ML20062D811

Loose Thermal Sleeve Safety Evaluation

ML20062D811
Person / Time
Site:	North Anna
Issue date:	08/31/1982
From:	WESTINGHOUSE ELECTRIC COMPANY, DIV OF CBS CORP.
To:
Shared Package
ML20062D798	List: ML20062D795 ML20062D811
References
2616Q:1, NUDOCS 8208060226
Download: ML20062D811 (41)

Text

,

.a

. r NORTH ANNA UNIT 2 LOOSE THERMAL SLEEVE SAFETY EVALUATION 1

AUGUST 1982 l

l WESTINGHOUSE ELECTRIC CORPORATION DObK O hohh$9 PDR 2616Q: 1

. f LOOSE THERMAL SLEEVE SAFETY EVAlllATION 1.0 Sumary l,

2.0 Introduction i

• 2.1 Purpose 2.2 Hi story 2.3 Thermal Sleeve Inventory 2.4 Assumptions 3.0 Nozzle Integrity 3.1 Introduction 3.2 Stress Analysis 3.3 Conclusions 4.0 Mechanical Effects of Loose Objects 4.1 Reactor Coolant Pipe 4.2 Steam Generator 4.3 Reactor Internals 4.4 Reactor Vessel 4.5 Fuel 4.6 Reactor Coolant Pump 4.7 Pressurizer 4.8 Primary Loop Stop Valves 4.9 Other RCS Components 4.10 Auxiliary Systems

.4.11 Materials 5.0 Flow Blockage Effects of Loose Objects 5.1 Normal Operating 5.2 Local Core Flow Distribution 5.3 Non LOCA Transients

~ ~

5.4 LOCA Evaluation 2616Q:1

1.0

SUMMARY

Extensive evaluations were performed to determine the effects of loose reactor coolant pipe thermal sleeves at the Virginia Electric and Power Company (Vepco) North Anna Unit 2 plant. These evaluations assumed all thermal sleeves similar in design to that shown in Figure 3.1 becoule loose and are transported in the reactor coolant system as single units or fragments. The evaluations considered nozzle integrity without thermal sleeves, the mechanical effacts of loose sleeves in the reactor coolant system, and flow blockage effects of loose sleeves during normal cperation and transient conditions.

~

?

a

, a

2.0 INTRODUCTION

2.1 PURPOSE As a result of the discovery of failed welds on certain thermal sleeves, a safety evaluation was performed on the effects of loose and missing reactor coolant pipe nozzle thermal sleeves. This report summarizes and documents that safety evaluation.

2.2 HISTORY Westinghouse was recently informed by one of its operating plant cus-tomers that an underwater television inspection had revealed a loose metal piece under the reactor internals lower core plate. Subsequent investigations by Westinghouse and the utility resulted in the discovery of additional loose parts in the reactor vessel and an eventual conclu-sion that the sources of the parts were the thermal sleeves from the 10 inch RHR/ SIS line nozzles; That conclusion has been verified by radio-graphic examination of all four such nozzles on the affected unit. The sleeves traveled through the cold leg into the reactor vessel where all missing parts have been accounted for and recovered. Radiographic examination of other similarly designed sleeves on the affected unit have revealed one broken weld and a very slight movement of the 14 inch surge line nozzle sleeve as well as an indication of a possible crack of a thermal sleeve weld in one of the two 3 inch charging lines.

A similar failure was also discovered at another operating plant where one 10 inch nozzle thermal sleeve of the same design was found to be

missing and has been demonstrated to be in the reactor vessel lower plenum. The welds of the remaining sleeves of the subject design at this unit were shown to be intact by radiographic inspection.

On North Anna Unit 2, a radiographic examination by Vepco has indicated

'that four of the subject design thermal sleeves appear to have~ cracked l

l 26160:1 l

, , welds at the attachment to the pipe; and that th2 remaining seven; including the 14 inch pressurizer surge line thermal sleeve; have welds that are intact.

2.3 THERMAL SLEEYE INVENTORY

. Thermal sleeves are utilized at several locations in the North Anna Unit 2 plant Reactor Coolant System (RCS) to reduce thermal stresses on RCS ,

pipe nozzles. Table 1 provides locations; sizes; and number of the reactor coolant pipe thermal sleeves of the design which have exhibited cracked welds.

1 l

2616Q

1 I

' ~

, , TABLE 1 NORTH ANNA 2 i

THERMAL SLEEVE STATUS N0ZZLE SIZE LOOP iiELD CONDITION CO M NTS Surge Line 14" C Intact Accumulator 12" A Intact B Cracked To be removed C Cracked To be removed SI 6" A Intact B Intact C Cracked To be removed Charging 3" B Cracked To be removed I

l l

l t

l l

l l

i 2616Q:1

The material of construction of the themal sleeves is SA 376 stainless steel type 316 or SA 240 stainless steel, type 304.

Themal sleeves of a different design are also present at the surge line and spray line nozzles at the pressurizer and in the CVCS fill lines on

,the RCS crossover leg. These sleeves employ welds of 45*, 45' and 360*

respectively and have a counter bore geometry to prevent movement of the sleeve. There has been no evidence of failure of these types of sleeves, and they are not considered in this safety evaluation.

2.4 ASSUMPTIONS To complete the safety evaluation for North Anna Unit 2 certain assump-tions were made. These assumptions are based on facts gathered from the first operating plant to discover missing themal sleeves, engineering judgement, and recommended actions for continued operation. The assumptions are as follows: ,

1. All reactor coolant piping themal sleeves of the subject design are assumed to come loose and are transported through the RCS system.

2. The sleeves are assumed to remain intact or split into quarter sections, whichever case provides the most conservative evaluation. The sleeves are attached by two welds at 180* in line with the loop flow on the upstream end. Field examination indicates cracking can occur at the welds allowing an intact sleeve to come loose. Another failure mode which has been observed at another plant is cracking of the sleeve along its length, beginning at one of the notches along the upstream end of the sleeve. Both of these failure modes produce large objects. The ductile i

- nature of the sleeve material also makes it unlikely that t

small pieces would be generated by impacts within the reactor coolant system. This evaluation specifically considered objects ranging in size from a complete 14 inch l 2616Q:1 l

O E

• ' sleeve to one quarter sections of the 3 inch sleeves'.

Smaller fragments were also addressed in the nuclear fuel evaluation.

3. The plant operators are aware of the potential for loose parts and will monitor plant operations and pertinent

~

equipment characteristics.

l l

l l

l . _.

l t

l 2616Q:1

'

• 3.0 N0ZZLE INTEGRITY

3.1 INTRODUCTION

This section sumarizes the stress evaluation of the 3" charging noz-zies, the 12" accumulator nozzles; the 6" SI nozzles; and the 14" pres-

'surizer surge nozzle on the main reactor coolant loop piping; to insure the structural integrity of the nozzles assuming certain failures of the themal sleeves. The specific themal sleeve failure discovered during inspection of the subject nozzles; and considered in this evaluation; included; a three inch charging and 6 inch safety injection themal sleeve weld failure and rotation of the sleeve; and a 12 inch accumula-tor injection nozzle themal sleeve weld failure with the sleeve becoming lodged in the nozzle between 1.0 and 6.0 inches below (down-stream) its installed location.

The analysis included an evaluation of the subject nozzles without a themal sleeve and a " bounding" evaluation of the nozzle at the location of the failed sleeve / nozzle attachment weld. Even though the themal sleeves have been removed on certain nozzles included in this evalua-tion; a " bounding" analysis was still perfomed on all nozzles for conservatism. This evaluation which considered all design transients and mechanical loads specified in the piping design specification demonstrates the structural integrity of the subject nozzles without themal sleeves.

Due to the similarities in the geometry of all subject nozzles; and the similarities in the themal sleeve designs (see Figure 3.1) the same analytical techniques were applied to all nozzles. The evaluation was separated into the following three basic regions on the nozzle; (see Figure 3.1); 1) the location of the nozzle to pipe field weld at the

" safe-end" of the nozzle; 2) the location of the original sleeve weld to nozzle wall and 3) the remaining body of the nozzle including the crotch

' region.

2616Q:1

, , 3.2 STRESS ANALYSIS The stress analysis perfomed on the subject nozzles can be summarized as follows. The detailed geometry and material of the nozzle; without a themal sleeve, was obtained from the appropriate specifications. (For example, the previously mentioned figure and the plant specific drawings and equipment specifications). Then a detailed 2-dimensional finite element model was developed for the nozzle and appropriate representa-tive portions of the large header pipe and attached branch pipe (Figures 3.2 and 3.3).

Using piping design specifications containing operating transient des-criptions developed on the basis of the systems design criteria, the temperature transients; fluid velocities, number of occurrences, etc.

were summarized for all applicable transients; and appropriate loading conditions were developed for the heat transfer analysis using the finite element model. The analysis included a time-history thermal loading for a sufficient duration of time to insure the maximum stress intensities were calculated for all locations.

Using the same finite element model, stress intensities were calculated from the pipe wall temperature distribution obtained from the heat transfer analysis for all critical locations. The actual fatigue evalu-ation of the component incorporates the methods and guidelines specified in the ASME ANSI B31.7 Nuclear Power Piping Code, USA Standard for Pres-sure Piping; 1969 Edition; including the 1970 and 1971 Addenda.

This rigorous treatment has been applied to the 3" charging nozzle, the 6" safety injection nozzle, and the 14" surge line nozzle without themal sleeves. Due to design modifications for later plants, the 90*-12" accumulator nozzle was changed to a 45* inclined injection l nozzle without a themal sleeve. A complete set of themal transient stress analysis was perfomed for this inclined injection nozzle for the same loading conditions as specified for the 90* injection nozzle. In addition; analysis was also perfomed on a geometrically similar nozzle 2616Q: 1

, (6-inch) without a themal sle ve with siailar design transients. Th3 results of these two analyses were used in the qualification of the 12-inch accumulator injection nozzle without a thermal sleeve.

In the analysis of the nozzle without themal sleeves, two locations were found where maximum peak stress intensity and fatigue usage occurred,1) the th.ck part of the nozzle near the crotch region end 2) the nozzle to the branch pipe field weld. This second region was found to be critical after stress intensification factors were applied to the weld location; as specified in the ANSI Code. Assuming the as-welded conditions; a stress concentration factor of 1.7 was applied o,n top of the calculated values. At the crotch region, a factor of only 1.1. was applied; due to the ground flush condition at the weld location.

To complete the fatigue calculation, the external loadings on the noz-zie, as calculated for the North Anna Unit 2 plant were incorporated; and a usage factor was calculated for each nozzle.

Finally, an evaluation of the failed fillet weld region on the nozzle was performed. Because of the close proximity of the fillet weld loca-tion to the pipe / nozzle butt weld (1.0-1.5 inches); the evaluation of the safe-end location could be shown to yield the same usage factor, once the following was considered. An appropriate stress intensifica-tion factor was required to simulate the inside surface of the nozzle at this location. Factors of 1.4 for K3 and 1.5 for K2 were conserva-tively used. This was based upon the relative severity of the condi-tions which resulted in the factors (K 3=1.7 and K2 = 1.8;) for an as-welded butt weld, (i.e., affected inside surface; thin-walled pipe, l

misalignment of the butted pipe walls,) and the condition actually l present at the fillet weld location (affected inside surface, thick wall pipe, perfect alignment). This difference in stress intensification factors more than offset the small increase in stress intensity due to the location being closer to the thick part of the nozzle and resulted

'in no significant change in stress. -

1 2616Q: 1

, 3.3. CONCLUSIONS The cumulative usage factors calculated on the basis described in the previous sections and the external loadings based on North Anna Unit 2 specific as-built information indicates that all critical locations meet the ANSI Code requirements. Therefore, it is concluded that the nozzles are qualified to withstand all applicable design transients and will maintain their structural integrity without thermal sleeves for the plant design life.

f i

F i

2616Q:1

TER%L SLEEVE BlDT WELD N0ZZLE l PATERIAL: SA240 OR SA312 l 1

TP3C4 OR TP306 1/8" WIE SLOTS DP. 6 PLAES l

\ \ 7.

JL h p

/\'

2 FILLET hELDS AT 180" 562 2 d ,

- (2)562" DIA. VEffF

%A Ujl M HOES AT 180

\  : 4 N ~ ii -

,3" a j N0ZZLE f k j 4 hED MPOSITS AT

90* - GP.IND FOP. TIGH

^ '

FIT TO SLEEVE

^ -

REALTOR C00!MT PIPEWALL i:

7, -

,4 4 i s: e i

FIGUE 3.1

,. . 9 i t.1 u

l. ,'2 s t ."., ' (

ti 54

.. . . .v .

i,.i 4

4 - .

in., ,

g .

N .

s t B ,

t

'w (:

i

,. . r. .. - ..

. .,g 4- . . , ,, c. -

3 gute (maattet, nothi vet 188*** '4fI'I I8N I ' u m..un to at=

l I

FIGURE 3.2

e

• l l

l 1

P9.9 ,

t E

M3 2d 21.3 E"-3" -

{!d 2 23.0 '

'2

!

CU 8

30.8 ,

u,.3

.1*

t P1 .

p.-'

' ~ ~ ~

39.9 ,

• N 34 88.3 , r,i  %

$ X

• it. ,

f ,

/

11.8 ,

1 14.5 l ,

g

["'y)M -

33.3 '

s I

8 ' 's'. i 4.. a.t ** +' '*

- '* *8 66 6 63 3 '3 3 3 fus (masEM tel$ki g#9 MIINAL RI(V( (&W l WAI EE N I MM FIGURE 3.3

. . 4.0 MECHANICAL EFFECTS OF LOOSE OBJECTS 4.1 REACTOR COOLANT PIPE The effect of the loose thernal sleeves on the primary system piping, either through impact or erosion, is expected to be negligible due to the limited impact energy created by the low radial flow velocities in the piping. The ductile material of the piping and the themal sleeve would also preclude any sharp impact marks on the piping, thus e ' min-ating any concern regarding possible stress concentration points.

The locations of the RTD bypass scoops and thermowells in the reactor coolant piping are upstream of the thermal sleeves, except for the 14" surge line themal sleeve which is upstream of the hot leg RTD scoops and T hot themowell. The effect of a loose 14" thermal sleeve impacting these components during operation has been considered and is discussed below.

There are three RTD bypass scoops at 120* locations which protrude 6 to 8 inches into the flow stream. Upon impact of the themal sleeve, the scoop is assumed to be sheared off or defomed sufficiently to make it l ineffective. The RCS pressure boundary will not be violated. An additional loose part could be generated, but it will be captured in the steam generator channel head along with the sleeve. Damage to one or more scoops could affect the flow rate in the bypass line, thereby increasing the delay time in the temperature measurement. However, a significant change in the flow velocity would actuate a low flow alarm to alert the plant operator.

l l

Defoming or shearing the hot leg RTD scoops has two effects: 1) it l will increase the RTD loop flow transport time, and 2) it could cause the RTD to indicate a different coolant temperature than the other loops due to radial temperature distribution in the RCS pipe.

2616Q:1 l

l

• Tha maxirum flow transport time under the worst postulated conditions would be 1.3 seconds. The safety analysis reported in the FSAR assumes a delay of 2 seconds for flow transport time and 2 seconds for the RTD sensor. Thus, the predicted flow transport time is bounded by the safety analysis. Should the scoops be sheared off, the RTD's would be biased by less than 2*F. The overtemperature AT trip requires a 2 out

'of 3 input. Thus a bias in one RTD will not cause a safety concern since two channels remain unaffected.

Damage to the RTD scoops can affect the perfomance of control systems which use this temperature measurement as input; e.g. rod control sys-tem. The perfemance of these systems could be affected, however, no safety concern is created.

In sumary, potential damage to the RTD scoops does not create a safety concern.

hot themowell is located at the 0* centerline of the RCS piping The T and protrudes approximately 3 inches into the flow stream. Although the profile of the thermowell is small relative to the total flow area, if the 14" themal sleeve did impact the thermowell it could result in the severance of the well at the RCS pipe wall. An additional loose part would be generated and a leak would result (approximately 15.7 lb/sec) through the connection. The Thot measurement would become unavailable and the potential for a missile (Thot probe) and jet impingement would resul t.

l l The RCS safety would not be jeopardimd since this sim leak is detect-able and also is within the makeup capability of one charging pump.

This does not present an unanalymd event nor does it challenge the plant safety systems.

l 4.2 STEAM GENERATOR 4.2.1 Introduction This evaluation considers the potential effects of an 11.5" outside diameter themal sleeve entering the primary side of the Series 51 steam 2616Q: 1

gsneraters at North Anna Unit 2 from the connection of the pressurizsr surge line. Affected components of the steam generator may include the tube sheet; divider plate; channel head; tube-to-tubesheet welds; tube-sheet-to-divider plate weld; and the divider plate-to-channel head weld.

Damage as a result of the 34 pound sleeve impacting on these components is considered separately in the following sections.

4.2.2 Tube Sheet and Tubes The tubesheet of Series 51 steam generators is clad with Inconel 600 which is,quite ductile. Repeated impacting on the cladding by the 304 stainless steel themal sleeve; which is also a ductile material, would not be expected to cause the cladding to crack and break loose. The design of the Series 51 steam generator incorporates tube ends which extend approximately 0.22 inches below the primary face of the tubesheet cladding. This configuration exposes the tube ends to potential impacts from the presence of a loose themal sleeve.

Evidence from a previous incident with a loose part of comparable mass in a Series 51 steam generator shows damaged tube ends. The damage was mainly bending and deformation of the tube ends rather than peening as has been seen resulting from smaller loose parts in the primary side of the steam generator at another unit. The evidence of comparable mass indicates tube end flow restriction due to this ductile deformation of the tube ends. If the pressurizer surge line thermal sleeve were to enter the steam generator the damage would be expected to be similar to the bending and defomation explained above. The resulting flow restriction would have a minimal affect on the primary flow through the loop. A margin would still exist to the conservatively low thermal design flow. If a large flow mismatch did occur; it would be detected by the operator who could then take appropriate actions. The ductility l

of the inconel material makes it unlikely that small pieces of the tube ends would break off under short-term exposure to impacts by the loose

. parts. .,

i t

2616Q: 1

.o .

Operation for a short period of time in the presence of a loose thermal sleeve would be expected only to produce bent and defomed tube ends but not to generate any tube end pieces. Operation for longer time periods could generate tube end pieces that would not affect the steam generator but may affect other components of the reactor coolant system.

.The thermal sleeve design contains notches at the upper ends for stress relief. These notches are 90* apart and experience indicates some cracking at these notches on failed sleeves. Thus, if the sleeve were to break apart it is anticipated to break at the notches, feming large sections. No piece small enough to f!t into the 0.775" inside diameter opening of the tube is expected to be formed from the break-up of the thermal sleeve.

Therefore, it is concluded that potential damage to the tubesheet and tubes resulting from short-term impacts by a failed themal sleeve would not violate the integrity of the steam generator components.

4.2.3 Tube-to-Tubesheet Weld l

Impacting of the tube-to-tubesheet (TTS) weld by a loose thermal sleeve is considered unlikely due to the presence of the tube-ends extending l

0.22" beyond the primary face of the tubesheet, thus protecting the welds from absorbing a large number of direct impacts. Thus, disinte-gration of the weld from impacting is considered unlikely due to the ductility of the materials and the geometry of the welo.

l

[

One design feature of the Series 51 steam generator is the explosive f expansion (WEXTEX) of the tube over the entire 21 inch depth of the

! tubesheet. This feature provides added strength to the tubes in the I tubesheet hole and provides an additional margin against primary to l

I secondary leakage.

~

I If it is assumed that some of the welds do completely disintegrate and primary to secondary leakage occurs, the amount of leakage would be l .

2616Q: 1

. . . , i low. Such leakaga would bo detectable by normal radiation monitoring; and the extent of the leakage could be monitored. This leakage would be expected to be within the allowable technical specification limits and would present no safety concern. Monitoring of the leakage would be l possible so that if an increase is detected the plant could be shut down l in an orderly manner.

'424 Divider Plate The 2 inch thick Inconel 600 divider plate is welded both to the channel head and the tubesheet to fom a barrier separating the hot leg and cold 1eg of the steam generator. The rigidity of the plate is highest closest to the welds; and it becomes more flexible toward the middle of -

the plate. Impacting of a thermal sleeve could be expected to occur in the flexible region of the plate. The geometry of the channel head limits access to areas closest to the welds.

The flexibility of the plate in the most likely impact region along with the flexibility of the themal sleeve will cause the impact loadings to be sufficiently distributed so as to be of no concern to the integrity of the divider plate.

As mentioned previously; the ductility of the sleeve material would l reduce the likelihood that sharp edges would be created. Therefore; any marks that result from themal sleeve / divider plate impacts would most likely be round-bottom; rather than sharp-pointed. It is therefore unlikely that stress riser areas would be created.

The effect of impacts near the welds of the divider plate are discussed in the next section.

4.2.5 Tubesheet-to-Divider Plate and Divider Plate-to-Channel Head Welds

.Because of the location of these welds, the number of direct impacts they would receive from a loose themal sleeve is low. InprEvious l

2616Q:1

circumstances involving loos 3 parts on th2 pricary side of the steam generator at other plants; inspection of these welds showed no evidence of degradation due to impact forces.

Long-term fatigue induced by forces being transmitted to the welds by continual impacting of the divider plate and/or channel head in the region close to the welds is of no concern, due to the flexibility of

'the divider plate and the low stresses induced in the welds.

4.2.6 Channel Head The inside of the channel head is weld clad with a ductile, austenitic stainless steel. Impacting of the thermal sleeve on the channel head would thus not cause any sharp dents where a point of stress concentra-tion would form. The ductility of the clad material makes it unlikely that enough impacts will occur on a particular spot to cause cracking and loose cladding. Therefore; impacting of the thermal sleeve would not be expected to adversely affect the channel head and cladding.

4.2.7 Conclusions The potential entry of a 14 inch thermal sleeve from the pressurizer surge line into the primary side of the steam generator is not expected to adversely affect the continued safe operation of the steam genera-tor. Short-term operation with the sleeve present in the steam genera-tor will not create loose tube end pieces.

I t

2616Q:1

• 4.3 REACTOR INTERNALS The reactor internals were evaluated to determine the effects of impact and wedging loads on reactor guide and support structures due to the presence of loose thennal sleeves in the reactor coolant system.

'.3.1 4 Upper Internals The 3 inch charging injection line thennal sleeve, the 6 inch safety injection line thermal sleeves and the 12 inch accumulator injection line thermal sleeves will be confined between the lower core plate in the reactor vessel, and the steam generator cold leg plenum. As such, these thermal sleeves will have no impact consideration on reactor upper internals. The thermal sleeve located in the 14 inch hot leg pres-surizer surge line does have the capability of becoming lodged in the upper internals. In a back flow or alternate leg blowdown situation, a loose surge line thermal sleeve could travel back through the hot leg into the upper internals. The following assessment utilizes plastic analysis to determine impact loads on support columns and guide tubes in the reactor upper internals.

Support Columns Length 78.77" 0.0. 7.49" I.D. 6.53" Thickness = 0.4875" A = 10.72 in 2 l

Material : ASTM A 479 Type 304 stainless steel, cold finished.

l 2616Q:1

Guide Tubes 17 x 17 Length 125" Thickness 0.25" Size 7.34" x 7.34" k=7.09in2 Back Flow Velocity Mass back flow 935 lbm/sec Density 9.34 lbm/ft 3 Area 4.587 ft 2 Y - 21.8 ft/sec UPPER INTERNAL STRESS

SUMMARY

LOAD STATIC COLLAPSE DEFLECTION (KIP) LOAD (KIP) (INCH)

Support 17.45 22.1 0.266 Column b

Guide 12.3 25.9 0.3 60 Tube As seen from this table, the loads exerted are less than the static collapse load, therefore, impact loadings on reactor internals upper support columns and guide tubes are acceptable.

Objects in the bottom of the reactor vessel would not be expected to reach the upper internals due to the filtering action of the lower

)

2616Q: 1

The close spacing of the rods, the con-internals and fuel assemblies.

figuration of the grids and the flow deflectors, and the configuration of the nozzles should prevent large particles and most other particles from reaching the upper internals. Small particles which could pass through the fuel assemblies are likely to pass through the upper internals or to be forced clear during operation of the drive line. In order for a foreign object to cause interference, it would have to be preferentially oriented.

As part of the normal startup tests, control rod drop times are recorded and evaluated to confirm proper driveline performance. In the unlikely '

event that a foreign object would become lodged in the upper package during operation and cause a driveline to become inoperable, the existing FSAR analyses assumption of one stuck control rod assembly would not be exceeded.

4.3.2 LOWER INTERNALS The reactor vessel and lower internals were analyzed for structural integrity with thermal sleeves from the 3 inch charging line, 6 inch safety injection line and 12 inch accumulator lines within the reactor vessel. The thermal sleeve from the 14 inch pressurizer surge line is unable to reach the reactor vessel lower internals.

4.3.2.1 Core Barrel It was assumed that a complete 12" sleeve strikes the core barrel at the inlet nozzle velocity. Since the sleeve is thin it will deform before the core barrel deforms. Therefore, the load applied to the core barrel is determined by the load capacity of the piece. Assuming an ultimate 2

strength of 63.5 ksi for the piece, and an impact area of 5.84 in for the end of the sleeve, the maximum load applied to the core barrel is 371 kips. Assuming the core barrel responds as a cantilever beam, the impact stresses in the core barrel are calculated to be small.

( q,,x = 1120 psi and Tmax = 435 psi).

2616Q: 1

l

. i The method used for the minimum missile energy required to perforate a target plate per WCAP 9934 results in a maximum depth of dent equal to

.034 in.

Due to the low magnitude of the impact stresses and the short time

, duration of impact loads, the core barrel is unaffected by impacting loose parts.

4.3.2.2 IRRADIATION SPECIEN GUIDES The irradiation specimen guides are welded to the outside of the thermal shield panels. The top portion is welded using a .38" groove joint,'

4.56" long on each side. The middle portion is intermittent 0.11" bevel welds totaling 70.57" long on each side. The contact area is calculated by assuming the face of a quarter section of a 12" sleeve strikes the top of the specimen guide at 34 ft/sec. The impact force is calculated to be 63,400 lb. The maximum shear stress is 3,340 psi. In view of the small magnitude of the shear stress, the specimen guide will not be affected by the impact.

4.3.2.3 BOTTOM MOUNTED INSTRUENTATION TUBES The instrumentaticn tubes in the bottom head of the vessel were evalu-ated for impacting of thermal sleeves or thermal sleeve sections which may be loose in the system. The cases evaluated were for an impact at the tube / bottom head intersection (shear strike) and for an impact at i the highest point on the instrument tube which could be struck without

. first striking the internals. Resulting values were compared to appro-priate shear and collapse load allowables.

t f

2616Q:1

The shear strike was evaluat:d only for the largsst themal sleeve (one-half of a 12 inch themal sleeve) which could impact the instrument 1

tubes. The maximum shear stress was found to be only 1.13 KSI which gave a margin of safety of 36.8 compared to the allowable of 0.6 Sm.

The loads on the instrument tubes resulting from the bendinc strike of a

' half section of the 12 inch themal sleeve were evaluated as exceeding the instrumentation tube collapse load. This result indicates that plastic defomation of an instrumentation tube could result if the tube were struck in an unfavorable manner by the loose thermal sleeves.

However, due to the ductility of the Ni-Cr-Fe alloy tube, defomation could occur, but the tubes will not rupture and will continue to protect the thimble guide tubes. The thimble tubes would therefore not rupture and the pressure boundary will not be violated.

In the unlikely event that the failure of a bottom mounted instrumenta-tion tube leads to leakage, the double ended break of this tube results

! in a leak area of 0.00024 Ft2 . Assuming a discharge coefficient of 1.0 and using the Zaloudek subcooled critical flow model which over-predicts leak flow, one charging pump in the nonnal charging mode can provide makeup for at least 3 broken tubes.

This would be classified as a leak, not a LOCA, and RCS pressure would l be maintained at 2250 psia. If both charging pumps were available, additional tube leaks could be tolerated.

Small break LOCA analyses with minimum safeguards SI have demonstrated that full instrument line breaks in as many as 5 instrument tubes will not result in core uncovery. RCS depressurization and automatic SI initiation will occur, however, this small break LOCA will maintain fon:ed or natural cin:ulation, and the RCS will reach equilibrium condi-tions.

Therefore, the loose themal sleeves striking the instrumentation tubes in the bottom head of the vessel does not constitute a safety hazard.

I 2616Q:1

=_ - - -- - - -

4.4 REACTOR VESSEL During plant heatup, the gap between the reactor vessel bottom head inside surface and the bottom of the secondary core support structure will decrease. A foreign object present in this area could impose mechanical loadings on the vessel. Due to the size of the gaps a full "3" sleeve could not enter the gap. A quarter section of a 3" sleeve could enter the gap, and the force necessary to deflect the piece to the minimum gap size was calculated to be approximately 6,460 pounds. This load is acceptable.

The effect of impacts on the radial key was also evaluated. The largest piece that could enter the outer annulus of thermal shield is determined to be one half of a 12" thermal sleeve. The impact velocity is assumed to be 36 ft/sec. and the impact force is determined to be 42,300 lo.

Assuming all the impact load is carried by the six 1" dia. dowel pins, the resulting stresses is 8980 psi. Comparing to the allowable stress intensity, this gives a margin of safety of approximately 4.7.

4.5 NUCLEAR FUEL Foreign objects in the primary system have two potential effects on the nuclear fuel: 1) partial flow blockage of fuel assemblies due to an object becoming wedged in the fuel assembly flow paths, and 2) clad wear due to pieces becoming lodged in the assembly or between two assem-blies. Flow blockage effects are discussed in Section 5 of this report.

From a fuel mechanical design viewpoint, loose pieces should not pose an operational problem when the fuel assemblies are seated properly on the core plate. The loose pieces should be stopped by the bottom nozzle or the lower core plate due to dimensional considerations. Although highly unlikely, it is possible for a very small piece to wedge between fuel assemblies and cause fretting and/or grid damage. This is highly impro-

'bable due to the fact that space between fuel assemblies is approxi-mately 40 mils, i.e. approximately one third the thickness of the 2616Q:1

. thtrmal slcove material. Should a fratting mechanism causs clad failure on a fuel rod it is unlikely that any radiation release woulti approach the technical specification limit, and as such no safety concern would exist.

Due to the relatively large fragments expected from the themal sleeves, the transport of loose pieces into and through the fuel assemblies is unlikely.

4.6 REACTOR COOLANT PUMP There are no thermal sleeves of the subject design located in piping connections between the reactor coolant pump (RCP) and the steam genera-tor. A loose thermal sleeve can enter the RCP only when a reverse flow condition occurs, in which case the plant is not operating at power. If this occurs a thermal sleeve or portion of one will not affect the pres-sure boundary integrity due to the geometry, mass and low impact energy of the pieces.

An intact 3 inch themal sleeve or similar size fragments of a larger thermal sleeve can pass through the pump internals without significant deformation.

The larger thermal sleeves would not pass through the pump diffuser and impeller during a non rotating impeller condition. During RCP startup i the forward flow would eject any fragments.

If thermal sleeve fragments did lodge between the impeller and diffuser in such a way as to cause interference, the material may be pinched or sheared between the impeller and diffuser vanes due to the very high torque of the RCP. A consequence may be an increase in shaft vibration with continued RCP operation, i.e., no locked rotor or pressure boundary violation is expected to occur. Increased vibration could be observed

'by the operator and corrective action could be taken. ~

2616Q:1 i . . - -

. ' A steilar safety evaluation of largsr caterial (1 1/16 inch thick, 304 SS) that was postulated to enter the RCP in various size fragments was previously performed, and it also concluded that there was no safety Concern.

4 In summary, the loose thermal sleeves are not considered a safety con-

! cern for RCP integrity and operation.

4.7 PRESSURIZER The thermal sleeves in the 4 inch spray line and tha 14 inch surge line connections in the pressurizer proper are attached in a different manner than the reactor coolant piping nozzle thermal sleeves. On the pres-surizer thermal sleeves the upstream end of each sleeve is welded over an arc of 45 degrees. The sleeves themselves are of larger diameter than the nozzle safe ends, thus preventing sleeve movement away from the pressurizer. The flow distribution screen inside the pressurizer at the i surge line connection prevents that sleeve from entering the pres-surizer. Similarly, the spray header traps the sleeve on the spray line connection.

Due to their method of attachment, it is also very unlikely that these sleeves would become loose within the reactor coolant system. In addi-tion operating experience has indicated no evidence of failure of these sleeves, and thus they are not considered in this loose sleeve safety evaluation.

Based on the most probable movement of any dislodged thermal sleeves l from the 12 inch SI lines or the 3 inch charging line it is extremely unlikely that any piece would cause mechanical damage or become lodged in the pressurizer inlet piping or the pressurizer.

i l 4.8 PRIMARY LOOP STOP VALVES The effect of loose thermal sleeves on the primary loop stop valves either through impact or erosion is expected _to be negligible since l

l l

l I 2616Q:1 l

1 - . -_.. - . _- -- .- _ _ . - _ - -_ . . _ . ___ __ -.

• thera are low radial flow velocities and no appurtenances ext nding into the flow path during plant operation. The remote possibility exists that the disc guides, located outside the flow path, could deform if impacted by a thermal sleeve. If this were to occur, the valve may not reach its fully closed position; however, the primary coolant pressure Doundary would not be violated. The loop stop valve has no safety function, and any restrictions to closing would not present a safety l con::ern.

4.9 OTHER REACTOR COOLANT SYSTEM COMPONENTS l

Due to the physical separation from the remainder of the reactor coolant system of such components as control rod drive inechanisms and safety, relief and block valves, no adverse effect is expected to result from loose thermal sleeves in the reactor coolant system.

4.10 AUXILIARY SYSTEMS The possibility of'the potentially loose themal sleeves affecting the i operation of other systems connected to the RCS was also investigated in this safety evaluation. The evaluation belcw-considers each themal sleeve location and the possible paths to systems or components inter-facing the RCS.

4.10.1 SURGE LINE THERMAL SLEEVE If the surge line thermal sleeve came loose during operation, it would be moved by the loop flow to the steam generator inlet plenum. The sleeve could impact the RTD bypass line scoops or the themowell. The potential effects on these components is discussed in Section 4.1. The SIS, drain line, and loop stop valve bypass line connections and the pressure tap in the hot leg do not protrude into the loop flow and are not vulnerable to impact damage. Entry of a piece of sleeve into the SI nozzle would be highly unlikely due to the location, orientation and stagnant flow conditions of the line.

2616Q:1

l 4.10.2 NORML CHARGING LINE THERMAL SLEEYES The 3" charging line enters loop B downstream of the 12" accumulator discharge line and the 6" safety injection line. The normal flow is toward the reactor vessel. There are no other connections to the RCS j piping between this line and the vessel. Thus the thermal sleeve or i parts thereof if dislodged would be expected to migrate to the reactor l vessel. With reverse flow in loop B it could be postulated that the sleeve from the charging line, or parts thereof, might enter the accumulator discharge line or the safety injection line. However, the parts would not migrate up these lines due to the geometry and stagnant flow conditions in the lines. In order to enter the lines the sleeve would have to travel against gravity since both connections are in the upper half of the RCS pipe. This would create no safety concern.

4.10.3 ACCUMULATOR DISCHARGE THERMAL SLEEVES The 12" accumulator discharge line is located upstream of the 3" charging line and downstream of the 6" safety injection line. In loop B a dislodged accumulator line sleeve or parts thereof would be expected to migrate toward the reactor vessel. Since the charging line has an ID of about 2.1" with a sleeve and 2.7" without a sleeve it is impossible for the entire sleeve or fragments to enter this line with its sleeve intact. Even if the sleeve were missing it is improbable that fragments at most .1" smaller would enter the line.

In loop A and C the accumulator discharge line enters the RCS upstream of the 6" safety injection line and the 4-inch spray line. There is no other connection downstream between the accumulator discharge and the vessel. Thus dislodged sleeves or fragments would migrate toward the vessel. If reverse flow is pstulated, it would be impossible for an entire accumulator discharge line themal sleeve to enter either a

. safety injection line or a spray line since the outside diameter of the sleeve is about 10.5" while the ID's of the SI lines and spray lines (without sleeves) are on the order of 5.1" and 3.4", respectively. It 2616Q:1

- could be postulated that fragments might enter thm SI line or tha spray line. Entry into the SI line would be of little consequence since there is no flow in the line and the fragments would be dislodged on safety injection. Should a fragment enter the spray line it could be postu-lated that the fragment could migrate against gravity and possibly damage the pressure control valve. Should the valve stick open a' pre-mature plant shutdown would be the worst consequence. There would not be any safety significance.

4.10.4 SAFETY INJECTION LINE THERMAL SLEEVES The 6" safety injection discharge line enters the RCS in Loop B upstream of the accumulator discharge line and the normal charging line and upstream of the accumulator discharge and spray line in loops A and C.

A dislodged sleeve or part would normally tra'v'el to the reactor vessel if it fell into the RCS. If it simply became loose, it would be unlikely to migrate up the SI lines as there is no flow in these lines.

On safety injection such parts would be expected to migrate with the flow toward the reactor vessel. It could be postulated that a sleeve or part thereof enters an accumulator discharge line. From this line it would be flushed toward the reactor vessel on accumulator discharge. It would be impossible for. an entire safety injection line thennal sleeve to enter the normal charging line and improbable for a fragment to enter the line due to the size and construction of the line. For the same reason it is also improbable that a. fragment would enter the spray l

li ne. As described above, should this occur it would be of no safety l signi ficance; ,

To summarize; it is unlikely that failed sleeves or parts thereof would '

cnter SI, accumulator discharge or spray lines since for this to occur the parts would have to travel against gravity as all of the connections l are in the upper half of the RC piping. If parts did enter any of the lines, they would'not migrate to places that would adversely impact plant safety. -

N I

2616Q:1 ,

s ,

4.11 MATERIALS No unacceptable material would be introduced into the reactor systems as a result of the failure of a thermal sleeve. Minor clad damage could occur on the surfaces of carbon steel components, however, this would present no safety or operational concern due to the very slow corosion rate of the carbon steel in the reactor coolant environment. l 2616Q: 1

. 5.0 FLOW BLOCKAGE

5.1 INTRODUCTION

In postulating the presence of loose thermal sleeves in the reactor coolant system, an evaluation was made of the effect of the sleeves or parts of the sleeves blocking flow in the core and in various locations in the reactor coolant system. The evaluation considered that all thermal sleeves come loose in the reactor coolant system and are moved by RCS flow to the following locations:

A. The 3", 6" and 12" sleeves protrude into the cold leg flow. This case bounds the case where they lodge in the lower internals and block flow at tne lower core plate.

B. The 14" sleeve from the pressurizer surge line blocks flow at the steam generator tube sheet. (The case of the intact 14" sleeve partially blocking flow in the hot leg was also analyzed, however, blockage at the steam generator tube sheet was determined to be more conservative).

The evaluations considered the effect of blockage on reactor coolant system total flow, local flow distributions in the core during normal operation, and the effect on LOCA and non-LOCA accident analyses.

5.1 REACTOR COOLANT SYSTEM TOTAL FLOW For the analysis of reactor coolant system flow reduction, the loose 3",

6" and 12" thermal sleeves were modeled as protruding fully into the cold leg flow.

The 14" sleeve segments in the steam generator were assumed to com-pletely block flow in 10 percent of the tubes. This is a very conser-vative assumption since it is extremely likely that the segnents will retain their curvature and only cause a flow restriction rather than total flow blockage.

2616Q:1

Th2 results of this constrvativa analysis sh:wed that the total reduc-tion in RCS flow was approximately 1.13 percent. This still results in the RCS flow being greater than themal design flow, which is a conser-vatively low value of flow rate upon which the core thermal-hydraulic design is based. Thus, this flow reduction will have no effect on the thermal-hydraulic design and DNB margin in nomal operation at rated power. Based on the above evaluations it was concluded that the reduc-tion in RCS flow would not affect design margins in nomal operation.

5.2 LOCAL CORE FLOW DISTRIBUTION Although at North Anna 2 no fragments have been generated, experience at other plants has indicated fragments can occur. Due to the wide distri-bution in the size of the potential pieces, the evaluation involved three postulated conditions: 1) the effects of material entrapped by the lower core plate, 2) the effects of material entrapped by the bottom nozzle plate, and 3) the effects of material carried upward into the assemblies. Infomation and discussions pertinent to each condition are given below. Note that the response given in Section 5.2.2 is consis-tent with the North Anna FSAR Chapter 4.4 dealing with the flow blockage.

5.2.1 Material Entrapped by the Lower Core Plate The segments from the sleeves remaining below the lower core plate would result in greater core blockage than the smaller segments reaching the fuel nozzles, since the smaller pieces could only reach the fuel nozzles in a lengthwise orientation. In perfoming this evaluation, it was assumed that the sleeve segments remain curved, and thus do not com-pletely block flow, but do cause restrictions in the flow to the core.

The information available on themal effects due to flow blockage indi-cates that there will be no significant increase in the likelihood of DNB at nomal operating conditions. WCAP-7956 shows results from a blocked assembly flow recovery test and WCAP-8054 shows that a-10 per-cent flow reduction in the hot assembly and its 8 surrounding assemblies reduces DNBR by only 0.3 percent. Since the themal sleeve pieces will 2616Q:1

-. = . _ , . _ = - .:  :- - - - - - -

remain curvad, there will always b2 some ficw thrt: ugh all of th2 lcwer core plate holes. This, along with the fact that the total core thermal design flow will remain unchanged, will insure that the DNBR will not be reduced by more than a few percent.

Thus, the effect of blockage on local core flow distribution and DNBR is dudged to be insignificant.

5.2.2 Material Entrapped by the Bottom Nozzle Plate Because of the limiting flow holes in the bottom nozzle plate, only small pieces could pass through the bottom nozzle plates and up into the fuel assembly. The size and shape of the smallest thermal sleeve sections would prevent them from moving completely through the lower core plate, but could allow them to locate against the bottom nozzle adapter plate in an upright position.

It is considered unlikely that pieces which are small enough to be trapped by the bottom nozzles in this manner would totally block the flow to any one assembly. However, THINC IV predictions (Reference 1) indicate that, even when blockage covers the complete nozzle, full recovery of flow occurs about 30 inches down stream of the blockage.

Thus inlet blockage effects would be limited to the lower portion of the active core, where DN8 and LOCA are not limiting concerns.

5.2.3 Material Within the Fuel Assembly Because of the size of the majority of loose parts considered, most of the parts could not pass through :he bottom nozzle plate. Those pieces that could pass through the bottom nozzle would not pass through the lower grid. This would not affect the DNB evaluations for this core.

Tests (Reference 2) on open lattice fuel assemblies indicate that a 41 percent blockage is acceptable, with disappearance of the stagnant zone behind the flow blockage after 1.65 L/DE. These types of local blockages have little effect on subchannel enthalpy rise and cause only 2616Q:1

,, .. l l

cinsr psrturbatiens in local mass velocity. In r:ality, a local ficw blockage is expected to promote trarbulence, and thus, would likely not affect DNB.

REFERENCES

1. Hockref ter, L. E., Chelemer, H. and Chu, P. T., "THINC IV - An Improved Program for Thermal-Hydraulic Analysis of Rod Bundle Cores," WCAP-7956. June 1973.

2. Basmer, P., Kirsh, D. And Schultheiss, G. F., " Investigation of the Flow Pattern in the Recirculation Zone Down Stream of Local Coolant Blockages in Pin Bundles, "ATOMWIRTSHAFT,17, No. 8, 416-417, (1972). 1 I

i l

2616Q: 1

5.3 NON-LOCA TRANSIENT ANALYSIS Flow blockage by loose thermal sleeves in the reactor coolant system potentially affects non-LOCA transients only in that there is a slight reduction in total RCS flow, as discussed previously in Section 5.1.

An evaluation was performed on the effect of the RCS flow reduction on the non-LOCA transients. In non-LOCA transient analysis, it is conser-vatively assumed that accidents are initiated with the reactor coolant system operating at thermal design flow (TDF). A reduction of 1.13 percent due to the thermal sleeve flow blockage effect on RCS flow still results in a measured flow greater than TDF. This assures that all the current safety analyses remain valid.

5.4 LOCA EVALUATION One may postulate that thermal sleeve material may in the future become located beneath the lower core plate or in the upper plenum of the North Anna reactor vessel. An evaluation of the impact of such material in the lower and upper plenums on the limiting case break ECCS performance analysis (CD = 0.4 DECLG) for North Anna follows:

A. Overall system thermal performance at 100 percent power has been shown to be negligibly different with large pieces present in the RCS. Since thermal design RCS flow can still be demonstrated for North Anna, the ECCS performance analysis previously performed j remains applicable with regard to RCS flow conditions.

l l

B. Thermal sleeves impinged against the lower core plate will remain curved, so there will always be some flow through all of the lower core plate holes, and no assembly will be starved of flow.

i l

l l

l 2616Q: 1

' . - - - . - _ - = :  ::,_== ---

, . _ . - . = -

1

, . Flow redistribution ab:vn a p:stulat:d sle:ve location will occur in the first several inches of the fuel during nomal operation, and )

that therefore reduced minimum DNBR is not of concern in the hot  !

assembly. In a LOCA analysis, post-LOCA thermal-hydraulics pre-dicted for the hot assembly directly define the calculated PCT.

Core flow post-LOCA is characterized by positive (normal direction)

. and negative core flow periods, in that order. From the above, during positive core flow when RCP perfomance determines flow magnitude and direction as during nomal operation, thermal-hydrau-lics should be equivalent to those computed in tne existing LOCA analysis. When the flow reverses any pieces impinged against the core plate will fall off into the lower plenum and thus not be in a position to impact the calculated core flow. Thus, calculated perfomance of the ECCS system will not be impaired by the presence of loose themal sleeve material in the vessel lower plenum.

C. One might postulate, in an ECCS perfomance evaluation, breakup of sleeve material into smaller pieces which become lodged within the fuel and provide additional blockage during core reflood following a LOCA event.

The limiting case ECCS perfomance analysis for North Anna exhibits its maximum calculated PCT when the core flooding rate is less than one inch /second. Appendix K requires a fuel blockage flow penalty to be considered during reflood at such flooding rates so any postulated added blockage from themal sleeve material within the North Anna Plant hot assembly will adversely impact the calculated PCT.

The currently-docketed large break LOCA analysis for North Anna utilizes the February,1978 Westinghouse ECCS Evaluation Model as amended with an evaluation of the impact of NUREG-0630 fuel rod model s. In that evaluation a flow reduction penalty based on 75 percent blockage in the hot assembly is assessed. In fact; the maximum blockage possible with NUREG-0630 is 71.5 percent, so a l

2616Q:1

, , blockaga level 3.5 parcent in excess of tha NUREG-0630 maximum is presently being considered for North Anna. If a small particle of material from a failed themal sleeve were conservatively postulated to enter the hot assembly at North Anna during a LOCA and become lodged at the coplanar locus of blockage from fuel rod ruptures, the -

maximum additional blockage over and above the NUREG-0630 defined

, value would be that due to completely closing the limited flow area remaining between two adjacent rows of rods in the assembly or roughly 1.8 percent. This added blockage from a postulated thermal sleeve particle is more than accomodated by the excess blockage assumed in the existing NUREG-0630 evaluation for North Anna; no further PCT penalty need be imposed.

! To summarize the above, predicted core themal-hydraulics post-LOCA are independent of the postulated presence of thermal sleeve pieces against the lower core plate at time zero. The presence of small pieces within the hot assembly could adversely affect the calculated PCT in the l 10CFR50.46 Appendix K analysis, but such an effect has been shown to be accounted for in the existing North Anna analysis as docketed.

The hot leg might also contain loose parts caused by a breakup of the pressurizer themal sleeve. Due to the plethora of guide tubes, support columns, etc. in the upper plenum it is highly unlikely that any piece could orient itself in such a way as to significantly block flow exiting i any particular fuel assembly. The pieces in the hot leg are not of concern from the standpoint of ECCS perfomance.

One might postulate that a pressurizer surge line themal sleeve propelled by post-LOCA blowdown forces might damage a particular guide tube in the North Anna Unit 2 upper internals as discussed in Section 4.3.1. In the Westinghouse large break LOCA Evaluation Model the conservative assumption is made that no credit be taken for insertion of

[ the control rods. Thus, failure of a guide tube will have no

.significant impact on the North Anna Plant ECCS performance analysis

limiting case.

I 2616Q:1

=- .-- - =: == =:: - =- -

= - = -_a.

~. ._

. s Finally, thm blockage of steaa g:nerator tubes in tha loop containing the pressurizer surge line was also considered. In this case, it is assumed that segments of the pressurizer surge line thermal sleeve are held against the steam generator tube sheet by reactor coolant pump flow prior to a LOCA. In this situation, during the initial part of the LOCA transient when the RCS is still in forward flow due to the influence of the RCPs, the core thermal-hydraulics should be equivalent to the existing LOCA analysis. When the steam generator channel head voids, the thermal sleeve pieces held against the tube sheet will fall off into the channel head and n'ot be in a position to affect flows in the reactor coolant system. Thus, the effect of the postulated sleeve segments at the steam generator tubesheet will nat significantly affect the 10CFR50.46 Appendix K ECCS analysis for North Anna.

I

(

l 2616Q:1

- ._,- . - . - .