Nonlinear Static (Pushover) Analysis

Anindya Dutta, Ph.D., S.E.

Topics
z What

is nonlinear static analysis? z How is it done? z What are the drawbacks to its use z When can (should) it be used?

What is nonlinear static analysis

z One

of four analysis procedures embodied in FEMA 356 / ASCE 41 and commonly used in performance-based design approaches
Linear static z Linear dynamic z Nonlinear static (pushover) z Nonlinear dynamic
z

Advantages
z Intuitive
z

and visual method

Allows engineers to understand structures nonlinear behavior and progression of damage with increasing ground motion intensity

z Simple

to perform (with or without nonlinear analysis software) z Does not require selection and scaling of ground motions

Disadvantages
z Static
z

Can not directly account for dynamic structural behavior

Nevertheless, nonlinear static analysis can be effectively used for performance assessment and design of many types of structures.

How is it done
z

Basic Steps
z z z z z z

Model structure (and important nonstructural components) Perform a series of linear analyses on sequentially degraded models as damage is predicted Develop pushover curve Determine effective dynamic properties Determine demand lateral displacement for design ground motion Check adequacy of elements for force and deformation demands at design lateral displacement

Getting Started
z Understand

the structure before attempting a nonlinear analysis Identify:

critical elements z probable yield/failure modes z importance of foundation to building response z importance of torsional behavior and need for 3D modeling
z

Identify Critical Elements

z

Nonlinear analysis is complex and it is desirable to minimize the number of elements modeled Sufficient elements must be included to capture important behavior modes

Critical elements:
z

all members of lateral force resisting system any members sensitive to large lateral displacements any elements that affect the above

Typical Critical Elements

z

Steel frame structures

z

Concrete structures
z z

moment connected frames

z

Beams, columns, panel zones

z z

z z z

braces and link beams masonry & concrete infills foundations

girders - column frames flat slab - column frames shear walls architectural walls, spandrels & parapets cast integrally with structure foundations

Identify Probable Yield/Failure Modes

z

Joint Shear Structures can exhibit a variety of Column Shear inelastic behaviors It is critically important to model those that are likely to occur in the structure analyzed Wall Shear Wall Flexure Foundation rocking

Identify Importance of Torsional Response

z

Two dimensional analyses can be used for torsionally regular buildings Three dimensional analyses are required for irregular buildings

or

Modeling Element Properties & Strengths

z Best

estimates (expected values) rather than design or specified strengths should be used in determining material capacities
z

use of artificially low nominal strengths will result in under estimates of strength demands on some elements
Expected Steel yield - 1.25Fy z Expected concrete compressive strength 1.33fc
z

Modeling Steel Moment Frames

z

Before building the pushover model, determine where the yielding will occur:
z

Girder
z z

column face interior span

z z

Column Panel Zone

Modeling Steel Braced Frames

z

Tension Braces
Py K 0.05K

Compression Braces
Pc 0.3Pc

P
K

c
K = EA L Pc = 17 . Fa A FL c = 1.7 a E

K = EA L y = Fy L E u = 5 y

u = 3 c

Modeling Concrete Structures

z

Before proceeding determine critical behavior mode for each element:

z z z

flexure shear end connection

Except for lowtensile demand elements, use effective cracked section properties

Modeling Concrete Structures Flexure

z

Cracked section properties-EIcr:

z z

Beams 0.5EIg Prestressed Beams EIg Columns (no net tension) 0.7 EIg Columns (net tension) 0.5EIg Walls uncracked 0.8EIg Walls cracked 0.5EIg

Use of MM- relations obtained from BIAX and straight mechanics of materials approaches will underestimate stiffness.

Modeling Concrete Structures Flexure

z z

Use elastic-plastic model Ultimate deformation determined based on concrete strain

z

s c db
. c u = 0.005 + 01 or: fy

u dependent on
z

f 'c < 0.02

confinement provided lateral support for longitudinal steel

u = 0.005 for s d 16 b u = 0.02 for s d 8 b

Modeling Concrete Structures Flexure

0.05K Mp K=4EIc/L y u lp p

y=ylp u=ulp d lp=d/2 u=uc neutral axis depth c can be evaluated at strain of 0.003

Sequential Linear Analysis

1.

Select a loading pattern for the structure

z

Loading pattern should produce a deflected shape in the structure similar to that it would undergo in earthquake response

Sequential Linear Analysis

z Loading
z

Pattern Alternatives

Inverse triangular z Rectangular z First mode z Modal dynamic z Modal dynamic variant z Multi-mode

10

Inverse triangular
z

Approximates first mode response shape of regular structures Forces distributed with height by:

Fx

Fx =

wx hx V wi hi

Rectangular
z

Approximates first mode response of structure with very soft stories, or postyield response of structrues with weak first stories

Fx

Fx =

wx V wi

11

First Mode
z

First mode response obtained from a modal analysis

w Fx = x x V i hi

Fx

Modal dynamic
z

Deformed shape of combined modes

Fx =

wx x V h ii

Fx

12

Loading Pattern
z FEMA
z

356 requires use of at least:

Inverse triangular or first mode z Rectangular

z FEMA

440 found that there is not substatnial difference in the accuracy produced by the various load patterns

Develop Pushover Curve

Applied Shear V V5 V4 V3 V6 V7 V8 V2 V1 8 7 1 2 3 4 5 6 9 V9

Displacement at Reference Point

13

Determine Effective Dynamic Properties

z Initially,

perform elastic modal analysis to determine fundamental period of structure, T z Determine initial stiffness, ki from pushover curve as V1/1 z Determine effective stiffness, ke at 60% of yield force from pushover curve

Determine Initial Stiffness

Applied Shear V V5 V4 V3 V7 V8 V9

ki 1

V6 0.6Vy-j

V2 V1

8 7 1 2 3 4 5 6

Displacement at Reference Point

14

Determine Effective Stiffness

Applied Shear V V5 V4 V3 Judgmentally selected yield force level- Vy-j V7 V8 V9

ki 1 ke 1
1 2 3

V6 0.6Vy-j

V2 V1

8 7 4 5 6

Displacement at Reference Point

Determine Effective Period

ke T Te = ki
15

Determine Demand Lateral Displacement

z

FEMA 356 permits two alternative methods

z z

Coefficient (displacement modification) method Capacity spectrum (equivalent linearization) method

z z

As presented both are significantly flawed FEMA 440 suggests important improvements to both methods

Only coefficient method as modified by FEMA 440 presented here

Determine Demand Lateral Displacement

T T = C0C1C2 e 2 S a (Te ) g 4
C0 = modal shape factor obtained from modal analysis C1 = modifier to account for difference between elastic and elastic-perfectly plastic response C2 = modifier to account for effects of hysteretic pinching (stiffness degradation) Sa = spectral acceleration from design spectrum at effective period of structure g = 386.4 in/sec2
2

16

Coefficient C1
C1 = 1 + R= R 1 aTe2
For periods of 1 second or greater, C1 may be taken as having a value of unity

Vy j

S a (Te )W

= modal mass participation factor a = coefficient dependent on site class:

130 for site class B 90 for site class C 60 for site class D

Coefficient C1

From FEMA 440

17

Coefficient C2

For periods greater than 0.7 seconds, C2 may be taken as unity.

From FEMA 440

Check element adequacy at t

z

Once the target displacement for a given spectrum is determined, each element must be evaluated at the demand displacement to assure that neither its strength demand or deformation demand exceeds permissible capacities for the desired performance level Capacities can be obtained from FEMA 356 hysteretic tables, or directly off laboratory test data

18

What are limitations of nonlinear static analysis

z An
z

approximate procedure z Can not directly account for:

Cyclic behavior z Certain types of strength degradation z Dynamic collapse z Behavior of multiple degree of freedom structures z Randomness of response related to ground motion characteristics

Randomness of response
2.5 AZA270 I-ELC180 ELC270 B-ICC000 CLW-T050 DZC-180 DZC-270 HAGR003 Average 2.0 1.5

Sa - g

1.0

0.5

0.0 0 1 Period Seconds 2 3

8 motions, all Class D sites, large strike slip events Scaled to same spectral acceleration at 1.1 second

19

When is pushover analysis appropriate

z Generally,
z

structures with initial periods of 1 second or less.

Only use if significant nonlinearity expected z If torsion is significant, pushover analysis may give an overly pessimistic understanding of structural behavior

20