ML20055J219

Safety Evaluation Re Review of Cen 387-P, C-E Owners Group Pressurizer Surge Line Flow Stratifiction Evaluation. Rept Found Inadequate

ML20055J219
Person / Time
Site:	Millstone
Issue date:	07/25/1990
From:	NRC
To:
Shared Package
ML20055J213	List: ML20055J212 ML20055J219
References
IEB-88-011, IEB-88-11, IEIN-88-080, IEIN-88-80, NUDOCS 9008010212
Download: ML20055J219 (10)

Text

, . ._. __ . . _ _ _ _ _. . _ _. _. _. _ _ _ _ _ _ _ . _

Enclosures 1.

m. .

Lage<lif7 P t

,s

, j

- ~ '

- ,, '~

, , ENCLOSURE,1 ,f C ' ' '

.m ~ *

[ , , . . .

' ^ '

.. REVIEW 0F J ' '

S COMBUSTION ENGINEERING OWNERS GROUP ~(CEOG)" J .

• s i

' PRES $URIZER SURGE LINE FLOW 5TRATIFICATION EVALUATION'4 <

l c o CEN-357-Pt JULY 1959 ,.

' I 4 . . .

( r

^

- INTRODUCTI'ONL I ,

, i'

. .. a, Thepressurizersurgeiline'(PSL)41n'thepressurizedwaterreactors " .

-(PWRs),isastainlessisteel' pipe,connectingthebottomofthe-w.'

pressurizer vessel of theihotJ1 eg;of. the coole - Toop. The out flow of the pressurizer water'is generally warmerithan : .;; hot leg flow, ;Such 4

. temperature differential (delta T) varies with plant operational .

activ-ities and can be'as high as1320?F during the' initial' plant.heatup. .

Thermal ~ stratification;is.the separation of the hot / cold flow stream in-the. horizontal portion of'the PSL"resulting in temperature differences at_

the top and bottom of:the: pipe.' Since thermal stratification is the

' direct. result of the difference in densities between the. pressurizer and' l

the hot.-leg water, the: potential for stratification 11s increased as'- ~

system delta-~T increa'ses and as-the.insurge orfoutsurge; flow decreases; Stratification:in PSLs'was'found recently and confirmed by data measured from several PWR plants. ' '

Original design analyses did notiinclude any stratified flow -loading.

conditions. Instead it' assumed complete sweep of fluid along the line during-insurges or outsurges resulting41n' uniform thermal loading st any .

particular piping location. Such analyses did not. reflect PSL' actual ,

'o

. thermal condition and potentially may: overlook undesirable line .

~ deflection and its actual; stresses may exceed design limits.-In <

1; ' addition, the. striping phenomenon, which is the oscillation of the hot and- cold stratified boundary,t may, induce high' cycle fatigue to the inner  !

pipe wall, 'needs. also to be analyzed. Thus assessment of-stratification. ~

effects.on PSLs is necessary to ensure piping integrity"and ASME Code Section--Ill.conformance.

STAFF EVALUATION Since-stratification in PSL is a generic concern to all PWRs an NRC Information Notice 88-80 . s issued on October 7, 1988 and then=an NRC-Bu11etin 88-11 for the same concern was also issued on December 20, 1988.- Combustion Engineering on behalf c'f the Combustion Engineering Owners Group (CEOG),<has performed a. generic bounding evaluation report.

CEN'387-P-(Reference 1), which documents'the results of the PSL stratification effects. The following is the staff's evaluation of the Combustion Engineering efforts and information provided in the report.

1 P.

, ----w,.,.,,,,-,e , - , - - , , . ,----m= .,--..-...-w* . . . - , . u r- - - + .w--*---.r,.vc.m.w. --M

g Page 2 of'7 JA). Plant monitoring an'd update of design transients.

.As'a: result of the INPO Safety Evaluation Report, which was issued'in-September 1987DandLidentified concerns associated with the stratified flow in the PSL, the CEOG initiated surge line temperature collection '

data at [ ,

K - Concurrently- with this effort.. [:

i data on PSL at []Linite ted efforts also '

for the

3. This wascollection'of later folded temperature

,into-the-CEOG effort. Inaddition,'[~;. ~ .

)alsocollected-similardatafor[- . .), < af ter the. CE06 Task " Reduction ~

' and analysis of Pressurizer Surge < Line data collected from CEOG plants'

'had comenced, and= submitted them to Combustion Engineering for review)

-and comparison with-the data already collected from the first two CE0G plants.. '

,With the exception' of.[ . .  ; ),whichwasabletoretain the temperature distribution data only.after the bubble was formed in,the -

pressurizer, the other two plants were able to retain the tem >erature '

distribution data _d0 ring heatup and until normal operation. ; )

obtained displacement readings also, in addition to temperature. ,

The Owners Group'is ' going to ' decide on a proposed task to collect data 1and during)thenextcooldownatboth[-The staff requests that monitoring shou cycle. Data should be obtained'and evaluated to determine whether the observed thermal: transients are bounded by the transients assumed.

Due to similar design featu'res of all the CEOG plants (10. plants.,15 units), the data obtained were deemed adequate and CEOG met with NRC staff on February 13, 1989, to discuss the scope of the ' Task" and how the Bulletin's-requirements-will be addressed.

All CEOG p5Ls are similar in< layout. They consist-of a 12" (except for

[ which is a 10") stainless steel schedule 160 pipe, with averticaldropf] rom =thepressurizertothehorizontalrunofpipednda-verticaldroptothehotlegnozzle(exceptfor[ ]whichisat-a 60' vertical angle drop).

A review of the-data, which measures. pipe wall outside temperature variation with-time, indicated that'the largest surge line top-to-bottom-temperature differentials were similar for the three plants and caused either by an insurge or_an outsurge of tae. pressurizer. Therefore

' emphasis was given to these transients for evaluation. Surge-line movementsin[ . .),werecalculatedandcomparedtoactual i pipe movements measured at three locations.

The deflections predicted by the analysis model were based on a stratified flow model with a pipe top-tc-bottom delta T=320'F. The actual measureddatacollectedat[,dwhenthefluidinsidethepipe],were.obtainedd top-to-bottom delta'T=181'F an -

approximated a uniform temperature distribution model. Even though the

...,.--r.--..,-..,,-.~.,...,..,,.,_.-....w- -

.,v . , - - ...m.m., - _ .....,,,,m.

- ., .wa ...

o t  !

1

'Page 3 of 7- ]

L a'nalysis rodel predicted the:same general shape as the measured data the L

fluid conditions inside the pipe were not similar. The staff feels that

- further investigation and/or comparisons,are required to predict PSL L displacement' behavior.

l The data obtained from all=three plants reco'rded outside pipe wall-surface l temperature distribution about the longitudinal,and circumferential axis

' of the pipe. In order to determine-fluid conditions for the design' basis events at the inside' surface of the. pipe wall, a 2-D finite ~ element heat-l transfer analysis of the pipe cross section was1 performed.

Two. bounding analytical heat transfer mocels' with various inside fluid- I conditions were developed, with~an attempt to reproduce the recorded .

outside pipe wall surface temperature distrib.ution.

-1)Astratifiedflowmodel.

2) A uniform temperature gradient model The stratified flow model assumed the hot (pressurizer temperature) fluid m the upper half of, the pipe, and the cold (hot ' leg temperature) fluid in the lower half of the pipe, with a sharp. inter, face in between. During the outsurge it was assumed-that flow occurred in the upper portion-of' the pipe only, while during the insurge it was-assumed that flow occurred

. in the lower portion of-the pipe only. 'For a given transient, a flow L

rate was calculated based on the pressurizer level change vs. t% plots, and a heat transfer coefficient was then determined ~  %

For the uniform temperature gradient model the pipe cross sectional area-was divided into a finite number of water layers to approximate a continuous temperature gradient. The uppermost layer was considered _the hot fluid (pressurizer temperature), and'the lowest: layer was considered the cold fluid (hot leg temperature), with the intermediate layers having.

a uniform temperature gradient. It was' assumed that flow. occurs at the l

full pipe cross section _during an outsurge:or an insurge. During.a given transient, a flow ~ rate was calculated' based on=the pressurizer level change vs. time plots and a heat transfer coefficient was then determined.

Based on the above coefficients, and using the in'-house CEMARC computer code, a'2-D finite element model was developed to determine the inside pipe wall temperature distribution for both the stratified flow and the uniform temperature gradient models. The temperatures at selected nodes-were calculated and compared with the thermocouple data. The uniform temperature distribution model more closely approximated the measured results. This indicates that it does not appear to be a sharp hot / cold interface, and it is more likely that there is some mixing of the hot and.

cold fluids with-a uniform temperature gradient from top to bottom of pipe. Changes were made to the' stratified flow model to better match the measured data. These changes tended to better match the. measured data for the outside pipe wall temperature distribution, but CE could not i I

explain why these would be valid _ assumptions. Since a unique solution could not be derived, assumptions were used for the thermal striping, stress and fatigue evaluations utilizing the stratified flow model.

1 l

4 'j

. l l

Page'4 of'7 1 ASME Code compliance for Stress and Fatigue <

l

1) CodeComplianceinStress(Inelastic'Analysisi.

' Each plant specific surge line was reanalyzed by the SUPERPIPE? I 1

computer code using a_ bounding generic stratified flow loading.

Elastic analyses were performed on the plant specific pi>ing layout and support configuration for each_ plant, considering that tse maximum delta T *

, for's given transient, occurs ~along the entire horizontal-length of.

pipe. These results were used to: choose'a specific surge line for the~

bounding inelastic analysis.u The elastic, analyses pred' cted stress-intensity' levels in excess of the 35' allowable limit of the ASME Code -

. Section !!!, NB-3600, equation 12. Thusan#inelasticshakedownanalysis I was performed as per NB-3228.4 to determine if,af ter a few cycles 'of ' load,

. application, racheting and progressive inelastic deformation ceases.'

However, the PSLnozzle1 moments:were calculated from the.SUPERPIPE elastic l analysis.

g ASME Code ' stress indices were used for_ each pipe- component for- the plant.

specific elastic analyses. The bounding inelastic' analysis was based on a Finite Element shell model and therefore, the stress indices were:

' inherently included in the analysis..

^

The $UPERPIPE computer code was used to performed the initial dlastic analysis, which' considered thermal' effects of the- stratified f. low overL the entire horizontal length ~of pipe, for delta T=32'F, delta T=90'F and delta T=320'F. For each structural model, a uniform fluid: temperature -

loading:and a stratified flow loading were applied. Three types of ,

stratified flow effects were investigated.

L a) 1.ocal stress due to temperature gradientLin the pipe wall. )

b) Thermal gradient stress across pipe wall'due to transient condition.

c) Thermal pipe bending moment generated by the restraining effects of supports.

Actual support stiffnesses were used considering a'* 2' limit of sprirg l motion,' beyond which springs will act as rigids.~ The maximum ~ movement l 1 -based on delta T=320'F, pipe top to bottom stratified flow, was calculatedfor[ ] and [ _,

3,bothat location'H2. ,

The staff feels that since no plant specific support data and displacement limitations were considered, further evaluations are required to justify the [ ,

) inelastic analysis as the worst case.

In addition, it is the staff's opinion that the assumption on-spring motion may not be conservative, in that, upward movement of a spring: i which e'xceeds it's travel range will cause the spring to unload and redistribution of stresses will occur.

i-

, -,. <.m. -- -

y ,

i ...

l 1

4 Page 5 of 7 The[ . . . 1PSLconfiguratiionLwaschosenforlthe,ineldsticevaluation, ,

.since it predicted the highest stress levels =under the elastic, analysis ~. ,

While each line will; behave differently under a given stratified flow-  !

loading, it'was concluded that the surge line with'the highest elastiet stresses will provide an upper bound for all other. lines. -This was:

verified by;the~ fact that the most highly stressed region is the rue location for both~the. elastic'and the inelastic evaluation. For this i line, the elbow.under the: pressurizer was determined to be the.most critical location.

- Material properties as T=650'F were used consider _ing the strain hardening behavior of-the material. The stress strain curve used was developed by Combustion-Engineering based on the ASME code minimum yield stress value and plastic strain.

4Three complete cycles of heatup.. steady state and cooldown were analyzed.

For fatigue-evaluations, the maximum principal-strain range values were ,

calculated from the maximum and minimum principalJstrains. The maximum positive principal strain was -calculated for three cycles and-extrapolated to be less than 25 after 500 heatup/cooldown cycles, based on the decreasing rate ~of strain increases with additional cycles. The analysis:results demonstrated that the first cycle undergoes significant permanent strain.with subsequent cycles having smaller: accumulation. The' strain range from'the first two-cycles was considered in the fatigue analysis with the strain range from the-third cycle used for the-remaining 498 cycles.

Review of. Fig. 3.6.2-8 and Fig. 3.6.2-9 of the report could not clearly demonstrate that strains were stabilized after'the three heatup/cooldown cycles and that progressive distortion.does'not exist.

l Changes in plastic s' trains showed some decrease with _each cycle but the staff concluded that additional investigation was required to-demonstrate that the decreasing rate-of plastic strain will' approach zero. Since there are'no maximum strain limits prescribed in the ASME Section 111 code,'the value of 2% was obtained from the High Temperature Code Case N47 and'it was used as a guide.for-the maximum positive-principal strain limit. The staff concluded that the use of 25 strain limit in this case needs further justification.. .

Il Code Compliance in Fatigue.

To determine stresses at the.insi.de face of the pipe wall due to fluid oscillation at the interface of the. hot to cold boundary (strping), a 1-D finite element analysis was performed. The input assumptions used in

'. this analysis were based on the measured data from the CEOG plants, and i

other.information available in the public domain. The thermal striping Is l

)

. l

l, Page 6 of 7 model considered the. hot fluid at the Pressurizer temperature, the cold fluid at the Hot Leg temperature, and a sharp interface with no mixing of the hot and cold fluid.- A sawtooth fluid ~ oscillation was assumed to occur across the interface region.

1 Results indicated that fatigue damage due to striping is insignificant when compared to all the other causes of fati ,1.e. static thermal I stratification, thermal transients etc.)gue damage. . The CE report indicated that -l based on the stress-levels caleviated, an infinite number of allowable cycles exist and thermal strising is not a concern.- Since maximum stress duetostripIngoccursatthetot/coldinterface,'whichisnearthe horizontal axis of the pipe, and maximum stress due'to fatigue occurs at the top and bottom of the pipe, these stresses do not occur at the same location and are not additive. The staff feels that further investigation should be provided for the use of a fraction of the striping amplitude.

In addition, data based on. measurement outside the pipe may be inconclusive for the purpose of defining the striping phenomena.

Analysis for cyclic operation (fatigue) was performed, in addition to the shakedown analysis. Using-the results of the inelastic analysis, the maximum principal total strain range which occurs from shakedown analysis was multiplied by one half the elastic modulus to determine the equivalent alternating stress, as per NB-3228.4 (c). This maximum strain range occurs af ter cycle 3, and this value was assumed for the remaining cycles.

For the first two heatup-cooldown cycles, the larger of cycle 1 and 2 strain range was used. .

The cumulative usage factor for this generic bounding analysis was determined to be 0.21 for .] . The maximum cumulative usage f actor, when the eff[ct e of the :t 2 displacement limitation was' considered,was0.36for[l  :) . The staff feels that further evaluation is required to justify the [ ]inelasticanalysisis the worst case.

CONCLUSION Based on our review, we conclude that the.information provided by Combustion Engineering in References 1 and 2 is not adequate to justify continued operation for the.40 year plant life. However, the staff believes that there is no immediate or short term safety concerns associated with the stratification effects for continued plant operation until final: resolution of the Bulletin 88-11 is issued. This is scheduled to be completed by the end of 1990 and should also address the Code acceptance criteria of ASME NB-3600.

Concerns that the staff has are the following:

a) The ASME code acceptance criteria of Section NB-3600 Equations 9-14 need to be satisfied as applicable.

- -, . - - . _ - - , - - . - + _ - , - ,-. ~ .....-.-_,.,-,v. , . vw..-c., .- ..__,,y.-..-.w, , . - -

't 1 1 1: - '

, , o -

/ i. J

,,s 7 ,, , o , ,

,a y, '

y da -

, page,7 of 7: .j w y; +

c r 4

+ l 4 0 ' " ?; ^

,, . _ c

h);.'g11.

supp.ortsdnclu'dingpipesh'iprestriIn'ts,be;consideredforthe effects:of providing_any additional constraints.to the. Surge Line,-lIn,

, f"' (the plant' specific-or.the bounding; pipe stress

• evaluation. :c

l S,'

c) _A11/ supports.includingpipewhiprestraints,;requNeplant" specific.1 -

f^ confirmation of:their capabilities',' including ~ clearances,land that they" 4

?

.fa11'within-thel bounds.ofothe analysis.; ' '

s- ",i

.. '. . ~ '

Justifythe[- a

. d )'. 1 stressandfatiguefora]11CE0G.' plants,jincluding[1 inelastic ,

.]. a'nalysis ,

as t t

e) PSLdisplacementbaSivior.predic'tedb the1 analysis:model '

and;the Justify 'use of a fraction'of the; striping; amp itude..

I REFERENCE.S. .

l

1. . Combustion Engineering ' Report' CEN 387-P (Proprietary), fCombus' tion;

. Engineering Owners Groilp' pressurizer surgeLline. flow: stratification:.

evaluation." July.1989.4 '

2. Draft meeting minutes ofithe.NRC audit-oniSeptember 25 and 26' 1989 ,

regarding the CEOG Report CEN.387-P.MPS-89-1048, dated: October 17,-

1989.

• l 3 L

l 5

H 1

i l

l l

1-l

^

y, > 7 -

u

, J ~

w y . _

'pageticof 3'

> [, .

' ENCLOSURE 2'. ' ' '

y -

,;o '

, . ~,i

' 4 '

~

Staff review oft he t CE responses regarding the NRC Audit on-5eptember 25 and 26. 1959 Ref: UE05 Report GEN 357-P MP5-s9-1048 ',

an -

o . dated October 17. 1989.

. j 1

Se'etion-2.0 II)

The staff! requests that monitoring should continue for,a full fuel . '

cycle; Data should be obt'ained and evaluated to determin~e whether the c  !

observed thermal; transients are bounded by-the-transients assumed.

~

l2)

The ktaff" feels that furth'er investigation;is required to predict pSt. i displacement behavior, considering-the: stratification effects. The- , j

- 'def,lection predicted by the analysis model were'. based on'a stratified f, low model:with-a pipe top-to-bottom delta T=320'F. The actual measured

' Jdata collected at [. . ' 1 were obtained during aipipe top-to-bottom * '

< " delta:T=181'F-and when the fluid inside the-pise approximated a uniform temperature distribution model. . Even though tie: analysis model predicted

,the same general shape as the measured data, the fluid' conditions inside . ,

the pipe were not similar. i

.3) Closed.

, a

4) Closed. '

'5) ' Closed.

c

6) Closed..

,, ~)

.7 Closed.

B)

The staff requests that further investigation'is required to demonstrate

~

that strains were stabilized after the.three heatup/cooldown cycles and that progressive distortion does not exist. It is required to. demonstrate

.that the decreasing rate'of' plastic strain will' approach.zero and.the '

peak value will not exceed'a maximum strain acceptance criteria of 2.

The staff feels that the inelastic analysis will be accepted as Justification.for Continued Operation and that'the ASME Code acceptance ' s criteriatofby required section NB-3600 equations 9-14-need to be satisfied; as the' Bulletin.

1 i

____________.------h--

.-s ,

i.

l 9

page'2 sf 3~

, .: ,o-9);LClosed. [ .

. h0)); Closed. '

~

4

11) The staff feels that all supports be considered for the effects of p, including roviding pipe whip additional. constraintsrestraints, in the plant specific or the bounding evaluationt < *

12) Theistaff feels that'all supports.. including p'ipe whip restraints > ,

m, ' ~ requireplantspecificconfirmationoftheircapabilities,. including -

clearances, and that they fall wit.hin the bounds of the analysis.:: ,

, Section~3.'0~ ,,

, 1) Closed. '

~'

2)= ' Closed.

1

3) Closed.

t

4) Closed. 4 l

5) Closed.

6) -Closed. . ,

Table 2.7-1. Closed "

Table 3.2.2-4. See response of Section 2.0-Item 2.

Table 3.4.3-2. Closed. <

, t Table 3.6.2-1. Closed.

- Table-3.6.'3-2. The staff feels'that further evaluation is required to

, justify .th'e maximum cumu[lativegefactor'for.[. usa ] inelastic analysis as the worst case'. The is 0.36.when the-effectsofthe2"displacementlimitationsarecon]sidered.

Figure 3.1.2-5. Closed.

O f'

t )

9 I

e

' .g.

a v t

=

, ' -~f

, . 1 ,

_ r

~- . ,

i , ,.

+ -

, .. .. 1

=-

m o ' Page 3_ of 3- . s, J

9

.4. ,

, e, , ,

+ -

o . .

9

0 *

' Sectio'n44 .

-n oi

' i) . No specific review-was_ performed.

See response of'Section 2.0 Item B.1, ,

n,

'2) -

{

R ,

a$

3);LNospecific;reviewwasperformed.,, . ,

I I r l 1 3 s . i Questions;during< meeting".

c. ,

, I f 1 >- I

_l t g r s c Closed' .

D -

f w .j' s

• 4 4 t g g # 1  %

e I (  %

g 4

I

< 4 e. 3 4 ' d s

e , ( e -g

' ' 4 ' '

9, , i 3! n

'. 4- t 2

, e '$

,'- i

% o c :-

I e 14 k #

t . .

4 g , t 4

r J t < 4 -

• 6 e l f i

t

'x 5 f

L

\ d 1 4 4

4 1

I a \

i .

s . .

f

(

1 e

^\ l f

\

1 1 g 4 i i

i J

d g- II

' g.

L t

s i

i /

f f

I t

.. _ . . ., ,.,c ,. - # . . . . .

& . , , ,