ML20011F450

Forwards Util 900131 Draft Ltr Containing Steam Generator Tube Insp Results,In Preparation for 900222 Meeting

ML20011F450
Person / Time
Site:	Millstone
Issue date:	02/05/1990
From:	Vissing G Office of Nuclear Reactor Regulation
To:	Cheng C, Murphy E, Wichman K NRC
References
NUDOCS 9003060050
Download: ML20011F450 (20)

Text

{{#Wiki_filter:._ _ _ _ _ ____ e' February 5, 1990 To E. Wichman i E. Murphy C. Y. Cheng F. Witt J. Strosnider FRON: Guy S. Vissing, Project Manager - Millstone 2

SUBJECT:

NORTMEAST UTILITIES DRAFT SUBMITTAL ON MILLSTONE 2 STRAM GENERATOR TUBE INSPECTION RESULTS The attached draft submittal on the Millstone 2 steam generator tube inspection results is provided to give you an opportunity to prepare for the Feb. 22, 1990 aceting. Although this is a draft, I have been informed that the final signed letter, which is forthcomming, will not represent a chan i s1 ' ' jact Manager P jact Dire I-4 Division of ctor Projects I/II MNU Qg s l h i f l i D 9003060000 90020". IJ ADOCK 05000336 h'DR PDC ic I

g 0,. Othefel Offices e $0lO9h street, Berlm. C0hhDCticut c PREL MINARY rro Re: ECT Inspection Mr. William T. Russell Regional Administrator, Region ! U.S. Nuclear Regulatory Commission 475 Allendale Road King of Prussia, PA 19406

Dear Mr. Russell:

Millstone Nuclear Power Station, Unit No. 2 Steam Generator Insoettion On June 5, 1989,II) NRC Region I requested that Northeast Nuclear Eneroy Company (NNECO) perform an in cycle steam generator eddy current testing (ECT) examination for M 12,1989,gstone Unit No. 2 during the current Cycle 10 operation. On July NNECO responded by letter to the Region I request and committed to perfor:n an ECT inspection of the Millstone Unit No. 2 steam generators with an estimated shutdown time frame between September and October of 1989. A more definitive 30,1909.gr}edule ano detailed work scope was provided by 3 1etter dated August In continuing dialogue with the NRC Staff, NNECO has been requested to present the latest steam generator CCf results and an overall safety assessment of the steac generat9rs to the Staff in a meeting to be hold on February 8,1990, in the NRC Headquarters in White Flint, Maryland. In preparation for that meeting, NNECO submits the attachet/ information to allow the NRC sufficient time for review of this matter. Attachment I to this letter provides.NNECO's safety assessment for Millstone Unit No. 2's steam generators with an overall conclusion, based upon our analysis of current Millstone Unit No. 2 testing (1) W. T. Russell letter to E. J. Mroczka, " Millstone Unit No. 2 Mid-Cycle Steam Generator Inspection," dated June 5, 1989. (2) E. J. Mroczka letter to W. T. Russell, " Millstone Nuclear Power Station, Unit No. 2, Steam Generator Inspection," dated July 12, 1989. (3) E. J. Mroczka letter to W. T. Russell,

• Millstone Nuclear Power Station, Unit No. 2, Steam Generator Ins tion ",dat,ed g u 89.

h3 4 g Os34u REv.4-se

1, o Mr. William T. Russell B13443/Page 2 i January 31, 1990 .i results and other industry data, that continued operation through the current Cycle 10 is fully justified. Should you need additional information or clarification prior to the February 8, 1989, meeting, please contact us. Very truly yours, NORTHEAST NUCLEAR ENERGY COMPANY l PRE [lplMADV no c es Attachment cc: G. S. Vissing, NRC Project Manager, Millstone Unit No. 2 W. J. Raymond, Senior Resident Inspector, Millstone Unit Nos. 1, 2, and 3 U.S. Nuclear Regulatory Commission, Document Control Desk i L L l 1 w,

i -1, '= Docket No. 50 336 B13443 1 i L Attachment ! e Millstone Nuclear Power Station, Unit No. 2 Steam Generator Safety Assessment PRELIMINARY b b I / January 1990 e t 4-w m.- -...-m..-..----.-m _...w,,_-es-.--.. ,.,-.v,....-._.-... . - -. - - -. -. - - -,.. -.,... ~.. e

~- Mr. William T. Russell B13443/ Attachment I/Page 1 January 31, 1990 SAFETY ASSES $NDrr OF TEE MILLSTONE 2 STEAM CENERATORS INTRODUCTION The purpose of this assessment is to evaluate the structural integrity of the Millstone Unit No. 2 (MP2) steam generators (SGs) with regard to safety and reliability. The SG tube bundle forms a portion of the primary pressure ) boundary. Corrosion induced degradation of the tubes and tube supports has occurred as a result of past operation of the unit. This degradation is identified by nondestructive inspection of the SG tubes. A in-cycle inspec-tion of the SG tubes has been recently completed. The intent of this addi-i tional inspection was to identify any stress corrosion cracks which may have developed and to ascertain whether the corrosion process was under control. The following discussion addresses the effeet of known and potential degra-dation on the integrity of the pressure boundary and its implications with regard to safety and reliability for the remainder of the current operating cycle. DISCUSSION Corrosion induced degradation of the tubes and supports has been identified in the MP2 SGs by various nondestructive and destructive inspections including eddy current testing (ECT), ultrasonic testing, radiography, 11beroptics a.sd removed tube examination. Corrosion of the tubes has occurred as a result of concentrations of corrosive cherical species on the secondary side of the l

tubes, primarily in sludge areas within 13 inches from the top of the tube-i r; hee t.

Tvo corrosive mechanisms, pitting ami stress corrosion, are pr!.ncipal-ly responsible for the existing tube degradation. Both the hot leg and cold l leg sides of the SG have been affected by each mechanism ts varlovs degrees. Tube support degradation has occured principally on the hot leg vide of all rupport levels. o Pitting Pitting has occurred in the MP2 SG tubes as a result of acidic chemistry conditions and sludge build up on the tubesheet early in the life of the SGs. The pitting has occurred in all four SG plenums. Tubes with pits identified by ECT as having depths of greater than or equal to 40 percent throughvall are either repaired or removed from service. Improvements in k secondary side water chemistry and chemical cleaning in 1985 which g removed a large portion of the sludge pile effectively stopped the progression of pitting. Prior to 1985, a large inventory of pits less than 40 percent throughvall vas created. These pits have not shown any significant growth over the last several cycles, indicating that the conditions and environment which originally promoted the pitting no longer exists. l' PRELIMINARY

f..

q.. j..

X* Mr. William T; Russell '813443/ Attachment 1/Page 2 January 31, 1990 c-Tube Support Degredation Tube support degradation of the MP2 SGs has occurred as a result of acidic chemistry conditions early in the operating life of the SGs. The aci3 conditions caused the carbon steel material to corrode and form magnetite.. The build up of the magnetite corrosion product in-turn caused denting of the tubes and shifting of the supports. Elimination of the corrosive acidic environment has stopped the further degradation of the supports. The progression of support degradation, as measured-by tube denting, has been approximately zero since Cycle 7. A fiber optic inspection of the first eggerate tube support was performed during the EOC9. Refueling Ootage (RFO). This inspection revealed an eggerate con-dition much better than laboratory testing of the model boiler eggerates and better than assumed in previous evaluations. Although tube boving was noted, no loss of support was identified by the fiberoptic inspec-tion. No examination of the tube supports was performed during this ou t age',' o Stress Corrosion Cracking Stress corrosion crack N has developed in the MP2 SG tubes at the top of the tubesheet on bot-hot leg and cold leg sides. The cracking is OD initiated and cire v.ntially oriented. The cracking was first dis-covered in a leakit.,, ioe (SG IH L25/R19) in-January 1987. At the time the safety significance of. a circumferential1y oriented crack vas assessed.and operation of the unit was concluded to be safe with a reduc-tion of the primary-to-secondary leak rate to 0.15 gpm. Subsequent ECT inspections during the E008 RF0 identified 26 additional

cracks, all in the SG-1 hot leg. In addition, during the outage, Tube L25/R19 vas removed for examination. Based on the results of the exami-nation of L25/R19 and the ECT inspection results, the safety significance of the cracking was reassessed and found to be acceptable with a further reduction of the primary-to-secondary leak-rate to 0.10 gpm.

Destructive examination of the leaking Tube L25/R19, removed after almost a year in operation while the tube was plugged, confirmed the presence of e .an OD initiated stress corrosion circumferential crack. The tube con-L tained a continuous circumferential crack which was throughwall over l 190', and greater than 50 percent throughwall over 260'. The remaining 100' of the tube circumference contained several individual microcracks less than 50 percent throughwall. Based on-examination of the fracture surface and review of MP2 SG chemistry, the best estimate of the causative chemical species was determined to be caustic. Evaluation of the observed crack size indicated that the tube should not have been able to withstand operating loads and therefore the crack must l-have grown during the time the tube was plugged. Careful evaluation of the fracture surface showed that the continuous segment of the crack L PRELIMINARY

-1 d

3.,

. es. PRELIMINARY !L actually initiated as approximately 20 separate microcracks which even-tually linked up by stress corrosion. It was postulated that the crack developed by the following scenario 't 1. Caustic concentrating in the sludge area initiated circumferentially oriented microcracks on the OD surface of the tube. 1 2. Existing microcracks propagated as additional microcracks initiated. l 3. As the microcracks propagated, the ID of the tube was breached when a crack vas propagated between 14' and 21* around the circumference. This aspect ratio is typical for stress corrosion of SG tubes. 4. The throughwall portion of the crack enlarged and the tube began to L leak. The remaining ligaments between the microcrack provided sub-L stantial structural strength. The initial 0.10 gpm leak rate indi-cated a throughvall opening of approximately 35'. 5. The tube was plugged. 6. The remaining ligaments between microcracks stress corroded during the operating period while the tube was plugged, forming the observed continuous crack. The postulated cracking scenario was reviewed with Westinghouse Electric Corporation.. Vestinghouse agreed that the postulated scenario. vas i reasonable and closely matched results observed for tests of circumfer-ential-cracks in North Anna SG tubes. Using the postulated cracking -scenario and the observed leak rate, calculations were performed which demonstrated that the crack in Tube L25/R19 shoved acceptable " leak before break" behavior and at the initial leak rate of 0.10 gpm main-tained the structural margins required by Regulatory Guide 1.121. Corrective actions to control the stress corrosion cracking-of the tubes were instituted in Cycle 9. These actions consisted of on-line boric acid addition, begun approximately two months into Cycle 9, and further l improvement in bulk veter chemistry impurity levels. The addition of L boric acid was expected to slow the propagation of incipient caustic L stress corrosion cracks and halt the formation of new caustic stress corrosion cracks. The unit suaessfully operated the entire Cycle 9 without the development of any primary-to-secondary leaks. l. An inspection of the MP2 SGs was performed during the EOC9 RF0 using the l best. equipment and' techniques available in the industry at the time. j; Improvements in the equipment and techniques had been made since the previous refueling outage inspection. Following the completion of all E l EOC9 nondestructive testing, 309 tube ends were identified as being cracked at the top of the tubesheet. Cracks were identified in all four plenums with 85 per crac n n th a

a., o 14 1 mg.: PRELIMINARY i 60 percent /40 percent split between the hot rg and cold leg, respec.- tively. The location of cracking within the SG tube bundle corresponds with both a corrosive environment and high tube stress. The sludge pile provides.the corrosive environment, while high tube stress develops as a result of tubesheet denting and support corrosion. The support corrosion causes shifting of the tube bundle and boving of the tube. Tube bowing has been demonstrated by the inability to successfully-pass long rigid objects (such as sleeves and stabilizers) through the top of tubesheet region of some tubes. The fact that support corrosion as measured by tube denting has been inactive over the last three operating cycles pro-vides reason to suspect that the tube stresses at the top of the tube-sheet have been present since 1985 or earlier and are not increasing. The identification of 309 cracks during the EOC9 RF0 created the impres-sion that the number of cracks was rapidly increasing.

However, increased rotating probe testing (468 tube ends EOC8 versus 12,556 tube ends 'EOC9),

reduced probe speed and a tighter crack determination criteria contributed to the large number of cracks detected during the EOC9 RFO. In addition, the criteria for determining which tubes to test with the rotating probe technique changed from bobbin Distorted Roll Indications (DRI) EOC8 to 100 percent of the crack boundary region E0C9. An added effect was that the E0C9 rotating probe test was run at a slover speed than during EOC8 (200 rpm and 0.1 inch per second versus 300 rpm i and 0.2 inch per second). The slover speed reduced the level of noise generated by the rotation of the probe past the tube transition and l allowed for cracks. of smaller signal amplitude to be more easily detected. A comparison of the limited E0C9 ultrasonic test results with the EOC9 rotating probe results demonstrated that the rotating probe data analysis criteria used EOC8 did-not identify certain cracks, therefore, the analysis criteria was tightened. lL The number and apparent growth of cracks on the cold leg did not appear to be consistent with the expected Arrhenius behavior with temperature typical of caustic induced stress corrosion. Normally, the hot leg would be expected to show a larger growth rate. One possible explanation was that the cold leg cracks which developed to a state identifiable by bobbin coil ECT during E0C9 RF0 actually initiated during Cycle 8 while caustic chemistry conditions existed in the SGs, and propagated slower L due to the lower temperature. Variation in sludge consistency between l-the hot leg and the cold leg, variations in diffusion parameters with temperature, solubility variations with temperature, variations in boric acid effectiveness with temperature and other unknown factors also may i have contributed to the observed cold leg cracking. This indicated that new cold leg cracking could be expected to occur during Cycle 10. Based on review and interpretation of the EOC9 inspection results, the L-corrective actions to mitigate stress corrosion of the tubes were con-l sidered to be effective. The historical ECT review indicated a downward trend in the rate of cracking. The cracks which were identified as new

..-. *(

sq i ~ Mr. Villiam T. Russell '.B13443/ Attachment I/Page 5 January 31, 1990 e cracks in a boric acid environment. The -safety significance of the cracking was reassessed following the EOC9 RF0 inspection and found to be acceptable. Projections of the number of new cracks expected to form in Cycle 10 were made assuming effective corrective actions. On a '.,ast-estimate basis, 60 new cracks were expected by the in-cycle inspection i and 75 new cracks.by EOC10. Upper bound estimates were 200 new cracks by the in-cycle inspection and 250 new cracks by E0010. If an active cor-rosion process were assumed, over 1,000 new cracks vould have been 1 expected by E0010. During October 1989, MP2 was shut down to perform an in-cycle inspection of the SGs. The purpose of this inspection was to identify stress cor-rosion cracks which had developed to a detectable size since the EOC9 RTO inspection and to ensure that the corrosion process responsible for the cracking was under control. The inspection was performed with the best -equipment and techniques available at the time and reflected improvements which " had been made since the EOC9 RF0 inspection. The inspection pro-gram consisted of an examination for cracks by rotating pancake -coil probe testing of all,ubes within the region where cracks vould be expected to occur, plus bobbin coil probe testing for pits. In addition, ultrasonic tests vere performed on tubes which RPC indicated that cracks were possibly present to obtain additional information on crack length and depth. A total of 104 tube ends were identified as being cracked at the top of u the tubesheet. Cracks were identified in all four plenums, with 90 percent of the cracking taking place in SG1, with an approximate 60 percent /40 percent split between the hot leg and cold leg, respec-tively. Figure 1 illustrates _ the location and number of cracks in a representative SG plenum. The location and distribution of the cracking is almost identical to the EOC9 RF0 inspection results. All tubes with identified cracks were staked and plugged. Determination of whether a tube was cracked was made conservatively and it is possible that some of the repaired tubes may not actually contain cracks. An additional 11 tubes vere preventatively staked and plugged, since a positive determi. nation could not be made that a crack vas not present in these tubes. A comparison of.the February 1989 (E0C9) and October 1989 rotating probe data.vas performed for all but six tubes identified as cracked during the October examination. No rotating pancake coil probe data was collected in February on six cracked tubes. The extrapolated results from this comparison identified 35 of the 104 cracks (35 percent) as not being detectable by ECT in February. A total of 69 cracks (65 percent of the total) were present in February. Identification of the cracks using February 1989 data was possible by lowering the threshold for crack determination to the level equivalent to that used for the October 1989 examination. The reasons for identifying cracks in October which vere present (but not identified) in February included: PRELIMINARY

l '.7 ..z [

Mr. Villiam T. Russell 813443/ Attachment I/Page 6

. January 31,.1990 t 1. Improved training of the analysts prior to the October examination, 2. Independent ' review of data tapes in October versus an independent review of RPC graphics only, in February, 3. Lovering the initial reporting threshold to include all "possible" cracks versus "best estimate" cracks. 4. Ultrasonic testing of all tubes with "possible" cracks as identified by rotating pancake coil probe. The small number and downward trend of new cracks present during the October 1989 inspection is consistent with the best estimate prediction. This prediction was based on the observed effect of boric acid treatment on caustic stress corrosion in laboratory tests. The addition of boric acid to the MP2 SGs was expected to halt the initiation of new cracks, and slov'the propagation of existing incipient cracks. The 35 new cracks are likely cracks which were at an incipient stage prior to the boric acid addition and have now propagated to a detectable size. I Based on results from the previous three inspections, the cracks in the MP2 SG tuber can be characterized as follows: the cracks are located at the top of the tubesheet, within approximately 0.2 inches from where the expansion transition meets the nominal tube diameter, and are circumfer-l L entially oriented. The macrocrack, as defined by rotating probe ECT, is made up of several discontinuous microcracks which are separated by lig,- ments of sound material. Circumferential extents of the macrocracks ranged from 14' to 329' by RPC and up to 360' by ultrasonic testing (UT). Crack depths up to 100 percent throughwall were estimated by UT. The discontinuous nature of the array of microcracks was confirmed by UT and removed tube examinations. In general, the. microcracks were located at different axial planes, typically within a 0.1 inch band. A tube with a discontinuous array of microcracks is substantially stronger than a tube with a continuous crack of equivalent circumferential. extent and depth. The actual circumferential extent and depth of the macrocrack is rela-tively independent of the operating time. The circumferential extent is-L controlled by the number of microcracks which have initiated around the circumference and the uniformity of the stresses in the tube at the top of the-tubesheet. The depth of the individual microcracks is also con-trolled by the stress field. Once a microcrack has passed the incipient i

stage, the crack vill propagate to the depth supported by the stress 1-field fairly rapidly compared to the length of an operating cycle.

Projection of the number of cracked tubes expected by the E0010 were made i L by extrapolating the current cracking trends. On this basis, 45 tubes [ vill be identified as cracked during the E0C10 RF0 inspection. The I cracks are expected to be the result of either the propagation of incip-ient cracks, which developed during the time period caustic conditions PRELIMINARY

'l. a ~ ~Mr. Villiam T. Russell l l B13443/ Attachment I/Page 7 l. January 31,-1990 existed in the SGs, or the identification of' existing cracks which were previously below NDE detection thresholds. A distribution of the circumferential extent of'the macrocracks, as indi-cated by RPC, has been present for each inspection where cracks were observed. The distribution of circumferential extents has remained the

same, regardless of the operating period between inspections. This is illustrated in Figure 2.

This indicates that once macrocracks have

formed, the overall population vill not show any significant growth-in i

circumferential extent. The distribution of the cracks observed during the October 1989 inspection can be modeled by a gamma distribution with a median of 90', at mean of 102', and a standard deviation of 67*, Figure 3. It is expected that the new cracks projected to be identified during the 4 EOC10.. refueling outage vill continue to follov a gamma distribution of the circumferential extents. Cracks with large circumferential extents are potentially more challenging from a structural standpoint. Using the-gamma distribution model, statistical inferences can be made with regard to the number of observed macrocracks which will have large circumfer-ential extents. Given a projection of 45 cracks, the number of macro-cracks which will exceed 240' at the 95 percent confidence level is four and the expected number is two, Figure 4. To date 36 cracks have been found which exceed 240' in circumferential extent, including the tube which leaked at 0.1 gpm in January 1987 (Tube L25/R19 254' circumfer-ential extent). Three tubes were removed.from the MP2 SG during the October 1989 shutdown for burst testing and destructive examinations.. The tubes contained the 'following circumferential cracks: TUBE RPC EXTENT UT EXTENT UT MAX DEPTH SG2 Cold L118/R14 2918 340' Cluster 80% SG1 Cold L52/R22 191' 155' 80% 110' -60% SG1 Cold L145/R23 88' 80' 40% 65' 40% Tube SG 2C L118/R14 was considered a worst case crack based on the combi-nation of circumferential extent and depth as indicated by nondestructive I inspections. This tube was pressurized with water at increesing pressure until bursting at a pressure of 11,200 psi. The failure was axial, coin-ciding with a groove created by the removal process. Tube SG IC L52/R22 was also tested, developing a leak at 8500 psi. .This tube could not be burst because of test equipment limitations. A virgin tube vould be expected to fail axially at a burst pressure between 11,000 psi and 12,000 psi. 1 PRELIMINARY ..,y

g 4 s Mr. William T. Russell-B13443/ Attachment I/Page 8 ' January 31, 1990 The burst pressures of the removed tubes compare quite favorably with the normal operating differential pressure between the primary and secondary side of the. tube -of-1375 psi and the maximum accident pressure, with dynamic effects considered, of 3150 psi. Applying the Regulatory Guide. 1.121 safety factors of 3 on normal operating loads and 1.4 on accident loads produce pressures of 4125 psi and 4450 psi respectively.- Despite the large extent of cracking and potential weakening of the tube as a result of removal from the SG, both of the tested tubes fully met-the j Regulatory Guide requirements. For -circumferential cracks of the type end location found in the MP2 SG, a substantial portion of.the cross-sectional area must be corroded before the tube would no longer fail axially at virgin tube pressure in a burst test. The presence of the tubesheet provides restraint to bulging of the L tube in the region of the crack and effectively increases the corroded area which can be tolerated without loss of margin to burst. In addi-tion, 'the stiffness of the tube, the tube supports, and the presence of adjacent tubes restrain bending of asymmetric circumferential cracks, j such that the corroded area averaged over the circumference of the tube can be used to provide a reasonable approximation of the burst strength of the tube. It is estimated that a tube vill continue to fail axially with no loss of margin to burst with circumferential cracks up to approx-imately 58 percent throughvall averaged over 360'. The Regulatory Guide 1.121 safety factors to burst continue to be met by a uniform 77 percent throughwall 360' circumferential crack. These estimates do not. consider the additional strength provided by ligaments between the microcracks and the presence of microcracks at separate elevations. Regulatory Guide 1.121 requires an operational allowance for tube degra-dation that may occur before the next scheduled tube inspection. SG tube rupture scenarios are restricted to single tube rupture events. There-

fore, the operational allowance must incorporate both leak-before-break and frequency challenge criteria. The allowable operating period before the next inspection outage vill be the length of time required to accumu-late not more than one crack which could potentially exceed Regulatory Guide 1.121 margins.

l L Following the in-cycle inspection, UT depth information was analyzed to determine the depth profile of the cracks. This information was then l used to calculate an equivalent throughwall degradation when averaged over the 360' circumference of the tube. Equivalent depths were also calculated for the tube which leaked in January 1987 and the tubes removed for destructive examination. Similar to the distribution of circumferential extents, the distribution of equivalent depths was modeled by a gamma distribution with a mean of 22.3 percent throughwall and a standard deviation of 15.4 percent throughvall, Figure 5. A similar distribution of equivalent depths is expected for any cracks which are identified in the future. Using this distribution, statistical inferences can be made with regard to the number of tubes with'large equivalent depth crack G o to 5e t

I g g} ,/ se PRELIMINARY i two cracks are expected to have equivalent depths greater than 58 percent throughwall, Figure 6. In addition, not more than one crack vill have an equivalent depth greater than 77 percent at the 95 percent confidence level. Even this postulated low probability crack would not erode - Regulatory Guide 1.121 margins due to mitigating factors known to be present such as the discontinuous nature of actual cracks and the strengthening provided by the tubesheet. 1 The tube with the array of cracks with largest equivalent depth was tube 1H L25/R19, which leaked at 0.1 gpm in January 1987. This tube had an equivalent depth of 73 percent-throughwall when averaged over 360' as p indicated.by the destructive exam. Tube 2C L118/R14, which burst axially in testing at 11,200 psi, had the largest equivalent crack of the tubes removed in October 1989, estimated at 55 percent throughwall by UT and 49 percent throughwall by the destructive ' examination. All of the remaining tubes for which UT data was available had equivalent depths less 'than 55 percent throughvall, with the exception of tube IC L45/R13 and tube IC L28/R36 which had estimated equivalent depths of 70 percent throughvall and 60 percent throughvall, respectively. In all cases, the Regulatory Guide 1.121 margins to burst were satisfied, and in all but I three cases, the tube vould fail axially in a burst test without loss of margins when compared to a virgin tube. The results of all nondestructive and destructive testing performed to

date, including the burst tests, fully support the assumptions regarding the corrosion mechanism and crack morphology used to determine the allow-H

- able leak rate of 0.1 gpm. The vast majority of cracks found to date l' have been relatively small in circumferential extent and shallov in equivalent depth, such that margins to burst have not been affected..The few ~ cracks with large extents and equivalent depth either exhibited or i, would be expected to-exhibit a leak before break behavior. The number of cracks identified by the inspections has decreased by a factor of 3 demonstrating that the corrosion process is understood and under control. The' reduction in the number of cracks further reduces the risk of a tube rupture and enhances safety. Improved crack detection methods and con-servative plugging criteria assure that existing cracks, which could l potentially compromise Regulatory Guide 1.121. requirements, have been detected and plugged. The risk of a tube rupture is considered to be extremely lov. In summary, based on our conservative assessment of the in-cycle steam generator tube inspection results (UT and ECT) and the available burst test data, it is concluded that the observed stress corrosion cracking ..z meets the intent of the applicable structural integrity and primary-to-secondary leakage criteria of Regulatory Guide 1.121. CONCLUSIONS Continued operation of the MP2 SGs for Cycle 10 with regard to concerns associated with degraded tube and support conditions has been evaluated and PRELIMINARY

-.. ~ ). ~ Mr. Villiam T.-Russell' B13443/Attachsent I/Page 10 January 31, 1990.- y found to'be safe in accordance with -applicable regulatory guidelines. ECT performed :during the October 1989 outage has identified tube defects due to corrosion which could potentially compromise the margins of safety required by a Regulatory Guide _1.121. These identified defects have been removed from service. Pitting and cracking of MP2 SG tubes have been~ analyzed and found to exhibit acceptable leak before break behavior. If ' additional pitting or cracking vere to develop during the remainder. of the operating cycle, struc-tural integrity.of the tube primary pressure boundary, with the appropriate i margins of safety required by Regulatory Guide 1.121, is maintained provided the 0.1 gpm primary-to-secondary leakage rate of the MP2 technical specifica-tions is not exceeded. Leakage-in excess of the technical specification limits requires shutdown of the unit, ensuring safety. i o l PRRIMINARY

PREUMINARY OCTOBER 1989 RPC INSPECTION RESULTS IEELLSTQW 4dE3T 3 SG 4. Isef LES i o.m.. ,r , s 33, m,s r,r, . 'J. r,pt. ~,,.. t * .r. r. 330 p

=

- <awwat

e ammm. w

( s. a

' d.r ".o r LGIM-:'4prh'/.MGO iG 'e '.G id t.1 MdpC.Gi i:. 3.e .p,p,w-.,3 -,3. :....~.,,..w...,, #... p...,,:,.

,:, w

w..,y,,#.. w 4,l,e L

. :m w :n w ew ream a w

w. 9:: e vs. m g.> :::.

St. 1 .:::sc,ce: x. s a,. +:m

.i.:.:J:.:.:.:.:J:.;c... +: pgn7p:;;:

t . :.:,o:wwc-e9t.:e, ;.5 ;55 55:55FM55 :tJ-VlWC.1,'?. li'i

",e 5

,: E,w

.- op;e:g.g:;g;;;s :;s ;j;t :t;P;t$s ;$i K r'c 'tk c5 M,x M',:1 :,' ':,': E d$b ,_s,:::: 5<: :it.,:.M.2:5N.s.:53ES3...:@..i):ii,?.$.,.T.. U,,c,$..U,,. rm$ 1 'i,. s. r . m..+v.~.- y.s,,,, 30 . (. ... + .....s..- .y v a

m......_

, =, =. g .o.Gi? e,.r;-.r;r/.c -.3r.: .., ;.3 g 4:~.g-;. t;. ";.. m,,,, m. .~ c. u, s

v,,.

e. ?r,.,,. ,~, w, .t. ". ~. g.. ;.;s .g3.<;.,' ;g. y,,

g.

map a zpco:q;g;+- ;;';;.3:": c.;a 7; g:g<gi.g g g g;, .;;;.:,;3.g g:33.g:.ew. ae - g .m aawe46: 3:..:..A : c. ,y ;.:. 6.;. .. +.m-.cm.oc.m.s :.:! 4::: ::::Js3.',i:.: 'i,i$$;- is s-x:t.; M.2fiq: b. .. ;, is. .. ;, ;s. q. .coc z . -i.ti.e:i'...:. t q, .4 g, _ x..~. .....v,..at .x.,, .v. ,v.,w ,. v....... .c m .m c,

.;. s.

. _.,,.c. M@ps:rE.;3 :st ':::: :M_ %,1;m,. .: r.. ?-: rv *.z ';c v..:s.. :..:";. ". -,c c.c. .= ,,, e- ~ t,,

y....

.r m 69P. M 4MM: g ;;.'l :M;: ;:?l ;gi 9.:"? :!;] ;??$Tsifjy.::: :::."!??;9 J3 2::.'9;;59 R::- M:kk ogR:- eamRV;.s

g: g; wr::.:: ::5 ::::i ::m:!:,:g-: s;::sss :: i;t: '::c;;:: 5:J:i! ?;:. ::: g:a:.c ma. 9:wne

E ::j: M$ :, :.i. $;.c;?8e;,;$;i:s;3$ $< ;i, :.;J$; i:ill l;:1 j;i:, J:;;,;;:1.;;. WN

.ewc-Mh i mm

35 ggc!.;:: :i:t i:py.qg :-:: j:v?: :s;:.:::j:(-:i:!:iti:33 :::- :.gf: ;te;it: si;;;i;:@;il.w:np

.;s. -. s.x, a;. *.,.- ; ss."; .:.: gg M .yr , rn. e. ,g,,m.,;;g. o.. m.pg e, r<;. ~, c.- s-... ,.;.5 a.;;a,. - s-v n.o r.g re g

o.

..s. s, s r r.'. _ r. p.., y 7.~., .g -

n.Rr:3 q:!:1:3h -:,:w iti: !35 i;p :55 q:

:.:-";;i;;wgs;igi:ve ;:!u:::;:!.;ng gre

man:gg

pq ;5 w ;-

,t :":sm:-

gr-;<gi:p:

!:!.:5.:si:5w:: a. w a :

.Q:eti ;. :; s.s: iWi?!4;S$5.i:: -;.;. w 3.?WRm:g ;i:2.;:i tr$r;;st; 5:::lfi 1:i:it;:~

gx?,F:5 e:gs:5::pr;3ssti ;:f:;;i:. ;:-

3it, ' :S i3i:g:,;.:. 2:"f:;c35i:i:;ri:i%

. U..

;:~:.:w.t:.a.s.:4,:.:U.:.:17.c:re

• - pw::3:.r..:m.:.:s.:a., :35 ;:.:a..w.:a.r

a.;:F.:;.:;

. ~.. .a. ?. c .y. -.c rr ...... t; e p rr wr='r.-

f. <

c',&'.r.r r; t. -;rli c'; ^ ', r. f.r.r.r; f,r., '.ryql. ',r.;> ' r ' f. r',r; ..'t ..:.6.. ,t.: r.r.rtst!,f.r t. r;r.

r;-..e 3

a 43343e78 19 a lt t s t e s 33333333333333333n3444444444451555555SM6666666667777777777^^^^ ^^ e stts tal31stI331381133311ste a stat statIt alettiet att t 3:4331at attat allt ala 33 64 334SS30sel3345670ses 334367eest 33436Fenes 3345670se s 3345670ses 3345670sel334M70sel334M78.^^^^^^^^^^ t 3 3 3131111333333333333333333334444444444555555555566666666667 e33343678 012345674e0533456748es3345670e013345670 eel 3343570sel3345670se LIW ISddie ~. FICURE 1 sta aa.= cea er are me caem.m. se ce.a,e.r. arc - cm 1a cu e urw e l .e.. S a

L .PR.F..i..l.M...I.N..A.R..Y. .CIRCUMFERENTIAL CRACK EXTENTS i MILLSTONE 2 STEAM GENERATOR TUBES j numeen or caacus E ALL CRACKS TO DATE 3 C JAN 87 LEAK L24/R10 j Q JAN 84 INSPECTION E FES 1949 INSPECTION G OCT 1989 INSPECTION 3 l 4e - o l l 'Q ~ ~ { f, SG - P A 6 l s j g7pj e y. '/T7/}/ / /////////// 9 J 7.x

1. -

e : 7/,e/yyy i a /b! b..Y........$...$....$....h....././ G $9 40 SG -SG See 126 14 0 10 0 144 200 220 240 See 280 300 320 See See MACROCRACK CIRCUMFERENTIAL EXTENT FIGURE 2 .. j =

L PR.Fi lM...I.N..A.R.Y. CIRCUMFERENTIAL CRACK EXTENTS l MILLSTONE 2 STEAM GENERATOR TUBES NUMBER OF CRACKS .ss - SA448dA D8.TSt.UTload um ocr in.....1. i i 4_ k 1 I I I I f. !iii i.i i i.i i i ! i l i i i l : i Y" i i i i Tl iiii i i: : = = a. u. u. n. m. no

n. - u.

n. MACROCRACK CIRCUMFERENTIAL EXTENT l FIGURE 3 i l ^

l DDEIIMihlADV I ERE.LBETIER1rtR1I PROJECTED CIRCUMFERENTIAL CRACK EXTENTS MILLSTONE 2 STEAM GENERATOR TUBES l FOR A POPULATION OF 46 CRACKS EOC10. l l NUtdSER OF CRACKS 10 Si + 95% CONFIDENCE f I 90% CONFIDENCE .g_ O EXPECTED NUMBER 7 e 6 ~ [ 6 4 3 l i 2 w ~ 1 -{l 0 180 -210 240 270 300 CIRCUMFERENTIAL EXTENT (DEGREES) FIGURE 4 g. s i*- - 9 = = A=

MI.I.M...I.N.A.D.Y. CIRCUMFERENTIAL CRACK DEPTHS MILLSTONE 2 STEAM GENERATOR TUBES UT EXAM i NUMBER OF CRACKS 20 GAMMA DISTRIBUTION M L26/R19 (LEAK) 15 M-OCT-1989-INSPECTION 4 i MEAN = 22.3% 10 STANDARD ~ DEVIATION T15.~4% 5 i i I 0 1 O 5 '1015 20 25 3'O 35 40 45 50 55 6G 65 70 75 80 85 90 95100 MERAGE CRACK DEPTH (% TW) FIGURE 6 --A = ^ =. - - - -

L DDF11MihlADV E BIBueBasM BW a s sim us u a L PROJECTED CIRCUMFERENTIAL' CRACK' DEPTHS MILLSTONE 2 STEAM GENERATOR TUBES: FOR A POPULATION OF 46 CRACKS EOC10 l-. 8 NUMBER OF CRACKS 7; + 95%' CONFIDENCE 90% CONFIDENCE ) [ 6-O EXPECTED NUMBER i 5 .' 4 c 3 L' Y )( 0 ~ 45 50 55 80 66 70 77 AWERAGE DEPTH (% THROUGHMLL) - FIGURE 6 ...., ^ ~ .}}