ML20070N834
| ML20070N834 | |
| Person / Time | |
|---|---|
| Site: | Peach Bottom  |
| Issue date: | 04/29/1994 |
| From: | Hunger G PECO ENERGY CO., (FORMERLY PHILADELPHIA ELECTRIC |
| To: | NRC OFFICE OF INFORMATION RESOURCES MANAGEMENT (IRM) |
| References | |
| REF-GTECI-A-46, REF-GTECI-SC, TASK-A-46, TASK-OR GL-87-02, GL-87-2, NUDOCS 9405090355 | |
| Download: ML20070N834 (11) | |
Text
-
Station Support Dspartrnent GL 87-02 Supp.1 m
g PECO ENERGY
= n=#:L, 905 Chethntavok Doulaud Wayne, PA 19087 5691 April 29,1994 Docket Nos. 50-277 50-278 License Nos. DPR-44 DPR-56 U. S. Nuclear Regulatory Com. mission ATTN: Document Control Desk Washington, DC 20555
Subject:
Peach Bottom Atomic Power Station, Units 2 and 3 l
Response to Request for Additional Information on Supplemental Response to Generic Letter 87-02, Supplement 1, " Verification of Seismic Adequacy of Mechanical and Electrical Equipment in Operating Reactors, Unresolved Safety Issue A-46."
References:
1)
Letter from G. J. Beck [PECO Energy Company (formerly Philadelphia Electric Company)] to NRC dated September 18,1992 2)
Letter from J. W. Shea (NRC) to G. J. Beck dated November 17,1992 3)
Letter from G. A. Hunger, Jr. (PECO Energy Company) to NRC dated January 24,1994 4)
Letter from S. Dembek (NRC) to G. A. Hunger, Jr. dated March 29,1994 l
Dear Sir:
l By letter dated September 18,1992, (Reference 1) PECO Energy Company (PECO Energy) submitted a response to Generic Letter (GL) 87-02, Supplement 1,
" Verification of Seismic Adequacy of Mechanical and Electrical Equipment in Operating Reactors, Unresolved Safety Issue (USI) A-46," for the Peach Bottom Atomic Power Station (PBAPS), Units 2 and 3. In the submittal PECO Energy outlined its plan to follow the Seismic Qualification Utility Group's (SOUG) " Generic i
Implementation Procedure (GIP) for Seismic Verification of Nuclear Plant Equipment," to resolve the seismic verification issues associated with USI A-46.
940509035s 940429 s
DR ADOCK 05000277 l
p Ak
April 29,1994 Page 2 The NRC responded to the September 18,1992, submittal by letter dated November 17,1992, (Reference 2) stating that the staff had evaluated the response and i
concluded that the procedures and criteria used to generate the licensing basis in-
]
structure response spectra are adequate for the resolution of USI A-46.
By letter dated January 24,1994, (Reference 3), PECO Energy supplemented Reference 1 to inform the NRC of its plans to develop and implement realistic, median-centered in-structure response spectra as outlined in the GIP for resolution of USI A-46 for equipment in the Radwaste Building.
The NRC staff performed a preliminary review of the January 24,1994, letter and by letter dated March 29,1994, (Reference 4) informed PECO Energy that additional information was needed to complete their review. The NRC requested that the additional information be provided within 30 days from receipt of the March 29, 1994, letter.
. Restated below are the NRC staff's questions followed by the PECO Energy response:
i Question 1.
Provide the details of the suite of earthquake acceleration time histories that will be selected from historical earthquakes or generated artificially, including their characteristics such as duration of strong ground motion.
Response 1.
A suite of thirty different acceleration time histories were selected.
The following information concerning these time histories is provided.
I l
Table 1 identifies the characteristics of the historical records which are part of the 30 records used in the time history analyses.
Table 2 identifies the strong motion characteristics of all the time histories used in the time history analyses.
Figure 1 is a plot of the 5% damped response spectra of all the 30 time histories and their median at all frequencies.
Figure 2 is a comparison of the target NUREG/CR-0098 84th percentile spectral shape (anchored to 0.12g) and the median and 84th percentile response spectra obtained from the 30 time histories.
Figure 3 plots the ratio of the 84th percentile spectral accelerations and the median spectral accelerations of the 30 time histories at all frequency ranges.
d
{
April 29,1994 Page 3
]
Duestion 2.
Describe, in detail, the procedure used to develop the in-structure response spectra.
J l
Response 2.
This discussion outlines the methodology for developing in-structure i
response spectra for use in the USI A-46 program evaluations of equipment in the Seismic Class I portion of the Peach Bottom Radwaste Building. The methods described below comply with j
guidelines given in section 4.2.4 of the SQUG GIP. Specifically, the GIP (p. 4-17) states that realistic, median-centered in-structure response spectra may be compared to 1.5 times the Bounding i
Spectrum as a valid comparison of seismic capacity to seismic demand for USI A-46 equipment evaluations.
i l
The proposed method makes use of an existing model of the.
i Radwaste/ Turbine Building developed for the Individual Plant j
Examination of External Events (IPEEE) in-structure spectra j
generation but incorporates a more elaborate method for generating
}
the in-structure spectra. The A-46 spectra will be generated using a suite of earthquake time history inputs to the structure model, while randomly varying key properties of the structure model. The use of a suhe of time histories and variation of structure properties incorporates, in a statistically correct manner, the variability i
inherent in the input motion as well as the modeling of the j
structure.
f Radwaste/ Turbine Building Model The structural model developed for IPEEE in-structure spectra i
generation is a full three-dimensional modeling, incorporating horizontal eccentricities between centers of mass and centers of rigidity at each major elevation. The analytical model is shown in Figure 4. The model has realistic concrete properties to account for j
concrete aging and stiffness reduction due to micro cracking. The j
median value for structural damping that will be selected for all modes of vibration will be 7% of critical damping.
Input Motions l
i A suite of at least 30 different earthquake acceleration time histories in 15 pairs will be selected from historical earthquakes or J
j generated artificially. The 5% damping response spectra for all time history components will be generated. The median (50 percentile) response spectrum will be calculated from all the individual time histories and compared with a A-46 GIP recommended ground response spectrum shape, the NUREG/CR-0098 (84 percentile) response spectrum shape with a peak ground acceleration (PGA) of 0.12g.
i
,,... - ~,
April 29,1994 Page 4 The earthquake time histories will be scaled and possibly modified until a suitable match is obtained with the target spectrum.. A suitable variability in the suite of time histories will be maintained by showing that the resulting response spectra have a coefficient of variation (COV) of about 0.2 to 0.3 over the range of important structure frequencies.
Each horizontal earthquake pair will be used twice. The second use of each earthquake pair will switch the horizontal components, resulting in a total of 30 horizontal time history pairs for use in 30 analyses. For each of the 30 pairs of horizontal earthquake time histories, a vertical time history will be selected randomly from the remaining 28 time histories. This approach avoids the correlation between the vertical and horizontal motions that would result if the same component were reused. The vertical component will be scaled down by an additional factor of 2/3.
Variation of Structure Parameters Variation in the structural response due to variation in structural damping and frequency will be included in the following manner. A -
Latin Hypercube simulation will be used to select random variables (model parameter values) to be used in each of the 30 time history analyses. The earthquake time histories are assumed to be equally likely so that the sample size for the simulation will be set equal to
- 30. Damping ratios and structural frequencies are assumed to be random variables that are log normally distributed with the medians and variabilities shown below. The damping variability implies that a i one standard deviation is 5% to 10% damping. The variability i
in the frequency ratio implies that a one standard deviation is 0.78 to 1.28 times the median frequency value, i
Parameter Median Variability (B)
Structural Damping 0.07 0.35 Structure Frequency Ratio 1.00 0.25 The domain of each model parameter is divided into 32 strata such that each strata is of equal probability. -Parameter values within the first and the 32nd strata (that is, the tails of the probability distribution function) are considered to be extreme, unrealistic values. The sampling is then limited to the remaining 30 strata.
The Latin Hypercube simulation then randomly selects (from a log normal distribution) a damping and a frequency ratio with a specified median and a variability and combines each value with one of the 30 equally likely time history sets used for analysis.
.= -
April 29,1994 j
Page5 For each analysis, the modal frequencies will be scaled and the modal damping assigned according to the values selected. A mode superposition time history analysis is then performed for each of the 30 earthquake /model parameter value sets.
Median In Structure Response Spectra In-structure response spectra at 5% damping will be generated for each response time history at each model response point and for the -
two horizontal and the vertical directions. The 30 response spectral will be combined and the median response spectrum for each location and direction will be calculated.
As stated in. Reference 3, PECO Energy plans to have the realistic, median-centered in-structure response spectra developed and available for use prior to the Radwaste Building walkdowns. The utilization of this response spectra is not expected to impact the projected completion date of November 20,.1995.
Should the staff require any additional information, please contact us soon as possible, thereby avoiding the possibility of impacts to the completion schedule.
Sincerely, W..c. Erg pc j
G. A. Hunger, Jr., Director Licensing Attachments: Tables 1,2; Figures 1,2,3,4 cc:
T. T. Martin, Administrator, Region I, USNRC W. L Schmidt, USNRC Senior Resident Inspector, PBAPS m
. ~.
9 i
e U
Chotocteristics of 16 Historical Time Histories used in Peach Bottom Rodwoste/ Turbine Building Anotyses l
Time History Earthquake Magnitude RecordingStation Component Site Conditions i
Nome Date Distance l
tobos-tr Tobos. fron 7.4 Tobos ir-Stiff ottuvium/ rock i
tobos-In 16-Sep-78 3km Ln goz!!eos Gozii. U.S.S.R 6.8 Korokyr Point East Rock / stiff ottuvium j
gozlinor 17-May-76 3km North i
i i
Ivdan00e Imperto! Valley, CA 6.5 Differentlo! Array N00E Deep o'tuvium ivdon9(ht 15-Oct-79 5km N90W
[
ivec-40e Imperial Valley, CA 6.5 El Centro No. 4 S40E Deep o!!uvium 1
Ivec-50w 15-Oct-79 4km S50W pvpp-045 Coolingo. CA 6.5 Pleasant Vo!!ey Pump Station N45E Stiff attuvium/ rock pvpp-135 2-May-83 (Switchyard) 10 km S45E 4
sfth-21e Son Fernando. CA 6.6 Loke Huges No.12 N21E
. Rock j
sflh-69w 9-Feb-71 20 km N69W 4
sfpac-16 Son Fernando. CA 6.6 Pocolmo Dom S16E Rock sfpac-74 9-Feb-71 3km S74W I
i dayhookt Tobos. tron 7.4 Doyhook NICE Rock i
dayhook!
16-Sep-78 17 km N80W Table 1 i
i k
(. -- -..
.__ -. ~... -. -
1 i
~
y l
I i
l Chorocteristics of 30 Time Histories used in
]
Peach Bottom Rodwoste/ Turbine Building Analyses l
1 Time-History Time Step Total No. of Duration
- Strong Motion i
Nome (sec)
Points (sec)
(sec) l
]*
tobos-tr 0.01 2900 29 7.89 tobas-In 0.01,
2900 29 7.48.
gozlieos 0.00651 2048 13.46 6.15 j
gozlinor 0.00657 2048 13.46 5.62-ivdon00e 0.01 4096 40.96 5.03
}
ivdon90w 0.01 4096 40.96 5.5 i
ivec-40e.
0.01 4096 40.96 6.2 i
ivec-50w 0.01 4096 40.96 5.28 pvpp-045 0.005 4000 20 7,9 1
pvpp-135 0.005 4000 20 6.6 I
sflh-21e 0.02 1838 36.76 -
5 l
sflh-69w 0.02 1838 36.76 5
]
sfpac-16 0.02 2092 41.84 6.6 i
sfpac-74 0.02 2092 41.84 7.38
)
dayhookt '
O.02 2000 40 7,72.
(
dayhook!
0.02 2000 40 7.84 i
Artificial T/H 1 0.005 4000 20 9.24 j
Artificlot T/H 2 0.005 4000 20 11.22 l
Artificial T/H 3 0.005 4000 20 9.63 j
Artificial T/H 4 0.005 4000 20 9.06 l
ArtificialT/H 5 0.005 4000 20 9.57 Artificial T/H 6 0.005 4000 20 10.2 l
Artificial T/H 7 0.005 4000 20 8.57 j
Artificial T/H 8 0.005 4000 20 10.02 i
ArtificialT/H 9 0.005 4000 20 9.88 I
Artificial T/H 10 0.005 40C0 20 9.82 l
ArtificialT/H 11 0.005 4000 20 9.39 d
ArtificlolT/H 12 0.005 4000 20 -
9.83 ArtificialT/H 13 0.005 4000 20 10.34 Artificial T/H 14 0.005 4000 20 10.52 i
- The duration used in the structural time-history analyses Is 20 teconds or the total duroflon of the record for records with less than 20 seconds of motion l
l Table 2 i
b J
il' l
Peach Bottom Radwaste/ Turbine Building,30 Response Spectra for 30 Horizontal T/H Inputs,5% Damping 0.8 0.7
)M 0.6
}l
? !
0.5 I
^
O-
^
0.4 l'
A 0.3 E
0.2 2
= -- ? -
+ --
0.1
-- m w
w m-gmM.
W~
~
0 t
0.1 1
10 100 Frequency (Hz)
Figure 1 6
Peach Bottom Radwoste/ Turbine Building Horizontalinput 5% Damping 0.8 Target Spectrum (NUREG/CR-0.7 _
0098,84 Percentile @ 0.12g PGA) 50 Percentile of 30 T/H 0.6 -
84 Percentile of 30T/H 3 0.5 he t'/5 g0.4 v
E O
an
[%
M
.3 m
0
\\
0.2 o
j y
x\\x 7
w__
0.1 g/
0 0.1 1
10-100 Frequency (Hz)
Figure 2 1
Peach Bottom Radwaste/ Turbine Building, Ratio of 84th Percentile Spectral Acceleration to 50th Percentile Spectral Acceleration for 30 T/H 2
1.9 1.8 1.7 i
J 61.6
\\
/ \\
j Il 1
/
\\
IM J l V e
i
- \\
13 I-#
/\\
I t
\\_ /
~
\\/
J
~
~
)
1.1 1
0.1 1
10 100 Frequency (Hz)
Figure 3
1 l
l e
IL El 190 g,,,
e so El 165 j
b El 150 x, N l
t 7
1 5
s El 135 1
4 7/d a
y%x 9 Mass point l
L l
Figure 4 Radwaste Building Analysis Model (Y= West) t I