Structural Analysis

Methods of structural analysis; Imperfections; Material non-

linearity; Effects of deformed geometry on structures; Structural
stability of frames.

Dr. H C Ho & Prof. K F Chung

Department of Civil and Environmental Engineering
The Hong Kong Polytechnic University

1
 Quiz
Q1 Linear elastic design assumes:
1. representative (characteristic) loads
2. material strengths are not exceeded, i.e. use of a factor of safety
3. actions  resistances with all materials remain linear and elastic

Q2 Limit state design assumes

1. factored design loads, and factored material strengths
2. steel strengths fully utilized with significant yielding
3. actions  resistances with large deformations in steel members
and connections

Q3 In limit state design, only linear elastic analyses on structures are

conducted while section resistances with significant yielding are used
because
1. non-linear analyses are very difficult to be performed
2. with sufficient ductility, load re-distribution take places
3. by negligence and ignorance, this is the engineering !

2
Overview

 Introduction

 Analysis types

 Second order effects

 Imperfections

3
General approach

 Choose an appropriate analysis.

 Assume an appropriate model.

 Consider all actions (loads) as well as combinations of actions.

 Check for resistances and deformations of cross-sections,

members and joints.

4
Types of analysis
Conventional analysis
Actions:
 Linear elastic behaviour.
 Superposition assuming no adverse interaction.
 In typical cases, the results are acceptable.
 Geometrical imperfections, and initial out-of-straightness are
ignored.
 Simplified failure mechanisms
 Moment re-distribution.
 Global plastic analysis.

Resistances:
 Elastic values increased to plastic values with substantial yielding.
 Does the structure still behave in a linear elastic manner?

Advanced analysis
 Material and geometrical non-linearity is incorporated.
5
 Conventional analysis

M D

q = qel,max
s
D
Mhog = My,Ed = My,el,Rd
Elastic moment resistance
e

M Plastic hinge
q = qpl,max
s
Mhog = My,Ed = My,pl,Rd
D
Plastic moment resistance
6
e
Stress – strain curves
Full deformation range
Stress-strain curve
. . D
E

. .
B
.
σ .
A
C
F

ε
Development of stress blocks at various strain levels from A to
F

A B C D E F
7
 Conventional analysis

M Plastic hinge
q = qpl,max
s
Mhog = My,Ed = My,pl,Rd
D
Plastic moment resistance
e
Can we still adopt elastic structural analysis method to
evaluate the internal force?
M
Elasto-Plastic hinge

D
Plastic
hinge 8
 2nd order effect to be accounted for
P P

Initial
imperfection of Initial imperfection
the straightness d of the alignment of
of the both end supports
compression
member

Initial Initial
geometry geometry

P-d effect P-D effect

9
Importance of 2nd order effect

P 25.0 kNm
A vertical load of 50 kN
is applied at an 25.0 kNm
eccentricity of 0.5 m to
a 6 m high signpost of
150x150x6 mm SHS
with S275 steel. s = 171 N/mm2
The maximum bending
stress is grossly
underestimated. Why?

25.0 kNm
Initial
BMD
geometry

10
 Importance of 2nd order effect

D = 0.18 m 25.0 kNm

25.0 kNm

s = 251 N/mm2
i.e. 47% increase!

PD = 12.7 kNm 25.0 kNm Pd = 37.7 kNm

Deformed
BMD BMD BMD
geometry

Second First
order effect order effect 11
Frame stability

Frame stability is assured by checking:

 Cross-sections (Plastic section vs Slender section ?)
 Members (Straightness ?)
 Joints (Simple / continuous / semi-continuous?)

Proven structural adequacy:

 A correct frame model.
 Realistic loading conditions.
 An analysis fully recognizing both local and overall deformations.

12
 Second order effects on a portal frame
P P P P

First order
– linear
BMD

P P

The magnitude of d depends on: d d Second order

 Load level
– nonlinear
 Member slenderness Deformation

 Joint rigidity

Pd Pd Total

BMD 13
 Second order effects on a portal frame

First order
– linear
BMD

The magnitude of D depends on: Second order

 Load level
– nonlinear
 Member slenderness Deformation

 Joint rigidity
PD
 Sway of the frame

Total

BMD 14
Effects of deformed geometry

EN 1993-1-1 Clause 5.2.1(2) states that deformed geometry

(second order effects) shall be considered:

 if they increase the action effects significantly

or
 If they modify significantly the structural behaviour

15
 Quiz
Q1 Second order effects
1. are often neglected in commonly used structural forms
2. are usually small in steel building structures
3. should be considered in all practical design of steel structures

Q2 Second order effects are often critical in

1. steel columns with large axial shortenings
2. steel beams with large deflections
3. steel beam-columns with large deformations

16
Limits for ignoring deformed geometry

For elastic analysis:

Fcr
cr   10
FEd
where
cr is the factor by which the design load would have to be increased
to allow for elastic instability in a global mode of a structure (cr in
BS 5950-1).
FEd is the design load on the structure.
Fcr is the elastic critical buckling load for global instability based on
the initial elastic stiffness of the structure.

17
 Limits for ignoring deformed geometry

Assuming FEd = Afy and Fcr = Ncr,

Reduction factor, c

Afy FEd
 
Ncr Fcr
1

cr
0.2

Non-dimensional slenderness, ഥ
For elastic analysis,
Column buckling curve cr = 10   = 0.316

So, when cr ≥ 10, the effect of deformed geometry in the structure may
be safely ignored. A first order analysis will be sufficient.
18
Limits for ignoring deformed geometry

For plastic analysis,

Fcr
cr   15    0.258
FEd

 A stricter limit for plastic analysis due to the loss of stiffness

associated with material yielding.

 So, when cr ≥ 15, the effect of deformed geometry in the

structure may be safely ignored. A first order analysis will be
sufficient.

19
Simple estimate for cr

Simple estimate for cr may be applied to:

 Portal frames with shallow roof slopes
 Beam=column frames

 HEd   h 
cr     for each storey
 VEd   dH,Ed 
where
HEd is the horizontal reaction at the top of the storey
VEd is the total vertical load at the bottom of the storey
dH,Ed is the storey sway when loaded with horizontal loads such as
wind and equivalent horizontal forces
h is the storey height

20
Simple estimate for cr

Definition of symbols:

where
HEd is the horizontal reaction at the top of the storey
VEd is the total vertical load at the bottom of the storey
dH,Ed is the storey sway under horizontal loads
h is the storey height
21
Frame stability

Fcr
Limits for treatment of second order
cr 
effects depend on cr: FEd

Limits on cr Action Achievement

cr ≥ 10 First order analysis First order only

Second order
First order analysis plus
effects by
10 > cr ≥ 3 amplification or effective
approximate
length method
means
Second order
3 > cr Second order analysis effects more
accurately

22
Global initial sway imperfections

Global initial sway imperfections:

  0hm 2 2
h  but   h  1.0
where h 3
0 is the basic value = 1/200
h and m are reduction factors  1
m  0.5  1  
 m

h: height of the structure

in meters
m: the number of
columns in a row

23
Individual bow imperfections of members

The relative initial local bow imperfections of members for

flexural buckling:
e0
L
where L is the member length.

24
Equivalent horizontal forces

The effects of initial sway imperfection and local bow imperfections

may be replaced by systems of equivalent horizontal forces,
introduced for each column.

25
Global imperfections for frames

 Much easier to apply as equivalent horizontal forces NEd, where

NEd is the design compressive force in the column.

 Saves changing the model for opposite direction in asymmetric

buildings.

 Many buildings have such complicated arrangements that it will be

best to ignore the h and m reductions and use 1/200.

 Don’t forget them.

26
 Quiz
Q1 Approximation methods are used because:
1. they give accurate solutions
2. second order effects are often not critical in common structures
3. exact methods are too complicated to practising engineers

Q2 Equivalent horizontal forces are able to

1. simulate effects of wind loads
2. introduce additional moments equivalent to those of second order
effects
3. assess lateral stiffness of the structures under consideration

27
Actions to be specified

 EN 1991-1-1: Densities, self-weights, imposed loads

 EN 1991-1-2: Fire
 EN 1991-1-3: Snow loads
 EN 1991-1-4: Wind actions
 EN 1991-1-5: Thermal actions
 EN 1991-1-6: Actions during execution / erection
 EN 1991-1-7: Impact and explosions

28
Other Actions

 Equivalent horizontal forces

- unless using initial imperfection model

 Derived from imperfections

 Applied in ALL combinations

(only in gravity combinations in BS 5950)

29
Checks

 Analyse structure

 Classify sections using clause 5.5

- for plastic global analysis, check clause 5.6

 Check cross-sectional resistance to clause 6.2

 Check buckling resistance to clause 6.3

- check built-up members to clause 6.4

30