NONLINEAR ANALYSES OF

STRUCTURAL SYSTEMS

Main Aspects of Nonlinear

Behavior
D
Spring 2015

Prof.Dr. Bülent AKBAŞ

Departmant of Earthquake and Structural Engineering
 Outline

• Important aspects of nonlinear behavior

• Strength-based vs. deformation-based design

• Performance-based design

• Capacity design

• Modeling for practical nonlinear analyses

• Limit states and demand/capacity ratios

• Dynamic analyses vs. static push-over analyses

 Outline

• Important aspects of nonlinear behavior

• Strength-based vs. deformation-based design

• Performance-based design

• Capacity design

• Modeling for practical nonlinear analyses

• Limit states and demand/capacity ratios

• Dynamic analyses vs. static push-over analyses

 Important Aspects of Nonlinear Behavior

Main aspects of Force (F) – Deformation (D) Relationship:

Force (F) Ultimate

There can be substantial Strength
uncertainty in the strength
Strain Ductile
and other properties
Hardening Limit

Strength
First Yield Loss

Initial Residual
Stiffness Strength

Deformation (D)
Hysteresis Loop
 Important Aspects of Nonlinear Behavior

Uncertainity in strength and other properties: Cyclic Degradation

Force (F)
in the first cycle,
there may be little or
no degradation

Initial After a few cycles, the

Stiffness strength and stiffness
may degrade

Deformation (D)
 Important Aspects of Nonlinear Behavior

Uncertainity in strength and other properties: Cyclic Strength Gain

Force (F)
the strength may increase with
cyclic deformation

Initial
Stiffness

Deformation (D)
 Important Aspects of Nonlinear Behavior

Uncertainity in strength and other properties: Effect of Strength Loss

Force (F)

deformation reverses after

strength loss

Deformation (D)

Strength in opposite direction

may or may not be reduced
 Important Aspects of Nonlinear Behavior

Uncertainity in strength and other properties: Loop Shape

F F

D D

non-degrading model stiffness degrading model

F F

D D

pinched buckling
 Important Aspects of Nonlinear Behavior

What does F-D mean for a structural component and structure?

Lateral Distribution of D
Seismic Force

F
D

F
F
F – D Relationship for a Component
(F=axial load, moment, shear force,
D=corresponding deformation)
Seismic Force (should be obtained experimentally and must
be known prior to structural analyses)
F – D Relationship for a Structure
(F=load, D=deflection)
(can be obtained by structural analyses)
 Important Aspects of Nonlinear Behavior

Structural Components in a Moment Frame:

 Important Aspects of Nonlinear Behavior

Structural Components in a CBF:

 Important Aspects of Nonlinear Behavior

Structural Components in an EBF:

 Important Aspects of Nonlinear Behavior

Structural Components in a Special Plate Shear Wall

 Important Aspects of Nonlinear Behavior

Structural Components in a R/C Shear Wall

 Outline

• Important aspects of nonlinear behavior

• Strength-based vs. deformation-based design

• Performance-based design

• Capacity design

• Modeling for practical nonlinear analyses

• Limit states and demand/capacity ratios

• Dynamic analyses vs. static push-over analyses

 Strength-Based vs. Deformation-Based Design

Strength-Based Design:
F Nominal Strength, Fn
F= force (axial force, bending
(Capacity)
moment, shear)

Basic Design Equation:

Design Strength, fFn
Demand  Capacity
Demand at factored load
Demand at working load

D
Basic Steps in Strength-Based Design:

a. Check demand vs. capacity at the component level.

b. If the demand at factored load level is smaller than the design strength, the design is O.K.
c. We can use linear analyses to calculate demands, because the structure should remain
essentially linear at factored load level.
 Strength-Based vs. Deformation-Based Design

Deformation-Based Design:

How does the structure behave when a major earthquake ground motion hits
the structure:

Demand on an elastic structure can

significantly exceed the capacity

Deformation
demand on an
inelastic structure
may not be that
excessive

D
 Strength-Based vs. Deformation-Based Design

Deformation-Based Design:
Design of components with significant ductile behavior:
F

Deformation demand (should

be calculated) Deformation capacity
based on available
ductility

D
Basic Steps in Strength-Based Design:

a. Calculate the deformation demand at the component and/or structure level.

b. If the deformation demand is smaller than the deformation capacity for all components, the
design is O.K.
c. We have to use nonlinear analyses to calculate deformation demands, because the structure
can yield.
 Strength-Based vs. Deformation-Based Design

Deformation-Based Design:
Design of components with little or no ductile behavior (brittle behavior):
F

D
a. Components with brittle behavior have little or no ductile capacity beyond yield.
b. Design of such components must be based on strength-based design.
 Strength-Based vs. Deformation-Based Design

Deformation-Based Design:

General Design Principles:

a. Deformation capacities for ductile components should be defined.

b. Strength capacities for brittle components and the components that should remain
essentially elastic should be defined.
c. Nonlinear structural analyses should be carried out to estimate the deformation and
strength demands.
d. The performance of the components and the structure is said to be O.K. if all Demand (D) /
Capacity (C) ratios are less than or equal to 1.0.

Appearances sometimes may be deceiving !...

These very straigtforward design principles seem to be simple, but when it comes to practice, we
do have difficult times.
 Outline

• Important aspects of nonlinear behavior

• Strength-based vs. deformation-based design

• Performance-based design

• Capacity design

• Modeling for practical nonlinear analyses

• Limit states and demand/capacity ratios

• Dynamic analyses vs. static push-over analyses

 Performance-Based Design

General Principles:

a. Traditional design codes do not explicity establish a performance level.

b. Seismic design codes imply that if you follow the seismic design guidelines as set
forth by them, the structure would not collapse during a strong earthquake
ground motion. This is not explictly stated in any seismic design codes.
c. Performance-based design tries to make sure that a design satisfies a specific
performance level with a reasonable margin of safety.
d. The most common structural performance levels are:
-Immediate Occupancy (IO) : little or no damage is expected.
-Life Safety (LS) : some damage, few or no injuries are expected.
-Collapse Prevention (CP) : more damage, but no collapse is expected.
 Performance-Based Design

LS CP
IO
Note: Structural engineers
might define other
Substantial More damage, performance levels.
inelastic but no collapse
Little (just before the
behavior
inelastic onset of ductile
behavior capacity of the
component)

D
 Performance-Based Design

Main Steps for Performance-Based Design:

a. Select the performance level and the design loads.

b. Define the demand and capacity measures.
-drift, plastic hinge rotation, shear strength, etc.
c. Find the deformation and strength capacities:
-Strength : AISC, ACI, TS500, etc. Codes.
-Deformation: ASCE 41 (FEMA 356, TEC), experiments.
d. Calculate the deformation and strength demands:
Use structural analyses to come up with the deformation and strength demands.
e. Calculate demand / capacity (D/C) ratios. If D/C > 1:
-change the design,
-try to find a way to reduce the demand,
-increase the capacity.
 Performance-Based Design

ASCE 41 (FEMA 356):

a. It is used for seismic rehabilitation of exsiting buildings based on performance-

based design principles.
b. It can also be used for new designed buildings.
c. It provides useful information for modeling nonlinear analysis.
d. It provides deformation capacities for a variety of structural components for the
IO, LS and CP performance levels.

It should be noted that the modeling guidelines in ASCE 41 are quite simplistic and
the proposed deformation capacities might be considered quite conservative. It is,
however, a very useful resource, but not a definiteve set of rules.
 Outline

• Important aspects of nonlinear behavior

• Strength-based vs. deformation-based design

• Performance-based design

• Capacity design

• Modeling for practical nonlinear analyses

• Limit states and demand/capacity ratios

• Dynamic analyses vs. static push-over analyses

 Capacity Design

Some aspects of engineering demand parameters:

a. We have to use engineering demand parameters and establish demand-capacity

measures.
b. Performance assessment should be based on those measures. Other analyses
(modal, dead, etc.) results shoud also be examined very carefully.
c. There may be subjective judgements in choosing demand/capacity meaures,
establishing capacity values, and calculating demand values. However, the overall
process must be objective.
d. We should not seek for a perfect performance assessment procedure, because
there is no such way.
 Capacity Design

Capacity Design Concept:

a. Some components in a structure should be selected as fuse elements, e.g. they should be
allowed to yield. Other structural members and components should remain elastic.
b. If we do not know which components yield and which do not, then we have to rely on
analysis to tell us. This is not an educated engineer’s approach and can be quite dangerous.
The model we are analyzing is not a perfect model. We should alway remember that we are
not analyzing the actual structure.
c. An educated structural engineer decides, in advance, which components can yield and which
must remain elastic.
d. After this decision, we must design the yielding components to have suficient ductility, and
the elastic components to have sufficient strength.

Capacity design allows :

-setting up more reliable nonlinear models,
-calculating more reliable demands and D/C ratios,
-the final design is likely to be more forgiving.
We may not able to achive thsese goals all the time, but we can get there.
 Capacity Design

Capacity Design on Frame Structures:

a. Specify the components that can yield. Use deformation D/C.

- beams in bending.
-columns at base and at roof.
b. The remaining components should not yield. Use strength D/C .
-columns except at base and roof.
-connections.
-beams and columns in shear.
-foundation.
 Capacity Design

Capacity Design on Shear Walls:

a. Specify the components that can yield. Use deformation D/C.

- Flexural hinging at base of wall.
-Shear in coupling beams.
b. The remaining components should not yield. Use strength D/C .
-bending in wall at all levels except the base.
-shear in wall at all levels, especially in hinge region.
-foundation.
 Capacity Design

Do we need nonlinear analyses for capacity design?

a. We do not have to carry out nonlinear analyses for capacity design. The basic requirement is
that all D/C ratios be < 1.
b. Fuse elements (components that are allowed to yield) can be detailed to provide deformation
capacities that the structural engineer believes are sufficient, using judgement of simplified
analyses.

However, nonlinear analyses gives more accurate deformation and strength demands. Judgement
or simple analyses can lead to over-design, or if the judgement is faulty, to under-design.
 Outline

• Important aspects of nonlinear behavior

• Strength-based vs. deformation-based design

• Performance-based design

• Capacity design

• Modeling for practical nonlinear analyses

• Limit states and demand/capacity ratios

• Dynamic analyses vs. static push-over analyses

 Modeling for Practical Nonlinear Analyses

What is our goal of structural analyses?

a. Always keep in mind that structural analyses is only a tool for use in design. The design does
not end there.
b. Our goal is only to simulate the behavior of the structure. Exact simulation is impossible.
c. Our goal is to get D/C ratios that are accurate enough for use in decision making.

Structural engineer never seeks for an exact simulation. Accurate enough simulation is more than
enough in decision making and much easier.
 Modeling for Practical Nonlinear Analyses

Main aspects of nonlinear behavior:

Force (F) Ultimate

Strength
Strain Ductile
Hardening Limit

Strength
First Yield Loss

Initial Residual
Stiffness Strength

Deformation (D)
Hysteresis Loop
 Modeling for Practical Nonlinear Analyses

F-D relationhip for analysis:

Since earthquake causes cyclic loading on fuse

Force (F) Initial relationship elements, we use degraded relationhip in
for monotonic load
current practice, e.g. we allow for a reasonable
amount of cycling.
In the future, we may have ‘initial’ and ‘fully
degraded’ relationships, and transition
Degraded between them as the amount of cycling
relationship
after cyclic increases.
loading

Deformation (D)
 Modeling for Practical Nonlinear Analyses

Practical F-D relationhips:

F F

C
U L X
B Y
Hardening Hardening
stiffness stiffness
Initial Initial
stiffness D E stiffness
R
X
A
D D
ASCE 41 Model Perform Model
 Modeling for Practical Nonlinear Analyses

Typical ASCE 41 capacities:

F
ASCE 41 gives capacity values for a variety of
CP capacity components. Experimental results can be used
LS capacity
IO capacity instead. Perform-3D use the term ‘capacity’,
ASCE 41 uses the term ‘acceptance criteria’.
Usually Usually at or
about close to
75% of ductile limit
Usually at CP
onset of capacity
visible
damage

D
ASCE 41 Model
 Modeling for Practical Nonlinear Analyses

Steel vs. Concrete in ASCE 41:

F F
Deformation capacities Deformation capacities
are multiples of Dy are D - Dy

Expected
strength (e.g.
ACI 318)
Expected Upto 6% Upto 10%
strength (e.g. strain strain
FyZx for hardening hardening
bending)

Dy
D D
Steel Concrete
 Modeling for Practical Nonlinear Analyses

Example: R/C Beam:

Performance Assessment:
Bending behavior is ductile.
Use hinge rotation D/C ratio

Deformation capacities
are multiples of Dy
M
IO LS CP
M V

M
V

0.010 0.020 0.025

q (hinge rotation – q
elastic or plastic)
ASCE 41 hinge rotation capacities
 Outline

• Important aspects of nonlinear behavior

• Strength-based vs. deformation-based design

• Performance-based design

• Capacity design

• Modeling for practical nonlinear analyses

• Limit states and demand/capacity ratios

• Dynamic analyses vs. static push-over analyses

 Limit States and Demand/Capacity Ratios

Limit Sates in PERFORM-3D:

a. It is not practical to examine each component individually in a large structure, because there
will be a huge number of components and component D/C ratios.
b. D/C ratios for the components can be organized based on
-element (member) type,
- component type and behavior,
- performance level.

Example of a single limit state in PERFORM:

-Beams and girders in all frames,
- All plastic hinge components in those beams and girders,
-Plastic hinge rotations of those hinges,
-Plastic hinge rotation capacities for any Life Safety Level.
 Limit States and Demand/Capacity Ratios

Limit Sates in PERFORM-3D:

c. PERFORM defines the Usage Ratio for a limit state based on the D/C ratios that are included in
that limit state.
d. The largest D/C ratio can be accepted as the usage ratio for that limit state (very simple, but
conservative).
e. An alternative method to determine the usage ratio for that limit state might be:
-running analyses for at least 7 earthquakes,
-considering each component in turn,
-calculating the mean D/C ratio over 7 earthquakes,
-assuming the D/C ratio for the component as being equal to the calculated mean value,
- assuming the usage ratio for the limit state as being equal to the maximum of the mean
values (this approach is allowed in ASCE 41).
Note that for some components the D/C ratio may become larger than 1.0.
 Limit States and Demand/Capacity Ratios

Limit Sates in PERFORM-3D:

The concept of usage ratio for a limit state simplifies the results of complex computations to
simple values for decision making.

In each limit state,, we will have certain elements, components, different modes of behavior and
performance levels.

The usage ratio for a limit state is calculated from the D/C ratios for the considered elements and
components.

In general, if the usage ratio is less than 1.0 for the limit state being considered, the limit state is
accepted to be satisfied.
 Outline

• Important aspects of nonlinear behavior

• Strength-based vs. deformation-based design

• Performance-based design

• Capacity design

• Modeling for practical nonlinear analyses

• Limit states and demand/capacity ratios

• Dynamic analyses vs. static push-over analyses

Dynamic Analyses vs. Static Push-over Analyses

Basic Steps for Nonlinear Dynamic Time History Analysis:

a. Apply gravity loads. The structure is expected to remain essentially linear under gravity
loads.
b. Define the earthquake ground motions. Usuall horizontal component of the ground motions
(H components) are considered rather than the vertical components (V component).
c. Carry out the nonlinear dynamic time history analyses with 0.01 or 0.02 time steps.
d. At each step, the demands should be calculated. At the end of the analyses, we can calculate
the component D/C ratios and usage ratios for the limit state.
e. The performance requirements for the limit state can be considered satisfied if all the usage
ratios are less than 1.0.
 Dynamic Analyses vs. Static Push-over Analyses

Basic Steps for Push-over Analysis:

V Apply first V loads, then H loads and

load pattern plot the push-over curve.
D
H

H
load
pattern

Use a response D due to

Seismic Force spectrum and H loads
estimate the
performance point
(target displacement)

Calculate usage ratios for the limit

state corresponding to this point. If
all usage ratios are less than 1.0, the
performance is said to be satisfied.
Dynamic Analyses vs. Static Push-over Analyses

Basic Steps for Push-over Analysis:

In push-over analysis, we need a response spectrum or pre-defined lateral load

pattern rather than a set of earhtquake ground motions. However, we carry out a
static analyses on the structural model which is quite simple than a nonliner dynamic
time history analysis. In addition, the computational time will be substantially less.

However, we should always keep in mind that push-over analyses is an approximate

static analyses to simulate the dynamic loads and cyclic deformations. For high-rise
and tall structures, push-over analyses will not work at all.
Dynamic Analyses vs. Static Push-over Analyses

A note on the Performance Evaluation:

Current methods, as indicated, use a deterministic approach to design a new building

and evaluate an existing structure. However, probabilistic approaches seem to be
implemented in the near future, e.g. we can define a usage ratio for a limit state with a
certain probability of exceedance, say 5%.
 References

• Graham Powell, A Presentation of CSI Educational Services,

Performance Based Design Using Nonlinear Analysis: Perform
3D.
• Shen,J., Seismic Retrofit and Earthquake Hazard Reduction,
Class Notes.
• Bathe, KJ, Introduction to Finite Element Analyses.