ML20073Q452

Forwards Ltr Repts Re Seismic Confirmatory Program, Explosion & Vibration Tests & Computer Program for Evaluating Equation 58

ML20073Q452
Person / Time
Site:	Summer
Issue date:	04/25/1983
From:	Algermissen S INTERIOR, DEPT. OF, GEOLOGICAL SURVEY
To:	Rolonda Jackson Office of Nuclear Reactor Regulation
Shared Package
ML20073Q454	List: ML20073Q452 ML20073Q455
References
NUDOCS 8304290126
Download: ML20073Q452 (1)

Text

.-.- .__.. . _ . . . .

t y-f United States Department of the Interior H gl-.

=

_._ j GEOLOGICAL SURVEY BOX 250-16 '- M.S o DENVER FEDERAL CENTER DENVER, COLORADO 80225

. IN Rt Pl.T Rtti R 'In Branch of Engineering' Geology and Tectonics Mail Stop-966 April 25, 1983-p .67' N Dr. Robert E. Jackson, Chief Geosciences Branch Division-of Engineering .

Office of-Nuclear Reactor Regulation Nuclear Regulatory Commission' Washington, D.C.'.20555

Dear Bob:

Enclosed are three letter reports, one from W. B. Joyner and two from Noel Bycroft, concerning the. Virgil C. Summer Nuclear Station Unit 1.

Please let me know if you have any questions regarding these reports.

Sincerely, 1

S. T. Algeraissen 1 Enclosures 1.

J 4

i

@D7 s ,

c. 8304290126 830425 PDR ADOCK O$000395

p. PDR w-

- ~ . . _ _ . . _ _ _ - .. . _ . - _ . . ~ . . , _ . . . _ . . . . - . . _ . . . - . . . _ . _ _ . _ _ . _ _ _ . _ . _ . _ . _ . _ ,

/,bl's

. ., 1 United States Department of the Interior

h. GE01DGICAL SURVEY OFFICE OF EARTHOUAKES, VOLCAN0ES. . AND Ef'GIf;EERIPG Eranch of Engineering Seismology and Geology 345 Middlefield Road, MS 77 Henlo Park, CA 94025 April 8,1983 Dr. Jerry L. King U. S. fiuclear Regulatory Cmmission Mail Stop P-514 Washington, D. C. 20555

Dear Jerry:

In preparation for our scheduled neeting with the Licensee later this month, I am setting down ny preliminary comnents on the report " Seismic Confirnatory Program, Virgil C. Summer Nuclear Station Unit 1", including the Adder: dun to Appendix B, dated March 1993.

floel Bycrof t will cement on the theory used in interpreting the pad tests, and I will leave that subject largely to him. I.wauld like to suggest, however, that since the modal anproach, which was used to get the final answer, is not conpletely rigorous, it would be of great interest to see a comparison between the observed displacement anolitude and phase frm the low-force-level test (e.g., Figure 4 of Appendix A) and the predictions of the model obtained by the modal analysis. That should be a relatively simple thing to do.

With respect to the explosion tests as a means of determining foundation ef fects, I don't believe the submissions demonstrate the applicability of the explosion tests to earthquakes. They show that the explosion seisnograms are dominated by what they call S-wave and higher node surface waves. They claim to show that the spectral ratios are similar whether cmputed for a P-wave window, an S-wave window (including the surface waves), or for the whole record. They show that fundamental node surface waves do not contribute to the spectral ratios. All of this, however, does not demonstrate that the relative response of foundation and free-field sites is the same for earthquakes as for explosions. If we accept the description of the dominant portion of the selsnograms as a cmbination of S-waves and higher node surface waves, we have to presume that the relative excitation of the various components of this cmbination is sensitive to the depth of the source, and further, that the relative response of foundation and free-field sites is sensitive to the relative excitation of the coeponents. Looked at from a s - dif ferent point of view, it ray be more realistic to consider the dominant portion of the seismogram to be a combination of scattered S-waves, scattered not only in the vicinity of the recording site, but also in the vicinity of the source a in the vicinity of any surface reflection points between the f .

pr go

source and the recording site. _ From this point of view, one would expect the character of the motion, and in particular, the relative response at

. - foundation and free-field sites to be sensitive to the depth of the source.

- Either way, the source depth may be a significant variable. The average depth of the explosions must be significantly less than the average depth of the earthquakes, and all of the. explosion depths may in fact be less than all of g  : the earthquake depths. Two _ shots were at depths comparable to the computed l hypocentral' depth of the October 1979 earthquake, but no data from those shots

- is shown and, as I interpret Table V.C.I. no data from those shots was. used in cmputing horizontal spectral ratios. In any case, the uncertainty of the computed hypocentral depth forces us to acknowledge that the real depth could be substantially greater than that for all 'of the explosions. This leaves me with substantial doubt concerning the applicability of the explosion tests for determining relative. response in an earthquake. (My conern on this point is increased by comparing Figure IV.C.1 with Figures IV.C.3 and IV.C.6.. The comparison is difficult to make because of the scale change, but the' records-

, of the shots seem to _ have higher coda amplitudes at free-field sites than the L record of the earthquake, and the particle-motion diagrams show a different

[ pattern for the earthquare compared to the explosions.)

l I am-also concerned about the use of a zero-phase-shift filter to L represent the transfer function!between foundation and free-field sites. The i effect of this 'is difficult to assess and could be quite significant, Relative _ phase controls the. extent to which 'an increase in modulus represents an : increase in peak amplitude and the extent to which it represents an increase in duration. The effect of an increase in amplitude on damped response spectra may-'be very different from the effect of an increase in l duration, depending on ae character of the signal. To help us assess this

- aspect of their work, I think the Licensee should be asked to show comparisons of typical observed free-field records with synthatic free-field records formed by filtering Auxiliary Building records with the corresponding zero-phase-shift filter.

Another major problon in the interpretation of the explosion tests is the treatment of variablility, The consultants use the-mean spectral ratio. I question this cnoice in spite of the ingenious argument offered by Robin McGuire. As you remember, he argued that, if the mean ratio had turned out to be 1.0, NRC would not have required that the spectrum be raised; so no extra margin over the mean ought to be required in the case of a nean less than i 1.0. I disagree. I think ti',4t if the mean had (surprisingly) turned out to be I.0, prudence would have required giving serious consideration to raising the spectrum.- From the standpoint of safety, the question is not whether the i spectrum is exceeded on the average, but, rather, whether it is exceeded anywhere. If the Envelope Spectrum embodied a large degree of conservatism, it might be argued that the use of the mean spectral ratio was justified, but

i. the envelope of all events recorded in approximately a five-year period cannot be considered conservative without some debatable assumption about the rate at which the earthquake sequence is dying off.

4 e mee e-+e ,---v

-. .  ?.

.t- ;a.

. I cannot find a clear explanation of how the " band pass spectral ~ ratios"-

for the various windows were calculated. I'may just.have missed it,but, if

not', the consultants should supply it. -In comparing the band pass ratios-shown in Figures VI.C.55 and _VI.C.58 with the ratios in Figures VI.C.20 and VI.C.24, it does not seem to me that-the comparison is very good,-particularly L in the; vicinity of 20 Hz. Similarly, I don't think the agreement -is very good '

between the ratios shown in Figures VI.C.47 and VI.C.48. .The consultants apparently consider disagreerents by a factor of two as matters of trivial

.importance. That being the case, they will presumably not object to raising _

their final answer by a factor of two.

The problems discussed.above are the most important issues I see at this time. -In going through the report I noted some minor. problems. At the bottom of page 1, it _is stated that the results of the pad tests comfirm testimony at the 1982 ASI.B hearings. -I do not.believe that is the -case. You-have access

.to the transcripts and _could check. If the statement is .not correct, perhaps

it should be deleted; otherwise, the record gives a false impression of consistency. I don't understand the last paragraph on page 8.- I expect you "do. but, if not, it should be clarified.

It is not clear to me that the reasons they give on page 9 (paragraph 1) are adequate to'show that the Auxiliary Bullding motions-are representative _of motions in the Intemediate Building; an engineer should look at this. On page 10, paragraph 2, -I don't .

understand why motions recorded in the sump would overestimate motions

• expected on the main floor of the Diesel. Generator Building. Either Figures

! IV.A.7 and -IV.A.8 are mislabeled or Figure IV.C.3 is mislabeled. The labeling ,

i . of Figure V.C.1'.is inconsistent. Does it represent data recorded at site F3 '

or' site FR7 In Table C.V.1 the tem " Code Revision #" should be defined. On

,, page 43 of Appendix B, line 13 up from the bottom, a reference is made to Figure _VI.C.7' when Figure VI.C.6. is obviously intended. The discussion in the first paragraph'of page _43 of the differences between spectral ratios

-computed for different time _ windows is entirely in terms of the vertical

, component. It-would have been more relevant if it have been done for the horizontal components.

If you want to discuss any of these problems before the meting, give me a cal l .

Best regards, William B. Joyner Geophysicist Copy to:

yMed Algemissen

r

/s.A~s United States Department of the Interior

- GEOLOGICAL SURVEY OFFICE OF EARTHQtlAKES, VOLCAN0ES ' AND ENGINEERING '

Branch of Engineering Seistology and Geology 345 Middlefield Road, MS 77

'Fenlo Park, CA 94025 April 8, 19P.3 Dr. Jerry L. King U. S. Nuclear Regulatory Commission Mail Stop P-514 Washington, D. C. 20555

Dear Jerry:

Bill Joyner and I agreed that he would review the explosion tests, and I would review the forced vibration tests on the pad.

I have found no analytical errors in either the stiffness or r'odal approach. The stif fness rethod which was correctly elected to be the h

fundamental nethod of analysis is a rigorous method and ivolves no assumptions, except that of linearity. However, it is obvious from the-results that something is amiss. This is shown by the neaative damping obtained, by the large value of the transfer functions, by the fact that the resonance of the hut appears in the value of toe stiffness functions Kyg, K12' K21, and especially in K22, by the strange behavior of K12 and K21, and by the fact that the real part of K, should be approximately a constant value. It is hard to suggest why this is all so. Suggestions about the phases being difficult to determine are a possibility. Certainly, it is difficult to l

separate out the needed information from cmplex data where all the modes are I

being excited at the same time, unless the reasurements are very clean and the equations well conditioned. Possibly, it would be better experimentally if i-

~

the hut was removed and treated separately and a pure torque obtained from two l

out of phase vibrators, symnetrically placed, be applied. Similarly, a translation notion with ninimum rotation could be obtained by exciting the pad

(

through its center of gravity by a horizontal rod connected to a vibrator j 1 hk L

=

. N ,

a 6

placed on the adjacent ground. This would be experimentally more difficult, but may praduce more tractable measurements.

The essential assumption in the modal approach is that classic normal modes exist. This is' not strictly the case in this application because the

~

compliance. functions are frequency dependent.- This frequency dependence is shown-in Figures 4 and 5 of the enclosed paper for the case of translation and rotation of a circular plate' on the surface of an elastic halfspace. If we assume a shear wave velocity of 600 ft/sec then for an effective r value o of 2.3 ft., the value of tha non-dimensional frequency factor a o is 'about 1.2 for the resonant frequency of, say, 50 Hz. For this value of ao it is to be noticed that.the values of fj -and fj are not too different from their values at a o= 0 and consequently, the modal approach is a reasonable approxi-mation. C. B. Crouse, in a private communication, has pointed out that-this is an embedded foundation (18" embedment) and that Dominguez (referenced in the Ertex Appendix A) has shown that this degree of embedment substantially l' changes the compliance functions- from those of a foundation on the surface. I l have not yet seen this reference. Presumably Dominguez assumes that the l

sides of the foundation can take tension as well as compression. I assume

-that compaction takes place at the sides and so the effective embedment is somewhat less than 9" and that the plate on the surface is'a reasonable approximation ~. The results obtained from the modal approach appear entirely feasible. We have programmed Eq. 58 of the enclosed paper in order to nake an l analytical assessment of the problem and initially obtained values of the I free-field similar to those of Appendix A. The effect of the hut was not included on the assumption that we-are most interested in the region of the second resonance. However, at the last minute of writing this note, we suspect our computer program and are currently checking it out. We used the following values of the parameters:

e r- ,- ,,m-r l.,-,..,-e- - - , , - - - . . - , - ,,-,a - ,en- , , - - - - - , ~ -e-, , , -- ,

Shear wave velocity = 600 ft/sec by = 3.0 b2 = 1.0 A y = 0.45 A2 " 1*2 The value of R discussed in the enclosed paper for these parameters has not yet been calculated. Figure 8 in the enclosed paper shows R for values of t'.e r parameters somewhat. different from the above values but roughly in the ,

same region. Remembering that a, = 1.2 corresponds to' about 50 Hz a comparison between Figures 8 and 9 of the enclosed paper and Figures 4 and 13' of Appendix A shows that they fall in the same ballpark for the second

- resonance. The relevant larger value of by = 3.0 will increase the value of R shown in Figure 8.

Bill Joyner notes that as the modal approach is not completely rigorous, it would be useful to make a comparison between the observed displacement amplitude and phase such as shown in Figure 4 of Appendix A and the predictions of the model obtained by the modal approach.

. Conclusions

1. The results of the stiffness method are invalid.

2. The results of the modal method appear entirely acceptable.

Sincerely, G. N. Bycroft Physical Scientist topy to:

T. Algenaissen