View metadata, citation and similar papers at core.ac.

uk brought to you by CORE

provided by Missouri University of Science and Technology (Missouri S&T): Scholars' Mine

Missouri University of Science and Technology

Scholars' Mine

International Specialty Conference on Cold- (1984) - 7th International Specialty Conference

Formed Steel Structures on Cold-Formed Steel Structures

Nov 13th, 12:00 AM

Load and Resistance Factor Design of Cold-formed Steel

Structural Members
Theodore V. Galambos

Wei-wen Yu
Missouri University of Science and Technology, [email protected]

Follow this and additional works at: https://scholarsmine.mst.edu/isccss

Part of the Structural Engineering Commons

Recommended Citation
Galambos, Theodore V. and Yu, Wei-wen, "Load and Resistance Factor Design of Cold-formed Steel
Structural Members" (1984). International Specialty Conference on Cold-Formed Steel Structures. 2.
https://scholarsmine.mst.edu/isccss/7iccfss/7iccfss-session10/2

This Article - Conference proceedings is brought to you for free and open access by Scholars' Mine. It has been
accepted for inclusion in International Specialty Conference on Cold-Formed Steel Structures by an authorized
administrator of Scholars' Mine. This work is protected by U. S. Copyright Law. Unauthorized use including
reproduction for redistribution requires the permission of the copyright holder. For more information, please
contact [email protected].
 Seventh International Specialty Conference on Cold-Formed Steel Structures
St. Louis, Missouri, U.S.A., November 13-14, 1984

LOAD AND RESISTANCE FACTOR DESIGN OF COLD-FORMED STEEL

STRUCTURAL MEMBERS

by

Theodore V. Galambos l and Wei~Wen Yu 2

I. INTRODUCTION

In the United States and some other countries, the current

method of designing cold-formed steel structural members is based
on the allowable stress design method. In this approach, the
stresses in structural members are computed by using accepted
methods of structural analysis for the specified service loads.
These stresses should not exceed the allowable stresses per-
mitted by the applicable design specification (Refs. I and 2).

Recently, the load and resistance factor design method for

designing hot-rolled and built-up members has been proposed in the
United States (Ref. 3). The limit states design method has been
used in Canada and Europe for sometime as an alternate to the
allowable stress design criteria (Refs. 2 and 4). In this method,
separate load and resistance factors are applied to specified loads
and nominal resistance to ensure that the probability of reaching
a limit state is acceptably small. These factors reflect the un-
certainties of analysiS, design, loading, material properties and
fabrication.

In order to develop load and resistance factor design (LRFD)

criteria for cold-formed steel structural members, a joint re-
search project was sponsored by American Iron and Steel Institute
under the direction of the authors.

I Professor in Civil Engineering, Department of Civil and Mineral

Engineering, University of Minnesota, Minneapolis, Minnesota.

2 Curators' Professor of 'Civil Engineering, University of Missouri-

Rolla, Rolla, Missouri.

519
520 SEVENTH SPECIALTY CONFERENCE

Subsequently, tentative recommendations with a commentary have

been prepared for consideration of the American Iron and Steel
Institute (Ref. 5). This proposed document contains six sections
for designing cold-formed steel structural members and connections.
The background information for developing the proposed design
criteria for structural members is discussed in this paper. For
connections, additional information can be found .in Ref. 5.

II. DESIGN PROCEDURE

A. Load and Resistance Factor Design

Ae,discussed in the Introduction, the current method of

designing cold-formed steel structural members, as presented in the
1980 AISI Specification (Ref. 1), is based on the allowable stress
design method. The allowable stress is determined by dividing a
stress at a limit state by a factor of safety. Usual factors 0f
safety inherent in the AISI Specification for the Design of Cold-
Formed Steel Structural Members are 5/3 for beams and 23/12 for
columns.

A limit state is the condition at which the structural useful-

ness of a loa.d-carrying element is impaired to such an extent that
it becomes unsafe for the occupants of the structure, or the element
no longer performs its intended function. Typical limit states for
cold-formed steel members are excessive deflection, yielding, buck-
ling and attainment of maximum strength after local buckling (i.e.,
post-buckling strength). These limit states have been established
through experience in practice or in the laboratory, and they have
been thoroughly investigated through analytical and experimental
research. The background for· the establishment of the limit states
is extensively documented in the Commentary on the AISI Specification
(Refs. 6 and 7), and a continuing research effort provides further
improvement in understanding them.

The factors of safety are provided to account for the uncertainties

and variabilities inherent in the loads, the analysis, the limit state
model, the material properties and the geometry. Through experience it
has been established that the present factors of safety provide satis-
factory design.

The allowable stress design method employs only one factor of

safety for a limit state. The use of multiple load factors provides a
refinement in the design which can account for the different. degrees
of the uncertainties and variabilities of the design parameters. Such
a design method is called load and resistance factor design, and its
format is expressed by the following design criterion:
 LOAD AND RESISTANCE FACTOR DESIGN 521

(1)

where Rn the nominal resistance

~ resistance factor
Yi load factors
Qi load effects

The nominal resistance is the strength of the element or member

for a given limit state, computed for nominal section properties
and for minimum specified material properties according to the
appropriate analytical model which defines the strength. For a
column, for example, ~ = AF cr ' where A is the cross-sectional area
and Fcr is the buckling stress. The resistance factor ~ accounts
for the uncertainties and variabilities inherent in Rn , and it is
usually less,.than unity. The load effects Qi are the forces on the
cross section (bending moment, axial force, shear force) determined
from the specified minimum loads by structural analysis, and the
Yi's are the corresponding load factors which account for the un-
certainties and variabilities of the loads. The load factors are
greater than unity.

The advantages of LRFD are: (1) the uncertainties and the

variabilities of different types of loads and resistances are differ-
ent (e.g., dead load is less variable than wind load), and so these
differences can be accounted for by use of multiple factors, and (2)
by using probability theory all designs can achieve ideally a uniform
reliability. Thus LRFD provides the basis for a more rational and
refined design method than is possible with the allowable stress
design method.

B. Probabilistic Concepts

Factors of safety or load factors are provided against the uncer-

tainties and variabilities which are inherent in the design process.
Structural design consists of comparing nominal load effects Q to
nominal resistances R, but both Q and R are random parameters. A
limit state is violated if R<Q. While the possibility of this event
ever occurring is never zero, a successful design should, nevertheless,
have only an acceptably small probability of exceeding the limit state.
If the exact probability distributions of Q and R were known, then
the probability of R - Q < 1 could be exactly determined for any design.
In general the distributions of Q and R are not known, and only the
means, Qm and Rm, and the standard deviations, 0Q and oR are available.
Nevertheless it is possible to determine relative reliabilities of
several designs by using the concept of the "reliability index " 6,
which is extensively discussed in Refs. 8 through 11. This reliability
index can be expressed by the equation

R.n(Rm/Qm)
6 (2)
Iv2+v2
R Q
522 SEVENTH SPECIALTY CONFERENCE

where VR = 0R/Rm and VQ = 0Q/Qm' the coefficients of variation of

Rand Q, respectively. The "reliability index" B is a relative
measure of the safety of the design. When two designs are compared,
the one with the larger B is more reliable.

The concept of the reliability index can be used in deter-

mining the relative reliability inherent in current design, and it
can be used in testing out the reliability of new design formats,
as illustrated by the following example of simply supported braced
beams with stiffened flanges subjected to dead and live loading.

The design requirement of the 1980 AISI Specification for

such a beam is

Se ff Fy 2
~ (D
n + Ln)
(3)
(FS) 8
where S the section modulus based on the effective cross section
eff
FS 5/3 = the factor of safety
F the specified yield point
y
Q, the span length and s = the beam spacing
D and L are, respectively, the code specified dead and
l~ve loa~ intensities.
The mean resistance is defined as (Ref. 8)

(4)
In this equation Rn is the nominal resistance, which in this case
is
(5)

that is, the ultimate moment predicted on the basis of the post-
buckling strength of the compression flange. The mean values P ,
M , and F , and the corresponding coefficients of variation V ,mv
a~d VF , a~e the statistical parameters which define the variaRilify
of the resistance:
P the mean ratio of the experimentally determined
m
ultimate moment to the predicted ultimate moment
for the actual material and cross-sectional pro-
perties of the test specimens;
M mean ratio of the yield point to the minimum
m
specified value;
F mean ratio of the section modulus to the Hand-
m
book (nominal) value.
The coefficient of variation of R equals

(6)
 LOAD AND RESISTANCE FACTOR DESIGN 523

The values of these data were obtained from examining all

available tests on beams with stiffened compression flanges, and
from analyzing data on yield point values from tests and cross-
sectional dimensions from many measurements. This information
is developed in Refs. 12 and 13, and it is given below:
Pm = 1.08, Vp = 0.10; Mm = 1.10, VM = 0.10; Fm = 1.0, VF = 0.05
and thus Rm = 1.19 Rn and VR = 0.15.

The mean load effect is equal to

(7)

and

~(Dm VD)2 + (Lm VL )2

vQ = D + L
(8)
m m
where Dm and Lm are the mean dead and live load intensities,
respectively, and VD and VL are the corresponding coefficients of
variation.

Load statistics have been analyzed in Ref. 9, where it was

shown that
Dm = 1.05 Dn , VD = 0.1; Lm = Ln , VL = 0.25.
The mean live load intensity equals the code live load intensity if
the tributary area is small enough so that no live load reduction
is included. Substitution of the load statistics into Eqs. 7 and 8
gives

t 2s (1.05 Dn+
Qm =-8- Ln 1) Ln (9)

_IC·~:Dn)2 VD2 + VL2

VQ
C.05 L
Dn + 1)
(10)

n
Q and V thus depend on the dead-to-live load ratio. Cold-formed
b~ams tY~ically have small Dn/Ln, and for the purposes of checking
the reliability of these LRFD criteria it will be assumed that Dn/Ln
1/3, and so Qm = 1.35Ln (t2 s/8) and VQ = 0.19.

From Eq. 3 we obtain the nominal design capacity for D /L 1/3

and FS = 5/3. Thus n n

R 1.19 x 2.22 x L (t2 s/8)

m n
Qm = 1.96
1. 35L (t 2 s/8)
n
524 SEVENTH SPECIALTY CONFERENCE

and, from Eq. 2:

8 £n 1.96 = 2.78
/0.15 2 + 0.19 2
Of itself S = 2.78 for beams with stiffened compression flanges
designed by the 1980 AISI Specification means nothing. However,
when this is compared to 8 for other types of cold-formed members,
and to 8 for designs of various types from hot-rolled steel shapes
or even for other materials, then it is possible to say that this
particular cold-formed steel beam has about an average reliability
(Ref. 10).

C. Basis for LRFD of Cold-Formed Steel Structures

A great deal of work has been performed for determing the

values of the reliability index 8 inherent in traditional design as
exemplified by current structural design specifications such as
the AISC Specification for hot-rolled steel, the AISI Specification
for cold-formed steel (Ref. 1), the ACI Code for reinforced concrete
members, etc. The studies for hot-rolled steel are summarized in
Ref. 8, where also many further papers are referenced which contain
additional data. The determination of 8 for cold-formed steel element,
or members is presented in Refs. 12 through 16, where both the basic
research data as well as the 8' s inherent in the AISI Spe'cification
are presented in great detail.

The entire set of data for hot-rolled steel and cold-formed steel
design. as well as data for reinforced concrete, aluminum, laminated
timber, and masonry walls was re-analyzed in Refs. 9 through 11 by
using a) updated load statistics and b) a more advanced level of pro-
bability distributions which describe the true distributions more
realistically. The details of this extensive reanalysis are presented
in Refs. 9 through 11 and so only the final conclusions from the
analysis are summa.r ized here:

1) The values of the reliability index 8 vary considerably for

the different kinds of loading, the different types of construction,
and the different types of members within a given material design
specification. In order to achieve more consistent relia.bility,
it was suggested that the folloWing values of S would provide this
improved consistency while at the same time give, on the average,
essentially the same design by the new LRFD method as is obtained by
current design for all materials of construction. These target
reliabilities So for use in LRFD are:
Basic case: Gravity loading, 80 = 3.0
For wind loading: 80 = 2.5
 LOAD AND RESISTANCE FACTOR DESIGN 525

2) The following load factors and load combinations were

developed in Refs. 9 and 11 to give essentially the same S's as
the target S 's, and are recommended for use with the 1980 ANSI
Load Code (Rgf. 17) for all materials, including cold-formed steel:

1.4Dn
1.2Dn + 1.6 (Ln or Sn)
1.2D n + 0.5 (Ln or Sn) + 1.3Wn
1.2Dn + 0.5 (Ln or Sn) + 1.5En
1.2D n + 0.5Ln + 1.6S n
1.2Dn + 1.6Sn + 0.8Wn
0.9Dn - 1.3Wn
0.9Dn - 1.5En
1. 2Dn + 1.2Pn
where Dn nominal dead load
Ln nominal live load due to occupancy
Sn nominal snow load
WtJ. nominal wind load
En nominal earthquake load
Pn nominal ponding load, including the increase due
to ponded liquid.

Deflection calculations for serviceability criteria are to

be made with the appropriate unfactored loads.

The load factors and load combinations given above have been
recommended for use with the LRFD criteria for cold-formed steel.
The following portions of this paper present the background for the
resistance factors ¢ listed in Table 1, and which are recommended
for use in the AISI LRFD Specification. These ¢ factors were deter-
mined in conformance with the load factors given above to approximately
provide a target S of 3.0 for members and 4.0 for connections,
respectively, for ~he load combination 1.2Dn + 1.6L n . For practical
reasons it is desirable to have relatively few different resistence
factors, and so the actual values of S will differ from the derived
targets. This means that
¢Rn = c(1.2D n + 1.6Ln ) = (1.2Dn/Ln + 1.6) cLn (11)
where c is the deterministic influence coefficient translating load
intensities to load effects.

By assuming Dn/Ln = 113, Eqs. 11 and 9 can be rewritten as follows:

Rn 2.00(cL n /¢) (12)
Qm 1. 350cLn (13)
Therefore,

(14)
526 SEVENTH SPECIALTY CONFERENCE

TABLE 1

Resistance Factors

(Section 8.3.5 of the Proposed Tentative Recommendations)

Type of Strength Resistance Factor, ¢

(a) Tension members 0.90

(b) Flexural members

Section strength 0.90

Laterally unbraced beams 0.90

Web design

Shear strength* 0.90

Flexural strength 0.90
Web crippling 0.80

(c) Axially loaded compression members 0.80

(d) Beam - columns

¢c 0.80
¢s 0.90
¢ 0.90

(e) Cylindrical tubular members

Flexural strength 0.90

Axial compression 0.80

*When h/t < 1.0

The ¢ factors can be computed from Eq. 14 and the following equation
by using VQ = 0.19:
tn(Rm/~)
Target So = (15)
hR2 + VQ2
In the above calculation, the values of (Rm/Qm) and VR can be obtained
from Refs. 12 through 16.
 LOAD AND RESISTANCE FACTOR DESIGN 527

The resistance factors ¢ can also be determined for the

desired S and the resistance statistics ~/Rn and VR from charts
provided ~n Ref. 9. For example, for the cold-formed beams with
stiffened compression flanges, for which Rm/Rn = 1.19 and VR = 0.15,
for a D/L ratio of 1/3 and So = 3.0, ¢ = 0.86 f~om the charts in
Ref. 9.

III. DESIGN OF MEMBERS

A. Yield Point

The following statistical data (mean values and coefficients

of variation) on material and cross-sectional properties were
developed in Refs. 12 and 13 for use in the derivation of the re-
sistance factors ¢:
(Fy)m = 1.10 Fy; Mm = 1.10; VFy = VM = 0.10
(Fya)m= 1.10 Fya; Mm = 1.10; VFya = VM = 0.11
(Fu)m = 1.10 Fu; Mm= 1.10; VFu = VM = 0.08
Fm = 1.00; VF = 0.05
The subscript m refers to mean values. The symbol V stands for
coefficient of variation. The symbols M and F are, respectively,
the ratio of the mean-to-the nominal material property or cross-
sectional property; and Fy ' Fya ' and Fu are, respectively, the
specified minimum yield point, the average yield point including
the effect of cold forming, and the specified minimum tensile
strength.

These data are based on the analysis of many samples, and

they are representative properties of materials and cross sections
used in the industrial application of cold-formed steel structures.

B. Tension Members

The resistance factor of ¢ = 0.90 used for tension member

design was derived from the procedure described in Section II.A of
this paper and a selected S value of approximately 3.0. In the
determination of the resistgnce factor, the following formulas were
used for Rm and Rn:

R A (F ) (16)
m n y m
R A F (17)
n n y

(18)
in which A is the net area of the cross section, (F) is equal to
1.10 Fy asndiscussed in Section III.A of the paper. YB~ using
VM = 0.10, VF = 0.05 and Vp = 0, the coefficient of variation VR is:
528 SEVENTH SPECIALTY CONFERENCE

(19)

Based on Vs= 0.19, the resistance factor is approximately 0.85

for S = 3. , which is rounded off to ¢ = 0.90, glvlng S = 2.7
which is about the same reliability as that of beams.

C. Flexural Members

Flexural members are differentiated according to whether or

not the member is laterally braced. If such members are laterally
supported, then they are proportioned according to the strength of
the cross section. If they are laterally unbraced, then the limit
state is lateral-torsional buckling. Cross section strength depends
on whether or not the compression flange is composed of stiffened
or unstiffened elements.

i) Section Strength

a) Flexural Members with Stiffened Compression Flange

The strength of beams with a compression flange having stiffened

elements is based on the post-buckling strength of the member, and
use is made in LRFD of the effective width concept in the same way
as in the 1980 AISI Specification (Ref. 1). References 6 and 7
provide an extensive treatment of the background research.

The experimental basis for the post-buckling strength of cold-

formed beams is examined in Ref. 12, where Table 3 gives the calculation
of the predicted strength according to Winter's effective width
formulas. A total of 43 tests were examined, and the statistics are
summarized as follows:

Pm = 1.08, Vp = 0.10
The symbol P is the ratio of the experimental strength to the strength
predicted by the effective width theory for the material and cross-
sectional properties of the test specimens. The mean and coefficient
of variation of the resistance are equal to:
R
m
1.08 x 1.10 x 1.0 R 1.19 R
n n
and
va = !Gp 2 + vM2 + VF2 = ~0.102 + 0.11 2 + 0.05 2 = 0.16
The nominal strength Ru is based on the nominal effective cross section
and on the specified minimum yield point, i.e., Rn = SeffFy.

The value of S, as determined from the charts in Ref. 9 for the

selected value of ¢ = 0.9 for a dead-to-live load ratio of 1/3 is
approximately 2.73.
 LOAD AND RESISTANCE FACTOR DESIGN 529

b) Flexural Members with Unstiffened Compression Flanges

The basis for the prediction of the strength of beams with

unstiffened compression flanges in these LRFD criteria and in the
1980 AISI Specification is the plate buckling theory. The data
of the tests are given in Table 3 of Ref. 13, and they are summarized
as follow:

for 63.3/~ < wit < 25; Pm = 1.24; Vp = 0.13 (for 24 tests)
for 25 < wit < 60; Pm = 1.76, Vp = 0.21 (for 26 tests)
where wit is the width/thickness ratio of the unstiffened flange
element. If all 50 tests are averaged, Pm = 1.51 and Vp = 0.26.
It is evident from these data that the theory underestimates the
capacity considerably. This has long been noted, and a generalized
effective-width theory, including both stiffened and unstiffened
compression flanges, has been proposed (Ref. 18). The same 50 test
results with this improved theory give Pm = 1.04 and Vp = 0.14.
Since the intent of these LRFD criteria is to provide only a trans-
lation from the 1980 Allowable Stress Design criteria (Ref. 1)
into a LRFD format, no change in the basic treatment of the under-
lying theory was made. The ¢-factor is derived as follows:
for 63.3/~ < wit < 25:
Rm/Rn = Pm Mm Fm = 1.24 x 1.10 x 1.0 1.36

VR = tG p2 + VM 2 + VF2 = vO~132 + 0.10 2 + 0.05 2 0.17

for 25 <,wit < 60:
Rm/Rn = 1.76 x 1.0 x 1.0 = 1.76

VR = ;6.21 2 + 0.06 2 + 0.05 2 = 0.22

In the latter case the limit state is elastic buckling. M = 1.0 and
Vm = 0.06 have been used to account for the basic materiaT variable,
the elastic modulus, E.

The ranges of Rm/Rn and VR in both instances are beyond the

charts provided in Ref. 9. The procedure will thus be to select
a value of ¢ and then to determine the resulting reliability index
using Eq. 2. For a dead-to-live load ratio of Dn/Ln = 1/3, the load
effect data is Qm = (1.05 Dn/Ln + 1) cLn = 1.35 cLn and VQ = 0.19.
According to the LRFD load factors
¢ Rn = c (1.2Dn + 1.6Ln ) = cLn (1.2Dn/Ln + 1.6) = 2.0 cLn
or 2.0cL n
Rn = - ¢ - -
530 SEVENTH SPECIALTY CONFERENCE

Table 2 summarizes the results. The selected value of $ = 0.9 in

the LRFD criteria thus furnishes a greater reliability than the
target value of 60 = 3.0.

TABLE 2

Reliability Index Values

Rm/Rn VR ¢ 6
1.36 0.17 0.90 3.2
1.36 0.17 0.95 3.0
1. 76 0.22 0.90 3.7
1. 76 0.22 0.95 3.5
1. 76 0.22 1.00 3.3

ii) Laterally Unbraced Beams

There are not many test data on laterally unsupported cold-

formed beams. The available test results are summarized in Ref.
15, and they are compared with predictions from elastic buckling
theory which states that for a simply supported 1- or Channel-
shaped beam bent about the major axis by a uniform moment, the
buckling moment is equal to:

2
+GJL
--
2
(20)
1f ECw
where L = unbraced length
Iy = minor axis moment of inertia
J = torsion constant
Cw = warping constant
G = shear modulus
E = elastic modulus
The statistical data from Ref. 15 are the following:
Pm = 1.15 and Vp = 0.15
MID= 1.0 and Vm 0.06
Fm = 1.0 and VF = 0.05
and thus
Rm/~ = 1.15 x 1.0 x 1.0 = 1.15 and VR = 10.15 2+0.06 2+0.05 2=0.17
The symbol P is the ratio of the test capacity to the lateral-torsional
buckling strength predicted by Eq. 20, M is the ratio of the actual to
the specified value of the modulus of elasticity, and F is the ratio
of the actual to the nominal sectional properties.

Using the values of Rm/Rn = 1.15 and VR = 0.17, the recommended

resistance factor ¢ = 0.90 gives 6 = 2.5. This relatively low value
of 6 can be justified by noting that a simplified:and conservative form
of Eq. 20 is the basis of the design criteria (same as in the allow-
able stress design rules in the AISI Specification of 1980; the second
square root in Eq. 20 is taken to be unity).
 LOAD AND RESISTANCE FACTOR DESIGN 531

iii) Web Strength

For the design of beam webs, consideration should be given to

the shear strength, bending strength, combined bending and shear,
and web crippling.

a) Shear Strength of Beam Webs

The shear strength of beam webs is governed by either yielding

.or buckling, depending on the hIt ratio and the mechanical properties
of steel. For beam webs having small. hIt ratios, the shear strength
is governed by shear yielding, i.e.: '

(21)

in which Aw is the area of the beam web computed by (hxt), and Ty is

the yield point of steel in shear.

For beam webs having large hIt ratios, the shear strength is
governed by elastic shear buckling, i.e.:
k 'IT2EA
Vu = AwTcr = v w (22)
12 (l-i) (h}t) 2
in which Tcr is the critical shear buckling stress in the elastic range,
k is the shear buckling coefficient, E is the modulus of elasticity,
~vis the Poisson's ratio, h is the web depth, and t is the web thickness.
By using E = 29,500 ksi and ~ = 0.3, the shear strength, Vu , can be
determined as follows:
26,700 kv Aw
v u
(23)
(hIt) 2
For beam webs having moderate hIt ratios, the shear str~ngth is
based on inelastic buckling, i.e.:

(24)

In the above equation, the maximum shear stress is based on the allow-
able shear stress specified in Section 3.4.1 of the AISI Specification
and a safety factor of 5/3.

In view of the fact that the appropriate test data on shear are
not available, the ~ factors were derived from the condition that the
nominal resistance for the· LRFD method is the same as the nominal
resistance for the allowable stress design method. Thus,
~Rn)LRFD = (Rn)ASD (25)
532 SEVENTH SPECIALTY CONFERENCE

Since (Rn)LRFD ~ c(1.2Dn+l.6L n )/¢ (26)

(Rn)ASD ~ c(F.S.) (Dn+Ln) (27)
the resistance factors can be computed from the following formula:

1.2(D /L ) + 1.6
¢ = (F.S.~(Dn/L
n n
+ 1) (28)

By using a dead-to-live load ratio of D IL = 1/3, the ¢ factors

computed from the above equation are li~tea in Table 3 for three
different ranges of hit ratios. The factors of safety are adopted
from the AISI Specification for allowable stress design. It should
be noted that the use of a small safety factor of 1.44 for yielding
in shear is justified by long-standing use and by the minor consequences
of incipient yielding in shear compared with those associated with
yielding in tension and compression.

TABLE 3

Resistance Factors for S~ Strength of Beam Webs

F. S. for ¢ Factor Recommended

Range of hit Ratio Allowable Stress Computed ¢ Factor
Design by Eq. 28

h/t< l71/k7F 1.44 1.04 1.00

- v Y

l7l/k7F ~ h/t< 243~ 1.67 0.90 0.90

v y- - v Y
hit > 243/k7F 1.71 0.88 0.90
v y

b) Flexural Strength of Beams Governed by Webs

The flexural strength of beams is governed by either yielding or

buckling of beam webs.

The bending strength of beams governed by webs was studied in

Refs. 16 and 19 by comparing the experimental data and the predicted
results. Based on a study made on beams having stiffened and unstiff-
ened flanges, the statistical data are as follows (Table III of Ref.
19) :
(a) Beams having stiffened flanges
Pm = 1.00; Vp 0.08
Mm = 1.10; Vm 0.10
Fm = 1.00; VR 0.05
Rm/Rn = 1.10; V){ 0.14
(b) Beams having unstiffened flanges
Pm 0.99; Vp 0.09
Mm = 1.10; VM 0.10
Fm = 1.00; VF 0.05
lim/Rn = 1.09; VR 0.14
 LOAD AND RESISTANCE FACTOR DESIGN 5::33

For ¢ = 0.90, the computed S's are 2.52 and 2.48 for beams
having stiffened flanges and beams having unstiffened flanges,
respectively.

On the basis of the statistical analysis of the available test

data on web crippling, the values of Pm,Mm,Fm,Vp ' VM, and VF were
computed and selected. These values are presented in Table 4 (See
Table III of Ref. 19). By using So = 3.0 and the computed values
of VR for different conditons, the resistance factors, ¢, were
calculated by using Eq. 11 for various conditions as listed in Table
4. For the purpose of simplicity, the value of ¢ = 0.80 is used.
The values of S corresponding to this value of ¢ are also given in
Table 4.

D. Axially Loaded Compression Members

The available experimental data on cold~formed steel axially

loaded compression members were evaluated in Ref. 14, The test
results were compared to the predictions based on the same math-
ematical models on which the AISI Specification (Ref. 1) was based.
The design provisions in these LRFD criteria are also based on the
same mathematical models.

a) Cross-Sectional Strength

Axially loaded columns are designed against overall instability

and local instability. This latter effect is included through the
use of the Q-factor in the column equations where this is appropriate.
For columns the resistance factor ¢ thus includes both types of in-
stability. Beam-columns are designed both against an overall stability
limit state and against a member strength limit state separately.
Therefore it is necessary to derive a value of ¢ for member strength
to be used in beam-column design. The basis for the determination
of ¢ for the limit state of member strength is the capacity of a
compressed short member. Stub column strength is predicted from the
effective-width concept for members with stiffened elements, and the
theory of plate buckling is used for the prediction of the capacity
of members with unstiffened elements. This latter theory is overly
conservative, and a generalized effective-width formula has been
developed for use with both stiffened and unstiffened elements (Ref.
18). However, the new recommendations have not yet been incorporated
into the AISI Specification as of this date (1984), and so the buckling
limit state is retained here for unstiffened elements. It should be
noted that the statistical evaluation of the test results in Refs.
12, 13, and 14 also includes the comparisons with the generalized
effective-width approach. Thus the necessary information to develop
new ¢-factors when the specification is changed is already developed.

Stiffened Elements

Stub-column strength was analyzed in Ref. 12 by comparing the

experimental strength to the prediction from the effective-width
(post-buckling strength) theory. A total of 44 tests were reported,
 01
TABLE 4 CI:I

Computed ~-Factors for Web Crippling """

R
m 13 for
Case R P M VM F VF P Vp VR cp value
n u m m m R for
n
130 '" 3.0 cp = 0.80
(a) Single, Unreinforced Webs
- End One-Flange Eq.* (9.3.3-9) 1.10 0.10 1.00 0.05 1.00 0.12 1.10 0.16 0.77 2.9
Loading
(Stiffened Flanges) ~
- End One-Flange Eq. (9.3.3-10) 1.10 0.10 1.00 0.05 1.00 0.16 1.10 0.20 -0.71 2.6 ~
Loading
(Unstiffened Flanges) ~
00
- Interior One-Flange Eq. (9.3.3-11) 1.10 0.10 1.00 0.05 0.99 0.11 1.09 0.16 0.77 2.8
Loading
~
o
- End Two-Flange Eq. (9.3.3-12) 1.10 0.10 1.00 0.05 0.99 0.09 1.09 0.14 0.80 3.0
:;
Loading ~
.....::I
- Interior Two-Flange Eq. (9.3.3-13) 1.10 0.10 1.00 0.05 0.98 0.10 1.08 0.15 0.79 2.9 o
Loading o
(b) I-Sections ~
- End One-Flange Eq. (9.3.3-14) 1.10 0.10 1.00 0.05 1.10 0.19 1.21 0.22 0.75 2.8 ~
Loading
Z
o
t.%J
Interior One-Flange Eq. (9.3.3-15) 1.10 0.10 1.00 0.05 0.96 0.13 1.06 0.17 0.73 2.7
Loading

- End Two-Flange Eq. (9.3.3-16) 1.10 0.10 1.00 0.05 1.01 0.13 1.11 0.17 0.77 2.8
Loading

- Interior Two-Flange Eq. (9.3.3-17) 1.10 0.10 1.00 0.05 1.02 0.11 1.12 0.16 0.79 2.9
Loading
*Equ.ation Numbers refers to the corresponding equations in the AISI LRFD Specification.
 LOAD AND RESISTANCE FACTOR DESIGN 535

and the statistical data are as follows (Table 4 in Ref. 12):

Pm = 1.10; Vp 0.10
~ = 1.10; VF = 0.10
Fm = 1.0; VF = 0.05
Rm/Rn = 1. 21 ; VR = 0.15
The reliability index S. as determined from the charts in Ref.
9 for a Dn/Ln of 1/3 is 2.9 for the selected value of ~s = 0.9.

Unstiffened Elements

The strength of stub-columns with unstiffened elements was

analyzed in Ref. 13 according to the plate buckling theory. and the
statistical data from Table 4 in Ref. 13 are as follows:

a) width-thickness ratios < 25

Number of data: 22*
Pm 1.08; Vp = 0.11
Mm = 1.10**; VM 0.10
Fm = 1.0; VF 0.05
R;;/Rn = 1.19; VR 0.16
For the recommended ~s 0.9 the reliability index is S 2.7
when Dn/Ln = 1/3 (Ref. 9).
b) width-thickness ratios > 25
Number of data:
Pm = 1.69; 0.18
~ = 1.00**; 0.06
Fm = 1.0; 0.05
Rro/Rn = 1.69; 0.20
For the recommended ~s 0.9. the reliability index S = 3.7 when
D /L = 1/3. This is considerably above the target of So 3 and a
vR1u~ of ~s = 1.0 could have been justified. However. ~s = 0.9 is
recommended for the sake of consistency.

Column Strength

Column capacity in these LRFD criteria is based on the same pre-

diction models as were employed in the formulation of the AISI Specifica-
tion: elastic buckling theory for the case of slender columns. and the

* Last test from Table 4b in Ref. 13 is included in the data for

w/t > 25.
** Limit state is inelastic buckling. and so the statistics of the
yield stress are used here.
*** This includes the last test point from Table 4b and all data from
Table 4c of Ref. 13. except that the last two tests were omitted.
These stub columns had w/t of about 60 and their inclusion would
have biased the results unduly.
536 SEVENTH SPECIALTY CONFERENCE

tangent modulus theory for columns of intermediate and short length.

Two types of limit states are considered: flexural buckling in the
plane perpendicular to the minor principal axis (FB) and torsional-
flexural buckling (TFB). In the latter case it is required that
the cross section is compact. i.e •• Q = 1.0. while in the case of
FB the cross-sectional strength of noncompact shapes is accommodated
through the use of Q < 1.0. same as in the 1980 AISI Specification.

The resistance factor ~ = 0.80 was selected on the basis of

the statistical data given ia Ref. 7. The summary of the informa-
tion is given in Table 5.

The reliability index S was determined from the charts in Ref.

9 for a DnlRn ratio of 1/3. The target of So = 3.0 is not entirely
satisfied. and different ~-factors could have been used for the differ-
ent cases.

E. Beam-Columns

With the exception of one set of beam-column tests (see Ref. 15)
for hat shapes for which the limit state was torsional-flexural buck-
ling. there are no tests of cold-formed s,teel beam-columns. The LRFD
design criteria provides the same interaction equa'tions as the 1980
Edition of the AISI Specification (Ref. 1). with ~c = 0.80 (i.e ••
as recommended for columns) when the limit state is overall member
instability. and ~s = 0.90 (i.e •• as recommended for laterally braced
beams) when the limit state is section strength. In the calculation
of the factored nominal beam strength. ~u. the ~-factor is taken as
0.90 when the limit state is section strength and for laterally unbraced
members. the ~-factor is also 0.90.

IV. SUMMARY
The tentative recommendations for load and resistance factor design
of cold-formed steel have been developed. This paper presents a
discussion of the reasoning behind. and the justification for. various
provisions being proposed for designing various types of cold-formed
steel structural members. Additional publications are mentioned in
the discussion for future reference.

V. ACKNOWLEDGMENTS
This project was sponsored by the American Iron and Steel Institu~e
The technical guidance provided by the A~SI Task Group on Lo~d and.Res1s
tance Factor Design under the chairmansh1p of Mr. Karl H. K11ppste1n
and the AISI Staff. Dr. Albert L. Johnson. is gratefully acknowledged.
Thanks are also expressed to, Drs. M. K. Ravindra. T. N. Rang. and B.
Supornsilaphachai for their contributions to the development of load
and resistance factor design criteria.
 TABLE 5

Column Statistics from Ref. 14

t""
Rm 0
Table No. Number Limit State P M F Vp VM VF VR i3 for
m m m >
in Ref. 14 of Tests Rn cp = O.SO t:l
>
Z
13 9 Elastic FB 0.97 1.0 1.0 0.04 0.06 0.05 0.97 0.09 2.S t:l
14 10 Inelastic FB, compact 1.09 1.1 1.0 0.05 0.11 0.05 1.20 0.13 3.5 ~
r:n
.....
15 lS Inealstic FB, stiffened* 0.97 1.1 1.0 0.16 0.10 0.05 1.07 0.20 2.5 r:n
>-3
16 31 Inelastic FB, unstiffened 1.53 1.1 1.0 0.24 0.10 0.05 1.6S 0.26 3.5 ~
0
17 12 Inelastic FB, stiffened* 1.10 1.1 1.0 O.OS 0.11 0.05 1.21 0.14 3.4 t:I:j
"".:l
20 S Elastic TFB loll 1.0 1.0 0.11 0.06 0.05 loll 0.13 3.1 >
0
>-3
21 30 Inelastic TFB 1.20 1.1 1.0 0.14 0.10 0.05 1.32 O.lS 3.4 0
;:d
t:l
t:I:j
r:n
.....
* These two data sets differ in that the predictions for the tests from Table 17 include the effect of 0
cold forming on the average yield stress. Z

en
c,;,
"'I
538 SEVENTH SPECIALTY CONFERENCE

APPENDIX I - REFERENCES

1. American Iron and Steel Insitute, "Specification for the Design

of Cold-Formed Steel Structural Members", 1980 Edition.

2. Canadian Standards Association, "Cold Formed Steel Structural

Members", CSA Standard S136-l974.

3. American Institute of Steel Construction, "Proposed Load &

Resistance Factor Design Specification for Structural Steel
Buildings", September 1, 1983.

4. European Convention for Constructional Steelwork, "European

Recommendations for the Design of Profiled Sheeting", ECCS-
TC7-l983, CONSTRADO, London, April 1983.

5. Galambos, T. V., and Yu, W. W., "Revised Tentative Recommendations

-Load and Resistance Factor Design Criteria for Cold-Formed Steel
Structural Members With Commentary", Seventh Progress Report,
University of Missouri-Rolla, September 1983.

6. Winter, G., "Commentary on the 1968 Edition of the Specification

for the Design of Cold-Formed Steel Structural Members", published
by American Iron and Steel Institute, 1970.

7. American Iron and Steel Institute, "Commentary on the September

3, 1980 Edition of the Specification for the Design of Cold-
Formed Steel Structural Members", Cold-Formed Steel Design Manual
Part II, 1983.

8. Ravindra, M. K., and Galambos, T. V., "Load and Resistance Factor

Design for Steel", Journal of the Structural Division, ASCE,
Vol. 104, No. ST9, September 1978.

9. Ellingwood, B., Galambos, T. V., MacGregor, J.G., and Cornell,

C. A., "Development of a Probability Based Load Criterion for
American National Standard A58: Building Code Requirements for
Minimum Design Loads in Buildings and Other Structures", NBS
Special Publication 577, June 1980.

10. Galambos, T. V., Ellingwood, B., MacGregor, J. G., and Cornell,

C. A., "Probability Based Load Criteria: Assessment of Current
Design Practice", Journal of the Structural Division, ASCE, Vol.
108, No. STS, May 1982.
11. Ellingwood, B., MacGregor, J. G., Galambos, T. V., and Cornell,
C. A., "Probability Based Load Criteria: Load Factors and Load
Combinations", Journal of the Structural Divisibn, ASCE, Vol.
108, No. ST5, May 1982.
 LOAD AND RESISTANCE FACTOR DESIGN 539

12. Rang, T. N., Galambos, T. V., and Yu, W. W., "Study of Design
Formats and Safety Index Combined with Calibration of the AISI
Formulas for Cold Work and Effective Design Width", First Pro-
gress Report, University of Missouri-Rolla, January 1979.

13. Rang, T. N., Galambos, T. V., and Yu, W. W., "Statistical

Analysis of Mechanical Properties and Thickness of Materials
Combined with Calibrations of the AISI Design Provisions on
Unstiffened Elements and Connections", Second Progress Report,
University of Missouri-Rolla, January 1979.

14. Rang, T. N., Galambos, T. V., and Yu, W. W., "Calibration of the
Design Provisions on Connections and Axially Loaded Compression
Members", Third Progress Report, University of Missouri-Rolla,
January 1979.

15. Rang, T. N., Galambos, T. V., and Yu, W. W., "Calibration of the
Design Provisions on Laterally Unbraced Beams and Beam-Columns",
Fourth Progress Report, University of Missouri-Rolla, January
1979.

16. Supornsilaphachai, B., Galambos, T. V., and Yu, W. W., "Calibration

of the Design Provisions on Beam Webs", Fifth Progress Report,
University of Missouri-Rolla, September 1979.

17. American National Standards Institute, "Building Code Require-

ments for Minimum Design Loads in Buildings and Other Structures",
ANSI A58.l-l980.

18. Kalyaranaman, V., Pekoz, T., and Winter, G., "Unstiffened Com-
pression Elements", Journal of the Structural Division, ASCE,
Vol. 103, No. ST9, September 1977.

19. Supornsilaphachai, B., "Load and Resistance Factor Design of Cold-

Formed Steel Structural Members", Ph.D. Thesis, University of
Missouri-Rolla, 1980.
540 SEVENTH SPECIALTY CONFERENCE

APPENDIX II - NOTATION
The following symbols are used in this paper:

cross-sectional area
net area
area of beam web
deterministic influence coefficient translating load
intensities to load effect
warping constant
mean dead load
nominal dead load
= modulus of elasticity
= nominal earthquake load
= buckling stress
= mean ratio of the actual section modulus to the nominal value
specified minimum tensile strength
specific yield point
average yield point
shear modulus
clear distance between flanges measured along the plane of the web
= moment of inertia about y-axis
torsional constant
= shear buckling coefficient
= unbraced length
= mean live load
nominal live load
span length
mean ratio of the yield point to the minimum specified value
nominal ultimate bending moment
= mean ratio of the experimentally determined ultimate moment to the
predicted ultimate moment of test specimens
Pn = nominal ponding load
Pu = nominal ultimate concentrated load
Q load effect
Qm mean load effect
R resistance
Rm mean value of resistance
Rn nominal resistance
Seff section modulus based on the effective cross section
Sn nominal snow load
s beam spacing
t thickness
V coefficient of variation
Vu shear strength
Wn nominal wind load
w flat width of compression element
S reliability index
80 target reliability index
 LOAD AND RESISTANCE FACTOR DESIGN 541

y = load factor
~ = resistance factor
~ = Poisson's ratio
o = standard deviation
Ty = yield point in shear