ML20138D067

Nonproprietary WCAP 10950, Tubesheet Region Plugging Criterion for Full Depth Hardroll Expanded Tubes

ML20138D067
Person / Time
Site:	Summer
Issue date:	09/30/1985
From:	WESTINGHOUSE ELECTRIC COMPANY, DIV OF CBS CORP.
To:
Shared Package
ML19344C066	List: ML20138D021 ML20138D035 ML20138D048 ML20138D067
References
WCAP-10950, NUDOCS 8510230162
Download: ML20138D067 (33)

Text

mgu vr n e r _ - _ _ _ _ _ _ _ _ ---

41 i 1h , l a

• 'ESTINGHOUSE PROPRIETARY CLASS 3 i

10950

@ MAT put51 WBESHEE"i Fi::0N P1DCGING CF. ITER:0N FOR F2'U. DEPTH HAP.DRO:.l. EXPANDED W5ES September, 1985 -

WESTINGHOUSE ELECTRIC CORPORATION STEAM CENERATOR TECHNOLDCY DIVISION P. O. 301 835 P:.~5'"RCH, PENNSYLVANIA 15230 I ,

1 A

'j 8510230162 851016

.o PDR ADOCK 05000395 Y p PM

.E ,

WESTINGHOUSE PROPRIETARY CLASS 3 s

C0h7DTS TOPIC , PAGE 1.0 INTROD"CTION 2 2.0 EVA:11ATION 3 2.1 Maintennner of a FLxed Tube End Condition 3 2.2 Limitation of Primary to Secondary Leakage 6 2.2.1 Postulated Accident Condition 7 Leak Consideratinns 2.2.2 Operating Condita.n Leak Considerations 7 2.2.2.1 Tube to Tubesheet Hardroll Effect 8 on Leak Race 2.2.2.2 Operating Plant Leakage Experience 10 for Tube Tubesheet Cracks v

2.3 Tube Integrity Under Postulated Limiting 10 Conditions 2.4 Exclusions 1.1 3.0 SUMP.ARY 12 1

Table 1. Tub'e sheet Pullout Summary for P* Criterion. 14 Table 2. 'En'gege=ent Langth Required to Resist Pressure 15

• Pullout Forces.

Appendix I - Tubing U Bend Cap Study, Vestinghouse 14

, Preheater Sten Generators.

Appendix II - Disposition of Tubes with Indications 23 Above P*.

Appendix III - Tube Vall Te:peratures of Plugged Tubes. 26

• Appendix IV - Preload Testing of Rardrolled Tube to 28 Tubesheet Joints. .

Table IV.1 - Model D Sten: Generator 30 Tube Roll Preload Test.

1 Page 1 of 33 I  ;

WESTINGHOUSE PROPRIETARY CLASS 3 TUBESP!W Plc!Ou P1PGGING CF2EF'DN FOR Mril DFvTy HAn?:*1 E7?ACET TFES 1.0 1r? D":"!OV The purpose of this document is to address the issue of repairing or plugging full depth hardroll expanded stes= generator tubes when degradation has been detected in that portion of the tube which is within the tubesheet. Existing plant Technical Specification tube repairing / plugging criteria apply throughcu:

the tube length and do not take into account the reinforcing effect of the tubesheet on the sxternal surface of the tube. The presence of the tubeshee:

vili constrain the tube sn:' vili cc:ple=ent its integrity in that refi=~ :?

essentially precluding tube defor=.ation beyond its expanded outside diameter.

The resistance to both tube rupture and tube collapse is significantly strengthened by the tubesheet. In addition, the proxicity of the tubesheet significan 17 s!fects the leak bahavior of through wall cube cracks in this -

region , i .e. , no significant leakage reistive to plant technical specificaticn allovables is to be expected. Based on these considerations, the use of c alternate criterion for establishing plugging margin is justified.

This evalvetion forms the basis for the development of a criterion for obvisting the need to repair a tube (by sleeving) or to re.ncve a tube fro =

service (by plugging) due to detection of indications, e.g. , by addy current testing (ECT), in a region extending over avst of the length of tubing within the tubesheet. This evaluation app 11es to Westinghouse model D2, D3, DL, E, and a fev model 31 steam generators and assesses the integrity of the tube bundle with ECT indications on tukas within the tubesheet under normal operating and postuisted accident conditions. Three aspects of bundle integrity sze addressed: 1) aaintenance of a fLxed tube and condition in the tubesheet with the limiting case of a circumferancial Indication near the top of the tubesheet, 2) 1Laitation of primary to secondary leakage consistani vich accident analysis assu ptions, and 3) maintenance of tube integrity under postuisced limiting conditions of primary to-secondary and secondary to primary diifevential pressure.

The result of the evaluation is the identification of a distance, designated Pc and icentified as the P* criterion, below the scp of the cubesheet for which tube degradation of any extent does not necessitate recedisi action, e.g. ,

plugging or sleeving. The P* criterion provides the same level of protection for tube degradation in the tubesheet region as that afforded by Regulscory Guide (RG) 1.121 for degradation located outside the tubesheet region.-Is is siso to be noted that while the P* criterion was derived based on svallable clearances for tube motion in the absence of restraining forces due to the tube.

to tubesheet hardroll preload, it siso provides for su!!1cient angsgeoent fcr that preload to successfully resist the pressure induced pullout forces that could be developed during norr.nl or accident operating conditions. Table 1 provides a list of P* values applicable to Westinghouse model D (D2,D3 & D4.'.

E, and 31 s t en: generatort  !.icitations on the usage of the P= critericn are discussed in the evaluation section of this report.

Page 2 of 33

9 WESTINGHOUSE PEOPRIETARY CLASS 3 2.0 EVA'l'A' ION 2.1 MAKEMAVCE OF A THED Tv?? END CONMTION is The criterion for maintaining a fLxed tube and condition in the tubesheet P*, below the top of the deter =ined by evaluating the siniana distance, that vill provide margin against pullout of the tube for a postulate:

tubesheet This evaluation conservatively "circumferentia1* separation of the tube.

considers 1Laiting loading conditions associated with both normal operation sn postuisted design basis accidents.

In the developme':t of the P* criterion, the separated tube is conservatively postuisted to travel upwards to a position where the extrados surface of the U. bend spex is in contact with the tube directly above it (one rov higher in

~ The vertics1 dispiscenent of the separated tube disthe limited che same column). b intset to the actus1 asnufacturing clostance between the separated tu e anStudie tube above it in the U bend.

clearance between ad) scant U-tubes any exceed the nominsi drawing b d gap value; therefore, the geometric clearance has been based on the nominsi U- en Appendix 1 to this report plus annufacturing tolerance considerations.

provides a complete description of the methodology used to determine the manuf.~cturing colorsnce values which have been added to the nominsi ciestsace.

On this basis, the vertical displacements shavn on Table 1 were developed.

Once the sepstated tube is assumed to have noved into contscc vich the identintact tube directly above it, its further vertical displacement is bothThecoincpotentisi with and 111ticed by any vertics1 extension of the intact tube.

for any superimposed effect(s) of thermal and differentisi pressuretube o displacement of the intact tube are considered in the following paragraphs:

1.) 7t.ermal effects do not cause any net change 2n the Pk to its outboard clearanc neighbor

&alue for the inboard the tubeinboard relativetube is considered to be separatec regaralers of whether on either the cold leg (CL), the hot leg (HL) or on both legs.

Thermal elongation of a tube is a function of the temperature a11 erential and the tube length, in this case, both the cold leg or the intact (outboard) tube and the separated (inboarc) tube are at the same temperature. Both hot legs are also at the same temperature. Furthermore, the lengths of both the separated inboarc and intact outboard tube are identical except for the slightly larger.overall length of the intact tube attributable to the larger U-bend radius. Therefore, separated tube elongatzon attributable to thermal effects does not contribute increase its value beyond the to the determ2 available geometric clearance.nation of P I

Page 3 of 33

s a l

l vistzscxovst raorRIE:An cuss 3 l l

2.) Axisi displacement due to the pricary to secondary differentini pressure vill result in axisi stretching of the Lntact tube one row above the separated tube. The lialting condition of primary to-secondary differentisi pressure occurs during a postuinted feed 1Lne break (T:2) accident and has been conservatively determined to be 2650 psi. Under this loading, the intact tube vould stretch, siloving a presu=ed equivalent vertical displacement of the separated tube end as specified in Table 1.

Considering the model D3 stess generator as an exs=ple, the tube ID is nomins117 [0.664)s,c.e Inch, thus the cross section ares is (0.3463)s,c.e square inch. The langth of the straight leg section of the tube is approxiantely [213)s c.e Laches. Thus, for a T12 pressure load of [2650)s,c.e psi che extension of the leg would be [0.091)s,c.e inch.

using the folicving formuis*

P 1. D12 d=

E Do 2-D 12 Rovaver, if the tuoe is considered to be of anxisum CD r.nd minir.:=

thickness, [0.734 and 0.039)s,c.e inch respectively, the ID chen beco=es (0.680]s,c.e inch. Further considering the design allowance for erosion and corrosion of [0.001 Inch for the outside of the tube and 0.002 inch for the inside of the tube)s,c.e yields diamete:s of [0.680 and 0.732)s c.e for the ID and CD respectively. The elongation of the tube due to the TLJ pressure then becomes (0.110)s,c.e Lnch. For the model D.'.

stes= generator this value was found to be (0.121)s,c.e Inch due to the longer straight les langth.

The offacts of 1Laiting bending loads and tube rocking on vertical displace =en:

of the tube end in the tubesheet have siso been considered. For this case. the loading condition vich the limiting effect at the tubesheet is safe shutdown

• earthquake (SSE). For a tube vich a anzlaua span length betveen supports, tube

} uplift effects due to SSE loading have been determined to be negligible (on the order of [0.001)s c.e Inch for the combined effects of direct uplift and sxis.'

j contraction due to bending.

} In addition the the bending loads associated with SSE, an axisi contraction of the tube due to 1 sterni deflection can siso occur due to the cross flow of fluid in the stes= generator. The loads will be the highest where the fluid is discharged through the lover vrapper opening. The latersi loading associated with the fluid flow was found to be (0.239 lbfinch)s,c.e fa, , ,,g,3 p3 ,ges:

Page 4 of 33 l

1 1

8 b'ESTINGHOUSE PROPRIETARY CLASS 3 generator. The bending deflection of the tube vould then be on the order c!

[ Ja,c.e inch and the associated axisi contraction vould be [ Ja.c.e inch, which is considered to be negligible. On the basis of the results for the I mdel D3 evaluation, neither of these effects have been included in the )

deter =ination of P* for the generators considered in this study. l Dented tube conditions are bounded by the criterion evaluation in that the constriction (s), hence, restraint, at the tube support piste(s) would act to \

licit the axisi movement of an individus1 cube to a value determined by the axisi mve ent of the tctal bundle. Thus, the Initisi reistive ution between the separated tube and the next adjacent higher tube vould not be expected !ct a scen: generator with danced tubes. The resultant axial move =ent of the separated tube vould be that associated with a:cident condition loading only.

and is less than the P* criterion. No credic vas taken, however, for restricted vert l cal displacement due to possible denting conditions of the l tubea l

l Combining the various components of possible vertical displacement of a separated tube results in a 1Laiting value. This value is the P* criterion !ct minia.:= scceptable distance below the top of the tubesheet of a presuced tube separation. If the distsace of the tube degradstion below the top of the tubesheet is greater than the above criterion, remedial action such as plugging or sleeving need not be taken. -

The P* criterion for repairing or ramoving tubes from service is considered to be conservative. The basis for the analysis is the assu=ption that the degradstion is oriented circu=!erentisily; however, circu=!erential cracking ,

has not been observed or confirmed by Westinghouse in the tubesheet region of l docestic steam generators. It is noted, however, that severs 1 Indications reported for one domestic plant in one sten: generator could have been interpreted, using either the standard bcbbin coll ECT probe or an 8 x 1 probe.

as either being axially or circu=!erentis117 oriented. One kovn occurrence c!

circu=!erentisi cracking in a stess generator Ln a non domestic, non.Vestinghouse plant was due to the incorrect application of a hardrolling technique known as

• kiss ro11 Lng.*

i There have been five kovn occurrences of circu=!erentisi cracking being reported for the tubesheet or transition stes. Of these, four have been reported in foreign plants and one instance in a docestic plant. These are as l fc11cvs:

i l 1. 14o indications were found in the roll transition region of a single tube pulled from the French Da= pierre plant. One of the Indications had progressed 90 percent through the tube vali and had an astmuchal extent of 360 degrees. Another sas11er Indication was found siso in the transition region of the same tube. This occurrence of circunferential cracking was in a non Westinghouse plant and was due to the incorrect application of a hardrolikng technique h avn as

• kiss-ro11 Lng.*

2. Two tubes with Indication in the transition region vere found in the Doel plant in Belgiu=. In both cases the cracks were predo=inantly Page 5 of 33 l

Tr.STINGHOUSE PROPRIETARY CLASS 3 sxially oriented with very sas11, approximately 0.030 inch, extensions in the azinachs1 direction at one and of the indication.

3. One Lndication within the tubesheet . region of a do=estic plant has bee .

In this esse, however, the reported as being circu=!erentisi in nature.

Inspection technique was not that required to absolutely identL!y the indication as being circu=!erentis1. That is , i t wa s found wi th an S 1 1 probe, in which five of the coils indicated the presence of an indication. It is just as likely that the eddy current indications cou:d have been Lnterpreted as being five sxisi Indications. In this case the reported orientation of the indication (s) is at the descretion of the inspecting agency or utiiity representative.

On the basis of known occurrences of ci:cu=!erentisi cracking within the tubesheet region, the criterion is considered to be addressing an extre:ely low probabllity phenomenon. The usefulness of the criterion sce:s from the repisce=ent of any applicable structurs1 integrity criteria by a single parameter, P*,

a nessurei value, and eitninstion of the need to deter =ine the crientation of indications located within the tubesheet. Reistive to the assumptions azplicit in the development of the criterion In addition to the postuisced tube gap geometry and conservative loading conditions considered, additions 1 margin is provided by the resistance to vertics1 displacement of a separated tube that develops due to interference betveen the tube and the tubesheet (this effect is discussed in AppendLx IV to this report), the tube support pistes and the anti-vibration bars. Rovever, no credic was taken for this in the development analysis for the F* criterion.

In using the P* criterion for determining whether or not remedisi actions need to be taken for tubes with Lndications in the tubesheet region, the following considerations must be taken into account:

a) As discussed in the description of the develop =ent of the criterion, asintenance of a fLxed tube end condition within the tubesheet is an issue only for a poscuisced separated tube condition. If an Indication on a tube within the tubesheet has been determined to be axis 1, pullout from the

' cubesheet is not an issue. The disposition of tubes with axisi or circu=!erentisi indications located above the P* criterion should be made according to plant technical specification requirements, which are based on USNR: Reguistory Guide 1.121, " Basis for Plugging Degraded PVR Stes:

Generator Tubes.*

b) A second consideration in the application c! the ?* criterion reistes tc i che assu=ption in the develop =ent that the tube one row above the postuinted separated tube is structurs117 intact, i.e. , an intact tube does not contain indications which would result in its inability to resist the

}

load Lcposed by the separated tube. The case of two ad) scent tubes :n the i ss=e column with significant circumferentisi Indications has not been considered in the development of the criterion due to the low probability of the simultaneous occurrence of the postuinted limiting conditions, i.e. ,

the maximu= cube gap between adjacent U-bends. *circu:!erential* 11 citing Page 6 of 33 I

_ ^" 7 3 WESTINGHOUSE PROPRIETARY CLASS 3 degradation on adjacent tubes, postulated separation of adjacent tubes, ang zero frictional restraint of separated tube ends . On t' ' asis. the consideration of the outboard adjacent neighbor tube nc. ~1ng intact is judged to be of very low probability. This issue is further discusse following these subsection paragraphs.

c)

A third consideration relates to the uncertainty in the ECT determinaticn of thethe location of an axial or circu !erential indication below the tcp ci tubesheet. To maintain conservatism in the criterion, it must be applied to a reported location, of a presumed "circu=!erential" indicatien, below the top of the tubesheet which conservatively accounts for Ec7 sig a:

Interpretation uncertainty as determined to apply at the ti=e of the inspection. For exa=ple, for a standard bobbin coil probe a conservative eddy current uncertainty could be postuinted to be [ ]* C**, or a program could be undertaken to make a more exact determination, i.e. , l

\

corresponding to the type of ECT probe used, of the appropriate uncertainty value to be included in the determination P*.

A potential limitation to the use of the ?* criterion could be postulated to occur Ln the case of adjacent tubes in the saae column. The assu=ption of circu=!erential cracking in each of the tubes leads to the consideration that ,

for the Inboard tube, the neighbor affords no restriction to vertical travel.

This is not, however, considered to be a credible assumption. Reiterating the discussion on the conservatism of the criterion, the operating experience of domestic plants has revealed, as observed or confirmed by Vestinghouse, no circu=!erential cracking Ln alli annesled Inconel 600 full depth rolled tubes.

In developing the criterion, it has been considered prudent to postulate th, existence of extensive circumferential cracking for the single tube under consideration. At the same tLae, it is not considered credible to postulate l

circu:!erentisi cracking of two adjacent tubes. Any cracking that has been observed in operating units has generally been located in the non expanded to expanded transition of the tube hard roll and/or at skip roll locations in the tubesheet.

The observed cracking has sivsys been typified as short and axially oriented. On this basis, also, an intact neighboring tube can be defined as a tube'which has not been degraded beyond Les' structural itait over an azicuchal extent of 360* in the free span, i.e. , outside of the tubesheet region.

2.21.1MITATIOV 0.r PP1M AfY.70.SECOMAJY IIA. TACE

The allowable amount of pris.ary to secondary leakage in a steam generator

, during normal generally plant to 0.35 gpaoperation

. is ilmited by plant technical specifications, a This limit, based on plant radiological release considerations and Laplicitly enveloping the leak before break consideration h for a through vail crack in the free span of a tube, is also applicable to a l 1eak source vichLn the tubesheet. In evaluating the primary to secondary p

leakage aspect of the P* criterion, the relationship between the tubesheet region leak race at postuinted F12 or steam 1Lne break (51.3) conditions is assessed relative to that at normal plant operating conditions. The analysis was perforced by assuming the existance of a leak path, however, no actus1 leak path would be expected due to the hardro11 Lng of the tubes into the tubesheet.

\ '

\

( Page 7 of 33

(

l s

l WESTINGHOUSE PROPRIETARY CLASS 3 2.2.1 Postulated Accident Condition 1.esk Considerations Tor the postulated leak source within ch'e tubesheet, increasing the tube '

di!!erentisi pressure increases the driving head for the insk. It also l decreases the length of the lesk path annulus, due to the pullout of a

{

postuinted separated tube end, and increases the tube to tubesheet loading. i Cf these e!!ects, only the last tvo are significant to a leakage source vichin the tubesheet. The separated tube and pullout distances due to the i increased differential pressure associated with a F12 for the various cedel stes: generators included in the P* criterions are given in Note 1 c!

Table 1. For an initial location of a leak source below the icp of the l cubesheet equal tc **, the F1.3 pullout e!!ect results in apptcxit.ately a 10

{

percent increase in the leak rate reistive to that which could be associated -

with normal plant operation. This san 11 effect is reduced by the increased tube-to-cubesheet loading associated with the. increased differential b pressure.

Thus, for a circu=!erentisi indication within the cubesheet regicn {

vhich is lef t in service in accordance with the pullout criterian (P*), the existing technical specification prLaary to secondary leakage criterion is  ;

su!!icient to maintain conditions consistent with accident analysis sssumptions. (

Tor axial indications in the tubesheet region, the tube end remains structurally intact, minimizing any sacunt of pulicut due to the previously i identified cechanis=.s. For this case, the leak rate due to F13 differential pressure vould be bounded by the leak race for a free span leak source with ,

the ss=e crack length, which is the basis for the accident analysis assu=ptions.  :

2.2.2 Operating Condition 1.eak Considerations i

in actus11ty, the hardrolled joint would be expected to be lenktight, i.e.,

the plant would not be expected to experience leak sources a=anating belw '

P*. Since the presence of the tubesheet tube indicstions is not expected to  ;

i incrasse the likelthood that the plant would experience a significant nu=ber of leaks, it could siso be axpected, that if a primary to secondary leak is i detected in a staan generator it is not in the tube region below P*. Thus, no l l

significant radiation exposure for personnel lookLng for tube tubesheet leaks should be anticipated, i.e. , the use of the P* criterion is consistent vich e

ALAPA considerations. As an additional benefit reistive to A1APA considerations, precluding the need to install plugs below the P* criterion

\

vould result in a significant reduction of unnece'ssary radiation exposure to installing personnel.

1 The issue of leakage for the consideration of postuisced accident conditions included consideration that the viciation of the tube vall is very extensive, i.e. , that no materisi is required at s11 below P*. It vss siso noted as part of the justification t} t the possibility of two adjacent tubes having the separation type of de:

is incredible. Based on operating plant and laboratory experience occur, is axial.

ene expected configuration of any cracks, should they The existence of significant circua!erentisi cracking is considered to be of very lov probability. Thus, consideration of whether or Page 8 of 33 i

I" WESTINGHOUSE PROPRIETARY CLASS 3 not a plane vili come sff-lin e :o search for leaks a significnnt nu:ber of claes should be based an the type of degradstion that alght be expect,g g, occur, i.e. , nxisi eracks. AzLs1 cracks have been found both in pinn:

operation and Ln laboratory experLoents to be short, approximately 0.$ iney in length, and tight. In addition, for both the field and laboratory

[ experience, once the cracks ha're grown so that the crack front is out c! the skipro11 or transition stess, they nrrest. Such a crack in the free synn portion of the tube, with no superimposed thinning, would leak at approximately the technical specification acceptable leak rate. Tcr a cro:k in the tubesheet portion of the cube, at a distance of P* or greater below the top of the cubtsheet, the leakage should be significantly reduced. )

Lankage through cracks in tubes has been Lnvestigated experl=entally within Westinghouse for a significnnt nu=her of tube vall thicknesses and thinning lengths. In general, the amount of leakage through a crack for a particular size tube has been found to be approxLaately proportional to the fourth power of the crack length . Analyses have siso been performed which show, on nn approzicate basis for both elastic and elastic plastic crack behavior, that the expected dependency of the crack opening ares for an unrestrained tube is on the order of the fourth power (reference: NUREC CR-3464). The amount of leskage through a crack vill be proportions 1 to the sres of the opening.

thus, the enslytic results substantiste the cast results. The presence of the tubesheet vill preclude deformation of the tube valls ad) scent to the crack, i.e., the crack flanks, and the crack opening ares any be considered to be direcc17 proportional to the length. The additions 1 dependency, i.e. , fourth power reistive to first power, is due to the dilatation of the unconstrained tube in the vicinity of the crack and the bending of the side faces or fis=ks of the crack. For a tube crack located withln the tubesheet, the dilatation cf the tube and bending of the side faces of the crack are suppressed. Thus.

a 0.5 inch crack located below P* would be sxpected to leak, vichout considering the flow path between the tube and tubesheet, at a race of one eighth that of a sLc11st crack Ln the free span, i.e. , it would leak at a race abcut equal to that fro: a 0.0625 Inch free span crack. Additions 1 resistance provided by the tube to tubesheet annulus vould reduce this accunt even further, and Ln fact, vould be expected to e1Lainste it.

2.2.2.1 Tube to Tubesheet Hardro11 Effect on Leskrste For all of the plants under consideration the Lnsts11 scion of the tube into the tubesheet included a full depth hardroll of the tube. The accepted i nethod of measuring the magnitude of the hardroll for Westinghouse pla=ts is the amount of applied torque delivered to the rolling tool. For model 51 scen= generators with full depth hardrolled tubes and s11 of the model D 6

( E steam generators the set torque for tube hardro11 Lng was [

)s,c.e. The procedures involved in applying the hardro11 included i provision for disposition of cases where a local section of the tube cisht Y

have been skipped.

The basis used for establishment of the hardrolling to be applied was a series of test progts:s involving the variation of the torque parn=eter coupled with mets 11urgics1 exs=ination of the interface of sectioned rolled Page 9 of 33

WESTINGHOUSE PROPRIETARY CLASS 3 tube samples. The testing downstrated that a significant amount of microscopic smoching of the tube material takes place. Rovever, no  !\

previous casting was identified that yes atced at evaluating the perfor .ance of the hardroll relative to tube to tubesheet joint l e a k.ag e , in order to qus1Lfy the offact of the hardroll on leakage through the tube to tubesheet joint, a series of tests were perforced at the Westinghouse  !

Forest Hills che=1stry laboratory (which has rolling equipment used fcr preparing ss=ples for metallurgical tests). A detailed discussion of the tests and the results are provided in appendix IV of this report. The test approach for this purpose was to nessure tubs strains in order to deternine the amunt of elastic preload present in a hardrolled joint. It was found that. In the region away from the entry and of the rolling tool, a pre 1 cat pressure c! about [

]s.c.e exists at the Lnterface of the tube and tubesheet at room comperature prior to operation of the stes: generatcr.

I During plant operation, however, the amount of preload vill change i depending on the pressure and te:perature conditions experienced by the  !

tube. The room te=perature preload stresses, L.e. , radial, circu=!erential y and axial, are such that the materisi is nearly in the yield state. Since '

the coefficient of thermal expansion of the tube is greater than that of the tubesheet, heacup of the plant vill result in an increase in the preload and could result in some yielding of the tube. In addition, the yield strength of the tube material decreases with temperature. Both of these effects vill result in the preload being reduced upon return to '

n=bient te=perature condL:.:ns. The plant operating pressure influences the preload the tube, directly based on the application of the pressure load to the ID of thus increasing the amount of interface loading. The pressure  ;

also acts indirectly the tubesheet to bov to decrease the amount of interface loading by causing upward. This bow results Ln a dilatation of the '

tubesheet holes, thus, reducing the amount of tube to tubesheet preload.

Each of these effects any be quantitatively created.

The maximum amunt of tubesheet bov loss of preload vill occur at the top of the tubesheet. Since F* 1s nessured from the top of the tubesheet, and leakage is to be restricted by the portion of the tube above P*, the potentLa1 for the tube section above P* to experience a net loosening during e:eration is considered for evaluation. The effects of the three icentiiled cechanis=s n!!ecting the preload are considered as follows:

1  !

2. Ther=al Expansion Tightening The mean coefficient of thermal expansion for the Inconil tubing betveen sabient conditions and 600*T is 7.80*10'6 infin/CT. That for the stes= generator tubesheet is 7.28*10 6 infin/*T. Thus, there is a net difference of 0.52*10 6 Infin/*T in the expansion property of the tvo materials. Considering a te:perscure difference of 350*T betvoen ambient and operating conditions results in a net tightening of 2.86*10d* Ln/Ln or 0.03 i strain at the joLnt. This results in a preload radial stress of about 1000 psi.

2. Internal Pressure Tightening - The normal operating differential f ressure fro = the prLaery to secondary side of the stes: generator is Page 10 of 33

, y --

WESTINGHOUSE PROPRIETARY C1. ASS 3

~

on the order of 1500 psi. Considering the tube dimensions repor:eg in AppendL1 IV for the preload test progran shows that the internai pressure results Ln a radial tube CD preload of about 1400 psi.

3. Tubesheet Bov inosening - An analysis of the tiodel D3 tubesheet was performed to evaluate the loss of preload stress that vould occur as a result of tubesheet bow. The analysis was based on the classica:

theory of the deflection of fist pistes, including correcticns which l account for the shear of layers parallel to the plane of the piste, \

e.g. " Theory of Plates and Shells

• by S. Tincshenko and S.

Voinovski-Krieger, supplemented by the work of T. Sort and V.

O'Donnell relative to perforated plates. This spproach has been axnstrated to correlate veil with results found by using finite ele =ent techniques. Basically the deflection of the tubesheet was used to find the stresses active on the top surface and then the presence of the holes was accounted for. For the location where the loss of preload is a maximus, the radini stress vould be reduced by

[ Ja,c.e p,3, Co=b bing the loss of load vich the preload due to hardroll, thermal and pressure offacts results in a not operating preload of [ ja,c.e y,g.

This preload shculd offactively retard leakage from indications in the tubesheet region of the tubes.

It is siso to be noted that this preload force acts to restrain the tube in the tubesheet during operating and postuinted faulted conditions.

Considering the actual conditions for the most limiting mdel D plant, an engage =ent distance, F*, adequate for the radial preload force to resist the tube pull out force du* to the unbalanced pressure condition resulting frc: postuisting the severance of a tube can be calculated. The results of the :sluistion, succ.arised in Table 2 based on the mdel D results reported in AppendL1 IV, show that for the 1Laiting case of Til, an engagement length of ( )s,c.e Inch below the top of the tubesheet is sufficient to i resist tube pullout. This value is equivalent to the required engage =ent to preclude pullout based on clearance considerations, P*, thus, e= ploying the h

P* criterion is also sufficient to provide for the engagement length necessary to preclude pullout based on axial force considerations.

j 2.2.2.2 Operating Plant 1.eakage Experience for Tube Tubesheet Cracks i>

A significant nu=ber of tube tubesheet indications has been reported fcr h 1 so=e non docestic stesa generator units. The ac'titude covard operation with these indications present has been to colerate them with no recedial action relative to plugging or sleeving. No significant nu=ber of shutdovns occuring due to leaks through these indications have been reported.

,\

i 2.3 TUBE IWEGP1TY ifNDEP POST"'WED LIMITING CONDITIONS The final aspect of the evaluation is to demonstrate tube integrity under the i postulated 1DCA condition of secondary to primary differential pressure. A Page 11 of 33

=.

E '

• VESTZNGHOUSE PROPREETARY CLASS 3 review of tube collapse scrength characteristics indicaces chac the constrainc provided to che cube by the cubesheet gives a significantimargin becueen cube l pressure collapse strength and the limiting secondary-to-primary differenc a condition, even in the presence of circumferential or axial indicacions.

2,4 fxCLUSIONS The analysis for the develop =ent of P* The is based on consideracion peripheral of thedo tubes in the bundle rescrainc afforded by a neighboring tube. he nec have outboard neighbors; however, restraint isEven afforded generally so, chere may be by t presence of the anci-vibracion bars recaining rings. exisc. These may be several cubes in the bundle where such restraint does not considered as follows:

a. Peripheral Tubes - Since che tube bundle in the U-bend region does not form a smooch, uniform surface not all cubes on the periphery are tubes Also, the contacted by the anti-vibration bars retaining rings.

cuo columns may be beyond the extent of the in the firsc and last Since peripheral tubes may be bridged by the retaining rings.

recaining ring such chac a full tube pitch exists between tothe tube and exclude all the recaining ring, it is therefore considered prudent peripheral cubes from the P* criterion.

b. Incerior Tubes - There are cuo cases where neighbor restraint vill not be afforded within che tube bundle; across the T-slot in Model D4 and Model E steam generators, and the interior tubes adjacencThe to and on specific the cube lane side of the tube support plate stayrods.

tubes cay be identified by referring to a tube sheec map for the For chese cubes, the P* criterion also particular model of interest.

does not apply.

3.0 SUMMA?Y

! cubes with eddy current On the basis of this evaluac.on, it is determined that indicacions in the cubesheet region below the P* pullout criterion shown on in service. Tubes with circumferential17 orienced eddy Table 1 can be left current indicacions of pluggable magnitude and located less chan P* below che cop of the cubesheet should be removed from service by plugging or repaired in accordance with the plant technical specificacion plugging limic.

)

l

! there is no need to

[ In addition, it has been decemained, see Appendix II, that indications stabilize cubes which are removed from service due to eddy current in the region between the top of the cubesheet and P*.

N 1

1 Page 12 of 33

k'ESTINGHOUSE PROPRIETARY CLASS 3 t

TABLE 1 n'?E cP ct"T PL'W'. M'"" A3Y TOP P* CF 'Tr? 'ON S~~TA" CTV??A'OP N0 DEL

& D L

Tube End Displacenent (inches):

Nc=inal Clearance l Manufacturing Tolerance l s,c,e l

Ther=al E!!ects l l

l Pressure E!!ects (1) l l

Total Displace =ent or ,

l Unadjusted Value for P* l l

l Uncertainty Allowance (2) l pe l m l '

FCE.c -

(1) Contribution to this l vertical displacement due l*C*

to feed 1Lne break. .

(2) The actual uncertainty varies on a plant specific basis based on the eddy current probe used. The value of 0.250 inch was used here for nrnple purposes only.

(3) The reported manufacturing value was ( Ja,c.e, this was increased to ( Ja,c.e to be acre unifor: relative to the Model D and E values.

l' J,

r  :.

lt

. [

} !l 5 ,

0 ..

l

.f 1

Page 13 of 33 I

_ _ - . -, _ _ _ _ . - - - -~ ' ^~

"q . zz& - J WESTINGHOUSE PROPRIETARY CLASS 3 l

TABLE 2 l l

ENGA:EMEh7 irVG*? Pr0"1Pr TO PrsisT PP_rss"P_r PULT'.T TOCTES Limiting Model D Stes= cenerator -

Ccnditien Formal Fault'.2

1. Hardroll Preload l l**C *

2. Ther:s1 Exp. Pre 1 cad l l

3. Internal Pres. Pre 1 cad l l .

4 Tubesheet Bov Preload .l l l 1 NET RADIAL STRESS PPI w AD l l l 1 NET RADIAL FORCE PREwADI l l 1 l AIIAL RESISTANCE FORCE 2 l }

l 1

3. Axis 1 Pressure Force l l l 1 FIQUIRED RACRD1.1. ENGAGEMENT l l l l

6. End Effects Distance 3 l l l l

7. Rardroll Tolerance l l l 1 1 I TOTAL ENGAGEMENT REQUIRED l l v

i

[ ' Notes: 1. This load is the force per inch of engage =ent. Lt is found by l multiplyLng the radisi stress by the tube circunference.

4

2. .'his value is derived considering a coe!!1cient of friction value l of 0.15.

3. This value is derived based on the distance required to citigs:e end effcces in accordance with usage based on the ASME B & P" Code, i.e. , 2.3 times the square rc:t of the radius tLces the.

• chickness.

1, i

J Page 14 of 33 4

w -

l .

WESTINGHOUSE PROPRIETARY CLASS 3 APPEVDIX i - TUB!NG U.?EXD GLD S"YDY FOR VES'INGVOUSE STEA" GTVEFK OP3 VITM R*2 DEPTH ML OPOU ErPAVTED T?!!S I.1 IVT?OD?CTIOV A detailed review of Tn=pa Division drawings nnd process spacifications fcr the procure =ent and asse=bly of Westinghouse models $1, D and E sten generate:

heat trnnsfer tubing was perforced to determine the final asse: bled tubing U bend gap clearances. The review included consideration of annufacturing assembly operations and/or rework which may have Increased the final size c!

che U bend gaps in the as shipped unit.

Using anxim allovable tolerances, or conservative estimates where certain tolerances were not specified, an estimate of the anximun U bend tube gap in Westinghouse models $1, D and E stes: generators a:p1oying full depth hard cil

~

expansion of the tubes into the tubesheet was deter =ined. The review activity was restricted to nn nsn=ination of VerCLnghouse Nuclear Camponents Division (NCD) angineering records and drawings.

Z.2 DIS:?SSIOM The Westinghouse preheater stens generatcts are a vertical U-tube unit in which '

the tube bundle heat transfer tubing is formed by a plurality of vertical U-tubes. Each U tube has a 180 degrees *U* band and two equal length leg sections.

Each of several thousand preformed U tubes are installed through '

several tube supports and a thicker tubesheet to for= the tube bundle.

All WestLnghouse prehester steam generators utilise a square tube pitch. In square tube pitch units, Lncreasingly larger band radius U. tubes are asse: bled directly over previously installed U-tubes until the last or outermost U tubes in a tube colu=n are installed. The tube pattern is such that the outside shape of the final U-band section is approximately spherical. Revover, the tube' hole pattern and the many different tube band radius sizes do not for: a smooth or true spherical shape. There are anny dips or valleys formed by the various size U bends and tube colu=ns as the top of the tube bundle is traversed along the center 1Lne of the tubeinne. E enver, any full section taken through U bend bundle perpendicular to the tubeinne vill show a series c!

concentrically arrnnged 180 degrees tube bend radii.

6 i

The previous general description demonstrates that 'the exact deter:ination c!

the absolute anximu= U-bend gap would be very difficult, if nct L=possible, fcr the as shipped condition of the sten = generators. This vould be true even if

\ an in process gap inspection had been ande during fabrication, because

! subsequent operations such as tube expnnsion within the tubesheet holes n!!ects i

che U-bend gaps. In addition, the maximum U-bend gaps could not have been easily acasured, nor were such measurecents taken and recorded for these units.

The objective of the study was to determine nn estimated maximua U-bend tube gap value which would exceed an actual gap size that might exist in the field 1

sten: generators. In order to determine this anximu=, which is considered to be Page 15 of 33

=.

~^=%

WESTINCHOWI rR0PRIETARY CLASS 3 a dsparture or increase from th* draving no: ins 1 tube gap, all the variables in the product geometry and in opetet lons directly on the cubLng which a change in the no=insi U bend ser were considered. may s!!ect The folicuing factors a.nfluence, nc inal, the as fabri:sted U band sap betveen tube anybeng, tvo (2) adjacent 1,

Tube to Tubesheet Fit up.

2. Tube Expansion.

3.

Tube Schedule Tolerances.

A.

Unincrement ed U. bends.

(This is dis ting 21shed from incra=ented U-bends. which B. were employed in units not under consideration.)

4 Tube Outside Diameter.

Tube Trim: Lng and U. Bend Cap Adjustment.

To deter =ine the effect or angnitude of uch of the above factors on the U bend -

gap relative to the nocinsi drawing dLaension, a detailed reviev was conducted to decercine the applicable drawings and process specifications !ct the stes:

generators for each of approximately eventy three preheater stesa generator and {'

four mdel 31 stes= generator power plants. A detailed QA records search of shop routings and_ inst.'uctions, which might Indicate the exset revisions used for the process specifications (PL) for each steam generator unit , was not perforced based on the results of tursory comparisons of the earliest revisions versus the 1stest demns tra t e:' ;~.s t draving littic*

and prc*t ess specification revisions which processes c: If an1, basic changes were made in tube asse=bly requirasents for a given stesa generator model series. The only significant variable was special U. bend rework and tube end triac:ing done on .

scoeTherefore, gap. D4 and E mdels to maintain a greater than minimum value for the U bend throughout this study, conservative assumptions on gap incr ease reistive to no: ins) were ande, using the available records and drawing and l process specifications revisions and/or were esclasted from discussions with l personnel with a first hand knowledge of the manufacturing methods or operations used on some of the Cider units.

The following sections provide a detailed on discussion

'the as-fabricated U-bendofgaptheforindividas!

the stes: effect of each of theorsidentified lfact generators of interest. The succ:ary section of this report dessribes hov chose individual effects were

\,

co=binedstes:

prehester to obtain a maxis.uc estiasteg U. bend gap for each of the Model D and generators. ' E 5 1.2.1 Tube-to Tubesheet Ficup vers-s Veld Process Specification .

L i

4 1.2.1.1 Extended Type Tube CG Tu%esheet Veld ' Process.

4 f '

Process last specification (PS) n. ter 82127KR, Issue 1 through issue . the12 to issue, tubesheet dated veld.

6 3 For 74, this was type Lead of tube for veld the extended or protruding type tube \ 't l,

such that the tube ends prcjected [

the tubes were installed 1.e., belev, Ja,c.e beyond, (

the face of the prisery side of the tubesheet prioro t perforcing tube rolling and veldir.g operations. All D3 stesa generntors

\\

vich the exception of one plact vsth four sten: generators and s11 model 51

!sg* 14 of 33 1

1 Lamm -

R

~

, , WESTINGHOUSE PROPRIETARY CLASS 3 cess: genera: cts with the exception of one three loop p1A=: have this type of weld. Considering the tolerance band spread, therefore, a[ \

Ja.c.e rnximu= increase in the no=insi U bend gap could be attrib;_:e to the extended type tube to cubesheet veld fit-up and was conside:ed in i this ev.slua:lon.

1.2.1.2 Recessed Type Tube to Tubesheet Veld Process.

The recessed type tube to tubesheet veld was used on sl wdel : stes:

generators, four m del D3 steam generators, four m del E sten = genera: cts, and three codel 31 steam generators. For this type of tube to tubeshee:

veld, the end of the tube is recessed [ Ja.c.e gnge, i.e., above, the prL=.ary side of the tubesheet prior to the tube rolling and velding operations. PS nu=ber 82127UM, first issued 3 14-73 and las:

1ssued, revision 4, dated 1-24-77 was the recessed tube veld procedure used for all c! t?. stesa generators vich the recessed type tube to tubeshee:

veld a::sch ent . A reviev nach of the four issues shoved tha: tner all spe:1!;ed n [ )**C

• cube recess, i.e., a [ .

Ja.c.e 7,ng, or , [ Ja,c.e blisteral colerr:.ce on a nc=inal [ Ja,c.e recess.

For this evaluation, an additional [ Ja.c.e tolerance was conservatively considered possible to account for gaging error and/or tube move =ent a!cer gaging, thus, a total blisteral tolerance of [

Ja.c.e on the no:inal recess has been considered.

1.2.1.3 T1ush Type Tube to Tubesheet Veld Process.

The above recessed tube veld was subsequently re laced by a

• flush" type veld wherein the tube end was set flush vich the primary face of the tubesheet. PS nu=her 82127VI was used for this type tube veld. It was first issued on 2 19 79 and la:t issued, revision 4, on 5-8 80. All four issues of this specification required [ ]*=C** tolerance c=

positioning the tube end flush with the machined face of the tubeshee:.

1.2.2 Tube Expansion Effects on the Bend Cap.

The tube expansicn process affects the no:insi U bend gap betveen tubes via the tube axial elongation or contraction associated with the specific type of tube expansion. A 'hard" roll or mechanical expansion via torque

} controlled rollers progressively elongates each leg of the U tube during

{ the tube dia=etrical expansion step rolling process. The equip =ent PS for cechanical expansion, nu=her 81007]A, reports an expected tube elongaticn t iro: [ Ja,c,e agnja._u, to g ja,c,e ,,,3 ,a,fo,fu33 g,p:y

\ \ tubesheet rolling in a [ Ja,c.e inches thick tubesheet. To consider a l

' conservative increase in the U bend gay, it was assumed that the first tube h

of an adjacent pair, in the ss=e colum, would expand the minimum amoun; h ([' )* ****) and the second (overlapping) tube of the adjacent pair b

in the tube colum would expand the c.nximum s: cunt ([ Ja.c.ey.

Thus, the no=inal U bend gap betveen these tubes was considered to be i

1

Page 17 of 33 i

J

WESTINGH01'SE PROPRIETARY CLiSS 3 Increased by the di!!erence between the anxt== and the cini== elongaticn.

[ Ja.c.e, 1.2.3 Tube Schedule Tolersn:es.

The tube ordering draving tolerances on tube overall lent ' U bend radius, tube 0.D., etc., affect the as built U bend gap between .'. *

..t ?c.

Independent of the other variables such as tube fit up, szpansion, etc.

Variations fro = true profile in the U-bend region any involve an in!Lnite nu=ber of possible U-bend gap conditions, however, it is possible to decer:ine the conservative or controlling ma.xic= pap which v .i 1=t c:be vertical mve=ent after a postuinted 360 0 tt.be leg separacicn within the tubesheet. The apex location of the U-bend was first investigated based cn ease of deter =inscion.*It is noted that if the anxt== calculated spex gap is greater than the vertics1 cc=ponent ci an estimated gap in the U bend region o!! the spex, the apex gap vill still control the maxi == cube un st! move:ent (with a postuinted 360 degrees separation in thetube leg' sLn:e the tube vili close the sas11 *off spex* gap and then be ts!1etted along the inside of the outer tube until the apex gap Ls closed.

Conversely, if the apex gap is determined to be less than the anzi==

'off-spex* gap values, the upward innermost tube movnsent vill be limited by the sasiler spex gap. Therefore, the apex gap alone appears to govern vertics1 tube mvament regardless of the gap conditions of the U-bend either side of the apex region.

The avdel 31, D3, D4, and E steam generator tube schedule dravings show that these models s11 have unincrecented U bands, i.e., those in which the tangent line of sil U bends are at the ss=e elevation and s11 straight leg sections are of the ss=e length. Thus, s11 cube band radLL originate at a single point on the center 1Lne of the bundle.

All of the cube schedule dravings give a tolerance on the tube straight leg length of ( )*** *. The bend radius, R, dicension is no=inally tabulated but band radius tolerances were not given on the

' drawings nor vere they called out in the drawing notes. The tubing supplier did have to anke U band geometry checks using a checklng fLxtute on all these acdel 51 7/8' tube sets and on the model D3, D4 and E 3/4" tube sets.

Later D5 tube schedule drsvings for 3/4" tubes spectiicsily identify a

[ Ja,c e colerance on the radius. Since the same basic tube rolling procedure was used for the 7/8" and s113/4* tube sets, it was assumed that this [ ]^*C** colerance applied to s11 model 31, D and E sten: generscor c111 annesied tubes. Howeve'r, some of the inst D4 units

{ had U bend / leg tolerance conditions which required criming corrections to

increase scall U-bend gaps, large gaps were tact considered a design issue.

t which is addresed in a later section titled " Tube Triming and U Bend Cap

? Adjustaent.*

I n

D The anxiw= spex gap increase between any two column adjacent tubes for the the model D and E units associated with the tube colerances is equal the sum of the leg tolerance, [ Ja,c,e, plus tvice the bend tolerance.

[ ]*****, a totsi of ( Ja.c.e. However, during shcp Page 18 of 33

k'ESTINGHOL'SE PROPRIETARY CLASS 3 assembly, the angnitude of this increase vould affect the stackup of the gap betveen the acst s!! acted tube and the next installed tube. zn c:g,,

vords, the next gap formed by the addition of a third tube vould Lcx:ediately be red.:ced by the [ Ja,c.e higher position of the mst affected tube. This happened on acce units ,-g was addressed by additional manufacturing operations, such as tube er.g trl=ing ct U bend pin check and rework discussed la ter and in the su=:. .-..

secticn of this appendix.

l For the model $1 series generators, leg spacing of [ Ja c,e y,,

checked with a ( )* C *

• U. bend te=pis te. Thus U-bend radius tolerance could be as large as ( )* C

• and the U-tube vould fit th e t e=pis t e . A[ Ja,c.e y,, ye y,eg ,,y,,7 esti=ste for all $1 cube schedules which co not spec...cally state the h-radius tolerance. Therefore, the ass!~'~ apex gap increase between any cyc (2) cubes fcr model 31 units associated with these tube tolerances is equa:

to [ )^ *.'

I.2.3.2 Tube Outside Dianeter.

The equations developed thus far for the spex gaps assume a perfect nominal tube 0.D. exists at the apex of each tube of the tvo tube, or adjacent pair, co=hination involved. The apex gap vill be increased if the dia:eter of each cube in the combination vere at mLninua size in the vertical, i.e. ,

the 6 to 12 0' clock direction. The effect of tube 0.D. tolerances was evaluated for their effect on the determination of the saxicu= gap size to be considered.

The tube schedule drawings specify a [ )**'** or

[ Ja.c,e Inch requirement for the tube 0.D. , for the mdel 31 and the model D and E generators respectively, in the tube straight leg sections. Notes on the tubebend region allov [-

ja e,e respective ovality in the U bends. This converts to a rather

_1arge [ Ja,c,e and [ . Ja,c.e.

respective range between the ninima and the anzimum tube 0.D. The actual minim: 0.D. is the relevant number in the determination of the maxicu:

gap. The *ninir.n=" 0.D. in the tube bends could not be exactly deter =ined since few actual values were recorded during fabrication of the. sten:

generators. Dimensions for the tube 0.D.'s in the straight leg sections were recorded but are not considered directly applicable to the band areas, particularly in the snail bend radius tubes. In the absence of available

. da ta , the dravi.sg [' Ja,c.e ainlaum 0.D. for model 31's and the

[ ja.c.e ainlaum 0.D. for model D's and E's, i.e. , [ ) *

• C ;'

inch l Ja,e,e respectively under nominal, was judged to be k appropriate.

I

i Thus, all of the anximum gap, formulas were adjusted by [ ja c,e

\

for the mode: 31 generators and by [ Ja .c , e f,7 gy, ,,g,3 p ,ng g generators by conservatively assuming a reduction'of [ Ja.c.e or i

[ Ja,c,e respectively occurs on the 0.D. of both of the adjacent i

U tubes along the 6 to 12 0' clock dina'ter as follows:

3 Page 19 of 33 l

i 1

1 L__ .

WESTINGHOUSE PROPRIETARY CLASS 3 l la.c.e l I l l l l l 1 I I 1.2.4 Tube Trt:cing sad U-bend Cap Adjustment Tube Tric=Lrg.

1.2.4.1 Tube Trin:ing.

The tube asse=bly PS, nu=ber 80314PT, slioved crLocing of the tube ends d:.:rLng bundle arse =bly, and required quality assurance, QA, to record the n =:ers cf tubes which were trin:ed. The sacunt of triccing on each tube was not required to be recorded. Bovever, trLacing was extra work, i.e. ,

work beyond the nor=.a1 pisnned operations for the stes= generator fabrication. Tri=ciz; any siso have been performed, judged to be a rare Anciden t , if the and of a particulst tube was accidents 117 s!fected to the excent that a

• clean up* vas required to make the tube to tubesheet veld.

Trt:cing associated with maintaining minium U-bend gaps was done on so:e of the inst D4 *s and one mdel E. The U-bends of the other E mdels did not have a y significant trLacing. The s.= cunt c! tr =ing as reported by c.anufacturing engLneers, with a first hand knowledge of the shop fabrication of the stes= generscor models under consideration, varied frc:

as lictie as [ )s,c.e ,ggyoug ,3gngfje,ng disposition consideration. Maximum criar tng was reported to be on the order of [ Js c.e, if it is considered that the triccing was done correctly, the finite a: cunt of corrective triscing should not be directly added to the no:Lnsi gap in a maxiw= gap study. The sacs.mt of *over trincing,* mre than the necessary amoun t , which not only corrected a minimum gap condition, but, siso l

produced a 1stge gap in excess of the nominsi gap is the value of interest.

l Considering the annuiscturing technique of and cutting, vich a I.D. tube cutter and the scale of measurement, a best estimate factor of [ l

)s c.e was considered to be an approprince slicvance for over trl=ing to be applied to the nominsi gap for the U-bend tubes for the andel D and E l

units, and [ )* ** *

• vas considered fbr the nedel 31 units.

+ 1 7

l 1.2.4.2 U Bend Cap Adjustment.

y l For the model D and E steam genersters, physical s'iftLnr and pulling of 1 the U-bends to s!jur: z:111 e n e gaps to meet min 1=.:c U band gap criterio was done on an as needed basis. PS number 80314NT specified approprie:e p

' protection to prevent dsange to the tubes at the top support piste when the

' U bend was adjusted. In general, a horizontal shift was ande to open up a small gap on one side of the U bend above the top support piste elevation.

When the small gap on one side was opened up, the gap on the other side c!

Page 20 of 33 -

l 4

] __

WESTINGHOUSE PROPRIETARY CLASS 3 the tube band canded to close up. Therefore, the tube U-bend adjustment acted to nonnalize the total U-bend gap. The apex gep, which is vertical, vould see little change caused by a hcrizontal shift, unless there was e eccentric U-bend on the inner tas. Is .ss estic.ated, for this evalust:c,,

that pulls ande during horisontal shif ting, or direct pulls to the apex area itself, for gap correction could increase the apex gaps by [

' , c . e over nc=!=s ' T.' e r - ever p..11ed tutes would ha.e been left as. .,

because the gap that would occur to the next asse= bled tube vould still meet the minia.un gap criteria.

On the unincra:ented model 31 units there was no identification of any extensive U bend adjustment work. This was probably because of the large, realstive to the model D and E steam generators, [ Ja c,e no:inal gap on these units and no drawing requirement for a minir.us U-bend gap.

Therefore, for the acdel 31 units no U bend adjustment factor was included

, in the evaluation.

It must be noted that on units for which the tube ends were trinced, the. e

, was probably not any U band adjustment required. Considering this to t e tsue for the majority of rework cases, the estimated [ )***** tri:

factor and the [

Ja c,e band adjustment factor should not be l spplied slaultaneously to the U band gap estLaste. One or the other is considered applicable depending on whether the unit was extensively tricced or simply had tube U-bands annus117 adjusted.

Z.5 Kol!d"" CAD SUMMAVY.

The individual anximu= gap increase, or delta, estinates were cochined to determine the estLanted anzia.;: U-band tube gap increase as follows:

KA77?]M CAD INCPIAc? ES*TMA'ES l '

Models D & E Model 31 l

l la.c.e l

I I

I I

I I

I I

I I

I I

I i l i

It 1

is to be noted that the second figure in the crabination for the model D a-d E units includes the 0.140 inch U-bend adjustaent tolerance but does not' i

include the 0.09 inch triac:ing factor.

The anximu= estimated tube U band gap was obtained by adding these total deltas to the drawing nocinal tube gap. SLnce the above delta values exceed the Page 21 of 33

WESTINGHOUSE PROPRIETARY CLASS 3 no=Ln21 gap in sw we ccses, it is obvious if these large deltas occur next 1siger U-bend gap would have a significant decrease reg,relative gy, to th no=Lnsi gap. e been adjusted and sc forth until gs; equilibri= again,y,was oMaIf inte

. e .

I h .

1 I

Ik 1

!ste 22 of 33

=

y _. ,. ,

kT.STINGHOUSE PROPRIETARY CLASS 3 1

DTICAL U-BD'D CA? D7ICAL BEND RADI"5

- - U BEND TAVCE.'7 tygg .

N I

l c

1AST TUBE IN COMrJ A A A  %  % w TYPICAL SECTION i PEPJENDICULAR TO TURELANE s'

NON. INCREMENTED U BEND CE0r2TRY

$ FIGURE 1.1.

Page 23 of 33 L

WESTINGHOUSE PROPRIEIARY CLASS 3 A??Ev?!X !! - DISPO!!T!04 0! TUP!5 {

VITH INDICATIONS A?0E P*

l Co:ple:entary to the criterion for leaving a tube in service with nxial c:

)

circa:!erential Lndica::c=s below the top oz the tubesheet is a criterion !ct de:er=ining the need to stabilise tubes which are removed froa service due to circu=!erential indications below the tcp of the tubesheet.

As uns previously stated, ECT Lndications located above the F* criter::= are to be disposi: cne in accordance with the pinnt technical specification plugging 1L it which is based on USNRC RC 1.121, which does not distinguish between circa:!eren: a1 nnd nxial cracks. Moreover, RC 1.121 is concerned vich the depth of pene:rs:Lon of tube vail degradation, L.e. , when the plugging 1 Lait is reached, the tube is either cubes. plugged or sleeved. RC 1.121 does not require stabilisation of plugged ~

7 = kinetics of strest :rrrr. n cracking c! :lli annealed inecnel 600 in pr.=:ry .s:r-

.s b.gnly in=perature dependen:. Righ ce=peratures accelers:e rates of cracking. Laboratory measurements of Arrhenius relation type activation energies typically range from 30 to 75 kcal per mole. Field experience with row 1 U bends Ln domestic sten: generators nnd roll crnnsitions in foreign units indicate an activation energy of 85 kcal per mole.

Conditions in tubes leading to lower tube metal temperatures greac1 7 re:ard the kinetics of any subsequent cracking even Li appiled or residual stresses are' anintained. Belov an assumed te=perature, Thot, of 620 degrees Y, cracking is retarded by a factor of 4 at 600 degrees F, a factor of 15.5 at 580 degrees T, and a factor of 64 at 560 degrees F.

Moreover, the presence of hydrogen in primary water is another Laportant consideration relative to the kLnetics c!

cracking of inconel 600.1.aboratory measursaants show that standard concentrations of hydrogen Ln primnry water accelerates cracking by approxinately hydrogen. a factor of 2 to 5 compared to control tests in the absence c!

For eddyuse in a materials evaluation, in determinLng whether a tube plugged fcr an current Lndica:Lon above the P* criterion should be stabilized due to :he p=:ential for continued grovch of nn ID stress corrosion crack, tube te:peratures within nnd above the tubesheet region were assessed , refer to Appendix 111. A plugged tube was postuinted to axist in a variety of environ =ents that would Lnfluence tube temperature, including the buildup of sludge around the cube, as the sludge may act as an insuistor nnd alter the 1

hest conduction patterns nnd surface metal camperature of the tube.

, For itselfconservatis=,

except at active tubes adjacent to the plugged cubes, and the tubesheet

} te:perature. For this the tubesheet secondary surface are assumed to be at primary fluid deposition cases were hypothesised:tamperature condition, several sludge CASE 1 considers no sludge buildup adjacent to the tube and the tube does not have a through vall penetration prior to plugging. Certain conditions in Page 24 of 33

WESTINCHOUSE PROPRIETARY CLASS 3 tubes (such as vet valis prior to plugging) may lead to the presence of superheated stes= axisting withLn the cube. Limited dats on inconel 600 at high temperatures is consistent with genersi observations on alu=inu= and steel alloys In ice ta=perature water vapor. At lov superheat, i.e., high relative humidity, the cracking response in vster vapor is essentially equivalent to that in the 11guld phase at the ss=e te=perature, while a t high superhast, i .e. , Joe relative hu=1dity, the cracking kinetics are much reduced. A plugged cube is essentially dry on its ID when plugged; i chere!cre, sichough the ID camperature of the tube in the region within the a

tubesheet vculd most likely be equivalent to That, the ratio of the vapcr pressure of any water trapped in the tube during pluggi. g er the pressure c!

satu:sted water vapor would be lov, i.e., high superhest, thus greatly red cing the cracking kinetics. Also, as previously discussed, the 1sek c!

the presence of hydrogen in a plugged tube significantly retards further cracking. Therefore, combining the above two effects, the probst lity a plugged tube with degradation that has not progressed through vs11 vould continue to degrade is san 11 in this environment and vc:.1d nct ~e ire stabilisation. It is noted that this case is really inuev. c. . . c: whethe'r or not sludge is postulated to be present, i.e., the ta=pe.zcure inside the tube Ln the tubesheet region vill be near That regardless ci the preser. e c!

sludge.

1 CASE 2 considers no sludge buildup either ad) scent to a plugged tube or in a plugged tube. A through vail Indication is postuisted and the tube is filled with water due to the ingress of secondary side water through the penettstion. The water contained in the tube in the tubesheet area boils.

This rapid heat transfer mechanism astntains the tube inner disneter metal surface at or slightly above T,ag for the portion of the tube in the l

cubesheet. ne reistively lov secondary side temperature vill significantly inhibit the continuation and/or initiation of stress corrosion cracking.

The:afore, considering both the n!! acts of the reduced secondary side es=perature and the Jack of the presence of any hydrogen concentration on l continued stress corrosion cracking, the probability of a plugged tube with a through vail penettstion continuing to degrade is very sas11 and would not l

j require re=edisi action ether than plugging or sleeving.

1 1

CASE 3 considers the effect of sludge buildup on the tubesheet ad) scent to a plugged cube with a through vall penetration. Since the sludge acts as a poor conductor, the mechanisms for cooling the tube are not as efficient as for the previous two cases. If the secondary side unter ingress re= sins pri=.arily in liquid form with some locaissed boiling at the cube vall, and the sludge pile depth is less than about 4 inches, the to=perature on the

inner dis =eter of the tube vill probably be slightly above 7,,g. As

[ discussed previcasly, certsin conditions in plugged tubes any lead to the

{ presence of superheated steam rather than 11guld vster the through vall

\ degraded tube. It was siso stated that at lov superhest, the crack i response in water vapor is essentis117 equivalsnt to that in the . >

phase se the same ta=perature. At sludge depths greater than about . sches, the tube metal temperature in the^ tubesheet approache_s plant hot leg te=perature. The offact of lov superhest and higher tecperatures could result in additions 1 crack grovch. However, the above two scenarios are not Page 25 of 33

[ . _ _ _ _ _ _ _ _ _ _ _ _ _ . _ _ _ _ . _ _ _

k'ESTINGHOUSE PROPRIETARY CLASS 3 expected to occur. In stesa generators with a flow distribution bs!!1e, sludge buildup to a height of 8 Inches is precluded by geometry constraints.

Moreover, as the postulated crack would most likely 1 Lait the ingress of the secondary side water in the through vali degraded tube, the m st likely scenario would be that the tube is essentially dry on the inside and the ratio of the vapor pressure of the water to its saturation pressure is relatively low, thereby greatly reducing the crack kinetics. With the lack of the presence of any hydrogen concentrations, the pctential for additional crack growth would be significantly reduced; therefore, tube stabilization is not required.

An extension to CASE 3 could be postulated such that the through vall

\ penetration is of such size as to initially admit water into the plugged tube. The water subsequently boils and the internal pressure prevents any further water from entering the tube. In this case the stesa vould be at the seco=dary side pressure, i.e. , high superheat conditions, and for reasons cited in consideration of case 1, further crack growth would not be ,

expected.

1 1

t 3

il 1

Page 26 of 33

) I, ---- "

WESTINGH0l'SE PROPRIETARY CLASS 3 ADPENT!X I!! - n'EE V E! TEMPEPA~TPIS DE PL"CSET R'EES To assess whether further degradstion due to postulated PV5CC can occur in a plugged tube and to disposition tubes with indications above the P* criterien as to whether they should be stabilized when plugged, the metal temperature of the tube inside diameter at elevations above the P* criterion, but, below the top of the tubesheet, was evaluated. Active tubes adjacent to the plugged tube, and the tubesheet itself axcept at the secondary surface, are at the prir.ary fluid ce=perature. For a tubesheet temperature condition equivalent to Tyce, five different sludge depcsition cases were hypothesized.

1. An intact tube without sludge deposition on the tubesheet

2. A perforsted cube without sludge on the tubesheet

3. 'An int:ct tube with sludge deposition on the tubesheet. -

4. A perforated tube with sludge deposition on the tubesheet.

5. A perforsted tube vich/vithout sludge deposition on the tubesheet without secondary water ingress.

An intsce tube is defined as a plugged tube with no through us11 penetration i.e. , no secondary water comes in contset with the tube inner vall, while a per! crated tube is de!Lned as a tube with a througn us11 penetration L.e. ,

secondary water comes in contsct with the tube inner vs11.

1.1 Intact Tube Without Sludge Deposition With the exception of a shallov layer at the tubesheet surface, the tubesheet metal te:perature ad) scent to active tubes just below the top surface of the tubesheet is expected to be at primary coolant inlet ta=perature (i.e. Thot) for the hot leg side of the tube bundle. Therefore, the outer vail te=perature can be as high as That for a full depth hardroll expanded tube.

For an intact tube, the inner vail of the tube is essentially dry. The inne' vall anxicum tube ce=perature along the length of the cubesheet vould approach Thot*

1.2 A Perforated Tube Vithout Sludge Deposition Once a plugged tube is perforated, secondary vster can ingress into the*

primary side of the inactive tube. The water contained in that portion of the tube within the tubesheet boils. This rapid heat transfer mechaniss keeps the inner tube vail comperature at approximately T,,e + 5 *T (siloving for a localized vail superheat effect).

h! '

4 y .

i.

15 Page 27 of 33

,r  :

i

)

1L_ .. - -

v

• WESTINGHOUSE PROPRIETARY CLASS 3 r 1.3 Intact Tube Vith Sludge Depo:Gon With sludge accumuistion on the top of the tubesheet, the whole depth of the tubesheet is expected to be at Thot. As is the case without sludge deposition, the inner vail of the inactive intact dry with a anxt=u= ca=pezature of T tube would be essentially  ;

tubesheet. hat anticipated siong the length of the  ;

1.4 A Perforated Tube With Sludge Deposition Similst to the case of a perforated tube with sludge deposition, secondary water can ingress into the primary side of the inactive tube. The heat transfer mechanism for cooling the tube inner vail metal ta=perature vould be the same as with the case of no sludge deposition on the tubesheet. The inner vall to:perature occuring inside the would be at approxis.stely 7,,e + 5 *F because of the boiling tube. ~

1.5 A f eriorated Tube Vithout Concunication A situation could develop such that only a linited amount of secondary water vould initially postulated thatleak into a cube with a through vail penetration. It can be the small amount of water Lagressing Lnto the tube inner diameter could evaporate and form superheated stes: vichLn the depth of the tubesheet or the tubesheet plus the height of the sludge. This case vould be .

l similst to an intact tube as the superheated scess would prevent the water  !

fic:

antering into the primary. side of the tube. The inner vall of the tube would essentisily be in a dry condition and ths anzinu= inner vs11 metal ta perature vould be T hat.

In su=:ary, of the the inner vail te perature for the perforated tube within the depth cubasheet, both with or without sludge deposition, is essentisily at 7,,e + 5 *T vhen there is water cor:nication due to a through vs11 penetration. The inner vs11 te=perature for a perforsted tube without water com:nication could be as high as T an intact hot. TLns11, 7 the inner vali ce:perature for tube with or without sludge deposition could be as high as Tho:. l f

l k

k i

Page 28 of 33

• k 5

• ' + '

WESTINGHOUSE PROPRIETARY CLASS 3

1. PMH N - PPr'. DAD TEs'!Nc or RArmpm_1 r9 ME TO TUBEsyrr~ yo;y 5 IV.1 IV7 POD?cTioN Tubes are Lnstalled in the scen= generator by a hardrolling process which expands the cube to bring the outside surface into intLaste contact vich the tubesheet hole. The roll process and roll corque are specified to provide metal to metal contact between the tube and the cubesheet. Recently, concerns have arisen as a result of the discovery of addy current Lndications on the inside dinceter of the tubes in the tubesheet hardroll area. The tube Indications represent a potential leak path across the primary pressure boundary through the interface at the tube and the tubesheet joint. The assessment of this condition has led to the establishment of an acceptance criterion known as P*. P* refers to a dimension sensured from the secondary face of the tubesheet,belov which degraded areas may occur on the inside diameter of the tube without compromising its performance. No specific leak test data exists with which to quantify the functional performance of the Pc

' criterion reintive to leakage. Therefore the testing described herein was performed to identify the Laterfacial displacements as a result of the hardrolling operation and facilitate analytical support of the P* criterion.

A test program was conducted by VestLnghouse to quantify the degree of Interference fit between the tube and the tubesheet provided by the full depth nachanical hardrolling operation. The data generated in these tests has been .

used to support the acceptability of the application of the P* criterion. The amunt of interference was determined by insta11 Lag tube specimens in a tubesheet mockup using representative hardroll torques as used during scen:

generator manaisecure. Once the hard rolling was complaced, the test co~.!!guration simulating the tubesheet was removed and the springback of the tube was measured. The

• DELTA

• thus obtained L.e. , the swunt of spring ba:k.

vas considered representative of the interference fit between the rube and the tubesheet and could therefore be used to determLne analytically the residual tubeftubesheet radial load acting to restraLn the tube in the tubesheet and provide leak limiting behavior stallar to a hybrid expnnsion joint (REJ) used in sleeve installation.

IV.2 QLSCUSSION The test progrs= vas designed to simulate a tube.co, tubesheet full depth hardroll for a model D stanz generator. The test configuration consisted of a 7

six cylindrical collars, approximately 6 inches in length, 2 inches in outside din =eter, and 0.766 inch Ln inside dinceter. A mill annealed, Inconel 600 (ASME SB 163),

tubing spe.imen, approxicately 8 inches long with a nominal 0.750 outside dinceter before railing, was hard rolled into each colist. The outside dicension of the collar was determined such that the upper six inches of a unit cell in the tubesh:r t was simulated. The hard rolling process simuisted actua:

$ tube insta11s tic. canditions. .

The collars were fabricated from AISI 1010 carbon steel similar in archanical Page 29 of 33 S

4 L__m.-_M _tE:---- "#-*---"-- _..---.v_: -

=m_..== m. =- ___ - - ,

. , , . WESTINGHOUSE PROPRIETARY CLASS 3 properties to the cctual tubssheet anterial. The collar asse=bly was cin= ped in a vise during the ro11 Lng process and for the post roll measurements of the tube ID. Following the taking of all post to11 seasurements, the collars were saw cut to vichin a small distnnce from the tube vall. The collars vez e then split for rnmoval from the tube and tube ID and CD measure =ents repeated. In addition, the axial length of the tube within the co11st was measured both before and after colist removal.

Two end boundary conditions were Laposed on the tube specimen during rolling.

The end was restrained from axisi action in order to perform a tack roll at the 1 botto: end, and was allowed to expand freely during the final roll.

IV.3 Pls"LTS

• The results of the test are tabulated in Table TV-1. The data recorded was that necessary to determine the Lnterfacial conditions of the tubes and collars.

These consisted of the ID nnd CD of the tubes prior to and nfter rolling an'd removal from the collars as veil as the inside and outside dLoensions of each collar before and n!cer tube ro11 Lng. The remainder of the data of particular interest was calculated from these specific dimensions. The calculated dLaensions included vail chickness, change in vall thickness for both rolling and removal of the tubes from the collars, and percent of spring back.

In addition to calculating the changes in physical dimensions, the sacunt of preload radini stress was also deterutned, as ve11 as the anticipated changes in the preload due to thermal expansion and pressure loading, as specificed in the naLn body of this report. On the average, the radini preload stress due to the hardroll is on the order of ~5800 psi at room enaperature. The difference in thermal expansion coefficients betveen the tube and tubesheet vill result in and additional 1000 psi stress during operation. Internal pressure during operation vill incrzase this value by another 1400 psi. Thus, during operation the totsi preload value vill be on the order of 8200 psi. This value vill be reduced to approximately 6200 psi by di11tation of the holes in the tubeshee:

due_ to deflection of the tubesheet caused by the prLaary to secondary pressure differential. A discussion of this effect is contained in the main body of this report.

Based on the results of the testLng , it is concluded that following the installation of a tube by the senndard hardrolling process, a residual radial preload stress exists due to the plastic deforasticn of the tube. This residusi stress tends to restrain the tube in the cubesheet and provides a leak 11=iting seal condition.

k 4

Page 30 of 33

• • • • WESTINGHOUSE PROPRIETARY CLASS 3

_s . C. .

• i l

l

= .

t k

'. h Wr page 31 of 33

=

==

M

- . - ~- _,_ ___

.. .. ESTINGHOUSE PROPRIETARY CLASS 3

.c .e I

/

}

page 32 of 33  !

I r