ML19341D400

Review of Piping Seismic Analysis Methods.

ML19341D400
Person / Time
Site:	Mcguire, McGuire
Issue date:	03/02/1981
From:	DUKE POWER CO.
To:
Shared Package
ML19341D398	List: ML19341D397 ML19341D400
References
NUDOCS 8103050533
Download: ML19341D400 (16)

Text

_ . - _ _ .- -

O ,

l i

lO Duke Power Company '

McGUIRE NUCLEAR STATION REVIEW OF PIPING SEISMIC AN ALYSIS METHODS O MARCfl 2, 1981 i

(a> DUKE POW [R ,

l o &

P103050c4>

o

INTRODUCTION O-1.0 On January 27, 1981, the USNRC requested by letter (Attachment 1) that Duke Power justify the piping seismic design approach stated in Section 3.7.2.1.2.l(a) of the McGuire FSAR because it is incon-sistent with the staff's position of Section 3.7.3 of the S.R.P.

In response to this request, Duke Power has reviewed the McGuire seismic analysis criteria and their implementation in the safety-1 related piping analysis calculations. This report describes

• results of that review and responds directly to the USNRC request.

The selection of the proper "in-structure" response spectra is an important step in assuring that the results of piping analysis (stresses, support loads, etc.) are conservative. However, the total conservatism or cumulative margin of safety in the results of piping analysis is a result of the series accumulation of margins of safety for parameters used in the complex analysis process.

Parameters and methods such as damping, response spectra develop-

O ment, the response spectra analysis technique, and modeling techniques each have conservatisms which when accumulated into a single analysis process produce conservatisms which are considerable. This report addresses these margins of safety and their incorporation into the piping analysis 'or the McGuire Nuclear Station.

The commitments stated in the McGuire FSAR have been incorporated into the piping ar.alysis as minimum requirements. As is noted in this report, seismic analysis of Me suire piping has included conservatisms which significantly exceed FSAR requirements in certain areas. .

2.0 FSAR COMMITMENTS l The McGuire Nuclear Station FSAR details commitments which directly or indirectly influence the results of piping seismic analysis. )

The following discussion of selected commitments and their incorpor-p ation into the McGuire piping analysis demonstrates conservatisms V included beyond those specifically required.

1

2.1 "In-Structure" Response Spectra The "in-structure" response spectra used in piping analysis were 2

developed using 5% critical damping rather than the 7% recommended  !

in USNRC Regulatory Guide 1.61 which, according to the Seismic Safety-  ;

. Margins Research program (SSMRP) performed for the USNRC by Lawrence Livermore Laboratory (Reference 3), are in all cases lower than those obtained from actual test data. This in itself produces approximately 25% overall conservatism in the analysis input.

The "in-structure" spectra were further modified in a conservative j manner to produce curves for 4% damping as reconsnended by Newmark j in 1973.

j The McGuire "in-structure" spectra were peak-broadened per USNRC Regulatory Guide 1.122. However, the peak was also amplified

! 10% which is in excess of the guide. Attachment 2 is an example of a typical broadened end amplified "in-structure" response spectrum used as input to the piping analysis problem.

l O

2.2 Critical Damping Values Section 3.7.1.3 of the McGuire FSAR states the critical damping j for piping as follows

" Equipment and large diameter piping systems (pipe diameter greater than 12 inches) are analyzed using 2% damping data.

Small diameter piping systems (less than or equal to 12 inches in diameter) are analyzed using 1% damping data."

~These are values applicable t'o the Operating Base Earthquake (OBE).

j McGuire piping analysis applied these values for both OBE and I the Safe Shutdown Earthquake (SSE) by applying the ratio of 15/8 I to the OBE analysis results to obtain SSE data. USNRC Regulatory ,

4 Guide 1.61 allows the use of 2% damping for piping less than or equal to 12-inch nominal 0.D. and 3% damping for piping greater than L -

12-inch nominal 0.D..for the SSE. llence, for McGuire, this method of analysis _ produces 30 to 60% increase in peak accelerations

,' for tne faulted Condition (SSE) earthquake above that required.

2

f 4

2.3 Spectra Application to Piping Analysis Section 3.7.2.1.2.l(a) of the FSAR is a discussion of seismic criteria for system piping. This section specifies that for piping ' systems scanning between two or more elevations, the "in-structure" spectra used should be the one closest to, or higher than, the center of mass of the piping system. In general, the method used for selection of spectra in the McGuire analysis was the envelope method. This will be discussed in more detail in Section 3.0. .

3.0 McGUIRE PIPING SEISMIC ANALYSIS METHODS / DATA Duke Power has reviewed actual McGuire Nuclear Station piping analysis calculations relative to the response spectra utilization.

Both in-house analysis and analysis performed by consultants were

. reviewed for compliance and conservatism with respect to current S.R.P.

requi rements.

Duke's consultant, EDS Nuclear, who was responsible for all safety-related piping analysis in the Reactor Building and selected

, systems in the Auxiliary Building, used the envelope of all applicable "in-structure" response spectra for analysis problems which spanned two or more elevations. This is in accordance with Section 3.7.3 of the S.R.P. and is part of their seismic criteria for piping analysis.

The in-house seismic criteria for McGuire piping analysis was based on Section 3.7.2.1.2.l(a) of the McGuire FSAR. However, af ter a thorough review of each piping analysis calculation,

• it was determined that the enveloping method of Section 3.7.3 of the S.R.P. was generally employed. Any modification of this envelope method has been addressed on a case-by-case basis, and calculations were performed to verify the validity of each of these deviations. Table 1 summarizes the results of the review in terms of numbers of math models which used the different spectra h application methods. From this table it is observed that all math models in the EDS scope were enveloped as well as 89 of the 111 math models in the in-house scope.

3

= . - _ . -

The six math models in Table 1 which appear in the " Multiple Spectra" category are math models which anchor at a Steam Generator

, nozzle (Main Steam, feedwater and Auxiliary Feedwater Systems).

A modified multiple spectra analysis was performed on these math models. The first part of the analysis determined the region of the piping system which was influenced by the Steam Generator spectrum. Next, the Steam Generator spectrum was enveloped with other building sr.ectra and used in the seismic analysis. This method is termed " modified multiple spectra" because instead of applying the spectra at only the supports attaching to the structure represented by the spectra (i.e. , the Steam Generator nozzle),

the Steam Generator spectra was applied to all supports in the entire math model. Stresses, support loads and valve accelerations were then evaluated for the region which was determined to be influenced by the Steam Generator spectra.

The 16 math models in the Penetration category of Table 1 are the math models which are restrained in the Auxiliary Building a with the exception of an anchor at the penetration on the Reactor Building exterior structure. These math models were analyzed O' using an envelope of the Auxiliary Building spectra but excluded the Reactor Building spectra. This was done because a review of the two buildings' spectra concluded that the Reactor Building spectra peak was too far into the flexible range to significantly influence the Auxiliary Building piping which is generally designed i

in the rigid range (> 33 hz). Analyses of four randomly selected

! math models were performed to confirm these conclusions. The Reactor Building was enveloped with the Auxiliary Building spectra and applied at all supports in the Auxiliary Building rather than just at the Reactor Building penetration anchor which would have been the case if'the multiple spectra method had been used. The significant frequencies of each math model fell above the peak frequency of the Reactor Building spectra. A comparison of the results (stresses, support loads, etc.) showed no significant increases above those generated from the envelope of the Auxiliary Building spectra in regions where the Reactor Building spectra would influence pipe response.

4

(34.0 MARG!flS OF C0f4SERVATISM IfiHERENT TO McGUIRE p!Pif4G Af1ALYSIS O

The margins of conservatism which exist in the McGuire piping seismic analysis and are reflected in results of the analysis, such as support / restraint design loads, equipment nozzle loads and stress evaluations, result from an accumulation of conservatism in analysis parameters and the analy. sis process itself. A qualitative review of parameters and methods pertinent to response spectrum usage was included in the overall review of seismic analysis methods.

This review addresses safety margins which have occurred as' a natural outgrowth of the analysis methods used and those margins existing in parameters which were input directly or indirectly into the analysis. Industry practice and special studies were considered and are referenced in the following paragraphs where appropriate.

4.1 C_ritical Damp h Q The Seismic Safety Margins Research Program (Reference 1) concluded El that two items which significantly influence calculated seismic response of a subsystem are modeling accuracy and damping. Relative to damping, several studies have been conducted throughout the industry to verify the Regulatory Guide 1.61 values. The SSMRP study along with a study by flewmark and Hall for the USNRC (Reference 2) concluded that, although damping is a parameter which is difficult to confirm due to its variability, the values for damping in Regulatory Guide 1.61 are very conservative.

For McGuire piping analysis, only the OBE values of Regulatory Guide 1.61 have been used. The piping analysis was performed for the OBE load case using the values of 1% for small pipe (< 12 in. 0.D.) and 2% for large pipe (> 12 in. 0.0.). The OBE analysis results were multiplied by the ratio of 15/8 (SSE/0BE earthquake ratio) to obtain SSE results. These include support / restraint loads, nozzle loaus, and stresses for ASME Code equations for the Faulted Conditior,. Hence, the higher values for SSE were not incorporated in the analysis procedures.

5

g The margins recognized by not using the SSE values can best be seen from examining typical spectra curves for a single elevation in McGuire Auxiliary Building. Attachment 3 is a spectra plot of the 750+0 elevation. Curve 1 is for 1% damping which is used for the OBE analysis and is essentially multiplied by 15/8 to obtain the SSE curve. If the higher damping values were used in the McGuire analysis, Curve 2 (times 15/8) would have been used to obtain the SSE input. A differential peak value of 0.5 g's exist for this

% curve and is typical for most McGuire spectra curves. This differential represents a 30 to 60% increase in peak accelerations for the faulted Condition (SSE) carthquake for the McGuire analysis.

4.2 Math Modeling Generally, industry-accepted practices have been employed in McGuire piping analysis math modeling, however, some notable differences, which represent improvements in accuracy or increased conserva-tism, have been incorporated. These can be supinarized as follows:

1) Multiple mass point valve models to account for flexible valves. Models are test verified.

2) Sophisticated computer programs - among the best in the industry.

Duke was a leader in recognizing valve flexibility concerns and performed extensive testing and math model development in-house.

Through its own personnel and consultant personnel, Duke has maintained state-of-the-art or better math modeling in McGuire piping analysis, which has been performed over a period of rapidly changing methodology in piping analysis (1970's).

4.3 Response Spectrum Development Significant margins were introduced into the piping analysis from the way the "in-structure" or floor response spectra were developed. The McGuire spectra are peak broadened per Regulatory Guide 1.122 as discussed in Section 2.0. Piping systems are generally

d ,.

j flexible and possess many modal frequencies below the rigid range (33 bz) which leads to the occurence of multiple piping frequencies

! coinciding with peak frequencies of the input spectra due to the peak bro'adening. This implies that multiple frequencies of the pipe system are in resonance with a single building frequency.

1 This- of course, is physically impossible. Additional margin was introduced at this step'of the analysis since McGuire piping analysis l permitted multiple piping frequency excitation by the broadened >

f  % input response spectra. This margin is particularly pronounced i

because the response spectra peak to which each coincident piping mode is excited was developed with conservative structural damping and was then amplified by 10% as discussed in Section 2.0. If these 4 modes are closely spaced, in addition, then modal responses were combined absolutely in lieu of the square root of the sum of the i

squares.

4

]

4.4 Design parameters l

O "cce're nini"9 ' isis s ce aectea s st 8 ra nr cti<e "s'"9 the design temperature and pressure condition in lieu of the normal

operating temperatures and pressures, except for Class 1 piping requiring operational mode evaluation. This results in stresses,

support / restraint loads and nozzle loads which are more conservative than required.

1 i

4.5 Rye,sponse Spectrum Analysis Methods i piping analysis employing the single response spectrum method is very conservative, especially when the piping is actually supported from more than one building (spectra) elevation. The SSMRP con-cluded (Refereve 1) that major conservatism arises when a subsystem is supported from elevations which have significantly different responses and the envelope of all of the "in-structure" spectra I

is used for analysis. When this enveloping is employed, phase and

-amplitude relationships are lost and conservatisms are introduced into the analysis input. The SSMRP also concluded that the time-l history analysis is more applicable for this application and is

. generally considered to be more realistic. Ilowever, the time-7

e a

history method is much more costly in both manhours and computer charges. This nethod is generally reserved for more sophisticated applications such as for the analysis of- coolant loops and generation of "in-structure" spectra.

I Multiple response spectrum methods (methods where the "in-structure" r

spectrim is applied only at supports actually attaching at that elevation) are gaining acceptance in the industry and are generally less conservative than the single response spectrum method and less expensive to apply than the time-history method. These nethods are presently being used in the industry where computer capability exists and for cases where there are significantly different spectra I involved such as Steam Generator spectra versus the Reactor Building spectra.

I j

5.0 CONCLUSION

S

} Section 2.0 provides a brief suninary of pertinent commitments outlined in the McGuire FSAR for piping seismic analysis.. Each

consnitment was reviewed against actual application to show that each was met properly and conservatively.

Section 3.0 summarizes techniques used in McGuire piping seismic analysis and, specifically, the application of the response spectra.

Dasically, the envelope method in Section 3.7.3 of the S.R.p.

l has been employed either directly or with some slight modifications

) as justified by additional analysis.

5 l

Section 4.0 outlines additional conservatisms inherant in the McGuire piping analysis. One of the most significant conservatisms employed has to do with the response spectra actually used, due to the damping used in its development along with the 10% increase in peak amplitude.

-Another significant margin exists due to damping values used in the Faulted Condition evaluation. Considering values in Regulatory Guide 1.61, which are already considered to be conservative, increased conservatism is employed by not using *he allowed higher values for SSE. This produces Faulted Condition results such 8

1

( as support loads and pipe stresses with a significant amount of

, f^);

\

conservatism not required.

This review of the McGuire piping seismic analysis methods shows that analysis incorporates conservatisms which meet or exceed those required. After review of the calculations, it is apparent that Section 3.7.3 of the S.R.P. has been met or adequately justified otherwise. Overall, the level of conservatism is found

% to exceed S.R.P. requirements, hence adequacy relative '.o require-ments is clearly demonstrated.

nm 9

N

{s/ REFERENCES

1. Kennedy, R. P., R. D. Campbell, D. A. Wesley, H. Kamil, A. Gantoyat, and R. Vasudevan, Seismic Safety Margins Research Program,

" Subsystem Response Review", NUREG/CR-1706, 1979.

. 2. Newmark, N. M. , W. J. Hall, "Developnent of Criteria for Seismic Review of Selected Nuclear Power Plants", NUREG/CR-0098, May, 1978.

3. Healey, J. J. , S. T. Wu, M. Murga, Seismic Safety Margins Research Program, " Structural Building Response Review", NUREG/CR 1423, Vol. I.

O o

V 10

f O

TABLE 1 Summary of

' Math Model Review for Response Spectra Application .

TOTAL TYPE OF SPECTRA APPLICATION SCOPE '

MATH MODELS ENVELOPE MULTIPLE SPECTRA PENETRATION EDS 244 244 0 o IN-HOUSE 111 39 6 16 O

11

)){'l7UW f'I?), Attachment 1

'[ k [ N UNITED STATES

{ ) y 'a,; rtu 2 jogdUCLEAR REGULATORY COMMISSION

.* \ r e

. mIII WASHING TON. D. C. 20$55 j

~'s (V  %, v.....#u . . , -

o.

a m ... . n w ,: m ; m 2's a 4

Docket Nos.: 50-369 and 50-370 ,

Duke Power Company .

ATTH: Mr. William O. Parker, Jr. -

Vice President - Steam Production P. O. Box 33189

' 422 South Church Street Charlotte, North Carolina 28R2

SUBJECT:

JUSTIFICATION OF PIPING SEISMIC DESIGN APPROACH (MCGUIRE NUCLEAR STATION UNITS 1 & 2)

It is stated in Section 3.7.2.1.2.1(a) of the McGuire Nuclear Station FSAR that "For a piping system spanning between two or more elevations (spectra),

the nectrum curve associated wii.'. thG elevation closest to, or higher than, the center of mass of the piping system is used".

This approach is inconsistent with the staff's position delineated in Section 3.7.3 of the SRP. The staff's position requires that an envelope of all pertinent response spectra be used for the analysis of systems spanning between two or more supports.

It is requested that you provide technical justifications for the deviation and demonstrate that your approach would result in an equivalent or more conservative analysis than is required. The j Jstifications may consist of a comparison of inputs, a comparison of selective resuits, or a combination of both.

Sincerely.

• Mc.&

Robert L. Tedesco, Assistant Director for Licensing l

Division of Licensing cc: See next page O

12

.- - 4 -- --

Attachment 1

. ., Sheet 2 of 2 l

r . William O. Parker , Jr .

ice President, Steam Production uke Power Company P. O. Box 2178 422 South Church Street Charlotte, North Carolina 28242 cc: Mr. W. L. f'orter Dr. Cadet H. Hand, Jr. , Director Duke Power Company Bodega Marine Lab of California P. O. Box 2178 .

P. Q. Box 247 422 South Church Street Bodega Bay, California 94923 Charlotte, North Carolina 28242 Richard P. Wilson, Esa.

Mr. R. S. Howard Assistant Attorney General Power Systems Division State of South Carolina Westinghouse Electric Corporation 2600 Bull Street P. O. Box 355 Columbia, South Carolina 29201 Pittsburgh, Pennsylvania 15230 Office of Intergovernmental Relations Mr. E. J. Keith 116 West Jones Street EDS Nuclear incorporated Raleigh, North Carolina 27603 220 Montgomery Street San Francisco, California 94104 County Manager of Mecklenburg County 720 East Fourth Street Mr. J. E. Houghtaling Charlotte, North Carolina 28702 NUS Corporation e' s 2536 Countryside Boulevar6  : U. S. Environmental Protection Agency

) Clearwater, Florida 33515 ATTN: EIS Coordinator Region IV Office Mr. Jesse L. Riley, President 345 Courtland Stree.t. N. W. -

The Carolina Environinental Study Group Atlanta, Georgia 30308 854 Henley Place Charlotte, North Carolina 28207 Mr. Tom Donat Resident Inspector McGuire HPS J. Michael McGarry, !!!, Esq. c/o USNRC Debevoise & Liberman Post Office Box 216 1200 Seventeenth Street, N. W. Cornelius, North Carolina 28031 Washington, D. C. 20036 .

Shelly Blum, Esquire Rober t M. Lazo, Esq. , Chairman 1402 Vickers Avenue Atomic Safety and Licensing Board Durham, North Carolina 27707 U. S. Nuclear Regulatory Commission Washington, D. C. 20555 Dr . Emmeth A. Luebke Atomic Safety and Licensing Board U. S. Nuclear Regulatory Commission Washington, D. C. 20555 13

o)

L, q

Qj ME , IRE NUCLEAR STATIGN. .

RESPONSE ACCELERRIION SPECTRA. DAMPING = 0.020 l' AUXILIARY BUILDING. E-W DIRECIION, EL. 750+0 SPECTRA MODIFIED FOR o .

4% Structural Damping

~

Response Spectra Construction O IF, F-7, (Newmark 1973)

US ..

l 3 i

d w l g, ZS  %  ;

w I \

c \

k -' ' ,

t

\

AEJUSTE] SPECTR JM y$

... l- j(

\

a j 'n -

% I W ~.' TCc m

[

E f CALCULATED SP CTRUM '

Vy V '(

a-

/

4 .

,/ v a q

o' 7  %< \

.e a

g

<d o

o

%  % n=

1. 1 ). 05 p.10 0'.20 0.50 l'.00 2'.00 5'.00 10.00

.- PEfl100 IN SECONOS

.e ,e ame , ., e.ee

• eee e * * * * * '

• O O O McGUIRE NUCLEAR STATION i

RESPONSE ACCELERATION SPECTRA AUXILIARY BUILDING, E-W DIRECTION, EL. 750 + 0

? r, d l n; j l.

E SPECTPA M03IFIE0 FOR

, , 4% Structural Damping N < t s J I Response Spectra Construction I )l i

{ ~ [I (Newmark 1973)

S. *1 5 -

w 'l E z ll I! I

~

~

E ik Ib

, <2 l# \ 't

^

lI I \\

y j

1. \

s C o

b,,i \

r . ,( urve1('% l

. _ Damping)i l l .

5 w

D \, '-J_l ,

~

j li it ~ , .

, w y, .. 9 S, f l l' ' l' ' , .

2 o ' s N l/ 1 g

( C arve 2 (2% Damping)

\ ls i's f l L / Q\ E w

=

E/ l' st "r

=

5 ~

r,' R s s 5

' - ' , -e ,, w a ** %

o s g i , * -

0.05 0.10 0.20 0.50 1.00 2.00 5.00 10.00 PERIOD IN SECONDS

. - _ _ _ _ - -