ML13330A031
| ML13330A031 | |
| Person / Time | |
|---|---|
| Site: | San Onofre  |
| Issue date: | 04/11/1980 |
| From: | Baskin K Southern California Edison Co |
| To: | Ziemann D Office of Nuclear Reactor Regulation |
| References | |
| TASK-03-06, TASK-3-6, TASK-RR NUDOCS 8004220503 | |
| Download: ML13330A031 (36) | |
Text
Southern California Edison Company A
P. 0. BOX 800 2244 WALNUT GROVE AVENUE ROSEMEAD. CALIFORNIA 91770 K. P. BASKIN TELEPHONE MANAGER, NUCLEAR ENGINEERING (213) 572-1401 AND LICENSING April 11, 1980 Director of Nuclear Reactor Regulation Attention:
Mr. D. L. Ziemann, Chief Operating Reactors Branch No. 2 Division of Operating Reactors U. S. Nuclear Regulatory Commission Washington, D. C. 20555 Gentlemen:
Subject:
Docket No. 50-206 Seismic Backfit Project and Site Specific Earthquake Program San Onofre Nuclear Generating Station Unit 1 On November 14, 1979 we met with the NRC Staff and NRC consultants to discuss the results of their review of SCE's Seismic Backfit Project as documented in our sub mittal of April 29, 1977.
As a result of that meeting seven open items were identified. Responses to these items are provided in Enclosure 1.
At a meeting with the NRC Staff on March 20, 1980, SCE and our consultant TERA Corporation outlined additional work which we have initiated with respect to the Site Specific Earthquake Program. Specifically, the impact of strong motion data from the 1979 Coyote Lake and 1979 Imperial Valley earthquakes on the site specific computer model will be evaluated. A description of SCE's program for evaluating these earthquakes is provided as Enclosure
- 2. Results from this program are anticipated in July of this year.
If you have any questions on any of the enclosed information please let me know.
Very truly yours, 152A' Enclosures 0o 22 09D3
ENCLOSURE 1 Responses to Open Items From.November 14, 1979 Meeting ITEM 1 Provide information which demonstrates the adequacy of the time-histories to generate realistic seismic response.
Particular emphasis should be given to RCL non-linear analysis results.
RESPONSE
The following approach is taken to show that the time-history components employed in the RCL seismic re-evaluation produce applicable seismic response consistent with the defined Housner ground spectra.
- 1. The..-building DBE seismic deflections associated with the Northeast building corner and the foundation point obtained from time-history analysis and response spectra analysis are compared. It is concluded from this comparison that the deflections of the interior concrete structure obtained from the time-history analysis are consistent with the results from the Housner spectra analysis.
- 2. A comparison of steam generator upper seismic support loads determined from the RCL time-history analysis and elastic non-linear analytical load formulas with sinusoidal input is made.
It is concluded from this comparison that the impact forces produced in the seismic time-history analysis performed are consistent with the forces associated with the defined Housner seismic levels.
Items 1 and 2 are discussed in detail below.
RESPONSE SPECTRA ANALYSIS Response spectra analyses were performed using the coupled building and RCS model with support modifications.
Three NSSS system model cases were considered:
open gap, closed gap and mixed gap.
These cases are discussed in Section 3.7.2.2.2 of the April 29, 1977 submittal.
A 3-D response spectra analysis was performed considering the 0.67g horizontal and 0.44g vertical Housner response spectra.
Seismic deflections are obtained for the operating deck Northeast corner and the foundation point (node 4 and node 1, respectively as shown in Figure 3.7.2-5 of the April 29, 1977 submittal document). A separate analysis was performed for each shock direction (X,
Y, Z).
Two cases of modal displacement combination were considered.
One case was square-root-sum-of
- squares, SRSS.
The other case considered closely spaced mode combinations, CSM.
The significant modes which contribute to nearly 100% of the total building seismic displacements, at nodes 4 and 1 as defined above, are in the frequency range of 3.5 to 4.0 Hz.
A small response is at 3.22 Hz for the open gap case.
The significant building modes are given in Table 1.
For the CSM
- case, the modal displacements are summed absolutely since the significant frequencies are all considered closely spaced in the frequency range 3.5 to 4.0 Hz.
-2 The total 3-D seismic response is obtained by combining common displacements obtained from the analysis of each shock direction by the root mean square method. As given in the April 29, 1977 submittal document, Table 3.7.1.1, the modes corresponding to building soil rocking are considered to have 10% of critical damping; the modes corresponding to building soil vertical translation are considered to have 18% of critical damping.
In Table 2, the response spectra seismic displacements associated with the northeast building corner and foundation are compared to the time-history seismic displacements given in Table 3.7.2-4, page 3.7.2-47 of the April 29,.
1977 submittal document. For the two cases considered, SRSS and CMS, deflec tion ranges are given which are established from the open, closed and mixed
-gap cases.
As seen from this table, the time-history displacements are comparable to the response spectra displacements.
Therefore, the time histories used in the San Onofre Unit 1 reactor coolant loop seismic re evaluation produce. seismic responses consistent with the Housner ground response spectra for the reactor building.
IMPACT FORCE EVALUATION During a seismic event, building motion can approach sine beat as well as sinusoidal motion for short-time segments.
Therefore, it is possible to de termine if impact loads are produced having magnitudes consistent to the defined seismic input levels by comparison of the non-linear impact forces from the time-history analysis to impact forces calculated from elastic non linear analytical load formulas with sinusoidal input.
The upper steam generator seismic stops, described in Section 3.7.2.1.3.4 of the April 29, 1977 submittal document, act in an elastic non-linear manner because of the gaps between the steam generator and supports.
The loads in these supports are used in the impact load evaluation because they are produced from elastic non-linear gap
- behavior, and are at different orientations in the reactor building depending on the location of the steam generator.
Two analytical load formulas are used which are developed using the model shown in Figure 1 and the two assumptions given below.
- 1. The supports do not move and the mass responds in tune with the horizontal building motion (approximately 3.9 Hz).
- 2. The supports respond in a sinusoidal manner and the mass is in resonance with the support motion.
The single mass non-linear model employed in the evaluation of impact load effects is a simple representation of the steam generator dynamic response characteristics at the upper steam generator support.
This model will not yield exact impact loads; however, it will provide a good measure of the general magnitude of the seismic impact loads to be expected from the defined seismic event.
rY9
-3 The development of the expression for the velocity at impact associated with assumption 1 and assumption 2 is given in Figures 2 and 3, respectively.
Impact damping is accounted for by using the coefficient of restitution.
Expressions for the impact velocity after adjustment to account for impact damping are given in Figure 4. The impact force, F, is determined by relating maximum kinetic energy to maximum strain energy. The expression is:
F = V I'MK Where F = impact force V = velocity of mass M after adjustment to account for impact damping K = impact stiffness M = (T /2r) 2K 0
Table 3 shows the steam generator upper support parameters needed to determine the associated impact load.
In Table 4 the associated upper steam generator impact forces for assumption 1 and 2 are given, as well as the impact forces from the time-history analysis.
Also given are the impact loads associated with the average of the assumption 1 and assumption 2 impact loads.
The time history results are close to the average impact loads.
This should De expected since the resonance conditions in assumption 2 represent an upper bound and the conditions defined in assumption 1 represent a lower bound.
The time-history results would be lower than the assumption 2 loads since:
- 1. The reactor building does not respond as a perfect sinusoid during the seismic event.
Its motion can be sinusoidal for short-time segments or sine-beat in character or random.
Motions other than sinusoidal can significantly reduce the impact loads.
- 2. The steam generators do not respond in perfect resonance with the reactor building.
- 3. The reactor building motion does not have a
constant acceleration throughout the seismic event.
If the impact loads developed above, a
maximum acceleration level was used.
- 4. The analytical one-mass model employed will not yield exact results, but does provide a good measure of the general magnitude of expected impact loads.
Noting the above, it is concluded that the synthesized time-histories used in the San Onofre Unit 1 reactor coolant loop reanalysis produce impact loads that are representative of the defined seismic event.
-4 GENERAL CONCLUSIONS The set of synthesized seismic time-history components used to reevaluate the reactor coolant loop, components, and associated supports produce results which are representative of a true seismic event defined by the Housner ground response spectra.
This statement is based on the information previously presented which showed that:
- 1. The time-history analysis of the reactor building results in seismic response levels consistent with those from the Housner response spectra analysis.
- 2. The time-histories produce non-linear impact loads with a magnitude consistent with that expected from the defined Housner event.
ITEM 2 Quantify the influence of the Enclosure Building on the dynamic response of the containment/reactor building considering appropriate soil -
structure interaction effects.
RESPONSE
The steel containment sphere, the concrete internal structures and the surrounding soil media were modeled and reanalyzed during the Seismic Reevaluation Program.
The reanalysis was performed with a 2/3g Housner response spectrum input.
The reanalysis did not include the sphere enclosure building since the program was nearly complete before the design of that building was initiated.
The Axi-symmetric model which was developed for and used in the Seismic Reevaluation Program was, therefore, an uncoupled model.
The design of the sphere enclosure building was performed with a response spectra input equivalent to the 2/3g San Onofre Units 2 and 3 DBE.
The sphere enclosure model was essentially the same model as that used in the Seismic Reevaluation Program with the exception that the enclosure building was added.
The addition of the sphere enclosure building to the original model made possible the investigation of coupling effects.
The first model, which did not include the enclosure building, was designated as the uncoupled model.
The second model was designated as the coupled model since it included the enclosure building as well as the structural elements present in the first (the uncoupled) model.
In order to compare the coupled and uncoupled models, the coupled model was reanalyzed using a 2/3g Housner response spectra.
Points on the primary shield, the operating deck and containment-reactor building foundation which were common to both the coupled and uncoupled models were chosen for comparison.
The comparison of displacements due to dynamic response showed a reduction in the coupled model.
Accelerations in the coupled model also were less than the accelerations experienced in the uncoupled model. SRSS values of displacements and accelerations are tabulated in Tables 5 and 6.
On the basis of the results obtained in the comparative analysis, it is apparent that the addition of the sphere enclosure building produced no increase in dynamic response and that the uncoupled model produced conservative results.
-5 ITEM 3 Provide an example design calculation for both a drilled and grouted anchor bolt installation.
Clarify which anchoring methods were utilized in the modifications.
RESPONSE
Drilled anchors were used only for compression or shear loadings.
An example of this type of anchorage is the knee brace added for vertical support of steam generator E-1C lower seismic stop which was connected to the secondary shield wall by the use of 4-1" 0 Hilti Quick Bolt Wedge type concrete anchors at each support.
Since drilled anchors were not used for tension loadings, detailed calculations were not required.
It was only necessary to select anchor bolt sizes with sufficient capacity to resist the calculated shear or compression loads.
Supports subject to tension loadings were connected to reinforced concrete members using grouted anchors.
An example of this type of anchorage is the anchorage for the lateral restraints attached to the reactor pressure vessel head lifting mechanism.
In this anchorage, tensile loads were resisted by 1-1/41" diameter bolts grouted into 31? diameter core-drilled holes.
The anchorage geometry and the design considerations are outlined in the Calculation Summary below.
CALCULATION
SUMMARY
In order to provide increased lateral support for the reactor pressure vessel head lifting mechanism, additional restraints were installed.
These restraints consisted of stainless steel bars anchored to the surrounding concrete walls.
Anchorage was by means of grouted anchor bolts.
Sketches showing the support details are provided in Figures 5 and 6.
The design of the supports is documented in calculation number SBP-C-10 of Bechtel job 10719-002 and is summarized as follows.
The materials used were 3000 psi concrete and type 304L stainless steel (Fy = 25 ksi).
High strength non-shrink grout was used to grout the bolts into place.
The number of bolts was determined using maximum design loads and shear tension interaction.
The tension forces used for design took into consider ation prying action in the plate. The anchorage of the bolt into the concrete was designed based on the maximum allowable forces in the bolt itself. Extra embedment was provided in order to assure the steel governed the design. The embedment calculations took into consideration concrete shear cone development and the grout bond capacity.
-6 ITEM 4 Provide an example stress calculation including the method of deriving the rebar stress from the finite element analysis results showing how a representative concrete section was either checked or designed.
RESPONSE
A summary of Calculation No.
C.7.8, which was performed during the period November 1974 -
February 1977 to investigate the adequacy of the reinforcement in the reactor building secondary shield walls, is provided below.
The calculation demonstrated that the most critical wall sections were adequate to resist applied loads based on the criteria described in the report "Seismic Reevaluation and Modification," April 29, 1977.
CALCULATION
SUMMARY
A reanalysis was performed in 1974 to determine the structural adequacy of the reactor building secondary shield walls.
The secondary shield walls were originally designed in 1965 under the codes and standards in effect at that time.
The reanalysis was to ascertain if the walls were adequate under the more recent codes and standards.
These calculations; found under calculation number C.7.8, Bechtel job number 1304-899;. are summarized as follows:
The materials used were Grade 40 reinforcing steel (f = 40 ksi) and 3000 psi concrete.
The ultimate strength design method was used in the reanalysis of the walls.
The maximum forces on any point in the walls was determined by computer analysis.
These forces were then checked against the walls' capacity, as determined by the 1971 ACI code.
The walls were checked for both horizontal and vertical out-of-plane bending.
Extreme fiber tension as a result of this bending was also calculated and it was determined that the stress due to bending would not exceed the modulus of rupture.
Both axial compression and tension were also checked.
The axial loads were considered to occur concurrently with the flexural loads as expressed in the computer output.
The sections were then checked to assure in-plane shear and moment forces were adequately resisted.
ITEM 5 Clarify how the stress quantities Pm' PL and Pb were determined.
-7
RESPONSE
In response to this item, the method of stress calculation at the steam generator primary nozzle reactor coolant pipe juncture will be explained. For this portion of the nozzle, strength of materials equations were used for the stress evaluation.
The primary membrance (Pm) stresses caused by internal pressure are calculated with the following equations, p ri
'A 2t p r.
'H t
a = -p/2
- where, GA is axial stress aH is hoop stress a is radial stress p is internal pressure
- r. is inside radius of nozzle at analysis section (nominal) t is thickness of nozzle at analysis section (nominal)
The Pm stress caused by a direct axial load on the nozzle from the pipe is calculated as follows, aA Fx/A
- where, F is direct axial load A is nozzle cross section area
-8 The shear stresses that contribute to Pm are calculated in the following
- manner, Tr T
= V/A +
where T
is shear stress AH V
is direct shear load from pipe T
is direct torsional load from pipe rm is mean radius of nozzle cross section J
is torsional moment of inertia When primary bending (Pb) stresses are needed for checking the PM (PL)
+
limit, one of the terms in the shear stress calculation method is modified as
- follows, Tr T
0
'AH J
where r
is outside radius of nozzle cross section.
- For Pb calculations, the bending stress, caused by bending moments imposed on the nozzle from the attached piping, is calculated with the following equation Mr o
A I
where M is direct bending load from pipe I is second moment of inertia of nozzle section The stresses calculated as above are used to evaluate the nozzle per ASME Section III, NB3200.
ITEM 6 Provide the basis for neglecting the horizontal soil spring in the analysis.
Since the horizontal to vertical stiffness ratio is about one, the horizontal translation response would be expected to be significant.
-9
RESPONSE
The translational soil stiffness that was applied to the foundation center of gravity in the three dimensional finite element analysis was developed after several trial and error iterations, which attempted to match the axisymmetric analysis results as closely as possible for the soil structure interaction modes.
The stiffness, as determined in this manner, was then applied at the foundation node of the detailed 3 dimensional finite element model of the reactor building. The-final stiffness values are given in Table 7.
The detailed reactor building model was reduced to 24 degrees-of-freedom, including six at the foundation node. This reduced model was then transmitted to Westinghouse for use in a three dimensional non-linear coupled analysis of the NSSS.
The equations of motion in the Westinghouse non-linear analysis were solved by direct integration.
During the initial checkout of the Westinghouse non-linear model it was deter mined that deleting the horizontal translational stiffness of the soil at the foundation node (i.e.,
assuming it rigid) would result in a more stable integration.
In view of this, a review was made of the 3 dimensional linear model and it was concluded that this was a valid assumption for the horizontal translational stiffness given in Table
- 7.
Therefore, to facilitate the integration of the equations of motion the horizontal translational stiffness given in Table 7 was deleted from the Westinghouse non-linear model.
The stiffness values listed in Table 7 indicate a horizontal stiffness that is much greater than the vertical stiffness.
By contrast, the relationship between horizontal and vertical stiffness for a flat circular foundation klocated near the ground surface of a uniform subgrade, as described by the equations given in BC-TOP-4A, Rev. 3, imply approximately equal horizontal and vertical stiffness coefficients. These equations may be expressed as follows:
Horizontal K
32
-u) GR x
7-8u 5 GR for u=0.35 4 GR Vertical K = -
z 1-u 6 GR for u=0.35 8 GR3 Rocking K =
ip 3 (1-u) 4 GR3 for u=0.35 The application of the translational coeffients in conjunction with the rocking stiffness is described by Figure 7.
It should be noted that a force applied at the ground surface in any of the Ax, A,A directions will be resisted only by the corresponding stiffness colinea* with the applied force. For the case of an embedded foundation with the geometry of a
-10 spherical segment these equations do not apply directly.
It is for this reason that an iterative solution, based upon a trial and error attempt to match the mode shapes and frequencies of the axisymmetric analysis, was used.
Figure 8 illustrates the relationship between ground surface displacements and the stiffnesses applied at the center of gravity of the foundation at some depth below the ground surface.
For a rigid foundation the following relations hold:
A A'
z z
A =A' - A xH x
x A
A' Based upon these relationships, which were derived from the two coordinate systems shown in Figures 7 and 8, it is possible to obtain a transformation T relating the A' coordinates to the A coordinates as follows:
A = TA' where A
A' z
x A
{A and A' =A>
X X
A A'
and T 1,-H Using the transformation T it is also possible to transform the stiffness matrix in the A' coordinate system to the A coordinate system as follows:
T K = T K' T Performing this transformation on K' where KK X
results in K =K HK '
HK' K' +H2 K' X.
x
-11 K represents the foundation stiffness matrix in the A coordinate system and provides a basis to compare the horizontal translational stiffness due to K' with theoretical values of translational stiffness for flat circular slabs located near the ground surface.
The translational stiffness due to K may be determined by applying a load P in the AX direct at the ground surface.
The load vector in the A coordinate system may be expressed as:
P P
0 Forming the product Kx = P results in the following equations K' A= 0 z
z K'
A
-HK'
=P xx x
and
-HK' A + K' A =H2K' A =0 xx 44 x p Solving the above equations yield displacements in the A coordinate system as follows:
A = 0 z
P + PH2 X
x
/
A = PH/K Expressing the effective translational stiffness at the ground surface as K :
such that Kx x = P results in the following relationships:
P = P +PH2 Kx K'x K' 1 = 1 + H2 or K
K K xK'X K x x
'p Substituting the values for K' and K' from Table 7 results in an effective x
)
horizontal translational stiffness at the ground surface of:
1 =
1
+ (360)2 K
6.4x1O 9 8.8x10 12 x
or 7
Kx = 6.7x10 lb/in
-12 This surface effective stiffness may be compared to the theoretical value for a flat circular slab Kx = 5GR, for G=1500 KsF (see Figure 3.7.2-19) and R=70 feet.
Kx = 5(1500x1OO/144)7Ox12
= 4.4x1O 7 lbs/in The theoretical value of effective translational stiffness at the surface does not include the additional lateral resistance due to embedment.
This is for an embedded foundation.
Kx 5GR + C where C
additional stiffness due to embedment A rough estimate of the added stiffness due to embedment indicates a value of Kx between 5 and 6x10 7 lb/in; which is very close to the value obtained by performing a coordinate transformation on the values of K', and K given in Table 7.
It is also interesting to note that the value of K'g given in Table 7 is very close to the theoretical value of 6GR=5.3x10 7 lb/in.
By performing a similar' coordinate transformation on the K'z will demonstrate that this value is unaffected by coupling with the rotational stiffness and, therefore, K'
should be close to 6GR.
As has been described above, a translational stiffness comparison can only be made in equivalent coordinate systems.
In the case of the reactor building analysis, the location of the foundation node below grade introduces a
coupling between horizontal translational stiffness at the ground surface and the rotation stiffness applied at the foundation node.
When the effective horizontal translational stiffness at the ground surface is computed by performing a
coordinate transformation an approximate agreement with the theoretical values calculated in accordance with BC-TOP-4A Rev. 3 is demonstrated.
Exact agreement should not be a basis for reviewing the validity of the stiffness values given in Table 7 since the actual foundation geometry and embedment cannot be accounted for with these theoretical formulas.
The axisymmetric finite element model did account for these additional parameters and as such provided an excellent basis for the stiffness coefficient given in Table 7.
Therefore, although the horizontal and vertical stiffness may be comparable at the ground surface, this is not the case at the foundation location where the horizontal stiffness is essentially rigid and can be neglected.
ITEM 7 Provide the basis for the allowable fuel assembly grid impact load of 6660 lbs which is discussed in Sections 3.9.1.4.4 and 3.9.1.4.7 of the April 29, 1977 submittal.
-13
RESPONSE
Experimental impact strength data for the San Onofre Unit 1 fuel assembly grid design was not available, consequently the impact strength for that design was obtained by extrapolating test information from a 14 x 14 grid of similar design.
The structural and strength differences between the two designs result primarily from the differences in, the inner strap thickness.
The tested grid had a nominal inner strap thickness of.0135 inches, whereas the San Onofre Unit 1 grid has a nominal inner strap thickness of.0155 inches.
The impact strength for the tested grid used in the extrapolation was reported as 6660 lbs.
This value was obtained using a pendulum hammer type test fixture.
The grid impact strength was defined as the maximum attainable load prior to grid local permanent deformation or buckling.
The estimated allowable grid impact strength value of 7300 for the San Onofre Unit 1 grid was extrapolated from the above tests and was based on 80% of the nominal estimated grid strength.
It is noted that our understanding of current NRC practice is to use 95% of the mean strength as an allowable grid strength criteria. This provides additional conservatism in the value used in the reanalysis.
I/
TABLE 1 SIGNIFICANT BUILDING FREQUENCIES Building Frequencies *-Hz Mode Closed Gap mixed Gap Open Gap Description Case Case Case oil Vertical Translation 3.5 3.5 3.55 oil Rocking Motion 3.7 3.65 3.71 3.9 3.9 3.79 4.03
- Slight variation in reactor building frequencies is the result of different modal effective mass, mode shapes, and frequencies associated with the primary loop due to the different gap conditions assumed.
- TABLE 2 SEISMIC REACTOR BUILDING DEFLECTION COMPARISON Displacements -
Inches Time-History SRSS CSM Location Direction Displacement Range Range Operating Deck X
0.36 0.2 to 0.3 0.4 to 0.5 Northeast Building Y
0.30 0.3 to 0.4 0.5 to 0.6 Corner - Node 4 Z
0.55 0.4 to 0.6 0.7 to 0.8 Foundation Node 1 X
Y Z
0.4 0.3 0.4
\\~T~MKGfl~A~
9UPERTABLE 3PA WTR PRfF'GENER
, UPPER SUPPORT IMPACT LOAD PARA Total Impact o
To Steam Upper Gap Stiffness 2
(2)
Generator Support (in)
K(lb/in)
(sec)
(sec)
A Upper NW 0.2 3.73 x 106
.046
.09 Upper SE 4.1 x 106
.046 B
Upper E 1.93 x 106
.069 Upper W 0.27 1.58 x 106
.065
.135 C
Upper SW 0.24 4.82 x 106
.04
.08 Upper NE 4.6 x 106 (1) For all supports A
= 0.9g = 348 in/sec 2 1
=
= 0.256 sec.
M K
(2)
T0 is determined from averaging the associated values and rounding off.
TABLE 4 STEAM GENO OR UPPER SUPPORT IMPACT LOADSW SG A SG B SG C NW SE W
E SW NE No.
Description*
Support Support Support Support Support Support
- 1.
Impact Forces -
101 112 120 145 129 123 Assumption 1
- 2.
Impact Forces -
678 746 536 655 773 738 Assumption 2
- 3.
Average Impact Forces 390 429 328 400 451 430 From 1 and 2
- 4.
Impact Forces from 410 435 438 400 460 550 Time-History Analysis Assumption 1
- Supports do not move and mass responds in tune with the horizontal building motion.
Assumption 2
- Supports respond in a sinusoidal manner and the mass is in resonance with the support motion.
TABLE 5 -
Displacements UNCOUPLED MODEL COUPLED MODEL SRSS DISPLACEMENT SRSS DISPLACEMENTS COMPONENT (FEET)
(FEET)
Primary Shield Wall
.0648
.0554 Foundation
.0654
.0582 Operating Deck
.1080
.0877 TABLE 6-Accelerations UNCOUPLED MODEL COUPLED MODEL COMPONENT SRSS ACCELARATION SRSS ACCELERATIONS Primary Shield Wall 1.09g 0.82g Foundation 1.09g 0.83g Operating Deck 1.77g 1.16g
Table 7 -
Soil Structure Interaction Stiffness Horizontal Translatin 6.4 x 10 lbs/in Vertical Translation 5.4 x 107 lbs/in Rocking1 8.8 x 1012 lbs/radian Torsion1 5.2 x 1013 lbs/radian
- 1. Applied at the center-of-gravity of the foundation.
- e e
FIGURE 1 - Analytical Load Model
N K (1) 'r
= 27rfi (2) 2G Vi Where
= Period associated with mass M V1 = Velocity of mass M between impacts G = Total gap size between supports Solving equation (2) for V, yields:
2G FIGURE 2 - Development of Assumption I Velocity at Time of Impact
K
~
P4 X ~A iw
\\AVV\\t-J Simas leavng impact configuration of model at I~~v~vvIzero x
e--
/4 impact at other side mass impact z qw,- support motion mass mpact
-f A X "
Sinet Terms are defined as:
To
- 21r vN/il 0
T
= Period associated with mass M A
= Acceleration of support W
= Frequency of support motion = 27 T
V2 -= Velocity of mass M between impacts G
= Total gap size between supports
= Total displacement of supports between impacts Development of Velocity before Impact
(
G+A' _
A' 1
x = 2 at t = (r-ro) then (2) A-Sin
)
Solving (1) and (2) for V2 yields V
2G+ -Z Sin
(
To 2
(T To FIGURE 3 - Development of Assumption 2 Velocity at Time of Impact
V e' = V +V s e
coefficient of restitution For Assumption 1 e
Where VS and Vs are the support velocities before and after impact = zero V1 = Velocity of mass M just before impact.
Vi = Velocity of mass M adjusted to account for impact damping.
Vi = e'V 1
1 For Assumption 2 V' + V' V2 + V 2
s Where V2
= Velocity of mass M just before impact.
V2 Velocity of mass M adjusted to account for impact damping.
Vs = Velocity of support at time of impact Vs
= Velocity of support after impact.
= e' (V2+V ) - V s2 s2 From the definition of support motion in Figure 3, expressions for V and Vcan be determined to be:
s2 A
Cos (1 -
- 0)
Vs A
Cos 1+
)
FIGURE 4 - Development of Adjusted Impact Velocities to Account for Impact Damping
~~~~~1 In ui~
'4.
/
'-4
/
/
U,
/I
- 1.
C" 4
I.
- 0.
S'~r I
- 1.
I 4'
j
6--ELEVATiotj e',A IIh" r>
-A LOCATiort:
-Tieii,.
A-___
Ad tl Clw I!V E-M 1
o
-C-Ile 11Y 14" 1~~
FIGURE 6
Az A9, Figure 7: Impedance stiffness analog for near surface foundations.
Ax R
Figure 8: Impedance stiffness analog for an embedded foundation.
ENCLOSURE 2 PROGRAM FOR ASSESSMENT OF THE IMPACT OF THE 1979 COYOTE LAKE AND THE 1979 IMPERIAL VALLEY EARTHQUAKES ON THE SITE SPECIFIC COMPUTER MODEL
- 1.
INTRODUCTION The recent Imperial Valley and Coyote Lake earthquakes provide new and important data on the characteristics of ground shaking in the near vicinity to major strike-slip earthquakes. These data will be assessed to determine what, if any, impact they have on the site specific computer model.
For the case of the 1979 Imperial Valley earthquake, the rupture process will be modeled so as to simulate ground motions at several of the strong motion recording stations using earth structure representative of the Imperial Valley. For the case of the 1979 Coyote Lake earthquake, the lack of information on underlying earth properties precludes the possi bility of a similar modeling study at this time. Alternately, the strong motion data will be examined for the Coyote Lake earthquake within the context of strong motion recordings of other earthquakes for which computer simulations have been made. Comparisons will be made to see if the Coyote Lake data fit within previous experiences in modeling other data under similar configurations with respect to the earthquake rupture.
It should be noted that the Imperial Valley simulation is predictive in nature. With the exception of fault offset, the rupture parameters have been previously established by mechanistic considerations and calibrations against earthquake data. Parameters describing the earth structure at Imperial Valley have also been assigned prior to the recent Imperial Valley earthquake. Furthermore, the Imperial Valley earthquake is quite similar to the hypothesized earthquake that is being simulated along the offshore zone of deformation, some 8 km west of the San Onofre Nuclear Generating Station. Consequently, by comparing computer-generated response spectra with spectra obtained from strong motion data recorded in the near vicinity to the recent Imperial Valley rupture, the predictive capabilities of the earthquake model can be tested under conditions quite similar to those for which it is being applied.
1
- 2.
BACKGROUND The development of TERA/DELTA's earthquake modeling capabilities was based on the necessity for extrapolating strong ground motion to sites involving specific fault geometries and earth structures. The computer model is based upon a representation of the physics of earthquake rupture and upon Green's Functions which serve to propagate the seismic energy through the earth. The earthquake model has been calibrated and tested against strong motion data recorded during several historical earthquakes.
The results of modeling studies of ground motions recorded at five stations during the 1966 Parkfield earthquake (Ms - 6.4) and at one station during the 1940 Imperial Valley earthquake (M S
7.1) have been documented in reports dated May 1978 and July 1979. Both of these reports predict response spectra at the San Onofre Nuclear Generating Station resulting from a major earthquake rupture hypothesized to occur eight kilometers offshore.
- 3.
COYOTE LAKE EARTHQUAKE The Coyote Lake earthquake of August, 1979 occurred along the Calaveras Fault near Gilroy, California with a local magnitude of ML ' 5.7.
The motion was predominantly right-lateral strike-slip along an apparently vertical fault plane. As illustrated in Figure 1, this event was recorded by six strong motion instruments located within.10 km of the fault trace.
Recorded values for the largest acceleration at each station are presented in Table 1. The earth structure in the vicinity of this earthquake is not well understood at this time;.however, the geology is known to exhibit lateral heterogeneity along a direction perpendicular to the fault.
Due to insufficient information on the earth properties in the vicinity of Coyote Lake, the impact of this earthquake on the computer model will be assessed by examining these new recordings within the context of similar data for other strike-slip earthquakes. Comparisons will be made for response spectra and peak values for displacement, velocity and acceleration. Recording stations for the 1966 Parkfield, the 1940 Imperial Valley, and the 1979 Impearial Valley earthquakes will be.selected 2
for purposes of comparison with Coyote Lake data based on distance to the rupture and orientation with respect to rupture direction (focussing).
For example, recordings at Coyote Lake Station 6 (Figure 1) should be compared with Parkfield Station 2 (not pictured) and with Imperial Valley Stations 6 and 7 (Figure 2).
All of these stations are quite close to the fault and nearly directly in the path of rupture. Similar associations can be made for the remaining five stations within 10 km of the Coyote Lake earthquake.
- 4.
IMPERIAL VALLEY EARTHQUAKE The Imperial Valley earthquake of October, 1979 was located on the Imperial Fault in approximately the same location as the 1940 Imperial Valley earthquake. This event generated more strong motion recordings than any other strike-slip earthquake to date. There were more than 25 strong motion recordings in California and about eight additional recordings in Mexico. Most of the U.S. data have been processed and are not available.
The Mexican stations were operated in part by IGPP of the Scripps Institution of Oceanography. Data from the Mexican stations should be available within a few weeks.
Instrumentation for this earthquake is so extensive as to establish important new information on the attenuation of ground motion from within a few hundred meters of the fracture surface. Earthquake focussing has been recorded over a wide range of geometric configurations. The data provide scientifically important constraints on:
rupture sequences and fault slip; ground motion resulting from rupture focussing; and wave propagation effects from the rupture source to the many recording stations.
The propensity of near-field recordings provides a suitable test of the predictive capabilities of the site-specific computer model under condi tions quite similar to those for which it is being applied, that is, for estimating ground motion at the San Onofre Nuclear Generating Station in the event of a major strike-slip earthquake rupturing 8 km to the west.
3
The suitability of the computer model will be tested by simulating ground motion over the range of source-receiver configurations that were recorded within about 16 km of the rupture surface for the 1979 Imperial Valley earthquake. As seen in Figures 2 and itemized in Table 2, there are nineteen stations within this distance range, fifteen in California and four in Mexico. Because of the close spacing between some stations and symmetric positioning about the rupture between other stations, it appears that the simulation of about twelve stations will serve to adequately model the seventeen near-field recording stations. The twelve stations currently being considered for explicit simulation are marked in Table 2.
It should be noted that these twelve stations include the complete range of distances (within 16 km) and orientations with respect to the rupture.
Also, the station with the largest recorded acceleration was generally selected when selecting between two stations of comparable distance and orientation.. The sited selections for explicit modeling are provided as illustration of our intent at this time; subsequent information may lead to alternate selections.
4
TABLE 1 Coyote Lake Strong Motion Stations Distance*
Acceleration Station Coordinates (km)
Azimuth (g)
Coyote Creek 37.120
.2 70
.24 121.540 340
.14 Up
.10
- 6 37.030
.3 140
.31 121.480 50
.41 Up
.15
- 4 37.000 5.2 270
.23 121.530 360
.25 Up
.41
- 3 36.990 6.6 50
.25 121.540 140
.24 Up
.14
- 2 36.980 8.7 50
.19 121.560 140
.25 Up
.16
- 1 36.970
- 10.
320
.13 121.570 230
.1 Up
.08 Closest distance to fault trace.
TABLE 2 Imperial Valley Strong Motion Stations Distance+
Acceleration Station Coordinates (km)
Azimuth (g)
- 7 32.850 1
230
.52 115.500 140
.36 Up
.65 0-2 km
- 6*
32.840 1
230
.45 115.490 140
.72 Up 1.74 Aeropuerto*
32.650 2
45
.32 115.330 315
.24 Up
.18 Bond's Corner*
32.690 3
230
.81 115.340 140
.66 Up
.47 Agrarius*
32.620 4
3
.28 115.300 273
.23 Up
- 8*
32.810 4
230
.50 115.530 140
.64 Up
.55 3-6 km
- 5 32.860 4
230
.40 115.470 140
.56 Up
.71 Diff. Array 32.800.
5 360
.51 115.54 270
.37 Up
.93 Brawley 32.990 7
315
.22 Airport*
115.510 225
.7 Up
.18
- 4*
32.860 7
230
.38 115.430 140
.61 Up
.32 Holtville 32.810 8
315
.22 7-11 km 115.380 225
.26 Up
.31
- 10*
32.780 9
50
.20 115.570 320
.23 Up
.15
Table 2 (continued)
Distancet Acceleration Station Coordinates (km)
Azimuth (g)
Calexico*
32.830 11 315
.22 115.490 225
.28 Up
.21 Mexicali 32.620 13 0
.31 SA Hop 115.420 90
.46 Up
.33
- 11*
32.750 13 230
.38 115.590 140
.38 Up
.16
- 3 32.890 13 230
.22 12-16 km 115.380 140
.27 Up
.15
- 2 32.920 16 230
.43 115.370 140
.33 Up
.17 Cucapah*
32.550 14 85
.31 115.230 355 Up
.12 Parachute*
32.930 15 315
.20 115.700 225
.11 Up
.18 Stations proposed for modeling.
tDistance to nearest point on the 1940 Imperial Fault trace.
37010' Coyote Creek EPICENTER.
\\*#6 Gilroy
- #4 37000'
- 3
- 2
- 1 II I
0 5
10 (km) 121030' Figure 1. Strong motion stations for the 1979 Coyote Lake earthquake.
Dashed-line shows approximate position of Calavaras Fault trace.
DEL MAR TECHNICAL ASSOCIATES
Calipatria 33000' Brawley Superstition Airport Mountain
- #1
- S Parachute
- #2
- 3
- #5A #4
- 6
- 8
- 0 Holtville
- Plaster City DA DA 10 EPICENTER 1940 32045'.
- 11
- #12 Bonds Corner
- 13 Calexico U.S.A.
Aeropuerto EPICENTER 1979 Mexico Igrarius Mexicali Compuertas S A Hop Cucapah 320301 0* -
Chihuahua Cerro Prieto 0
10 20 30 (km)
Delta 11504 115030 1150153.
Victoria Figure 2. Strong motion stations for the 1979 Imperial Valley earthquake.
Fault trace is for the 1940 Imperial Valley earthquake.
DEL MAR TECHNICAL ASSOCIATES