2009 NEHRP RECOMMENDED SEISMIC
PROVISIONS FOR NEW BUILDINGS AND
OTHER STRUCTURES:
PART 2, COMMENTARY
TO ASCE/SEI 705
This part of the 2009 NEHRP Recommended Seismic Provisions for New Buildings and Other Structures presents
commentary to ASCE/SEI 705 utilizing the chapter and section numbers of that standard. Commentary to the modifications
of the standard that appear in Part 1 of this Provisions volume is presented at the end of each chapter of modifications and
can be used to replace or add to this Part 2 Commentary (e.g., this Part 2 Commentary addresses the maps that appear in
ASCE/SEI 705, not the new risktargeted maps and procedures presented in Part 1 of this volume).
This commentary is intended primarily for design professionals and members of the codes and standardsdevelopment
community. However, an understanding of the basis for the seismic regulations contained in the nation’s building codes and
standards is important to many outside this technical community including elected officials and other decision makers
responsible for aspects of the built environment, the financial and insurance communities, and individual business owners
and other citizens. These individuals and others who do not have indepth technical knowledge may find a complementary
report that presents a brief overview of the 2009 Provisions of interest. This overview is published as FEMA P749,
Concepts of Earthquakeresistant Design: An Introduction to the NEHRP Recommended Seismic Provisions for New Buildings and Other
Structures.
Page intentionally left blank.
COMMENTARY TO CHAPTER 11,
SEISMIC DESIGN CRITERIA
C11.1 GENERAL
C11.1.1 Purpose. When prescribed wind loading governs the stress or drift design, the resisting system still must conform
to the special requirements for seismicforceresisting systems. This is required in order to resist, in a ductile manner,
potential seismic loads in excess of the prescribed wind loads. A proper, continuous load path is an obvious design
requirement, but experience has shown that it often is overlooked and that significant damage and collapse can result. The
basis for this design requirement is twofold:
1. To ensure that the design has fully identified the seismicforceresisting system and its appropriate design level and
2. To ensure that the design basis is fully identified for the purpose of future modifications or changes in the structure.
Detailed requirements for analyzing and designing this load path are given in the appropriate design and materials chapters.
C11.1.2 Scope. The scope statement establishes in general terms the applicability of ASCE/SEI 705. Certain structures are
exempt for the following reasons:
Exemption 1 – Detached one and twofamily dwellings in Seismic Design Categories A, B, and C, along with those located
where Ss < 0.4g, are exempt because they represent low seismic risks.
Exemption 2 – Structures constructed using the conventional lightframe construction requirements in Section 12.5 are
deemed capable of resisting the anticipated seismic forces. While specific elements of conventional lightframe construction
may be calculated to be overstressed, typically there is a great deal of redundancy and uncounted resistance in such
structures. Detached one and twostory woodframe dwellings generally have performed well even in regions of higher
seismicity. Section 12.5 adequately provides the level of safety required for such dwellings without imposing any additional
requirements.
Exemption 3 – Agricultural storage structures generally are exempt from most code requirements because of the
exceptionally low risk to human life involved.
Exemption 4 – Bridges, transmission towers, nuclear reactors, and other structures with special configuration and uses are not
covered. The regulations for buildings and buildinglike structures presented in this document do not adequately address the
design and performance of such special structures.
ASCE/SEI 705 is not retroactive and usually applies to existing structures only when there is an addition, change of use, or
alteration. Minimum acceptable seismic resistance of existing buildings is a policy issue normally set by the authority having
jurisdiction. Appendix 11B of the standard contains rules of application for basic conditions. ASCE/SEI 31, Seismic
Evaluation of Buildings, and ASCE/SEI 41, Seismic Rehabilitation of Existing Buildings, provide technical guidance but do
not contain policy recommendations. A chapter in the International Building Code (IBC) applies to alteration, repair,
addition, and change of occupancy of existing buildings, and the International Code Council maintains the International
Existing Building Code (IEBC) and an associated commentary.
C11.1.4 Alternate Materials and Alternate Means and Methods of Construction. It is not possible for a design standard
to provide criteria for the use of all possible materials and their combinations and methods of construction, either existing or
anticipated. While not citing specific materials or methods of construction currently available that require approval, this
section serves to emphasize that the evaluation and approval of alternate materials and methods require a recognized and
accepted approval system. The requirements for materials and methods of construction contained within the document
represent the judgment of the best use of the materials and methods based on wellestablished expertise and historical seismic
performance. It is important that any replacement or substitute be evaluated with an understanding of all the ramifications of
performance, strength, and durability implied by the standard.
It also is recognized that until needed standards and agencies are created, authorities having jurisdiction need to operate on
the basis of the best evidence available to substantiate any application for alternates. If accepted standards are lacking, it is
strongly recommended that applications be supported by extensive reliable data obtained from tests simulating, as closely as
is practically feasible, the actual load and deformation conditions to which the material is expected to be subjected during the
service life of the structure. These conditions, when applicable, should include several cycles of full reversals of loads and
deformations in the inelastic range.
C11.4 SEISMIC GROUND MOTION VALUES1
The approach adopted in Section 11.4 is intended to provide for a uniform margin against collapse at the design ground
motion. In order to accomplish this objective, ground motion hazards are defined in terms of maximum considered
earthquake (MCE) ground motions, which are based on a set of rules that depend on the seismic hazard of a region. Design
ground motions are based on a lower bound estimate of the margin against collapse inherent in structures designed to the
seismic provisions in the standard. This lower bound was judged, based on experience, to correspond to a factor of about 1.5
in ground motion. Consequently, the design earthquake ground motion was selected at a ground shaking level that is 1/1.5
(or 2/3) of the MCE ground motion.
For most regions of the nation, the MCE ground motion is defined with a uniform probability of exceedance of 2 percent in
50 years (return period of about 2500 years). While stronger shaking than this could occur, it was judged that it would be
economically impractical to design for such very rare ground motions and that the selection of the 2 percent probability of
exceedance in 50 years as the MCE ground motion would result in acceptable levels of seismic safety.
In regions of high seismicity, such as in many areas of California, the seismic hazard is typically controlled by largemagnitude
events occurring on a limited number of welldefined fault systems. Probabilistic ground motions calculated at a
2 percent probability of exceedance in 50 years can be much larger than deterministic ground motions computed based on the
characteristic magnitudes of earthquakes on these known active faults. These probabilistic motions are greater if these major
active faults produce characteristic earthquakes every few hundred years. For these regions, it is considered more appropriate
to determine MCE ground motions directly by deterministic methods based on the characteristic earthquakes of these defined
faults. In order to provide an appropriate level of conservatism in the design process when the deterministic approach is used
to calculate MCE ground motion, the median ground motion estimated for the characteristic event is multiplied by 1.5.
1 Note that this section focuses on the methods and design procedures of ASCE/SEI 705 and the 2003 edition of the Provisions;
commentary on the new risktargeted maps and design procedures is presented in Part 1 of this volume following the modifications to
ASCE 7 Section 11.4 and Chapter 22.
C11.4.1 Mapped Acceleration Parameters. In the general procedure, these motions are computed from mapped values of
the spectral response acceleration at short periods, S
http://earthquake.usgs.gov/designmaps.
S
S , and at 1 second, S1 , for Class B sites. These Ss and S1 values may be
obtained directly from Figures 221 through 2214 (in Chapter 22). Development of these maps is explained in detail in
Appendix A of the Part 2 – Commentary volume of the 2003 NEHRP Recommended Provisions. The 2003 Ss and S1 values
also can be obtained from the U.S. Geological Survey (USGS) website:
S is the mapped value of the 5percentdamped MCE spectral response acceleration for shortperiod structures founded on
Site Class B (firm rock) sites. The shortperiod acceleration has been determined at a period of 0.2 second because it was
concluded that 0.2 second was reasonably representative of the shortest effective period of buildings and structures that are
designed using the standard, considering the effects of soil compliance, foundation rocking, and other factors typically
neglected in structural analysis.
Similarly, S1 is the mapped value of the 5percentdamped MCE spectral response acceleration at a period of 1 second on
Site Class B. The spectral response acceleration at periods other than 1 second typically can be derived from the acceleration
at 1 second. Consequently, for MCE ground shaking on Site Class B sites, these two response acceleration parameters, SS
and S1, are sufficient to define an entire response spectrum for the period range of importance for most buildings and
structures.
C11.4.3 and C11.4.4 Site Coefficients and Adjusted Acceleration Parameters. Using the general procedure to obtain
acceleration response parameters that are appropriate for sites with a classification other than Site Class B, the SS and S1
values must be modified as indicated in Section 11.4.3. This modification is performed using two coefficients, Fa and Fv,
that respectively scale the SS and S1 values determined for Site Class B to values appropriate for other site classes. The MCE
spectral response accelerations adjusted for site class are designated SMS and SM1, respectively, for shortperiod and 1secondperiod
response. As described above, structural design in ASCE/SEI 705 is performed for earthquake demands that are 2/3
of the MCE response spectra. As set forth in Section 11.4.4, two additional parameters, SDS and SD1, are used to define the
acceleration response spectrum for this design level event. These parameters are 2/3 of the respective SMS and SM1 values and
define a design response spectrum for sites of any characteristics and for natural periods of vibration less than the transition
period, TL. Values of SMS, SM1, SDS, and SD1 can also be obtained from the USGS website cited above.
The site coefficients, Fa and Fv, presented respectively in Tables 11.41 and 11.42 for the various site classes are based on
the results of empirical analyses of strongmotion data and analytical studies of site response.
The amount of groundmotion amplification by a soil deposit relative to bedrock depends on the wavepropagation
characteristics of the soil, which can be estimated from measurements or inferences of shearwave velocity and in turn the
shear modulus for the materials as a function of the level of shaking. In general, softer soils with lower shearwave velocities
exhibit greater amplifications than stiffer soils with higher shearwave velocities. Increased levels of ground shaking result in
increased soil stressstrain nonlinearity and increased soil damping which, in general, reduces the amplification, especially
for shorter periods. Furthermore, for soil deposits of sufficient thickness, soil amplification is generally greater at longer
periods than at shorter periods.
An extensive discussion of the development of the Fa and Fv site coefficients is presented by Dobry, et al. (2000). Since the
development of these coefficients and the development of a community consensus regarding their values in 1992, earthquake
events have provided additional strongmotion data from which to infer site amplifications. Analyses conducted on the basis
of these more recent data are reported by a number of researchers including Crouse and McGuire (1996), Dobry et al. (1999),
Silva et al. (2000), Joyner and Boore (2000), Field (2000), Steidl (2000), RodriquezMarek et al. (2001), Borcherdt (2002),
and Stewart et al. (2003). Although the results of these studies vary, the site amplification factors are generally consistent
with those in Tables 11.41 and 11.42.
C11.4.5 Design Response Spectrum. The design response spectrum (Figure 11.41) consists of several segments. The
constantacceleration segment covers the period band from To to Ts; response accelerations in this band are constant and
equal to SDS . The constantvelocity segment covers the period band from Ts to TL, and the response accelerations in this
band are proportional to 1/T with the response acceleration at 1sec period equal to SD1. The longperiod portion of the
design response spectrum is defined on the basis of the parameter, TL, the period that marks the transition from the constantvelocity
segment to the constantdisplacement segment of the design response spectrum. Response accelerations in the
constantdisplacement segment, where T = TL, are proportional to 1/T 2. Values of TL are provided on maps in Figures 2215
through 2220.
The TL maps were prepared following a twostep procedure. First, a correlation between earthquake magnitude and TL was
established. Then, the modal magnitude from deaggregation of the groundmotion seismic hazard at a 2second period (1
second period for Hawaii) was mapped. Details of the procedure and the rational for it are found in Crouse et al. (2006).
C11.4.7 SiteSpecific Ground Motion Procedures. The objective of a sitespecific groundmotion analysis is to determine
ground motions for local seismic and site conditions with higher confidence than is possible using the general procedure of
Sections 11.4.
Nearsource effects on horizontal response spectra for periods of vibration greater than approximately 0.5 second include
directivity, which increases ground motions for fault rupture propagating toward the site, and directionality, which increases
ground motions normal (perpendicular) to the strike of the fault. These effects are discussed in Somerville et al. (1997) and
Abrahamson (2000).
C11.5 IMPORTANCE FACTOR AND OCCUPANCY CATEGORY
Large earthquakes are rare events that will include severe ground motions. Such events are expected to result in damage to
structures even if they were designed and built in accordance with the minimum requirements of the standard. The
consequence of structural damage or failure is not the same for the various types of structures located within a given
community. Serious damage to certain classes of structures, such as critical facilities (e.g., hospitals), will disproportionally
affect a community. The fundamental purpose of this section and subsequent requirements that depend on this section is to
improve the ability of a community to recover from a damaging earthquake by tailoring the seismic protection requirements
to the relative importance of that structure. That purpose is achieved by requiring better performance of those structures that:
1. Are necessary to response and recovery efforts immediately following an earthquake,
2. Present the potential for catastrophic loss in the event of an earthquake, or
3. House a very large number of occupants or occupants less able to care for themselves than the average.
The first basis for seismic design in the standard is that structures will have a suitably low likelihood of collapse in the very
rare event defined as the maximum considered earthquake (MCE) ground motion. A second basis is that life threatening
damage, primarily from failure of nonstructural elements in and on structures, will be unlikely in an unusual but less rare
earthquake ground motion, which is given as the design earthquake ground motion (defined as twothirds of the MCE).
Given the occurrence of ground motion equivalent to the MCE, a population of structures built to meet these design
objectives will probably still experience substantial damage in many structures, rendering these structures unfit for occupancy
or use. Experience in past earthquakes around the world has demonstrated that there will be an immediate need to treat
injured people, to extinguish fires and prevent conflagration, to rescue people from severely damaged or collapsed structures,
and to provide sustenance to a population deprived of its normal means. Experience also has shown that these needs are best
met when structures essential to response and recovery activities remain functional.
The standard addresses these objectives by requiring that each structure be assigned to one of the four occupancy categories
presented in Chapter 1 and by assigning an importance factor to the structure based upon that occupancy category. (The two
lowest categories, Ordinary and Low Hazard, are combined for all purposes within the seismic provisions). The occupancy
category is then used as one of two components in determining the Seismic Design Category (see Section C11.6) and is a
primary factor in setting drift limits for building structures under the design earthquake ground motion (see Section C12.12).
Figure C11.51 shows the combined intent of these requirements for design. The vertical scale is the likelihood of the ground
motion with the MCE being the rarest considered. The horizontal scale is the level of performance of the structure and
attached nonstructural components from collapse prevention at the low end to operational at the high end. (These
performance levels are discussed further at other locations in the commentary.) The basic objective of collapse prevention at
the MCE for ordinary structures (Occupancy Category II) is shown at the lower right by the solid triangle; protection from
lifethreatening damage at the design ground motion (defined by the standard as twothirds of the MCE) is shown by the open
triangle. The performance implied for higher occupancy categories is shown by square and circles. The performance
anticipated for less severe ground motion is shown by dotted symbols. The three (net) classes and the numerical values
assigned are far too coarse to assure the portrayed outcome for all structures, but it is judged to be adequate for the purpose
given present limitations of knowledge and tools.
Figure C11.51 Expected performance as related to occupancy category (OC)
and level of ground motion.
C11.5.1 Importance Factor. The importance factor is used throughout the standard in quantitative criteria for strength. In
most of those quantitative criteria, the importance factor is shown as a divisor on the factor R or Rp in order to send a
message to designers that the objective is to reduce damage for important structures in addition to preventing collapse in
larger ground motions. The R and Rp factors adjust the computed linear elastic response to a value appropriate for design; in
many structures, the largest component of that adjustment is ductility (the ability of the structure to undergo repeated cycles
of inelastic strain in opposing directions). Inelastic strain damages a structure so, for a given strength demand, reducing the
effective R factor (by means of the importance factor) increases the required yield strength, thus reducing ductility demand
and related damage.
C11.5.2 Protected Access for Category IV Structures. Those structures considered essential facilities for response and
recovery efforts must be accessible to carry out their purpose. For example, if the collapse of a simple canopy at a hospital
could block ambulances from the emergency room admittance area, the canopy must meet the same structural standard as the
hospital. This requirement must be considered in the siting of essential facilities in densely built urban areas.
Equation
PERFORMANCE LEVEL
Immediate
Occupancy
Collapse
Prevention
Frequent
MCE
Design
GROUND
MOTION
Operational
Life Safety
OC IV: Essential
OC III: High
Occupancy
OCII: Ordinary
C11.6 SEISMIC DESIGN CATEGORIES
Seismic design categories (SDCs) provide a means to step progressively from simple, easily performed design and
construction procedures and minimums to more sophisticated, detailed, and costly requirements as both the level of seismic
hazard and the consequence of failure escalate. The SDCs are used to trigger requirements that are not scalable; such
requirements are either on or off. For example, the basic amplitude of ground motion for design is scalable – the quantity
simply increases in a continuous fashion as one moves from a low hazard area to a high hazard area. However, a requirement
to avoid weak stories is not particularly scalable. Requirements such as this create step functions. There are many such
requirements in the standard, and the SDCs are used systematically to group these step functions. (Further examples include
whether seismic anchorage of nonstructural items is required or not, whether particular inspections will be required or not,
and height limits applied to various structural systems.)
In this regard, SDCs perform one of the functions of the seismic zones used in earlier U.S. building codes and still in use
throughout much of the world. However, SDCs also are dependent on a building’s occupancy and, therefore, its desired
performance. Further, unlike the traditional implementation of seismic zones, the ground motions used to define the SDCs
include the effects of individual site conditions on probable groundshaking intensity.
In developing the groundshaking limits for the various Seismic Design Categories and the design requirements for each, the
equivalent modified Mercalli intensity (MMI) of various shaking spectra were considered. There are now various
correlations of the qualitative MMI with quantitative characterizations of ground. The reader is encouraged to consult any of
a great many sources that describe the MMIs. The following list is a very coarse generalization:
MMI V No real damage
MMI VI Light nonstructural damage
MMI VII Hazardous nonstructural damage
MMI VIII Hazardous damage to susceptible structures
MMI IX Hazardous damage to robust structures
When the current design philosophy was adopted (the 1997 edition of the NEHRP Recommended Provisions, FEMA 302,
and Commentary, FEMA 303), the upper limit for SDC A was set at roughly onehalf of the lower threshold for MMI VII
and the lower limit for SDC D was set at roughly the lower threshold for MMI VIII. However, the lower limit for SDC D
was more consciously established by equating that design value (twothirds of the MCE) to onehalf of what had been the
maximum design value in building codes over the period of 1975 to 1995. As more correlations between MMI and
numerical representations of ground motion have been created, it is reasonable to make the following correlation between the
MMI at MCE ground motion and the Seismic Design Category (all this discussion is for ordinary occupancies):
MMI V SDC A
MMI VI SDC B
MMI VII SDC C
MMI VIII SDC D
MMI IX SDC E
An important change was made to the determination of SDC when the current design philosophy was adopted.
Earlier editions of the Provisions utilized the peak velocityrelated acceleration, Av, to determine a building’s
Seismic Performance Category. However, this coefficient does not adequately represent the damage potential of
earthquakes on sites with soil conditions other than rock. Consequently, the 1997 Provisions adopted the use of
response spectral acceleration parameters SDS and SD1, which include site soil effects for this purpose.
Except for the lowest level of hazard (SDC A), the SDC also depends on the occupancy categories. For a given
level of ground motion, the SDC is one category higher for Occupancy Category IV structures than for lowerrisk
structures. This has the effect of increasing the confidence that the design and construction requirements will
deliver the intended performance in the extreme event.
Note that the tables in the standard are at the design level, defined as twothirds of the MCE level. Also recall that
the MMIs are qualitative by their nature and that the above correlation will be more or less valid depending on
which numerical correlation for MMI is used. The numerical correlations for MMI roughly double with each step so
correlation between design earthquake ground motion and MMI is not as simple or convenient.
In sum, at the MCE level, SDC A structures should not see motions that are normally destructive to structural systems,
whereas the MCE level motions for SDC D structures can destroy vulnerable structures. The grouping of step function
requirements by SDC is such that there are a few basic structural integrity requirements imposed at SDC A graduating to a
suite of requirements at SDC D based upon observed performance in past earthquakes, analysis, and laboratory research.
The nature of ground motions within a few kilometers of a fault can be very different from more distant motions. For
example, some near fault motions will have strong velocity pulses, associated with forward rupture directivity, that tend to be
highly destructive to irregular structures even if they are well detailed. For ordinary occupancies, the boundary between
SDCs D and E is set to define sites likely to be close enough to a fault that these unusual ground motions may be present.
Note that this boundary is defined in terms of mapped bedrock outcrop motions affecting response at 1 second, not site
adjusted values, in order to better discriminate between sites near and far from faults. Shortperiod response is not normally
as affected as the longer period response. The additional design criteria imposed on structures in SDCs E and F specifically
are intended to provide acceptable performance under these very intense nearfault ground motions.
For most buildings, the SDC is determined without consideration of the building’s period. Structures are assigned to a SDC
based on the more severe condition determined from 1second acceleration and shortperiod acceleration. This is done for
several reasons. Perhaps the most important of these is that it is often difficult to estimate precisely the period of a structure
using default procedures contained in the standard. Consider, for example, the case of rigid wall/flexible diaphragm
buildings including lowrise reinforced masonry and concrete tiltup buildings with either untopped metal deck or wood
diaphragms. The formula in the standard for determining the period of vibration of such buildings is based solely on the
height of the structure and the length of wall present. These formulas typically indicate very short periods for such structures,
often on the order of 0.2 second or less. However, the actual dynamic behavior of these buildings often is dominated by the
flexibility of the diaphragm – a factor neglected by the approximate period formula. Large buildings of this type can have
actual periods on the order of 1 second or more. In order to avoid misclassifying a building’s SDC by inaccurately estimating
the structural period, the standard generally requires that the more severe SDC determined on the basis of short and longperiod
shaking be used.
Another reason for this requirement is a desire to simplify building regulation by requiring all buildings on a given soil
profile in a particular region to be assigned to the same SDC regardless of the structural type. This has the advantage of
permitting uniform regulation of structural system selection, inspection and testing requirements, seismic design
requirements for nonstructural components, and similar aspects of the design process regulated on the basis of SDC, within a
community.
Notwithstanding the above, it is recognized that classification of a building as SDC C instead of B or D can have significant
impact on the cost of construction. Therefore, the 2005 edition of the standard includes an exception permitting the
classification of buildings that can reliably be classified as having short structural periods on the basis of shortperiod shaking
alone.
Local or regional jurisdictions enforcing building regulations may desire to consider the effect of the maps, typical soil
conditions, and Seismic Design Categories on the practices in their jurisdictional areas. For reasons of uniformity of practice
or reduction of potential errors, adopting ordinances could stipulate particular values of ground motion, particular site classes,
or particular Seismic Design Categories for all or part of the area of their jurisdiction. For example:
1. An area with a historical practice of high seismic zone detailing might mandate a minimum SDC of D regardless of
ground motion or site class.
2. A jurisdiction with low variation in ground motion across the area might stipulate particular values of ground motion
rather than requiring use of the maps.
3. An area with unusual soils might require use of a particular Site Class unless a geotechnical investigation proves a better
Site Class.
C11.7 DESIGN REQUIREMENTS FOR SEISMIC DESIGN CATEGORY A
Seismic Design Category A is assigned when the MCE ground motions are well known to be below those normally
associated with hazardous damage. Damaging earthquakes are not unknown or impossible in such regions, however, and
ground motions close to such events may be large enough to produce serious damage. Providing a minimum level of
resistance reduces both the radius over which the ground motion exceeds structural capacities and resulting damage in such
rare events. There are reasons beyond seismic risk for minimum levels of structural integrity.
The requirements for SDC A are all minimum strengths for structural elements stated as forces at the level appropriate for
direct use in the strength design load combinations. The two fundamental requirements are a minimum strength for a
structural system to resist lateral forces and a minimum strength for connections of structural members.
For many buildings the wind force will control the strength of the lateralforceresisting system but, for lowrise buildings of
heavy construction with large plan aspect ratio, the minimum lateral force specified here may control. Note that the
requirement is for strength and not for toughness, energy dissipation capacity, or some measure of ductility. The force level
is not tied to any postulated seismic ground motion. The boundary between SDCs A and B is based on a spectral response
acceleration of 25 percent of gravity (MCE level) for shortperiod structures; clearly the 1 percent acceleration level
(Equation 11.71) is far smaller. For ground motions below the A/B boundary, the spectral displacements generally are on
the order of a few inches or less depending on period. Experience has shown that even a minimal strength is beneficial in
providing resistance to small ground motions, and it is an easy provision to implement in design. The low probability of
motions greater than the MCE is a factor in taking the simple approach without requiring details that would produce a ductile
response. Another factor is that larger design forces are specified for connections between main elements of the lateral force
load path.
The minimum connection force is specified in three ways: a general minimum horizontal capacity for all connections; a
special minimum for horizontal restraint of beams and trusses in line, which also includes the live load on the member; and a
special minimum for horizontal restraint of concrete and masonry walls perpendicular to their plane. The 5 percent
coefficient used for the first two is a simple and convenient value that provides some margin over the minimum strength of
the system as a whole. The value for anchorage of concrete and masonry walls is simply scaled upward from the value of
200 pounds per linear foot traditionally used in past building codes for allowable stress design.
C11.8 GEOLOGIC HAZARDS AND GEOTECHNICAL INVESTIGATION
In addition to this commentary, Part 3 of the 2009 NEHRP Recommend Provisions includes additional and more detailed
discussion and guidance on evaluation of geologic hazards and determination of seismic lateral pressures.
C11.8.1 Site Limitation for Seismic Design Categories E and F. Because of the difficulty of designing a structure for the
direct shearing displacement of fault rupture and the relatively high seismic activity of SDCs E and F, locating a structure on
an active fault having the potential to cause rupture of the ground surface at the structure is prohibited.
C11.8.3 Additional Geotechnical Investigation Report Requirements for Seismic Design Categories D through F. The
dynamic lateral earth pressure on basement and retaining walls during earthquake ground shaking is considered to be an
earthquake load, E, for use in design load combinations. This dynamic earth pressure is superimposed on the preexisting
static lateral earth pressure during ground shaking. The preexisting static lateral earth pressure is considered to be an H load.
Liquefaction potential should be evaluated for design earthquake ground motions consistent with peak ground accelerations
of SDS/2.5. The occurrence and consequences of geologic hazards for MCE ground motions also should be considered when
evaluating structural stability and other pertinent performance criteria.
REFERENCES
Abrahamson, N.A. 2000. “Effects of Rupture Directivity on Probabilistic Seismic Hazard Analysis” in Proceedings of the
6th International Conference on Seismic Zonation, Palm Springs, California.
Borcherdt, R. D. 2002. “Empirical Evidence for Site Coefficients in Buildingcode Provisions,” Earthquake Spectra,
18(2):189217.
Crouse, C. B., and J. W. McGuire. 1996. “Site Response Studies for Purposes of Revising NEHRP Seismic Provisions,”
Earthquake Spectra, 12(3).
Crouse, C. B., E. V. Leyendecker, P. G. Somerville, M. Power, and W. J. Silva. 2006. “Development of Seismic Ground
Motion Criteria for the ASCE/SEI 7 Standard,” Paper 533 in Proceedings 8th U.S. National Conference on Earthquake
Engineering, April 1822, 2006, San Francisco, California.
Dobry, R., R. Ramos, and M. S. Power. 1999. Site Factors and Site Categories in Seismic Codes, Technical Report
MCEER990010. Multidisciplinary Center for Earthquake Engineering Research.
Dobry, R., R. Borcherdt, C. B. Crouse, I. M. Idriss, W. B. Joyner, G. R. Martin, M. S. Power, E. E. Rinne, and R. B. Seed.
2000. “New Site Coefficients and Site Classifications System Used in Recent Building Seismic Code Provisions,”
Earthquake Spectra, 16(1):4167.
Field, E. H. 2000. “A Modified Ground Motion Attenuation Relationship for Southern California that Accounts for Detailed
Site Classification and a Basin Depth Effect,” Bulletin of the Seismological Society of America, 90:S209S221.
Joyner, W. B., and D. M. Boore. 2000, “Recent Developments in Earthquake Ground Motion Estimation in Proceeding of
the 6th International Conference on Seismic Zonation, Palm Springs, California.
RodriguezMarek, A., J. D. Bray, and N. Abrahamson. 2001. “An Empirical Geotechnical Site Response Procedure,”
Earthquake Spectra, 17(1):6587.
Silva, W., R. Darragh, N. Gregor, G. Martin, N. Abrahamson, and C. Kircher. 2000. Reassessment of Site Coefficients and
Nearfault Factors for Building Code Provisions, Program Element II, Report 98HQGR1010 to the U.S. Geological
Survey.
Somerville, P. G., N. F. Smith, R. W. Graves, and N. A. Abrahamson. 1997. “Modification of Empirical Strong Ground
Motion Attenuation Relations to Include the Amplitude and Duration Effects of Rupture Directivity,” Seismological
Research Letters, 68:199222.
Stewart, J. P, A. H. Liu, and Y. Choi. 2003. “Amplification Factors for Spectral Acceleration in Tectonically Active
Regions,” Bulletin of the Seismological Society of America, 93(1):332352.
Steidl, J. H. 2000. “Site Response in Southern California for Probabilistic Seismic Hazard Analysis, Bulletin of the
Seismological Society of America, 90:S149S169.
COMMENTARY TO CHAPTER 12,
SEISMIC DESIGN REQUIREMENTS
FOR BUILDING STRUCTURES
C12.1 STRUCTURAL DESIGN BASIS
The performance expectations for structures designed in accordance with ASCE/SEI 705 are described in Sections C11.1
and C11.5. Structures designed in accordance with the standard are likely to have a low probability of collapse but suffer
serious damage if subjected to the maximum considered earthquake (MCE) or stronger ground motion. The uncertainty in
performance results from variability of both ground motion and structural characteristics.
Earthquakes load structures indirectly. As the ground displaces, a structure follows and vibrates. The vibration produces
structural deformations with associated strains and stresses. Computation of dynamic response to earthquake ground shaking
is complex. The basic methods of analysis in the standard employ the common simplification of a response spectrum. A
response spectrum for a specific earthquake ground motion approximates the maximum value of response to that ground
motion for simple structures without reflecting the total time history of response. The design response spectrum specified in
Section 11.4 and used in the basic methods of analysis in Chapter 12 is a smoothed and normalized approximation for many
different ground motions.
Although the seismic requirements of the standard are stated in terms of forces and loads, there are no external forces applied
to the aboveground portion of a structure during an earthquake. The design forces are intended only as approximations to
generate internal forces suitable for proportioning the strength of structural elements and for estimating the deformations
(when multiplied by the deflection amplification factor, Cd) that would occur in the same structure in the event of designlevel
(not MCE) ground motion.
C12.1.1 Basic Requirements. Chapter 12 of the standard sets forth a set of coordinated requirements that must be used
together. The basic steps in structural design for acceptable seismic resistance are as follows:
1. Select gravity and seismicforceresisting systems appropriate to the anticipated intensity of ground shaking. Section
12.2 sets forth limitations depending on the Seismic Design Category.
2. Lay out these systems to produce a continuous, regular, and redundant load path so that the structures act as integral units
in responding to ground shaking. Section 12.3 addresses configuration and redundancy issues.
3. Analyze a mathematical model of the structure subjected to lateral seismic motions and gravity forces. Sections 12.6 and
12.7 set forth requirements for the method of analysis and for construction of the mathematical model.
4. Proportion members and connections to have adequate lateral and vertical strength and stiffness. Section 12.4 specifies
how the effects of gravity and seismic loads are to be combined to establish required strengths, and Section 12.12
specifies deformation limits for buildings.
One to threestory structures with shear wall or braced frame systems of simple configuration may be eligible for design
under the simplified alternative contained in Section 12.14. Any other deviations from the requirements of Chapter 12 are
subject to approval and must be rigorously consistent as specified in Section 11.1.4.
The baseline seismic forces for proportioning structural elements (individual members, connections, and supports) are static
horizontal forces derived from a linear elastic response spectrum procedure. A basic requirement is that horizontal motion
can come from any direction, with detailed requirements being provided in Section 12.5. For most structures, the effect of
vertical ground motions is not analyzed specifically; it is included in an approximate fashion by adjusting the load factors for
dead load up and down, as described in Section 12.4. Certain conditions requiring more detailed analysis of vertical response
are defined in Chapters 13 and 15 for nonstructural components and nonbuilding structures, respectively.
Higher levels of seismic analysis are permitted (and encouraged) for any structure and are required for some structures (see
Section 12.6), but lower limits based on the equivalent lateral force procedures apply. The basic procedure uses response
spectra that are representative of, but substantially reduced from, the anticipated ground motions. As a result, at the MCE
level of ground shaking, structural elements are expected to yield, buckle, or otherwise behave inelastically.
This approach has substantial historical precedent. In past earthquakes, structures with appropriately ductile, regular,
continuous systems designed for reduced forces have performed acceptably. In the standard, such design forces are
computed by dividing the forces that would be generated in a structure behaving linearly when subjected to the design ground
motion by the response modification coefficient, R, and the design ground motion is taken as twothirds of the MCE ground
motion.
The elastic deformations calculated under these reduced design forces are multiplied by the deflection amplification factor,
Cd, to estimate the deformations likely to result from the design ground motion. As set forth in Sections 12.12 and 13, the
amplified deformations are used to assess story drifts and to determine seismic demands on elements of the structure that are
not part of the seismicforceresisting system and on nonstructural components within structures. Where Cd is substantially
less than R, the system is considered to have damping greater than the nominal 5 percent of critical damping.
The seismicforceresisting system is expected to reach significant yield for forces in excess of the design forces. Significant
yield is the point where complete plastification of the most critical region of the structure (e.g., formation of a first plastic
hinge in the structure) occurs, not the point where first yield occurs in any member. Figure C12.11 shows the lateral force
versus deformation relation for a typical structure. Significant yield is shown as the lowest yield hinge on the forcedeformation
diagram. With increased lateral loading, additional plastic hinges form and the resistance increases (following
the solid curve) until a maximum is reached. The maximum resistance developed along the curve is substantially higher than
that at first significant yield, and the margin is referred to as the overstrength capacity.
The provisions of the standard contemplate a seismicforceresisting system with redundant characteristics wherein
significant structural overstrength above the level of significant yield can be obtained by plastification at other points in the
structure prior to the formation of a complete mechanism. The overstrength obtained by this continued inelastic action
provides the reserve strength necessary for the structure to resist the extreme motions of the actual seismic forces that may be
generated by the design ground motion.
The structural overstrength described above results from the development of sequential plastic hinging in a properly
designed, redundant structure. Several other sources will further increase structural overstrength. First, material overstrength
(i.e., actual material strengths higher than the nominal material strengths specified in the design) may increase the structural
overstrength significantly. For example, a recent survey shows that the mean yield strength of A36 steel is about 30 to 40
percent higher than the minimum specified strength used in design calculations. Second, member design strengths usually
incorporate a strength reduction (or resistance) factor, F, to produce a low probability of failure under design loading. Third,
designers themselves introduce additional overstrength by selecting sections or specifying reinforcing patterns that exceed
those required by the computations. Similar situations occur where prescriptive minimums of the standard, or of the design
standards referenced from it, control the design. Finally, the design of many flexible structural systems (e.g., moment
resisting frames) often is controlled by the drift rather than strength limitations of the standard with sections selected to
control lateral deformations rather than to provide the specified strength.
The result is that structures typically have a much higher lateral resistance than that specified as a minimum by the standard,
and first significant yielding of structures may occur at lateral load levels that are 30 to 100 percent higher than the prescribed
design seismic forces. If provided with adequate ductile detailing, redundancy and regularity, full yielding of structures may
occur at load levels that are two to four times the prescribed design force levels.
Most structural systems have some components or limit states that cannot provide reliable inelastic response or energy
dissipation. Such components or limit states must be designed considering that the actual forces in the structure will be larger
than those at first significant yield. The standard specifies an overstrength factor, O0, to amplify the prescribed forces for
use in design of such components or limit states. This specified overstrength factor is neither an upper nor a lower bound; it
is simply an approximation specified to provide a nominal degree of protection against undesirable behavior.
Figure C12.11 illustrates the significance of design parameters contained in the standard including the response modification
coefficient, R; the deflection amplification factor, Cd; and the system overstrength factor, O0. These design values, provided
in Table 12.21, as well as the criteria for story drift and Pdelta effects, have been established considering the characteristics
of typical properly designed structures. The actual structural overstrength, O, often will be less than the tabulated factor, O0.
This means that the required ductility, Rd, usually will exceed R/O0. If excessive “optimization” of a structural design is
performed with lateral resistance provided by only a few elements, the successive yield hinge behavior depicted in Figure
C12.11 will not be able to form, the actual overstrength (O) will be small, and use of the design parameters in the standard
may not provide the intended seismic performance.
The response modification coefficient, R, represents the ratio of the forces that would develop under the specified ground
motion if the structure had entirely linearelastic response to the prescribed design forces (see Figure C12.11). The structure
must be designed so that the level of significant yield exceeds the prescribed design force. The ratio R, expressed as R = VE
/VS, is always larger than 1.0; thus, all structures are designed for forces smaller than those the design ground motion would
produce in a structure with completely linearelastic response. This reduction is possible for a number of reasons. As the
structure begins to yield and deform inelastically, the effective period of response of the structure lengthens which, for most
structures, results in a reduction in strength demand. Furthermore, the inelastic action results in a significant amount of
energy dissipation (hysteretic damping) in addition to other sources of damping present below significant yield. The
combined effect, which is also known as the ductility reduction, explains why a properly designed structure with a fully
Figure C12.11 Inelastic forcedeformation curve.
Lateral
seismic
force, V
VE
Lateral deformation
(drift),
Fully yielded strength
.
Design force level
E
Successive yield
hinges
Elastic response
of structure
.
. , S
drift
under
design
forces
Vy
VS
E
S
R V
V
=
E
d
y
R V
V
=
y
S
V
V
O =
D
d
S
C
= .
.
. , D
design
drift
yielded strength (Vy in Figure C12.11) that is significantly lower than the elastic seismic force demand (VE in Figure C12.1
1) can be capable of providing satisfactory performance under the design ground motion excitations.
Figure C12.11 Inelastic forcedeformation curve.
The energy dissipation resulting from hysteretic behavior can be measured as the area enclosed by the forcedeformation
curve of the structure as it experiences several cycles of excitation. Some structures have far more energy dissipation
capacity than others. The extent of energy dissipation capacity available depends largely on the amount of stiffness and
strength degradation the structure undergoes as it experiences repeated cycles of inelastic deformation. Figure C12.12
shows representative load deformation curves for two simple substructures such as a beamcolumn assembly in a frame.
Hysteretic curve (a) in the figure is representative of the behavior of substructures that have been detailed for ductile
behavior. The substructure can maintain nearly all of its strength and stiffness over several large cycles of inelastic
deformation. The resulting forcedeformation “loops” are quite wide and open, resulting in a large amount of energy
dissipation. Hysteretic curve (b) represents the behavior of a substructure that has not been detailed for ductile behavior. It
loses stiffness rapidly under inelastic deformation, and the resulting hysteretic loops are quite pinched. Such a substructure
has much less energy dissipation than that for the substructure (a) but has a greater change in response period. The structural
response is determined by a combination of energy dissipation and period modification.
The R values in the standard are based largely on engineering judgment of the performance of the various materials and
systems in past earthquakes. The R factor for a specific project should be chosen and used with care. For example, lower
values should be used for structures possessing a low degree of redundancy wherein all the plastic hinges required for the
formation of a mechanism may be formed essentially simultaneously and at a force level close to the specified design
strength. This situation can result in considerably more detrimental Pdelta effects. Since it is difficult for individual
designers to judge the extent to which R factors should be adjusted based on the inherent redundancy of their designs, Section
12.3.4 provides a coefficient, ., that is calculated based on the removal of individual seismicforceresisting elements.
C12.1.2 Member Design, Connection Design, and Deformation Limit. Given that key elements of the seismicforceresisting
system will likely yield in response to ground motions as discussed in Section C12.1.1, it might be expected that
structural connections would be required to develop the strength of connected members. Although that is a logical procedure,
it is not a general requirement. The actual requirement varies by system and generally is specified in the standards for design
of the various structural materials cited by reference in Section 14. Good seismic design requires careful consideration of this
issue.
Figure C12.12 Typical hysteretic curves.
Force Force
Displacement Displacement
a. Ductile hysteresis loops b. Pinched hysteresis loops
Figure C12.12 Typical hysteretic curves.
C12.1.3 Continuous Load Path and Interconnection. In effect, Section 12.1.3 calls for the seismic design to be complete
and in accordance with the principles of structural mechanics. The loads must be transferred rationally from their point of
origin to the final point of resistance. This should be obvious, but it often is overlooked by those inexperienced in earthquake
engineering. Design consideration should be given to potentially adverse effects where there is a lack of redundancy. Given
the many unknowns and uncertainties in the magnitude and characteristics of earthquake loading, in the materials and
systems of construction for resisting earthquake loadings and in the methods of analysis, good earthquake engineering
practice has been to provide as much redundancy as possible in the seismicforceresisting system of buildings. Redundancy
plays an important role in determining the ability of the building to resist earthquake forces. In a structural system without
redundant components, every component must remain operative to preserve the integrity of the building structure. On the
other hand, in a highly redundant system, one or more redundant components may fail and still leave a structural system that
retains its integrity and can continue to resist lateral forces, albeit with diminished effectiveness.
While a redundancy requirement is included in Section 12.3.4, overall system redundancy can be improved by making all
joints of the vertical loadcarrying frame moment resisting and incorporating them into the seismicforceresisting system.
These multiple points of resistance can prevent a catastrophic collapse due to distress or failure of a member or joint. (The
overstrength characteristics of this type of frame are discussed in Section C12.1.1.) The designer should be particularly
aware of the proper selection of R when using only one or twobay rigid frames in one direction for resisting seismic loads.
A single, onebay frame or a pair of such frames provides little redundancy so the designer may wish to consider a reduced R
to account for a lack of redundancy if the calculated redundancy is considered to be too low. As more onebay frames are
added to the system, however, overall system redundancy increases. The increase in redundancy is a function of frame
placement and total number of frames.
The minimum connection forces are not intended to be applied simultaneously to the entire seismicforceresisting system.
C12.1.4 Connection to Supports. The requirement is the same as given in Section 11.7.4 for Seismic Design Category A.
See Section C11.7.
C12.1.5 Foundation Design. Most foundation design criteria are still stated in terms of allowable stresses, and the forces
computed in the standard are all based on the strength level of response. When developing strengthbased criteria for
foundations, all the factors cited in Section 12.1.5 require careful consideration. Section C12.13 provides specific guidance.
C12.1.6 Material Design and Detailing Requirements. The design limit state for resistance to an earthquake is unlike that
for any other load within the scope of the standard. The earthquake limit state is based on overall system performance, not
member performance, where repeated cycles of inelastic straining are accepted as an energy dissipating mechanism.
Provisions that modify customary requirements for proportioning and detailing structural members and systems are provided
to produce the desired performance.
C12.2 STRUCTURAL SYSTEM SELECTION
C12.2.1 Selection and Limitations. For purposes of these seismic analyses and design requirements, seismicforceresisting
systems are grouped into categories as shown in Table 12.21. These categories are subdivided further for various
types of vertical elements used to resist seismic forces. In addition, the sections for detailing requirements are specified.
Specification of R factors requires considerable judgment based on knowledge of actual earthquake performance as well as
research studies. The factors in Table 12.21 continue to be reviewed in light of recent research results. R values for the
various systems were selected considering observed performance during past earthquakes, the toughness (ability to dissipate
energy without serious degradation) of the system, and the amount of damping typically present in the system when it
undergoes inelastic response. FEMA P695, Quantification of Building Seismic Performance Factors (Applied Technology
Council, 2009) has been developed with the purpose of establishing and documenting a methodology for quantifying building
system performance and response parameters for use in seismic design. While the response modification coefficient (R
factor) is a key parameter being addressed, related design parameters such as the system overstrength factor (O0) and
deflection amplification factor (Cd) also are addressed. Collectively, these terms are referred to as “Seismic Performance
Factors” (SPFs). Future systems will likely derive their SPFs using this methodology and existing system SPFs also may be
reviewed in light of this new procedure.
Building height limits have been specified in codes and standards for over 50 years. The structural system limitations and
building height limits specified in Table 12.21 evolved from these initial limitations and were further modified by the
collective expert judgment of the PUC and the ATC3 project team (the forerunners of the PUC). They have continued to
evolve over the past 30 years based on observations and testing, but the specific values are based on subjective judgment.
In a bearing wall system, major loadcarrying columns are omitted and the walls carry a major portion of the gravity (dead
and live) loads. The walls supply inplane lateral stiffness and strength to resist wind and earthquake loads as well as other
lateral loads. In some cases, vertical trusses are employed to augment lateral stiffness. In general, this system has
comparably lower values of R than other systems due to the frequent lack of redundancy for support of vertical and horizontal
loads.
In a building frame system, gravity loads are carried primarily by a frame supported on columns rather than by bearing walls.
Some portions of the gravity load may be carried on bearing walls, but the amount carried should represent a relatively small
percentage of the floor or roof area. Lateral resistance is provided by shear walls or braced frames. Lightframed walls with
shear panels are intended for use only with wood and steel building frames. Although gravityloadresisting systems are not
required to provide lateral resistance, most of them do. To the extent that the gravityloadresisting system provides
additional lateral resistance, it will enhance the building’s seismic performance capability, so long as it is capable of resisting
the resulting stresses and undergoing the associated deformations.
In a momentresisting frame system, momentresisting connections between the columns and beams provide lateral
resistance. In Table 12.21, such frames are classified as ordinary, intermediate, or special. In high Seismic Design
Categories, the anticipated ground motions are expected to produce large inelastic demands so special moment frames
designed and detailed for ductile response in accordance with Chapter 14 are required. In low Seismic Design Categories, the
inherent overstrength in typical structural designs is such that the anticipated inelastic demands are reduced somewhat, and
less ductile systems may be employed safely. Since these less ductile ordinary framing systems do not possess as much
toughness, lower R values are specified.
The R, O0, and Cd values for the composite systems in Table 12.21 are similar to those for comparable systems of structural
steel and reinforced concrete. Use of the tabulated values is allowed only when the design and detailing requirements in
Section 14.3 are followed.
In a dual system, a threedimensional space frame made up of columns and beams provides primary support for gravity loads.
Primary lateral resistance is supplied by shear walls or braced frames, and secondary lateral resistance is provided by a
moment frame complying with the requirements of Chapter 14.
Where a beamcolumn frame or slabcolumn frame lacks special detailing, it cannot act as an effective backup to a shear wall
subsystem so there are no dual systems with ordinary moment frames. Instead, Table 12.21 permits the use of a shear wallframe
interactive system with ordinary reinforced concrete moment frames and ordinary reinforced concrete shear walls. Use
of this defined system, which requires compliance with Section 12.2.5.10, offers a significant advantage over a simple
combination of the two constituent ordinary reinforced concrete systems. Where those systems are simply combined, Section
12.2.3.2 would require use of design parameters for an ordinary reinforced concrete moment frame.
In a cantilevered column system, stability of mass at the top is provided by one or more columns with base fixity acting as a
singledegreeoffreedom system.
Cantilever column systems are essentially a special class of momentresisting frame except that they do not possess the
redundancy and overstrength that most momentresisting frames derive from sequential formation of yield or plastic hinges.
Where a typical momentresisting frame must form multiple plastic hinges in members in order to develop a yield
mechanism, a cantilever column system develops hinges only at the base of the columns to form a mechanism. As a result,
their overstrength is limited to that provided by material overstrength and by design conservatism.
It is permitted to construct cantilever column structures using any of the systems that can be used to develop moment frames
including ordinary, intermediate, and special steel and concrete detailing systems as well as timber frames. The system
limitations for cantilever column systems reflect the type of moment frame detailing provided but with a height limit of 35
feet.
The R factor for cantilever column systems is derived from momentresisting frame values where R is divided by O0 but is
not taken as less than 1 or greater than 3. This accounts for the lack of sequential yielding in such systems. Cd is taken as
equal to R, recognizing that damping is quite low in these systems and inelastic displacement of these systems will not be less
than the elastic displacement.
C12.2.2 Combinations of Framing Systems in Different Directions. Different systems can be utilized along each of the
two orthogonal directions as long as the respective R, O0, and Cd values are used. Depending on the combination selected, it
is possible that one of the two systems will limit the extent of the overall system with regard to use and height. The more
restrictive of the limitation systems governs.
C12.2.3 Combinations of Framing Systems in the Same Direction.
C12.2.3.1 R, O0, and Cd Values for Vertical Combinations. The intent of the provision requiring us of the more stringent
seismic design parameters (R, O0, and Cd) is to prevent mixed systems that could concentrate inelastic behavior in the lower
stories. Exceptions to these requirements exist for conditions that do not affect the dynamic characteristics of the structure or
that will not result in concentration of inelastic demand in critical areas.
For the past several decades, building codes have allowed twostage static analysis for certain structures with a vertical
combination of dynamically uncoupled systems. While this approach may be used for any structure that meets the
requirements, it is most often used for the design of lightframed construction built on a rigid concrete base. The design
process requires that the “flexible” upper structure and “rigid” lower structure be designed separately with the reactions from
the upper portion amplified by the ratio of respective R/. values. This ratio, which must be taken as no less than 1, produces
demands for the “rigid” lower portion that are commensurate with its inelastic capability.
C12.2.3.2 R, O0, and Cd Values for Horizontal Combinations. For nearly all conditions, the least value of R of different
structural systems in the same direction must be used in design. This requirement reflects the expectation that the entire
system will undergo the same deformation with its behavior controlled by the least ductile system. However, where the listed
conditions are met, the R value for each independent line of resistance can be used. This exceptional condition is consistent
with lightframe construction that utilizes the ground for parking with residential use above.
C12.2.4 Combination of Framing Detailing Requirements. This requirement is provided so that the higher R value
system has the necessary ductile detailing throughout. The intent is that details common to both systems be designed to
remain functional throughout the response in order to preserve the integrity of the seismicforceresisting system.
C12.2.5 System Specific Requirements.
C12.2.5.1 Dual System. The moment frame of a dual system must be capable of resisting at least 25 percent of the design
seismic forces; this percentage is based on judgment. The purpose of the 25 percent frame is to provide a secondary lateral
system with higher degrees of redundancy and ductility in order to improve the ability of the building to support the service
loads (or at least the effect of gravity loads) after strong earthquake shaking. The primary system (walls or bracing) acting
together with the moment frame must be capable of resisting all of the design seismic forces. The following analyses are
required for dual systems:
1. The moment frame and shear walls or braced frames must resist the design seismic forces considering fully the force and
deformation interaction of the walls or braced frames and the moment frames as a single system. This analysis must be
made in accordance with the principles of structural mechanics considering the relative rigidities of the elements and
torsion in the system. Deformations imposed upon members of the moment frame by their interaction with the shear
walls or braced frames must be considered in this analysis.
2. The moment frame must be designed with sufficient strength to resist at least 25 percent of the design seismic forces
including torsional effects.
C12.2.5.2 Cantilever Column Systems. Cantilever column systems are singled out for special consideration because of
their unique characteristics. These structures often have limited redundancy and overstrength and concentrate inelastic
behavior at their bases. As a result, they have substantially less energy dissipation capacity than other systems. A number of
apartment buildings incorporating this system experienced very severe damage and, in some cases, collapse in the 1994
Northridge earthquake. Because the ductility of columns having large axial stress is limited, cantilever column systems may
not be used where column axial demands exceed 15 percent of their axial strength.
Elements providing restraint at the base of cantilever columns must be designed with overstrength so that the strength of the
cantilever columns is developed.
C12.2.5.3 Inverted PendulumType Structures. Inverted pendulumtype structures do not have unique entry in Table
12.21 since they can be formed from many structural systems. Inverted pendulumtype structures have more than half of
their mass concentrated near the top (producing one degree of freedom in horizontal translation) and rotational compatibility
of the mass with the column (producing vertical accelerations acting in opposite directions). Dynamic response amplifies this
rotation; hence, the bending moment induced at the top of the column can exceed that computed using the procedures of
Section 12.8. The requirement to design for a top moment that is onehalf of the base moment calculated in accordance with
Section 12.8 is based on analyses of inverted pendulums covering a wide range of practical conditions.
C12.2.5.4 Increased Building Height Limit for Steel Braced Frames and Special Reinforced Concrete Shear Walls.
The first criterion for an increased building height limit precludes extreme torsional irregularity since premature failure of
one of the single walls or frames could lead to excessive inelastic torsional response. The second criterion, which is similar
to the redundancy requirements, is to limit the height of systems that are too strongly dependent on any single line of walls or
braced frames. The inherent torsion resulting from the distance between the center or mass and center of stiffness must be
included, but accidental torsional effects are neglected for ease of implementation.
C12.2.5.5 Special Moment Frames in Structures Assigned to Seismic Design Categories D through F. Special moment
frames, either alone or as part of a dual system, are required to be used in Seismic Design Categories D through F where the
building height exceeds 160 feet (or 240 feet for buildings that meet the provisions of Section 12.2.5.4) as indicated in Table
12.21. In shorter buildings where special moment frames are not required to be used, the special moment frames may be
discontinued and supported on less ductile systems as long as the requirements for system combinations are followed.
For the situation where special moment frames are required, they should be continuous to the foundation. In cases where the
foundation is located below the building’s base, provisions for discontinuing the moment frames can be made as long as the
seismic forces are properly accounted for and transferred to the supporting structure.
C12.2.5.6 SingleStory Steel Ordinary and Intermediate Moment Frames in Structures Assigned to Seismic Design
Category D or E. Ordinary and intermediate moment frames are less ductile than special moment frames; consequently,
restrictions are placed on their use in higher Seismic Design Categories. The height limit of 65 feet and the limitations on
roof and wall dead load are intended to restrict the use of such systems to metal buildings and similar onestory structures, the
design of which is often controlled by wind forces, and which have generally evidenced acceptable performance in past
seismic events.
C12.2.5.7 Other Steel Ordinary and Intermediate Moment Frames in Structures Assigned to Seismic Design
Category D or E. Compared to the limits in Section 12.2.5.6, this section imposes a stricter height limit because higher
loads and additional stories are permitted. Lowrise lightframe structures that are commonly used in residential construction
generally have evidenced adequate performance in past seismic events due to their light weight, abundance of lateral forceresisting
elements, and general resilience.
C12.2.5.8 SingleStory Steel Ordinary and Intermediate Moment Frames in Structures Assigned to Seismic Design
Category F. See Section C12.2.5.6.
C12.2.5.9 Other Steel Intermediate Moment Frame Limitations in Structures Assigned to Seismic Design Category F.
The intent of this section is to prohibit the use of steel ordinary moment frames in lightframe construction that does not
comply with Section 12.2.5.8.
C12.2.5.10 Shear WallFrame Interactive Systems. For structures assigned to Seismic Design Category A or B (where
seismic hazard is low), it is usual practice to design shear walls and frames of a shear wallframe structure to resist lateral
forces in proportion to their relative rigidities, considering interaction between the two subsystems at all levels. As discussed
in Section C12.2.1, this typical approach would require use of a lower R factor than that defined for shear wallframe
interactive systems. Where the special requirements of this section are satisfied, more reliable performance is expected,
justifying a higher R factor.
C12.3 DIAPHRAGM FLEXIBILITY, CONFIGURATION IRREGULARITIES, AND REDUNDANCY
C12.3.1 Diaphragm Flexibility. Most seismicforceresisting systems have two distinct parts: the horizontal system that
distributes lateral forces to the vertical elements and the vertical system that transmits lateral forces between the floor levels
and the base of the structure.
The horizontal system may consist of diaphragms or a horizontal bracing system. For the majority of buildings, diaphragms
offer the most economical and positive method of resisting and distributing seismic forces in the horizontal plane. Typically,
diaphragms consist of metal deck (with or without concrete), concrete slabs, and wood sheathing/decking. While most
diaphragms are flat, consisting of the floors of buildings, they also may be inclined, curved, warped, or folded configurations,
and most diaphragms have openings.
The diaphragm stiffness relative to the stiffness of the supporting vertical seismicforceresisting system ranges from flexible
to rigid and is important to define. Provisions defining diaphragm flexibility are given in Sections 12.3.1.1 through 12.3.1.3.
If a diaphragm cannot be idealized as either flexible or rigid, explicit consideration of its stiffness must be included in the
analysis.
The diaphragms in most buildings braced by wood lightframe shear walls are semirigid. Because semirigid diaphragm
modeling is beyond the capability of available software for wood lightframe buildings, it is anticipated that this requirement
will be met by evaluating force distribution using both rigid and flexible diaphragm models and taking the worst case of the
two. While this is in conflict with common design practice, which typically includes only flexible diaphragm force
distribution for wood lightframe buildings, it is one method of capturing the effect of the diaphragm stiffness.
Further detailed discussion of diaphragms can be found in Delebi, et al. (1980) and in an Applied Technology Council report
on diaphragms (1981).
C12.3.1.2 Rigid Diaphragm Condition. Span length is included in the deemedtocomply condition as an indirect measure
of the flexural contribution to diaphragm stiffness.
C12.3.2 Irregular and Regular Classification. The configuration of a structure can significantly affect its performance
during a strong earthquake producing the ground motion contemplated in the standard. Configuration can be divided into
two aspects: horizontal and vertical. Most seismic design provisions were derived for buildings having regular
configurations, but earthquakes have shown repeatedly that buildings having irregular configurations suffer greater damage.
This situation prevails even with good design and construction. There are several reasons for this poor behavior of irregular
structures. In a regular structure, the inelastic response produced by strong ground shaking, including energy dissipation and
damage, tends to be well distributed throughout the structure. However, in irregular structures, inelastic behavior can be
concentrated by irregularities and result in rapid failure of structural elements in these areas. In addition, some irregularities
introduce unanticipated demands into the structure, which designers frequently overlook when detailing the structural system.
Finally, the elastic analysis methods typically employed in the design of structures often cannot predict the distribution of
earthquake demands in an irregular structure very well, leading to inadequate design in the areas associated with the
irregularity. For these reasons, the standard encourages regular configurations and prohibits gross irregularity in buildings
located on sites close to major active faults where very strong ground motion and extreme inelastic demands are anticipated.
C12.3.2.1 Horizontal Irregularity. A building may have a symmetric geometric shape without reentrant corners or wings
but still be classified as irregular in plan because of its distribution of mass or vertical seismicforceresisting elements.
Torsional effects in earthquakes can occur even where the centers of mass and resistance coincide. For example, ground
motion waves acting on a skew with respect to the building axis can cause torsion. Cracking or yielding in an asymmetric
fashion also can cause torsion. These effects also can magnify the torsion due to eccentricity between the centers of mass and
resistance. Torsional irregularities are defined to address this concern.
A square or rectangular building with minor reentrant corners would still be considered regular, but large reentrant corners
creating a crucifix form would produce an irregular configuration. The response of the wings of this type of building
generally differs from the response of the building as a whole, and this produces higher local forces than would be
determined by application of the standard without modification. Other winged plan configurations (e.g., Hshapes) are
classified as irregular even if symmetric due to the response of the wings.
Significant differences in stiffness between portions of a diaphragm at a level are classified as irregularities since they may
cause a change in the distribution of seismic forces to the vertical components and create torsional forces not accounted for in
the distribution normally considered for a regular building. Figure C12.31 illustrates plan irregularities.
Where there are discontinuities in the path of lateral force resistance, the structure cannot be considered to be regular. The
most critical discontinuity defined is the outofplane offset of vertical elements of the seismicforceresisting system. Such
offsets impose vertical and lateral load effects on horizontal elements that are difficult to provide for adequately.
Where vertical elements of the lateralforceresisting system are not parallel to or symmetric about major orthogonal axes, the
equivalent lateral force procedure of the standard cannot be applied appropriately so the structure is considered to be
irregular.
Figure C12.31 Building plan irregularities.
Rigid Flexible Open
1. Torsional 2. Reentrant corner
3. Diaphragm discontinuity
4. Outofplane offset 5. Nonparallel system
X
Xp
Yp
Y
Irregular:
X
X
p > 0.15
and
Y
Y
p > 0.15
Wall below
Wall above
Irregular:
1
A > 2 XY
max avg
Seismic
force
min
X
Y
Irregular:
>1.4
Extreme:
max
avg
>1.2
= max
avg 2 max
avg
+ min
open
Figure C12.31 Building plan irregularities.
C12.3.2.2 Vertical Irregularity. Vertical configuration irregularities affect the responses at the various levels and induce
loads at these levels that differ significantly from the distribution assumed in the equivalent lateral force procedure given in
Section 12.8. A momentresisting frame building might be classified as having a vertical irregularity if one story is much
taller than the adjoining stories and the design did not compensate for the resulting decrease in stiffness that normally would
occur. Figure C12.32 illustrates vertical irregularities.
A building is classified as irregular where the ratio of mass to stiffness in adjacent stories differs significantly. This might
occur where a heavy mass (e.g., an interstitial mechanical floor) is placed at one level. Irregularity Type 3 in Table 12.32
applies regardless of whether the larger dimension is above or below the smaller one. Buildings with a weakstory
irregularity tend to develop all of their inelastic behavior and consequent damage at the weak story, possibly leading to
collapse. Section 12.3.3.2 provides an exception for Seismic Design Category B or C structures where essentially elastic
response of the weak story is expected.
C12.3.3 Limitations and Additional Requirements for Systems with Structural Irregularities.
C12.3.3.1 Prohibited Horizontal and Vertical Irregularities in Seismic Design Categories D through F. The
irregularity prohibitions of this section stem from poor performance in past earthquakes and the potential to concentrate large
inelastic demands in certain portions of the structure. Even when such irregularities are permitted, they should be avoided
whenever possible in all structures.
Figure C12.32 Building vertical irregularities.
K i+3
K i+2
K i+1
K i
K i+3
K i+2
K i+1
K i
K i+3
K i+2
K i+1
K i
Mi+1
Mi
Mi1
Li+1
Li
2. Weight (Mass) 3. Geometric
4. InPlane Discontinuity 5. Lateral Strength  Weak Story
Irregular:
< 0.7K
or
i+1 Ki
K < i (K + K + K ) 3 i+1
0.8
i+2 i+3
Extreme:
< 0.6K
or
i+1 Ki
K < i (K + K + K ) 3 i+1
0.7
i+2 i+3
1. Stiffness  Soft Story
Stri+1
Stri
Labove
Lbelow
offset
Irregular:
offset > L
or
below
offset > Labove
Irregular:
Str < 0.8Str
Str < 0.65Str
i
i
i+1
i+1
Extreme:
Irregular:
L > 1.3L i i+1
Irregular:
M > 1.5M
or
i i+1
M i > 1.5Mi1
Figure C12.32 Building vertical irregularities.
C12.3.3.2 Extreme Weak Stories. Since extreme weak story irregularities are prohibited for buildings located in Seismic
Design Categories D, E and F, the limitations and exceptions in this section apply only to buildings assigned to Seismic
Design Category B or C.
C12.3.3.3 Elements Supporting Discontinuous Walls or Frames. The purpose of this requirement is to protect the
supporting elements from overload caused by overstrength of a discontinued seismicforceresisting element. Columns,
beams, slabs, or trusses may be subject to such failure so all are included in the design requirement. Overload may result
from forces in either the downward or upward direction; therefore, both possibilities must be considered. Such load reversals
may be especially problematic for reinforced concrete beams, weaker top laminations of glulam beams, unbraced flanges of
steel beams, and steel trusses.
The connection between the discontinuous element and the supporting member must be adequate to transmit the forces for
which the discontinuous element is designed. For example, where the discontinuous element must be designed using the load
combinations of Section 12.4.3, as is the case for a steel column in a braced frame or moment frame, its connection to the
supporting member must be designed using the same load combinations. Since concrete shear walls are not required to be
designed using the load combinations of Section 12.4.3, the connection between a discontinuous shear wall and the
supporting member may be designed using the loads associated with the shear wall and not the load combinations with
overstrength factor.
C12.3.3.4 Increase in Forces Due to Irregularities for Seismic Design Categories D through F. The irregularities listed
may result in loads that are distributed differently than assumed in the equivalent lateral force procedure of Section 12.8,
especially as related to the interconnection of the diaphragm with vertical elements of the seismicforceresisting system.
The 25 percent increase in force is intended to account for this difference. Where the load combinations with overstrength
apply, no further increase is warranted.
C12.3.4 Redundancy. The desirability of redundancy, or multiple lateralforceresisting load paths, has long been
recognized. The redundancy provisions of this section reflect the belief that an excessive loss of story shear strength or
development of an extreme torsional irregularity may lead to structural failure. The redundancy factor determined for each
direction may differ.
C12.3.4.1 Conditions Where Value of . is 1.0. This section provides a convenient list of conditions where . is 1.0.
C12.3.4.2 Redundancy Factor, ., for Seismic Design Category D through F. There are two approaches to establishing a
redundancy factor of 1.0. Where neither condition is satisfied, . is taken equal to 1.3. It is permitted to take . equal to 1.3
without checking either condition.
The first approach is a check of the elements outlined in Table 12.33 for cases where the story shear exceeds 35 percent of
the base shear. Parametric studies (conducted by Building Seismic Safety Council Technical Subcommittee 2 but
unpublished) were used to select the 35 percent value. Those studies indicated that stories with at least 35 percent of the base
shear include all stories of lowrise buildings (buildings up to 5 to 6 stories) and about 87 percent of the stories of tall
buildings. The intent of this limit is to exclude penthouses and the uppermost stories from the redundancy requirements.
This approach requires the removal (or loss of moment resistance) of an individual lateralforceresisting element to
determine its effect on the remaining structure. If the removal of elements, onebyone, does not result in more than a 33
percent reduction in story strength or an extreme torsional irregularity, . may be taken as 1.0. For this evaluation, the
determination of story strength requires an indepth calculation. The intent of the check is to use a simple measure (elastic or
plastic) to determine whether an individual member has a significant effect on the overall system. If the original structure
has an extreme torsional irregularity to begin with, the resulting . is 1.3. Figure C12.33 presents a flowchart for
implementing the redundancy requirements.
As indicated in the table, braced frame, moment frame, shear wall, and cantilever column systems must conform to
redundancy requirements. Dual systems also are included but, in most cases, are inherently redundant. Shear walls or wall
piers with a heighttolength aspect ratio greater than 1.0 within any story have been included; however, the required design
of collector elements and their connections for O0 times the design force may address the key issues. In order to satisfy the
collector force requirements, a reasonable number of shear walls usually is required. Regardless, shear wall systems are
addressed in this section so that either an adequate number of wall elements is included or the proper redundancy factor is
applied. For wall piers, the height is taken as the height of the adjacent opening and generally is less than the story height.
The second approach is a deemedtocomply condition wherein the structure is regular and has a specified arrangement of
seismicforceresisting elements to qualify for . of 1.0. As part of the parametric study, simplified braced frame and moment
frame systems were investigated to determine their sensitivity to the analytical redundancy criteria. This simple deemedtocomply
condition is consistent with the results of the study.
Figure C12.33 Calculation of the redundancy factor, ..
Perform linear analysis
with all elements
Define story X above
which no more than 35% of
base shear is resisted
Below X is item b of
Section 12.3.4.2 satisfied?
Extreme torsional
irregularity?
Does the seismic
forceresisting system
comprise only shear walls
or wall piers with a
heighttolength ratio not
greater than 1.0?
Prioritize elements
based on highest force or
force/story shear
Select an element (below X ) to
remove, and perform linear analysis
without that element
Does the demand in any
remaining element (below X )
increase by more than 50%?
Does plastic mechanism
analysis show that element
removal decreases story
strength by more than 33%?
Have all likely
elements been
considered?
Extreme torsional
irregularity?
. = 1.0 . = 1.3
No*
No
No
No
No
Yes*
Yes
No
Yes
No
Yes
Yes
Yes
* or not considered
Yes
p
p
p
p
Figure C12.33 Calculation of the redundancy factor, ..
Figure C12.34 Shear wall and wall pier heighttolength ratios.
hwall
Lwall
wp
hwp
L
Story level 2
Story level 1
Shear wall
heightto length ratio = wall h
Lwall
Wall pier
heightto length ratio = wp h
Lwp
= height of shear wall
= height of wall pier
= length of shear wall
= length of wall pier
wall h
Lwall
wp h
Lwp
Figure C12.34 Shear wall and wall pier heighttolength ratios.
C12.4 SEISMIC LOAD EFFECTS AND COMBINATIONS
C12.4.1 Applicability. Structural elements designated by the engineer as part of the seismicforceresisting system
typically are designed directly for seismic load effects. None of the seismic forces associated with the design base shear are
formally assigned to structural elements that are not designated as part of the seismicforceresisting system, but such
elements must be designed using the load conditions of Section 12.4 and must accommodate the deformations resulting from
application of seismic loads.
C12.4.2 Seismic Load Effect. Section 12.4 presents the required combinations of seismic forces with other loads. The
load combinations are taken from the basic load combinations of Chapter 2 of the standard with further elaboration of the
seismic load effect, E. The seismic load effect includes horizontal and vertical components. For strength design, the effect of
vertical seismic forces, Ev, is based on an assumed effective vertical acceleration of 0.2SDS times gravity.
It may be helpful to recognize that the quantities Eh and Ev are the effects of loads, not the loads themselves. They can be
tension or compression axial forces, shear, bending moments, or torsional moments. For a onestory shear wall, application
of the horizontal seismic forces from V causes overturning moment and shear in the wall, both of which are Eh effects. The
factor 0.2 SDS times gravity dead load corresponds to an Ev load effect that increases or decreases the axial force in the wall.
In this simple example, an Eh force or moment is never added directly to an Ev force or moment because the former affects
only moment and shear, while the latter affects only axial force.
While the shear and moment are independent of the axial force, the capacity check of the wall may need to include all three
terms (or certainly moment and axial force) simultaneously.
For a diagonal brace that carries earthquake and gravity load, application of the horizontal seismic forces from V causes a
brace force that has both horizontal and vertical components, and the factor 0.2 SDS times dead load produces a load effect
that also affects both the horizontal and vertical components of axial force. In this case the brace force is based on Eh ± Ev.
Section 12.4.2.3 presents the load combinations written using the separate horizontal and vertical load effects that constitute
E.
The 0.2SDS vertical acceleration effect is required to be considered in the design of all members of a structure—even those
that are not part of the seismicforceresisting system. For example, design of a gravity loadresisting prestressed concrete
girder may be governed by the dead and earthquake condition, where 0.2SDSD is subtracted from the dead load. This could be
the controlling condition for tension at the top of the girder.
C12.4.3 Seismic Load Effect Including Overstrength Factor. Certain structural elements or actions, such as collectors in
Seismic Design Categories C through F or columns supporting discontinuous walls, are required to be designed for seismic
load combinations with overstrength. In such cases the seismic load effect, Em, has its horizontal component multiplied by
the overstrength factor O0, as indicated in Section 12.4.3.
C12.4.4 Minimum Upward Force for Horizontal Cantilevers for Seismic Design Categories D through F. In Seismic
Design Categories D, E, and F, horizontal cantilevers are designed for an upward force that results from an effective vertical
acceleration of 1.2 times gravity. This is to provide some minimum strength in the upward direction and to account for
possible dynamic amplification of vertical ground motions resulting from the vertical flexibility of the cantilever. The
requirement is not applied to downward forces on cantilevers, for which the typical load combinations are used.
C12.5 DIRECTION OF LOADING
Seismic forces are delivered to a building through ground accelerations that may approach from any direction relative to the
orthogonal directions of the building; therefore, seismic effects are expected to develop in both directions simultaneously.
The standard requires structures to be designed for the most critical loading effects from seismic forces applied in any
direction, and the procedures outlined in this section are deemed to satisfy that requirement.
The orthogonal combination procedure combines the effects from 100 percent of the seismic load applied in one direction
with 30 percent of the seismic load applied in the perpendicular direction. Combining effects for seismic loads in each
direction and accidental torsion results in 16 load combinations as follows:
Orthogonal load combinations
where :
QE = +/ QE_X+AT +/ 0.3QE_Y
QE_Y = effect of Ydirection load at the center of mass
(Section 12.8.4.2)
QE = +/ QE_XAT +/ 0.3QE_Y
QE_X = effect of Xdirection load at the center of mass
(Section 12.8.4.2)
QE = +/ QE_Y+AT +/ 0.3QE_X
AT = accidental torsion computed in accordance with
Section 12.8.4.2
QE = +/ QE_YAT +/ 0.3QE_X
For horizontal structural elements such as beams and slabs, orthogonal effects may be minimal; however, for vertical
elements of the seismicforceresisting system that participate in both orthogonal directions, the design likely will be
governed by these combinations.
Orthogonal combinations should not be confused with modal combinations such as the square root of the sum of the squares
(SRSS) or complete quadratic combination (CQC) technique.
The maximum effect of seismic forces, QE, from orthogonal load combinations must be modified by the redundancy factor,
., or the overstrength factor, O0, and consider the effects of vertical seismic forces, EV, in accordance with Section 12.4, to
obtain the seismic load effect, E.
C12.6 ANALYSIS SELECTION PROCEDURE
Table 12.61 applies only to buildings without seismic isolation (Chapter 17) or passive energy devices (Chapter 18).
The procedures addressed in Table 12.61 are equivalent lateral force (ELF) analysis (Section 12.8), modal response
spectrum (MRS) analysis (Section 12.9), linear response history (LRH) analysis, and nonlinear response history (NRH)
analysis. Requirements for performing response history analysis are provided in Chapter 16. Nonlinear static (pushover)
analysis is not addressed in the standard.
The value of Ts (= SD1/SDS) depends on the site class because SDS and SD1 include such effects. Where ELF is not allowed,
analysis must be performed using modal response spectrum or response history analysis.
ELF is not allowed for buildings with the listed irregularities because it assumes a gradually varying distribution of mass and
stiffness along the height and negligible torsional response. The 3.5Ts limit recognizes that higher modes are more
significant in taller buildings (Lopez and Cruz, 1996; Chopra, 2007) such that the ELF method may underestimate the design
base shear and may not predict correctly the vertical distribution of seismic forces.
C12.7 MODELING CRITERIA
C12.7.1 Foundation Modeling. Structural systems consist of three interacting components: the structural framing (girders,
columns, walls, diaphragms), the foundation (footings, piles, caissons), and the supporting soil. The ground motion that a
structure experiences, as well as the response to that ground motion, depends on the complex interaction between these
components.
Those aspects of ground motion that are affected by site characteristics are assumed to be independent of the structurefoundation
system as these effects would occur in the freefield in the absence of the structure. Hence, site effects are
considered separately (Sections 11.4.2 through 11.4.4 and Chapters 20 and 21).
Given a sitespecific ground motion or response spectrum, the dynamic response of the structure will depend on the
foundation system and on the characteristics of the soil that support the system. The dependence of the response on the
structurefoundationsoil system is referred to as soilstructure interaction. Such interactions will usually, but not always,
result in a reduction of base shear. This reduction in shear is due to the flexibility of the foundationsoil system and an
associated lengthening of the period of vibration of the structure. In addition, the soil system may provide an additional
source of damping. However, that total displacement typically increases with soilstructure interaction.
If the foundation is considered to be rigid, the computed base shears usually will be conservative, and it is for this reason that
rigid foundation analysis is allowed. The designer may ignore soilstructure interaction or may consider it explicitly in
accordance with Section 12.13.3 or implicitly in accordance with Chapter 19.
C12.7.2 Effective Seismic Weight. During an earthquake, the structure accelerates laterally, and these accelerations of the
structural mass produce inertial forces. These inertial forces, accumulated over the height of the structure, produce the design
base shear.
When a building vibrates during an earthquake, only that portion of the mass or weight that is physically tied to the structure
needs to be considered as effective. Hence, live loads (e.g., loose furniture, loose equipment, and human occupants) need not
be included. However, certain types of live loads such as storage loads may develop inertial forces, particularly where they
are densely packed.
Also considered as effective weight is all permanently attached equipment (e.g., air conditioners, elevator equipment, and
mechanical systems), movable partitions (a minimum of 10 psf is required), and 20 percent of significant roof snow load.
The full snow load need not be considered because maximum snow load and maximum earthquake load are unlikely to occur
simultaneously and loose snow does not move with the roof.
C12.7.3 Structural Modeling. The development of a mathematical model of a structure is always required because the
story drifts and the design forces in the structure cannot be computed without such a model. In some cases, the mathematical
model can be as simple as a freebody diagram as long that model can appropriately capture the strength and stiffness of the
structure.
The most realistic analytical model is threedimensional, includes all sources of stiffness (and flexibility) of the structure and
the soilfoundation system as well as Pdelta effects, and allows for nonlinear inelastic behavior in all parts of the structurefoundation
soil system. Development of such an analytical model is very time consuming, and such analysis is rarely
warranted for typical building designs performed in accordance with the standard. Instead of performing a nonlinear
analysis, inelastic effects are accounted for indirectly in the linear analysis methods by means of the response modification
factor, R, and the deflection amplification factor, Cd.
Using modern software, it often is more difficult to decompose a structure into planar models than it is to develop a full
threedimensional model so threedimensional models now are commonplace. Increased computational efficiency has
reduced the motivation to model rigid diaphragms, allowing for easy and efficient modeling of diaphragm flexibility. Threedimensional
models are required where the structure has torsional irregularities, outofplane offset irregularities, or
nonparallel system irregularities.
In general, the same threedimensional model may be utilized for equivalent lateral force, modal response spectrum, and
linear response history analysis. The response spectrum and linear response history models require a realistic modeling of
structural mass, and the response history method also requires an explicit representation of inherent damping. Five percent
critical damping is automatically included in the modal response spectrum approach. See Chapter 16 and the related
commentary for additional information on linear and nonlinear response history analysis.
It is well known that deformations in the panel zones of the beamcolumn joints of steel moment frames are a significant
source of flexibility. Two different mechanical models for including such deformations are summarized in Charney and
Marshall (2006). These methods apply to both elastic and inelastic systems. For elastic structures, centerline analysis
provides reasonable, but not always conservative, estimates of frame flexibility. Fully rigid end zones should not be used, as
this will always result in an overestimation of lateral stiffness in steel momentresisting frames. Partially rigid end zones may
be justified in certain cases such as where doubler plates are used to reinforce the panel zone.
Including the effect of composite slabs on the stiffness of beams and girders may be warranted in some circumstances.
Where composite behavior is included, due consideration should be paid to the reduction in effective composite stiffness for
portions of the slab in tension (Schaffhausen and Wegmuller, 1977; Liew, et al., 2001)
Figure C12.71 Undesired interaction effects.
h
H
Expected plastic hinge capacity = M
Expected column shear = 2M /H
p
p
Actual column shear = 2M p /h
Expected
hinging
region
Unexpected
hinging
region
For reinforced concrete buildings, it is important to address the effects of axial, flexural, and shear cracking in modeling the
effective stiffness of the structural components. Determining appropriate effective stiffness of the structural components
should take into consideration the anticipated demands on the components, their geometry, and the complexity of the model.
Recommendations for computing cracked section properties may be found in Paulay and Priestley (1992) and similar texts.
C12.7.4 Interaction Effects. The interaction requirements are intended to prevent unexpected failures in members of
momentresisting frames. Figure C12.71 illustrates a typical situation where masonry infill is used, and this masonry is
fitted tightly against reinforced concrete columns. Since the masonry is much stiffer than the columns, column hinges form
at the top of column and at the top of the masonry rather than at the top and bottom of the column. If the column flexural
capacity is Mp, the shear in the columns increases by the factor H/h, and this may cause an unexpected nonductile shear
failure in the columns. Many building collapses have been attributed to this effect.
Figure C12.71 Undesired interaction effects.
C12.8 EQUIVALENT LATERAL FORCE PROCEDURE
The equivalent lateral force (ELF) procedure provides a simple way to incorporate the effects of inelastic dynamic response
into a linear static analysis. This procedure is useful in preliminary design of all structures and is allowed for final design of
the vast majority of structures. The procedure is valid only for structures without significant discontinuities in mass and
stiffness along the height, where the dominant response to ground motions is in the horizontal direction without significant
torsion.
The ELF procedure has three basic steps:
1. Determine the seismic base shear,
2. Distribute the shear vertically along the height of the structure, and
3. Distribute the shear horizontally across the width and breadth of the structure.
Each of these steps is based on a number of simplifying assumptions. A broader understanding of these assumptions may be
obtained from any structural dynamics textbook that emphasizes seismic applications.
C12.8.1 Seismic Base Shear
C12.8.1.1 Calculation of Seismic Response Coefficient. Equation 12.81 simply expresses the base shear as the product of
the effective seismic weight, W, and a response coefficient, Cs. The response coefficient is a spectral pseudoacceleration, in
g units, which has been modified by R and I to account for inelastic behavior and to provide for improved performance for
high occupancy or essential structures.
There are five equations for determining the response coefficient Cs; the first three are plotted in Figure C12.81.
Equation 12.82, representing the constant acceleration part of the spectrum, controls where 0.0 < T < Ts. As shown in Table
C12.61 (which provides values of 3.5Ts), Ts is a function of seismicity and site. It may be as low as 0.2 seconds for low
hazard regions on Site Class B or as high as 0.9 seconds in high hazard regions on Site Class E.
Figure C12.81 Seismic response coefficient versus period.
T0 TS TL
Seismic R esponse Coefficient, Cs
Period, T
Constant
acceleration
[Eq. 12.82]
Constant velocity
[Eq. 12.83]
Constant
displacement
[Eq. 12.84]
Transition to peak
ground acceleration
[not used for ELF]
The true pseudoacceleration response spectrum transitions to the peak ground acceleration as the period approaches zero.
This transition is not used in the ELF method. One reason is that simple reduction of the response spectrum by (1/R) in the
very short period region would exaggerate inelastic effects.
Figure C12.81 Seismic response coefficient versus period.
Equation 12.83, representing the constant velocity part of the spectrum, controls where Ts < T < TL. In this region, the
seismic response coefficient is inversely proportional to period, and the pseudovelocity (pseudoacceleration divided by
circular frequency, lower case omega, is constant. TL, the longperiod transition period, is provided in Figures 2215 through 2220. TL
ranges from 4 seconds in in the northcentral conterminous states and western Hawaii to 16 seconds in the Pacific Northwest
and in western Alaska.
Equation 12.84, representing the constant displacement part of the spectrum, controls where T > TL. Given the current
mapped values of TL, this equation only affects tall and flexible structures.
Equation 12.85 is the minimum base shear and provides a (working stress) strength of approximately 3 percent of the weight
of the structure (Seismology Committee, Structural Engineers Association of California, 1996). This minimum base shear
was originally enacted in 1933 by the state of California’s Riley Act.
Equation 12.86 applies to sites near major active faults (as reflected by values of S1) where pulse effects can increase longperiod
demands.
C12.8.1.2 SoilStructure Interaction Reduction. Soilstructure interaction, which can influence significantly the dynamic
response of structures to earthquakes, is addressed in Chapter 19.
C12.8.1.3 Maximum Ss Value in Determination of Cs. The maximum value of Ss was created as hazard maps were
revised in 1997. The cap on Ss reflects engineering judgment about performance of codecomplying buildings in past
earthquakes so the height, period, and regularity conditions required for use of the limit are very important qualifiers.
C12.8.2 Period Determination. The fundamental period of the structure, T, is used to determine the design base shear as
well as the exponent k that establishes the distribution of the shear along the height of the structure. Equation 12.87 is an
Figure C12.82 Variation of fundamental period with building height.
0
1
2
3
4
5
6
7
0 100 200 300 400 500 600
Fundamental period, T (s)
Building height, hn (ft)
measured
values
[mean minus one
standard deviation]
[mean]
0.028 0.80 a n T = h
0.035 0.80 a n T = h
empirical relationship determined through statistical analysis of the measured response of buildings in California. Figure
C12.82 illustrates such data for various structures with steel moment resisting frames.
Since the empirical expression is based on the lower bound of the data, it produces a lower bound for the period of a building
of given height. This lower bound period, used in Equations 12.83 and 12.84, provides a conservative estimate of base
shear.
The fundamental period determined from a rational analysis may be used in design unless it exceeds the approximate period
times the coefficient provided in Table 12.81. This period limit prevents the use of unusually low ELF base shear for design
of buildings (or computational models) that are overly flexible. The coefficients in the table have two effects. First, the
conservatism of lower bound empirical formulas for Ta is removed. Second, the period is increased in regions of lower
seismicity as buildings in such areas generally are more flexible (and, hence, have longer periods) than buildings in regions
of higher seismicity.
Figure C12.82 Variation of fundamental period with building height.
C12.8.3 Vertical Distribution of Seismic Force. Equation 12.812 is based on the simplified first mode shape shown in
Figure C12.83. In the figure, Fx is the inertial force at level x, which is simply the total acceleration at level x times the mass
at level x. The base shear is the sum of these inertial forces, and Equation 12.8 simply gives the ratio of the force at level x to
the total base shear.
The deformed shape of the structure of Figure C12.83 is a function of the exponent k, which is related to the fundamental
period of vibration of the structure. The variation of k with T is illustrated in Figure C12.84. The exponent k is intended to
approximate the effect of higher modes, which are generally more dominant in structures with a longer fundamental period of
vibration. Lopez and Cruz (1996) discuss the factors that influence higher modes of response. Although the actual first
mode shape for a structure is also a function of the type of seismicforceresisting system, that effect is not reflected in these
equations.
The horizontal forces computed using Equation 12.812 do not reflect the actual inertial forces imparted on a structure at any
particular time. Instead, they are intended to provide design story shears that are consistent with enveloped results from more
accurate analysis (Chopra and Newmark, 1980).
Building height, hn (ft)
Figure C12.83 Basis of Equation 12.812.
hx
k
x
wx
h Figure C12.83 Basis of Equation 12.812.
2
2
1
k x
x x
n
k i
b i
i
k
x x x
vx n
b k
i i
F h w
g
V h w
g
C F w h
V wh
. a
. a
=
=
=
= =
S
S Figure C12.84 Variation of exponent k with period T.
k
k = 0.75 + 0.5T
T (seconds)
1.0
2.0
0.5 2.5
Figure C12.83 Basis of Equation 12.812.
C12.8.4 Horizontal Distribution of Forces. Within the context of an elastic ELF analysis, the distribution of lateral forces
to various seismicforceresisting elements depends on the type, geometric arrangement, and vertical extents of the resisting
elements and on the shape and flexibility of the floor diaphragms. Because seismicforceresisting elements are expected to
respond inelastically to design ground motions, the distribution of forces to the various elements also depends on the strength
of the elements and their sequence of yielding. Clearly, such effects cannot be captured accurately by a linear elastic static
analysis (Paulay, 1997). Nonlinear dynamic analysis is too cumbersome to be applied to the design of most buildings so
other approximate methods are used.
Figure C12.84 Variation of exponent k with period T.
Of particular concern is the torsional response of the structure during the earthquake. This response has been observed in
structures that are designed to be nearly symmetric in plan and layout of seismicforceresisting systems (De La Liera and
Chopra, 1994). This torsional response is due to a variety of “accidental” eccentricities that exist due to uncertainties in
quantifying the mass and stiffness distribution of the structure, as well as torsional components of ground motion that are not
included explicitly in codebased designs (Newmark and Rosenbleuth, 1971).
C12.8.4.1 Inherent Torsion. When lateral forces in a particular direction are applied statically at each story of a building
with rigid diaphragms, torsional displacement (twisting about the vertical axis) occurs if the centers of stiffness and mass of
each story are not perfectly coincident in plan. When threedimensional analysis is used, this inherent torsion is included
automatically. When planar analysis is used, the centers of mass and rigidity for each story must be determined explicitly.
Unfortunately, it is difficult to determine the center of rigidity for a multistory building to compute the inherent torsion; the
center of rigidity for a particular story depends on the configuration of the seismicforceresisting elements above and below
that story and may be load dependent (Chopra and Goel, 1991).
For buildings with fully flexible diaphragms (as defined in Section 12.3), vertical elements are assumed to resist inertial
forces from the mass that is tributary to the elements, but with no explicitly computed torsion. No diaphragm is perfectly
flexible, so some torsional forces always develop even when they are ignored.
Figure C12.85 Amplification factor for symmetric rectangular buildings.
L/B = 4
L/B = 1
cap
floor
0
1
2
3
0.1 0.2 0.3 0.4
Torsional amplification factor, Ax
Dimensional coefficient, a
curves for L/B = 1, 1. , 2, 4
L
B
aL
aB
V
C12.8.4.2 Accidental Torsion. Even for perfectly symmetric buildings, the true locations of the centers of mass and rigidity
are uncertain. As discussed in Section C12.8.4, other effects also may produce torsion. The requirement to consider
accidental torsion is intended to address this concern.
Accidental and inherent torsions result in forces that must be combined with those obtained from the application of the lateral
story forces; all components must be designed for the maximum effects determined considering positive accidental torsion,
negative accidental torsion, and no accidental torsion.
C12.8.4.3 Amplification of Accidental Torsion. Equation 12.814 was developed by the SEAOC “seismology committee
to encourage buildings with good torsional stiffness” (Structural Engineers Association of California, 1999).
In calculating the torsional amplification factor, Ax, the applied loads include inherent and accidental torsion, but with no
further amplification; the calculation is not iterative.
Figure C12.85 illustrates the effect of Equation 12.814 for a symmetric rectangular building with various aspect ratios (L/B)
where the seismicforceresisting elements are positioned at a variable distance (defined by a) from the center of mass in
each direction. Each element is assumed to have the same stiffness. The structure is loaded parallel to the short direction
with an eccentricity of 0.05L.
Figure C12.85 Amplification factor for symmetric rectangular buildings.
For a equal to 0.5, these elements are at the perimeter of the buildings, and for a equal to 0.0, they are at the center
(providing no torsional resistance). For a square building (L/B = 1.00), the torsional amplification factor is greater than 1.0
where a is less than 0.25 and increases to the maximum of 3.0 where a is equal to 0.11. For a rectangular building with L/B
equal to 4.00, the amplification factor is greater than 1.0 where a is less than 0.34 and increases to 3.0 where a is equal to
0.15. For the range of aspect ratios investigated, Ax is equal to 1.0 where a is greater than 0.34 and Ax reaches its maximum
value of 3.0 where (a < 0.11 to 0.15).
C12.8.6 Story Drift Determination. Equation 12.815 is used to estimate inelastic deflections, which are then used to
calculate design story drifts. These story drifts must be less than the allowable story drifts of Table 12.121. For buildings
without torsional irregularity, computations are performed using deflections at the centers of mass of adjacent stories. For
Seismic Design Category C, D, E, or F structures that are torsionally irregular, Section 12.12.1 requires that drifts be
computed along the edges of the structure.
Figure C12.86 Displacements used to compute drift.
Force,
V
Displacement,
Elastic
response
Actual inelastic
response
Idealized inelastic
response
VE
V=V /R
Analysis
domain
E
d d E d
The term Cd in Equation 12.815 amplifies the displacements from elastic analysis at design level forces, which are reduced
by R.
Figure C12.86 illustrates the relationships between elastic response; response to reduced designlevel forces; and the
expected inelastic response. If the structure remained elastic during an earthquake, the force would be VE, and the
corresponding displacement would be d E. Note that VE does not include the reduction factor, R, which accounts primarily
for ductility and overstrength. According to the equal displacement “rule” of seismic design, the maximum displacement of
an inelastic system is approximately equal to that of an elastic system with the same initial stiffness. This condition has been
observed for structures idealized with bilinear inelastic response and a fundamental period greater than Ts. For shorter period
structures, peak displacement of an inelastic system tends to exceed that of the corresponding elastic system. Since the forces
used for design include the response modification coefficient, R, the resulting displacements are too small and must be
amplified.
Figure C12.86 Displacements used to compute drift.
This analysis domain is shown in Figure C12.86. Because of overstrength and associated stiffness increases, the actual
inelastic response differs from the idealized inelastic response; the actual displacement of the system may be less than R
times d. The standard accounts for this difference by multiplying the fictitious (designlevel) elastic displacements d by the
factor Cd, which is usually less than R.
The design forces used to compute d xe include the importance factor, I, so Equation 12.815 includes I in the denominator.
This is appropriate since the allowable story drifts (except for masonry shear wall structures) in Table 12.121 are more
stringent for higher occupancy categories.
C12.8.6.1 Minimum Base Shear for Computing Drift. Except for period limits (as described in Section C12.8.6.2), all of
the requirements of Section 12.8 (including minimum base shears and force distributions) must be satisfied where computing
drift for ELF analysis.
C12.8.6.2 Period for Computing Drift. Where the response spectrum of Section 11.4.5 or the corresponding equations of
Section 12.8.1 are used and the structural period is less than TL, displacements increase with increasing period (even though
forces may decrease). Section 12.8.2 applies a period limit so that design forces are not too low, but if the lateral forces used
to compute drifts are inconsistent with the forces corresponding to the computed period, displacements will be overestimated.
Therefore, the standard allows the determination of drift using forces that are consistent with the computed period of
vibration of the structure.
Computed periods greater than CuTa are common, particularly for moment frames. In such cases the seismic design forces
used to proportion strength may produce displacements that violate drift limits, whereas displacements based on the
computed period will satisfy drift limits.
Equation
0
G 0y
0y
K P
K V h
d
. = = Equation
(1 ) 1y 0y V =V . Equation
0
1 1
d
d
.
=
 Figure C12.87 Pdelta effect on a simple structure.
Displacement,
Force,
V
d
y
(= d
0y = d
1y)
V0 y
V1 y
Slope = K0
Slope = K
1 = K
0 + K
G
Excluding Pdelta
Including Pdelta
Slope = KG
d
d
0
d
1
The more flexible the structure, the more likely it is that Pdelta effects will ultimately control the design. Computed periods
that are significantly greater than (perhaps more than 1.5 times) CuTa may indicate a modeling error.
C12.8.7 Pdelta Effects. Pdelta effects influence both the stiffness and strength of structures. Figure C12.87 shows
idealized static forcedisplacement responses for a simple, onestory structure (such as a cantilevered column). The stiffness
and strength of the structure without considering Pdelta effects (condition 0) are represented by K0 and V0. When Pdelta
effects are considered (condition 1), the related quantities are K1 and V1. Since the two model conditions are for the same
structure, inherent capacity of the structure is the same in either condition, the yield displacement is the same (d 0y = d 1y = d y).
The geometric stiffness of the structure, KG, is equal to P/h, where P is the total gravity load and h is the story height. KG is
negative where gravity loads cause compression in the story.
The stability coefficient, ., is defined as the absolute value of the geometric stiffness divided by the elastic stiffness. From
Figure C12.87, K0 = V0y / d 0y. Hence,
C12.81
Given the above, and the geometric relationships shown in Figure C12.87, it can be shown that the force producing yield in
condition 1 (with Pdelta effects) is
C12.82
and that for an applied force, V, less than or equal to V1y
C12.83
As . approaches 1.0, d 1 approaches infinity and V1 approaches zero, defining a state of static instability.
Figure C12.87 Pdelta effect on a simple structure.
The intent of Section 12.8.7 is to determine whether Pdelta effects are significant, and if so, to modify the strength and
stiffness of the structure to account for such effects. Also, maximum permitted values of . are established.
Equation 12.816 is used to determine the stability coefficient of each story of a structure. Where the stability coefficient
exceeds 0.1, Pdelta effects must be considered using one of two approaches. Displacements and member forces are either
multiplied by 1/(1. ) to reflect the conditions shown in Figure C12.87 in accordance with the equal displacement rule or
determined by rational analysis. Two types of rational analysis are envisioned. First, a nonlinear static (pushover) analysis
could be performed to show that the postyield slope of the pushover curve is continuously positive up to the target
displacement. Second, a nonlinear dynamic response history analysis could be repeated with and without Pdelta effects to
determine if the behavior including Pdelta meets all performance criteria.
Although the Pdelta procedures in the standard reflect the simple static idealization shown in Figure C12.87, the real issue
is one of dynamic stability. For that reason, nonlinear response history analysis is appealing. Such analysis should reflect
variability of ground motions and system properties, including initial stiffness, strain hardening stiffness, initial strength,
hysteretic behavior, and magnitude of gravity load. Unfortunately, the dynamic response of structures is highly sensitive to
such parameters, causing considerable dispersion to appear in the results (Vamvatsikos, 2002). This dispersion, which
increases dramatically with stability coefficient ., is due primarily to the incrementally increasing residual deformations
(ratcheting) that occur during the response. Residual deformations may be controlled by increasing either the initial strength
or the secondary stiffness. See Gupta and Krawinkler (2000) for additional information.
Equation 12.817 establishes the maximum stability coefficient permitted. The intent of this requirement is to protect
structures from the possibility of stability failures triggered by postearthquake residual deformation.
C12.9 MODAL RESPONSE SPECTRUM ANALYSIS
In the modal response spectrum analysis method, the structure is decomposed into a number of singledegreeoffreedom
systems, each having its own mode shape and natural period of vibration. The number of modes available is equal to the
number of mass degrees of freedom of the structure, so the number of modes can be reduced by eliminating mass degrees of
freedom. For example, rigid diaphragm constraints may be used to reduce the number of mass degrees of freedom to one per
story for planar models, and to three per story (two translations and rotation about the vertical axis) for threedimensional
structures. However, where the vertical elements of the seismicforceresisting system have significant differences in lateral
stiffness, rigid diaphragm models should be used with caution as relatively small inplane diaphragm deformations can have a
significant effect on the distribution of forces.
For a given direction of loading, the displacement in each mode is determined from the corresponding spectral acceleration,
modal participation, and mode shape. Because the sign (positive or negative) and the time of occurrence of the maximum
acceleration are lost in creating a response spectrum, there is no way to recombine modal responses exactly. However,
statistical combination of modal responses produces reasonably accurate estimates of displacements and component forces.
The loss of signs for computed quantities leads to problems in interpreting force results where seismic effects are combined
with gravity effects, produces forces that are not in equilibrium, and makes it impossible to plot deflected shapes of the
structure.
C12.9.1 Number of Modes. The key motivation to perform modal response spectrum analysis is to determine how the
actual distribution of mass and stiffness of a structure affects the elastic displacements and component forces. Where at least
90 percent of the model mass participates in the response, the distribution of forces and displacements is sufficient for design.
The scaling required by Section 12.9.4 controls the overall magnitude of design values so that incomplete mass participation
does not produce unconservative results.
The number of modes required to achieve 90 percent modal mass participation is usually a small fraction of the total number
of modes. See Lopez and Cruz (1996) for further discussion of the number of modes to use for modal response spectrum
analysis.
C12.9.2 Modal Response Parameters. The design response spectrum (whether the general spectrum from Section 11.4.5
or a sitespecific spectrum determined in accordance with Section 21.2) is representative of linear elastic structures. Division
of the spectral ordinates by R accounts for inelastic behavior, and multiplication of spectral ordinates by I provides the
additional strength needed to improve the performance of important structures. The displacements that are computed using
the response spectrum that has been modified by R and I (for strength) must be amplified by Cd and reduced by I to produce
the expected inelastic displacements. (See Section C12.8.6.)
C12.9.3 Combined Response Parameters. Most computer programs provide for either the SRSS or the CQC method
(Wilson, et al., 1981) of modal combination. The two methods are identical where applied to planar structures, or where zero
damping is specified for the computation of the crossmodal coefficients in the CQC method. The modal damping specified
in each mode for the CQC method should be equal to the damping level that was used in the development of the response
spectrum. For the spectrum in Section 11.4.5, the damping ratio is 0.05.
The SRSS or CQC method is applied to loading in one direction at a time. Where Section 12.5 requires explicit
consideration of orthogonal loading effects, the results from one direction of loading may be added to 30 percent of the
results from loading in an orthogonal direction. Wilson (2000) suggests that a more accurate approach is to use the SRSS
method to combine 100 percent of the results from each of two orthogonal directions where the individual directional results
have been combined by SRSS or CQC, as appropriate.
C12.9.4 Scaling Design Values of Combined Response. The modal base shear, Vt, may be less than the ELF base shear, V,
because: (a) the calculated fundamental period may be longer than that used in computing V, (b) the response is not
characterized by a single mode, and (c) the ELF base shear assumes 100 percent mass participation in the first mode, which is
always an overestimate. The scaling required by Section 12.9.4 provides, in effect, a minimum base shear for design. This
minimum base shear is provided because the computed period of vibration may be the result of an overly flexible (incorrect)
analytical model. The possible 15 percent reduction in design base shear may be considered as an incentive for using a
modal response spectrum analysis in lieu of the equivalent lateral force procedure.
Displacements from the modal response spectrum are not scaled because the use of an overly flexible model will result in
conservative estimates of displacement that need not be further scaled.
C12.9.5 Horizontal Shear Distribution. Accidental torsion must be included in the analysis as specified in Section 12.8.7.
For modal analysis there are two basic approaches to include accidental torsion.
The first approach is to perform static analyses with accidental torsions applied at each level of the structure, and then add
these results to those obtained from the modal response spectrum analysis. Where this approach is used, torsional
amplification in accordance with Section 12.8.4.3 is required.
The second approach, which applies only to threedimensional analysis, is to offset the centers of mass of each story
5 percent in each direction, thus requiring four separate models. The advantage of this method is that the effects of direct
loading and accidental torsion are combined automatically. A practical disadvantage is the increased bookkeeping for
multiple analyses.
Where this approach is used, further amplification of accidental torsion is not required because repositioning the center of
mass in a dynamic analysis changes the natural mode shapes and frequencies, producing torsions larger than the static
accidental torsion.
C12.9.6 Pdelta Effects. The requirements of Section 12.8.7, including the stability coefficient limit, . max, apply to modal
response spectrum analysis.
Amplification of displacements and member forces as a result of Pdelta effects may be accomplished through use of the
geometric stiffness. For the purpose of dynamic analysis, the linearized geometric stiffness, which includes the storywise P
. effect, is usually sufficient. Using the consistent geometric stiffness (Pd effect), which is associated with the deflected
shape of the individual elements of the structure, slightly improves accuracy. Including Pdelta effects directly in dynamic
analysis lengthens of the periods of vibration of each mode of response and increases lateral displacements.
C12.10 DIAPHRAGMS, CHORDS, AND COLLECTORS
C12.10.1 Diaphragm Design. Diaphragms are generally treated as horizontal deep beams or trusses that distribute lateral
forces to the vertical elements of the seismicforceresisting system. As deep beams, diaphragms must be designed to resist
the resultant shear and bending stresses. Diaphragms are commonly compared to girders, with the roof or floor deck
analogous to the girder web in resisting shear, and the boundary elements (chords) analogous to the flanges of the girder in
resisting flexural tension and compression. As in girder design, the chord members (flanges) must be sufficiently connected
to the body of the diaphragm (web) to prevent separation and to force the diaphragm to work as single unit.
Diaphragms may be considered flexible, semirigid, or rigid. The flexibility or rigidity of the diaphragm determines how
lateral forces will be distributed to the vertical elements of the seismicforceresisting system. See Section C12.3.1. Once
the distribution of lateral forces is determined, shear and moment diagrams are used to compute the diaphragm shear and
chord forces. Where diaphragms are not flexible, inherent and accidental torsion must be considered in accordance with
Section 12.8.4.
Diaphragm openings may require additional localized reinforcement (subchords and collectors) to resist the subdiaphragm
chord forces above and below the opening and to collect shear forces where the diaphragm depth is reduced. (See Figure
C12.101.) Collectors on each side of the opening drag shear into the subdiaphragms above and below the opening. The
subchord and collector reinforcement must extend far enough into the adjacent diaphragm to develop the axial force through
shear transfer. The required development length is determined by dividing the axial force in the subchord by the shear
capacity (in force/unit length) of the main diaphragm.
Chord reinforcement at reentrant corners must extend far enough into the main diaphragm to develop the chord force through
shear transfer. (See Figure C12.102.) Continuity of the chord members also must be considered where the depth of the
diaphragm is not constant.
Figure C12.101 Diaphragm components.
Main diaphragm
chords
Subdiaphragm
Collector
elements
Shear
wall
Shear
wall
Subdiaphragm
Direction of loading
Subchords Figure C12.102 Diaphragm with a reentrant corner.
Main diaphragm
chords
Shear
wall
Shear
wall
Direction of loading
Chord force
development
length
In wood and metal deck diaphragm design, framing members are often used as continuity elements, serving as subchords
and collector elements at discontinuities. These continuity members also are often used to transfer wall outofplane forces to
the main diaphragm, where the diaphragm itself does not have the capacity to resist the anchorage force directly. For
additional discussion, see Section C12.11.2.2.3.
Figure C12.101 Diaphragm components.
Figure C12.102 Diaphragm with a reentrant corner.
C12.10.1.1 Diaphragm Design Forces. Diaphragms must be designed to resist inertial forces, as specified in Equation
12.101, and to transfer design seismic forces due to horizontal offsets or changes in stiffness of the vertical resisting
elements. Inertial forces are those seismic forces that originate at the specified diaphragm level, while the transfer forces
originate above the specified diaphragm level. The redundancy factor, ., used for design of the seismicforceresisting
elements also applies to diaphragm transfer forces, thus completing the load path.
C12.10.2.1 Collector Elements Requiring Load Combinations with Overstrength Factor for Seismic Design
Categories C through F. The overstrength requirement of this section is intended to keep inelastic behavior in the ductile
elements of the seismicforceresisting system (consistent with the R factor) rather than in collector elements.
C12.11 STRUCTURAL WALLS AND THEIR ANCHORAGE
As discussed in Section C11.7, structural integrity is important not only in earthquakeresistant design but also in resisting
high winds, floods, explosion, progressive failure, and even such ordinary hazards as foundation settlement. The detailed
requirements of this section address walltodiaphragm integrity.
C12.11.1 Design for OutofPlane Forces. Because they are often subjected to local deformations caused by material
shrinkage, temperature changes, and foundation movements, wall connections require some degree of ductility in order to
accommodate slight movements while providing the required strength.
Although nonstructural walls are not subject to this requirement, they must be designed in accordance with Chapter 13.
C12.11.2 Anchorage of Concrete or Masonry Structural Walls. One major hazard in past earthquakes is the separation
of heavy masonry or concrete walls from floors or roofs. The forces defined in this section apply only to the anchorage or
connection of the wall to the structure, and not to overall wall design. The anchorage force should be considered both for
tension (outofplane) and sliding (inplane) directions.
Where the lateral spacing of connections used to resist the wall anchorage force are spaced further apart than 4 feet (1219
mm) as measured along the length of the wall, the section of wall that spans between the anchors must be designed to resist
the local outofplane bending caused by this force.
C12.11.2.1 Anchorage of Concrete or Masonry Structural Walls to Flexible Diaphragms. Diaphragm flexibility can
amplify outofplane accelerations so the wall anchorage forces in this condition are twice those defined in Section 12.11.1.
C12.11.2.2 Additional Requirements for Diaphragms in Structures Assigned to Seismic Design Categories C through
F.
C12.11.2.2.1 Transfer of Anchorage Forces into Diaphragm. This requirement, which aims to prevent the diaphragm
from tearing apart during strong shaking by requiring transfer of anchorage forces across the complete depth of the
diaphragm, was prompted by failures of connections between tilt up concrete walls and wood panelized roof systems in the
1971 San Fernando earthquake. An exception is provided for modestly proportioned diaphragms of lightframe construction,
which have not performed poorly.
Depending upon diaphragm shape and member spacing, numerous suitable combinations of subdiaphragms and continuous
tie elements and smaller subsubdiaphragms connecting to larger subdiaphragm and continuous tie elements are possible.
The configuration of each subdiaphragm (or subsubdiaphragm) provided must comply with the simple 2.5to1 lengthtowidth
ratio, and the continuous ties must have adequate member and connection strength to carry the accumulated wall
anchorage forces.
C12.11.2.2.2 Steel Elements of Structural Wall Anchorage System. A multiplier of 1.4 has been specified for strength
design of steel elements in order to obtain a fracture strength of almost 2 times the specified design force (where ft is 0.75 for
tensile rupture).
C12.11.2.2.3 Wood Diaphragms. Material standards for wood structural panel diaphragms permit the sheathing to resist
shear forces only; use to resist direct tension or compression forces is not permitted. Therefore, seismic anchorage forces
from walls must be transferred into framing members (such as beams, purlins, or subpurlins) using suitable straps or anchors.
For wood diaphragms, it is common to use local framing and sheathing elements as subdiaphragms to transfer the uniform
lateral wall forces into more concentrated lines of drag or continuity framing that carry the forces across the diaphragm and
hold the building together. Figure C12.111 shows a schematic plan of typical roof framing using subdiaphragms.
Fasteners to wood framing are intended to transfer shear forces only along the wood framing; any forces acting transverse to
the framing tend to induce splitting (due to crossgrain tension). Fasteners into wood ledgers attached to concrete or masonry
walls are designed to resist shear forces only; separate straps or anchors generally are provided to transfer outofplane wall
forces into perpendicular framing members.
C12.11.2.2.4 Metal Deck Diaphragms. In addition to transferring shear forces, metal deck diaphragms often can resist
direct axial forces in at least one direction. However, corrugated metal decks cannot transfer axial forces in the direction
perpendicular to the corrugations and are prone to buckling if the unbraced length of the deck as a compression element is
large. To manage diaphragm forces perpendicular to the deck corrugations, it is common that metal decks are supported at 8
to 10foot intervals by joists that are connected to walls in a manner suitable to resist the full wall anchorage design force and
to carry that force across the diaphragm. In the direction parallel to the deck corrugations, subdiaphragm systems are
considered near the walls; if the compression forces in the deck become large relative to the joist spacing, small compression
reinforcing elements are provided to transfer the forces into the subdiaphragms.
C12.11.2.2.6 Eccentrically Loaded Anchorage System. Wall anchors often are loaded eccentrically, either because the
anchorage mechanism allows eccentricity, or because of anchor bolt or strap misalignment. This eccentricity reduces the
anchorage connection capacity and hence must be considered explicitly in design of the anchorage. Figure C12.112 shows a
onesided rooftowall anchor that is subjected to severe eccentricity due to a misplaced anchor rod. If the detail were
designed as a concentric twosided connection, this condition would be easier to correct.
C12.11.2.2.7 Walls with Pilasters. The anchorage force at pilasters must be calculated considering twoway bending in
wall panels. It is customary to anchor the walls to the diaphragms assuming oneway bending and simple supports at the top
and bottom of the wall. However, where pilasters are present in the walls, their stiffening effect must be taken into account.
Each panel between pilasters is supported on four sides. The reaction at the pilaster top is the result of twoway action of the
Figure C12.111 Typical subdiaphragm framing.
Girder line
(typically also
continuity ties)
Subdiaphragm
chords
Subdiaphragm
Main diaphragm #1
Main diaphragm #1
chords
Purlins
(typical)
Subdiaphragm
chords Main diaphragm #2
Typical subdiaphragm
for outofplane forces
Opening Opening Figure C12.112 Plan view of wall anchor with misplaced anchor rod.
Roof joist
Shim added due to
misplaced anchor rod
Holddown anchor
Castinplace
anchor rod
Tiltup
wall panel
Alternate solution to one
sided connection: use
twosided connection
panel and is applied directly to the beam or girder anchorage at the top of the pilaster. The anchor load at the pilaster
generally is larger than the typical uniformly distributed anchor load between pilasters. Figure C12.113 shows the tributary
area typically used to determine the anchorage force for a pilaster.
Anchor points adjacent to the pilaster must be designed for the full tributary loading, conservatively ignoring the effect of the
adjacent pilaster.
Figure C12.111 Typical subdiaphragm framing.
Figure C12.112 Plan view of wall anchor with misplaced anchor rod.
C12.12 DRIFT AND DEFORMATION
As used in the standard, deflection is the absolute lateral displacement of any point in a structure relative to its base, and story
drift is the difference in deflection across a story (i.e., the deflection of a floor relative to that of the floor below).
The drifts and deflections are checked for the design earthquake ground motion, which is twothirds of the maximum
considered earthquake (MCE) ground motion.
Figure 12.113 Tributary area used to determine anchorage force at pilaster.
Tributary area of
wall on pilaster for
pilaster anchorage design
Wall yield line
45°
Top of parapet
Roof line
Pilaster
anchorage
There are many reasons to control drift; the most significant are to address the structural performance concerns of member
inelastic strain and system stability and to limit damage of nonstructural components, which can be lifethreatening. Drifts
provide a direct but imprecise measure of member strain and structural stability. Under small lateral deformations, secondary
stresses due to the Pdelta effect are normally within tolerable limits. (See Section C12.8.7.) The drift limits provide indirect
control of structural performance.
Figure 12.113 Tributary area used to determine anchorage force at pilaster.
Buildings subjected to earthquakes need drift control to restrict damage of partitions, shaft and stair enclosures, glass, and
other fragile nonstructural elements. The drift limits have been established without regard to economic considerations such
as a comparison of present worth of future repairs with additional structural costs to limit drift. These are matters for building
owners and designers to address.
The drift limits of Table 12.121 reflect consensus judgment taking into account life safety and damage control objectives
described above. Since the displacements induced in a structure include inelastic effects, structural damage in the designlevel
earthquake is likely. This may be seen from the seismic drift limits stated in Table 12.121. For ordinary structures
(Occupancy Category I or II), the drift limit is 0.02hsx, which is about ten times the drift ordinarily allowed under wind loads.
If deformations well in excess of the seismic drift limits were to occur repeatedly, structural components could lose so much
stiffness or strength that they compromise the safety and stability of the structure.
To provide better performance for Occupancy Category IV essential facilities, their drift limits generally are more stringent
than those for Occupancy Categories II and III. However, those limits are still greater than the damage thresholds for most
nonstructural components. Therefore, while the performance of Occupancy Category IV buildings should be better than that
of lower Occupancy Category buildings, there still can be considerable damage in the design earthquake.
The drift limits for lowrise structures are relaxed somewhat, provided that the interior walls, partitions, ceilings, and exterior
wall systems have been designed to accommodate story drifts. The type of steel building envisioned by the exception to the
table would be similar to a prefabricated steel structure with metal skin.
The limits set forth in Table 12.121 are for story drifts and apply to each and every story. For some structures, satisfying
strength requirements may produce a system with adequate drift control. However, the design of momentresisting frames
and of tall, narrow shear walls or braced frames often is governed by drift considerations. Where design spectral response
accelerations are large, seismic drift considerations are expected to control the design of midrise buildings. Where design
spectral response accelerations are small or the building is very tall, design for wind generally will control.
C12.12.3 Building Separation. The intent of this section is to address separations (also called seismic joints) between
adjacent structures or portions of the same structure (with or without frangible closures) for the purpose of permitting
independent response to earthquake ground motion. For irregular structures that cannot be expected to act reliably as a unit,
seismic joints should be used to produce separate units whose independent response to earthquake ground motion can be
predicted.
The standard does not give a precise formulation for the separations, but it does require that the distance be “sufficient to
avoid damaging contact under total deflection.” It is recommended that the distance be no less than the square root of the sum
of the squares of the lateral deflections, which represent the anticipated maximum inelastic deformations including torsion, of
the two units assumed to deflect toward each other (thus increasing with height). If the effects of impact can be shown not to
be detrimental, these distances can be reduced. For very rigid shear wall structures with rigid diaphragms whose lateral
deflections cannot be reasonably estimated, it is suggested that older code requirements for structural separations of at least 1
inch (25 mm) plus 1/2 inch (13 mm) for each 10 feet (3 m) of height above 20 feet (6 m) be followed.
C12.12.4 Deformation Compatibility For Seismic Design Categories D Through F. The purpose of this section is to
require that the seismicforceresisting system provide adequate deformation control to protect elements of the structure that
are not part of the seismicforceresisting system. In regions of high seismicity, many designers apply ductile detailing
requirements to elements that are intended to resist seismic forces but neglect such practices in nonstructural elements or
elements intended to resist only gravity forces. Even where elements of the structure are not intended to resist seismic forces
and are not detailed for such resistance, they can participate in the response and suffer severe damage as a result.
In the 1994 Northridge earthquake, such participation was a cause of several failures. A preliminary reconnaissance report of
that earthquake (EERI, 1994) states:
Of much significance is the observation that six of the seven partial collapses (in modern precast concrete parking
structures) seem to have been precipitated by damage to the gravity load system. Possibly, the combination of large
lateral deformation and vertical load caused crushing in poorly confined columns that were not detailed to be part of the
lateral load resisting system. . . . Punching shear failures were observed in some structures at slabtocolumn connections
such as at the Four Seasons building in Sherman Oaks. The primary lateral load resisting system was a perimeter ductile
frame that performed quite well. However, the interior slabcolumn system was incapable of undergoing the same lateral
deflections and experienced punching failures.
This section addresses such concerns. Rather than relying on designers to assume appropriate levels of stiffness, this section
explicitly requires that the stiffening effects of adjoining rigid structural and nonstructural elements be considered and that a
rational value of member and restraint stiffness be used for the design of components that are not part of the seismicforceresisting
system.
This section also includes a requirement to address shears that can be induced in structural components that are not part of the
seismicforceresisting system, since sudden shear failures have been catastrophic in past earthquakes.
The exception in Section 12.12.4 is intended to encourage the use of intermediate or special detailing in beams and columns
that are not part of the seismicforceresisting system. In return for better detailing, such beams and columns are permitted to
be designed to resist moments and shears from unamplified deflections. This reflects observations and experimental evidence
that welldetailed components can accommodate large drifts by responding inelastically without losing significant vertical
loadcarrying capacity.
C12.13 FOUNDATION DESIGN
C12.13.3 Foundation LoadDeformation Characteristics. This section of the standard provides guidance on modeling
loaddeformation characteristics of the foundationsoil system (foundation stiffness) for linear analysis procedures. The
further guidance contained herein addresses both linear and nonlinear analysis methods. Where linear analysis procedures
are used with the methodology given below, the earthquake forces should not be reduced by R.
Modeling of the loaddeformation characteristics of foundations should be in accordance with ASCE/SEI 41. For nonlinear
analysis of piles that may form plastic hinges, the lateral loaddeformation characteristics of piles may be taken from Song, et
al. (2005).
For load combinations including seismic load effects, the vertical, lateral, and rocking load capacities of foundations as
limited by the soil should be sufficient to resist loads with acceptable deformations, considering the short duration of loading,
the dynamic properties of the soil, and the ultimate load capacities, Qus, of the foundations.
Ultimate foundation load capacities should be determined by a qualified geotechnical engineer based on geotechnical site
investigations that include field and laboratory testing to determine soil classification and soil strength parameters or on insitu
testing of prototype foundations. For competent soils that do not undergo strength degradation under seismic loading,
strength parameters for static loading conditions may be used to compute ultimate load capacities for seismic design. For
sensitive cohesive soils or saturated cohesionless soils, the potential for earthquakeinduced strength degradation should be
considered.
Ultimate foundation load capacities, Qus, under vertical, lateral, and rocking loading should be determined using accepted
foundation design procedures and principles of plastic analysis. Calculated ultimate load capacities, Qus, should be bestestimated
values using soil properties that are representative average values for individual foundations. Bestestimated
values of Qus should be reduced by resistance factors (f) to reflect uncertainties in site conditions and in the reliability of
analysis methods. The factored foundation load capacity, fQus, should be used both to check acceptance criteria and as the
foundation capacity in nonlinear loaddeformation models.
If ultimate foundation load capacities are determined based on geotechnical site investigations including laboratory or insitu
tests, f factors equal to 0.8 for cohesive soils and 0.7 for cohesionless soils should be used for vertical, lateral, and rocking
resistance for all foundation types. If ultimate foundation load capacities are determined based on fullscale fieldtesting of
prototype foundations, f factors equal to 1.0 for cohesive soils and 0.9 for cohesionless soils are recommended.
For both linear and nonlinear analysis procedures, a model incorporating a combined superstructure and foundation system is
necessary to assess the effect of foundation deformations on the superstructure elements.
For linear analysis methods, the linear loaddeformation behavior of foundations should be represented by an equivalent
linear (secant) stiffness using soil properties that are compatible with the soil strain levels associated with the design
earthquake motion. The straincompatible shear modulus, G, and the associated straincompatible shear wave velocity, vs,
needed for the evaluation of equivalent linear stiffness are specified in Chapter 19 of the standard or can be based on a sitespecific
study. ASCE/SEI 41 is an acceptable alternative to that contained in the standard and may provide more realistic
results.
For nonlinear analysis procedures, the nonlinear loaddeformation behavior of the foundationsoil system may be represented
by a bilinear or multilinear curve having an initial equivalent linear stiffness and a limiting foundation capacity. The initial
equivalent linear stiffness should be determined as described above for linear analysis methods. The limiting foundation
capacity should be taken as the factored foundation load capacity, fQus. Parametric variations in analyses should include:
(a) a reduction in stiffness of 50 percent combined with a limiting foundation capacity, f Qus, and (b) an increase in stiffness
of 50 percent combined with a limiting foundation capacity equal to Qus multiplied by 1/f.
For linear analysis procedures, factored foundation load capacities, f Qus, should not be exceeded for load combinations that
include seismic load effects.
For the nonlinear analysis procedures, if the factored foundation load capacity, f Qus, is reached during seismic loading, the
potential significance of associated transient and permanent foundation displacements should be evaluated. Foundation
displacements are acceptable if they do not impair the continuing function of Occupancy Category IV structures or the life
safety of any structure.
C12.13.4 Reduction of Foundation Overturning. Since the vertical distribution of forces prescribed for use with the
equivalent lateral force procedure is intended to envelope story shears, overturning moments are exaggerated. (See
Section C12.13.3.) Such moments will be lower where multiple modes respond, so a 25 percent reduction is permitted for
design (strength and stability) of the foundation using this procedure. This reduction is not permitted for inverted pendulum
or cantilevered column type structures, which typically have a single mode of response.
Since the modal response spectrum analysis procedure more accurately reflects the actual distribution of shears and
overturning moments, the permitted reduction is only 10 percent.
C12.13.5 Requirements for Structures Assigned to Seismic Design Category C.
C12.13.5.1 PoleType Structures. The high contact pressures that develop between pole and soil as a result of lateral loads
make poletype structures sensitive to earthquake motions. Bending in the poles and soil lateral capacity and deformation are
key considerations in the design. For further discussion of polesoil interaction, see Section C12.13.6.7.
C12.13.5.2 Foundation Ties. One important aspect of adequate seismic performance is that the foundation acts as a unit,
not permitting one column or wall to move appreciably with respect to another. To attain this performance, it is common to
provide ties between footings and pile caps. This is especially important where the use of deep foundations is driven by the
existence of soft surface soils.
Multistory buildings often have major columns that run the full height of the building adjacent to smaller columns that
support only one level; the calculated tie force is based on the heavier column load.
The standard permits alternate methods of tying foundations together. Lateral soil pressure on pile caps is not a
recommended method because motion is imparted from soil to structure and during displacement under dynamic conditions.
C12.13.5.3 Pile Anchorage Requirements. The pile anchorage requirements are intended to prevent brittle failures of the
connection to the pile cap under moderate ground motions. Moderate ground motions can result in pile tension forces or
bending moments that could compromise shallow anchorage embedment. Loss of pile anchorage could result in increased
structural displacements from rocking, overturning instability, and loss of shearing resistance at the ground surface. A
concrete bond to a bare steel pile section usually is unreliable, but connection by means of deformed bars properly developed
from the pile cap into concrete confined by a circular pile section is permitted.
C12.13.6 Requirements for Structures Assigned to Seismic Design Categories D through F.
C12.13.6.1 PoleType Structures. See Section C12.13.5.1.
C12.13.6.2 Foundation Ties. See Section C12.13.5.2. For Seismic Design Categories D through F, the requirement is
extended to spread footings on soft soils.
C12.13.6.3 General Pile Design Requirements. Design of piles is based on the same R factor used in design of the
superstructure; since inelastic behavior will result, piles should be designed with ductility similar to that of the superstructure.
When strong ground motions occur, inertial structure pilesoil interaction may produce plastic hinging in piles near the
bottom of the pile cap, and kinematic soilpile interaction will result in bending moments and shearing forces throughout the
length of the pile, being higher at interfaces between stiff and soft soil strata. These effects are particularly severe in soft
soils and liquefiable soils so Section 14.2.3.2.1 requires special detailing in areas of concern.
The shears and curvatures in piles caused by inertial and kinematic interaction may exceed the bending capacity of
conventionally designed piles, resulting in severe damage. Analysis techniques to evaluate pile bending are discussed by
Margason and Holloway (1977) and Mylonakis (2001), and these effects on concrete piles are further discussed by Shepard
(1983). For homogeneous, elastic media and assuming the pile follows the soil, the freefield curvature (soil strains without a
pile present) can be estimated by dividing the peak ground acceleration by the square of the shear wave velocity of the soil;
considerable judgment is necessary in using this simple relationship for a layered, inelastic profile with pilesoil interaction
effects. Norris (1994) discusses methods to assess pilesoil interaction.
Where determining the extent of special detailing, the designer must consider variation in soil conditions and driven pile
lengths, so that adequate ductility is provided at potential high curvature interfaces. Confinement of concrete piles to provide
ductility and to maintain functionality of the confined core pile during and after the earthquake may be obtained by use of
heavy spiral reinforcement or use of exterior steel liners.
C12.13.6.4 Batter Piles. Partially embedded batter piles have a history of poor performance in strong ground shaking, as
shown by Gerwick and Fotinos (1992). Failure of battered piles has been attributed to design that neglect loading on the piles
from ground deformation or that assumes that lateral loads are resisted by axial response of piles without regard to moments
induced in the pile at the pile cap (Lam and Bertero, 1990). Because batter piles are considered to have limited ductility, they
must be designed using the load combinations with overstrength. Momentresisting connections between pile and pile cap
must resolve the eccentricities inherent in batter pile configurations. This concept is illustrated clearly by EQE Engineering
(1991).
C12.13.6.5 Pile Anchorage Requirements. Piles should be anchored to the pile cap to permit energy dissipating
mechanisms, such as pile slip at the pilesoil interface, while maintaining a competent connection. This section of the
standard sets forth a capacity design approach to achieve that objective. Anchorages occurring at pile cap corners and edges
should be reinforced to preclude local failure of plain concrete sections due to pile shears, axial loads, and moments.
C12.13.6.6 Splices of Pile Segments. A capacity design approach, similar to that for pile anchorage, is applied to pile
splices.
C12.13.6.7 Pile Soil Interaction. Short piles and long slender piles embedded in the earth behave differently when subject
to lateral forces and displacements. The response of a long slender pile depends on its interaction with the soil considering
the nonlinear response of the soil. Numerous design aid curves and computer programs are available for this type of analysis,
which is necessary to obtain realistic pile moments, forces, and deflections and is common in practice (Ensoft, 2004). More
sophisticated models, which also consider inelastic behavior of the pile itself, can be analyzed using generalpurpose
nonlinear analysis computer programs or closely approximated using the pilesoil limit state methodology and procedure
given by Song, et al. (2005).
Short piles (with lengthtodiameter ratios no more than 6) can be treated as a rigid body simplifying the analysis. A method
assuming a rigid body and linear soil response for lateral bearing is given in the current building codes. A more accurate and
comprehensive approach using this method is given in a study by Czerniak (1957).
C12.13.6.8 Pile Group Effects. The effects of groups of piles, where closely spaced, must be taken into account for vertical
and horizontal response. As groups of closely spaced piles move laterally, failure zones for individual piles overlap, and
horizontal strength and stiffness response of the pilesoil system is reduced. Reduction factors or “pmultipliers” are used to
account for these groups of closely spaced piles. For a pile centertocenter spacing of three pile diameters, reduction factors
of 0.6 for the leading pile row and 0.4 for the trailing pile rows are recommended by Rollins, et al. (1999). Computer
programs are available to analyze group effects assuming a nonlinear soil and elastic piles (Ensoft, 2004a).
C12.14 SIMPLIFIED ALTERNATIVE STRUCTURAL DESIGN CRITERIA FOR SIMPLE BEARING WALL OR
BUILDING FRAME SYSTEMS
C12.14.1 General. In recent years, engineers and building officials have become concerned that the seismic design
requirements in codes and standards, while intended to make structures perform more reliably, have become so complex and
difficult to understand and to implement that they may be counterproductive. Since the response of buildings to earthquake
ground shaking is very complex (especially for irregular structural systems), realistically accounting for these effects can lead
to complex requirements. There is a concern that the typical designers of small, simple structures, which may represent more
than 90 percent of construction in the United States, have difficulty understanding and applying the general seismic
requirements of the standard.
The simplified procedure presented in this section of the standard applies to lowrise, stiff structures. The procedure, which
was refined and tested over a fiveyear period, was developed to be used for a defined set of structures deemed to be
sufficiently regular in configuration to allow a reduction of prescriptive requirements. For some design elements, such as
foundations and anchorage of nonstructural systems, other sections of the standard must be followed, as referenced within
Section 12.14.
C12.14.1.1 Simplified Design Procedure. Reasons for the limitations of the simplified design procedure of Section 12.14
are as follows:
1. The procedure was developed to address adequate seismic performance for standard occupancies. Since it was not
developed for higher levels of performance associated with Occupancy Category III and IV structures, no importance
factor (I) is employed.
2. Site Class E and F soils require specialized procedures that are beyond the scope of the procedure.
3. The procedure was developed for stiff, lowrise buildings, where highermode effects are negligible.
4. Only stiff systems, where drift is not a controlling design criterion, may employ the procedure. Because of this
limitation, drifts are not computed. The response modification coefficient, R, and the associated system limitations are
consistent with those found in the general Chapter 12 requirements.
5. In order to achieve a balanced design and to achieve a reasonable level of redundancy, two lines of resistance are
required in each of the two major axis directions. Because of this stipulation, no redundancy factor (.) is applied.
6. To reduce the potential for dominant torsional response, at least one line of resistance must be placed on each side of the
center of mass.
7. Large overhangs for flexible diaphragm buildings can produce response that is inconsistent with the assumptions
associated with the procedure.
8. A system that satisfies these layout and proportioning requirements avoids torsional irregularity, so calculation of
accidental torsional moments is not required.
9. An essentially orthogonal orientation of lines of resistance effectively uncouples response along the two major axis
directions, so orthogonal effects may be neglected.
10. Where the simplified design procedure is chosen, it must be used for the entire design, in both major axis directions.
11. Since inplane and outofplane offsets generally create large diaphragm, collector, and discontinuous element demands
that are not addressed by the procedure, these irregularities are prohibited.
12. Buildings that exhibit weakstory behavior violate the assumptions used to develop the procedure.
C12.14.3 Seismic Load Effects and Combinations. The seismic load effect and combination equations for the simplified
design procedure are consistent with those for the general procedure, with one notable exception: the overstrength factor
(corresponding to O0 in the general procedure) is set at 2.5 for all systems as indicated in Section 12.14.3.2.1. Given the
limited systems that can use the simplified design procedure, specifying unique overstrength factors was deemed
unnecessary.
C12.14.7 Design and Detailing Requirements. The design and detailing requirements outlined in this section are similar to
those for the general procedure. The few differences include the following:
1. Forces used to connect smaller portions of a structure to the remainder of the structures are taken as 0.20 times the shortperiod
design spectral response acceleration, SDS, rather than the general procedure value of 0.133 (Section 12.14.7.1).
2. Anchorage forces for concrete or masonry structural walls for structures with diaphragms that are not flexible are
computed using the requirements for nonstructural walls.
C12.14.8 Simplified Lateral Force Analysis Procedure
C12.14.8.1 Seismic Base Shear. The seismic base shear in the simplified design procedure, as given by Equation 12.1411,
is a function of the shortperiod design spectral response acceleration, SDS. The value for F in the base shear equation
addresses changes in dynamic response for two and threestory buildings. As in the general procedure (Section 12.8.1.3),
SDS may be computed for short, regular structures with SS taken no greater than 1.5.
C12.14.8.2 Vertical Distribution. The seismic forces for multistory buildings are distributed vertically in proportion to the
weight of the respective floor. Given the slightly amplified base shear for multistory buildings, this assumption, along with
the threestory height limit, produces results consistent with the more traditional triangular distribution without introducing
that more sophisticated approach.
C12.14.8.5 Drift Limits and Building Separation. For the simplified design procedure, which is restricted to stiff wall and
braced frame structures, drift need not be calculated. Where drifts are required (such as for structural separations and
cladding design) a conservative drift value of 1 percent is specified.
REFERENCES
American Society of Civil Engineers. 2006. Seismic Rehabilitation of Existing Buildings, ASCE/SEI 41. ASCE, Reston,
Virginia.
Applied Technology Council. 2009. Quantification of Building Seismic Performance Factors, FEMA P695. Federal
Emergency Management Agency, Washington, D.C.
Applied Technology Council. 1978. Tentative Provisions for the Development of Seismic Regulations for Buildings, ATC 3
06. ATC, Redwood City, California.
Bernal, D. 1987. “Amplification Factors for Inelastic Dynamic Pdelta Effects in Earthquake Analysis,” Earthquake
Engineering and Structural Dynamics, 18: 635681.
Building Seismic Safety Council. 2004. NEHRP Recommended Provisions for Seismic Regulations for New Buildings and
Other Structures, FEMA 450. Federal Emergency Management Agency, Washington, D.C.
Charney, F. A. 1990. “Wind Drift Serviceability Limit State Design of Multistory Buildings,” Journal of Wind Engineering
and Industrial Aerodynamics, 36:203212.
Charney, F. A., and J. Marshall. 2006. “A Comparison of the Krawinkler and Scissors Models for Including BeamColumn
Joint Deformations in the Analysis of MomentResisting Frames,” AISC Engineering Journal, 43(1):3148.
Chopra, A. K. 2007. Structural Dynamics. Prentice Hall.
Chopra, A. K., and R. K. Goel. 1991. “Evaluation of torsional provisions in seismic codes, Journal of Structural
Engineering, 117(12):37623782
Chopra, A. K., and N. M. Newmark. 1980. “Analysis,” Chapter 2 of Design of Earthquake Resistant Structures, edited by E.
Rosenblueth. John Wiley and Sons.
Czerniak, E. 1957. "Resistance to Overturning of Single Short Piles" ASCE Journal of the Structural Division, 83(ST2).
Degenkolb, Henry J. 1987. “A Study of the Pdelta Effect,” Earthquake Spectra, 3(1).
De La Liera, J. C., and A. K. Chopra. 1994. “Evaluation of Code Accidental Torsion Records from Building Records,”
Journal of Structural Engineering, 120(2):697616.
EQE Engineering. 1991. Structural Concepts and Details for Seismic Design, UCRLCR106554. Department of Energy,
Washington, D.C.
Ensoft, Inc. 2004a. Computer Program GROUP, Version 6.0, A Program for the Analysis of a Group of Piles Subjected to
Axial and Lateral Loading, User’s Manual and Technical Manual. Ensoft, Austin, Texas.
Ensoft, Inc. 2004b. Computer Program LPILE Plus Version 5.0, A Program for the Analysis of Piles and Drilled Shafts
under Lateral Loads, User’s Manual and Technical Manual. Ensoft, Austin, Texas.
Gerwick, Jr., B., and G. Fotinos. 1992. "Drilled Piers and Driven Piles for Foundations in Areas of High Seismicity,"
SEAONC Fall Seminar, October 29, San Francisco, California.
Griffis, Larry. 1993. "Serviceability Limit States Under Wind Load," Engineering Journal, American Institute of Steel
Construction, First Quarter.
Gupta, A., and H. Krawinkler. 2000. "Dynamic Pdelta Effects for Flexible Inelastic Steel Structures," Journal of Structural
Engineering, 126(1):145154.
Lam, I., and V. Bertero. 1990. “Aseismic Design of Pile Foundations for Port Facilities,” in Proceedings of the POLA
Seismic Workshop on Seismic Engineering, March 2123, San Pedro, California, Port of Los Angeles.
Liew, J. Y. 2001. “Inelastic Analysis of Steel Frames with Composite Beams,” Journal of Structural Engineering,
127(2):194202.
Lopez, O. A., and M. Cruz. 1996. "Number of Modes for the Seismic Design of Buildings," Earthquake Engineering and
Structural Dynamics, 25(8):837856.
Margason, E., and M. Holloway. 1977. “Pile Bending During Earthquakes,” in Proceedings of the 6th World Conference on
Earthquake Engineering, New Delhi.
Mylonakis, G. 2001. “Seismic pile bending at soillayer interfaces,” Soils and Foundations, 41 (4), pp. 4758.
Newmark, N. A., and E. Rosenbleuth. 1971. Fundamentals of Earthquake Engineering. Prentice Hall.
Norris, G. M. 1994. “Seismic Bridge Pile Foundation Behavior,” in Proceedings, International Conference on Design and
Construction of Deep Foundations, Federal Highway Administration, Vol. 1.
Paulay, T. 1997. “Are Existing Seismic Torsion Provisions Achieving Design Aims?” Earthquake Spectra, 13(2):259280.
Paulay, T., and M. J. N. Priestly. 1992. Seismic Design of Reinforced Concrete and Masonry Structures. John Wiley and
Sons.
Rollins, Kyle M., K. T. Peterson, T. J. Weaver, and Andrew E. Sparks. 1999. “Static and dynamic lateral load behavior on a
fullscale pile group in clay,” Brigham Young University, Provo, Utah, and the Utah Department of Transportation, Salt Lake
City, June 23.
Schaffhausen, R., and A. Wegmuller. 1977. “Multistory Rigid Frames with Composite Girders under Gravity and Lateral
Forces,” AISC Engineering Journal, 2nd Quarter.
Seismology Committee, Structural Engineers Association of California. 1996. Recommended Lateral Force Requirements
and Commentary. SEAOC, Sacramento, California.
Shepard, D. A. 1983. “Seismic Design of Concrete Piling,” PCI Journal (March/April).
Song, S. T., Y. H. Chai, and T. H. Hale. 2005. “Analytical model for ductility assessment of fixedhead concrete piles,”
ASCE Journal of Structural Engineering 131(7):10511059.
Vamvatsikos, D. 2002. “Seismic Performance, Capacity and Reliability of Structures as Seen Through Incremental
Dynamic Analysis,” Ph.D. Dissertation, Department of Civil and Environmental Engineering, Stanford University, Palo Alto,
California.
Wilson, E. L., A. Der Kiureghian, and E. P. Bayo. 1981. "A Replacement for the SRSS Method in Seismic Analysis,"
Earthquake Engineering and Structural Dynamics, Vol. 9.
Wilson, E. L. 2000. Threedimensional Static and Dynamic Analysis of Structures, Computers and Structures, Inc.,
Berkeley, California.
Figure C13.11 Image of Hospital imaging equipment that fell from overhead mounts.
COMMENTARY CHAPTER 13,
SEISMIC DESIGN REQUIREMENTS FOR
NONSTRUCTURAL COMPONENTS
C13.1 GENERAL
Chapter 13 defines minimum design criteria for architectural, mechanical, electrical, and other nonstructural systems and
components recognizing structure use, occupant load, the need for operational continuity, and the interrelation of structural
and architectural, mechanical, electrical, and other nonstructural components. Nonstructural components are designed for
design earthquake ground motions as defined in Section 11.2 and determined in Section 11.4.4 of the standard. In contrast to
structures, which are implicitly designed for a low probability of collapse when subjected to the maximum considered
earthquake (MCE) ground motions, there are no implicit performance goals associated with the MCE for nonstructural
components. Performance goals associated with the design earthquake are discussed in Section C13.1.3.
Suspended or attached nonstructural components that could detach either in full or in part from the structure during an
earthquake are referred to as falling hazards and may represent a serious threat to property and life safety. Critical attributes
that influence the hazards posed by these components include their weight, their attachment to the structure, their failure or
breakage characteristics (e.g., certain types of glass), and their location relative to occupied areas (e.g., over an entry or exit, a
public walkway, an atrium, or a lower adjacent structure). Architectural components that pose potential falling hazards
include parapets, cornices, canopies, marquees, glass, large ornamental elements (e.g., chandeliers), and building cladding.
In addition, suspended mechanical and electrical components (e.g., mixing boxes, piping, and ductwork) may represent
serious falling hazards. Figures C13.11 through C13.14 show damage to nonstructural components in past earthquakes.
Figure C13.11 Hospital imaging equipment that fell from overhead mounts.
Figure C13.12 Image of Collapsed light fixtures.
Figure C13.13 Image of Collapsed duct and HVAC diffuser.
Figure C13.12 Collapsed light fixtures.
Figure C13.13 Collapsed duct and HVAC diffuser.
Figure C13.14 Image of Damaged ceiling system.
Figure C13.14 Damaged ceiling system.
Components whose collapse during an earthquake could result in blockage of the means of egress deserve special
consideration. The term “means of egress” is used commonly in building codes with respect to fire hazard. Consideration of
egress may include intervening aisles, doors, doorways, gates, corridors, exterior exit balconies, ramps, stairways, pressurized
enclosures, horizontal exits, exit passageways, exit courts, and yards. Items whose failure could jeopardize the means of
egress include walls around stairs and corridors, veneers, cornices, canopies, heavy partition systems, ceilings, architectural
soffits, light fixtures, and other ornaments above building exits or near fire escapes. Examples of components that generally
do not pose a significant falling hazard include fabric awnings and canopies. Architectural, mechanical, and electrical
components that, if separated from the structure, will fall in areas that are not accessible to the public (e.g., into a mechanical
shaft or light well) also pose little risk to egress routes.
For some architectural components such as exterior cladding elements, wind design forces may exceed the calculated seismic
design forces. Nevertheless, seismic detailing requirements may still govern the overall structural design. Where this is a
possibility, it must be investigated early in the structural design process.
The seismic design of nonstructural components may involve consideration of nonseismic requirements that are affected by
seismic bracing. For example, accommodation of thermal expansion in pressure piping systems often is a critical design
consideration and seismic bracing for these systems must be arranged in a manner that accommodates thermal movements.
Particularly in the case of mechanical and electrical systems, the design for seismic loads should not compromise the
functionality, durability, or safety of the overall design; this requires collaboration between the various disciplines of the
design and construction team.
For various reasons (e.g., business continuity), it may be desirable to consider higher performance than that required by the
building code. For example, to achieve continued operability of a piping system, it is necessary to prevent unintended
operation of valves or other inline components in addition to preventing collapse and providing leak tightness. Higher
performance also is required for components containing substantial quantities of hazardous contents (as defined in Section
11.2). These components must be designed to prevent uncontrolled release of those materials.
The requirements of Chapter 13 are intended to apply to new construction and tenant improvements installed at any time
during the life of the structure, provided they are listed in Table 13.51 or 13.61. Further, they are intended to reduce (not
eliminate) the risk to occupants and to improve the likelihood that essential facilities remain functional. While property
protection (in the sense of investment preservation) is a possible consequence of implementation of the standard, it is not
currently a stated or implied goal; a higher level of protection may be advisable if such protection is desired or required.
C13.1.1 Scope. The requirements for seismic design of nonstructural components apply to the nonstructural component as
well as to its supports and attachments to the main structure. In some cases as defined in Section 13.2, it is necessary to
consider explicitly the performance characteristics of the component. The requirements are intended to apply only to
Figure C13.15 Image of Toppled storage cabinets.
permanently attached components – not to furnishings, temporary items, or mobile units. Furnishings such as tables, chairs,
and desks may shift during strong ground shaking but generally pose minimal hazards provided they do not obstruct
emergency egress routes. Storage cabinets, tall bookshelves, and other items of significant mass do not fall into this category
and should be anchored or braced in accordance with this chapter.
Figure C13.15 Toppled storage cabinets.
Temporary items are those that remain in place for short periods of time (months, not years). Components that, while
movable or relocatable, are expected to remain in place for periods of a year or longer should be considered permanent for
the purposes of this section. Modular office systems are considered permanent since they ordinarily remain in place for long
periods. In addition, they often include storage units of significant capacity which may topple in an earthquake. They are
subject to the provisions of Section 13.5.8 for partitions if they exceed 6 feet in height. Mobile units include components that
are moved from one point in the structure to another during ordinary use. Examples include desktop computers, office
equipment, and other components that are not permanently attached to the building utility systems (Figure C13.15).
Components that are mounted on wheels to facilitate periodic maintenance or cleaning but that otherwise remain in the same
location (e.g., server racks) are not considered moveable for the purposes of anchorage and bracing. Likewise, skidmounted
components (as shown in Figure C13.16) as well as the skids themselves are considered permanent equipment.
In all cases, equipment must be anchored if it is permanently attached to utility services (electricity, gas, and water). For the
purposes of this requirement, “permanently attached” should be understood to include all electrical connections except
NEMA 515 and 520 straightblade connectors (duplex receptacles).
C13.1.2 Seismic Design Category. The requirements for nonstructural components are based in part on the Seismic Design
Category to which they are assigned. As the Seismic Design Category is established considering factors not unique to
specific nonstructural components, all nonstructural components occupying or attached to a structure are assigned to the same
Seismic Design Category as the structure.
C13.1.3 Component Importance Factor. Performance expectations for nonstructural components often are defined in
terms of the functional requirements of the structure to which the components are attached. While specific performance goals
for nonstructural components have yet to be defined in building codes, the component importance factor (Ip) implies
performance levels for specific cases. For noncritical nonstructural components (those with an importance factor, Ip, of 1.0)
the following behaviors are anticipated for shaking having different levels of intensity:
1. Minor earthquake ground motions – minimal damage; not likely to affect functionality
2. Moderate earthquake ground motions – some damage that may affect functionality
3. Design earthquake ground motions – major damage but significant falling hazards are avoided; likely loss of
functionality.
Figure C13.16 Image of Skidmounted components.
Figure C13.16 Skidmounted components.
Components with importance factors greater than 1.0 are expected to remain in place, sustain limited damage, and, when
necessary, function following an earthquake (see Section C13.2.2). These components can be located in structures that are
not assigned to Occupancy Category IV. For example, fire sprinkler piping systems have an importance factor, Ip, of 1.5 in
all structures since these essential systems should function following an earthquake.
The component importance factor is intended to represent the greater of the lifesafety importance of the component and the
hazardexposure importance of the structure. It indirectly influences the survivability of the component via required design
forces and displacement levels as well as component attachments and detailing. While this approach provides some degree of
confidence in the seismic performance of a component, it may not be sufficient in all cases. For example, individual ceiling
tiles may fall from a ceiling grid that has been designed for larger forces. This may not represent a serious falling hazard if
the ceiling tiles are made of lightweight materials, but it may lead to blockage of critical egress paths or disruption of the
facility function. When higher levels of confidence in performance are required, the component is classified as a designated
seismic system (Section 11.2), and, in certain cases, seismic qualification of the component or system is necessary. Seismic
qualification approaches are provided in Sections 13.2.5 and 13.2.6. In addition, seismic qualification approaches presently
in use by the Department of Energy (DOE) can be applied.
Occupancy Category IV structures are intended to be functional following a design earthquake; critical nonstructural
components and equipment in such structures are designed with Ip equal to 1.5. This requirement applies to most
components and equipment since damage to vulnerable unbraced systems or equipment may disrupt operations following an
earthquake even if they are not directly classified as essential to life safety. The nonessential/nonhazardous components
themselves are not assessed, and requirements focus solely on the supports and attachments. UFC 331004 has additional
guidance for improved performance.
C13.1.4 Exemptions. Some nonstructural components either possess inherent strength and stability, are subject to lowlevel
earthquake demands (accelerations and relative displacements), or both. Since these nonstructural components and systems
are expected to achieve the performance goals described earlier in this commentary without explicitly satisfying additional
requirements, they are exempt from the requirements of Chapter 13.
Chapter 13 does not apply to Seismic Design Category A due to its very low level of seismic hazard. (See Section C11.7.)
With the exception of parapets supported by bearing walls or shear walls, all components in Seismic Design Category B are
exempt due to the minimal level of seismic risk. Parapets are not exempt because experience has shown these items can fail
and pose a significant falling hazard even at low shaking levels.
Mechanical and electrical components in Seismic Design Category C with an importance factor (Ip) equal to 1.0 are exempt
because they are subject to low levels of seismic hazard, they do not contain hazardous substances, and their function is not
required to maintain life safety following an earthquake. Small components with Ip equal to 1.0 in Seismic Design
Categories D, E, and F also are exempt since they do not contain hazardous substances and are not large enough to pose a
lifesafety hazard if they fall, slide, or topple. Failures of unbraced distribution systems at or near the point of connection to
nonstructural components have been observed in past earthquakes. For this reason, flexible connections such as expansion
loops, braided hose, or expansion joints are required to allow for the larger relative displacements associated with unbraced
components. Note that the stiffness of flexible connections may be sensitive to internal pressure and length of the
connection.
C13.1.5 Applicability of Nonstructural Component and Requirements. At times, a nonstructural component should be
treated as a nonbuilding structure. When the physical characteristics associated with a given class of nonstructural
components vary widely, judgment is needed to select the appropriate design procedure and coefficients. For example,
cooling towers vary from small packaged units with an operating weight of 2,000 pounds or less to structures the size of
buildings. Consequently, design coefficients for the design of “cooling towers” are found both in Table 13.61 and Table
15.42. Small cooling towers are best designed as nonstructural components using the provisions of Chapter 13 while large
ones are clearly nonbuilding structures that are more appropriately designed using the provisions of Chapter 15. Similar
issues arise for other classes of nonstructural component (e.g., boilers and bins). Guidance on determining whether an item
should be treated as a nonbuilding structure or nonstructural component for the purpose of seismic design is provided in
Bachman and Dowty (2008).
There are practical limits on the size of a component that can be qualified via shake table testing. Components too large to be
qualified by shake table testing need to be qualified by a combination of structural analysis and qualification testing or
empirical evaluation through a subsystem approach. Subsystems of a large, complex component (e.g., a very large chiller)
can be qualified individually and the overall structural frame of the component evaluated by structural analysis
Premanufactured modular mechanical units are considered nonbuilding structures supporting nonstructural components. The
entire system (all modules assembled) can be shake table qualified or qualified separately as subsystems. Modular
mechanical units house various nonstructural components that are subject to all the design requirements of Chapter 13.
The specified weight limit for nonstructural components (25 percent relative to the combined weight of the structure and
component) relates to the condition at which dynamic interaction between the component and the supporting structural
system is potentially significant. Section 15.3.2 contains requirements for addressing this interaction in design.
C13.1.6 Reference Documents. Professional and trade organizations have developed nationally recognized codes and
standards for the design and construction of specific mechanical and electrical components. These documents provide design
guidance for normal and upset (abnormal) operating conditions and for various environmental conditions. Some of these
documents include earthquake design requirements in the context of the overall mechanical or electrical design. It is the
intent of the standard that seismic requirements in referenced documents be used. The developers of these documents are
familiar with the expected performance and failure modes of the components; however, the documents may be based on
design considerations not immediately obvious to a structural design professional. For example, in the design of industrial
piping, stresses due to seismic inertia forces typically are not added to those due to thermal expansion.
There is a potential for misunderstanding and misapplication of reference documents for the design of mechanical and
electrical systems. A registered design professional familiar with both the standard and the reference documents used should
be involved in the review and acceptance of the seismic design.
Even when reference documents for nonstructural components lack specific earthquake design requirements, mechanical and
electrical equipment constructed in accordance with industrystandard reference documents have performed well historically
when properly anchored. Nevertheless, it is expected that manufacturers of mechanical and electrical equipment will
consider seismic loads in the design of the equipment itself even when not explicitly required by this chapter.
While some reference documents provide requirements for seismic capacity appropriate to the component being designed, the
seismic demands used in design may not be less than those specified in the standard.
Specific guidance for selected mechanical and electrical components and conditions is provided in Section 13.6.
C13.1.7 Reference Documents Using Allowable Stress Design. Many nonstructural components are designed using
specifically developed reference documents that are based on allowable stress loads and load combinations and permit
increases in allowable stresses for seismic loading. Although Section 2.4.1 of the standard does not permit increases in
allowable stresses, Section 13.1.7 explicitly defines the conditions for their use in the design of nonstructural components.
C13.2 GENERAL DESIGN REQUIREMENTS
C13.2.1 Applicable Requirements for Architectural, Mechanical, and Electrical Components, Supports, and
Attachments. Compliance with the requirements of Chapter 13 may be accomplished by projectspecific design or by a
manufacturer’s certification of seismic qualification of a system or component. In each case, the evidence of compliance is
submitted to the authority having jurisdiction. When compliance is by manufacturer's certification, the items must be
installed in accordance with the manufacturer’s requirements.
Components addressed by the standard include individual simple units and assemblies of simple units for which reference
documents establish seismic analysis or qualification requirements. Also addressed by the standard are complex
architectural, mechanical, and electrical systems for which reference documents either do not exist or exist for only elements
of the system. In the design and analysis of both simple components and complex systems, the concepts of flexibility and
ruggedness often can assist the designer in determining the necessity for analysis and, when analysis is necessary, the extent
and methods by which seismic adequacy may be determined. These concepts are discussed in Section C13.6.1.
C13.2.2 Special Certification Requirements for Designated Seismic Systems. While the goal of design for most
nonstructural components is to prevent detachment or toppling that would pose a hazard to life safety, designated seismic
systems (with Ip = 1.5) are intended to meet higher performance goals. In some cases, failure of mechanical or electrical
equipment itself poses a significant hazard. This section addresses the design and certification of designated seismic system
components and their supports and attachments. The goals of this section are to improve survivability and achieve a high
level of confidence in postearthquake functionality, and they require additional considerations.
Examples of designated seismic systems include fire protection piping, uninterruptible power supplies for hospitals, and
certain vessels or piping that contain highly toxic or explosive substances.
Using an importance factor, Ip, equal to 1.5 to increase design forces can reduce the possibility of detachment or toppling, but
this directly affects only structural integrity and stability; function and operability of mechanical and electrical components
may be affected only indirectly by increasing design forces. For complex components, testing or experience may be the only
reasonable way to improve the confidence of function and operability. For many types of equipment, past earthquake
experience has shown that maintaining structural integrity and stability provides postearthquake function and operability.
On the other hand, mechanical joints in containment components (e.g., tanks, vessels, and piping) may not remain leaktight
in an earthquake. Avoiding this condition may require a design that is more conservative than that required by the standard.
Evaluating postearthquake operational performance by analysis is impractical for active mechanical and electrical equipment
with components that rotate or otherwise move mechanically during operation. Active equipment includes pumps and
electric motors. In many cases, such equipment is inherently rugged, and an evaluation of experience data together with
analysis of the component anchorage is adequate to demonstrate compliance (see Section 13.6). In other cases (e.g., motor
control centers and switching equipment), shake table testing may be required. Components that contain hazardous materials
(e.g., tanks, piping, and vessels) typically are qualified by analysis, but it may be necessary to qualify certain operational
valving or mechanical equipment within the system by other means.
C13.2.3 Consequential Damage. Although the components identified in Tables 13.51 and 13.61 are listed separately,
significant interrelationships exist and must be considered. Consequential damage occurs due to interaction between
components and systems. Even “braced” components displace and the displacement between lateral supports can be
significant in the case of distributed systems such as piping systems, cable and conduit systems and other linear systems. It is
the intent of the standard that the seismic displacements considered include both relative displacement between multiple
points of support (addressed in Section 13.3.2) and, for mechanical and electrical components, displacement within the
component assemblies. Impact of components must be avoided unless the components are fabricated of ductile materials that
have been shown to be capable of accommodating the expected impact loads. With protective coverings, ductile mechanical
and electrical components and many more fragile components are expected to survive all but the most severe impact loads.
Flexibility and ductility of the connections between distribution systems and the equipment to which they attach is essential
to the seismic performance of the system.
The determination of the displacements that generate these interactions are not addressed explicitly in Section 13.3.2.1. That
section concerns relative displacement of support points. Consequential damage may occur due to displacement of
components and systems between support points. For example, in older suspended ceiling installations, excessive lateral
displacement of a ceiling system may fracture sprinkler heads that project through the ceiling. A similar situation may arise
if sprinkler heads projecting from a small diameter branch line pass through a rigid ceiling system. While the branch line
may be properly restrained, it may still displace sufficiently between lateral support points, to impact other components or
systems. Similar interactions occur where a relatively flexible distributed system connects to a braced or rigid component.
Figure C13.21 Schematic plans illustrating branch line flexibility.
Sufficient flexibility may be
achieved in northsouth and
vertical directions only
Flexibility may be achieved in
northsouth, eastwest, and
vertical directions
Distribution system
Component
or
structure
Component
or
structure
Anchors
Component
or
structure
Distribution
system
Component
or
structure
Anchor
Anchor
N
The potential for impact between components that are in contact with or in close proximity to other structural or nonstructural
components must be considered. However, where considering these potential interactions, the designer must determine if the
potential interaction is both credible and significant. For example, the fall of a ceiling panel located above a motor control
center is a credible interaction because the falling panel in older suspended ceiling installations can reach and impact the
motor control center. An interaction is significant if it can result in damage to the target. Impact of a ceiling panel on a
motor control center may not be significant, due to the light weight of the ceiling panel. Special design consideration is
appropriate where the failure of a nonstructural element could adversely influence the performance of an adjacent critical
nonstructural component, such as an emergency generator.
C13.2.4 Flexibility. In many cases, flexibility is more important than strength in the performance of distributed systems,
such as piping and ductwork. A good understanding of the displacement demand on the system as well as its displacement
capacity is required. Components or their supports and attachments must be flexible enough to accommodate the full range
of expected differential movements; some localized inelasticity is permitted in accommodating the movements. Relative
movements in all directions must be considered. For example, even a braced branch line of a piping system will displace, so
it needs to be connected to other braced or rigid components in a manner that will accommodate the displacements without
failure (see Figure C13.21). For another example, cladding units (such as precast concrete wall units) while often very rigid
inplane, if supported at more than one level, require connections capable of accommodating story drift. (See Section
C13.5.3 for an illustration.)
If component analysis assumes rigid anchors or supports, the predicted loads and local stresses can be unrealistically large, so
it may be necessary to consider anchor and/or support stiffness.
Figure C13.21 Schematic plans illustrating branch line flexibility.
C13.2.5 Testing Alternative for Seismic Capacity Determination. Testing is a well established alternative method of
seismic qualification for small to mediumsize equipment. Several national reference documents have testing requirements
adaptable for seismic qualification. One such reference document, ICCES AC156 (2007), is a shaketable testing protocol
that has been adopted by the ICC Evaluation Service. It was developed specifically to be consistent with acceleration
demands (that is, force requirements) of the standard.
The development or selection of testing and qualification protocols should at a minimum include the following:
1. Description of how the protocol meets the intent for the projectspecific requirements and relevant interpretations of the
standard.
2. Definition of a test input motion with a response spectrum that meets or exceeds the Design Earthquake spectrum for the
site.
3. Accounting for dynamic amplification due to abovegrade equipment installations. Consideration of the actual dynamic
characteristics of the primary support structure is permitted, but not required.
4. Definition of how shaketable input demands were derived.
5. Definition and establishment of a verifiable pass/fail acceptance criterion for the seismic qualification based upon the
equipment importance factor and consistent with the building code and projectspecific design intent.
6. Development of criteria that can be used to rationalize test unit configuration requirements for highly variable equipment
product lines.
To aid the design professional in assessing the adequacy of the manufacturer’s certificate of compliance it is recommended
that certificates of compliance include:
1. Product family or group covered
2. Building code(s) and standard(s) for which compliance was evaluated
3. Testing standard used
4. Performance objective and corresponding importance factor (Ip = 1.0 or Ip = 1.5)
5. Seismic demand for which the component is certified, including code and/or standard design parameters used to
calculate seismic demand (such as values used for ap, Rp, and site class)
6. Installation restrictions, if any (grade, floor, or roof level)
Without a test protocol recognized by the building code, qualification testing is inconsistent and difficult to verify. The use
of ICCES AC156 simplifies the task of compliance verification since it was developed to address directly the testing
alternative for nonstructural components, as specified in the standard. It also sets forth minimum test plan and report
deliverables.
Use of other standards or adhoc protocols to verify compliance of nonstructural components with the requirement of the
standard should be considered carefully and used only where projectspecific requirements cannot be met otherwise.
Where other qualification test standards will be used, in whole or in part, it is necessary to verify compliance with this
standard. For example, IEEE 693 indicates that it is to be used for the sole purpose of qualifying electrical equipment
(specifically listed in the standard) for use in utility substations. Where equipment testing has been conducted to other
standards (for instance, testing done in compliance with IEEE 693), a straightforward approach would be to permit
evaluation, by the manufacturer, of the test plan and data to validate compliance with the requirements of ICCES AC156,
because it was developed specifically to comply with the seismic demands of this standard.
The qualification of mechanical and electrical components for seismic loads alone may not be sufficient to achieve high
performance objectives. Establishing a high confidence that performance goals will be met requires consideration of the
performance of structures, systems (fluid, mechanical, electrical, instrumentation, etc.), and their interactions (for example
interaction of seismic and other loads) as well as compliance with installation requirements.
C13.2.6 Experience Data Alternative for Seismic Capacity Determination. An established method of seismic
qualification for certain types of nonstructural components is the assessment of data for the performance of similar
components in past earthquakes. The seismic capacity of the component in question is extrapolated based on estimates of the
demands (force, displacement) to which the components in the database were subjected. Procedures for such qualification
have been developed for use in nuclear facility applications by the Seismic Qualification Utility Group (SQUG) of the
Electric Power Research Institute.
The SQUG rules for implementing the use of experience data are described in a proprietary Generic Implementation
Procedure (GIP) database. It is a collection of findings from detailed engineering studies by experts for equipment from a
variety of utility and industrial facilities.
Valid use of experience data requires satisfaction of rules that address physical characteristics, manufacturer’s classification
and standards, and findings from testing, analysis, and expert consensus opinion.
Four criteria are used to establish seismic qualification by experience, as follows:
1. Seismic capacity versus demand (a comparison with a bounding spectrum)
2. Earthquake experience database cautions and inclusion rules
3. Evaluation of anchorage
4. Evaluation of seismic interaction
Experience data should be used with care, since the design and manufacture of components may have changed considerably
in the intervening years. The use of this procedure is also limited by the relative rarity of strong motion instrument records
associated with corresponding equipment experience data.
C13.2.7 Construction Documents. Where the standard requires seismic design of components or their supports and
attachments, appropriate construction documents defining the required construction and installation must be prepared. This
facilitates the special inspection and testing needed to provide a reasonable level of quality assurance. Of particular concern
are large nonstructural components (such as rooftop chillers) whose manufacture and installation involves multiple trades and
suppliers, and which impose significant loads on the supporting structure. In these cases, it is important that the construction
documents used by the various trades and suppliers to satisfy the seismic design requirements are prepared by a registered
design professional.
The information required to prepare construction documents for component installation includes the dimensions of the
component, the locations of attachment points, the operating weight, and the location of the center of mass. For instance, if
an anchorage angle will be attached to the side of a metal chassis, the gage and material of the chassis must be known so that
the number and size of required fasteners can be determined. Or, when a piece of equipment has a base plate that will be
anchored to a concrete slab with expansion anchors, the drawings must show the base plate’s material and thickness, the
diameter of the bolt holes in the plate, and the size and depth of embedment of the anchor bolts. If the plate will be elevated
above the slab for leveling, the construction documents must also show the maximum gap permitted between the plate and
the slab.
C13.3 SEISMIC DEMANDS ON NONSTRUCTURAL COMPONENTS
The seismic demands on nonstructural components, as defined in this section, are acceleration demands and relative
displacement demands. Acceleration demands are represented by equivalent static forces. Relative displacement demands
are provided directly and are based on either the actual displacements computed for the structure or the maximum allowable
drifts that are permitted for the structure.
C13.3.1 Seismic Design Force. The seismic design force for a component depends on the weight of the component, the
component importance factor, the component response modification factor, the component amplification factor, and the
component acceleration at a point of attachment to the structure. The forces prescribed in this section of the standard reflect
the dynamic and structural characteristics of nonstructural components. As a result of these characteristics, forces used for
verification of component integrity and design of connections to the supporting structure typically are larger than those used
for design of the overall seismicforceresisting system.
Certain nonstructural components lack the desirable attributes of structures (such as ductility, toughness, and redundancy)
that permit the use of greatly reduced lateral design forces. Thus values for the response modification factor, Rp, in Tables
13.51 and 13.61 generally are smaller than R values for structures. These Rp values, used to represent the energy absorption
capability of a component and its attachments, depend on both overstrength and deformability. At present these potentially
separate considerations are combined in a single factor. The tabulated values are based on the collective judgment of the
responsible committee.
The 2005 edition of the standard includes significant adjustments to tabulated Rp values for certain mechanical and electrical
systems. For example, the value of Rp for welded steel piping systems is increased from 3.5 to 9. The ap value increased
from 1.0 to 2.5, so while it might appear that forces on such piping systems have been reduced greatly, the net change is
negligible, as Rp/ap changes from 3.5 to 3.6. The minimum seismic design force of Equation 13.33, which governs in many
cases, is unchanged.
The component amplification factor (ap) represents the dynamic amplification of component responses as a function of the
fundamental periods of the structure (T) and component (Tp). When components are designed or selected, the structural
fundamental period is not always defined or readily available. The component fundamental period (Tp) is usually only
accurately obtained by shaketable or pullback tests and is not available for the majority of components. Tabulated ap
values are based on component behavior that is assumed to be either rigid or flexible. Where the fundamental period of the
component is less than 0.06 seconds, dynamic amplification is not expected, and the component is considered rigid. The
tabulation of assumed ap values is not meant to preclude more precise determination of the component amplification factor
where the fundamental periods of both structure and component are available. The NCEER formulation shown in Figure
C13.31 may be used to compute ap as a function of Tp/T.
Dynamic amplification occurs where the period of a nonstructural component closely matches that of any mode of the
supporting structure, although this effect may not be significant depending on the ground motion. For most buildings, the
primary mode of vibration in each direction will have the most influence on the dynamic amplification for nonstructural
components. For longperiod structures (such as tall buildings), where the period of vibration of the fundamental mode is
greater than 3.5 times Ts, higher modes of vibration may have periods that more closely match the period of nonstructural
components. For this case, it is recommended that amplification be considered using such higher mode periods in lieu of the
higher fundamental period. This approach may be generalized by computing floor response spectra for various levels that
reflect the dynamic characteristics of the supporting structure to determine how amplification will vary as a function of
component period. Calculation of floor response spectra can be complex, but simplified procedures are presented in Kehoe
and Hachem (2003). Consideration of nonlinear behavior of the structure greatly complicates the analysis.
Figure C13.31 NCEER formulation for ap as function of structural and component periods.
T /T
a
2.5
1.4 2.0 p
1.0
p
0.5 0.7
NCEER study Figure C13.32 Lateral force magnitude over height.
Figure C13.31 NCEER formulation for ap as function of structural and component periods.
Equation 13.31 represents a trapezoidal distribution of floor accelerations within a structure, varying linearly from the
acceleration at the ground (taken as 0.4SDS) to the acceleration at the roof (taken as 1.2SDS). The ground acceleration
(0.4SDS) is intended to be the same acceleration used as design input for the structure itself, including site effects. The roof
acceleration is established as three times the input ground acceleration based on examination of recorded instructure
acceleration data for short and moderate height structures in response to large California earthquakes. Work by Miranda and
Singh suggest that, for taller structures, the amplification with height may vary significantly due to higher mode effects.
Where more information is available, Equation 13.34 permits an alternate determination of the component design forces
based on the dynamic properties of the structure.
Equation 13.33 establishes a minimum seismic design force, Fp, that is consistent with current practice. Equation 13.32
provides a simple maximum value of Fp that prevents multiplication of the individual factors from producing a design force
that would be unreasonably high, considering the expected nonlinear response of support and component. Figure C13.32
illustrates the distribution of the specified lateral design forces.
Figure C13.32 Lateral force magnitude over height.
For elements with points of attachment at more than one height, it is recommended that design be based on the average of
values of Fp determined individually at each point of attachment (but with the entire component weight, Wp) using Equations
13.31 through 13.33.
Alternatively, for each point of attachment a force Fp may be determined using Equations 13.31 through 13.33, with the
portion of the component weight, Wp, tributary to the point of attachment. For design of the component, the attachment force
Fp must be distributed relative to the component’s mass distribution over the area used to establish the tributary weight. To
Figure C13.33 Displacements over less than story height.
Flexible glazing
system
Rigid spandrels
hx
x
dy
hsx
hy
illustrate these options, consider a solid exterior nonstructural wall panel, supported top and bottom, for a onestory building
with a rigid diaphragm. The values of Fp computed, respectively, for the top and bottom attachments using Equations 13.31
through 13.33 are 0.48SDSIpWp and 0.30SDSIpWp. In the recommended method, a uniform load is applied to the entire panel
based on 0.39SDSIpWp. In the alternative method, a trapezoidal load varying from 0.48SDSIpWp at the top to 0.30SDSIpWp at
the bottom is applied. Each anchorage force is then determined considering static equilibrium of the complete component
subject to all the distributed loads.
Cantilever parapets that are part of a continuous element should be checked separately for parapet forces. The seismic force
on any component must be applied at the center of gravity of the component and must be assumed to act in any horizontal
direction. Vertical forces on nonstructural components equal to ±0.2SDSWp are specified in Section 13.3.1 and are intended
to be applied to all nonstructural components and not just cantilevered elements. Nonstructural concrete or masonry walls
laterally supported by flexible diaphragms must be anchored outofplane in accordance with Section 12.11.2.
C13.3.2 Seismic Relative Displacements. The equations of this section are for use in design of cladding, stairways,
windows, piping systems, sprinkler components, and other components connected to one structure at multiple levels or to
multiple structures. Two equations are given for each situation. Equations 13.35 and 13.37 produce structural
displacements as determined by elastic analysis, unreduced by the structural response modification factor (R). Since the
actual displacements may not be known when a component is designed or procured, Equations 13.36 and 13.38 provide
upperbound displacements based on structural drift limits. Use of upperbound equations may facilitate timely design and
procurement of components, but may also result in costly added conservatism.
The standard does not provide explicit acceptance criteria for the effects of seismic relative displacements, except for glazing.
Damage to nonstructural components due to relative displacement is acceptable, provided the performance goals defined
elsewhere in the chapter are achieved.
C13.3.2.1 Displacements within Structures. Seismic relative displacements can subject components or systems to
unacceptable stresses. Nonstructural components designed with no intended structural function, such as infill walls, may
interact with structural framing elements as a result of building deformation. The resulting stresses may exceed acceptable
limits for the nonstructural components, the structural elements, or both. Consideration of this interrelationship is likely to
govern the clearance between such components and the ductility and strength of their supports and attachments.
Where nonstructural components are supported between, rather than at, structural levels, as frequently occurs for glazing
systems, partitions, stairs, veneers, and mechanical and electrical distributed systems, the height over which the displacement
demand, Dp, must be accommodated may be less than the story height, hsx, and should be considered carefully. For example,
consider the glazing system supported by rigid precast concrete spandrels shown in Figure C13.33. The glazing system will
be subjected to full story drift, Dp, although its height (hx – hy) is only a fraction of the story height. The design drift must be
accommodated by anchorage of the glazing unit, the joint between the precast spandrel and the glazing unit, or some
combination of the two. Similar displacement demands arise where pipes, ducts, or conduit that are braced to the floor or
roof above are connected to the top of a tall, rigid, floormounted component.
Figure C13.33 Displacements over less than story height.
For ductile components, such as steel piping fabricated with welded connections, the relative seismic displacements between
support points can be more significant than inertial forces. Ductile piping can accommodate relative displacements by local
Figure C13.34 Displacements between structures.
Case A Case B
A
B
A
B
d
xA
h
y
h
x
h
y
h
x
d
xB
d
xA
d
xB
yielding with strain accumulations well below failure levels. However, for components fabricated using less ductile
materials, where local yielding must be avoided to prevent unacceptable failure consequences, relative displacements must be
accommodated by flexible connections.
C13.3.2.2 Displacements between Structures. A component or system connected to two structures must accommodate
horizontal movements in any direction, as illustrated in Figure C13.34.
Figure C13.34 Displacements between structures.
C13.4 NONSTRUCTURAL COMPONENT ANCHORAGE
Unless exempted in Section 13.1.4, components must be anchored to the structure, and all required supports and attachments
must be detailed in the construction documents. To satisfy the load path requirement of this section, the detailed information
described in Section C13.2.7 must be communicated, during the design phase, to the registered design professional
responsible for the design of the supporting structure.
Unanchored components often rock or slide when subjected to earthquake motions. Since this behavior may have serious
consequences, is difficult to predict, and is exacerbated by vertical ground motions, positive restraint must be provided for
each component.
The effective seismic weight used in design of the seismic forceresisting system must include the weight of supported
components. To satisfy the load path requirements of this section, localized component demand must also be considered.
This may be accomplished by checking the capacity of the first structural element in the load path (for example, a floor beam
directly under a component) for combined dead, live, operating, and seismic loads, using the horizontal and vertical loads
from Section 13.3.1 for the seismic demand, and repeating this procedure for each structural element or connection in the
load path until the load case including horizontal and vertical loads from Section 13.3.1 no longer governs design of the
element. The load path includes housekeeping slabs and curbs, which must be adequately reinforced and positively fastened
to the supporting structure.
Since the exact magnitude and location of loads imposed on the structure may not be known until nonstructural components
are ordered, the initial design of supporting structural elements should be based on conservative assumptions. The design of
the supporting structural elements must be verified once the final magnitude and location of the design loads have been
established.
Tests have shown there are consistent shear ductility variations between bolts installed in drilled or punched plates with nuts
and connections using welded shear studs. The need for reductions in allowable loads for particular anchor types to account
for loss of stiffness and strength may be determined through appropriate dynamic testing. Although comprehensive design
recommendations are not available at present, this issue should be considered for critical connections subject to dynamic or
seismic loading.
C13.4.2 Anchors in Concrete or Masonry. Design capacity for anchors in concrete must be determined in accordance with
ACI 318 Appendix D. Design capacity for anchors in masonry is determined in accordance with ACI 530. Anchors must be
designed to have ductile behavior or to provide a specified degree of excess strength. In either case, design forces are
multiplied by 1.3 or based on the capacity of the component or its supports. The anchorage criteria provided in Chapter 13
specifically address the issue of nonductile response and force amplification. Since the capacity of anchors in masonry is
rarely governed by steel capacity, and failure in the masonry is nonductile, an Rp of 1.5 should be used for design.
Depending on the specifics of the design condition, ductile design of anchors in concrete may satisfy one or more of the
following objectives:
1. Adequate load redistribution between anchors in a group
2. Allowance for anchor overload without brittle failure
3. Energy dissipation
Achieving deformable, energyabsorbing behavior in the anchor itself is often difficult. Unless the design specifically
addresses the conditions influencing desirable hysteretic response (adequate gauge length, anchor spacing, edge distance,
steel properties, etc.), anchors cannot be relied upon for energy dissipation. Simple geometric rules, such as restrictions on
the ratio of anchor embedment length to depth, are not adequate to produce reliable ductile behavior. For example, a single
anchor with sufficient embedment to force ductile tension failure in the steel body of the anchor bolt may still experience
concrete fracture (a nonductile failure mode) if the edge distance is small, the anchor is placed in a group of tensionloaded
anchors with reduced spacing, or the anchor is loaded in shear instead of tension. In the common case where anchors are
subject primarily to shear, response governed by the steel element may be nonductile if the deformation of the anchor is
constrained by rigid elements on either side of the joint. Designing the attachment so that its response is governed by a
deformable link in the load path to the anchor is encouraged. This approach provides ductility and overstrength in the
connection while protecting the anchor from overload. Ductile bolts should only be relied upon as the primary ductile
mechanism of a system if the bolts are designed to have adequate gauge length (unbonded strained length of the bolt) to
accommodate the anticipated nonlinear displacements of the system at the design earthquake. Guidance for determining the
gauge length can be found in Part 3 of the Provisions.
Postinstalled expansion and undercut anchors must be qualified in accordance with ACI 355.204, Qualification of Post
Installed Mechanical Anchors in Concrete. The ICCES acceptance criteria AC193 and AC308, which include specific
provisions for screw anchors and adhesive anchors, also reference ACI 355.2. Reference to adhesives (such as in
Section 13.5.7.2) apply, not to adhesive anchors, but to steel plates and other structural elements bonded or glued to the
surface of another structural component with adhesive; such connections are generally nonductile.
Anchors used to support towers, masts, and equipment are often provided with double nuts for leveling during installation.
Where baseplate grout is specified at anchors with double nuts, it should not be relied upon to carry loads since it can shrink
and crack or be omitted altogether. The design should include the corresponding tension, compression, shear, and flexure
loads.
C13.4.3 Installation Conditions. Prying forces on anchors, which result from a lack of rotational stiffness in the connected
part, can be critical for anchor design and must be considered explicitly.
For anchorage configurations that do not provide a direct mechanism to transfer compression loads (for example, a base plate
that does not bear directly on a slab or deck but is supported on a threaded rod), the design for overturning must reflect the
actual stiffness of baseplates, equipment, housing, and other elements in the load path when computing the location of the
compression centroid and the distribution of uplift loads to the anchors.
C13.4.4 Multiple Attachments. While the standard does not prohibit the use of single anchor connections, it is good
practice to use at least two anchors in any loadcarrying connection whose failure might lead to collapse, partial collapse, or
disruption of a critical load path.
C13.4.5 Power Actuated Fasteners. The capacity of power actuated fasteners in concrete often varies more than that of
drilled postinstalled anchors. The shallow embedment, small diameter, and friction mechanism of these fasteners make
them particularly susceptible to the effects of concrete cracking. The suitability of power actuated fasteners to resist tension
in concrete should be demonstrated by simulated seismic testing in cracked concrete.
Where properly installed in steel, power actuated fasteners typically exhibit reliable cyclic performance. Nevertheless, they
should not be used singly to support suspended elements. Where used to attach cladding and metal decking, subassembly
testing may be used to establish design capacities since the interaction between the decking, the subframe, and the fastener
can only be estimated crudely by currently available analysis methods.
C13.4.6 Friction Clips. Friction clips, such as beam clamps, may loosen under cyclic loading, resulting in slippage or loss
of connection capacity. Where friction clips are used, they may not be relied upon for seismic resistance. Fasteners that
provide a positive mechanical connection have more reliable seismic performance. Clips that provide marginal mechanical
connection, such as beam clamps that “dimple” the flange of the steel support may still rely chiefly on friction. These may
not provide adequate cyclic capacity and should be qualified by seismic testing.
C13.5 ARCHITECTURAL COMPONENTS
For structures in Occupancy Category I through III, the requirements of Section 13.5 are intended to reduce property damage
and lifesafety hazards posed by architectural components due to loss of stability or integrity. When subjected to seismic
motion, components may pose a direct falling hazard to building occupants or to people outside the building (as in the case of
parapets, exterior cladding, and glazing). Failure or displacement of interior components (such as partitions and ceiling
systems in exits and stairwells) may block egress.
For structures in Occupancy Category IV, the potential disruption of essential function due to component failure must also be
considered.
Architectural component failures in earthquakes can be caused by deficient design or construction of the component,
interrelationship with another component that fails, interaction with the structure, or inadequate attachment or anchorage. For
architectural components, attachment and anchorage are typically the most critical concerns related to their seismic
performance. Concerns regarding loss of function are most often associated with mechanical and electrical components.
Architectural damage, unless very severe, can be accommodated temporarily. Very severe architectural damage is often
accompanied by significant structural damage.
C13.5.1 General. Suspended architectural components are not required to satisfy the force and displacement requirements of
Chapter 13, where prescriptive requirements are met. The requirements were relaxed in the 2005 edition of the standard to
better reflect the consequences of the expected behavior. For example, impact of a suspended architectural ornament with a
sheet metal duct may only dent the duct without causing a credible danger (assuming the ornament remains intact). The
reference to Section 13.2.3 allows the designer to consider such consequences in establishing the design approach.
C13.5.2 Forces and Displacements. Partitions and interior and exterior glazing must accommodate story drift without
failure that will cause a lifesafety hazard. Design judgment must be used to assess potential lifesafety hazards and the
likelihood of lifethreatening damage. Special detailing to accommodate drift for typical gypsum board or demountable
partitions is unlikely to be costeffective, and damage to these components poses a low hazard to life safety. Damage in these
partitions occurs at low drift levels, but is inexpensive to repair.
If they must remain intact following strong ground motion, nonstructural fireresistant enclosures and firerated partitions
require special detailing that provides isolation from the adjacent or enclosing structure for deformation equivalent to the
calculated drift (relative displacement). Inplane differential movement between structure and wall is permitted. Provision
must be made for outofplane restraint. These requirements are particularly important in steel or concrete moment frame
structures, which experience larger drifts. The problem is less likely to be encountered in stiff structures, such as those with
shear walls.
Differential vertical movement between horizontal cantilevers in adjacent stories (such as cantilevered floor slabs) has
occurred in past earthquakes. The possibility of such effects should be considered in the design of exterior walls.
C13.5.3 Exterior Nonstructural Wall Elements and Connections. Nonbearing wall panels that are attached to and enclose
the structure must be designed to resist seismic (inertial) forces, wind forces, and gravity forces and to accommodate
movements of the structure resulting from lateral forces and temperature change. The connections must allow wall panel
movements due to thermal and moisture changes, and be designed so as to prevent the loss of loadcarrying capacity in the
event of significant yielding. Where wind loads govern, common practice is to design connectors and panels to allow for not
less than two times the story drift caused by wind loads determined using a return period appropriate to the site location.
Design to accommodate seismic relative displacements often presents a greater challenge than design for strength. Story
drifts can amount to 2 inches (50 mm) or more. Separations between adjacent panels are intended to limit contact and
resulting panel misalignment or damage under all but extreme building response. Section 13.5.3(a) calls for a minimum
separation of 1/2 inch (13 mm). For practical joint detailing and acceptable appearance, separations typically are limited to
about 3/4 inch (19 mm). Manufacturing and construction tolerances for both wall elements and the supporting structure must
be considered in establishing design joint dimensions and connection details.
Cladding elements, which are often very stiff inplane, must be isolated so that they do not restrain and are not loaded by drift
of the supporting structure. Slotted connections can provide isolation, but connections with long rods that flex achieve the
desired behavior without requiring precise installation. Such rods must be designed to resist tension and compression in
addition to induced flexural stresses, brittle, lowcycle fatigue failure.
Fullstory wall panels are usually rigidly attached to and move with the floor structure nearest the panel bottom and isolated
at the upper attachments. Panels also can be vertically supported at the top connections with isolation connections at the
bottom. An advantage of this configuration is that failure of an isolation connection is less likely to result in complete
detachment of the panel, since it will tend to rotate into the structure rather than away from it.
To minimize the effects of thermal movements and shrinkage on architectural cladding panels, connection systems are
generally detailed to be statically determinate. Since the resulting support systems often lack redundancy, exacerbating the
consequences of a single connection failure, fasteners must be designed for amplified forces and connecting members must
be ductile. The intent is to keep inelastic behavior in the connecting members while the more brittle fasteners remain
essentially elastic. To achieve this intent, the tabulated ap and Rp values produce fastener design forces that are about 3 times
those for the connecting members.
Limited deformability curtain walls, such as aluminum systems, are generally light and can undergo large deformations
without separating from the structure. However, care must be taken in design of these elements so that low deformability
components(as defined in Section 11.2) that may be part of the system, such as glazing panels, are detailed to accommodate
the expected deformations without failure.
In Table 13.51, veneers are classified as either limited or low deformability elements. Veneers with limited deformability,
such as vinyl siding, pose little risk. Veneers with low deformability, such as brick and ceramic tile, are highly sensitive to
the performance of the supporting substrate. Significant distortion of the substrate results in veneer damage, possibly
including separation from the structure. The resulting risk depends on the size and weight of fragments likely to be dislodged
and on the height from which the fragments would fall. Detachment of large portions of the veneer can pose a significant
risk to life. Such damage can be reduced by isolating veneer from displacements of the supporting structure. For structures
with flexible lateral forceresisting systems, such as moment frames and bucklingrestrained braced frames, approaches used
to design nonbearing wall panels to accommodate story drift should be applied to veneers.
C13.5.5 OutofPlane Bending. The effects of outofplane application of seismic forces (defined in Section 13.3.1) on
nonstructural walls, including the resulting deformations, must be considered. Where weak or brittle materials are employed,
conventional deflection limits are expressed as a proportion of the span. The intent is to preclude outofplane failure of
heavy materials (such as brick or block) or applied finishes (such as stone or tile).
C13.5.6 Suspended Ceilings. Suspended ceiling systems are fabricated using a wide range of building materials with
differing characteristics. Some systems (such as lath and plaster or gypsum board, screwed or nailed to suspended members)
are fairly homogeneous and should be designed as lightframe diaphragm assemblies, using the forces of Section 13.3 and the
applicable materialspecific design provisions of Chapter 14. Others comprise discrete elements laid into a suspension
system and are the subject of this section.
Seismic performance of ceiling systems with layin or acoustical panels depends on support of the grid and individual panels
at walls and expansion joints, integrity of the grid/panel assembly, interaction with other systems (such as fire sprinklers),
and support for other nonstructural components (such as light fixtures and HVAC systems). Observed performance problems
include dislodgement of tiles due to impact with walls and water damage (sometimes leading to loss of occupancy) due to
interaction with fire sprinklers. Extensive shake table testing performed at the State University of New York at Buffalo
addresses seismic performance of suspended ceiling systems at various ground motion levels. That work is reported by Yao
(2000) and by Bidillo, et al. (2003, 2006, and 2007).
The performance of ceiling systems is affected by the placement of seismic bracing and the layout of light fixtures and other
supported loads. Dynamic testing has demonstrated that splayed wires, even with vertical compression struts, may not
adequately limit lateral motion of the ceiling system due to straightening of the end loops. Construction problems include
slack installation or omission of bracing wires due to obstructions. Other testing has shown that unbraced systems may
perform well where the system can accommodate the expected displacements, by providing both sufficient clearance at
penetrations and wide closure members which are now required by the standard.
C13.5.6.1 Seismic Forces. Where the weight of the ceiling system is distributed nonuniformly, that condition should be
considered in the design, since the typical Tbar ceiling grid has limited ability to redistribute lateral loads.
C13.5.6.2 Industry Standard Construction. Industry standard construction relies on ceiling contact with the perimeter wall
for restraint. The key to good seismic performance is sufficiently wide closure angles at the perimeter to accommodate
relative ceiling motion and adequate clearance at penetrating components (such as columns and piping) to avoid
concentrating restraining loads on the ceiling system.
C13.5.6.2.1 Seismic Design Category C. While there is no direct equivalency between Seismic Design Categories and
seismic zones, application of CISCA requirements for Seismic Zones 0 to 2 produces reasonable results for Seismic Design
Category C. ASTM E580 is currently being revised for consistency with the IBC and ASCE/SEI 705. When updated, it is
expected to replace the CISCA requirements.
C13.5.6.2.2 Seismic Design Categories D through F. Where certain prescriptive requirements are met, lateral restraints
may be omitted for small areas of suspended ceiling. The behavior of an unbraced ceiling system is similar to that of a
pendulum; therefore, the lateral displacement is a function of the level of ground motion and the square root of the
suspension length. The default displacement limit is based on anticipated damping and energy absorption of the suspended
ceiling system without significant impact with the perimeter wall.
The requirements set forth in this section of the standard for Seismic Design Categories D through F are in addition to the
CISCA requirements for Seismic Zones 3 and 4. Therefore, seismic requirements for ceilings are triggered where ceiling
areas exceed 256 square feet, and additional requirements apply where ceiling areas exceed 1,000 square feet and 2,500
square feet. The alternative to provide swing joint connections or flexible devices (such as hoses) for sprinkler drops is
included in the latest edition of NFPA 13.
C13.5.6.3 Integral Construction. Ceiling systems utilizing integral construction are constructed of modular preengineered
components, which integrate lights, ventilation components, firesprinklers, and seismic bracing into a complete system.
They may include aluminum, steel, and PVC components and may be designed using integral construction of ceiling and
wall. They often use rigid grid and bracing systems, which provide lateral support for all the ceiling components, including
sprinkler drops. This reduces the potential for adverse interactions between components, and eliminates the need to provide
clearances for differential movement.
C13.5.7 Access Floors
C13.5.7.1 General. In past earthquakes and in cyclic load tests, some typical raised access floor systems behaved in a brittle
manner and exhibited little reserve capacity beyond initial yielding or failure of critical connections. Testing shows that
unrestrained individual floor panels may pop out of the supporting grid unless mechanically fastened to supporting pedestals
or stringers. This may be a concern, particularly in egress pathways.
For systems with floor stringers, it is accepted practice to calculate the seismic force, Fp, for the entire access floor system
within a partitioned space and then distribute the total force to the individual braces or pedestals. For stringerless systems,
the seismic load path should be established explicitly.
Overturning effects subject individual pedestals to vertical loads well in excess of the weight, Wp, used in determining the
seismic force, Fp. It is unconservative to use the design vertical load simultaneously with the design seismic force for design
of anchor bolts, pedestal bending, and pedestal welds to base plates. “Slip on” heads that are not mechanically fastened to the
pedestal shaft and thus cannot transfer tension are likely unable to transfer to the pedestal the overturning moments generated
by equipment attached to adjacent floor panels.
To preclude brittle failure, each element in the seismic load path must have energy absorbing capacity. Buckling failure
modes should be prevented. Lower seismic force demands are allowed for special access floors that are designed to preclude
brittle and buckling failure modes.
C13.5.7.2 Special Access Floors. An access floor can be a “special access floor” if the registered design professional opts to
comply with the requirements of Section 13.5.7.2. Special access floors include construction features that improve the
performance and reliability of the floor system under seismic loading. The provisions focus on providing a reliable load path
for seismic shear and overturning forces. Special access floors are designed for smaller lateral forces, and their use is
encouraged at facilities with higher nonstructural performance objectives.
C13.5.8 Partitions. Partitions subject to these requirements must have independent lateral support bracing from the top of
the partition to the building structure or to a substructure attached to the building structure. Some partitions are designed to
span vertically from the floor to a suspended ceiling system. The ceiling system must be designed to provide lateral support
for the top of the partition. An exception to this condition is provided to exempt bracing of light (gypsum board) partitions
where the load does not exceed the minimum partition lateral load. Experience has shown that partitions subjected to the
minimum load can be braced to the ceiling without failure.
C13.5.9 Glass in Glazed Curtain Walls, Glazed Storefronts, and Glazed Partitions. The performance of glass in
earthquakes falls into one of four categories:
1. The glass remains unbroken in its frame or anchorage.
2. The glass cracks but remains in its frame or anchorage while continuing to provide a weather barrier, and to be otherwise
serviceable.
3. The glass shatters but remains in its frame or anchorage in a precarious condition, likely to fall out at any time.
4. The glass falls out of its frame or anchorage, either in shards or as whole panels.
Categories 1 and 2 satisfy both immediateoccupancy and lifesafety performance objectives. Although the glass is cracked
in Category 2, immediate replacement is not required. Categories 3 and 4 cannot provide for immediate occupancy, and their
provision of life safety depends on the postbreakage characteristics of the glass and the height from which it can fall.
Tempered glass shatters into multiple, pebblesize fragments that fall from the frame or anchorage in clusters. These broken
glass clusters are relatively harmless to humans when they fall from limited heights, but they could be harmful when they fall
from greater heights.
C13.5.9.1 General. Equation 13.51 is derived from Earthquake Safety Design of Windows, published in November 1982 by
the Sheet Glass Association of Japan and is similar to an equation in Bouwkamp and Meehan (1960) that permits calculation
of the story drift required to cause glasstoframe contact in a given rectangular window frame. Both calculations are based
on the principle that a rectangular window frame (specifically, one that is anchored mechanically to adjacent stories of a
structure) becomes a parallelogram as a result of story drift, and that glasstoframe contact occurs when the length of the
shorter diagonal of the parallelogram is equal to the diagonal of the glass panel itself. The value .fallout represents the
displacement capacity of the system and Dp represents the displacement demand.
The 1.25 factor in the requirements described above reflect uncertainties associated with calculated inelastic seismic
displacements of building structures. Wright (1989) states that “postelastic deformations, calculated using the structural
analysis process, may well underestimate the actual building deformation by up to 30 percent. It would therefore be
reasonable to require the curtain wall glazing system to withstand 1.25 times the computed maximum interstory displacement
to verify adequate performance.”
The reason for Exception 2 to Equation 13.51 is that the tempered glass, if shattered, would not produce an overhead falling
hazard to adjacent pedestrians, although some pieces of glass may fall out of the frame.
C13.5.9.2 Seismic Drift Limits for Glass Components. As an alternative to the prescriptive approach of Section 13.5.9.1,
the deformation capacity of glazed curtain wall systems may be established by test.
C13.6 MECHANIAL AND ELECTRICAL COMPONENTS
These requirements, focused on design of supports and attachments, are intended to reduce the hazard to life posed by loss of
component structural stability or integrity. The requirements increase the reliability of component operation but do not
address functionality directly. For critical components where operability is vital, Section 13.2.2 provides methods for
seismically qualifying the component.
Traditionally, mechanical equipment without rotating or reciprocating components (such as tanks and heat exchangers) is
anchored directly to the structure. Mechanical and electrical equipment with rotating or reciprocating components often is
isolated from the structure by vibration isolators (such as rubberinshear, springs, or air cushions). Heavy mechanical
equipment (such as large boilers) may not be restrained at all, and electrical equipment other than generators, which are
normally isolated to dampen vibrations, usually is rigidly anchored (for example, switchgear and motor control centers).
Two distinct levels of earthquake safety are considered in the design of mechanical and electrical components. At the usual
safety level, failure of the mechanical or electrical component itself due to seismic effects poses no significant hazard. In this
case, design of the supports and attachments to the structure is required to avoid a lifesafety hazard. At the higher safety
level, the component must continue to function acceptably following the design earthquake. Such components are defined as
designated seismic systems in Section 11.2 and may be required to meet the special certification requirements of Section
13.2.2.
Not all equipment or parts of equipment need to be designed for seismic forces. Where Ip is specified to be 1.0, damage to,
or even failure of, a piece or part of a component does not violate these requirements as long as a lifesafety hazard is not
created. The restraint or containment of a falling, breaking, or toppling component (or its parts) by means of bumpers,
braces, guys, wedges, shims, tethers, or gapped restraints to satisfy these requirements often is acceptable, although the
component itself may suffer damage.
Judgment is required to fulfill the intent of these requirements; the key consideration is the threat to life safety. For example,
a nonessential air handler package unit that is less than 4 feet (1.2 meters) tall bolted to a mechanical room floor is not a
threat to life as long as it is prevented from significant displacement by having adequate anchorage. In this case, seismic
design of the air handler itself is unnecessary. On the other hand, a 10foot (3.0 meters) tall tank on 6foot (1.8 meters) long
angles used as legs, mounted on a roof near a building exit does pose a hazard. The intent of these requirements is that the
supports and attachments (tank legs, connections between the roof and the legs, and connections between the legs and the
tank), and possibly even the tank itself be designed to resist seismic forces. Alternatively, restraint of the tank by guys or
bracing could be acceptable.
It is not the intent of the standard to require the seismic design of shafts, buckets, cranks, pistons, plungers, impellers, rotors,
stators, bearings, switches, gears, nonpressure retaining casings and castings, or similar items. Where the potential for a
hazard to life exists, it is expected that design effort will focus on equipment supports including base plates, anchorages,
support lugs, legs, feet, saddles, skirts, hangers, braces, or ties.
Many mechanical and electrical components consist of complex assemblies of parts that are manufactured in an industrial
process that produces similar or identical items. Such equipment may include manufacturer's catalog items and often are
designed by empirical (trialanderror) means for functional and transportation loadings. A characteristic of such equipment
is that it may be inherently rugged. The term “rugged” refers to an ampleness of construction that provides such equipment
with the ability to survive strong motions without significant loss of function. By examining such equipment, an experienced
design professional usually should be able to confirm such ruggedness. The results of an assessment of equipment
ruggedness may be used in determining an appropriate method and extent of seismic design or qualification effort.
C13.6.1 General. The exception allowing unbraced suspended components has been clarified, addressing concerns about the
type of nonstructural components allowed by these exceptions as well as the acceptable consequences of interaction between
components. In previous editions of the standard, certain nonstructural components that could represent a fire hazard
following an earthquake were exempt from lateral bracing requirements. In the revised exception, reference to Section 13.2.3
addresses such concerns while distinguishing between credible seismic interactions and incidental interactions.
The seismic demand requirements are based on component structural attributes of flexibility (or rigidity) and ruggedness.
Table 13.61 provides seismic coefficients based on judgments of the component flexibility, expressed in the ap term, and
ruggedness expressed in the Rp term. It may also be necessary to consider the flexibility and ductility of the attachment
system that provides seismic restraint.
Entries for components and systems in Table 13.61 are grouped and described to improve clarity of application.
Components are divided into three broad groups, within which they are further classified depending on the type of
construction or expected seismic behavior. For example, mechanical components include “airside” components (such as fans
and air handlers) that experience dynamic amplification but are light and deformable; “wetside” components that generally
contain liquids (such as boilers and chillers) that are more rigid and somewhat ductile; and very rugged components (such as
engines, turbines, and pumps) that are of massive construction due to demanding operating loads, and generally perform well
in earthquakes, if adequately anchored.
A distinction is made between components isolated using neoprene and those that are spring isolated. Spring isolated are
assigned a lower Rp value since they tend to have less effective damping. Internally isolated components are classified
explicitly to avoid confusion.
C13.6.2 Component Period. Component period is used to clarify components as rigid (T = 0.06s) or flexible (T > 0.06s).
Determination of the fundamental period of a mechanical or electrical component using analytical or test methods can
become very involved. If not properly performed, the fundamental period may be underestimated, producing unconservative
results. The flexibility of the component’s supports and attachments typically dominates response and thus fundamental
component period. Therefore, analytical determinations of component period must consider those sources of flexibility.
Where determined by testing, the dominant mode of vibration of concern for seismic evaluation must be excited and captured
by the test setup. This dominant mode of vibration cannot be discovered through insitu tests that measure only ambient
vibrations. To excite the mode of vibration with the highest fundamental period by insitu tests, relatively significant input
levels of motion are required (that is, the flexibility of the base and attachment must be exercised). A resonant frequency
search procedure, such as that given in ICCES AC156, may be used to identify the dominant modes of vibration of a
component.
Many types of mechanical components have fundamental periods below 0.06 seconds and may be considered to be rigid.
Examples include horizontal pumps, engine generators, motor generators, air compressors, and motor driven centrifugal
blowers. Other types of mechanical equipment are very stiff, but may have fundamental periods up to about 0.125 seconds.
Examples include vertical immersion and deep well pumps, belt driven and vane axial fans, heaters, air handlers, chillers,
boilers, heat exchangers, filters, and evaporators. These fundamental period estimates do not apply where the equipment is
mounted on vibration isolators.
Electrical equipment cabinets can have fundamental periods of about 0.06 to 0.3 seconds, depending upon the supported
weight and its distribution, the stiffness of the enclosure assembly, the flexibility of the enclosure base, and the load path
through to the attachment points. Tall, narrow motor control centers and switchboards lie at the upper end of this period
range. Low and mediumvoltage switchgear, transformers, battery chargers, inverters, instrumentation cabinets, and
instrumentation racks usually have fundamental periods ranging from 0.1 to 0.2 seconds. Braced battery racks, stiffened
vertical control panels, benchboards, electrical cabinets with top bracing, and wallmounted panelboards have fundamental
periods ranging from 0.06 to 0.1 seconds.
C13.6.3 Mechanical Components and C13.6.4 Electrical Components. Most mechanical and electrical equipment is
inherently rugged and, where properly attached to the structure, has performed well in past earthquakes. Since the
operational and transportation loads for which the equipment is designed typically are larger than those due to earthquakes,
these requirements focus primarily on equipment anchorage and attachments. However, Designated Seismic Systems, which
are required to function following an earthquake or which must maintain containment of flammable or hazardous materials,
must themselves be designed for seismic forces or be qualified for seismic loading in accordance with Section 13.2.2.
The likelihood of postearthquake operability can be increased where the following measures are taken:
1. Internal assemblies, subassemblies, and electrical contacts are attached sufficiently to prevent their being subjected to
differential movement or impact with other internal assemblies or the equipment enclosure.
2. Operators, motors, generators, and other such components that are functionally attached to mechanical equipment by
means of an operating shaft or mechanism are structurally connected or commonly supported with sufficient rigidity
such that binding of the operating shaft will be avoided.
3. Any ceramic or other nonductile components in the seismic load path are specifically evaluated.
4. Adjacent electrical cabinets are bolted together and cabinet lineups are prevented from impacting adjacent structural
members.
Components that could be damaged, or could damage other components, and are fastened to multiple locations of a structure
must be designed to accommodate seismic relative displacements. Such components include bus ducts, cable trays, conduit,
elevator guide rails, and piping systems. As discussed in Section C13.3.2.1, special design consideration is required where
full story drift demands are concentrated in a fraction of the story height.
C13.6.5 Component Supports. The intent of this section is to require seismic design of all mechanical and electrical
component supports to prevent sliding, falling, toppling, or other movement that could imperil life. Component supports are
differentiated here from component attachments to emphasize that the supports themselves, as enumerated in the text, require
seismic design even if fabricated by the mechanical or electrical component manufacturer. This is regardless of whether the
mechanical or electrical component itself is designed for seismic loads.
C13.6.5.1 Design Basis. Standard supports are those developed in accordance with a reference document (Section 13.1.6).
Where standard supports are not used, the seismic design forces and displacement demands of Chapter 13 are used with
applicable materialspecific design procedures of Chapter 14.
C13.6.5.2 Design for Relative Displacement. For some items, such as piping, seismic relative displacements between
support points are of more significance than inertial forces. Components made of high deformability materials such as steel
or copper can accommodate relative displacements inelastically, provided the connections also provide high deformability.
Threaded and soldered connections exhibit poor ductility under inelastic displacements, even for ductile materials.
Components made of less ductile materials can accommodate relative displacement effects only if appropriate flexibility or
flexible connections are provided.
Detailing distribution systems that connect separate structures with bends and elbows makes them less prone to damage and
less likely to fracture and fall, provided the supports can accommodate the imposed loads.
C13.6.5.3 Support Attachment to Component. As used in this Section, “integral” relates to the manufacturing process, not
the location of installation. For example, both the legs of a cooling tower and the attachment of the legs to the body of the
cooling tower must be designed, even if the legs are provided by the manufacturer and installed at the plant. Also, if the
cooling tower has an Ip=1.5, the design must address not only the attachments (welds, bolts, etc.) of the legs to the
component but also local stresses imposed on the body of the cooling tower by the support attachments.
C13.6.5.5 Additional Requirements. As reflected in this Section of the standard and in the footnote to Table 13.61,
vibration isolated equipment with snubbers is subject to amplified loads as a result of dynamic impact.
Use of expansion anchors for nonvibration isolated mechanical equipment rated over 10 hp is prohibited based on
experience with older anchor types. The ASCE/SEI 7 Seismic Subcommittee developing the 2010 edition of the standard is
considering a proposal that would allow anchors qualified by simulated seismic testing and longterm vibration testing to also
be exempt.
C13.6.6 Utility and Service Lines. For essential facilities (Occupancy Category IV), auxiliary onsite mechanical and
electrical utility sources are recommended.
Where utility lines pass through the interface of adjacent, independent structures, they must be detailed to accommodate
differential displacement computed in accordance with Section 13.3.2 and including the Cd factor of Section 12.2.1.
As specified in Section 13.1.3, nonessential piping whose failure could damage essential utilities in the event of pipe rupture
are deemed Designated Seismic Systems.
C13.6.7 HVAC Ductwork. Experience in past earthquakes has shown that HVAC duct systems are rugged and perform
well in strong ground shaking. Bracing in accordance with the Sheet Metal and Air Conditioning Contractors National
Association ANSI/SMACNA 001 has been effective in limiting damage to duct systems. Typical failures have affected only
system function, and major damage or collapse has been uncommon. Therefore, industry standard practices should prove
adequate for most installations. Expected earthquake damage is limited to opening of duct joints and tears in ducts.
Connection details that are prone to brittle failures, especially hanger rods subject to large amplitude cycles of bending stress,
should be avoided.
The amplification factor for ductwork has been increased from 1.0 to 2.5, because even braced duct systems are relatively
flexible. The Rp values also have been increased so that the resulting seismic design forces are consistent with those
determined previously.
Ductwork systems that carry hazardous materials or must remain operational during and after an earthquake, are assigned a
value of Ip =1.5, and require a detailed engineering analysis addressing leaktightness.
C13.6.8 Piping Systems. In earthquakes, piping systems rarely collapse but often cause nonstructural damage due to
leaking. Industry standards and guidelines address a wide variety of piping systems and materials. Construction in
accordance with referenced national standards is effective in limiting damage to and avoiding loss of fluid containment in
piping systems under earthquake conditions.
ASHRAE’s A Practical Guide to Seismic Restraint, while not an ANSI standard, is in common use and may be an
appropriate reference document for use in the seismic design of piping systems.
The prescriptive conditions provided in the standard under which seismic bracing for piping may be omitted are based on
observed performance in past earthquakes.
C13.6.8.1 ASME Pressure Piping Systems. The Rp values tabulated for ASME B31 compliant piping systems reflect the
stringent design and quality control requirements as well as the intensified stresses used in ASME design procedures.
C13.6.8.4 Other Piping Systems
Piping not designed in accordance with ASME B31 typically is assigned lower Rp values. Piping component testing suggests
that the ductility capacity of carbon steel threaded and grooved joint piping component joints ranges between 1.4 and 3.0.
Therefore, these types of connections have been classified as having limited deformability. Grooved couplings and other
articulating type of connections may demonstrate free rotational capacity that increases the overall rotational design capacity
of the connection. When considered in design, this increase should not exceed 50 percent of the total demonstrated design
capacity. The free rotational capacity is the maximum articulating angle where the connection behaves essentially as a
pinned joint. The remaining rotational capacity of the connection is where it behaves as a conventional joint whose design
force demands are determined by traditional means.
C13.6.9 Boilers and Pressure Vessels
Experience in past earthquakes has shown that boilers and pressure vessels are rugged and perform well in strong ground
motion. Construction in accordance with current requirements of the ASME Boiler and Pressure Vessel Code (ASME
BPVC) has been shown to be effective in limiting damage to and avoiding loss of fluid containment in boilers and pressure
vessels under earthquake conditions. It is, therefore, the intent of the standard that nationally recognized codes be used to
design boilers and pressure vessels provided that the seismic force and displacement demands are equal to or exceed those
outlined in Section 13.3. Where nationally recognized codes do not yet incorporate force and displacement requirements
comparable to the requirements of Section 13.3, it is nonetheless the intent to use the design acceptance criteria and
construction practices of those codes.
C13.6.10 Elevator and Escalator Design Requirements
The ASME Safety Code for Elevators and Escalators (ASME A17.1) has adopted many requirements to improve the seismic
response of elevators; however, they do not apply to some regions covered by this chapter. These changes are to extend force
requirements for elevators to be consistent with the standard.
C13.6.10.3 Seismic Switches
The purpose of seismic switches as used here is different from that of ASME A17.1, which has incorporated several
requirements to improve the seismic response of elevators (such as rope snag point guards, rope retainer guards, and guide
rail brackets) and which does not apply to some buildings covered by the standard. Building motions that are expected in
areas not covered by the seismic provisions of ASME 17.1 are sufficiently large to impair the operation of elevators. The
seismic switch is positioned high in the structure where structural response will be the most severe. The seismic switch
trigger level is set to shut down the elevator where structural motions are expected to impair elevator operations.
Elevators in which the seismic switch and counterweight derail device have triggered should not be put back into service
without a complete inspection. However, in the case where the loss of use of the elevator creates a lifesafety hazard, an
attempt to put the elevator back into service may be attempted. Operating the elevator prior to inspection may cause severe
damage to the elevator or its components.
The building owner should have detailed written procedures in place defining for the elevator operator/maintenance
personnel which elevators in the facility are necessary from a postearthquake life safety perspective. It is highly
recommended that these procedures be inplace, with appropriate personnel training, prior to an event strong enough to trip
the seismic switch.
C13.6.10.4 Retainer Plates
The use of retainer plates is a very low cost provision to improve the seismic response of elevators.
C13.6.11 Other Mechanical and Electrical Components. The material properties set forth in Item 2 of this Section are
similar to those allowed in ASME BPVC and reflect the high factors of safety necessary for seismic, service, and
environmental loads.
REFERENCES
ANCO Engineers, Inc. 1983. Seismic Hazard Assessment of NonStructural Components —Phase I, Final Report for
National Science Foundation from ANCO Engineers, Inc., Culver City, California, September.
American National Standards Institute/ Sheet Metal and Air Conditioning Contractors’ National Association. 2000. Seismic
Restraint Manual: Guidelines for Mechanical Systems, ANSI/SMACNA 0012000.
American Society of Heating, Refrigeration, and Air Conditioning Engineers. 2000. A Practical Guide to Seismic Restraint.
ASTM International. 2006. Standard Practice for Application of Ceiling Suspension Systems for Acoustical Tile and Layin
Panels in Areas Requiring Seismic Restraint, ASTM E580
Bachman, R. E., and S. M. Dowty. 2008. “Nonstructural Component or Nonbuilding Structure?” Building Safety Journal
(AprilMay).
Bachman, R. E, and R. M. Drake. 1996. “A Study To Empirically Validate the Component Response Modification Factors
in the 1994 NEHRP Provisions Design Force Equations for Architectural, Mechanical, and Electrical Components,” letter
report to the National Center for Earthquake Engineering Research, July.
Bachman, R. E., R. M. Drake, and P. J. Richter. 1993. 1994 Update to 1991 NEHRP Provisions for Architectural,
Mechanical, and Electrical Components and Systems, letter report to the National Center for Earthquake Engineering
Research, February 22.
Badillo, H. 2003. Seismic Fragility Testing of Suspended Ceiling Systems, Masters Thesis, Department of Civil, Structural,
and Environmental Engineering, State University of New York at Buffalo, Buffalo New York, September.
Badillo, H., A. S. Whittaker, A. M. Reinhorn, and G. P. Cimmarello. 2006. Seismic Fragility of Suspended Ceiling Systems,
MCEER Report060001. Multidisciplinary Center for Earthquake Engineering Research, State University of New York at
Buffalo.
Badillo, H., A. S. Whittaker, and A. M. Reinhorn. 2007. “Seismic Fragility of Suspended Ceiling Systems,” Earthquake
Spectra, 23(1).
Behr, R. A., A. Belarbi, and A. T. Brown. 1995. “Seismic Performance of Architectural Glass in a Storefront Wall System,”
Earthquake Spectra, 11(3):367391.
Behr, R. A., and A. Belarbi. 1996. “Seismic Tests Methods for Architectural Glazing Systems,” Earthquake Spectra,
12(1):129143.
Bouwkamp, J. G., and J. F. Meehan. 1960. “Drift Limitations Imposed by Glass,” Proceedings of the Second World
Conference on Earthquake Engineering, Tokyo, Japan, 17631778.
Ceiling and Interior Systems Construction Association. 2004. Guidelines for Seismic Restraint for Directhung Suspended
Ceiling Assemblies (zones 34).
Ceiling and Interior Systems Construction Association. 2004. Recommendations for Directhung Acoustical Tile and Layin
Panel Ceilings (zones 02).
Drake, R. M., and R. E. Bachman. 1996. “NEHRP Provisions for 1994 for Nonstructural Components,” ASCE Journal of
Architectural Engineering, March.
Drake, R. M., and R. E. Bachman. 1995. “Interpretation of Instrumented Building Seismic Data and Implications for
Building Codes,” in Proceedings of the 1995 SEAOC Annual Convention.
Drake, R. M., and R. E. Bachman. 1994. “1994 NEHRP Provisions for Architectural, Mechanical, and Electrical
Components,” in Proceedings of the 5th United States National Conference on Earthquake Engineering.
Earthquake Engineering Research Institute. 1994. “Northridge Earthquake, January 17, 1994: Preliminary Reconnaissance
Report,” edited by John F. Hall, pp. 5657. EERI, Oakland, California.
Gates, W. E. and G. McGavin. 1998. “Lessons Learned from the 1994 Northridge Earthquake on the Vulnerability of
Nonstructural Systems,” in Proceedings, Seminar on Seismic Design, Retrofit, and Performance of Nonstructural
Components, ATC291, January 2223, 1998, San Francisco, California. Applied Technology Council, Redwood City,
California, pp. 93101.
Housner, G. W., and M. A. Haroun. 1980. “Seismic Design of Liquid Storage Tanks” in ASCE Convention Proceedings.
Institute of Electrical and Electronics Engineers, Inc. 2005. IEEE Recommended Practices for Seismic Design of
Substations, IEEE 6932005.
International Code Council Evaluation Service. 2007. Seismic Qualification by Shaketable Testing of Nonstructural
Components and Systems, ICCES AC156. International Code Council Evaluation Service, Whittier, California.
Kehoe, B., and M. Hachem. 2003. "Procedures for Estimating Floor Accelerations" by ATC 292, 2003
Pantelides, C. P., K. Z. Truman, R. A. Behr, and A. Belarbi. 1996. “Development of a Loading History for Seismic Testing
of Architectural Glass in a ShopFront Wall System,” Engineering Structures, 18(12):917935.
Unified Facilities Criteria, Department of Defense. 2007. Seismic Design for Buildings, United States Department of
Defense, UFC 331004.
Wright, P. D. 1989. The Development of a Procedure and Rig for Testing the Racking Resistance of Curtain Wall Glazing,
BRANZ Studey Report 17. Building Research Association of New Zealand (BRANZ).
Yao, G. C. 2000. “Seismic Performance of Direct Hung Suspended Ceiling Systems,” Journal of Architectural Engineering,
6(1):611.
Page intentionally left blank.
COMMENTARY TO CHAPTER 14,
MATERIAL SPECIFIC SEISMIC DESIGN AND DETAILING
REQUIREMENTS
Because seismic loading is expected to cause nonlinear behavior in structures, seismic design criteria require not only
provisions to govern loading, but also provisions to define the required configurations, connections, and detailing to produce
material and system behavior consistent with the design assumptions. Thus, while ASCE/SEI 705 is primarily a loading
standard, compliance with Chapter 14, which covers material specific seismic design and detailing, is required. In general,
Chapter 14 adopts material design and detailing standards developed by industry material standards organizations. These
materials standards organizations maintain complete commentaries covering their standards and such material is not
duplicated here.
The refinements, additions, and recommended changes to the material standards produced by the Provisions Update
Committee appear in Part 1 of the 2009 NEHRP Recommended Seismic Provisions as exceptions to ASCE/SEI 705 along
with associated commentary.
C14.0 SCOPE
The scoping statement in this section clarifies that foundation elements are subject to all of the structural design requirements
of the standard.
C14.1 STEEL
C14.1.1 Reference Documents. This section lists a series of structural standards published by the American Institute of
Steel Construction (AISC), American Iron and Steel Institute (AISI), American Society of Civil Engineers (ASCE/SEI), and
Steel Joist Institute (SJI) that are to be applied in the seismic design of steel members and connections in conjunction with the
requirements of ASCE/SEI 7. The AISC references are available free of charge in electronic format at www.aisc.org.
C14.1.2 Seismic Design Categories B and C. For the lower Seismic Design Categories B and C, the engineer is allowed a
choice in the design of a steel lateral force resisting system. The first option is to design the structure to meet the design and
detailing requirements for structures assigned to higher Seismic Design Categories, with the corresponding seismic design
parameters (R, O0, and Cd ). The second option is to use a lower R factor of 3 (and higher resulting base shear), an O0 of 3,
and a Cd value of 3 but without specific seismic design and detailing requirements. The concept of this option is that design
for a higher base shear force will result in essentially elastic response that will compensate for the limited ductility of the
members and connections, resulting in performance similar to that of more ductile systems.
C14.1.3 Seismic Design Categories D through F. For the higher Seismic Design Categories, the Engineer is not given a
choice, but must follow the seismic design provisions of either AISC or AISI using the seismic design parameters specified
for the chosen structural system. It is not considered appropriate to design structures without specific design and detailing for
seismic response in these high Seismic Design Categories.
C14.1.4 ColdFormed Steel. This section adopts two standards by direct reference: AISI NAS, North American
Specification for the Design of ColdFormed Steel Structural Members, and ASCE/SEI 8, Specification for the Design of
Cold Formed Stainless Steel Structural Members.
Both of the adopted reference documents have specific limits of applicability. AISI NAS (Section A1.1) applies to the design
of structural members that are coldformed to shape from carbon or lowalloy steel sheet, strip, plate, or bar not more than
oneinch in thickness. ASCE/SEI 8 (Section 1.1.1) governs the design of structural members that are coldformed to shape
from annealed and coldrolled sheet, strip, plate, or flat bar stainless steels. Both documents focus on loadcarrying members
in buildings; however, allowances are made for applications in nonbuilding structures, if dynamic effects are considered
appropriately.
Within each document, there are requirements related to general provisions for the applicable types of steel; design of
elements, members, structural assemblies, connections, and joints; and mandatory testing. In addition, AISI NAS contains a
chapter on the design of coldformed steel structural members and connections undergoing cyclic loading. Both standards
contain extensive commentaries for the benefit of the user.
C14.1.4.1 LightFramed ColdFormed Steel Construction. This subsection of coldformed steel relates to lightframed
construction, which is defined as a method of construction where the structural assemblies are formed primarily by a system
of repetitive wood or coldformed steel framing members or subassemblies of these members (ASCE/SEI 705 Section 11.2).
Not only does this subsection repeat the direct adoptions of AISI NAS and ASCE/SEI 8, but it also allows the user to choose
from an additional suite of standards that address different aspects of construction, including the following:
1. AISI GP, Standard for ColdFormed Steel Framing— General Provisions, applies to the design, construction, and
installation of structural and nonstructural coldformed steel framing members where the specified minimum base metal
thickness is between 18 mils and 118 mils (Section A1).
2. AISI WSD, Standard for ColdFormed Steel Framing – Wall Stud Design, applies to the design and installation of coldformed
steel studs for both structural and nonstructural walls in buildings (Section A1).
3. AISI Lateral, Standard for ColdFormed Steel Framing – Lateral Design, contains design requirements for shear walls,
diagonal strap bracing (as part of a structural wall), and diaphragms (Section A1).
The requirements of AISI GP apply to all lightframed coldformed steel and, consequently, the standard is adopted by direct
reference in both AISI WSD and AISI Lateral. In addition, all of these documents include commentaries to aid the user in
the correct application of their requirements.
C14.1.5 Prescriptive Framing. This section adopts AISI PM, Standard for ColdFormed Steel Framing – Prescriptive
Method for One and Two Family Dwellings, which applies to the construction of detached oneand twofamily dwellings,
townhouses, and other attached singlefamily dwellings not more than two stories in height using repetitive inline framing
practices (Section A1). This document adopts AISI GP by direct reference and includes a commentary to aid the user in the
correct application of its requirements.
C14.1.6 Steel Deck Diaphragms. Design of steel deck diaphragms is to be based upon recognized national standards or a
specific testing program directed by a person experienced in testing procedures and steel deck. All fastener design values
(welds, screws, power actuated fasteners, button punches) for attaching steel deck sheet to steel deck sheet or for attaching
the steel deck to the building framing members must be per recognized national design standards or specific steel deck testing
programs. All steel deck diaphragm and fastener design properties must be approved for use by the authorities in whose
jurisdiction the construction project occurs. Steel deck diaphragm inplane design forces (seismic, wind, or gravity) must be
determined per ASCE/SEI 705 Section 12.10.1. Steel deck manufacturer test reports prepared in accordance with this
provision can be used where adopted and approved by the authority having jurisdiction for the building project. The
diaphragm design manual produced by the Steel Deck Institute (2004) is also a potential reference for design values.
Steel deck is assumed to have a corrugated profile consisting of alternating up and down flutes that are manufactured
in various widths and heights. Use of flat sheet metal as the overall floor or roof diaphragm is permissible where designed by
engineering principles, but is beyond the scope of this section. Flat or bent sheet metal may be used as closure pieces for
small gaps or penetrations or for shear transfer over short distances in the steel deck diaphragm where diaphragm design
forces are considered.
Steel deck diaphragm analysis must include design of chord members at the perimeter of the diaphragm and around interior
openings in the diaphragm. Chord members may be steel beams attached to the underside of the steel deck designed for a
combination of axial loads and bending moments due to acting gravity and lateral loads.
Where diaphragm design loads exceed the bare steel deck diaphragm design capacity, then either horizontal steel trusses or a
structurally designed concrete topping slab placed over the steel deck must be provided to distribute lateral forces. Where
horizontal steel trusses are used, the steel deck must be designed to transfer diaphragm forces to the steel trusses. Where a
structural concrete topping over the steel deck is used as the diaphragm, the diaphragm chord members at the perimeter of the
diaphragm and edges of interior openings must be either: (a) designed flexural reinforcing steel placed in the structural
concrete topping or (b) steel beams located under the steel deck with connectors (that provide a positive connection) as
required to transfer design shear forces between the concrete topping and steel beams.
C14.1.7 Steel Cables. These provisions reference ASCE/SEI 1996, Structural Applications of Steel Cables for Buildings,
for the determination of the design strength of steel cables. ASCE/SEI 19 uses service level load combinations with a safety
factor relative to the cable design strength. The service level load combinations specified in ASCE/SEI 19 are adjusted in
two ways. First, the prestress loading is multiplied by a factor of 1.1 to account for any over prestressing that may occur in
the field. Second, the safety factor for load combinations including seismic effects is reduced from 2.0 to 1.5 to account for
the dynamic nature of seismic loading and the ductility of the system. While T3 and T4 in ASCE/SEI 19 may be calculated
using either wind or seismic loads, the modifications of this section apply only to load combinations including seismic
loadings.
C14.1.8 Additional Detailing Requirements for Steel Piles in Seismic Design Categories D through F. Steel piles used
in higher Seismic Design Categories are expected to yield just under the pile cap or foundation due to combined bending and
axial load. Design and detailing requirements of AISC 341 for Hpiles are intended to produce stable plastic hinge formation
in the piles. Since piles can be subjected to tension due to overturning moment, mechanical means to transfer such tension
must be designed for the required tension force, but not less than 10 percent of the pile compression capacity.
C14.2 CONCRETE
The section adopts ACI 31805, Building Code Requirements for Structural Concrete (ACI 318), by reference for structural
concrete design and construction. In addition, modifications to ACI 318 are made to coordinate the provisions of that
material design standard with the provisions of ASCE/SEI 7.
C14.2.2.1 ACI 318 Section 7.10. The reinforcement details for ties in compression members prescribed in ACI 318 Section
7.10.5 are appropriate for SDC A and B structures. This modification prescribes additional details for ties around anchor
bolts of structures assigned to SDC C, D, E, or F.
C14.2.2.2 ACI 318 Section 10.5. This provision affects ordinary moment frames. It is intended to improve continuity, and
thereby lateral force resistance and structural integrity, compared to that of frames designed to the provisions of Chapters1
through 18 of ACI 318 only. The provision does not apply to slabcolumn moment frames.
C14.2.2.3 ACI 318 Section 11.11. This requirement is intended to provide additional toughness to resist shear for columns
of frames in SDC B. Otherwise the proportions of those columns make them more susceptible to shear failure under
earthquake loading.
C14.2.2.4 Definitions. The first four definitions relate the wall types of ASCE/SEI 705 with detailing requirements of ACI
318 and distinguish between ordinary reinforced concrete structural walls and ordinary precast structural walls. These
definitions are essential to the proper interpretation of the R and Cd factors for each wall type specified in Table 12.21.
A wall pier is recognized as a separate category of structural element in this document but not in ACI 318.
C14.2.2.5 Scope. ACI 318 uses the terminology of low, moderate, and high seismic risk for structures assigned to SDC A
and B, SDC C, and SDC D through F, respectively. The modifications of this provision show how the ACI 318 provisions
should be interpreted for consistency with the ASCE/SEI 705 provisions.
C14.2.2.6 Reinforcement in Members Resisting EarthquakeInduced Forces. ACI 318 does not allow the use of
prestressing tendons in special and intermediate moment frames. This provision and ASCE/SEI 705 Sections 14.2.2.7 and
14.2.2.8 impose conditions that have been demonstrated to permit the safe use of such tendons.
These provisions are intended to apply to frames containing unbonded tendons only. The average prestress in plastic hinge
regions is restricted to limit the strain in the prestressing steel under the design displacement to not greater than 1 percent.
The strain in the prestressing steel at the design displacement should be calculated considering the anticipated inelastic
mechanism of the structure.
C14.2.2.7 Anchorages for Unbonded Posttensioning Tendons. Fatigue testing for 50 cycles of loading between 40 and
80 percent of the specified tensile strength of the prestressing strand has been an industry practice of long standing (ACI
423.6, Specification for Unbonded SingleStrand Tendons). The 80 percent limit is increased to 85 percent for seismic
applications in order to correspond to a 1 percent limit, and therefore the effective start of yielding, in the prestressing steel.
Testing over this range of stress conservatively simulates the effect of a severe earthquake on structures prestressed in
accordance with the requirements of ASCE/SEI 705 Sections 14.2.2.6 and 14.2.2.8.
C14.2.2.8 Flexural Members of Special Moment Frames. The restrictions on the flexural strength provided by the
tendons are based on the results of analytical and experimental studies (Ishizuka and Hawkins, 1987; Park and Thompson,
1977). Although satisfactory seismic performance can be obtained with greater amounts of prestressing steel, this restriction
is needed to allow the use of the same response modification and deflection amplification factors as those specified for
special moment frames without prestressing steel.
C14.2.2.9 Wall Piers and Wall Segments. Wall piers are typically segments between openings in walls that are thin in the
direction normal to the face of the wall. In current practice these elements are often not regarded as columns or as part of the
special structural walls. If not properly reinforced these elements are vulnerable to shear failure, and that failure prevents the
wall from developing the assumed flexural hinging. ACI 318 Section 21.7.10 is written specifically to preclude such preemptive
shear failure. The required shear strength in ACI 318 Section 21.4.5.1 is based on the probable shear strength, Ve ,
under the probable moment, M pr. Wall segments with a horizontal lengthtothickness ratio less than 2.5 and a clear heightto
length ratio of at least 2 are required to be designed as columns in compliance with ACI 318 Section 21.4 if they are used
as part of the lateralforceresisting system even though the shortest crosssectional dimension may be less than 12 inches in
violation of Section 21.4.1.1. Such wall segments may be designed to comply with ACI 318 Section 21.11 if they are not
used as part of the lateralforceresisting system. Wall segments with a horizontal lengthtothickness ratio larger than or
equal to 2.5, which do not meet the definition of wall piers (ASCE/SEI 705 Section 14.2.2.4), must be designed as special
structural walls or as portions of special structural walls in full compliance with ACI 318 Section 21.7.
C14.2.2.12 Members Not Designated as Part of the LateralForceResisting System. ACI 318 Section 21.4.3.2 permits
lap splices only within the center half of the column. Section 21.11.2 applies where the magnitude of the moments induced
in the column by the design displacement are explicitly checked. Section 21.11.3 applies where the effects of the design
displacement are not explicitly checked. Section 21.11.2.2, if not modified, would permit lap splices to be placed at any
location over the height of the column if the column is expected to yield. If, however, the column is not expected to yield the
wording effectively requires the splice to be located near midheight. This is not rational and the modification results in a
more rational provision.
C14.2.2.13 Columns Supporting Reactions from Discontinuous Stiff Members. Discontinuous shear walls and other
stiff members can impose large axial forces on supporting columns. The specified transverse reinforcement is to improve
column toughness under anticipated seismic demands.
C14.2.2.14 Intermediate Precast Structural Walls. ACI 318 Section 21.13 imposes requirements on precast walls for
moderate seismic risk applications. The intent is to produce ductile behavior by yielding of the steel elements or
reinforcement between panels or between panels and foundations. The 2003 IBC restricted yielding to steel reinforcement
because of concern that steel elements in the body of a connection could fracture due to strain demands.
Several steel element connections have been tested under simulated seismic loading and the adequacy of their loaddeformation
characteristics and strain capacity of yield has been demonstrated (Schultz and Magana, 1996). One such
connection was used in the fivestory building test that was part of the PRESSS Phase 3 research. The connection was used
to provide damping and energy dissipation, and demonstrated a very large strain capacity (Nakaki et al., 2001). Since then
several other steel element connections have been developed that can achieve similar results (Banks and Stanton, 2005;
Nakaki et al., 2005). In view of these results it is appropriate to allow yielding in steel elements that have been shown
experimentally to have adequate strain capacity to maintain at least 80 percent of their yield force of through the full design
displacement of the structure. This provision requires the designer to determine the deformation in the connection
corresponding to the earthquake design displacement, and then to check for experimental data that the connection type used
can accommodate that deformation without significant strength degradation.
The wall pier requirements in the modified ACI 318 Section 21.13.5 are less stringent than those for wall piers for special
structural walls as specified in the modified Section 21.7.10. Where intermediate precast structural walls are used in SDCs
D, E and F, wall piers should satisfy the requirements of ASCE/SEI 705 Section 14.2.2.9 rather than 14.2.2.14.
C14.2.2.15 Detailed Plain Concrete Shear Walls. Design requirements for plain masonry walls have existed for many
years, and the competing type of concrete construction is the plain concrete wall. To allow the use of such walls as the
lateralforceresisting system in SDC A and B, this provision requires such walls to contain at least the minimal
reinforcement specified in ACI 318 Section 22.6.7.2.
C14.2.2.16 Plain Concrete in Structures Assigned to Seismic Design Category C, D, E, or F. Modifications are made to
ACI 318 Section 22.10 that restrict markedly the use of ordinary and detailed structural plain concrete walls in SDC C, D, E,
and F.
C14.2.2.17 General Requirements for Anchoring to Concrete. ACI 318 uses the terminology of regions of moderate or
high seismic risk and structures assigned to intermediate or high seismic performance or design categories. In this
modification, the only changes to ACI 318 in Sections D3.3.3 through D3.3.4 are the replacement of that terminology with
the SDC terminology.
There are two changes to the provisions in ACI 318 Section D3.3.5. The first is the use of the SDC terminology, and the
second is the addition of the last phrase of the provision referring to the minimum design strength of the anchors. The last
phrase requires an anchor strength that is at least the maximum likely O0 value (2.5) times the design force calculated as
being transmitted to the attachment by the lateralforceresisting system.
C14.2.2.18 Strength Requirements for Anchors. ACI 318 requires laboratory testing to establish the strength of anchor
bolts greater than 2 inches in diameter or exceeding 25 inches in tensile embedment depth. This modification makes the ACI
318 equation giving the basic concrete breakout strength of a single anchor in tension in cracked concrete applicable
irrespective of the anchor bolt diameter and tensile embedment depth.
Korean Power Engineering (KPE) has made tension tests on anchors with diameters up to 4.25 inches and embedment depths
up to 45 inches and found that the diameter and embedment depth limits of ACI 318 Section D4.2.2 for the design procedure
for anchors in tension (Section D5.2) can be eliminated. KPE has also made shear tests on anchors with diameters up to 3.0
inches and embedment depths as large as 30 inches and found no effect of the embedment depth on shear strength. However,
the diameter tests showed that the basic shear breakout strength equation (ACI 318 Section D24) needed some modification
for the complete elimination of the 2 inch limit to be fully appropriate. Analytical work performed at the University of
Stuttgart supports the need for some modification to the ACI 318 Equation D24. Changes consistent with the Korean and
Stuttgart findings have already been made to the FIB Design Guide for anchors and a change proposal consistent with those
changes has been submitted to ACI 318 for consideration.
C14.2.3.1.2 Reinforcement for Uncased Concrete Piles (SDC C): The transverse reinforcing requirements in the
potential plastic hinge zone of uncased concrete piles in Seismic Design Category C is a selective composite of two ACI 318
requirements. In the potential plastic hinge region of an intermediate momentresisting concrete frame column, the
transverse reinforcement spacing is restricted to the least of: (a) 8 times the diameter of the smallest longitudinal bar, (b) 24
times the diameter of the tie bar, (c) onehalf the smallest crosssectional dimension of the column, and (d) 12 inches.
Outside of the potential plastic hinge region of a special momentresisting frame column, the transverse reinforcement
spacing is restricted to the smaller of: 6 times the diameter of the longitudinal column bars and 6 inches.
C14.2.3.1.5 Reinforcement for Precast Nonprestressed Concrete Piles (SDC C): Transverse reinforcement requirements
inside and outside of the plastic hinge zone of precast nonprestressed piles are clarified. The transverse reinforcement
requirement in the potential plastic hinge zone is a composite of two ACI 318 requirements (see Section C14.2.3.1.2).
Outside of the potential plastic hinge region the eight longitudinalbardiameter spacing is doubled. The maximum 8in. tie
spacing comes from current building code provisions for precast concrete piles.
C14.2.3.1.6 Reinforcement for Precast Prestressed Piles (SDC C): The transverse and longitudinal reinforcing
requirements given in ACI 318 Chapter 21 were never intended for slender precast prestressed concrete elements and will
result in unbuildable piles. The requirements are based on the 1993 Recommended Practice for Design, Manufacture and
Installation of Prestressed Concrete Piling by the PCI Committee on Prestressed Concrete Piling.
ASCE/SEI 705 Equation 14.21, originally from ACI 318, has always been intended to be a lowerbound spiral
reinforcement ratio for larger diameter columns. It is independent of the member section properties and therefore can be
applied to large or small diameter piles. For castinplace concrete piles and precast prestressed concrete piles, the resulting
spiral reinforcing ratios from this formula are considered to be sufficient to provide moderate ductility capacities.
Full confinement per Equation 14.21 is required for the upper 20 feet of the pile length where curvatures are large. The
amount is relaxed by 50 percent outside of that length in view of lower curvatures and in consideration of confinement
provided by the soil.
C14.2.3.2.5 Reinforcement for Precast Concrete Piles (SDC D through F): The transverse reinforcement requirements
for precast nonprestressed concrete piles are taken from current building code requirements and are intended to provide
ductility in the potential plastic hinge zones.
C14.2.3.2.6 Reinforcement for PrecastPrestressed Piles (SDC D through F): The last paragraph provides minimum
transverse reinforcement outside of the zone of prescribed ductile detailing.
C14.3 COMPOSITE STEEL AND CONCRETE STRUCTURES
This section provides guidance on the design of composite and hybrid steelconcrete structures. Composite structures are
defined as those incorporating structural elements made of steel and concrete portions connected integrally throughout the
structural element by mechanical connectors, bond, or both. Hybrid structures are defined as consisting of steel and concrete
structural elements connected together at discrete points. Composite and hybrid structural systems mimic many of the
existing steel (moment and braced frame) and concrete (moment frame and wall) configurations, but are given their own
design coefficients and factors in Table 12.21. Their design is based on the same ductility and energy dissipation concepts
used in conventional steel and reinforced concrete structures, but requires special attention to the interaction of the two
materials as it affects the stiffness, strength, and inelastic behavior of the members, connections, and systems.
C14.3.1 Reference Documents. Seismic design for composite structures assigned to Seismic Design Category D, E, or F is
governed primarily by Part II: Composite Structural Steel and Reinforced Concrete Buildings of ANSI/AISC 341. Part II of
ANSI/AISC 341 is less prescriptive than Part I and provides flexibility for designers to utilize analytical tools and results of
research in their practice. Composite structures assigned to Seismic Design Category A, B, or C may be designed according
to principles outlined in ANSI/AISC 360 and ACI 318. ACI 318 and ANSI/AISC 360 provide little guidance on connection
design; therefore, designers are encouraged to review ANSI/AISC 341 Part II for guidance on the design of joint areas.
Differences between older AISC and ACI provisions for crosssectional strength for composite columns have been
minimized by changes in the latest ANSI/AISC 360. However, there is not uniform agreement between the provisions in
ACI 318 and ANSI/AISC 360 regarding detailing, limits on material strengths, stability, and shear design for composite
columns. The composite design provisions in ANSI/AISC 360 are considered to be current.
C14.3.2 MetalCased Concrete Piles. Design of metalcased concrete piles, which are analogous to circular concrete filled
tubes, is governed by ASCE/SEI 705 Sections 14.2.3.1.3 and 14.2.3.2.4. The intent of these provisions is to require metalcased
concrete piles to have confinement and protection against longterm deterioration comparable to that for uncased
concrete piles.
C14.4 MASONRY
Seismic design for masonry structures is governed primarily by two documents produced by the Masonry Standards Joint
Committee (MSJC): ACI 53005/ASCE/SEI 505/TMS 4025, Building Code Requirements for Masonry Structures (the
MSJC Code), and ACI 530.105/ASCE/SEI 605/TMS 60205, Specification for Masonry Structures (the MSJC
Specification).
C14.4.2 R Factors. Where intermediate and special reinforced masonry shear walls are designed using the allowablestress
provisions of the MSJC Code, these additional requirements are intended to produce a level of inelastic flexural deformation
capacity consistent with that of intermediate and special reinforced masonry shear walls designed using the strengthdesign
provisions of the MSJC Code. The additional requirements are discussed in ASCE/SEI 705 Section C14.4.6.
C14.4.3 Classification of Shear Walls. Section 1.14 of the 2005 MSJC Code can be interpreted as permitting, in SDCs A
and B, masonry walls that need not be considered part of the lateralforceresisting system and that do not need to be isolated.
ASCE/SEI 705 Section 14.4.3 is intended to preclude that interpretation.
C14.4.5.1 Separation Joints. This section is intended to address force transfer across interfaces between masonry and other
materials, but it is redundant. Article 3.2B of the MSJC Specification requires that the interface between concrete and
masonry be cleaned and acceptable for laying of units. Further, Section 1.9.4.2.4 of the 2005 MSJC Code addresses the
design and transfer of shear at interfaces, and Section 1.7.5.2 requires that a load path and force transfer between a foundation
and the masonry above be maintained.
C14.4.5.2 Flanged Shear Walls. Section 1.9.4.2.3 of the MSJC Code contains the compression requirement (lesser of 6
times the flange thickness or the actual flange). The principal effect of the tension provision in ASCE/SEI 705 Section
14.4.5.2 is to establish the amount of tensile reinforcement used in calculating flexural capacity and maximum permitted
reinforcement, but this provision is not well established technically. Research in masonry, and analogous design provisions
for concrete (ACI 318 Section 21.7.5.2), suggest that effective flange widths in tension are more logically related to the total
wall height rather than the floortofloor height. The 2005 MSJC Code and ASCE/SEI 705 are working together to resolve
this issue and add appropriate requirements to TMS 402.
C14.4.6 Modifications to Chapter 2 of ACI 530/ASCE/SEI 5/TMS 402. Chapter 2 of the MSJC Code deals with
allowablestress design.
C14.4.6.1 Stress Increase. The MSJC Code permits allowable stresses to be increased by onethird for allowablestress
loading conditions that include wind or earthquake, provided that the legally adopted building code so permits. While the
alternate allowablestress loading combinations of the 2006 IBC do so permit, the allowablestress loading combinations of
ASCE/SEI 705 do not.
C14.4.6.2 Reinforcement Requirements and Details.
C14.4.6.2.1 Reinforcing Bar Size Limitations. The intent of this requirement is to prevent splitting of masonry due to the
presence of reinforcement. A similar requirement is appears in Chapter 3 (Strength Design) of the 2005 MSJC Code. The
MSJC is working to move that requirement to Chapter 1 (General Requirements) so that it would apply to all masonry
construction.
C14.4.6.2.2 Splices. In general, the first portion of this section, which prohibits splices in plastic hinge zones, is intended to
produce adequate inelastic deformation capacity in those regions. In general, the presence of splices in plastic hinge zones
reduces inelastic deformation capacity because the area of steel is doubled at the splice, reducing the extent of yielding.
However, there is some controversy concerning the technical validity and necessity for this requirement for masonry walls.
Similar requirements apply to plastic hinge zones of reinforced concrete frames, they do not apply to plastic hinge zones of
reinforced concrete walls. Also, this requirement does not distinguish between shearcritical and flexurally dominated shear
walls. The MSJC is continuing to discuss related requirements for flexurally dominated, highly ductile shear walls.
The remaining portions of this section (requirements for splices) are intended to provide adequate capacity of welded splices
and mechanical connections. The MSJC is developing similar provisions.
C14.4.6.2.3 Maximum Area of Flexural Tensile Reinforcement. The intent of this section is to produce adequate inelastic
flexural deformation capacity in flexurally dominated masonry shear walls by placing an upper limit on flexural
reinforcement, so that behavior is dominated by yielding of reinforcement rather than by crushing of the compression toe.
Similar provisions appear in Chapter 3 (Strength Design) of the MSJC Code and are being developed for Chapter 2
(AllowableStress Design).
C14.4.7 Modifications to Chapter 3 of ACI 530/ASCE/SEI 5/TMS 402.
C14.4.7.2 Splices in Reinforcement. See Section C14.4.6.2.2.
C14.4.7.3 Coupling Beams. The intent of this requirement is to produce adequate inelastic flexural deformation capacity in
coupling beams. The section is somewhat redundant with Section 3.1.3 of the MSJC Code, which requires capacity design of
masonry elements for shear.
C14.4.7.4 Deep Flexural Members. The intent of this requirement is to require that the design of deep flexural members
correctly addresses the presence of distributed flexural reinforcement in capacity design for shear, and that crack widths are
adequately controlled.
C14.4.7.5 Shear Keys. The intent of this requirement is to increase resistance to sliding shear at the foundation level of
flexurally dominated masonry shear walls. The original proposal was based on laboratory research (Leiva et al., 1990)
involving isolated shear walls. In subsequent research (Seible et al., 1993), flanged walls without shear keys did not show
sliding.
C14.4.7.6 Anchoring to Masonry. The intent of this requirement is to guard against brittle failure of masonry anchorages
that are part of the seismic forceresisting system.
C14.4.7.7 Anchor Bolts. ASCE/SEI 705 Sections 14.4.7.7 and 14.4.7.8 augment the current anchor bolt provisions of
MSJC Code Chapter 3 (Strength Design) to address pryout and to include an appropriate f factor.
C14.4.8 Modifications to Chapter 6 of ACI 530/ASCE/SEI 5/TMS 402. There is an apparent difference in the treatment
of corrugated sheet metal anchors in different chapters of the MSJC Code. Chapter 6 of that document, dealing with masonry
veneer, permits corrugated sheetmetal anchors. Chapters 2 and 3 of that document do not permit multiwythe, noncomposite
masonry (functionally identical to veneer) to be bonded by corrugated sheetmetal anchors.
C14.4.9 Modifications to ACI 530.1/ASCE/SEI 6/TMS 602.
C14.4.9.1 Construction Procedures. This requirement was introduced originally as a result of the TCCMaR program as a
way to address volume loss as a result of plastic shrinkage of grout. The original provision required the use of a particular
admixture (Sika’s Grout Aid®) in the grout. The MSJC Specification requires both consolidation and reconsolidation of
masonry grout, which in combination with today’s masonry construction materials can minimize grout shrinkage without the
requirement of a proprietary grout admixture available from a single source.
C14.5 WOOD
C14.5.1 Reference Documents. Two national consensus standards are adopted for seismic design of engineered wood
structures: the National Design Specification (NDS), and the Special Design Provisions for Wind and Seismic (SDPWS)
Supplement to the NDS. Both of these standards, published by the American Forest and Paper Association (AF&PA), are
presented in dual allowable stress design (ASD) and load and resistance factor design (LRFD) formats. Both standards
reference a number of secondary standards for related items such as wood materials and fasteners. SDPWS addresses general
principles and specific detailing requirements for shear wall and diaphragm design and provides tabulated nominal unit shear
capacities for shear wall and diaphragm sheathing and fastening. The balance of member and connection design is to be in
accordance with the NDS. A commentary to the NDS is published by AF&PA (2005b); commentary to the SDPWS is
included in the SDPWS publication (AF&PA, 2005c).
C14.5.2 Framing. This section provides specific guidance on two general topics related to detailing. First, vertical loads on
columns and posts must be transferred in and out by end bearing only or by connectors only; mixing the capacity of end
bearing and connectors is prohibited due to a potential lack of deformation compatibility. Second, load path continuity for
top plates, which often function as collectors, is addressed.
C14.5.3.1 ASCE/SEI 705 Modification to SDPWS Section 4.3.3.2, Summing Shear Capacities. This amendment to the
SDPWS does not provide additional clarity; therefore, it is expected to be deleted ASCE/SEI 710.
REFERENCES
American Forest and Paper Association. 2005a. National Design Specification (NDS) for Wood Construction,
ANSI/AF&PA NDS2005. AF&PA, Washington, D.C.
American Forest and Paper Association. 2005b. National Design Specification Commentary, 2005 ed. AF&PA,
Washington, D.C.
American Forest and Paper Association. 2005c. Special Design Provisions for Wind and Seismic (Wind & Seismic),
ANSI/AF&PA SDPWS2005. AF&PA, Washington, D.C.
American Institute of Timber Construction. 2005. Timber Construction Manual, 5th ed. John Wiley and Sons, Inc.
APA  The Engineered Wood Association. 2004. Diaphragms and Shear Walls Design/Construction Guide, L350. APA,
Tacoma, Washington.
APA  The Engineered Wood Association. 1994. Northridge California Earthquake, T945. APA, Tacoma, Washington.
Applied Technology Council. 1981. Guidelines for the Design of Horizontal Wood Diaphragms, ATC7. ATC, Redwood
City, California.
Banks, G., and J. Stanton. 2005. “PaneltoPanel Connections for HollowCore Shear Walls Subjected to Seismic Loading,”
in Proceedings, 2005 PCI Convention, Palm Springs, California.
Breyer et al. 2006. Design of Wood Structures ASD/LRFD, Sixth Edition. McGrawHill Book Company, New York, New
York.
Building Seismic Safety Council. 2003. NEHRP Recommended Provisions for Seismic Regulations for New Buildings and
Other Structures, FEMA 450. Federal Emergency Management Agency, Washington, D.C.
Canadian Wood Council. 2005. Wood Design Manual 2005. Canadian Wood Council, Ottawa.
Canadian Wood Council. 1995. Wood Reference Handbook. Canadian Wood Council, Ottawa.
Cobeen, K. 2004. “Recent developments in the Seismic Design and Construction of Woodframe Buildings,” Chapter 18 of
Earthquake Engineering from Engineering Seismology to PerformanceBased Engineering, edited by Yousef Bozorgia and
Vitelmo Bertero. CRC Press, LLC, Boca Raton, Florida.
Consortium of Universities for Research in Earthquake Engineering. 2004. Recommendations for Earthquake Resistance in
the Design and Construction of Woodframe Buildings, CUREE W30. CUREE, Richmond, California.
Department of the Army, Navy, and Air Force. 1992. Seismic Design for Buildings, TM580910 (TriServices Manual).
U.S. Government Printing Office, Washington, D.C.
Dolan, J. D. 2003. “Wood Structures,” Chapter 15 of Earthquake Engineering Handbook, edited by WaiFah Chen and
Charles Scawthorn. CRC Press, LLC, Boca Raton, Florida.
Earthquake Engineering Research Institute. 1996. “Northridge earthquake reconnaissance report,” Earthquake Spectra,
Chapter 6 of Supplement C to Volume 11.
Faherty, Keith F., and T. G. Williamson. 1989. Wood Engineering and Construction Handbook. McGrawHill, New York,
New York.
Federal Emergency Management Agency. 2005. Coastal Construction Manual, Third Edition, FEMA 55. FEMA,
Washington, D.C.
Forest Products Laboratory. 1986. Wood: Engineering Design Concepts. Materials Education Council, The Pennsylvania
State University, University Park.
Goetz, KarlHeinz, Dieter Hoor, Karl Moehler, and Julius Natterer. 1989. Timber Design and Construction Source Book: A
Comprehensive Guide to Methods and Practice. McGrawHill, New York, New York.
Hoyle and Woeste. 1989. Wood Technology and Design of Structures. Iowa State University Press.
International Code Council. 2006. “ICC Standard on the Design and Construction of Log Structures,” Third Draft. ICC,
Country Club Hills, Illinois.
Ishizuka, T., and N. M. Hawkins. 1987. Effect of Bond Deterioration on the Seismic Response of Reinforced and Partially
Prestressed Concrete Ductile Moment Resistant Frames, Report SM 872. Department of Civil Engineering, University of
Washington, Seattle.
Karacabeyli, E., and M. Popovsky . 2003. “Design for Earthquake Resistance,” Chapter 15 of Timber Engineering, edited
by H. Larsen & S. Thelandersson. John Wiley & Sons.
Keenan, F. J. 1986. Limit States Design of Wood Structures. Morrison Hershfield Limited.
Masonry Standards Joint Committee (MSJC). 2005a. Building Code Requirements for Masonry Structures, ACI 530
05/ASCE/SEI 505/TMS 4025.
Masonry Standards Joint Committee. 2005b. Specification for Masonry Structures, ACI 530.105/ASCE/SEI 605/TMS
60205.
Nakaki, S., R. Becker, M. G. Oliva, and D. Paxson. 2005. “New Connections for Precast Wall Systems in High Seismic
Regions,” in Proceedings, 2005 PCI Convention, Palm Springs, California.
Nakaki, S., J. F. Stanton, and S. Sritharan. 2001. “The PRESSS FiveStory Precast Concrete Test Building, University of
California, San Diego, La Jolla, California,” PCI Journal, 46(5):2026.
Park, R., and K. J. Thompson. 1977. “Cyclic Load Tests on Pretsressed and Partially Prestressed BeamColumn Joints,”
PCI Journal, 22(3):84110.
Schultz, A. E., and R. A. Magana. 1996. “Seismic Behavior of Connections in Precast Concrete Walls,” Proceedings, Mete
A. Sozen Symposium, SP162, pp. 273311. American Concrete Institute, Farmington Hills, Michigan.
Steel Deck Institute. 2004. Diaphragm Design Manual, 3rd Edition, No. DDMO3. SDI, Fox River Grove, Illinois.
Structural Engineeers Association of California. 1999. Recommended Lateral Force Requirements and Commentary.
SEAOC, Sacramento, California.
Structural Engineeers Association of Northern California. 2005. Guidelines for Seismic Evaluation and Rehabilitation of
Tiltup Buildings and Other Rigid Wall/Flexible Diaphragm Structures. SEAONC, Sacramento, California.
Sherwood and Stroh. 1989. “WoodFrame House Construction” in Agricultural Handbook 73. U.S. Government Printing
Office, Washington, D.C.
Somayaji, Shan. 1992. Structural Wood Design. West Publishing Co., St. Paul, Minnesota.
Stalnaker, Judith J., and E. C. Harris. 1996. Structural Design in Wood, Second Edition. McGrawHill, New York, New
York.
U.S. Department of Agriculture, National Oceanic and Atmospheric Administration. 1971. San Fernando, California,
Earthquake of February 9, 1971. NOAA, Washington, D.C.
Page intentionally left blank.
Figure C15.11 Image of Steel multilegged water tower.
COMMENTARY TO CHAPTER 15,
SEISMIC DESIGN REQUIREMENTS FOR
NONBUILDING STRUCTURES
C15.1.1 Nonbuilding Structures. Building codes traditionally have been perceived as minimum standards for the design of
nonbuilding structures, and building code compliance of these structures is required by building officials in many
jurisdictions. However, requirements in the industry reference documents are often at odds with building code requirements.
In some cases, the industry documents need to be altered while in other cases the building codes need to be modified.
Registered design professionals are not always aware of the numerous accepted documents within an industry and may not
know whether the accepted documents are adequate. The intent of Chapter 15 of the standard is to bridge the gap between
building codes and existing industry reference documents.
Differences between the ASCE/SEI 705 design approaches for buildings and industry document requirements for steel
multilegged water towers (Figure C15.11) are representative of this inconsistency. Historically, such towers have
performed well when properly designed in accordance with American Water Works Association (AWWA) standards and
industry practices. Those standards and practices differ from the ASCE/SEI 705 treatment of buildings in that tensiononly
rods are allowed, upset rods are preloaded at the time of installation, and connection forces are not amplified.
Figure C15.11 Steel multilegged water tower.
Chapter 15 also provides an appropriate link so that the industry reference documents can be used with the seismic ground
motions established in the standard. It should be noted that some nonbuilding structures are very similar to buildings and can
be designed employing sections of the standard directly, whereas other nonbuilding structures require special analysis unique
to the particular type of nonbuilding structure.
Note that building structures, vehicular bridges, electrical transmission towers, hydraulic structures (e.g., dams), buried utility
lines and their appurtenances, and nuclear reactors are excluded from the scope of the nonbuilding structure requirements.
The excluded structures are covered by other well established design criteria (e.g., electrical transmission towers and
vehicular bridges), are not under the jurisdiction of local building officials (e.g., nuclear reactors, and dams), or require
technical considerations beyond the scope of the standard (e.g., buried utility lines and their appurtenances).
C15.1.2 Design. Nonbuilding structures and building structures have much in common with respect to design intent and
expected performance, but there are also important differences. Chapter 15 relies on other portions of the standard where
possible and provides special notes where necessary.
There are two types of nonbuilding structures: those with structural systems similar to buildings, and those with structural
systems not similar to buildings. Specific requirements for these two cases appear in Sections 15.5 and 15.6.
Figure C15.12 Image of Steel pipe rack.
C15.1.3 Structural Analysis Procedure Selection. Nonbuilding structures that are similar to buildings are subject to the
same analysis procedure limitations as building structures. Nonbuilding structures that are not similar to buildings are subject
to those limitations and are subject to procedure limitations prescribed in applicable specific reference documents.
For many nonbuilding structures supporting flexible system components, such as pipe racks (Figure C15.12), the supported
piping and platforms generally are not regarded as rigid enough to redistribute seismic forces to the supporting frames.
Figure C15.12 Steel pipe rack.
For nonbuilding structures supporting rigid system components, such as steam turbine generators (STGs) and heat recovery
steam generators (HRSGs) (Figure C15.13), the supported equipment, ductwork, and other components (depending on how
they are attached to the structure) may be rigid enough to redistribute seismic forces to the supporting frames. Torsional
effects may need to be considered in such situations.
Section 12.6 presents seismic analysis procedures for building structures based on the Seismic Design Category; the
fundamental period, T; and the presence of certain horizontal or vertical irregularities in the structural system. Where the
fundamental period is greater than or equal to 3.5Ts (where Ts = SD1/SDS), the use of the equivalent lateral force procedure is
not permitted in Seismic Design Categories D, E, and F. This requirement is based on the fact that, unlike the dominance of
the first mode response in case of buildings with lower first mode period, higher vibration modes do contribute more
significantly in situations when the first mode period is larger than 3.5Ts. For buildings that exhibit classic flexural
deformation patterns (such as slender shear wall or braced frame systems), the second mode frequency is at least 3.5 times
the first mode frequency, so where the fundamental period exceeds 3.5Ts, the higher modes will have larger contributions to
the total response as they occur near the peak of the design response spectrum
It follows that dynamic analysis (modal response spectrum analysis, or responsehistory analysis) is required for buildinglike
nonbuilding structures if the first mode period is larger than 3.5Ts and that the equivalent lateral force analysis is sufficient
for nonbuilding structures that respond as singledegreeoffreedom systems such as singlepedestal elevated water tanks.
The recommendations for nonbuilding structures provided below are intended to supplement the designer’s judgment and
experience. The designer is given considerable latitude in selecting a suitable analysis method for nonbuilding structures.
Figure C15.13 Image of Heat recovery steam generators.
Figure C15.13 Heat recovery steam generators.
Buildinglike Nonbuilding Structures. Table 12.61 is used in selecting analysis methods for buildinglike nonbuilding
structures, but, as illustrated in the following three conditions, the relevance of key behavior must be considered carefully:
1. Irregularities: Table 12.61 requires dynamic analysis for Seismic Design Category D, E, and F structures having certain
horizontal or vertical irregularities. Some of these building irregularities (defined in Section 12.3.2) are relevant to
nonbuilding structures. The weakand softstory vertical irregularities (Types 1a, 1b, 5a, and 5b of Table 12.32) are
pertinent to the behavior of buildinglike nonbuilding structures. Other vertical and horizontal irregularities may or may
not be relevant as described below.
a. Horizontal irregularities: Horizontal irregularities of Type 1a and 1b affect the choice of analysis method, but these
irregularities apply only where diaphragms are rigid or semirigid and some buildinglike nonbuilding structures have
either no diaphragms or flexible diaphragms.
b. Vertical irregularities: Vertical irregularity Type 2 is relevant where the various levels actually support significant
loads. Where a buildinglike nonbuilding structure supports significant mass at a single level, while other levels
support small masses associated with stair landings, access platforms, and so forth, dynamic response will be
dominated by the first mode, so the equivalent lateral force procedure may be applied. Vertical irregularity Type 3
addresses large differences in the horizontal dimension of the seismic forceresisting system in adjacent stories,
since the resulting stiffness distribution can produce a fundamental mode shape unlike that assumed in the
development of the equivalent lateral force procedure. Since the concern relates to stiffness distribution, it is the
horizontal dimension of the seismic forceresisting system, not of the overall structure, that is important.
2. Arrangement of supported masses: Even where a nonbuilding structure has buildinglike appearance, it may not behave
like a building, depending on how masses are attached. For example, the response of nonbuilding structures with
suspended vessels and boilers cannot be determined reliably using the equivalent lateral force procedure because of the
pendulum modes associated with the significant mass of the suspended components. The resulting pendulum modes,
while potentially reducing story shears and base shear, may require large clearances to allow pendulum motion of the
supported components and may produce excessive demands on attached piping. Dynamic analysis should be performed
in such cases, with consideration for appropriate impact forces in the absence of adequate clearances.
Figure C15.14 of Multiple lateral supports.
3. Relative rigidity of beams: Even where a classic building model may seem appropriate, the equivalent lateral force
procedure may underpredict the total response if the beams are flexible relative to the columns (of moment frames) or
the braces (of braced frames). This is because higher modes associated with beam flexure may contribute more
significantly to the total response (even if the first mode response is at a period less than 3.5Ts). This situation of
flexible beams can be especially pronounced for nonbuilding structures since the “normal” floors common to buildings
may be absent. Therefore, the dynamic analysis procedures are recommended for buildinglike nonbuilding structures
with flexible beams.
Nonbuilding Structures Not Similar to Buildings. The (static) equivalent lateral force procedure is based on classic
building dynamic behavior, which is an inappropriate characterization for many nonbuilding structures not similar to
buildings. As discussed below, several issues should be considered for selecting either an appropriate method of dynamic
analysis or a suitable distribution of lateral forces for static analysis.
1. Structural geometry: The dynamic response of nonbuilding structures with a fixed base and a relatively uniform
distribution of mass and stiffness, such as bottomsupported vertical vessels, stacks, and chimneys, can be represented
adequately by a cantilever (shear building) model. For these structures the equivalent lateral force procedure provided in
the standard is suitable. This procedure treats the dynamic response as being dominated by the first mode. In such cases,
it is necessary to identify the first mode shape (using, for instance, the RayleighRitz method or other classical methods
from the literature) for distribution of the dynamic forces. For some structures, such as tanks with low heighttodiameter
ratios storing granular solids, it is conservative to assume a uniform distribution of forces. Dynamic analysis is
recommended for structures that have neither a uniform distribution of mass and stiffness nor an easily determined first
mode shape.
2. Number of lateral supports: Cantilever models are obviously unsuitable for structures with multiple supports. Figure
C15.14 shows a nonbuilding braced frame structure that provides nonuniform horizontal support to a piece of
equipment. In such cases, the analysis should include coupled model effects. For such structures an application of the
equivalent lateral force method could be used depending on the number and locations of the supports. For example,
most beamtype configurations lend themselves to application of the equivalent lateral force method.
Figure C15.14 Multiple lateral supports.
3. Method of supporting dead weight: Certain nonbuilding structures (such as power boilers) are supported from the top.
They may be idealized as pendulums with uniform mass distribution. In contrast, a suspended platform may be idealized
as a classic pendulum with concentrated mass. In either case, these types of nonbuilding structures can be analyzed
adequately using the equivalent lateral force method by calculating the appropriate frequency and mode shape. Figure
C15.15 shows a nonbuilding structure containing lug supported equipment with WP greater than 0.25(WS + WP). In
such cases, the analysis should include a coupled system with the mass of the equipment and the local flexibility of the
supports considered in the model. Where the support is located near the nonbuilding structure’s vertical location of the
center of mass, a dynamic analysis is recommended.
Figure C15.15 of Unusual support of dead weight.
Figure C15.16 of Mass irregularities.
W1
W2
W1
W2
W3
W4
(a) (b)
Figure C15.15 Unusual support of dead weight.
4. Mass irregularities: Just as in the case of buildinglike nonbuilding structures, the presence of significantly uneven mass
distribution can render the structures unsuitable for application of the equivalent lateral force method. The dynamic
analysis methods are recommended in such situations. Figure C15.16 illustrates two such situations. In part (a), a mass
irregularity exists if W1 is greater than 1.5W2 or less than 0.67W2. In part (b), a mass irregularity exists if W3 is greater
than either 1.5W2 or 1.5W4.
Figure C15.16 Mass irregularities.
Figure C15.17 of Torsional irregularity.
Figure C15.18 of Softstory irregularity.
5. Torsional irregularities: Structures in which the fundamental mode of response is torsional or in which modes with
significant mass participation exhibit a prominent torsional component may also have inertial force distributions that are
significantly different from that predicted by the equivalent lateral force method. In such cases dynamic analyses should
be considered. Figure C15.17 illustrates one such case where a vertical vessel is attached to a secondary vessel with W2
greater than about 0.25(W1 + W2).
Figure C15.17 Torsional irregularity.
6. Stiffness/strength irregularities: Just as for buildinglike nonbuilding structures, abrupt changes in the distribution of
stiffness or strength in a nonbuilding structure not similar to buildings can result in substantially different inertial forces
that differ substantially from those indicated by the equivalent lateral force method. Figure C15.18 represents one such
case. For structures having such configurations, consideration should be given to use of dynamic analysis procedures.
Even where dynamic analysis is required, the standard does not define in any detail the degree of modeling; an adequate
model may have a few dynamic degrees of freedom or tens of thousands of dynamic degrees of freedom. The important
point is that the model captures the significant dynamic response features so that the resulting lateral force distribution is
valid for design. The designer is responsible to determine whether dynamic analysis is warranted and, if so, the degree
of detail required to address adequately the seismic performance.
Figure C15.18 Softstory irregularity.
Figure C15.19 of Coupled system.
WS
WP
(a) (b)
Nonbuilding
structure
Support
structure
7. Coupled Response: Where the weight of the supported structure is large compared to the weight of the supporting
structure, the combined response can be affected significantly by the flexibility of the supported nonbuilding structure.
In that case, dynamic analysis of the coupled system is recommended. Examples of such structures are shown in
Figure C15.19. Part (a) shows a flexible nonbuilding structure with Wp greater than 0.25(Ws + Wp), supported by a
relatively flexible structure; the flexibility of the supports and attachments should be considered. Part (b) shows flexible
equipment connected by a largediameter, thickwalled pipe and supported by a flexible structure; the structures should
be modeled as a coupled system including the pipe.
Figure C15.19 Coupled system.
C15.2 REFERENCE DOCUMENTS
Chapter 15 of the standard makes extensive use of reference documents in the design of nonbuilding structures for seismic
forces. The documents referenced in Chapter 15 are industry documents commonly used to design specific types of
nonbuilding structures. The vast majority of these reference documents contain seismic provisions that are based on the
seismic ground motions of the 1997 UBC or earlier editions of the UBC. In order to use these reference documents, Chapter
15 modifies the seismic force provisions of these reference documents through the use of “bridging equations.” The standard
only modifies industry documents that specify seismic demand and capacity. The bridging equations are intended to be used
directly with the other provisions of the specific reference documents. Unlike the other provisions of the standard, if the
reference documents are written terms of allowable stress design, then the bridging equations are shown in allowable stress
design format. In addition, the detailing requirements referenced in Tables 15.41 and Table 15.42 must be followed, as well
as the general requirements found in Section 15.4.1. The usage of reference documents in conjunction with the requirements
of Section 15.4.1 are summarized below in Table C15.21.
Table C15.21 Usage of Reference Documents in Conjunction with Section 15.4.1
Subject
Requirement
R, O0, and Cd values, detailing
requirements, and height limits
Use values and limits in Tables 12.21, 15.41, or 15.42 as appropriate.
Values from the reference document are not to be used.
Minimum base shear
Use the appropriate value from Equation 15.41 or 15.42 for nonbuilding
structures not similar to buildings. For structures containing liquids, gases,
and granular solids supported at the base, the minimum seismic force
cannot be less than that required by the reference document.
Importance factor
Use the value from Section 15.4.1.1 based on Occupancy Category.
Importance factors from the reference document are not to be used unless
they are greater than those provided in the standard.
Vertical distribution of lateral
load
Use requirements of Section 12.8.3 or Section 12.9 or the applicable
reference document.
Seismic provisions of
reference documents
The seismic force provisions of reference documents may be used only if
they have the same basis as Section 11.4 and the resulting values for total
lateral force and total overturning moment are no less than 80 percent of
the values obtained from the standard.
Load combinations
Load combinations specified in Section 2.3 (LRFD) or Section 15 (includes
ASD load combinations of Section 2.4) must be used.
Currently, only two reference documents have been revised to meet the seismic requirements of the standard. AWWA D100
05 and API 650 10th Edition Addendum 4 (2005) have been adopted by reference in the standard without modification except
that height limits are imposed on “elevated tanks on symmetrically braced legs (not similar to buildings)” in AWWA D100
05. Both of these reference documents apply to welded steel liquid storage tanks.
C15.3 NONBUILDING STRUCTURES SUPPORTED BY OTHER SRUCTURES
There are instances where nonbuilding structures not similar to buildings are supported by other structures or other
nonbuilding structures. This section specifies how the seismic design loads for such structures are to be determined and the
detailing requirements that are to be satisfied in the design.
C15.3.1 Less than 25 Percent of Combined Weight Condition. In many instances, the weight of the supported
nonbuilding structure is relatively small compared to the weight of the supporting structure such that the supported
nonbuilding structure will have a relatively small effect on the overall nonlinear earthquake response of the primary structure
during designlevel ground motions. It is permitted to treat such structures as nonstructural components and use the
requirements of Chapter 13 for their design. The ratio of secondary component weight to total weight of 25 percent at which
this treatment is permitted is based on judgment and was introduced into code provisions in the 1988 Uniform Building Code
by the SEAOC Seismology Committee. Analytical studies, typically based on linear elastic primary and secondary
structures, indicate that the ratio should be lower, but the SEAOC Seismology Committee judged that the 25 percent ratio is
appropriate where primary and secondary structures exhibit nonlinear behavior that tends to lessen the effects of resonance
and interaction. In cases where a nonbuilding structures (or nonstructural component) is supported by another structure, it
may be appropriate to analyze in a single model. In such cases it is intended that seismic design loads and detailing
requirements be determined following the procedures of Section 15.3.2. Where there are multiple large nonbuilding
structures, such as vessels supported on a primary nonbuilding structure, and the weight of an individual supported
nonbuilding structure does not exceed the 25 percent limit but the combined weight of the supported nonbuilding structures
does, it is recommended that the combined analysis and design approach of Section 15.3.2 be used. It is also suggested that
dynamic analysis be performed in such cases, since the equivalent lateral force procedure may not capture some important
response effects in some members of the supporting structure.
Where the weight of the supported nonbuilding structure does not exceed the 25 percent limit and a combined analysis is
performed, the following procedure should be used to determine the Fp force of the supported nonbuilding structure based on
Equation 13.34:
1. A modal analysis should be performed in accordance with Section 12.9. The base shear of the combined structure and
nonbuilding structure should be taken as no less than 85 percent of the equivalent lateral force procedure base shear.
2. For a component supported at level i, the acceleration at that level should be taken as ai, the total shear just below level i
divided by the seismic weight at and above level i.
3. The elastic value of the component shear force coefficient should next be determined as the shear force from the modal
analysis at the point of attachment of the component to the structure divided by the weight of the component. This value
is preliminarily taken as aiap. Since ap cannot be taken as less than 1.0, the value of ap is taken as aiap / ai, except that
the final value ap need not be taken as greater than 2.5 and should not be taken as less than 1.0. The final value of aiap
should be the final value of ai determined in Step 2 multiplied by the final value of ap determined earlier in this step.
4. The resulting value of (aiap) should be used in Equation 13.34; the resulting value of Fp is subject to the maximum and
minimum values of Equations 13.32 and 13.33, respectively.
C15.3.2 Greater Than or Equal to 25 Percent Combined Weight Condition. Where the weight of the supported
structure is relatively large compared to the weight of the supporting structure, the overall response can be affected
significantly. The standard sets forth two analysis approaches, depending on the rigidity of the nonbuilding structure. The
determination of what is deemed rigid or flexible is based on the same criteria used for nonstructural components.
Where the supported nonbuilding structure is rigid, it is acceptable to treat the supporting structure as a nonbuilding structure
similar to a building and to determine its design loads and detailing using the requirements of Section 15.5. The design of the
rigid nonbuilding structure and its anchorage is determined using the requirements of Chapter 13 with the amplification
factor, ap, taken as 1.0. However, this is a relatively rare condition since the flexibility of any directly supporting members in
the primary structure, such as floor beams, must be considered in determining the period of the component.
In the usual case, where the supported nonbuilding structure is flexible, a combined model of the supporting structure and the
supported nonbuilding structure is used. The design loads and detailing are determined based on the lower R value of the
supported nonbuilding structure or supporting structure.
Although not specifically mentioned in Section 15.3.2, another approach is permitted. A nonlinear response history analysis
of the combined system can be performed in accordance with Section 16.2, and the results can be used for the design of both
the supported and supporting nonbuilding structures. This option should be considered where standard static and dynamic
elastic analysis approaches may be inadequate to evaluate the earthquake response (such as for suspended boilers). This
option should be used with extreme caution since modeling and interpretation of results requires considerable judgment. Due
to this sensitivity, Section 16.2 requires independent design review.
C15.4 STRUCTURAL DESIGN REQUIREMENTS
This section specifies the basic coefficients and minimum design forces to be used to determine seismic design loads for
nonbuilding structures. It also specifies height limits and restrictions. As with building structures, it presumes that the first
step in establishing the design forces is to determine the design base shear for the structure.
There are two types of nonbuilding structures: those with structural systems similar to buildings and those with structural
systems not similar to buildings. Specific requirements for these two cases appear in Sections 15.5 and 15.6.
C15.4.1 Design Basis. Separate tables are provided in this section that identify the basic coefficients, associated detailing
requirements, and height limits and restrictions for the two types of nonbuilding structures.
For nonbuilding structures similar to buildings, the design seismic loads are determined using the same procedures used for
buildings as specified in Chapter 12 with two exceptions: fundamental periods are determined in accordance with Section
15.4.4, and Table 15.41 provides additional options for structural systems. Although only Section 12.8 (the equivalent
lateral force procedure) is specifically mentioned in Section 15.4.1, Section 15.1.3 provides the analysis procedures that are
permitted for nonbuilding structures.
In Table 15.41, seismic coefficients, system restrictions, and height limits are specified for a few nonbuilding structures
similar to buildings. The values of R, O o, and Cd, the detailing requirement references, and the structural system height
limits are the same as those in Table 12.21 for the same systems, except for ordinary moment frames. In Chapter 12
increased height limits for ordinary moment frames structural systems apply to metal building systems, while in Chapter 15
they apply to pipe racks with end plate bolted moment connections. The seismic performance of pipe racks was judged to be
similar to that of metal building structures with end plate bolted moment connections, so the height limits were made the
same as those specified in previous editions.
Table 15.41 also provides lower R values with less restrictive height limits in Seismic Design Categories D, E, and F based
on good performance in past earthquakes. For some options, no seismic detailing is required if very low values of R (and
corresponding high seismic design forces) are used. The concept of extending this approach to other structural systems is the
subject of future research using the methodology developed by the ATC 63 project.
For nonbuilding structures not similar to buildings, the seismic design loads are determined as in Chapter 12 with three
exceptions: the fundamental periods are determined in accordance with Section 15.4.4, the minima are those specified in
Section 15.4.1.2, and the seismic coefficients are those specified in Table 15.42.
Some entries in Table 15.42 may seem to be conflicting or confusing. For example, the first major entry is for elevated
tanks, vessels, bins, or hoppers. A subset of this entry is for tanks on braced or unbraced legs. This subentry is intended for
structures where the supporting columns are integral with the shell (such as an elevated water tank). Tensiononly bracing is
allowed for such a structure. Where the tank or vessel is supported by buildinglike frames, the frames are to be designed in
accordance with all of the restrictions normally applied to building frames. The entry for tanks or vessels supported on
structural towers similar to buildings assumes that the operating weight of the supported tank or vessel is less than 25 percent
of the total weight; if the ratio is greater than 25 percent, the proper entry is that most closely related to the subject vessel or
bin.
C15.4.1.1 Importance Factor. The importance factor for a nonbuilding structure is based on the occupancy category
defined in Chapter 1 of the standard or the building code being used in conjunction with the standard. In some cases,
reference standards provide a higher importance factor, in which case the higher importance factor is used.
If the importance factor is taken as 1.0 based on a Hazard and Operability (HAZOP) analysis performed in accordance with
Chapter 1, the third paragraph of Section 1.5.2 requires careful consideration; worstcase scenarios (instantaneous release of a
vessel or piping system) must be considered. HAZOP risk analysis consultants often do not make such assumptions, so the
design professional should review the HAZOP analysis with the HAZOP consultant to confirm that such assumptions have
been made in order to validate adjustment of the importance factor. Clients may not be aware that HAZOP consultants do
not normally consider the worstcase scenario of instantaneous release but tend to focus on other more hypothetical limitedrelease
scenarios, such as those associated with a 2inch square hole in a tank or vessel.
C15.4.2 Rigid Nonbuilding Structures. The definition of rigid (having a natural period of less than 0.06 second) was
selected judgmentally. Below that period, the energy content of seismic ground motion is generally believed to be very low,
and therefore the building response is not likely to be excessively amplified. Also, it is unlikely that any building will have a
first mode period as low as 0.06 second, and it is even unusual for a second mode period to be that low. Thus, the likelihood
of either resonant behavior or excessive amplification becomes quite small for equipment having periods below 0.06 second.
The analysis to determine the period of the nonbuilding structure should include the flexibility of the soil subgrade.
C15.4.3 Loads. As for buildings, the seismic weight must include the range of design operating weight of permanent
equipment.
C15.4.4 Fundamental Period. A significant difference between building structures and nonbuilding structures is that the
approximate period formulas and limits of Section 12.8.2.1 may not be used for nonbuilding structures. In lieu of calculating
a specific period for a nonbuilding structure for determining seismic lateral forces, it is of course conservative to assume a
period of Ts (= SD1/SDS) which results in the largest lateral design forces. Computing the fundamental period is not
considered a significant burden, since most commonly used computer analysis programs can perform the required
calculations.
C15.4.8 SiteSpecific Response Spectra. Where sitespecific response spectra are required, they should be developed in
accordance with Chapter 21 of the standard. If determined for other recurrence intervals, Section 21.1 applies, but Sections
21.2 through 21.4 apply only to MCE determinations. Where other recurrence intervals are used, it should be demonstrated
that the requirements of Chapter 15 also are satisfied.
C15.5 NONBUILDING STRUCTURES SIMILAR TO BUILDINGS
C15.5.1 General. Although certain nonbuilding structures exhibit behavior similar to that of building structures, their
functions and occupancies are different. Section 15.5 of the standard addresses the differences.
C15.5.2 Pipe Racks. Freestanding pipe racks supported at or below grade with framing systems that are similar to building
systems are designed in accordance with Section 12.8 or 12.9 and Section 15.4. Singlecolumn pipe racks that resist lateral
loads should be designed as inverted pendulums.
Based on good performance in past earthquakes, Table 15.41 sets forth the option of lower R values and less restrictive
height limits for structural systems commonly used in pipe racks. The R value versus height limit tradeoff recognizes that
the size of some nonbuilding structures is determined by factors other than traditional loadings and results in structures that
are much stronger than required for seismic loadings. Therefore, the ductility demand is generally much lower than that for a
corresponding building. The intent is to obtain the same structural performance at the increased heights. This option will
prove to be economical in most situations due to the relative cost of materials and construction labor. The lower R values and
increased height limits of Table 15.41 apply to nonbuilding structures similar to buildings; they cannot be applied to
building structures. Table C15.51 illustrates the R values and height limits for a 70foothigh steel ordinary moment frame
(OMF) pipe rack.
Table C15.51 R Value Selection Example for Steel OMF Pipe Racks
SDC
R
ASCE/SEI 705 Table
System
Seismic Detailing
Requirements
C
3.5
12.21 or 15.41
Ordinary steel moment frame
AISC 341
C
3
12.21
Structural steel systems not
specifically detailed for seismic
resistance
None
D or E
2.5
15.41
Steel OMF with permitted height
increase
AISC 341
(AISC Seismic)
D, E, or
F
1
15.41
Steel OMF with unlimited height
None
C15.5.3 Steel Storage Racks. The two approaches to the design of steel storage racks set forth by the standard are intended
to produce comparable results.
These recommendations address the concern that storage racks in warehousetype retail stores may pose a greater seismic risk
to the general public than exists in lowoccupancy warehouses or more conventional retail environments. Under normal
conditions, retail stores have a far higher occupant load than an ordinary warehouse of a comparable size. Failure of a
Figure C15.51 of Merchandise restrained by netting.
storage rack system in a retail environment is much more likely to cause personal injury than a similar failure in a storage
warehouse. To provide an appropriate level of additional safety in areas open to the public, an importance factor of 1.50 is
specified. Storage rack contents, while beyond the scope of the standard, may pose a potentially serious threat to life should
they fall from the shelves in an earthquake. It is recommended that restraints be provided, as shown in Figure C15.51, to
prevent the contents of rack shelving open to the general public from falling during strong ground shaking.
Figure C15.51 Merchandise restrained by netting.
C15.5.4 Electrical Power Generating Facilities. Electrical power plants closely resemble building structures, and their
performance in seismic events has been good. For reasons of mechanical performance, lateral drift of the structure must be
limited. The lateral bracing system of choice has been the concentrically braced frame. In the past, the height limits on
braced frames in particular have been an encumbrance to the design of large power generating facilities. Based on acceptable
past performance, Table 15.41 permits the use of CBRs with both lower R values and less restrictive height limits. This
option is particularly effective for boiler buildings that generally are 300 feet or more in height. A peculiarity of large boiler
buildings is the general practice of suspending the boiler from the roof structures; this results in an unusual mass distribution
as discussed in Section C15.1.3.
C15.5.5 Structural Towers for Tanks and Vessels. The requirements of this section apply to structural towers that are not
integral with the supported tank. Elevated water tanks designed in accordance with AWWA D10006 are not subject to
Section 15.5.5.
C15.5.6 Piers and Wharves. Current industry practice recognizes the distinct differences between the two categories of
piers and wharves described in the standard. Piers and wharves with public occupancy, described in Section 15.5.6.2, are
commonly treated as the “foundation” for buildings or buildinglike structures; design is performed using the standard, likely
under the jurisdiction of the local building official. Piers and wharves without occupancy by the general public are often
treated differently and are outside the scope of the standard; in many cases, these structures do not fall under the jurisdiction
of building officials, and design is performed using other industryaccepted approaches.
Design decisions associated with these structures often reflect economic considerations by both owners and local, regional, or
state jurisdictional entities with interest in commercial development. Where building officials have jurisdiction but lack
experience analyzing pier and wharf structures, reliance on other industryaccepted design approaches is common.
Where occupancy by the general public is not a consideration, seismic design of structures at major ports and marine
terminals often uses a performancebased approach, with criteria and methods that are very different from those used for
buildings, as provided in the standard. Design approaches most commonly used are generally consistent with the practices
and criteria described in the following documents:
1. Seismic Design Guidelines for Port Structures, Working Group No. 34 of the Maritime Navigation Commission
(PIANC/MarCom/WG34), A. A. Balkema, Lisse, Netherlands, 2001.
2. Seismic Criteria for California Marine Oil Terminals, Vol. 1 and Vol. 2, Technical Report TR2103SHR, Naval
Facilities Engineering Service Center, Ferritto, J., Dickenson, S., Priestley N., Werner, S., Taylor, C., Burke D., Seelig
W., and Kelly, S., Port Hueneme, CA, 1999.
3. Seismic Design and Retrofit of Bridges, Priestley, N.J.N., Siebel, F., and Calvi, G.M., New York, 1996.
4. Seismic Guidelines for Ports, by the Ports Committee of the Technical Council on Lifeline Earthquake Engineering,
ASCE/SEI, edited by Stuart D. Werner, Monograph No. 12, published by ASCE, Reston, Virginia, March 1998.
5. MOTEMS, 2005, “Marine Oil Terminal Engineering and Maintenance Standards”, 2001 Title 24, Part 2, California
Building Code, Chapter 31F, January 31, 2005.
These alternative approaches have been developed over a period of many years by working groups within the industry, and
they reflect the historical experience and performance characteristics of these structures, which are very different from those
of building structures.
The main emphasis of the performancebased design approach is to provide criteria and methods that depend on the
economic importance of a facility. Adherence to the performance criteria in the documents listed above does not seek to
provide uniform margins of collapse for all structures; their application is expected to provide at least as much inherent lifesafety
as for buildings designed using the standard. The reasons for the higher inherent level of lifesafety for these
structures include the following:
1. These structures have relatively infrequent occupancy, with few working personnel and very low density of personnel.
Most of these structures consist primarily of open area, with no enclosed structures that can collapse onto personnel.
Small control buildings on marine oil terminals or similar secondary structures are commonly designed in accordance
with the local building code.
2. These pier or wharf structures typically are constructed of reinforced concrete, prestressed concrete, or steel and are
highly redundant due to the large number of piles supporting a single wharf deck unit. Tests done at the University of
California at San Diego for the Port of Los Angeles have shown that very high ductilities (10 or more) can be achieved in
the design of these structures using practices currently employed in California ports.
3. Container cranes, loading arms, and other major structures or equipment on piers or wharves are specifically designed
not to collapse in an earthquake. Typically, additional piles and structural members are incorporated into the wharf or
pier specifically to support such items.
4. Experience has shown that seismic “failure” of wharf structures in zones of strong seismicity is indicated not by collapse
but by economically irreparable deformations of the piles. The wharf deck generally remains level or slightly tilting, but
has shifted out of position. Complete failure that could endanger lifesafety due to earthquake loading has never
occurred historically where the structure in the marine environment has been maintained properly.
5. The performancebased criteria of the listed documents address reparability of the structure, which is much more
stringent criteria than collapse prevention and results in a greater margin for lifesafety.
Lateral load design of these structures in low, or even moderate, seismic regions often is governed by other marine
conditions.
C15.6 GENERAL REQUIREMENTS FOR NONBUILDING STRUCTURES NOT SIMILAR TO BUILDINGS
Nonbuilding structures not similar to buildings exhibit behavior markedly different from that of building structures. Most of
these types of structures have reference documents that address their unique structural performance and behavior. The
ground motion in the standard requires appropriate translation to allow use with industry standards.
C15.6.1 EarthRetaining Structures. Section C11.8.3 presents commonly used approaches for the design of nonyielding
walls and yielding walls for bending, overturning, sliding, etc., taking into account the varying soil types, importance, and
site seismicity.
C15.6.2 Stacks and Chimneys. The design of stacks and chimneys to resist natural hazards generally is governed by wind
design considerations. The exceptions to this general rule involve locations with high seismicity, stacks and chimneys with
large elevated masses, and stacks and chimneys with unusual geometries. It is prudent to evaluate the effect of seismic loads
in all but those areas with the lowest seismicity. Although not specifically required, it is recommended that the special
seismic details required elsewhere in the standard be considered for application to stacks and chimneys.
Guyed steel stacks and chimneys generally are lightweight. As a result, the design loads due to natural hazards generally are
governed by wind. On occasion, large flares or other elevated masses located near the top may require indepth seismic
analysis. Although it does not specifically address seismic loading, Chapter 6 of Troitsky (1982) provides a methodology
appropriate for resolution of the seismic forces defined in the standard.
Figure C15.61 Showing Wall forces.
Hydrodynamic
force
Hydrodynamic
force
Wall
inertia
force
Wall
inertia
force
Wall
inertia
force
C15.6.4 Special Hydraulic Structures. The most common special hydraulic structures are baffle walls and weirs that are
used in water treatment and waste water treatment plants. Because there are openings in the walls, during normal operations
the fluid levels are equal on each side of the wall, exerting no net horizontal force. Sloshing during a seismic event can exert
large forces on the wall, as illustrated in Figure C15.61. The walls can fail unless they are designed properly to resist the
dynamic fluid forces.
C15.6.5 Secondary Containment Systems. This section reflects the judgment that designing all impoundment dikes for the
MCE ground motion when full and sizing all impoundment dikes for the sloshing wave is too conservative. Designing an
impoundment dike as full for the MCE assumes failure of the primary containment and occurrence of a significant
aftershock. Such significant aftershocks (of the same magnitude as the MCE ground motion) are rare and do not occur in all
locations. While explicit design for aftershocks is not a requirement of the standard, secondary containment must be
designed full for an aftershock to protect the general public. The use of twothirds of the MCE ground motion as the
magnitude of the design aftershock is supported by Bath’s Law, according to which the maximum expected aftershock
magnitude may be estimated to be 1.2 scale units below the main shock magnitude.
The risk assessment and risk management plan described in Section 1.5.2 are used to determine where the secondary
containment must be designed full for the MCE. The decision to design secondary containment for this more severe
condition should be based on the likelihood of a significant aftershock occurring at the particular site, considering the risk
posed to the general public by the release of hazardous material from the secondary containment.
Figure C15.61 Wall forces.
Secondary containment systems must be designed to contain the sloshing wave where the release of liquid would place the
general public at risk by exposing them to hazardous materials, by scouring of foundations of adjacent structures, or by
causing other damage to adjacent structures.
C15.6.6 Telecommunication Towers. Telecommunication towers support small masses, and their design generally is
governed by wind forces. Although telecommunication towers have a history of experiencing seismic events without failure
or significant damage, seismic design in accordance with the standard is required.
Typically bracing elements bolt directly (without gusset plates) to the tower legs, which consist of pipes or bent plates in a
triangular plan configuration.
C15.7 TANKS AND VESSELS
C15.7.1 General. Methods for seismic design of tanks, currently adopted by a number of reference documents, have
evolved from earlier analytical work by Jacobsen, Housner, Veletsos, Haroun, and others. The procedures used to design
flatbottom storage tanks and liquid containers are based on the work of Housner, Wozniak, and Mitchell. The reference
documents for tanks and vessels have specific requirements to safeguard against catastrophic failure of the primary structure
based on observed behavior in seismic events since the 1930s. Other methods of analysis, using flexible shell models, have
been proposed but at present are beyond the scope of the standard.
The industryaccepted design methods employ three basic steps:
1. Dynamic modeling of the structure and its contents. When a liquidfilled tank is subjected to ground acceleration, the
lower portion of the contained liquid, identified as the impulsive component of mass, Wi, acts as if it were a solid mass
rigidly attached to the tank wall. As this mass accelerates, it exerts a horizontal force, P
(C15.71)
An examination of those terms as used in the different references reveals the following:
i, on the wall; this force is
directly proportional to the maximum acceleration of the tank base. This force is superimposed on the inertia force of
the accelerating wall itself, Ps. Under the influence of the same ground acceleration, the upper portion of the contained
liquid responds as if it were a solid mass flexibly attached to the tank wall. This portion, which oscillates at its own
natural frequency, is identified as the convective component, Wc, and exerts a horizontal force, Pc, on the wall. The
convective component oscillations are characterized by sloshing whereby the liquid surface rises above the static level on
one side of the tank and drops below that level on the other side.
2. Determination of the period of vibration, Ti, of the tank structure and the impulsive component; and determination of the
natural period of oscillation (sloshing), Tc, of the convective component.
3. Selection of the design response spectrum. The response spectrum may be sitespecific or it may be constructed on the
basis of seismic coefficients given in national codes and standards. Once the design response spectrum is constructed,
the spectral accelerations corresponding to Ti and Tc are obtained and are used to calculate the dynamic forces Pi, Ps,
and Pc.
Detailed guidelines for the seismic design of circular tanks, incorporating these concepts to varying degrees, have been the
province of at least four industry reference documents: AWWA D100 for welded steel tanks (since 1964); API 650 for
petroleum storage tanks; AWWA D110 for prestressed, wirewrapped tanks (since 1986); and AWWA D115 for prestressed
concrete tanks stressed with tendons (since 1995). In addition, API 650 and API 620 contain provisions for petroleum,
petrochemical, and cryogenic storage tanks. The detail and rigor of analysis prescribed in these documents have evolved
from a semistatic approach in the early editions to a more rigorous approach at present, reflecting the need to include the
dynamic properties of these structures.
The requirements in Section 15.7 are intended to link the latest procedures for determining designlevel seismic loads with
the allowable stress design procedures based on the methods in the standard. These requirements, which in many cases
identify specific substitutions to be made in the design equations of the reference documents, will assist users of the standard
in making consistent interpretations.
ACI has published ACI 350.301, “Seismic Design of LiquidContaining Concrete Structures.” This document, which
addresses all types of concrete tanks (prestressed and nonprestressed, circular and rectilinear), has provisions that are
consistent with the seismic criteria of the 2000 Provisions. The document serves as both a practical “hoto” loading reference
and a guide to supplement application of ACI 318 Chapter 21.
C15.7.2 Design Basis. In the case of the seismic design of nonbuilding structures, standardization requires adjustments to
industry reference documents to minimize existing inconsistencies among them, while recognizing that structures designed
and built over the years in accordance with these documents have performed well in earthquakes of varying severity. Of the
inconsistencies among reference documents, the ones most important to seismic design relate to the base shear equation. The
traditional base shear takes the following form:
Equation
w
V ZIS CW
R
=
1. ZS: The seismic zone coefficient, Z, has been rather consistent among all the documents since it usually has been
obtained from the seismic zone designations and maps in the model building codes. On the other hand, the soil profile
coefficient, S, does vary from one document to another. In some documents these two terms are combined.
2. I: The importance factor, I, has varied from one document to another, but this variation is unavoidable and
understandable owing to the multitude of uses and degrees of importance of tanks and vessels.
3. C: The coefficient C represents the dynamic amplification factor that defines the shape of the design response spectrum
for any given ground acceleration. Since C is primarily a function of the frequency of vibration, inconsistencies in its
derivation from one document to another stem from at least two sources: differences in the equations for the
determination of the natural frequency of vibration, and differences in the equation for the coefficient itself. (For
example, for the shell/impulsive liquid component of lateral force, the steel tank documents use a constant design
spectral acceleration [constant C] that is independent of the “impulsive” period, T.) In addition, the value of C will vary
depending on the damping ratio assumed for the vibrating structure (usually between 2 percent and 7 percent of critical).
4. Where a sitespecific response spectrum is available, calculation of the coefficient C is not necessary except in the case
of the convective component (coefficient Cc) which is assumed to oscillate with 0.5 percent of critical damping and
whose period of oscillation is usually long (greater than 2.5 seconds). Since sitespecific spectra are usually constructed
for high damping values (3 percent to 7 percent of critical) and since the sitespecific spectral profile may not be well
defined in the longperiod range, an equation for Cc applicable to a 0.5 percent damping ratio is necessary in order to
calculate the convective component of the seismic force.
5. Rw: The response modification factor, Rw, is perhaps the most difficult to quantify, for a number of reasons. While Rw
is a compound coefficient that is supposed to reflect the ductility, energydissipating capacity, and redundancy of the
structure, it is also influenced by serviceability considerations, particularly in the case of liquidcontaining structures.
In the standard the base shear equation for most structures has been reduced to V = CsW, where the seismic response
coefficient, Cs, replaces the product ZSC/Rw. Cs is determined from the design spectral response acceleration parameters SDS
and SD1 (at short periods and at a period of 1, respectively) which, in turn, are obtained from the mapped MCE spectral
accelerations Ss and S1. As in the case of the prevailing industry reference documents, where a sitespecific response
spectrum is available, Cs is replaced by the actual values of that spectrum.
The standard contains several bridging equations, each designed to allow proper application of the design criteria of a
particular reference document in the context of the standard. These bridging equations associated with particular types of
liquidcontaining structures and the corresponding reference documents are discussed below. Calculation of the periods of
vibration of the impulsive and convective components is in accordance with the reference documents, and the detailed
resistance and allowable stresses for structural elements of each industry structure are unchanged, except where new
information has led to additional requirements.
It is expected that the bridging equations of Sections 15.7.7.3 and 15.7.10.7 will be eliminated as the relevant reference
documents are updated to conform to the standard. The bridging equations previously provided for AWWA D100 and API
650 already have been eliminated as a result of updates of these documents.
C15.7.3 Strength and Ductility. As is the case for building structures, ductility and redundancy in the lateral support
systems for tanks and vessels are desirable and necessary for good seismic performance. Tanks and vessels are not highly
redundant structural systems, and therefore ductile materials and welldesigned connection details are needed to increase the
capacity of the vessel to absorb more energy without failure. The critical performance of many tanks and vessels is governed
by shell stability requirements rather than by yielding of the structural elements. For example, contrary to building
structures, ductile stretching of anchor bolts is a desirable energy absorption component where tanks and vessels are
anchored. The performance of crossbraced towers is highly dependent on the ability of the horizontal compression struts
and connection details to develop fully the tension yielding in the rods. In such cases, it is also important to preclude both
premature failure in the threaded portion of the connection and failure of the connection of the rod to the column prior to
yielding of the rod.
C15.7.4 Flexibility of Piping Attachments. Poor performance of piping connections (tank leakage and damage) due to
seismic deformations is a primary weakness observed in recent seismic events. While commonly used piping connections
can impart mechanical loads to the tank shell, proper design in seismic areas results in only negligible mechanical loads on
tank connections subject to the displacements shown in Table 15.71. API 650 treats the values shown in Table 15.71 as
allowable stress based values and therefore requires that these values be multiplied by 1.4 where strengthbased capacity
values are required for design.
In addition, interconnected equipment, walkways, and bridging between multiple tanks must be designed to resist the loads
and accommodate the displacements imposed by seismic forces. Unless connected tanks and vessels are founded on a
common rigid foundation, the calculated differential movements must be assumed to be out of phase.
C15.7.5 Anchorage. Many steel tanks can be designed without anchors by using annular plate detailing in accordance with
reference documents. Where tanks must be anchored due to overturning potential, proper anchorage design will provide both
a shell attachment and an embedment detail that will yield the bolt without tearing the shell or pulling the bolt out of the
foundation. Properly designed anchored tanks have greater reserve strength to resist seismic overload than do unanchored
tanks.
Where anchor bolts and attachments are misaligned such that the anchor nut or washer does not bear evenly on the
attachment, additional bending stresses in threaded areas may cause premature failure before anchor yielding.
C15.7.6 GroundSupported Storage Tanks for Liquids
C15.7.6.1 General. The response of ground storage tanks to earthquakes is well documented by Housner, Mitchell and
Wozniak, Veletsos, and others. Unlike building structures, the structural response of these tanks is influenced strongly by the
fluidstructure interaction. Fluidstructure interaction forces are categorized as sloshing (convective) and rigid (impulsive)
forces. The proportion of these forces depends on the geometry (heighttodiameter ratio) of the tank. API 650, API 620,
AWWA D100, AWWA D110, AWWA D115, and ACI 350.3 provide the data necessary to determine the relative masses
and moments for each of these contributions.
The standard requires that these structures be designed in accordance with the prevailing reference documents, except that the
height of the sloshing wave, ds
, must be calculated using Equations 15.713. Note that API 650 and AWWA D100 include
this requirement in their latest editions.
Equations 15.710 and 15.711 provide the spectral acceleration of the sloshing liquid for the constantvelocity and constantdisplacement
regions of the response spectrum, respectively. The 1.5 factor in these equations is an adjustment for
0.5 percent damping.
Smalldiameter tanks and vessels are more susceptible to overturning and vertical buckling. As a general rule, a greater ratio
of H/D produces lower resistance to vertical buckling. Where H/D is greater than 2, overturning approaches “rigid mass”
behavior (the sloshing mass is small). Largediameter tanks may be governed by additional hydrodynamic hoop stresses in
the middle regions of the shell.
The impulsive period (the natural period of the tank components and the impulsive component of the liquid) is typically in
the 0.25 to 0.6 second range. Many methods are available for calculating the impulsive period. The Veletsos flexibleshell
method is commonly used by many tank designers. For example, see Velestos (1974) and Malhotra, Wenk, and Wieland
(2000).
C15.7.6.1.1 Distribution of Hydrodynamic and Inertia Forces. Most of the reference documents for tanks define reaction
loads at the base of shellfoundation interface, without indicating the distribution of loads on the shell as a function of height.
ACI 350.3 specifies the vertical and horizontal distribution of such loads.
The overturning moment at the base of the shell in the industry reference documents is only the portion of the moment that is
transferred to the shell. The total overturning moment also includes the variation in bottom pressure, which is an important
consideration for design of pile caps, slabs, or other support elements that must resist the total overturning moment. Wozniak
and Mitchell (1978) and TID 7024 (1963) provide additional information.
C15.7.6.1.2 Sloshing. In past earthquakes, sloshing contents in ground storage tanks has caused both leakage and noncatastrophic
damage to the roof and internal components. Even this limited damage, and the associated costs and
inconvenience, can be significantly mitigated where the following items are considered:
1. Effective masses and hydrodynamic forces in the container.
2. Impulsive and pressure loads at
a. The sloshing zone (that is, the upper shell and edge of the roof system),
b. The internal supports (such as roof support columns and traysupports), and
c. The internal equipment (such as distribution rings, access tubes, pump wells, and risers).
3. Freeboard (which depends on the sloshing wave height).
A minimum freeboard of 0.7ds is recommended for economic considerations but is not required.
Tanks and vessels storing biologically or environmentally benign materials typically do not require freeboard to protect the
public health and safety. However, providing freeboard in areas of frequent seismic occurrence for vessels normally operated
at or near top capacity may lessen damage (and the cost of subsequent repairs) to the roof and upper container.
The sloshing wave height specified in Section 15.7.6.1.2 is based on the design earthquake defined in the standard. For
economic reasons, freeboard for tanks assigned to Occupancy Category I, II, or III may be calculated using a fixed value of
TL equal to 4 seconds (as indicated in Section 15.7.6.1, Note d) but using the appropriate importance factor taken from Table
11.51. Due to lifesafety concerns, freeboard for tanks assigned to Occupancy Category IV must be based on the mapped
value of TL. Because use of the mapped value of TL results in the theoretical maximum value of freeboard, the calculation of
freeboard in the case of Occupancy Category IV tanks is based on an importance factor equal to 1.0 (as indicated in Section
15.7.6.1, Note c).
If the freeboard provided is less than the computed sloshing height, ds, the sloshing liquid will impinge on the roof in the
vicinity of the rooftowall joint, subjecting it to a hydrodynamic force. This force may be approximated by considering the
sloshing wave as a hypothetical static liquid column having a height, ds. The pressure exerted at any point along the roof at a
distance ys above the atrest surface of the stored liquid may be assumed equal to the hydrostatic pressure exerted by the
hypothetical liquid column at a distance ds – ys from the top of that column. A better approximation of the pressure exerted
on the roof is found in Malhotra (2005 and 2006).
Another effect of a lessthanfull freeboard is that the restricted convective (sloshing) mass “converts” into an impulsive mass
thus increasing the impulsive forces. This effect should be taken into account in the tank design. A method for converting the
restricted convective mass into an impulsive mass is found in Malhotra (2005 and 2006). It is recommended that sufficient
freeboard to accommodate the full sloshing height be provided wherever possible.
C15.7.6.1.4 Internal Components. Wozniak and Mitchell (1978) provides a recognized analysis method for determining
the lateral loads on internal components due to sloshing liquid.
C15.7.6.1.5 Sliding Resistance. Historically, steel groundsupported tanks full of product have not slid off foundations. A
few unanchored, empty tanks or bulk storage tanks without steel bottoms have moved laterally during earthquake ground
shaking. In most cases, these tanks may be returned to their proper locations. Resistance to sliding is obtained from the
frictional resistance between the steel bottom and the sand cushion on which bottoms are placed. Because tank bottoms
usually are crowned upward toward the tank center and are constructed of overlapping, filletwelded, individual steel plates
(resulting in a rough bottom), it is reasonably conservative to take the ultimate coefficient of friction as 0.70 (AISC, 1986),
and therefore a value of tan 30o (= 0.577) is used in design. The value of 30o represents the internal angle of friction of sand.
The vertical weight of the tank and contents, as reduced by the component of vertical acceleration, provides the net vertical
load. An orthogonal combination of vertical and horizontal seismic forces, following the procedure in Section 12.5.3, may be
used.
C15.7.6.1.6 Local Shear Transfer. The transfer of seismic shear from the roof to the shell and from the shell to the base is
accomplished by a combination of membrane shear and radial shear in the wall of the tank. For steel tanks, the radial (outofplane)
seismic shear is very small and usually is neglected; thus, the shear is assumed to be resisted totally by membrane (inplane)
shear. For concrete walls and shells, which have a greater radial shear stiffness, the shear transfer may be shared. The
ACI 350.3 commentary provides further discussion.
C15.7.6.1.7 Pressure Stability. Internal pressure may increase the critical buckling capacity of a shell. Provision to include
pressure stability in determining the buckling resistance of the shell for overturning loads is included in AWWA D100.
Recent testing on conical and cylindrical shells with internal pressure yielded a design methodology for resisting permanent
loads in addition to temporary wind and seismic loads. See Miller, Meier, and Czaska (1997).
C15.7.6.1.8 Shell Support. Anchored steel tanks should be shimmed and grouted to provide proper support for the shell
and to reduce impact on the anchor bolts under reversible loads. The high bearing pressures on the toe of the tank shell may
cause inelastic deformations in compressible material (such as fiberboard), creating a gap between the anchor and the
attachment. As the load reverses, the bolt is no longer snug and an impact of the attachment on the anchor can occur. Grout
is a structural element and should be installed and inspected as an important part of the vertical and lateralforceresisting
system.
C15.7.6.1.9 Repair, Alteration, or Reconstruction. During their service life, storage tanks are frequently repaired,
modified, or relocated. Repairs often are related to corrosion, improper operation, or overload from wind or seismic events.
Modifications are made for changes in service, updates to safety equipment for changing regulations, or installation of
additional process piping connections. It is imperative these repairs and modifications be designed and implemented
properly to maintain the structural integrity of the tank or vessel for seismic loads as well as the design operating loads.
The petroleum steel tank industry has developed specific guidelines in API 653 that are statutory requirements in some states.
It is recommended that the provisions of API 653 also be applied to other liquid storage tanks (water, wastewater, chemical,
etc.) as it relates to repairs, modifications, or relocation that affects the pressure boundary or lateral forceresisting system of
the tank or vessel.
C15.7.7 Water Storage and Water Treatment Tanks and Vessels. The AWWA design requirements for groundsupported
steel water storage structures use allowable stress design procedures that conform to the requirements of the
standard.
C15.7.8 Petrochemical and Industrial Tanks and Vessels Storing Liquids
C15.7.8.1 Welded Steel. The American Petroleum Institute (API) uses an allowable stress design procedure that conforms
to the requirements of the standard.
The most common damage to tanks observed during past earthquakes includes the following:
1. Buckling of the tank shell near the base due to excessive axial membrane forces. This buckling damage is usually
evident as “elephant foot” buckles a short distance above the base or as diamondshaped buckles in the lower ring.
Buckling of the upper ring also has been observed.
2. Damage to the roof due to impingement on the underside of the roof of sloshing liquid with insufficient freeboard.
3. Failure of piping or other attachments that are overly restrained.
4. Foundation failures.
Other than the above damage, the seismic performance of floating roofs during earthquakes has generally been good, with
damage usually confined to the rim seals, gage poles, and ladders. However, floating roofs have sunk in some earthquakes
due to lack of adequate freeboard or the proper buoyancy and strength required by API 650. Similarly the performance of
open tops with top wind girder stiffeners designed per API 650 has been generally good.
C15.7.8.2 Bolted Steel. Bolted steel tanks are often used for temporary functions. Where use is temporary, it may be
acceptable to the jurisdictional authority to design bolted steel tanks for no seismic loads or for reduced seismic loads based
on a reduced return period. For such reduced loads based on reduced exposure time, the owner should include a signed
removal contract with the fixed removal date as part of the submittal to the authority having jurisdiction.
C15.7.9 GroundSupported Storage Tanks for Granular Materials
C15.7.9.1 General. The response of a groundsupported storage tank storing granular materials to a seismic event is highly
dependent on its heighttodiameter (H/D) ratio and the characteristics of the stored product. The effects of intergranular
friction are described in more detail in C15.7.9.3.1 (increased lateral pressure), C15.7.9.3.2 (effective mass), and C15.7.9.3.3
(effective density).
Longterm increases in shell hoop tension due to temperature changes after the product has been compacted also must be
included in the analysis of the shell; Anderson (1966) provides a suitable method.
C15.7.9.2 Lateral Force Determination. Seismic forces acting on groundsupported liquid storage tanks are divided
between impulsive and convective (sloshing) components. However, in a groundsupported storage tank for granular
materials, all seismic forces are of the impulsive type and relate to the period of the storage tank itself. Due to the relatively
short period of a tank shell, the response is normally in the constant acceleration region of the response spectrum, which
relates to S
(C15.72)
where V, S
DS. Therefore, the seismic base shear is calculated as follows:
DS, I, and R have been previously defined, and WEffective is the gross weight of the stored product multiplied by an
effective mass factor and an effective density factor, as described in Sections C15.7.9.3.2 and C15.7.9.3.3, plus the dead
weight of the tank. Unless substantiated by testing, it is recommended that the product of the effective mass factor and the
effective density factor be taken as no less than 0.5 due to the limited test data and the highly variable properties of the stored
product.
C15.7.9.3 Force Distribution to Shell and Foundation
C15.7.9.3.1 Increased Lateral Pressure. In a groundsupported tank storing granular materials, increased lateral pressures
develop as a result of rigid body forces that are proportional to ground acceleration. Information concerning design for such
pressure is scarce. Trahair et al. (1983) describes both a very simple, conservative method and a very difficult, analytical
method using failure wedges based on the MononobeOkabe modifications of the classical Coulomb method.
C15.7.9.3.2 Effective Mass. For groundsupported tanks storing granular materials, much of the lateral seismic load can be
transferred directly into the foundation, via intergranular shear, before it can reach the tank shell. The effective mass that
loads the tank shell is highly dependent on the H/D ratio of the tank and the characteristics of the stored product.
Quantitative information concerning this effect is scarce, but Trahair et al. (1983) describes a very simple, conservative
method to determine the effective mass. That method presents reductions in effective mass, which may be significant, for
H/D ratios less than 2. This effect is absent for elevated tanks.
C15.7.9.3.3 Effective Density. Granular material stored in tanks (both groundsupported and elevated) does not behave as a
solid mass. Energy loss through intergranular movement and graintograin friction in the stored material effectively reduces
the mass subject to horizontal acceleration. This effect may be quantified by an effective density factor less than 1.0.
Based on Chandrasekaran and Jain (1968) and on shaketable tests reported in Chandrasekaran et al. (1968), ACI 313
recommends an effective density factor of not less than 0.8 for most granular materials. According to Chandrasekaran and
Jain (1968), an effective density factor of 0.9 is more appropriate for materials with high moduli of elasticity, such as
aggregates and metal ores. Equation
Effective
DS W
I
R
V S
..
.
..
.
=
C15.7.9.3.4 Lateral Sliding. Most groundsupported steel storage tanks for granular materials rest on a base ring and do not
have a steel bottom. To resist seismic base shear, a partial bottom or annular plate is used in combination with anchor bolts
or a curb angle. An annular plate can be used alone to resist the seismic base shear through friction between the plate and the
foundation, in which case the friction limits of Section 15.7.6.1.5 apply. The curb angle detail serves to keep the base of the
shell round while allowing it to move and flex under seismic load. Various base details are shown in Figure 13 of Kaups and
Lieb (1985).
C15.7.9.3.5 Combined Anchorage Systems. This section is intended to apply to combined anchorage systems that share
loads based on their relative stiffnesses, and not to systems where sliding is resisted completely by one system (such as a steel
annular plate) and overturning is resisted completely by another system (such as anchor bolts).
C15.7.10 Elevated Tanks and Vessels for Liquids and Granular Materials
C15.7.10.1 General. The three basic lateral loadresisting systems for elevated water tanks are defined by their support
structure:
1. Multileg braced steel tanks (trussed towers, as shown in Figure C15.71),
2. Smalldiameter singlepedestal steel tanks (cantilever columns, as shown in Figure C15.72), and
3. Largediameter singlepedestal tanks of steel or concrete construction (loadbearing shear walls, as shown in Figure
C15.73).
Unbraced multileg tanks are uncommon. These types of tanks differ in their behavior, redundancy, and resistance to
overload. Multileg and smalldiameter pedestal tanks have longer fundamental periods (typically greater than 2 seconds)
than the shear wall type tanks (typically less than 2 seconds). The lateral load failure mechanisms usually are brace failure
for multileg tanks, compression buckling for smalldiameter steel tanks, compression or shear buckling for largediameter
steel tanks, and shear failure for largediameter concrete tanks. Connection, welding, and reinforcement details require
careful attention in order to mobilize the full strength of these structures. To provide a greater margin of safety, R factors
used with elevated tanks typically are less than those for other comparable lateral loadresisting systems.
C15.7.10.4 Transfer of Lateral Forces into Support Tower. The vertical loads and shears transferred at the base of a tank
or vessel supported by grillage or beams typically vary around the base due to the relative stiffness of the supports,
settlements, and variations in construction. Such variations must be considered in the design for vertical and horizontal
loads.
C15.7.10.5 Evaluation of Structures Sensitive to Buckling Failure. Nonbuilding structures with little structural
redundancy for lateral loads may exhibit total failure when loaded only slightly beyond the design loads. Tanks and vessels
supported on shell skirts or pedestals that are governed by buckling require evaluation for this critical condition.
The design spectral response acceleration, Sa, used in this evaluation includes site factors. The I/R coefficient is taken as 1.0
for this critical check. The structural capacity of the shell is taken as the critical buckling strength (that is, the factor of safety
is 1.0). Vertical and orthogonal combinations need not be considered for this evaluation, since the probability of peak values
occurring simultaneously is very low.
While the standard requires this evaluation only for structures assigned to Occupancy Category IV, it may be applied to any
bucklingsensitive structure. Where such optional evaluations are performed, an R value of 2 or 3 can be used. In most
cases, the design of the structure will be governed by this additional evaluation.
C15.7.10.7 Concrete Pedestal (Composite) Tanks. A composite elevated waterstorage tank is comprised of a welded
steel tank for watertight containment, a single pedestal concrete support structure, a foundation, and accessories. The lateral
loadresisting system is a loadbearing concrete shear wall. Since the seismic provisions in ACI 371R98 are based on an
older edition of ASCE/SEI 7, appropriate bridging equations are provided in Section 15.7.10.7.
C15.7.11 Boilers and Pressure Vessels. The support system for boilers and pressure vessels must be designed for the
seismic forces and displacements presented in the standard. Such design must include consideration of the support, the
attachment of the support to the vessel (even if “integral”), and the body of the vessel itself, which is subject to local stresses
imposed by the support connection.
C15.7.12 Liquid and Gas Spheres. The commentary in Section C15.7.11 also applies to liquid and gas spheres.
C15.7.13 Refrigerated Gas Liquid Storage Tanks and Vessels. Some refrigerated storage tanks and vessels, such as those
storing LNG, are required to be designed for ground motions and performance goals in excess of those found in the standard,
so such structures are outside the scope of the standard. All other welded steel refrigerated storage tanks and vessels must be
designed in accordance with the requirements of the standard, the requirements of API 620, and the seismic requirements of
API 650. Note that the seismic requirements of API 620 (10th Edition, Addendum 1) are note used as they are inconsistent
with the requirements of the standard.
Figure C15.71 Image of Multileg braced steel tank.
Figure C15.72 Image of Smalldiameter singlepedestal steel tank.
C15.7.14 Horizontal, Saddle Supported Vessels for Liquid or Vapor Storage. Past practice has been to assume that a
horizontal, saddle supported vessel (including its contents) behaves as a rigid structure (with natural period, T, less than 0.06
seconds). For this situation, seismic forces would be determined using the requirements of Section 15.4.2. For large
horizontal, saddlesupported vessels (lengthtodiameter ratio of 6 or more), this assumption can be unconservative, so
Section 15.7.14.3 requires that the natural period be determined assuming the vessel to be a simply supported beam.
Figure C15.71 Multileg braced steel tank.
Figure C15.72 Smalldiameter singlepedestal steel tank.
Figure C15.73 Image of Largediameter singlepedestal tank.
Figure C15.73 Image of Largediameter singlepedestal tank.
(a) Steel
(b) Concrete
Figure C15.73 Largediameter singlepedestal tank.
REFERENCES
American Institute of Steel Construction, Inc. 1986. Load and Resistance Factor Design Specification for Structural Steel
Buildings.
Anderson, P. F. 1966. “Temperature Stresses in Steel GrainStorage Tanks,” ASCE Civil Engineering (January):74.
Chandrasekaran, A. R., and P. C. Jain. 1968. “Effective Live Load of Storage Materials Under Dynamic Conditions,” Indian
Concrete Journal (Bombay), 42(9):364365.
Chandrasekaran, A. R., S. S. Saini, and I. C. Jhamb. 1968. “Live Load Effects on Dynamic Behavior of Structures,” Journal
of the Institution of Engineers (India), 48:850859.
Kaups, Taavi, and John M. Lieb. 1985. A Practical Guide for the Design of Quality Bulk Storage Bins and Silos. Chicago
Bridge & Iron Company, Plainfield, Illinois.
Malhotra, P. K., T. Wenk, and M. Wieland. 2000. “Simple Procedure for Seismic Analysis of LiquidStorage Tanks,”
Journal of Structural Engineering International, IABSE, 10(3):197201.
Malhotra, P. K. 2005. “Sloshing Loads in LiquidStorage Tanks with Insufficient Freeboard,” Earthquake Spectra,
21(4):11851192.
Malhotra, P. K. 2006. “Earthquake Induced Sloshing in Cone and Dome Roof Tanks with Insufficient Freeboard.” Journal
of Structural Engineering International, IABSE, 16(3):222225.
Miller, C. D., S. W. Meier, and W. J. Czaska. 1997. “Effects of Internal Pressure on Axial Compressive Strength of
Cylinders and Cones,” paper presented at Structural Stability Research Council Annual Technical Meeting, June.
Department of Energy. 1963. Nuclear Reactors and Earthquakes, TID 7024. Department of Energy, Atomic Energy
Commission, Division of Reactor Development, Washington, D.C
Trahair, M. S., A. Abel, P. Ansourian, H. M. Irvine, and J. M. Rotter. 1983. Structural Design of Steel Bins for Bulk Solids.
Australian Institute of Steel Construction, Ltd., Sydney.
Troitsky, M. S. 1982. Tubular Steel Structures – Theory and Design. The James F. Lincoln Arc Welding Foundation.
Veletsos, A. S. 1974. “Seismic Effects in Flexible LiquidStorage Tanks,” in Proceedings of the Fifth World Conference on
Earthquake Engineering, Rome, Italy, pp. 630639.
Wozniak, R. S., and W. W. Mitchell. 1978. “Basis of Seismic Design Provisions for Welded Steel Oil Storage Tanks,”
presented at the Session on Advances in Storage Tank Design, American Petroleum Institute, Refining, 43rd Midyear
Meeting, Toronto, Canada, May 9.
Page intentionally left blank.
COMMENTARY TO CHAPTER 16,
SEISMIC RESPONSE HISTORY PROCEDURES
C16.1 LINEAR RESPONSE HISTORY PROCEDURE
The standard does not require the use of linear response history analysis. However, the use of such analysis may be useful in
validation of the results of the analysis methods presented in Chapter 12, or as a step in a series of analyses that culminate in
a nonlinear response history analysis. While not commonly used in the past to design typical structures, this technique is
seeing increased use in the design of some structures including structures that are neither damped nor base isolated.
The purpose of the linear response history procedure is to determine design forces for structural components and to compute
displacements and story drifts, which must be within the limits specified by Table 12.121. In this sense, the linear response
history procedure shares the forcebased philosophy of the Equivalent Lateral Force (ELF) procedure and the Modal
Response Spectrum (MRS) analysis procedure (both of which are specified in Chapter 12). Response history analysis offers
several advantages over modal response spectrum analysis: it is more accurate mathematically, signs of response quantities
(such as tension or compression in a brace) are not lost as a result of the combination of modal responses, and story drifts are
computed more accurately. The principal disadvantages of response history analysis are the need to select and scale an
appropriate suite of ground motions, and the necessity to perform analysis for several (usually seven) such motions. See
Section C16.1.3 for discussion of ground motion selection and scaling techniques.
C16.1.1 Analysis Requirements. In response history analysis, the seismic hazard is characterized by a number of ground
acceleration records. Using these records and a detailed mathematical model of the structure, nodal displacements and
component forces are computed, stepbystep, by integration of the equations of motion. Two basic approaches for solving
the equations may be used. In the first approach, called direct analysis, all the equilibrium equations for the entire system are
solved simultaneously in each step. The number of equations solved equals the number of degrees of freedom in the
structure.
In the second approach, called modal analysis, the equilibrium equations are transformed, by change of coordinates, into a
number of singledegreeoffreedom (SDOF) systems. The maximum number of SDOF systems that can be formed is equal
to the number of mass degrees of freedom in the structure. The SDOF equations are solved individually in time, and then the
computed displacement histories are transformed back to the original coordinates and superimposed to obtain the system
response history. The transformation of coordinates in the modal analysis approach is usually based on the undamped natural
mode shapes of the structure. Other bases, such as a set of orthogonal loaddependent Ritz vectors, may be preferable in
certain cases (Wilson et al., 1982).
Where modal analysis uses the full set of mode shapes and the damping ratios in each mode are identical to those obtained
from the equations of motion used in the direct analysis, the two approaches produce identical results. A distinct advantage
of the modal analysis approach is that a limited number of modes may be used to produce reasonably accurate results. While
some accuracy is sacrificed where fewer modes are used, the computer resources required to perform the analysis are
significantly less than those required for direct analysis. The number of modes required for a “reasonably” accurate analysis
is discussed in Section C12.9.1.
C16.1.2 Modeling. The mathematical model used for linear response history analysis is usually identical to that used for
modal response spectrum analysis, and it often reflects a preliminary design developed using the ELF procedure. The main
modeling difference between response history analysis and modal response spectrum analysis is that the inherent damping
(taken as 5 percent of critical) is included in the design response spectrum for response spectrum analysis, while it must be
assigned explicitly for response history analysis.
In the modal analysis approach to response history analysis, damping is simply assigned to each mode that is included in the
response (Wilson and Penzien, 1970). Although not specified in the standard, the damping used for each mode should be 5
percent of critical for consistency with the design response spectrum.
Direct response history analysis requires an explicit damping matrix. However, such a matrix cannot be formed from first
principles; it is common to use a damping matrix that is proportional to the mass, the stiffness, or a linear combination of the
two:
(C16.11)
where C is the damping matrix, M is the mass matrix, K is the stiffness matrix, and a and ß are scalar constants of
proportionality. Such damping is often referred to as Rayleigh damping. Equation
C =aM +ß K
Figure C16.11 Example of Rayleigh damping.
mass proportional
stiffness
proportional
total
.a
.b
0.00
0.05
0.10
0 10 20 30
Damping ratio, .
Frequency, . (radians/second)
The proportionality constants are determined as follows:
(C16.12)
where and are the desired damping ratios at any two system circular frequencies, and , where > . It is
common, but not necessary, for the two specified frequencies to correspond to two of the system’s lower natural frequencies
(such as the first and third mode frequencies).
If both damping values are the same (. = .
(C16.13)
The advantage of Rayleigh damping is that it is simple to implement because all the analyst has to do is to specify the two
proportionality constants a and ß, and these can be established using Equation C16.12 given the two desired damping ratios
and corresponding frequencies. The disadvantage is that the damping ratios increase with frequency and may cause the
higher mode contributions to response to be overdamped. This effect is shown in Figure C16.11, where the damping ratios
. have been set at 0.05 at frequencies of 4.2 and 12.5 radians per second. The damping at all other frequencies is given by
the curve marked “Total”. For frequencies above approximately 32 radians per second, the damping is greater than 10
percent of critical and may be excessive.
a = .b), which is usually the case, the mass and stiffness proportionality constants
may be determined as follows: Equation
1 1/
2
1/
a a a
b b b
a . . .
ß . . .
 . . . . . .
. . = . . . .
. . . . . . Equation
a . Equation
b . Equation
a . Equation
b . Equation
b . Equation
a . Equation
2
2
a b
a b
a b
. .
a .
. .
ß .
. .
=
+
=
+
Figure C16.11 Example of Rayleigh damping.
C16.1.3 Ground Motion. One of the most demanding aspects of response history analysis is the selection and scaling of an
appropriate suite of ground motions (Anderson and Bertero, 1987). It is considered appropriate to select records that have
magnitudes, fault distances, source mechanisms, and soil conditions that are characteristic of the site. This poses quite a
challenge even for sites in the western United States, where numerous records from largemagnitude earthquakes are
available; it is virtually impossible in the central and eastern United States, where there are no recorded ground motions from
largemagnitude events. (The web site for the Pacific Earthquake Engineering Research Center (PEER) provides a large
number of ground motion acceleration records that may be used in response history analysis. In addition to the ground
motions, the PEER site provides detailed background information on the source characteristics of the ground motions and on
the instrument and site characteristics of the particular station that recorded the acceleration record.)
Because of the scarcity of available recorded motions, use of simulated ground motions is permitted. To this end, available
records may be modified for site distance and soil conditions. Such modification is considered part of the ground motion
selection.
The standard requires that at least three ground motions (or ground motion pairs, in the case of threedimensional analysis) be
used, and it provides an incentive for using at least seven motions (as discussed in Section C16.1.4).
The scaling technique specified in Sections 16.1.3.1 and 16.1.3.2 is one of several that have been proposed. See Shome and
Cornell (1998), Shome et al. (1998), Somerville et al. (1998), Mehrain and Naiem (2003), and Iervolino and Cornell (2005)
for background on ground motion selection and scaling. (Applied Technology Council (ATC) Projects 58 and 63 are also
investigating scaling techniques.)
C16.1.3.1 TwoDimensional Analysis. This scaling method begins with ground motions that have been selected (and
modified as necessary) to have magnitude, distance, and site conditions compatible with the maximum considered
earthquake. The 5 percent damped pseudoacceleration response spectra for these records are scaled for consistency with the
design ground motion spectrum shown in Figure 11.41. For twodimensional analysis, the ground motion spectra must be
scaled such that the average of the spectra is not less than the design spectrum in the period range from 0.2T to 1.5T, where T
is the fundamental period of vibration of the structure being designed. The short period of the range (0.2T) is set to capture
higher mode response, and the long period of the range (1.5T) is set to allow for period lengthening that would be associated
with inelastic response.
C16.1.3.2 ThreeDimensional Analysis. Approaches to scaling ground motions for threedimensional analysis are similar to
those for twodimensional analysis. The two orthogonal components within each pair must have the same scale factor, but
the individual pairs may have different scale factors. Within 10 kilometers of a fault, ground motion components often are
selected to represent faultnormal and faultparallel directions, but this is not required. For certain structures, the response
under both horizontal and vertical ground motions should be considered. It is noted, however, that vertical ground motion
spectra are not readily available, so the scaling of the vertical components of ground motion would be problematic.
The 1.3 factor in the comparison of the average of the SRSS spectra to the design spectrum is intended to compensate for the
increase associated with taking the SRSS of the two components of each ground motion pair. If the two components are
perfectly correlated (identical response spectra in both directions), the SRSS would be larger than the average by the square
root of 2. Because real ground motions are not perfectly correlated, a smaller factor is acceptable. The judgment of the
writers (after two cycles of revision) is to allow a factor of 1.3 and to allow the results to be low by as much as 10 percent
(producing an effective factor of 1.18).
Given a set of appropriate ground motions, there are an infinite number of scaling factors that may be applied to the
individual motions to meet the requirements of Sections 16.1.3.1 and 16.1.3.2. Thus, two analysts, working with the same set
of ground motions, are likely to produce a different set of scale factors. While this difference in scaling would have little
impact in linear analysis, it may lead to vastly different results in nonlinear analysis. For this reason, the process of selection
and scaling of ground motions should be included in the design review (Section 16.2.5 of the standard) that is required
wherever nonlinear response history analysis is used.
Both amplitude scaling and spectral matching procedures can be used to satisfy the scaling technique specified in Sections
16.1.3.1 and16.1.3.2. Both procedures provide reasonable estimates of mean response for the individual response
parameters. Spectral matching can provide mean estimates with a smaller suite of motions, although seven suites are still
required as outlined in Section 16.1.4. Neither scaling approach, however, is adequate to give an accurate estimate of the
variability, although amplitude scaling gives a better understanding of the potential variability than spectral matching.
C16.1.4 Response Parameters. The responses derived from the response history analysis are multiplied by I to provide
enhanced strength and stiffness for more important facilities, and are divided by R to account for inelastic behavior. For
consistency with the ELF procedure and the MRS analysis procedure, the displacements computed from the response
histories that have been further modified by I/R should be multiplied by Cd to obtain the displacement histories to use for
computing the story drift histories. (The requirement to multiply displacements by Cd was incorrectly omitted in the
standard.)
If for any ground motion the peak base shear is less than that computed from Equation 12.85 or 12.86, the entire response
history must be scaled up such that the peak base shear is not less than that computed from Equation 12.85 or 12.86, as
applicable. The base shear typically is computed from component elastic forces. A slightly different shear would be
computed from the total inertial forces, with the difference being due to damping. Note that while the results of MRS
analysis must be scaled up such that the corresponding base shear is not less than 85 percent of the base shear that would be
computed from an ELF analysis (see Section C12.9.4), the scaling for linear response history analysis considers only the
applicable minimum base shear coefficient.
If seven or more ground motions are used, the design values may be taken as the average of the scaled values from the
response history analysis. This provides some difficulty for components for which the capacity depends on multiple values.
For a column, for example, both the axial force and the concurrent bending moment are needed to compare demand and
capacity. In that instance, if seven or more ground motions are used, the column is deemed suitable if the average of the
seven peak demandtocapacity ratios for the column is less than 1.0. Where fewer than seven ground motions are used, the
column is deemed suitable if the maximum demandtocapacity ratio is less than 1.0.
The direction of loading requirements of Section 12.5 and the modeling requirements of Section 12.7 apply to response
history analysis, but Chapter 16 of the standard does not address additional requirements such as accidental torsion,
amplification of accidental torsion, or detailed consideration of Pdelta effects. These effects should be included in a manner
consistent with the requirements of Section 12.9.
C16.2 NONLINEAR RESPONSE HISTORY PROCEDURE
Nonlinear response history analysis is not used as part of the normal design process for typical structures. In some cases,
however, nonlinear analysis is recommended, and in certain cases required, to obtain a more realistic assessment of structural
response and verify the results of simpler methods of analysis. Such is the case for systems with frictionbased passive
energy dissipation devices, nonlinear viscous dampers, seismically isolated systems, selfcentering systems, or systems that
have components with highly irregular forcedeformation relationships.
The principal aim of nonlinear response history analysis is to determine if the computed deformations of the structure are
within appropriate limits. Strength requirements for the designated lateral loadresisting elements do not apply because
element strengths are established prior to the analysis. These initial strengths typically are determined from a preliminary
design using linear analysis.
The nonlinear response history analysis may also provide useful information on the strength requirements for nonstructural
components, which are often assumed to remain elastic in the analysis.
Where displacements computed from the nonlinear response history analysis are excessive, a typical remedy is to increase the
stiffness of the structure, which is likely to affect the computed strength.
Nonlinear response history analysis offers several advantages over linear response history analysis, including the ability to
model a wide variety of nonlinear material behaviors, geometric nonlinearities (including large displacement effects), gap
opening and contact behavior, and nonclassical damping, and to identify the likely spatial and temporal distributions of
inelasticity. Nonlinear response history analysis has several disadvantages, including increased effort to develop the
analytical model, increased time to perform the analysis (which is often complicated by difficulties in obtaining converged
solutions), sensitivity of computed response to system parameters, and the inapplicability of superposition to combine live,
dead, and seismic load effects.
C16.2.1 Analysis Requirements. Nonlinear response history analysis of structures with widely distributed inelastic
behavior is usually carried out using the direct analysis approach (described in Section C16.1.1), wherein all equations are
solved simultaneously at each time step. In some cases, it is possible to use a highly efficient nonlinear modal analysis
approach called Fast Nonlinear Analysis, or FNA (Wilson, 2004). The class of nonlinear structures that may be analyzed by
the FNA approach consists of structures with a very limited number of discrete sources of welldefined nonlinear behavior.
Such structures include seismically isolated structures and structures with damping systems. Because of the limited
applicability of FNA, this commentary discusses only the direct analysis approach.
The sensitivity of nonlinear response history analysis may be evidenced by results that appear to be chaotic or even counterintuitive,
although they may be correct. For example, it is possible for the analysis to predict that a structure collapses when
subjected to a given ground motion, while surviving at higher intensity of the same motion. Similarly, the results from
analyses of the same structure for several ground motions with similar spectral shapes and amplitudes often differ
substantially. A systematic approach to assess the sensitivity of structures to different ground motions and structural system
parameters, using Incremental Dynamic Analysis (IDA), is reported by Vamvatsikos (2002). The IDA method has become
an important tool in earthquake engineering research.
C16.2.2 Modeling. Nonlinear response history analysis requires a mathematical description of the hysteretic behavior of
those portions of the structure that are expected to exhibit inelastic behavior during an earthquake. Such models must reflect
the expected properties, accounting for the following effects as appropriate:
1. Material overstrength and strain hardening
2. Cyclic degradation of stiffness and strength
3. Incycle degradation of stiffness and strength (Applied Technology Council, 2009)
4. Pinching
5. Buckling
6. Axialflexuralshear interaction
Most of the available mathematical models are phenomenological and represent yielding portions as distinct elements (such
as plastic hinges). More exact analysis may be performed by subdividing yielding portions into a number of slices or fibers.
This more exact approach is preferable but is more computationally demanding.
An inelastic threedimensional analysis is particularly useful for buildings that are prone to torsional response in plan, even
where the main seismic forceresisting systems resist loads predominantly in their own plane. If only twodimensional
software is available, a “pseudo” threedimensional analysis may be performed (Mehrain and Naeim, 2003).
In moment resisting steel frames, the elastic and inelastic behavior of beamcolumn joint regions should be modeled
explicitly. Pdelta effects should be considered explicitly in the analysis. Nonstructural components also should be included
in the model if it is expected that their stiffness and strength have a significant effect on the response.
Nonlinear response history analysis requires that inherent damping be set for the structure. As for linear response history
analysis, nonlinear response history analysis typically is performed assuming inherent damping of 5 percent of critical. Some
analysts and designers advocate the use of lower levels of inherent damping (perhaps 2 percent of critical), especially for
steel frames, but there is no widespread agreement on this point.
The mechanism used to include inherent damping in the analytical model is critically important to the accuracy of the
computed response. Most nonlinear analysis programs use a form of Rayleigh damping, wherein the damping matrix (used
for direct integration of the equations of motion) is represented as a linear combination of the mass and stiffness matrices.
(See Section C16.1.2.) If the damping matrix is based on the initial stiffness of the system, artificial damping may be
generated by system yielding. In some cases, the artificial damping can completely skew the computed response (Chrisp,
1980; Carr, 2004; Charney, 2006; Hall, 2006). One method to counter this occurrence is to base the damping matrix on the
mass and the instantaneous tangent stiffness. (Where basing the damping on tangent stiffness, care must be taken so that the
damping is not negative when the tangent stiffness is negative.) Other approaches have been suggested, such as capped
Rayleigh damping (Hall, 2006) and hysteretic damping (Charney, 2006). Several commercial programs, including SAP2000,
Perform 2D, and Ruaomoko, provide for tangent stiffnessbased damping.
Threedimensional analysis must be used where certain plan irregularities are present. For structures composed of twodimensional
seismic forceresisting elements connected by floor and roof diaphragms, the diaphragms should be modeled as
flexible inplane, particularly where the vertical elements of the seismic forceresisting system are of different types (such as
moment frames and walls). Where structures are modeled in three dimensions, axial forcebiaxial bending interaction should
be considered for corner columns, nonrectangular walls, and other similar elements.
As mentioned above, Pdelta effects should be included where significant. The significance of Pdelta effects on the overall
response may be assessed by performing analyses with and without Pdelta effects, and comparing story drift response
histories. Destabilizing effects of gravity loads are often manifested by accumulated residual deformations, and these
deformations, if not controlled, can lead to dynamic instability of the structure.
C16.2.3 Ground Motion and Other Loading. Since linear superposition cannot be used with nonlinear analysis, each
response history analysis must begin with an initial gravity load, consisting of the expected dead load and live load. The live
load may be as little as 25 percent of the unfactored design live load because multiple transient loads are unlikely to attain
their maxima simultaneously.
C16.2.4 Response Parameters. As discussed above, the principal aim of nonlinear response history analysis is to determine
deformation demands in structural and nonstructural components for comparison with accepted limits. Where at least seven
ground motions are used, the member and connection deformations may be taken as the average of the values computed from
the analyses. If fewer than seven motions are used, the maximum values among all analyses must be used. It is very
important to note, however, that assessment of deformations in this manner should not be done without careful inspection of
the story displacement histories of each analysis. It is possible that the maximum displacement or drift may be completely
dominated by the response from one ground motion, and such dominance, when due to ratcheting (increasing deformations in
one direction resulting in a high residual deformation), may be a sign of imminent dynamic instability. Where these kinds of
dynamic instabilities are present, the analyst should attempt to determine the system characteristics that produce such effects.
The ground motion that produces dynamic instability should not be replaced with one that does not.
C16.2.4.1 Member Strength. The strength design load combinations of Section 12.4 need not be assessed because linear
combinations of load are not applicable in nonlinear analysis. Overstrength effects are evaluated directly since hysteretic
forcedeformation relationships are modeled explicitly and the material properties so used include overstrength and strain
hardening (as required by Section 16.2.2).
C16.2.4.2 Member Deformation. This section requires that member and connection deformations be assessed on the basis
of tests performed for similar configurations.
C16.2.4.3 Story Drift. The 25 percent increase in allowable story drift is provided because the nonlinear analysis is
generally more accurate than linear analysis and because member deformations are assessed explicitly.
C16.2.5 Design Review. As discussed above, nonlinear response history analysis is quite complex, and the results may be
strongly influenced by subtle changes in ground motion or system properties. Hence, such analysis must only be conducted
by experienced professionals with training in engineering seismology, earthquake engineering, structural dynamics, stability,
nonlinear analysis, and inelastic behavior of structures. Regardless of the level of expertise of the individual or individuals
who perform the analysis and design, a design (peer) review of the structural system and the nonlinear analysis is required
wherever the design is based on the nonlinear response history procedure.
REFERENCES
Anderson, J. A, and V. V. Bertero. 1987. “Uncertainties in Establishing Design Earthquakes,” Journal of Structural
Engineering, 113(8):17091724.
Applied Technology Council. 2009. The Effects of Strength and Stiffness Degradation on Seismic Response, FEMA P440A.
FEMA, Washington, D.C.
Carr, A. J. 2004. Ruaumoko Users Manual, Volume 1. Department of Civil Engineering, University of Canterbury, New
Zealand.
Charney, F. A. 2006. “Unintended Consequences of Modeling Damping in Structures,” in Proceedings of the 17th Analysis
and Computation Conference, ASCE, St. Louis, Missouri.
Chrisp, D. J. 1980. Damping Models for Inelastic Structures, Master of Engineering Report, University of Canterbury, New
Zealand.
Hall, J. F. 2006. “Problems Encountered from the Use (or Misuse) of Rayleigh Damping,” Earthquake Engineering and
Structural Dynamics, 25(5):525545.
Iervolino, I., and A. Cornell. 2005. “Record Selection for Nonlinear Seismic Analysis of Structures,” Earthquake Spectra,
21(3):685714.
Mehrain, M., and F. Naeim. 2003. “Exact Three Dimensional Linear and Nonlinear Seismic Analysis of Structures Using
TwoDimensional Models,” Earthquake Spectra, 19(4):897912.
Naeim, F., A. Alimoradi, and S. Pezeshk. 2004. “Selection and Scaling of Ground Motion Time Histories for Structural
Design using Genetic Algorithms,” Earthquake Spectra, 20(2):413426.
Shome, N., and A. Cornell. 1998. “Normalization and Scaling Accelerograms for Nonlinear Analysis,” in Proceedings of
the 6th U.S. National Conference on Earthquake Engineering. Earthquake Engineering Research Institute.
Shome, N., A. Cornell, P. Bazzurro, and J. E. Carballo. 1998. “Earthquakes, Records, and Nonlinear Response,” Earthquake
Spectra, 14(3):469500.
Somerville, P., Anderson, D., Sun, J., Punyamurthula, S., and Smith, N., 1998 Generation of Ground Motion Time Histories
for Performance Based Seismic Engineering,” Proceedings of the 6th U.S. National Conference on Earthquake Engineering,
Earthquake Engineering Research Institute.
Vamvatsikos, D. 2002. Seismic Performance, Capacity and Reliability of Structures As Seen Through Incremental Dynamic
Analysis, Ph.D. Dissertation, Department of Civil and Environmental Engineering, Stanford University, Palo Alto,
California.
Wilson, E. L., and J. Penzien. 1970. “Evaluation of Orthogonal Damping Matrices,” International Journal for Numerical
Methods in Engineering, 4:510.
Wilson, E. L., A. Der Kiureghian, and E. Bayo. 1981. “A Replacement for the SRSS Method in Seismic Analysis,”
Earthquake Engineering and Structural Dynamics, 9:192.
Wilson, E. L., M. Yuan, and J. Dickens. 1982. “Dynamic Analysis by Direct Superposition of Ritz Vectors,” Earthquake
Engineering and Structural Dynamics, 10:813823.
Wilson, E. L. 2004. Static and Dynamic Analysis of Structures, 4th Edition. Computers and Structures, Inc., Berkeley,
California.
Figure C17.11 Idealized forcedeflection relationships for isolation systems (stiffness effects of sacrificial windrestraint systems not shown for clarity).
B: Hardening A: Linear
C: Softening
D: Sliding
DD
k
Force
Displacement
COMMENTARY TO CHAPTER 17,
SEISMIC DESIGN REQUIREMENTS FOR
SEISMICALLY ISOLATED STRUCTURES
C17.1 GENERAL
Seismic isolation, commonly referred to as base isolation, is a method used to decouple substantially the response of a
structure from potentially damaging earthquake motions. This decoupling can result in response that is reduced significantly
from that of a conventional, fixedbase building.
The potential advantages of seismic isolation and the recent advancements in isolationsystem technology have led to the
design and construction of a large number of seismically isolated buildings and bridges in the United States.
Design requirements for seismically isolated structures were first codified in the United States as an appendix to the 1991
Uniform Building Code, based on “General Requirements for the Design and Construction of SeismicIsolated Structures”
developed by the Structural Engineers Association of California State Seismology Committee. In the intervening years, those
provisions have developed along two parallel tracks into the design requirements in Chapter 17 of the standard and the
rehabilitation requirements in Section 9.2 of ASCE/SEI 41, Seismic Rehabilitation of Existing Buildings. The design and
analysis methods of both standards are quite similar, but ASCE/SEI 41 permits more liberal design for the superstructure of
rehabilitated buildings. The AASHTO Guide Specification for Seismic Isolation Design provides a systematic approach to
determining bounding values of mechanical properties of isolators for analysis and design. Rather than addressing a specific
method of seismic isolation, the standard provides general design requirements applicable to a wide range of possible seismic
isolation systems. Because the design requirements are general, testing of isolationsystem hardware is required to confirm
the engineering parameters used in the design and to verify the overall adequacy of the isolation system. Use of isolation
systems whose adequacy is not proved by testing is prohibited. In general, acceptable systems: (a) remain stable when
subjected to design displacements, (b) provide increasing resistance with increasing displacement, (c) do not degrade under
repeated cyclic load, and (d) have quantifiable engineering parameters (such as forcedeflection characteristics and damping).
The forcedeflection behavior of isolation systems falls into four categories, as shown in Figure C17.11, where each
idealized curve has the same design displacement, DD. A linear isolation system (Curve A) has an effective period
independent of displacement, and the force generated in the superstructure is directly proportional to the displacement of the
isolation system.
with increasing displacement, the procedures of the standard cannot be applied, and use of the system is prohibited.
Figure C17.11 Idealized forcedeflection relationships for isolation systems
(stiffness effects of sacrificial windrestraint systems not shown for clarity).
A hardening isolation system (Curve B) is soft initially (long effective period) and then stiffens (effective period shortens) as
displacement increases. Where displacements exceed the design displacement, the superstructure is subjected to larger forces
and the isolation system to smaller displacements than for a comparable linear system.
A softening isolation system (Curve C) is stiff initially (short effective period) and then softens (effective period lengthens)
as displacement increases. Where displacements exceed the design displacement, the superstructure is subjected to smaller
forces and the isolation system to larger displacements than for a comparable linear system.
The response of a purely sliding isolation system (Curve D) is governed by the friction force at the sliding interface. For
increasing displacement, the effective period lengthens, and loads on the superstructure remain constant. For isolation
systems governed solely by friction forces, the total displacement due to repeated earthquake cycles is highly dependent on
the characteristics of the ground motion and may exceed the design displacement, DD. Since such systems do not have
increasing resistance
Chapter 17 provides isolator design displacements, shear forces for structural design, and other specific requirements for
seismically isolated structures. All other design requirements, including loads (other than seismic), load combinations,
allowable forces and stresses, and horizontal shear distribution, are the same as those for conventional, fixedbase structures.
C17.1.1 Variations in Material Properties. For analysis, the mechanical properties of seismic isolators generally are based
on values provided by isolator manufacturers. Values of the mechanical properties should be in the range that accounts for
natural variability and uncertainty, and variability of properties among isolater of different manufacturers. Examples may be
found in Constantinou et al. (2007b). Prototype testing is used to confirm the values assumed for design. Unlike conventional
materials whose properties do not vary substantially with time, the materials used in seismic isolators have properties that
generally will vary with time. Because mechanical properties can vary over the life of a structure and the testing protocol of
Section 17.8 cannot account for the effects of aging, contamination, scragging (temporary degradation of mechanical
properties with repeated cycling), temperature, velocity effects, and wear, the designer must account for these effects by
explicit analysis. One approach to accommodate these effects, introduced in Constantinou et al. (1999), is to use property
modification factors. Information on variations in material properties of seismic isolators and dampers is reported in
Constantinou et al. (2007).
C17.2 GENERAL DESIGN REQUIREMENTS
Ideally, most of the lateral displacement of an isolated structure will be accommodated by deformation of the isolation
system rather than distortion of the structure above. Accordingly, the seismicforceresisting system of the structure above
the isolation system is designed to have sufficient stiffness and strength to avoid large, inelastic displacements. Therefore,
the standard contains criteria that limit the inelastic response of the structure above the isolation system. Although damage
control is not an explicit objective of the standard, design to limit inelastic response of the structural system also will reduce
the level of damage that would otherwise occur during an earthquake. In general, isolated structures designed in accordance
with the standard are expected:
1. To resist minor and moderate levels of earthquake ground motion without damage to structural elements, nonstructural
components, or building contents and
2. To resist major levels of earthquake ground motion without failure of the isolation system, significant damage to
structural elements, extensive damage to nonstructural components, or major disruption to facility function.
Isolated structures are expected to perform much better than fixedbase structures during moderate and major earthquakes.
Table C17.21 compares the performance expected for isolated and fixedbase structures designed in accordance with the
standard.
Table C17.21 Performance Expected for Minor, Moderate, and Major Earthquakesa
Performance Measure
Earthquake Ground Motion Level
Minor
Moderate
Major
Life safety: loss of life or serious injury is not expected
F, I
F, I
F, I
Structural damage: significant structural damage is not expected
F, I
F, I
I
Nonstructural damage: significant nonstructural or contents damage
is not expected
F, I
I
I
a F indicates fixed base; I indicates isolated.
Loss of function is not included in Table C17.21. For certain fixedbase facilities, loss of function would not be expected
unless there is significant structural damage causing closure or restricted access to the building. In other cases, a facility with
only limited or no structural damage would not be functional as a result of damage to vital nonstructural components or
contents. Isolation would be expected to mitigate structural and nonstructural damage and to protect the facility against loss
of function.
C17.2.4 Isolation System
C17.2.4.1 Environmental Conditions. Environmental conditions that may adversely affect isolation system performance
must be investigated thoroughly. Related research has been conducted since the 1970s in Europe, Japan, New Zealand, and
the United States.
C17.2.4.2 Wind Forces. Lateral displacement over the depth of the isolator zone resulting from wind loads must be limited
to a value similar to that required for other story heights.
C17.2.4.3 Fire Resistance. While fire may adversely affect the lateral performance of the isolation system, its gravityload
resistance must be maintained as required for other elements of the structure.
C17.2.4.4 Lateral Restoring Force. The restoringforce requirement is intended to limit residual displacement as a result of
an earthquake, so that the isolated structure will survive aftershocks and future earthquakes.
C17.2.4.5 Displacement Restraint. The use of a displacement restraint is discouraged. Where a displacement restraint
system is used, explicit analysis of the isolated structure for maximum considered earthquake response is required to account
for the effects of engaging the displacement restraint.
C17.2.4.6 Verticalload Stability. The vertical loads to be used in checking the stability of any given isolator should be
calculated using bounding values of dead load and live load and the peak earthquake demand of the maximum considered
earthquake. Since earthquake loads are reversible in nature, peak earthquake load should be combined with bounding values
of dead and live load in a manner which produces both the maximum downward force and the maximum upward force on
any isolator. Stability of each isolator should be verified for these two extreme values of vertical load at peak maximum
considered earthquake displacement of the isolation system.
C17.2.4.7 Overturning. The intent of this requirement is to prevent both global structural overturning and overstress of
elements due to local uplift. Isolator uplift is acceptable so long as the isolation system does not disengage from its
horizontalresisting connection detail. The connection details used in some isolation systems are such that tension is not
permitted on the system. Where the tension capacity of an isolator is used to resist uplift forces, design and testing in
accordance with Sections 17.2.4.6 and 17.8.2.5 must be performed to demonstrate the adequacy of the system to resist
tension forces at the total maximum displacement.
C17.2.4.8 Inspection and Replacement. Although most isolation systems will not need to be replaced after an earthquake,
access for inspection and replacement must be provided, and periodic inspection is required. After an earthquake, the
isolation system should be inspected and any damaged elements replaced or repaired.
C17.2.4.9 Quality Control. A testing and inspection program is necessary for both fabrication and installation of the
isolator units. Because seismic isolation is a rapidly evolving technology, it may be difficult to reference standards for
testing and inspection. Reference can be made to standards for some materials, such as elastomeric bearings (ASTM D
4014). Similar standards are yet to be developed for other isolation systems. Special inspection procedures and load testing
to verify manufacturing quality must be developed for each project. The requirements will vary depending on the type of
isolation system used.
C17.2.5 Structural System
C17.2.5.2 Building Separations. A minimum separation between the isolated structure and rigid obstructions is required to
allow free movement of the superstructure in all lateral directions during an earthquake.
C17.2.6 Elements of Structures and Nonstructural Components. To accommodate the differential movement between
the isolated building and the ground, flexible utility connections are required. In addition, stiff elements crossing the
isolation interface (such as stairs, elevator shafts, and walls) must be detailed to accommodate the total maximum
displacement without compromising life safety.
C17.3 GROUND MOTION FOR ISOLATED STRUCTURES
C17.3.1 Design Spectra. Seismically isolated structures located on Site Class F sites and on sites with S1 = 0.6 must be
analyzed using response history analysis. For those cases, the response spectra must be sitespecific in order to account, in
the analysis, for nearfault effects and for soft soil conditions, both of which are known to be important in the assessment of
displacement demands in seismically isolated structures.
C17.3.2 Ground Motion Histories. The selection and scaling of ground motions for response history analysis requires
fitting to the response spectra in the period range of 0.5TD to 1.25TM, a range that is different from that for conventional
structures 0.2T to 1.5T. The following sections provide background on the two period ranges:
1. Period Range – Isolated Structures. The effective (fundamental) period of an isolated structure is based on
amplitudedependent, nonlinear (pushover) stiffness properties of the isolation system. The effective periods, TD
and TM, correspond to design earthquake displacement and maximum considered earthquake displacement,
respectively, in the direction under consideration. Values of effective (fundamental) periods, TD and TM, are
typically in the range of 2 to 4 seconds, and the value of the effective period, TD, typically is 15 to 25 percent less
than the corresponding value of effective period, TM.
The response of an isolated structure is dominated by the fundamental mode in the direction of interest. The
specified period range, 0.5TD to 1.25TM, conservatively bounds amplitudedependent values of the effective
(fundamental) period of the isolated structure in the direction of interest, considering that individual earthquake
records can affect response at effective periods somewhat longer than TM, or significantly shorter than TD.
2. Period Range – Conventional, FixedBase Structures. The fundamental period, T, of a conventional, fixedbase
structure is based on amplitudeindependent, linearelastic stiffness properties of the structure. In general, response
of conventional, fixedbase structures is influenced by both the fundamental mode and higher modes in the direction
under consideration. The period range, 0.2T to 1.5T, is intended to bound the fundamental period, considering some
period lengthening due to nonlinear response of the structure (that is, inelastic periods up to 1.5T) and periods
corresponding to the more significant higher modes (that is, second and possibly third modes in the direction of
interest).
C17.4 ANALYSIS PROCEDURE SELECTION
Three different analysis procedures are available for determining designlevel seismic loads: the equivalent lateral force
procedure, the response spectrum procedure, and the response history procedure. For the first procedure simple, lateralforce
formulas (similar to those for conventional, fixedbase structures) are used to determine peak lateral displacement and design
forces as a function of spectral acceleration and isolatedstructure period and damping. For the second and third procedures,
which are required for geometrically complex or especially flexible buildings, dynamic analysis (either the response spectrum
procedure or the response history procedure) is used to determine peak response of the isolated structure.
The three procedures are based on the same level of seismic input and require a similar level of performance from the
building. Where more complex analysis procedures are used, slightly lower design forces and displacements are permitted.
The design requirements for the structural system are based on the design earthquake, taken as twothirds of the maximum
considered earthquake. The isolation system—including all connections, supporting structural elements, and the “gap”—is
required to be designed (and tested) for 100 percent of maximum considered earthquake demand. Structural elements above
the isolation system are not required to be designed for the full effects of the design earthquake but may be designed for
slightly reduced loads (that is, loads reduced by a factor of up to 2) if the structural system is able to respond inelastically
without sustaining significant damage. A similar fixedbase structure would be designed for loads reduced by a factor of 8
rather than 2.
This section delineates the requirements for the use of the equivalent lateral force procedure and dynamic methods of
analysis. The limitations on the simplified lateralforce design procedure are quite restrictive. Limitations relate to the site
location with respect to major, active faults; soil conditions of the site; the height, regularity, and stiffness characteristics of
the building; and selected characteristics of the isolation system. Responsehistory analysis is required to determine the
design displacement of the isolation system (and the structure above) for the following isolated structures:
1. Isolated structures with a “nonlinear” isolation system including, but not limited to, isolation systems with effective
damping values greater than 30 percent of critical, isolation systems incapable of producing a significant restoring force,
and isolation systems that restrain or limit extreme earthquake displacement; and
2. Isolated structures located on a Class E or Class F site (that is, a soft soil site that amplifies longperiod ground motions).
Lowerbound limits on isolation system design displacements and structuraldesign forces are specified by the standard in
Section 17.6 as a percentage of the values prescribed by the equivalent lateral force procedure, even where dynamic analysis
is used as the basis for design. These lowerbound limits on key design parameters provide consistency in the design of
isolated structures and serve as a “safety net” against gross underdesign. Table C17.41 provides a summary of the lowerbound
limits on dynamic analysis specified by the standard.
C17.5 EQUIVALENT LATERAL FORCE PROCEDURE
C17.5.3 Minimum Lateral Displacements. The lateral displacement given by Equation 17.51 approximates peak design
earthquake displacement of a singledegreeoffreedom, linearelastic system of period, TD, and equivalent viscous damping,
ß D. Similarly, the lateral displacement given by Equation 17.53 approximates peak maximum considered earthquake
displacement of a singledegreeoffreedom, linearelastic system of period, TM, and equivalent viscous damping, ß M.
Table C17.41 LowerBound Limits on Dynamic Procedures Specified in Relation to
ELF Procedure Requirements
Design Parameter
ELF Procedure
Dynamic Procedure
Response
Spectrum
Response
History
Design displacement – DD
DD = (g/4p2)(SD1TD/BD)
—
—
Total design displacement  DTD
DTD 1.1DD Equation
=
0.9DTD Equation
=
0.9DTD Equation
=
Maximum displacement – DM
DM = (g/4p2)(SM1TM/BM)
—
—
Total maximum displacement  DTM
DTM 1.1DM Equation
=
0.8DTM Equation
=
0.8DTM Equation
=
Design shear – Vb
(at or below the isolation system)
Vb = kDmaxDD
0.9Vb Equation
=
0.9Vb Equation
=
Design shear – Vs
(“regular” superstructure)
Vs = kDmaxDD/RI
0.8Vs Equation
=
0.6Vs Equation
=
Design shear – Vs
(“irregular” superstructure)
Vs = kDmaxDD/RI
1.0Vs Equation
=
0.8Vs Equation
=
Drift (calculated using RI for Cd)
0.015hsx
0.015hsx
0.020hsx
Equation 17.51 is an estimate of peak displacement in the isolation system for the design earthquake. In this equation, the
spectral acceleration term, SD1, is the same as that required for design of a conventional, fixedbase structure of period, TD.
A damping term, BD, is used to decrease (or increase) the computed displacement where the equivalent damping coefficient
of the isolation system is greater (or smaller) than 5 percent of critical damping. Values of coefficient BD (or BM for the
maximum considered earthquake) are given in Table 17.51 for different values of isolation system damping, ß D (or ß M).
A comparison of values obtained from Equation 17.51 and those obtained from nonlinear timehistory analyses are given in
Kircher et al. (1988) and Constantinou et al. (1993).
Consideration should be given to possible differences in the properties used for design of the isolation system and those of
the isolation system as installed in the structure. Similarly, consideration should be given to possible changes in isolation
system properties due to different design conditions or load combinations. If the true deformational characteristics of the
isolation system are not stable or if they vary with the nature of the load (being rate, amplitude, or timedependent), the
design displacements should be based on deformational characteristics of the isolation system that give the largest possible
deflection (kDmin), the design forces should be based on deformational characteristics of the isolation system that give the
largest possible force (kDmax), and the damping level used to determine design displacements and forces should be based on
deformational characteristics of the isolation system that represent the minimum amount of energy dissipated during cyclic
response at the design level.
The isolation system for a seismically isolated structure should be configured to minimize eccentricity between the center of
mass of the superstructure and the center of rigidity of the isolation system. In this way, the effect of torsion on the
displacement of isolation elements will be reduced. As for conventional structures, allowance must be made for accidental
eccentricity in both horizontal directions. Figure C17.51 illustrates the terminology used in the standard. Equation 17.55
(or Equation 17.56 for the maximum considered earthquake) provides a simplified formula for estimating the response due
to torsion in lieu of a more refined analysis. The additional component of displacement due to torsion increases the design
displacement at the corner of a structure by about 15 percent (for one perfectly square in plan) to about 30 percent (for one
very long and rectangular in plan) if the eccentricity is 5 percent of the maximum plan dimension. These calculated torsional
Figure C17.51 Displacement terminology.
Total maximum displacement
(maximum considered earthquake
corner of building)
Maximum displacement
(maximum considered earthquake
center of building)
Design displacement
(design earthquake
center of building)
TM
M
D
D
D
D
Plan view
of building
displacements are for structures with an isolation system whose stiffness is uniformly distributed in plan. Isolation systems
that have stiffness concentrated toward the perimeter of the structure, or certain sliding systems that minimize the effects of
mass eccentricity, will have smaller torsional displacements. The standard permits values of DTD as small as 1.1DD, with
proper justification.
Figure C17.51 Displacement terminology.
C17.5.4 Minimum Lateral Forces. Figure C17.52 illustrates the terminology for elements at, below, and above the
isolation system. Equation 17.57 specifies the peak seismic shear for design of all structural elements at or below the
isolation system (without reduction for ductile response). Equation 17.58 specifies the peak seismic shear for design of
structural elements above the isolation system. For structures that have appreciable inelasticdeformation capability, this
equation includes an effective reduction factor of up to 2 for response beyond the strengthdesign level.
The reduction factor is based on use of strengthdesign procedures. A factor of at least 2 is assumed to exist between the
designforce level and the trueyield level of the structural system. An investigation of 10 specific buildings indicated that
this factor varied between 2 and 5 (ATC, 1982). Thus, a reduction factor of 2 is appropriate to produce a structural system
that remains essentially elastic for the design earthquake.
In Section 17.5.4.3, the limits given on VS provide at least a factor of 1.5 between the nominal yield level of the
superstructure and (a) the yield level of the isolation system, (b) the ultimate capacity of a sacrificial windrestraint system
that is intended to fail and release the superstructure during significant lateral load, or (c) the breakaway friction level of a
sliding system.
These limits are needed so that the superstructure will not yield prematurely before the isolation system has been activated
and significantly displaced.
Figure C17.52 Isolation system terminology.
Structure above the
isolation system
Structural elements that transfer
force between isolator units
Isolator
unit
Isolator
unit
interface
Isolation
Figure C17.52 Isolation system terminology.
The design shear force, VS, specified in this section results in an isolated structural system being subjected to significantly
lower inelastic demands than a conventionally designed structural system. Further reduction in VS, such that the inelastic
demand on a seismically isolated structure would be the same as the inelastic demand on a conventionally designed structure,
was not considered during development of these requirements but may be considered in the future.
Using a smaller value of RI in Equation 17.58 will reduce or eliminate inelastic response of the superstructure for the designbasis
event, thus further improving the structural performance.
C17.5.5 Vertical Distribution of Forces. Equation 17.59 produces a vertical distribution of lateral forces consistent with a
triangular profile of seismic acceleration over the height of the structure above the isolation interface. Kircher et al. (1988)
and Constantinou et al. (1993) show that Equation 17.59 provides a conservative estimate of the distributions obtained from
more detailed, nonlinear analysis studies for the type of structures for which use of Equation 17.59 is allowed.
C17.5.6 Drift Limits. The maximum story drift permitted for design of isolated structures is constant for all Occupancy
Categories, as shown in Table C17.51. For comparison, the drift limits prescribed by the standard for fixedbase structures
also are summarized in that table.
Table C17.51 Comparison of Drift Limits for FixedBase and Isolated Structures
Structure
Occupancy Category
FixedBase
Isolated
Buildings (other than masonry)
four stories or less in height with
component drift design
I or II
0.025hsx/(Cd/R)
0.015hsx
III
0.020hsx/(Cd/R)
0.015hsx
IV
0.015hsx/(Cd/R)
0.015hsx
Other (nonmasonry) buildings
I or II
0.020hsx/(Cd/R)
0.015hsx
III
0.015hsx/(Cd/R)
0.015hsx
IV
0.010hsx/(Cd/R)
0.015hsx
Drift limits in Table C17.51 are divided by Cd /R for fixedbase structures since displacements calculated for lateral loads
reduced by R are multiplied by Cd before checking drift. The Cd term is used throughout the standard for fixedbase
structures to approximate the ratio of actual earthquake response to response calculated for “reduced” forces. Generally, Cd
is 1/2 to 4/5 the value of R. For isolated structures, the RI factor is used both to reduce lateral loads and to increase
displacements (calculated for reduced lateral loads) before checking drift. Equivalency would be obtained if the drift limits
for both fixedbase and isolated structures were based on their respective R factors. It may be noted that the drift limits for
isolated structures generally are more conservative than those for conventional, fixedbase structures, even where fixedbase
structures are assigned to Occupancy Category IV.
C17.6 DYNAMIC ANALYSIS PROCEDURES
This section specifies the requirements and limits for dynamic procedures. The design displacement and force limits on
response spectrum and response history procedures are shown in Table C17.41.
A more detailed or refined study can be performed in accordance with the analysis procedures described in this section. The
intent of this section is to provide procedures that are compatible with the minimum requirements of Section 17.5. Reasons
for performing a more refined study include:
1. The importance of the building.
2. The need to analyze possible structure/isolationsystem interaction where the fixedbase period of the building is greater
than onethird of the isolated period.
3. The need to explicitly model the deformational characteristics of the lateralforceresisting system where the structure
above the isolation system is irregular.
4. The desirability of using sitespecific groundmotion data, especially for very soft or liquefiable soils (Site Class F) or
for structures located where S1 is greater than 0.60.
5. The desirability of explicitly modeling the deformational characteristics of the isolation system. This is especially
important for systems that have damping characteristics that are amplitudedependent, rather than velocitydependent,
since it is difficult to determine an appropriate value of equivalent viscous damping for these systems.
Section C17.4 discusses other conditions that require use of the response history procedure. As shown in Table C17.41, the
drift limit for isolated structures is relaxed where story drifts are computed using nonlinear response history analysis.
Where response history analysis is used as the basis for design, the design displacement of the isolation system and design
forces in elements of the structure above are computed from not fewer than three separate analyses, each using a different
ground motion selected and scaled in accordance with Section 17.3.2. Where the configuration of the isolation system or of
the superstructure is not symmetric, additional analyses are required to satisfy the requirement of Section 17.6.3.4 to consider
the most disadvantageous location of eccentric mass. As appropriate, nearfield phenomena may also be incorporated. As in
the nuclear industry, where at least seven ground motions are used for nonlinear response history analysis, it is considered
appropriate to base design of seismically isolated structures on the average value of the response parameters of interest.
C17.7 DESIGN REVIEW
Review of the design and analysis of the isolation system and design review of the isolator testing program is mandated by
the standard for two key reasons:
1. The consequences of isolator failure could be catastrophic.
2. Isolator design and fabrication technology is evolving rapidly and may be based on technologies unfamiliar to many
design professionals.
The standard requires review to be performed by a team of registered design professionals that are independent of the design
team and other project contractors. The review team should include individuals with special expertise in one or more aspects
of the design, analysis, and implementation of seismic isolation systems.
The review team should be formed prior to the development of design criteria (including sitespecific ground shaking criteria)
and isolation system design options. Further, the review team should have full access to all pertinent information and the
cooperation of the design team and regulatory agencies involved with the project.
C17.8 TESTING
The design displacements and forces determined using the standard assume that the deformational characteristics of the
isolation system have been defined previously by comprehensive testing. If comprehensive test data are not available for a
system, major design alterations in the structure may be necessary after the tests are complete. This would result from
Figure C17.81 The effect of stiffness on an isolation bearing.
Force
Displacement
Fmax
+
F min
+
Kmax
Kmin
. I .+I
F min

Fmax

K m a x =
K m in =
F m a x F
+
max

F m i n F
+
min

.+ I  .I
.+ I  .I
variations in the isolationsystem properties assumed for design and those obtained by test. Therefore, it is advisable that
prototype systems be tested during the early phases of design, if sufficient test data are not available on an isolation system.
Typical forcedeflection (or hysteresis) loops are shown in Figure C17.81; also illustrated are the values defined in Section
17.8.5.1.
Figure C17.81 The effect of stiffness on an isolation bearing.
The required sequence of tests will verify experimentally the following:
1. The assumed stiffness and capacity of the windrestraining mechanism;
2. The variation of the isolator’s deformational characteristics with amplitude (and with vertical load, if it is a vertical loadcarrying
member);
3. The variation of the isolator’s deformational characteristics for a realistic number of cycles of loading at the design
displacement; and
4. The ability of the system to carry its maximum and minimum vertical loads at the maximum displacement.
Forcedeflection tests are not required if similarly sized components have been tested previously using the specified sequence
of tests.
The variations in the vertical loads required for tests of isolators which carry vertical, as well as lateral, load are necessary to
determine possible variations in the system properties with variations in overturning force.
C17.8.5 Design Properties of the Isolation System
C17.8.5.1 Maximum and Minimum Effective Stiffness. The effective stiffness is determined from the hysteresis loops as
shown in Figure C17.81. Stiffness may vary considerably as the test amplitude increases but should be reasonably stable
(within 15 percent) for more than three cycles at a given amplitude.
The intent of these requirements is that the deformational properties used in design result in the maximum design forces and
displacements. For determining design displacement, this means using the smallest damping and effective stiffness values.
For determining design forces, this means using the smallest damping value and the largest stiffness value.
C17.8.5.2 Effective Damping. The determination of equivalent viscous damping is reasonably reliable for systems whose
damping characteristics are velocity dependent. For systems that have amplitudedependent energydissipating mechanisms,
significant problems arise in determining an equivalent viscousdamping value. Since it is difficult to relate velocity and
amplitudedependent phenomena, it is recommended that where the equivalentviscous damping assumed for the design of
amplitudedependent energydissipating mechanisms (such as puresliding systems) is greater than 30 percent, the designbasis
force and displacement be determined using the response history procedure, as discussed in Section C17.4.
REFERENCES
Applied Technology Council. 1982. An Investigation of the Correlation between Earthquake Ground Motion and Building
Performance, ATC Report 10. ATC, Redwood City, California.
Association of State Highway and Transportation Officials. 1999. Guide Specification for Seismic Isolation Design.
AASHTO, Washington, D.C.
Constantinou, M. C., C. W. Winters, and D. Theodossiou. 1993. “Evaluation of SEAOC and UBC analysis procedures, Part
2: Flexible superstructure,” in Proceedings of a Seminar on Seismic Isolation, Passive Energy Dissipation and Active
Control, ATC Report 171. ATC, Redwood City, California.
Constantinou, M. C., P. Tsopelas, A. Kasalanati, and E. D. Wolff. 1999. “Property modification factors for seismic isolation
bearings,” MCEER990012. Multidisciplinary Center for Earthquake Engineering Research, Buffalo, New York.
Constantinou, M. C., A. S. Whittaker, Y. Kalpakidis, D. M. Fenz, and G. P. Warn. 2007. “Performance of seismic isolation
hardware under service and seismic loading,” MCEER070012. Multidisciplinary Center for Earthquake Engineering
Research, Buffalo, New York.
International Code Council. 2006. International Building Code. ICC, Washington, D.C.
Kircher, C. A., B. Lashkari, R. L. Mayes, and T. E. Kelly. 1988. “Evaluation of nonlinear response in seismically isolated
buildings,” in Proceedings of a Symposium on Seismic, Shock and Vibration Isolation, ASME PVP Conference.
Lashkari, B., and C. A. Kircher. 1993. “Evaluation of SEAOC & UBC analysis procedures, Part 1: stiff superstructure,” in
Proceedings of a Seminar on Seismic Isolation, Passive Energy Dissipation and Acti.: ATC.
COMMENTARY TO CHAPTER 18,
SEISMIC DESIGN REQUIREMENTS FOR STRUCTURES
WITH DAMPING SYSTEMS
C18.1 GENERAL
The requirements of this chapter apply to all types of damping systems including both displacementdependent damping
devices of hysteretic or friction systems and velocitydependent damping devices of viscous or viscoelastic systems (Soong
and Dargush, 1997; Constantinou et al., 1998; Hanson and Soong, 2001). Compliance with these requirements is intended to
produce performance comparable to that for a structure with a conventional seismicforceresisting system, but the same
methods can be used to achieve higher performance.
The damping system (DS) is defined separately from the seismicforceresisting system (SFRS), although the two systems
may have common elements. As illustrated in Figure C18.11, the damping system may be external or internal to the
structure and may have no shared elements, some shared elements, or all elements in common with the seismicforceresisting
system. Elements common to the damping system and the seismicforceresisting system must be designed for a
combination of the two loads of the two systems. When the DS and SFRS have no common elements, the damper forces must
be collected and transferred to members of the SFRS.
C18.2 GENERAL DESIGN REQUIREMENTS
C18.2.2 System Requirements. Structures with a damping system must have a seismicforceresisting system that provides
a complete load path. The seismicforceresisting system must comply with all of the height, Seismic Design Category, and
redundancy limitations and with the detailing requirements specified in this standard for the specific seismicforceresisting
system. The seismicforceresisting system without the damping system (as if damping devices were disconnected) must be
designed to have not less than 75 percent of the strength required for undamped structures having that type of seismicforceresisting
system (and not less than 100 percent if the structure is horizontally or vertically irregular). The damping systems,
however, may be used to meet the drift limits (whether the structure is regular or irregular). Having the SFRS designed for a
minimum of 75 percent of the strength required for undamped structures provides safety in the event of damping system
malfunction and produces a composite system with sufficient stiffness and strength to have controlled lateral displacement
response.
The damping system must be designed for the actual (unreduced) earthquake forces (such as, peak force occurring in
damping devices) and deflections. For certain elements of the damping system ( such as the connections or the members into
which the damping devices frame ), other than damping devices, limited yielding is permitted provided such behavior does
not affect damping system function or exceed the amount permitted for elements of conventional structures by the standard.
C18.2.4 Procedure Selection. Linear static and response spectrum analysis methods can be used for design of structures
with damping systems that meet certain configuration and other limiting criteria (for example, at least two damping devices
at each story configured to resist torsion). In such cases, additional nonlinear response history analysis shall be used to
confirm peak responses when the structure is located at a site with S1 greater than or equal to 0.6. The analysis methods
damperd structures are based on nonlinear static “pushover” characterization of the structure and calculation of peak response
using effective (secant) stiffness and effective damping properties of the first (pushover) mode in the direction of interest.
These are the concepts used in Chapter 17 to characterize the forcedeflection properties of isolation systems, modified to
incorporate explicitly the effects of ductility demand (postyield response) and highermode response of structures with
dampers. Like conventional structures, dampered structures generally yield during strong ground shaking, and their
performance can be influenced strongly by response of higher modes.
The response spectrum and equivalent lateral force procedures presented in the standard have several simplifications and
limits, as outlined below:
1. A multidegreeoffreedom (MDOF) structure with a damping system can be transformed into equivalent singledegreeof
freedom (SDOF) systems using modal decomposition procedures. This assumes that the collapse mechanism for the
structure is a singledegreeoffreedom mechanism so that the drift distribution over height can be estimated reasonably
using either the first mode shape or another profile, such as an inverted triangle. Such procedures do not strictly apply to
either yielding buildings or buildings that are nonproportionally damped.
2. The response of an inelastic singledegreeoffreedom system can be estimated using equivalent linear properties and a
5percentdamped response spectrum. Spectra for damping greater than 5 percent can be established using damping
Figure C18.11 Damping system (DS) and seismicforceresisting system (SFRS) configurations.
Internal damping devices common elements
Internal damping devices some shared elements
Internal damping devices no shared elements
External damping devices
SFRS
SFRS
SFRS
SFRS
DS
DS
DS
DS
Damper
Damper
Damper
Damper
coefficients, and velocitydependent forces can be established either by using the pseudovelocity and modal information
or by applying correction factors to the pseudovelocity.
3. The nonlinear response of the structure can be represented by a bilinear hysteretic relationship with zero postelastic
stiffness (elastoplastic behavior).
4. The yield strength of the structure can be estimated either by performing simple plastic analysis or by using the specified
minimum seismic base shear and values of R, O0, and Cd.
5. Higher modes need to be considered in the equivalent lateral force procedure in order to capture their effects on velocitydependent
forces. This is reflected in the residual mode procedure.
Figure C18.11 Damping system (DS) and seismicforceresisting system (SFRS) configurations.
FEMA 440 (Applied Technology Council, 2005) presents a review of simplified procedures for the analysis of yielding
structures. The combined effects of the simplifications mentioned above are reported by Ramirez et al. (2001) and Pavlou
and Constantinou (2004) based on studies of 3story and 6story buildings with damping systems designed by the procedures
of the standard. The response spectrum and equivalent lateral force procedures of the standard are found to provide
conservative predictions of drift and predictions of damper forces and member actions that are of acceptable accuracy when
compared to results of nonlinear dynamic response history analysis. When designed in accordance with the standard,
structures with damping systems are expected to have structural performance at least as good as that of structures without
damping systems. Pavlou and Constantinou (2006) report that structures with damping systems designed in accordance with
the standard provide the benefit of reduced secondary system response, although this benefit is restricted to systems with
added viscous damping.
C18.3 NONLINEAR PROCEDURES
For designs in which the seismicforceresistingsystem is essentially elastic (assuming an overstrength of 50 percent), the
only nonlinear characteristics that must be modeled in the analysis are those of the damping devices. For designs in which
the seismicforceresisting system will yield, the postyield behavior of the structural elements must be modeled explicitly.
C18.4 RESPONSE SPECTRUM PROCEDURES and C18.5 EQUIVALENT LATERAL FORCE PROCEDURE
Effective Damping
In the standard the reduced response of a structure with a damping system is characterized by the damping coefficient, B,
based on the effective damping, ß , of the mode of interest. This is the same approach as that used for isolated structures.
Like isolation, effective damping of the fundamentalmode of a damped structure is based on the nonlinear forcedeflection
properties of the structure. For use with linear analysis methods, nonlinear properties of the structure are inferred from the
overstrength factor, O 0, and other terms.
Figure C18.41 illustrates reduction in design earthquake response of the fundamental mode due to increased effective
damping (represented by coefficient, B1D). The capacity curve is a plot of the nonlinear behavior of the fundamental mode in
spectral accelerationdisplacement coordinates. The reduction due to damping is applied at the effective period of the
fundamental mode of vibration (based on the secant stiffness).
In general, effective damping is a combination of three components:
1. Inherent Damping ( ßI
)—Inherent damping of the structure at or just below yield, excluding added viscous damping
(typically assumed to be 5 percent of critical for structural systems without dampers).
2. Hysteretic Damping (ß H)—Postyield hysteretic damping of the seismicforceresisting system and elements of the
damping system at the amplitude of interest (taken as 0 percent of critical at or below yield).
3. Added Viscous Damping (ß V)—Viscous component of the damping system (taken as 0 percent for hysteretic or frictionbased
damping systems).
Both hysteretic damping and added viscous damping are amplitudedependent, and the relative contributions to total effective
damping change with the amount of postyield response of the structure. For example, adding dampers to a structure
decreases postyield displacement of the structure and, hence, decreases the amount of hysteretic damping provided by the
seismicforceresisting system. If the displacements are reduced to the point of yield, the hysteretic component of effective
damping is zero, and the effective damping is equal to inherent damping plus added viscous damping. If there is no damping
system (as in a conventional structure), effective damping simply equals inherent damping (typically assumed to be 5 percent
of critical for most conventional structures).
Linear Analysis Methods
The section specifies design earthquake displacements, velocities, and forces in terms of design earthquake spectral
acceleration and modal properties. For equivalent lateral force (ELF) analysis, response is defined by two modes: the
fundamental mode and the residual mode. The residual mode is a new concept used to approximate the combined effects of
higher modes. While typically of secondary importance to story drift, higher modes can be a significant contributor to story
velocity and, hence, are important for design of velocitydependent damping devices. For response spectrum analysis, higher
modes are explicitly evaluated.
For both the ELF and the response spectrum analysis procedures, response in the fundamental mode in the direction of
interest is based on assumed nonlinear (pushover) properties of the structure. Nonlinear (pushover) properties, expressed in
terms of base shear and roof displacement, are related to building capacity, expressed in terms of spectral coordinates, using
mass participation and other fundamentalmode factors shown in Figure C18.42. The conversion concepts and factors
shown in Figure C18.42 are the same as those defined in Chapter 9 of Seismic Rehabilitation of Existing Buildings
(ASCE/SEI 41), which addresses seismic rehabilitation of a structure with damping devices.
Figure C18.41 Effective damping reduction of design demand.
SD1/T1
SDE
SD1/T1D
SD1/(T1DB1D)
SD1D
Spectral acceleration
Spectral displacement
BV+I
B1D
Effective damping:
(added viscous,
inherent, and hysteretic)
Design earthquake demand spectrum
(5% damping)
Design earthquake demand spectrum
(ßV + ßI
damping)
Capacity
curve
T1
T1D
1 1
1 2
1
2
1
2
1
4
4
D D
D
D
DS D
D
g S T SD
B
g S T
B
p
p
. .
=. .
. .
. .
=. .
. . Figure C18.42 Pushover and capacity curves.
(SD1D, SA1D)
(D1D, F1/W)
Acceleration
Displacement
Capacity
curve
T1
T1D
Pushover curve
(normalized by
weight)
2
1 4 2 1 1 D D D
SD g SA T
p
. .
=. .
. .
1 1 1
1
1
1
1
D D
n
i i
i
D SD
W
wf
=
= G
G =
S
1 1
1
2
1
1
1
2
1
1
D
n
i i
i
n
i i
i
F W
SA
W W
w
W
w
f
f
=
=
=
. .
. .
= . .
S
S
Figure C18.41 Effective damping reduction of design demand.
Figure C18.42 Pushover and capacity curves.
Figure C18.43 Showing Idealized elastoplastic pushover curve used for linear analysis.
V1
VY
DY D1D
Base shear
Roof displacement
Pushover curve
Idealized elastoplastic
pushover curve
2
1 1 1
1 2 1 2 1
1 1 4 4
D D DS D
D
D D
g S T g S T D
p B p B
. . . .
= . .G = . .G
. . . .
0 d C
R
O
D µ
2
1 1
2 max
1
D D
D
Y
D T
D T
µ = ˜ = µ
Where using linear analysis methods, the shape of the fundamentalmode pushover curve is not known, so an idealized
elastoplastic shape is assumed, as shown in Figure C18.43. The idealized pushover curve is intended to share a common
point with the actual pushover curve at the design earthquake displacement, D1D. The idealized curve permits definition of
the global ductility demand due to the design earthquake, µ D, as the ratio of design displacement, D1D, to yield displacement,
DY. This ductility factor is used to calculate various design factors; it must not exceed the ductility capacity of the seismicforce
resisting system, µ max, which is calculated using factors for conventional structural response. Design examples using
linear analysis methods have been developed and found to compare well with the results of nonlinear time history analysis
(Ramirez et al., 2001).
Elements of the damping system are designed for fundamentalmode design earthquake forces corresponding to a base shear
value of VY (except that damping devices are designed and prototypes tested for maximum considered earthquake response).
Elements of the seismicforceresisting system are designed for reduced fundamentalmode base shear, V1, where force
reduction is based on system overstrength (represented by O 0), multiplied by Cd /R for elastic analysis (where actual
pushover strength is not known). Reduction using the ratio Cd /R is necessary because the standard provides values of Cd
that are less than those for R. Where the two parameters have equal value and the structure is 5 percent damped under elastic
conditions, no adjustment is necessary. Because the analysis methodology is based on calculating the actual story drifts and
damping device displacements (rather than the displacements calculated for elastic conditions at the reduced base shear and
then multiplied by Cd), an adjustment is needed. Since actual story drifts are calculated, the allowable story drift limits of
Table 12.121 are multiplied by R/ Cd before use.
Figure C18.43 Idealized elastoplastic pushover curve used for linear analysis.
C18.6 DAMPED RESPONSE MODIFICATION
C18.6.1 Damping Coefficient
Values of the damping coefficient, B, in Table 18.61 for design of damped structures are the same as those in Table 17.51
for isolated structures at damping levels up to 20 percent, but extend to higher damping levels based on results presented in
Ramirez et al. (2001). Table C18.61 compares values of the damping coefficient as found in the standard and various
resource documents and codes. FEMA 440 and the draft of Eurocode 8 present equations for the damping coefficient, B,
whereas the other documents present values of B in tabular format.
The equation in FEMA 440 is
Equation
4
5.6 ln(100 )
B
ß
=

The equation in Eurocode 8 is
Table C18.61 Values of Damping Coefficient, B Equation
0.05
0.10
B +ß
=
Effective Damping,
(%) Equation
ß
Table 17.51,
1999 AASHTO, 2001
CBC (seismically
isolated structures)
Table 18.61
(structures with
damping systems)
FEMA 440
EUROCODE 8
2
0.8
0.8
0.8
0.8
5
1.0
1.0
1.0
1.0
10
1.2
1.2
1.2
1.2
20
1.5
1.5
1.5
1.6
30
1.7
1.8
1.8
1.9
40
1.9
2.1
2.1
2.1
50
2.0
2.4
2.4
2.3
C18.6.2 Effective Damping. The effective damping is calculated assuming the structural system exhibits perfectly bilinear
hysteretic behavior characterized by the effective ductility demand, µ, as described in Ramirez et al. (2001). Effective
damping is adjusted using the hysteresis loop adjustment factor, qH, which is the actual area of the hysteresis loop divided by
the area of the assumed perfectly bilinear hysteretic loop. In general, values of this factor are less than unity. In Ramirez et
al. (2001) expressions for this factor (which they call Quality Factor) are too complex to serve as a simple rule. Equation
18.65 provides a simple estimate of this factor. The equation predicts correctly the trend in the constant acceleration domain
of the response spectrum, and it is believed to be conservative for flexible structures.
C18.7 SEISMIC LOAD CONDITIONS AND ACCEPTANCE CRITERIA
C18.7.2.5 Seismic Load Conditions and Combination of Modal Responses. Seismic design forces in elements of the
damping system are calculated at three distinct stages: maximum displacement, maximum velocity, and maximum
acceleration. All three stages need to be checked for structures with velocitydependent damping systems. For displacementdependent
damping systems, the first and third stages are identical, whereas the second stage is inconsequential.
Force coefficients CmFD and CmFV are used to combine the effects of forces calculated at the stages of maximum displacement
and maximum velocity to obtain the forces at maximum acceleration. The coefficients are presented in tabular form based on
analytic expressions presented in Ramirez et al. (2001) and account for nonlinear viscous behavior and inelastic structural
system behavior.
REFERENCES
American Association of State Highway and Transportation Officials. 1999. Guide Specifications for Seismic Isolation
Design, Washington, D.C.
Applied Technology Council. 2005. Improvement of Nonlinear Static Seismic Analysis Procedures, FEMA 440. Federal
Emergency Management Agency, Washington, D.C.
California Buildings Standards Commission. 2007. California Building Code. Sacramento, California.
Constantinou, M. C., T. T. Soong, and G. F. Dargush. 1998. Passive Energy Dissipation Systems for Structural Design and
Retrofit, Monograph 1. Multidisciplinary Center for Earthquake Engineering Research, University of Buffalo, State
University of New York, Buffalo.
European Committee for Standardization. 2005. Design of Structures for Earthquake Resistance. Part 2: Bridges, Eurocode
8, EN19982, draft, August.
Hanson, R. D., and T. T. Soong. 2001. Seismic Design with Supplemental Energy Dissipation Devices, MNO8. Earthquake
Engineering Research Institute, Oakland, California.
Newmark, N. M., and W. J. Hall. 1969. “Seismic Design Criteria for Nuclear Reactor Facilities,” in Proceedings of the 4th
World Conference in Earthquake Engineering, Santiago, Chile.
Pavlou, E., and M. C. Constantinou. 2004. “Response of Elastic and Inelastic Structures with Damping Systems to Near
Field and SoftSoil Ground Motions,” Engineering Structures, 26:12171230.
Pavlou, E., and M. C. Constantinou. 2006. “Response of Nonstructural Components in Structures with Damping Systems,”
ASCE Journal of Structural Engineering, 132(7):11081117.
Ramirez, O. M., M. C. Constantinou, C. A. Kircher, A. Whittaker, M. Johnson, J. D. Gomez and C. Z. Chrysostomou. 2001.
Development and Evaluation of Simplified Procedures of Analysis and Design for Structures with Passive Energy
Dissipation Systems, Technical Report MCEER000010, Revision 1. Multidisciplinary Center for Earthquake Engineering
Research, University of Buffalo, State University of New York, Buffalo.
Ramirez, O. M., M. C. Constantinou, J. Gomez, A. S. Whittaker, and C. Z. Chrysostomou. 2002a. “Evaluation of Simplified
Methods of Analysis of Yielding Structures With Damping Systems,” Earthquake Spectra, 18(3):501530.
Ramirez, O. M., M. C. Constantinou, A. S. Whittaker, C. A. Kircher, and C. Z. Chrysostomou. 2002b. “Elastic and Inelastic
Seismic Response of Buildings with Damping Systems,” Earthquake Spectra, 18(3):531547.
Ramirez, O. M., M. C. Constantinou, A. S. Whittaker, C. A. Kircher, M. W. Johnson, and C. Z. Chrysostomou. 2003.
“Validation of 2000 NEHRP Provisions Equivalent Lateral Force and Modal Analysis Procedures for Buildings with
Damping Systems,” Earthquake Spectra, 19(4):981999.
Soong, T. T., and G. F. Dargush. 1997. Passive Energy Dissipation Systems in Structural Engineering. J. Wiley & Sons,
London, UK.
Whittaker, A.S., M. C. Constantinou, O. M. Ramirez, M. W. Johnson, and C. Z. Chrysostomou. 2003. “Equivalent Lateral
Force and Modal Analysis Procedures of the 2000 NEHRP Provisions for Buildings with Damping Systems,” Earthquake
Spectra, 19(4):959980.
Page intentionally left blank.
COMMENTARY FOR CHAPTER 19,
SOIL STRUCTURE INTERACTION FOR SEISMIC DESIGN
C19.1 GENERAL
The response of a structure to earthquake shaking is affected by interactions between three linked systems: the structure, the
foundation, and the geologic media underlying and surrounding the foundation. A seismic SoilStructure Interaction (SSI)
analysis evaluates the collective response of these systems to a specified freefield ground motion. The term “freefield”
refers to motions not affected by structural vibrations and represents the condition for which the design spectrum is derived
using the procedures given in Chapter 11.
SSI effects are absent for the theoretical condition of rigid foundation and soil conditions. Accordingly, SSI effects reflect
the differences between the actual response of the structure and the response for the theoretical, rigid base condition.
Visualized within this context, three SSI effects can significantly affect the response of building structures:
1. Foundation stiffness and damping. Inertia developed in a vibrating structure gives rise to base shear, moment, and
torsional excitation, and these loads in turn cause displacements and rotations of the foundation relative to the free field.
These relative displacements and rotations are only possible because of compliance in the soilfoundation system, which
can significantly contribute to the overall structural flexibility in some cases. Moreover, the relative foundationfree
field motions give rise to energy dissipation via radiation damping (i.e., damping associated with wave propagation into
the ground away from the foundation, which acts as the wave source) and hysteretic soil damping, and this energy
dissipation can significantly affect the overall damping of the soilfoundationstructure system. Since these effects are
rooted in the structural inertia, they are referred to as inertial interaction effects.
2. Variations between freefield and foundationlevel ground motions. The differences between foundation and freefield
motions result from two processes. The first is known as kinematic interaction and results from the presence of stiff
foundation elements on or in soil, which cause foundation motions to deviate from freefield motion as a result of base
slab averaging, wave scattering, and embedment effects. Procedures for modifying design spectra to account for these
effects are given in FEMA 440 and ASCE/SEI 41. The second process is related to the structure and foundation inertia
and consists of the relative foundationfree field displacements and rotations described above.
3. Foundation deformations. Flexural, axial, and shear deformations of foundation elements occur as a result of loads
applied by the superstructure and the supporting soil medium. Such deformations represent the seismic demand for
which foundation components should be designed. These deformations can also significantly affect the overall system
behavior, especially with respect to damping.
Chapter 19 treats only the inertial interaction effects (the first item above). Inertial interaction in buildings tends to be
important for stiff structural systems (such as shear walls and braced frames), particularly where the foundation soil is
relatively soft (i.e., Site Classes C to E). Kinematic interaction effects are neglected in these provisions. Foundation design
is covered in Section 12.13.
The procedures in Chapter 19 are used to modify the fixedbase properties (period and damping) of a structural system. If
fixedbase properties are obtained from an analytical model of the structure, the fixedbase properties correspond to a
condition without soil springs. If soil springs are included in the analytical model of the structure, then the procedures given
in Chapter 19 should not be used to modify the building period. The damping procedures in Chapter 19 could still be used in
this case if the foundation springs are linear (thus introducing no damping) and there are no dashpots in parallel with the
springs. In the remainder of this commentary, it is assumed that the structural period and damping ratio that are being
modified for SSI effects correspond to a fixedbase condition.
In design procedures that utilize response spectra to establish design values of base shear (i.e., forcebased methods such as
those given in Chapter 12), the effects of inertial SSI on the seismic response of buildings is represented as a function of the
ratio of flexible to fixedbase firstmode natural period, , and system damping, , attributable to foundationsoil
interaction. The flexiblebase firstmode damping ratio, , is calculated using Equation 199. Figure C191 illustrates two
possible effects of inertial SSI on the peak base shear, which is commonly computed from spectral acceleration at the firstmode.
The spectral acceleration for a flexiblebased structure ( ) is obtained by entering the spectrum drawn for
effective damping ratio, , at the corresponding elongated period, . For buildings with periods greater than about 0.5 s,
using in lieu of Sa (=Cs/g) typically reduces base shear demand, whereas in very stiff structures SSI can increase the base Equation
1 1
T~ T Equation
0 ß Equation
ß ~ Equation
S C g a s
~ = ~ Equation
ß ~ Equation
T~ Equation
a S~
Figure C191 Schematic showing effects of period lengthening and foundation damping on design spectral accelerations.
0
0 Period
Spectral acceleration
T T. T T.
a S.
a S.
a S
a S
ß.
ß
, Fixedbase period, damping ratio
(neglects SSI effects)
, Flexiblebase period, damping ratio
(includes SSI effects)
T
T
ß
ß
=
. . =
shear. Most equivalent lateral force methods feature a flat spectral shape at low periods that, when coupled with the
requirement that , results in modeling of inertial SSI that can only decrease the base shear demand. Equation
ß~ > ß
Figure C191 Schematic showing effects of period lengthening and foundation damping on design spectral accelerations.
The method given in Chapter 19 for evaluating inertial SSI effects is optional and has rarely been used in practice. There are
several reasons for this. First, because the guidelines were written such that base shear demand can only decrease through
consideration of SSI, SSI effects are ignored in order to be conservative. Second, many design engineers who have attempted
to apply the method on projects have done so for major, highrise buildings for which they felt evaluating SSI effects could
provide cost savings. Unfortunately, inertial interaction effects are negligible for these tall, flexible structures, and hence the
design engineers realized no benefit for their efforts and thereafter discontinued use of the procedures. The use of the
procedures actually yield the most benefit for shortperiod (T < 1 sec), stiff structures with stiff, interconnected foundation
systems (i.e., mats or interconnected footings) founded on soil.
C19.2 EQUIVALENT LATERAL FORCE PROCEDURE
This procedure considers the response of the structure in its fundamental mode of vibration and accounts for the contributions
of the higher modes to story shears implicitly through the choice of the effective weight of the structure and the vertical
distribution of the lateral forces. The effects of soilstructure interaction are accounted for on the assumption that they
influence only the contribution of the fundamental mode of vibration.
C19.2.1 Base Shear. Base shear is reduced for the effects of SSI as indicated in Equation 19.21 and 19.22. As indicated
in Equation 19.22, the change in base shear is related to the change in seismic coefficient (or spectral acceleration, as shown
in Figure C191). The term (0.05 )
.
C19.2.1.1 Effective Building Period. The fixed base period, T, is lengthened to the flexiblebase period, , using Equation
19.23, which was derived by Veletsos and Meek (1974). Terms Ky and K. represent the translational and rocking stiffnesses
of the foundation, respectively. The standard does not provide guidance on the evaluation of these stiffness terms. Equations
for Ky and K. are given by Gazetas (1991), and a number of practical considerations associated with the analysis of these
terms are reviewed in FEMA 440 (2005). For convenience, simplified guidelines are presented below for these stiffness
terms for use with the standard.
For building foundation systems having lateral continuity, such as mats or footings interconnected with grade beams,
stiffnesses Ky and K. can often be approximated as:
(C192)
0.4 in Equation 19.22 represents the reduction in spectral ordinate associated with a
change of damping from the fixed base value of ß = 0.05 to the flexible base value of Equation
ß. Equation
ß. Equation
T~ Equation
y a K Gr
.
=
2
8
(C193)
where: ra = an equivalent foundation radius that matches the area of the foundation, A0 (i.e., ra = v(A0/p)); rm = an equivalent
foundation radius that matches the moment of inertia of the foundation in the direction of shaking (i.e., ); G = the
straindependent shear modulus, as defined in the standard; . = the soil Poisson’s ratio (generally taken as 0.3 for sands and
0.45 for clays); and a. = a dimensionless coefficient that depends on the period of excitation, the dimensions of the
foundation, and the properties of the supporting medium (Luco, 1974; Veletsos and Verbic, 1973; Veletsos and Wei, 1971).
A similar coefficient exists for translation (ay), but can be taken as 1.0 without introducing significant error, and hence is not
shown in Equation C192.
As noted in the standard, shear modulus G is evaluated from smallstrain shear wave velocity as G = (G/Go)Go =
(G/Go).v
< 0.05 1.0
0.15 0.85
0.35 0.7
0.5 0.6
Foundation embedment has the effect of increasing the stiffnesses Ky and K.. For embedded foundations for which there is
positive contact between the side walls and the surrounding soil, Ky and K. may be determined from the following
approximate formulas (Kausel, 1974):
(C194)
(C195)
Experimental studies and field performance data (Stokoe and Erden, 1975; Stewart et al., 1999) indicate that the effects of
foundation embedment are sensitive to the condition of the backfill and that judgment must be exercised in using Equations
C194 and C195. For example, if contact is lost between the soil and basement walls, stiffnesses Ky and K. should be
determined from the formulas for surfacesupported foundations. More generally, the quantity d above should be interpreted
as the effective depth of foundation embedment for the conditions that would prevail during the design earthquake ground
motion.
The formulas for Ky and K. presented above can be applied to most soil profiles in which soil shear wave velocity, vs0,
changes with depth. However, if the soil profile consists of a surface stratum of soil underlain by a much stiffer deposit with
a shear wave velocity more than twice that of the surface layer, Ky and K. may be determined from the following two
generalized formulas in which G is the shear modulus of the surface soil and Ds is the total depth of the stratum:
(C196)
(C197)
2
so/g (all terms defined in the standard). Shear wave velocity, vs0, should be evaluated as the average smallstrain
shear wave velocity within the effective depth of influence below the foundation. The effective depth should be taken as
0.75ra for horizontal vibrations of the foundation and 0.75rm for rocking vibrations (Stewart et al., 2003). Methods for
measuring vs0 (preferred) or estimating it from other soil properties are summarized elsewhere (e.g., Kramer, 1996).
The dynamic modifier for rocking, a., can significantly affect the computed response of some building structures. In the
absence of more detailed analyses, for ordinary building structures with an embedment ratio d/rm < 0.5 (where d = depth of
embedment, measured from ground surface to base of foundation), the factor a. can be estimated as follows (Stewart et al.,
2003):
Equation
. ( ) . a
.
3
3 1
8
m K Gr

= Equation
4
0 4 mr = I p Equation
r (v T ) m s 0 Equation
. a Equation
8 1 2
2 3
a
y
a
K Gr d
. r
. . .. .. = +.. .. .. . . .. .. Equation
( )
8 3 1 2
3 1
m
m
K Gr d
r . .
. . ..
= . + . ..  . . .. Equation
..
.
..
.
. ..
.
. ..
.
..
.
..
+ . ..
.
..
.
. ..
.
. ..
.
..
.
..
+ . ..
.
.. .
. ..
.
. ..
.
..
.
..
+ .

=
s s
a
a
a
y D
d
D
r
r
K Gr d
4
1 5
2
1 1
3
1 2
2
8
. Equation
( ) ..
.
..
.
. ..
.
. ..
.
+ ..
.
..
.
. ..
.
. ..
.
..
.
..
+ . ..
.
..
.
. ..
.
. ..
.
+

=
s s
m
m
m
D
d
D
r
r
K Gr d 1 0.7
6
1 2 1 1
3 1
8 3
. .
symbol
ß ~ symbol
ß ~ symbol
ß ~ symbol
ß ~
The above formulas are based on analyses of a stratum supported on a rigid base (Elsabee et al., 1977; Kausel and Roesset,
1975) and apply for r/Ds < 0.5 and d/r < 1 (r taken as either ra or rm). The applicability of those rigid base solutions to
practical situations (nonrigid geologic media) was evaluated by Stewart et al. (2003), resulting in the recommendations
provided above.
For buildings supported on footing foundations, the above formulas can generally be used with ra and rm calculated using the
full building footprint dimensions, provided that the footings are interconnected with grade beams. An exception can occur
for buildings with both shear walls and frames, for which the rotation of the foundation beneath the wall may be independent
of that for the foundation beneath the column (this is referred to as weak rotational coupling). In such cases, rm is often best
calculated using the dimensions of the wall footing. Very stiff foundations, which provide strong rotational coupling, are
best described using an rm value that reflects the full foundation dimension. Regardless of the degree of rotational coupling,
ra should be calculated using the full foundation dimension if foundation elements are interconnected or continuous. Further
discussion can be found in FEMA (2005). The use of discrete (noninterconnected) spread footing foundations in seismic
regions is not recommended.
For buildings supported on pile foundations, lateral stiffness, Ky, can be taken as the sum of the lateral head stiffnesses of the
supporting piles. These stiffness values are generally calculated using a beam on Winkler foundation model, which is
discussed in detail elsewhere (e.g., Salgado, 2006). Rotational stiffness, K., can be calculated from the vertical stiffness of
the individual piles, kzi, as follows:
(C198)
where yi = horizontal distance from the foundation centroidal axis to pile i measured in the direction of shaking. The
approximation in Equation C198 assumes an infinitely rigid pile cap and neglects the rotational stiffness of individual piles,
which is typically a small contribution. Quantity kzi can be calculated for an individual pile using wellestablished methods,
such as discrete element modeling with tz curves (e.g., Salgado, 2006).
The alternate approach in the standard, represented by Equation 19.25, was derived using Poisson’s ratio . = 0.4, and is
generally sufficient for nonembedded foundations that are laterally continuous across the building footprint and for which
there is no “rigid” layer at depth in the profile (which would require the use of Equations C196 and C197 to calculate
foundation stiffness). The value of relative weight parameter, a (defined in the standard), can be taken as approximately 0.15
for typical buildings. Equation
˜S
i
zi i K k y2
.
C19.2.1.2 Effective Damping. Bielak (1975, 1976) and Veletsos and Nair (1975) expressed the flexiblebase firstmode
damping ratio, , as indicated in Equation 19.29. This equation is based on analyses of the harmonic response of singledegree
offreedom oscillators supported on a viscoelastic medium with hysteretic damping. Foundation damping factor ß0
incorporates the effects of energy dissipation into the soil due to radiation damping and hysteretic damping in the soil.
Figure 19.21 shows ß0 as a function of period lengthening ratio and was derived from the analytical solution presented in
Veletesos and Nair (1975) for the condition of zero foundation embedment. Additional damping can be realized for
embedded foundations, and the use of damping values from Figure 19.21 is conservative for such conditions. More exact
solutions can be obtained using procedures given in FEMA (2005).
Equation 19.29, in combination with the information presented in Figure 19.21, may lead to damping factors for the soilfoundation
structure system, , that are smaller than the fixed base structural damping, ß (assumed to be 0.05). However, it
is recommended that never be taken as less than 0.05 for design applications. The use of values of > ß is welljustified
from field casehistory data (Stewart et al., 1999, 2003).
The presence of a stiff layer at depth in the soil profile can impede radiation damping, rendering the values in Figure 19.21
too high. If a site consists of a relatively uniform layer of depth, Ds, overlying a very stiff layer with a shear wave velocity
more than twice that of the surface layer, damping values should be reduced as indicated by Equation 19.212.
C19.2.2 Vertical Distribution of Seismic Forces. The vertical distributions of the equivalent lateral forces for flexibly and
rigidly supported structures are similar, and it is recommended that the same distribution be used in both cases, changing only
the magnitude of the forces to correspond to the appropriate base shear. A greater degree of refinement in this step would be
inconsistent with the approximations embodied in the requirements for rigidly supported structures.
With the vertical distribution of the lateral forces established, the overturning moments and the torsional effects about a
vertical axis are computed as for rigidly supported structures. The above procedure is applicable to planar structures and,
with some extension, to threedimensional structures.
C19.2.3 Other Effects. In addition to its effect on base shear, inertial SSI also can increase the horizontal displacements of
the structure relative to its base (because of rocking). This can increase the required spacing between structures and
secondary design forces associated with Pdelta effects. Such effects can be significant for stiff structural systems (e.g., walls
and braced frames).
C19.3 MODAL ANALYSIS PROCEDURE
The procedure outlined above in Section C19.2 is applicable to a modal analysis by adjusting the modal period and damping
ratio of the fundamental mode only. Higher modes are relatively unaffected by SSI (e.g., Bielak, 1976; Chopra and
Gutierrez, 1974; Veletsos, 1977). Hence, the contributions of higher modes are computed as if the structure were fixed at the
base, and the maximum value of a response quantity is determined as for fixedbase structures but with the adjusted firstmode
responses.
REFERENCES
American Society of Civil Engineers/Structural Engineering Institute. 2006. Seismic Rehabilitation of Existing Buildings,
ASCE/SEI 41. ASCE/SEI, Reston, Virginia.
Bielak, J. 1975. "Dynamic Behavior of Structures with Embedded Foundations," Earthquake Engineering and Structural
Dynamics, 3:259274.
Bielak, J. 1976. "Modal Analysis for BuildingSoil Interaction," Journal of the ASCE Engineering Mechanics
Division 102(EM5):771786.
Chopra, A. K., and J. A. Gutierrez. 1974. "Earthquake Analysis of Multistory Buildings Including Foundation Interaction,"
Journal of Earthquake Engineering and Structural Dynamics, 3:6567.
Elsabee, F., I. Kausel, and J. M. Roesset. 1977. "Dynamic Stiffness of Embedded Foundations," in Proceedings of the ASCE
Second Annual Engineering Mechanics Division Specialty Conference, pp. 4043.
Federal Emergency Management Agency. 2005. Improvement of Nonlinear Static Seismic Analysis Procedures, FEMA 440.
Federal Emergency Management Agency, Washington, D.C.
Gazetas, G. 1991. “Formulas and Charts for Impedances of Surface and Embedded Foundations,” Journal of Geotechnical
Engineering, 117(9):13631381.
Kausel, E. 1974. Force Vibrations of Circular Foundations on Layered Media, Report R7411. Department of Civil
Engineering, Massachusetts Institute of Technology, Cambridge.
Kausel, E., and J. M. Roesset. 1975. "Dynamic Stiffness of Circular Foundations," Journal of the Engineering Mechanics
Division 101(EM6):771785.
Kramer, S.L. 1996. Geotechnical Earthquake Engineering. Prentice Hall, Upper Saddle River, New Jersey.
Luco, J. E. 1974. "Impedance Functions for a Rigid Foundation on a Layered Medium," Nuclear Engineering and Design,
(31):204217.
Salgado, R. 2006. The Engineering of Foundations. McGrawHill.
Stewart, J. P., R. B. Seed, and G. L. Fenves. 1999. "Seismic SoilStructure Interaction in Buildings. II: Empirical
Findings,” ASCE Journal of Geotechnical & Geoenvironmental Engineering, 125(1):3848.
Stewart, J. P., S. Kim, J. Bielak, R. Dobry, and M. Power. 2003. "Revisions to Soil Structure Interaction Procedures in
NEHRP Design Provisions," Earthquake Spectra, 19(3):677696.
Stokoe, II, K. H., and S. M. Erden. 1975. “Torsional Response of Embedded Circular Foundations,” in Proceedings of the
5th European Conference on Earthquake Engineering.
Veletsos, A. S., and V. V. Nair. 1975. "Seismic Interaction of Structures on Hysteretic Foundations," Journal of the ASCE
Structural Division, 101(ST1):109129.
Veletsos, A. S. 1977. "Dynamics of StructureFoundation Systems," in Structural and Geotechnical Mechanics, A Volume
Honoring N. M. Newmark, edited by W. J. Hall, pp. 333361. PrenticeHall, Englewood Cliffs, New Jersey.
Veletsos, A. S., and J. W. Meek. 1974. "Dynamic Behavior of building Foundation Systems," Earthquake Engineering and
Structural Dynamics, 3(2):121138.
Veletsos, A. S., and B. Verbic. 1973. "Vibration of Viscoelastic Foundations," Earthquake Engineering and Structural
Dynamics, 2(1):87105.
Veletsos, A. S., and Y. T. Wei. 1971. "Lateral and Rocking Vibration of Footings," Journal of the ASCE Soil Mechanics
and Foundations Division, 97(SM9):12271248.
COMMENTARY FOR CHAPTER 20,
SITE CLASSIFICATION PROCEDURE FOR SEISMIC DESIGN
C20.1 SITE CLASSIFICATION
Site classification procedures are given in Chapter 20 for the purpose of classifying the site and determining site coefficients
and siteadjusted maximum considered earthquake ground motions in accordance with Section 11.4.3. Site classification
procedures are also used to define the site conditions for which sitespecific site response analyses are required to obtain site
ground motions in accordance with Section 11.4.7 and Chapter 21.
C20.3 SITE CLASS DEFINITIONS
C20.3.1 Site Class F. Site conditions for which the site coefficients Fa and Fv in Tables 11.41 and 11.42 may not be
applicable and for which siteresponse analyses are required by Section 11.4.7. For shortperiod structures it is permissible to
determine values of Fa and Fv assuming that liquefaction does not occur, because ground motion data obtained in liquefied
soil areas during earthquakes indicate that shortperiod ground motions generally are attenuated due to liquefaction whereas
longperiod ground motions may be amplified. This exception does not affect the requirements in Section 11.8 to assess
liquefaction potential as a geologic hazard and to develop hazard mitigation measures as required.
C20.3.2 through C20.3.5. These sections and Table 20.31 provide definitions for Site Classes A through E. Except for the
additional definitions for Site Class E in Section 20.3.2, the site classes are defined fundamentally in terms of the average
smallstrain shear wave velocity in the top 100 feet (30 meters) of the soil or rock profile. If shear wave velocities are
available for the site, they should be used to classify the site. However, recognizing that in many cases shear wave velocities
are not available for the site, alternative definitions of the site classes also are included. These definitions are based on
geotechnical parameters: standard penetration resistance for cohesionless soils and rock, and standard penetration resistance
and undrained shear strength for cohesive soils. The alternative definitions are intended to be conservative since the
correlation between site coefficients and these geotechnical parameters is more uncertain than the correlation with shear wave
velocity. That is, values of Fa and Fv will tend to be smaller if the site class is based on shear wave velocity rather than on
the geotechnical parameters. Also, the site class definitions should not be interpreted as implying any specific numerical
correlation between shearwave velocity and standard penetration resistance or undrained shear strength.
Although the site class definitions in Sections 20.3.2 through 20.3.5 are straightforward, there are aspects of these
assessments that may require additional judgment and interpretation. Highly variable subsurface conditions beneath a
building footprint could result in overly conservative or unconservative site classification. Isolated soft soil layers within an
otherwise firm soil site may not affect the overall site response if the predominant soil conditions do not include such strata.
Conversely, site response studies have shown that continuous, thin, soft clay strata may affect the site amplification.
The site class should reflect the soil conditions that will affect the ground motion input to the structure or a significant portion
of the structure. For structures receiving substantial ground motion input from shallow soils (for example, structures with
shallow spread footings, with laterally flexible piles, or with basements where substantial ground motion input to the
structure may come through the side walls), it is reasonable to classify the site on the basis of the top 100 feet (30 meters) of
soils below the ground surface. Conversely, for structures with basements supported on firm soils or rock below soft soils, it
may be reasonable to classify the site on the basis of the soils or rock below the mat, if it can be justified that the soft soils
contribute very little to the response of the structure.
Buildings on sites with sloping bedrock or having highly variable soil deposits across the building area require careful study
since the input motion may vary across the building (for example, if a portion of the building is on rock and the rest is over
weak soils). Sitespecific studies including two or threedimensional modeling may be used in such cases to evaluate the
subsurface conditions and site and superstructure response. Other conditions that may warrant sitespecific evaluation
include the presence of low shear wave velocity soils below a depth of 100 feet (30 meters), location of the site in a
sedimentary basin, or subsurface or topographic conditions with strong two and threedimensional siteresponse effects.
Individuals with appropriate expertise in seismic ground motions should participate in evaluations of the need for and nature
of such sitespecific studies.
C20.4 DEFINITION OF SITE CLASS PARAMETERS
Section 20.4 provides formulas for defining Site Classes in accordance with definitions in Section 20.3 and Table 20.31.
Equation 20.41 is for determining the effective average smallstrain shearwave velocity, , to a depth of 100 feet (30 Equation
v s
meters) at a site. This equation defines as 100 feet (30 meters) divided by the sum of the times for a shear wave to travel
through each layer within the upper 100 feet (30 meters), where travel time for each layer is calculated as the layer thickness
divided by the smallstrain shear wave velocity for the layer. It is important that this method of averaging be used as it may
result in a significantly lower effective average shear wave velocity than the velocity that would be obtained by directly
averaging the velocities of the individual layers. Equation
v s
For example, consider a soil profile having four 25footthick layers with shear wave velocities of 500, 1,000, 1,500, and
2,000 ft/s. The arithmetic average of the shear wave velocities is 1250 ft/s (corresponding to Site Class C), but Equation
20.41 produces a value of 960 ft/s (corresponding to Site Class D). The Equation 20.41 value is appropriate as the four
layers are being represented by one layer with the same wave passage time.
Equation 20.42 is for classifying the site using the average standard penetration resistance blow count, , for cohesionless
soils, cohesive soils, and rock in the upper 100 feet (30 meters). A method of averaging analogous to the method of Equation
20.41 for shear wave velocity is used. The maximum value of N that may be used for any depth of measurement in soil or
rock is 100 blows/foot. For the common situation where rock is encountered, the standard penetration resistance, N, for rock
layers is taken as 100. Equation
N
Equations 20.43 and 20.44 are for classifying the site using the standard penetration resistance of cohesionless soil layers,
Nch, and the undrained shear strength of cohesive soil layers, su, within the top 100 feet (30 meters). These equations are
provided as an alternative to using Equation 20.42 for which Nvalues in all geologic materials in the top 100 feet (30
meters) are used. Where using Equations 20.43 and 20.44, only the respective thicknesses of cohesionless soils and
cohesive soils within the top 100 feet (30 meters) are used.
COMMENTARY FOR CHAPTER 21,
SITESPECIFIC GROUND MOTION PROCEDURES
FOR SEISMIC DESIGN
GENERAL
Sitespecific procedures for computing earthquake ground motions include dynamic site response analyses and probabilistic
and deterministic seismic hazard analyses (PSHA and DSHA), which may include dynamic site response analysis as part of
the calculation. Use of sitespecific procedures may be required in lieu of the general procedure in Sections 11.4.1 through
11.4.6; Section C11.4.7 explains the conditions under which the use of these procedures is required. Such studies must be
comprehensive and incorporate current scientific interpretations. Because there is typically more than one scientifically
credible alternative for models and parameter values used to characterize seismic sources and ground motions, it is important
to formally incorporate these uncertainties in a sitespecific analysis. For example, uncertainties may exist in seismic source
location, extent and geometry; maximum earthquake magnitude; earthquake recurrence rate; groundmotion attenuation; local
site conditions, including soil layering and dynamic soil properties; and possible two or threedimensional wavepropagation
effects. The use of peer review for a sitespecific groundmotion analysis is encouraged.
Sitespecific groundmotion analysis can consist of one of the following approaches: (a) PSHA and possibly DSHA if the site
is near an active fault, (b) PSHA/DSHA followed by dynamic siteresponse analysis, and (c) dynamic site response analysis
only. The first approach is used to compute ground motions for bedrock or stiff soil conditions (not softer than Site Class D).
In this approach, if the site consists of stiff soil overlying bedrock, for example, the analyst has the option of either (a)
computing the bedrock motion from the PSHA/DSHA and then using the sitecoefficient (Fa and Fv) tables in Section 11.4.3
to adjust for the stiff soil overburden or (b) computing the response spectrum at the ground surface directly from the
PSHA/DSHA. The latter requires the use of attenuation equations for computing stiff soilsite response spectra (instead of
bedrock response spectra).
The second approach is used where softer soils overlie the bedrock or stiff soils. The third approach assumes that a sitespecific
PSHA/DSHA is not necessary, but that a dynamic site response analysis should or must be performed. This analysis
requires the definition of an outcrop ground motion, which can be based on the 5 percent damped response spectrum
computed from the PSHA/DSHA or obtained from the general procedure in Section 11.4. A representative set of acceleration
time histories are selected and scaled to be compatible with this outcrop spectrum. Dynamic site response analyses using
these acceleration histories as input are used to compute motions at the ground surface. The response spectra of these surface
motions are used to define a maximum considered earthquake (MCE) ground motion response spectrum.
The approaches described above have advantages and disadvantages. In many cases, user preference governs the selection,
but geotechnical conditions at the site may dictate the use of one approach over the other. On the one hand, if bedrock is at a
depth much greater than the extent of the site geotechnical investigations, the direct approach of computing the groundsurface
motion in the PSHA/DSHA may be more reasonable. On the other hand, if bedrock is shallow and a large impedance
contrast exists between it and the overlying soil (i.e., density times shearwave velocity of bedrock is much greater than that
of the soil), the twostep approach might be more appropriate.
Use of peak ground acceleration as the anchor for a generalized sitedependent response spectrum is discouraged because
sufficiently robust groundmotion attenuation relations are available for computing response spectra in western United States
and eastern United States tectonic environments.
C21.1 SITE RESPONSE ANALYSIS
C21.1.1 Base Ground Motions. Ground motion acceleration histories that are representative of horizontal rock motions at
the site are required as input to the soil model. Where a sitespecific ground motion hazard analysis is not performed, the
MCE response spectrum for Site Class B (rock) is defined using the general procedure described in Section 11.4.1. If the
model is terminated in material of Site Class A, C, or D, the input MCE response spectrum is adjusted in accordance with
Section 11.4.3. The United States Geological Survey national seismic hazard mapping project website
(http://earthquake.cr.usgs.gov/research/hazmaps/) includes hazard deaggregation options that can be used to evaluate the
predominant types of earthquake sources, magnitudes, and distances contributing to the probabilistic groundmotion hazard.
Sources of recorded acceleration time histories include the databases of the Consortium of Organizations for Strong Motion
Observation Systems (COSMOS) Virtual Data Center website (db.cosmoseq.org) and the Pacific Earthquake Engineering
Research Center (PEER) Strong Motion Data Base website
(http://peer.berkeley.edu/products/strong_ground_motion_db.html/), and the United States National Center for Engineering
Strong Motion Data (NCESMD) website (http://www.strongmotioncenter.org). Ground motion acceleration histories at these
sites generally were recorded at the ground surface and hence apply for an outcropping condition and should be specified as
such in the input to the site response analysis code (see Kwok et al., 2007, for additional details).
C21.1.2 Site Condition Modeling. Modeling criteria are established by sitespecific geotechnical investigations that
should include: (a) borings with sampling, (b) standard penetration tests (SPTs), cone penetrometer tests (CPTs), and/or
other subsurface investigative techniques, and (c) laboratory testing to establish the soil types, properties, and layering. The
depth to rock or stiff soil material should be established from these investigations. Investigation should extend to bedrock or,
for very deep soil profiles, to material in which the model will be terminated. While it is preferable to measure shear wave
velocities in all soil layers, it is also possible to estimate shear wave velocities based on measurements available for similar
soils in the local area or through correlations with soil types and properties. A number of such correlations are summarized
by Kramer (1996).
Typically, a onedimensional soil column extending from the ground surface to bedrock is adequate to capture firstorder site
response characteristics. For very deep soils, the model of the soil columns may extend to very stiff or very dense soils at
depth in the column. Two or threedimensional models should be considered for critical projects when two or threedimensional
wave propagation effects may be significant (for example, sloping ground sites). The soil layers in a onedimensional
model are characterized by their total unit weights and shear wave velocities from which lowstrain (maximum)
shear moduli may be obtained, and by relationships defining the nonlinear shear stressstrain behavior of the soils. The
required relationships for analysis are often in the form of curves that describe the variation of soil shear modulus with shear
strain (modulus reduction curves) and by curves that describe the variation of soil damping with shear strain (damping
curves). In a two or threedimensional model, compression wave velocities or moduli or Poisson ratios also are required. In
an analysis to estimate the effects of liquefaction on soil site response, the nonlinear soil model also must incorporate the
buildup of soil pore water pressures and the consequent reductions of soil stiffness and strength. Typically, modulus
reduction curves and damping curves are selected on the basis of published relationships for similar soils (for example,
Vucetic and Dobry, 1991; Electric Power Research Institute, 1993; Darendeli, 2001; Menq, 2003; Zhang et al., 2005). Sitespecific
laboratory dynamic tests on soil samples to establish nonlinear soil characteristics can be considered where published
relationships are judged to be inadequate for the types of soils present at the site. Shear and compression wave velocities and
associated maximum moduli should be selected based on field tests to determine these parameters or, if such tests are not
possible, on published relationships and experience for similar soils in the local area. The uncertainty in the selected
maximum shear moduli, modulus reduction and damping curves, and other soil properties should be estimated (see
Darendeli, 2001; and Zhang et al., 2008). Consideration of the range of stiffnesses prescribed in Section 12.13.3 (increasing
and decreasing by 50 percent) is recommended.
C21.1.3 Site Response Analysis and Computed Results. Analytical methods may be equivalent linear or nonlinear.
Frequently used computer programs for onedimensional analysis include the equivalent linear program SHAKE (Schnabel et
al., 1972; Idriss and Sun, 1992) and the nonlinear programs FLAC (Itasca, 2005), DESRA2 (Lee and Finn, 1978), MARDES
(Chang et al., 1991), SUMDES (Li et al., 1992), DMOD_2 (Matasovic, 2006), DEEPSOIL (Hashash and Park, 2001), TESS
(Pyke, 2000), and OpenSees (Ragheb, 1994; Parra, 1996; Yang, 2000). If the soil response induces large strains in the soil
(such as for high acceleration levels and soft soils), nonlinear programs may be preferable to equivalent linear programs. For
analysis of liquefaction effects on site response, computer programs incorporating pore water pressure development (effective
stress analyses) should be used (for example, FLAC, DESRA2, SUMDES, DMOD, TESS, DEEPSOIL, and OpenSees).
Response spectra of output motions at the ground surface are calculated as the ratios of response spectra of groundsurface
motions to input outcropping rock motions. Typically, an average of the response spectral ratio curves is obtained and
multiplied by the input MCE response spectrum to obtain the MCE groundsurface response spectrum. Alternatively, the
results of siteresponse analyses can be used as part of the PSHA using procedures described by Goulet et al. (2007) and
programmed for use in OpenSHA (www.opensha.org; Field et al., 2005). Sensitivity analyses to evaluate effects of soilproperty
uncertainties should be conducted and considered in developing the final MCE response spectrum.
C21.2 GROUND MOTION HAZARD ANALYSIS
Uncertainties in the characterizations of the key seismic sources (tectonic provinces, zones of seismicity, and active faults),
with respect to location, earthquake recurrence, and maximum earthquake magnitude, must be considered in the ground
motion hazard analysis. Uncertainties in the groundmotion models are typically included by incorporating more than one
groundmotion attenuation equation. However, these equations may underestimate the intermediate and longperiod motion
from large earthquakes on nearby active faults due to directivity and directionality effects mentioned in C11.4.7. The
probabilistic seismic hazard analysis code can be modified to account for these effects in a consistent probabilistic manner, or
a deterministic adjustment can be made to the probabilistic MCE response spectrum using methods in Somerville et al.
(1997) and Abrahamson (2000) or more recent procedures. If the deterministic adjustment is used, then judgment must be
exercised in selecting the parameters comprising these methods. The worstcase scenario yielding the maximum possible
increase in motion from directivity/directionality effects is acknowledged to be conservative, but it offers an upperbound
solution to help gauge the appropriate level for the MCE response spectrum.
Siteresponse effects in PSHA generally should be evaluated by using the site term in the groundmotion prediction
equations. This term is generally a scale factor or a function of Vs30 = average shearwave velocity in the upper 30 meters.
Sitespecific dynamic response analyses can also be performed as described in Section C21.1.
C21.2.1 Probabilistic MCE. PSHA methods are sufficient to define the MCE ground motion at all locations except those
near highly active faults. Descriptions of current PSHA methods can be found in McGuire (2004).
C21.2.2 Deterministic MCE. Ground motions for the deterministic MCE shall be based on characteristic earthquakes on all
known active faults in a region. The magnitude of a characteristic earthquake on a given fault should be a best estimate of
the maximum magnitude capable for that fault but not less than the largest magnitude that has occurred historically on the
fault. The maximum magnitude should be estimated considering all seismicgeologic evidence for the fault, including fault
length and paleoseismic observations. For faults characterized as having more than a single segment, the potential for rupture
of multiple segments in a single earthquake should be considered in assessing the characteristic maximum magnitude for the
fault.
For consistency, the same attenuation equations used in the PSHA should be used in the DSHA. Adjustments for
directivity/directional effects should also be made, when appropriate. In some cases, groundmotion simulation methods may
be appropriate for the estimation of longperiod motions at sites in deep sedimentary basins or from great (M = 8) or giant
(M = 9) earthquakes, for which recorded groundmotion data are lacking.
As a point of clarification, the deterministic lower limit spectrum on the MCE (Figure 21.21) extends to zero period in the
same manner as the design response spectrum of Figure 11.41. The spectrum in Figure 21.21 is simply a schematic
illustrating the lower bounds for the constant spectral acceleration (SaM = 1.5Fa) and constant spectral velocity (SaM = 0.6Fv/T)
portions of the spectrum. The transition in the deterministic lower limit spectrum from the 1.5Fa plateau to zero period
occurs at a period (in seconds) of 0.08Fv/Fa which is derived in the same manner as T0 in Section 11.4.5. From this period to
zero period, where the ordinate is 0.6Fa, the deterministic lower limit spectrum is a straight line, similar to the design
response spectrum in the period band, 0 to T0.
C21.3 DESIGN RESPONSE SPECTRUM
Eighty percent of the design response spectrum determined in accordance with Section 11.4.5 was established as the lower
limit to prevent the possibility of sitespecific studies generating unreasonably low ground motions from potential
misapplication of sitespecific procedures or misinterpretation or mistakes in the quantification of the basic inputs to these
procedures. Even if sitespecific studies were correctly performed and resulted in groundmotion response spectra less than
the 80 percent lower limit, the uncertainty in the seismic potential and groundmotion attenuation across the United States
was recognized in setting this limit. Under these circumstances, the allowance of up to a 20 percent reduction in the design
response spectrum based on sitespecific studies was considered reasonable.
C21.4 DESIGN ACCELERATION PARAMETERS
The 90 percent lower limit rule, which can affect the determination of SDS, was inserted because it was recognized that sitespecific
studies could produce response spectra with ordinates at periods greater than 0.2 second that were significantly
greater than those at 0.2 second. Similarly, the rule that requires that SD1 be taken as the larger of the spectral acceleration at
a period of 1 second and two times the spectral acceleration at a period of 2 seconds accounts for the possibility that the
assumed 1/T proportionality for the constant velocity portion of the design response spectrum begins at periods greater than 1
second or is actually 1/T n (where n < 1). Thus, this rule leads to more accurate spectral ordinates at periods around 2 seconds
and conservative estimates at shorter periods. However, the conservatism is unlikely to be excessive.
REFERENCES
Abrahamson, N. A. 2000. “Effects of Rupture Directivity on Probabilistic Seismic Hazard Analysis,” in Proceedings of the
Sixth International Conference on Seismic Zonation, Palm Springs, California.
Chang, C.Y., C. M. Mok, M. S. Power, and Y. K. Tang. 1991. Analysis of Ground Response at Lotung LargeScale Soil
Structure Interaction Experiment Site, Report NP7306SL. Electric Power Research Institute, Palo Alto, California.
Darendeli, M. 2001. Development of a New Family of Normalized Modulus Reduction and Material Damping Curves,”
Ph.D. Dissertation, Department of Civil Engineering, University of Texas, Austin.
Electric Power Research Institute. 1993. Guidelines for Determining Design Basis Ground Motions, Report EPRI TR
102293. EPRI, Palo Alto, California.
Field, E. H., N. Gupta, V. Gupta, M. Blanpied, P. Maechling, and T. H. Jordan. 2005. “Hazard Calculations for the
WGCEP2002 Forecast Using OpenSHA and Distributed Object Technologies,” Seism. Res. Letters, 76:161167.
Goulet, C. A., J. P. Stewart, P. Bazzurro, and E. H. Field. 2007. “Integration of SiteSpecific Ground Response Analysis
Results into Probabilistic Seismic Hazard Analyses,” Paper 1486 in Proceedings of the 4th International Conference on
Earthquake Geotechnical Engineering, Thessaloniki, Greece.
Hashash, Y. M. A. and D. Park. 2001. "NonLinear OneDimensional Seismic Ground Motion Propagation in the
Mississippi Embayment," Engineering Geology, 62(13):185206.
Idriss, I. M., and J. I. Sun. 1992. User's Manual for SHAKE91. Center for Geotechnical Modeling, Department of Civil and
Environmental Engineering, University of California, Davis.
Itasca Consulting Group. 2005. FLAC, Fast Langrangian Analysis of Continua, Version 5.0. Itasca Consulting Group,
Minneapolis, Minnesota.
Kramer, S. L. 1996. Geotechnical Earthquake Engineering. Prentice Hall.
Kwok, A. O., J. P. Stewart, Y. M. A. Hashash, N. Matasovic, R. Pyke, Z. Wang, and Z. Yang. 2007. “Use of Exact
Solutions of Wave Propagation Problems to Guide Implementation of Nonlinear Seismic Ground Response Analysis
Procedures,” ASCE Journal of Geotechnical & Geoenvironmental Engineering, 133(11):13851398.
Lee, M. K. W., and W. D. L. Finn. 1978. DESRA2, Dynamic Effective Stress Response Analysis of Soil Deposits with
Energy Transmitting Boundary Including Assessment of Liquefaction Potential, Soil Mechanics Series 36. Department of
Civil Engineering, University of British Columbia, Vancouver, Canada.
Li, X. S., Z. L. Wang, and C. K. Shen. 1992. SUMDES, A Nonlinear Procedure for Response Analysis of Horizontally
Layered Sites Subjected to MultiDirectional Earthquake Loading. Department of Civil Engineering, University of
California, Davis.
Matasovic, N. 2006. “DMOD_2 – A Computer Program for Seismic Response Analysis of Horizontally Layered Soil
Deposits, Earthfill Dams, and Solid Waste Landfills,” User’s Manual. GeoMotions, LLC, Lacey, Washington, p. 20 (plus
Appendices).
McGuire, R. K. 2004. Seismic Hazard and Risk Analysis, Monograph, MNO10. Earthquake Engineering Research
Institute, Oakland, California, p. 221.
Menq, F. 2003. Dynamic Properties of Sandy and Gravely Soils, Ph.D. Dissertation, Department of Civil Engineering,
University of Texas, Austin.
Parra, E. 1996. Numerical Modeling of Liquefaction and Lateral Ground Deformation Including Cyclic Mobility and
Dilation Response in Soil Systems, Ph.D. Dissertation, Department of Civil Engineering, Rensselaer Polytechnic Institute,
Troy, NY.
Pyke, R. M. 2000. TESS: A Computer Program for Nonlinear Ground Response Analyses. TAGA Engineering Systems &
Software, Lafayette, California.
Ragheb, A. M. 1994. Numerical Analysis of Seismically Induced Deformations in Saturated Granular Soil Strata, Ph.D.
Dissertation, Department of Civil Engineering, Rensselaer Polytechnic Institute, Troy, New York.
Schnabel, P. B., J. Lysmer, and H. B. Seed. 1972. SHAKE: A Computer Program for Earthquake Response Analysis of
Horizontally Layered Sites, Report EERC 7212. Earthquake Engineering Research Center, University of California,
Berkeley.
Somerville, P. G., N. F. Smith, R. W. Graves, and N. A. Abrahamson. 1997. “Modification of Empirical Strong Ground
Motion Attenuation Relations to include the Amplitude and Duration Effects of Rupture Directivity” Seismological Research
Letters, 68:199222.
Vucetic, M., and R. Dobry. 1991. “Effect of Soil Plasticity on Cyclic Response,”
ASCE/SEI Journal of Geotechnical Engineering, 117(1):89107.
Yang, Z. 2000. Numerical Modeling of Earthquake Site Response Including Dilation and Liquefaction, Ph.D. Dissertation,
Department of Civil Engineering and Engineering Mech., Columbia University, New York, New York.
Zhang, J., R. D. Andrus, and C. H. Juang. 2005. “Normalized Shear Modulus and Material Damping Ratio Relationships,”
ASCE Journal of Geotechnical and Geoenvironmental Engineering, 131(4):453464.
Zhang, J., R. D. Andrus, and C. H. Juang. 2008. “Model Uncertainty in Normalized Shear Modulus and Damping
Relationships,” ASCE Journal of Geotechnical and Geoenvironmental Engineering, 134(1).
COMMENTARY TO CHAPTER 22,
SEISMIC GROUND MOTION AND
LONGPERIOD TRANSITION MAPS
SEISMIC GROUND MOTION MAPS
ASCE/SEI 705 continues to use contour maps of spectral response acceleration (Figures 221 through 2214). The spectral
acceleration design maps were prepared by the U.S. Geological Survey (USGS) based on USGS probabilistic maps of the 48
conterminous states (2002), Alaska (1998), Hawaii (1998), and Puerto Rico/Virgin Islands (2003) with modifications based
on the 1997 recommendations of the Building Seismic Safety Council. The maps of the 48 states and Puerto Rico/Virgin
Islands have been updated from the 2002 edition of the standard but the maps of Alaska, Hawaii, Guam, and Tutuila are
unchanged. The USGS also has developed a companion software program that calculates locationspecific spectral values
based on latitude and longitude or zip code; use of zip codes is discouraged in regions where groundmotion values vary
substantially over a short distance. The calculated values are based on the data used to prepare the maps. The spectral values
should be adjusted for Site Class effects using the Site Classification Procedure in Section 20 and the site coefficients in
Section 11.4. Latitude and longitude for a given address can be found at a variety of websites. The companion software
program may be accessed at the USGS website (http://earthquake.usgs.gov/designmaps). The software program should be
used to establish spectral values for design because the maps found in ASCE/SEI 705 are too small to provide accurate
spectral values for many sites.
LONGPERIOD TRANSITION MAPS
The maps of the longperiod transition period, TL, (Figures 2215 through 2220) were introduced in ASCE/SEI 705. They
were prepared by the USGS in response to BSSC recommendations and subsequently included in the 2003 edition of the
Provisions. See Section C11.4.5 for a discussion of the technical basis of these maps. The value of TL obtained from these
maps is used in Equation 11.47 to determine values of Sa for periods greater than TL.
The exception in Section 15.7.6.1, regarding the calculation of Sac, the convective response spectral acceleration for tank
response, is intended to provide the user the option of computing this acceleration with three different types of sitespecific
procedures: (a) the procedures in Chapter 21, provided they cover the natural period band containing Tc, the fundamental
convective period of the tankfluid system, (b) groundmotion simulation methods using seismological models, and (c)
analysis of representative accelerogram data. Elaboration of these procedures is provided below.
With regard to the first procedure, attenuation equations have been developed for the western United States (Next Generation
Attenuation, Power et al., 2006, 2008) and for the central and eastern United States (e.g., Somerville et al., 2001) that cover
the period band, 0 to 10 seconds. Thus, for Tc = 10 seconds, the fundamental convective period range for nearly all storage
tanks, these attenuation equations can be used in the same PSHA/DSHA procedures described in Chapter 21 to compute Sa
(Tc). The 1.5 factor in Equation 15.711, which converts a 5 percent damped spectral acceleration to a 0.5 percent damped
value, could then be applied to obtain Sac. Alternatively, this factor could be established by statistical analysis of 0.5 percent
damped and 5 percent damped response spectra of accelerograms representative of the ground motion expected at the site.
In some regions of the United States, such as Pacific Northwest and southern Alaska, where subductionzone earthquakes
dominate the groundmotion hazard, attenuation equations for these events only extend to periods between 3 and 5 s,
depending on the equation. Thus, for tanks with Tc greater than these periods, other sitespecific methods are required.
The second sitespecific method to obtain Sa at long periods is simulation through the use of seismological models of fault
rupture and wave propagation (Graves and Pitarka, 2004; Hartzell and Heaton, 1983; Hartzell et al., 1999; Liu et al., 2006;
Zeng et al., 1994). These models could range from simple seismic sourcetheory and wavepropagation models, which
currently form the basis for many of the attenuation equations used in the central and eastern United States for example, to
more complex numerical models that incorporate finite fault rupture for scenario earthquakes and seismic wave propagation
through 2D or 3D models of the regional geology, which may include basins. These models are particularly attractive for
computing longperiod ground motions from great earthquakes (Mw = ~ 8) because groundmotion data are limited for these
events. Furthermore, the models are more accurate for predicting longerperiod ground motions because: (a) seismographic
recordings may be used to calibrate these models and (b) the general nature of the 2D or 3D regional geology is typically
fairly well resolved at these periods and can be much simpler than would be required for accurate prediction of shorter period
motions.
A third sitespecific method is the analysis of the response spectra of representative accelerograms that have accurately
recorded longperiod motions to periods greater than Tc. As Tc increases, the number of qualified records decreases.
However, as digital accelerographs continue to replace analog accelerographs, more recordings with accurate longperiod
motions will become available. Nevertheless, a number of analog and digital recordings of large and great earthquakes are
available that have accurate longperiod motions to 8 seconds and beyond. Subsets of these records, representative of the
earthquake(s) controlling the groundmotion hazard at a site, can be selected. The 0.5 percent damped response spectra of the
records can be scaled using seismic source theory to adjust them to the magnitude and distance of the controlling earthquake.
The levels of the scaled response spectra at periods around Tc can be used to determine Sac. If the subset of representative
records is limited, then this method should be used in conjunction with the aforementioned simulation methods.
REFERENCES
Graves, R. W., and A. Pitarka. 2004. “Broadband Time History Simulation using a Hybrid Approach,” Paper 1098 in
Proceedings of the 13th World Conference on Earthquake Engineering, Vancouver, Canada.
Hartzell, S., and T. Heaton. 1983. "Inversion of Strong Ground Motion and Teleseismic Waveform Data for the Fault
Rupture History of the 1979 Imperial Valley, California Earthquake," Bulletin of the Seismological Society of America,
73:15531583.
Hartzell, S., S. Harmsen, A. Frankel, and S. Larsen. 1999. "Calculation of Broadband Time Histories of Ground Motion:
Comparison of Methods and Validation Using Strong Ground Motion from the 1994 Northridge Earthquake," Bulletin of the
Seismological Society of America, 89:14841504.
Liu, P., R. J. Archuleta, and S. H. Hartzell. 2006. “Prediction of Broadband GroundMotion Time Histories: Hybrid
Low/HighFrequency Method with Correlated Random Source Parameters,” Bulletin of the Seismological Society of
America, 96:2118–2130.
Power, M., B. Chiou, N. Abrahamson, Y. Bozorgnia, T. Shantz, and C. Roblee. 2008. “An Overview of the NGA Project,”
Earthquake Spectra Special Issue on the Next Generation of Ground Motion Attenuation (NGA) Project.” Earthquake
Engineering Research Institute, March.
Power, M., B. Chiou, N. Abrahamson, and C. Roblee. 2006. “The Next Generation of Ground Motion Attenuation Models,”
in Proceedings of the 100th Anniversary Earthquake Conference Commemorating the 1906 San Francisco Earthquake, San
Francisco, California.
Somerville, P. G., N. Collins, N. Abrahamson, R. Graves, and C. Saikia. 2001. Earthquake Source Scaling and Ground
Motion Attenuation Relations for the Central and Eastern United States, Final Report to the USGS under Contract
99HQGR0098.
Zeng, Y., J. G. Anderson, and G. Yu. 1994. "A Composite Source Model for Computing Synthetic Strong Ground
Motions," Geophys. Research Letters, 21:725728.