ML20217E543

Safety Evaluation Re Flaw Evaluation of Core Spray Internal Piping for Plant,Unit 3

ML20217E543
Person / Time
Site:	Browns Ferry
Issue date:	09/22/1997
From:	NRC (Affiliation Not Assigned)
To:
Shared Package
ML20217E534	List: ML20217E530 ML20217E543
References
NUDOCS 9710070079
Download: ML20217E543 (9)

Text

_ .-_

e e f

gj * ,

UNITED STATES s* }

2 NUCLEAR REGULATORY COMMISSION WASHINGTON, D.C. 30MHo01 SAFETY EVALUATION BY THE OFFICE OF NUCLEAR REACTOR REGULATION RELATED TO THE FLAW EVALVATION OF THE CORE SPRAY INTERNAL PIPING TENNESSEE VALLEY AUTHORITY BROWNS FERRY NUCLEAR POWER STATION. UNIT 3 DOCKET NUMBER 50-296

1.0 INTRODUCTION

During the 1997 Browns Ferry Nuclear Plant (BFN) Unit 3 refueling outage, the Tennessee Valley Authority (the licensee) inspected core spray piping within the reactor vessel pursuant to commitments associated with IE Bulletin 80-13.

"Cra t ng In Core Spray Spargers," dated May 12. 1980.

Crack-like indications were identified at two locations in the core spray (internal) C downcomer piping. The internal core spray piping is a 6-inch diameter, schedule 40 pipe composed of Type 304 stainless steel material.

The two flawed locations are on the collar-to-shroud weld (P8b) and the 31pe to elbow weld (P4d). The collar-to shroud weld is not associated wit 1 the piping pressure boundary but l

rather is a separate sleeve that attaches the core spray aiping to the shroud.

The functions of the collar are to provide stiffness to tie piping to resist flow-induced vibration and to prevent leakage from the inside to the outside of the core shroud at the location where the core spray piping enters the shroud. The flaws on weld P8b were found using an automated ultrasonic testing (UT) system available on a remote-operated vehicle. For weld P8b, four separate circumferential indications were identified.

of these indications as measured by UT were 19.8 inches (284').The Two total minor length indications were visually observed on weld P4d, The lengths of these indications in weld P4d were 4 inch and 4 inch, respectively. The locations and characteristics intergranular of these indications stress corrosion indicated that the cause was cracking (IGSCC).

core spray piping or sparger were identified. No other indications on the By letter dated March 7. 1997, the licensee submitted to the U.S. Nuclear Regulatory Commission (NRC) its flaw evaluation of the core spray internal piping. Additional information regarding the flaw evaluation was provided by the licensee on March 9 and 10, 1997. The licensee concluded that the indications on welds P8b and P4d are acceptable for continued operation during the next fuel cycle without repair. By letter dated June 11, 1997, the licensee provided NRC plant-specific system and risk assessments assuming the complete failure of two welds (P8b and P9) in the core spray piping. The results of the licensee's assessments showed that the increase in the limiting peak clad temperature (PCT) from the bounding Loss of Coolant Accident (LOCA) analysis case is well below the acceptable PCT limit specified in 10 CFR 50.46(b)(1) and that the increase in core damage frequency is not risk significant. The staff's evaluation and conclusions are provided below.

ENCLOSURE okU ooN ob$o$$96 p PDR j

+

7 On March 11, 1997, the NRC informed the licensee that no technical issues had been outage.identified that would prevent restart of BFN Unit 3 from its refueling This safety evaluation provides details of the NRC staff's assessment  ;

to support restart of BFN Unit 3 and its operation for the current fuel cycle.

2.0 EVALUATION 2.1 Scone of Insoection The licensee stated that the performance of the inspection of core spray  !

internal piping and spargers during the last outage followed the guidelines provided in the Boiling Water Reactor Vessel and Internals Project (BWRVIP) document. " Core Spray Internals Inspection and Flaw Evaluation Guidelines  !

(BWRVIP-18)." dated July 26, 1996. Using the BWRVIP-18 guidance for the inspection and scope of the core spray internal downcomer piping, a baseline inspection of all welds would be would not be accessible using UT. performed using UT. For those welds that the licensee would follow the BWRVIP-18 guidance for enhanced visual examination (EVT-1) which is capable of achieving a resolution of 0.0005 inches (0.0127 mm). Because the inspection scope proposed in BWRVIP-18 focused on areas of the core spray piping which are more likely to experience-intergranular-stress corrosion cracking and the proposed inspectioh methods for the core spray internal downcomer piping are more stringent than those recommended in IEB 80-13, the staff finds that the scope and inspection methods used for the inspection of the core spray internal piping at BFN Unit 3 are acceptable for this outage.

It should be noted that the NRC staff is 3resently reviewing the acceptability of using BWRVIP-18 generically for all BWRs. While the staff has not identified any major deficiencies in the BWRVIP's technical assessment at this time, neither has the staff made a determination as to its generic acceptability. Therefore. the-licensee should be aware that if concerns are identified during the staff's generic review of BWRVIP-18. and if the licensee intends to follow the BWRVIP-18 guidance in the future the NRC staff may request basis.

that the licensee also address these concerns o,n a plant-specific 2.2 Flaw Evaluation The results of the inspections of the core spray piping welds indicate that the identified cracks in the collar-to-shroud weld (P8b) are of IGSCC origin.

-Because IGSCC is known to be initiated from the piping inside surface, visual examination can only find flaws that are through-wall. Therefore, the licensee's UT examination of the flaws provides reasonable assurance that the flaws have been adequately detected and sized.

Because the flaws are located where the collar is welded to the shroud wall, it was inconclusive whether the flaws were on the shroud side or the collar side. The licensee was not able to-detect the flaw using visual ins)ection (EVT-1) of the weld. The licensee, therefore. . conservatively assumed t1e flaws to be through-wall.

In its crack gr rate of 5.0x10',0wth calculation, the licensee used a bounding crack in/hr as recommended by the BWRVIP in the document. BWRgrowth Core Spray Internals Inspection and Flaw Evaluation Guidelines." dated June 1996 (General Electric Report No. GE-NE-B13-01805-21). The licensee stated that the crack growth rate is conservative because the reactor water conductivity

at BFN Unit 3 has been better than the BWR fleet average. The coolant conductivity at BFN Unit 3 has averaged 0.09 US/cm over the past cycle compared to a conductivity of approximately 0.303 pS/cm during the first 5 years of operation. The staff finds that the bounding value used by the licensee for its crack growth rate is acceptable.

By using the bounding crack growth rate, the licensee calculated the remaining ligament for weld P8b at the end of the next fuel cycle of 18 months to be 2.5 inches. With this calculated remaining ligament, this weld would not have a momentto ligament load actcarrying capability, The licensee assumed the remaining like a hin support only axial loads (ge i.e.,and theevaluated the capability moment carrying capacityof of thethe weld to and weld 31 ping at that location were released in the piping analytical model).

Evaluated in this manner, the weld would possess a safety factor greater than 5.0 for axial loads compared to the required safety factor of 1.4 for the emergency / faulted plant condition. However, because the calculations did not clearly demonstrate that the remaining ligament would be able to withstand the moments induced by required load combinations, the stafi finds that the structural integrity of weld P8b is not assured for the next fuel cycle. The licensee also performed a flaw evaluation by assuming that the flaws at weld PBb are located in the shroud side and propagated in the direction of the wall l

thickness'. Based on these assumptions, the licensee's evaluation showed that the structural integrity of weld P8b will be maintained during the next cycle of o)eration. However, the staff Lelieves that the propagation of the flaws in t1e shroud would follow the heat-affected zone adjacent to the collar to shroud weld and, therefore, an evaluation that assumes crack growth not following the heat-affected zone is not credible. If the flaws propagated in that direction, the structural integrity of weld P8b might not be maintained during the next operating cycle.

To evaluate the acce)tability of observed flaws, the licensee had prepared a flaw evaluation hand)ook for BFN Units 2 and 3 that provided a set of allowable flaw lengths at various core spray piping locations and leak rate calculations for postulated through wall indications. The method used in the analysis was based on a finite-element analysis of the core saray piping and sparger using the ANSYS computer program to determine the mem)rane and bending stresses from the postulated loadings. Applying the stresses at several key locations in the piping system, the licensee used limit load methods of paragraph IWB-3640 of the American Society of Mechanical Engineers (ASME)

Boiler and Pressure vessel Code.Section XI as a guide to determine the allowable flaw lengths. The Handbook also provided a leak rate calculation methodology and results of leak rate calculations.

Weld P9 is a full penetration girth weld. This weld is not accessible for inspection because it is hidden inside the collar. When the collar welds (P8a and P8b) are intact, the integrity of weld P9 is not essential for the performance of the designed function of the core spray piping system.

Therefore, for a core spray piping system with no sign of degradation, there is no concern regarding the integrity of weld P9 and inspection of this weld

-is not required.

In the licensee's core spray piping structural analysis, it is assumed that the structural integrity of weld P9 is maintained. This assumption is derived from the licensee's evaluation of weld P9 which stated that the subject weld

does not have a true crevice condition and that its susceptibility to IGSCC is similar to other girth butt welds in the core spray pi Since no cracking was found in the other girth butt welds with the exce) ping.

tion of minor cracking found on weld P4d the licensee concludes that the 11telihood of finding significant cracking at weld P9 is expected to be low. Furthermore, weld P9 is a nonflux weld with high flaw tolerances.

The licensee's flaw evaluation has shown that this weld can withstand a through wall flaw greater than 9.8 inches in length. However, the licensee has taken exception to BWRVIP-18 classification of the weld P9 as creviced and highly susceptible to cracking.

The licensee interior contends core spray piping.that weld P9 issimilar to other butt welds in the from grinding of the core shroad to facilitate re-installation of the collarThe lic (weld P8b) is partly related to the cracking of the P8b weld. However, the licensee did not address the fact that P9 weld was also ground. Therefore, since weld P9 was not inspected, its integrity is in doubt and it may not be appropriate to assume that this portion of the core spray s,ystem would be functional during this fuel cycle. Therefore, on May 28, 1997, the staff requested the licensee to assess the risk associated with a potential failure of welds P8b and P9.

In a res weld P9.ponse to the staff's concern regarding the structural integrity of the licensee performed plant-saecific system and risk assessments assuming the complete failure of welds of June 10. 1997. In this scenario, it 28b is and P9. as describe in its letter alping would not sup)1y flow to the vessel. postulated that The licensee this section performed the of 30unding cases for tie loss-of-coolant accident (LOCA). using the best estimate code. SAFER /GESTR. In this analysis, no credit was taken for core spray loop II. The worst case single failure including the recirculation discharge line break was assumed. According to the licensee, the peak cladding temperature (PCT) calculated is the licensing basis PCT, which is referred to as PCT: APP.K in the NRC's June 1. 1984 Safety Evaluation Report on Topical Report NEDE-23785. "GESTR-LOCA and SAFER Models for the Evaluation of the loss of Coolant Accident." The results of this analysis showed that a nominal increase in the PCT occurred, from less than 1591*F to less than 1609*F. The staff finds that this value is below the 2200*F PCT limit specified in 10 CFR 50.46(b)(1), and is acceptable. The licenses also performed a probabilistic safety assessment assuming the complete failure of welds P8b and P9 and the loss of core spray loop II. The assessment was 3erformed in accordance with the Electric Power Research Institute (EPRI) 3robabilistic Safety Assessment (PSA) Applications Guide (EPRI PSA Applications Guide. TR-105396, dated August 1995). The results of this assessment 9.17 x 10',showed that t,he core damage frequency would increase from to 1.15 x 10'.

frequency is small and is acceptable for one cycle of operation.The staff finds The licensee also evaluated the potential leakage from the inside shroud to the outside of the shroud via the core spray piping penetration when the collar is com)letely separated from the shroud assuming the weld P9 is intact.

By assuming tie collar separation to be 1/16 inch based on a maximum geometric allowance, the calculated leakage from the inside of the shroud is less than 0.02% of rated core flow. The licensee concluded that the small leakage would not have any significant impact on the safe operation of BFN Unit 3.

___ J

j.'

' Weld P4d is a groove weld connecting the core shroud penetrating pipe to the lower elbow in the core spray C downcomer.

Minor crack indications were visually found on the outside surface-(00) of this weld. The reported cracking consists of two small indications with a length.of k inch and 4 inch, respectively. The crack indications were located on the elbow side of the weld. The two indications were about 1/8 inch apart at the closest point, and i

were oriented about 30 degrees from the pipe axis. The-licensee believes that these crack indicatices are most likely the product of cold work and were initiated from the OD surface, because the subject weld does not contain a crevice and the OD surface of the elbow was ground during the sparger rework performed during construction.

The surface cold work is known to accelerate the initiation of IGSCC. Because of the limited accessibility and the curvature of the elbow. the licensee could not depth size the visually observed flaws using UT or perform inspection of the inside diameter (ID) surface of this weld.

In evaluating the )otential crack growth in weld P4d, the licensee assumed the initial crack dept 1 to be through-wall and the length to be 1.4 inches in the circumferential direction. Based on the licensee's flaw evaluation handbook.

the total allowable effective crack length with one cycle of crack growth for weld P4d.is 9.8 inches.

Based on the core spray piping inspection results and the licensee's flaw evaluation, the staff believes that it is unlikely that there is extensive cracking at-the ID surface of weld P4d: nevertheless, this conclusion needs to be confirmed by inspection.

term operation of B Therefore. the staff concludes that the short-N Unit 3 with this weld in an "as-is" condition uritil the next refueling outage is acceptable. However, the staff recommends that at the next refueling inspection of the ID surface or repair of weld P4d be performed. The staff's concerns are based on the considerations that weld P4d is a closure weld during reinstallation of the spargers, and there is a possibility that significant cold spring and fit up stresses may be present which would accelerate the initiation and growth of IGSCC. Serious cracking at this weld joint could jeopardize the effectiveness of the core spray system.

2.3 Core Soray Pioino Analysis The core saray (CS) piping structural analysis assuming weld P9 is intact was performed ay the licensee to demonstrate that, with a completely severed P8b weld at the shroud / collar interface, the m6ximum stresses in the piping during normal operation and under design basis accident conditions remain within ASME Code.Section III allowable limits. A finite element model consisting of one loop of the internal core spray piping was developed to analytically determine the structural integrity of piping under operational and design basis loads, and with consideration of certain postulated degraded conditions.

The geometry of the internal core spray line and sparger in the analytical model is based on as-built drawings listed in Reference 1. Besides deadweight and inertial loads, other applied loads on the CS line are due to anchor displacements. fluid drag and CS flow initiation.

~

-2.3.1 Evaluation of the-loads The licensee 3rovided an evaluation of the CS piping load and stress analysis as discussed )elow, which was reviewed by the staff. The inertial loadings are due to seismic and flow-induced excitation of the CS line including interaction with the reactor pressure vessel (RPV) and the core shroud.

Seismic excitation, which may cause anchor displacements, is imparted to the CS lines

.tolthe at the vessel nozzle, the support brackets and the attachment points shroud. Anchor displacements of 0.10 inches and 0.136 inches during an OBE at the RPV nozzle and the shroud respectively were obtained from previous analyses (Reference 2), which have been approved by the staff.

The drag loads are due to the fluid flow past the CS lines. The flow in the

. annulus piping. region during normal operation exerts a downward drag force on the CS The magnitude of this loading was determined based on a conservative value of 5 ft/second for the fluid velocity in the vessel annulus region.

During upset conditions. core spray operation is assumed with no feedwater flow, therefore the drag loads are insignificant. Drag loads could occur during a postulated double-ended break of either the recirculation line or the main steam line. The pressure and flow loads which occur only durin operation, are based on an internal pressure of 79 psi (Reference 3)g . CSWater a

hammer lo' ds are insignificant, since the core spray inlet valve ramps open over a period of time upon system actuation, Additionally, the piping is full of water the T-box.during actuation due to the presence of the vent hole on the top of The two anchor points of the internal core spray line at the RPV and shroud attachment locations expand vertically and horizontally at different rates primarily due to differences in the thermal expansion properties of the RPV

-low alloy steel and the shroud stainless steel. However, there are a number of-other factors which affect differential anchor point motion, including RPV thermal and pressure cycles during-various transients and the interval between the postulated LOCA event and the CS initiation.

The resultant stresses due to differential anchor motior, were treated as secondary stresses in the analysis.

Appropriate bounding anchor displacements during normal operation, loss of feedwater pump transient, and during LOCA conditions were considered in the analysis. .The anchor displacement values used in the faulted condition analysis include the shroud vertical dis)lacement during the combined safe-shutdown earthquake plus steam line breac event.

The-staff has reviewed the design input loads for the CS piping stress

-analysis as discussed above and found them acceptable.

2.3.2 Evaluation of the CS oicina stress analysis The licensee evaluated the CS piping with postulated-degraded conditions of the P8b collar welds using a finite element analytical model. The collar weld is assumed completely cracked, in which case the core spray annulus piping is capable of displacing up to 0,0625 inches radially and up to G.23 inches vertically and horizontally. These displacements are limited by clearances

-between the core spray 31 ping, sparger and core shroud.

When the weld is assumed completely cracted, displacement is specified along the three axes and rotation is allowed in all directions. Although the indications of weld cracking at the collar location (P8b) are believed by the licensee to exist in I

7-o the shroud itself, for the purposes of conservatively evaluating the capability of the core spray piping for continued operation, a bounding assumption was made that the indications are located in the collar and that the collar is completely disengaged from the shroud.

an evaluation of the resultant stresses on the core spray performed.

)iping wasUsing these assumptions, An assumed failure of the groove weld P8a. eitler by itself or in combination with P8b. does not alter the bounding assumption of a disengaged collar stated above. in addition, a structural assessment was performed for condition where the indications are located in the shroud. For all the degraded weld cases analyzed, the stresses in the CS system annulus piping were determined to be within ASME Code.Section III allowable limits.

The licensee pi>ing. alsoweld If the collar evaluated the flow-induced is assumed to be completelyvibration (FIV) severed, the effects on the CS li celihood of rattling in the piping section near the collar due to flow induced forces during normal plant operation was evaluated. The resulting moments from displacements of 0.0625 inches in the radial direction and 0.23 inches in the hcrizontal and vertical directions were combined by the absolute sum method and the peak stress range at each of the welds was calculated at the fillet and groove welds respectively.

alternatin The maximum value of the calculated 1910 psi, psi.'g Thisstress (one-half less value is considerabl of the eak than the stresslimit endurance range) of 10000was determine indicating that even if the col ar weld is com)letely severed and the core spray pipe could move to the extent permitted )y the clearances, it would not pose fatigue concerns due to flow induced vibration. The fundamental frequency of the degraded core spray piping was determined to shift only

-slightly from the fundamental frequency of the original piping and was considered negligible.

A primary plus secondary stress evaluation showed that when the FIV stresses are added to the other accident condition stresses, the largest stress range is still lessmargin sufficient than the ASME Code allowable value of 3S, which indicates that exists, Based on its review of the licensee's structural analysis, assuming weld P9 is intact, of the potentially degraded core spray piping including loads development, analytical methodology, boundary conditions and evaluation of flow induced vibration effects the staff concludes that the licensee has demonstrated that the core spra,y piping with the collar weld assumed to be com)letely wit severed will maintain its functionality during the next fuel cycle 1out repairs.

3.0 LONCLUSION Based on the staff's review of the licensee's flaw evaluations, the staff concludes that the structural integrity of the piae to elbow weld (P4d) would be maintained during the next fuel cycle for BFN Jnit 3 on the basis that the final flaw size at the end of the next fuel cycle will not exceed the Code allowable value.

For the collar-to-shroud weld (P8b). the staff concludes that the core spray internal piping would maintain its functionality if-it is assumed that the subject collar-to-shroud weld loses its structural integrity and the weld P9 retained its structural integrity. Further, the licensee s system and risk assessment has demonstrated that, if both welds P8b and P9 failed, adequate core cooling would be maintained under postulated design

n.

8-bases accidents and'the increase in risk would be acceptable-for one cycle of-operation. The staff notes that the integrity of core spray piping should be based on qualified examinations of critical welds in the system and that 6ppropriate inspection and repair method should be developed.

in a response to the staff's concerns regarding the structural integrity of weld F4d beyond the next fuel cycle. TVA, in a-letter dated April 7,1997, has committed to work with the Boiling Water Reactor Vessel and Internal Projects (BWRVIP) committee to develop ultrasonic methods and re) air criteria for the P4d weld and perform other efforts to demonstrate that tie observed indications at the weld P4d originate at the _outside diameter (00) surface-during the next refueling outage.

The staff notes that a repair or replacement of the weld P8b or a qualified volumetric inspection of the weld P9 whose results could be evaluated to 7 demonstrate its structural-integrity should be planned by the licensee for the next refueling.

Principal Contributors: W. Koo, J. Rajan, K. Kavanagh l

Dated: September 22, 1997

9-i

4.0 REFERENCES

1. Browns Ferry Shroud Drawings (a) Reactor Drawing No. 104R935.

(b) Reactor Vessel Purchase Part Drawing No. 8860439. (c) Core Spray 1.ine Drawing No. 920DB24, (d) Shroud Drawing No. 729E458, (e) Nozzle Thermal Cycles Drawing No. 135B9990.

2. Sumary Report on Small Task G056.212: Recalculation of Seismic Responses for Reactor Building Drywell and Internals of Browns l

Ferry Nuclear Power Plant Using El Centro Time History input (Revised Damping Values, Prepared for TVA by Bechtel. August 1989.

3. " Process Diagram Core Spray System " GE Drawing No. 161F257. Rev. 9.

M