ML113210370

Rev 1 to GE-NE-637-0005-0393, Core Spray Crack Analysis for Monticello Nuclear Generating Plant.

ML113210370
Person / Time
Site:	Monticello
Issue date:	03/30/1993
From:	Booth R, Plaxton S, Torbeck J General Electric Co
To:
Shared Package
ML113110884	List: ML113110883 ML113210370
References
GE-NE-637-0005, GE-NE-637-0005-0393, GE-NE-637-5, GE-NE-637-5-393, NUDOCS 9303190028
Download: ML113210370 (41)

Text

MAR 08 '93 88: 85PM GENE ENGRG J2455P. P. 2

-a GE Nuclear Energy Te'cat~ Ser'4ces Businessr GE-NE-637-0005-O$93, Rev. I 175 CwMer Awnuw DRP A00-05W4 San Jone, CA 95125 March 1993 CORE SPR~AY CR~ACK ANALYSIS FOR MON'fCU.LO NUCLEAR GENERATING PLANT Prepared jeoR.H. Both, Engineer Plant Performance Analysis Projects Prepared:-

Structural Mechanics Projects Verified :'

HS M P ii ninee Structural Mechaffics Proj=r Approvod Torbeck, Project Manager E'iCS and Containmnent Analysis Projects 9303190028 PDR ADOCK 930.309 05000263 a PDR

MAR 88 '93 08:05PM GENE ENGRG J2455 P. 3 GE-NE 637-0005-0393, Rev. 1 IMPORTANT NOTICE REGARDING CONTENTS OF THIS REPORT Please Read Careull The only undertakings of the General lectric Company (GE) respecting Infomation in this document are contained In the contract between the Northern States Power and GE, and nothing contalned Inthis documed shall be construed as changing the contract. The use of this information by anyone other than the Northern States Power, or for any purpose other than that for which It is intended under such contract is not authorized; and with respect to any unauthorized use, GE makeA no representation or warranty, expressed or implied, and assumes no liability as to the completeness, accuracy, or usefalness of the information contained Inthis document, or that it use nay not Infringe privately owned rights.

I

MAR 08 '93 08:06PM GENE ENGRG J2455 P.4 GENE-637-0005-0393, Rev. I TABLE OF CONTENTS

1.0 INTRODUCTION

1 AND

SUMMARY

1-1 1.1 CRACK LEAKAGE ESTIMATE 11 1.2 STRUCTURAL ANALYSIS 1-1 1.3 LOST PARTS ANALYSIS 1-2 1.4 EFFECT ON LOCA ANALYSIS 1-2

1.5 CONCLUSION

S 1-2 2.0 CRACK LEAKAGE ESTIMATE 2-1 2.1 CURRENT LEAKAGE RATE 2-1 2.2 MAXIMUM ESTIMATED CRACK LEAKAGE 2-2 3.0 CORE SPRAY PIPE STRUCTURAL INTEGRITY 3-1 3.1 POTENTIAL CAUSE OF CRACKING AND LIKELIHOOD 3-1 OF CRACK ARREST 3-1 3.1.1 Cracking Mechanism 3-1 3.1.2 Thmnal Fatigue 3-1 3,1,3 Inergranular Stress Corrosion Cracking 3-2 3.2 STRUCTURAL INTEGRITY 3-2 3.2.1 Sumnary 3-2 3.2.2 Crack Arrest Evaluation 3-3 3.2.3 Allowable Flaw Size Determination 3-3 3.2.3.1 Analysis and Results 3-4 3.2.3.2 Temperature Gradient Evaluation 3-5 3.2.3.3 Flow Induced Vibration Evaluation 3-6 3.3

SUMMARY

AND CONCLUSIONS 3-6 4.0 LOST PARTS ANALYSIS 4-1

4.1 INTRODUCTION

4-1 4.2 LOOSE PIECE DESCRIPTION 4-1 4.3 SAFETY CONCERNS 4-1 4,4 EVALUATION .4-1 4.4.1 General Description 4-1 4.4.2 Postulated Loose Pieces 4-2 4.4.2.1 Core Spray Pipe 4-2 4.4.22 SmaR Pieces 4.2 i

MAR 08 '93 08:06PM GENE ENGRG J2455 P.5 GE-NE-637-0005-0393, 1 Ay. 1

4.5 CONCLUSION

S 4-4 5.0 IMPACT ON ECCS ANALYSIS 5-1 5.1 - IMPACT OF CORE SPRAY LINE LEAK ON SAFER/GESTR LOCA ANALYSIS 5-1 5.2 IMPACT OF POSTULATED FAILURE OF CORE SPRAY LINE AT CRACK LOCATION 5-1

6.0 REFERENCES

6-1 APPENDIX A: STRUCTURAL ANALYSIS OF THE MONTICELLO PLANT CORE SPRAY PIPE A-1 A.1 CRACK GROWTH DUE TO IOSCC A-2 A.2 FATIGUE.CRACK GROWTH A-6 w

MAR 08 '93 08: 07PM GENE ENGRG J2455 P. 6 GB-NE-637-0005-0393, Rev. 1 1.0 INIRODUCTION AND

SUMMARY

During the current refheling and maintenance outage, the vessl in-service inspection identified a crack indicadon (Figure 1-1) on the core spray line at Manticello Nuclear Generating Plant. 'The indication was identfied using an under water camera during the inspection in response to IE Bulletin 80-13 (Reference 1). The crack indication is located outside the sbroud where the piping and junction bo meet in the heat afifated zone (HAZ) of the weld. The following additional information was promvided by Northerm States Power (NSP):

a) The crack was verified by UT inspection to be through-wall.

b) The crack is approximately 3.5 inches in length along the outside diameter ofthe pips based a4 visual mmuremeuts.

GE Nuclear Energy has performed an evaluation to address the safty significance ofthe throughWall amck The technical basis to support the continued structaral integrity ofthe core spray line for all normal and injection conditions is provided. A discussion of the possible consequences of potential loose pieces from a cracked pipe is also presented. Finally, the consequences of a postulated Loss-of-Coolant Accident (LOCA) with a crack in the core spray piping are discussed.

1.1 CRACK LEAKAGE ESTIMATE Abounding calculation to estimate the leakage through the crack. presented in Section 2, demonstratud that the total leakage is well within the margin inherent inthe core spray system design and performance evaluations. The results indicate that for this crack configuration including the postulated crack growth, the total flow leakage is conservatively esinmated to be 24 gpm..

1.2 STRUCTURAL ANALYSIS The structural analysis, precuted inSection 3, concludes that the integrity of the core spray piping will be malitained for all conditions of operation over the next operating cycle. In addition, potential causes of cracking are discussed, and based on the infnnation available, it is expected that the most likely cause isIntergranular Stress Corrosion Cracking (IGSCC).

1-1

MAR 08 '93 08:07PM GENE ENGRG 32455 P.7 GE-NE-637-0005-0393, Rev. 1 1.3 LOST PART ANALYSIS Because continued sparger structural integrity was demonstrated, lost parts (loose pieces) are not expected. Nevertheless, a lost parts analysis has been performed and is presented in Section 4. It is concluded that the probability of unacceptable flow blockage ofa fbel assembly or unacceptable control rod intedbreace due to lost parts is negligible. The potential for corrosion or other enical reactims with reactor materials does not exist because the piping material is designed for in-vesel use. It is also shown that loose pieces are not expected to ause damage to the other reactor pressure vessel internals.

1.4 EFFECT ON LOCA ANALYSIS Section 5 presents th results of the LOCA analysis. The results show that the inherent conservatisms present in currnt LOCA analyses mare than offset the small amount of leakage estimated through the crack. It is concluded that no change to the present Maximum Average Plant Linear Heat Generation Pat (MAPLHGR) fbr Manticello is required.

1.5 CONCLUSION

S A detailed evaluation of the Monticello core spray crack has been performed. Ithis evaluation included structural, lost parts and LOCA analyses to determine the impact on plant operation with the crack inthe core spray piping. Based on t analysis, it is concluded at Monticello can safely operate in this condition during the next fuel cycle, and that no operational changesor restrictions are required during that period.

1-2

I

.s CRACK fEss SPRAY uns C)>

O 4-. o I

q~I to Figure 1-1 T-tus with crack 4 Vent Bele

MAR 08 '93 08:08PM GENE ENGRG J2455 P. 9 GE-1E637-0005-0393, Rev, 1 2.0 CRACK LEAKAGE ESTIMATE There are no direct measurements of leakage fm the crack during the operatian of the core spray system However, from prvious analyses and tests performed fbr the cracks observed in other BWRs, it Ispossible to establish an upper bound leakage for the crack Identified at the Montiello Plant.

The gniflance ofprevious crack occurrences at other BWRs has been assessed by both visual inspections and air-bubble tests. Based upon these inspections and tests, the upper bound leakage was estimated to be Ices than halfthe leakage through the 1/4 inch vent hole presc-t in the T-box. (The vent hole is part of the original piping design and is included to allow the release of any von-condensable which could collect inthe core spray piping). The video fon the Monficello inspection indicated that the crack inthe Monticello core spray line isconservatively estimated to be 3.5 inches in length. Consequently, itis conservative to assume that the maximurn leakage fram the Core Spray Line crack is approximately 24 gpm, assuming a 180s (48.5") through-wall 2.1 CURRENT LEAKAGE RATE The vent hole is a 1/4 inch hole present inthe T-box. The leakage rate through the vent hole is estimated assuming incompressible Bomoulli flow through the hole:

Q= CA,/2gOAP/p where, C flow coflicient (assumed to be0,6 for an abrupt conanction)

A area p = mass dwsity of fluid AP = pressure difference across the pipelvent The flow rate through the vent hole was detemnined utilizing a bounding pressureof 125 psig acs the core spray line (actual pressure - 111 psig). This corresponds to the differential pressure expected during the rated core spray flow conditions. Utilizing the equation above, tbo estimated leakage rate through the vent hole during a LOCA was determined to be less than 13 gpm. Therefore, during the core spray injection phase of a LOCA, the total leakage through the crack is expected to be less than 5 am (less than one'balf of the vent hole leakage).

2-1

MAR 08 '93 08:09PM GENE ENGRG J2455 P110 P.

GE-NE-637-0005-0393, Rev. 1 2.2 MAXIMUM ESTIMATED CRACK LEAKAGE In order to estimate the rnadmu leakage expected through the crack, the configuration for a 180o-hrough-wall crack was used. This configuration was considered to be the upper bound based an the crack arrest results of Section 3.0. Acrack width of 0.01 inch was conservatively assumed based an the results of Linear Elastic Fracture Mechanics (LEFM) methods which showed the crack opening to be < 0.01 inch under the applied loads described in Section 3.0.

Using the methods of Section 3.2 for these loads and the 180' through-wall crack coafigurtion, the leakage was detenmined to be 24 gpm, It was also estimated that the current crack size is expected to grow less than 1.2 inchca during the next 18 month cycle (4.7 inches in length). This result isbased on the consideration of both IGSCC and thtigue crack growth. For IGSCC, conservative crack growth rates at moderate conductivity for 304 stainless steel were assumed (4x10 5 in/hour), and crack growth from both ends of the crack was considered. The assumed IOSCC crack growth rate is considered conservative for two reasons:

1) The assumed value of crack growth is based on normalwater chemiistry conditions.

The Monticello plant isexpected to operate with hydrogen water chemistry. Although the lctro chemical potential (ECP) in the area of the core spray line is not epected to meet the necessary lowl for full IGSCC protection, the crack growth rate is likely to be substantially lover than the assumed value of 4x10 5 in/hour.

2) Using the NRC curve, the assumed crack growth rate of 4x05 inbour is predicted at intensity factor, K,of 23 ksi(in) 1'2 . Sinca the subject crack is through wall a stress (thus, the weld residual stress induced Kisexpected to be very low) and the applied strs are low, tho K valucs are xpected to be less than 25 ksi(in)112 for realistio crack geometries.

llms, een after 18 months of additional operation, the crack length is expected to be less than 100" of the pipe circumncrance. Therefore, the leakage estimate of 24 gpm for a 1800 crack length is consvative for th next cycle.

2-2

MAR 08 '93 08:10PM GENE ENGRG J2455 P.11 GE-NE.637-0005-0393, Rev. 1 3.0 CORE SPRAY PIPE STRUCTURAL INTEGRITY The structural Integrity aspects of the core spray piping have been reviewed to assess- a) the potential cracking mechanism, and b) the impact the crack could have on the structural integrity ofthe piping. Structural analyses were performed to determine the potential sources of

-stre in the piping, the potential causes of cracking, and the likelihood of crack propagation.

Although there is carrantly not enough infonmation to definitively determine the mode of cracindg it is expected that the crack is due to an Intergranular Stess Corrosion Cmadkig (1GSCC) mechanism. T1h results of thesea ments are discussed below:

3.1 POTENTIAL CAUSE OF CRACKING AND LIKELIHOOD OF CRACK ARREST 3.1.1 Craclng Mechanism Cracks in the core spray line could be due to either thermal fktigue or intergranular stress corrosion cracking (IGSCC). At this point, there isnot enough infomation to determine the mode oferacking, definitively. Consequcatly, the following discussion addresses the implications ofthe observed cracking assuming each mecchanism separately.

3.1.2 Tberma E T1e feedwater (FW) sparger islocated above the core spray piping in the annulus. Some high frequency thermal cycling could occur near the core spray piping because of the turbulent dixing ofthe ecoler feedwater from the FW sparger and the hotter downcomer flow. The magnitude oftemperature cycling isdependent on the feedwater temperature and flow rate, Fatigue initiation duo to thermal cycling isnot only a function ofths temperature difrence between the fedwater and downcomer flow, but also depends on the time duration over which the cycling occum. With this cyclic mechanism, crack initiation is most likely to occur =ar a weld because of the high residual stresses present. Even if fatigue initiation do occur because of rapid thermal cycling, the cracking is likely to be confined to the outside surface ofthe pipe since the thermal stresses attenuate rapidly through the thickmess ofte pipe. Thus, ifthermal fttiguo is the initiation mechanism, extensive fatigue crack growth is unle(fhrther growth can occur by 10SCC since the fatigue crack acts as a crevice). Therefore, th IGSCC Induced growth analysis described below Isa bounding crack growth assessment as shop later in this section.

3-1

MAR 08 '93 8:10PM GENE ENGRG J2455 P. 12 GE-NE-637-0005-0393, Rev. I 3.1.3 Interra a StmAs Corrsionr The core spray line in the Monticello plant where the crack is located Ismade of type 304 stainless sleel. Type 304 stainless steel can snshiz, leading to IOSCC inthe weld HAZ. Local cold wade which could have occurred during thricatlon can also contribute to IGSCC initiation.

'Te Mondcello plant has been operating for sonc time with hydrogen water chemistry (fWC).

Howve the HWC isnot very effective inthe area in which the core spray line is located. The electro chemical potential (ECP) at the core spray line isexpected to be above the threshold (-230 mV), below which full IGSCC protection isassured. Thus, if the observed cracking is of IGSCC origin, some crack growth during futur operation can not be uled out. Aconservative estimate of this potential crack growth and its effect on the structural integrity of the core spray line is discussed in Section 3.2.

3.2 STRUCTURAL INTEGRITY 3.2.1 Summar All idetifiod stresses expected during normal reactor operation were fond to be small.

Based upon a review of these stmses, it is concluded that the structural integrity of the piping with the crack will be maintained during core spray injection, Te stresses considenrd include those due to downoomer flow impingeent load, scismic loading, pressure, weight and thennally induced loads.

Although the normal operating loads by themselves do not result in stresses which arc suffcient to cause IGSCC initiation, the addition ofthe weld residual stresses coupled with local cold work could result inexceeding the initiation threshold. Once initiated, the nornal operating load stresses and the residual stresses could cause subsequent growth of the induced cracks.

In order to deternine the integrity of the core spray line with the crack, a crack arrest evaluation was performed. The stresses due to pipe restraint were also includoe in this evaluation.

Because the applied normal loading ispredominantly displacement controlled, the stresses relax as the crack grows and the compliance (or flexibility) of the pip inceases. The results of the analysis showed that when the crack reaches 180' of the cirumwferoae, the compliance is reduced sufficietly to relieve almost all of the displacement controlled stresses. Consequently, the crack growth is expected to be negligibla or at virtual arrest prior to reaching 180s. (The current through-wall affected area is less than 900 of the piping ciorcuferrene.)

3-2

MAR 08 '93 08:11PM GENE ENGRG J2455 P. 13 GE-NE-637-0005-0393, Rev. 1 3.2.2 Crack Arrest Evaluation Stresses in the core spray piping due to bracket sustraint ass govmed by the applied displacement and the ocenpliance of the pipe, Since the displacement is fixed, the compliance chanpo with crack growth culd lead to crack arrest. This is comparable to crack arrest in a bolt loaded wedopening-loading (WOL) specimen in stress corrosion tests.

Figure 3-1 shows the variation ofoompliance with crack length fbr a pipe subjected to bending Tte compliance was determined using the relationship between strain energy release rato, 0; and the compliance chango per unit area of crack extension dc/dA (Rerance 4). For the cracls in the core spray line, LD is expected to be inthe range 0 < IJD < 40. Figure3-1 shows that the compliance ofthe pipe increases by a ftctor of ten when more than 30% of the pipe is cracked.

Therefore, for the given initial displacement, the stress in the core spray line and the applied stress intensity Actor would decrease by a Actor of ten when more than 30% of the pipe circumference is cracked. Cleady, when the crack Ingth exceeds this value, the restraint stresses become negligible and crack arrest is expected. Therefore, crack arrest is expected before the crack grows to 1800.

3 .2.3 Allowable Flaw Sie Dtemnio Even though the cracks are expected to self arrest at 180' under the sustained displacement controlled loading as discussed in Section 3.2.2, an evaluation was perforned to determine the maximum allowable circumferential through-wall flaw size in the core spray pipe. This analysis will therefore provide an assessment of the safety margin in the pipc due to primary loads such as deadweight, pressure, flow impingement and scismic.

The acceptable tbrough-wall flaw size of the core spray line is determined utilizing the not section colapse formulation of Reference 5. To apply this methodology, primary menmne stresses in the longitudinal direction and primary bending stresses were determined fbr the T-box region ofthe pipe. A finite element model ofth core spray pipe was developed to obtain the stresses due to deadweight, seismic nd reactor vessel downcomer flow impingement on the pipe at the location of interest. The resulting stresses were then combined with tbo stresses due to pressure and core spmy flow loads inorder to get the total stresses acting on the pipe. Stresses due to water hammer loads were considered insipificant and neglected inthis analysis. This isbased on tbe tht that the core spray inlet valve ramps open over a period of twenty seconds upon system 3-3

MAR 08 '93 08: 12PM GENE ENGRG J2455 P.14 I GE-NE-637-0005-0393, Rev. 1 actuation. Additionally, the piping is IB1 of water during actuation because of the presence of the vent hole on the top.of the T-box. Previous analyses have shown that the water hammer loads in the ore spray line were calculated to be less than 20 pounds of axial load on the pipe. Str es due to thermal mismatch were evahmated and found to be insignificant. By applying these resulting primary stresses, it was shown that the core spray pipe can tolerate a crack up to 240 through wall at the T-box location without incipient fMlum.

31-3.1 Analysis and Results Afinite element model of the core spray line configuration was constructed using the ANSYS computer code (Reference 6). A sketch of the finite element model is shown in Figure 3-2.

The following boundary conditions were applied to the model:

Nodes 1, 49, 54; completely fixed Nodes 13,37 : fixed in vessel radial direction to account fbr bohed vessel clamps.

Loads due to the weight of the pipe (including captured water inthe pipe) were applied to the model along with vertical and horizontal seismic loads and reactor vessel downconer flow impingement loads. Calculations of these loads are given in Appendix A. The largest resulting stresses in the region ofthe T-box (nodes 24-26) were used from the finite emeat model results.

These stresses we= then combined with the stresses due to pressure and core spray flow loads.

The resulting total stresses are shown inTable 3.1. Note that loads due to thernal mismatch of the corn spray line and reactor vessel need not be included as they are secondary in nature.

TABLE 3.1 RESULTING PRIMARY STRESSES AT TEE BOX REGION Membrane Stress, Pin 1306 psi Bending Stress, Pb 1606 psi The stresses of Table 3.1 are utilized to determine the acceptable through-wall flaw size based on the methods of Rebrence 5. The acceptable flaw sie isdetenined by requiring a suitable design margin on the critical flaw conditions. 'Te critical flaw size is determined by using limit load concepts. It is assumed that the pipe with a circumferential crack isat the point of 3-4

MAR 08 '93 08:12PM GENE ENGRG J2455 P.15s GE-NE-637-0005-0393, Rv. I incipicut failure when the net section at the crack develops a plastic hinge. Plastic flow is assumed to occur at a critical strees level, q, called the flow stress ofthe material. For ASME Code analysis, oymay be taken as equivalent to 3Sm. Tis results in considerable simplification of the analyals.

Consider a cirumfvrcial crack of length, 1= 2Res, and constraint depth, d,located as shown in Figure 3-3. Inorder to determine the point at which collapse occurs, it is necessary to apply the equations of equilibrium assuming that the cracked section behaves lile a hinge. For this condition, the assumed stress state at the cracked section is as shown in Figure 3-3 where de maxinmm stress is the flow stress of the material, q. Equilibrium of longitudinal forces and moments about the axis gives the following equations:

(Forneural ads locaed such that a + 0 <%)

S= [(a - adit) -(P/g]2 Pb = (2ng/w) (2 sin A- d/t sin ca) where, t pipe thiclkness, inches.

a- crack half-angle as shown in Figure 3-3.

- angle that defines the location of the neutral axis.

Using the stresses of Table 3.1 and a d/t ratio of 1.0 (through-wall flaw), the allowable through-wall crack for which failure by collapse might occur is240'.

3.2.3.2 Temperature Gradient Evanation A (htigue crack growth analysis due to thermal gradients across the core spray piping was conducted using conservative values for the temperature differences expected between the inside and outside surfaces of the core spray pipe. Two events were conservatively considered for this analysis. The first was aHPCI injection inwhich cold water (100*F) from the feedwater spargers impinges on the bt core spray line (550 0F). Conservatively assuming the temperature at the bottom of the core spray piping remains at 550PF, a thermal bending stress results across the pipe cross section. The second event considered was the actuation ofthe core spray system. In this event, cold water (50.F) is injected through the hot core spray line (5501F) which induces a unifhrm thermal mernino stress throughout the pipe cross section. This analysis showed that a 3-5

MAR 08 '93 08:13PM GENE ENGRG J2455 P. 16

. GE-NE-637-0005-0393, Rv, I conservative estimate of-the fatigue cack growth due to HPCI and core spray injections is about 0.14 inches. Details of this analysis are provided in Appendix A. This growth is minimal when compared to the crack growth predicted due to the IGSCC mechanism (predicted to be 1.05 inches). Based on this conservative evaluation, fatigue crack propagation as a result of severe thermal transionts isnegligible when compared to that from IGSCC.

3.2.3.3 Flow Induced Vibration Evaluation A flow induced vibration (FIV) evaluation was conducted considering field measured data 90m a similarly designed core spray system. Inorder to eliminate FIV concerns, it is required that the natura fraquency of the system be greater than three times the vortex shedding frequccy The vortox shedding frequency for this system due to d ' flow was calculated to be 5.1 Hz A natural frequency of 27.5 Hz was obtained for 4 -prayline from the limited fidd data. To asss the potential change in these values as a result of a cracked core spray line, additional analysis was conducted assuming a 180' through-wall crack. The ratio of the compliance of the the uncracked line to the cracked line was calgulated to be 0.649. Givc that the natural frequency is proportional to the square root of the stiffness (stiffnese = inverse of the compliance), this leads to a predicted 20% decrease in the natural fraquency (22 Hz). Since this adjusted value afthe natural frequency still remains greater than three times the vortex shedding lquency (15.2 Hz), the results of this evaluation show that no degradation as a result of FIV Iseipected.

3.3

SUMMARY

AND CONCLUSIONS The potential sources of stress inthe piping resulting from noratal operation and operation during postulated Loss of Coolant Accidents were reviewed. Potential causes of racking, themal fatigue and IOSCC, and the likelihood of crack propagation were also evaluated. It is expected that the crack was caused by IGSCC.

Because of the predominant secondary stresses, the crack can be expected to arrest prior to reaching 180*. An assessment was made to determine the critical flaw aim of the core spray pipe by treating stresses associated with the design loadings as primary stresses and performing a not section collapse evaluation. The results of this evaluation confim that a through-wall crack ofup to 240' around the circumference would not cause pipe filute. This length is much greater thar the maximum estimated crack length at the end of the next fuel cycle (predicted to be 4.7 inches, 100), Therefort, it is concluded that the structural integrity of the piping with a crack will be maintained for all conditions of normal operation for the next operating cycle.

3-6

MAR 08 '93 08:14PM GENE ENGRG J2455 P. 17 GE-NF,637-0005-0393, Rev. 1

• F~Ac1~OM 0 c~aCEI~ ~1A@.J&IPpIIkCt UD' FrMcRE 3-1; Compliance Change, Cracked Pips 3-7

ANSVS 4.4A1 FEB 25 1993 11:43:16 3 PREP? ELEMENTS TVPE NUM TDIS

.. RDZS FORC NV -1' 37 YV --I 54 ?V -1 26 DIST-115.178 0 5 VF -74.88 I' 24 ZF- --29.625 (A AMGZ--58 m J49 13 Figure 3-2:.

MOMTICELLO CORE SPRAY LINE ANALYSIS

Nominal Stress Ic L in the Uncraked Section of Pipe Crack Length - 2Rts Pm+Pb

-I Flow Stress., o CA A)

-6 rin I

I I

~0 I

Pm Stress Distributlon In Pm - Applied Uembrane Stress in Uncrucked Section the Cracked Section at the Point of Collopse Pb - Applied Banding Stress in Uncracked Sctaion FIGURE 3-3 STRESS DISTRIBUTION IN A CRACKED PIPE AT THE POINT OF COLLAPSE

-u I

w

MAR 08 '93 08:15PM GENE ENGRG J2455 P. 20 4&-NE-637-0005-0393, Rev. I 4.0 LOST PARTS ANALYSIS 4.1 INTRODUCION Basd on the struct"ua analysis gien in Section 3, it is aspected that the Monticello Nuclear Generating Plant core spray pipe will not break and consequently, will not result in loose pieces in the reactor. However, an evaluation of the possible consequeces of a potntial loose pic ispresented inthis sectaion.

4.2 LOOSE PIECE DESCRIPTION Since a piece has not been lost, it cannot be uniquely described. Two ditreat types of loose piec are postulated:

I) a section of care spray pipe, and,

2) a small piece of the core spray pipe.

4.3 SAFETY CONCERNS The following saftty concerns are addressed in this analysis:

1) Potential fbr corrosion or other chemical reaction with reactor materials.

2) Potential for fuel bundle flow blockage and subsequent fuel damage.

3) Potential fbr interference with control rod opeation.

4) Potential for damage to the reactor intemals.

4.4 EVALUATION The above safety concerms for the postulated loose pieces are addressed in this section.

The effect ofthese concerns on saf reactor operation is also addressed.

4.4.1 General Df4cription Since the core spray pipe with the crack is inthe annular region of the reactor pressure vessel, this evaluation assumes that any potential loose piece generated fron the core spray pipe will most likely sink into the downcorner region.

4-1

MAR 08 '93 08:16PM GENE ENGRG J2455 P. 21 GE-NE-637-0005-0393, Rev. I For a loose part to reach, and potentially block the inlet of a fbel assembly (Figure 4-1), It would have to be carried into the lower plnum. To accomplish thi, it would have to be carried by the recirculation flow through the jet punp nozzle into the lower plenum, then make a 18O turn and be caraed upward to the Iel assembly inlet orifices.

For a piece ofthe core spray pipe to reach a control rod it must first migrate to the lower plenum, pass through the fuel inlet orifice, and traverse the fuel bundle. Then, it must either ll through the restrictive passage between two fuel channels, or fill through an opening between the peripheral bundles and the core shroud. Both of these potential paths am unlikely.

The core spray pipe is fibicated from Typo-304 grade stainless steel and all parts of the core spray pipe are designed for in-reactor service. Consequently, there is no postulated loos part that will cause any corrosion or other chemical reaction with any reactor material.

4,4.2 Postulated Loose Pieces 4.4,2.1 Core Spray Pipe The core spray pipe is 5 inch Schedule 40S pipe. In order to gmerate a loose piece of pips, a minimum of two through-wall cracks would have to propagate 3600 around the pipe.

Ifa pipe segment were postulated to break off it would sink into the downcomer region.

Since it cannot fit through the jet pump, it cannot enter the lower plenum, and therefore will not cause any flow blockage at the fuel inlet orifice, Since it is too large to fit between fuel channels, it cannot cause any interference with control rod operations. Nevertheless, due to the slow propagation rate of potential cracks, and based on previous experience with cracks incore spray spargcs, it isjudged that a piece of the piping will not break off and bcome lose.

4.41.2 Small Picoes In order to generate small pieces of the core spray pipe, both longitudinal and aircubrntial through-wall cracking must occur. A small piece could then sink, be carried into the down caer annulus, pass through thojet pump and enter the lower plenum. A piece that antered the lower plc=u would probably be driven by the jet pump flow to the bottom of the retor pressure vessel where it would be expected to remain. However, a small picc < 0.4 inches 4-2

MAR 08 '93 08:16PM GENE ENGRG J2455 P. 22 GE-NE-637.0005-0393, Rev. I could be carried by the flow up to the thel inlet orifices. The orifice aims Inthe Monticello Plant vary from approximately 1.4 to 2.1 inches indiameter (Pdrance Figure 4-2).

Given the dimensions, the piece would pass through the inlet orifices and be trapped at the lower tic plate grid and cause some bundle flow blockage. However, the flow blockage is much less than that nquired to initiate critical boiling transition in the bundle. Multiple pieces migrating to the same bundle may result in critical flow blockage, but the probability for such an occurr e is extremely low.

It is also very unlikely that a small piece could lift and migrate frm the lower plenum through the fhel bundle and fall into the control rod guide tube. In order to do this, the piece would have to be so small that it pass through all the bundle spacers and out through the top of the bundle. Such a small pcwould not present any potential for control rod interirnce.

Figure 4-3 shows a typical unit cell of four fuel assemblies and one control rod. The control rod moves in the gap betwcca the fuel channels. Thom is a small possibility that a piece small enough to fit inthe gap between the channel wall and control blade could sink and pass through the cavity betweenfthe control blade and the fuel support casting and migrate into the control rod guide tube. Should this happen the piece will most likely come to rest on the top ofthe velocity limiter where it is expected to remain and move only with the movement of the velocity limiter as the control rod isjnserted or withdrawn. If the piece is small enough to pass between the velocity limiter and the guide tube wall it will most likely sink and come to rest at to bottom of the guide tube. Due to the hardware geometry of the control blade drive mechanism It ishighly unlikely tha any piece wouM be smail enough to migrate into the control blade drive system.

Thus, any potential small piece which migrates to the control rod guide tube is not expected to pose any concen for potential interference with control rod operation.

One of the licensing bases of the reactor isthat with the highest worth control rod fully withdrawn tbo reactor can be brought to cold shutdown. Thus, unacceptable control rod interbef would require multiple precisely sized pieces interfering uinmtaneously with control rods that am in close proshity to each other. The probability of this isjudged to be insignificant.

4-3

MAR 08 '93 08:17PM GENE ENGRG J2455 P. 23 GE-NE-637-0005-0393, Rev. I

4.5 CONCLUSION

S The core spray pipe at Monticalo Plant is expected to remain intact; therefore, it is highly unlikely that pieces of the com spmy pipe will break oft From the above evaluation it is concluded that the probability for unacceptable corrosion or other chemical reaction due to looso pieces is zero. The potential for =anccptable low blockage or other damage to the fhel assemblis is ngligible. The potential fbr iuacceptable control rod intedbrance is negfliibly small, Thorfore, it is concluded that there is no safety concern posed by any postulated loose parts.

4-4

MAR 08 '93 08:18PM GENE ENGRG J2455 P.24 GE-NE-637-0005-0393, Rev. 1 FIGuRE 4-1 LOOSE PIECE POTtNTIAL UPWARD FL,0j PATS 4-5

MAR 08 '93 08: 18PM GENE ENGRG J2455 P.25

• .E-NE-637-0005-0393, Rev. 1 CONTROL ROD BLADE IN. CORE GUIDE TUBE FIGURE 4-2 DRIFTICED FUEL SUPPORT 4-6

MAR 8 '93 08:18PM GENE ENGRG J2455 P. 26 GE-NE-637-0005-0393, Rev. 1 FUEL ASSEMBLIES

& CONTROL ROD MODULE' vOP FUEL, UIDE 2.CMANNEL.

PAS"I'NER

&UPPER TIE PLATS 4.gXPANUION

• PRING

'lLOCKING TAS 0.CMANNEL 7.CONTROL ROD B.FUEL ROD ii.CORE g.SPACUlt PLATE 11.LOWER

,ne PLATE 12,PtJUL SUPPORT 13.FUEL PELLETS 14.END PLUS ICMANNEL SPACER 1g5.PI.NUM

$PF41NG GENERAL ELECTRIC FIGURE 4-3: UEL, ASSEMBLIES AND C ROD MODULE 4-7

MAR 08 '93 08:19PM GENE ENGRG J2455 P. 27 GE-NE-637-0005-0393, Rev. 1 S. IMPACT ON ECCS ANALYSIS 5.1 IMPACr OF CORE SPRAY LINE LEAK ON SAFER/OESTR-LOCA ANALYSIS For Monticello Nucear Generating Station anly two single filure camlidates am potentially limiting fbr ECCS system performance following a LOCA. These are assoclated with the limiting break which is a recriulation pipe break (Reference 2). These limiting cases are A. Battery Failure. This postulated failure leaves I Core Spray + 2 Lw Pressure Coolant Injection (LPCI) + the Automatic Depressurization System (ADS) operable; B. LPCI njection Valve Failure (LPCI IV). This postulated fhilure leaves 2 Core Spray +High Pressure Coolant Jijection (HPCI) + the ADS operable.

The CS flow rate assumed for each loop in the above referenced Monticallo SAFER oauations was 2700 gpm (Reference 3, p. 5-2). This value reflects the CS flow rate which is asumed to actually inject inside the core shroud, and this is 320 gpm less than the expected system performance af3020 gpm per core spray system (Refbrance 3, p. 5-4). This margin of 320 gpm in the assuned CS flow in the SAFER evaluation is much greater than the total estimated leak flow of 24 pm fkom a 180s crack inthe CS line plus the 13 gpm leak through the vent hole located in the top of the T-box. Thus, the SAFER/OESTR-LOCA analysis for both the nominal and Appendix K assumptions as documented in Refrences 2 and 3 covers, with significant margin, the estimated leak through the CS line crack.

5.2 IMPACT OF POSTULATED FAILURE OF CORE SPRAY LINE AT CRACK LOCATION Additional analyses for the above limiting single f ropaseqwere perfbrmed considering the unlikly failure of the CS line at the crack location. With t postulated thilure of the core spray line all core spray flow for this loop was assumed to drain into the downcomer region. These analyses were performed far the recirculation suction line break with nominal assumptions (Reference 3, Table 3-1). The Peak Cladding Temperature (PC ftbese cases were calculated to bless than 1330OF inthe case of battery failure and 1650aF in tncase of LPCI injectin valve failure. The LPCI injection valve failure istbc more limiting ca beuse it has amuch lower ECCS makeup flow rate than the battery failure. Other po UIures are not specifically fulaed 5.1

MAR 08 '93 08:20PM GENE ENGRG J2455 P. 28 GE*NE-637)005-0393, Rev. 1 considered because they all result in more ECCS capacity than the above assumed failmes (Refeence 2,Table 4-2). For smaller breaks with the additional CS line failmr, the PCT is expected to be lower than the above caculatod values since the core inventory loss will be less and hence the ftm to reflood will be less compared to to larte break senarios. In addition, at some break sizes tbe CS makeup flow into the aniulus with this assumed failure will exceed the break flow rate, allowing the dowacomer to refill and the CS water iqjected into the downcomer region to reach the core via thejet pumps/lower plenu.

5-2

MAR 08 '93 08:20PM GENE ENGRG J2455 P. 29 GE-NE-637-0005-0393, Rev. I

6.0 REFERENCES

1. USNRC IEBulletin No. 80-13, Craidng in Core Spray Sparger, May 12, 1980

2. Haman, D.A., A.D. Unruh, W.M. Wan& M tello,5Efl L QES Lw Coolat Acdent Anal Table 4-2 and Table 51, NEDC*31786P, Class I December 1990.

3. Haman, D.A., Monticello Nuclear Gmrating Plant SAFER/GESTR-LOCA Analysis Bas Documetan. GE-NE-187-02.0392, March 1992.

4. Kiss, E,, Heald, DA and Hale, D.A., Low Cyc Vatiga of Poy Pipi GEAP-10133, Jamury 1970.

5. Ranganath, S. and Mobta, LS., "Engineering Methods for the Assessment of Ductile Fracture Margin in Nuclear Power Plant Piping," Elastic-Plastic Fracture: Second Symposium, Volume II - Fracture Resistance Curves and Engineering Applications, ASTM STP 803, CF. Shih and J.P. Oudas, Eds., American Society for Testing and Materials, 1983, pp.11-309 - II-330.

6. DeSalvo, G.J., Ph.D. and Swanson, JA., Ph.D., ANSYS Enineerina Analysis System User's Manual Revision 4.1 Swanson Analysis Systems, Inc., Houston, PA, March 1, 1983.

7. Paris, P.C. and Sib, G.C., "Stress Analysis of Cracks," Fracture Toughness testing and its Applcations, ASTM STP 381, American Society for Testing and Materials, 1965, pp. 30 - 83.

8. Rooke, D.P. and Cartwright, D.J., om dium of Stress Intensity Egr, Her Majesty's Stationary Office, The Hillingdon Press, 1976.

9. Hale, D.A., Yuen, J.L., and Gerbor, T.L., Faidue Crack Growth In Pipni and RPV StIs in Siulated BWR Water Enom e GEAP-24098, January 1978.

6-1

MAR 08 '93 08:21PM GENE ENGRG J2455 P. 30 GB-NFA37-0005-0393, Rev. 1 STRUCTURAL ANALYSIS OF TE MONTICELLO PLANT CORE SPRAY PIPE A-1

MAR 08 '93 08:21PM GENE.ENGRG J2455 P. 31 I GE-NE-637-0005-0393, Rev. I A.1: CRACK GRO DUE TO IOSCC The stress rsults given inTable 3-1 of Section 3.2.3.1 ofthis report are developed in this Section othe Appendiz. The stresses woer determined by applying dead weight, seismic and flow impingement loads to the finite element model developed for ft core spray line (see Fgure 3-2).

The calculation of these loads isgiven here. Also incinded are the calculations of the stresses due to pressure and.flow loads as well as the total combined primazy streses given in Table 3.1.

Wgbt ofL An equivalat density was input to the ANSYS unite clemet model to incude both the weights of the pipe and capfured water. This equivalent density is calculated below Metal density - 0.2879 lb/in Water density 62,4 lb/ftV - 0,0361 lb/in2 Pipc sizeo5 inch schedule 40S Stainless Steel OD 5563", t= 0.258", ID= 5.047" Metal area ; (r/4) (5.5632 - 5.0472) = 4.3 in2 Water area= (x/4) (5.0472)= 20.0 ir?

Metal weight (0.2879 lb/i) (4.3 In)= 1.238 lin Water weigt (0.0361 lb/inP) (20.0 in - 0.722 lb/in Adjusted density.- (total weight)/(metal area)

- (1.238 + 0.722)/4.3

-0.456 Iblin u 0.0012 slugaina A-2

MAR 08 '93 08:22PM GENE ENGRG J2455 P. 32 GE-NE-637-0005-0393, Rev. 1 Imninsemen Land (90" N ction of low):

F= PA = pVfDIg D - 5.563"112 - 0.464 ft Asune downoomer flow, V - 5 fthes (coniervative)

For wktr, p - 62.4 Idft P/L = pV2D/g = (62.4)(52)(0.464)132.2 = 22.5 b/ift 1.87 lWin

'Its nodes of the fite eleenat model are spaced 50 apart. Thus, the following load will be applied to all nodes omprising the horizontal arms of the core spray line (nodes 10 - 40):

Node spacing - RO - (99.25") (5) (z/1800) = 8.66" Load per node = (1.87 lb/in) (8.66") = 16,2 lb Seismic Lads From the Monticallo FSAR, the vertical accelerationfor a Operating Basis Earthquakc (OBE) is 0,04,g and the horizontal acceleration is 0.13 g. In order to provide conservative and bounding results, the seismic coefficients were doubled to obtain results for a Safe Shutdown Earthquake (SSE) and a safety factor of 1.5 was added. Based on a review ofthose coeients, the following were selected for use inthis analysis:

Vertical= 0.12 g Horizontal = 0.39 g The following accelerations were therefore applied to the finite element model:

Total vertical accelerAtion - Weight + Seismic

- L.0 g + 0.12 g.

= 1.12 g=4323 io/secF Total horizontal acceleration - 0.39 g= 130.5 in/se A-3

MAR 08 '93 08:23PM GENE ENGRG 32455 P.33 GE-NE-637-0005-0393, Rev, 1 The horizontal acceleration was appliad in both the X and Y direcdans of the model mch that the resultant was 150.5 in/sc 2 . Thus, the horizontal accderaton applied to both directions of tb model was:

150.5 / = 106.4 in / soo Pressure/flow Load&

Assumed flow - (Rated fiow)*2 - 3020 pm *2 - 6040 gpm through the core spray line Q- (6040 Spm) (1 im/60 sec) (1 ftl/7.48 al= 13.49 f/sco F pQ/2(V) - pQ/2(Q/A) = (62.4) (13.492)/[(32.2X/4)(5.047/12)zj = 2,538 lb AP - 150 psi @ 6040 gpm StrSp Dueto Loads-:

Pressur:v .

crp - (150) (n/4) (5.0472) / [(iT/4) (5.5632 - 5.0472)]

- 698 psi Flow Lon cr = F/A= 2,538/(al4) (5.5623 - 5.047)]

= 590 psi A-4

MAR 08 '93 08:23PM GENE ENGRG J2455 P. 34 OB-NE-637-0005-0393, Rev. I Imingement. isht and seismic:

Stress due to the above loading was determined using tbo Onite dement model of the internal orie spray piping, From the ANSYS results, the maximum of the stresses at nodes 24-26 were used since they are in the area of the cracks. The maximm stresses are given below:

cUD = Axial Stres - 17.9 psi c=- Beding tss - 1602 psi M - Torsioal Streas= 106.6 psi qTAU - Shear Stns= 142.4 psi Combining all of the prinry stresses, the following values amtobtained*

Primary Membrane = P F + ct +oDR

= 590 + 698 + 11.9

= 1306 psi Primary Bending - P6 = ci% > + VA

=4160 +106.6 = 1,606 psi The shear sts TAU, is small and its effoct is negligible so it is not included.

A-5

MAR 08 '93 08:24PM GENE ENGRG J2455 P. 35 GE-NE-637-0005-0393, Rev. I A.2: FATIGUE CRACK GROWTH Tbc teults ofthe atigue cruack growth analysis due to thenmal gmdlents across the care spray piping peseated in Section 3.2.3.2 are developed in this section of the Appendx The irst event considered is a HPCI injection in which cold water (IOO*F) ftkm the feedwater spargers impinges on the hat core spray line (550*F). Conservatively assuming the tmperature at the bottomn of the core spray piping remains at 550F, a therna bending stress is applied across the pipe cross sectlan. This condition is approximated by the following sketch*

Tout = 100 'F Vout = 5. ftlaec Tin= 550 F Qin - 7 pm To=550oFZ The second event couisidered isthe actuation ofthe core spray system. In this event, cold water (50"F) is injected through tho hot core spray line (550"F) which induces a unifbrm thermal membrano stess throughout the the pipe cross section. This condition is approximated by the following sketch Tout = 550 T Vout = 5. ft/sec Tin 50 F Qin 2250 Spm A-6

MAR 08 '93 08:24PM GENE ENGRG J2455 P. 36 I

GE-NE..637-0005,0393, Rev, I The heat trmnsfbt coeffcients for each event were computed usizng classical heat trnsfr methods.

For the HICI injection event, the outer heat transfer coefficient is calculated assuming a cylinder in turbuaent cross flow. The heat transfer coefficient for a cross fow luid tCpcrature of 1000 F is, hout - 992 thr WOF The inner surface heat transbr coefficient assuming fully developed laminar flow (@550*F) due to lekage through the vent hole is calculated to be, hin =2.83 BAr ft OF For the core spray injection event, the the outer heat transfer coefficient is again calculated assuming a cylinder in turbulent cross flow. Thiz heat transr coefficient for a cross flow fluid temperature of.550-F-is, bout 1497 Bt/hr fl"F The inner surface heat transfer coefficient is calculated assuming turbulentflow through a cyclinder (@507, bin - 4251 Btulhr f OF Using the heat transfer coefficicats calculated, the temperature dhnge across the pipe was calculated assuming one-dimensional, steady state conduction. For the IPCI infection event, the inner and outer tnempatures of the core spray piping at the top of the pipe was calculated to be, Ti s 1000P T9 s 1001F Assuming the temperature of t bottom of the pipe remains at 5500, AT 550 - 1000 = 4500 I

A-7

MAR 08 '93 08:25PM GENE ENGRG J2455 P.37 OE-NE-6370005-0393, Rev. I For the core spray o icto ent, the imer and outer tnperae of the wall were calculated to be, T . 50*F Tom

. 55OP Eased en these results, the Aveage change intcmperature acrss the wall is,

. AT (500 +O/2= 250 The thermal stesses for both events were calcula using the expression, EaA The thenal bending stress aue to the HPCI injectio is calculated to be, a*76.9 Jl The thermal membrane stress due to the core spray iection Iscalculated to to,

• o=42.7 lie Stmas intensity fctor were alculwed using the thermal stresses calculated for ach evvnt. For the HPCI injection, the expmssion used for K1 is (Reforunco 7) o 1 v) where e- Bendin stress (ksi) = 76.9 ksi A-3

'1

MAR 08 '93 08:25PM GENE ENGRG J2455 P. 38 GE-NE-637-0005-0393, Rev. I a - Crack kngth (in)= 2.35" v - Poissos ratio -. 287

. 76.9,14.35(1+0.287) f (3+0.287)

For the core spray infection, the expression used fbr K, is (Reference 8):

K, = O4 (Gm -G) a = Bending stress (ksi) = 42.7 ksi a- Crack length (in) - 2,35" Om -Membrane stres contribution - 1.62 Gb Bending stress contribution= -. 14 IC,=42.7;W75 (1.62 - (-0,14)) - 202 ksit/

Cack gm~latie Using the values of K, calculated above, the tigue crack growth rate was detrmined uAng expedmentally determined curves for 304 tainless steel in a simulated BWR environment (Reference 9). Conservatively assuming K = 0, A is equal to K . From Figure 4-1. of Reference 9, the ack growth rates for the events considered are:

Aa HIjection- -= 0.0054 in / cycle Core Spray Injection -= 0.135 in/cycle An A-9

- STRUCTUBAL INTEGRITY ASSOCIATES, INC.

3150 Almaden Expressway Fossil Plant Operations Suite 145 66 South Miller Road San Jose, CA 95118 March 3, 1993 Suite 206 (408) 978-8200 AJG-93-01 5/PCR-93-032 Akron, Ohio 44333 FAX: (408) 978-8964 (216) 864-8886 or (408) 978-0438 FAX: (216) 864-5705 Mr. Peter Kissinger Northern States Power Company Monticello Nuclear Generating Plant 2807 West Hwy 75 Monticello, Minnesota 55362-9637

Subject:

Third Party Review of General Electric Evaluation Approach for Continued Operation of Cracked Core Spray Pipe at Monticello Nuclear Generating Plant

Dear Pete:

.performedAs authorized by you on February 19, 1993, Structural Integrity Associates, Inc. (SI) has a third party review of the General Electric Nuclear Energy Division evaluation approach used to justify the technical suitability of the subject core spray piping for continued operation for one cycle at Monticello without repair. The SI review team consisted of Peter Riccardella, Anthony Giannuzzi and Hal Gustin.

Our review consisted of:

1. a review of the of the drawings supplied by you of the tee box and core spray arms in the vicinity of the cracking,

2. a review of the video tape (also supplied by you) illustrating the cracking location and the surrounding vicinity, and

3. participating in a meeting with General Electric personnel to review their approach to the stress and fracture mechanics analyses used to justify continued operation of the core spray system without repair for one additional cycle.

Note that our review concentrated on GE's overall technical approach and methodology, but did not include detailed verification or independent calculations. The following paragraphs describe the results and conclusions of our review.

Based upon the crack location in the heat affected zone of the pipe to tee box weld, as observed in the video tape, plus the tee pipe cover plate weld in the immediate vicinity of this attachment weld, the 304 stainless steel material at the crack location is expected 1983 Celebrating 10 Years of Engineering Excellence 1993

to be highly sensitized and in a state of high welding residual stress. Also recognizing that the BWR coolant in this region is highly oxidizing, both on the OD due to the downcomer flow and on the ID due to stagnant conditions in the core spray line, it is concluded that the most likely failure mechanism is intergranular stress corrosion cracking (IGSCC) of the weld sensitized material. Furthermore, it is highly probable, although not conclusively so, that the cracking initiated on the outside surface of the pipe, as two separate cracks in separate arc strikes observed in the videotape. Weld solidification puddles associated with arc strikes are highly susceptible sources for IGSCC initiation because of additional sensitization, residual stresses and altered physical and mechanical properties. A likely scenario is that two separate cracks initiated in the two arc strikes on the OD of the pipe, and grew towards one another by IGSCC in the direction of greatest primary bending stress and largest expected residual stress.

The GE analytical approach in justifying continued operation of the core spray system without repair for an addition operating cycle included the following stress and fracture mechanics considerations:

1. Membrane and bending stress were determined using an ANSYS beam model due to differential thermal expansion between the piping and vessel, downcomer flow forces, and anticipated pressure drop in the core spray piping and spargers under core spray operation. Calculated stress due to these conditions were quite low (approximately 1.5 ksi membrane plus 1.5 ksi bending).

2. The critical crack length was determined under these loads for a through-wall crack and was found to be on the order of 235* of the circumference.

3. The currently observed crack length is approximately 3.5", or 66* of the circumference, indicating considerable margin to critical crack length (-9"). GE also added approximately 1"to the observed crack length in their evaluation to allow for possible greater length on the inside surface, since ultrasonic examination (UT) of the crack had not yet been performed. (Subsequent UT indicated this to be conservative.)

4. Potential leakage through the crack under core spray operation was estimated, assuming the crack to be a .01" wide slot 1800 of circumference, or -9"long. The resulting leakage flow rate is on the order of 23 gpm, which is negligible compared to overall core spray delivery capacity of 4000 to 6000 gpm. The 23 gpm value seems to be a very conservative estimate for a crack of this length based on SI's independent experience in computing leakage rates through cracks.

5. A conservatively predicted crack growth rate during the next fuel cycle would produce only an additional 1 to 1.5" of crack length. The crack growth rate was assumed to be an upper bound of 4 x 10- in/hr, independent of stress intensity, during the entire subsequent operating cycle. This estimate of crack growth rate is extremely conservative as illustrated by examining the NRC recommended curve in Figure 1, taken from NUREG-0313, Rev. 2.

ATETY CASSOCITS,INC.

7 Although Structural Integrity Associates personnel have not performed detailed independent check calculations for the above items due to time constraints, the technical approach appears to us to be quite sound. and the results appear reasonable, based on our experience with these types of evaluations. In our view, the GE analysis demonstrates a justifiable basis for operation of the core spray system for one additional cycle at Monticello without repair.

If you have any questions regarding our review comments, or if we can be of any further service, please feel free to call. Thank you for this opportunity to be of service to you.

Very truly yours, Giannuzzi

/ms attachment cc: B. Day (Northern States Power)

ASSOCIATES INC.

• I00-NRC CURVE Iin./yr 10

-5'

~io c:-t 100C W0.2(* we 02152Fz4 a sensitized (C/2 at 1150*F/24 h C.D C* 0.2 ppmGE0; seistize KIatXTACMXI

£ 6 0 2"o; I.~ttllzv at IIZCF/24 h DT (1150F/2 mi C.EMED w A"L

• 010 HClI* .2 JO 0 0.2 pa (EV GECA O -GE Pop.2 Oe.; sensitized pscz; iestze byat wieldings; 02; severtly snsitized (useZrn0s a CI ) 0.04 n./yrLTS CL)8 ppe at 932*F/24 h (ski)

CZ- %rsnsitizM at (12$Z*F/10 mift)"

O s* (92*F/24

. ar . sesxa k) (Ef n* 0.944 C/M2 nt 8Pp" 02; sensitized at (125t*F/10 minr)

• (932*F/24 h) (Ert

• 4 C/cs )

f

• 0.1 5. I

• 0.94 3"p 0 z; sensitized at (12l*F/10 min)

• (942*F/257 %) (EMt

• 15 C/cs )

3 ppE C ; seusitized at (129t*F/10mOin)

- (842F/25 k) (EPt

• 15 C/m2Z RMX If *0.1 Mt. a

• 0.94&N RACK R;ssitized atDARAT (EPM* 20 C/c) 0.00 KZ a

• 0.95 8 pX sensitized at Int*F/14 (EPR *20 C/cs f *0.08 KZ.

I

• 0.95 0 10 20 30 40 50 60 70 STRESS INTENSITY, K(ksiv-.)

lie Figure 1 CRACK GROWTH RATE DATA

-A.6