ML19340C470
| ML19340C470 | |
| Person / Time | |
|---|---|
| Site: | San Onofre  |
| Issue date: | 08/27/1980 |
| From: | Herrmann R ST. LOUIS UNIV., ST. LOUIS, MO |
| To: | Reiter L Office of Nuclear Reactor Regulation |
| References | |
| NUDOCS 8011170453 | |
| Download: ML19340C470 (2) | |
Text
._
y SAINT LO UIS UNIVERSITY DEPARTMENT OF EARTil AND ATMOSPIIERIC SCIENCES 3507 L ACLEDE AVENUE MAILING ADDRESS:
S AINT LOUIS, Mf SSOURI 63103 P.O. BOX 8099-LACLEDE ST ATION S AINT LOUIS, MISSOURI 63156 August 27, 1980 1CN 1'C Dr. Leon Reiter i
Geology & Seismology Section Geophysics Branch U.S. Nuclear Regulatory Commission Washington, D.C.
20555
Dear Dr. REiter:
I have reviewed Supplements II and III to the report " Simulation of Earthquake Ground Motions for San Onofre Nuclear Generating Station Unit 1."
These reports represent a major new effort at' calibrating the model used for esticating response spectra. The authors have demonstrated the validity of their model for predicting the mean response spectra at short distances from the causative fault.
I have no reservatione in accepting the results of their analysis.
As always, there is the tendency to compare their results to that of others, in this case the Housner spectrum.
It should be recalled that the response spectra being compared are smoothed, or somewhat rian, spectra. Thus the shape depends on the data set.
In particular it is doubtful if the response spectra recorded at the same site for two carthquakes of the same size and location would have the same response spectra in all detail because of the inherent variability of the earthquake source. Given an enlarged data set from real earthquakes, it is likely that the Housner spectrum might be changed. The advantage of the technique used in the presett study is that a whole suite of earthquake motions can be used to obtain food estimates of the mean response spectrum shape and its variance.
Some specific comments on the report are as follows:
- 1) p. 14, Supplement II.
Reference is made to independent random number sequences. This is not adequately described. Presum-ably this refers to the micro-randomness introduced in chapter 5 of Supplement III.
W 011170 Lf}
q
- 2) The. validity of the modeling above 10 Hz may be in doubt because of the sampling _ interval of 1/40th second. Even though,the Nyquist frequency is 20 Hz, it takes at least 4 time samples
'j per cycle to properly define a sinousoidal waveform. Similarly, j
the Housner spectrum may have problems above 10 Hz for similar reasons. The model results at periods less than a few seconds are to be preferred over the Housner spectrum because of recognized problems with strong motion data at low frequencies.
The overview presented in Section 1.1 of Supplement III adequately and concisely describes the procedure used, mentioning all the critical features of the development. The acquisition of new close-in strong motion in 1977 provided the required observational i
evidence to support their method.
Si
- rely, wt&~
Robert B. Herrmann Associate Prdfessor of Geophysics RBH:leh
\\
e i
.