Sie sind auf Seite 1von 14


James O. Malley
Senior Principal, Degenkolb Engineers

* RSA procedure developed in 1950s (or earlier) as a means of

addressing dynamic characteristics of the structure through modal
analysis approach
* Code History
* 1985 UBC only mentioned for irregular structures, with no direction
* 1988 UBC first time dynamic provisions introduced. Many of the concepts
in ASCE 7-10 date back to this time
* Required for irregular buildings
* Reduced forces when compared to ELF designs for regular buildings
* 90% rule for mass participation
* A lot has changed since 1988 (or 1950)!
* Computing power
* Better understanding of structural response, especially in the nonlinear
* Issue 90% Mass Participation Rule
* Instituted many years ago when computing power was
fraction of present
* This may result in not properly capturing local demands on
stiff elements
* Not a major issue for overall response in most cases
* Approach Follow ASCE 4 Procedures
* Capture all modes down to T=0.05 seconds
* Calculate Residual Mass for shorter period modes
* Apply Sa at T=0.5 sec to Residual Mass
* Sa=Sds(0.4 + 0.15/Ts )
* Exception allows old 90% rule to be used
* Result All mass is included in the design
* Avoids underestimating design forces on some elements
* Implication Better captures full dynamic response of
* Not critical for most building structures
* May apply more for non-building structures
* Modern software already can perform Residual Mass
Revise Section 12.9.1 as follows:
12.9.1 Number of Modes
An analysis shall be conducted to determine the
modes of vibration for the structure. The analysis shall include
sufficient modes to capture participation of 100% of the
structures mass. For this purpose, It shall be permitted to
represent all modes with periods less than 0.05 seconds in a
single rigid body mode having a period of 0.05 seconds.
EXCEPTION: Alternatively, the analysis shall be permitted to
include a sufficient number of modes to obtain a combined
modal mass participating of at least 90 percent of the actual
mass in each orthogonal horizontal direction of response
considered in the model.

*Issue Existing Commentary is out of date

* Missing updated references
* Does not include updated procedure (CQC3)
*Approach Reflect updates and new procedure
* CQC3 Method (Menun and Der Kiureghian, 1998)
provides critical orientation of structure
* Endorsed by Wilson for multi-component analyses

* Result Brings commentary up to date with modern

software applications that incorporate this approach for
modal combination
* . 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 et al.
(20001995) suggested 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. Menun and Der Kiureghian (1998) proposed an
alternate method, referred to as CQC3, that provides the
critical orientation of the earthquake relative to the structure.
Wilson (2000) endorsed the CQC3 method for combining the
results from multiple component analyses.
*Issue 85% Rule for Scaling to ELF Base Shear
*Instituted many years ago to encourage use of
* Thought to provide better distribution of forces,
and therefore better performing structures
*Recent studies by ATC for FEMA P695 and NIST
GCR10-917-8 and NIST GCR12-917-20 indicate
increased collapse probability for RSA design
when compared to ELF
* Effect exacerbated for long period structures
* Equivalent performance indicated when scale to 100%
*Approach Scale to 100% of ELF for both forces
and drifts
*ELF design will be used more often (RSA still
required for irregular structures)
*More uniform collapse probability
*Implication May drive the more use of NLRH
analyses to reduce structural frame costs
Revise Section 12.9.4 as follows:

12.9.4 Scaling Design Values of Combined Response

A base shear (V) shall be calculated in each of the two orthogonal horizontal
directions using the calculated fundamental period of the structure T in each
direction and the procedures of Section 12.8. Scaling of Forces
Where the calculated fundamental period exceeds CuTa in a given direction, CuTa
shall be used in lieu of T in that direction. Where the combined response for the
modal base shear (Vt) is less than 100 percent of the calculated base shear (V)
using the equivalent lateral force procedure, the forces shall be multiplied by
V/Vt, where:
V = the equivalent lateral force procedure base shear, calculated in
accordance with this section and Section 12.8
Vt = the base shear from the required modal combination Scaling of Drifts
Where the combined response for the modal base shear (Vt) is less than CsW,
and where Cs is determined in accordance with Eq. 12.8-6, drifts shall be
multiplied by CsW/Vt.

* Issue Use of 3D Models for RSA

* Modern analysis tools can pick up 3D effects automatically
* Use of rigid and semi-rigid diaphragms to capture 3D response
* Approach Requires 3D model for all RSA designs, either
rigid or semi-rigid diaphragms
* Result Better overall representation of structural
* Use of 2D models limited to ELF or preliminary analyses
* Implication May result in more use of ELF for simple

12.9.8 Structural Modeling

A mathematical model of the structure shall be constructed in accordance
with Section 12.7.3, except that all structures designed in accordance with
this Section shall be analyzed using a 3-D representation. Where the
diaphragms have not been classified as rigid in accordance with Sections
12.3.1, the model shall include representation of the diaphragms stiffness
characteristics and such additional dynamic degrees of freedom as are
required to account for the participation of the diaphragm in the structures
dynamic response.

Proposed new Commentary:

Using modern software, it often is more difficult to decompose a structure
into planar models than it is to develop a full three-dimensional model. As a
result, three-dimensional models are now commonplace. Increased
computational efficiency also allows efficient modeling of diaphragm
flexibility. As a result, when modal response spectrum analysis is used, a
three-dimensional model is required for all structures, including those with
diaphragms that can be designated as flexible.
* Divide only T1 response by R (other modes respond elastically)
* Adjust Cs in the short period range to account for higher
displacement demands (equal displacement rule doesnt apply)
* Should we scale only to base shear? (also to OTM, e.g.)
* Should a factor be applied to account for multi-mode effects?
* Revise spectra development to multi-point approach
* Incorporate accidental torsion into procedure
* Modify to better approximate 1% in 50 year collapse probability
* Should RSA be re-worked to allow de-coupling from ELF?
* Should RSA be replaced by LRHA???
* Another proposal was also approved to update LRHA Provisions to
make them more widely applicable.