Introduction to Design of Cold Formed

Steel

by

Prof Emad Gad

Dean of Engineering
Swinburne University of Technology, Australia
[email protected]
 Outline

• Background, standards and resources

• Differences to hot rolled steel
• Design methods
• Local buckling and postbuckling behaviour
• Design of tension members
• Design of compression
• Design of flexural members
• Design of connections
• Appendix B – Example for compression member
design using the effective with method
• Appendix A - Direct strength method
Design standards
 Design standards

• USA (Standard & commentary) • Europe

Eurocode 3- Design of steel

structures –
Part 1-3: General rules -
Supplementary rules for
cold-formed members and
sheeting
 Helpful resources
• From Australia • From USA
 Helpful software
• CUFSM – free US software tutorials for use with Direct
Strength Method (by Prof Ben Schafer at Johns Hopkins
University): www.ce.jhu.edu/bschafer/cufsm/

• THINWALL, ColdSteel/4600, PURLIN4600 & PURLIN –

Softwares from Australia for general and specific sections
covering both effective width and direct strength methods by
Prof Greg Hancock at Sydney University:
www.sydney.edu.au/engineering/civil/research/case/software

• SwinChannel and Roll Former – Simple spreadsheet programs

developed for C sections based on effective width design
according to AS/NZS4600-2005 developed by students at
Swinburne University.
 Design standards
Cold formed steel in framing applications (NASH Australia)

Based on
AS/NZS4600
 Design standards
Cold formed steel in framing applications (NASH Australia)

For NASH documents:

Mr Ken Watson
Executive Director
[email protected]
www.nash.asn.au
Differences to Hot Rolled Steel
 Differences to hot rolled steel

Major points of distinction between cold formed steel and

hot rolled steel.
1- Local buckling & post buckling behaviour
2- Propensity for twist
3- Distortional buckling
4- Cold work due to forming
5- Web crippling
6- Connection details
7- Corrosion
 1. Local buckling

• Elements are wide compared to their thickness

(high slenderness)
• Local buckling may occur before yielding

A-A
 1. Local buckling

• Local buckling mode

– deformation of the cross-section without translation of the
“fold lines” (element edges)
 1. Local buckling

• A thin elastic plate does not fail soon after it

undergoes local buckling. Elements can carry an
increasing load after local buckling.
• Post-buckling reserve is allowed for in the design to
achieve economic sections. This can be done through
the effective width concept.
 2. Twist

• Monosymmetic sections
have their shear centre
eccentric to their centroid.
• Beams may require
torsional restraints (covered
in Cl 4.3)
• Columns may experience
flexural-torsional mode at
lower load than flexural.
 2. Twist

• Flexural-torsional buckling
– translation and/or rotation of the entire cross-section without
deformation of the cross-section
 3. Distortional buckling
• Member in compression or bending with bracing against lateral or
flexural-torsional buckling can experience distortional buckling
• Design checks are covered in Cl 3.3.3.3 (flex) and 3.4.6 (comp).
 3. Distortional buckling

• Flange-distortional bucking
– deformation of the cross-section with translation of the “fold
lines” (element edges)
 4. Cold work

• Cold work increases yield and ultimate strength and

decreases ductility. Covered in Cl 1.5.1.2.
• The increased strength is used for design. For hollow
sections the strength is obtained from formed tubes.

- SHS grade C350

 5. Web crippling

• Web depth to thickness ratio is high.

• Webs of beams may cripple due to the high local intensity
of the load or reaction.
• Normally no load bearing or end bearing web stiffeners.
For sheeting and decking, the web is often inclined.
• Design covered in AS/NZS4600 in CL3.3.6 & 3.3.8.
 6. Connection details

• Failure modes of connections may be different in cold

formed sections (e.g. tearing of thin plates in fillet
welds). These are covered in Section 5.
• Different types of fasteners are used (e.g. screws, blind
rivets, clinches).
 7. Corrosion protection
• Protection is governed by type and thickness of treatment not base
metal thickness.
• Most common coating is Zinc (Z), Aluminium-Zinc (AZ) and
Aluminium Magnesium/Zinc (AM).
• In specification, the letters indicate coating type and the number is
the mass in grams per m2 on 2 sides (for steel made to AS1397).
For example, Z275 is zinc coating with 275grams per m2.
• The selection of suitable coating is dependent on:
– Atmospheric corrosively category
– Design life
– Accessibility for maintenance
• Some guidance is given in Appendix C.
Design Methods
 Design methods

• There are two alternative design approaches to cold-

formed steel members:
• Effective Width Method (EWM)
• Direct Strength Method (DSM)
• The EWM uses the effective cross-section for capacity
calculations
• The DSM uses the full (gross) cross-section for
capacity calculations
 Design methods

• The EWM uses the effective cross-section for capacity

calculations

Effective section in compression

 Design methods

• The EWM uses the effective cross-section for capacity

calculations

Effective section in bending

 Design methods

• The EWM uses hand calculations to determine the

element buckling stresses, the element slenderness,
then the effective section – the capacity is then
calculated using the effective section
• The DSM uses elastic buckling software to determine
the section buckling stresses - the capacity is then
calculated directly using the full (gross) section
 Design methods
• DSM softwares typically use Finite Strip Method (FSM)
• FSM divides the member into finite longitudinal strips
• The strips are free to deform in-plane and out-of-plane, and are
restrained from cross-sectional deformation at their ends
• Each strip has uniform longitudinal stress along its length, and the
stress varies linearly from one strip to the next
 Design methods

• Buckling is assumed to have sinusoidal form and the

buckling length is taken as half wavelength

One half-wave

One half-wave
One
half-wave
 Design methods
– FSM provides the buckling stress for all half-wavelengths
defined by the user
– FSM therefore provides all buckling modes and stresses for a
cross-section for all possible buckling half-wavelengths
– The plot of buckling stresses vs half-wavelengths is therefore a
complete solution for buckling of a particular cross-section
geometry – termed the signature curve
– The buckling stresses of importance for design are the minima
of the signature curve, which identify the different buckling
modes and their corresponding minimum bucking stress values
(and the half-wavelengths at which they occur)
 Design methods
• The signature curve
– The relevant buckling mode minima occur at different half-
wavelengths i.e. the member will buckle in different modes
depending on its length
– Local buckling occurs at the shortest half-wavelength,
followed by distortional then flexural-torsional

Distortional
Flexural-torsional
Local
 Design methods
• The signature curve - Compression

Local Distortional Flexural-torsional

Buckling stress (MPa)

Buckle half-wavelength (mm)

 Design methods
• The signature curve - Bending
Distortional

Local
Buckling stress (MPa)

Flexural-torsional

Buckle half-wavelength (mm)

 Design methods

• Advantages of the DSM:

– No effective width equations
– Versatile and can be applied to complex sections
– No iterations (to establish the effective width for members in
bending)
– Uses the gross section properties
– Distortional buckling is accounted for without the need for
the equations in Appendix D
 Design methods
• Limitations of the DSM:
– No provisions for shear, web crippling nor members with holes
– Limited number/geometry of pre-qualified members
– No provisions for strength increase due to cold-work of
forming
– Overly conservative if very slender elements are used
– Shift in the neutral axis is ignored
– Cross-section cannot vary along the length
– Loads cannot vary along the length (e.g. no moment gradient)
– Global boundary conditions at the member ends are simply-
supported
– Assignment of modes sometimes difficult, particularly for
distortional buckling
Behaviour Under Compression
 Local buckling
Local buckling occurs at a load below yielding, but does not lead to
immediate failure like flexural buckling
 Local buckling
There is a change in stiffness after local buckling

Nol = local
buckling load
 Local buckling
The elastic critical stress for local buckling is given by
2
πE k
f cr =
12(1 − υ 2 ) (b / t ) 2 (1)

b/t=plate slenderness (b=width, t=thickness)

ν= Poisson’s ratio (0.3 for steel)
K= plate local buckling coefficient
 Local buckling
b Bucking Half
Coefficient Wavelength
l (k)
ss
ss ss 4.0 b
ss

ss
ss ss
free
0.43 L or 2b

built-in
ss
free
ss
1.25 1.64b

built-in

ss ss
built-in 6.97 0.66 b
 Postbuckling
After local buckling plate elements will continue to carry load but
with reduced stiffness (<50% of initial elastic stiffness)

Note stress
redistribution
from uniform
 Postbuckling
Stiffened element is a flat
compression element of
which both edges parallel
to the direction of stress
are stiffened by a web,
flange, edge stiffener,
intermediate stiffener, or
the like.

Unstiffened element is a
flat compression element
which is stiffened at only
one edge parallel to the
direction of stress
Postbuckling

S
 Postbuckling
Von Karman’s simplification – Effective width concept

fmax*be*t = actual applied load

Effective width
Effective width
 Effective width
- When elastic critical stress (fcr) = fy, failure occurs:
2
πE k
fy = (2)
12(1 − υ 2 ) (be / t ) 2

(Divide 1 by 2) be f cr
=
b fy

To allow for geometric imperfections and residual stresses from cold

forming (Winter, 1947 & 68).
be f cr  f cr 
ρ= = 1 − 0.22
b f y  f y 
 Effective width
- The equation for the effective width factor (ρ) can apply to any
design compression stress (f*) in an element calculated based on
the effective width.
- Hence, fy is replaced by f*
f*
- Assuming λ = is non-dimensionless plate slenderness,
f cr
0.22
(1 − )
be
ρ= = λ ≤ 1.0
b λ
- This is adopted in AS/NZS4600 Cl 2.2.1.2
be = b For λ ≤ 0.673:
be = ρb For λ > 0.673:
 Effective width
b = flat width of element excluding radii
ρ = effective width factor ≤ 1.0

 1.052  b  f *
λ =   
 k  t  E 
t = thickness
k = plate buckling coefficient (eg. 4, 0.425)
E = Young’s modulus (200 GPa)
f* = design stress in the compression element
For strength limit state:
- if failure occurs by yield f*=fy
- if failure occurs by other buckling modes (e.g., flexural), f*
equals the design stress for this buckling mode.
 Effective width

• The expression for calculating the effective width presented so

far is for strength calculations for elements under uniform
compressive stress.
• The effective widths for deflection calculations; for elements
with circular holes; and for elements under stress gradient
(associated with bending) are covered in Cl 2.2 and 2.3.
Effective width
Effective width
Design of Tension Members
 Tension members
Cl 3.2
N*≤φt Nt
Nt = lesser of
Ag fy and
0.85 kt An fu
N* = design tensile force
Nt = nominal section capacity
Ag = gross sectional area
An = net cross sectional area
φt = 0.9 (Table 1.6)
kt = correction factor for force distribution (0.75 – 1.0)
Design of Compression Members
 Compression members

Design complexity due to:

- local buckling of compression elements,
- flexural, torsional and flexural-torsional modes of
buckling of the whole member, and
- distortional buckling.
 Compression members

Flexural buckling (lateral buckling about a principal axis)

 Compression members
Torsional buckling (twisting about the shear centre)

Centroid

Shear
centre
 Compression members

Flexural-torsional buckling (Lateral buckling and twisting

simultaneously)
u

Centroid

Shear
centre
v
φ
 Compression members

Distortional buckling
 Compression members

Section with multiple stiffeners

Local Buckling Distortional Buckling

 Compression members
Member under concentric compression

t=1.5mm

1800mm

Hancock, 2007
 Compression members
Member under concentric compression

t=1.5mm

1500mm

Hancock, 2007
 Compression members
Member under concentric compression

Distortional buckling

Hancock, 2007
 Compression members
Failure modes
• Section failure due to yielding
- Calculate section capacity (Ns)

• Member failure due to buckling

- Calculate the member capacity (Nc) based on the governing
buckling mode (flexural, torsional, torsional-flexural or
distortional)
 Compression members
Cl 3.4
N*≤ φc Ns
N*≤φc Nc
φc = 0.85
Ns = Nominal SECTION capacity in compression
= Ae fy
Ae = effective area at yield stress (fy)
Nc = Nominal MEMBER capacity in compression
= Ae fn
Ae = effective area at the critical stress (fn)
 Compression members

(
f n = 0.658 λ2c
)f y
For λc ≤ 1.5 Equation 3.4.1(3)

 0.877 
f n =  λ2  f y For λc > 1.5 Equation 3.4.1(4)
 c 
fy
λc = = Slenderness used to determine λc
foc

foc = The least stress due to flexural, torsional

and flexural-torsional buckling.
 Compression members
Classification of sections: Not all sections are
susceptible to all buckling modes

Shear
centre (S)
 Compression members

Doubly symmetric sections and closed cross-sections

(not subjected to torsional or flexural-torsional buckling)

Only
flexural Equation 3.4.2(1)
buckling
failure

Both principal axes ( X &Y) should be checked

 Compression members

Doubly- or singly-symmetric sections subject to torsional

or flexural-torsional buckling.
foc is the smaller of:
• flexural buckling - equation 3.4.2(1); and
• flexural torsional buckling foxz - equation 3.4.3(1)

Equation 3.4.3(1)

Equation 3.3.3.2(12)
 Compression members

Point-symmetric sections subject to torsional or flexural

buckling.

foc is the smaller of:

• flexural buckling - equations 3.4.2(1); and
• torsional buckling – equation 3.3.3.2(12).
 Compression members

Non-symmetric sections

Possible failure mode is flexural-torsional buckling

about two axes.

foc is the smallest to satisfy equation 3.4.5

Equation 3.4.5
 Compression members

Singly-symmetric sections subject to distortional

buckling
Nc is the lesser of the:

fod can be calculated using appendix D for specific sections, or obtained

from elastic buckling analysis. A is the area of full cross section.
 Compression members

The Standard also covers the case of channel- or Z-

sections concentrically loaded along their longitudinal
axis, with one flange attached to sheeting by screw
fasteners (Cl 3.4.7).
Design of Flexural Members
 Flexural members

Design of beams for strength is governed by one of the following:

(a) Yielding, including allowances for local buckling and post
buckling of elements
(b) Elastic or inelastic buckling of whole beam in flexural-torsional
mode (called lateral buckling in the Standard)
(c) Elastic on inelastic buckling of a length of the beam in
distortional buckling
(d) Buckling and/or yielding of the web due to shear or combined
shear and bending
(e) Yielding and/or buckling of the web due to bearing (web
crippling) or combined bearing and bending.
 Flexural members

Hancock, 2007
 Flexural members

Hancock, 2007
 Flexural members
Cl 3.3 M*≤ φb Ms
M*≤φb Mb
φb = 0.9 (but see Table 1.6 for further details)
Ms = Nominal moment SECTION capacity
= Ze fy
Ze = effective section modulus at yield stress (fy)
Mb = Nominal moment MEMBER capacity
= Zc fc both are calculated for “lateral buckling” and
distortional buckling. Additional criteria are provided for C and
Z sections fastened to sheeting on one flange.
 Flexural members
Zc = effective section modulus calculated at a stress fc in the
extreme compression fibre
fc = Mc/Zf
Mc = critical moment
Zf = full unreduced section modulus for the extreme
compression fibre
Stocky

Intermediate

Slender
 Flexural members
For singly-symmetric sections bent about the symmetry axis, for
doubly symmetric sections bent about the x-axis and for Z-sections
bent about an axis perpendicular to the web

Coefficient for moment distribution within unrestrained segment. Can

be taken as 1.0 or

Bending moment
diagram

M3 M4 M5

Mmax

= = = =
 Flexural members
The Standard provides various expressions for Mo which include
mono-symmetric sections, members bending about their minor axis
and purlins attached to sheeting along one flange.

The standard also provides expressions for determining the

distortional buckling stress for standard sections.

Beware, ………… the computation can be demanding due to

need to iterate to find the effective section modulus.
 Flexural members
The location of the neutral axis is a function of the effective
widths of the various elements which are in turn are influenced
by the stress on the elements.
Design of Connections
 Connections
Common Types of Connections

Covered by the Standard Not covered

• Clinches
• Welding
• Self piercing rivets
• Bolts
• Nails
• Screws
• Powder actuated fasteners
• Blind rivets
(typically for steel to
concrete)
 Welds

Weld types covered

(a) Butt
(b) Fillet
(c) Arc spot
(d) Arc Seam
(e) Flare bevel

Generally the design is

similar to hot rolled steel,
except for additional failure
modes.
 Welds

For fillet welds, tearing (rupture) of plate around the weld may
occur.

Fillet welds under transverse loading (Cl 5.2.3.3)

 Welds
Fillet welds under longitudinal loading (Cl 5.2.3.2)
 Welds
For Spot welds, the shear failure modes are similar to a bolt
connection except for mode (b) below (Cl 5.2.4).

The Standard also covers tension load on spot welds.

 Bolts

• The Standard only applies if plate thicknesses are less than

3mm, otherwise use hot rolled steel provisions
• Grades X.Y (e.g. 4.6 & 8.8)

• Failure modes and capacity calculations are similar to those of

hot rolled steel, except there is an influence for when washers
are provided on both sides.

• When washers are provided on both sides, the stress

concentration is reduced and capacity is increased for relevant
modes.
 Bolts
Failure modes of a bolted connection in shear

(i) Tearing (ii) Bearing (plates

may also curl)

(iii) Net section (iv) Bolts shearing

The Standard also covers tension loading on bolted connections.

 Screws
• The Standard applies to cold-formed steel using self-tapping
screws of nominal diameter 3.0 to 7.0 mm.
• Self tapping screws create the threads through the thickness of
the steel. The screws can be with or without a self-drilling point.
• The design model is based on minimum edge distance of 1.5df.

Self tapping without self

Self drilling and self tapping drilling point
 Screws

Screws have similar shear failure modes as bolts (ie.,

tearing, bearing, net section failure and shearing) plus
Tilting

The Standard also covers tension loading on screwed connections.

The failure modes are: (i) screw pull out; (ii) screw tensile failure;
(iii) sheeting pull over screw head and washer.
 Screws
Specification of screws

Diameter

Length

Length
 Blind rivets
Also known as POP rivets

Pull-stem
rivets

Drive-pin
rivets
 Blind rivets
• Common for sheet metal construction.
• Require pre-drilled holes, but only need access from one side.
• Failure modes in shear are similar to screws .
• The shear capacity of the rivets need to be determined by
testing.
• The tension capacity of rivets is generally low and need to be
developed by testing.
• Blind rivets are also available in aluminum which have very
limited capacity.
 Self piercing rivets

Henrob rivets (www.henrob.com)

• No holes required
• Neat finish with no damage to surface
• Need access from both sides
• Reportedly more expensive
• Equipment more suited for fabrication workshops
Self piercing rivets
Clinches
 Clinches

• Force one layer into another.

• No new materials introduced.
• Need access from both sides
• Capital cost high, but competitive for volume
production.
• Two types are available, rectangular and round.
• Rectangular clinches pierce the material and is
directional but takes less energy to form.
Clinches
Clinches
 Compression members

Appendix A
Example for Compression Design using
Effective Width Method
Based on Hancock (2007)
 Compression members
Determine the design compression capacity of the following SHS
column with lex to ley = 3000mm and yield stress of 350MPa.

1. Sectional Properties of Full Section

 Compression members

< 200 Ok – Cl 3.4.1

2. Critical Stress (fn) (Cl 3.4.2)

 Compression members

3. Effective area (Ae) at fn

be = b for λ ≤ 0.673 & be = ρb for λ > 0.673
 1.052  b  f *
λ =   
 k  t  E 
 Compression members

(flanges of SHS are stiffened elements)

f* = fn = 166.6 MPa

4. Effective area (Ae) at fy

f* = fy = 350 MPa
 Compression members

λ= 0.726 > 0.672

0.22 0.22
(1 − ) (1 − )
ρ= λ = 0.726 = 0.96
λ 0.726
be = ρb = 0.96*66 = 63.4mm
Ae =557 mm2
5. Design Compression Capacity (Cl 3.4.1)

Ns= Aefy = 557 * 350 * 10-3= 195.0 kN

Nc= Aefn = 578 * 166.6 * 10-3= 96.3 kN

Design capacity = φc * 96.3 = 81.9 kN

 Appendix B

Direct Strength Examples

 Direct strength method

• The effective width method is used to account for local buckling

and postbuckling by using reduced area of compression
elements.
• The effective width method has been extended indirectly to
cover distortional buckling.
• For complex sections, the calculations of the effective widths
become very demanding.
• Shafer and Peköz in 1998 proposed the Direct Strength Method
(DSM) to overcome the complexities of the effective width
method.
• DSM uses elastic buckling solutions for the entire member
rather than individual elements.
• DSM is included in AS/NZS in Section 7.
 Direct strength method
• Developed for columns and beams.
• Not developed explicitly for beam-columns
• Not developed explicitly and calibrated for shear.
• DSM uses a software (typically finite strip analysis) to determine
the elastic buckling solutions.
• From finite strip analysis, the following are determined:
– local buckling stress due to compression or bending
– distortional buckling stress due to compression or bending
• Finite strip analysis does not take into account member boundary
conditions (apart from simply supported), hence, it is not used
for determining the flexural and flexural-torsional (overall)
buckling stresses. Section 3 of the Standard is used for this.
 Direct strength method

• Softwares such as CUFSM and Thinwall perform the elastic

buckling analysis and produce the necessary “Signature Curve” for
the analysed section.
• Example:- Determine the signature curve for the following purlin
section in compression
 Direct strength method
Signature curve from Thinwall for purlin in compression
 DSM - Compression
• The nominal compression member capacity (Nc) is the lesser of:
(a) minimum of nominal member compression capacity (Nce) for: (i)
flexural buckling; (ii) torsional buckling or (iii) flexural-torsional
buckling. The lowest of these three buckling stresses is (foc).
Nce α Noc = Afoc ( no need to calculate effective area)
(b) nominal member capacity compression for local buckling(Ncl).
Ncl α Nol = Afol (no need to calculate effective area)
(c) nominal member capacity compression for distortional buckling (Ncd).
Ncd α Nod = Afod (no need to calculate effective area)
• The design member capacity = φcNc,
• φc = 0.85 for qualified sections or 0.8 otherwise
 DSM - Bending
• The nominal moment member capacity (Mb) is the lesser of:
(a) nominal member moment capacity for lateral-torsional buckling
(Mbe).
Mbe α Zffy (no need to calculate effective modulus)
(b) nominal member moment capacity for local buckling (Mbl)
Mbl α Mol = Zffol
(c) nominal member moment capacity for distortional buckling
(Mbd).
Mbd α Mod = Zffod
• The design member moment capacity = φbMb,
• φb = 0.9 for qualified sections or 0.8 otherwise
 DSM – Compression example
• For nominated purlin section shown earlier, determine Nc assuming
lex = 2m, ley = 1m and lez = 1m and fy = 300MPa. (Hancock, 2007)

A. From the section signature curve in compression

(Full cross sectional area)

DSM – Compression example

(based on Cl 3.4)
 DSM – Compression example

Nc from the effective width method is 67.75kN. This can be obtained

using Swin Channel spreadsheet.
 DSM – Bending example
• For nominated purlin shown below, determine Mb assuming the
loading as below (Hancock, 2007)
 DSM – Bending example
A. Signature curve from Thinwall for purlin in bending
DSM – Bending example

(full section modulus)

(from Cl 3.3.3.2)

(from Cl 3.3.3.2.12)
DSM – Bending example
 DSM – Bending example

Mb = 7.14kNm

Mb from the effective width method is 6.67kNm. This can be

obtained using the provided Swin Channel spreadsheet.
Thank You