ML20086T100

Adequacy of Seismic Margins Assuming Increase in Amplitude of Diablo Canyon Seismic Long-Term Program Horizontal & Vertical Ground Motions Described in Sser 34

ML20086T100
Person / Time
Site:	Diablo Canyon
Issue date:	12/31/1991
From:	PACIFIC GAS & ELECTRIC CO.
To:
Shared Package
ML17083C388	List: ML20086T100
References
NUDOCS 9201060135
Download: ML20086T100 (39)

Text

1

\ \ .

1 n,

1l f

DIABLO CANYON L LONG TERM SEISMIC PROGRAM PACIFIC GAS AND ELECTRIC COMPANY Diablo Canyon Power P! ant Docket Nos. 50-275 and 50-323 t

N I *<

[4'7< nj f) f'** % :e e bl .:99f-ig.%%<.D-

- 4 - -

v.g p.. y 6 '

+ ' , D:;ss 3 f$ggJy"Ij@@ .-

.{ m'Ng{, "

~ , .w

,_ g , Qg g ..n . .n-grv> en.

. m3,'I J'f f N; w' % ,. .

. r. A x ,4, -

, 4 , -

.I +

," ~> 6 p .,e ' ,,-. , , . .[A, 7

m

,ca

~ ~

[ Y ' ..e'*I $ ~' '

r v

w w w w"" We bgg (8.5. re: .n . n. . s~ % 4 .

x :i. <

p: c; mw yw .g 4...%

1 , ps , y%

.v.

s  :.

n.. ,gw m w. ,gg %sq.r,,.4, y-_, ; ,.mwn

.,.,q., .f., .. o, .e. ,

j; 4. .  ;

. ; ,e , ':. .e, gy .. ,

,,

• ls *

, > . ,: f

" jg e,79 C., g . v.

x,, ,,,, .

.e= )% +a . ,4,: j,,

• g'.,

f. , . ,

"  ;. g' Ng,[*9 b J.r e

] ,'. *,;. ?fM ,+, %g % f.M A <

, ,, sinii

> mJ.:, ct -.J-y,._ ~

- -- i .,. .

[! e .: ~'- f; 9 ,. ,,

,y- e s=v. ...,+ [1,4. v. , .. yw+; g.,. .c . -,

"y

%.a . .,* . a y , . . ,, .

" 1ly, 'y, ,n ~ ~.,

~

' * * ;-  : 1

• y f, 4* l? N , *

'.,]f w a..3 t g,j;

'. c;

.,) ,.- -

y=pg.,y..ks$'.{k.p$frn;f,,,lcq,;:.,?;%,",>.xQg.jw:y D

- , - . ~.

,. .Q ,%; .

3

, pr r, ,

e.. 9 , p c.,,._ ay,,.s. fi .i .

' =

. , ., y

• . ,

• e 23 ,; p. ,

hLAyjy, m g -Q _

g,Qf f,,4iWGJ g) ) &q a ,Q - {r) y . . n ;4 ,, fg y

b

't  ;

. . 3 -w, %- w,!M 'M

-. w w w@&p %g;ysg,.

dDR$dt y

M _y ~geacK_

,v0000 m , '.6 "

._,. I

[= ,;e __

ff . h. y p p q W Y

W r

l l

l l ADEQUACY OF SEIShilC h1ARGINS ASSUh11NG AN INCREASE IN AMPLITUDE OF TIIE DIABLO CANYON I LONG TERh1 SEIShilC PROGRAh!

IlORIZONTAL AND VERTICAL GROUND hlOTIONS AS DESCRIBED IN SSER 34 l

l Deamber 1991 l This volume responds to confirmatory analysis asked of PG&E by the U.S. Nuclear Regulatory Commission (NRC) staff in their Safety Evaluation Report, Supplement No. 34, dated June 1991, to confirm PG&E's conclusions that the plant seismic margins are adequate to accommodate the horizontal and vertical spectral exceedances that result from use of the staff's estimates of horizontal and vertical spectra.

Diablo Canyon "ower Plant

. ' Pacific Gas and El:ctric Company Long Term seismic Program

Confirmatory Analysis, December 1991 Page1 ADEQUACY OF SEIShilC hlARGINS ASSUh11NG AN INCREASE IN AhlPLITUDE OF TIIE DIAtlLO CANYON I.ONG TERhl SEIShilC PROGRAh!

IlORIZONTAL AND YvRTICAL GROUND h10TIONS AS DESCRIBED IN SSER 34

] At the request af the U. S. Nuclear Regulatory Commission (NRC) Staff, PG&E submitted discussions

,1 of the effects on the Diablo Canyon Power Plant structures and equipment of an assumed increase in I amplitudes of the Diablo Canyon Long Term Seismic Program (LTSP) 84th-percentile site-specific ground l motions (PG&E,1988) in the horizontal and vertical directions at certain discrete frequency ranges

[ (PG&E,1991 and PG&E,1991a). The NRC Staff reviewed these submittals and accepte;t PG&E's conclusions that the plant seismic margins remained adequate to accommodate the postulated ground motion exceedances. However, m the Supplemental Safety Evaluation Report (SSER) No. 34, the NRC I Staff required that PG&E perform analyses to confirm that the seismic margins are adequate to accommodate the Staff's estimate of spectral acceleration levels. We have performed the analytical evaluations; a summary of the results is provided below. Note that the evaluations consider simultaneous effects of horizontal and vertical components of earthquake motions.

INCREASE IN IDRIZONTAL SPECTRAL AhlPLITUDE Figure I shows the 5 percent damped free-field horizontal ground motions as postulated by the NRC Staff in SSER 34 and as developed in PG&E,1988. Both the median and the S4th-percentile nonexceedance probability spectra are shown. In both the SSER 34 and the PG&E,1988, only the 84th-percentile site-I specific ground motion was considered in the evaluation of plant seismic margins.

As shown in Figure 1, the PG&E,1988 spectrum completely envelops the SSER 34 spectrum at frequencies greater than I hertz. However, below I hertz, the SSER 34 spectrum shows an increase in spectral amplitudes of about 10 to 20 percent over the LTSP spectrum. As stated in PG&E,1991, no

! essential equipment or components have fundamental frequencies lower than I hertz. Therefore, an increase in the ground motion spectrum in the range of I hertz and below would have no impact on the seismic margins of essential equipment and co nponents.

Likewise, none of the essential building structures or structural elements are in the frequency range below I hertz. However, the sloshing modes of outdoor water storage tanks have low frequency responses (in the range of 0.2 ta 0.4 hertz), which caused PG&E to examine what effect, if any, could be expected by the increase in spectral amplitude. The refueling water storage tank, which is a typical example of outdoor water storage tanks, was analyzed to examine the effect on seismic margin of the S3ER 34 spectral amplitude increases.

Refueling Water Storage Tank The refueling water storage tank has a relatively high median capacity of 9.92 g* (Table 6-23 of PG&E,1988), governed by overturning moment at the concrete / bedrock interface. The base overturning moment of tank structures results from the combined effects from two modes; the impulsive mode, which I includes the tank inertia and the pa.t of the fluid that is moving in unison with the tank, and the sloshing mode. The impulsive mode typically has a much higher frequency (7.6 hertz) response and contributes significantly more to the overall demand than does the sloshing mode. A horizontal spectral amplitude increase in the range of 1 hertz and below has no effect on the dominant impulsive mode.

• Capacities of plant structures and equipment are expressed in terms of site-specific horizontal spectral acceleration averaged over 3 to 8.5 hertz frequency range ($, m m) g otabio canyon rower riant m Pacific Gas and Electric Company tong Term seismic Program ,

Confirmatory Analysis, December 1991 Page 2 Revised fragility estirnates for the refueling water storage tank were made by considering the increased effect due to the tank sloshing mode. These revised fragilities are compared with the original estimates (PG&E,1988) in Table 1. There is a small reduction in the High-Confidence-of-Low-Probabilitymf-Failure (HCLPF) value; however, the seismic margin remains high and the overall plant margin is therefere not affected.

Table 1 CO51PARISON OF FRAGILITY ESTIMATES Refueling Water Storage Tank Median S. HCLPFo 3 HCLPFu Scale Factori LTSP 1988 9.92 g 3.40 g 4.08 g 2.10 SSER 34 9.54 g 3.27 g 3.92 g 2.02 Note:

I ' Demand S,3,3,,, = 1.94 g l

i INCREASE IN VERTICAL SPECTRAL AMPLITUDE I

Figure 2 shows the 5 percent damped free-field vertical ground motion developed by PG&E in PG&E,1988. Both the median and the 84th-percentile nonexceedance probability spectra are shown.

The NRC Staff postulated a vertical ground motion that, for the 84th-percentile spectrum, exceeds the PG&E,1988 spectrum in the frequency range I to 10 hertz by 15 percent. The NRC Staff spectrum (SSER 34) is shown as a broken line in Figure 2 and extends between the frequencies of 1 and 10 hertz.

EOUIPMENT AND COMPONENTS There are two factors that determine which essential equipment and components are affected by the vertical spectral increases First, components having vertical natural frequencies in the 1 to 10 hertz range, and, second, components supported at locations where the structural floor slab vertical frequency falls within the 1 to 10 hertz range. Based on this component screening criteria, the following nine components were identified for assessment:

• NSSS Piping

• Main Steam PORV

• Diesel Generator Fuel Oil Day Tank l

• 4.16 Kv Switchgear

• 4.16 Kv Potential Transformer

• Safeguard Relay Panel A = 1mpulse Lines l

• Balance of Plant Piping and Supports

• Conduits, Cable Trays, and Supports Diablo canyon Power Plant

i. El Pacific Gas and Electric Cornpany Long term seismec Program

Confirmatory Analysis, December 1991 Page 3 A summary discussion of ana:ytical evaluations of the effect of the SSER 34 spectral amplitude increase on these components is given below:

NSSS Piping I The fragility evaluation of NSSS piping established that the median spectral acceleration capacity was "high", i.e., greater than 10 g (PG&E,1988). The critical seismic stresses in NSSS piping are due to g bending moment in the pipes associated with loadings from the attached equipment components (the steam j generator and the reactor coolant pump) resulting from seismic effects in the horizontal direction, The vertical natural frequencies of the steam generator and the reactor coolant pump are 23 hertz and 20.6 hertz, respectively, which are well beyond the frequency range where the spectral increases are identified. Therefore, a 15 percent increase of the PG&E,1988 vertical ground spectrum in the I to 10 hertz frequency range does not impact the fragility of the NSS'; piping.

Main Steam Power Operated Relief Valves Piping attached to the main steam power operated relief valves (PORVs) has vertical frequencies in the 10 to 20 hertz range. Thus, vertical seismic loads generated from attached piping and transmitted to the PORVs are unlikely to be affected by spectral acceleration increases in SSER 34. In addition, the main

. steam PORVs are mounted in a vertical position and are primarily sensitive to horizontal accelerations.

Therefore, an increase in the PG&E,1988 vertical ground spectrum in the I to 10 hertz frequency range does not affect the fragility of the main steam PORVs.

Diesel Generator Fuel Oil Day Tank The fragility evaluation of the diesel generator fuel oil day tank showed the median spectral acceleration capacity greater than 10 g (PG&E,1988). The critical failure mode is associated with bending of the tank flat bottom plate. The vertical frequency of the tank is 10.4 hertz and the critical failure mode is governed entirely by the vertical excitation. In the fragility evaluation, the seismic load was conservatively based on a vertical spectral acceleration at the peak frequency, further increased by a factor of 1.5 to allow for variations in the frequency of the tank due to the depth of the fuel oil and possible higher modes. While an increase in the vertical spectrum in the I to 10 hertz range does increase the seismic stress, the median spectral acceleration capacity of the tank still remains higher than 10 g. Therefore, we conclude that a 15 percent increase over the PG&E,1988 vertical ground spectrum in the I to 10 hertz range has an insignificant effect on the seismic capability of the diesel generator fuel oil day tank.

4,16 Kv Switchgear The vertical frequencies of 4.S Kv sw ' ;hgear are greater than 33 hertz for the structural failure mode and between 19 and 21 hertz foi de functional failure mode. The switchgears are, however, located at j turbine building elevation 119 feet, where the flexible floor slab has a fundamental vertical frequency in

! the 7 to 8 hertz range. Thus, an increase in the vertical ground spectrum in the 1 to 10 hertz range affects the switchgear through the local floor response. The structural failure mode of the switchgear corresponds to a bending failure of the switchgear carriage rod, which is controlled only by the horizontal response and which has a negligible contribution from the vertical excitation. Therefore, the structural failure mode is not affected by any increase in spectral accelerations of the vertical ground spectrum.

On the other hand, the furetional failure mode, associated with chatter of the General Electric IAC 53 protective relay, is attributable to the vertical excitation alone. The functional failure mode fragility Dubio Canyon Power Plant ldM Pacific Gas ana Electric Company Long Term Seismic Program

_ _ _ _ _ _ - . . =. -.. ...

l'onfirmatory Analysis, December 1991 Page 4 capacity factor was evaluated by comparing the Generic Equipment Ruggedness Spectra for the IAC 53 with the demand vertical spectral acceleration obtained at the location of the relay.

The functional failure mode fragility for the 4.16 Kv switchgear, as revised to include the effect due to the SSER 34 vertical ground spectrum, is compared below with the fragility value previously reported iti PG&E,1988.

LTSP 1988 SSER 34 Median 9, = 3.53 g Median S, = 3.06 g

[ #a = 0.35 #a = 0.35 l Bu = 0.25 Bu = 0.25 HCLPF, = 1.31 g HCLPF, = 1.14 g HCLPFu = 1.57 g HCLPFu = 1.37 g We note that the functional failure mode capacity is reduced as a result of an increase of 15 percent in I spectral accelerations of the PG&E,1988 vertical ground response spectrum. However, since the I functional failure of the 4.16 Ky switchgear is recoverable by operator action, the plant seismic margin remains unchanged in the discussion of the turbine building floor system at elevation i19 feet later in thL document, the effect of deflection of the floor on the functionality of the switchgear is described.

4.16 Ky Potential Transformers Similar to the 4.16 Kv switchgear, the fundamental vert: cal frequencies of the 4.16 Kv potential transformers are also very high. This equipment is also supported on the turbine building Door slab at l elevation i19 feet, which has a vertical frequency in the range of 7 to 8 hertz. As a result, the increase in the SSER 34 vertical ground spectrum over the PG&E,1988 spectrum at the frequency of the floor slab would increase the floor slab response, which in turn would affect the equipment fragility, in the fragility evaluation, two configurations were included: Dus F, which has three cabinets on a single stand

, and Buses G and H, which have two cabinets on a single stand. The vertical frequency of the three-cabinet stand is 33.7 hertz, while the vertical frequency of the two-cabinet stand is greater than 50 hertz.

For both configurations, the critical failure mode corresponded to failure of the fillet weld anchoring each support leg to the floor embedment plate, for which the contribution of the vertical earthquake component is relatively small. The revised fragility evaluation showed that the fragility capacity remained very high, i.e., above 10 g. Therefore, the increased vertical ground spectral accelerations do not affect the fragility value of the 4.16 Kv potential transformers reported in PG&E,1988.

Safeguard Relay Panel The safeguard relay panel is also supported on the turbine building elevation 119 feet floor slab, similar to the 4.16 Ky switchgear and the potential transformers. Thus, the increased vertical ground spectral f accelerations affect the panel through the local structure response, in the fragility evaluation reported I in PG&E,1988, the critical failure mode is associated with failure of the anchor welds.

! By incorporating the 15 percent increase in the vertical spectral acceleration, the seismic demand load is increased. However, since the vertical response of this component makes a small contribution to the critical stress, the net effect on the median spectral acceleration and HCLPF capacities is very small. A t

comparison of th trag:iity reported in PG&E,1988 witn the values obtained using the 15 percent increase in the vertical groui J spectral accelerations is shown below:

g lat Pacific Gas and Electric Company Diablo Canyon Power Plant Long Term Seisnde Program l

1 Confirmatory Analysis, December 1991 Page5 LTSP 1988 SSER 34 x

Median S. = 10.76 g Median S. = 10.70 g

1. a = 0.34 #a = 0.34 Bu = 0.36 #u = 0.36 HCLPFw = 3.39 g HCLPF, = 3.37 g HCLPFu = 4.07 g HCLPFu = 4.04 g I Impulse Lines l The potential failure of impulse lines is related to their excessive deflections as a result of interactions F with equipment and components to which the lines are connected, Since the lines typically have small diameters, failure is not associated with inertial loads on the lines. The components in the vicinity of the impulse lines, which were included in the fragility evaluation, have median spectral acceleration capacities in the range of 7 g to 9 g. In addition, these components have vertical frequencies that are greater than 10 hertz and failure modes that have negligible contributions for the vertical excitation. Therefore, we

) conclude that since the failure of impulse lines is not controlled by their inertial loads and the nearby I components are not sensitive to the vertical excitation, the fragility of impulse lines is not affected by vertical spectral accelerations exceeding the LTSP 84th percentile spectrum in the 1 to 10 hertz range.

9 f Balance of Plant Piping und Supports As discussed in PG&E,1988, a generic fragility was developed for the balance of plant piping based on pipe support design specifications for allowable stresses corresponding to various loading conditions and failure modes. The critical failure mode le associated with failure of fillet welds in pipe supports. As discussed in PG&E,1988, a frequency range of 6 to 20 hertz was judged to envelope the probable piping f frequencies. In addition, in the fragility evaluation, it was found that the piping systems with a frequency of 8 hertz show the least relative factor of safety. Therefore, the fragility for the balance of plant piping is affected by the increase in the vertical ground spectral accelerations ir ~ l to 10 hertz range.

By incorporating the increased vertical spectral accelerations, the median spectral acceleration capacity and the HCLPF value are reduced approximately 5 to 6 percent. The results reported in PG&E,1988 L are compared with evaluations for SSER 34 below.

LTSP 1988 SSER 34 Median S. = 11.03 g Median S. = 10.34 g

1. a = 0.40 #, = 0.40 Bu = 0.39 du = 0.38 HCLPF, = 3.00 g HCLPFw = 2.85 g HCLPFu = 3.60 g HCLPFu = 3.42 g l

Cable Trays, Conduit, and Supports i

As discussed in PG&E,1988, flexible cable tray systems will not fail. For cable trays, conduit, and supports, a generic fragility was developed based on the seismic qualification analyses for typical cable i tray supports located at various elevations in the auxiliary building and the containment building, shown in Table D-1 of Kipp,1989, the contribution of the vertical excitation gggr,igsgejs 14 w Pacific Gas and Electric Company Long Term Seismic Program

Confirmatory Analysis, December 1991 Page 6 ranged up to 95 percett and the vertical frequencies ranged from 6 to 35 Swa However, for each of the typical supports, the vertical frequency ranged from 18.9 bertz to gaater than 33 hertz. The low frequencies reported in PG&E,1988 correspond to vertical nodes of the cable trays themselves rather than the suppons; the important frequencies for the cable tray supports fall outside of the I to 10 hertz

[ range. While some vertical frequencies of the cable trays may fall into the I to 10 hertz range, the trays I themselves are expected to perform well ss demonstrated by previous tests en cable tray systems, which stayed intact and remained functional well above the design amplitude of motion (Linderman,1981).

Therefore, vertical ground spectral accelerations exceeding the LTSP 1988 spectrum in the I to 10 hertz range will have minor effect on the fragility values in PG&E,1988 and Kipp,1989, where it was reported that cable trays and supports have high capacity, i.e., greater than 10 g.

STRUCTURES AND STRUCTURAL ELEMENLS The plant seismic margin assessment based on probabilistic methoc described in PG&E,1988, and later bench-marked by deterministic assessments in PG&E,1990, established that shear walls are the governing elements that control sei:mic capacities of major civil structures. These walls are primarily horizontal earthquake resisting elements; also, the shear walls are rigid in the vertical direction having frequencies much greater than 10 hertz. We, therefore, conclude that there would be no effect of an increase in the LTSP 1988 vertical ground spectrum in the 1 to 10 hertz frequency range on the seismic capacity of major structures.

Vertical earthquake motions, however, may have some effect on parts of structures, such as flexible Hoors and Door beams. These vertical load carrying elements are usually controlled by ductile bending behavior, which is accompanied by a large inelastic energy absorption capability. Thus, these types of structural elements are highly unlikely to fail; however, it is necessary to examine supported equipment its . tant to plant safety to see the effects of the attendant motions. Based upon these considerations, we have identined two vertically Rexible slab systems for analysis to examine the effeci of the SSER 34 spectrum. These are the control room roof slab located in the auxiliary building and the 4.16 Kv switchgear area Door slab located at elevation i19 feet in the turbine building.

Auxiliary Building Control Room Roof Slah The auxiliary building has 11 vertically nexible slabs; however, the control room roof slab located at elevation 163 feet is the only slab that has a fundamental frequency in the vertical direction in the range of I to 10 hertz. Figuru 3 :.nd 4 show respectively the plan and section of the control room roof slab.

As shown in Figure 3, the 3 feet-4 inch thick reinforced concrete slab is supported by embedded structural steel beatra with a span of 57 fat and with end moment restraints provided by reinforced concrete shear walls. Attachments to the slab consist of light-weight lighting fixtures and ceiling tiles.

Dere components are suspended from the underside of the slab via a grid of unistrut steel channels w-Ided to insert plates embedded in the slab. The insert plates are anchored to the slab by steel Cat bars that are welded to the bottom Range of the embedded steel beams. A few HVAC ducts are also attached to the slab by concrete expansion anchors.

The seismic margin evaluation of the slab was performed by using the conservative deterministic failure margin (CDFM) approach discussed in attachn mt DE-Q6A of PG&E,1990. As described in PG&E,1990, the seismic margin factor, which represents the amount by which the deterministic spectrum (in this case SSER 34) can be scaled to produce a demand equal to the HCLPF capacity of a structure or component, is given by the elastic scale factor times a factor F,. The elastic scale factor is the factor by which the deterministic spectrum can be scaled to produce a demand equal to the yield capacity of the structure or component, and Fy is the CDFM inelastic energy absorption factor.

Diablo canyon Power Plant 14 la Pacific Gas and Electric Compny long Term setsmic Program

Confirmatory Analysis, December 1991 Page 7 Progressive softening of the slab due to concrete cracking and development of plastic hinges at the end restraints of the slab together with formation of a plastic hinge near the midspan results in a fundamental frequency of the slab of about 5 hertz. The elastic nale factor was computed from the slab response just before a mechanism is formed. Capacities wea based upon composite construction provisions of the Load and Resistance Factor Design Specification of the AISC.

The inelastic energy absortM factor was determined as a function of target displacement beyond yield

[ displacement (i.e., ductility, p) by using the Riddell-Newmark approach. For the control room roof slab, l the ductility of the system is directly related to the ultimate displacement of the slab near midspar The HCLPF value of the ultimate displacement of the slab is based upon structural capability consideratwns.

The inelastic rotation capacity limit specified in the ACI 349 code is consioered to be a HCLPF value.

For floors s.ibjected to a significant gravity load, however, it is necessary to account for ratcheting behavior of the slab or beam when estimating the system ductility. Ratcheting refers to the progressive downward displacement that occurs following multiple seismic load reversals in the inelastic range and has the effect of reducing the available ductility of the system. To account for this behavior, a reduced effective ductility value was used in the inelastic energy absorption calculation. The process is described

' Appendix A.

A sensitivity study was performed to investigate the variation of the seist. tic margin with displacement.

, The results are shown in Figure 5. We note that, at a seismic margin factor of 1.64 that corresponds to the auxiliary building seismic margin reported in PG&E,1988, the displacement of the slab is only about 4 inches near the center of the 57-feet span. This displacement is less than about 0.6 percent of the slab span (i.e., span over 120) and corresponds to a member ductility of about 3. This can be compared to a ductility limit of about 9 determined from the rotation capacity spccified in the ACI 349 code. At such small displacements, no significant degradation of the capability of anchored components is expected.

We, therefore, conclude that the control room roof slah has an acceptable seismic margin over the SSER 34 vertical ground motion spectrum.

Turbine Building Floor System at Elevation 119 feet -

The turbine building houses portions of three major safety-related systems: the emergency diesel

[

gen:.cators, 4.16 kV switchgear, and the component cooling water heat exchangers. The emergency diesel generators and the component cooling heat exchangers are supported on the basemat The diesel generator silencers at elevation 107 feet and the 4.16 kV switchgear at elevation 119 feet are located on floor systems that consist of reinforced concrete slabs supported by structural steel beams. Both floors have similar frequency responses in the vertical direction, with the fundamental mode in the I to 10 Hz range, and have seismic demands that are dominated by vertical ground motion. Of these two slabs, the 4.16 kV switchgear tloor is located at a higher elevation, and hence has higher amplification of response.

The switchgear floor also supports equipment that are more displacement sensitive, mo- massive, and is, therefore, subjected to greater seismic demands. Thus, the switchgear floor at elevation 119 feet was selected for evaluation.

The switchgear floor consists of a 10-inch thick reinforced concrete slab supported by compact wide flange structural steel beams and columns (Figures 6 and 7). Welded studs are used to anchor the slab to the steel beants. Thus, the concrete slab provides continuous lateral support to the compression flanges of the supporting beams. The switchgear panels are arranged in three rows that are separated by masonry walls, la addition to the building columns shown in Figure 7, the columns supporting masonry walls along column lines 3 and 4, from E.2 to G, and connected to the slabs at Elevations 119 feet and 140 feet by through bolts, also provide mechanisms for load transfer between the two floors. Note that the Diablo canyon Power Plant W Pacific Gas and Electric Company tong Term seismic Program

1 , .

Confirmatory Analysis, December 1991 Page 8 connections of columns along column lines 3 and 4 to the slabs at the top and the bottom, were made l during the Hosgri reevaluation in order to connect the two floors, and thus, to reduce the response of the I slab at elevation 119 feet.

l The seismic margin evaluation was performed by using the (CDFM) approach. It was determined that l the beams along column lines 2 and 3, from E.2 to G, that support masonry partition walls on each side of Bus "G" panels, were the controlling structural elements, whose failure was governed by bending

[ (ductile behavior). Since the steel eclumns supporting the masonry wall along column line 2 are not l connected rigidly to the floors above and below, the beam along column line 2 does not receive any support from the beam above. However, the floor beam along column line 3 acts in unison with the floor beam at elevation 140 feet; thus, the elastic scale factor for this combined system was calculated by considering the total seismic demand and their combined capacities.

The evaluation of the structural steel beams, and consequently the switchgear floor system, consisted of performing a sensitivity study wherein seismic margins (i.e., the scale factors by which the input vertical ground motion would have to be raised) corresponding to specific target dis,lacements (or, pre-assigned ductilities) were calculated. The calculations accounted for the ratcheting effect due to dead load in accordance with the procedure described in Appendix A.

f Figure 8 shows the results of the sensitivity study where the seismic margin factor is plotted against the ultimate target displacement of the floor. We observe tha:, for a seisu.ic margin factor of 1.45, which is approximately the seismic margin of the turbine building reported in FG&E,1990, the maximum displacement of the switchgear floor is about 1.7 inches. This displacement is equivalent to a member ductility of about 2 which can be compared to a value of 10, corresponding to a structural capability limit given by the AISC commentary to the Load and Resistance Factor Design specification. We have evaluated the switchgear cabinets for such a postulated displacement of the supporting floor of 1.7 inches.

The evaluation is described below.

IMPACT OF TimBINE BMlWING FtDOR IWHICDON ON 4.16 KV SWITCllGEAR t

{ The 4,16 kV switel. ar is an assembly of 4.16 kV magneblast circuit breakers, buses, insulators, wiring, control, and protecting and monitoring devices that are contained in a metal-clad enclosure. Power is

[ made available to all load circuits through three electrically contiguous and high conductivity buses

! running through the entire length of the switchgear. At each cubicle a vertical tap is made for connection to the circuit breaker. The breaker is held in place by a cradle that prevents it from falling under severe conditions of vertical and horizontal seismic conditions.

j The maximum displacement of 1.7 inches is reduced by the dead load displacement of 0.3 inch to obtain the differential displacement of 1.4 inches affecting the supported equipment. The 1.4 inches maximum sag of the switchgear floor, applicable to bus G, is equivalent to 0.2 inch dip of one edge of the 26-inch wide switchgear cubicle. A review and walkdown of the control and protective devices located in the switchgear cubicle, including relays, switches, fuses, terminal blocks and wiring, indicates that this deflection has no significant functional impact on these devices. Certam switchgear mounted relays could potentially trip the breaker due to chattering. Following the trip the breaker will stay open. Control room annunciation and indication of the breaker status will remain opewional. After the event,

, realignment of the breakers can be made through operator control. .Also, the copper buses have adequate 2 flexibility so that small deflections will not affect the integrity of the bus and conductivity will be 2 maintained.

m

Diablo canyon Power Plant

t. r , Pacific Gas and Electric Company Long Term Seismic Program

. . . - - . - ~ . -

Confirmatory Analysis, December 1991 Page 9

( Re structural integrity of the 4.16 kV switchgear was evaluated for the combined horizontal and ver al seismic loads as well as the effect due to turbine ouilding floor displacement, he results of the evalua' ion indicate that the switchgeat wih maintain its stmetural integrity with acceptable margin for loadings due to SSER 34 seismic spectra.

Nate that each row of switchgcar cabinets (see Figures 6 and 7) consist of several individual cabinets cor.nected to one another in series by a line of bolts along their heights. These cabinets are thus structurally affected by the deflection of the supporting floor. In contrast, the other two equipment items, considered essential in LTSP risk assessment and margin studic,, namely, the 4.16 kV potential transformer and the safeguard relay panel are individa units, and consequently, the effect of Door displacement on these equipmear items is irsignificant.

Fuel llandling Ilullding Crane

{

in PG&E 1990, we raported results of the deterministic evaluation of the fuel h,. ng building crane.

As noted in that submittal, the crane was chosen for eva'uation because its roponse is primarily governed by the vertical component of ground motion. The failure modes identitied in PG&E 1990 lie within the frequency range from 1 to 10 hertz, and therefore, these failure modes are affected by the SSER 34 vertical spectrum.

[ We have reevaluated the fuel handling building crane for the SSER 34 <pectra. The results are reported I in Tabs 3, which shows that the fuel handling building crane continues to have an acceptable seismic margin over the SSER 34 vertical ground motion spectrum. Please note that, as stated in PG&E,1990, the margins reported for the lifted case are very conservative because of the analysis methodology auumptions and low probability of a .25-ton litt concurrent with a signincant seismic event.

Structural Stw! Frames and Truss Systems Pipeway 5tructure i

The last class of structures that may potentially be affected by the vertical ground motion in the frequency range of 1 to 10 Hz are structural steel frame and truss systems. Effects of vertical seismic excitation on these types of structures are generally small compared to horizontal ground motion effects.

Furthermore, steel framing systems a e generally highly redundant and consist of elements that are usually controlled by ductile bendi@ ehavior, Ahich is accompanied by large inelastic energy absorption capabilit y.

We have selected the pipeway structu'e among the class of structures composed of structural steel framing u and truss system for evaluation of increased vertical ground motion. In contrast to the structural steel roof truss system o"er the tuihine building, certain safety-related systems are directly attached to the

)

pipeway structure. For example, the pipeway structure provides support to the auxiliary feedwater pipe line as well ac smaller piping and safety relateJ instrumentation tt.bing and conduits. The pipeway structure also supports the main steam and feedwater pipe lines between the containment and the auxiliary building, in addition, the pipeway holds pipe w',ip restraints d-signed to restrain pipes from whipping after a postulated pipe break.

The pipeway structure is a threedimensional stiuctural steel frame attached to the outside d %e containment shell, the auxiliary building and the turbine building. Figure 9 shows locations of .' mti and Unit 2 pipeway structures, which are basically similar to one another. The structure has five wvels of plane frames with tne main platform a.t elevation 109 feet (shown in Figure 10), and includes ten radial r bents that cantilever out fiom the containment wall tj horizontal beams and/or inclined beam columrts.

Connections between the pipeway structure and the auxiliary and the turbine buildings are provided with DiaMo caw Powet Plant 14 6 4 PaClfic Ca$ and ElectrlC Compn, tong Term seismic Program l l

1

i' 1

' Confirmatory Analysis, December 1991 _

Page 10 slotted holes oriented such that horizontal forces cannot be transtmttu between the pipeway structure and these adjacent buit ' ,.

Is a manner ' i - t ..at applied to the control room roof slab and the 4.16 LV switchgear floor, the pineway strwt. semale margin was evaluated by using the CDFM approach. Critical structural e%nents were identified for margin assessmerit on the basis of seismic capacity-to-demand ratios

,termined during the llosgri reevaluation, as well as the importance of the metr.bers to the overall structural system. Capacities were determined by usmg the Load and Resistance Factor Design (LRFD) specification of the AISC. In some cases, CDFM capacities were taken to correspond to 1.7 times the AISC allowable stress, which is approximately at the first yield of a ductile member, s

As noted previously, for ductile structural systems beyond the elastic yieh limit, the selimic margin

[ factor can be assessed by limiting the ultimate target displacement from consideration of functionality of

( suppomd systems and components. For the pipe' ay structun., however, we calculated the displacement corresponding to a preassigned seismic margin f actor. Thus, to reach a seismic margin factor of 1.76.

which is the seismic margin for balance of plant piping, beams along 11ents 211 and 6.611 (see Figure 10) that become inelastic, dbplace vertically by up to about one inch. Long-span, tiexible piping systems that a.e typically attached to the pipeway st ueture are expected to accommodate displacements of such small magnitude.

CONCLUSIONS

[ Sased upon the confirmatory analysis described above., we conclude that the essential structures,

{ equipment and components important to plant safety have adequate seismic margins to accommodate the i NRC S:af t's estimate of the horizontal and vertical ground motions shown in the SSER 34. We h;va also g

shown that the SSER 34 speetra have minor effects on tne seismic fragility estimates of the s'ructures and components. We can thus conclude that the assumed increase in amplitude of the Diablo Canyon LTSP 84th-percentile site 4pecific ground motions in the horizontal and vertical directions, sho% in SSER 34, will not alter the conclusions of the probabilistic risk assessment (PG&E,1988).

I m

h Diablo canyon Power Plant i.....

a u Pacific Gas and Electric Company Long Term schmic Program

i

) ,

Confirmatory Analysis, Dec rc.ver 1991 Page 11 REFERE,NCES U. S. Nucley Regulatory Commission, June 1991, Safety Evaluation Repon, related to the operation of Diablo Canyon Nuclear Power Plant, Units I and 2, Docket Nos. 50 275 and 50-323, NUREG-0675, Supptement V . 34.

Pacific Gas and Electric Company,1991, Adequacy of seismic margins assuming an increase in arnplitude of the Diablo Canyon Long Term Seismic Program horizontal ground motion in the frequency range below 2.5 hertz: PG&E Letter No. DCL 91 108.

Pacific Gas and Electric Company,1991a, Adequacy of seismic margins assuming an increase in amplitude of the Diablo Canyon Long Term Seismic Program vertical grourd motion in the frequency range from 2 to 10 hertz: PG&E Letter No. DCl 91-131.

.acific Gas and Electric Company,1988, Pinal Report of the Diablo Canyon Long Term Seismic Program: Docket Nos. 50 275 and 50-323.

Kipp, T. R., Wesley, D. A., and Nakaki, D. K.,1989, Seismic fragilities of civil structures and equipment components at the Diablo Canyon Power Plant, NTS Engineering Report No. 1643.02, Revision 0, Prepared L.. ?G&E.

Pacific Gas and Electric Company,1990 Additionaldeterministic evaluations performed to assess seismic margins of the Diablo Canyon Power Plant, Units 1 and 2: PG&E Letter No. DCL-90 226.

Linderman, R.B. and Hadjian, A.H.,1981, Development of Bechtel's electilcal raceway system test program, Proceedings of the American Power Conference, Vol. 43.

J Diablo canyon Powet Plant I. sa Pacthe Cas and Dectric Company tong Term seismic Program

1 Conf rmatory Analysis. Deceniber 1991 Page 12 Table 2

, EFFECT OF INCitEASE OF SPECTilAE ACCEL.EllATIONS SIIOWN IN SSEll 34 l FitAGil.lTIES OF AITECTED EQUIPhlENT AND C051PONENTS (Ita ed on liarard Defined Over 3 to 8.5 liertz llange)

SPECTRAL ACCI:LI: RATION CAPAC11T L10P IMN SSI:R 34 LQUlthtLNT/ VI:RTICAL COMPONENT l~RI:OUENCY S, (g) n, no IICLPr S, (r,) n, no licLPF IN'I (A) (g)

NS$$ hping 7-9 > 10 00 > 10.00 ISping Main Stesin IORV 10- 0 11.50 0.34 0.38 3.51 Unchanged Diesel Generator Fuct oil 10.4 > 10 00 > 10 00 l Day Tank 4.16 LV $witchgear* > 33 3.53 0.35 0.25 1,31 3.06 0.35 0.25 1.14 4.16 LV Potential > 33 10.83 0.31 0.38 3 47 > 10.00 T:ansformer*

Safeguard ttelay Pancl* > 33 10.76 0.34 0.36 3.39 10.70 0 34 0.36 3.37 Impulac Lines 5 20 7.09 0.28 0 32 2.63 Unchanged Balance of Plant ISping and 420 11.03 0.40 0.39 3.00 10.34 0.40 0.38 2 85 Supsurts Conduits. Cable Trays, and 4 35 > 10 00 > 10 00 Supports

• Equipenent sup;mrted on vertically flexible floor slab having vertical frequency less than 10 hertL

. Diablo Canyon Power Plant c6 Pacific Cas and Ilectric Company tong term seismic Program

l Confirmatory Analysis, December 1991 Page 13 Table 3 r Fuel llandling 11ullding Crane l Seismic hlargin Factors l

Crane Element Ufted lead Seismic htargin Factors or (tom)

Failure htode LTSP 1988 SSER 34 Cable 125 2.90 2.52 Ilridge Girder 125 1.55 1.35 Bridge Resistance 0 3.2 2.7 to Global Uplift Trolley Resistance 0 2.4 2.0 to Uplift i

i Diablo Canyon Power Plant if b PacillC Cas and Dectric Company Long Term $eismic Program

) .

, ' Confirmatory Analysis, December 1991 Page 14 l

1 3 i i , , i i s ii i i i i TTi i s i i i i i ii i

~ ~

= LTSP,84th Percentile 5% Damping

~

LTSP, Median -

- 0 SSER 34,84th Percentile -

~

+ SSER 34, Median _

$2 c

E - 0 -

3 - e s -

o -

g -

N o o 3u -

U - + + _

a ) ~ o .

O _ e

~

~

c + -

+ -

,. e + ,,,

' ' ' I ' ' ' I ' ' ' 'I 0

0.1 1 10 100 Frequency (Hz)

Figure 1 Site-speciife horizontal ground motion response spectra.

Diablo Canyon vower Plant

.& Pacific Gas and Electric Company Long term seismic Program

, Confirmatory Analysis, December 1991 l , l' age 15 i

3 , , ,i,,,,, , , , , , , , , , , , , , , ,,

~

_ = a LTSP V/H Ratio,84th Percentile 5% Damping

[ LYSP V/H Ratio, Medlan

~

LTSP Regression,84th Percentile -

+ LTSP Regression, Median ,

B2

- - ""* SSER 34 ,

~

~

' * ~

E - ,./ o n 2 -

2 -

je .

~

~

N ,Y o n m

- . ~

2 ~

/ o ~

U / e E1

~

" ~

-  : w va  ;

. .+ ++4 _

+ _

_ + ,

' +

c /. .

+ + d

- c _

0 i i i i n i t il i i i i i r i il i i i i i ii?

0.1 1 10 100 Frequency (Hz)

Figure 2 Site-specific vertical ground motion response spectra.

g.- Diablo Canyon Power Plant is'%1 Pacific Gat and Electric Company Long Term 5etsmic Prograni

'Confinnatory Analysis, December 1991 Page 16 I

17" 6" 52' 6" g ~ !'

3 I I dk s@ Key Plan 9\C

'o N

Critical strip L-

[ of slab

= @

, 7 -A 7==

= ll

• A o N

g

+A

==: @

'o 4,

i 4  :

~5T Figure 3 Auxiliani building control room roof slab.

g Diablo Lanyon Power Plant 1.O Pacific Gas and Electric Company Long Term Seismic Program

. .Confirmator'y Analysis. December 1991 Page 17 l

l s'. n" .

lh LW 3G x 170 Beams l

Welded Anchor

, El.163' 4" 3 3/4" 0 x 7"long

..... i; .,..n

. .. a

... .. A o

> M.

o-py T.O.S. El.167' 0"3 iru'; -

.. .- i .. -

9

.e

.6.,. Q** 0, e 0 ., %, b

., p .

o:.e.

• 6 ', ,. , . .o - * ,

. y , . *o '..o. . e 6, .. ,,

i ,n - '

i--

n V .

!~=p2 d ir (

\/ u insett plate I attached to W36 / - El.158 - 8..

Section A A Ilgure 4 Auxiliary building control room roof slab. Section A - A

. 04ahlo Canyon Power Plant

.6 Pacific Gas and Electrk Cc:npany tong lerm schmic reogram

' Confirmatory Analysis, December 1991 Page 18 1

l 3.0 -

Existing scismle 2.0 -

[ margin factorfor y Auxillary Building u v 3 1.04 - - - - - - - - - - - - -

5 1 e

(O E ,

0 g E 1.0 --

n l 3 I I

I i

1 1

I l

i a ._ _

l i l I l 0 2 4 6 8 10 Ultimate displacement near midspan (in.)

Figure 5 Ultimate displacement near midspan of control room roof slab vs. seismic margin factor.

g r&i Pacific Gas and [lectric Company 04ablo Canyon Pour Plant Long Term Seismit Program

Confirmatory Analysh, December 1991 Page 19 1

1 1.9 3.5 4.8 5.7 6.6 8 9 14 15 16 17 18 k9

, (10.6 (2.2 EXP,

. n, . JT. g" G'-

i. t A ,~ ,,, 1 .x .

- 1

.-I[-

id'IS . Mr '"W D.'s t&..,m;:$q::'ynje,'e.ep; y,r t '.sg. g.#35 a .m('fSg sj'g' 1 pq e '3:,

f,.1,;* ';R' 5- .v..s. -

, .m,c s.: ,  ;~.-- m, : %y u. u ..~. u. ..m. w. .x ;,a s . .

c . - .

~, ,.. c,. .s . .

. s e:- 1. n: . I

%- - pm . . 4, c q. ,. , ,. ,. -, i

. . i-,l. 'g4,; i ' ![ n%pM;'%Mik:g'RUN'J'if 2

cNv w. , w ;a

?::/O'i3d$$II!B;ps6!El TD .F.. I', / .

A W. r c m. ~;m7-.

e,s a J. %-

c:s: ug..v, .. i t.

s s

-s , , s v .

3. -

,, ,. . . . .. . ,, y .. . c;]'Qg.,

Q :.\: 0... '; t;x,s e. .

99, -

. . ,g;.ggg ws e. ' - - -

,* . 3 .G....*g. .

I'..,' 4. <

EF *- i; a

. AI,to ..tl.

,! & - Turblne Pedestal , . , '. .,

v Al .

4'p./& 'W.:'. Not Shown ',i eS/. I 9(a*c.

',l:

s. y h :]y K .m. -

'1 h ><' 'Y iki; y?iVllm.. :e= y --e w WF,M A ,' .' ,0

n. Xi.%,.r.1i:-. Vu,,,weae w. +, s . .

.2 ~s:..e kh wm., n.,

- J i

[ '. ,. M j. M l. .. 3  %%)~:s.3/h j%'/yd4 I 7:%-[ls13fI$$s

f. . .'.'. s. .  :.y*pl.;t

.w Q,sd  ;?ng,i~ n

,g '? t@a.:. b; .-'w';J, l

m. . .Td M~r [..', 4 t- -a ;w- y'Jf b y A W ---- s ; it iap y .;

H H N H H H H H H

.Mm ) .I I 1 I I I .1 .1 I

] Concrete

] Grating

Steel Checkered Plate Figure 6 Unit I turbine biuilding plan at elevation 119 feet.

Diablo Canyon Power Plant ab Pacific Gas and Dectric Cornpany Long Inm Seismic Program

'Canfirmatory Analyns, December 1991 Pace 20 s

I

,18' ti 3/4" 16' 6" 16* 6" _

\

Controlling Beams 7

-[ E 1 O t

s s m c

t -

t F' v

-l - - -  : - -

a .,

. 4 s c'

h 4 9

scN

[

'n 2$k

@b ~

h x

.N

-1_

'*( -1_

s q

dx r

h' s

's '

[' p..

D, -

D' h h -

i I I-O = Masonry wallcolumn linking e!evation 119 ft.

and 140 ft, floors.

4.16 kV Switchgear Figure 7 Turbine building 4kV switchgear floor partial plan.

Diablo Canyon Power Piant 46 Pacif1C Gas and Dectric Company Long Term 5eismic Program

Confirmatory Analysis, December 1991 Page 21 s

l

}

2.5 -

Beam at @ >

Center line of switch gear location fa 2.0 -

i E

j f'

• Beam at @

-.m

@ 1.5 -

i i i i 0 1.0 2.0 3.0 4.3 Ultimate displacement (in.)

Figure 8 Ultimate displacement of floor beams vs. seismic margin factor.

Diablo Canyon Power Plant 4 a Pacific Cas and Electric Company long Term seismic Pro 0 ram

Confirmatory Analysis, Dece:nber 1991 Page 22 s

504 ft 8 :

duxillary Dullding 5 l _ __ ]

w Unit 1 Unit 2 Unit 2 Pipeway Unit 1 Containment Containment  ;,

Pipeway Structure Structure dilj)

.Ma; . l?/

118ft 89 ft E Turbine Building n

, 748 ft l

l Figure 9 Power block stnictures with pipeway structures of units 1 and 2 highlighted.

Diablo Canyon Power Plant .

1 4 El PatillC Gas and EltCtrlC Company Long Term sehmic Program

• ~ Continnatory Analysis, Dec:mber 1991 _ Page 23 4s

\ -g

\

f\ \

h- 'h eN y'N ~'

x f

/

N 's gN .

y .

79 '

N

/% , g., (6k

}t -

6'S 6 g af .. @

s Figure 10 Unit pipeway structure, platform at c'evation 109 feet.

Pacific Cas and Electric Con pany ngleen Se sn Pro a

i

- \

(

(

(

Al'I'ENDIX A INYFSTIGATION OF TIIE EFFECTS OF DEAD LOAD ON DUCTILITY OF FIDOlt SYSTEM December 1991

-) Diablo Canyon Power Plant I.:6i PacMc Gas and Electric Company Long 1erm seme Program

, ' Appendix A, Deceil>er 1991 Page A 1 r

L APPENDIN A INVESTIGATION OF Tile EITirl'S Ol' DEAD LOAD ON DUCTILITY OF FLOOR SYSTEM s

introduction This report presents results of a study to investigate the effects of dead load on the inelastic energy ,

absorption factor of floor systems. Previous testing and studies by others have demonstrated that both reinforced concrete and steel structures have significant capability to absorb carthquake energy inelastically at ground motion levels above the static strength. This increased capacity can be quantified by the ductility scale factor, F y, which scales the earthquake time history from the point at which static strengtn is reached to the target Icwl ofductility.

The ductility scale factor was studied in Kennedy,1984 for wrtically mounted shear walls subjeued to I horizontal motions. Nonlinear analyses were performed and simplified procedures were developed and investigated in that report. Otner simplified procedures for obtaining Fy for wnically-mounted elements can be found in Riddell,1979 and Reed,1991.

Because of the constant downward pull on a horizontally oriented structural element due to gravity, the

( nonlinear vertical displacement of a floor structure, when subjected to vertical motion, is biased downward. This ratcheting phenomenon reduces the ductility scale factor for a horizontal element compared to the case where the same element is mounted vertically and is subjected to horizontal motion.

The objective of this study was to develop a simple modification to the procedure for calculating the ductility scale factor for wrtically-mounted components to calculate the F, factor for .'wrt:ontal cIcments.

This was achieved by first perfoaming nonlinear time history analyses for hort:ontally- and vertically-mounted elements and then using the results to obtain a simplified procedure for including the effects of dead load.

Model and Procedure The response of a concrete floor structure can be idealized as a one-degree-of freedom (1 DOF) system since its response to seismic motion is predominantly due to the response of its fundamental mode.

Figure A 1 shows the 1 DOF model mounted both vertically and horizontally. Also shown in Figure A 1 is the force-deformation relationship assumed for the model. This relationship is used for both orientations of the model.

The capacity of typical floor slabs in nuclear power plan's, comt ased of both reinforced concrete and structural steel members, is due primarily to ductile flexural behavior. Since the slabs are under-reinforced, the capt:ity is controlled by the strength of reinforcing steel, not due to crushing of concrete.

This leads to full hysteretic loops where there is very little, if any, cyclic strength or stiffness degradation. Thus, the force-deformation curve is assumed to be bilinear with a small amount of strain-hardening to reflect the characteristics of both structural steel members and reinforcing steel, as shown in Figure A-1.

The time history used in the analysis is the modified Pacoima Dam vertical record used in the Diablo g Diablo Canyon Power Plant IP6i Pacific Gas and Dectric Company Long Term seumk Program

. +,

Appendix A, Decemoc ~. J 1 Page A 2 l Canyon Long Term Seismic Program (PG&E,1988). Figures A-2 and A 3 show plots of the time history and the 5 percent damped response spectrum for this time history, respectively. The absolute level of the input time history is not significant since F gis equal to the earthquake scale factor required to reach the capacity corresponding to the desimi ductihty level divided by the scale factor to reach yield.

( Rus, the absolute level of the input time history cancels out.

Tv" fundamental frequencies for the models, one at 4 hertz and the other at 7 hertz, were analyzed to reach three levels of ductility, p. The values assigned to p were 2,5 and 10, p being the ratio of the displacement at the desired ductility level to the displacement at yield (see Figure A 1). The damping was assumed to be 7 percent of critical, although additional sensitidty studies were performed to show that the ductility scale factor was only weakly influenced by the clastic damping value. Tangent stif fness-proportional damping was assumed, which implied that, beyond yield, the damping was almost equal to re,o (i.e.,0.03 x 7 percent equals 0.2 percent). Finally, for all cases where the model wu mounted horizontally the dead load stress wee fixed at 30 percent of the yield level. This level is representative of the dead load stress level for typic. floor slabs at the Diablo Canyon Flant.

Procedure and Results of Nonlinear Analpi.;

Nonlinear analyses were performed using the computer program DRAIN 2D (Kannan,197.1, revised 1985). The bilinear truss model was used to simulate the inelastic behavior of the model, and a time step of 0.001 tesond was used for all computer runs.

A target displacement corresponding to an assigned value of ductility p, was chosen and an iterative procedure was followed to determine the level of earthquake time history 'equired to create the target displacement. Cenain features in DRAIN 2D ensured convergence of the iterative process in about 5 to 10 iterations.

Table A-1 shows results of the nonlinear analyt.es. Scale factors are shown for 4 hertz and 7 hertz models, and for both the cases of no dead load and dead load equal to 30 percent of yield. The " Yield FS" column is the factor used to scale the earthquake time history to reach yield level (i.e., the point defined by Py and A, in Figure A le). Similarly, the " Ultimate FS" column is the factor required to scale the time history to reach the target displacement level (i.e., the displacement equal to pA, in Figure A-Ic). Finally, the ductility scale factor, F,, is the ratio of the corresponding values in these two columns.

Note that the yield scale factor for the " dead load equal to 30 percent Py" cases do net exactly equal to 70 percent of the yield scale factors for the corresponding "no load" cases. This is because peak response for the dead lead is sensitive to the sgu of the input motien (i.e., plus or minus). For the "no load" case, the peak of the two direction responses is always used, while for the " dead load equal to 30 percent Py case the doivnward component was used. it turned out that the signs were opposite for the cases considered. The effect of reversing the input direction on the ductility scale factor is discussed below.

Table A 1 shows that the ductility scale factor is reduced for the case where dead load is present.

Figures A-4 and A-5 show example response time histories for the "no dead load" case and the " dead load equal to 30 percent Py " case, respectively. The effect of the dead load can be clearly seen by comparing these two figures.

The effect of rev:rsing the earthquake time history sign was also studied. Table A-2 gives results of changing the earthquake direction and compares the results to the original case (i.e., from Table A-1).

Although the scale factors to reach yield and ultimate are both less for the reversed sign case, which Detdo canyon Power Plant i, u Pacific Cas and Hect.lc Company Long lew se sm Program

' Aypendit A, December 1991 Pape A 3 produced the greatest response elastically, the ductility scale factors are very close. This implies that the earthquake sign does not have a signineant effect on F,.

Deselopment of Simplified Procedure A simplified procedure w as developed for estimating F, by first computing an "ctfective" system ductility

( ), which takes into account the effects of ratcheting. The Riddell Newmark method (Riddell,1979) was then directly applied to compute F,.

The effective ductility p, was determined by adjusting the system ductility as follows:

F,

• 1*"~' n Where:

p, = Effective system ductitity by coraldering effect of ratcheting y = System ductility without ratcheting effects n = Adjustment factor to accoant for ratcheting (n=3 was used in this evaluation)

The value of "n* may be interpreted as the number of strong motion ratcheting cycles to which the structure is subjected, and may depend upon, among other variables, duration of ground motion, difference in the available resistances in the upward vs. downward directions before yield, and input ti.ne history of ground motion. Such an interpretation of "n" is illustrated in Figure A-6.

In the Riddell Newmark method, the ductility scale factor is predicted from equations that are defined for the amplified acceleration, velocity and displacement regions of the input respense spectrum. For a damping ratio of 7 percent of critical, the following equations predict F, for systems in different frequency ranges. For systems considered in this study, only the acceleration and velocity regions are of interest.

Amplified Acceleration Region: F, = (2.673p - 1.673)"" (1)

Amplified Velocity Region: F, = (2.24 1.24)"" (2)

Rigid Range Limit: Fy = p 0"(S,/ZPA) (3)

Where:

S, = 7 percent damped spectral acceleration ZPA = 7ero period acceleration For the amplified acceleration region, F,is defined as the lesser of either equation (1) or equation (3).

Diablo canton Power Plant

1. M Pacific Gas Bild Electric Co.npany tong Term schmk Program

Appendix A December 1991 Page A 4 Also, since both the 4 hertz and 7 heru oscillators have frequencies that lie below the peak of the input spectrum, the rigid range limit w's not applied.

Note that, for this study, the simplified procedure was established by modifying the above Riddell-Newmark equations by substituting effective ductility, p,, for the value of system ductility, p.

Comparimns of Results of Simplified Procedure with Nonline.ir Analyses ne $1mplified procedure was used to calculate values of ductility scale factor, F ,y assuming an earthquake having 3 strong motion ratchetir.g cycles. These were compared with the corresponding quantities from nonlinear time history analyses and the results are shown in Table A 3.

As shown in the table, Fyfor the 4 Hz system agree well with the prediction by the Riddell Newmark method using the velocity region equation . Since the 4 Hz frequency lies far below the peak of the input spectrum, use of velocity region equation is appropriate. For a ductility of 10, however, the Riddell-Newmark method under predicts the F, value. For the 7 Hz systern, the results agree well when, for the Riddell-Newmark method, the geometric n'ean of the aceleration and the velocity region equations are used.

Conclusions The study was performed to examine the effect of the constant downward pull on a horizontally-oriented structural element due to giavity on the nonlinear vertical displacement of a floor being subjected to vertical motion. We observe from the results of nonlinear time history analyses that the ratcheting phenomenon created by the dead load reduces the ductility scale factor for a horizontal element eumpared to the case where the same element is mounted vertleally and is subjected to horizontal motion i.e., with no dead load.

A simplified procedure is developed by modifying the Riddell-Newmark method, Th. procedure incorporates the elfect of ratcheting by dead load and is achieved by using an " effective ductility", p,,

that is a function of the system ductility, p, and the number of strong motion cycles, n. For LTSP application using the LTSP vertical ground motion spectral shape, the simplified procedure is considered to be valid when a value of n = 3 is used, and for structures suojteted to gravity loads 3 :he range of about 30 percent of the yield strength of the system.

Dubio Canyon Pow er Plant is k

• Pacific Cas and Etectric Compary Long term seismic Ping-am

Appendix A. December 1991 Page A 5

( References KennWy, R. P., et al. Engineering characterization of ground motion - Task I, Effects of characteristics of free-Geld motion on structural response, NUREG/CR 3805,1984, prepared for the U. S. Nuclear

[ Regulatory Commission.

Riddell, R., and Newmark, N. M.,1979, C,.tistical analysis of the response of nonlinear systems subjected to canhquake, SRS 468, Department of Civil Engineering, University of Illinols.

Reed, J, W., Kennedy, R. P., and Lashkari, B.,1991 Analysis of high-frequency seismic effects, Jack R.13enjamin and Associates Report, prepared for Electric Power Research Institute.

Pac!0c Gas and Electric Company,1988, Final Report of the Diablo Canyon Long Term Seismic Program: Docket Nos. 50-275 and 50-323.

Kannan, A. E. and Powell, G. H., DRAIN 2D: A general computer program for dynamic analysis of inelastic plane structures,1973 (Revised 1985), distributed by NISEE/ Computer Applications, EERC, Urdversity of California, Berkeley, Reports No. EERC 734 and EERC 73 22.

Diablo Canyon Power Plant

i. J Pact!Ic Gas and Electric Company tong Term seismic Program

• Appendix A, December 1991 Page A-6 Table A 1

(

Scale Factor for No Dead imad and Dead lead Casm 4 Itz Model 7 Itz Model Ductility, n Yleid Ultimate F, Yleid Ultimate F, FS FS FS FS b'0_Dnd_ lad 2 4.83 8.52 1.76 - - -

5 4.83 17.77 3.68 3.78 8.57 2.27 10 - - - - - -

Dead lead Equal to 30 PercenLPy 2 3.71 5.29 1.43 3.03 4.09 1.35 5 3.71 8.27 2.23 3.03 5.55 1.83 10 3.71 18.22 4.91 3.03 9.16 3.02 Table A 2 Comparison of Scale Factor Based on Earthquake Sign 4 Ilz Model 7 liz Model ..

Yield Ultimate

• F, Yield Ultimate

• F, FS l'S FS FS Original Sign (+) 3.71 8.27 2.23 3,03 5.55 1.83 Reversed Sign (-) 3.38 7.75 2.29 2.65 5.08 1.92

• Ductility, p, equi to 5 Diablo Canyon Pcwer Plant is '6 Pacific C35 and Llettric Company long Term Setsmic Program

' Appendix A, December 1991 Page A-7 Table A-3

(

Comparison of Nonlinear Analysis Raults with Modified Riddell-Newmark Afethod Ductility Scale Factor, F, Frequency Target Nonlinear Modified Riddell Newmark Method (117) Ductility Analysis' 8 (see Vek>chy Acceleration Geometric Table A 1) Region Region Mean 2 1.43 1.41 1.30 1.35 5 2.23 2.33 1.87 2.09 4

5" 2.29 2.33 1.87 2.09 10 4.91 3.49 2.47 2.94 2 1.35 1.41 1.30 1.35 5 1.83 2.33 1.87 2.09 7

5" 1.92 2.33 1.87 2.09 10 3.02 3.49 2.47 2.94

• Dead load equal to 30 percent P y

" Denotes case where the input time history is applied with a reverse sign.

Diable Canyon Power Plant

!.'ai Pacific Gas cnd Electric Comp ny Long Term seam 6r. r agram

l o Appendix A, December 1991 Prge A 8 L

yfi"

'd 8

i

'A

.2 ~~.

S

_s =-

K h K g . ra NNNNvv//

1a) Vertical model 1b) Hori..ontal model Force 0.03K Py - -

l l 4  : Displacement

-May -ay oy pay K

0,03K .a

.. Py 1c) Force-deformation relationship Figure A-1 Models used in the nonlinear analysis, Diablo Canyon Power Plant

i. 6 Pacific Cas and Electric Company Long Term Selsmic Program

' Ay ndix A December 1991 l' age A 9 l

}.

0.400 -

0.200 -

S S

\

f

~

c U i ,  !

l 0.000 - .

J '

g 0

Lf![ I .

WNf Y

hW4wl%Wh 8 - J A

I i ,- i l y i l l j

0.200 -

I

}j 0.400 -

i iiiiiiiii iii.iiiiii,,,,,,,,,iii,,ii,,

D 40 5.00 10.00 15.00 20.00 Time (sec.)

Figure A-2 Input time history used in the nonlinear analysis.

Diablo Canyon Power Plant 4 ki Pat:fic Gas and Electric Company teng term seismic Program

• *

• gpendi A$ December 1991 Page A 10 l

2.00 , , , , ,,,,, , , , ,,,,,

5 - -

So - -

% I y - -

8 1.00 - -

M - -

~ ~

E @'

g - -

0.00 - , , , ,,,,,, , , , ,,,,,

1 10 100 Frequency (Hz)

Figure A-3 Response spectrum of the input time history.

Diablo Canyon Power Plant

. I Pacific Gas and Uectric Company Long Term seismic Program

• *

• AJpendi$ A, December 1991 Page A-ll l

l 11.0 -

~

)

!g 7.00 - g r \

- t  : (

i j 3.00

, (l I 0.00 - $ (

u l

$ ~

l a +3.00 -

)

e _

7.00 -

11.0 i ii i iiiiii,iiiii,,iiiiiii,,,,iiiii,,,t ,i 0.00 5.00 10.00 15.00 20.00 Time (sec.)

Figtire A-4 Response time history for the case without dead load, Diablo Canyon Power Plant

.6 Pacific 0a5 and Ut; b k Company long Term teismic Program

[.

4.

, Appendix A, December 1991 Page A 12

(

r L

f f 11.0 -

~

( 7.0 -

7 c.

3.0 -

E ..

E 0.0 -

f

~

.e o

4hr , l )

3.0 - l

- l 1 c

7.0 - l

.' ij IQ , W , = :::: n: . ::

-11.0 , ie i i i iiiii,,,ii ,,,,,,T, , , , , , , , , , , , ,,,,,

0.00 5.00 10.00 15.00 20.00 Time (sec.)

Figure A-5 Response time history for the case with dead load equal to 30 percent yield.

Diablo Canyon Power Plant 4 6 Pacif!C Gas and (lettric Company long lerm Seit.mit Program

_ _ _ _ _ _ _ _ _ _ _ _ _ - - - _ _ - - - - - _ - - - - - - = - - - - - - - - - - - - - - - - - - - - ~

g ,.j . Ay,< ;liI .' L1: cember ' el Page A 13 V

ij -

1 * .

. trong motion cycles

{ r

~

A l '

1

, T f

? h

/ g g e

I I I I I (

l - 1 I y

, e s. 43y o Ay Y (Egl')3y

( U

n/\ b.

s + ay Ductility r,er cycle =

OY Etlective ductility, p= ( _1) + 3 e

Figure A-6 Effective ductility considering multiple strong motion cycles.

?

Dtablo Canyon Pow =t P; ant

' 4i 44 Pacific Gas and Electric Cornpr.ny Long Term Seismic Prfgram