ML20137E056
Text
'
k I y
%o UNITED STATES [8M80 Y
- i. n NUCLEAR REGULATORY COMMISSION OFFICIAL USE ONLY
- U" ' E ADVISORY COMMITTEE ON REACTOR SAFE!UARDS PREPARED FOR ACRS USE ONLY wAssincros, o. c. 20sss F0IA EXEMPTION 5
\....*/
January 27, 1982 MEMORANDUM FOR: R.F. Fraley, Execut ye Director, ACRS FROM: /rufhF %J.F. Donoghue, AE Fellow
SUBJECT:
Commissioner Gilinsky request for review of changes in seismic design methodology as proposed by P.C. Jennings
References:
- 1. P.C. Jennings ltr to Gilinsky, 10/5/81
- 2. Victor Gilinsky memo to J.C. Mark, 12/3/81
- 3. R. Fraley memo to ACRS, 12/3/81
- 4. SECY Paper 79-30, 4/27/79
- 5. Regulatory Guide 1.60, Design response spectra for seismic design of nuclear power plants
- 6. NUREG-0800, Standard Review Plan, July 1981 At the request of Commissioner Gilinsky's office, P.C. Jennings, a seismic engineer, has written a letter (ref. 1) outlining his views of how earth-quake-re d stant design criteria should be developed for nuclear plants.
Commissioner Gilinsky requested that the ACRS review and comment on these recommendations (ref. 2). The matter was referred to the Extreme External Phenomena Subcommittee for follow-up (ref. 3). As background for this discussion you asked that I prepare a sumary of current practice re seismic design of nuclear plarts. This memo presents my findings.
In his letter Dr. Jennings recommends a five-step process for developing seismic design criteria: 1) determination of possible sources of ground motion and faulting, using geologic and seismologic methods; 2) development of frequency-response curves for the site; 3) specification of the allowable response for the structure in the case of the Safe Shutdown Earthquake (SSE) and of the Operating Basis Earthquake (OBE); 4) implementation of design criteria; 5) resolution of the site-specific questions concerning layout, location of structural members, choice of materials, placement of reinforcing steel, etc.
The letter briefly details each of these steps, and states that the major differences between this method and the NRC procedure is that some over-simplified definitions and practices of the NRC are not used, and the judgments are made differently. It is difficult to discern much difference between the methods described in the letter and those presently in use by NRC, as stipulated in 10 CFR 100, Appendix A, which was adopted in 1973.
Appendix A is currently in the process of being updated to reflect advances 0D00230120 e50722 PDR FOIA DELLD5-363 PDR
/
)
in the geosciences (ref. 4), but the revisions have not yet been finalized.
A brief comparison is given below. The' numbers in parentheses refer to sections,of Appendix A.
- 1) Determination of possible sources of ground motion and faulting. The letter recommends seismologic and geologic studies for the purpose of identifying and cataloging nearby faults, and making an estimate of their activity rates and most recent movement. The expected result would be an estimate of the magnitude of both the 200-year earthquake and the approxi-mately 2000-year earthquake near the site. These two estimates would define the OBE and SSE, respectively. An estimate would also be made of the largest potential earthquake that could occur anywhere near the site, not clearly linked with any of the faults that have been identified. .
Appendix A of 10 CFR 100 requires the identification and evaluation of all tectonic structures in the site region (IV.2.1). All historic earthquakes affecting the site must be cataloged as to date, magnitude, epicenter, estimated ground acceleration and duration of ground shaking (IV.2.5).
Epicenters of historic earthquakes must be correlated, where possible, with tectonic structures or tectonic provinces within 200 miles of the site (IV.2.6). All capable faults (i.e., those which have moved once within 35,000 years or on a recurring basis within 500,000 years) within 200 miles must also be identified (IV.2.7). An investigation into the potential for surface faulting at the site must also be conducted (IV.b).
Appendix A defines the SSE by reference to the historic earthquake record, where that record is reliable and complete (V.a.1). The largest historic earthquake is assumed to occur at the point on its associated tectonic structure or tectonic province that is closest to the site. Alternatively, where the largest potential earthquake based on geologic evidence (e.g., f
-length of the fault) is greater than the largest historic earthquake, this would define the SSE. Thus, the SSE must be at least equal to the largest earthquake in the historical record. As a minimum, the acceleration at the plant foundation due to the SSE is defined as at least 0.19 The OBE is defined as the maximum earthquake actually expected during the 40-year life of the plant and must be of a nature that it will produce a maximum ground acceleration of at least half that of the SSE (V.a.2).
- 2) Development of frequency-response curves for the site. The letter next suggests that seismic engineers would take the design earthquakes, as defined above, and develop a response spectrum (in terms of acceleration, velocity or displacement) over the entire frequency range of interest. One or two earthquakes would govern in most cases, each in different parts of the frequency range. The design response spectrum would be drawn conser-vatively, probably being set above the mean response plus one standard deviation. 10 CFR 100 likewise requires that more than one SSE be evaluated t
L
4 -
. I in order to develop the maximum response over the entire frequency range (V.2.1). The regulations state that "in view of the limited data available on vibratory ground motions of strong earthquakes, it usually will be appropriate that the response spectra be smoothed design spectra developed from a series of response spectra related to the vibratory motions caused by more than one earthquake." (VI.2).
Supplementing these regulations, Regulatory Guide 1.60 (ref. 5) details an acceptable method for defining the design response spectra due to the SSE and OBE. It provides a generic response curve developed from the normalized mean-plus-one-standard-deviation responses of records from 33 earthquakes of various magnitudes, recorded at various distances, and on varying site conditions. The curve is plotted for an earthquake with a 1.0g maximum horizontal acceleration. For sites with different values specified for the design earthquake, the design spectra can be linearly scaled from the given curve. The generic curves provided in the Regulatory Guide are in-tended for use in sites that are not close to the epicenter of an expected earthquake and which are not underlain by poor soil deposits. If either of these conditions exist, a site-specific design response spectrum is called for. NRC regulations thus differ slightly from the suggestion in the referenced letter, by not requiring site-specific response spectra for all sites.
As described in the 1979 SECY paper on NRC seismic design policy (ref. 4), the staff attempted in the early 1970's to develop a site-specific method for deriving response spectra but was unsuccessful due to data limitations and the difficulty of obtaining general acceptance. At present the staff encourages the use of site-dependent spectra, especially for plants in the west (ref. 6 ). The SECY paper notes, however, that seismic data may be insufficient to develop site-specific spectra for reactors in the east.
Conversely, Regulatory Guide 1.60 (ref. 5) is not readily applicable to eastern sites, since the generic spectra were developed using western earth-quakes exclusively.
- 3) Specification of the allowable response for the structure. The letter recomends that, af ter the design spectra are defined, the allowable response for the structure be set for the OBE and SSE cases. For the OBE, only elastic response would be permitted. For the SSE, some response beyond the elastic limit would be allowed provided that structural integrity against collapse or catastrophic failure is retained.
Appendix A states that all components of the plant necessary for continued operation without undue risk to the public must remain within the applicable stress and deformation limits when subjected to the vibratory effects of the OBE(VI.2.2). For the SSE, however, the regulations permit design strain limits beyond the yield strain (IV.2.1).
l
\
l
- 4) Implementation of design criteria. The next step in seismic design !
l recommended by the letter is proper implementation of the criteria by three means: a) resolution of problems by a consulting board as they arise; l b) design review by an independent firm; and c) in-process inspection during construction.
Appendix A makes no specific mention of how the design criteria are to be implemented, other than to require that a dynamic analysis or other suitable qualification test be performed in order to insure that safety-related com-ponents can withstand an OBE (VI.2.2).
- 5) Resolution of site-specific construction and materials questions. The final step recommended by the letter is to increase the seismic capacity of the structure through judicious choice of materials, plant layout, rein-forcing rod placement, location of structural members, etc. Since these are engineering decisions which may vary from one plant to the next Appendix A imposes no requirements in this regard.
Conclusion Dr. Jennings' recommendations differ from present NRC seismic procedures in two respects:
a) The SSE and OBE would be determined by reference to recurrence intervals. The 2000-year earthquake would be the SSE and the 200-year earthquake would be the OBE. This method would require extrapolation beyond the historic record, since no U.S. earthquake histories extend back more than a few centuries. This technique would be similar to that proposed by Dr. Michael Chinnery, of MIT, during the ACRS hearings on the Seabrook operating license in 1974. Dr. Chinnery stated that larger earthquakes have a longer recurrence interval and could not adequately be predicted by simply referring te the historic record. He developed a log-linear relationship between recurrence interval and intensity, and showed that the maximum earthquake for a site must be determined by extending the historical record thousands of years into the future.
Dr. Chinnery stated flatly that the historic record provides no upper limit for size of an earthquake that may be expected at a site. NRC pro-cedures, on the other hand, relate the SSE to the largest historic earth-auake in the region cr, alternatively to the largest potential earthquake based on geologic evider.ce, b) The design response spectra would be developed on a site-specific 3 asis. In the western United States, where there is a more complete
'l
-S-historic record of earthquakes, ar.d where faults and tectonic structures are generally not deeply buried, site-specific response spectra might be a reasonable alternative to Regulatory Guide 1.60. In the east, however, and in particular on the Atlantic and Gulf coasts, the earthquake record is more sparse. Faults and tectonic structures are typically more deeply buried. In such a case design spectra might involve large uncertainties and be no more valuable that the generic criteria provided in Regulatory Guide 1.60.
Without a specific set of calculations to compare Dr. Jennings' technique with NRC-approved technology, it is not possible to determine if his system is more or less conservative than that now in use. Based on a discussion with him, however, he feels that the seismic design standards may be too low for some eastern sites and too rigid for western sites.
He believes that a minimum of 0.25g should be the design acceleration applied to eastern plants. Sone of the older eastern and midwestern sites have been designed to 0.05g. Rather than being specific recanmendations for procedural improvements, his remarks were intended to describe a general methodology used successfully in the past by advisory panels re-viewing the design of large structures.
The ACRS Subcommittee on Extreme External Phenomena is holding a meeting on power plant seismic design on January 28-29, involving presentations by members of the academic community and USGS. Results from the meeting and Dr. Jennings' comments may* Le of use to the staff in revising the present seismic design methods.
cc: ACRS Members T.G. McCreless R. Savio ACRS Fellows
N
[~ o UNITED STATES O
jq f n
i NUCLEAR REGULATORY COMMISSION OFFICIAL USE ONLY ADVISORY COMMITTEE ON REACTOR SAFECUARDS PREPARED FOR ACRS USE ONLY rASHINITON,0. C. 20555 FOIA EXEMPTION 5
% ...* /
January 27, 1982 MEMORANQUM FOR: R.F. Fraley, Execut ve Director, ACRS EJ.F. "h F'P sAtR
)Donoghue, or FROM: Fellow
SUBJECT:
Commissioner Gilinsky request for review of changes in seismic design methodology as proposed by P.C. Jennings 1
References:
T. P.C. Jennings 1tr to Gilinsky, 10/5/81
- 2. Victor Gilinsky memo to J.C. Mark, 12/3/81
- 3. R. Fraley memo to ACRS, 11/3/81
- 4. SECY Paper 79-30, 4/27/79
- 5. Regulatory Guide 1.60, Design response spectra for seismic design of nuclear power plants
- 6. NUREG-0800, Standard Review Plan, July 1981 At the request of Conrnissioner Gilinsky's office, P.C. Jennings, a seismic engineer, has written a letter (ref. 1) outlining his views of how earth-quake-resistant design criteria should be developed for nuclear plants.
Commissioner Gilinsky requested that the ACRS review and comment on these recommendations (ref. 2). The matter was referred to the Extreme External Phenomena Subcommittee for follow-up (ref. 3). As background for this discussion you asked that I' prepare a summary of current practice re seismic design of nuclear plants. This memo presents my findings.
In his letter Dr. Jennings recommends a five-step process for developing seismic design criteria: 1) determination of possible sources of ground motion and faulting, using geologic and seismologic methods; 2) development of frequency-response curves for the site; 3) specification of the allowable response for the structure in the case of the Safe Shutdown Earthquake (SSE) and of the Operating Basis Earthquake (OBE); 4) implementation of design criteria; 5) resolution of the site-specific questions concerning layout, location of structural members, choice of materials, placement of reinforcing steel, etc.
The letter briefly details each of these steps, and states that the major differences between this method and the NRC procedure is that some over-simplified definitions and practices of the NRC are not used, and the judgments are made differently. It is difficult to discern much difference between the methods described in the letter and those presently in use by NRC, as stipulated in 10 CFR 100, Appendix A, which was adopted in 1973.
Appendix A is currently in the process of being updated to reflect advances f*7~/?CNNfAII' 3
l in the geosciences (ref. 4), but the revisions have not yet been finalized. A brief comparison is given below. The~ numbers in parentheses refer to I sections of Appendix A.
- 1) Determination of possible sources of ground motion and faulting. The letter recommends seismologic and geologic studies for the purpose of identifying and cataloging nearby faults, and making an estimate of their activity rates and most recent movement. The expected result would be an estimate of the magnitude of both the 200-year earthquake and the approxi-mately 2000-year earthquake near the site. These two estimates would define the OBE and SSE, respectively. An estimate would also be made of the largest potential earthquake that could occur anywhere near the site, not clearly linked with any of the faults that have been identified.
Appendix A of 10 CFR 100 requires the identification and evaluation of all tectonic structures in the site region (IV.2.1). All historic earthquakes affecting the site must be cataloged as to date, magnitude, epicenter, estimated ground acceleration and duration of ground shaking (IV.2.5). Epicenters of historic earthquakes trust be correlated, where possible, with I tectonic structures or tectonic provinces within 200 miles of the site (IV.2.6). All capable faults (i.e., those which have moved once within 35,000 years or on a recurring basis within 500,000 years) within 200 miles must also be identified (IV.2.7). An investigation into the potential for surface faulting at the site must also be conducted (IV.b). Appendix A defines the SSE by reference to the historic earthquake record, where that record is reliable and complete (V.a.1). The largest historic earthquake is assumed to occur at the point on its associated tectonic structure or tectonic province that is closest to the site. Alternatively, f where the largest potential earthquake based on geologic evidence (e.g., length of the fault) is greater than the largest historic earthquake, this would define the SSE. Thus, the SSE must be at least equal to the largest earthquake in the historical record. As a minimum, the acceleration at the plant foundation due to the SSE is defined as at least 0.19 The OBE is defined as the maximum earthquake actually expected during the 40-year life of the plant and must be of a nature that it will produce a maximum ground acceleration of at least half that of the SSE (V.a.2). #
- 2) Development of frequency-response curves for the site. The letter next suggests that seismic engineers would take the design earthquakes, as defined above, and develop a response spectrum (in terms of acceleration, velocity or displacement) over the entire frequency range of interest. One or two earthquakes would govern in most cases, each in different parts of the frequency range. The design response spectrum would be drawn conser-
'~~ vatively, probably being set above the mean response plus one standard deviation. 10 CFR 100 likewise requires that more than one SSE be evaluated M Nk ,' E )0 0 $ W L~.:25 m ( w--[< a a. c3 f,s ,-j S&6 M " \
A-ca m ec,L .AZ, & CL A m er A= y n.e en . ..% /~ n =,/ per A- Q $$SL 9%o r4e. w^#.d,p A< d.e.JMA
in order to develop the maximum response over the entire frequency range i (V.2.1). The regulations state that "in view of the limited data available l on vibratory ground motions of strong earthquakes, it usually will be appropriate that the response spectra be smoothed design spectra developed from a series of response spectra related to the vibratory motions caused by more than one earthquake." (VI.2). Supplementing these regulations, Regulatory Guide 1.60 (ref. 5) details an acceptable method for defining the design response spectra due to the SSE and OBE. It provides a generic response curve developed from the normalized mean-plus-one-standard-deviation responses of records from 33 earthquakes of various magnitudes, recorded at various distances, and on varying site conditions. The curve is plotted for an earthquake with a 1.0g maximum I horizontal acceleration. For sites with different values specified for l the design earthquake, the design spectra can be linearly scaled from the I given curve. The generic curves provided in the Regulatory Guide are in- l tended for use in sites that are not close to the epicenter of an expected I earthquake and which are not underlain by poor soil deposits. If either l of these conditions exist, a site-specific design response spectrum is called , for. NRC regulations thus differ slightly from the suggestion in the referenced letter, by not requiring site-specific response spectra for'all sites. As described in the 1979 SECY paper on NRC seismic design policy (ref. 4), the staff attempted in the early 1970's to develop a site-specific method for deriving response spectra but was unsuccessful due to data limitations and the difficulty of obtaining general acceptance. At present the staff encourages the use of site-dependent spectra, especially for plants in the west (ref. 6 ). The SECY paper notes, however, that seismic data may be insufficient to develop site-specific spectra for reactors in the east. Conversely, Regulatory Guide 1.60 (ref. 5) is not readily applicable to eastern sites, since the generic spectra were developed using western earth-quakes exclusively.
- 3) Specification of the allowable response for the structure. The letter recommends that, after the design spectra are defined, the allowable response for the structure be set for the OBE and SSE cases. For the OBE, only elastic response would be permitted. For the SSE, some response beyond the elastic limit would be allowed provided that structural integrity against collapse or catastrophic failure is retained.
Appendix A states that all components of the plant necessary for continued operation without undue risk to the public must remain within the applicable stress and deformation limits when subjected to the vibratory effects of the OBE(VI.2.2). For the SSE, however, the regulations permit design strain limits bey 6nd the yield strain (IV.2.1).
- 4) Implementation of design criteria. The next step in seismic design recenmended by the letter is proper implementation of the criteria by three means: a) resolution of problems by a consulting board as they arise; b) design review by an independent firm; and c) in-process inspection during construction.
Appendix A makes no specific nention of how the design criteria are to be implemented, other than to require that a dynamic analysis or other suitable qualification test be performed in order to insure that safety-related com-ponents can withstand an OBE (VI.2.2).
- 5) Resolution of site-specific construction and materials questions. The final step recommended by the letter is to increase the seismic capacity of the structure through judicious choice of materials, plant layout, rein-forcing rod placement, location of structural members, etc. Since these are engineering decisions which may vary from one plant to the next. Appendix A imposes no requirements in this regard.
Conclusion Dr. Jennings' recommendations differ from present NRC seismic procedures in two respects: a) The SSE and OBE would be determined by reference to recurrence intervals. The 2000-year earthquake would be the SSE and the 200-year earthquake would be the OBE. This method would require extrapolation beyond the historic record, since no U.S. earthquake histories extend back more than a few centuries. This technique would be similar to that proposed by Dr. Michael Chinnery, of MIT, during the ACRS hearings on the Seabrook operating license in 1974. Dr. Chinnery stated that larger earthquakes have a longer recurrence interval and could not adequately be predicted by simply referring to the historic record. He developed a log-linear relationship between recurrence interval and intensity, and showed that the maximum earthquake for a site must be determined by extending the historical record thousands of years into the future. Dr. Chinnery stated flatly that the historic record provides no upper limit for size of an earthquake that may be expected at a site. NRC pro-cedures, on the other hand, relate the SSE to the largest historic earth- ) quake in the region or, alternatively to the largest potential earthquake l based on geologic evidence. b) The design response spectra would be developed on a site-specific basis. In the western United States, where there is a more complete I
l
\
historic. record of earthquakes, and where faults and tectonic structures i are generally not deeply buried, site-specific response spectra might be I a reasonable alternative to Regulatory Guide 1.60. In the east, however, and in particular on the Atlantic and Gulf coasts, the earthquake record is more sparse. Faults and tectonic structures are typically more deeply buried. In such a case design spectra might involve large uncertainties and be no more valuable that the generic criteria provided in Regulatory Guide 1.60. Without a specific set of calculations to compare Dr. Jennings' technique with NRC-approved technology, it is not possible to determine if his system is more or less conservative than that now in use. Based on a discussion with him, however, he feels that the seismic design standards may be too low for some eastern sites and too rigid for western sites. He believes that a minimum of 0.25g should be the design acceleration applied to eastern plants. Some of the older eastern and midwestern sites Y have been designed to 0.05g. Rather than being specific recommendations for procedural improvements, his remarks were intended to describe a general methodology used successfully in the past by advisory panels re-viewing the design of large structures. The ACRS Subcommittee on Extreme External Phenomena is holding a meeting on power plant seismic design on January 28-29, involving presentations by members of the academic community and USGS. Results from the meeting and Dr. Jennings' comments may' be of use to the staff in revising the present seismic design methods. cc: ACRS Members T.G. McCreless R. Savio ACRS Fellows-
-54 #/ # f jO ~
Y ( CA ~:r o~ e. /oS a L., , % gQ}}