Vous êtes sur la page 1sur 30

CEE 243B – RESPONSE AND DESIGN OF RC STRUCTURAL SYSTEMS

4a. Nonlinear Structural Analysis


for Seismic Design

NIST GCR 10-917-5

CEE 243B - Spring 2018


Nonlinear Structural Analysis for Seismic Design
Response and Modeling of Structural Components

2
G. Deierlein - 2014 EERI Technical Seminar Series
Nonlinear Structural Analysis for Seismic Design
Model Types

Idealized structural component models


..enforce predefined behavioral modes ..model underlying physics of materials
Phenomenological Fundamental

RC Column Test

macro-models micro-models

NEHRP Seismic Design Technical Brief No. 4, NIST GCR 10-917-5


PEER ATC 72-1 3
Nonlinear Structural Analysis for Seismic Design
Model Types

Idealized structural component models

Phenomenological Fundamental

• Simplest models..
• Concentrate inelastic deformations at ends of the element..
• Condensed, numerically efficient formulations..

4
Nonlinear Structural Analysis for Seismic Design
Model Types

Idealized structural component models

Phenomenological Fundamental

• Designated hinge zones at member ends..


• Inelastic hinge zones can be defined by nonlinear M-φ relations
• Integration of deformations along hinge length captures spread
of yielding more realistically..

5
Nonlinear Structural Analysis for Seismic Design
Model Types

Idealized structural component models

Phenomenological Fundamental

• Plasticity distributed through member cross section and along member length..
• Uniaxial material models (stress-strain) relations defined - Uniaxial material
“fibers” integrated over the cross section to obtain stress resultants (axial force
and moments)
• Plane sections remain plane enforced..
6
Nonlinear Structural Analysis for Seismic Design
Model Types

Idealized structural component models

Phenomenological Fundamental

• Most complex models, numerous input parameters..


• Modeling at the fundamental level..
• Discretize continuum along member length and cross section into small
(micro) finite elements with nonlinear hysteretic constitutive properties.

7
Nonlinear Structural Analysis for Seismic Design
Analysis Software

Commonly used commercially available software:


• PERFORM 3D, Nonlinear Analysis and Performance Assessment for 3D Structures (CSI)
• SAP2000, Integrated Software for Structural Analysis and Design (CSI)
• ETABS, Extended Three Dimensional Analysis of Building Systems (CSI)
• LARSA 4D, Advanced Software for the Analysis and Design of Bridges and Structures (LARSA)
These programs generally support use of concentrated plasticity models (i.e., plastic hinges,
inelastic springs) and, to some extent, fiber-type beam-column and flexural wall models.
Although not common, some consulting firms utilize more advanced commercially available
software, including:
• Abaqus (Dassault), ANSYS (ANSYS), DIANA (TNO), LS-DYNA (LSTC).

Commonly used research software :


• OpenSees, Open System for Earthquake Engineering Simulation, UC Berkeley (OpenSees)
• RUAUMOKO, University of Canterbury, New Zealand
• IDARC-2D, University of Buffalo
• SeismoStruct, SeismoSoft Read the
Manual!
8
Nonlinear Structural Analysis for Seismic Design
Model Types

Macro-scale phenomenological models (concentrated plasticity models)


… simplified representation of overall force-deformation component response
… based on observed behavior / test data

Advantages :
➢ conceptually simple, easily parameterized
➢ easily calibrated to test data
➢ uniaxial phenomenological hysteretic hinge models have been a mainstay approach in
seismic response analysis for many years, and have been developed to the point of
capturing cyclic strength and stiffness degradation fairly well
➢ less computationally demanding than fiber and finite element models, relatively robust.

Limitations:
➢ Although such models are able to capture nonlinear cyclic degradation well, it is difficult
to extend their rule-based formulations to multi-axial response.
➢ P-My-Mz and P-M-V interaction with cyclic degradation difficult to capture.
➢ Limitations associated with ASCE 41 force-displacement relationships commonly used
in definition of concentrated hinge models.

NIST GCR 14-917-27 9


Nonlinear Structural Analysis for Seismic Design
Model Types

Macro-scale phenomenological models (concentrated plasticity models)

capping point
monotonic
vs. cyclic
response
envelope

Inelastic hinge model for RC beam-column elements:


 Response idealized by a backbone curve that relates moment to rotation in the concentrated hinges
 Backbone curve generally calibrated to test data (e.g. ASCE 41, PEER database: 250+ columns)
 Key parameters of inelastic hinge model:
• Strength
• Initial stiffness
• Post-yield stiffness
• Plastic rotation (capping) capacity
• Post-capping slope
• Cyclic behavior (hysteresis model)
PEER ATC 72-1 10
Nonlinear Structural Analysis for Seismic Design
Model Types

Macro-scale phenomenological models (concentrated plasticity models)

capping point
monotonic
vs. cyclic
response
Backbone
envelope
curve in model

Do we use the monotonic curve


or the cyclic envelope as the
backbone curve in a plastic
Definitions… hinge model??
 Backbone Curve: Relationship between generalized force
and deformation of a structural component that is used to
characterize response in a nonlinear analysis model.
 Monotonic Curve: Curve of generalized force vs. deformation
data obtained from monotonic loading of a structural component
 Cyclic Envelope: Curve of generalized force vs. deformation
that envelopes response data obtained from cyclic loading of a
structural component
Why are monotonic and cyclic
curves different??
PEER ATC 72-1 & NIST GC 10-917-5 11
Nonlinear Structural Analysis for Seismic Design
Model Types

Macro-scale phenomenological models (concentrated plasticity models)

capping point
monotonic
vs. cyclic
response
Backbone
envelope
curve in model

Monotonic Curve vs. Cyclic Envelope…


Except at small deformations, the cyclic envelope curve
generally falls below the monotonic loading curve.
The cyclic envelope curve is loading history dependent. The
number of cycles used in the loading protocol, the amplitude of
each cycle, and the sequence of loading cycles can result in
different envelopes. While the monotonic curve for a given
component is nearly unique, the cyclic envelope curve is not.
 Standardized loading protocols exist to facilitate
Normalized Δ

interpretation and comparison of test data.

PEER ATC 72-1 & NIST GC 10-917-5 12


Number of Cycles
Nonlinear Structural Analysis for Seismic Design
Model Types

Effect of Loading History on Component Response

NIST GCR 14-917-27 13


Nonlinear Structural Analysis for Seismic Design
Model Types

Effect of Loading History on Component Response

Different hysteretic responses for identical structural members


subjected to different earthquake time histories 14
G. Deierlein - 2014 EERI Technical Seminar Series
Nonlinear Structural Analysis for Seismic Design
Model Types

Macro-scale phenomenological models (concentrated plasticity models)

capping point
monotonic
vs. cyclic
response
Backbone
envelope
curve in model
Do we use the monotonic curve
or the cyclic envelope as the
backbone curve in a plastic
Depends on the analytical model.. hinge model??

Option 1 Cyclic deterioration is explicitly incorporated in the analytical model; use the monotonic
curve as a reference (initial) backbone curve; backbone curve updated (moving “inward”-towards the
origin) during analysis as a function of the loading history (e.g. Ibarra-Krawinkler).
Options 2-3Use of cyclic envelope curve as the backbone curve; backbone curve is fixed- not
updated during the analysis; cyclic envelope based on measured data (Option 2) or approximate
modification of the monotonic curve (Option 3).
Option 4 Analytical model does not capture post-capping (negative tangent stiffness) behavior; use
the cyclic envelope but limit ultimate deformation to that associated with 80% of the capping
strength on the descending branch of the cyclic envelope obtained using Option 2 or 3.
PEER ATC 72-1 & PEER/TBI 2010 15
Nonlinear Structural Analysis for Seismic Design
Model Types

Macro-scale phenomenological models (concentrated plasticity models)

Backbone Curves for Alternative Model Types


PEER ATC 72-1 16
Nonlinear Structural Analysis for Seismic Design
ASCE/SEI 41 Generalized Force-Deformation Curves

17
Nonlinear Structural Analysis for Seismic Design
ASCE/SEI 41 Generalized Force-Deformation Curves

The generalized force-deformation


curves in ASCE 41 are calibrated to
the cyclic response envelope for
components subjected to a
standardized, symmetric loading
history.

Generalized force-deformation curves in ASCE 41:


• primarily intended for use with nonlinear static analysis  lack information for defining cyclic
loading effects.
• primarily intended for existing building components  parameters and acceptance criteria not
necessarily appropriate for performance expectations in new building design.
• related almost exclusively to generalized hinge or nonlinear spring type models unclear how
criteria relate to more detailed fiber or continuum models.
• almost exclusively uniaxial models, except that in some cases, model parameters differentiate
certain response modes as a function of other force components (e.g., hinge rotation parameters in
concrete beams are specified as a function of the shear force in the member) in recognition of
interaction effects.
• criteria do not include consideration of the expected variation (dispersion) in modeling parameters
or response prediction. 18
Nonlinear Structural Analysis for Seismic Design
ASCE/SEI 41 Generalized Force-Deformation Curves

Comparison of effective stiffness Note variability


values of RC beam-columns in test results
from Haselton et al. (2008)
and Elwood et al. (2007)
with test results
Ke [PEER ATC 72-1]

(same in ASCE 41-13)

Note that effective flexural


stiffness values in ASCE 41 are also
intended to take into account the
deformations associated with
bond-slip into the member end
anchorages.

For columns with low axial loads


(<0.1Agfc’) deformations caused by
bar slip can account for as much as
0.35EcIg 50% of total deformations at yield
in ASCE 41-17
19
Nonlinear Structural Analysis for Seismic Design
Model Types

Macro-scale phenomenological models (concentrated plasticity models)

Examples of
uniaxial
hysteretic
response
models

Selection based on:


-- software used
-- engineering judgment

20
NEHRP Seismic Design Technical Brief No. 4, NIST GCR 10-917-5
Nonlinear Structural Analysis for Seismic Design
Model Types

Macro-scale phenomenological models (concentrated plasticity models)


when yielding
initiates further upon reloading
elastic loading stiffness
deformation occurs at deformation occurs at
and yield strength
positive, non-zero a reduced stiffness
remain constant
stiffness

Examples of
uniaxial
hysteretic
response
each successive cycle
initiates yielding at a
models
lower force level

in successive cycles of
yielding, an increasingly
negative post-yield stiffness
occurs

Recommended Modeling Parameters and Acceptance Criteria for Nonlinear Analysis


in Support of Seismic Evaluation, Retrofit, and Design , NIST GCR 17-917-45 21
Nonlinear Structural Analysis for Seismic Design
Model Types

Cyclic versus In-cycle strength degradation

negative
post-
Same
yield
backbone stiffness
curves but
different
behavior..

..apparent loss in .. loss in strength


strength at a under increasing
given deformations within
deformation one loading
level under excursion of a cycle
reversed cyclic or under monotonic
loading loading
22
NIST GCR 10-917-5 & G. Deierlein - 2014 EERI Technical Seminar Series
Nonlinear Structural Analysis for Seismic Design
Model Types

Sources of Deterioration
In reinforced concrete components, the following phenomena can cause deterioration:

• Concrete tensile cracking


• Concrete crushing and spalling
• Rebar buckling and fracture
• Bond slip
• Reduction in confinement due to fracture of transverse reinforcement
• Loss of anchorage of transverse reinforcement
• Reduction in aggregate interlock
• Diagonal tension and horizontal shear cracking
• Sliding at cracked interfaces and construction joints
• Insufficient crack closure

Modeling of deterioration is a complex issue. In most practical cases reliance must be placed on
experimental evidence, and simplifications must be made. It is a matter of engineering judgment to
decide when a source of deterioration is a sufficiently important contributor to response, and must
be incorporated in the analytical model. 23
Nonlinear Structural Analysis for Seismic Design
Model Types

Sources of Deformation/Limits on ductile behavior


Reinforced Concrete Moment Frames:
Reliable nonlinear behavior can occur through:
• ductile flexural hinging of beams and limited ductile flexural hinging of columns accompanied
by bar slip at joints
Limits on ductile behavior include:
• buckling, and under repeated cycles, fracturing of longitudinal reinforcing steel
• crushing and spalling of cover concrete
• at extreme deformations, fracture of confining reinforcement followed by crushing and
spalling of member cores.

Undesirable, non-ductile failure modes that prevent significant nonlinear frame response
include:
• member and joint shear failure
• reinforcing splice failure
• reinforcing pull-out at joints
24
• slab column punching shear failure
Nonlinear Structural Analysis for Seismic Design

The 2010 Tests: Full scale 4-Story RC and PT Buildings


RC
PT

RC PT

X Y

▪ Longitudinal direction LFRS :


3.0m
two-bay moment frames
▪ Transverse direction LFRS :
structural walls coupled to corner
25
columns by beams
Nonlinear Structural Analysis for Seismic Design

The 2010 Tests: Full scale 4-Story RC Building


g

26
Nonlinear Structural Analysis for Seismic Design

The 2010 Tests: Full scale 4-Story RC Building

27
Nonlinear Structural Analysis for Seismic Design

The 2010 Tests: Full scale 4-Story RC Building

28
Nonlinear Structural Analysis for Seismic Design

The 2010 Tests: Full scale 4-Story PT Building

29
RC Wall δ δ
UPT Wall

Lateral load (F)


Lateral load (F)
F F

Residual
displacement

Lateral displacement (δ) Lateral displacement (δ)

Appropriate
hysteretic model
to use depends
RC Wall on actual UPT Wall
characteristics of
element being
represented

Marked difference in performance 30

Vous aimerez peut-être aussi