube within the tubesheet was removed. The tubes were then taken out of service by installing welded plugs in the tubesheet openings.
4.
Mechanically rolled plugs.
The licensee has evaluated the effect of the kinetic expansion on the mechanical integrity and leak tightness of these plugs. Tests (Topical Report 008, Rev. 2) of the kinetic expansion process in steam gener-ator model test blocks with
(
V
t.
37 -
conditions simulating those in the TMI-1 steam generator show that the kinetic expansion does not produce any permanent tubesheet ligament deformation. Test data (Topical Report 008, Rev. 2) obtained in the laboratory ana in actual OTSG's confirm that there is no damage to the welded and rolled plugs of the type discussed earlier due to explosive expansions adjacent to the plugged tube.
Leak rate and axial load t'sts e
performed,after the expansion verified that the plugs continued to meet acceptance criteria to which they were originally qualified for use, f, ir,.
Leak Testing of Repaired Steam Generators A preoperational post-kinetic expansion pressure test of each generator was made to verify the integrity of the primary to secondary pressure boundary thus providing added assurance that no degraded tubes or plugs will go into service. By letter dated May _, 1983, the licensee provided the results of the OTSG leakage tests.
Leak testing of both repaired OTSG, utilizing 150 pounds of nitrogen pressure, revealed a total of approximately twenty-four tubes with minor leakage. These tubes were repaired to the same criteria listed in Section 3.4.3 below. All of the leaking tubes were located within the UTS, with the exception of two tubes which were leaking at two inches below the UTS.
In addition, a total of approximately 23 plugs were. leaking.
The plugs were repaired by either removal and replacement or grinding and re-welding, as appropriate.
Based on staff review of the licensee's qualification tests and analysis, and indep;ndent verification tests and analyses by the staff consultant, (Attachment 1) the staff finds that:
1.
The kinetic expansion procedure is an effective and reliable method to repair the cracked tubes;
(
O w
e H
.I
, (-
2.
The kinetic joint meets the qualification requirements in terms of load carrying capability, tube preload and residual stresses.
Leakage of the laboratory test blocks, while somewhat exceeding the qualification goal, is well within the plant Technical Specifications limit of 1 gpm and is acceptable. Nitrogen pressure testing of the repaired OTSG has shown fewer leaking tubes than was anticipated based on the laboratory test blocks, and is, therefore, acceptable.
The structural and leak-tight integrity of the expanded joint has been demonstrated for at least a five year period and an extended life qualification program is in progress to qualify the expanded joint for time periods in excess of 5 years; l
3.
The differences betwee'n the test assemblies and actual steam
~
generators are not considered significant enough to affect the validity of the test data.
Sufficient margins of safety exist to account for some uncertainties that may exist in the simili-(
tude of the test environment; 4.
The hot functional tests planned as part of the start-up program will provide added assurance that the repair process meets the qualification requirements; 5.
Calculations utilizing linear elastic fracture mechanics, including loads due to high cycle flow-induced vibrations, as well as mechanical and thermal effects were performed to determine the maximum crack size that would remain stable during a MSLB accident.
Such cracks can be detected by ECT and would result in a primary-to-secondary leak rate which is detectable. These cracks are required to be repaired by the license condition; and 6.
The kinetic expansion process does not produce any permanent tube-sheet ligament deformation or degrade the effectiveness of the existing plugs.
smo
. b The staff will condition the license to require submittal of the extended life cyclic program qualification test results by the first refueling.
3.4.3 Evaluation of Plugging Repair a.
Discussion Those tubes which have defects below a distance of 16 in. from the primary surface of the UTS and cannot be returned to service by the kinetic expansion repair procedure have been removed from service by plugging. This included approximately 625 tubes in A and 185 tubes in B OTSG. There were a total of 259 tubes in A and 88 tubes in B OTSG that have previously been plugged with e'ither Westinghouse rolled plugs or B&W welded and explosive plugs.
Portions of 19 tubes in A and 10 tubes in B OTSGwereremovedformetallurgicalexaAlination. The remaining tubesheet hole was plugged with a B&W welded tapered cap on the top and an explosive
(,
plug at the lower tubesheet (LTS). Defective tubes in the area of high steam cross flow (16th span) have been stabilized to the 14th tube support plate. Approximately 475 tubes are stabilized. The purpose of tube stabilization is to reduce the risk of a plugged tube failing due to flow-induced vibration.
Failure of a plugged tube can result in circumferential tube severance and consequent damage to adjacent tubes or generation of loose parts. The lower tube end has been plugged with an explosive plug.
The following paragraphs provide an evaluation of the tube plugging and l
l stabilization techniques and criteria.
b.
Metheds for Tube Plugging A variety of plugging techniques are being used at TMI-1.
B&W Welded Tapered Plugs are used to plug the bare UTS hole where the tube has been removed. B&W's welded cap is used to seal the upper tube end for those tubes which will be plugged and stabilized.
Prior to installing the weld cap, the existing tube end will be machined off, leaving a portion of the I.
tube end and the existing weld protruding above the tubesheet surface.
The weld of the nail cap will fuse with the existing seal weld, providing the desired pressure boundary.
B&W Explosive Plug is used to plug LTS tube ends where B&W plugs are used in,UTS. Both B&W. plug types have been qualified for OTSG tube
~
plugging and used in other operating B&W units.
B&W standard design stabilizer rods will be threaded onto the welded cap to form a stabilizer assembly of the desired length. The stabilizer is a multi piece assembly of solid rod made of Inconel 58-166. Joint tightness is maintained by crimping the pieces together beyond t.he threaded sections. The segment ler.gth is dependent upon the tube bundle location.
the B&W welded tapered plugs, welded cap explosiva plugs and stabilizer rods have been previously q'ualified.(10-15) Westinghouse plugs were designed for a primary pressure of 2500 psi and at 650 F, and a secondary pressure of'1050 psi at 600*F. Cracks in the roll transition or the area of the seal weld do not exclude the use of the Westinghouse
(
rolled plug because of its extended sealing length.
The Westinghouse rolled plug is machined from bar stock that has received a thermal heat treatment which has been demonstrated by laboratory testing to have improved resistance to intergranular attack in caustic and polyth. ionic acid environments, compared to treatments at different temperatures and times (Topical Report 008, Rev. 2).
The Westingbouse Roll Plug Qualification Program has been completed for a 5 year life, and results are documented in Westinghouse Report WCAP-10084.
c.
Pluoging and Stabilization Criteria The OTSG tubes were divided into several areas for the purposes of plugging and stabilizing.
For reference purposes, the secondary face
(
of the upper tubesheet is defined as US zero(0), and elevations going
~
(-
into the upper tubesheet are designated as positive i.e., US + 5 is five' inches above the tubesheet secondary face or nineteen inches below the i
top of tubesheet primary face.
Similarly, the secondary face of the lower tubesheet is defined as LS zero (0).
Tubes With Defects Betwecn US + 4 and US + 8:
Tubes having defects between US + 4 and US + 8 cannot be effectively repaired by the 22-inch kinetic expansion. A minimum kinetically expanded length of 6-inches is required to establish a leak-tight, load-carrying joint to assure that the OTSG integrity is retained under design basis conditions. Therefore, those tubes with defects between US + 4 and +8 will be remov'ed from service. These have already been kinetically expanded and plugged.
Even if the existing crack would propagate in the future to a 360' circumferential crack and eventually the tube should sever at US + 4 elevation, there is still a 3-inch length of joint below the severance, which will provide enough engagement to maintain the preload in the tube..A 3-inch kinetic expansion joint below the defect will assure that the tube is still in tension under the most severe transient during normal operation. The natural frequency of a severed tube with a 3-inch expansion below the defect is about the same as an intact tube because tensile preload is maintained by the expansion.
Therefore, its potential for flow induced vibrations (FIV) failure is about i
l l
the same as for an intact tube. On this basis, it was concluded that there is no need to stabilize these tubes.
Both welded cap and Westinghouse rolled plugs can be used to plug these defective tubes in addition to a 22-inch expansion.
Tubes With Defects Between US + 0 and US + 4:
i As discussed above, a minimum length of 3-inch expanded joint is considered necessary to prevent disengagement of the tube due to flow-induced vibrations.
Tubes with defect at or below US + 4 would not have enough' joint length even with a 22-inch expansion, which is the l
l
.w.--
-w-
,--31, p.---
+ - - - - - - - - - - -. -., -. -. -,
7
, w
m
m.m
. j, maximum depth being used to assure that the tube will not slip under the most severe transient during normal operation, i.e., 100*F/hr. cooldown.
As the tube load decreases to zero, the natural frequency of the tube may reduce to a critical value that results in dynamic instability.
For this reason, tubes in this category have been plugged and stabilized.
Tubes With Defects Between 15th Tube Support Plate (TSP) and US+0:
Tubes having defects in the 16th span have been plugged and stabilized.
,The progressive degradation of the defect has the potential to cause severance of the tube in the high steam cross flow region. The severed tube could damage the adjacent tubes if it is not stabilized.
The stabilizer will function as a damping and capturing device which will inhibit tuba fatigue failure due to flow-induced vibration in the vicinity of high steam cross-flow.
k
~
Tubes with Defects from 15th Support Plate to LS-4 in Lane Wedge Area B&W plants have a history of corrosion and vibration problems in the areas of the untubed inspection lane. As a precautionary measure, an area of potential problems has been defined one row on either side of the lane, widening to a wedge shape as the lane nears the periphery.
For tubes in this area, any I.D. eddy current indication, regardless of type, through-wall measurement, or circumferential extent, will be treated as a defect.
These tubes will receive a 17-inch expansion, then be plugged and stabilized through the 14th support plate unless other criteria require stabilization through the defect in the lower spans.
Tubes with Defects Between the 15th Support Plate and LS-4 After a 17-inch expansion, tubes with defects greater than or equal to 40%
through wall and 8 x 1 greater than 2 coils were stabilized through the span with the lowest defect for that tube. Tubes with greater than 40%
through wall defects and 8 x 1 indications on 2 coils or fewer were l
expanded to 17 inches or 22 inches as appropriate and plugged with the i
Westinghouse rolled plug.
l t
~~
This criterion.provides a means for determining the tube arc length extent of a defect in order to decide if the tube should be stabilized.
The tube would be stabilized at least within the tube span containing the ECT indication if there is any substantial size or arc length it.volved in the ECT indication.
If an ECT indication is seen on less than three coils on the 8 x 1 ECT probe it means that the arc length of the degraded area of the tube is a maximum of about 0.41 inch long at
~
the inside diameter of the tube. Because of the " thumbnail" shape of the insiae diameter cracks found at TMI-1, this means that the average are length of 19e largest two-coil ECT crack would be about 0.25 inch.
This size crack is acceptable without stabilizing and is not expected to propagate to failure by mechanical means during operation.(16)
Tubes with Defects in the Lower Tubesheet Below LS-4 Tubes with pluggable defects in the lower 20-inch region of the LTS were removed from service using a Westinghouse rolled plug or an explosive plug.
A welded plug was used in lieu of mechanical rolled plugs if the defect was in the rolled area.
Several post-repair tests will be performed to verify plug integrity.
These will include a bubble test at 150 psig, a drip test at 150 psig and at: operational leak test at 1400 psi. Details of the test are provided in Appendix A of Reference 1.
These test procedures have been used successfully in the past to identify OTSG tube leakage at rates which are orders of magnitude below plant Technical Speciff-cations and are, therefore, acceptable. Tubes which have defects in the 16th span will be stabilized, thus preventing tube severance due to flow induced vibration.
Conclusion Based on the above evaluation of the plugging and stabilization criteria, the staff concludes that the proposed OTSG Tube Plugging Plan will restore the pressure boundary integrity of the steam generators by removing defec-
(I 44 -
tive tubes from service and stabilizing tubes which are subject to sever-ance from excessive flow-induced vibration and therefore, is acceptable.
a d
(
L 9
l
_.m..._.
[.' 3.5 To be provided by RSB O
(
l l
l s
i i
.' 3.6 Cleanup of the Contaminant All RCS tanks, components and piping, which had contained sulfur impurities, have been flushed to remove soluble sulfur contaminants. Based on Topical Report 008, Rev. 2, sulfate concentrations remaining in the coolant have been reduced to less than 0.1 ppe.
In this concentrativii range, sulfate ion does not have a significant corrosive effect.
If the sulfur is present as thiosulfate ion, S 0 =, testing conducted by the staff con-23 sultant (Attachment 2) has shown that the threshold concentration for the initiation of SCC in aerated borated water is about 75 ppb for sensitized 304 stainless steel and about 1 ppm for sensitized Inconel 600.
The s.taff consultant (Attachment 2) has also demonstrated that the initiation and propagation of SCC in these metals is suppressed by the addition of LiOH. A Li/S ratio of 10 in the solution is sufficient to achieve this suppression. The staff consultant further states (Ath;hment 2) that thiosulfate ion affects SCC by electromigration into the propagating crack; partial neutralization of boric acid by LiOH provides competing borate anions which exclude thiosulfate anions from the crack. The licensee proposes to maintain the lithium concentration in the RCS at i to,2 ppm which is a factor of.10 to 20 in excess of the maximum allowable sulfur concentration of 0.1 ppm. The staff finds that this measure is consistent with the staff consultant's recommendation (Attach-ments 2-4) and, therefore, is acceptable.
After the removal of dissolved sulfur from the RCS, a concern remains that sulfur trapped in the oxide corrosion film on reactor surfaces may be converted by some sequence of operating conditions in the future to more corrosive and active species. The staff consultant (Attachment
- 4) indicates that metallic sulfides (MS) can react with oxygen to form thiosulfate:
2 2+
2 MO + S 0
+ 2H+
2 MS + H O + 20 23
==
i
I
( Only. approximate surface sulfur concentrations on steam generator tube and other RCS 'urfaces are availab1'e for estimating the potential of s
regenerating thiosulfate from sulfide.
Estimates of surface sulfur concentrations, based mainly on swipe tests, range over more than an 2
order of magnitude, with the larger values as high as 10 pg sulfur /cm.
If all of this sulfur is oissolved in the coolant volume, the sulfur concentration would amount to a few ppa.
In a state of intermediate valence, this concentration of sulfur would have the potential to reinitiate the corrosion mechanism.
The licensee has carried out an extensive series of stress corrosion tests on sections of sulfur-contaminated steam generator tubes from TMI-1 under conditions" simulating those that resulted in the original failur_es of these tubes.
Even in the absence of added LiOH, no initiation of SCC and no propagation of existing cracks were observed.
(
These negative results, however, do not provide adequate assurance that some untested combination of exposure conditions would not liberate aggressive sulfur species.
To reduce the likelihood of corrosion problems from the sulfur remaining on the RCS pressure boundary component and piping surfaces,
.the licensee proposes to desulfurize these surfaces by oxidation with a dilute solution of hydrogen peroxide (H 0 ).
The treatment conditions 22 are summarized in the following table.
Boron (boric acid) 2300 ppm pH (ambient temperature) 8.0 to 8.2 H0 concentration 15 to 20 ppm 22 Temperature 130*F Cover Gas N2 Lithium ion concentration 2 to 2.2 ppm Duration of Treatment 2 to 3 weeks
(
\\
l
~
The pH of the system will be maintained within the desired range by the addition of ammonium hydroxide. The' concentration of H 02 2 "III D' kept constant by the injection of concentrated H 0 into the RCS using 22 positive displacement pumps. The extent of clean-up w111 be assessed on a continuous basis by analyzing for sulfate in the reactor coolant at least daily.
The licensee's tests on contaminated tubes from the THI-1 steam generators have demonstrated that the peroxide treatment will remove 50 to 80% of the sulfur in 2 to 3 weeks. Tests have shown that the desulfurization process will be slowed in the kinetically expanded portion of the steam generator tubes because the expansion process leaves a thin polypropylene film on the tube surface.
Sulfur removal on the remainder of the RCS can be anticipated to be more effective than for the OTSG tubes because the polypropylene film does not exist outside of the OTSG.
A beneficial aspect of the slowing of the reaction of sulfur with the alkaline peroxide before complete removal is the indication that the remaining sulfur is tightly bound in the oxide corrosion film and will be released only slowly during reactor operation.
The slow release, monitored by the daily sulfate analysis, would provide ample time i
for actfon to prevent the buildup of corrosive sulfur concentrations.
Examination of the TMI-1 steam generator tube samples by the licensee after the desulfurization tests showed no signs of corrosive attack by the alkaline peroxide reagent. The staff finds that this is co'nsistent with the absence of any observations of corrosive effects from the peroxide which is normally formed in PWR plants during shut-down and from the additional peroxide added during shutdowns at some PWR plants to dissolve deposited activation products.
(
ee m -.
~
Sulfur contamination was also found on 304 stainless steel surfaces which had been in contact with thiosulfate-containing coolant during the cracking incident. A large number of RCS surfaces and components were examined, including regions known to be sensitized by welding.
Deposits of sulfur or sulfur-containing materials were observed on a few surfaces, but there was no sign cf SCC on RCS components in contact with the coolant. The absence of SCC in these regions is attributed to the lack of dissolved oxygen (Attachment 2).
Two cases of sulfur-assisted SCC were ob, served on RCS components exposed to the gas phase above the coolant solution: the PORV above the pressurizer and the heat affected zone (HAZ) of a weld in the waste gas' system piping. A staff consultant suggests (Attachment 4) that a different corrosion mechanism involving H S r gaseous polysulfides 2
may be operative in these cases.
Desulfurization would also minimize the source of the corrosive agent for this type of corrosion.
The nature of the occlusion of sulfur is similar in the oxide corrosion films on 304 stainless steel and Inconel 600. Therefore, similar desulfurization efficiencies by alkaline peroxide are expected. With regard to the corrosion of 304 stainless steel and the other alloys in the RCS, (17) the alkaline peroxide reagent does slot have significant adverse effects.
In the absence of hydrogen overpressure during normal reactor shutdowns, low concentrations of H 0 are radiolytically generated 22 by decaying activities in PWR coolant. There is no evidence that this H02 2.has caused corrosion at any operating PWR's.
No indications of corrosion of RCS components were reported following H 0 additions up to 22 10 ppe in some PWR plants during shutdowns. The staff, therefore concurs with the licensee that there is reasonable assurance that stainless steel components will not be significantly corroded by the desulfurization treatment.
(
,.---y-.,-
-r-
- - - - +
,w
---v
-v---+--- - + -- - --- -
_.u_
o
. A-series of " lead tests" are currently underway in which synthetically sulfurized Inconel 600 tubes will be tested in mock desulfurization experiments in which all of the scheduled reactor operations from desulfurization through hot functional testing and early power operation will be simulated. Samples of 304 SS and the other RCS alloys will be exposed to the coolant during these operations. These long-term tests are designed to' detect any significant degradation of the Inconel 600 and other specimens in time to take appropriate mitigative action, if necessary.
In addition, the licensee has performed tests in which highly susceptible sensitized Type 304 SS and Inconel 600 tubing samples, along with the TMI-1 OTSG tubing specimens, were exposed to the cle~aning solution.
No IGSCC was observed after 400 hours0.00463 days <br />0.111 hours <br />6.613757e-4 weeks <br />1.522e-4 months <br />. The staff finds that these results provide reasonable assurance that no adverse effect on the RCS and the OTSG tubes would occur during the chemical cleaning period of about 200 hours0.00231 days <br />0.0556 hours <br />3.306878e-4 weeks <br />7.61e-5 months <br />. Therefore, the staff does not agree that an additional corrosion test in a cold, high oxygen, and high concentra-tion sulfate (10 ppm) environment is needed, as suggested by a staff consultant (Attachment 3).
Basec: on the above information, the staff finds,that there is reasonable assurance that the proposed hydrogen peroxide treatment of the RCS will effectively remove an appreciable quantity of the sulfur occluded in the corrosion films on the steam generator tube and RCS surfaces, and that the alkaline peroxide treatment will not have a significant adverse effect on the Inconel 600 steam generator tubes or the 304 SS and other RCS alloys.
(
e m-e em one we
--e 7, -.-,-
51 -
('-
. Conclusions The chemical factors causing corrosion of the steam generator tubes have been identified; the corrosion-causing contaminants in the coolant and on the RCS surfaces will be removed, and procedures have been implemented to prevent their re-introduction. Therefore, the staff' finds that GDC 1, 14, 15 and 31 have been met as-they apply to reducing the possibility of a rapidly propagating failure of the RCS pressure boundary and that reasonable assurance exists that the public health and safety will be protected.
Because the long term corrosion " lead tests" can provide information which should be fa.ctored into plant operations, the license will be conditioned to require routine reporting of test results on a quarterly basis as well as more timely notification if adverse corrosion test results are discovered.
I
(
ea
=em.
r-.
p.
w--9
52 -
(*
3.7 Procedures to Prevent Re-Introduction of Contaminants The following' measures have been implemented to prevent re-introduction of contaminants to the RCS.
1.
The sodi.:::: thiosulfate tank has been drained and the piping connecting it to the RCS has been physically severed.
2.
All RCS piping, tanks, valves, the reactor vessel and other components which had contacted thiosulfate solutions were flushed to remove soluble sulfur compounds to a concentration of less than 0.1 ppia sulfate in the coolant.
3.
Administrative controls have been instituted on all pathways by Twhich foreign chemicals might be injected into the RCS. These pathways include the Lithium Hydroxide Mix Tank, the Boric Acid
(.
Mix Tank, the Reactor Coolant Bleed Tanks,-the Borated Water Storage Tank and the Sodium Hydroxide Tank.
4.
New analytical procedures have been implemented to detect the ingress of deleterious chemicals. The coolant will be sampled daily for sulfur analysis and continuously monitored for pH and conductivity.
l l
5.
New limits have been placed on primary water chemistry to prevent the development of an aggressive coolant environment. These changes are summarized in Table 3.7-1.
l l
6.
The RCS will be treated with an alkaline peroxide to remove a I
large fraction of the sulfur occluded in the oxide corrosion film on RCS surfaces. The tightly bound remaining sulfur will l
not be subject to sudden release to the coolant in corrosive L
concentrations.
l I
e oe.
m--
w
-e--
v--
%w--
w--,e-
--,-me--emerwee m
-4 r++-
w-----
v-r-,
w-+-
v v-wt-w7--
r-
W"
- ('
8 The staff cuncludes that the above listed measures provide reasonable
- assurances that sulfur-containing contaminants will not be re-introduced to the RCS.
Table 3.7-1 Three Mile Island Unit 1 Primary Water Chemistry Administrative Limit Changes OLD NEW SAMPLING SAMPLING PARAMETER FREQUENCY FREQUENCY OLD LIMIT NEW LIMIT Lithium NONE Daily' O.2-2.0(ppm) 1.0-2.0(ppm)
Varies with boron concentration Chlorides 5X/wk SX/wk F0.15 (ppm)
Sulfate, (50 =)
NONE DAILY NONE F100 (ppb) 4 Sodium NONE 2X/wk NONE
_F1.0 ppm
(
pH SX/wk 5X/wk 4.8-8.5-4.6-8,5 Conductivity SX/wk SX/wk NONE Check for Con-sistency With Boric Acid and LiOH Con-centration l
+.--eo m
g,
.n...
p,
,___,_g,
,y,
,e
.--3
_y
,.--m.-
(-
.3. 8 Crack Arrest Considerations In Section 3.1 of this evaluation, the staff concluded that the OTSG tube degradation was caused by thiosulfate intrusion, and the sulfur induced IGSCC occurred during cooldown or cold shutdown conditions, in a low pH and high oxygen environment, which occurred subsequent to the hot.
functional testing period.
Literature information on sensitized stainless steels (1) and laboratory testing conducted by the staff consultant (Attachment 2) show that sulfur-induced IGSCC of stainless steel and Inconel 600 can be prevented by base additives such as lithium hydroxide.
Laboratory results obtained by the staff consultant (Attachment 2) also demonstrated that greater than 1.0 ppm of sulfur as thiosulfate is required in the bulk water to initiate and to propagate cracks.
Furthermore, the staff consultant's testing also demonstrated that the addition of lithium hydroxide to the solution inhibits crack initiation and propagation.
Cracking can be preventcd by keeping the lithium concentration ten times
(
that of sulfur.
In Section IV of Topical Report 008, Rev. 2, the licensee described the steps taken to mitigate additional damage in the RCS from an aggressive environment.
Prevention of direct injection of contaminants will be accomplished by the rem. oval of the sodium thiosulfate tank and by administrative controls.
Chemistry changes have been made to include an analysis for sulfur, a conductivity consistency check which will indicate the need for reanalysis of samples, and an increase in the lithium concen-tration specification because of its inhibiting effect on crack initiation.
In addition, to preclude' reactivation of the sulfur which is presently in the OTSG and RCS, a chemical cleaning program will be conducted to remove or oxidize residual sulfur.
Specific measures which have been instituted by the licensee to reduce further sulfur-induced corrosion ere listed below.
(
v e
(-..a.
Removal of the sodium thiosulfate tank has eliminated the source of sulfur contamination. By instituting a chemical cleaning program to remove or oxidize residual sulfur remaining in the RCS, there is reasonable assurance that the remaining sulfur in the RCS will be reduced to significantly'less than the minimum 1.0 ppm concentration required to initiate or propagate SCC.
b.
By administratively controlling substances which are placed into the Lithium Hydroxide Tank and the Boric Acid Mix Tank, the licensee has reduced the likelihood of any direct injection of corrosion-enhancing foreign chemicals into the RCS during operation. The staff finds that these administrative control measures provide additional assurance that tube degradation will not recur.
c.
Although it is difficult for the sodium hydroxide in the Caustic
(
Mix Tank to enter the RCS, if it occurred, damage would not be expected because an increase in pH results in a more benign corrosive condition. Based on the laboratory testing and analyses by staff consultants (Attachments 2-4) which substantiate the inhibitive effect of high pH on sulfur-induced IGSCC, the staff agrees with the licensee's conclusion that introduction of sodium hydroxide into the RCS will not cause additional corrosion' d.
The licensee has implemented an administrative limit change in primary water chemistry as given in Table IV-1 of Topical Report 008, Rev. 2.
Specifically, the lower limit for lithium content has been increased from 0.2 ppm to 1.0 ppe. The chloride concentration has been decreased from F 0.15 ppm to F 0.1 ppm.
In addition, analysis for sulfur at a frequency of five times per week has been instituted with a maximum limit of 0.1 ppe sulfur in the RCS.
(
o-o
.56 -
k" According to laboratory testing conducted by the staff consultant (Attachment 2), lithium exhibits an inhibitive effect on sulfur-induced crack initiation and propagation, when the lithium concen-5 tration is maintained at a concentration ten times that of sulfur.
Therefore, the staff concludes that the increase in lithium concen-tration, in conjunction with keeping the sulfur concentration below 0.1 ppm in the RCS, provides additional assurance that the sulfur-induced IGSCC will not recur.
As a part of the new administrative limit change on primary water e.
chemistry, the licensee, in Table IV-1 of their Topical Report, stated that pH and conductivity are checked five times per week to confirm that the conductivity reading is consistent with the pH, boric acid, lithium hydroxide, and ionic species concentrations which
~
are being measured. Because these measurements are effective and
(-
reliable in detecting ingress of foreign chemicals, the staff finds t
that the licensee has provided additional assurance that possible intrusion of chemicals would be detected and corrodants identified in case administrative controls fail to prevent the introduction of foreign chemicals into the RCS.
f.
The licensee stated that a hot functional. test of the OTSG will be conducted prior to normal precritical hot functionals. The OTSG hot functional will take approximately thirty days and include extensive leak testing and transients which will maximize stresses on the tubing.
Following the hot functional testing, a cooldown will be conducted.
Cracking of the OTSG tubes occurred during the cooling period or cold shutdown conditions following the previous hot functional test. The licensee's leak testing measures would detect SCC if initiated, l.
o L
9
~ 57 -
-(
Anormalpre-criticalhotfunctioAaltestwillbeperfermedwhich g.
provides a second opportunity to check for crack propagation and leakage.
h.
A very gradual power escalation will be conducted, over a period of months, including holds'at various power levels.
This would mitigate the effects of any potential OTSG tube leaks.
i.
The plant will be shutdcwn during the middle of the fuel cycle for ECT to determine if cor:osion is continuing.
s j.
Administrative limits will be implemented for primary to secondary leakage which will require plant shutdown prior to reaching the plant Technical Specifications limit of 1.0 gpm.
N Conclusions Based'on the above evaluation, the staff finds that the licensee has 1) eliminated the principal source of sulfur corrodanti by removing the sodium thiosulfate tank; 2) removed the sulfur in the bulk water by purification, anc reduced residual sulfur species on RCS surfaces by peroxide treatment; 3) instituted administrative controls to prevent the introduction of sulfur or other contaminants into the RCS; 4) implemented new administrative limits on primary coolant chemistry control to maintain a benign environment by increasing the lithium concentration to ten times the maximum sulfate limit'.of 0.1 ppm and; 5) instituted frequent RCS sulfate analysis, pH and conductivity checks to detect any inadvertent contamination by sulfur or other potential corrodants. The staff finds that these measures provice reasonable assurance that dracking of the OTSG tubes will not recur.
In addition, if cracking due to sulfur does recur after return to operation, it would occur when the systnm is in cold condition.
Post-repair leak tests
(
s
_:_...._.______.___'._____u..r.-
m
_ 2 _.
_____.Z__..
c' G
References 1.
Dravnieks A. and Samans C. H., " Corrosion Control 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", Corrosicn, 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.
TMI-1 Steam Generator Recovery Program Task 7 - RCS Inspections and Requalification. April 16, 1982.
5.
Literature Search on Laboratory Cracking of Materials (Other than Inconel 600) in Solutions Ccntaining Sulfur Species.
D. Cubicciotti, EPRI, Memo Report, April 1982.
6.
TMI-1 OT5G Repair Kinetic Expansion Technical Report, GPUN-TDR-077
\\
November 1, 1982.
7.
TMI-1 Steam Generator Preliminary Test for Kinetic Tube Expansion
" Residual Stress Measurement of Test Blocks at Penn State University" 5054-PT-2 Forster Wheeler Corp. September 16, 1982.
8.
EPRI-NP838 LEFM "BIGIF" Program, 1978 9.
Fracture Analysis of Steam Generator Tubes Part II, Stress Intensity Factor'and Crack Opening Displacement (COD) by F. Erdegen, Lehigh University.
Prepared for GPUNC Dated September 15, 1982.
- 11. Welded Taper Plug Stress "eport, B&W 1002581C-02
- 12. Stress Report for MK-1, B&W 32-1127439-00 13.
Stress Report for MK-3, B&W 32-1127439-01 14.
OTSG Stabi'11zer Design Review, B&W 80-0150-00]
- 16. GPUN Topical Report 010, TMI-1 OTSG Adequacy of Tube Plugging and Stabilizing Repair Criteria.
(
- 17. Safety Evaluation of TMI-1 Reactor Coolant System Cleaning, Topical Report TR-010, Revision 0, March 3, 1983.
G
- e e-4hw n-
~
e t
1;
~~-
' I'T 'L/. TC.;Y CC W. !SSION i
...;n..G7ii*.3.C.;0335
+
Ju.ie 17,1983 s
2,,-
I n
hr. Dona h' E. Hossler 501 Vine Street
. Middletown, Pennsylvar.f a 17057
Dear Mr. Hossier:
Thank you for your letter of April 24, 1983 which asked several questions about the' Tlil-1 steam senerator repair program. We hope that the enclosed answers to jour questions provide the informtion jou are seeking.
Sincerely,
/Y
/
Harold R. Denton, Ofrector Office of Nuclear Reactor Regulation En:losure:
As Stated e
e r*
f f
i
"~
c
'}
e
[
s t
)
A
.e
~
~
~~
GPU tiuclear Corporation 50-289, TMi-1 Mr. R. J. Toole Jordan D. Cunningham, Esq.
Manager, TMI-1 Fox, Farr and Cunningham
' GPU Nuclear Corporation 2320 North 2nd Street P. O. Box 480 Harrisburg, Pennsylvania 17110 Middletown, Pennsylvania 17057 Ms. Louise Bradford TMIA Board of Directon 1011 Green Street P.A.N.E.
Harrisburg, Pennsylvania 17102 p, O. Box 268 Middletown, Pennsylvania 17057 ta. Marjorie M. Aamodt s
R. D. #5 Coatesville, Pennsylvania 19320
- 0ccketing and Service Section Earl B. Hoffman U. S. Nuclear Regulatory Commission Dauphin County Comissioner Washington, D. C.
20555 Dauphin County Courthouse Fmnt and Market Streets Chauncey Kepford Harrisburg, Pennsylvania 17101 Judith H. Johnsrud Envimnmental Coalition on Nuclear Power Union of Concerned Scientists 433 Orlando Avenue e/o - Harmon & Weiss State College, Pennsylvania 16801 1725 I Street, N. W.
Suite 506
- Judge Reginald L. Gotchy Washington, D. C.
20006 Atomic Safety & Licensing Appeal Board U.St Nuclear Regulatory Commission Mr. Steven C. Sholly Washington, DC 20555 Union of Concerned Scientists 1346 Connecticut Avenue, N. W.
J. B. Lieberman, Esq.
Dupont Circle Building, Suite 1101 Berlock, Israel & Lieberman Washington, D. C.
20036 26 Broa&ay New York, New York 10004 Regional Adininistrator( l ~
U. S. H. R. C., Region I 631 Park Avenue King of Prussia, Pennsylvania 19406
+ 9ary J. Edl es, Chairman Atomic Safety & Licensing Appeal Board ANGRY /TMI PIRC U.S. Nuclear Regulatory Commission 1037 Maclay Street Washington, DC 20555 Harrisburg, Pennsylvania 17103
- Or. John H. Buck Atomic Safety & Licensing Appeal Board John Levin, Esq.
U.S. Nuclear Regulatory Commissio.'
Pennsylvania Public Utilities Washington, DC 20555 Commission Box 3255 Harrisburg, Pennsylvania 17120
)
c
GPU Nuc1 ear Corporation
" General CouM el.
Federal Em arfcy Management Agency Mr. Thcmas Gerusky ATTN: Docket Clerk Bureau of Radiation Protection 1725 I Street, NW Department' of Environmental Resources Washington, DC 20472 P. O. Box 2063 Harrisburg, Pennsylvania 17120 Karin W. Carter, Esq.
505 Executive House
'P. O. Box 2357
~
Harrisburg, Pennsylvania 17120 G. F. Trowbridge, Esq.
Dauphin County Office Emergency Shaw, Pittman, Potts & Trowbridge Preparedness 1800 M Street, N.W.
Court House, Room 7
- Washington, D. C.
20036 Front & Market Streets Harrisburg, Pennsylvania 17101 Mr. E. G. Wallace Licensing Manager GPU Nuclear Corporation 100 Interpace Parkway Parsippany, New Jersey 07054 William S. Jordan, III, Esq.
Ms. Lennie Prough Harmon & Weiss U. S. fl. R. C. - t1I Site s
1725 I Street, NY, suite 505 p, o, gox 3il Washington, DC 20006 Middletown, Pennsylvania 17057 Ms. Virginia Southard, Chairman Citizens for a Safe Environment 264 Walton Street Lemoyne, Pennsylvania 17043 Mr. Robert B. Borsum f
- Babcock & Wilcox Nuclear Power Generation Division Suite 220, 7910 Woodmont Avenue Bethesda, Maryland 20814 Mr. David D. Maxwell, Chairman Board of Supervisors Londonderry Township RFD#1 - Geyers Church Rcad
~
Middletown, Pennsylvania 17057 Mr. C. W. Smyth Supervisor of Licensing U1I-l GPU Nuclear Corporation Regional Radiation Representative P. O. Box 430 EPA Recion III Middletown, Pennsylvania 17057 Curtis' Building (Sixth Floor) 6th and Walnut Streets Philadelphia, Pennsylvania 19106 Mr. Richard Conte Governor's Office of State Planning Senior Resident Inspeci:or (TMI-l) and Development U.S.N.R.C.
ATTN: Coordinator, Pennsylvania P. O. Box 311 State Clearinghouse Middle' town, Pennsylvania 17357 P. O. Box 1323 Harrisburg, Pennsylvania 17120 e.-
~-
- 2..
CUESTION 41:
What was the allowable steam tube leak rate for TMI-1 on March 1, 19797 (single and multiples)
ANSWER:
The allewable steam generator tube leak rate for TMI-1, on March 1, 1979, was one gallon oer minute ecmbined leakage from all tubes of both steam generators.
QUESTION #2:
What is the current allowable tube leak rate for TMI-17 (single and multipk)
If changes in the allowable rate occurred between March 1, 1979 and April 22, 1983, please list the date and rate.
ANSWER:
The current allowable steam generator tube leak rate for TMI-1 has not changed.
It is one gallon per minute ccmbined leakage from all tubes of both steam generators.
OUESTION #3:
In " lay-person language," how, when, and how often' does a utility decermine a leak rate? Please provide specific NRC procedures.
ANSWER:
The TMI-l licensee is required to determine total re' actor coolant system (RCS) leakage daily while operating.
This is typically accomplished by performing a' system mass balance.
Significant increases in RCS leakage (through tha steam generator tubes, or other RCS boundaries) would, thus, be detected o'n a daily basis.
To specifically determine steam generator tube leak rates, licensees typically perfom activity balances which ratio the radioactivity of primary coolant to the radioactivity of secondary coolant.
Frem this ratio, one can then determine leak rates.
Since licensees must remain within steam generator tube leakage limits at all times, the level of radioactivity discharged from the secondary system is continuously monitored.
An increase in the level of radioactivity being discharged from the secondary system, indicating an increase in steam generator tube leakage, would actuate an alarm to alert plant operators of the situation so that appropriate actions could be initiated.
Additionally, the TMI-l licensee is required to determined the gross radioactivity of the secondary coolant at least weekly.
An increase in gross radioactivity would provide another indication of increased steam generator tube leakage.
CUESTION #4:
How and when would the utility check to determine if a tube wall is worn partially through?
ANSWER:
The TMI-l licensee is required to periodically perform inservice inspection of steam generator tubing by eddy-current testing or other equivalent techniques.
Eddy-current testing is a means c
U' L
2-whereby the electrical conductivity of a tube is checked i
by passing a coil with an induced voltage along the tube.
If some form of tube degradaticc has occurred wnich has separated the tube metal, an electrical discontinuity will exist.
The discontinuity will be proportional to the amount of missing metal.
The specific tube inspection criteria and inspection frequency requirements are complex.
Generally, inspection frequency varies from 20 to 40 months depending on the results of previous inspections. The crecise steam generator tube inspection criteria including sample size and frequency requirements for TMI-l are set forth in Section 4.19 of Appendix A, the Technical Specification: to the TMI-l Operating License which is available for review j.
at the Government Publications Section, State Library of Pennsylvania, Education Building, Commonwealth and Walnut Streets, Harrisburg, Pennsylvania 17125.
OUESTION #5:
What % of a tube wall at TMI-l would have to be worn before plugging /sleevino would have to take place?
ANSWER:
' The TMI-l plugging _ limit is 40% of ncminal tube wall thickness.
All tubes with defects to a death ecual to or orgater than this limit must be plugged.
This limit is consistent with the plugging ifmit of other operating reactors.
QUESTION #6:
In TMI-1, how many (total number) steam tubes in each of the two steam generators are plugged?
ANSWER:
885 of 15,531 tubes are plugged in the "A" steam generator.
273 of 15,531 tubes are plugged in the "B" steam generator.
We understand that some additional. plugging may be necessary in both steam generators upon comoletion of certain testing recently canpleted.
00ESTION #j;:
What % of power output is lost by the number plugged, as listed in #6?
ANSWER:
The TMI-l plant has been analyzed for design basis transiente and accidents assuming operation with 1500 plugged tubes (about 5% of the total number of tubes), and it was determined that existing operating limits provide adequate margins for the analyzed events. Othcr pressurized water reactors have operated with 10% of the steam generator tubes plugged without derating.
Thus, no operating restrictions such as derating appear necessary for TMI-1.
.--c
,Q.--.
QUESTION #8:
Please give me a scenario of leak rates and number of pluggings that would create a situation where a S&W reactor (sic) may not be economically useful.
ANSWER:
The intent of our review of the TMI-l steam generator repair has been to assure continued public health and safety.
To this end, we have neither performed, nor are se currently performing any analyses of the type +. hat would be required to respond to your question.
Of course, as you may know, a licensee faced. with severe steam generator problems vould probably evaluate all aspects of a number of repair options, including steam generator replacement such as that performed at Surry Power Station and Point Peach recently.
b 4
1 l
l O
O.