Basic Concepts: Performance-Based Earthquake Engineering

Basic Concepts: Performance-based
Earthquake Engineering
Seismic Performance
What are our goals?
A design framework for expressing performance goals
Performance vs. Engineering Response parameters
Nonlinear response - Is it desirable feature or a problem to overcome?
Some engineering approaches to improve performance
Quantifying performance
CEE 227 - Earthquake Engineering

U.C. Berkeley Spring 2003 ©UC Regents 1-1
Performance Expectations
Model codes
Current codes - What are their stated
objectives?
Vision 2000
Ideal situation - A simple limit states
framework for design.
Current directions - Vision 2000 FEMA 273/356
(SEAOC), SAC LRFD approach, etc.

Future directions - reliability based FEMA 350-353
approaches, PEER performance-based

evaluation strategy PEER PBEE
References
FEMA PBEE

1
References: Performance-Based Design Codes
Hamburger, R.O., Performance-Based Analysis and Design
Procedure for Moment Resisting Steel Frames, Background
Document, SAC Steel Project, Sept. 1998.
SEAOC, Vision 2000: Performance Based Seismic Engineering of
Buildings, San Francisco, April 1995.
Recommended Seismic Design Criteria for New Steel Moment-
Frame Buildings, FEMA 350, Federal Emergency Management
Agency, Washington DC, July 2000
FEMA, Guidelines for Seismic Rehabilitation of Buildings, Vol. 1:
Guidelines, FEMA 356, Washington DC, 2002 (formerly FEMA 273).
Earthquake Engineering Research Center, Performance-based
Seismic Design of Buildings: An Action Plan , U.C., Berkeley, 1995.
FEMA/EERI, Action Plan for Performance -Based Seismic Design,
FEMA 349, Washington DC, 2000.
ATC, Development of Performance-based Earthquake Design
Guidelines, ATC-58, Redwood City, 2002.
Current Model Codes

CBC, IBC and UBC
Stated purpose:
Provide minimum provisions
for design and construction of
structures to resist effects of Structurally undamaged building astride fault
seismic ground motions

“…to safeguard against major
structural failures and loss of
life, not to limit damage or
maintain function.”
(UBC, 1997 ed., Section 1626) Shear failures in short “captive” columns

2
SEAOC “Blue Book” Recommendations
Commentary states:
Earthquake F requency of Desired
Intensity Occurrence Performance
1 Minor Several times No damage to
during service structure or
life nonstructural
contents
Three One or more

2 Moderate Limited damage to
times during nonstructural
Tiers service live components and

no significant
damage to
structure
3 Major Rare and No collapse of
(Catastrophic) unusual event structure or other
(10% as large as any damage that
exceedence experienced in would create a life
in 50 years) vicinity of site. safety hazard.
(After: Lateral Force Recommendations and Commentary, SEAOC.)

Current code goals are ambiguous

Definitions are non-
Ea rthquake Fre que ncy of Desired quantitative (e.g., limited
1
I ntensity
Minor
Occurre nce
Several times
Performa nce
No damage to
damage, one times,
or more
during service structure or
etc.)
Three tiers, but…
life nonstructural
contents
Only one design earthquake

2 Moderate One or more Limited damage to
times during nonstructural
Provisions not specifically

service live components and
no significant
damage to
structure
3 Major Rare and No collapse of
associated with any particular
performance level.
(Catastrophic) unusual event structure or other
as large as any damage that
Leads to wide variation in

(10%
exceedence experienced in would create a life
in 50 years) vicinity of site. safety hazard.
interpretation and
(After: Lateral Force Recommendations and Commentary, SEAOC.)
performance.
3
Vision 2000 - Trends toward
Performance-Based Seismic Engineering
of Buildings
Seminal Document - some powerful new concepts
The definitions of performance states developed are:
incorporated in the appendices of the SEAOC “Recommended
Lateral Force Requirements and Commentary”
Refined by other groups in later documents
Focuses on:
defining what constitutes a frequent, rare or very rare
earthquake, and
describing in detail what are the performance states that one
wants for different types of events and structures.
Vision 2000 - Basic Approach

Relationship developed Performance objective
between:
Performance objective increases (i.e., less damage):
Type of facility for a high probability earthquake
(one that may occur several times
Probability of earthquake during the life of a structure), or
and for an important structure or
dangerous occupancy (i.e., a
Response parameters related hospital or dynamite plant)
to each performance Conversely, more damage is
objective. acceptable:
Specific demand parameters for a rare, severe earthquake,
identified, and
Initial acceptance criteria are for less critical or temporary
facilities.
established.

4
Vision 2000 - Performance States
Fully operational Continuous service. Negligible structural
and non-structural damage.
Operational Most operations and functions can resume
immediately. Structure safe for occupancy.
Essential operations protected, non-essential
operations disrupted. Repair required to restore
some non-essential services. Damage is light.
Life Safe Damage is moderate, but structure remains
stable. Selected building systems, features or
contents may be protected from damage. Life
safety is generally protected. Building may be
evacuated following earthquake. Repair possible,
but may be economically impractical.
Near Collapse Damage severe, but structural collapse
prevented. Non-structural elements may fall.

Occupancy or Use of Building Considered

Three occupancy types considered in Vision
2000.
Safety Critical Facilities:
Large quantities of hazardous materials (toxins, radioactive
materials, explosives) with significant external effects of damage to
building.
Essential/Hazardous Facilities
Critical post-earthquake facilities (hospitals, communications
centers, police, fire stations, etc.)
Hazardous materials with limited impact outside of immediate
vicinity of building. (Refineries, etc.)
Basic Facilities One can argue with
All other structures. or adapt these definitions.

5
Quantitative Indexing of Earthquake
The earthquake intensity is now described quantitatively
in probabilistic terms for Vision 2000.
Earthquake Recurrance Interval Probability of Occurance
Claassification
Frequent 43 years 50% in 30 years
Occasional 72 years 50% in 50 years
Rare 475 years 10% in 50 years
Very Rare 970 years* 10% in 100 years
* need not exceed mean + 1 standard deviation

for the maximum deterministic event

Schematic Relation Between Performance

Objective and Earthquake Probability
Performance Objective
Earthquake Fully Operational Life Safe Near Collapse
Probability Operational
Frequent
Ess Ba s Unacceptable
Occasional ent i cF Performance
ial/ ac i
Saf Haz litie
e ard s
Rare
ty C ous
riti Fac
ca l iliti
F aci es
Very Rare li ti
es
adapted from Vision 2000, SEAOC

6
Comments on Relationship
Thus, a building
would be expected to
Performance Objective suffer more damage if
it were subjected to a
Earthquake Fully Operational Life Safe Near Collapse
Probability Operational
Frequent
Unacceptable more severe, less
Ess
e
Ba
si c Performance likely earthquake.

Occasional ntia Fac
l /Ha il itie
A more critical
Sa zar s
fety dou
Rare Cr i s Fa c
ti
building would be
cal ilitie
F aci s
litie
expected to have less
Very Rare s
damage for the same

adapted from Vision 2000, SEAOC
earthquake
probability.

Comments on Approach.
A basic structure would be expected to:

have essentially no damage if subjected to an
earthquake with a 30% probability of occurrence
in 30 years, whereas it would be

be near collapse if subjected to an event with a
10% probability within 100 years.
One can substitute more appropriate numbers for a

particular project, or upgrade the characterization of
the structure (to an essential facility, for instance,
where the structure would be designed to remain
life safe during the very rare event.)

7
Some More Comments
This method removes some of the ambiguity from
current recommendations.
Geotechnical engineers (seismologists and structural
engineers) are able to and do regularly develop
estimates of peak ground motion parameters
(acceleration, velocity, etc.), elastic response spectrum
and even time histories corresponding to:
x% probability of occurrence in “y” years
We will look at how this is done later in the course.

Quantification of Earthquake Hazard

8
Acceptance Criteria
Vision 2000 introduces engineering response
parameters to consider (drift, stress, plastic
hinge rotation angle, acceleration, etc.) and
what limits are acceptable for a particular
performance objective.
These criteria were for the most part based on
consensus, rather than on test data or
quantitative field observation.
For example, ...
Drift Limits in Vision 2000

NEW
Permissible Ma ximum Permissible Pe rma ne nt
Drift, % D rift, %
Fully Ope ra tiona l 0.2 ne gligible
Ope ra tiona l 0.5 ne gligible
Life S a fe 1.5 0 .5
Ne a r C olla pse 2.5 2 .5
After, Vision 2000, SEAOC
Vision 2000 does not describe acceptable analysis methods. So,

how do we calculate the maximum drift (or maximum permanent
drift) and prove we satisfy these criteria?
Why are these criteria selected? Will a building at 2.6% drift
collapse? Can all buildings with drifts of 0.4% remain operational?

9
Damage to Steel Moment Frames
Damage Description
Fully Operational Negligible
Operational Minor local yielding at a few places. No
observable fractures. Minor buckling or
observable permanent distortion of
members.
Life Safe Hinges form. Local buckling of some
Big jump
beam elements. Severe joint distortion.
Isolated connection failures. A few
elements may experience fracture.
Near Collapse Extensive distortion of beams and
column panels. Many fractures in
connections.
May need to add intermediate limit state related to reparability where

damage is limited to make repair quick and/or economically feasible.
Since damage difficult to quantify and economics issues are owner-
sensitive, these intermediate states are difficult to incorporate in a code.

Extending the Vision 2000 approach

The Vision 2000 approach does not suggest analytical
approaches nor methods to assure reliability of structure.
Intermediate limit states difficult to quantify.
The Vision 2000 is an uncoupled approach. That is, we end up
with a deterministic procedure based on a probabilistically
determined spectrum. Load and resistance factors still remain
to be determined to provide desired reliability.
Identification of limit states by subjective name (continued
operation) may lead to legal problems if goal is not realized
following an earthquake. Some codes use a letter system (i.e.,
performance objective A, B, C, etc.). Probabilistic specification
of response parameters may be better.

10
Several Major Advances in FEMA-273/356
Guidelines for Seismic Rehabilitation of Buildings
Four Performance Goals:
Collapse Prevention, Life Safe, Continued Occupancy, Operational
National Seismic Hazard Maps developed by USGS
Spectral ordinants (5% damping) for Sa
SD1/T
different probabilities of occurrence and
soil conditions at T= 0.2sec and T=1sec. SDS
Displacement-Based Approach with subjective factors to
T
assess uncertainty
d = C C C C C S
Defines Nonlinear Dynamic and Static Pushover methods
roof 0 1 2 3 4 delastic
in addition to conventional elastic methods

Severity of Damage
Joe’s Joe’s Joe’s

Beer! Beer! Beer!
Food! Food! Food!
Operational Immediate Life Collapse

Occupancy Safety Prevention
0% Damage 99%
(R. Hamburger)

11
Structural/Nonstructural/Element Criteria
From FEMA 356

Damage related to demand parameters
Joe’s
Beer!
Joe’s
Beer!
Food!
Member Capacity
Force
Food!
Lateral Resistance
Deformation
Joe’s
Beer!
Food!
Structural Displacement ∆

12
Relate Probabilities of Exceedence to
Damage States
Occasional
(72 years)
Very Rare
Joe’s Joe’s (2500 years)
Beer! Beer!
Food! Food!
Frequent Rare
(25 years) (500 years)
10
Pea k Ground Acceleration - g

1
Joe’s 0.1 0.01 0.001 0.0001
Beer!
0.1
Food!
0.01
Annual Probability of Exceedance

Some limitations of FEMA 356

While ground motion is defined in probabilistic terms,
uncertainty and randomness not considered related to:
structural demands, and
capacities.
Evaluation is made on a member by member basis…the
failure of a few elements might not lead to the failure of
the system.
Performance goals are defined in absolute, but
subjective terms. Structure is either life safe or it is not.

13
Extending the FEMA 356 concept
Basic Limit States Design Format
Common format in Europe and in other industries.
Explicit list of performance goals, criteria, and usually, a
acceptable probability of reaching or exceeding the goal;
Direct relation between goal and what engineering
demand parameter is checked (and acceptance criteria).
Explicit recognition and consideration of randomness and
uncertainty (e.g., LRFD format implementation)
Probability of
Limit State Performance Objective Evaluation Criteria for Exceeding
Engineering Parameters Performance Criteria
Name
Goal you are trying to
achieve
Response
parameter(s) measured
and acceptance criteria
x 1 % in y 1 years Many
For each limit state:

Need to recognize and
manage randomness and Frequency of Occurrence
uncertainty.
Not adequate to say Demand
Dmedian Cmedian
Capacity
Dmedian < Cmedian Failure
Need probability of failure Probability
less than a specified

amount. Response Parameter
X % probability in y years (often, y
is the assumed service life)
Cmedian γ For a given
>
In LRFD format Dmedian φ probability of failure
in ‘y’ years
γ Dmedian < φ Cmedian

14
Large randomness and uncertainty in
earthquake-resistant design
Randomness in both demand and capacity.
Earthquake motions inherently random. Even with increased knowledge there will
be large randomness in excitation and response.
Structural behavior effected by random variations in material properties,
deterioration and construction quality. Capacity is also affected by loading history
and duration which are influenced by randomness of excitation.
Uncertainty in demand has components related to:
Seismology (what earthquake intensity is expected during a given interval of time) -
various methods available to improve estimates
Ground motion characteristics (what response spectrum corresponds to an
earthquake motion corresponding to a given intensity and soil conditions)
Structural characteristics (what is the structure’s actual mass, stiffness, strength,
damping, foundation condition, etc.?)
Modeling (have we accurately modeled the structures: completeness, etc.)
Structural Analysis Method (Elastic, Inelastic; dynamic, static?)

Uncertainty and randomness in capacity

In the past, strength was
generally primary criterion
related to capacity
Now, focus is increasingly on

strain, deformation and energy
dissipation (fatigue) capacities.
From Marc Eberhard, UW

15
Capacity estimates may be a problem
Even Slabforcontributions
flexural strength, there are difficulties:
(composite action)
Connections (panel zone deformations, welds, bar pull out, etc.)
Shear (in members, connections and structural walls)
Non-compliant or marginally ductile elements (existing structures).

Inconsistent development of capacity equations

In complete tests or inadequate documentation
Non-structural components (cladding and other architectural
features may actually behave like structural elements, or alter the
behavior of structural elements)
Both strength & deformation capacity sensitive to:

Loading history (low-cycle fatigue)
Rate of loading effects (effects on strength and deformability)
Ultimately, seismic capacity is related to dynamic aspects of

response of a complete structural system

Probabilistic PBE Approaches

In general one would want to state the problem as:
w% chance of exceeding performance objective in ‘y’

years (life of structure)
This is complex, computationally intensive reliability

problem.
Solve rigorously as a reliability problem.
Results tend to be dominated by uncertainty in ground
motions

16
Seismic Hazard and Performance Level
Now more common to uncouple the problem, as:
x% chance of exceeding performance level for an earthquake
with an z% probability of occurrence in ‘y’ years
Treat ground motion and structure separately:
Probabilistic response spectrum used with deterministic “conservative” selection of
seismic hazard for design.
Develop calibrated load and resistance factors using reliability analysis or Monte Carlo
simulation to have appropriate overall reliability.
Thus, Performance Objective has three parts:

Definition of Performance Level
Statement of associated Seismic Hazard
Statement of desired confidence

Significance of Confidence Level

Now, engineer can state:
We have a high, moderate, or low confidence that the
performance objective can be met for an earthquake
with a “x”% probability of occurring in “y” years
We would say, for instance: there is a 90% (10%)
confidence that a structure will remain stable in earth-
quakes having a median probability of exceedence of 2% in
50 years.
Powerful evaluation tool, and one that is understandable by
clients and other professionals
17
FEMA/SAC Steel Project
Builds on FEMA 273/356, but covers new & existing buildings
Two Performance Objectives
Continued Occupancy - damage permitted, so long as it does not reduce future
confidence in building’s ability to achieved performance objectives
Collapse Prevention - local plastic rotations, global instability, avoid premature failure
modes
NEHRP seismic hazard data: 50%, 10% and 2% in 50 years
Consistent system-level reliability approach used. Treats
randomness and uncertainty to focus on confidence of
achieving performance goal during specified period.
Designer/owner can select confidence level.
Rational load, resistance and analysis bias factors developed for various
forms of uncertainties.
Extensive Monte Carlo simulation used.
Performance-based Design: SAC Approach

LRFD-type format often utilized
Demand uncertainty divided into several parts;
Ground motion (randomness and uncertainty treated using probabilistically-based
response spectrum and load factors corresponding to geographical location and
soil conditions),
Structural response -- Even for a family of ground motions with “similar”
characteristics, structural response will have large variations.
Analysis method - ESP, EDP, NSP, NDP
Modeling --Variations in mechanical and dynamic characteristics will make these
uncertainties in response demand larger.
Seismic capacity related to three main components:
Element level effects (stress, plastic hinge rotations)
Global behavior (drifts, static and dynamic instability)
Brittle failure modes (premature column fracture or buckling)
See: R. Hamburger, Performance-Based Analysis and Design Procedure
for Moment-Resisting Steel Frames, SAC Background Report

18
SAC Approach
Basic approach
γ 1γ 2 γ 3 ... γ n D ≤ φ1φ2 φ3 .. .φn C
or, combining terms and adding confidence level:
γ γ con D ≤ φC
Multi-level design approach
Standard default code approach with specified demand, capacity
and confidence values
90% confidence for global instability response parameters

Explicit
50% confidence for local stability and continued occupancy
methods allowing nonlinear analysis and testing to
develop demand and capacity values, or to specify different
target confidence levels

SAC Approach for New Buildings

1. Select performance objective, confidence levels and earthquake hazard,
e.g. Collapse Prevention, 90% confidence and 2% in 50 year hazard.
2. Determine design seismic earthquake for hazard, e.g. spectral displacement
at the fundamental periods of the building, time histories...
3. Develop a mathematical model of the building.
4. Analyze mathematical model to determine the values of the key design
parameters: maximum and permanent inter-story drift; column load.
5. Apply demand and bias factors to the computed response parameter values
to compensate for the various biases and uncertainties inherent in the
predictive methodology as well as the randomness inherent in seismic
structural response . Apply additional demand factor to achieve desired
confidence level.
6. Compare the factored demand against the factored acceptance criteria value
for the response parameter.
γ γ con D ≤ φC

19
Fragility curves describe reliability of structure
1
Minor Dcol = 60", ar = 6, Pr = 0.2
Damage
0.9
(Park&Ang) Spalling
0.8
0.7
Significant
Damage
(Park&Ang)
0.6
Probability
0.5
Spalling
Park Ang > 0.4
0.4 Park Ang > 1.0
Fatigue Index> 0.5
Fatigue Index> 1.0
0.3
Fatigue
0.2
Failure
0.1
0
0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2
Sa/SaARS

Rapid Evolution of Performance-oriented

Codes and Guidelines expected.
These PBEE guidelines (FEMA 273, SAC, etc.) are being
routinely used for many new and existing buildings.
These are only first steps in developing Performance
Based Codes. Much work is needed to evaluate and
validate methodologies. Lots of changes will be made in
the next few years.
Improved tools needed for analysis and design.

20
Structural Engineering

Tools Improve
Improved proportioning strategies
Capacity
Design
Greater demands for quantitative Improved earthquake characterization
design and evaluation methods Probability
that realistically and explicitly Hazard

account for performance Model

Sd
Improving analysis tools Improving control of uncertainties
Probability
Demand
Analysis Capacity Reliability
Model Fails
Engine
δ
Improving characterization of Improved assessment of losses
performance
Damage
Loss
Models
$
Models

Drift is not ... performance

Engineers must use quantitative engineering demand
parameters (such as drift, plastic rotation, stress) as a
measure of performance.
These are generally not performance indices of interest
to an owner who is concerned about repair cost, loss of
revenue, injuries to occupants, down time, etc.
Many intermediate performance objectives related to
reparable damage or minimization of economic or social
impacts.

21
Structural Engineering

Tools Improve
Improved proportioning strategies
Capacity
Design
Greater demands for quantitative Improved earthquake characterization
design and evaluation methods Probability
that realistically and explicitly Hazard

account for performance Model

Sd
Improving analysis tools Improving control of uncertainties
Probability
Demand
Analysis Capacity Reliability
Model Fails
Engine
δ
Improving characterization of Improved assessment of losses
performance
Damage
Loss
Models
$
Models

Economic, social & operational impacts

In general, we need to consider economic and related
impacts as well. Initial costs need to be compared with
“life cycle” costs to determine the design that performs
best. Evaluation of economic costs also depend on who
pays for damage (owner, insurance, government).
Perspective varies
* Developer * Large institutional owner
* Insurance company * Government decision makers
* Engineer

22
Performance-Based Design Approaches
Capacity
Design
Damage
Models
Analysis Loss
Engine Demand Probability Models $
Capacity Reliability
Model Fails?
δ
Probability
Hazard
Model
Sd

PEER – Probability Framework Equation
v (DV ) = ∫∫∫ G DV DM | dG DM EDP | dG EDP IM | dλ ( IM )

Impact Performance (Loss) Models and Simulation Hazard
IM – Intensity Measure
EDP –Implementation
Engineering DemandThrough
Parameter
DM – Damage Measure
LRFD-like Format:
DV – Decision Variable
γ D|Sa ⋅ D̂
ν(DV) – Probabilistic Description
P ≤ φC ⋅ Ĉ
of Decision Variable
Saf0
(e.g., Mean Annual Probability $ Loss > 50% Replacement Cost)
23
FUTURE of PBEE
FEMA is funding a new project to implement a PBEE
framework for all structural systems, not just retrofit or
steel… ATC-58 project
PBEE is currently being implemented on many
conventional and important structures
Answers need for more reliable, quantitative information
on performance, utilizing modern capabilities for
characterizing seismic hazard, simulating seismic
response, and assessing impact of response on owner
and society
Validation and refinement needed
How do we design a structure to attain our objectives
reliably? NEXT….
24

Basic Concepts: Performance-Based Earthquake Engineering

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Basic Concepts: Performance-Based Earthquake Engineering

Uploaded by

Copyright:

Available Formats

Basic Concepts: Performance-based

CEE 227 - Earthquake Engineering

(SEAOC), SAC LRFD approach, etc.

approaches, PEER performance-based

CEE 227 - Earthquake Engineering

Current Model Codes

seismic ground motions

CEE 227 - Earthquake Engineering

Three One or more

Tiers service live components and

(After: Lateral Force Recommendations and Commentary, SEAOC.)

CEE 227 - Earthquake Engineering

Current code goals are ambiguous

Only one design earthquake

Provisions not specifically

Leads to wide variation in

Vision 2000 - Basic Approach

CEE 227 - Earthquake Engineering

CEE 227 - Earthquake Engineering

Occupancy or Use of Building Considered

CEE 227 - Earthquake Engineering

Frequent 43 years 50% in 30 years

Occasional 72 years 50% in 50 years

Rare 475 years 10% in 50 years

Very Rare 970 years* 10% in 100 years

* need not exceed mean + 1 standard deviation

CEE 227 - Earthquake Engineering

Schematic Relation Between Performance

CEE 227 - Earthquake Engineering

damage for the same

CEE 227 - Earthquake Engineering

One can substitute more appropriate numbers for a

CEE 227 - Earthquake Engineering

CEE 227 - Earthquake Engineering

Quantification of Earthquake Hazard

CEE 227 - Earthquake Engineering

Drift Limits in Vision 2000

Fully Ope ra tiona l 0.2 ne gligible

Ope ra tiona l 0.5 ne gligible

Ne a r C olla pse 2.5 2 .5

After, Vision 2000, SEAOC

Vision 2000 does not describe acceptable analysis methods. So,

CEE 227 - Earthquake Engineering

May need to add intermediate limit state related to reparability where

CEE 227 - Earthquake Engineering

Extending the Vision 2000 approach

CEE 227 - Earthquake Engineering

in addition to conventional elastic methods

CEE 227 - Earthquake Engineering

Joe’s Joe’s Joe’s

Operational Immediate Life Collapse

CEE 227 - Earthquake Engineering

From FEMA 356

CEE 227 - Earthquake Engineering

Damage related to demand parameters

CEE 227 - Earthquake Engineering

Pea k Ground Acceleration - g

CEE 227 - Earthquake Engineering

Some limitations of FEMA 356

CEE 227 - Earthquake Engineering

For each limit state:

Need probability of failure Probability

less than a specified