ML20147H519
| ML20147H519 | |
| Person / Time | |
|---|---|
| Site: | Oyster Creek |
| Issue date: | 02/26/1988 |
| From: | Kakai S, Sato T, Yamazaki F OHSAKI RESEARCH INSTITUTE, INC. |
| To: | |
| Shared Package | |
| ML20147H481 | List: |
| References | |
| NUDOCS 8803080511 | |
| Download: ML20147H519 (62) | |
Text
_
SOIL STRUCYURE INTERACTION ANALYSIS OF QUARTER SCALE MODEL USING AXERA CODE T.Sato', 5. Nakal" F. Yamazakl**, A. Mita** and T. lshil8 INTRODUCTION I Lotung SSI research project was initiated by the Electric Power Research Institute
)
(EPRI), the US Nuclear Regulatory Comission (NRC) and the Taiwan Power Company '
(TPC). Based on the technical exchange agreement between EPRI and the Tokyo Electric Power Company (TEPCO) TEPCO sponsored the analytical investigation of the j Lotung experiment in Japan. Ohsaki Research Institute (ORI) participated in this l project as one of the contractors of TEPC0 and conducted the prediction analysis and correlation assessment whose flow chart is shown in Fig. 1. This workshop paper presents research results and findings obtained through the analysis using thecomputercodeAXERA(A1spetric{arthquakeResponseAnalysis).
5 Note that ORI also employed another analysis method using the computer code HYBAX. The detailed '
flow charts of the prediction and correlation studies for the forced vibration test (FVT) and the actual earthquake records (May 20, 1986 and November 14, 1956) are shown in Figs. 2 and 3, respectively. A schematic view of the analysis method used in this paper is depicted in Fig. 4. AXERA based on an axis p etric finite element method can deal with both substructure and direct approaches. In the present i study, the direct approach is chosen. AXERA is similar to ALUSH except that AXERA allows an introduction of viscous boundaries (dashpots) on the bottom surface.
t The analysis matrix of the current work is shown in Table 1 in which only the properties of the upper portion of the layered soil model and the backfill are ;
depicted since the lower portion was not changed throughout the analysis. Based on the soil data packages, model 8 has been constructed for FVT blind prediction.
Since the elastic wave velocities of the backfill are not given, they were assumed i i in tha modeling. Through the comparison between the prediction and FVT
. measurement, model C has been constructed. These models B and C have been used for the blind prediction of responses subjected to May 20, 1986 earthquake. In the l earthquake response analysis, the computer code SHAKE (1) was utilized to obtain I the free-field in-ground motion and to construct equivalent linear soil models.
Model C' has further constructed thrt, ugh the comparison between the earthquake t i
i response prediction and recorded motions. The response prediction for November 14, l
, 1986 earthquake has been finally carried out using the model C'. '
l l
I Chief Research Engineer, Ohsaki Research Institute, Inc.
I
- Chief Research Engineer, Ohsaki Research Institute Inc., Ph. D./Dr. Eng, j f Research Engineer, Ohsaki Research Institute, Inc. '
SBR P
3%BBu Bi88b -
t I
Modeling of Structure f..,....
l l Modeling of Soil l , ,,l l 0 1 1
4 Structure Data l
l l l Soil Data from Data Packages ! !
l from Data Packages ll l
l l ,...............u................., !
j l
l '
!! Eigenvalue Analysis l for Three-Dimensional Finite Element
. . .Model
!l l
l l
. . . . . . . . . ,! ;l l
l
- 6................ ; ; ;
i ! l l Axis m tric Finite Element l ! Axisymetric Finite Element l l Model for AXERA l l Model for AXERA l l l l l
....................................l .!...................................!
l
. o Modeling of Complete Soil-Structure System (Model B) for AXERA l
l +
FVT Blind Prediction by Model B l
FVT Response Data l
Comparise of Predicted and Measured Results l
Refined Model (Model C) l l by HYBAX 1r Modeling of Complete Soil. Structure System (Model C) for AXERA
,: l
- I
- Comparison of Predicted and Measured Results l
Figurs 2. Flow Chart of FVT Prediction and Correlation Analysis 1
l i
1 I
L _ __ _ _ _ _ _
I T
A 1 I I
I t Forsed Vibrouse ,
+ 23~ es n.4 myy-
~_
m e.
m ,o
'- - m,u -+--
f-
\1 .
7 _ _ -
g - -
r - Energy I
a - TrenemJtung _
1 Boundary g __,,
SHAKE
/ i h D" 8v I . % %3 a.o Properum 1
y1 s leem l
[
i i
1 l
L Analysis of Forced Vibration Test - -
Analysis of Earthquate Response -
Figure 4. Analysis Metnod of AXERA i 1
c l
1 l
J
i l MODEL DEVELOPMENT AND DESCRIPTION '
i
)
' J Soil Modelinc !
Soil properties of Lotung SSI experiment site were susnarized out of the geological (
explorationandsoiltestingreports(DataPackagesa,eandf). The soil modeling ;
(Table 2) was determined primarily based on the result of cross-hole shootings (set i of L2, L5, L6 and L7 holee). Since there are no measured data for the backfill, i
! the S-wave velocity V, and the P-wave velocity V, were assumed for the blind i
predictknofFVT. The saturated unit weight is employed for all the soil, layers :
j (exce;t for the backfill in model 8) because the water table is considered to be l high (W.!..= GL20).
A nonlinear soil model was constructed based on the soil testing report (Data Package h) which contains results of resonant column tests and cyclic triaxial tests for undisturbed and remolded soil specimens. These test data were plotted in
)
i the equivalent linear representation (shear modulus reduction ratio G/GO vs. shear )
strain y, and damping ratio h vs. shear strain y) for five confining pressures,
, o,=0.35, 0.5, 1.0, 2.0, 3.5 kg/cm2 . Then the following hype bolic relationship !
between G/Go and y was introduced: l G 1 1 {0 1+y .......... ( 1 )
1 Yr i
] = ., in which y, is the reference strain when G/Go=0,5. It is well-known that the i j reference strain becomes large as the confining pressure increases. Thus three ,
values of y, were used in the modeling according to the confining pressure as shown in Fig. 5. The damping ratio A was modeled as ,
y '
g .......... ( 2 ) l l A= h,( 1 y ) + h, = h,- +h o a
1 1._
Y, in which A is the damping ratio at infinitely large strain andoh is the initial damping ratio added to the linear relationship in order to avoid infinitesimal '
. value of A. In this study, these parameters for the damping ratio were determined j j as A =0.3 and h o =0.015 for all the soils based on the laboratory test results.
i The soil region below GL-80s was assumed to be a linear elastic half-space.
i i
i 4
- 1. . _ _ _ _ _. -
_ _ . . _ _ _ _ _ _ _ __ _ O
.g
\l.
,,/ /.
I'
, 7.
GL 0 - -5 (GL10- -8forModelC')
7 .1,,. GL
. -1: -~' -31 (GL -31 for Model C')
j GL -80l 1.0 -, , --... g .a. '
a'+ i i "
, s
- l- ' ' , s' ' ' "
. i ,.
,N' .
s .'( .
.1 "i .
> s,\?
<; ,\ -
bg -
i
- g o '
\T 2 0.5 '
\
0
\\'\
'N \ -
\
N,N .
q
\ \ -
\s, .. N,s ,
O.
, , , , , , , , , , ,,,,n , i e"
08 10 5 4 10 10-8 10 4 Shear Strain y 0.4 . , ,,,,o , ,,,,,o , , ,,,,o , , ,,,,o
.c -
o . s * ...
4J a
/ .,-
/
/. ./ -
cs 0.2 ./ /*/
F
,/
/
a t -
/' /
/ .
~ * ,/ /./ ' .
. -&.1' O.0 ' ' ' ' ' ' " ' ' ' ' ' ' " ' ' ' ' ' ' "
4 4 4 10 10 10 10-s 10 4 Shear strain y Figure 5. Model for Strain-Dependent Soil Properties
l 11.0 unit : m l- 4
' 5.25 , 5.75 !
i I
- l r
i M 11
' ll C; I ll g . ll
\
1
- li l l I'
. l w
Bkdkhkid 1 g
_ 7 1 g 1
--.- 2 8 si 3 3 8 m
2:
)
2: --
1
- 1
} 1 5
2: 4 c l 2 ..
U o 4 8 g
I A - '
2.
. ha a g 2:
,3 s n
3 t
5 5 8 2: .
c3 2:
2:
3:
6 6 8 2:
j 3:
. t 5! '
2:- -
7 7 8 a
1-b $
l l
. 12.00 i
Figure 6. Axisy metric Model Layout for AXERA l
1
Table 4 Comparison of System Frequency and Damping System Frequency Damping Coefficient
/.(Hz) A (%)
FVT (radial) 3.80 9.5 (tangential) 3.81 11.0
'iind Prediction (modelB) 4.25 11.6 Correlation Analysis (model C) 3.98 14.0 Vibration Modes The vibration modes at the system frequency of the first rocktr.g modes for the radial and tangential excitation tests have been identified and they are presented in Fig. 11 using a vector representation, which includes both the amplitude and phase information. The rocking components due to the rotation of the base were obtained from the vertical response values recorded at the edge of the base assuming it to be rigid.
Comparing these mode shapes, it has been confirmed that the vibration modes predicted by the Phase I analysis agree well with those identified from the FVT
-n uits and that at the system frequency of the first rocking mode, the containment structure na del vibrates almost like a rigid body.
Refined Model C Comparing the blind prediction and vibration test results, it was found that the model B gave slightly higher system frequency for the first rocking mode.
Inspection of the vibration modes and response amplitudes indicated tb.t 6ne containment model might need no modification. Therefore the possible modifications were restricted to the soil properties.
The elastic wave velocities of the soil layers of the model B were determined from the results of the cross-hole shootings as described before. However, for the top soil layer which extends GLt0m to GL-Sm and the backfill, it is expected that their surface portions cre softer than to assumed soil profile. In addition, since there are no data available to determine the elastic wave velocities of the backfill, the assumed values for the backfill in the blind prediction (Phase I) analysis can be modified. Considering all these, it has been decided to modify
- FVT . FVT Model B Model B Model C Model C 2.o , , , , , , , , , ,,,,, 2.o , , ,,,,
i -
p g- 1.5 -
.gs > -- g I.5 ... -
i.\
1\ 4 ,
n Q,,.." " .. - l 8.. -
t,
, 3 ,, "
5 g i
./ .i.&..
5 a5 - ,e gi -
e5 .-
=
\.~4,*
- ..- ./ '\\
v.ss e.
ggl n s . e e n . a n . . __
g,g I a e a . e i e B 2 3 4 5 6 7 e 5 10 lt 82 83 14 85 0 2 3 4 5 5 7 e 1 3 to 11 82 13 84 15 ruum tai rauwvtwa 380.0 , , , , , , , , , , , , , , IED , , , , , , , , , , , , , ,
Rg - -
90.0 -
0 N N M o.o o.o E -
I T
\ -
I
-me -
't -
w ~. ~. .. ~. ::.-:. =;,
i g - %,.
-: ,- ?
0 8 2 3 4 5 6 7 8 9 le II 12 83 le 15 0 2 3 4 5 6 7 8 1 8 to it 82 13 14 15 rune ses rame tes Channel 2 Channel 3 Figure 7. Response to the Radial Excitation (Channel 2,3)
"!T . FVT hodel B - - - Model 8 Model C 2.o Model C
, , , , , , , , , ,,,,, 2.0 , , ,
. . e T.
i.5 g% D i.5 h ii b h NS s.a -
- .1 a =\
1 L ,
%q', , " " ' .
g n.o L ,
. g
% ,=
5 t:
_ I ~\ d I5 -_. ,. / "ki. ss I5 -
.,s-.\
gy-2 9 8 3 4 5 6 7 8 9 IS II 62 83 14 15 0 2 8 3 4 5 6 7 e s 80 II 82 13 84 15 Fil[0LA.NCY (NZ) iMOUENCY (H2) imo . . . . . . 8mo MO -
30.0 -
N N
. # .. n a0 s
s I '
- \
.go _ \ \-
-90.0 -
k%%%'.WW'*""**~",*.*.*.*'.',*
w = % _. ,, 't.y% es , . - _ _ __
- C' ,,ij.j 0 8 2 3 4 5 6 7 0 8 It 88 82 83 54 85 0 t 2 3 4 5 6 7 e 9 Is is 82 33 se 35 FM0tE8CT (HZ) FMOLKICT (H2)
Channel 1 Channel 5 Figure 9. Response to the Tangential Excitation (Channel 1.5) 4
=r =r I [ g
=
Y 4
i l Ur i Ur
! l /
- ewl " ! ': U- i s.i i_ u. _ s e,n i w i
- i g i e
I
- l l 's- 'a
, g , .,
- I 7
l m/ ..
l i
l i...}.. !,i *. ,/
i' /
l . .: ,3 ;:
,! l
, ,. .,,-,,,,,,,,,,,M,,,,,,.
l l "l ..t ,<.
I i s :
7 i : .
w\v, /
.. .__=== _r - 4 .p
\u, ) \
w4 i
Vibration Test Result Vibration Test Result l for Radial Excitation for Tangential Excitation '
(fo=3.80Hz) (f0=3.81Hz)
~ I
- b =r i
! ,' N' ,
o I, ,
n l l :
.! ! i E o
/' V - ' u' E l */
I Nl I e e = 8
=
m i l
i
,e i I
a o 8
0 l
I I
I i I E i
- =
u -
r =
s . --
\
u.A_n: v. . L.n- r Model B for Radial Excitation Model C for Radial Excitation (f0=4.25Hz) (f0=3.98Hz)
Figure 11. Vector Representation of Total Response at the Top
i V, was used instead of V,. All the other conditions of the analysis are the same as those for the N-S and E-W components. The equivalent linear soil properties for
)
the N-S and E-W components obtained by the deconvolution analyses are shown in Fig.
- 13. In Fig.14, in-ground distribution of the maximum accelerations to the May 20, 1986 earthquake is depicted. Acceleration response spectra with h=5% at DH86 and DH847 are calculated and shown in Figs. 15 and 16, respectively.
Equivalent linear Soil Models for AXERA !
The computer code AXERA is based on a frequency domain analysis and does not deal with nonlinear problems. Hence, equivalent linear soil models were constructed from the analysis results by SHAKE to consirier the soil nonlinearity during the earthquake. In the SHAKE analysis two sets of G/G and h values for both N-S and ,
o E-W components were obtained. For making the equivalent linear soil models, they were averaged in each soil layer and the resulting models are explained in Table 5 for the models B and C and Table 6 for the medel C'. In the above process, the P- I wave velocities were kept unchanged over all layers. Though the damping ratios for vertical motion are different from those for the horizontal motion, the damping values for the horizontal motion were employed for the models B and C. The G/G value of the backfill was determined by assuming that the shear stress of the o l backfill is equal to that of the top soil layer. Therefore the strain of the l backfill y3 was determined by the following equation:
1 01 y
g = Q*" y g .............. @ {
where y3, Go , and Go3 represent the strain of the top soil layer, the initial shear l modulus of the backfill and that of the top soil layer, respectively. Substituting this y, into Eqs.1 and 2, G/Goand h for the backfill were obtained.
Response Analysis Usina AXERA Using the equivalent linear soil models, the frequency response analysis was perforn.ad to obtain transfer functions with respect to the control point (FAl-5).
The seismic excitations were chosen to be vertically incident plane S-waves for the horizontal excitations and a vertically incident plane P-wave for tne vertical excitation. The transfer functions were obtained for the N-S (vertically incident plane S-wave with N-S particle motion), E-W (vertically incident plane S-wave with l E-W particle motion) and U-D (vertically incident plane P-wave) excitations. I i
Time Histories and Acceleration Response Spectra By use of the transfer functions, the time histories of the response to the May 20, 1986 earthquake were calculated through the FFT algorithm. Since the steam generator was not modeled, the response values at F4GU and F4GL were not obtained.
The 5% acceleration response spectra for the calculated motions and the recorded l
, - - , - - , ,,a w
i Layer Depth Sublayer Loestion of Vs No. G.L. (m) Accelerometers (m/sec) 1 b
V - 5.' 120/85*
2 v -8.0 GL-6 140+
3 V -13.0 GL-11 190 i GL-17 1
4 220 l
V -31.0 '
5 7 - 34.0 280 6 250 V -48.0 GL-47 7 270 V -60.0 l
I l
8 320 l 1
i V -80.0
////////////
g Elastic 480 Halfspace V, values of sublayers in Layer 1 (and 2) are assumed to increase according to depth (See detail in Table 1).
Figure 12. Analysis Model For SHAKE
2 Max. Acceleration (cm/s) Max. Acceleration (cm/s2 ) Max. Acceleration (cm/s?)
N-S Component E-W Component U-D Component 0.0 100.0 200.0 300.0 400.0 0.0 100.0 200.0 300.0 400.0 0.0 25.0 50.0 75.0 100.0 0.c F 0.0 -
0.0 e l > \\
l
.; .l .
. l ,, l .
-20.0 l
-20.0 l -20.0 l h?
! l n l S I m -40.0 -40.0 g' -40.0 j
%. s
.L 1
1
\ >
( '.
-60.0 ) -60.0 j -60.0 i
{. I t
- l. I j
= Recorded
- Model B 1
! t Model C
-80.0
-80.0 l ' 'k -80.0 Model C' Figure 14. Maximum Ground Acceleration for May 20. 1986 Earthquake
Table 6 Equivalent Linear Soll Model for May 20, 1986 Earthquake (Model C')
Layer Depth Reduction V, vp Unit Weight Shear Modulus I Poisson's Damping No. G.L. (m) Ratio G/G, (m/s) (m/s) y,(t/m3) G (t/m2) Ratio v istias A 0.0 1
? -5.0 0.50 60 262 1.87 690 0.472 0.105 2 -8.0 0.50 99 810 1.87 1870 V 0.492 0.105 3 V -13.0 0.75 165 1270 1.87 5160 0.491 0.070 4 0.70 184 1330 1.87 6460 0.490 0.080 V -31.0 5 V -34.0 0.90 266 1330 1.90 13670 0.479 0.050 6 0.85 230 1250 1.90 10290 0.483 0.055 V -48.0 7 0.85 249 1220 1.90 12000 0.478 0.055 V -.
C *
? -5.0 0.33 60 323 2.24 822 0.482 (inner) 0.115 9 _S.0 0.65 85 323 2.24 1640 0.463 0.085
M Recorded "
W Model B Mode 1 C
[4
'r'u".i ru.:(
, , , ou, ,,
Model C' ,, ~ouin ,,onii 0.5 , , , , , , , ,
oni, onst S
0.4 * *"A "
60.3 C
E w 0.2 - > -
' ~
, ,f '}\ '
0.0 : ' ' ' '
O. 2 0.5 1.0 2.0 5.0 10.0 20.0 FREQUENCY (HZ) 05 e i iiii i i e i i l 1
n 0. 4 DHB47-E l O
E 0.3 A
s EE F
w n,y ** 's /
/'i %.
gj ,, '( \
d cm o,g .e 2
O
's, N
-m- &j., s s,,, - -
- s. _
)
, v -c s....
... j l
0.0 ' ' ' ' '
0.2 l 0.5 1.0 2.0 5.0 10.0 20.0 FRECUEtCY CH2) 0.25 , , , , , , , , , , , ,
g, g DHB47-V S
E 0.15 b 0.10 b 0.05 '- -
M r. = .....
0.00 ' ' ' '
O.2 0.5 1.0 2.0 5. 0 10.0 20.0 FREQUENCY (HZ)
Figure 16. Acceleration Response Spectra at DHB47 for May 20, 1986 Earthquake (h=5%)
A
Recorded r.us W r4w
<y Model B f 7* "i -' ' "
re Model C Mode l C ' @@}t ras
- "^2'
'c 5 1.0 ,
. ,3 i3 , , 6 ,,i
- 0.8 F46-N onA4' 'o"*47 o
w 60.5 t
E g
5 * f,,1 '
" 02 s . -
~..v.:.=... k.:. h 5 m 0.0 I ,,,, , , ,,1
,l 0.2 0.5 1.0 2. 0 5.0 10.0 20.0 FREQUENCY (HZ) 1.0 , , , , , , ,
- 0.8 F4LS-E o
w 5
i 0.6 E
3 0.4 ,
- 0. 2 0.0 ' ' ' ' '
O.2 0.5 1.0 2.0 5.0 10.0 20.0 FREQUENCY (HZ) 0.5 , , , , , , , , , , , ,
O.4 F4LS-V e
w E 0. 3 E
g 0. 2 -
U r s.
ll 7s,7, \'/i' Y ,, ! '. n.*;NA VW %?)~~x 0.0 -- >! ,,,, , , , , , ,
0.2 0.5 1.0 2. 0 5.0 10.0 20.0 FREQUENCY (HZ)
Figure 18. Acceleration Response Spectra at F4LS for May 20, 1986 Earthquake (h=5%)
1 i
l
Recorded 1 Model B l Model C l Model C '
j 2.0 , , , , , , , , , , , , , ,
i is F4UN
, ,t ., ~.- ,
,3 s .
' / F4UE 1 3C -r$.c'*us l.s - /\ I #
re.s m l\l\[g
'\* <w .ons i
, e i & jeun 'enn l / . ! " . lean
~ ;/ .i / \\
$ l.,' / \ \
..cun loun .ono l
- 1.0 { -
'\ ,,\
{ '\',/'\\ l \
s tA t 0.5 \. b, f
\ \,N ,
\
'N \- s
'Ns/(' 't'[,\,r
,---._.s/\g,.,,,^.,,
g,g , , , , , , , , 'T- r------- --- ~ r 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 FREQUENCY (HZ)
F4US-N / FA',-5-N 2.0 , , , , , , , ,
\
l')\ ll \
1.s - / '
S, j ,
w
,I \, \ 'y \
3 1.0
\/ ./' \\
\i i,
% \. s '
Md i.\, \,\, ,a 0.5 -
\. \ )'\
\ ' ,'4i I.\w lg, N 'N /
'N . ~ ~,,. ~\. -,. ,,f A , ~/'" , /
g,g , , , , , , r - r------- g - o r 0 1 2 3 4 5 6 7 8 9 to 11 12 13 14 15 FREQUEC Y CHZ) l F4US-E / FAl-5-E l 1
Figure 20. Transfer Functions for May 20, 1986 Earthquake (F4US-E/FAl-5-E)
t cM/s/s ) ex 200.02 Recorded 200 -
0 -
- .' ' ' ~,0 yj.Y'y^(.y'^ ^^ ^--^^
- =
-200 - -
~
Mx=162. 93 2o0 . .
'[ Mode 1B o
o _ . , . .y,A ! i > . .gM,.
. .^; _ .,_ m... _ . . ___- _,
- l,,
-200 -
psx153.71 goo . .
[ModelC f
0 ' "
"?. '^$.c'[!y^
,, ,l,
^
^-^-
)
-200 -
-Mx=167.35 200 - [Model C '
o ""
" "! ,' kl,', ' L '[/ y^r . " ~ ' - ^ -
-v- -
.N, 200 -
b 1'o 50 30 $0 (SEC3 Figure 23. Acceleration Time Histories of F4US-E for May 20, 1986 Earthquake C CM/s/s ) mx=156.93 200 - Recorded O .4.,- '{; ^ - ,: ~:_ _ - - - V
-200 - -
- t
- - -e x=155.79 - --Model B-200- - l 1
- f. l 0
- -- . ] g,h q, ' : ^ l y ^ ,. : - : - .- .- -
} .
l -200 - - --
-e x 150.57 ,Model C 200 . .
t 0 ^4 ,,. A a y ^- -- - - -^ , - -
!Y.yra'/t
. goo . . .
. . . .ex 154.58. . . . . . . . . .,Model C' 200 -
.f. A 0
^
^? . . y 'v"-- ^[i y^ - ^~ ^ - ^
i -200 - - d t'o do 30 io (SEC3 l l Figure 24. Acceleration Time Histories of FAl-1-E for May 20, 1986 Earthquake I
- . , -l
l l l l 1 l C eR 3 MRX 35.37 80 - Recorded 0 ^' y-
'^
-^^^-
^'~~ ~ - ~ ~
~
-gRx=45.00" so . .
"Mode 1 B O ^ ^
^li'., //' 'v #/v^
--~c' --
^ -
-so - ~
-MRx.52.91' 'Model C 33 .
O ^
.4 ,l,1;//' 'vd e v ^ --
^
^
40'
'MRx EB.06 ~ Model'C
so . . t 0 ./ ^/ glf,,",# ^ '[."- 9^ ^^ ^ - - -'v-~
^ --
3a . h l'O NO $a /0CSEC3 Figure 27. Pressure Time Histories of P4WS for May 20, 1986 Earthquake ( en ) MRx=48.95 so - Recorded 0 _t, . A .A . . n
' 1pqvv vN M 'vv w- i l
go . .
%xso.52 Model"B~
~
0 -
>l'/,^,^ p y^ y~.y- _^ - -_-- --
- - - -MRx-es.Os-so . . . .
..M. odel .' .'C' -
. , 1 0
'^^T.
.'((--^y"v' v1^ y" ^v~vt' -^ -^
gJ .
. --MRX=79.69 ='odel M C' so . .
0 . '/ ^ " Y ~yv ^ ^ 'yb ^ ^y ~ ^v~v'# ^^ ~~ - - ^ -
.co . ..
b 1O io 50 /0 CEC 3 Figure 28. Pressure Time Histories of P45W for May 20, 1986 Earthquake
9 RESPONSE PREDICTION AND RESULT COMPARISON FOR NOV. 14,1986 EARTHQUAKE ! l The blind earthquake response analysis has also been carried out for the November 14, 1986 earthquake following the same procedure as for the May 20, 1986 j earthquake. The refined model C' was utilized in this prediction. The maximum l ground response and equivalent linear soil model obtained by SHAXE are shown in Fig. 29 and Table 8. respectively. The recorded maximum ground acceleration is also depicted in Fig. 29. A good agreement is seen between the predicted and I recorded maximum ground acceleration. 1 Acceleration response spectra at DHB6 obtained by SHAKE and those at F4US, F4LS and FAl-1 obtained by AXERA are plotted in Fig. 30 to 33. Acceleration response l spectra at DHB47 are not shown here since no record was available at this point. ; In these response spectra, a very good agreement is observed in the wide range of l frequency. The transfer functions between the points, F4US and FAl-5, are plotted I in Fig. 34 in the similar manner as for Fig. 20. The predicted transfer functions by the model C' excellently match those obtained from the actual records of i November 14 earthquake. It is also observed that the system frequency by this November 14 event is almost the same as that by the May 20 event: f=1.8 to 2.5 Hz. Finally, acceleration time histories at F4US and FAl-1 are shown in Fig. 35 to 40, the pressure time histories at P4WS and P45W are shown in Figs. 41 and 42, and the maximum acceleration and pressure values at the selected points are sumarized in Table 9. In these figures and table, a good agreement between the prediction and observation is again confirmed except for the interface pressure. Hence the SSI model construction and modification procedure developed for May 20 earthquake may be valid for the other earthquakes of similar intensity. From the comparison between the earthquake response analysis and recorded responses for the May 20 and the November 14 earthquakes, it may be concluded that the correlated model C' explain the response to the November 14 earthquake better than that to the May 20 earthquake. This is probably due to the fact that the recorded I ground motions for the May 20 earthquake contain surface wave components, which are - not considered in the current analysis, more than those for the November earthquake l does. l l
4 Table 8 Equivalent Linear 5011 Model for Nov. 14. 1986 Earthquake (Model C') Layer Depth Reduction V, Yp Unit Weight Shear Modulus Poisson's Damping No. G.L. (m) Ratio G/Go (m/s) (m/s) y,(t/m3) G (t/m2) Ratio y Ratio A 0.0 1 V -5.0 0.55 63 262 1.87 760 0.469 0.100 . 2 7 -8.0 0.55 104 810 1.87 2060 0.492 0.100 3 V -13.0 0.80 170 1270 1.87 5510 0.491 0.065 4 0.75 191 1330 1.87 6960 0.489 0.070 V -31.0 5 V -34.0 0.90 266 133J 1.90 13670 0.479 0.050 6 0.85 230 1250 1.90 10290 0.483 0.055 V -48.0 7 0.85 249 1220 1.90 12000 0.478 0.055 V -= 0.35 62 323 2.24 0.481 (1 r) V -5.0 878 0.105 Backfill 0.0 0.70 88 323 2.24 1770 0.460 0.080 (outer) V -5.0 4
S
- Recorded
[4!w ra M ru.i ru.: res N, re Ious . ons ; 1.0 Model C' Y "o m : out "oun
-o m )
l m 0. 8 ri
^^
,ri. F4US-N .o m , .o m ,
0.6 ? \
,I%)'.' Y p .\ i a s \
5 0.4 ''' '^
& of' $ Uq"-1, ' a ..
0 E f; ---________ l
- 0. 2 a -, y' ' '
0.0 O.2 0.5 1.0 2.0 5.0 10.0 20.0 FREQUENCY (HZ) 1.0 , , , ,,, , , , ,,,
- 0.8 F4US-E S l's 50.6 ,, l' C j \' ! ,% n.
f o,4 ll N, t.kI __
\..
e . ~
, , g. s 8
1
,\f,% ,; 3 -
-g Q,c i 0.2 l
;;;..n
- 0. 0 0.2 0.5 1.0 2. 0 5.0 10.0 20.0 FREQUENCY (H2) l 0.5 , , , ,,, , , ,
,/\, ,
F4US-V l - 0.4 ,L
,- i
)
E 0.3 c !E ,!,t),w ,u , / xi w 0. 2 g r.l'M. y
~
01 tv v l
~
J ' 0.0 O.2 0.5 1.0 i.0 5.0 10.0 20.0 I FREQUENCY (HZ) Figure 31. Acceleration Response Spectra at F4US for Nov. 14, 1986 Earthquake (A 5%) i i
r<ve M y vm [ 4's, y34: f fay
------ Recorded d re low om
_..._.. Model C' T f**' iDW1 1.0 , , , , , , 4DMAlf *DK3 t f
- 0.8 FAl-1-N ow? .o m t g 0.6 - ,f, .
E i .^ / \. s
}VI 'l
%,,,l.
Eo2 yy w~~. . .- ~<.=.: -;
~~ ' ' '
0.0 ' ' ' ' ' ' O.2 0.5 1.0 2. 0 5.0 10.0 20.0 FREQUENCY (HZ) 1.0 -
, O.8 FAl-1-E E 0.6 E
w 0. 4 l Q*rQ *?' ,'s\__ 0.2
,,j s,j
.a,y,,,_,Q _
0.0 0.2 0.5 1.0 2. 0 5.0 10.0 20.0 FREGUENCY (HZ) 0.5 , , , ,,, , , , . , . - 0. 4 I FAl-1-V
. f.
E 0.3 - !'. \ t,' ' , ',' C e V \ 'r. \.\ (..\ E .I$ %W b)/(*s,'\)/l \
,'ll '
\'\)r.
J
/v u'
_ .,e Y 0.0 -- 0.2 0.5 1.0 2. 0 5.0 10.0 20.0 FREQUENCY (HZ) Figure 33. Acceleration Response Spectra at FAl-1 for Nov. 14, 1986 Earthquake (A=5%)
C CM/S/S ) max =247.87 240 - "' Recorded
.. A u ,i83 A s.4 II ./ iA l o -
. - . . . . . . , r ipp i.ge aim,qA M.y.c. .' A . e. c- ., A , &
~# ~ ~
M'Ax=189.69 ,,Model" C s~ 240- - o , . h k,[ A , (.,,.,r A , t ,, 4.....s. .hi) q. , ,A h f jl.l, k,4,, q.q, y
-240 -
b l'O $0 3'O 4O CSEC) Figure 35. Acceleration Time Histories of F4US-N for Nov. 14, 1986 Earthquake t cx/s/s ) Max leo.8e 240 - . Recorded
.'."w".','," n. p' . . y' 'i y y'y ."p' q", , ' ,. ',. '-.' .'. ^. ^-
0 - --
^ '
N n. U[ ~~ max 15s.72'
.Model C' I c ._.........~ ....... L 4 ,,3, # ,,s.
.,. , , . 7 , .q p y , , , , .~ . .
-240 -
l'0 20 da /0 CSEC) l Figure 36. Acceleration Time Histories of F4US-E for Nov. 14, 1986 Earthquake i cM/s/s ) max =1:1.ot im. . . ., Recorded I g _ . _ m
~~ ~ ~ ' m!rir u,. O Ndt.hYII.!IVI.ill.lk A. a n a J 'v'"- .
A..u < m - -~ i., pig.1 9 q.'q y, -( r r" 'r w l -120 ' ' HIRX 91.98" .Mofg1 Cs1' 120 -
' D ^ '" "'
0 ~-
'"r]ijpe p r"'. T'"r
^
"". r' 'PU q Ibir ' i
-120- - b l'0 20 io 40CSEC) Figure 37. Acceleration Time Histories of F4US-V for Nov. 14, 1986 Earthquake
( pa ) max =28,19 so - Recorded 0 --^ ^ ^ ' ' Y+'f?'8, f .^. -^ * ' ~ ^ ' ' ' " i
-80 '
l sx 84.se" ' so - [Model.[C 0 a
-. ..,. A s . , l i s . 6 h . .I l i. M i k.hl .H A.. ,... e u .
.. l
. ....,.y y 9 .p.q ' p w v'ir' ' ' 1 80 .
6 l'c io da lo CSEC) Figure 41. Pressure Time Histories of P4WS for Nov. 14, 1986 Earthquake C ca ) Max 4e.es 80 - Recorded g ...,n.u.Ae*.I.td. 't sfhl d! A .A ,.A, .
") ') )( * ' b g'ip' g up' lg y'}i Y' V W *"'W V" %
80 . sr76.11" "Model C' 0 ^
' b - '
. . . . . m '.".',g
. ,, .m', , Y. '.'q'
~ ' '- b 80 i b
l'O $0 $a /0CSEC:l Figure 42. Pressure Time Histories of P45W for Nov. 14, 1986 Earthquake l l i l 1 1 I
-- -=-
l i METHOD EVALUATION AND CONCLUSION l Forced Vibration Test (FVT) l The blind prediction analysis based on the data packages gave a good estimation for the system frequency and damping, and the mode shape of the response of the containment model to the forced vibration excitations. Hence the modeling of the l soil and structure employed in this stage may be fairly close to the actual system. l Although the elastic wave velocities of the backfill still remain unknown, the backfill might be harder than the top soil layer from the comparison between the i analysis and measured results. The realistic modeling capability of AXERA, i.e., viscoelastic layers with viscous dashpots on the bottom surface and the inhomgeneous soil region near the foundation, has also contributed to give the accurate estimates. The code AXERA is therefore considered to be able to simulate the soil-structure interaction system accurately when the soils are considered to be elkstic media. l Earthcuake Response The equivalent linearization algorithm was employed in the deconvolution analysis for both May 20, 1986 and Nov. 14, 1986 earthquakes. The conventional value of effective strain factor a, 0.65 or 0.7, did not give convergence in the process to find strain compatible soil parameters. The shear strain induced in the upper soil layers during the earthquakes is considered to be large enough to exhibit the strong nonlinearity. When such large strain occurs, the deconvolution analysis based on the equivalent linearization algorithm may not be appropriate. It is also pointed out that frequency-independent complex damping ratios for soils may cause unrealistic increase of high frequency contents in ground responses when it is employed in deconvolution analyses of strong earthquakes. A correlation analysis has been performed for the recorded ground motions of the May 20, 1986 and the November 14, 1986 earthquakes at FAl-5, FA2-5 and FA3-5. It showed that the horizontal phase velocity is very large, so that the assumption of the vertical incidence of the plane elastic waves may be reasonable. The estimated angle of incidence is less than 5' from the vertical axis at the top soil layer for the May 20 and the November 14 earthquakes. In the equivalent linear algorithm, a careful attention should be paid for determining the equivalent damping ratio of soil. Although the conventional soil-structure analysis codes usually assume that the damping ratio associated with P-wave propagation is the same with that associated with S-wave propagation, it has been shown that such an assumption may reduce especially the vertical component of the response. Therefore it is suggested that the computer codes should be capable of defining different damping ratios for S- and P-waves, respectively. The maximum pressures at P4WS and P45W calculated using the model C' are somewhat s
MD f
?
l l l i SOIL-STRUCTURE INTERACTION ANALYSES OF QUARTER-SCALE CONTA MODEL EXPERIMENT IN LOTUNG. TAIWAN i l Bechtel Group, Inc. San Francisco, California l l PREFACE The work described in this paper was performed by Bechtel Power Corporation, San Francisco, under a contract agreement between Electric PowerThe Research BechtelInstitute project (EPRI), Palo Alto, and Bechtel Group, Inc., San Francisco. Drs. W.S. : team for various phases of the project consisted of Bechtel engineers: , Tseng, T. Takayanagi, S. Vahdani, K. Lilhanand, S.Y. Tuann, and F. Ostadan. W.S. Tseng was the principal investigator. A.H. Hadjian served as the project l consultant. l l l I l l l 1 i l l l
, - - - . - , - , - - -, . ~ . , . - . -.
l l l l l l PART 3: FLUSH METHOD l CENERAL DESCRIPTION OF TE FLUSH METHOD ; Methodology and Computer Program ! The FLUSH SSI analysis method is a two-dimensional (2-D) finite element soil-structure interaction analysis method using the frequency domain solution procedure. The computer program FLUSH was developed at the University of California, Berkeley (Ref. 3-1). The program has been widely applied in the U.S. , j nuclear industry for seismic SSI analysis of embedded structures. The version of I FLUSH applied for this project is the public domain version which is available' ' comercially at the CDC computer service bureau. This version of FLUSH has capabilities and limitations as sumarized below: o The soil medium of the SSI system must be horizontally layered I soil strata overlaying a rigid base rock. No wave transmitting boundary is used at the rigid base boundary. 1 o The seismic input motion is assumed to be vertically propagating plane seismic 5 and P waves. The motion can be prescribed at any : elevation within the free-field soil medium. The program can perform free-field deconvolution analysis of the input motion to ; generate the base boundary motion for the interaction analysis. ' 1 o The soil material is assumed to be linear viscoelastic material characterized by the complex modulf with ccnstant hysteresis damping. The soil shear modulus and damping ratio can be specififed as functions of shear strain. The strain-dependent soil properties can be specified differently for every soil element in the SSI model. o The program is a linear analysis program. However, it has the capability of automatically or manually perfoming a series of iterative, equivalent linear analyses to ensure strain compatibility of strain-dependent soil properties. ) o The program has the options of specifying the horizontal wave transmitting boundary along either one or both side boundaries of the soil model and the viscous wave transmitting boundary at any point in the plane of the soil model to simulate 3-D wave radiation effect. o The 2-D soil model is modelled with plane strain finite elements and the 2-D structural model can be modelled with 2-D beam elements. Special void elemer.ts can be used to simulate the basement void space of ersedded structures, l i _ _ . . _ _ . _ _ _ _ _ _.. _ _ _ _ _ _ _ J
Since the internal structures are located off the NS and EW axes of the containment, the structural model for the internal structures cannot be directly included in the 2-D FLUSH model. Thus it was decoupled from the containment model and not explicitly included in the 2-D FLUSH model. This decoupling is reasonable because of the relatively light internal structures. For detemining the response of the internal structures, the containment base response motions at the internal structure support location were obtained from the 2-D FLUSH SSI analyses. These motions were properly combined to give the 3-component base i motions which were then, applied as the base input motions for 3-D fixed-base response analyses for the internal structures. For these decoupled analyses, the 3-D lumped-mass stick models of Model B and Model C for the internal structures described in Part I were directly used. Thus, the difference between Model B and Model C for FLUSH analyses was only in the internal structure models. FLUSH Model The configuration of the FLUSH SSI model developed is shown in Fig. 3-2. The model is a half model with anti-spimetric boundary condition at the certerline for the horizontal NS and EW analyses, and with symetric boundary condition for the vertical analyses. The model consists of 2-D lumped-mass beam stick model for the containment and 2-D plane strain finite elements for the soil medium. The depth of soil medium included in the FLUSH model above rigid base was chosen from the result of a sensitivity study in which the depth of the soil model was varied and the sensitivity of the containment response was evaluated. The finite element mesh of the FLUSH model was sized to pass seismic S and P waves up to a maximum frequency of 10 cps. The horizontal wave transmitting boundary was used for the right-hand side boundary of the model. The viscous wave transmitting boundary was applied to every soil node in the model. Since the 2-D containment stick model is the same for the NS and EW directions, ; the FLUSH model shown in Fig. 3-2 is applicable for both the horizontal NS and EW analyses. The initial low-strain soil profile used for the FLUSH model was the i l best-estimate low-strain soil properties as given in Tables 1-1 and 1-2 of Part 1. I EARTHQUAKE RESPONSE PREDICTIONS AND RESULT C(MPARISONS Earthquake respense predictions have been made using the FLUSH computer program for both earthquake events, LSST07 dated 5/20/86 and LSST16 dated 11/14/86. The respense predictions for the internal structures have oeen made for both earthquake events using both Model B and Model C internal structure models. The input motions to the FLUSH analyses were the free-field surface recorded motions at the recording station FAl-5. These motions were assumed to be prescribed at the surface of the FLUSH soil model. FLUSH free-field' deconvolution analyses were performed to obtain the strain-compatible soil properties and the free-field soil response motions at the free-field downhole recording station DHB6 which is located at 6 m below grade. The free-field soil response motions at DHB47, located 47 m below grade, were not calculated since the FLUSH model did not exter,d to this depth. The strain-compatible soil properties as obtained were very similar to those obtained from SHAKE analyses presented in Figs. 1-15 of Part 1.
- _ - .- -- -- - - - - - - - , - - - - - - - - - - - - - - - ---- ---- A
i l response amplitudes agree well with the corresponding recorded amplitudes. The decoupled internal structure response provides acceptable results due to its relative small mass. From the result of thi: study, it may be concluded that the FLUSH method of SSI analysis with the associated model parameters applied in this project, in particular the modeling of the equivalent strip, is adequate but slightly on the unconservative side. This may be caused partly by the criteria for developing the equivalent strip and partly by the viscous dampers used for simulation of 3-D wave radiation which may have over estimated the damping for the FLUSH soil model. REFERENCES 3-1. Lysmer, J. , Udaka, T. , Tasi, C. F. , and Seed, H. B., "FLUSH - A Computer Program for Approximate 3-D Analysis of Soil-Structure Interaction Problems," Report No. EERC-75-30, University of California, Berkeley, 1975. : 3-2. Hadjian, A. H., Luco, J. E., and Wong, H. L., "On the Reduction of the Three-Dimensional Analysis of Soil-Structure Interaction Problems to Two-Dimensional Models," Proceedings of the 2nd ASCE/EM Specialty Conference on the Dynamic Response of Structures, Atlanta, Georgia, January 1981. i l I l 1 l
l l l 4 ! MIPt(a)
' g7 10.67 (128) 1
- 7. 5 o
.46 2.13 4.28 5.81 9.78 '
1 2.90 133 11 28 1 0 (134 lu: < ' ' ' l 138 (138) 137 :03 81 59 *4.57 (138' 1 10s, (st) ,40) .g
-4 13 8***e ry Transeitting M Boundary ;
wrtnnut Analysis l l Vertical I Analysts j l l 31 14 154 !!4 102 80 le 21 164 163 !$2 til 163 til 37 m vm,=e g _ _.r I algte sese Figure 3-2. Configuration of FLUSH 2-D Finite Element Models B & C
t i B9 i . . 1.9 .
--e ar.t,e.nts m s goe.ee inest6 e a et sa samles--. weer sa semeses
- t 8tCaso wo. g
- sre
= esse (6 e . g
,. . e ,. . e
& E 8 3
- .. : .. _ . _ _ . g....
U P 7 0 Y.. .- Zl W .. I /d N b_\ I Y t u ) u g ., -%- g ..
/;
Q
.e ' .e 19-8 se
- le a se e ge o 3, . se o gg a FDt0WtsC,-C*S F8t M sCf. css FLUlw p(tse00 L55707 Fev$ kl FLuga settM00 LS5707 74L5 al
( / l N 1.0 - . . - -
, - . . 3 . 0 ,---
54 Samle6
--86w arte.*urs4eBWWW ee immag 6 e e ts === WIWWE3,WWest
-sw$. g wg samet6 e a g3 to Desprot e .J .9 A a 5
0 .s C
/ E .e
- 5 5 .. -
.? $ .. -
1 I
- T
- 2 w
u I ... ~
=
I ., - 4 8 48 ' se 8 le 8 se s see se o gg e F #E M mC V-CM Fet M uCT CPS PLUSN mtTM00 LSlf07 FevS gm FLugw agtee00 LllTO? F 45.5 tu i.. - -- - . . .... , . . ... ...__ -
= F2'"!,3?X 6 . . ,, = f 3 " t m ,. . ,, a
... . .e J J 8 I J
- , =
W d V.. I i w E u I .. K ..
, i
% 4 i . --
f i.. w
.. / y l i .- . ... ... ...
FM0utett CPS Pgtututt-C'S PLUSM M 71800 Lllig? F4W1 y Flutes MTM00 Lllit? PeLS v Figure 3-4. Response Spectra at F4US and F4LS for Models B & C (LSST07 Surface Input)
. l l
l
3.0 . .
. . . 3.0 . . . . . . .
- seemeu,- - = . e. u *** i s
- .,4- cw ga,=w w.mer
.mes si u e.=in
. .ar si
. i .e . .e
& 5 h l.4 h .0 5
- d w d w
E d
.0 Y.. '*'
j
$ 0 #
.~/g-I .. e e E ',
/. < g N~ 'N ._
J, - 10 40
- 10
- 10 8 80-8 le 8 it
- 10 8 Pet M utt-C*S PLWSM = tN00 Lllit? F.5CW wt FwSM ettn00 p et M cC blite
' *C ol , F.tCL as 3.0 .
1.9
--artmets or se oneses - str-n et.paer t6we murspo,a0066
' ee g
- p 64. g1seg ise0(6 CD 64 passies 3 s I 5
- i.: :. .e 5 5 t 5 9 ..
d 1 d l' w w (s .
..- - - - ,/
.e, N'
.4
./. .1 a t-e gg a gg a gg 3 . ,1g-e :q e gg a gg a rws wr= FD($VEMCT*C#$
oite , e.uu .. rws wre F W M ECT.Cpt uite , r ua c.
- ==
.. - .c
--..e..... - ,- .. s. .e.,
. ,e, ..-i i i 5 5 h i.: :
5 a .. l c 2 i V .. - V .. < I ;
! )
E .. .. J _ ., . . : E ., ' l - i .-- p a a ...* - ;. 1
~ . . . . . . . : . . ., ..,
./ ...
l Pte M " "C .0 ... rw. =re . . . .v. , w = r. Fet M cC 9uit *C DS ,
. . us .
Figura 3-6 D'. cue Spectra a
'.5CU and F4SCL for Model C (LSST07 Sur' e Input) t i
i i i
Unit: CH/SEC2 gal-0 4', pgg.g gg
~ .-
I
= ; ::: ,v w - =.,iW 0 4 0 = 9 -i d 96l-8 tg i
.o - -
W A _ z t -- -_: -
- A _- _ :-:
~0 C=
848-B 9 FAl-l T O- *'
.O - -^ ^-' ^ - - - ~ - - ~
V-
. .en **
-. e
~"*- ',.
. i. . i. . A. s. 4. 4. . 4 .
- i. . a. A.
- 4. A. .:
.. v . w s r:o men ( -nar av scit e.in- e a wa 4. . t a wn ... e r. ' e m e.i .* e i . .' o -
F:;ure 3-8. Response Acceleration Time Histories at FAl-1 for Models 8 & C (LSST07 Surface Input) 1 4 3
i l
= F,5::"J,L"*.'. . . n " " " ' "
r f"t"l. M ,..n " ""'" l A
. .. e ..
g ,
! ((l 1 E.
e .i t :. a f 2 3 \\ i ,, $/ S i g .. -f "- K .. F '\ 1... ... ... ... 1. . i.l . ... i.. FAC M uC'* CPS FLWla M f aC0 Fet M uC'*CPt 6551 5 F4WS kl FLWS" aff e Lllfl6 F469 al r ".!""'.L*""" , . . n
" ""'" r f'.'t!l,L"E., . . n " ""'"
8 8
- a a l'LJ !$ ..
D 3 j' P 5 v ., / _. y ,, A i , r c. t
! j +
2 . u u \ .- .. K .. -- I ., ' ' '
..- see i. i., 1.T i. ii i..
F.f D t uCf CP. F.t M uCT-CPt FLWln ottwDO Lllf88 F4VS Em PLUlm PtfdOO Lllfl5 F465 tu i.. . . .-. ...
= '"t"!,' r?S., . . n " **'" r :"2 " l,3 r5 5.,. . o " *-*'"
S S 8 7 E e
.. i ..
a ' 3 3 ' i .. A i .. A [ ,, ,N
.\ [ ,, V- \
1.-. ... ... ... 1. . ... ... ... F.t M E T+ CPS F at M mC T CP. FLula MTM00 L55788 #8WE v FLUSm sattuoD L55786 Fas t v Figure 3-10. Response Spectra at F4US and F4LS for Models B & C (LSST16 Surface Input)
w .nspis '
~
- agamm'.gves' immeb ti
~
g , ,g,
- 6. mig 8 5 j i.e j .. n.j s 3 s -
s ".' \ /, ;
- I fu . t 3
b y u Ia ' I' f s wsx
^ ..gg'.
e.. gg . gg . 3. . g.. g.. gg.
. Col f MM ut t-C*l Flutes M? 00 FateutuC'tle Lll F. law al Pwla att.s00 Llltal F.lG6 al 3,. . . . . . . . , . . . . ..
4.. . . . . . . . . . ,
- Fuen artes is.mts se as saaeses w 7 77 'aee aee t Hwu l
i , I i i i i i.' . i' e s ii e p
! .. ! .. b a.
,, p Y \, .
Ny '?f .~ s .-. I.* 8. i 68 8 .** F at M ut '
- C #1
- 8. ' t, 8 n. -
F.C M utT-Cpl FWSee nt?n00 bli?34 F.lGu (m Falas MtiMOD Lll,46 7.l:4 tu
- = .r, .... . . .p ..
- m ,y . _., ,,
i. k k \ s a
=.. =
8 2 1 .. V .. h V .. i ; I b 1 % t ', ^ < I ', nd
/ #~ W
/~
Q ,. l e uw,-c r wo =? no a.ssu . en M.C.c ! Lliti. ews =? oo 6tst.: e.ics . Figure 3-12. Response Spectra at F45GU and F45GL for Model C (LSST16 5.;.t' ace Input) I
l i Unit: CP1/SEC2
. . r . ...
. .....s
=4
=w w-8 41 9 - 8 J
. FAl-4 fW o ;. . ; [ u- h t "i ,dow i
- m. . . . .
one s . un ., , e
.c ," ,% 7. i^- M 1- _w 1
~"*
I, . '.. i. . A. A. A. 4. 4. A .. '
...A. A.
.. .m . A. 4. 4. 4
. . .e ,su .
S L ULH NE IMO9 tt39t t - RW AV .SOI L L L',816 (PR$ PHGF-4 RICOR0f D R[$f'. fINF HISI . .F Al- 1.tS$f 9 6 Figure 3-14. Response Acceleration Time Histories at FAl-1 for Models B & C (LSST16 Surface Input) 4
ENCLOSURE 3' COMPUTATIONAL PARAMETERS AND DIEIR LIMITATIONS l l Introduction ! i I The following paragraphs describe the enveloping, broadening, and smoothing ' 4 i
- proceduree that have been followed to obtain final floor response spectra i
! (FRS) for Oyster Creek project. This is in response to ites 3 of NRC memo (Ref. Memo from Mr. A. Droserick (NRC) to Mr. P. B. Fiedler (CPU), dated
)
December 16, 1987). Enveloping Procedure 1 Seismic responses for the Oyster Creek reactor building were obtained by 1 three soil-structure interaction (SSI) analyms assuming three different l soil profiles - best estimate, upper bound, und lower bound. Raw floor acceleration response spectra we re computed from each of the three I 1 I analyses. The three raw spectra at each selected location were then ' enveloped to obtain the final raw spectra. Figure 1 shows the three raw horizontal spectra at the baseast level (elevation -19'-6") in the reactor building. Figure 2 presents the enveloped raw spectra. Broadening /Saoothing Procedure i The spectral peaks of the enveloped raw spectra were broadened *15% at fixed-base structural frequencies following USNRC Reg Guide 1.122. This broadening was done to account for any variations in structural f requencies due to any possible uncertainties in building structural properties. The , soil-structure peaks were enveloped and not broadened, since variations in soil properties were accounted for in the analysis by considering the best estimate, upper bound, and lower bound soil shear modulus cases. Figure 3 l 2 4
, presents the broadened /ssooth horizontal spectra at baseast elevation. l l For the purposes of peak broadening, the fixed base structural frequencies
- were obtained f rom the eigenvalue analysis of the model. Table 1 presents the frequencies for the reactor building model. The lowest horizontal fre-quency, for example, is 5.4 Hz (0.184 see). In our broadening procedure, all the s tructural peaks at this frequency or at higher frequencies were
~I-23 %
broadened il5%. Peaks at lower frequencies are primarily due to soil effects and were enveloped from the three soil profile cases. Figures 4, 5, and 6 present individual raw, enveloped raw, and smooth / broadened final vertical spectra at a typical slab location (eleva-tion 51'-3") in the ructor building. j Conclusion The enveloping, peak broadening, and smoothing procedures, adopted for obtaining the final floor response spectra (FRS) for Oyster Creek reactor building, have rigorously accounted for any peak variations in the FRS that , may occur due to possible variations in the soil and structure properties used in the analyses. Soil property variations are considered by envelop-ing results from three sets of soil properties. Structural property varia-tions are accounted for by broadening structural peaks by *15%. e i 1 I l i l ' l ym i 1 J
TABLE 1 FREQUENCIES AND H0DAL PARTICIPATION FACTORS FOR THE FIXED-8ASE REACTOR BUILDING MODEL Mode Freq. Period No. (CPS) (Sec) X-Dir. Y-Dir. Z-Dir. 1 5.42 0.184 33.6 -0.4 -2.0 2 5.45 0.183 0.0 -0.5 31.8 3 6.22 0.161 27.1 -0.5 -8.1 4 6.27 0.159 -8.4 0.9 -27.9 5 7.32 0.136 -0.0 0.0 0.0 6 8.05 0.124 17.6 0. 2 0. 8 7 8.11 0.123 0.1 0.3 -17.9 8 9.47 0.106 -3.1 -0.1 8.2 9 12.52 0.080 0.0 0.4 0.0 10 14.39 0.069 -2.8 -1.7 -6.4 11 14.40 0.069 -6.2 -1.1 2.7 12 14.82 0.067 0.2 -0.0 -0.8 13 15.41 0.065 -0.0 -0.0 -0.0 14 15.58 0.064 2.9 48.7 2.2 15 18.46 0.054 1.8 -4.8 1.6 16 18.64 0.054 3.1 -0.7 -0.0 17 18.71 0.053 -0.1 -0.6 6.1 18 19.90 0.050 -0.0 -3.1 22.1 19 20.63 0.048 -21.8 5.7 0.1 20 23.06 0.043 1.3 -0.5 -0.0 21 23.06 0.043 0.0 -0.2 1.5 , 22 25.51 0.039 -3.5 -0.6 -0.9 23 28.82 0.035 -15.0 -1.2 3.3 24 28.97 0.035 3.2 1.8 13.7 25 29.66 0.034 -1.2 -0.1 -0.0 26 29.67 0.034 0.4 0.0 0.2 27 31.12 0.032 8.3 0.3 -1.2 28 31.23 0.032 1.6 2.3 7. 8 i 29 32.09 0.031 0.6 7.3 -2.8 ' 30 36.90 0.027 -0.3 7.5 -0.2 ' i X = North-south direction Z = East-west direction Y = Vertical direction l l l l l l Yb I
OYSTER CREEK - HORIZ. SSE 2% SPECTRA 0.5 NODE 199. REACTOR BLDG. (EL -19'-6") 4 0.4 - I
/ }![. . /s.
\'
a I b' \.sO' '\ s h f . 0.3 - o [! o N.
/
. g s I !l) /i p'
h~ %-
' m .~. ~ . .
ll N _N y f g
~
f %
~
o 0.2 - s h / N - 3 ,ff\/ \ t,, b_/
' O.1 -
I O O.2 0.4 0.6 0.8 y PERIOG(SEC) usr t.snmie son. " 80"- g' umEm some sori.
- g EI RAW ll0RIZONTAL SSE SPEN AT BASENAT !.EVEL OF lhE REACTOR Suggagg
OYSTER CREEK - HORIZ. SSE 2% SPECTRA 0.5 , NODE 199, REACTOR BLDG. (EL -19*-6") 0.4 - o
)
E O.3 - o
. n O.2 - '
l}{ O.1 O
' 8 s ' '
O ' s 0.2 0.4 0.6 0.8 PERKXXSEC) nmRE 2 l RAW MEMPED mRIZONTAL SSE SPECTRIM AT WNT LEVEL OF THE Am SUILDING
OYSTER CREEK - HORIZ. SSE 2% SPECTRA 0.5 NODE 199, REACTOR BLDG. (EL -19*-6") 0.4 - 8 7 y 0.3 - o
. s hs 8
4: 0.2 - I 3 0.1 - 4
- O i , , , , , , , ,
O O.2 0.4 0.6 0.8 1 l PERIOD (SEC)
; FIGURE 3 SHOGnl/3ROADENED llORIZONTAL SSE SPECTRIM AT BASENAT Ili/EL OF 118E REACTOR SUILDING i
_ _ _ _ _ _ _ _ _ - - - _ _ - . - - - - - ~ ~ ~
Y OYSTER CREEK - SSE VERT. 2% SPECTRA 0.5 NODE 283, REACTOR BU)G. (EL s1'-3") i I
. )
T 0.4 - l [\ \
') N I
B L
- y 0.3 -
; i l O -,
g\
- 3 I l s 0 0.2 -
l
.I 0.1 -
_bh))! 0 , , , 0 0.2 0.4 0.6 0.8 1 PERIOD (SEC)
- BEST ESTIETE SOIL --- UFPER SOUND SOIL - MhfEA 90016 SOIL FIGURE 4 RAW VERTICAL SSE SPECTRA FOR SIAS PANEL SETWEEN COLIMN LINES RA/RS AND R3/R4, EL. 51'-3" 0F THE REACTOR SUILDING
OYSTER CREEK - VERT. SSE 2% SPECTRA NODE 283, REACTOR BLDG. (EL51'-3") 0.5 f I l O.4 - O y 0.3 - o
- B W
i ' N 8 0.2 - 1 s 0.1 - i O : , i : i O O.2 0.4 0.6 0.8 1 PERIOD (SEC) FIGURE 6 SM00111/ AND R3/R4,BROADENED VERTICAL EL. 51'-3" 0F Tile REACTOR SSE SPECTRtfr4 FOR SLAB PANEL BETWEEN CollDIN LINES RA/ BUILDING
OYSTER CREEK - SSE VERT. 2% SPECTRA NODE 283, REACTOR BLDG. (EL 51'-3") { O.4 - F O.3 -
- s a
8 0.2 - 0.1 - i l 3 3 3 I I i i O : O.2 0.4 0.6 0.8 1 PERIOD (SEC) FIGURE 5 l' RAW ENVEIDPED R3/R4, VERTICAL EL. 51*-3" OF SSE SPECTRIM lilE REACTOR SUIISING FOR SIAS PANEL BETWEEN COLIMN LINES RA/RB AND I
ENCLOSNRE 5
.) - ,
l l EFFECT OF SATURATED S0IL ON OYSTER CREEK $SI ANALYSIS i i In:roduction l The following paragraphs describe the soil saturation condition at the Oyster Creek ' reactor building area and how the saturation effects were con-
~
sidered by URS/Blume in the reactor building SSI analyses. i Conditior at Oyster Creek Site Oyster Creek reactor hailding is embedded to elevation -29 f t 6 in. The ground surface is at elevation +23 ft 6 in. The water table varies between elevation +3 f t 6 in. to +10 f t as monitored in the well network at the site since 1984 (GPU Memo, 1987). Figure 1 presents the SSI model of the reactor building indicating water table elevations. Procedure Adopted in Oyster Creek SSI Analysis to Consider Soil Saturstion Effects We have adopted the fo1161ng procedures to account for soil saturation in SSI analyses for the Gyo:ar Creek plant. We have used field-measured soil properties at the Oyster Creek site, which include effects of soil satura-tion as manifested in higher P-vave velocities and Poisson's ratio for layers under the water table (Figure 2). In addition, we have monitored P-vave velocities in soil layers in the saturated soil zone of our soil-st ructure model (Figure 1) during SSI analysis and imposed the condition I that the P-vave velocities in these layers will not be less than the P-vave velocity of water. This has been implemented by adjusting the soil j Poisson's ratio of thase layers using the following equationt [1-2(Ve /V p )2] i 4 y , 2(1-(Vs /V p )2] j 1 l where j u = Poisson's ratio to be used for the soil layer in , saturated zone ' V, = Soil layer shear wave velocity l V, = P-vave velocity in water 0E@/Blume 1
4 This satisfies the measured field data which indicates P-wave velocity of saturated soil to be close to water P-wave velocity. It is also bonsistent with theoretical studies of wave propagation in saturated soil medium. Finally, this procedure has been used in SSI analyses of a 1/4-scale con-tainment model embedded in saturated soil and produces good correlation with responser measured in actual earthquakes. These items are discussed in more detai. in subsequent sections. Field Data on Saturated Soil Properties A large set of in-situ measurements of saturated soil properties using down-hole and cross-hole techniques have been accumulated by Japanese in-vestigators Yokota (1982), Ishihara (1971), Isai (1970), and others. It has been observed in the field surveys that in relatively loose saturated soils, the observed P-vave velocity is close to or slightly higher than the ) P-wave velocity in water (4,700 fps). In stiffer saturated soils, the P-wave velocity seems to be controlled by the stiffness of the soil frame; i.e., the P-wave travels mainly in the polid phase of the medium at veloc-ities higher than 4,700 fps. It can be inferred from these data that water ' is the dominant P-wave-transmitting medium in relatively loose saturated soils (i.e., when solid phase P-wave velocity is less than water P-wave velocity). On the other hand, the solid phasa (soil frame) becomes the dominant F-vave-transmitting medium in stif f saturated soils. Similar trends are observed in the field measurements of soil properties reported for the Oyster Creek site (Figure 2). It shows that the P-wave velocity changes from 1,400 fps for the top unsaturated or partially sat-urated layer to 5,200 fps for the lower fully saturated layer. The Poisson's ratio of the top layer is derived to be 0.39, while that for the second layer is 0.48, reflecting noil saturation effect on P-vave velocity. f l Results from Theoretical Studies The basic theoretical research on wave propagation in a saturated medium was conducted by Biot (1956 and 1962). It was assumed that the saturated soil may be represented by a two phase syster. consisting of an elastic skeleton uniformly distributsJ in space and compressive fluid filling the pores of the skeleton. Soil permeability ef f ects couple the two phases ES/Blume
together. Biot showed that three body waves can propagate in such a medium i
-- two P-waves and an S-wave. The P-wave of the first kind is' identified to be. a pure compressional wave. The second kind decays rapidly due to a diffusien-type process.
i J Studies by Biot and others have concluded that S-vaves are hardly influ-4 enced by saturation. The P-waves are af fected in loose saturated soils, which is consistent with field observations. Validation of Procedure in Actual SSI Experiments ' Electric Power Research Institute (EPRI) is conducting a research program , to nessure SSI ef fects in a 1/4-scale model of a reactor containment build-ing subjected to actual earthquakes (Tang, 1987). This 1/4-scale model is set up at 1.otung in Taiwan, in an area of high seismicity. To date, j 18 earthquakes with Richter magnitudes f rois 4.5 to 7.0 have been recorded since the seismic instrumentation was put into operation. Two of these ! earthquakes of magnitudes 6.5 and 7.0, with peak ground acceleration of l 0.20g at free field at the site, were used for extensive analyses using ; various SSI codes. I Lotung site has a deep sandy soil deposit with shear wave velocity varying ) f rom 300 fps to 1,000 fps in the top 200-f t soil deposit. The site has , saturated soils, as groundwater level is measured at 3-ft depth. The coa-pression wave velocity asasurements reflect the saturation effect with f almost constant F-wave velocity of 4,500 f t/see from 20 ft down, with a transition zone near the surface. The effect of soil saturation is re-
- flected in Poisson's ratios which are calculated to M in the 0.45 to 0.49 range.
I Several studies were conducted to pradict the structural responses of the containment model due to the two earthquakes, using s t a t e-o f-t he-a rt SSI j methodologies including the PLUSM group of programs. In one of the the j a studies (Tseng, 1987) with the F1.USH program, the Poisson's ratio was ad- I ) justed, similar to what is done by URS/Blume for Oyster Creek analyses, to ] keep the P-wave velocity to a minimum of 4.790 ft/sec. The analyses ) I
@ s2/ M
results show very good correlation with measured data as presented in Fig-ures 3 and 4.
- Another study (Berger, 1987) reported using ALUSH, an axisyneetric version of FLUSH, for the SSI analysis. In this study also, Poisson's ratio was adjusted to maintain P-wave velocity in soil equal to that of water. The correlation with measured data is again very good, as shown in Figures 5 and 6.
Conclusions Based on the above discussions, the following conclusions may be drawn for Oyster Creek SSI analyses:
- The soil saturation issue has been addressed in the state-of-the-art SSI method for Oyster Creek reactor building SSI analyses.
e The applied methodology is consistent with observed P-wave velocity data from field measurements. e Application of the same methodology in the EPRI-sponsored Lotung project has yielded analytical re-i sults showing good correlation with sessured re-sponses in a 1/4-scale reactor containment model situated in a saturated soil site subject to actual earthquakes. ! References Berger, E., H. Fiers, and D. Kluge (1987), "Predictive Response Computa-tions for Vibration Tests and Earthquake of May 29, 1986, Using an Axisyn- ! I metric Finite Element Fornualtion Based on the Complex Response Method and j Comparison with Measurements--A Swiss Contribution," EPRI/NRC/TPC Lotung SSI Workshop, Palo Alto. ' Biot, M. A. (1965), "Theory of Propagation of Elastic Waves in a Fluid- ! Saturated Porous Solid," J. Acoustical Society of America, Vol. 28(2). Biot, M. A. (1962), "Mechanics of Deformatios and Acoustic Propagation in j Porous Media." J. Appl. Physics, Vol. 33(4). l 1 1 J ES/Blume '
]
GPU Memo (1987), "Transmittal of Groundwater Data for Oyster Creek Site," October 19. - Isai, T., et al. (1970), "Velocities of Elastic Wave Propagation and ' Mechanical Characteristics of Sof t Soil Grounds," Tsuchi to Kiso, Vol. 18, No. 1, pp 17-22. Ishihara, K. (1971), "On Longitudinal Wave Velocity and Poisson's Ratio in Saturated Solids," Proc. Pourth Asian Regional Conf. on Soil Mech. and l Found. Engineering," Vol. 1, pp. 197-201, Bangkok, Thailand. l Tang, H. T. (1987), "1.a rge-s cale Soil-structure Interaction," NP-5513-SR, EPRI, Palo Alto, California. Tseng. W. S., et al. (1987), "Soil-structure Interaction Analyses of Quarter-Scale Containment Model Experiment in Lotung, Taiwan," EPP.1/NJRC/ TPC Lotung SSI Workshop, Palo Alto. Yokota, K., et al. (1982), "Geotechnical and Geophysical Methods for Evalu-ating Dynamic Soil Properties at Various Sites in Japan," Proc. third In-l ternational Conference on Microtonation, pp. 1129-1144, Seattle, U.S.A. l l 1 l
)
l l I l l I gpggggg l l ! l
l l l r 7 Lhe af Synynsky 31igr.g. L no ,% q i p Drywel - El 107 01 WP Reamer Pressure ye: Set B 97 S* , ,~ B tr 9* . . _ (rSNewWas
, (
t i q t ' Seheffwls of i '
, [', 4 , LWerTer Steet t , d M '
gehem g EL. *23'.4" _fJ
, s etr.,sieet
! y, . 900
. IL. +10'.0" Vs e
'I
; . s n. 3. 6 - ;600 IrATIA 12 VEL h
- i i
l Yp a 5600 - Vs e 1200 t i Yp e SD00 - Vs e 1400
+ + + + + + + + + -
4 < L ..ir. e a o g' g 9/ao . Superflush SSI Model, Oyster Creek Reactor Building . Figure 1. i L J f l
L C r I w g . h bb W WW w y .. e h U. .* ". S * --**
=N $
8 #E
) .'b[I !.. .$) * " . ,' '.hf 53 r,."*,O;f 's. -. '.S..7. '. .[^ ' +o
):: f. ,'.' h; ,h W -
L> aj . W. Q . !.9::#..c:.
. flb- ;* .
.l..:
3 .
- q. .;., f. ' :_ : r '.+
Ww i
. + : .'.y.-.; .*- **..j.g.. .+ ,
.s.- s...
s .. . ~
.o .
1 '
- d.- M l
.,; ./ . .4 [d M . ..
1- . i. + p.y> w.. . + <.
>, + .. ;
s L ..'a'
; g.-i. ..y:
m:5.'n:.. y; M i , '... ,e- . '. 1, :'v , yi .s .s. : : .y a:q.t:s. .n w y
*. ... . .4 hE ..7 -
. : '. .. v ,
.:.:#?.y;;i.
- f.
,%'.h.,
. h' *. *.* t, ' .~
*.. g 'f.M.i.v
; ,e.".'
- /h',. ;', :jQgy: .Wuh. i)>.T . N
- k. y;.ii,(:, o W
- '; a,.;.i.: -Xss.v.-
t cr:.... s . . : < . .r...:2 3 :.v y
. x
%^ _
1 H E u .j
, E b
$ b c
e g g
, il i g i II I I E -
3
=*
B a
., a R 5 3 3 E d 4 4 d E ..
~
N , s I g ! c N !
> s h m O 4 *
,8 E SEc_: c 3 h I G. ,
> 4 4
4 LEP/Blume
i
*"~'*
= l'55t.M. . . .. = fl'5Mr.3r?E.. . . s. " ""'"
k 3
- i. I ..
li a v ,, , o
.. )
.. y n_
l l 1.-. ... ... ... 1. . ... ... ... !
,. "" " ' M. , .. .. c " " " ' @. 6 , . . . ..
l
- a. Top of Structure, N/S Direction d. Basemat of Structure, N/S Direction l
" **'" " *~'"
= ! J.Ml 3r"&. . . i. = !"5Mr.3r% . . ..
. .. . .. l k k \
1 l
- l. .
/t..
- l. .
l . l <
}
I.-
. _ _ t .. _
1.-- ... ... ... 1. . . ... ...
, . = , d
'*""' i;'., .i. ., "" " ' @. 6 , ,e i.
- b. Top of Structure, E/W Direction e. Basemat of Structure, E/W Direction l
=l'E"t.3' E .... = PJE'sM . . ..
J k i
.. i.
I l , I i ' 1. f .
. . 1.-.
f . . .. l
.. ""*IGb .. c. ., "'""*'3,., ,e . 1
- c. Top of Structure, Vertical Direction f. Basemat of Structure, Vertical Direction Figure 3. Comparison of Hessured and FLUSH Computed Response Spectra at the l Top and Basemat of the 1/4 Scale Containment Model, Lotung Study !
! ( May 20, 1986 Earthquake, 0.20g ZPA ) ' ) i
.4 , . . ..-
i..
- rwup ir*e. amma is .e .mapeas g g,,y, ,,,t , t , O me l
... .4 b i
\
1 .. i .. .
' l 1
3 P's r s d w 4%
. ,. ... i.. ... .:... ...
,. =, ""JiiiP . . . , . = , ****uY;'i. ..m..
- a. Top of Structure, N/S Direction d. Basemat of Structure, N/S Direction i
'd pg,g
~
- w .r, ei u wi '
u -i
.... .a i f, i i !!
... .I .
l l\ l1 l.
- g. {\
- l. .
A f, y v. k d h ' u.g (
.r. ... i.. ...
ir. .
,w . ""Jii!P .. . i. . = , *** *G'?i . ,.m e.
- b. Top of Structure, E/W Direction e. Basemat of Structure, E/W Direction
'" ~
'd w ir A .i .' = gg , '
- = :
I.
.. i. .
I .. s I .. [ g. 1 i h la
- _. i O !
y go - u t .
. .r. . . ...
rw = tie.*"EiE , .m, . ,w = ,.e. " "
- I Elie e.m ,
- c. Top of Structure, Vertical Direction f. Basemat of Structure, Vertical Direction i
i Figure 4. Comparison of Haasured and F1.USH Computed Response Spectra at the , Top and Basemat of the 1/4 Scale Containment Model, Lotung tydv ' ( Nov. 14, 1986 Earthquake ) MMMUf116 l
- s. N/S -
1
- 1. <
,.?.
\
- t. ,, .! \,
.a t '.
.*' / \.,
.i 4
- b. E/W 4 4
r, j'-
'.s.
j \g t. d I
- c. Vertical ' -
1 1 4 'I j .- , j
- t. =. .
a 1
< i l
.,vg
- ~..
\ .
j r.......,t=1 4 1 Figure 5. Nessured ( ) and Predicted ( - -) Acceleration Response Spectra l on Top of Structure ( May 20. 1986 Earthquah: ) l UM/Blume l i i i
- a. N/S . ,
t
~
- =* ,
a*.
,.s ' .,
. ....'~ -- ~. - ... .......
s
- b. E/W ,
1 I 4.. . t W -..__
- c. Vortical ,
i
~8 4
- =. ,
1 l i - a v. i = i . e. . *
- 2 re es...e o ins 1 P
4 Figure 6. Measured ( ) and Predicted (....) Acceleration Response Spectra at Basonat of Structure ( May 20, 1930 Earthquake ) ' NI M 1 l _,___1
1144 OCTOBE31977 GT10 GT10 SEtSANC EFFECTS sophisticated forms of half-space analysis which introduct the essential concepts in the Refueling Buildi and' " * + 32) Some 330 I of a complete interaction approach are developed, it seems that both methods it (100 m) south of the B
- of analysis may ukimately deveksp to the point in which they give similar resuks The June 7,1973 cartefuel)g for embedded structures at comparable costs. had its epicenter a WE 5 m the RiclWer scale)
EI en s south of the plant site and triggered strong A major factor contributing to the continuing debate concerning the merits of any forrn of analytical approach has been the total absence of recorded -
' ~
f.cid performance by which the adequacy of such an approach might be w_ __ _ judged-making et necessary for engineers to adopt one approach or the other on the basis of slicir personal appraisals of factors swh as the degree of sophistication of the analysis, t*ne potential savings in design costs, the potential +-
;'j "-
"" 'F h "- . ____
losses in overall project costs, their degree of understanding of the nature of
.s. ,_f the phenomena and principles involved, etc. In the absence of known fickt !*-,= =
.u ?*'-
1 performance. all reasonable suggestions for design approaches must be considered
- , ( -
potentially applicable, and considerable insight, wisdom, and intellectual honesty #
}
h"
- w .. _
4- *- -
*~ . ,5 , r> ' l - -
Y .
**~
W
)*****"....
-(^ .=,- --=
^ ^~
} ) *
- m. ,
/ s*,Nl W g -
l g l 7o ,
,.e / NA s
. ,, . ~
an , M # )gs",.%
- m_ --
N / N , y m === m s . N i
'N
'N s
\ ('/ 4- fr -
'NPs s/ 5_ , ' ~
t i l Bm- e. 1 J 'h
)
'So- * -
1 RG.1.-Genersi Vleur e G tlamibeidt power Plant (After Sectiest Co.)
~
.rs_ . . "
are required to select a design metimmi that offers the greatest potential for *"** **"* ( <== ' combining adequate safety for critical structures with reasonable overall economy a_ ! ** l in the cost of the completed facibly. It is for these reasons that the motions recorded at the Humboldt Bay Power Plant in the Ferndaic carthquake of June m_
* - ', ( .
7.1973 are of major sigmficance. ** - , p A general view of the plant is shown in Fig. I. Units I and 2 are fossil ami . i 4 I fuel unste whereas Unit 3 is nuclear. The buried reactor structure within the
. . . . ; ; ; ,, ; ; ,h Refuchns Buildme of Unit 3 consists of a massive concrete caisson embedded at a depth of aperou 85 f (26 m) belowIhe ground surface.The various surrounding FIG. 2.-Geo id asetsen fleceeds at h. set % h %1 h = SJg5 m3 strisctures are hghtweight structures and are founded at or close to the ground surface. The facility was constructed in 1%3 and has been operating satisfactorify motion instruments in the "" a
. ng those located at the since that time. plumholdt Ba Pla a
Sirong motion inssruments at the plant have been in operation since Septendier. usnboldt Plant f are shown in Fi te. 2. Althou 1971. These are located at elevation + 12 (plant grade level) and clevation -66 3 sec-5 sec the baseline-corrected pea ac ion de
ENCLOSURE 6 TORSIONAL EFFECTS ON STRUCTURE Introduction The following paragraphs describe the study done by URS/Blume to assess the torsional effects of Oyster Creek reactor building mathematical model on the computed horizontal floor response spectra (FRS).
Background
A lumped-mass stick model (Figure 1) was develope.d to compute seismic FRS for Oyster Creek reactor building. The masses of a floor and of the top and bottoa valls tributary to the floor are lumped at that floor. Ihe ver-tical beams between the floors represent the lateral stiffness of the walls between the floors. Mass and stiffness calculations for the reactor build-2 ing has shown that the centers of mass at each floor and the centers of rigidity between the floors do not coincide. This is due to some asyssetry in mass and stiffness distribution in the reactor building. However, the eccentricities between the centers of masses and rigidities in the Oyster Creek reactor building are small, and frequency calculations showed that the mass participation in torsional modes are insignificant compared to the horizontal modes. In addition, the building dynamic prop-erties in the east-west and north-south directions are very similar. The computed FRS in the east-west and north-south directions f rom the mathemat-ical model of Figure 1 were very close. As a result, it was determined that a 2-dimensional representation of the reactor building model will be able to capture accurately the seismic responses. Such a model was used for SSI calculations and FRS generation. Horizontal FRS computed from such a model does not include the horizontal component that may occur due to torsion of the floor mass about the center of the rigidity due to eccen-tricities between the centers of acss and rigidity. The NRC staf f re-quested a study to demonstrate that the horizontal FRS component due to torsion is insignificant as assumed in the Oyster Creek reactor building FRS generation. , bdMN l
Parametric Study-To study the effects of the torsional component on horisontal FRS, two com-puter runs were made with the lumped-mass mathematical model of the reactor building (Figure 1). The first computer run computed the total horizontal FRS at an extreme corner point on the operating floor elab (E1. 119 ft 3 in.) from the slab center line (70 ft in west and 70 f t in south direc-tion) considering both translation and torsion of the structure. _This is the worst case slab location for torsion. The second run did not consider any contribution to the horizontal FRS due to torsion of the structure. The resultant FRS for 5% damping are presented in Figure 2. It shows insignificant (5% or less) contribution of torsion on the horizontal FRS of Oyster Creek reactor building. Conclusion The torsional effects, due to mass and stiffness asymmetry of Oyster Creek reactor building, have insignif t: ant impact on total horizontal FRS gener-ated for the building. U l l
I
}
M64CTO '
- u. se. ,9ww'A
(>9
@ I 04.Isele* <,. l (L.fi9Ia* # Node 7: Center of mass at E1. +119' 3" Node 17:
/ gpf 'II
/, Center of rigidity between E1. +93'.3" and El. +119'.3" Node 27: Center of rigidity between El. +119'.3" and E1. +134'.0"
@ Node 20: Extrese comer of the operating floor
#6 94' 8* sisb. 70' to the west and 70' to the south of slab conter line
/ + 1/
J1' I GL .14 '= 2 " ottl
- 94. tiis o y,'e+
l4 SL.ts!s'
/.4/ >
po r
,e O
CL. 19te" , rir w r, y o j -M(M/t) '
* (t/W) i Figure 1 3-D Lumped-mass Model of Reactor Building
@@Mf716
OYSTER CREEK SSE 5% HOR. SPECTRA,N-S T-H EL 119'-3" 1 0.9 - l 4 0.8 - 0.7 - o y 0.6 - 0 1 0.5 - s N O.4 - 2 ' O.3 - f 0.2 - 0.1 - O i 3 3 ' ' O O.2 0.4 M h PERIOD (SEC) WITHOUT TORSION ORSION CONSIDERED (NODE 20) g .... -.-.-~.. - ~ -}}