ML20150B720

Forwards Responses to Encls of Kn Jabbour 880524 Ltr Re Direct Generation Response Spectra

ML20150B720
Person / Time
Site:	Catawba
Issue date:	07/05/1988
From:	Tucker H DUKE POWER CO.
To:	NRC OFFICE OF ADMINISTRATION & RESOURCES MANAGEMENT (ARM)
References
NUDOCS 8807120143
Download: ML20150B720 (21)

Text

- .. . . .~. ,

l  ;-G

, y..; Q :

A:

DUKE Powen GOMPAhT

' P.O. BOX 33180 ~

CHA HLOTTE, N.C. 28949 HALH. TUCKER TELEPHONE m mpaamment (704) 373-4831 stuBAR pmODocTwm July 5, 1988 U. S. Nuclear Regulatory Commission Attention: Document Control Desk Washington, D. C. 20555

Subject:

Catawba Nuclear Station Docket Nos. 50-413 and 50-414 RAI, Direct Generation Response Spectra

Dear Sir:

i Attached are responses to the enclosures of K. N. Jabbcur's Ray 24, 1988 letter of the same subject.

1 Please let us know if further information on this subject is needed to complete your review.

Very truly yours, Hal B. Tucker PGL/33/sbn i

1. i Attachment  ;

l xc: Dr. J. Nelson Grace, Regional Administrator l U. S. Nuclear Regulatory Comission j Region II j 101 Marietta Street, NW, Suite 2900 i Atlanta,. Georgia 30323 Mr. P. K. Van Doorn NRC Resident Inspector Catawba Nuclear Station

~

8807120143 08070543 f PDR ADOCK 050 1 P

ENCLOSURE 1 STRUCTURAL AND GEOSCIENCES BRANCH DIVISION OF ENGINEERING AND SYSTEMS TECHNOLOGY REQUES1 FOR ADDITIONAL INFORMATION

1. G.ENERAL COMMENT

1. With reference to your February 24, 1988 letter (page 2, first paragraph), provide the technical basis for your statement that,

"!.he Direct Generation method results in more accurate Design 27.uponse Spectra." Additionally, a reference to IRC's NUREG/CR 3461 was made presuming that the Direct Generation method was recommended by NRC staff. It should be noted that positions or recommendations of a NUREG/CR report do not represent staff ,

positions and should not be construed as such. However, the proposed revised version (published for public comments) of SRP 3.7.2 (Ref.1) allows the use of direct solution methods but requires a review of such methods on a case by case basis.

Rasponse The direct generation method does not result in more accurate Design Response Spectra than the time history method. The two methods provide comparable results. The raathod also provides consistent, repeatable results, as compared to the time history method.

II. SPECIFIC REQUEST FOR INFORMATION

1. When proposing a generally unverified gt yet to be proven dynamic analysit method for use in qualifying setc'ic design adequacy of safety related structures, systems and 6autpient, a mere comparison of limited artlysis results with one set of time history analysis based on one single structural model appears to be inadequate. A more vigorous justification of the method is needed. Specifically. provide any applicable test verification data which support the general applicability of tha direct generation method.

Response

The direct method of response spectra generation is an accepted and ver'.fied analysis tool. References 1 through 6 provide examples of the research and verification using direct generation techniques.

In addition, direct generation techniques have been approved by the NRC for the generation of in-structure response spectra (Ref 9).

J Page 1 i

NVPEG/CR-1161 concludes that the "Standard Review plan should give equal weight to the use of both time-history analysis methods and direct solution methods for the generation of in-structure response spectra." A draft revision of the Standard Review Plan recommends that direct solution methods to be reviewed and accepted on a case-by-case basis. The two key points to be reviewed are:

o the theoretical basis of the technique must be demonstrated o selected comparisons are required between direct solution results and results from a time history approach The theoretical basis of the technique is discussed in NUREG/eR-3480, which includes a general discussion of direct generation methods as

well as comparisons between the method developed by Singh (Ref 1,2) and time-history results. The concept of a spectrum-consistent power spectral density function was used in the development of the Singh method. The report states that the "method produces excellent, consistent, and repeatable results as compared to time-history approaches." It also notes that direct generation methods are "based upon sound theoretical backgrounds and are suitable for adaptation on computers."

The methods used in EDASP arc based upon the use of a spectrum-consistent power spectral density function. The basic eauations and fundamental concepts used in EDASP are the same as those used by Singh. The procedures used in EDASP are identical to those recommended in NUREG/CR-3875, which were verified in the EDASP Theory and Verification Manual (Ref 7).

. Selected comparisons, in addition to those described in Reference 7, were preformed for tb3 specific application and structural model for which the direct generation method was used. (see response te questions 6 of Enclosure 2). These results showed a favorable  !

comparison between the direct generation method and the time-history '

method.

1 References. I i

1. M. P. Singh, ' Generation of Seismic Floor Spectra," Journal of Engineering Mecl;anics Division, ASCE EMS (October, 1975).

2. M. P. Singh, "Seismic Design Inout for Secondary Systems," ASCE Mini-conference on Civil Engineering and Nuclear power, Session 11, Boston, Massachusetts, April, 1979. Volume 11.

3. R. H. Scanlan and K. Sachs, ' Development of Compatible Secondary Spectra without Time-Histories, "Trnnssetions of the 4th International Conference on Structural Mechanics in Reactor Technology, San Francisco, California, August, 1977.

Page 2

4. H. Sato, M. Kamazaki, and M. 0hori, "An Extensive Study on a Simple Method Estimating Response Spectrum Based on A Simulated Spectrum,"

Transactions of the 4th International Conference on-Structural-Mechanics in Reactor Technology, San Francisco, California, August, 1977.

5. D. Schnitz and K. Peters, "Direct Evaluttion of-Floor Response Spectra from a Given Ground Respor.se Spectrum, "Transactions of the 4th International Conference on Structural Mechanics-in Reactor Technology, San Francisco, California, August, 1977.

l

6. K. Peters, D. Schnitz, an V. Wagner, "The Problem of Resonance in the Evaluation of Floor Response Spectra, "Transactions of the 4th .

International Conference on Structural Mechanics in Reactor Technology, San Francisco, California, August, 1977.

7. EDASP Version 1.1 Theory and Verification Manual, Stevenson &

Associates, Woburn, Massachusetts, September 1, 1986.

8. EDASP Version 1.1 Users Guide, Stevenson & Associates, Hoburn, Massachusetts, September 1, 1986.

2. The licensing basis for Catawba utilizes the Newmark Spectra, termed as Design Response Spectra (DRS), anchored at 0.15g peak ground acceleration (PGA) for design of structures. In the stsff's Safety Evaluation Raport (Ref 2), it was pointed out that the site specific spectra based on Perry and Wolf Creek site analyses exceed the Catawba DRS by 15 to 16% between tha frequencies of 3 to 10 Hz. However, the average spectrum of the four synthetic time-histories used in generating the floor response spectra exceeds the site specific spec.trum at all frequencies. Thus, the staff had accepted a dual approach for design purposes (DRS for structural design, average spectrum obtained from the synthetic time-histories for equipment design). Your proposed approach is based upon the direct generation of the floor response spectra from the DRS. Provide information on how the exceedance of site specific spectra is incorporated in the proposad approach.

Response

There is no site specific spectra for Catawba Nuclear Station. Our approach is based upon using the licensing basis Design Response Spectra. Because there is no s)te specific spectra for Catawba, there is no exceedance to be addressed.

In the SER, the staff indicated that a Reg. Guide 1.60 R S anchored at

.13 PGA would be appropriate for Catawba. A comparison at the PSDs for the Catawba DRS and the RG 1.60 0.13 PG4 R S was performed (Figure attached). The area under the PS0s were calculated giving a relative measure of the energy content of the earthquake. This value for the DRS for Catawba is 1.14 times larger than for RG 1.60.

Page 3 4

5 5 1 1 _

c .

c .

e Se Ts S s- Rs= Oe R .

L/

0 . P0 .

Ps P2d 02d .a S e 6 e D/

S2 _

R = e 1 = e P D c G c _

mo C nx R nx oE oE e _

i :i rn go S eab t .

eab t . oi rs R l ro l ro iur Pi v

D i ur FDP FDP Cei PF a _

0 b

wse 0 0

ai j i - - _ -

l g 1 _

t ti  ! I I

es n I

C e ' '

D ' -

l sar -

vt c 0 ).

e E 0 z 0 p l 1 H(

S l l 6 l I

. r i I

y _

e -

c 1 w o ' i n _

e _

\

P _

u

'.' q G -

- e r

._ F

. _ 0 0

R l I

1

(

l I l

' l l

\ i

' i i

' 'o -

j '; ' '

0

- - _ - 1 4 J 4 M 0 0 0 O 0O

- - - - 0 E E E E 0 0 5 0 0 0 2 5 5 3 2 1 7 7 NINn(9 >N ha ~ eyoea.ne te2O2 s .

1

?E^ -

3. Provide the basis for the frequency interval selection used in the direct generation method.

Response

The frequencies used in the direct generation method are selected to provide a good resolution in the Response Spectra and to attempt to avoid missing peak responses. The frequencies are spaced at even intervals on a logarithmic scale. In addition, significant building frequencies are inserted.

The following table provides a comparison for the frequency intervals 4

suggested in Reg. Guide 1.122 and those used in direct generation.

Frequency Reg. Guide 1.122 Direct Generation Range Increment Increment Range 2 - 3.0 .10 0.012 - 0.14 3.0 - 3.6 .15 0.14 - 0.17 3.6 - 5.0 .20 0.17 - 0.23 5.0 - 8.0 .25 0.23 - 0.41 8.0 - 15.0 .50 0.41 - 0.69 15.0 - 18.0 1.0 0.69 - 0.84 18.0 - 22.0 2.0 0.84 - 0.95 22.0 - 34.0 3.0 0.95 - 1.58 The number of frequencies for which a response is calculated in the l alrect generation is redccea from that originally committed to in the FSAR for time histories; however, this is acceptable because the input motion is a smooth curve and therefore the resulting spectra will be a much smoother curve than what the time history method would yield.

4. With reference to page 3.7-17, provide a detailed discussion of verification work done on the Equipment Oynamic Analysis Package (EDSAP). As applicable, provide an actual r-aparison and i verification results of EDSAP with respect t: observed floor l responses of a nuclear power plant with known ground responses  !

(e.g. Humbolt Bay records) to support the validitv of the EDSAP. )

I l

a s

Page 5

Response

The test problem used in the EDSAP Theory and Verification Manual was an eight noded eccentric frame. Modal' analyses and time-history dnalyses were performed using the STARDYNE finite element computer code. Ten separate response spectra analyses were performed using a variety of input excitation directions and combinations. In each case the direct generation results showed an excellent agreement with the time-history results.

Comparisons to actual earthquake responses would be impossible without an accurate math model of the structure (natural frequencies, mode shapes, and mass distribution). Therefore, as noted in the proposed Standard Review plan revision, the recommended method of verification is comparisons with time-history methods.

5. Pages 3.7-17 thru 3.7-17e primarily represent a direct copy of a referenced paper (Ref 33 in your February 24, 1988, submittal) without any discussion of the rationale, assumptions and limitations of the method. Specifically the impact on the reliability of the method due to the lack of an indepth study regarding the effect of varying the effective duration of the strong motion portion of the earthquake, "T", and the probability of exceedance "r" should be discussed. Also provide justificaticns regarding the EDASP's applicability to category I structure floor response analysis considering the fact that the method is primarily developed for component test / qualification work.

Response

In a discussion about probability of exceedance "r" and effective duration of the strong motion part of the earthquake "T", the EDSAP Users Guide notes that "the results are not sensitive to these two parameters if the user specifies response spectra as input, since the output RS will be consistent with the input RS."

Although the EDSAP program was originally developed for equipment qualification, the theory, equations, and application of the program  !

are not limited to equipment. The program is based on fundamental dynamic theory, which is identical for equipment, structures, and other i^ ems which can be adequately modeled using frequencies, mode shapes, id mass distribution.

l Page 6 j l

l

l

6. Provide clarification of the paragraph (page 3.7-17d) starting  !

with "A power spectral...EDSAP program." Also provide a step-by-step procedure used to obtain the values of Table 3.7.1-1.

Response

The paragraph was intended to provide the procedure used in developing the data in Table 3.7.1-1. Procedure follows:

Step 1 A PSD is generated for the time history ground response spectrum.

Step 2 The following results from the building model analysis are input:

a) Joint Coordinates b) Member Connectivity c) Masses d) Mode Shapes e) Model Participation Factors Step 3 Response locations and frequencies are selected.

Step 4 Structural dar. ping values are set and response damping values selected.

Step 5 Program uses information from steps 1, 2, 3, and 4 to generate an amplified floor PSD.

Step 6 The PSD From step 5 is converted to the floor response spectra.

The values in the columns headed EDASP are the selected frequency output points in step 3 above. Basis for these is Reg. Guide 1.122. The values in the column headed T.H. are from the R.S. generated by the Time His tory procedure from FSAR. They are limited to the maximum acceleration in proximity to frequencies in EDSAP column.

T.H. (AVE) column gives the value for acceleration corresponding to trcquency to T. H. column.

TH-EDAS column gives value from Amplified PSD to R. S.

conversion for frequency in EDASP column.

Page 7 t__

8 ENCLOSURE 2 MECHANICAL ENGINEERING BRANCH

' DIVISION OF ENGINEERING AND SYSTEMS TECHNOLOGY REQUEST FOR ADDITIONAL INFORMATION

1. State the condition under which the option is chosen to use the alternative set of damping values for piping, as shown in Section 3.7.1.3. Note that under the conditions stated in Regulatory Guide 1.84, Rev. 4, the alternative set of damping values as stated may not be acceptable to be used for piping analyses based on floor spectr? generated by the Direct Generation Method.

Response

The alternative set of damping values given in FSAR Section 3.7.1.3 may be used for reanalysis of any system designed for seismic loads for either modifications or support / snubber optimization. Either the original or alternatives damping values will be selected for each analysis. No combinations of the two damping criteria will be used. The alternative damping values will not be used in a time history analysis. Seismic displacements resulting from use of the alternative damping values will be reviewed to assure calculated displacements can be accommodated.

Spectra generated by the direct generation method has been shown to be consistent with those generated by the current time history method. The direct generation method is being used to economically generate response spectra for the alternative damping values. The method is equally applicable to RG 1.61 and the alternative damping values. All the restrictions of RG 1.84 have been satisfied.

2. ASCE Section III, Subsection NF, indicates that loads due to piping restraint of free end displacements should be included in the design of supports under Service Level D conditions. Stato why thermal loads are not included in the load combination for faulted conditio, s shown in Table 3.9.3-11.

Table 3.9.3-11 was submitted as part of the preliminary draft of a FSAR update. This revision to the table editorially modifies under the j upset condition the load ' Thermal Anchor Movement" to be called "Thermal ,

Transients' so as to agree with tables 3.9.3.7 and 3.9.3.8, load I combinations for Duke Classes A, B, & C. In response to request No. 2 on Enclosure 2, the faulted support load combination shown in Table i 3.9.3-11 meets the requirements of the ASME Code of record ASME Section III 1975 Ed. Summer 1974 Addenda). This load combination had been previously evaluated and accepted by the NRC in NUREG-0954, Safety Evaluation Report Related to the Operation of Catawba Nuclear Station, Units 1 & 2, docket nos. 50-413 and 50-414, February, 1983.

The submission of Table 3.9.3-11 is irrelevant and unrelated to the use of a response spectra genarated by the Direct Generation method.

Page 8 i

3. The Direct Generation Method is stated to having been proven an accurate method for generating floor response spectra. Provide a detailed discussion showing the basis for this assertion, and provide supporting experimental data if available.

Response

See response to I.1.

4. Indicate if peak broadening per Regulatory Guide 1.122 will be applied to the f'.oor spectra generated by the proposed method.

Response

Peak broadening will not be applied to the floor spectra per Regulatory Guide 1.122. The Peak will be broadened using the method presently stated in the FSAR.

5. Indicate how this procedure is used for a structure with different damping values.

Response

The input requirement of EDSAP is one structural damping value. This is consistent with the original time-history analysis.

6. Provide the data in Tables 3.7.1-1 in graphical form, from 1 to 40 Hz.

Response

l

See graphs on following pages.

I l Page 9 i

l

Direct Generation vs. Time History .

0.5% Caitical Damping at Eley. 562+0 100.0 , , ,, . .

i gg File : EDASP-1.rs Damping = .005 10.0 c,, -

o Z --

__. File : TH (av) 1.rs

$ -- ~

c Damping = .005

• 2

+3 m - ,

C.

m m . -

i 4: 1.00

,ma\

l ~ ' ' '

& f ',

~

's, --

's, -'

PC Pr ogr-en RS.J8 LOT 0.10 - - ' ' III ' III - ' ' 'III 0.10 1.00 10.0 100.0 Frequency (Hz .)

Direct Generation vs. Time History .

0.5.~ Critical Damping at Elev. 595+4 100.0..-. -

- i l lll

- i l lll i

s ig

~

File : EDASP-3.rs Damping = .005

" 50.0 2 Z E ___ File . TH (av) 3.rs a __

o

@ Damping = .005

~ -,

(

u e -

) . -

C- '.

e

,a -

e ,' \

u .

E i.00  ? ''"'

l

l gl'J

&, 2'

- '\ -.

PC FYogram RS_PL.0T 0.10 . - i Illl - - i 1111 - i Illl 0.10 1.00 10.0 100.0 Frequency (Hz .)

Direct Generation vs. Time History .

5% Critical Damping at Elev. 595+4 100.0 * '

' s '

8 l l ll Illl Ilg

~

File : EDASP-4.rs Damping = .05

. 10.0 (D =

i

? _ _

____ File : TH (av) 4.rs o

@ Damping = .05

~

w - -

.a e - -

C.

,e . .

e o

u ' )

< < 1.00 _

t

[ (% -

i 3 PC Program RS_ PLOT

' "# 3 o,101 . . . IIll . . . 1111 . . . till 0.10 1.00 10.0 100.0 Frequency (Hz . )

4 A

l

Direct Generation vs. Time History .

0.5% Critical Damping at Elev. 628+8 100.0 '

_ Illl Illl llg

__. File : EDASP-5.rs Damping - .005

- 10.0 .

to 2

e

.__ _ File : TH (av) 5.rs

• Damping = .005 E [c C * -

e ,

m _ -.

C )

e . .

g m ,,' i u .<v N i.00 #

l' ? '

m enem.

,n' ,,

~. _

PC Progt*em RS PLOT 0.10 . - i Illl - - i 1111 - - i IIll

0.10 1.00 10.0 100.0 Frequency (H

.)

i

Direct Generation vs. Time History .

5% Critical Damping at Elev. 628+8 100.0 , , , , , ,

_  ;,, , ,g Z -

File : EDASP-6,rs Damping = .05

- 10.0 to -

e Z -

____ File : TH (av) 6.rs "S c Damping = .05 i

O i -

• w _ --

s2 m . -

C.

e ~ .

H e i s o

o i < 1.00 -

4 m.

PC Program RS_Pt_0T 0.10 - - ' fil! - ' ' f lII- - ' ' IIII 0.10 1.00 10.0 100.0 Frequency (Hz.)

T Direct Generation vs. Time History .

0.5% Critical Desping at Elev. 662+0 100.0 . , , -

, igg

- i illt 'IH 3 __.

~

File : EDASP-7.rs Damping - .005

- W

. 10.0

~ __.

? -

- TH (av) 7.rs

%  ! ',) ~

g .

Damping = .005 G j -

m -

i

c. ,

N ,

u -

r '

M i.00 /

~_._'

,, \

PC Progrea RS_. PLOT igg]

Revision i 12/3/87 0.10 . , , illl - - ' ilII O.10 1.00 10.0 100.0 Frequency (Hz .)

Direct Generation vs. Time History 5% Critical Damping at Elev. 662+0 100.0 '

- i '

_ IIll iIII 3 IR'

~

File : EDASP-8.rs Damping - .05 '

i 10.0

'm 2 ~ _

C __ _ F . le : TH (av) 8.rs

% c r_

~

Damping = .05

• S _ _

.o m - _.

t. ,

e .

-e e

u 1

M 1.00 4

q -

- ~

PC Program RS_ PLOT

' III Revision i 12/3/87 0.10 - i n t] . . i till -

O.10 1.00 10.0 100.0 Frequency (Hz .)

]

Direct Generation vs. Time History 0.5% Critic 71 Damping at Elev. 691+2 100.0 - a l lll

- i I 11l

- i sig File : EDASP-9.rs Dampitig = .005

" 'l

! - 10.O'

! E - ,

e 2 _-

_ _ File : TH (av) 9.rs g

a

@ ,- Damping = .005 0 y -

I  ;.

e . l ,

C.

,e .

u

s. .

i e ,,

i u i*,

u >

4: 1.00 -

~~

Y.~.' .

i PC Progres RS_ PLOT

Revision i 12/3/87

. 0.10 - - a iIll - -

• IIII - - a lill

! 0.10 1.00 10.0 100.0 f Frequency (Hz .)

l 1

Direct Generation vs. Time History .

5% Critical Damping at Elev. 691+2 100.0 - -

= s

s I-ji-i IIll lll File : EDASP-10.rs Damping - .05

. 10.0 j { } ____ File : TH (av) 10.rs g Damping = .05 g - _ --

e e _ n

c. '

e .

g ..

e i o '

E 1.00

~

l -

% 5 pf _

OC Progr se RS_ PLOT 0.10 - i Ill! - - a lill - i' Ill!

0.10 1.00 10.0 100.0 Frequency (Hz .)

Direct Generation vs. Time History .

0.5% Critical Damping at Elev. 713+1 100.0 ,, . -

sigg

. i

,,II Ilit Z

~

File : EDASP-11.rs Damping = .005 10.0 )

l '

[* Z ,i

._ _ _ _ F i l e : TH (av) 11.rs f Damping = .005 E '

= -

+*

e - ,

c- ,

e .

r4 ,

e ,

o - , sly' U

< 1.00 , ' v, ,  %

, y -

PC Program RE_ PLOT 0.10 - ' 'III - ' ' 'I ^ '

O.10 1.00 10.0 100.0 Frequency (Hz . )

i

1 1

Direct Generation vs. Time History .

5% Critical Damping at Elev. 713+1 100.0 '

_ Illl IIll I i-ti File : EDASP-12.rs Damping - .05

. 10.0 -

e

~

[* { } ____ File : TH (bv) 12. rs Damping

@ =

.05 8 ; -

e -

) -

c.

e m . .

e \

u o

< 1.00 Z , Z f _

PC Program RS_ PLOT 0.10 - i Ill! - - i 1111 - - i IllI 0.10 1.00 10.0 100.0 Frequency (Hz .)

_ _ _ _ _ _ _ _ _ _ _ _ _ .__ __ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ - _ _ _ .