Protecting People and the EnvironmentUNITED STATES NUCLEAR REGULATORY COMMISSION
GL80109
DEC 15 1980
LETTER FOR ALL SEP LICENSEES
SUBJECT: GUIDELINES FOR SEP SOIL-STRUCTURE INTERACTION REVIEWS
Enclosed for your information are guidelines for performing soil-structure
interaction reviews for SEP facilities. Also included is a simplified
analytical approach for evaluating the effects of soil-structure interaction
using a lumped parameter model. The simplified approach presented does not
preclude the use of other procedures which would be reviewed and approved on
a case-by-case basis.
Sincerely,
Dennis N. Crutchfield, Chief
Operating Reactors Branch #5
Division of Licensing
Enclosure: SSRT Guidelines
for SEP Soil-Structure
Interaction Review
cc: See next page
DISTRIBUTION:
Docket Nos: 50-155 TERA RTedesco
50-10 & 237 OI&E (3) TNovak
50-213 ACRS (16) SEP File
50-409 Heltemes, AEOD JRoe
50-245 NRR Reading RDiggs
50-219 SEPB Reading
50-255 WRussell
50-244 DCrutchfield
50-206 HSmith
50-29 Project Managers
NRC PDR GCwalina
Local PDR RHermann
OELD DEisenhut
TERA RPurple
NSIC, JBuchanan GLainas
.
NATHAN M. NEWMARK CONSULTING ENGINEERING SERVICES
1211 CIVIL ENGINEERING BUILDING
URBANA, ILLINOIS 61801
8 December 1980
Mr. William T. Russell, Chief
Systematic Evaluation Program Branch
Division of Licensing
Office of Nuclear Reactor Regulation
U. S. Nuclear Regulatory Commission
Washington, D.C. 20555 (Mail Stop 516)
Re: SSRT Guidelines for SEP Soil-Structure
Interaction Review
Contract NRC-03-78-150
Dear Mr. Russell:
The Guidelines for SEP Soil-Structure Interaction Review, as prepared
by the Senior Seismic Review Team, are transmitted herewith with signature
approval.
We are appreciative of the help of the many individuals who contributed
to the preparation of these guidelines.
Sincerely yours,
N. M. Newmark
Chairman, SSRT
dp
Enclosure
Distribution:
W. T. Russell - 2
T. Cheng - 1
N. M. Newmark - 2
W. J. Hall - 1
R. P. Kennedy - 1
R. Murray - 1
J. D. Stevenson - 1
.
December 8, 1980
SSRT GUIDELINES FOR SEP SOIL-STRUCTURE INTERACTION REVIEW
Background
When a structure is founded within or on a base of soil, it interacts
with its foundation. The forces and displacements transmitted to the
structure and the feedback to the foundation regions are complex in nature;
the interactions that take place modify the free-field motions. Many methods
for dealing with soil-structure interaction have been proposed by a number
of writers. These methods can be classified in various ways and involve
generally: (1) procedures similar to those applicable to a rigid block on an
elastic half-space; (2) finite element or finite difference procedures
corresponding to various forcing functions acting on the combined structure
soil complex; and (3) substructure modeling techniques that may or may not
include use of the direct finite element method. Another, and perhaps more
convenient, classification of soil-structure interaction analysis procedures
is that of (a) direct solution techniques and (b) substructure solution
techniques as described in the report entitled "Recommended Revisions to
Nuclear Regulatory Commission Seismic Design Criteria", Report
NUREG/CR-1161, May 1980.
The elastic half-space theory considers a foundation plate resting on
an elastic medium with harmonic oscillation applied to the plate; the few
test results available to date in general have been obtained for this type
of model in this excitation condition. This concept is the basis for the
first of the three procedures described above, although for seismic
excitation the problem is the inverse of the original problem formulation.
.
2
in that the excitation originates in the earth. The other two methods noted
also involve modeling of the structure-soil system; as such the system has
intrinsic properties reflecting the make-up of the modeled system, physical
properties, and especially the boundaries (for example, as they affect
motion input, and reflection).
These analysis methods represent major advances in computational
ability, but unfortunately all the techniques have limitations, and in many
cases are not well understood. At present their use involves a great deal
of interpretive judgment.
One principal difficulty with all of the techniques is associated with
the handling of the ground input. Except for special long period waves, in
most cases the ground motion is noncoherent and nonuniform. Thus far it
appears that the analysis models may not be able to handle a broad spectrum
of complex wave motions. None of the techniques adequately handle nonlinear
effects, which are known to be of importance. As yet no good confirmatory
comparison basis exists between field observations and computations made
prior to an earthquake.
This entire topic is one that requires the most careful consideration.
Exercise of judgment as to the meaning of the results, in the light of the
comments given above, is required. Reliance on any sole approach is to be
avoided.
SEP Review Guideline Recommendations
In keeping with the SEP approach to review existing facilities, and as
reflected in the philosophy and criteria developed to date, it appears
.
3
desirable to outline briefly one technical procedure for estimating
soil-structure interaction effects. As a result of extensive discussions
between members of the SSRT and the NRC/LLL staff, and with recognition of
the many uncertainties and complexities of the topic under consideration,
the general approach presented below is recommended at this time as a
guideline. It will be appreciated that many decisions will have to be made
as a part of the calculational procedures described below and the exercise
of judgment obviously will be required. Justification and documentation are
necessary parts of the final analysis product.
At the outset it should be noted that the simplified approach described
below is not intended to preclude the use of any other procedures. The
structural input motions (at the foundation level), however developed and
justified, under no conditions shall correspond to less than 75 percent of
the defined control motions (normally taken as the free-field surface
motions); if a reduction in translational input motion is employed, then the
rotational components of motion also should be included. If other procedures
are employed they should be reviewed on a case-by-case basis.
For purposes of SEP review, one simplified approach for evaluating the
effects of soil-structure interaction, involving a lumped parameter model,
is deemed to be acceptable when employed under the following conditions.
1. The control motions are defined as the free-field surface motions
and are input at the structure foundation level.
2. The soil stiffness, as represented by springs anchored at the
.
4
foundation level, shall be modeled as follows.
i) To account for uncertainty in soil properties, the soil properties,
the soil stiffnesses (horizontal, vertical, rocking and torsional) employed
in analysis shall include a range of soil shear moduli bounded by (a) 50
percent of the modulus corresponding to the best estimate of the large
strain condition and (b) 90 percent of the modulus corresponding to the best
estimate of the low strain condition. For purposes of structural analysis
three soil modulus conditions generally will suffice corresponding to (a)
and (b) above, and (c), a best estimated shear modulus.
For structural capacity review the analyst generally should employ the
worst case condition. For equipment review the in-structure response spectra
shall be taken as a smoothed envelope of the resulting spectra from these
three analyses.
ii) When embedment is to be considered it is recommended that the soil
resistances (stiffnesses as noted above) shall correspond to 50 percent of
the theoretical embedment effects. This reduction is intended to account for
changes in soil properties arising from backfilling, and any gap effects.
iii) Where it is judged necessary to model the supporting soil media
as layered media, the stiffnesses are to be estimated through use of
acceptable procedures.
3. The radiation and material energy dissipation (i.e., the damping
values) are considered to be additive for computation convenience. Normally
the material damping can be expected to be about 5 to 8 percent.
The geometric damping (radiation energy dissipation) is recognized to
be frequency-dependent. However, in order to reduce the calculational
.
5
effort (at least initially), and to be sure that excessive damping is not
employed, it is recommended that values of damping be estimated
theoretically (on a frequency-independent basis) as follows.
i) Horizontal to be taken as 75 percent of the theoretical value.*
ii) Vertical to be taken as 75 percent of the theoretical value.*
iii) Rotation (rocking and torsional) to be taken at 100 percent of
the theoretical value.*
In the case of layered systems the approach employed in establishing
these values needs to be justified.
4. The following analysis approaches are considered to be acceptable.
i) When all composite modal damping ratios** are less than 20
percent, modal superposition approaches can be used without any validation
check.
ii) If in investigating the use of modal superposition approaches
it is ascertained that a composite modal damping ratio** exceeds 20 percent,
one must perform a validation analysis. To perform this validation, it is
generally acceptable to use a time-history analysis in which the energy
dissipation associated with the structure is included with the structural
elements, and that associated with the soil is included with the soil
elements.
*As calculated by generally accepted methods, as for example given in the
book Vibrations of Soils and Foundations, by F. E. Richart, Jr., J. R. Hall,
Jr., and R. D. Woods, Prentice-Hall Inc., 1970.
**As defined by generally accepted methods.
.
6
in-structure response spectra obtained from a superposition analysis
employing composite modal damping throughout the frequency range of interest
must be similar to or more conservative than those obtained from the
validation analyses.
It is emphasized that the aforementioned procedures are intended to be
guidelines and may be subject to revision as experience is gained under the
SEP Program in attempting to arrive at relatively economical and simplified
techniques for estimating the possible effects of soil-structure
interaction.
Respectfully submitted by the Senior Seismic Review Team:
N. M. Newmark, Chairman
W. J. Hall
R. P. Kennedy
R. C. Murray
J. D. Stevenson
