ML20062N798

Forwards Info Re Auxiliary Feedwater Header/Steam Generator Mod,In Response to NRC Questions in .Also Discusses Potential for Crack Propagation in Retired Header. Util Will Propose Changes to Tech Specs within 30 Days

ML20062N798
Person / Time
Site:	Davis Besse
Issue date:	08/16/1982
From:	Crouse R TOLEDO EDISON CO.
To:	Lainas G Office of Nuclear Reactor Regulation
References
849, TAC-48349, NUDOCS 8208230353
Download: ML20062N798 (24)

Text

• ,.

Docket No. 50-346 TOLEDO License No. NPF-3 EDISON RCHARO P. CAOUSE Serial No. 849 va %t Mita August 16, 1982 (4191259-5221 Director of Nuclear Reactor Regulation Attention: Mr. G. C. Lainas, Assistant Director Division of Operating Reactors United States Nuclear Regulatory Commission Washington, D. C. 20555

Dear Mr. Lainas:

This letter is to forward information related to the recent auxiliary feedwater header / steam generator modification at Davis-Besse Nuclear Power Station Unit 1.

Attached are responses to questions raised by your staff. This information is to clarify responses previously submitted in our letters of July 15, 1982 and August 6, 1982 (Serial Nos. 839 and 845).

Attachment A expands on responses to questions IA and 2 of your June 23, 1982 request. Attachment B discusses potential for crack propagation in i the retired header.

Within 30 days. Toledo Edison will propose changes to Davis-Besse Technical Specifications to reflect the extent and frequency of future inspections.

Very truly yours, ff p = -

l RPC:TJM:nif encl.

cc: DB-1 NRC Resident Inspector L

bool EDISON PLAZA 300 MADISON AVENUE TOLEDO, OHIO 43652 THE TOLEDO EDISON COMPANY 8208230353 820816 PDR ADOCK 05000346 P PDR , . -

_ = .

t

• t i

j l

Docket No. 50-346 I

License No. NPF-3 Serial No. 849 August 16, 1982 ATTACHMENT A i

I l

i

Answer t5 Questions IA and 2 1.0 Summary Analyses were performed to determine the adequacy of the header attachment design and the header structure. A three dimensional finite element model was utilized. Loads were combined according to ASME Code Criteria and applied to the structure. The resulting stresses were compared to allowablesalso in accordance with the ASME Code. The conclusion drawn from this analysis is that the header attachment design f is adequate for all anticipated loads and that the header structure has sufficient margin to accommodate a substantial amount of weld cracks or degradation.

i i

I

[

1 A-1

2.0 Method of Analysis The stabilized header is subjected to loads which cannot be simulated using axisymetric models. To provide adequate accuracy, the header, eight attachment points and an attenuation length of the shroud were modeled as a three dimensional structure using the ANSYS Finite Element Code. The header was modeled using quadrilateral plate elements to represent the four sides of the header. The circumference of the header was divided into 54 elements with nodes separated by an average of 6.70 . The shroud was also modeled using quadrilateral plate elements and included one dimensional elements at eight node points around the circumference to simulate the alignment pins and their interaction with the steam generator shell . The two structures are connected at eight locations by the use of tie-nodes to represent the welded attachments. In order to avoid excessive computer time the shroud was treated as a super element and thus specific results are not available for it. Figure 1 shows the full 360 model which was used.

A-2

1 a

i l

2 x/\ ,

I

%  : DQ  ;

_j5isi;3

( *<j ihiiSEE

I-es C; % ; ~C ~

C R s ss

~

~

s

% g \\\

//f f

$$ CI- Z ~ C s RND s

- -- ~ to e- <- -

~s :ss' ss p

b,P2,:s%{Ti- ccr6 1

i v

t ss *lQ -::yg

t \' Rt s 5

= s =w p - 33 ss\ x s

-so- - o-ss g\\\ ox*-- g N g

' - / j,///

g

/ f ///

gNN \ g %

% f

%% s

~

% -  %- $ - ' p /

//

s s _ - -

% } ,#

1 Rd/ CdMA'ETf %d

• HEAttOEL C .

- - - - - - - _ _ - . _. ._. . A-3 _ . _ . _ _ . _ _ . _ . ..

~

3.0 AttacNnent Weld Analysis 3.1 Attachment Design The internal header attachment design provides eight attachment points between the header and shroud. Each of the attachment points is located near one of the shroud alignment pins. The attachment is provided by a large fillet weld between the shroud and header in the corner formed by the two parts. In addition, a gusset plate is welded between the bottom of the header and the side of the shroud. The attachment design is shown in Figure 2.

3.2 Assumptions The model was created primarily to determine the loads imposed on the re-designed connections between the header and shroud. Because of the geometry.of the welded attachment, shown in Figure 2, the calculation of the stress intensities from the load and moment vectors required assumptions as to the way the welded attachments would carry the load.

1. Radial Horizontal Load Because the gusset is relatively flexible'in this direction compared ~ to the fillet _ weld it was conservatively assumed that only the fillet weld would carry this load in shear.-

2. Circumferential Horizontal Load-Both the gusset and fillet welds share this load in shear.

4 A-4  ;

O i nii mammis l'

N 1

v i

x

,,,, e fy,n h a Nedos i L .

graw ,

i

, i i

I i

J l

i W

~

' i g. z A-5

1 3. Vertical Load l Both the gusset and fillet welds share this load. l

4. Moment about Radial Axis Both the gusset and fillet welds share this moment with the centroid being at the center of the welds.

5. Moment about Circumferential Axis  ;

The gusset and fillet take this moment as a vertical couple. The centroid is between the two welds.

6. Moment about Vertical Axis The gusset and fillet weld share this moment with the centroid being between the welds at the center of the welds.

The weld area of the fillet weld is taken to be the theoretical throat times the length. The weld area of the. gusset welds is taken to be the thickness times the length. For both welds a weld quality factor of 0.5 is used as is recommended for a fillet weld in Section III Subsection NB paragraph 3356 of the ASME Code l

1977 Edition, Sumer of 1978 Addenda. '

l i

The analysis of the welds in the header itself used a weld l quality factor of .1 since these were' designed as full penetration i

! welds. The model is constructed such that the full stiffness of these corner welds is used. In considering the stresses in the-corner welds the use of the full stiffness is conservative since it maximizes the predicted loads on the weld.

A-6

3.3 Load Canbinations and Results Level A & B This analysis was performed for the combined Loads of Deadweight, Flow Induced Vibration, Operating Basis Earth-

, quake and thermal transients. Flow induced vibration due to random excitation was calculated and found not to exceed peak loads of 2880 lbs. horizontal and 77.4 lbs vertical l

i, once in 40 years. Flutter and Vortex Shedding were con-sidered and found to be negligible. The steady state drag load created a net downward force of less than 1700 lbs.

and a horizontal radial load of less than 60 lbs. The operating base earthquake for Rancho Seco, the plant with highest seismic loads, resulted in acceleration levels of 1.3g's horizontal and .29's vertical *. All of these loads result in low stresses in the header although they were added into the load combinations. The conditions which do produce significant stresses are two transient conditions, secondary side heatup and initiation of auxiliary feedwater.

Both of these are thermal transients which create secondary stresses in the shroud and header. All of the other transients considered did not result in a sufficient change in temperature in the generator to produce significant stresses. In a like manner the stresses in the attachment welds might be considered secondary; however to be conservative the stresses in these welds were treated as primary stresses.

• These are the accelerations for the internal header-due -

to steam generator motions calculated using lumped mass dynamic models.

A-7

The first condition, heatup, causes stress because of the interaction of the shroud alignment pins with the shell. During heatup the shroud and header follows the steam temperature more closely than the shell resulting in a maximum at of 700F. The shroud attempts to expand radially but is prevented by the alignment pins which contact the shell. The shroud deflects into an eight lobed shape. The header which is also at the steam temperature tends to remain round. The analysis was perfonned by imposing the calculated radial displacement caused by 70 F at, .026 inches, inward on the shtoud at the eight alignment pin locations. The maximum shear stress resulting in the most highly stressed bracket from this load combination was 4,800 psi compared to an allowable of 6,000 psi. The allowable stress is equal to

.6 Sm (Level A & B shear allowable) times a 0.5 weld quality factor or .3 S,.

The second transient condition, initiation of auxiliary i feedwater, causes stress by cooling the shroud by splash-back from the nozzle discharge. The splashback causes local cooling of the shroud at the 6 or 8 nozzle locations. The header is not cooled and tends to remain round thereby imposing loads on the attachments. The maximum shear stress resulting from the load combination including this transient is 3460 psi compared to 4740 psi allowable. The allowable is l

l lower than the heatup case because of the higher temperature.

I A-8 i

l i

A fatigue analysis of the Level A & B conditions shows that the header attachment welds are adequate for 360 heatup transients 29,000 initiation of a AFW transients and the full compliment of all other transient listed for the plant.

A fatigue stress concentration factor of 4 was used in the analysis.

Level C ,

l Level C analysis was performed considering Dead Weight, Flow j Induced Vibration, Thermal Transients and Safe Shutdown Earth-quake. All conditions for Level C are the same as analyzed for Levels A & B with the exception of the Safe Shutdown Earth-quake which has acceleration levels twice that of the Operating Basis Earthquake. The additional stress due to SSE is small resulting in the Level A&B margins being limiting.

Level D Two Load Combinations were considered: (1) Dead Weight, LOCA, and Safe Shutdown Earthquake; (2) Dead Weight, Main Steam

! Line Break (MSLB) and Safe Shutdown Earthquake. The -limiting case is the combination including Main Steam Line Break because of the lateral load resulting from the unsymetric steam flow 1

caused by the break. The lateral load was obtained from an analysis performed on a model representing a steam generator with a tall shroud rather than the combination of shroud and header. The side load taken from that analysis was a distri-buted pressure loading which when integrated over the header area yielded a load of 23,500 lbs. The header, because it A-9

reduces the steam annulus has a higher pressure drop than the tall shroud. A study was performed to access the affect this would have on the MSLB load. It was determined that a factor of 10 would conservatively bound the effect of the different geometry. This yielded a load of 235,000 lbs.

The application of this load plus deadweight and SSE yielded l

a shear stress of 10,250 in the most highly stressed attachment weld. This compares to an allowable of 10.500 psi which is equal to .175 Su or 0.7 Su times a 0.5 weld quality factor times 0.5 conversion factor to shear using maximum shear stress failure theory. Stresses such as stress due to the vertical load, which were not in the.

shear direction were multiplied by 2 and added to the shear stress. This is more conservative than calculating stress intensities but,because the non shear stresses were small, the effect is not great.

The load combination including LOCA is not limiting because the LOCA accelerations of 13.75g's horizontal and 8.25 vertical, although high, do not produce significant stresses due to the relatively low mass of the header.

The shear stress is the most highly stressed attachment weld is 1053 psi compared to 10,500 psi allowable.

i i

4 A-10

t 3.4 Conclusion The header attachment welds are adequate for all anticipated loads. The requirement for these attachments is to hold the header in place atop the shroud and'for Level A, B or C Conditions to prevent contact between the header and tubes. The attachments provide sufficient rigidity to satisfy this requirement. For Level D, the requirement is no tube rupture. The attachments by preventing the header from breaking loose avoid any potential for the header to cause tube rupture.

i L

1 f

I

. \

e i- ;f. 4 ' , u, j .'t

! li i

A-11 i- , _-.

9 v

4.0 Header Weld Analysis The same set of analyses was performed on the welds at the corners of the header. Because these were designed as full penetration welds the analysis was performed using a weld quality factor of one.

For Level A, B or C the significant stresses are primary plus secondary stresses where the peak stress intensity in any weld is 11,480 psi compared to an allowable of 47,400 psi which is equal to 3Sm. This yields a safety factor of 4.1.

For the Level D loads the combination including Main Steam Line Break is most limiting. The most highly stressed of any of the welds has a stress intensity of 17,200 compared to an allowable of 37,920 psi which yields a safety factor of 2.2.

The fatigue analysis fo'r the welds was performed using a stress concentration factor of 4'which is appropriate if cracks are present.

(A stress concentration factor of 1.0 would normally be used.)

This analysis yielded a fatigue usage factor of .86 for 360 heatup cycles and 29,000 AFW initiation cycles.

These analysis can be used to show that substantial margin exists to encompass the existence cracks in the weld. To meet the code limits for faulted condition only 45.4% of the weld would be required even if all of the weld were stressed at this peak value. For this to be true any cracks would have to be interspersed around the circumference. A reasonably conservative inference would be that 25%

of any weld could be fully degraded or cracked if the condition was intermittently distributed.

The stress averaged around the circumference for the corner welds due to the main steam line break-is much less than the peak values given. An analysis has been performed using the Main Steam Line A-12

Break Load assuming a 28 inch crack to exist in a irner corner weld to determine its effect on the header stress pattern. The result of the analysis was that the crack does cause a slight increase in local stresses in the corner welds but has no significant impact on the stresses elsewhere in the header. This leads us to conclude that the existence of some cracks does not invalidate the analysis reported here and supports the above conclusions.

4 I

A-13

Response'to Question Ib The minimum required clearance between the steam generator tubes and the header was first arrived at in a qualitative manner. There is a .250 inch clearance between steam generator tubes which has proved through many reactor years of operation to be adequate to prevent tube damage. The minimum tube to header gap was set at one-half this or .125 inches.

Qualitative analysis for Level A, B and C conditions have been performed to assure that the predicted tube and header motion is less than this minimum clearance of .125 inches and thus no contact will occur.

During Level A and B conditions the effects of dead weight, flow induced vibration, operating base earthquake, and thermal transients have been considered. Deadweight is not significant. Flow Induced vibration of the tubes has been addressed in analysis and test. The lane tube which is known to vibrate the most, has a vibratory amplitude of less than .015 inches. For OBE tube vibration is calculated to be 3 x 10-6 inches which is negligible. The header sees such small loads due to both FIV and OBE that its motion is less than .001 inches. During the heatup transient the shroud is restricted by the shell while the tubes move with the tubesheet which can result in a maximum radial relative motion of .026". This is the maximum shroud deflection and is a conservative estimate for the header. If these motions are assumed to occur simultaneously the total would be 0.042 inches which is approximately 1/3 the .125 inch requirement.

Level C conditions again vary only in that Safe Shutdown Earthquake is considered. The additional transients listed are either not significant in that they do not affect secondary side temperatures or they are similar to the transients considered for Level A & B. The doubling of the acceleration level for SSE has no significant affect on either tube or header relative motion.

There are two conditions considered for Level D conditions LOCA and Main A-14

_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ m.

Main Steam Lin'e Break. For both of these conditions tha requirements is that steam generator tubes not rupture. For LOCA, the accelerations do not cause sufficient motion to cause the tube to touch the header. The tube motion is calculated to be 1 x 10-5 inches and the header motion to be .002 .005 inches. For the main steam line break the drag force from the high velocity steam blowing across the tubes may be sufficient to cause the tubes to contact the header. This is acceptable because of the high ductility of the inconel tube which can accommodate as much as 50% strain without rupture. The plastic strain which would result if the tube were to deflect sufficiently to touch the header is less than 5%. This leaves a large margin of ductility to accommodate any local dentina which might occur because of contact with the corner of the header.

A-15 ,

P

.s Docket No. 50-346 License No. NPF-3 Serial No. 849 August 16, 1982 ATTACHMENT B

v'

.s RESPONSE TO AUGUST 13, 1982 NRC QUESTIONS INTERNAL AFW HEADER

1. S, tress Profiles With and Without Crack These MSLB stresses for respresentative elements around the internal header are shown in Table 1. Figures 1, 2 and 3 show the element orientation. Of these elements, the maximum stress intensities (14.8 kai without crack and 15.3 kai with crack) occur at element 6 which is approximately 100* from the center of the crack. The ASME Section III allowable stress intensity for this condition is 42.0 ksi.

2. Circumferential Crack Growth The largest loading on a corner weld is a moment about the tangential axis due to differential thermal expansion. If a through-wall crack were present, this moment would be relieved, lowering the thermal stresses.

A critical crack length in the circumferential direction cannot be determined by Section XI fracture mechanics. methods. However, a '

fatigue crack growth rate could be calculated using fracture mechanics methods if the proper material properties were known.

This would be expected to show a longer life than .the fatigue analysis already performed since the fracture analysis considers only the stress component driving the crack (2300 psi) while the fatigue, analysis considers the total stress intensity (11,000 psi) which was conservatively calculated assuming no relief in the thermally. induced stresses due to the presence of the through-wall crack. Therefore, the existing fatigue analysis using a fatigue strength reduction factor of four is deemed adequate to demonstrate that the crack will not propogate.

j h

~

Babcock &Wilcox' " ' * " ' " "

..a.. ...,

Nuclear Power Generation Division GENERAL CALCULATIONS SuMFMCY if~ rfGCTED Mto-BD 67 i~ME*Dr553 inr~ na 7/ M 'C N E NELG ( W <)

7f,gtr .L MSL Scrc R?m:.va -

__ APPP.oX. '

Q( . C t

' / yy l'At k% ~

M (P3d (.%ij (PD ) (Fsd RQ O F ti W 44 -13,829. -I4+C, 4766. ~ 7I- 14, 9 3 . ,

6 l 34 l E' .-it,S;e7 -/875. 412I. '

-80. '

- 15,3 5')-

l .-

Ho 2354. -332. 132 5, 4 tod . -lo8G. ,

/02 -

80.4

-p 32 '

Ms 33/2 -!!3 1913. -IIts.

no -12,783. -241, 843. ~ 185 -12,65') . ,

% -/3,O/4, - 141, 1204. - 31. -l2,12C.

,q%.

l Ma 14(,$, 71/. -3020. 47&I. - I 50 7. ,

f42 -

f/s/

% 3?//. /dC d. -3240 & 15 ' . - ??? .

th 1768, 67,8 -1501. 247). -640, 2N -

24a9'-

m -1206. /o 9. -57/. 322, , _j4,.9, 1 _

Ns 4977 31.6 -/141 C3I1 -34, 202 .

247.3,

% 2257 81C. -/492, 3z;(a, -(,4, No 7330, 20C, - /984, 7427. - M, l 3/O - - --

2 62.7,.

l 21 - 6982, ~M (. -/885, , I7I, - 7.Sog ,

1 - .

W l/ M 0df/ YrN: AEbAcrg sic AE0ACAN # n ,,y a y g numa u -Bh6) f . A M t' f am 8-/4-&'z. wl m _

1

=_ . . . . _ _ _

~

Babcock.m,..-,...&Wilcox '""a""*'"

Nuclear Poaor Gonaration Division GENERAL CALCULATIONS 5 TIM 6IA@ 0r_ E*/ R yfit Mic)- B 9/. K.:6_.* WAff." - CoAtr'D

7tjdE 1 6tSL 8& flFssot4s - ._ . _ _ _

64- wnD r e AWoK e - p ir p No. war N 'Y hy %g

"~

(MI) (PJ/) (petsii .n>

.~

(PrQ_ (psi) 95 lh -4 818. /6[. -49/. 233. -f 88(e .

342 --

gg/ g, 15 : -4704. - 48C,. - 977. - 2 71. '

-49 6, ik - 3 y14. //7 -&s 'i. 2:I . -3820 ,

EB - --

- 2268

's' -//27. 22, -Z87, Iq4. a169 lh i33C 2us' ~2020 4l12. - 11 2 .

392 -

3/8.4'

.m %3 2586 .100_? , ~ 2/&1, A *f,6. 42l.

O _.

Nu -st2 3C3. 144&. 197 -Hc'.

M -

.!LU '

wt -3.W. K f. /844, i s y,. - ;28! .

1

- ^

1 1 m -

W [o it f. fwJJ AMC8& % llEd$CSAl

,.. a ,e ,, _ 3*u 2 &cD&^ , a J PL_ m ., _ _ _.

Babcock &Wilcox m v ar.' w.='>

c- suci .r ea.7c~ 'n~e'r~a' tion oivision GtWERAL CALCULATIONS FqurE L L

.\ *

$ 8

\k$

~ e s

$9 ,

l  % m d

_, s - 3.~ ~~

Y .

t 1 e

\ \  % SK /

l .e $ (\ \ \ / \

g E \ \ le,fll ,1 /f jg \ t /

l 1 \ Y\X X X/ 3 )pj,jg gi > 3 :11 g

,d a gip gg /th .

l

~

vi 83 as kht Ws

, W. 4 3 l S l

l 0 -

Heta:E0 LY $/ OAM S f~B CC C, % . ~#

l

.x.

2 U3

& R m MOWCArion _Q}; Fut<TE t?LEHFar Hop 2L %_ REPeEs&Jr THC ~

Hf5ADER.

f ,

g MEtD CRAct !_4 T1+E SHira ftMr tToP oF HFASEe SSx l

__ _ 7 _ _ __. 7 _ ___ 7 _ -

T -- - y - - -- t s 2g p$

I ' i I I I 's p i i i i '

' i

'l o{E l ag i

l ( l i a l i i i 5 E-o}x i

4_ __ __ __ _{. .__ _ _ _ _. d __ _4____ - _ _ - -,

I ~

I im WALL i g I I I i I " ,

l I g i  !

i I I E j 4 ) 5.

\l J I l l I

I a I comR e.

o Iqw qu IV,e %7 l % f, #AL "

Mr  ! %4 i f I g s @ @ @ @ @ @ l c3

$ aqa es- Pc7 '1* ' ' 5*5~

45 iiso 1 m

2 wrma or M

1. \#NDSno dI cp NEdI

's te '417 NDhh8\fm

'4'i r., '4tr i

rg 3>

r-dfADER

.f y' C)

H

'l 5 f Mo% t tHG c\eaEO 14cs. AetG O h ELEHrar hks. - THE OTH6R h M. Ae6 kDE FitwT Mc5. .I x 5 n 3 . -

N -i t e,

' d' +2 O E 8 E&t1Mr CofKD. Sys. 2 z CO I I

l .

BabCOCkikWilCOX '* "'7'" '

. a.. . : :. . ,

c _

nuci.ar pow.r ceneration civision GENERAL CALCULATIONS 27fuc6 3 TEP V/EWfr /tb50u'A (NNER JA94L) GJ;byisy!G utconfr.etsvri.4t a c g ris y o f 7 m g i t.q , g o G E./._(E_Mrs _

-f f _a.c z

"* ' A Kimroe or .~ W~

Mpory /

'N O

. &. ?G.)

e-Su 49 h R-

/02 ru42 '

i i.

i EZ7' t udto 1 i

W ei l

.,' so. ll e= s i. f ~. #*

• 4 M47.3'* '

9. zdo.'f' 1

8a l(2. $

9 ~215.}

• 1 EZ..%7 i fl./98

.o

~

m ,... ,, _zra, e h e ... _ e - e v t- ,oc. so.

e-'

__ ___.-. _ . _ _ _ -