You are on page 1of 100

STRUCTURAL BEHAVIOR OF COMPRESSION ELEMENTS AND EFFECTIVE WIDTH DESIGN CRITERIA 93

e = Flat width between the first line of connector


and the edge stiffener. See Fig. 3.52(c).
D = Overall length of stiffener as defined in
Specification Section 1.3 (or Section 3.3.3.1 of
this book)

For the edge stiffener, ds and Ia shall be determined using w′ Figure 3.53 Section with single intermediate stiffener.
and f ′ in lieu of w and f, respectively.

𝑤′ = 2𝑒 + minimum of (0.75𝑠 and 𝑤1 ) (3.83)


compression flange of profiled steel decks were discussed

𝑓 = Maximum of (𝜌𝑚 𝑓 and 𝐹𝑐 ) (3.84) by the researchers.
As far as the design provisions are concerned, the 1980
where f ′ = Stress used in Specification Section 1.3(a) and earlier editions of the AISI Specification included the
for determining effective width of edge requirements for the minimum moment of inertia of the
stiffener
intermediate stiffener for multiple-stiffened compression
Fc = Buckling stress of cover plate determined
in accordance with Eq. 3.80
elements. When the size of the actual intermediate stiffener
w′ = Equivalent flat width for determining the did not satisfy the required minimum moment of inertia,
effective width of edge stiffener the load-carrying capacity of the member had to be deter-
w1 = Transverse spacing between the first and mined either on the basis of a flat element disregarding the
the second line of connectors in the intermediate stiffener or through tests. For some cases, this
compression element. See Fig. 3.52(c). approach could be unduly conservative.3.17
The AISI design provisions were revised in 1986 on the
The provisions of this section shall not apply to single flute
basis of the research findings reported in Refs. 3.75–3.77.
members having compression plates with edge stiffeners.
In that method, the buckling coefficient k for determining
the effective width of subelements and the reduced area
(b) Serviceability Determination. The effective width of the
of the stiffener was calculated by using the ratio 𝐼s ∕𝐼a ,
uniformly compressed element restrained by intermittent
where 𝐼s is the actual stiffener moment of inertia and 𝐼a is
connections used for computing deflection shall be deter-
mined in accordance with Section (a) except that: the adequate moment of inertia of the stiffener determined
(1) 𝑓𝑑 shall be substituted for f, where 𝑓𝑑 is the computed from the applicable equations. The same design requirements
compression stress in the element being considered at were retained in the 1996 edition of the AISI Specifica-
service load, and tion. Because a discontinuity could occur in those equations,
(2) The maximum extreme fiber stress in the built-up the design provisions were revised in the 2001 edition of
member shall be substituted for 𝐹𝑦 . the North American Specification by adopting Dinovitzer’s
expressions to eliminate the discontinuity.3.221
In the 2007 edition of the North American Specifica-
3.3.3.3 Uniformly Compressed Elements with Interme- tion, the design of uniformly compressed stiffened elements
diate Stiffeners with a single intermediate stiffener was merged with the
3.3.3.3.1 Uniformly Compressed Elements with Single stiffened elements having multiple intermediate stiffeners.
Intermediate Stiffener In the design of cold-formed steel Section 3.3.3.3.2 provides the AISI design requirements for
beams, when the width-to-thickness ratio of the stiffened this particular case by using the number of stiffeners equal to
compression flange is relatively large, the structural effi- unity (i.e., 𝑛 = 1) in Eqs. (3.92) and (3.93). See Example 4.7
ciency of the section can be improved by adding an inter- for the application of these equations.
mediate stiffener as shown in Fig. 3.53.
The buckling behavior of rectangular plates with central 3.3.3.3.2 Uniformly Compressed Elements with Mul-
stiffeners is discussed in Ref. 3.7. The load-carrying tiple Intermediate Stiffeners In beam sections, the normal
capacity of an element with a longitudinal intermediate stresses in the flanges result from shear stresses between
stiffener has been studied by Höglund,3.72 König,3.73 König the web and flange. The web generates the normal stresses
and Thomasson,3.74 Desmond, Pekoz, and Winter,3.75–3.77 by means of the shear stress which transfers to the flange.
Pekoz,3.17 and Yang and Schafer.3.223 The more remote portions of the flange obtain their normal
In the study of Bernard, Bridge, and Hancock,3.171,3.172 stress through shear from those close to the web. For this
both local buckling and distortional buckling in the reason there is a difference between webs and intermediate
94 3 STRENGTH OF THIN ELEMENTS AND DESIGN CRITERIA

stiffeners. The latter is not a shear-resisting element and


does not generate normal stresses through shear. Any normal
stress in the intermediate stiffener must be transferred to it
from the web or webs through the flange portions. As long
as the subelement between web and stiffener is flat or is only
very slightly buckled, this stress transfer proceeds in an unaf-
fected manner. In this case the stress in the stiffener equals
that at the web, and the subelement is as effective as a regular
single-stiffened element with the same w/t ratio. However,
for subelements having larger w/t ratios, the slight waves of
the subelement interfere with complete shear transfer and
create a “shear lag” problem which results in a stress distri-
bution as shown in Fig. 3.54.
In the 1996 edition of the AISI Specification, the design
requirements for uniformly compressed elements with
multiple intermediate stiffeners and edge-stiffened elements Figure 3.55 Buckling modes of multiple-stiffened elements with
with intermediate stiffeners included (a) the minimum longitudinal intermediate stiffeners3.176 : (a) distortional buckling
moment of inertia of the full stiffener about its own mode; (b) local buckling mode.
centroidal axis parallel to the element to be stiffened,
(b) the number of stiffeners considered to be effective, (c) and experimental results. It has been found that the 1996
the “equivalent element” of the entire multiple-stiffened AISI design requirements were nearly 20% unconservative
element for closely spaced stiffeners with an “equivalent for the 94 members studied.3.175,3.176 Based on the exper-
thickness,” (d) the reduced effective width of subelement imental and numerical studies, a method for calculating
having 𝑤∕𝑡 > 60, and (e) the reduced effective stiffener the ultimate strength of stiffened elements with multiple
area when the w/t ratio of the subelement exceeds 60. The intermediate stiffeners was proposed by Schafer and Pekoz
reasons for using the above requirements are discussed by in Ref. 3.176. This method involves the calculation of the
Yu in Ref. 1.354. critical local buckling stress for the subelement and the
In the past, the structural behavior and strength of distortional buckling stress for the entire multiple-stiffened
cold-formed steel members with multiple longitudinal element. Because the experimental and numerical data
intermediate stiffeners have been investigated by Papazian, revealed that the overall (distortional) buckling mode
Schuster, and Sommerstein,3.174 Schafer and Pekoz,3.175 3.176 usually dominated the behavior, a modified effective width
Acharya and Schuster,3.177,3.178 Teter and Kolakowski,3.224 equation was proposed for the entire multiple-stiffened
and Schafer.3.225 Some of these studies considered the element by using the proposed plate buckling coefficient to
distortional buckling of the entire stiffened elements as a determine the reduction factor.
unit (Fig. 3.55a) and local buckling of the subelements Consequently, in 2001, the design provisions were revised
between stiffeners (Fig. 3.55b). The AISI Specification and to reflect those additional research findings.1.336,3.176 The
the Canadian Standard have been compared with analytical same requirements are retained in Section 1.4.1 of the 2016
edition of the North American Specification for determining
the effective width of uniformly compressed stiffened
elements with single or multiple intermediate stiffeners as
given below:

1.4.1 Effective Widths of Uniformly Compressed Stiff-


ened Elements with Single or Multiple Intermediate
Stiffeners
The following notation shall apply as used in this section.
Ag = gross area of element including stiffeners
As = gross area of stiffener
Figure 3.54 Stress distribution in compression flange with inter- be = effective width of element, located at centroid of
mediate stiffeners.1.161 element including stiffeners; see Fig. 3.57
STRUCTURAL BEHAVIOR OF COMPRESSION ELEMENTS AND EFFECTIVE WIDTH DESIGN CRITERIA 95

Figure 3.56 Plate widths and stiffener locations.1.347

bo = total flat width of stiffened element; see Fig. 3.56 The plate buckling coefficient, k, shall be determined from
bp = largest subelement flat width; see Fig. 3.56 the minimum of Rkd and 𝑘loc , as determined in accordance with
ci = horizontal distance from edge of element to Specification Section 1.4.1.1 or 1.4.1.2, as applicable:
centerline(s) of stiffener(s); see Fig. 3.56 𝑘 = minimum of Rkd and 𝑘loc (3.89)
Fcr𝓁 = plate elastic buckling stress
F = uniform compressive stress acting on flat element R = 2 when 𝑏o ∕ℎ < 1
h = width of elements adjoining stiffened element (e.g., 11 − bo ∕h 1
depth of web in hat section with multiple 𝑅= ≥ when 𝑏o ∕ℎ ≥ 1 (3.90)
5 2
intermediate stiffeners in compression flange is equal
to h; if adjoining elements have different widths, use 1.4.1.1 Specific Case: Single or n Identical Stiffeners,
smallest one) Equally Spaced
Isp = moment of inertia of stiffener about centerline of flat For uniformly compressed elements with single, or multiple
portion of element; radii that connect the stiffener to identical and equally spaced, stiffeners, the plate buckling coef-
the flat can be included ficients and effective widths shall be calculated as follows:
k = plate buckling coefficient of element
kd = plate buckling coefficient for distortional buckling a. Strength Determination3.240
kloc = plate buckling coefficient for local subelement
𝑘loc = 4(bo ∕bp )2 (3.91)
buckling
Lbr = unsupported length between brace points or other (1 + β2 )2 + γ(1 + n)
restraints which restrict distortional buckling of 𝑘d = (3.92)
β2 [1 + δ(n + 1)]
element
where
R = modification factor for distortional plate buckling
𝛽 = [1 + γ(n + 1)] ∕4
1
(3.93)
coefficient
n = number of stiffeners in element where
t = element thickness 10.921sp
i = Index for stiffener “i” 𝛾= (3.94)
𝑏𝑜 𝑡3
𝜆 = slenderness factor
𝜌 = reduction factor 𝐴𝑠
𝛿= (3.95)
𝑏𝑜 𝑡
The effective width shall be calculated in accordance with If 𝐿br < 𝛽𝑏o , 𝐿br ∕𝑏o is permitted to be substituted for 𝛽
Eq. (3.85) as follows: to account for increased capacity due to bracing.
( )
𝐴𝑔 b. Serviceability Determination. The effective width, 𝑏d , used
𝑏e = ρ (3.85) in determining serviceability shall be calculated as in Spec-
𝑡
ification Section 1.4.1.1(a), except that 𝑓d is substituted
where
{ for f, where 𝑓d is the computed compressive stress in the
1 when 𝜆 ≤ 0.673 element being considered based on the effective section at
𝜌= (3.86)
(1 − 0.22∕𝜆)∕𝜆 when 𝜆 > 0.673 the load for which serviceability is determined.
where √
𝑓 1.4.1.2 General Case: Arbitrary Stiffener Size, Loca-
𝜆= (3.87)
𝐹cr𝓁 tion, and Number
where For uniformly compressed stiffened elements with stiffeners of
( )2
𝜋2𝐸 𝑡 arbitrary size, location, and number, the plate buckling coeffi-
𝐹cr𝓁 =𝑘 (3.88) cients and effective widths shall be calculated as follows:
12(1 − 𝜇 2 ) 𝑏𝑜
96 3 STRENGTH OF THIN ELEMENTS AND DESIGN CRITERIA

a. Strength Determination
𝑘loc = 4(bo ∕bp )2 (3.96)
∑ n
(1 + β2 )2 + 2 γi ωi
i=1
𝑘d = ( ) (3.97)
∑n
β2 1 + 2 δi ωi
i=1

where
( ) 1∕4

n
𝛽= 2 γi ωi + 1 (3.98)
i=1

where Figure 3.57 Effective width locations.1.347


10.92(Isp )i
𝛾i = (3.99)
bo t3
( ) 1.4.2 Edge-Stiffened Elements with Intermediate
c
𝜔i = sin2 π i (3.100) Stiffener(s)
bo
a. Strength Determination. For edge-stiffened elements with
(A ) intermediate stiffener(s), the effective width, 𝑏e , shall be
𝛿i = s i (3.101)
bo t determined as follows:
If 𝐿br < β𝑏o , 𝐿br ∕𝑏o is permitted to be substituted for 𝛽 • If 𝑏o ∕𝑡 ≤ 0.328𝑆, the element is fully effective and no
to account for increased capacity due to bracing. local buckling reduction is required.
b. Serviceability Determination. The effective width, 𝑏d , used • If 𝑏o ∕𝑡 > 0.328S, the plate buckling coefficient, k, is
in determining serviceability shall be calculated as in Spec- determined in accordance with Section 3.3.3.1, but with
ification Section 1.4.1.2(a), except that 𝑓d is substituted 𝑏o replacing w in all expressions.
for f, where 𝑓d is the computed compressive stress in the If k calculated from Section 3.3.3.1 is less than 4.0
element being considered based on the effective section at (𝑘 < 4), the intermediate stiffener(s) is ignored and the
the load for which serviceability is determined. provisions of Section 3.3.3.1 are followed for calcula-
tion of the effective width.
It should be noted that according to Eq. (3.85), the effec- If k calculated from Section 3.3.3.1 is equal to 4.0
tive width of the uniformly compressed stiffened elements with (𝑘 = 4), the effective width of the edge-stiffened
multiple intermediate stiffeners is determined from an overall element is calculated from the provisions of Section
equivalent flat width (𝐴g ∕𝑡), in which 𝐴g is the gross area 3.3.3.3.2, with the following exception: R calculated in
of the stiffened element including intermediate stiffeners. The accordance with Section 3.3.3.3.2 is less than or equal
equation used for computing the reduction factor, 𝜌, is the same to 1, where
as Eq. (3.34) , except that in the calculation of slenderness factor bo = total flat width of edge-stiffened element
𝜆, the plate buckling coefficient, k, is the lesser of Rkd and See Sections 3.3.3.1 and 3.3.3.3.2 for definitions of
𝑘loc , and the width-to-thickness ratio is based on 𝑏o ∕𝑡, in which other variables.
𝑏o is the total overall flat width of the stiffened element. See b. Serviceability Determination. The effective width, 𝑏d , used
Fig. 3.56. As shown in Fig. 3.57, the effective width is placed at in determining serviceability shall be calculated as in (a)
the centroidal line of the entire element including the stiffeners above, except that 𝑓d is substituted for f, where 𝑓d , is the
for the calculation of the effective sectional properties. computed compressive stress in the element being consid-
ered based on the effective section at the load for which
3.3.3.3.3 Edge-Stiffened Elements with Intermediate serviceability is determined.
Stiffeners For the design of edge-stiffened elements with In the above criteria, the modification factor (R) for the distor-
intermediate stiffeners, if the overall flat width-to-thickness tional plate buckling coefficient is limited to less than or equal to
√ 1.0 due to the fact that the edge-stiffened element does not have
ratio (𝑏o ∕𝑡) is small (i.e., 𝑏o ∕𝑡 ≤ (0.328𝑆 = 0.42 𝐸∕𝑓 ),
the same web rotational restraint along the side-supported edge
the flat subelements and intermediate stiffeners can be fully
stiffener.
effective. However, if the 𝑏o ∕𝑡 ratio is large, three buckling For the calculation of effective sectional properties, the effec-
modes are possible, as shown in Fig. 3.58.1.346,3.226 tive width (𝑏e ) of the edge-stiffened element with intermediate
In order to provide new requirements for computing the stiffeners is placed at the centroidal line as shown in Fig. 3.57.
effective width of edge-stiffened elements with intermediate The centroidal line is located on the basis of the gross areas of
stiffeners, Section 1.4.2 of the North American Specification subelements and intermediate stiffeners without using the edge
includes the following design provisions1.417 : stiffener.
PERFORATED ELEMENTS AND MEMBERS 97

Figure 3.58 Buckling modes in an edge-stiffened element with intermediate stiffeners.1.417

The adequacy of this approach was demonstrated by the


stub compression tests performed by Yang and Hancock in
2003.3.219

3.4 PERFORATED ELEMENTS AND MEMBERS


In cold-formed steel structural members, holes are some-
times provided in webs and/or flanges of beams and columns
for duct work, piping, bracing, and other construction
purposes. For steel storage racks (Fig. 1.10), various types
of holes are often used for the purpose of easy assembly.
The presence of such holes may result in a reduction of
the strength of individual component elements and of the
overall strength of the member depending on the size, shape, Figure 3.59 Effect of circular hole on buckling coefficient in
and arrangement of holes, the geometric configuration of compression.3.99
the cross section, and the mechanical properties of the
material used. circular holes on the buckling coefficients in compression
The exact analysis and the design of steel sections having is shown in Fig. 3.59. Figure 3.60 shows the effect of a
perforated elements are complex, in particular when the central square hole on the buckling coefficient for a simply
shapes and the arrangement of the holes are unusual. Even supported square plate, in which the top curve was computed
though limited information is available for relatively thick by the finite-element method developed by Yang.3.112 The
steel sections,1.148,1.165,3.84–3.86 on the basis of previous test data obtained from the testing of beams and columns are
investigations,3.87–3.90 these design criteria may not be also shown in these two figures.3.99
completely applicable to perforated cold-formed steel In Figs. 3.59 and 3.60, k is the buckling coefficient for
sections due to the fact that local buckling is usually a major square plates without holes, 𝑘c is the buckling coefficient
concern for thin-walled structural members. for perforated square plates having a circular hole, 𝑘s is the
For perforated cold-formed steel structural members the buckling coefficient for perforated square plates having a
load-carrying capacity of the member is usually governed square hole, d is the diameter of circular holes, h is the width
by the buckling behavior and the postbuckling strength of of square holes, and w is the width of the plate.
the component elements. The critical buckling loads for The postbuckling strength of perforated compression
perforated plates and members have been previously studied elements has also been studied by Davis and Yu in Ref.
by numerous investigators.3.91–3.111,3.227–3.233 The effect of 3.99. It was found that Winter’s effective width equation for
98 3 STRENGTH OF THIN ELEMENTS AND DESIGN CRITERIA

3.4.1 Uniformly Compressed Stiffened Elements


with Circular Holes
Based on the Cornell study presented in Ref. 3.100, limited
design provisions have been included in the AISI Speci-
fication since 1986. Section 1.1.1 of the North American
Specification1.417 includes the following provisions for deter-
mining the effective width of uniformly compressed stiffened
elements with circular holes (Fig. 3.62):

a. Strength Determination. The effective width, b, shall be


calculated by either Eq. (3.102) or Eq. (3.103) as follows:
For 0.50 ≥ 𝑑h ∕𝑤 ≥ 0, 𝑤∕𝑡 ≤ 70, and the distance between
centers of holes ≥ 0.50𝑤 and ≥ 3𝑑h ,
Figure 3.60 Effect of square hole on buckling coefficient in ⎧𝑤 − 𝑑 when 𝜆 ≤ 0.673 (3.102)
compression.3.99 ⎪ h
⎪ 𝑤[1 − 0.22∕𝜆 − 0.8𝑑h ∕𝑤 + 0.085𝑑h ∕(𝜆w)]
𝑏=⎨
⎪ 𝜆
⎪ when 𝜆 > 0.673 (3.103)
a solid plate [Eq. (3.26)] can be modified for the determina- ⎩
tion of the effective width of perforated stiffened elements. In all cases,
Even though the buckling load for the perforated stiffened 𝑏 ≤ 𝑤 − 𝑑ℎ
element is affected more by square holes than by circular
holes, the postbuckling strength of the elements with square where w = flat width
and circular holes was found to be nearly the same if the t = thickness of element
dh = diameter of holes
diameter of a circular hole was the same as the width of a
𝜆 = as defined in Section 3.3.1.1 with k = 4.0
square hole.
The effect of perforations on the design of industrial steel
storage racks has been accounted for by using net section b. Serviceability Determination. The effective width, 𝑏d , used
in determining serviceability shall be equal to b calculated
properties determined by stub column tests.1.165
in accordance with Eqs. (3.32)–(3.35) except that 𝑓d is
Considering the effect of holes on the shear buckling substituted for f, where 𝑓d is the computed compressive
of a square plate, the reduction of the buckling coeffi- stress in the element being considered.
cients has been studied by Kroll,3.113 Rockey, Anderson,
and Cheung,3.114,3.115 and Narayanan and Avanessian.3.101
Figure 3.61 shows the buckling coefficients in shear affected 3.4.2 Uniformly Compressed Stiffened Elements
by holes. with Noncircular Holes
For uniformly compressed stiffened elements with noncir-
cular holes such as the perforated web element of steel studs

Figure 3.61 Effect of circular hole on buckling coefficient in Figure 3.62 Uniformly compressed stiffened elements with
shear.3.114 circular holes.
PERFORATED ELEMENTS AND MEMBERS 99

Figure 3.63 Uniformly compressed stiffened elements with noncircular holes.1.417

shown in Fig. 3.63, the effective width of the perforated b. Serviceability Determination. The effective width, 𝑏d , used
web can be determined by assuming the web to consist of in determining serviceability shall be calculated in accor-
two uniformly compressed unstiffened elements with the flat dance with Eqs. (3.32)–(3.35).
width one on each side of the hole. The effective design width It should be noted that the effective area should be based
of these unsiffened compression elements can be calculated on the lesser of the total effective design width of two unstiff-
in accordance with Section 3.3.2.1 or the effective area of the ened elements and the effective design width determined for
perforated web can be determined from stub-column tests. the stiffened element with the flat width, w. The calculation
The unstiffened strip approach was studied by Miller and of the effective area for the steel stud having noncircular web
Pekoz at Cornell University in the 1990s.3.186 Test results perforations is illustrated in Example III-2 of the 2017 edition
indicated that this method is generally conservative for the of the AISI Design Manual.1.428
wall studs tested in the Cornell program. This approach has
long been used in the Rack Manufacturers Institute (RMI) 3.4.3 C-Section Webs with Holes under Stress
Specification for the design of perforated rack columns.1.156 Gradient
Since 1996, similar requirements were used in the AISI Spec-
ification for the design of wall studs under specific limita- In the past, numerous studies have been conducted to inves-
tions. The same requirements were moved from the previous tigate the structural behavior and strength of perforated
Section D4 to Section B2.2 of the 2007 edition of the Specifi- elements and members subjected to tension, compression,
cation , and was retained in Section 1.1.1 of the 2016 edition bending, shear, and web crippling.3.179–3.193,3.197–3.200 Based
of the Specification: on the research work conducted by Shan et al. at the Univer-
sity of Missouri–Rolla,3.184,3.197 the following requirements
have been included in Section 1.1.3 of the Specification for
a. Strength Determination. A uniformly compressed stiff- determining the effective depth of C-section webs with holes
ened element with noncircular holes shall be assumed to under stress gradient1.417 :
consist of two unstiffened strips of flat width, c, adjacent
to the holes (see Fig. 3.63). The effective width, b, of each a. Strength Determination. When 𝑑h ∕ℎ < 0.38, the effective
unstiffened strip adjacent to the hole shall be determined widths, 𝑏1 and 𝑏2 , shall be determined by Section 3.3.1.2
in accordance with Eqs. (3.32)–(3.35), except that the plate by assuming no hole exists in the web. When 𝑑h ∕ℎ > 0.38,
buckling coefficient, k, shall be taken as 0.43 and w as c. the effective width shall be determined by Section 3.3.2.1,
These provisions shall be applicable within the following assuming the compression portion of the web consists of
limits: an unstiffened element adjacent to the hole with 𝑓 = 𝑓1 ,
1. Center-to-center hole spacing, 𝑠 ≥ 24 in. (610 mm), as shown in Fig. 3.64.
2. Clear distance from the hole at ends, 𝑠end ≥ 10 in. b. Serviceability Determination. The effective widths shall be
(254 mm), determined by Section 3.3.1.2 by assuming no hole exists
3. Depth of hole, 𝑑h ≤ 2.5 in. (63.5 mm), in the web.
4. Length of hole, 𝐿h ≤ 4.5 in. (114 mm), and
Because the above requirements are based on the experi-
5. Ratio of the depth of hole, 𝑑h , to the out-to-out width,
mental study, these provisions are applicable only within the
𝑤𝑜 , 𝑑h ∕𝑤o ≤ 0.5.
following limits:
Alternatively, the effective width, b, is permitted to
be determined by stub-column tests in accordance with 1. 𝑑h ∕ℎ < 0.7
the test procedure, AISI S902. 2. ℎ∕𝑡 ≤ 200
100 3 STRENGTH OF THIN ELEMENTS AND DESIGN CRITERIA

Figure 3.64 C-section webs with holes under stress gradient.

Figure 3.65 Virtual hole method for multiple openings.1.431

3. Holes centered at mid-depth of the web 4.5 in. (114 mm) but still fits within an allowable circular
4. Clear distance between holes ≥ 18 in. (457 mm) virtual hole. For each case, the provisions apply to the
5. Noncircular holes, corner radii ≥ 2t geometry of the virtual hole, not the actual hole or
6. Noncircular holes, 𝑑h ≤ 2.5 in. (64 mm) and Lh ≤ holes.1.333
4.5 in. (114 mm)
For the effect of web holes on the shear strength and web
7. Circular hole diameter ≤ 6 in. (152 mm)
crippling strength of C-sections, see Section 4.3 on the design
8. 𝑑h ≥ 9∕16 in. (14 mm), where
of beam webs. Extensive studies of perforated elements
d = depth of web hole and members have been conducted by numerous investi-
h = depth of flat portion of web measured along gators. See Refs. 3.228–3.232, 3.242–3.248, 3.250–3.253.
the plane of the web Design provisions using the Direct Strength Method have
t = thickness of web been developed and will be discussed in Chapters 4 and 5.
Lh = length of web hole
b1 , b2 = effective widths defined by Fig. 3.30
3.5 DIRECT STRENGTH METHOD AND
CONSIDERATION OF LOCAL AND
Although these provisions are based on the tests of
DISTORTIONAL BUCKLING
C-sections having the web hole centered at mid-depth of
the section, the provisions may be conservatively applied to The Direct Strength Method provides a consistent design
sections for which the full unreduced compression region of procedure for determining cold-formed steel member
the web is less than the tension region. Otherwise, the web strengths under different buckling failure modes. This
strength must be determined by tests.1.333 method was developed by Schafer and Peoz3.254,3.255
The design provisions apply to any hole pattern that fits in 1990s and was continued developing by other
within equivalent virtual holes, as shown in Figs. 3.65 and researchers.3.281–3.293 In 2004, this method was adopted
3.66. Figure 3.65 shows the dimensions 𝐿h and 𝑑h for a into the North American Specification as Appendix 1.1.343
multiple-hole pattern that fits within a noncircular virtual In 2016, this method was incorporated into the main body
hole, while Fig. 3.66 illustrates the dimension 𝑑h for a of the Specification1.417 and is considered as an equivalent
rectangular hole that exceeds the limits of 2.5 in. (64 mm) × design method to the Effective Width Method.
DIRECT STRENGTH METHOD AND CONSIDERATION OF LOCAL AND DISTORTIONAL BUCKLING 101

Figure 3.66 Virtual hole method for opening exceeding limit.1.431

The Direct Strength Method is based on the same assump- members outside the limitations, the Specification permits
tion as the Effective Width Method3.255 : the member strength to use the rational engineering analysis to determine the
is the function of elastic buckling and the yielding of the member strengths and apply the safety and resistance
material. Therefore a good estimate of the elastic buckling factors provided in Specification Section A1.2(c)1.417 :
will result in a better prediction of member strength. To Ω = 2.00 (ASD) and 𝜙 = 0.80 (LRFD) and 0.75 (LSD).
realistically predict the member buckling strengths, the If test data available, the Specification also permits to use
Direct Strength Method analyzes the buckling of the whole the safety and resistance factors provided in the relevant
cross-section instead of individual elements, which ensures sections in Specification Chapters E through H provided
that the compatibility and equilibrium are maintained at the those tests which are performed per Specification Section
element junctures. To capture the postbuckling behaviors, the K21.417 , and the calculated resistance factor, 𝜙, is greater
method calibrated the strength expressions with numerous than that in Chapters E through F. Detailed provisions are
test data.3.254,3.255 Figures 3.67a and 3.67b show that the provided in Section B4.2 of the Specification.1.417
strength prediction expressions have a good agreement Figure 3.68a and 3.68b plotted the compression and
with the test data. Through these test data, the geometric bending strengths of a member that is laterally braced
limitations are established as provided in Table 3.1. For against global (lateral torsional) buckling. The curves

1.5
Local: Eq. (5.65)

Distortional: Eq. (5.70)

Local
1
Distortional

0.5

0
0 1 2 3 4 5 6 7 8
λ = λ =

Figure 3.67a Comparison between the test data and the nominal axial strengths calculated by DSM
for concentrically loaded pin-ended columns.1.431
102 3 STRENGTH OF THIN ELEMENTS AND DESIGN CRITERIA

1.5
Local: Eq. (4.92)

Distortional: Eq. (4.102)

Local
1
Distortional

0.5

0
0 1 2 3 4 5
λ = λ =

Figure 3.67b Comparison between the test data and the nominal flexural strengths calculated by
DSM for laterally braced beams.1.431

Local: Eq. (5.65)

Distortional: Eq. (5.70)

Figure 3.68a Local and distortional direct strength curves for a braced column.1.431

show the local and distortional post-buckling strengths as considered even if the member is braced against the global
compared to the elastic buckling, where the local buckling buckling.
strength possesses a higher post-buckling strength than the The application of the Direct Strength Method to deter-
distortional buckling. The figures also indicate that similar mine the member strengths due to different buckling modes
to the local buckling, the distortional buckling should be will be discussed in details in Chapters 4 and 5. This section
DIRECT STRENGTH METHOD AND CONSIDERATION OF LOCAL AND DISTORTIONAL BUCKLING 103

Inelastic Bending Reserve


Considered in Specification Sections
F2.4.2 and F3.2.3

Inelastic Bending
Reserve Ignored
in Sections
F2.1 and F3.2.1

Figure 3.68b Local and distortional direct strength curves for a laterally braced beam1.431 where
the referenced section numbers are those in the Specification.1.417

is focused on the determination of the elastic local and Figs. 3.69 and 3.70, which can be used to identify the
distortional buckling, which will be used in the Direct buckling modes and determine the buckling loads to be
Strength Method in later chapters. The global buckling will used in design. Open software that determines distor-
be discussed in Chapters 4 and 5. tional buckling of C- and Z-Section members3.261 can
be downloaded from (www.civil.ist.utl.pt/gbt/). The
3.5.1 Local Buckling finite strip method for cold-formed steel was pioneered
at the University of Sydney and the long-used program
The element local buckling has been fully discussed in (THIN-WALLED) is commercially available at (www
Sections 3.1 to 3.4. Expressions that are used for the Effective .civil.usyd.edu.au/case/thinwall.php).3.260 Through pa-
Width Method are provided. rtial research support of AISI, an open-source and
To consider the local buckling of the whole cross-section, free finite-strip method program (CUFSM) was
the following numerical and analytical solutions may be developed by Schafer et al.3.267,3.268 The software
employed.3.256,1.417,1.431 can be downloaded from (www.ce.jhu.edu/bchafer/
cufsm).3.259 The finite strip method can be used to
1. Numerical Solutions. Numerical methods such as determine the buckling loads and moments of pris-
the shell finite element method,3.262–3.264 the finite matic members with arbitrary cross-section. The
strip method,3.257–3.260 and generalized beam theory method has also been extended to determine the shear
(GBT)3.265,3.266,3.294 can be used to determine the buckling,3.269–3.271,5.109,3.272 generalized end boundary
member local, distortional, and global buckling. Even conditions,1.432 members with holes,3.273,3.274,3.275 and
though the shell finite element method provides a so forth.
flexible way to model members with different shapes Shown in Figs. 3.69 and 3.70 are the buckling
or support conditions, the method generally requires analysis signature curves for C-Section (9CS2.5×059)
the user to visually determine buckling modes and the obtained from CUFSM. The figures show that the
buckling modes are often coupled. The general beam C-Section member subjects a local buckling at short
theory was originally developed by Schardt3.265 and buckling wavelength. The local minimum buckling
extended by Davies et al.3.266 The method is capable of wavelength is at or near the outer dimensions of
generating the buckling signature curve, as shown in the member cross-section, the distortional buckling
104 3 STRENGTH OF THIN ELEMENTS AND DESIGN CRITERIA

Figure 3.69 Compression elastic buckling analysis of C-section (9CS2.5 × 059) with finite-strip
method.1.431

typically occurs between three and nine times the out extreme compression fiber, will be used to determine
dimensions, and the global buckling occurs at much the local buckling moment. Since the restraints at the
longer wavelength.1.383 The global buckling load or element junctures are not accurately modeled, this
moment can be selected from the signature curve based analytical approach could be very conservative.
on actual unbraced length.1.383
2. Analytical Solutions. The local buckling of an element 3.5.2 Distortional Buckling
can be determined by Eq. (3.45)
Since 1962, the distortional buckling problem of cold-
𝑘𝜋 2 𝐸
fcr𝓁 = (3.45) formed steel members has been studied by Douty,4.19
12(1 − 𝜇2 )(𝑤∕𝑡)2 Haussler,4.20 Desmond, Pekoz and Winter,3.76,3.77
where k is the buckling coefficient which can be deter- Hancock,1.69,4.163,4.164,1.358,4.223 Lim and Rhodes,4.293 Kwon
mined from Table 3.2 for typical boundary conditions, and Hancock,4.196 Hancock, Rogers, and Schuster,4.165
w is the flat width of the element, and t is the thickness Lau and Hancock,5.109–5.111 Serrette and Pekoz,4.158–4.162
of the element. For an interconnected element in a Buhagiar, Chapman, and Dowling,4.166 Davies and
cross-section, it is difficult to determine actual fixities. Jiang,4.167,4.1.68,4.197 Schafer and Pekoz,3.168,3.175,3.176,3.195
Therefore, the above equation only can provide an esti- Bambach, Merrick, and Hancock,3.173 Bernard,
mated critical buckling stress. In addition, depending Bridge, and Hancock,3.171,3.172 Ellifritt, Sputo, and
on the dimensions and the fixities, each element on Haynes,4.186 Kavanagh and Ellifritt,4.188 Ellifritt,
a cross-section may predict different critical buck- Glover, and Hren,4.169 Jonson,4.198,4.199 Bradford,4.200
ling stresses. The North American Specification1.417 Sarawit and Pekoz,4.201 Camotim, Silvestre, and
requires: for a compression member, the minimum Dinis,3.286,3.294,4.202,4.203,4.207,4.208,4.214,4.225 Nuttayasakul
𝑓cr𝓁 among all the elements on the cross-section be and Esterling,4.204 Cortese and Murray,4.205 Yu,4.209 Yu
used to determine the member local buckling force; and Schafer,4.206,4.210,4.217 Chodraui, Malite, Goncalves,
and for flexural member, the 𝑓cr𝓁 , which results in the and Neto,4.211,4.213 Schafer, Sarawit, and Pekoz,4.212
smallest stress level when linearly extrapolated to the Schafer and Adany,4.215 Yap and Hancock,4.216 Yu and
DIRECT STRENGTH METHOD AND CONSIDERATION OF LOCAL AND DISTORTIONAL BUCKLING 105

Figure 3.70 Bending elastic buckling analysis of C-section (9CS2.5 × 059) with finite-strip
method.1.431

Lokie,4.218 Javaroni and Goncalves,4.219 Mahaarachchi and Similar to the determination of local buckling, distortional
Mahendran,4.220 Georgescu,4.221 Pham and Hancock,4.222 buckling can be analyzed numerically and analytically.
Schafer, Sangree, and Guan,4.223 Yap and Hancock,4.222
Bambach,4.227 and others. Some of the past research find- 1. Numerical Solutions. Same as discussed in the numer-
ings and the development of the AISI design criteria for ical solutions for local buckling, all the methods
distortional buckling strength of cold-formed steel members mentioned for local buckling analysis in Section
are well summarized in the AISI commentary1.346,1.431 and 3.5.1(1) can be used for determining the distortional
direct-strength method design guide.1.383 Section 13.4.1 of buckling force and moments. See Section 3.5.1(1) for
the SSRC guide1.412 presents detailed discussions of the details.
available research work on distortional buckling. 2. Analytical Solutions. The following analytical expres-
According to Section 13.2.3 of the SSRC Guide, Lau and sions for C- or Z-Section members with simple or
Hancock’s analytical model5.109 is in widest use and is based complex stiffeners, are derived by Schafer3.280 and
primarily on the assumption that the flange acts as an isolated verified for complex stiffeners by Schafer et al.3.222
column undergoing flexural–torsional buckling, while the The following distortional buckling force and moment
web provides elastic restraint to the flange. This model was expressions are excerpted from the 2016 edition of the
subsequently improved to include more consistent treatment North American Specification.1.417
of the web. Their model is used in the Standards of Australia
and New Zealand.1.391 In 1999, Schafer and Pekoz further 2.3.1.3 Distortional Buckling (Fcrd , Pcrd )
developed the model to allow for the impact of applied The provisions of this section shall apply to any open
stresses on the web’s rotational stiffness, thus allowing for cross-section with stiffened flanges of equal dimension where
the case when distortional buckling is triggered by instability the stiffener is either a simple lip or a complex edge stiffener.
of the web as opposed to the flange.3.168 Schafer and Pekoz’s The elastic distortional buckling load, Pcrd , shall be calculated
model is used in the North American Specification,1.345,1.417 as follows:
and is enclosed item (2) Analytical Solutions, below. 𝑃crd = 𝐴g 𝐹crd (3.104)
106 3 STRENGTH OF THIN ELEMENTS AND DESIGN CRITERIA

where 𝐴g = Gross cross-sectional area Lm = Distance between discrete restraints that restrict distor-
k𝜙fe + k𝜙we + k𝜙 tional buckling (for continuously restrained members Lm = Lcrd )
𝐹crd = (3.105) Variables 𝐴f , 𝐽f , 𝐼xf , 𝐼yf , 𝐼xyf , 𝐶wf , 𝑥of , 𝑦of , and ℎxf are defined
̃
k𝜙fg + ̃
k𝜙wg in Table 3.4, and variables 𝐿x , 𝐿y are unbraced length for
where bending about the x and y axis, respectively; 𝐿t is the unbraced
length for torsion; E, G, 𝜇 are modulus of elasticity, shear
𝑘𝜙fe = Elastic rotational stiffness provided by the mdulys and Poisson’s ratio, respectively; and 𝐴g is the gross area
flange to the flange∕web juncture of the cross-section.
[ ]
( )4 I2xyf
π
= EIxf (xof − hxf ) + ECwf − E
2
(x − hxf ) 2
2.3.3.3 Distortional Buckling (Fcrd , Mcrd )
L Iyf of
The provisions of this section are permitted to apply to any
( )2
π open cross-section with a single web and single edge-stiffened
+ GJf (3.106) compression flange extending to one side of the web where the
L
stiffener is either a simple lip or a complex edge stiffener. The
𝑘𝜙we = Elastic rotational stiffness provided by the web to
elastic distortional buckling moment, 𝑀crd , shall be calculated
flange∕web juncture as follows:
Et3 𝑀crd = 𝑆𝑓 𝐹crd (3.111)
= (3.107)
6ho (1 − μ2 )
where
where 𝑘𝜙fe + 𝑘𝜙we + 𝑘𝜙
𝐹crd = 𝛽 (3.112)
ho = Out-to-out web depth (See Fig.3.30(c)) 𝑘𝜙fg + ̃
̃ 𝑘𝜙wg
t = Base steel thickness
k𝜙 = Rotational stiffness provided by restraining where 𝛽 = A value accounting for moment gradient, which is
elements (brace, panel, sheathing) to permitted to be conservatively taken as 1.0
flange/web juncture of member (zero if the
flange is unrestrained).If rotational stiffness = 1.0 ≤ 1 + 0.4(L∕Lm )0.7 (1 + M1 ∕M2 )0.7 ≤ 1.3
provided to the two flanges is dissimilar, the (3.113)
smaller rotational stiffness is used. where
̃
𝑘𝜙fg = Geometric rotational stiffness demanded by 𝐿 = Minimum of 𝐿crd and 𝐿m
flange from flange/web juncture
where
⎧ ( )2 { [
( )2 ⎪ ⎡ 𝐼 4𝜋 4 ℎo (1 − 𝜇 2 )
𝜋 ⎢(𝑥 − ℎ )2 xyf 𝐿crd = 𝐼xf (𝑥of − ℎxf )2 + 𝐶wf
= 𝐴𝑓
𝐿 ⎨ ⎪ ⎢ of xf
𝐼yf 𝑡3
⎩ ⎣
] } 1∕4
( ) 𝐼xyf
2
𝜋 4 ℎo 4
𝐼xyf − (𝑥of − ℎxf ) 2
+ (3.114)
−2𝑦of (𝑥of − ℎxf ) 𝐼yf 720
𝐼yf

⎫ Lm = Distance between discrete restraints that



⎪ restrict distortional buckling (for conti-
+ ℎ2xf + 𝑦2of ⎥+ 𝐼xf + 𝐼yf ⎬ (3.108)
⎥ ⎪ nuously restrained members Lm = Lcrd )
⎦ ⎭ M1 and M2 = Smaller and larger end moments,
̃ respectively, in the unbraced segment
𝑘𝜙wg = Geometric rotational stiffness demanded by
(Lm ) of the beam; M1 /M2 is positive
web from flange∕web juncture when the moments cause reverse
( )2 th3 curvature and negative when bent in
π o single curvature
= (3.109)
L 60 k𝜙fe = Elastic rotational stiffness provided by
where 𝐿 = Minimum of 𝐿crd and 𝐿m where the flange to the flange/web juncture,
{ [ given in Eq. (3.106)
6𝜋 4 ℎo (1 − 𝜇 2 )
𝐿crd = 𝐼xf (𝑥of − ℎxf )2 + 𝐶wf k𝜙we = Elastic rotational stiffness provided by
𝑡3 the web to the flange/web juncture
]} 1∕4 [ ]
𝐼xyf
2 ( )2 19ℎ ( )4 ℎ 3
Et3 3 𝜋 o 𝜋 o
− (𝑥of − ℎxf ) 2
(3.110) = + + (3.115)
𝐼yf 12(1 − 𝜇 2 ) ℎo 𝐿 60 𝐿 240
DIRECT STRENGTH METHOD AND CONSIDERATION OF LOCAL AND DISTORTIONAL BUCKLING 107

Table 3.4 Geometric Flange Plus Lip Properties for C- and Z-Sections1,2,3

b b
d θ
θ

h h

Af = (b + d)t Af = (b + d)t

Jf = 1∕3bt3 + 1∕3dt3 Jf = 1∕3bt3 + 1∕3dt3

t(t2 b2 + 4bd3 + t2 bd + d4 ) t(t2 b2 + 4bd3 − 4bd3 cos2 (θ) + t2 bd + d4 − d4 cos2 (θ))


Ixf = Ixf =
12(b + d) 12(b + d)
t(b4 + 4db3 ) t(b4 + 4db3 + 6d2 b2 cos(θ) + 4d3 bcos2 (θ) + d4 cos2 (θ))
Iyf = Iyf =
12(b + d) 12(b + d)
tb2 d2 tbd2 sin(θ)(b + d cos(θ))
Ixyf = Ixyf =
4(b + d) 4(b + d)
Cwf = 0 Cwf = 0

b2 b2 − d2 cos(θ)
xof = xof =
2(b + d) 2(b + d)
−(b2 + 2db) −(b2 + 2db + d2 cos(θ))
hxf = hxf =
2(b + d) 2(b + d)
−d2 −d2 sin(θ)
hyf = yof = hyf = yof =
2(b + d) 2(b + d)
Notes:
1. b, d, and h are mid-line dimensions of cross-section.
2. x–y axis system is located at the centroid of the flange with x positive to the right from the centroid, and y positive
down from the centroid. Table 3.4 does not include the effect of corner radius. More refined values are permitted.
3. Variables are defined as follows:

Af = Cross-sectional area of flange


t = Thickness of cross-section
Jf = St. Venant torsion constant of flange
Ixf = x-axis moment of inertia of flange
Iyf = y-axis moment of inertia of flange
Ixyf = Product of the moment of inertia of flange
Cwf = Warping torsion constant of flange
xof = x distance from centroid of flange to shear center of flange
yof = y distance from centroid of flange to shear center of flange
hxf = x distance from centroid of flange to flange/web junction
hyf = y distance from centroid of flange to flange/web junction
108 3 STRENGTH OF THIN ELEMENTS AND DESIGN CRITERIA

k𝜙 = Rotational stiffness provided by a restraining (5) 2 ≤ ho /bo ≤ 8, and


element (brace, panel, sheathing) to the (6) 0.04 ≤ D sin𝜃/bo ≤ 0.5
flange/web juncture of a member (zero if the
compression flange is unrestrained) where ho = Out-to-out web depth as defined in Figure
̃
𝑘𝜙fg = Geometric rotational stiffness demanded by 3.30(c)
the flange from the flange/web juncture, given bo = Out-to-out flange width as defined in
in Eq. (3.108) Figure 3.30(c)
̃
𝑘𝜙wg = Geometric rotational stiffness demanded by D = Out-to-out lip dimension as defined in
the web from the flange/web juncture Figure 3.48
t = Base steel thickness
⎧ ( )2 ⎫ 𝜃 = Lip angle as defined in Figure 3.48
⎪ [45360(1 − 𝜉web ) + 62160] 𝐿 ⎪
⎪ ( ℎ )2
ℎo ⎪
⎪ 4⎪ For compression members:
ℎo 𝑡𝜋 2 ⎪ + 448𝜋 + 𝐿 [53 + 3(1 − 𝜉web )]𝜋 ⎪
2 o
( )2
= π2 E
13440 ⎨ ( )2 ( )4 ⎬ (3.116) t
⎪ ⎪ Fcrd = 𝛼kd (3.117)
𝜋 4 + 28𝜋 2 ℎ𝐿 + 420 ℎ𝐿 12(1 − 𝜇 2 ) bo
⎪ 𝑜 o ⎪
⎪ ⎪ where 𝛼 = A value that accounts for the benefit of an
⎪ ⎪
⎩ ⎭ unbraced length, Lm , shorter than Lcr , but
where 𝜉 web = (f1 – f2 )/f1 , stress gradient in the web, where can be conservatively taken as 1.0
f1 and f2 are the stresses at the opposite ends = 1.0 for 𝐿𝑚 ≥ 𝐿cr
of the web, f1 > f2 , compression is positive,
tension is negative, and the stresses are = (𝐿𝑚 ∕𝐿cr )𝓁𝑛(𝐿𝑚 ∕𝐿cr ) for 𝐿𝑚 < 𝐿cr (3.118)
calculated on the basis of the gross section Lm = Distance between discrete restraints that
(e.g., pure symmetrical bending, f1 = –f2 , restrict distortional buckling
𝜉 web = 2 )
( )
bo Dsin𝜃 0.6
All other variables are defined in Specification Section 𝐿cr = 1.2ho ≤ 10ho (3.119)
ho t
2.3.1.3.
( )
bo Dsin𝜃 1.4
The AISI Cold-Formed Steel Design Manual1.428 has 𝑘𝑑 = 0.05 ≤ 0.1 ≤ 8.0 (3.120)
ho t
provided 𝐿crd , 𝑘𝜙fe , 𝑘𝜙we , ̃𝑘𝜙fg , ̃
𝑘𝜙wg , 𝐹crd , or 𝐹crd ∕𝛽 in
tables for stud/joist stock sections. These tables can be used 𝐸 = Modulus of elasticity of steel
in design. 𝜇 = Poisson’s ratio of steel

3. Simplified Analytical Solutions. For C- or Z-Section For flexural members:


( )2
members with simple lip stiffeners, the following 𝜋2E 𝑡
𝐹crd = 𝛽𝑘𝑑 (3.121)
simplified expressions can be used to predict the 12(1 − 𝜇 2 ) 𝑏o
distortional buckling stress. The following provisions where 𝛽 = A value accounting for moment gradient,
are adapted from Sections 2.3.1.2 and 2.3.3.3 of the which is permitted to be conservatively taken
Commentary1.431 on the 2016 edition of the North as 1.0
American Specification.
= 1.0 ≤ 1 + 0.4(L∕L𝑚 )0.7 (1 + M1 ∕M2 )0.7 ≤ 1.3 (3.122)

Simplified Method for Unrestrained C- and Z-Sections where L = Minimum of Lcr (per Eq. 3.119) and Lm
Lm = Distance between discrete restraints that
with Simple Lip Stiffeners
restrict distortional buckling (for continuously
For C- and Z-sections that have no rotational restraint of the
restrained members Lm = Lcr )
flange and that are within the dimensional limits provided in
M1 and M2 = Smaller and larger end moment, respectively,
this section, Eq. (3.117) can be used to calculate a conserva-
in the unbraced segment (Lm ) of the beam;
tive prediction of distortional buckling stress, 𝐹crd , provided the
M1 /M2 is positive when the moments cause
following dimensional limits are met:
reverse curvature and negative when bent in
(1) 50 ≤ ho /t ≤ 200, single curvature
(2) 25 ≤ bo /t ≤ 100, ( )
b Dsin𝜃 0.7
(3) 6.25 < D/t ≤ 50, 𝑘d = 0.5 ≤ 0.6 o ≤ 8.0 (3.123)
ho t
(4) 45∘ ≤ 𝜃 ≤ 90∘ ,
DIRECT STRENGTH METHOD AND CONSIDERATION OF LOCAL AND DISTORTIONAL BUCKLING 109
4. Determination of Rotational Restraints, k𝜙 L1 , L2 = One-half joist spacing to the first and second
In Eqs. (3.105) and (3.112) the rotational stiffness k𝜙 sides respectively, as illustrated in Fig. 3.71
accounts for the rotational restraints to the compression k𝜙c = Connection rotational restraint
flange(s). For example, in cold formed steel framing = Values as specified in Table 3.6 for fasteners
systems, structural sheathings that attached to studs or spaced 12 in. o.c. (305 mm) or closer
joist members will provide rotational restraints. Such
restraints can be calculated using the following expres-
sions, which was developed based on the research by Table 3.5(b) 1 Gypsum Board Bending Rigidity
Schafer et al.3.276, 4.223 and was adopted in Appendix 1 Effective Stiffness (Typical Range), EIw
of the 2015 edition of the North American Standard for
Cold-Formed Steel Structural Framing.1.432 Board Thickness EI (Lbf-in.2 /in.) of
(in.) (mm) width (N-mm2 /mm)
The rotational stiffness, k𝜙 , shall be determined in accordance
with the following: 0.5 1500 to 4000
−1 (12.7) (220,000 to 580,000)
𝑘𝜙 = (1∕𝑘𝜙w + 1∕𝑘𝜙c ) (3.124)
0.625 3000 to 8000
where (15.9) (440,000 to 1,160,000)
𝑘𝜙w = Sheathing rotational restraint
Note:
= EI w ∕𝐿1 + EI w ∕𝐿2 for interior members (joists or rafters) 1. Gypsum board bending rigidity is obtained from the Gypsum
with structural sheathing fastened on both sides (3.125) Association.
= EI w ∕𝐿1 for exterior members (joists or rafters) with
structural sheathing fastened on one side (3.126)
5. Distortional Buckling of C- or Hat Sections Subject to
where EIw = Sheathing bending rigidity
= Values as specified in Table 3.5(a) for Bending with Lips in Compression
plywood and OSB The following analytical solution was developed by
= Values as specified in Table 3.5(b) for Glauz3.277 for C-Section or Hat sections subject to
gypsum board permitted only for bending where both lips are in compression and the
serviceability calculations flanges are under the stress gradient.

Table 3.5(a) 1,2 Plywood and OSB Sheathing Bending Rigidity, EIw (lbf-in2 /ft)
Strength Parallel to Stress Perpendicular to
Strength Axis Strength Axis
Plywood Plywood
Span Rating 3-ply 4-ply 5-ply OSB 3-ply 4-ply 5-ply OSB
24/0 66,000 66,000 66,000 60,000 3,600 7,900 11,000 11,000
24/16 86,000 86,000 86,000 86,000 5,200 11,500 16,000 16,000
32/16 125,000 125,000 125,000 125,000 8,100 18,000 25,000 25,000
40/20 250,000 250,000 250,000 250,000 18,000 39,500 56,000 56,000
48/24 440,000 440,000 440,000 440,000 29,500 65,000 91,500 91,500
16oc 165,000 165,000 165,000 165,000 11,000 24,000 34,000 34,000
20oc 230,000 230,000 230,000 230,000 13,000 28,500 40,500 40,500
24oc 330,000 330,000 330,000 330,000 26,000 57,000 80,500 80,500
32oc 715,000 715,000 715,000 715,000 75,000 615,000 235,000 235,000
48oc 1,265,000 1,265,000 1,265,000 1,265,000 160,000 350,000 495,000 495,000
Notes:
1. To convert to lbf-in.2 /in., divide table values by 12.
To convert to N-mm2 /m, multiply the table values by 9.415.
To convert to N-mm2 /mm, multiply the table values by 9.415.
2. Plywood and OSB bending rigidity are obtained from APA.
110 3 STRENGTH OF THIN ELEMENTS AND DESIGN CRITERIA

Table 3.6 1 Connection Rotational Restraint 𝛽, L and Lm , k𝜙fe are defined in Specification Section 2.3.3.3
included in Section 3.5.2(2).
T t k𝜙c k𝜙c {
6(1 − 𝜇2 ) [
(mils) (in.) (lbf-in./in./rad) (N-mm/mm/rad) 𝐿crd = 𝜋ℎ𝑒 𝐶wf + 𝐼xf (𝑥of − ℎxf )2
𝑡3 ℎ3𝑒
18 0.018 78 348
( )] } 1∕4
27 0.027 83 367 𝐼xyf
2
1
30 0.03 84 375 × 1− + (3.127)
𝐼xf 𝐼yf 120
33 0.033 86 384
43 0.043 94 419 k𝜙we = Elastic rotational stiffness provided by
54 0.054 105 468 the web to the flange/web juncture
68 0.068 123 546 [ ( )2 ( )4 ]
97 0.097 172 766 Et3 1 𝜋ℎ𝑒 1 𝜋ℎ𝑒
= 1+ +
6ℎ𝑒 (1 − 𝜇2 ) 6 𝐿 120 𝐿
Note:
1. Fasteners spaced 12 in. (25.4 mm) o.c. or less. (3.128)

̃
𝑘𝜙fg = Geometric rotational stiffness demanded by
The elastic distortional buckling moment, Mcrd , is calcu-
the flange from the flange/web juncture
lated as follows:
( )2 { [
𝑀crd = 𝑆f 𝐹crd [Eq. (3.111)] 𝜋
= 𝐼xf + 𝐼yf + 𝐴𝑓 ℎ2xf + 𝑦2of − 2𝑦of (𝑥of − ℎxf )
𝐿
where ( )]} ( )2
𝐼xyf 𝜋
𝑘𝜙fe + 𝑘𝜙we + 𝑘𝜙 × 𝜓𝑓 + 𝐼yf 𝜉𝑓 (3.129)
𝐹crd = 𝛽 [Eq. (3.112)] 𝐼yf 𝐿
̃
𝑘𝜙fg + ̃
𝑘𝜙wg
̃
𝑘𝜙wg = 0
Sf = Gross elastic cross-sectional modulus referenced √
to the compression fiber of the flange/web 𝐼yf
juncture, the point at which ho is measured ℎe = 3.5 + ℎ2xf (3.130)
𝐴f

L
z x Sheathing

df
Lf

wtf

Joist or Wall
Joist or Wall
Spacing
Framing
Interior Joist Exterior Joist or
or Wall Example Wall Stud

Figure 3.71 Illustration of L1 and L2 for sheathing rotational restraint1.432


DIRECT STRENGTH METHOD AND CONSIDERATION OF LOCAL AND DISTORTIONAL BUCKLING 111

fcg
f1 (compression)
f2 (tension)
Shear center of
Flange/web stiffened flange
Centroid of
Juncture
stiffened flange
hxf xof

Figure 3.72 Flange stresses for bending about axis parallel to web.

𝜉 f = (f1 – f2 )/f1 , stress gradient in the flange,


where f1 is the stress at the extreme
compression fiber of the flange, f2 is the
stress at the flange/web juncture,
compression is positive, tension is negative,
and the stresses are calculated on the basis of
the gross section (see Fig. 3.72)
𝜓 f = fcg /f1 , stress ratio in the flange, where f1
is the stress at the extreme compression
fiber of the flange, fcg is the stress at the
centroid of the flange, compression is
positive, tension is negative, and the stresses
are calculated on the basis of the gross
section (see Figure 3.72)

Example 3.8 For the C-section shown in Fig. 3.73, deter-


mine the distortional buckling moment when bending about
x axis using the simplified analytical solution based on
Eq. (3.121) and the expressions given in Specification
Section 2.3.3.3. Figure 3.73 Example 3.8

SOLUTION Based on Eq. (3.120), the plate buckling coefficient for


A. Distortional Buckling Moment Determined by the Simpli- distortional buckling is
fied Method ( ) [ ]0.7
In order to use Eq. (3.121), the following geometric limits 𝑏o Dsinθ 0.7 3.5(0.72) sin 90∘
𝑘crd = 0.6 = 0.6 = 1.4
should be checked as the first step: ℎo t 10(0.075)
50 ≤ (ℎo ∕𝑡 = 133.33) < 200 OK Since 0.5 < 𝑘d < 8, use 𝑘d = 1.40. Because 𝑀1 and 𝑀2 are
not given in the problem, use 𝛽 = 1.0 as a conservative value.
25 ≤ (𝑏o ∕𝑡 = 46.67) < 100 OK Therefore,
6.25 < (𝐷∕𝑡 = 9.6) < 50 OK 𝜋 2 (29,500) ( 0.075 )2
𝐹d = (1)(1.40) = 17.14 ksi
45∘ ≤ (𝜃 = 90∘ ) = 90∘ OK 12(1 − 0.32 ) 3.5
The elastic modulus relative to the extreme compression
2 ≤ (ℎo ∕𝑏o = 2.86) < 8 OK fiber,which can be calculated using using linear method (Fig.
0.04 ≤ (𝐷 sin 𝜃∕𝑏o = 0.21) < 0.5 OK 1.32), is
𝑆f = 4.11 in.3
From Eq. (3.121), the elastic distortional buckling stress is
calculated as follows: The critical elastic distortional buckling moment
( )2 [(Eq.(3.104)] is
𝜋2E t
Fcrd = kd [Eq. (3.121)] 𝑀crd = 𝑆f 𝐹crd = (4.11)(17.14) = 70.45 in-kips
12(1 − 𝜇 ) bo
2
112 3 STRENGTH OF THIN ELEMENTS AND DESIGN CRITERIA

B. Distortional Buckling Moment Based on Specification According to Eq. (3.114), the critical unbraced length of
Section 2.3.3.3 distortional buckling, Lcrd , can be computed as follows:
Based on the equations listed in Table 3.4, the geometric { [
flange properties for the C-section can be computed as 4𝜋 4 ℎo (1 − 𝜇2 )
𝐿crd = 𝐼xf (𝑥of − ℎxf )2 + 𝐶wf
follows. The reason for these calculations is that the 𝑡3
mechanical model for prediction of distortional buckling ] } 1∕4
strength considers the flange itself as a “column” which 𝐼xyf
2
𝜋 4 ℎ𝑜 4
may undergo restrained flexural–torsional buckling, and the − (𝑥of − ℎxf ) 2
+
𝐼yf 720
restraint comes from the web and any additional attachments
{
ℎ = ℎ0 − 𝑡 = 10.000 − 0.075 = 9.925 in. 4𝜋 4 (10.000)(1 − 0.32 )
=
(0.075)3
𝑏 = 𝑏0 − 𝑡 = 3.500 − 0.075 = 3.425 in. [ ]
(0.00708)[1.428 − (−1.997)]2 + 0−
𝑑 = 𝐷 − 𝑡∕2 = 0.720 − 0.075∕2 = 0.6825 in. × 0.0249 2
[1.428 − (−1.997)]2
𝐴f = (𝑏 + 𝑑)𝑡 = (3.425 + 0.6825)(0.075) = 0.308 in.2 0.376
}1∕4
𝐼𝑥f = 𝑡[𝑡2 𝑏2 + 4bd3 + 𝑡2 bd + 𝑑 4 ]∕12(𝑏 + 𝑑) 𝜋 4 (10.000)4
+ = 27.07 in.
720
= (0.075)[(0.075)2 (3.425)2 + 4(3.425)(0.6825)3
Assume that the distortional buckling length L equals 𝐿crd ,
+ (0.075)2 (3.425)(0.6825)
𝐿 = 𝐿crd = 27.07 in.
+ (0.6825)4 ]∕12(3.425 + 0.6825) = 0.00708 in.4
From Eq. (3.106), the elastic rotational stiffness provided
𝐼𝑦f = 𝑡[𝑏4 + 4bd3 ]∕12(𝑏 + 𝑑) by the flange to the flange/web juncture, 𝑘𝜙fe , is
[ ]
= (0.075)[(3.425)4 + 4(0.6825)(3.425)3 ]∕12(3.425 ( )4 I2xyf
π
k𝜙fe = EIxf (xof − hxf ) + ECwf − E
2
(x − hxf ) 2
+ 0.6825) L Iyf of
= 0.376 in4 . ( )2
π
+ GJf
𝐼xyf = tb2 𝑑 2 ∕4(𝑏 + 𝑑) L

= (0.075)(3.425)2 (0.6825)2 ∕4(3.425 + 0.6825) ⎡(29,500)(0.00708) ⎤


( )4 ⎢[1.428 − (−1.997)]2 ⎥
π ⎢
= 0.0249 in4 . = (0.0249)2 ⎥
27.07 ⎢+(29,500)(0) − (29,500) ⎥
⎢ (0.376) ⎥
𝑥of = 𝑏2 ∕2(𝑏 + 𝑑) = (3.425)2 ∕2(3.425 + 0.6825) ⎣[1.428 − (−1.997)]2 ⎦
= 1.428 in. ( )2
π
+ (11,300)(0.000578)
𝑦of = −𝑑 2 ∕2(𝑏 + 𝑑) = −(0.6825)2 ∕2(3.425 + 0.6825) 27.07
= 0.429 in.-kips∕in.
= −0.0567 in.
From Eq. (3.115), the elastic rotational stiffness provided
ℎ𝑥f = −[𝑏2 + 2db]∕2(𝑏 + 𝑑)
by the web to the flange/web juncture, 𝑘𝜙we , is
= −[(3.425)2 + 2(0.6825)(3.425)]∕2(3.425 + 0.6825) [ ]
( )2 19h ( )4 h 3
Et3 3 π o π o
= −1.997 in. k𝜙we = + +
12(1 − μ2 ) ho L 60 L 240
𝐽f = [bt3 + dt3 ]∕3 ( )2
⎡ 3 π (19)(10.000) ⎤
= [(3.425)(0.075)3 + (0.6825)(0.075)3 ]∕3 (29,500)(0.075)3 ⎢ 10.000 + 27.07 60
+⎥
= ⎢ ( )4 ⎥
= 0.000578 in4 . 12(1 − 0.3 ) ⎢
2 π (10.000)3

⎣ 27.07 240 ⎦
𝐶wf = 0.0 in.6 = 0.391 in.-kips∕in.
DIRECT STRENGTH METHOD AND CONSIDERATION OF LOCAL AND DISTORTIONAL BUCKLING 113
From Eq. (3.108), the geometric rotational stiffness restraints are provided to the compression flange of the
demanded by the flange from the flange/web juncture, ̃
𝑘𝜙fg , beam, 𝑘𝜙 = 0. Therefore,
is 0.429 + 0.391 + 0.0
( )2 𝐹crd = 1.0 = 32.98 ksi
̃ π 0.0220 + 0.00286
k𝜙fg =
L which is 92% higher than the elastic distortional buckling
⎧ ⎡ ( )2 ⎤ ⎫ stress computed in item A above using the simplified provi-
⎪ ⎢(x − h )2 Ixyf − 2y (x − h ) ⎥ ⎪ sion on the basis of Section 2.3.3.3 of the Commentary1.431
⎪ ⎢ of xf Iyf of of xf
⎥ ⎪
× ⎨Af ( ) + Ixf + Iyf ⎬ on the Specification. From item A, Sf = 4.11 in.3 Based on
⎪ ⎢ ⎢ Ixyf ⎥ ⎪ Eq. (3.111), the critical elastic distortional buckling moment,
× I + h2xf + y2of ⎥
⎪ ⎣ yf ⎦ ⎪ 𝑀crd , is
⎩ ⎭
⎧ ( )2 ⎫ 𝑀crd = 𝑆f 𝐹crd = (4.11)(32.98) = 135.55 in.-kips
⎡ 2 0.0249 ⎤
⎪ ⎢[1.428 − (−1.997)] ⎥⎪
⎪ 0.376
( )2 ⎪ ⎢−2(−0.0567) ⎥⎪
π (0.308) ⎢ ( ) ⎥⎪
= Example 3.9 Determine the distortional buckling force
27.07 ⎨ ⎪
⎢[1.428 − (−1.997)] 0.0249 ⎥⎬
⎢ 0.376
⎥⎪ 𝑃crd and 𝑀crd of C-Section 9CS2.5×059 (designation per
⎪ ⎣+(−1.997)2 + (−0.0567)2 ⎦⎪ AISI Cold-Formed Steel Design Manual1.428 ) using the
⎪+0.00708 + 0.376 ⎪
⎩ ⎭ numerical and analytical solutions. The member yield stress,
𝐹y = 55 ksi.
= 0.0220 (in.-kips∕in)∕ksi
From Eq. (3.116), the geometric rotational stiffness
SOLUTION
demanded by the web from the flange/web juncture, ̃
𝑘𝜙wg ,
where ξweb = 2 for pure bending. Numerical Solution. The cross-section geometry and
gross section properties of 9CS2.5×059 can be obtained
⎧ ( )2 ⎫ from the 2017 edition of the AISI Cold-Formed Steel
⎪ [45360(1 − ξweb ) + 62160] L + 448π2 + ⎪
⎪ ( h )2
ho ⎪ Design Manual1.428 : 𝐴g = 0.881 in.2 ; Ix = 10.3 in.4 , and
⎪ o
[53 + 3(1 − ξ )]π 4 ⎪ Sx = 2.29 in.3 . From CUFSM analysis, the signature
ho tπ2 ⎪ web ⎪
̃
k𝜙wg =
L
curves under the compression and bending are as shown
13,440 ⎨⎪
( )2
L
( )4
L


π + 28π h
4 2 + 420 h in Figs. 3.70 and 3.69, respectively. From the ploted
⎪ o o ⎪
⎪ ⎪ deformed shapes, the distortional buckling occurred at
⎪ ⎪ Pcrd ∕Py = 0.26, and 𝑀crd ∕𝑀y = 0.85. The distortional
⎩ ⎭
buckling force and moment can be calculated as follows:
⎧ [45360(1 − 2) + 62160] ⎫
⎪ ( )2 ⎪ 𝑃y = 𝐴g 𝐹y = (0.881 in.2 )(55 ksi) = 48.455 kips
⎪ 27.07
+ 448π 2
+ ⎪
⎪ ( )2
10.000 ⎪ 𝑃crd = 0.26 𝑃y = (0.26)(48.455 kips) = 12.598 kips
⎪ 10.000
[53 + 3(1 − 2)]π 4 ⎪
2 ⎪ ⎪
(10.000)(0.075)π 27.07
𝑀y = 𝑆x 𝐹y = (2.29 in.3 )(55 ksi) = 125.95 kip-in.
= ⎨ ( )2 ( )4 ⎬
13440 ⎪ π4 + 28π2 27.07
+ 420 10.000 ⎪
27.07
𝑀crd = 0.85 My = (0.85)(125.95) = 107.058 kip-in.
⎪ 10.000 ⎪
⎪ ⎪
⎪ ⎪ Analytical Solution. Instead of using the expressions
⎪ ⎪ provided in Specification Sections 2.3.1.3 and 2.3.3.3 to
⎩ ⎭
determine the distortional buckling stresses, AISI Design
= 0.00286 (in.-kips∕in.)∕ksi
Manual Tables III-4 and II-7 can be used. This example
From Eq. (3.112), the elastic distortional buckling stress will illustrate how to use the Design Manual to determine
𝐹crd is the distortional buckling stresses.
𝑘𝜙fe + 𝑘𝜙we + 𝑘𝜙 Since there is no indication of rotational restraints to
𝐹crd = 𝛽
̃
𝑘𝜙fg + ̃𝑘𝜙wg members’ flanges in this example, 𝑘𝜙 = 0. From AISI Design
Manual Table III-4, the following values are obtained:
In the above equation, use 𝛽 = 1.0 as a conservative
value (i.e., ignoring moment gradient). Since no rotational
114 3 STRENGTH OF THIN ELEMENTS AND DESIGN CRITERIA

𝐿crd = 28.6 in.; 𝑘𝜙fe = 0.153 kips; ̃


𝑘𝜙fg = 0.00705 in.2 ; factor, i.e. 𝑀cr ∕𝑀y , where 𝑀y is the yield moment. The
𝑘𝜙we = 0.123 kips; ̃𝑘𝜙wg = 0.00867 in.2 ; 𝐹crd = 17.6 ksi. buckled cross-sections illustrate the buckling mode taking
The distortional buckling force: 𝑃crd = 𝐴g 𝐹crd = place at the corresponding minimum. As shown in the figure,
(0.881 in.2)(17.6 ksi) = 15.506 kips the local buckling occurs at the half-wavelength of 6.4 in.
From AISI Design Manual Table II-7, and the corresponding load factor is 2.39; the distortional
𝐿crd = 25.8 in.; 𝑘𝜙fe = 0.221 kips; ̃
𝑘𝜙fg = 0.00863 in.2 ; buckling occurs at the half-wavelength of 67 in. and load
factor is 0.86. From the numerical analysis output, the yield
𝑘𝜙we = 0.209 kips; ̃𝑘𝜙wg = 0.00181 in.2 ; 𝐹crd ∕β = 41.2 ksi.
moment 𝑀y = 161 in.-kips. Therefore, the local buckling
Assume moment distribution influence is not considered in
and distortional buckling moments can be calculated as
this example, 𝛽 = 1
follows:
𝑀crd = 𝑆f 𝐹crd = (2.29 in.3 )(41.2 ksi) = 94.348 kip-in
𝑀cr𝓁 = 6.4𝑀y = 384.8 in.kips
For cases where rotational restraints are provided to the
compressive flange, for example, structural sheathings 𝑀crd = 0.86𝑀y = 138.5 in.-kips
or metal sheets are attached to the flange, the rota- The results indicate that the distortional buckling moment
tional stiffness k𝜙 can be determined analytically using is much lower than the local buckling, and likely controls
Eq. (3.124) or experimentally (refer to AISI test standard, member strength.
AISI S9183.279 .) The distortional buckling moment can also be determined
analytically using the equations provided in Section 3.5.2(5).
Example 3.10 Determine the local and distortional buck- The following calculations are based on those equations
ling moment 𝑀cr𝓁 and 𝑀crd of U-shaped section as shown given in Section 3.5.2(5):
in Fig. 3.74 when the lips are in compression. Assume the
member is laterally braced and does not subject to global Calculate the centerline dimensions:
buckling. The yield stress 𝐹y = 33 ksi.
Note, the equations in Table 3.4 and the symbols in Section
SOLUTION 3.5.2(5) are based on a coordinate system with the x axis as
The numerical method of finite strip method is employed to the symmetric axis.
determine the local and distortional buckling moment. By ℎ = 8 − 0.105 = 7.895 in.; 𝑏 = 8 − 0.105 = 7895 in.;
using the CUFSM software, the following buckling signature 𝑑 = 1.34 − 0.105∕2 = 1.2875 in.
curve is obtained: Af = (b + d)t = 0.9642 in.2
In Fig. 3.75, two numbers are shown below the local
minima of the signature curve: the first number is the Jf = 1∕3bt3 + 1∕3dt3 = 0.003543 in.4
buckling half-wavelength, and the second number is the load t(t2 b2 + 4bd3 + t2 bd + d4 )
Ixf = = 0.0676 in.4
12(b + d)
t(b4 + 4db3 )
Iyf = = 6.117 in.4
12(b + d)
tb2 d2
Ixyf = = 0.2954 in.4
4(b + d)
Cwf =0
b2
xof = = 3.394 in.
2(b + d)
−(b2 + 2db)
hxf = = −4.501 in.
2(b + d)
−d2
hyf = yof = = −0.0903 in.
2(b + d)
From numerical analysis, the centroidal axis is 3.188 in.
from the bottom flange and 4.812 in. from the top of the lips:
Figure 3.74 Example 3.10 𝑐c = 4.812 in.; 𝑐t = 3.188 in.
DIRECT STRENGTH METHOD AND CONSIDERATION OF LOCAL AND DISTORTIONAL BUCKLING 115
6

4
Load Factor

6.4, 2.39
2

1
67.0, 0.86

0
1 10 100 1000
Length

Figure 3.75 U-shaped cross-section buckling signature curve

( )2
Since the centroid axis is closer to the bottom flange, the π
compression stress controls: + GJf = 0.421 kips
L
( )2 { [
𝑓1 = 𝐹y = 33 ksi; 𝑓2 = −𝑓1 × 𝑐t ∕𝑐c = −21.863 ksi ̃ π
k𝜙fg = Ixf + Iyf + Af x2hf + y2of − 2yof (xof − hxf )
The stress at flange centroid: L
( )]} ( )2
𝑓cg = 𝑓1 (|ℎxf | + 𝑡∕2 − 𝑐t )∕𝑐c = 8.647 ksi (Note: 𝑓cg is Ixyf π
positive since it is under compression.) × ψf + I = 0.0318 in2
Iyf L yfξf
ξf = (𝑓1 − 𝑓2 )∕𝑓1 = 1.662 ̃
k𝜙wg = 0
ψf = fcg ∕f1 = 0.262
√ Assume β, which accounts for moment gradient, is
Iyf ignored, i.e. β = 1
ℎe = 3.5 + h2xf = 18.05 in. The distortional buckling stress is calculated:
Af
{ [ ( )] 𝑘𝜙fe + 𝑘𝜙we + 𝑘𝜙
6(1 − μ2 ) I2xyf 𝐹crd = 𝛽 = 25.294 ksi
𝐿crd = πhe Cwf + Ixf (xof − hxf ) 1 −
2 ̃
𝑘𝜙fg + ̃
𝑘𝜙wg
t3 h3e Ixf Iyf
From numerical analysis,
} 1∕4 Moment inertia 𝐼 = 23.245 in.4
1 Section modulus 𝑆f = 𝐼∕𝑐c = 4.83 in.3
+ = 72.52 in.
120
𝑀crd = 𝐹crd 𝑆f = (9.957)(4.83) = 122.2 in.-kips
Assume no bracing for distortional buckling is provided,
therefore, 𝐿m > 𝐿crd , Compared with the numerical analysis result 𝑀crd =
𝐿 = minimum (𝐿crd , 𝐿m ) = 𝐿crd = 72.52 in. 138.5 in-kips. the analytical solution is reasonable close.
[ ( )2 ( )4 ]
Et3 1 πhe 1 πhe 3.5.3 Influence of Perforation on Local
𝑘𝜙we = 1+ + and Distortional Buckling
6he (1 − μ2 ) 6 L 120 L
Researchers have observed the influence of holes on
= 0.383 kips
[ ] the local buckling.3.273–3.275 Both numerical and analyt-
( )4 I2xyf ical approaches are provided in the North American
π
𝑘𝜙fe = EIxf (xof − hxf ) + ECwf − E
2
(x − hxf ) 2
Specification1.417 to consider the hole effects.
L Iyf of
116 3 STRENGTH OF THIN ELEMENTS AND DESIGN CRITERIA

3.5.3.1 Local Buckling 3.5.3.2 Distortional Buckling


(a) Numerical Solutions. The numerical shell finite (a) Numerical Solutions. Similar to the consideration
element method can be used to consider the hole of local buckling, an approximate method that can
effect. However, since shell finite element software be used in conjunction with the finite strip method
may not be readily available for regular users, the was developed.3.278 To implement the method, a
following approximate method was developed by finite strip analysis is performed with the gross
Moen and Schafer,3.278 which can be used in conjunc- cross-section to identify the critical distortional buck-
tion with the finite strip method to predict the local ling half-wavelength, 𝐿crd . Then the thickness of the
buckling stress. The method assumes that the local element with hole (not just the portion of the hole)
buckling occurs as either buckling of the unstiffened is modified from t to tr using the expression below.
strip adjacent to the hole at the net cross-section This revised thickness is to account for the reduced
(𝑃cr𝓁h ) or as local buckling of the gross cross-section restraints of the element with hole to other elements
between the hole (𝑃cr𝓁nh ), and: when distortional buckling occurs.
( )
𝑃cr𝓁 = min(𝑃cr𝓁nh , 𝑃cr𝓁h ) (3.131a) 𝐿h 1∕3
𝑡𝑟 = 𝑡 1 − (3.132)
Both 𝑃cr𝓁nh and 𝑃cr𝓁h can be determined using the 𝐿crd
finite strip method such as CUFSM. To ensure a where 𝐿h is the length of the hole. This simplification
consistent comparison of 𝑃cr𝓁nh and 𝑃cr𝓁h , the refer- is only appropriate for cases of flat-punched discrete
ence load applied in the analysis needs to be the holes in the web or flange (or both).
same. More detailed discussion can be found in the For patterned holes along web3.275 , the following
Commentary1.431 of the 2016 edition of the North reduction expression should be used:
American Specification. The same approach can ( )
𝐴web,net 1∕3
also be used to determine the beam local buckling 𝑡𝑟 = 𝑡 (3.133)
moment: 𝐴web,gross

𝑀cr𝓁 = min(𝑀cr𝓁nh , 𝑀cr𝓁h ) (3.131b) where t is the thickness of the web, 𝐴web,net is net area
of the web along the full member length; 𝐴web,gross is
where 𝑀cr𝓁nh and 𝑀cr𝓁h are the buckling moments the gross area of the web along the full member length.
of the gross cross-section (no hole) and the net The finite strip analysis is then performed using the
cross-section (with hole), respectively. modified cross section with the reference load consis-
(b) Analytical Solutions. For compression members, refer tent with the gross area. The determined critical distor-
to Section 3.5.1(2). For the element with hole: the tional buckling half-wavelength, 𝐿crd (based on the
local buckling stress, 𝐹cr𝓁 , should be calculated as gross cross-section), and the corresponding distor-
both unstiffened elements at the hole location (for tional buckling load, Pcrd , can then be obtained. This
determining Pcr𝓁h ) and as an element where the hole method has been validated for compressive members
is not located (for determining 𝑃cr𝓁nh ). For the unstiff- and is recommended for use for flexural members.
ened element, the buckling stress 𝐹cr𝓁 should be modi- (b) Analytical Solutions. For compression members
fied by multiplying a ratio 𝐴net ∕𝐴g to account for the meeting the requirements of Specification Section
net section. The smallerest 𝐹cr𝓁 from the element with 2.3.1.3, the analytical expressions given in Section
hole and the other element is then used to determine 2.3.1.3 can be used except the thickness, t, in Eqs.
𝑃cr𝓁 of the cross-section. (3.107) and (3.109) is replaced by tr determined per
For flexural members, refer to Section 3.5.1(2). For Eq. (3.121) for flat-punched holes, or thickness, t, in
the element with hole: the local buckling stress, 𝐹cr𝓁 , Eqs. (3.107), (3.109), and (3.110) is replace by 𝑡r per
should be calculated as both unstiffened elements at Eq. (3.122) for patterned holes.
the hole location (for determining 𝑀cr𝓁nh ) and as an For flexural members meeting the requirements of
element where the hole is not located (for determining Specification Section 2.3.3.3, the analytical expres-
𝑀cr𝓁nh ). For the unstiffened element, the buckling sions given in Section 2.3.3.3 can be used except the
stress 𝐹cr𝓁 should be modified by multiplying a ratio thickness t in Eqs. (3.115) and (3.116) is replaced by
𝑆fnet ∕𝑆fg to account for the net section. The smallest tr determined per Eq. (3.121) for flat-punched holes or
𝐹cr𝓁 extrapolated to the extreme compression fiber is t in Eqs. (3.114), (3.115) and (3.116) is replaced by 𝑡r
then used to determine 𝑀cr𝓁 of the cross-section. determined per Eq. (3.122) for patterned holes.
ADDITIONAL INFORMATION 117

AISI Cold-Formed Steel Design Manual 2017 edition1.431


contains design examples for members with holes.

3.6 PLATE BUCKLING OF STRUCTURAL SHAPES


Section 3.3 discussed the local buckling of stiffened and
unstiffened compression elements, for which the edges were
assumed to be simply supported. If the actual restraining
effects of adjoining cross-sectional elements are taken into
account, the plate buckling coefficient k for box sections,
channels, and Z-sections used as columns can be found from
Fig. 3.76. These curves are based on the charts developed
by Kroll, Fisher, and Hei-merl.3.116 Additional information
can be found from Refs. 1.94, 1.158, 3.8, 3.80, 3.117–3.123,
3.195, and 3.196 and in Chapter 5 on compression members.
The advantages of using a numerical solution for the
design of cold-formed steel members are discussed by
Schafer and Pekoz in Refs. 3.195 and 3.196. The element
interaction can be handled properly by the numerical
Figure 3.76 Plate buckling coefficient k for side h′ of columns.
solution.

can also be found from the publications included in the list


3.7 ADDITIONAL INFORMATION
of references for this chapter.
The strength of thin compression elements and the current The structural behavior of webs of beams subjected to
design criteria were discussed in this chapter on the basis of shear or bearing is discussed in Chapter 4 on the design of
the publications referred to in the text. Additional informa- flexural members. The buckling behavior of closed cylin-
tion on the strength of compression elements and members drical tubular members is discussed in Chapter 7.
3. Bracing requirements
4. Shear lag
5. Flange curling

In general, long-span, shallow beams are governed by


deflection and medium-length beams are controlled by
CHAPTER 4 bending strength. For short-span beams, shear strength may
be critical.
For design tables and charts, reference should be made to
Part II of the AISI Design Manual.1.428
Flexural Members
4.2 BENDING STRENGTH AND DEFLECTION
4.2.1 Introduction
In the design of flexural members, sufficient bending strength
must be provided, and at the same time the deflection of
the member under service loads should not exceed specific
limitations.
4.1 GENERAL REMARKS
A. ASD Method. According to Section 1.8.1.1, the
Beams are used to support transverse loads and/or applied ASD method requires the member flexural or bending
moment. Cold-formed steel sections such as I-sections, strength to satisfy the following requirement:
C-sections (channels), Z-shapes, angles, T-sections, hat
sections, and tubular members (Fig. 1.2) and decks and 𝑀 ≤ 𝑀a (4.1)
panels (Fig. 1.11) can be used as flexural members. where M is the required flexural strength or bending
In the design of cold-formed steel flexural members, moment for ASD computed from the load combina-
consideration should first be given to the moment-resisting tions discussed in Section 1.8.1.2 and 𝑀a is the allow-
capacity and the stiffness of the member, which may not be a able flexural strength or bending moment determined
constant value due to the noncompactness of the thin-walled by Eq.(4.2):
section and the variation of the moment diagram. Second, 𝑀
𝑀a = n (4.2)
the webs of beams should be checked for shear, combined Ωb
bending and shear, web crippling, and combined bending where Ωb = 1.67 is the safety factor for flexural or
and web crippling. In addition, flexural members must be bending strength where the bending strength is deter-
braced adequately to ensure their sufficient lateral–torsional mined in Section 4.2. For those members that their
buckling strength. Bracing to compression flange of a geometry or material property is outside the limits
flexural member can also increase member’s distortional provided in Table 3.1, rational engineering analysis
buckling strength. may be applied to determine the member strength. In
Unlike hot-rolled heavy steel sections, in the design of this case, Ωb = 2.00 should be used according to the
thin-walled cold-formed steel beams, special problems such North American Specification Section A1.2(c). 𝑀n in
as shear lag and flange curling are also considered to be Eq. (4.2) is the smallest nominal flexural strength or
important matters due to the use of thin material. Further- moment determined from the following four design
more, the design of flexural members can be even more considerations:
involved if the increase of steel mechanical properties due 1. Strength of initiation of yielding and global
to cold work is to be utilized. lateral–torsional buckling in accordance with
Based on the above general discussion, the following Section 4.2.2
design features are considered in this chapter with some 2. Strength of global buckling interacting with local
design examples for the purpose of illustration: buckling in accordance with Section 4.2.3
3. Strength of distortional buckling in accordance
1. Bending strength and deflection with Section 4.2.4
2. Design of webs for shear, combined bending and 4. Strength of beams having one flange attached to
shear, web crippling, and combined bending and web deck or sheathing determined in accordance with
crippling Section 4.2.8

119

$PME'PSNFE4UFFM%FTJHO 'JGUI&EJUJPO8FJ8FO:V 3PHFS"-B#PVCFBOE)FMFO$IFO


¥+PIO8JMFZ4POT *OD1VCMJTIFECZ+PIO8JMFZ4POT *OD
120 4 FLEXURAL MEMBERS

In addition to the above-listed four cases, consid- methods for determining the lateral–torsional buckling
eration should also be given to shear lag problems for strength of singly-, doubly-, and point-symmetric sections
unusually short span beams (see Section 4.2.10). Inter- according to the actual number and location of braces. The
action of bending with web shear, web crippling, and design of braces is discussed in Section 4.4.
torsions should be considered as well.
B. LRFD Method. According to Section 1.8.2.1, the 4.2.2.1 Doubly and Singly Symmetric Sections When a
LRFD method requires the member flexural or simply supported, locally stable I-beam is subject to a pure
bending strength to satisfy the following requirement: moment M as shown in Fig. 4.2, the following differential
equations for the lateral–torsional buckling of such a beam
𝑀u ≤ 𝜙b 𝑀n (4.3)
are given by Galambos in Ref. 2.45:
where 𝑀u is the required flexural strength or bending
EI 𝑦 𝑢iv + 𝑀𝜙′′ = 0 (4.6)
moment for LRFD computed from load combinations
′′
(see Section 1.8.2.2); EC𝑤 𝜙 − GJ𝜙 + 𝑀𝑢 = 0
iv
(4.7)
𝜙b = 0.90 is the resistance factor where the bending
strength is determined in Section 4.2. For member where 𝑀 =pure bending moment
strength determined in accordance with rational engi- 𝐸 =modulus of elasticity
neering analysis, 𝜙b = 0.80 should be used according 𝐺 =shear modulus, = 𝐸∕2(1 + μ)
to the North American Specification Section A1.2(c). 𝐼y =moment of inertia about the 𝑦 axis
C. LSD Method. According to Section 1.8.3.1, the LSD 𝐶w =warping constant of torsion of the cross
method requires that the member flexural or bending section (see Appendix B)
𝐽 = St. Venant torsion constant of cross section
strength to satisfy the following requirement ∑
approximately determined 13 𝑏i 𝑡3𝑖
𝑀f ≤ 𝜙b 𝑀n (4.4) 𝑢 = deflection of shear center in x direction
where 𝑀f is the bending moment for LSD computed 𝜙 = angle of twist
from load combinations (see Section 1.8.3.2); 𝑀n is
the nominal flexural resistance; and 𝜙b = 0.90 is the The primes indicate differentiation with respect to z.
resistance factor where the bending strength is deter- Considering the simply supported condition, the end
mined in Section 4.2. For member strength determined sections cannot deflect or twist; they are free to warp, and no
in accordance with rational engineering analysis, 𝜙b = end moment exists about the y axis. The boundary conditions
0.75 should be used according to the North American are
Specification. 𝑢(0) = 𝑢(𝐿) = 𝜙(0) = 𝜙(L) = 0 (4.8)
𝑢′′ (0) = 𝑢′′ (𝐿) = ϕ′′ (0) = ϕ′′ (𝐿) = 0 (4.9)
4.2.2 Yielding and Lateral–Torsional Buckling The solution of Eqs. (4.6) and (4.7) gives the following
Strength equation for the critical lateral buckling moment:
√ ( )
A cold-formed cross-section with small w/t ratio may fail by n𝜋 𝑛2 𝜋 2 ECw
yielding. The yielding moment is defined by Eq. (4.5): 𝑀cre = EI 𝑦 GJ 1 + (4.10)
𝐿 GJL2
𝑀y = 𝑆fy 𝐹y (4.5) where L is the span length and 𝑛 = 1, 2, 3, ⋯
where 𝑀y is the yield moment and 𝑆fy is the elastic section The deflected shape of the beam is
( )
modulus of cross-section relative to extreme fiber in first n𝜋z
𝜙 = 𝐶 sin (4.11)
yielding. As shown in Fig. 4.1, on a balanced section (Fig. 4.1 𝐿
(a)) or a section with neutral axis closer to the tension flange and the lateral deflection u can be determined by
(Fig. 4.1(c)), 𝑆fy is the section modulus about compression CML2 sin(n𝜋z∕𝐿)
flange; while for the section with neutral axis closer to the 𝑢= (4.12)
𝑛2 𝜋 2 EI 𝑦
compression flange (Fig. 4.1(b)), tension flange will yield
first, 𝑆fy should be the section modulus about the tension The deflection history of the I-beam is shown in Fig. 4.3.
flange. When 𝑀 ≤ 𝑀cre prior to lateral–torsional buckling taking
In general, cold-formed steel flexural members with open place, the beam deflects in the y direction. The vertical deflec-
cross-sections tend to twist and deflect laterally due to small tion 𝑣 can be obtained from Eq. (4.13) for in-plane bending,
lateral–torsional resistance. This section contains the design EI 𝑥 𝑣′′ = −𝑀 (4.13)
BENDING STRENGTH AND DEFLECTION 121

Figure 4.1 Stress distribution for yield moment: (a) balanced sections; (b) neutral axis close to
compression flange (initial yielding in tension flange); (c) neutral axis close to tension flange (initial
yielding in compression flange).

From Eq. (4.10), for 𝑛 = 1, the lowest critical moment for


lateral–torsional buckling of an I-beam is equal to
√ ( )
𝜋 𝜋 2 EC𝑤
𝑀cre = EI 𝑦 GJ 1 + (4.16)
𝐿 GJL2
Figure 4.2 Simply supported beam subjected to end moments.
For I-beams (Fig. 4.4)
Solving Eq. (4.13) and using the boundary conditions
𝑣(0) = 𝑣(𝐿) = 0, the deflection equation is2.45 𝑏3 td2
[( ) ( ) ] 𝐶w ≈ (4.17)
24
ML2 𝑧 𝑧 2
𝑣= − (4.14) 𝑏3 𝑡
2EI 𝑥 𝐿 𝐿 𝐼𝑥 ≈ (4.18)
6
When the beam buckles laterally, the section rotates about
the center of rotation 𝐶LB . This point is located at a distance Equation (4.16) can then be rewritten as follows:
of 𝑦LB below the shear center of the section as determined by
Eq. (4.15), √
𝑢 ML2 𝜋 𝐸 2 𝐼𝑦2 𝑑 2 ( 𝜋 )2
𝑦LB ≈ = 2 2 (4.15) 𝑀cre = EI 𝑦 GJ + (4.19)
𝜙 𝑛 𝜋 EI 𝑦 𝐿 4 𝐿
122 4 FLEXURAL MEMBERS

lateral–torsional buckling stress1.161,3.84,4.15 :


( √ )
𝜋 2 Ed 4GJ𝐿2
𝜎cre = 𝐼yc − 𝐼yt + 𝐼y 1 + 2 (4.21)
2𝐿2 𝑆xc 𝜋 𝐼y 𝐸𝑑 2
where 𝑆xc is the section modulus relative to the compres-
sion fiber and 𝐼yc and 𝐼yt are the moments of inertia of the
compression and tension portions of the full section, respec-
tively, about the centroidal axis parallel to the web. Other
symbols were defined previously. For equal-flange sections,
𝐼yc = 𝐼yt = 𝐼y ∕2 Eqs. (4.20) and (4.21) are identical.
For other than simply supported end conditions, Eq. (4.21)
can be generalized as given in Eq. (4.21a) as follows1.337 :
( √ )
𝜋 2 Ed 4GJ(𝐾t 𝐿t )2
𝜎cre = 𝐼yc − 𝐼yt + 𝐼y 1 + 2
2(𝐾y 𝐿y )2 𝑆xc 𝜋 𝐼y 𝐸𝑑 2
Figure 4.3 Positions of I-beam after lateral–torsional buckling.
(4.21a)
In the above equation, 𝐾𝑦 and 𝐾t are effective length
factors and 𝐿𝑦 and 𝐿t are unbraced lengths for bending about
the y axis and for twisting, respectively.
As previously discussed, in Eq. (4.21a) the second term
under the square root represents the St. Venant torsional
rigidity, which can be neglected without much loss in
economy. Therefore Eq. (4.21a) can be simplified as shown
in Eq.(4.22) by considering 𝐼y = 𝐼yc + 𝐼yt and neglecting
the term of 4GJ(𝐾t 𝐿t )2 ∕π2 𝐼y Ed2 :
Figure 4.4 Dimensions of I-beam.
𝜋 2 Ed𝐼yc
𝜎cre = (4.22)
(𝐾y 𝐿y )2 𝑆xc
Consequently the critical stress for lateral–torsional buck-
ling of an I-beam subjected to pure bending is given by Equation (4.22) was derived on the basis of a uniform
bending moment. It is rather conservative for the case of
𝑀cre 𝑀 𝑑
𝜎cre = = cre unequal end moments. For this reason it may be modified
𝑆x 2𝐼x by multiplying the right-hand side by a bending coefficient
√ 𝐶b 1.161,3.84 as given in Eq. (4.23):
√( )2 ( )(
√ )2
π2 𝐸 √ 𝐼y JI y 𝐿
= + 𝐶𝑏 𝜋 2 𝐸
2(𝐿∕𝑑)2 2𝐼x 2(1 + 𝜇)𝐼x2 𝜋d 𝜎cre = (4.23)
(𝐾𝑦 𝐿𝑦 )2 𝑆xc ∕(𝑑𝐼yc )
(4.20)
where 𝐶b is the bending coefficient, which can conserva-
where 𝑆x is the section modulus and 𝐼𝑥 is the moment tively be taken as unity. During the period from 1968 to
of inertia of the full section about the x axis. The unpub- 1996, the bending coefficient was calculated from 𝐶b =
lished data of 74 tests on lateral–torsional buckling of 1.75 + 1.05(𝑀1 ∕𝑀2 ) + 0.3(𝑀1 ∕𝑀2 )2 but must not exceed
cold-formed steel I-sections of various shapes, spans, 2.3. Here 𝑀1 is the smaller and 𝑀2 the larger bending
and loading conditions have demonstrated that Eq. (4.20) moment at the ends of the unbraced length, taken about the
applies to cold-formed steel sections with reasonable strong axis of the member. The ratio of end moments 𝑀1 ∕𝑀2
accuracy.1.161 is positive when 𝑀1 and 𝑀2 have the same sign (reverse
In Eq. (4.20) the first term under the square root represents curvature bending) and negative when they are of opposite
the strength due to lateral bending rigidity of the beam, and signs (single curvature bending).
the second term represents the St. Venant torsional rigidity. The above equation for 𝐶b was replaced by the following
For thin-walled cold-formed steel sections, the first term equation in the 1996 edition of the AISI specification and is
usually exceeds the second term considerably. retained in the North American Specification:
For simply supported I-beams with unequal flanges, the 12.5𝑀max
following equation has been derived by Winter for the elastic 𝐶b =
2.5𝑀max + 3𝑀A + 4𝑀B + 3𝑀C
BENDING STRENGTH AND DEFLECTION 123
where 𝑀max = absolute value of maximum moment in As shown in Fig. 4.5, assuming 𝐾y = 1 and 𝐿y = 𝐿, the
unbraced segment corresponding 𝐿2 𝑆xc ∕dI yc ratio for 𝜎cr = 𝜎pr is 1.8π2 ECb ∕𝐹y .
𝑀A = absolute value of moment at quarter point When the theoretical critical stress exceeds 𝜎pr , the crit-
of unbraced segment ical stress for inelastic buckling may be represented by the
𝑀B = absolute value of moment at centerline of following parabolic equation:
[ ( )]
unbraced segment 1 𝐹𝑦
𝑀C = absolute value of moment at three-quarter (𝜎cr )𝐼 = 𝐹𝑦 𝐴 − (4.27)
𝐵 𝜎cr
point of unbraced segment
where A and B are constants that can be determined by the
The above equation for bending coefficient 𝐶b was derived following conditions:
from Ref. 4.156. It can be used for various shapes of moment
diagrams within the unbraced segment and gives more accu- 1. When 𝐿 = 0,
10
rate results for fixed-end beams and moment diagrams which (𝜎cr )𝐼 = 𝐹
9 y
(4.28)
are not straight lines.
Consequently, the simplified, elastic critical moment for 2. When 𝐿2 𝑆xc ∕(dI yc ) = 1.8π2 EC𝑏 ∕𝐹𝑦 ,
lateral–torsional buckling of doubly symmetric I-beams can (𝜎cr )I = 0.56𝐹y (4.29)
be calculated from the elastic critical buckling stress given in
Eq. (4.23) and the section modulus relative to the compres- By solving Eq. (4.27), A and B are found as follows:
10
sion fiber as follows: 𝐴= 9
(4.30)
𝐶b 𝜋 2 EdI yc
𝑀cre = 𝜎cre 𝑆xc = (4.24) 𝐵 = 3.24 (4.31)
(𝐾y 𝐿y )2
Therefore Eq. (4.27) can be rewritten as
The above design formula was used in Section C3.1.2 [ ( )]
10 1 𝐹y
(b) of the 1996 edition of the AISI Specification for doubly 𝜎crI = Fy −
symmetric I-sections except that 𝐾y = 1 and 𝐿y = 𝐿. In 9 3.24 σcre
[ ( )]
the 2016 edition of the North American Specification, 10 10 𝐹y
Eq. (4.24) (or Eq. (4.23) for buckling stress) was limited = Fy − (4.32)
9 36 σcre
to doubly-symmetric I-Sections since the equation was
derived from I-Section members and was found that it may where 𝜎cre is the elastic buckling stress for lateral–
be unconservative for singly-symmetric sections.1.431 torsional buckling and (𝜎cr )I is the theoretical equation
It should be noted that Eq. (4.23) applies to elastic buck- for lateral–torsional buckling in the inelastic range.
ling of cold-formed steel beams when the computed theoret-
ical buckling stress is less than or equal to the proportional
limit σpr . However, when the computed stress exceeds the
proportional limit, the beam behavior will be governed by
inelastic buckling. For extremely short beams, the maximum
moment capacity may reach the full plastic moment 𝑀p for
compact sections. A previous study4.16 has indicated that for
wide-flange beams having an average shape factor of 10/9, 4.32
10 10
𝑀p = 𝑀 = 𝐹 𝑆 (4.25)
9 y 9 y x
4.23
where 𝑀p = full plastic moment
𝑀y = yield moment
This means that the stress in extreme fibers may reach
a hypothetical value of 10 𝐹 when 𝐿2 𝑆xc ∕dI yc ≈ 0 if the
9 y
elastic section modulus is used to compute the moment.
As in the previous design approach for compression
members (Ref. 1.4), the effective proportional limit (or the
upper limit of the elastic buckling) may be assumed to be
Figure 4.5 Maximum lateral–torsional buckling stress for
equal to one-half the maximum stress, that is, I-beams (𝐾𝑦 = 1 and 𝐿𝑦 = 𝐿).
( )
𝜎pr = 12 109
𝐹𝑦 = 0.56𝐹𝑦 (4.26)
124 4 FLEXURAL MEMBERS

Even though the maximum stress computed by Eq. (4.32) where 𝐾y , 𝐾t = effective length factors for bending about
as shown in Fig. 4.5 is larger than 𝐹y , a conservative approach the y axis and for twisting
has been used by AISI to limit the maximum stress to 𝐹y . 𝐿y , 𝐿t = unbraced length for bending about the
By using the inelastic critical buckling stress given in Eq. y√axis and for twisting
(4.32) and the section modulus relative to the compression
𝑟0 = 𝑟2𝑥 + 𝑟2𝑦 + 𝑥20
fiber, the inelastic critical moment for lateral–torsional buck-
ling of I-beams can be computed as follows: 𝑟𝑥 , 𝑟𝑦 = radii of gyration of the cross section about
the centroidal principal axes
(𝑀cr )I = (𝜎cr )I 𝑆xc ≤ 𝑀y 𝑥0 = distance from the shear center to the
[ ]
10 10 𝑀yc centroid along the principal x axis, taken as
= 1− ≤ 𝑀y (4.33) negative
9 36 𝑀cre
where 𝑀y and 𝑀yc are the yield moment and yield moment Other terms were defined previously. For singly symmetric
about compression fiber, respectively; and 𝑀cre is the elastic sections, the x axis is the axis of symmetry oriented such that
critical moment defined in Eq. (4.24). Equation (4.33) was the shear center has a negative x coordinate. The basis for
used in Section C3.1.2 of the 1996 edition of the AISI Spec- Eq. (4.34) is discussed by Pekoz in Ref. 3.17. A comparison
ification for (𝑀cr )e > 0.56𝑀y as shown in Fig. 4.6. Hill has of Eqs. (4.24) and (4.34) shows that these two equations give
demonstrated that the equations derived for I-sections can similar results for channels having 𝐼𝑥 > 𝐼𝑦 .3.17 However, for
also be used for channels with satisfactory accuracy.4.17 channel sections having 𝐼𝑥 < 𝐼𝑦 with large 𝐾𝑦 𝐿𝑦 ∕𝑟𝑦 ratios,
For cold-formed steel design, Eqs. (4.23) and (4.32) were the simplified Eq. (4.24) provides very conservative results
used in the 1968 and 1980 editions of the AISI Specification as compared with Eq. (4.34).
to develop the design equations for lateral–torsional buck- For singly symmetric sections bending about the
ling of I-beams and channels. In the 1986 and 1996 editions centroidal axis perpendicular to the symmetry axis, the
of the AISI Specification, in addition to the use of Eqs. (4.24) elastic critical moment based on the flexural–torsional
and (4.33) for determining the critical moment, new design buckling theory can be computed by using Eq. (4.37):
formulas for lateral–torsional buckling of singly and doubly √
symmetric sections bending about the symmetry axis perpen- C𝑠 𝐴𝜎ex [𝑗 + 𝐶𝑠 𝑗 2 + 𝑟20 (𝜎𝑡 ∕𝜎ex )]
dicular to the web3.17,6.11 were added: 𝑀er = (4.37)
√ 𝐶TF
𝑀cre = 𝐶𝑏 𝑟0 𝐴 𝜎ey 𝜎𝑡 (4.34)
where A is the full cross-sectional area and where 𝐶s = +1 for moment causing compression on shear
center side of centroid
𝜋2𝐸
𝜎e𝑦 = (4.35) 𝐶s = −1 for moment causing tension on shear center
(𝐾𝑦 𝐿𝑦 ∕𝑟𝑦 )2 side of centroid
[ ]
1 𝜋 2 ECw
𝜎t = 2 GJ + (4.36) 𝜋2𝐸
Ar0 (𝐾t 𝐿t )2 𝜎ex = (4.38)
(𝐾𝑥 𝐿𝑥 ∕𝑟𝑥 )2

𝐶TF = 0.6 − 0.4(𝑀1 ∕𝑀2 ), where 𝑀1 is the smaller


and 𝑀2 the larger bending moment at the ends
of the unbraced length and 𝑀1 ∕𝑀2 , the ratio
of end moments, is positive when 𝑀1 and 𝑀2
Eq. (4.33)
have the same sign (reverse curvature bending)
and negative when they are of opposite sign
(single curvature bending). When the bending
Eq. (4.24)
moment at any point within an unbraced length
is larger than that at both ends of this length
and for members subject to combined axial
load and bending moment, 𝐶TF shall be taken
as unity.
𝐾𝑥 = effective length factor for bending about the
x axis
Figure 4.6 Maximum lateral–torsional buckling moment for
𝐿𝑥 = unbraced length for bending about the 𝑥 axis
I-beams (𝐾y = 1 and 𝐿y = 𝐿).
BENDING STRENGTH AND DEFLECTION 125
and Table 4.1 Coefficients K in Eq. (4.41)3.3
( )
1
𝑗= 𝑥3 dA + xy2 dA − 𝑥0 (4.39) Loads Act At
2𝐼𝑦 ∫𝐴 ∫𝐴
GJ
𝐿2 Centroid Top Flange Bottom Flange
= 𝛽y ∕2 (see Appendix C for computation of 𝛽y ) (4.40) ECw

Other terms were defined previously. 0.4 86.4 51.3 145.6


The derivation of Eq. (4.37) is presented in Chapter 6 for 4 31.9 20.2 50.0
beam–columns. 8 25.6 17.0 38.2
It should be noted that Eqs. (4.34) and (4.37) can be used 16 21.8 15.4 30.4
only when the computed value of 𝑀cre does not exceed 24 20.3 15.0 27.2
0.56𝑀y , which is considered to be the upper limit for the 32 19.6 14.8 26.3
elastic buckling range. When the computed 𝑀cre exceeds 48 19.0 14.8 23.5
0.56𝑀y , the inelastic critical moment can be computed from 64 18.3 14.9 22.4
Eq. (4.33). The elastic and inelastic critical moments are 80 18.1 14.9 21.7
shown in Fig. 4.7. 96 17.9 15.1 21.1
The equations developed above for the uniform bending 160 17.5 15.3 20.0
moment can also be used for other loading conditions with 240 17.4 15.6 19.3
reasonable accuracy.1.161,4.18 If more accurate results are 320 17.2 15.7 18.9
desired, the theoretical critical value for a concentrated load 400 17.2 15.8 18.7
at the center of a simply supported beam can be computed
as3.3 √
EI 𝑦 GJ
𝑃cre = 𝐾 (4.41) Table 4.2 Coefficients K in Eq. (4.42)3.3
𝐿2
where K is a coefficient to be taken from Table 4.1 based Loads Act At
on the parameter GJL2 ∕ECw . For symmetrical I-sections, GJ
𝐿2 Centroid Top Flange Bottom Flange
𝐶w ≈ 𝐼y 𝑑 2 ∕4, where d is the depth of the section. ECw
For a uniformly distributed load, the critical load is
√ 0.4 143.0 92.9 222.0
EI 𝑦 GJ 4 53.0 36.3 77.3
𝑤cre = 𝐾 (4.42) 8 42.6 30.4 59.4
𝐿3
where K is to be taken from Table 4.2. 16 36.3 27.4 48.0
24 33.8 26.6 43.4
32 32.6 26.1 40.4
48 31.5 25.8 37.6
64 30.5 25.7 36.2
80 30.1 25.7 35.1
128 29.0 26.0 33.3
200 29.0 26.4 32.1
280 28.8 26.5 31.4
Eq. (4.33)
360 28.7 26.6 31.0
400 28.6 26.6 30.7

Eq. (4.24)
Eq. (4.34)
4.2.2.2 Point-Symmetric Sections Point-symmetric sec-
tions such as Z-sections with equal flanges will buckle later-
ally at lower strengths than doubly and singly symmetric
sections. A conservative design approach has been used in
Unbraced length the previous AISI Specification and is also used in the North
Figure 4.7 Elastic and inelastic critical lateral buckling moments American Specification, in which the elastic critical moment
for members bending about centroidal axis perpendicular to the is taken to be one-half of those permitted for I-beams or chan-
web. nels. Therefore instead of using Eq. (4.34), the following
126 4 FLEXURAL MEMBERS

equation is used for determining the elastic critical moment where 𝑏 = midline or centerline dimension of flange
for point-symmetric Z-sections bending about the centroidal 𝑑 = midline or centerline dimension of web
axis perpendicular to the web: 𝑡 = wall thickness
1 √ When a closed-box section is subject to a nonuniform
𝑀cre = 𝐶𝑏 𝑟0 𝐴 𝜎ey 𝜎𝑡 (4.43)
2 bending moment, the above equation for the elastic critical
In lieu of Eq. (4.43), the following simplified equation moment can be modified by a bending coefficient 𝐶b as
follows:
𝐶 𝜋√
can be used to calculate the elastic critical moment for
Z-sections: 𝑀cre = 𝑏 EI 𝑦 GJ (4.46)
𝐶𝑏 𝜋 2 EdI yc 𝐿
𝑀cre = (4.44) Consequently, the elastic critical lateral–torsional buck-
2(𝐾𝑦 𝐿𝑦 )2
ling stress 𝐹cre can be determined by Eq. (4.47):
All symbols used in Eqs. (4.43) and (4.44) are defined in 𝐶𝑏 𝜋 √
Section 4.2.2.1. 𝐹cre = EI 𝑦 GJ (4.47)
𝐾 𝑦 𝐿𝑦 𝑆 𝑓
in which 𝑆f is the elastic section modulus of the full unre-
4.2.2.3 Closed-Box Sections Closed sections such as
duced section relative to the extreme compression flange, 𝐾𝑦
box shapes have relatively larger torsional stiffness as
is the effective length factor, and 𝐿𝑦 is the unbraced length
compared with open sections such as I-beams, C-sections,
for bending about the y axis.
and Z-sections discussed in Sections 4.2.2.1 and 4.2.2.2. As
far as lateral–torsional buckling is concerned, these closed,
4.2.2.4 Lateral–Torsional Buckling with Hole Influence
double-web sections are more stable than single-web open
The lateral–torsional buckling stress, 𝜎cre or moment, 𝑀cre ,
sections, and therefore any use of closed-box sections will
discussed previously is based on beams without holes. The
result in an economical design if lateral–torsional stability
existence of holes along the member length will reduce
of the beam is essential.
the member bending rigidity, EI, and consequently reduce
In Ref. 4.18, Winter indicated that for closed-box beams
the buckling strength.3.252,3.278,4.294,4.295,6.39 The numerical
the bending strength is unaffected by lateral–torsional
shell finite element method can be used to determine the
buckling even when the length-to-width ratio is as high as
global buckling moment. However, the complication in iden-
100 for a steel having a yield stress of 33 ksi (228 MPa or
tifying the different buckling modes and software not readily
2320 kg/cm2 ).
available to regular design engineers make it not feasible in
Previous editions of the AISI Specification contained
design. The finite strip method is not applicable since the
in Section D3.3 a conservative design provision for
section properties affected by holes cannot be considered
lateral–torsional buckling of closed-box beams, in which simply by revising the thickness. Therefore, the analyt-
laterally unbraced box sections can be designed without any ical method developed by Moen and Schafer3.278,4.294,4.295
strength reduction for lateral–torsional buckling consider- is recommended. Based on the research and experiment
ation if the ratio of the unsupported length to the distance verifications,4.295 the lateral–torsional buckling moment
between the webs of the section does not exceed 0.086 of doubly- and singly-symmetric section members with
𝐸∕𝐹y . In 1999, this design requirement was replaced by patterned holes can be approximated by using the average
Section C3.1.2.2 in the Supplement to the 1996 edition of section properties in the buckling expression as developed
the AISI Specification.1.333 The same design provisions are in the previous sections:
retained in the North American Specification, except that √ ( )
the laterally unbraced length, L, was clarified to be 𝐾𝑦 𝐿𝑦 . 𝜋 𝜋2
𝑀cre = EI y,avg GJ avg + ECwnet
For a closed-box section subjected to a uniform bending 𝐾 𝑦 𝐿𝑦 (𝐾𝑡 𝐿𝑡 )2
moment as shown in Fig. 4.2, the elastic critical moment for (4.48)
lateral–torsional buckling is3.84 where 𝐼𝑦 ,avg and 𝐽avg are the average moment of inertia about
√ the y axis and average torsional constant, respectively; 𝐶wnet
𝜋
𝑀cre = EI 𝑦 GJ is the warping constant of the net section. The existence of the
𝐿
All terms are defined in Section 4.2.2.1, except that the hole creates a discontinuity that interrupts warping torsion
torsional constant J may be determined by the following resistance along the member. It is therefore a net warping
constant is used. 𝐶wnet can be determined by assuming thick-
equation for a closed-box section having a uniform
ness at the hole equals zero, it can also be determined using
thickness4.157 :
2𝑏2 𝑑 2 𝑡 software such as CUFSM by setting the thickness of the
𝐽= (4.45) element at the hole equals zero.
𝑏+𝑑
BENDING STRENGTH AND DEFLECTION 127
A complete list of the average properties of cross-section where 𝑆fy = Elastic section modulus of full unreduced cross-
is provided in Table 4.3, which can also be used to determine section relative to extreme fiber in first yielding
the global buckling loads for columns with holes later in 𝐹y = Yield stress
Chapter 5.
Fn shall be determined as follows:
The above analytical method was first adopted into the
2012 edition of the North American Specification1.416 and
For 𝐹cre ≥ 2.78Fy
is retained in 2016 edition of the Specification.1.417
The hole influence on the member strength could be
counterintuitive. The Commentary1.431 on the 2016 North 𝐹n = 𝐹y (4.51)
American Specification provided the following guidance:
“Rules of thumb on the influence of holes in both compres- For 2.78 𝐹y > 𝐹cre > 0.56𝐹y
sion and flexural members are: (1) rectangular or elongated
holes typically reduce local buckling strength more than ( )
10 10𝐹y
square and circular holes; (2) web holes always decrease 𝐹n = 𝐹y 1 − (4.52)
distortional buckling strength; (3) holes always reduce 9 36𝐹cre
global (Euler) buckling strength; (4) the more holes along a
member, the more the strength decreases; (5) hole patterns, For Fcre ≤ 0.56Fy
such as those typically present in storage rack columns,
can reduce strength as much as discrete holes; and (6) 𝐹𝑛 = 𝐹cre (4.53)
adding edge stiffeners to holes increases local buckling
strength more than distortional buckling and global buckling where 𝐹cre = Critical elastic lateral–torsional buckling stress,
strength.” determined in accordance with Specification
Sections F2.1.1 to F2.1.5, as applicable, or
Specification Appendix 2.
4.2.2.5 North American Design Criteria for
Lateral–Torsional Buckling Strength of Singly-, Doubly-,
and Point-Symmetric Sections Sections 4.2.2.1–4.2.2.4 F2.1.1 Singly or Doubly Symmetric Sections Bending
discussed how to determine the member lateral–torsional About Symmetric Axis
buckling moment and how to consider the inelastic moment The elastic buckling stress for singly or doubly symmetric
strength when the buckling moment exceeds 0.56𝑀y . sections bending about the symmetric axis shall be calculated
The North American Specification developed the design as follows:
provisions based on those expressions. The following C r A√
𝐹cre = b 0 𝜎ey 𝜎t (4.54)
excerpts are adapted from Sections F2.1 and F2.2 of the Sf
2016 edition of the North American Specification, which where
provides the needed design equations for computing the 12.5Mmax
𝐶b = (4.55)
critical lateral–torsional buckling stress.1.417 The applica- 2.5Mmax + 3MA + 4MB + 3MC
tions of the North American design criteria are illustrated in
Examples 4.1–4.3. where 𝑀max = Absolute value of maximum moment
in unbraced segment
𝑀𝐴 = Absolute value of moment at quarter point
F2.1 Initiation of Yielding Strength of unbraced segment
The nominal flexural strength [resistance], Mne , for yielding and 𝑀B = Absolute value of moment at centerline
global (lateral–torsional) buckling considering capacity up to of unbraced segment
first yield shall be calculated in accordance with Eq. (4.49). 𝑀C = Absolute value of moment at three-quarter
point of unbraced segment
𝑀ne = 𝑆f 𝐹n ≤ 𝑀y (4.49) 𝐶b = is permitted to be conservatively taken as unity
for all cases. For cantilevers or overhangs
where 𝑀ne = Nominal flexural strength [resistance] for where the free end is unbraced, 𝐶b shall be
yielding and global buckling taken as unity.
𝑆f = Elastic section modulus of full unreduced 𝑟0 = Polar radius of gyration of cross-section about
section relative to extreme compression fiber shear center

𝑀y = 𝑆fy 𝐹y (4.50) = 𝑟2𝑥 + 𝑟2𝑦 + 𝑥20 (4.56)
128 4 FLEXURAL MEMBERS

Table 4.3 Average Cross-Sectional Properties𝒂


Average Properties Formulas
𝐴g 𝐿g + 𝐴net 𝐿net
Cross-sectional area 𝐴avg =
L
𝐼g 𝐿g + 𝐼net 𝐿net
Moment of inertia about axis of buckling 𝐼avg =
L
𝐽g 𝐿g + 𝐽net 𝐿net
Saint-Venant Torsion constant 𝐽avg =
𝐿
𝑥0,𝑔 𝐿𝑔 + 𝑥0,net 𝐿net
Shear center x-coordinate relative to centroid 𝑥0,𝑎v𝑔 =
𝐿
𝑦0,𝑔 𝐿𝑔 + 𝑦0,net 𝐿net
Shear center y-coordinate relative to centroid 𝑦0,𝑎v𝑔 =
𝐿

𝐼𝑥,𝑎v𝑔 +𝐼𝑦,𝑎v𝑔
Polar radius gyration about shear center 𝑟0,𝑎v𝑔 = 𝑥20,𝑎v𝑔 + 𝑦20,𝑎v𝑔 + 𝐴𝑎v𝑔

𝑎
Definition of variables:
𝐴g , 𝐴net = Gross and net area, respectively
𝐿g = Segment length without holes
𝐿net = Length of holes or net section regions
𝐿 = Unbraced length about the axis of buckling
= 𝐿g + 𝐿net
𝐼g , 𝐼net = Moment of inertia of gross or net cross-section about the axis of buckling,
respectively
𝐽g , 𝐽net = Saint-Venant torsion constant of gross or net cross-section, respectively
𝑥0,g 𝑥0,net = Shear center x-coordinate relative to centroid for gross or net section, respectively
𝑦0 ,g , 𝑦0 ,net = Shear center y-coordinate relative to centroid for gross or net section, respectively
𝑟0,g , 𝑟0 ,net = Polar radius gyration about shear center of gross or net cross-section, respectively
𝐼xy,g , 𝐼xy,net = Product of inertia of gross or net cross-section, respectively

where 𝑟x , 𝑟y = Radii of gyration of cross-section about where 𝐺 = Shear modulus of steel


centroidal principal axes 𝐽 = Saint-Venant torsion constant
𝑥0 = Distance from centroid to shear center in of cross-section
principal x-axis direction 𝐶w = Torsional warping constant of cross-section
𝐴 = Full unreduced cross-sectional area 𝐾t = Effective length factor for twisting
𝑆f = Elastic section modulus of full unreduced 𝐿t = Unbraced length of member for twisting
cross-section relative to extreme compression
fiber For singly-symmetric sections, the x axis shall be the axis
of symmetry.
Alternatively, for doubly-symmetric I-sections, 𝐹cre is
𝜋2E
𝜎ey = (4.57) permitted to be calculated using the equation given
(Ky Ly ∕ry )2
Cb π2 EdI yc
𝐹cre = (4.59)
where 𝐸 = Modulus of elasticity of steel Sf (Ky Ly )2
𝐾𝑦 = Effective length factor for bending about y axis
where 𝑑 = Depth of section
𝐿y = Unbraced length of member for bending about 𝐼yc = Moment of inertia of compression portion
y axis of section about centroidal axis of entire section
[ ] parallel to web, using full unreduced
1 𝜋 2 𝐸𝐶w section
𝜎t = 𝐺𝐽 + (4.58)
Ar20 (Kt Lt )2
BENDING STRENGTH AND DEFLECTION 129
F2.1.2 Singly-Symmetric Sections Bending About F2.1.4 Closed-Box Sections
Centroidal Axis Perpendicular to Axis of Symmetry For closed-box section members, if the laterally unbraced length
The elastic buckling stress, 𝐹cre , for singly-symmetric sections of the member is less than or equal to 𝐿u , as calculated in Eq.
bending about the centroidal axis perpendicular to the axis of (4.66), the global buckling does not need to be considered, and
symmetry shall be calculated as follows, where the x axis is the the nominal stress, 𝐹n = 𝐹y .
symmetric axis of the cross-section oriented such that the shear
0.36𝐶b 𝜋 √
center has a negative x-coordinate: 𝐿u = EGJIy (4.66)
[ √ ] Fy Sf
C 𝐴𝜎
𝐹cre = s ex j + Cs j2 + r02 (σt ∕σex ) (4.60) where 𝐽 = Torsional constant of closed-box section
CTF Sf
𝐼y = Moment of inertia of full unreduced section about
where 𝐶s = +1 for moment causing compression on shear centroidal axis parallel to web
center side of centroid 𝐹y = Yield stress
= −1 for moment causing tension on shear center
side of centroid Other variables are defined in Specification Section F2.1.1.
If the laterally unbraced length of a member is larger than
𝜋2E Lu , as calculated in Eq. (4.66), the elastic buckling stress, 𝐹cre ,
𝜎ex = (4.61)
(Kx Lx ∕rx )2 for bending about the symmetric axis shall be calculated as
follows:
where 𝐾x = Effective length factor for bending about xaxis Cb 𝜋 √
𝐹cre = EGJIy (4.67)
𝐿x = Unbraced length of member for bending about Ky Ly Sf
x axis
F2.1.5 Other Cross-Sections
𝐶TF = 0.6 − 0.4 (M1 ∕M2 ) (4.62) For cross-sections other than those defined in Specification
Sections F2.1.1 through F2.1.4, the elastic buckling stress is
where 𝑀1 and 𝑀2 = The smaller and the larger bending
permitted to be determined in accordance with Specification
moment, respectively, at the ends of the
Section 2.2 of Appendix 2.
unbraced length in the plane of bending;
𝑀1 ∕𝑀2 , the ratio of end moments, is
positive when 𝑀1 and 𝑀2 have the F2.2 Beams With Holes
same sign (reverse curvature bending) For shapes whose cross-sections have holes, Fcre shall consider
and negative when they are of opposite the influence of holes in accordance with Specification
sign (single curvature bending). When Appendix 2.
the bending moment at any point within Exception: For the Effective Width Method, where hole sizes
an unbraced length is larger than that at meet the limitations of Specification Appendix 1.1.3, the provi-
both ends of this length, 𝐶TF shall be sions of this section shall not be required.
taken as unity
[ ] Even though the existence of holes does affect the
1 global buckling moment, North American Specification1.417
𝑗= x3 dA + xy2 dA − x0 (4.63)
2I𝑦 ∫A ∫A
provides some relief when the Effective Width Method is
where 𝑥0 = Distance from centroid to shear center in principal used and the hole dimensions are within the limitations of
x-axis direction, taken as negative Specification Appendix 1.1.3.
From Fig. 4.7, it can be seen that when a member’s
Other variables are defined in Specification Section F2.1.1. unbraced length is less than length, 𝐿u , the global buckling
F2.1.3 Point-Symmetric Sections
does not occur. 𝐿u can be derived from the stress changes
The elastic buckling stress, 𝐹cre , for point-symmetric Z-sections between Eqs. (4.51) and (4.52) by setting 𝐹cre = 2.78𝐹y . The
bending about x axis that is perpendicular web and through the following equations are given in Section 1.4 of Part II of the
centroid is permitted to be calculated as follows: AISI Cold-Formed Steel Design Manual1.428 for computing
C r A√ 𝐿u :
𝐹cre = b 0 σey σt (4.64)
2Sf
(a) For singly, doubly, and point-symmetric sections,
Alternatively, 𝐹cre is permitted to be calculated using
0.5
Eq. (4.65): ⎧ [ ( )2 ]0.5 ⎫
Cb π2 EdI 𝑦𝑐 ⎪ GJ 𝐶2 GJ ⎪
𝐹cre = (4.65) 𝐿𝑢 = ⎨ + + ⎬ (4.68)
2Sf (K𝑦 L𝑦 )2 ⎪ 2𝐶1 𝐶1 2𝐶1 ⎪
⎩ ⎭
Variables are defined in Specification Section F2.1.1.
130 4 FLEXURAL MEMBERS

where SOLUTION
1. For singly- and doubly-symmetric sections, As shown in Fig. 4.7, when 𝐾y 𝐿y ≤ 𝐿u , 𝑀ne = 𝑀y (or 𝐹n =
( )2 𝐹y ). The elastic critical lateral–torsional buckling stress of
7.72 𝐾𝑦 𝐹𝑦 𝑆f 𝜋 2 ECw
𝐶1 = C2 = the I-section is determined according to Section F2.1 of the
AE 𝐶b 𝜋𝑟𝑦 (𝐾t )2
North American Specification. For 𝐹n = 𝐹y ,
2. For point-symmetric sections,
𝐹cre ≥ 2.78𝐹y
( )2
30.9 𝐾𝑦 𝐹𝑦 𝑆𝑓 𝜋 2 EC𝑤
𝐶1 = C2 = Substituting Eq. (4.59) for 𝐹cre into the above expression
AE 𝐶𝑏 𝜋𝑟𝑦 (𝐾𝑡 )2
yields
(b) For I-, C-, or Z-sections bent about the centroidal axis 𝐶𝑏 𝜋 2 EdI yc
≥ 2.78𝐹𝑦
perpendicular to the web (x axis), in lieu of (a), the following 𝑆𝑓 (𝐾𝑦 𝐿𝑦 )2
equations may be used:
Therefore,
1. For doubly-symmetric I-sections: √
( )0.5 𝐾 y 𝐿y ≤ 𝐶b 𝜋 2 EdI yc ∕(2.78𝐹y 𝑆f )
0.36𝐶𝑏 𝜋 2 EdI yc
𝐿𝑢 = (4.69a)
𝐹𝑦 𝑆𝑓 (𝐾𝑦 )2
where 𝐸 = 29.5 × 103 ksi
2. For point-symmetric Z-sections: 𝐶b = 1.0 (assumed value)
( )0.5 𝐷 = 8.0 in.
36𝐶𝑏 𝜋 2 EdI yc
𝐿𝑢 = (4.69b) 𝐹𝑦 = 50 ksi
𝐹𝑦 𝑆𝑓 (𝐾𝑦 )2 𝐾y = 1.0
3. For closed-box sections: 𝐼yc = 0.724 in.4 (see the following calculation)
0.36𝐶b 𝜋 √ The section properties are calculated using the center-line
𝐿u = EGJI𝑦 (4.70)
𝐹y 𝑆 f dimensions: corner radius r = 3/16 in.+ t/2 = 0.255 in.,
flange flat width b = 2 in. − (𝑟 + 𝑡∕2) = 1.6775 in.; flat
web depth 𝑎 = 8 in –2(𝑟 + 𝑡∕2) = 7.355 in. Corner length =
In addition, Part II of the Design Manual provides beam
1.57𝑟 = 0.400 in.
design charts for determining the nominal flexural strengths
Calculation of A, Iyc , Sf :
of C-sections and Z-sections with lips. These charts were
prepared for 𝐹y = 33, 50 and 55 ksi (228, 345, and 379 MPa; Distance from y
2.32, 3.52, and 3.87 × 103 kg∕cm2 ) with Cb = 1.0. The Element Area Ai (in.2 ) Axis, xci (in.) Ai xci 𝟐 (in.4 )
torsional unbraced length (𝐾t 𝐿t ) is assumed to be equal to
Flanges 4(1.6775)(0.135) = 0.9059 2–b∕2 = 1.1613 1.2215
the unbraced length about the y axis (𝐾y 𝐿y ). Corners 4(0.400)(0.135) = 0.2162 0.363r + t∕2 0.0055
= 0.160
Example 4.1 An I-beam with cross-section shown Webs 2(7.355)(0.135) = 1.9859 𝑡∕2 = 0.0675 0.0090
in Fig. 4.8 is used as a simply supported beam with Total 3.108 1.2360
a span length of 10 ft to support a uniform load (see Iflanges = 4 × 1
0.135(1.6775)3 = 0.2124
12
Fig. 4.8). Determine what is the maximum unbraced length I𝑦 = 1.4484 in.4
such that the beam will not subject to lateral torsional
𝐼yc = 12 𝐼y = 0.724 in4
buckling.
𝑆𝑓 = 6.54 in.3 (see the following calculation)

Distance from
Element Area Ai (in.𝟐 ) Mid depth yi (in.) Ai y𝟐i (in.𝟒 )

Flanges 4(1.6775)(0.135) = 0.9059 3.9325 14.0093


Corners 4(0.05407) = 0.2163 3.8436 3.1955
Webs 2(7.355)(0.135) = 1.9859 0 0
Total 3.1081 17.2048
1
2𝐼web = 2 × 12
× (0.135)(7.355)3 8.9522
𝐼𝑥 = 26.1570 in.4
𝐼𝑥
Figure 4.8 Example 4.1. 𝑆f = 8∕2
= 6.54 in.3
BENDING STRENGTH AND DEFLECTION 131
The maximum unbraced length between lateral supports is

(1)𝜋 2 (29,500)(8)(0.724)
𝐿y = = 43.1 in.
2.78(50)(6.54)
Alternatively, the above unbraced length can be calculated
directly from Eq. (4.69a). Actually, the beam may be braced
laterally at one-third span length with an unbraced length
of 40 in., as shown in Fig. 4.9. For segment CD, 𝐶b = 1.01,
which is practically the same as the assumed value of 1.0.

Figure 4.10 Example 4.2.

12.5(wL2 ∕8)
𝐶b =
2.5(wL2 ∕8) + 3(7wL2 ∕128)
+ 4(12wL2 ∕128) + 3(15wL2 ∕128)
= 1.30
Using Eq. (4.59), with 𝐾𝑦 = 1.0,
Figure 4.9 Lateral supports.
𝐶𝑏 𝜋 2 EdI yc
𝐹cre =
Example 4.2 Determine the allowable uniform load 𝑆𝑓 (𝐾𝑦 𝐿𝑦 )2
considering the lateral–torsional buckling only if the I-beam
(1.30)𝜋 2 (29, 500)(8)(0.724)
used in Example 4.1 is braced laterally at both ends and = = 93.11 ksi
midspan. See Fig. 4.10. Use the value of 𝐶b determined (6.54)(5 × 12)2
by the formula included in the North American Specifica- 0.56𝐹y = 28.00 ksi
tion and 𝐹𝑦 = 50 ksi. Use the ASD method and the LRFD 2.78𝐹y = 139.00 ksi
method with an assumed dead load–live load ratio 𝐷∕𝐿 = 15 . Since 2.78 𝐹y > 𝐹cre > 0.56𝐹y , from Eq. (4.52)
( )
10 10𝐹𝑦
𝐹𝑛 = 𝐹 1−
SOLUTION 9 𝑦 36𝐹cre
A. ASD Method ( )
10 10(50)
1. Nominal Moment for Lateral–Torsional Buckling = (50) 1 − = 47.27 ksi
9 36(93.11)
Strength. From Example 4.1, 𝑆𝑓 = 6.54 in.3 and
Based on Eq. (4.49), the nominal moment for lateral–
𝐼yc = 0.724 in.4 Considering the lateral supports at
torsional buckling strength is
both ends and midspan and the moment diagram shown
in Fig 4.10, the bending coefficient 𝐶b for segment AB 𝑀ne = Sf 𝐹n = (6.54)(47.27) = 309.15 in.-kips
can be calculated by using Eq. (4.55) as follows:
12.5𝑀max 3. Allowable Uniform Load. The allowable moment based
𝐶b = on lateral–torsional buckling strength is calculated as
2.5𝑀max + 3𝑀1 + 4𝑀2 + 3𝑀3
follows:
where 𝑀max = wL2 ∕8 at midspan
𝑀1 = 7wL2 ∕128 at 1∕4 point of unbraced 𝑀ne 309.15
segment (𝑀𝑎 ) = = = 185.12 in.-kips
Ω𝑏 1.67
𝑀2 = 12wL2 ∕128 at midpoint of unbraced
segment The maximum moment at midspan is wL2 /8 ft-kips:
𝑀3 = 15wL2 ∕128 at 3∕4 point of unbraced 1 2
segment wL (12) = 185.12 in.-kips
8
132 4 FLEXURAL MEMBERS

Then the allowable uniform load is The allowable moment is


𝑤 = 1.234 kips∕ft 𝑀a = 𝑀D + 𝑀L = 30.24 + 151.22 = 181.46 in.-kips
It should be noted that the allowable load computed above The allowable uniform load can be calculated as
is based on the consideration of lateral–torsional buckling. It follows:
should also be checked for bending due to local and distor- 1 2
tional buckling. In addition, shear, web crippling, deflection, wL (12) = 181.46 in.-kips w = 1.21 kips∕ft
8
and other requirements should be checked, as applicable.
It can be seen that the allowable uniform load computed
B. LRFD Method on the basis of the LRFD method is similar to that computed
Using the same method employed above for the ASD
from the ASD method. The difference is only about 2%.
method, the governing nominal moment for lateral–torsional
buckling strength is
𝑀ne = 309.15 in.-kips Example 4.3 For the singly symmetric channel section
(8 × 2 × 0.06 in.) shown in Fig. 4.11, determine the nominal
The design moment is moment, 𝑀ne , for lateral–torsional buckling strength
𝜙𝑏 𝑀ne = 0.90(309.15) = 278.24 in.-kips according to Section F2.1 of the 2016 edition of the North
American Specification. Assume that the channel is used as
Based on the load combination of Eq. (1.5a), the required
a simply supported beam to support a concentrated load at
moment is
midspan and lateral supports are located at one-fourth of the
𝑀u = 1.4𝑀D
span length. Use 𝐹y = 33 ksi, 𝐾𝑦 𝐿𝑦 = 𝐾t 𝐿t = 2.5 ft.
where MD is the bending moment due to dead load. Similarly,
based on the load combination of Eq. (1.5b), the required SOLUTION
moment is
𝑀u = 1.2𝑀D + 1.6𝑀L = 1.2𝑀D + 1.6(5𝑀D ) = 9.2𝑀D 1. Sectional Properties. By using the design formulas
given in Part I of the AISI Design Manual,1.428 the
where 𝑀L is the bending moment due to live load. A compar-
following full-section properties can be calculated:
ison of the above computations indicates that for a given
member the load combination of Eq. (1.5b) allows a smaller 𝐴 = 0.706 in.2 𝑥0 = −0.929 in.
moment 𝑀D than the load combination of Eq. (1.5a). There-
𝑆f = 1.532 in. 3
𝑟0 = 3.14 in.
fore, the bending moment 𝑀D can be computed from 𝑀u =
𝜙b 𝑀ne as follows: 𝑟𝑥 = 2.945 in. 𝐽 = 0.000848 in.4
9.2𝑀D = 278.24 in.-kips 𝑟𝑦 = 0.569 in. 𝐶w = 2.66 in.6
Therefore, 2. Elastic Critical Lateral–Torsional Buckling
278.24 Stress, Fcre . Because the given singly symmetric
𝑀𝐷 = = 30.24 in.-kips
9.2 channel section is subject to a moment bending
ML = 5𝑀𝐷 = 151.22 in.-kips about the symmetry axis (x axis), the elastic critical

Figure 4.11 Example 4.3.


BENDING STRENGTH AND DEFLECTION 133
lateral–torsional buckling stress can be determined It should be noted that 𝑀ne is the nominal moment
according to Eq. (4.54) as follows: based on lateral–torsional buckling. The member
𝐶 𝑟 𝐴√ design strength should also consider the local buckling
𝐹cre = b 0 𝜎e𝑦 𝜎t and distortional buckling, as applicable.
𝑆f
For two central segments, the value of 𝐶b can be
4.2.2.6 Inelastic Reserve Strength The inelastic reserve
computed from Eq. (4.55) as follows:
strength in considering yielding and global buckling is shown
12.5𝑀max in Fig. 3.68b in the region with 𝑀n ∕𝑀y surpassing 1, where a
𝐶b =
2.5𝑀max + 3𝑀1 + 4𝑀2 + 3𝑀3 linear function between 𝑀p and 𝑀y is used. To determine the
where 𝑀max = PL∕4 at midspan inelastic reserve strength, the following design provisions,
𝑀1 = 5PL∕32 at 1/4 point of unbraced developed based on the research work by Shifferaw and
segment Schafer,4.296 were adopted in the 2012 edition of the North
𝑀2 = 6PL∕32 at midpoint of unbraced American Specification, and is retained in the 2016 edition
segment of the Specification:
𝑀3 = 7PL∕32 at 34 point of unbraced
segment F2.4.2 Direct Strength Method
The nominal strength (resistance), Mne , considering inelastic
Then flexural reserve capacity is permitted to be considered in accor-
dance with the provisions of this section:
12.5(PL∕4) For 𝑀cre > 2.78𝑀y
𝐶b = = 1.25
2.5(PL∕4) + 3(5PL∕32) √
+ 4(6PL∕32) + 3(7PL∕32) My ∕Mcre − 0.23
𝜋2E 𝜋 2 (29, 500) Mne = Mp − (Mp − My ) ≤ Mp (4.71)
𝜎ey = = 0.37
(Ky Ly ∕ry )2 (2.5 × 12∕0.569)2 where
= 104.73 ksi 𝑀cre = Critical elastic lateral–torsional buckling moment
[ ]
1 𝜋 2 ECw = 𝑆𝑓 𝐹cre (4.72)
𝜎t = GJ +
A𝑟20 (Kt Lt )2
[ where 𝑆𝑓 = Elastic section modulus of full unreduced
1 cross-section relative to extreme
= (11, 300)(0.000848)
(0.706)(3.14)2 compression fiber
] 𝐹cre = Critical elastic lateral–torsional buckling
𝜋 2 (29, 500)(2.66)
+ stress, determined in accordance with
(2.5 × 12)2
Specification Appendix 2 or
= 125.0 ksi Section F2.1
Therefore, the elastic critical lateral–torsional buck- 𝑀𝑦 = Member yield moment in accordance with
ling stress is Specification Section F2.1
(1.25)(3.14)(0.706) √
Fcre = (104.74)(125.0) 𝑀p = Member plastic moment
(1.532)
= 206.96 ksi = Zf Fy (4.73)
3. Critical Lateral–Torsional Buckling Stress, Fn .
0.56𝐹y = 18.48 ksi 2.78𝐹y = 91.74 ksi where 𝑍f = Plastic section modulus
𝐹y = Yield stress
Since 𝐹cre > 2.78𝐹y , the member segment is not
subject to lateral–torsional buckling. 𝐹n = Fy
4. Nominal Moment 𝑀ne . According to Eq. (4.49) the 4.2.3 Local Buckling Interacting With Yielding
nominal moment based on lateral–torsional buckling and Global Buckling
is: For members with larger w/t ratio, the local buckling inter-
𝑀ne = 𝑆f 𝐹n = (1.532)(33) = 50.56 in.-kips acting with the lateral–torsional buckling will result in
134 4 FLEXURAL MEMBERS

reduced member strength. Both the Effective Width Method the neutral axis is located at the middepth of the section,
and the Direct Strength Method can be used to determine as shown in Fig. 4.1a.
the reduced strength due to local buckling. 2. If the neutral axis is closer to the compression than
For members with small w/t ratio, like hot-rolled steel, to the tension flange, as shown in Fig. 4.1(b), the
it is possible for member cross-section to reach yielding in maximum stress of 𝐹y occurs in the tension flange.
compression and/or tension sides and the inelastic reserve The stress in the compression flange depends on the
capacity may be considered. location of the neutral axis, which is determined by
Both of above two situations will be considered by the the effective area of the section. The latter cannot be
Effective Width Method and the Direct Strength Method in determined unless the compressive stress is known.
the following subsections. The closed-form solution of this type of design is
possible but would be a very tedious and complex
4.2.3.1 Effective Width Method The Effective Width procedure. It is therefore customary to determine
Method considers the local buckling interacting with the the sectional properties of the section by successive
beam lateral–torsional buckling by reducing the section approximation.
modulus, Sf , in global buckling Eq. (4.49) to the effective
section modulus, Se , which is calculated based on stress The calculation of the nominal moment on the basis of
level Fn as determined in Specification Section F2.1. The initiation of yielding and the determination of the design
effective width of the compression flange and the effective moment are illustrated in Examples 4.4–4.7.
depth of the web can be computed from the design equations
presented in Chapter 3.
Therefore, the nominal strength for local buckling inter- 4.2.3.1.1 Members with Holes For members with a hole,
acting with the global buckling is calculated by using the effective widths of the flat elements adjacent to the hole
Eq. (4.74): are treated as unstiffened elements. Other elements are deter-
𝑀n𝓁 = 𝑆e 𝐹n ≤ 𝑆et 𝐹y (4.74) mined as described in Specification Appendix 1 or Chapter
3 of this book. Note, if the holes in a member are within the
limitations of Specification Section 1.1.3, the Specification
where 𝐹n = nominal stress considering lateral–torsional permits that the lateral–torsional buckling stress, 𝐹n , being
buckling per Eqs. (4.51) to (4.53) calculated based on gross cross-section, i.e., the hole influ-
𝑆e = elastic section modulus of effective section ence can be ignored in considering global buckling.
calculated with compression fiber at 𝐹n
𝑆e = elastic section modulus of effective section
with respect to the extreme compression Example 4.4 Determine the local buckling strength
fiber interacting with the lateral–torsional buckling using the
𝑆et = elastic section modulus of effective section Effective Width Method. Use the ASD and LRFD methods
calculated with tension fiber at 𝐹y to check the adequacy of the I-section with an unstiff-
ened compression flange as shown in Fig. 4.12. Assume
As discussed in Section 4.2.2, when the neutral axis the beam is simply supported with ends and midspan
is close to the compression flange, it is possible that the lateral bracing as shown in Fig. 4.10. The dead-load
yielding initiating from the tension flange controls the moment 𝑀D = 30 in.-kips and the live-load moment
design. Equation (4.74) therefore requires that the nominal 𝑀L = 140 in.-kips.
strength 𝑀n𝓁 be less than 𝑆et 𝐹y .
In cold-formed steel design, the effective section modulus,
𝑆e or 𝑆et , is usually computed by using one of the following
two cases:

1. If the neutral axis is closer to the tension than to


the compression flange, as shown in Fig. 4.1c, the
maximum stress occurs in the compression flange, and
therefore the plate slenderness factor 𝜆 and the effec-
tive width of the compression flange are determined by
the w/t ratio and 𝑓 = 𝐹n in Eq. (3.35). Of course, this
procedure is also applicable to those beams for which Figure 4.12 Example 4.4.
BENDING STRENGTH AND DEFLECTION 135
SOLUTION OK according to section 3.2.
A. ASD Method Since the compression flange is an unstiffened
1. Lateral–torsional buckling strength. From Example element and the neutral axis is either at middepth or
4.2, the nominal stress and nominal moment due to closer to the tension flange, use Eqs. (3.32)–(3.35)
lateral–torsional buckling are 𝐹n = 47.27 ksi, and with 𝑘 = 0.43 and 𝑓 = 𝐹n = 47.27 ksi. Therefore
𝑀ne = 309.15 in.-kips, respectively.
( )√𝑓
2. Local buckling strength interacting with the lateral– 1.052 𝑤
𝜆= √
torsional buckling. The nominal strength considering
𝑘 𝑡 𝐸
the local buckling can be determined per Section √
4.2.3.1. From Eq. (4.74), 1.052 47.27
=√ (12.426) [eq. (3.35)]
0.43 29,500
𝑀𝑛𝓁 = 𝑆𝑒 𝐹𝑛 ≤ 𝑆et 𝐹𝑦
= 0.798 > 0.673
The effective section modulus is calculated based on
the stress Fn , as shown below:
(a) Calculation of Sectional Properties. The 1 − 0.22∕𝜆
sectional properties of the corner element can 𝜌=
𝜆
be obtained from Table 4.4 or approximated using 1–0.22∕0.798
the linear method (see Fig. 1.32). For 𝑅 = 16 3
in. = = 0.907 [Eq. (3.34)]
0.798
and 𝑡 = 0.135 in.,
𝑏 = 𝜌w = (0.907)(1.6775)
𝐼𝑥 = 𝐼𝑦 = 0.0003889 in.4 𝐴 = 0.05407 in.2
= 1.523 in.
𝑥 = 𝑦 = 0.1564 in.
By using the effective width of the compression
For the unstiffened flange, flange and assuming the web is fully effective, the
4.000 location of the neutral axis, the moment of inertia
𝑤= − (𝑅 + 𝑡) = 1.6775 in.
2 𝐼𝑥 and the elastic section modulus of the effective
𝑤 section 𝑆e can be computed as shown in table on the
= 1.6775∕0.135 = 12.426
𝑡 following page.

Table 4.4 One 90∘ Corner, Dimensions and Properties

Dimensions Properties
Thickness t Inside Radius Moment of Inertia Centroid Coordinates Area A Blank Width
(in.) R (in.) 𝐼𝑥 = 𝐼𝑦 (in.4 ) 𝑥 = 𝑦(in.) (in.2 ) (in.)
0.135 0.1875 0.0003889 0.1564 0.05407 0.3652
0.105 0.1875 0.0002408 0.1373 0.03958 0.3495
0.075 0.0938 0.0000301 0.0829 0.01546 0.1865
0.060 0.0938 0.0000193 0.0734 0.01166 0.1787
0.048 0.0938 0.0000128 0.0658 0.00888 0.1724
0.036 0.0625 0.00000313 0.0464 0.00452 0.1170
Notes:
(1) Stock width of blank taken at t/3 distance from inner surface.
(2) 1 in. = 25.4 mm.
136 4 FLEXURAL MEMBERS

𝑘 = 4 + 2(1 + 𝜓)3 + 2(1 + 𝜓) [Eq. (3.46)]


Distance
from Top = 4 + 2(1 + 0.971) + 2(1 + 0.971) = 23.26
3
Fiber
Element Area 𝐴(in.2 ) y (in.) Ay(in.3 ) Ay2 (in.4 ) From Fig. 4.12,
Top flange 2(1.523)(0.135) = 0.4111 0.0675 0.0278 0.0019 ℎo = out-to-out depth of web
Top 2(0.05407) = 0.1081 0.1564 0.0169 0.0026
corners
= 8.00 in.
Webs 2(7.355)(0.135) = 1.9859 4.0000 7.9434 31.7746 𝑏o = out-to-out width of the compression
Bottom 2(0.05407) = 0.1081 7.8436 0.8479 6.6506
corners flange of each channel = 2.00 in.
Bottom 2(1.6775)(0.135) = 0.4529 7.9325 3.5928 28.500
flange Since ℎo ∕𝑏o = 4, then use Eq. (3.47a),
Total 3.0661 12.4288 66.930 𝑏𝑒

(Ay)
𝑏1 =
𝑦cg = ∑ =
12.4288
= 4.054 in. 3+𝜓
𝐴 3.0661
where 𝑏e is the effective width of the web deter-
mined in accordance with Eqs. (3.32)–(3.35) with
Since 𝑦cg > 𝑑∕2 = 4.00 in., initial yield occurs f1 substitued for f and 𝑘 = 23.26 as follows:
in the compression flange. Prior to computing the ℎ = 7.355 in.
moment of inertia, check the web for full effec-
tiveness by using Fig. 4.13 and Section 3.3.1.2 as ℎ 7.355
= = 54.48 < 200
follows: 𝑡 0.135
OK according to Section 3.2 [Eq. (3.44)]

( ) 1.052 43.51
𝑓1 = 47.27
3.7315
= 43.51 ksi (compression) 𝜆= √ (54.48) = 0.456 < 0.673
4.054 23.26 29,500
( )
3.6235
𝑓2 = 47.27 = 42.25 ksi (tension) 𝜌=1 [Eq. (3.33)]
4.054
| 𝑓 | | 42.25 | 𝑏𝑒 = ℎ = 7.355 in.
𝜓 = || 2 || = || | = 0.971
|
| 𝑓1 | | 43.51 |

47.27

3.7315” 4.054”

3.6235” 3.946”

Figure 4.13 Stress distribution in webs.


BENDING STRENGTH AND DEFLECTION 137
𝑏𝑒 B. LRFD Method
𝑏1 = [Eq. (3.47a)]
3+𝜓 1. Nominal and Design Moments. The nominal moment
7.355 for the LRFD method is the same as that used for the
= = 1.852 in. ASD method, that is,
3 + 0.971
Since 𝜓 > 0.236, 𝑀𝑛𝓁 = 297.23 in.-kips
The design moment is
1
𝑏2 = 𝑏𝑒 = 3.6775 in.
2 𝜙b 𝑀𝑛𝓁 = 0.9(297.23) = 267.51 in.-kips
b1 = 𝑏2 = 1.852 + 3.6775 = 5.5295 in. [Eq. (3.47b)]
2. Required Moment. According to the load factors and
Since 𝑏1 + 𝑏2 is greater than the compression the load combinations discussed in Section 1.3.3.3, the
portion of the web of 3.7315 in., the web is fully required moment for the given dead-load moment and
effective as assumed. Since the neutral axis (ycg = live-load moment can be computed as follows:
4.054 in.) is very close to the assumed location 𝑀u1 = 1.4𝐷
(𝑦cg = 4.00 in), no iteration is necessary. The total
𝐼𝑥 is determined as = 1.4(30) = 42.00 in. − kips

(Ay2 ) = 66.930 𝑀u2 = 1.2𝐷 + 1.6𝐿
( ) = 1.2(30) + 1.6(150)
1
2Iweb = 1 (0.135)(7.355)3 = 8.9522
12 = 276.00 in. − kips ⇐ controls
(∑ )
− A (y2cg ) = −(3.0661)(4.054)2 = −50.391 Since 𝑀u < 𝜙b 𝑀n , the I-section is also adequate for the
LRFD method.
Ix = 25.491 in.4
The elastic section modulus relative to the top
fiber is Example 4.5 For the C-section with an edge stiffener as
𝐼 shown in Fig. 4.14, determine the allowable moment (𝑀a )
25.4912
𝑆e = x = = 6.288 in.3 about the x axis for the ASD method and the design moment
𝑦cg 4.054
(𝜙b 𝑀n ) for the LRFD method. Assume that the yield stress
3. Nominal and Allowable Moments. The nominal
moment considering the local buckling interacting with
the lateral–torsional buckling is
𝑀𝑛𝓁 = 𝑆e 𝐹n = (6.288)(47.27) = 297.23 in.-kips
Compared with the nominal moment due to yielding
and lateral–torsional buckling (𝑀ne = 309.15 in.kips)
calculated in Example 4.2, the nominal moment due
to local buckling interacting with the global buckling
controls.
Therefore, the allowable moment is
𝑀 297.23
𝑀𝑎 = 𝑛𝓁 = = 178.0 in.-kips
Ω𝑏 1.67
4. Required Moment. Based on the ASD load combina-
tion discussed in Section 1.3.1.2, the required moment
for the given dead-load moment and live-load moment
is computed as follows:
𝑀 = 𝑀D + 𝑀L = 30 + 140 = 170 in.-kips
Since 𝑀 < 𝑀a , the I-section is adequate for the ASD
method. Figure 4.14 Example 4.5 (same as Fig 3.50).
138 4 FLEXURAL MEMBERS

of steel is 50 ksi and that lateral bracing is adequately Arc length:


provided. Determine the nominal moment due to local buck- 𝐿 = 1.57𝑅′ = 1.57(0.1313) = 0.2063 in.
ling with initiation of yielding. The linear method can be used
to determine the sectional properties. 𝑐 = 0.637𝑅′ = 0.637(0.1313) = 0.0836 in.
ii. Effective Width of the Compression Flange. For
SOLUTION the given C-section with equal flanges, the neutral
A. ASD Method axis is located either at the middepth or closer to
1. Calculation of Sectional Properties. In order to the tension flange. Therefore, use 𝑓 = 𝐹y = 50 ksi
simplify the calculation, line elements, as shown in to compute the effective width of the compres-
sion flange according to Section 3.3.3.1a. For the
Fig. 4.15a, are used for the linear method.
compression flange,
i. Corner Element. (Figs. 1.32 and 4.15a)
𝑤 = 3.50 − 2(𝑅 + 𝑡) = 3.1624 in.
1
𝑅′ = 𝑅 + 𝑡 = 0.131 in. 𝑤 3.1624
2 = = 42.17
𝑡 0.075
From Eq. (3.72)
√ √
𝐸 29,500
𝑆 = 1.28 = 1.28 = 31.09
𝑓 50
0.328𝑆 = 10.20
Since w/t > 0.328 S, use Eq. (3.73) to compute
the required moment of inertia of the edge stiffener
𝐼a as follows:
( )3
𝑤∕𝑡
𝐼a = 399𝑡4 − 0.328
𝑆
( )3
42.17
= 399(0.075)4 − 0.328
31.09
= 13.72 × 10−3 in.4
The above computed value should not exceed the
following value:
( )
𝑤∕𝑡
𝐼a ≤ 𝑡 115
4
+5
𝑆
(a) ( )
42.17
= (0.075)4 115 +5
31.09
= 5.093 × 10−3 in.4
Therefore, use 𝐼a = 5.093 × 10−3 in.4 For the
simple lip edge stiffener used for the given channel
section,
𝐷 = 0.720 in.
𝑑 = 𝐷 − (𝑅 + 𝑡) = 0.5512 in.
𝑑 0.5512
= = 7.35
𝑡 0.075
By using Eq. (3.75), the moment of inertia of the
(b)
full edge stiffener is
Figure 4.15 (a) Line elements. (b) Compression stresses f1 and f2 . 𝑑3𝑡
𝐼s = = 1.047 × 10−3 in.4
12
BENDING STRENGTH AND DEFLECTION 139
From Eq. (3.74), In the above equations, the compression stresses
𝐼s 𝑓1 and 𝑓2 (Fig 4.15b) are calculated on the basis of
𝑅I = = 0.206 < 1.0 OK the gross section as follows:
𝐼a ( )
4.8312
The effective width b of the compression flange 𝑓1 = 50 = 48.312 ksi
5.0
can be calculated as follows: ( )
4.28
𝐷 0.72 𝑓2 = 50 = 42.80 ksi
= = 0.228 5.0
𝑤 3.1624 Therefore
From Eq. (3.76), | 42.80 |
𝜓 = || | = 0.886
|
𝑤∕𝑡 42.17 | 48.312 |
𝑛 = 0.582 − = 0.582 −
4𝑆 4(31.09) and
0.578
= 0.243 <
1 𝑘= = 0.471
3 0.886 + 0.34
The k value of 0.471 calculated above for the
Use 𝑛 = 13 edge stiffener under the stress gradient is slightly
Since 𝐷∕𝑤 < 0.25 and θ = 90∘ for the simple larger than the k value of 0.43 for unstiffened
lip edge stiffened, from Table 3.3 elements under uniform compression.
𝑘 = 3.57(𝑅I )𝑛 + 0.43 The effective width of the edge stiffener can be
determined as follows:
= 3.57(0.206)1∕3 + 0.43 = 2.54 < 4.0 OK
𝑑 𝐷 − (𝑅 + 𝑡)
= = 7.35
Use 𝑘 = 2.54 to calculate the plate slen- 𝑡 𝑡
derness factor for the compression flange as 𝑓 = 𝑓1 = 48.312 ksi
( )
follows: ( )√𝑓
( ) 𝑑
( )√𝑓
1.052
𝜆= √
1.052 𝑤 𝑘 𝑡 𝐸
𝜆= √
𝑘 𝑡 𝐸 ( ) √
( ) √ = √
1.052
(7.35)
48.312
1.052 50 29,500
= √ (41.17) = 1.146 > 0.673 0.471
2.54 29,500
= 0.456 < 0.673
The effective width of the compression 𝜌 = 1.0
flange is
( ) The effective width of the edge stiffener is
1 − 0.22∕𝜆
𝑏 = 𝜌w = 𝑤 𝑑s ′ = 𝑑 = 0.551 in.
𝜆
( )
1 − 0.22∕1.146 The reduced effective width of the edge
= (3.1624) = 2.230 in.
1.146 stiffener is
From Eqs. (3.69) and (3.70), 𝑑s = 𝑑s ′ (𝑅I ) = 0.551(0.206) = 0.113 in.
( )
1 1 The above calculation indicates that
𝑏1 = bR𝐼 = (2.230)(0.206) = 0.230 in.
2 2 the compression stiffener is not fully
𝑏2 = 𝑏 − 𝑏1 = 1.146 − 0.230 = 2.00 in. effective.
iii. Reduced Effective Width of the Edge Stiffener.
iv. Location of Neutral Axis and Computation of 𝐼x
The effective width of the edge stiffener under
and Sx .
a stress gradient can be determined according to
a. Location of Neutral Axis. Assuming that the
Section 3.3.2.2. From Eq. (3.56), web element (element 7 in Fig. 4.16) is fully
0.578 |f | effective, the neutral axis can be located by
𝑘= where 𝜓 = || 2 ||
𝜓 + 0.34 | f1 | using the following table. See Fig. 4.16 for
dimensions of elements.
140 4 FLEXURAL MEMBERS

| f | | 41.05 |
𝜓 = || 2 || = || | = 0.847
|
| f1 | | 48.44 |
k = 4 + 2(1 + 𝜓)3 + 2(1 + 𝜓) = 20.30

From Fig 4.14, the out-to-out web depth h0 =


10.00 in. and the out-to-out compression flange
width b0 = 3.50 in. Since h0 ∕b0 = 10.00∕3.50
= 2.86 < 4, use Eq. (3.47a),
𝑏𝑒
𝑏1 =
3+𝜓
where 𝑏e is the effective width of the web deter-
mined in accordance with Eqs. (3.32)–(3.35)
with 𝑓1 substitued for f and 𝑘 = 20.30 as
follows:
ℎ 9.6624
= = 128.83 < 200 OK
𝑡 0.075

1.052 48.44
𝜆= √ (128.83)
20.30 29,500
Figure 4.16 Effective lengths and stress distribution using fully
effective web. = 1.219 > 0.673
1 − 0.22∕𝜆
𝜌= = 0.672
𝜆
𝑏𝑒 = 𝜌h = (0.672)(9.6624) = 6.4931 in.
𝑏𝑒 6.4931
Distance 𝑏1 = = = 1.6878 in.
Effective from Top 3+𝜓 3 + 0.847
Element Length L (in.) Fiber y (in.) Ly(in.2 ) Because 𝜓 > 0.236, Eq. (3.47b) is used to
compute b2 :
1 0.5512 9.5556 5.2670
2 2(0.206) = 0.4120 9.9148 4.0849 1
𝑏2 = 𝑏e = 3.2465 in.
3 3.1624 9.9625 31.5054 2
4 2(0.206) = 0.4120 0.0852 0.0351 𝑏1 + 𝑏2 = 4.9343 in.
5 2.2300 0.0375 0.0836
6 0.1130 0.2254 0.0255 Since the value of 𝑏1 + 𝑏2 is less than
7 9.6624 5.0000 48.3120 5.2302 in. shown in Fig. 4.16, the web element
Total 16.5430 89.3135 is not fully effective as assumed. The neutral
∑ axis should be relocated by using the partially
(Ly) 89.3135 effective web. The procedure is iterative as
𝑦cg = ∑ = = 5.399 in.
𝐿 16.5430 illustrated below.
b. Location of Neutral Axis Based on Ineffective
Web Elements. As the first iteration, the inef-
fective portion of the web can be assumed as
Use Section 3.3.1.2 in this volume or Section
follows:
1.1.2 of the North American Specification to
check the effectiveness of the web element. 5.2302 − (𝑏1 + 𝑏2 ) = 5.2302 − 4.9343
From Fig. 4.16,
= 0.2959 in.
( )
5.2302
f1 = 50 = 48.44 ksi (Compression) Therefore, the effective lengths of all
5.399
( ) elements are shown in Fig. 4.17 using partially
4.4322
f2 = 50 = 41.05 ksi (Tension) effective web.
5.399
BENDING STRENGTH AND DEFLECTION 141
1–0.22∕1.233
𝜌= = 0.666
1.233
𝑏e = 𝜌h = 6.4352 in.
𝑏e
b1 = = 1.6820 in.
3+𝜓
1
𝑏2 = 𝑏𝑒 = 3.2176 in
2
𝑏1 = 𝑏2 = 4.8996 in.
Because the above computed value of 𝑏1 + 𝑏2
is less than the previous value of 4.9343 in. by
0.7%, additional iterations are required.
For the second iteration, the ineffective
portion of the web is
5.2922 − (𝑏1 + 𝑏2 ) = 5.2922 − 4.8966
= 0.3926 in.
By using the same procedure shown above,
the neutral axis can be relocated as follows:
Figure 4.17 Effective lengths and stress distribution using
partially effective web (first iteration). Effective Distance from
Length Top Fiber
Element L (in.) y (in.) Ly(in.2 ) Ly2 (in.3 )
1 0.5512 9.5556 5.2670 50.3298
Effective Distance from
2 0.4120 9.9148 4.0849 40.5009
Element Length L (in.) Top Fiber y (in.) Ly(in.2 )
3 3.1624 9.9625 31.5054 313.8727
1 0.5512 9.5556 5.2670 4 0.4120 0.0852 0.0351 0.0030
2 0.4120 9.9148 4.0849 5 2.2300 0.0375 0.0836 0.0031
3 3.1624 9.9625 31.5054 6 0.1130 0.2254 0.0057 0.0057
4 0.4120 0.0852 0.0351 7 7.5878 6.0373 45.8098 276.5675
5 2.2300 0.0375 0.0836 8 1.6820 1.0098 1.6985 1.7151
6 0.1130 0.2254 0.0255 16.1504 88.5098 682.9977
7 7.6787 5.9919 46.0100 88.5098
8 1.6878 1.0127 1.7092 𝑦cg = = 5.481 in.
16.1504
16.2471 88.7207
88.7207 From Fig. 4.18,
𝑦cg = = 5.461 in.
16.2471
𝑓1 = 48.46 ksi (compression)

From Fig. 4.17, 𝑓2 = 39.68 ksi (tension)

𝑓1 = 48.45 ksi (compression) ℎ


𝜓 = 0.819 k = 19.68 = 128.83
𝑡
𝑓2 = 40.01 ksi (tension) √
ℎ 1.052 48.46
𝜓 = 0.826 k = 19.83 = 128.83 𝜆= √ (128.83)
𝑡 19.68 29,500

1.052 48.45 = 1.238 > 0.673
𝜆= √ (128.83)
19.83 29,500
1–0.22∕1.238
= 1.233 > 0.673 𝜌= = 0.664
1.238
142 4 FLEXURAL MEMBERS

𝐼𝑥′ = 234.6337 in.3


𝐼𝑥 = 𝐼𝑥′ 𝑡 = (234.6337)(0.075) = 17.598 in.4
17.598
𝑆𝑥 = = 3.211 in.3
5.481
2. Nominal and Allowable Moments. The nominal
moment for considering local buckling is
𝑀n𝓁 = 𝑆e 𝐹y = 𝑆x 𝐹y = 3.211(50)
= 160.55 in.-kips
The allowable moment is
𝑀 160.55
𝑀𝑎 = 𝑛𝓁 = = 96.14 in.-kips
Ω𝑏 1.67

B. LRFD Method
The nominal moment for the LRFD method is the same as
that computed for the ASD method. From item A above, the
nominal moment about the x axis of the C-section is
𝑀𝑛𝓁 = 160.55 in-kips

Figure 4.18 Effective lengths and stress distribution using The corresponding available moment:
partially effective web (second interaction). 𝜙b 𝑀𝑛𝓁 = 0.9(160.55) = 144.50 in.-kips

𝑏e = 𝜌h = 6.4158 in. Example 4.6 For the hat section with a stiffened compres-
sion flange as shown in Fig. 4.19, determine the allowable
𝑏e moment (𝑀a ) about the x axis for the ASD method and the
𝑏1 = = 1.6800 in.
3+𝜓 design moment (𝜙b 𝑀𝑛𝓁 ) for the LRFD method. Assume that
1 the yield stress of steel is 50 ksi, and the member does not
𝑏2 = 𝑏e = 3.2079 in.
2 subject to lateral–torsional buckling. Determine the nominal
𝑏1 = 𝑏2 = 4.8879 in. moment due to local buckling with initiation of yielding. The
linear method can be used to determine the member sectional
Because the above computed value of 𝑏1 + 𝑏2 properties.
is approximately equal to the value of 𝑏1 + 𝑏2 A. ASD Method
computed from the first iteration, it is accept-
1. Calculation of Sectional Properties. In order to use
able. Better accuracy can be obtained by using
the linear method, midline dimensions are shown in
additional iterations.
Fig. 4.20.
c. Moment of Inertia and Section Modulus. The
i. Corner Element. (Figs. 1.32 and 4.20)
moment of inertia based on line elements is
1
𝑅′ = 𝑅 + 𝑡 = 0.240 in.
𝐼1′ = 1
12
(0.5512)3 = 0.0140 2
Arc length:
𝐼6′ = 1
(0.113)3 = 0.0001
12
𝐿 = 1.57𝑅′ = 0.3768 in.
𝐼7′ = 1
(7.5878)3 = 36.4054
12 c = 0.637R′ = 0.1529 in.
𝐼8′ = 1
12
(1.6820)3
= 0.3965 ii. Location of Neutral Axis.
∑ a. First Approximation. For the compression
(Ly2 ) = 682.9977
flange,
𝐼𝑧′ = 719.8137 in.3 𝑤 = 15 − 2(𝑅 + 𝑡) = 14.415 in.
(∑ )
− 𝐿 (𝑦2cg ) = −(16.1504)(5.481)2 = −485.1800 𝑤
= 137.29
𝑡
BENDING STRENGTH AND DEFLECTION 143

Figure 4.19 Example 4.6

Figure 4.20 Line elements.

Using Eqs. (3.32)–(3.35) and assuming 𝑓 =


𝐹y = 50 ksi, Distance
from Top
√ Effective Fiber
1.052 50
𝜆 = √ (137.29) = 2.973 > 0.673 Element Length L (in.) y (in.) Ly(in.2 )
4 29,500
1 − 0.22∕2.973 1 2 × 1.0475 = 2.0950 9.9475 20.8400
𝜌= = 0.311 2 2 × 0.3768 = 0.7536 9.8604 7.4308
2.973
3 2 × 9.4150 = 18.8300 5.0000 94.1500
𝑏 = 𝜌w = 0.311(14.415) = 4.483 in. 4 2 × 0.3768 = 0.7536 0.1396 0.1052
5 4.4830 0.0525 0.2354
By using the effective width of the compres- Total 26.9152 122.7614
sion flange and assuming that the web is fully ∑
effective, the neutral axis can be located as (Ly) 122.7614
ycg = ∑ = = 4.561 in.
follows: L 26.9152
144 4 FLEXURAL MEMBERS

Because the distance 𝑦cg is less than the 𝑏0 = out-to-out width of compression flange
half depth of 5.0 in., the neutral axis is closer
= 15.00 in.
to the compression flange and, therefore, the
maximum stress occurs in the tension flange. Since ℎ0 ∕𝑏0 = 10∕15 = 0.667 < 4, then use Eq.
The maximum compressive stress can be (3.47a),
computed as follows: 𝑏𝑒
𝑏1 =
( ) 3+𝜓
4.561
𝑓 = 50 = 41.93 ksi where be is the effective width of the web deter-
10 − 4.561
Since the above computed stress is less than mined in accordance with Eqs. (3.32)–(3.35) with
the assumed value, another trial is required. 𝑓1 substitued for f and 𝑘 = 31.12 as follows:
b. Second Approximation. After several trials, ℎ 9.415
= = 89.67 < 200 OK
assume that 𝑡 0.105

𝑓 = 40.70 ksi 𝜆 = 2.682 > 0.673 1.052 38.04
𝜆= √ (89.67) = 0.607 < 0.673
29,500
𝑏 = 4.934 in. 31.12
𝑏e = ℎ = 9.415 in.
Distance 𝑏e
Effective from Top 𝑏1 = = 2.218 in.
3+𝜓
Element Length L (in.) Fiber y (in.) Ly(in.2 ) Ly2 (in.3 )
Since 𝜓 > 0.236,
1 2.0950 9.9475 20.8400 207.3059 1
𝑏2 = 𝑏e = 4.7075 in.
2 0.7536 9.8604 7.4308 73.2707 2
3 18.8300 5.0000 94.1500 470.7500 𝑏1 + 𝑏2 = 6.9255 in.
4 0.7536 0.1396 0.1052 0.0147
5 4.9340 0.0525 0.2590 0.0136 Because the computed value of 𝑏1 + 𝑏2 is
Total 27.3662 122.7850 751.3549 greater than the compression portion of the web
(4.1945 in.), the web element is fully effective.
122.7850 iv. Moment of Inertia and Section Modulus. The
𝑦cg = = 4.487 in.
27.3662 moment of inertia based on line elements is
( )
1
that is, 2𝐼3′ = 2 (9.415)3 = 139.0944
( ) 12
4.487 ∑
𝑓 = 50 = 40.69 ksi (Ly)2 = 751.3549
10 − 4.487
Since the above computed stress is close to 𝐼z′ = 890.4493 in.3
the assumed value, it is OK. (∑ )
iii. Check the Effectiveness of the Web. Use Section − 𝐿 (𝑦2cg ) = −27.3663(4.487)2 = −550.9683 in.3
3.3.1.2 in this volume or Section 1.1.2 of the North
American Specification to check the effectiveness 𝐼x′ = 339.4810 in.3
of the web element. From Fig. 4.21, The actual moment of inertia is
( )
4.1945 𝐼𝑥 = 𝐼𝑥′ 𝑡 = (339.4810)(0.105) = 35.646 in.4
𝑓1 = 50 = 38.04 ksi (compression)
5.513
( ) The section modulus relative to the extreme
5.2205
𝑓2 = 50 = 47.35 ksi (tension) tension fiber is
5.513
35.646
|𝑓 | 𝑆𝑥 = = 6.466 in.3
𝜓 = || 2 || = 1.245 5.513
| 𝑓1 | 2. Nominal and Allowable Moments. The nominal
𝑘 = 4 + 2(1 + 𝜓)3 + 2(1 + 𝜓) moment for section strength is
= 4 + 2(2.245)3 + 2(2.245) = 31.12 𝑀𝑛𝓁 = 𝑆e 𝐹y = 𝑆x 𝐹y = (6.466)(50) = 323.30 in.-kips
From Fig 4.19, The allowable moment is
𝑀 323.30
ℎ0 = out-to-out depth of web 𝑀a == 𝑛𝓁 = = 193.59 in.-kips
Ωb 1.67
= 10.00 in.
BENDING STRENGTH AND DEFLECTION 145

Figure 4.21 Effective lengths and stress distribution using fully effective web.

B. LRFD Method
The nominal moment for the LRFD method is the same as
that computed for the ASD method. From item A above, the
nominal moment about the x axis of the hat section is
𝑀𝑛𝓁 = 323.30 in.-kips
The corresponding design moment is:
𝜙b 𝑀𝑛𝓁 = 0.9(323.30) = 290.97 in.-kips

Example 4.7 For the section with an intermediate stiffener 2.435


2.2662
as shown in Fig. 4.22, determine the allowable moment (𝑀a )
2.3962
about the x axis for the ASD method and the design moment 2.565

(𝜙b 𝑀𝑛𝓁 ) for the LRFD method. Use the linear method with
𝐹y = 33 ksi. The nominal moment is determined by initiation
of yielding.
Figure 4.23 Example 4.7; line elements.
SOLUTION
A. ASD Method
1. Calculation of Sectional Properties. Using the linear
method as shown in Fig. 4.23.
i. Corner Element. (Figs. 1.32 and 4.23)
1
𝑅′ = 𝑅 + 𝑡 = 0.1313 in.
2
Arc length:
𝐿 = 1.57𝑅′ = 1.57(0.1313) = 0.2063 in.
𝑐 = 0.637R′ = 0.637(0.1313) = 0.0836 in.
ii. Location of Neutral Axis Based on Section
3.3.3.3.1 in this volume or Section 1.4.1 of the
Figure 4.22 Example 4.7. North American Specification.
146 4 FLEXURAL MEMBERS

a. First Approximation. For the top compression = 24.027


flange,
The modification factor for the distortional
𝑏o = 12 − 2(𝑅 + 𝑡) = 11.6624 in. plate buckling coefficient can be computed as
𝑏p = 𝑤 = 5.5686 in. follows:

𝐴s = [2(0.7) + 4(0.2063)](0.075) ℎ = 5 − 2(𝑅 + 𝑡) = 4.6625


𝑏o 11.6624
= 0.1669 in.2 = = 2.501 > 1
ℎ 4.6625
𝐴g = 2(5.5686)(0.075) + 0.1669 11 − 𝑏o ∕ℎ 11 − 2.501
𝑅= =
= 1.0022 in.2 5 5
1
The moment of inertia of the full intermediate = 1.700 > OK
2
stiffener (elements 7, 8, and 9 in Fig. 4.23) about
Since
its own centroidal axis is given as
[ ( ) ] Rkd = (1.700)(24.027)
1
2 12 (0.7)3 + 4(0.2063)×
𝐼s = (0.075) = 40.846 > (𝑘loc = 17.545)
(0.35 + 0.0836)2
use
= 15.923 × 10−3 in.4
the moment of inertia of the stiffener about 𝑘 = minimum of (Rkd and 𝑘loc ) = 17.545
the centerline of the flat portion of the element Then
as 𝑘𝜋 2 𝐸
𝑓cr𝓁 =
𝐼sp = 𝐼s + 𝐴s (0.35 + 0.0836 + 0.0477) 2
12(1 − 𝜇2 )(𝑏0 ∕𝑡)2
= 0.015923 + 0.1669(0.4813)2 (17.545)𝜋 2 (29, 500)
=
= 54.585 × 10−3 in.4 12(1 − 0.32 )(11.6624∕0.075)2
= 19.346 ksi
and the plate local buckling coefficient for the
subelement [Eq. 3.91)] as Assuming that 𝑓 = 𝐹y = 33 ksi yields
√ √
𝑘loc = 4(𝑏0 ∕𝑏p ) = 4(11.6624∕5.5686)
2 2
𝑓 33
𝜆= =
= 17.545 𝐹cr𝓁 19.346
From Eqs. (3.92)–(3.95), = 1.306 > 0.673
10.92𝐼sp 10.92(54.585 × 10−3 ) 1 − 0.22∕1.306
𝛾= = 𝜌= = 0.637
𝑏0 𝑡3 (11.6624)(0.075)3 1.306
0.596068 From Eq. (3.85), the effective width of
= = 121.150 elements 7, 8, 9, and 10 located at the centroid
0.00492
of the top flange, including the intermediate
𝐴 0.1669
𝛿= s = = 0.1908 stiffener, is given as
𝑏0 𝑡 11.6624 × 0.075 ( ) ( )
𝐴g 1.0022
𝛽 = [1 + 𝛾(𝑛 + 1)]1∕4 𝑏e = 𝜌 = 0.637
𝑡 0.075
= [1 + 121.150(1 + 1)]0.25 = 3.949 = 8.512 in.
𝑘d = plate buckling coefficient The location of the centroid from the top fiber
for distortional buckling of the flange is

(1 + 𝛽 2 )2 + 𝛾(1 + 𝑛) 2(5.5686)(0.057)(0.075∕2)
= +𝐴s (0.7∕2 + 3∕32 + 0.075)
𝛽 2 [1 + 𝛿(𝑛 + 1)] 𝑦7−10 =
𝐴g
(1 + 3.9492 )2 + 121.150(1 + 1)
= = 0.1176 in. from top fiber
3.9492 [1 + 0.1908(1 + 1)]
BENDING STRENGTH AND DEFLECTION 147
Since the computed stress is close to the
Distance from assumed value of 31.3 ksi, it is OK.
Effective Top Fiber To check if the web is fully effective, refer to
Element Length L (in.) y (in.) Ly(in.2 ) Fig. 4.23:
1 2 × 0.5812 = 1.1624 4.5406 5.2780 2.2662
𝑓1 = 33 = 29.156 ksi
2 2 × 0.2063 = 0.4126 4.9148 2.0278 2.565
3 2 × 3.1624 = 6.3248 4.9625 31.3868 𝑓2 = 33
2.3962
= 30.828 ksi
4 2 × 0.2063 = 0.4126 4.9148 2.0278 2.565
5 2 × 4.6624 = 9.3248 2.5000 23.3120 |f |
6 2 × 0.2063 = 0.4126 0.0852 0.0352 𝜓 = || 2 || = 1.126
| f1 |
7–10 8.5120 0.1176 1.001
Total 26.5618 65.0686 𝑘 = 4 + 2(1 + 𝜓)3 + 2(1 + 𝜓) = 27.47
∑ From Fig 4.22,
Ly 65.0686
𝑦cg = ∑ = = 2.4497 in. < 2.5 in.
𝐿 26.5618 ℎ0 = out-to-out depth of web
( )
2.4497 = 5.00 in.
𝑓= (33) = 31.70 ksi
5 − 2.4497
𝑏0 = out-to-out width of
Since the computed compression stress, f, is
compression flange
considerably less than 33 ksi, additional trials
are required. After several trials, it was found = 12.00 in.
that the stress should be about 31.3 ksi.
b. Additional Approximation. Assume Since ℎ0 ∕𝑏0 = 5∕12 = 0.417 < 4, then use
Eq. (3.47a)
𝑓 = 31.3 ksi 𝑏e
√ 𝑏1 =
3+𝜓
𝑓
𝜆= = 1.272 > 0.673 where be is the effective width of the web deter-
𝑓cr𝓁
mined in accordance with Eqs. (3.32)–(3.35)
1 − 0.22∕𝜆 with 𝑓1 substitued for f and 𝑘 = 27.47 as
𝜌= = 0.650
𝜆 follows:
( ) ℎ 4.6624
𝐴g
𝑏e = 𝜌 = 8.6857 in. = = 62.17 < 200 OK
𝑡 𝑡 0.075

𝑦7−10 = 0.1176 in.from top fiber 1.052 29.156
𝜆= √ (62.17) = 0.392 < 0.673
27.47 29,500
Effective Distance from 𝑏e = ℎ = 4.6624 in.
Length Top Fiber 𝑏e 4.6624
Element L (in.) y (in.) Ly(in.2 ) Ly2 (in.3 ) b1 = = = 1.13 in.
3+𝜓 3 + 1.126
1 1.1624 4.5406 5.2780 23.9653 Since 𝜓 > 0.236,
2 0.4126 4.9148 2.0278 9.9664 1
3 6.3248 4.9625 31.3868 155.7570 𝑏2 = 𝑏e = 2.3312 in.
2
4 0.4126 4.9148 2.0278 9.9664
b1 = 𝑏2 = 3.462 in. > 2.2662 in.
5 9.3248 2.5000 23.3120 58.2800
6 0.4126 0.0852 0.0352 0.0030 The web is fully effective as assumed.
7–10 8.6857 0.1176 1.0214 0.1201 iii. Total Ix and Sx .
Total 26.7355 65.089 258.058 1
2𝐼5′ = 2 × (4.6624)3 = 16.8918
65.089 12
𝑦cg = = 2.435 in.
26.7355 1
( ) 2𝐼1′ = 2 × (0.5812)3 = 0.0327
2.435 12
𝑓= (33) = 31.33 ksi
5 − 2.435 2(𝐼2′ + 𝐼4′ + 𝐼6′ ) = 0.0020
148 4 FLEXURAL MEMBERS

(Ly2 ) = 258.0345 load-carrying capacity of continuous beams and steel decks
have also been studied by some other investigators.4.6–4.9
274.9610 In order to utilize the available inelastic reserve strength
(∑ )
− 𝐿 (𝑦cg ) = −(26.7355)(2.435 ) = −158.521
2 2 of certain cold-formed steel beams, design provisions based
on the partial plastification of the cross section were added in
𝐼x′ = 116.440 in.3 the 1980 edition of the AISI Specification. The same provi-
sions were retained in the 1986 and the 1996 editions of the
𝐼x = 𝐼x′ 𝑡 = 116.440(0.075) = 8.7330 in.4
AISI specification and the North American Specification.
8.7330 According to F2.4.1 of the 2016 edition of the Specification,
𝑆x = = 3.4047 in.3
5 − 2.435 the nominal strengths 𝑀ne of those beams satisfying certain
2. Nominal and Allowable Moments. The nominal specific limitations can be determined on the basis of the
moment for due to initiation of yielding is inelastic reserve capacity with a limit of 1.25𝑆e 𝐹y . The
nominal moment 𝑀𝑛𝓁 is the maximum bending capacity
𝑀𝑛𝓁 = 𝑆e 𝐹y = 𝑆x 𝐹y of the beam by considering the inelastic reserve strength
= (3.4047)(33) = 112.355 in.-kips through partial plastification of the cross section. The
inelastic stress distribution in the cross section depends on
The allowable moment is the maximum strain in the compression flange, 𝜀cu . Based on
𝑀 112.355 the Cornell research work on hat sections having stiffened
𝑀a = 𝑛𝓁 = = 67.28 in.-kips
Ωb 1.67 compression flanges,4.1 the design provision included in
Section F2.4.1 of the North American Specification limits
B. LRFD Method the maximum compression strain to be 𝐶y 𝜀y , that is,
The nominal moment for the LRFD method is the same as
that computed for the ASD method. From the above calcu- 𝜀cu = 𝐶y 𝜀y (4.75)
lations, the nominal moment about the x axis of the section
is where 𝜀y = yield strain, = 𝐹y ∕E, in.∕in.
𝑀𝑛𝓁 = 112.355 in.-kips 𝐸 = modulus of elasticity of steel,
29.5 × 103 ksi (203 GPa or
The corresponding design moment is:
2.07 × 106 kg∕cm2 )
𝜙b 𝑀𝑛𝓁 = 0.90(112.355) = 101.12 in.-kips 𝐹y = yield stress of steel, ksi

4.2.3.1.2 Inelastic Reserve Capacity of Beams Prior and 𝐶y is the compression strain factor determined as
to 1980, the inelastic reserve capacity of beams was not follows:
included in the AISI Specification because most cold-formed
steel shapes have width-to-thickness ratios considerably in 1. Stiffened compression elements without intermediate
excess of the limits required by plastic design. Because of stiffeners:
the use of large width-to-thickness ratios for the beam flange a. When 𝑤∕𝑡 ≤ 𝜆1 ,
and web, such members are usually incapable of developing 𝐶y = 3.0 (4.76)
plastic hinges without local buckling.
In the 1970s research work on the inelastic strength of b. When 𝜆1 < 𝑤∕𝑡 < 𝜆2 ,
cold-formed steel beams was carried out by Reck, Pekoz, ( )
𝑤∕𝑡 − 𝜆1
Winter, and Yener at Cornell University.4.1–4.4 These studies 𝐶y = 3 − 2 (4.77)
𝜆 2 − 𝜆1
showed that the inelastic reserve strength of cold-formed
steel beams due to partial plastification of the cross section c. When 𝑤∕𝑡 ≥ 𝜆2 ,
and the moment redistribution of statically indeterminate 𝐶y = 1.0 (4.78)
beams can be significant for certain practical shapes. With
proper care, this reserve strength can be utilized to achieve where
more economical design of such members. In Europe, a study 1.11 1.28
𝜆1 = √ 𝜆2 = √
has been made by von Unger on the load-carrying capacity 𝐹y ∕𝐸 𝐹y ∕𝐸
of transversely loaded continuous beams with thin-walled
sections, in particular of roof and floor decks with trape- The relationship between 𝐶y and the w/t ratio of the
zoidal profiles.4.5 In addition, the buckling strength and compression flange is shown in Fig. 4.24.
BENDING STRENGTH AND DEFLECTION 149

Eq. (4.76)
Eq. (4.77)

Eq. (4.78)

Figure 4.24 Factor Cy for stiffened compression elements without intermediate stiffeners.

2. Unstiffened compression elements: For the above requirements, Eqs. (4.79a)–(4.79g) were
i. Unstiffened compression elements under a stress added in 2004 for sections having unstiffened elements under
gradient causing compression at one longitudinal a stress gradient.1.343 These added design equations are based
edge and tension at the other longitudinal edge: on the research work conducted by Bamback and Rasmussen
⎧3.0 at the University of Sydney on I- and plain channel sections
𝜆 ≤ 𝜆3 (4.79𝑎) in minor-axis bending.4.194,4.195 The 𝐶y values are dependent

⎪3 − 2[(𝜆 − 𝜆3 )∕(𝜆4 − 𝜆3 )] on the stress ratio 𝜓 and slenderness factor 𝜆 of the unstiff-
𝐶y = ⎨
⎪𝜆3 ≤ 𝜆 ≤ 𝜆4 (4.79𝑏) ened element with the stress gradient determined in Section
⎪1 𝜆 ≥ 𝜆4 (4.79𝑐) 3.3.2.2.
⎩ On the basis of the maximum compression strain εcu
where 𝜆 is the slenderness factor defined in Section allowed in Eq. (4.75), the neutral axis can be located by using
3.3.1.1, Eq. (4.81), and the nominal moment 𝑀ne can be determined
by using Eq. (4.82) as follows:
𝜆3 = 0.43 (4.79d)
𝜆4 = 0.673(1 + 𝜓) (4.79e) 𝜎dA = 0 (4.81)

and 𝜓 is defined in Section 1.2.2 of the Specifica-
tion. 𝜎ydA = 𝑀𝑛𝓁 (4.82)

ii. Unstiffened compression elements under a stress
gradient causing compression at both longitudinal where 𝜎 is the stress in the cross section.
edges: For hat sections Reck, Pekoz, and Winter gave the
𝐶y = 1 (4.79f) following equations for the nominal moments of sections
with yielded tension flange and sections with tension flange
iii. Unstiffened compression elements under uniform
not yielded:
compression:
𝐶y = 1 (4.79g)
a. Sections with Yielded Tension Flange at Nominal
3. Multiple-stiffened compression elements and compres- Moment.4.1 For the stress distributions shown in
sion elements with edge stiffeners Fig. 4.25, Eqs. (4.83)–(4.88) are used for computing
the values of 𝑦c , 𝑦t , 𝑦p , 𝑦cp , 𝑦tp , and Mn . For the purpose of
𝐶y = 1.0 (4.80) simplicity, midline dimensions are used in the calculation:
No limit is placed on the maximum tensile strain in the 𝑏t − 𝑏c + 2𝑑
North American Specification. 𝑦c = (4.83)
4
150 4 FLEXURAL MEMBERS

Figure 4.25 Stress distribution in sections with yielded tension flange at nominal moment.4.1

𝑦t = 𝑑 − 𝑦c (4.84) b. Sections with Tension Flange Not Yielded at Nominal


Moment.4.1 For the stress distribution shown in
𝑦c
𝑦p = (4.85) Fig. 4.26, 𝑦c is computed from the following quadratic
𝜀cu ∕𝜀y equation:
𝑦cp = 𝑦c − 𝑦p (4.86) ( )
1
𝑦2c 2 − − 𝐶y + 𝑦c (𝑏c + 2𝐶y 𝑑 + 𝐶y 𝑏t )
𝑦tp = 𝑦t − 𝑦p (4.87) 𝐶y

[ ( ) − (𝐶y 𝑑 2 + 𝐶y 𝑏t 𝑑) = 0 (4.89)
1 4
𝑀𝑛𝓁 = 𝐹y 𝑡 𝑏c 𝑦c + 2𝑦cp 𝑦p + 𝑦cp + (𝑦p )2
2 3
( ) ]
1
+2𝑦tp 𝑦p + 𝑦tp + 𝑏y 𝑦t (4.88)
2 Subsequently, the values of 𝑦t , 𝑦p , and 𝑦cp can be computed
from Eqs. (4.84),(4.85), and (4.86), respectively.

Figure 4.26 Stress distribution in sections with tension flange not yielded at nominal moment.4.1
BENDING STRENGTH AND DEFLECTION 151
If 𝑦p > 𝑦t , then the case in part (b) above applies and the SOLUTION
nominal moment Mn𝓁 is computed as follows: A. ASD Method
[ ( )
𝑦cp 2 1. Dimensions of Section. By using the midline dimen-
𝑀𝑛𝓁 = 𝐹y 𝑡 𝑏c 𝑦c + 2𝑦cp 𝑦p + + (𝑦p )2 sions and square corners, the widths of compression
2 3
( ) ( )] and tension flanges and the depth of webs are computed
2 𝜎t 𝜎t
+ (𝑦t )2 + 𝑏y 𝑦t (4.90) as follows:
3 𝐹y 𝐹y a. Width of compression flange:
In Eq. (4.90), 𝜎t = 𝐹y 𝐶y 𝑦t ∕𝑦c . 𝑏c = 3 − 0.105 = 2.895 in.
It should be noted that according to Section F2.4.1
of the North American Specification, Eqs. (4.88) and b. Width of tension flange:
(4.90) can be used only when the following conditions 𝑏t = 2(1.34 − 0.105∕2) = 2.576 in.
are met:
c. Depth of webs:
1. The member is not subject to twisting or to lateral, 𝑑 = 3 − 0.105 = 2.895 in.
torsional, or flexural–torsional buckling.
2. The effect of cold work of forming is not included in All dimensions are shown in Fig. 4.28a. Check the
determining the yield stress 𝐹y . effective width of the compression flange:
3. The ratio of the depth of the compression portion of the ( )
3
web to its thickness does not exceed 𝜆1 . w=3−2 + 0.105 = 2.415 in.
16
4. The shear force does not exceed 0.35𝐹y for ASD w 2.415
and 0.6𝐹y for LRFD and LSD times the web area = = 23
t 0.105
(ht for stiffened elements or wt for unstiffened k = 4.0
elements).
5. The angle between any web and the vertical does not f = Fy = 33 ksi
exceed 30∘ . √
1.052 33
λ = √ (23)
It should also be noted that, when applicable, effective 4 29,500
design widths should be used in the calculation of sectional = 0.405 < 0.673
properties.
b = w = 2.415 in.
Example 4.8 For the hat section (3 × 3 × 0.105 in.) shown Therefore, the compression flange is fully effective.
in Fig. 4.27, determine the allowable moment (𝑀a ) about the 2. Strain Diagram. The w/t ratio of the stiffened
x axis for the ASD method and the design moment (𝜙b 𝑀𝑛𝓁 ) compression flange is given as
for the LRFD method. Consider the inelastic reserve capacity 𝑤
according to Section F2.4.1 of the 2016 edition of the North = 23
𝑡
American Specification. Use 𝐹y = 33 ksi and assume that 1.11 1.11
lateral support is adequately provided. 𝜆1 = √ =√ = 33.2
𝐹y ∕𝐸 33∕29,500

Since 𝑤∕𝑡 < (𝜆1 = 33.2), according to Eq. (4.76),


𝐶y = 3.0. Therefore, 𝜀cu = 3 𝜀y , as shown in
Fig. 4.28b.
3. Stress Diagram. The values of 𝑦c , 𝑦t , 𝑦p , 𝑦cp , and 𝑦tp
are computed by using Eqs. (4.83)–(4.87) as follows:
𝑏t − 𝑏c + 2𝑑 2.576 − 2.895 + 2 × 2.895
𝑦c = =
4 4
= 1.368 in.
𝑦t = 𝑑 − 𝑦c = 2.895 − 1.368 = 1.527 in.
𝑦c 1.368
𝑦p = = = 0.456 in.
Figure 4.27 Example 4.8. 𝜀cu ∕𝜀y 3
152 4 FLEXURAL MEMBERS

Figure 4.28 Stress distribution: (a) midline dimensions; (b) strain; (c) stress.

𝑦cp = 𝑦c − 𝑦p = 1.368 − 0.456 = 0.912 in. B. LRFD Method


The nominal moment for the LRFD method is the same as
𝑦tp = 𝑦t − 𝑦p = 1.527 − 0.456 = 1.071 in. that computed for the ASD method. From item A above, the
All dimensions are shown in Fig. 4.28c. nominal moment about the x axis of the hat section is
4. Nominal Moment Mn𝓁 . In order to utilize the inelastic 𝑀𝑛𝓁 = 40.93 in.-kips
reserve capacity, the North American Specification
requirements must be checked: The corresponding design moment is:
𝑦c 1.368 𝜙𝑏 𝑀𝑛𝓁 = (0.9)(40.93) = 36.84 in.-kips
= = 13.03 < (𝜆1 = 33.2) OK
𝑡 0.105
Therefore, the nominal moment is Example 4.9 For the I-section with unequal flanges as
[ ( ) shown in Fig. 4.29, determine the allowable moment (𝑀a )
1 4
M𝑛𝓁 = Fy t bc yc + 2ycp yp + ycp + (yp )2 about the x axis for the ASD method and the design moment
2 3
( ) ] (𝜙b 𝑀n ) for the LRFD method. Consider the inelastic reserve
1
+2ytp yp + ytp + bt yt
2
[
= (33)(0.105) (2.895)(1.368) + 2(0.912)
( )
1 4
× 0.456 + × 0.912 + (0.456)2
2 3
( )
1
+ 2(1.071) 0.456 + × 1.071
2
]
+ (2 × 1.288)(1.527)

= 41.43 in.-kips
5. Based on the method illustrated in Example 4.6, 𝑆e for
the given hat section is 0.992 in.3 The nominal moment,
𝑀𝑛𝓁 should be limited to 1.25𝑆e 𝐹y according to the
Speccfication. Therefore
1.25𝑆e 𝐹y = 1.25(0.992)(33) = 40.93 in.-kips < 𝑀𝑛𝓁
6. Allowable Moment Ma . Because 𝑀𝑛𝓁 exceeds
1.25𝑆𝑒 𝐹𝑦 , use
M𝑛𝓁 = 1.25Se Fy = 40.93 in.-kips
M𝑛𝓁 40.93
Ma = = = 24.51 in.-kips Figure 4.29 Example 4.9.
Ωb 1.67
BENDING STRENGTH AND DEFLECTION 153
capacity and use 𝐹y = 50 ksi. Assume that the lateral support The depth of the webs is given as
is adequately provided to prevent lateral buckling.
𝑑 = 8.0 − 𝑡 = 8.0 − 0.135 = 7.865 in.
SOLUTION
All midline dimensions are shown in Fig. 4.30a.
A. ASD Method 2. Strain Diagram. For an unstiffened compression
1. Dimensions of Section. By using the midline dimen- flange under uniform compression, 𝐶y = 1.0. There-
sions and square corners, the widths of compression fore, 𝜀cu = 𝜀y , as shown in Fig. 4.30b.
and tension flanges and the depth of webs are computed 3. Stress Diagram. The values of 𝑦c , 𝑦t , 𝑦p , and
as follows: The flat width of the unstiffened compres- 𝑦tp are computed by using Eqs. (4.83)–(4.87) as
sion flange according to Section 3.2 is follows:
( )
3
𝑤 = 2.5 − (𝑅 + 𝑡) = 2.5 − + 0.135 𝑏t − 𝑏c + 2𝑑 1.865 − 3.754 + 2(7.865)
16 𝑦c = =
4 4
= 2.1775 in.
= 3.46 in.
𝑤 2.1775
= = 16.13
𝑡 0.135 𝑦t = 𝑑 − 𝑦c = 7.865 − 3.46 = 4.405 in.
For 𝑓 = 𝐹y = 50 ksi in the top fiber and k = 0.43 for 𝑦c
the unstiffened flange, 𝑦p = = 𝑦c = 3.46 in.
𝜀cu ∕𝜀y

1.052 50 𝑦cp = 0
𝜆= √ (16.13) = 1.065 > 0.673
0.43 29,500
( ) 𝑦tp = 𝑦t − 𝑦p = 4.405 − 3.46 = 0.945 in.
1 − 0.22∕𝜆
𝑏 = 𝜌w = 𝑤 = 1.622 in. All dimensions are shown in Fig. 4.30c.
𝜆
( ) ( ) 4. Nominal Moment. In order to satisfy the North Amer-
𝑏c 1 3 0.135
= 𝑏 + 𝑅 + 𝑡 = 1.622 + + ican Specification requirements for using the inelastic
2 2 16 2 reserve capacity, check the 𝑦c ∕𝑡 ratio against the limit
= 1.877 in. of λ1 :
𝑏c = 3.754 in. 𝑦c 3.46
= = 25.63
The width of the tension flange is determined as 𝑡 0.135
𝑏t 𝑡 0.135 𝜆1 = √
1.11
=√
1.11
= 26.96
=1− =1− = 0.9325 in.
2 2 2 𝐹y ∕𝐸 50∕29,500
𝑏t = 1.865 in.

Figure 4.30 Stress distribution: (a) midline dimensions; (b) strain; (c) stress.
154 4 FLEXURAL MEMBERS

Since 𝑦c ∕𝑡 < λl , OK. Therefore, the nominal (a) For 𝜆𝓁 ≤ 0.776


moment is determined as 𝑀𝑛𝓁 = 𝑀ne (4.91)
[ ( ) ]
4 1
𝑀𝑛𝓁 = 𝐹y 𝑡 𝑏c 𝑦c + (𝑦p )2 + 2𝑦tp 𝑦p + 𝑦tp + 𝑏t 𝑦t (b) For λ𝓁 > 0.776
3 2 [
[ ( ) ]( )
4 𝑀cr𝓁 0.4 𝑀cr𝓁 0.4
= 50(0.135) (3.754 × 3.46) + (3.46)2 + 2(0.945) 𝑀𝑛𝓁 = 1-0.15 𝑀ne (4.92)
3 𝑀ne 𝑀ne
( ) ]
1 √
× 3.46 + × 0.945 + 1.865(4.405) where λ𝓁 = 𝑀ne ∕𝑀cr𝓁 (4.93)
2
= 301.05 in. − kips 𝑀ne = Nominal flexural strength [resistance] for
lateral–torsional buckling as defined in
5. Following the same method illustrated in Example 4.4, Specification Section F2
the effective section modulus based on the yielding, 𝑀cr𝓁 = Critical elastic local buckling moment,
𝑆e = 6.247 in.3 The nominal moment, 𝑀𝑛𝓁 should be determined in accordance with Specification
limited to 1.25𝑆e 𝐹y according to the Specification: Appendix 2
1.25 𝑆e 𝐹y = (1.25)(6.247)(50)
4.2.3.2.1 Members with Holes. As discussed in
= 390.4 in. − kips > 𝑀𝑛𝓁
Section 3.5, the Direct Strength Method for considering
5. Nominal Moment and Allowable Moment. Because members with holes was developed by Moen and Schafer
𝑀𝑛𝓁 is less than 1.25𝑆e 𝐹y , use 𝑀𝑛𝓁 for the nominal et al.3.273–3.275 The design provisions was introduced into
moment, that is, Appendix 1 of the 2012 edition of the North American Spec-
𝑀𝑛 = 𝑀𝑛𝓁 = 301.05 in. − kips ification. In 2016, these design provisions were incorporated
into the main body of the Specification1.417 as Section
The allowable design moment is F3.2.2, and are excerpted as follows:
𝑀 301.05
𝑀a = 𝑛 = = 180.27 in.-kips F3.2.2 Members With Holes
Ω𝑏 1.67
The nominal flexural strength [resistance], 𝑀𝑛𝓁 , for local buck-
B. LRFD Method ling of beams with holes shall be calculated in accordance with
The nominal moment for the LRFD method is the same as Specification Section F3.2.1, except 𝑀cr𝓁 shall be determined
that computed for the ASD method. From item A above, including the influence of holes:
the nominal moment about the x axis of the I-section with
unequal flanges is 𝑀𝑛𝓁 ≤ 𝑀ynet (4.94)

𝑀n = 301.05 in. − kips where 𝑀ynet = Member yield moment of net cross-section

The corresponding design comment is: = Sfnet Fy (4.95)

𝜙𝑏 𝑀𝑛 = (0.9)(301.05) = 270.95 in.-kips where 𝑆fnet = Net section modulus referenced to the extreme
fiber at first yield
4.2.3.2 Direct Strength Method The Direct Strength 𝐹y = Yield stress
Method considers the strength due to local buckling as the
Design examples will be provided to illustrate the provi-
function of the local-buckling moment, 𝑀cr𝓁 , and the yield
sions.
stress, 𝐹y . How to obtain the local buckling moment using
readily available software or through analytical approach is
4.2.3.2.3 Members Considering Local Inelastic Reserve
discussed in Section 3.5. The following design provisions,
Strength The local inelastic reserve strength is plotted
developed by Schafer and Pekoz3.254,3.255 , were first adopted
in Fig. 3.68b in the region where 𝑀n exceeds 𝑀y . Based
into the North American Specification in 2004 as Appendix
1. In the 2016 edition, they are incorporated into the main on the research by Shifferaw and Schafer,4.296 the inelastic
body of the Specification:1.417 reserve strength can be considered when the predicted
lateral–torsional buckling strength, 𝑀ne is greater than yield
F3.2.1 Members Without Holes moment, i.e., 𝑀ne > 𝑀y . Similar to the Effective Width
The nominal flexural strength [resistance], 𝑀𝑛𝓁 , for considering Method to predict the inelastic reserve strength, the ratio
interaction of local buckling and global buckling shall be deter- of the maximum compressive strain to the yield strain is
mined as follows: limited to 3. For cross-sections with the first yielding in
BENDING STRENGTH AND DEFLECTION 155
tension, it is recommended that the ratio of the maximum determined according to Section 4.2.3. For the flange distor-
tensile strain to the yield strain be limited to 3 as well. The tional buckling mode, the flange and edge stiffener rotate
following design provisions are excerpted from the North about the flange–web junction with some rotational resis-
American Specification:1.417 tance provided by the web. This mode of failure occurs
at considerably longer wavelengths than local buckling but
F3.2.3 Members Considering Local Inelastic Reserve generally shorter wavelength than lateral–torsional buckling.
Strength The distortional buckling may also take place in sections
Inelastic reserve capacity is permitted to be considered as as shown in Fig. 3.2(b) and 3.2(c), where the portion of
follows, provided λ𝓁 ≤ 0.776 and 𝑀ne ≥ 𝑀y :
the flange (with intermediate stiffeners as shown in Fig.
3.2(b)) or the portion of the lip and the flange (as shown in
(a) Sections symmetric about the axis of bending or sections
with first yield in compression:
Fig. 3.2(c)) starts to rotate about the junction(s) between the
flange and web.
𝑀𝑛𝓁 = 𝑀𝑦 + (1 − 1∕𝐶𝑦𝓁
2
)(𝑀𝑝 − 𝑀𝑦 ) (4.96) Distortional buckling may occur simultaneously with
(b) Sections with first yield in tension: local buckling.1.358,1.358,4.208,4.223,4.224 Research work indi-
cated that the local–distortional interaction is generally
𝑀𝑛𝓁 = 𝑀yc + (1 − 1∕𝐶𝑦𝓁
2
)(𝑀𝑝 − 𝑀yc ) ≤ 𝑀yt3 (4.97) weak and that if this limit state is included in the design
√ requirements the resulting capacities are not consistent
where λ𝓁 = My ∕Mcrl (4.98) with observations.1.412 Therefore, no design provisions are
𝑀ne = Nominal flexural strength [resistance] as currently included in the North American Specification for
defined in Specification Section F2 this limit state. For detailed discussion of modal interactions,
√ see Section 13.4.4 of the SSRC guide.1.412
C𝑦𝓁 = 0.776∕𝜆𝓁 ≤ 3 (4.99) In earlier years, distortional buckling has not been specifi-
𝑀cr𝓁 = Critical elastic local buckling moment, cally considered for the design of cold-formed steel members
determined in accordance with Specification having edge-stiffened compression flanges. The AISI design
Appendix 2 provisions provided by Desmond et al.3.76 for uniformly
𝑀p = Member plastic moment as given in compressed element with a simple lip edge stiffener account
Eq. (4.73) for the inability of the stiffener to prevent flange buck-
𝑀y = Member yield moment in accordance with ling by reducing the local buckling coefficient k to less
Specification Section F2.1
than 4.0 for the partially stiffened compression flange. The
𝑀yc = Moment at which yielding initiates in
compression (after yielding in tension).
reduced buckling coefficient is then used to compute the
Myc = My may be used as a conservative effective width of the flange element. However, in 1992
approximation Kwon and Hancock found that the AISI approach is uncon-
servative for distortional buckling of C-sections composed of
Myt3 = My + (1 − 1∕C2yt )(Mp − My ) (4.100) high-strength steel using a yield stress of 80 ksi (550 MPa or
5624 kg/cm2 ).4.196 In 1999, Schafer and Pekoz indicated that
𝐶yt = Ratio of maximum tension strain to yield the AISI-reduced local buckling coefficient is only intended
strain to be used in conjunction with the specific effective width
= 3 expressions and is not actually the elastic buckling coeffi-
cient for distortional buckling.1.412,3.226 In addition, the tests
conducted by Yu and Schafer showed that the AISI effec-
4.2.4 Distortional Buckling Strength
tive width method is inadequate to account for distortional
The flexural strength of cold-formed steel beams bending buckling.4.206,4.210
about the major axis may be limited by local buckling, or The introduction of the Direct Strength Method3.254,3.255
lateral–torsional buckling. For members with edge-stiffened provided a means to determine the member strength due to
flanges, the flexural strength may also be limited by distor- distortional buckling. As described in Section 3.5, the key to
tional buckling. As shown in Fig. 4.31, the local buckling determine the member strength due to distortional buckling is
mode of a C-section for major-axis bending consists of buck- to obtain the distortional buckling moment, 𝑀crd . Numerical
ling of the compression portion of the web, the compres- and analytical solutions on how to obtain 𝑀crd are discussed
sion flange, and edge stiffener without movement of the in detail in Section 3.5.2.
line junction between the flange and edge stiffener. For this In 2004, the distortional buckling strength design provi-
type of limit state, the section strength of the member is sions developed by Shafer and Pekoz3.254,3.255 were adopted
156 4 FLEXURAL MEMBERS

Figure 4.31 C-section purlin buckling stress versus half wavelength for major-axis bending.1.69

in Appendix 1 of the Norther American Specification. In F4.2 Members With Holes


2012, the distortional buckling strength for members with The nominal flexural strength [resistance], 𝑀nd , for distortional
holes3.273–3.275 and inelastic reserve strength4.296 were added. buckling shall be calculated in accordance with Specification
These provisions are then incorporated into the main body Section F4.1, except 𝑀crd shall be determined including the
of the North American Specification in 2016. These design influence of holes, and when λd ≤ λd2 then:
provisions are excerpted below.
For λd ≤ λd1
𝑀nd = 𝑀ynet (4.106)
F4.1 Members Without Holes
The nominal flexural strength [resistance], 𝑀nd , shall be calcu- For λd1 > λd ≤ λd2
lated in accordance with Eq. (4.101) or Eq. (4.102). ( )
Mynet − Md2
For λd ≤ 0.673 𝑀nd = Mynet − (λd − λd1 )
𝑀nd = 𝑀y (4.101) λd2 − λd1
( )0.5 ( )0.5
For λd > 0.673 ⎡ Mcrd ⎤ M

≤ 1 − 0.22 ⎥ crd
My (4.107)
( )0.5 ( )0.5 ⎢ ⎥ My
⎡ 𝑀crd ⎤ 𝑀 ⎣
My

𝑀nd = ⎢1 − 0.22 ⎥ crd
𝑀y (4.102)
⎢ 𝑀y ⎥ 𝑀y
⎣ ⎦ where √
where √ λd = My ∕Mcrd (4.108)
λd = 𝑀y ∕𝑀crd (4.103)

𝑀y = 𝑆fy 𝐹y (4.104) where 𝑀crd = Distortional buckling moment including


influence of holes
where 𝑆fy = Elastic section modulus of full unreduced λd1 = 0.673 (Mynet ∕My )3 (4.109)
cross-section relative to extreme fiber in first
yielding λd2 = Limit of distortional slenderness transition
𝐹y = Yield stress
= 0.673 [1.7(My ∕Mynet )2.7 − 0.7] (4.110)
𝑀crd = Sf Fcrd (4.105)
𝑀d2 = [1-0.22(1∕λd2 )](1∕λd2 )My (4.111)
where 𝑆f = Elastic section modulus of full unreduced
cross-section relative to extreme compression 𝑀y = Member yield moment as given in
fiber Eq. (4.104)
𝐹crd = Elastic distortional buckling stress calculated in 𝑀ynet = Member yield moment of net cross-section
accordance with Specification Appendix 2 as given in Eq. (4.95)
BENDING STRENGTH AND DEFLECTION 157
F4.3 Members Considering Distortional Inelastic The yield moment [(Eq. (4.104)] is
Reserve Strength
Inelastic reserve capacity is permitted to be considered as 𝑀y = 𝑆f𝑦 𝐹y = 𝑆f 𝐹y = (4.11)(50) = 205.50 in.-kips
follows, provided λd ≤ 0.673:
The slenderness factor for distortional buckling
[(Eq. (4.103))] is
(a) Sections symmetric about the axis of bending or sections √ √
with first yield in compression: 𝑀y 205.50
𝜆d = = = 1.71
𝑀nd = 𝑀𝑦 + (1 − 1∕𝐶yd
2
)(𝑀𝑝 − 𝑀𝑦 ) (4.112) 𝑀crd 70.45
(b) Sections with first yield in tension: Since λd > 0.673, the nominal moment for distortional
buckling can be computed according to Eq. (4.102) as
𝑀nd = 𝑀yc + (1 − 1∕𝐶yd
2
)(𝑀𝑝 − 𝑀yc ) ≤ 𝑀yt3 (4.113)
follows:
where [ ( ) ]( )
√ 𝑀crd 0.5 𝑀crd 0.5
λd = My ∕Mcrd
𝑀nd = 1-0.22 𝑀y
(4.114) 𝑀y 𝑀y
√ [
𝐶yd = 0.673∕𝜆𝑑 ≤ 3 ( ) ]( )
(4.115) 70.45 0.5 70.45 0.5
= 1-0.22 (205.50)
205.50 205.50
𝑀crd = Critical elastic distortional buckling moment,
determined in accordance with Specification = 104.82 in.-kips
Appendix 2 or Section 2.3.3.3 in this volume
Based on ASD, the allowable moment for distortional
𝑀p = Member plastic moment as given in Eq. (4.73)
𝑀y = Member yield moment in accordance with
buckling is
Specification Section F2.1 𝑀nd 104.82
𝑀yc = Moment for yield in compression as defined in 𝑀a = = = 62.77 in.-kips
Ωb 1.67
Specification Section F3.2.3
𝑀yt3 = Maximum moment for yielding in tension as Based on LRFD, the design moment for distortional buck-
given in Eq. (4.100) ling is
𝜙𝑏 𝑀nd = 0.9(104.82) = 94.34 in.-kips

Example 4.10 For the C-section used in Example 4.5 B. Distortional Buckling Strength Based on Section 3.5.2(2)
(See Fig. 4.14 for cross-section dimensions), determine From Example 3.8, the distortional buckling stress based on
the available moment for distortional buckling according more accuracy method presented in Section 3.5.2(2) is
to ASD and LRFD. Use the elastic distortional buckling
stresses based on the simplified provision of Section 3.5.2(3) 𝐹crd = 32.98 ksi
and the more precise provision of Section 3.5.2(2). Deter- and
mine the member strength assuming the member is fully
𝑀crd = 135.55 in.-kips
braced against lateral–torsional buckling.
Also from item A, 𝑀y = 205.50 in.-kips. The slenderness
SOLUTION factor for distortional buckling [Eq. (4.103)] is
A. Distortional Buckling Strength Based on Section 3.5.2(3) √ √
𝑀y 205.50
The distortional buckling stress of the C-section has been 𝜆d = = = 1.231 > 0.673
calculated in Example 3.8. From the example, the distor- 𝑀crd 135.55
tional buckling stress based on the simplified method is: The nominal moment for distortional buckling according
𝐹crd = 17.14 ksi to Eq. (4.102) is
[ ( ) ]( )
The elastic section modulus relative to the compression 𝑀crd 0.5 𝑀crd 0.5
𝑀nd = 1-0.22 𝑀y
fiber is 𝑀y 𝑀y
𝑆f = 4.11 in.3 [ ( ) ]( )
135.55 0.5 135.55 0.5
and the distortional buckling moment = 1-0.22 (205.50)
205.50 205.50
𝑀crd = 70.45 in.-kips = 137.08 in.-kips
158 4 FLEXURAL MEMBERS

Based on ASD, the allowable moment for distortional SOLUTION


buckling is 1. Determine elastic local buckling moment:
137.08 Since the local buckling wavelength is short, the local
𝑀a = = 82.08 in.-kips buckling could occur within the hole length or between
1.67
the holes. Therefore, the local buckling of both the gross
Based on ASD, the design moment for distortional buck- and net cross-sections should be considered. The software
ling is CUFSM is used to determine the buckling moment. For
𝜙𝑏 𝑀nd = 0.9(137.08) = 123.37 in.-kips the net cross-section, the section properties can be calcu-
lated using CUFSM by setting element thickness at the hole
It is noted that the distortional buckling strength based on equals 0. The signature curves for gross cross-section and
the more accurate buckling analysis provides 31% higher cross-section with hole are shown in Figs. 4.33(a) and (b).
moment than the strength predicted based on simplified The yield moment based on gross and net cross-section
buckling stress prediction. can be obtained from the program output:

𝑀y = 25.648 in.-kips., and 𝑀ynet = 25.507 in.-kips


C. Member Strength
Since the member does not subject the lateral–torsional From Fig. 4.33(a), the local buckling between the holes,
buckling, the member strength is controlled by local 𝑀cr𝓁 = 1.45𝑀y with buckling wavelength 𝐿cr𝓁 = 3.2 in.,
and distortional buckling. From the calculation in and within the hole (Fig. 4.33(b)), 𝑀cr𝓁 = 0.78𝑀ynet and
Example 4.5, the allowable and design moments due to local 𝐿cr𝓁 = 4.4 in., which is less than the hole length (4.5 in.).
buckling are: Therefore, it is possible that local buckling takes place within
the hole length. Comparing both local buckling moments, the
ASD method: 𝑀a = 96.14 in.-kips local buckling within the hole controls: 𝑀cr𝓁 = 0.78𝑀ynet =
LRFD method: 𝜙b 𝑀𝑛𝓁 = 144.50 in.-kips 19.90 in.-kips.
The member strength based on local buckling can be deter-
Comparing with the member strengths due to distortional mined using Section 4.2.3.2:
buckling calculated under Item B, the member strength is Since the member is fully braced, 𝑀ne = 𝑀y .
controlled by distortional buckling. Therefore the member
strength is: √ √
λ𝓁 = 𝑀ne ∕𝑀crl = 25.648∕19.90 = 1.135 > 0.776

ASD method: 𝑀a = 82.08 in.-kips [ ( ) ]( )


LRFD method: 𝜙b 𝑀n = 123.37 in.-kips 𝑀crl 0.4 𝑀crl 0.4
𝑀𝑛𝓁 = 1-0.15 𝑀ne
𝑀ne 𝑀ne
In Tables II-7, II-8, and II-9 of the AISI Design Manual, [ ( ) ]( )
19.90 0.4 19.90 0.4
the computed distortional buckling properties are provided = 1-0.15 (25.648)
for the representative C-shapes, stock studs/joist, and 25.648 25.648
Z-shapes with lips, respectively. The values in these tables = 20.02 in.-kips < 𝑀ynet OK
have been calculated for use with Section 3.5.2(2). Examples
of using finite strip method to determine the local, distor- The allowable moment for ASD method:
tional buckling moments are also provided in Examples
II-2B, II-4B, II-6B, II-7B and II-15 of the 2017 edition of 𝑀a = 𝑀𝑛𝓁 ∕Ωb = (20.02)∕(1.67) = 11.99 in.-kips
the AISI Design Manual.1.428
The design moment for LRFD method:

ϕb 𝑀𝑛𝓁 = (0.9)(20.02) = 18.02 in.-kips


Example 4.11 A beam with geometry as shown in
Fig. 4.32 with web hole at mid-depth of the web. The
hole depth 𝑑h = 1.5 in., and hole length 𝐿h = 4.5 in. Hole 2. Determine distortional buckling moment, 𝑀nd :
spacing 𝑠 = 24.0 in. o.c. Assume that the beam is laterally
braced against lateral–torsional buckling. The member yield For a member with a hole, the distortional buckling
stress 𝐹y = 33 ksi. Determine member strength using the moment can be estimated using the method described in
Direct Strength Method. Section 3.5.3.2 by modifying the web thickness to take into
BENDING STRENGTH AND DEFLECTION 159

y y

1.625 in.
0.500 in.

0.0712 in.

0.0451 in.

6.000 in. x dh = 1.5 in. x

Section A-A Section B-B

Figure 4.32 Example 4.11.

2.5
Load factor

1.5 Lcrd = 15.3 in.


Lcrl = 3.2 in. L.F. = 1.65
1 L.F. = 1.45

0.5

0
100 101 102
Length

Figure 4.33 (a) Local and distortional buckling of gross cross-section.

consideration of the hole influence and then calculating the The modified thickness of the web is calculated per Eq.
distortional buckling at the wavelength where distortional (3.132):
( )1∕3 ( )
buckling takes place at its gross cross-section. 𝐿 4.5 1∕3
𝑡𝑟 = 𝑡 1 − ℎ = (0.0451) 1 −
From Fig. 4.33(a), the distortional buckling occurs at 𝐿crd 15.3
wavelength 𝐿crd = 15.3 in. = 0.04016 in.
160 4 FLEXURAL MEMBERS

3. Member Strength:
2 Comparing the available moments calculated under
items 1 and 2, the local buckling controls the design. The
member strength is therefore,
1.5
ASD method: 𝑀a = 11.99 in.-kips
Load factor

LRFD method: ϕb Mn = 18.02 in.-kips


1

Lcrl = 4.4 in. Example 4.12 A joist (800S200-97) with its cross-section
0.5 L.F. = 0.78 and section properties shown in Fig. 4.34. The beam is braced
at 24 in. o.c. Determine the member strength using the Direct
Strength Method.
0
100 101
Length SOLUTION
A. Global buckling strength:
Figure 4.33 (b) Local buckling of cross-section with hole. The global buckling strength can be determined in accor-
dance with Section 4.2.2. For C-section members, the global
Modify the thickness of the flat width of the web and calcu- buckling moment is determined by Eq. (4.54):
late the distortional buckling moment at the 𝐿crd = 15.3 in., Cb ro A √
𝐹cre = 𝜎ey 𝜎t
𝑀crd = 1.46𝑀y . Comparing with the distortional buckling Sf
moment (𝑀crd = 1.65𝑀y ) of the gross cross-section, the
distortional buckling moment is reduced about 12%.
The nominal moment due to distortional buckling can be
determined per Section 4.2.4 (Specification Section F4.2): y
√ √
𝑀𝑦 1 2.000 in.
𝜆𝑑 = = = 0.828
𝑀crd 1.46
√ √ 0.625 in.
𝑀𝑦 25.5 A = 1.27 in.2
𝜆𝑑 = = = 0.7785 R = 0.1525 in.
𝑀crd 42.07 Ix = 11.2 in.4

λd2 = 0.673 [1.7(My ∕Mynet )2.7 − 0.7] Iy = 0.577 in.4

= (0.673)[1.7(25.648∕25.507)2.7 − 0.7] = 0.212 Sf = 2.8 in.3


8.000 in.
J = 0.00438 in.4
Since λd > λd2 , the nominal moment due to distortional
buckling is determined using Specification Section F4.1 Cw = 7.68 in.6
x
except 𝑀crd includes the hole influence: rx = 2.97 in.
Since λd > 0.673, ry = 0.674 in.
[ ( ) ]( ) t = 0.1017 in.
𝑀crd 0.5 𝑀crd 0.5 ro = 3.28 in.
𝑀nd = 1 − 0.22 𝑀𝑦
𝑀𝑦 𝑀𝑦

= [1 − 0.22(1.46)0.5 ](1.46)0.5 (25.648) = 23.64 in.-kips


The allowable moment for ASD method:
𝑀a = 𝑀nd ∕Ωb = (23.64)∕(1.67) = 14.16 in.-kips
The design moment for LRFD method:
𝜙b 𝑀nd = (0.9)(23.64) = 21.28 in.-kips Figure 4.34 Example 4.12.
BENDING STRENGTH AND DEFLECTION 161

where 𝜎ey and 𝜎t are determined by Eqs. (4.57) and (4.58), 140∕1071 − 0.23
respectively: ≤ Mp = 169.85 − (169.85 − 140)
0.37
Assume the member unbraced length 𝐾y 𝐿y = 24 in. and
𝐾t 𝐿t = 24 in. and 𝐶b = 1.0 = 159.2 in.-kips < 169.8 in.-kips OK

π2 E π2 (29,500) B. Local buckling strength interacting with global buckling:


𝜎ey = = = 229.6 ksi Using CUFSM program, the following results are obtained
(Ky Ly ∕ry )2 (24∕0.674)2
(Fig. 4.35):
[ ] 𝑀y = 141.66 in.-kips (note the numerical analysis result
1 𝜋 2 EC𝑤 1
𝜎𝑡 = GJ + = may differ from the value calculated using the linear
Ar0 (𝐾𝑡 𝐿𝑡 )2 (1.27)(3.28)
[ ] method.)
𝜋 2 (29,500)(7.68) Load factor for local buckling (i.e., 𝑀cr𝓁 ∕𝑀y ), LF local =
× (11,300)(0.00438) +
(24)2 2.76
Load factor for distortional buckling (i.e., 𝑀crd ∕𝑀y ),
= 287.7 ksi
LF dist = 2.13
(1.0)(3.28)(1.27) √ Therefore, the local and distortional buckling moments are
𝐹cre = (229.6)(287.7) = 382.4 ksi
2.80 𝑀cr𝓁 = 2.76𝑀y = 2.76(141.66) = 390.98 in.-kips
𝑀cre = 𝐹cre 𝑆f = (382.4)(2.8) = 1071 in.-kips 𝑀crd = 2.13𝑀y = 2.13(141.66) = 301.74 in.-kips

Since 𝐹cre > 2.78𝐹y = 139 ksi, inelastic reserve strength The local buckling slenderness factor is calculated:
√ √
can be considered using Specification Section F2.4.2: 𝜆𝓁 = 𝑀𝑦 ∕𝑀cr𝓁 = 1∕2.76 = 0.602
The plastic section modulus is determined as follows:
Since the section is symmetric, the centroid x axis is at the Since 𝑀ne > 𝑀y and λ𝓁 < 0.776, inelastic reserve
mid-height. Determine the centroid of the section above the strength can be considered by using Specification Section
centroid x axis using linear method (Fig. 1.32): F3.2.3:
√ √
C𝑦𝓁 = 0.776∕λ𝓁 ≤ 3 = 0.776∕0.602 = 1.135 < 3 OK
Elements Length, 𝑙i Dist. To top fiber, 𝑦i 𝑙i 𝑦i For section symmetric about the axis of bending, the
nominal moment is determined by Eq. (4.96)
lip 0.3708 0.4396 0.1630
Flange 1.4916 0.0508 0.0758 M𝑛𝓁 = My + (1 − 1∕C𝑦𝓁
2
)(Mp − My ) = (141.66)
Web 3.7458 2.1271 7.9677
Corners 2x0.3193 0.1247 0.0796 + (1–1∕1.1352 )(169.85 − 141.66) = 148.0 in.-kips
Total 6.2467 8.2861 C. Distortional buckling strength:
As described in Item B, the distortional buckling moment
from CUFSM program:
Centroid of the portion above the cross-section neutral
axis, 𝑦′top = 4 − 8.2861∕6.2467 = 2.6735 in. 𝑀crd = 301.74 in.-kips
The plastic section modulus: The distortional buckling slenderness factor is calculated
𝑍f = (A∕2)(2y′c ) = (6.2467)(0.1017)(2 × 2.6735) using Eq. (4.103):
√ √
= 3.3969 in 3 λd = My ∕Mcrd = 1∕2.13 = 0.685

𝑀p = 𝑍f 𝐹y = (3.3969)(50) = 169.85 in.-kips Since λd > 0.673, inelastic reserve strength cannot be
considered, the distortional buckling strength is determined
𝑀y = 𝑆f 𝐹y = (2.8)(50) = 140 in.-kips by Specification Section F4.1:
[ ( ) ]( )
Since 𝑀cre > 2.78𝑀y , nominal moment considering 𝑀crd 0.5 𝑀crd 0.5
inelastic reserve is determined by Eq. (4.71): 𝑀nd = 1 − 0.22
𝑀𝑦 𝑀𝑦

My ∕Mcre − 0.23 𝑀𝑦 = [1 − 0.22(2.13)0.5 ](2.13)0.5 (141.66)
Mne = Mp − (Mp − My )
0.37 = 140.4 in.-kips
162 4 FLEXURAL MEMBERS

8
7
6
Load factor

5
4
3
2 Lcrl = 4.6,
L.F. = 2.76 Lcrl = 13.5,
1 L.F. = 3.13

0
100 101 102
Length

Figure 4.35 Signature Curve of 800S200-97.

D. Member strength:
By checking the cross-section dimensions (not shown), the
cross-section satisfies the limits given in Table 3.1. Therefore
the safety and resistance factors provided in Chapter F are
applicable.
Since the same safety factors for ASD method and the
same resistance factors for LRFD method are applied
to member strengths due to different buckling failures,
the member strength can be determined by the minimum
nominal strength as calculated in Items A to C.
𝑀n = minimum (𝑀ne , 𝑀𝑛𝓁 , 𝑀nd )
= minimum(159.2, 148.0, 140.4) = 140.4 in.-kips Figure 4.36 Three possible types of supporting elastic frame for
The available strengths are calculated: equivalent column.4.19

ASD method: 𝑀a = 𝑀n ∕Ωb = 140.4∕1.67 = 84.1 in.-kips Z-shaped sections for which the entire cross section rotates
LRFD method: 𝜙b 𝑀n = 0.9(140.4) = 126.3 in.-kips and deflects in the lateral direction as a unit. But this is not
the case for U-shaped beams and the combined sheet stiffener
In this example, the distortional buckling controls the sections as shown in Fig. 4.36. For the latter, when it is loaded
design. It is possible to increase the member distortional in such a manner that the brims and the flanges of stiffeners
buckling strength by restraining the compression flange from are in compression, the tension flange of the beams remains
rotating about the flange and web juncture, thus increase straight and does not displace laterally; only the compres-
the member distortional buckling moment. For example, by sion flange tends to buckle separately in the lateral direction,
attaching the structural sheathing to the compression flange, accompanied by out-of-plane bending of the web, as shown
the structural sheathing is capable to provide the rotational in Fig. 4.37, unless adequate bracing is provided.
stiffness, 𝑘ϕ , which can be determined in accordance with Prior to 2004, this buckling phenomenon was considered
Section 3.5.2(4). By including the term 𝑘ϕ in Eq. (3.105), or as lateral–torsional buckling. With the introduction of distor-
in the numerical analysis, the distortional buckling moment tional buckling design provisions in the Supplement 1 to
is expected to be increased. the 2001 North American Specification1.343 , the flange along
with web buckling is considered as a distortional buckling,
4.2.4.1 Laterally Unbraced Compression Flanges The and the member strength can be determined using the Direct
problems discussed in Sections 4.2.2 and 4.2.3 dealt with the Strength Method. In the following section, both approaches
type of lateral–torsional buckling of I-beams, C-sections, and are discussed:
BENDING STRENGTH AND DEFLECTION 163

Figure 4.38 Force normal to buckled web.4.19

where 𝐴web is the area of the web and 𝐶c and 𝐶t are the
distance from the neutral axis to the extreme compression
fiber and the extreme tension fiber, respectively (Fig. 4.38).
Consequently, the equation of equilibrium of the compres-
sion flange is
[ ] 2
𝑑4𝑥 𝐴web 𝑑 𝑥a
Figure 4.37 Force normal to buckled flange.4.19 EI f 4a + 𝜎cr 𝐴f + =0
dz 12𝐶 c ∕(3𝐶 c − 𝐶 t ) dz2
(4.119)
and the corresponding nontrivial eigenvalue leads to
4.2.4.1.1 Considering the Unstable Flange as Lateral
𝜋2𝐸
Torsional Buckling The precise analysis of the lateral– 𝜎cr = (4.120)
torsional buckling of U-shaped beams is rather complex. Not (𝐿∕𝑟)2
only to the compression flange and the compression portion where
of the web act like a column on an elastic foundation, but √
also the problem is complicated by the weakening influence 𝐼f
𝑟= (4.121)
of the torsional action of the flange. For this reason, the 𝐴f + 𝐴web ∕[12𝐶c ∕(3𝐶c − 𝐶t )]
design procedure for determining the allowable design stress
which is the radius of gyration of the effective column
for laterally unbraced compression flanges has been based
consisting of the compression flange and a part of
on the considerable simplification of an analysis presented
the compression portion of the web having a depth of
by Douty in Ref. 4.19. See Section 2 of Part V of the 2002
[(3𝐶c − 𝐶t )∕12𝐶c ]𝑑, where d is the depth of the beam.
edition of the AISI Cold-Formed Steel Design Manual.1.340
The above analysis is for the type of column supported on
When the compression flange of a U-shaped beam is
an elastic foundation where the elastic support is provided
subject to the critical bending forces 𝜎cr 𝐴f (𝜎cr being the crit-
by the remaining portion of the web and the tension flange
ical stress and 𝐴f the area of the flange), the component of
acting together as an elastic frame. The effect of torsional
these forces normal to the buckling flange is
weakening in the combined flexural–torsional stability of
𝑑 2 𝑥a the effective column can be determined by the theorem of
𝑞f = 𝜎cr 𝐴f (4.116) minimum potential energy4.19 :
dz2
See Fig. 4.37. In the same manner, the component on a unit 𝑈 = 𝑉1 + 𝑉2 + 𝑈w
strip of the buckled web as shown in Fig. 4.38 is L
1
= [EI 𝑦 (𝑢′′ )2 + ECw (𝜙′′ )2 + GJ(𝜙′ )2 ]dz
𝑑2𝑥 2 ∫0
𝑞w = 𝜎𝑡w (4.117)
dz2 1
𝐿
+ (𝐶1 𝑢2 − 2𝐶2 u𝜙 + 𝐶3 𝜙2 )dz (4.122)
As a result, the total lateral force 𝑅a transmitted to the 2 ∫0
compression flange by the buckled web is 𝐿 [ ( )
𝐼p
]
𝑃 ′ 2 ′ ′ ′ 2
− (𝑢 ) + 2𝑦0 𝑢 𝜙 + (𝜙 ) dz
𝐴web 𝑑 2 𝑥a 2 ∫0 𝐴
𝜎cr (4.118)
12𝐶c ∕(3𝐶c − 𝐶t ) dz2
164 4 FLEXURAL MEMBERS

where 𝑈 = change in entire potential energy of system and T, the torsional reduction factor, is determined as
consisting of effective column and its follows: {

supporting elastic frame 𝑇 = 𝑇0 = if 𝐿 ≥ 𝐿′ (4.125)
ℎ + 3.4𝑦0
𝑉1 = strain energy accumulated in bent and
( ) ( )( )
twisted column 𝐿 ℎ 𝐿
𝑇0 = if 𝐿 < 𝐿′ (4.126)
𝑉2 = strain energy of deflected supporting frame 𝐿′ ℎ + 3.4𝑦0 𝐿′
𝑈w = change in potential energy of external forces √ √
acting on system where 𝐿′ = 𝜋 4 2𝐼(ℎ∕𝑡)3 = 3.7 4 𝐼(ℎ∕𝑡)3
𝐼 = moment of inertia of column about its 𝑦0 = distance from centroid of equivalent column
vertical y axis to its shear center
ℎ = distance from tension flange to centroid of
𝑈 = horizontal displacement of shear center
equivalent column
Φ = rotation of column
𝐽 = torsional constant of column For beams with a large distance between bracing, the
following expression for 𝑃cr may be used:
𝑦0 = vertical distance between shear center and √
centroid of column 𝑃cr = 𝑇0 4𝛽EI (4.127)
𝐼p = polar moment of inertia of column about its From the value of 𝑃cr given above, the equivalent slender-
shear center ness ratio (𝐿∕𝑟)eq can then be determined as follows:
𝐶w = warping constant √
( )
𝐶1 = 𝛿𝜙 ∕(𝛿u 𝛿𝜙 − 𝛿u𝜙
2 ) 𝐿 𝜋2𝐸 490
=𝑘 =√ (4.128)
𝑟 eq 𝑃cr ∕𝐴c 𝑃 ∕𝐴
𝐶2 = 𝛿u𝜙 ∕(𝛿u 𝛿𝜙 − 𝛿u𝜙2 ) cr c

𝐶3 = 𝛿u ∕(𝛿u 𝛿𝜙 − 𝛿u𝜙 )
2 where k is an experimental correction factor for the postbuck-
δu = horizontal displacement of shear center due ling strength and equals 1/1.1 and 𝐴c is the cross-sectional
to unit load area of the equivalent column.
The allowable compression stress 𝐹a for the ASD method
δu𝜙 = horizontal displacement of shear center due can be computed from the column formula (Chapter 5) on the
to unit moment basis of this equivalent slenderness ratio. To obtain the allow-
𝛿ϕ = rotation of column due to unit moment able compression bending stress in the extreme compression
fiber 𝐹b′ , the axial stress 𝐹a may be extrapolated linearly from
By solving Eq. (4.122) and applying considerable simpli- the centroid level and adjusted for the different factors of
fications, the following expressions can be obtained for safety used for beam yielding and column buckling, that is,
( )
the stability of the effective column on an elastic foun- Ω 𝐶c
𝐹b′ = c 𝐹a (4.129)
dation taking the torsional weakening of the flange into Ωb 𝑦c
consideration4.19 :
( ) where Ωc = safety factor used for column buckling
⎧𝑇 1 + 𝛽𝐿2 𝑃 when 𝛽𝐿2 ∕𝑃 ≤ 30 (4.123)
⎪ 𝜋 2 𝑃e e e Ωb = safety factor used for beam yielding
⎪ ( √ ) 𝑦c = distance from neutral axis of beam to centroid
𝑃cr = ⎨𝑇 0.6 + 2 𝛽𝐿2
𝑃e of equivalent column
⎪ 𝜋 𝜋 2 𝑃e
⎪ The design method developed in Ref. 4.19 has been
⎩ when 𝛽𝐿2 ∕𝑃e > 30 (4.124)
compared with the results of more than 100 tests (Fig. 4.39).
where 𝑃cr =
critical load of equivalent column It has been found that discrepancies are within about
30% on the conservative side and about 20% on the
𝑃e =
Euler critical load, π2 EI∕𝐿2
nonconservative side.
𝛽 =
spring constant, 1∕𝐷 Based on the analysis and simplifications, the following
𝐷 =
lateral deflection of column centroid due to a 10-step design procedure has been included in the AISI
unit force applied to web at level of column design manual since 1962.1.159 1.349
centroid
𝐿 = unbraced length of equivalnt column 1. Determine the location of the neutral axis and define
as the “equivalent column” the portion of the beam
BENDING STRENGTH AND DEFLECTION 165

( )3

𝐿′ = 3.7
4
𝐼 (4.134)
𝑡

where 𝐼 = moment of inertia of equivalent


column about its gravity axis
parallel to web, in.4
𝐿 = unbraced length of equivalent
column, in.

If 𝐶 ≤ 30, compute
( )
𝛽𝐿2
Figure 4.39 Comparison between analysis and tests.4.19 𝑃cr = TPe 1 + 2 (4.135)
𝜋 𝑃e
If 𝐶 > 30, compute

from the extreme compression fiber to a level that is
𝛽𝐿2
a distance of [(3𝐶c − 𝐶t )∕12𝐶c ]𝑑 from the extreme 𝑃cr = TPe (0.60 + 0.635) (4.136)
𝑃e
compression fiber. In this expression, 𝐶c and 𝐶t are
the distances from the neutral axis to the extreme In both cases, if 𝐿 ≥ 𝐿′′ ,
compression and tension fibers, respectively, and d is
𝑇 = 𝑇0
the depth of the section.
2. Determine the distance 𝑦0 measured parallel to the and if 𝐿 < 𝐿′ ,
web from the centroid of the equivalent column to 𝑇0 𝐿
𝑇 = (4.137)
its shear center. (If the cross section of the equivalent 𝐿′
column is of angle or T shape, its shear center is at 5b. If the flange is braced at less than two points, compute
the intersection of the web and flange; if of channel √
𝑃cr = 𝑇0 4𝛽EI (4.138)
shape, the location of the shear center is obtained from
Section 4.4. If the flanges of the channel are of unequal 6. Determine the slenderness ratio of the equivalent
width, for an approximation take w as the mean of the column, ( )
KL 490
two flange widths, or compute the location of the shear =√ (4.139)
center by rigorous methods. See Appendix B.) 𝑟 eq 𝑃cr ∕𝐴c
3. To determine the spring constant 𝛽, isolate a portion where 𝐴c is the cross-sectional area of the equivalent
of the member 1 in. (25.4 mm) long, apply a force of column.
0.001 kip (4.45 N) perpendicular to the web at the 7. From Eqs. (5.51), and (5.52) compute the stress 𝐹n
level of the column centroid, and compute the corre- corresponding to (KL∕𝑟)eq .
sponding lateral deflection D of the centroid. Then the 8. The design compression bending stress based on
spring constant is previous factors of safety is
0.001 ( )
𝛽= (4.130) 𝐶c
𝐷 𝐹b2 = 1.15𝐹n ≤ 𝐹y (4.140)
𝑦c
4. Calculate
ℎ where 𝐶c = distance from neutral axis of beam to
𝑇0 = (4.131)
ℎ + 3.4𝑦0 extreme compression fiber
where h is the distance from the tension flange to the 𝑦c = distance from neutral axis of beam to
centroid of the equivalent column in inches. centroid of equivalent column

5a. If the flange is laterally braced at two or more points, The critical moment is 𝑀c = 𝐹b2 𝑆f .
calculate Use Eq. (4.74) to compute 𝑀n .
290,000𝐼
𝑃e = (4.132)
𝐿2 Example 4.13 Determine the design compression bending
𝛽𝐿2 stress in the compression flanges (top flanges of the U-shaped
𝐶= (4.133) section shown in Fig. 4.40. Assume that the compression
𝑃e
166 4 FLEXURAL MEMBERS

Figure 4.41 Equivalent column.

Figure 4.40 Example 4.13. and 4.44). The centroid of the equivalent column can
be located as follows:

flanges are laterally braced at the third points with unbraced


lengths of 48 in. The yield point of steel is 33 ksi. Distance
from
Top
SOLUTION Fiber
1. Location of Neutral Axis and Determination of Element Area 𝐴(in.2 ) y (in.) Ay (in.3 )
Equivalent Column (Fig. 4.40)
a. Location of Neutral Axis. 1 1.0475(0.105) = 0.1100 0.0525 0.0058
2 0.0396 (Table 4.4) = 0.0396 0.1373 0.0055
6 1.2655(0.105) = 0.1329 0.9253 0.1230
Distance
Total 0.2825 0.1343
from
Top 0.1343
𝑦cg = = 0.475 in.
Fiber Ay 0.2825
Element Area A (in.2 ) y (in.) (in.3 )
1 2(1.0475)(0.105) = 0.2200 0.0525 0.0116 From Appendix B it can be seen that the shear center
2 2(0.0396) (Table 4.4) = 0.0792 0.1373 0.0109 of an angle section is located at the intersection of two
3 2(7.415)(0.105) = 1.5572 4.0000 6.2288 legs. Therefore, the distance y0 between the centroid
4 2(0.0396) = 0.7920 7.8627 0.6227 and the shear center of the equivalent column is
5 7.415(0.105) = 0.7786 7.9475 6.1879 1
𝑦0 = 𝑦cg − 𝑡 = 0.475 − 0.0525 = 0.4225 in
Total 2.7142 13.0619 2
13.0619 3. Calculation of Spring Constant 𝜷. The spring
𝐶c = = 4.812 in.
2.7142 constant 𝛽 can be computed from Eq. (4.130) as
𝐶t = 8.0 − 4.812 = 3.188 in.
0.001
𝛽=
𝐷
b. Equivalent Column. Based on step 1 of the proce- for a portion of the member 1 in. in length. Here D is
dure, the equivalent column used in the design is an the lateral deflection of the column centroid due to a
angle section as shown in Fig. 4.41. The depth of the force of 0.001 kip applied to the web at the level of the
equivalent column can be determined as follows: column centroid (Fig. 4.42). Using the moment–area
( ) [ ] method (Fig. 4.43), the deflection D can be computed:
3𝐶c − 𝐶t 3(4.812) − 3.188
𝑑= × 8.00
12𝐶c 12(4.812) (7.4725)3 (7.4725)2 (7.895)
𝐷= +
= 1.558 in 3EI × 103 2EI × 103
where E = 29.5 × 103 ksi and
2. Determination of y0 (Distance from Centroid of 1
Equivalent Column to Its Shear Center). (Figs. 4.41 𝐼= (0.105)3 = 96.5 × 10−6 in.4
12
BENDING STRENGTH AND DEFLECTION 167

Figure 4.42 Force applied to web for computing spring constant.


Figure 4.44 Dimensions of equivalent column.

The I values of the individual elements about their


own centroidal axes parallel to the web are
1
𝐼1′ = (0.105)(1.0475)3 = 0.0101
12
𝐼2′ = 0.0002
𝐼6′ = 0.0000
𝐼1′ + 𝐼2′ + 𝐼6′ = 0.0103
Figure 4.43 Lateral deflection of equivalent column. ∑
(Ax2 ) = 0.0745

Therefore 𝐼𝑧 = 0.0848 in.4


359.50 (∑ )
𝐷= − 𝐴 (𝑥2cg ) = −0.2825(0.362)2 = −0.0370
(29.5 × 103 )(96.5 × 10−6 )103
= 0.1263 in. 𝐼𝑦 = 0.0478 in.4
1 1
𝛽= = = 7.918 × 10−3 and
𝐷 × 103 126.3 𝐴 = 0.2825 in.2

4. Computation of T0 [Eq. (4.131)] Since the compression flange is braced at the third
points, the values of 𝑃e , C, and L′ can be computed
ℎ 7.525
𝑇0 = = = 0.840 from Eqs. (4.132)–(4.134):
ℎ + 3.4𝑦0 7.525 + 3.4(0.4225)
𝐼 290,000(0.0478)
𝑃e = 290,000 2 =
𝐿 482
5. Determination of Pcr . In order to determine 𝑃cr , we
= 6.016 kips
should first compute the moment of inertia of the
equivalent column about its y axis parallel to the web 𝛽𝐿2 7.918 × 10−3 (48)2
𝐶= = = 3.032
(Fig. 4.44) as follows: 𝑃e 6.016
√ √
( )3 ( )
′ 4 ℎ 4 7.525 3
Distance 𝐿 = 3.7 𝐼 = 3.7 0.0478
𝑡 0.105
from = 42.61 in.
Area z Axis,
Element 𝐴 (in.2 ) x (in.) Ax (in.3 ) Ax2 (in.4 ) Since 𝐶 < 30 and 𝐿 > 𝐿′ , from Eq. (4.135),
( )
𝛽𝐿2
1 0.1100 0.8163 0.0898 0.0733 𝑃cr = 𝑇0 𝑃e 1 + 2
2 0.0396 0.1373 0.0055 0.0008 𝜋 𝑃e
[ ]
6 0.1329 0.0525 0.0070 0.0004 7.918 × 10−3 (48)2
Total 0.2825 0.1023 0.0745 = 0.840(6.016) 1 +
𝜋 2 (6.016)
0.1023 = 0.840(6.016)(1 + 0.307)
𝑥cg = = 0.362 in.
0.2825
= 6.605 kips
168 4 FLEXURAL MEMBERS

6. Determination of (KL∕r)eq . For the equivalent column expression for distortional buckling and method B used a
[Eq. (4.139)] modified version of Douty’s formulation discussed in this
( ) section. It was concluded that method A gives somewhat
KL 490
=√ better results than method B and is consistent with the present
𝑟 eq 𝑃 ∕𝐴 cr c formulation for flexural, torsional, and torsional–flexural
490 buckling.
=√ = 101.3
6.60∕0.2825 According to the 2008 edition of the AISI Cold-Formed
7. Determination of Compression Stress Fn . From Steel Design Manual, this type of buckling problem can be
Eq. (5.54), solved by using the direct-strength method. For this reason,
the above 10-step design procedure has been removed from
𝜋2𝐸 𝜋 2 (29, 500) the 2008 edition of the Design Manual.
𝐹cre = =
(KL∕𝑟)2eq (101.3)2
= 28.37 ksi 4.2.4.1.2 Considering the Unstable Flange as Distor-
√ √
𝐹y tional Buckling When the compressed flange buckles later-
33
𝜆c = = ally, the flange along with the web rotates about the juncture
𝐹cre 28.37 of the web and the tension (stable) flange as illustrated in
= 1.08 < 1.5 Fig. 4.37. The member strength can be determined using the
2 Direct Strength Method, while the buckling moment can be
𝐹n = (0.658𝜆c )𝐹y = (0.6581.08 )(33)
2

determined numerically or analytically in accordance with


= 20.25 ksi Chapter 3. The following example illustrates how to use the
Direct Strength Method to determine the member strength.
8. Design Compression Bending Stress [Eq. (4.140)].
( )
𝐶𝑐
𝐹𝑏2 = 1.15𝐹𝑛 Example 4.14 Determine the member strength of the
𝑦𝑐
U-shaped section with its dimension shown in Fig. 4.40.
4.812 Assume that the compression flanges are laterally unbraced.
= 1.15 (20.25) = 25.84 ksi
4.337 The yield point of steel is 33 ksi.
< (𝐹𝑦 = 33 ksi) OK
Once the design compression bending stress is computed, SOLUTION
the critical or nominal moment can be calculated as 𝑀n = 1. Buckling moments.
𝐹b2 𝑆e . From Example 3.10, the numerical analysis output
shows that the U-shaped section is susceptible to local
In 1964, Haussler presented rigorous methods for deter- and distortional buckling. The local buckling occurs at
mining the strength of elastically stabilized beams.4.20 In his the buckling wavelength Lcr𝓁 = 6.4 in. and the corre-
methods, Haussler also treated the unbraced compression sponding load factor = 2.39. The distortional buckling
flange as a column on an elastic foundation and maintained occurs at a buckling wavelength Lcrd = 67 in. and the
more rigor in his development. A comparison of Haussler’s corresponding load factor = 0.86. From the numerical
method with Douty’s simplified method indicates that the analysis, the yield moment My = 161 in.-kips. There-
latter may provide a smaller critical stress. fore, the local and distortional buckling moments are
In the early 1990s, the flexural behavior of standing seam obrained as:
roof panels with laterally unsupported compression flanges
was restudied by Serrette and Pekoz.4.158–4.162 Based on the 𝑀cr𝓁 = 6.4My = 384.8 in.-kips
available test data and the analytical results from elastic
𝑀crd = 0.86My = 138.5 in.-kips
finite-strip buckling analysis, the authors introduced two
design methods in Ref. 4.161 to estimate the maximum 2. Strength due to local buckling.
moment capacity of sections subjected to an interaction Per Section 4.2.3.2, the nominal moment due to local
between local and distortional buckling. It was assumed that buckling can be determined using Eqs. (4.91) and
distortional buckling may be taken as local overall buckling (4.92).
behavior. Both methods used the design philosophy currently Since the member is lateral stable,
used in the North American Specification for local–lateral
buckling interaction. Method A used a derived analytical 𝑀ne = 𝑀y = 161 in.-kips
BENDING STRENGTH AND DEFLECTION 169
Local buckling slenderness factor is calculated per Eq. virgin material. The effects of cold work were completely
(4.93): neglected.
√ √ When the effects of cold work are utilized in the deter-
λ𝓁 = Mne ∕Mcr𝓁 = 161∕384.8 = 0.647 < 0.776
mination of bending strength, the computation can be
Therefore, performed by one of the following two design approaches.
𝑀𝑛𝓁 = 𝑀ne = 161 in.-kips
1. Consider the increase in yield stress at corners due to
From cross-section dimension check not shown, the cold work and neglect the effects of cold work in all flat
section meets the limits given in Table 3.1. Therefore, portions of the section. As discussed in Chapter 2, the
the safety factor 1.67 for ASD and the resistance factor increase in yield stress can be found either by the use of
of 0.90 for LRFD can be used: Eq. (2.11) or by tests.
2. Consider the effects of cold work for corners and all flat
ASD method: 𝑀a = 𝑀𝑛𝓁 ∕Ωb = 161∕1.67 = 96.4 in.
elements. Equation (2.14) can be used to compute the
-kips
average yield stress of the entire section.
LRFD method: 𝜙b 𝑀𝑛𝓁 = (0.9)(161) = 144.9 in.-kips
3. Strength due to distortional buckling. In either design approach, the following procedures may be
The distortional buckling slenderness factor is calcu- used2.17 :
lated per Eq. (4.103):
√ √ 1. Subdivide the section into a number of elements.
λd = My ∕Mcrd = 161∕138.5 = 1.078 > 0.673 Assume a position of the neutral axis and the strain in
the top fiber. Compute the strains in various elements
Therefore, based on the assumed neutral axis and the top fiber
[ ( ) ]( )
Mcrd 0.5 𝑀crd 0.5 strain.
𝑀nd = 1 − 0.22 2. Determine the stresses from the stress–strain rela-
My My
tionship of the material in various elements for the
[ ( )0.5 ] ( )
138.5 138.5 0.5 computed strains.
My = 1 − 0.22 (161) 3. Locate the neutral axis by iteration until
161 161
= 118.86 in.-kips

The member strengths due to distortional buckling are 𝜎Δ𝐴 = 0
calculated is satisfied. Then the bending moment can be approxi-
ASD method: 𝑀a = 𝑀nd ∕Ωb = (118.86)∕(1.67) = mated by ∑
71.17 in.-kips 𝑀= 𝜎yΔ𝐴
LRFD method: 𝜙b Mn = (0.9)(118.86) = 106.97 in.-kip
where 𝜎 = stress
4. Member strength: ΔA = area for element
Since the member does not subject to lateral– 𝑦 = distance between center of gravity of each
torsional buckling the member strength is controlled element and neutral axis
by local and distortional buckling. By comparing the
member strengths due to local buckling (item 2) and
Results of the study by Winter and Uribe indicate that for
the strengths due to distortional buckling (item 3), the
the steels commonly used in thin-walled cold-formed steel
distortional buckling controls the design. Therefore
construction, considering the effects of cold work only in the
the member design strengths are:
corners of the formed sections, the moment capacities can
ASD method: 𝑀a = 71.17 in.-kips be increased by 4–22% compared with those obtained when
LRFD method: 𝜙b 𝑀n = 106.97 in.-kip neglecting cold work.2.17
If the effects of cold work are considered in both the flats
and the corners, the increase in bending strength ranges from
4.2.5 Effects of Cold Work on Bending Strength
17 to 41% above the virgin value.
The bending strength of cold-formed steel sections discussed It can be seen that a substantial advantage can be obtained
above was based on the mechanical properties of the by using the increase in strength of the material. Figure 4.45,
170 4 FLEXURAL MEMBERS

1. For unstiffened compression flanges,



𝐸
𝑤 = 0.43𝑡 (4.141)
𝑓
2. For stiffened compression flanges supported by a web
on each longitudinal edge,


𝐸
𝑤 = 1.28𝑡 (4.142)
𝑓

where 𝑤 = flat width for compression flange


𝑡 = thickness of steel
𝐸 = modulus of elasticity
𝑓 = maximum compressive edge stress in the
element without considering the safety factor

The economic design of continuous beams and long-span


purlins is discussed in Refs. 4.11 and 4.12.

Figure 4.45 Comparison of ultimate moments computed for three 4.2.7 Deflection of Flexural Members
different conditions.2.17 For a given loading condition, the deflection of flexural
members depends on the magnitude, location, and type of
the applied load, the span length, and the bending stiffness
reproduced from Ref. 2.17, shows a comparison of the EI, in which the modulus of elasticity in the elastic range is
ultimate moments computed for three different conditions. 29.5 × 103 ksi (203 GPa or 2.07 × 106 kg∕cm2 ) and I is the
It should be noted that the effects of cold work as shown in moment of inertia of the beam section.
Fig. 4.45 may not be directly applied to other configurations Similar to the bending strength calculation, the determina-
because the relative influence of corners or flats on the tion of the moment of inertia I for calculating the deflection
increase in bending strength depends mainly on the config- of steel beams can be calculated based on the either the Effec-
uration of the section and the spread between the tensile tive Width Method or the Direct Strength Method:
strength and yield stress of the virgin material. Attention
should be given to the limitations of Section A3.3.2 of the (a) The Effective Width Method is used: I is determined
North American Specification when the effects of cold work using the effective areas of the compression flange, edge
are used in design. stiffer, and beam web, for which the effective widths
are computed for the compressive stress developed from
4.2.6 Economic Design for Bending Strength the bending moment. If the compression flange, edge
stiffer, and beam web are fully effective, the moment
The above discussion and design examples are based on
of inertia is obviously based on the full section. In this
the fact that the allowable design moment is determined
case, the moment of inertia is a constant value along the
for a given section for which the dimensions are known.
entire beam length. Otherwise, if the moment of inertia
In the design of a new section, the dimensions are usually
is on the basis of the effective areas of the compression
unknown factors. The selection of the most favorable
flange, edge stiffener, and/or beam web, the moment
dimensions can be achieved by using the optimum design
of inertia may vary along the beam span because the
technique. This is a very complex nonlinear problem
bending moment usually varies along the beam length,
which can only be solved by computer analysis.1.247
as shown in Fig. 4.46.
However, if the depth and the thickness of the section
(b) The Direct Strength Method is used: I is considered
are known, previous study has shown that the maximum
linearly proportional to the strength of the section that
moment-to-weight ratio usually occurs in the neighborhood
is determined at the service stress of the interest:
of the flange width determined by Eq. (4.141) or (4.142) as
applicable: 𝐼 = 𝐼g (𝑀d ∕𝑀) ≤ 𝐼g (4.143)
BENDING STRENGTH AND DEFLECTION 171
SOLUTION
From Example 4.4, the allowable moment for the given
I-section is 178.0 in.-kips. The estimated compressive stress
in the top fiber under the allowable moment is
Mycg 178.0(4.063)
𝑓= = = 28.31 ksi
𝐼𝑥 25.382
The same stress of 𝑓 = 28.31 ksi will be assumed in the
calculation of the effective design width for deflection calcu-
lation.
By using Eqs. (3.32)–(3.34), and (3.36) and the same
procedure employed in Example 4.4, the effective width bd
of the unstiffened flange can be computed as follows:
𝑤 = 1.6775 in.
𝑤
= 12.426
Figure 4.46 Bending moment and variable moments of inertia for 𝑡
two-span continuous beam under uniform load.4.14 𝑘 = 0.43

where 𝐼g is the moment of gross cross-section; 𝑀d is the 𝑓d = 28.31 ksi


minimum of the member strengths determined according √
1.052 28.31
to Sections 4.2.2, 4.2.3.2, and 4.2.4 but with 𝑀y replaced 𝜆= √ (12.426) = 0.618 < 0.673
0.43 29,500
with M, and M is the moment at the service loads to be
considered. Detailed design provisions can also be found 𝜌 = 1.0
in Section L2 of the 2016 edition of the North American
𝑏d = 𝑤 = 1.6775 in.
Specification.
Using the full width of the compression flange and
In the design of thin-walled cold-formed steel sections, assuming the web is fully effective, the neutral axis is located
the method to be used for deflection calculation is based on at the middepth (i.e., ycg = 4.0 in.). Prior to computing the
the accuracy desired in the analysis. If a more exact deflec- moment of inertia, check the web for effectiveness as
tion is required, a computer program or a numerical method follows:
may be used in which the beam should be divided into a ( )
4 − 0.3225
relatively large number of elements according to variable 𝑓1 = 28.31 = 26.03 ksi (compression)
4
moments of inertia. The deflection calculation for such a ( )
4 − 0.3225
beam is too complicated for hand calculation. On the other 𝑓2 = 28.31 = 26.03 ksi (tension)
hand, if an approximate analysis is used, the deflection of 4
|𝑓 |
a simply supported beam may be computed on the basis of 𝜓 = || 2 || = 1.0
a constant moment of inertia determined for the maximum | 𝑓1 |
bending moment. The error so introduced is usually small 𝑘 = 4 + 2(1 − 𝜓)3 + 2(1 − 𝜓) = 24.0
and on the conservative side.3.13 For continuous spans, the
deflection of the beam may be computed either by a rational As in Example 4.4, ℎ0 ∕b0 = 4. Use Eq. (3.55a),
analysis4.13 or by a method using a conventional formula 𝑏e
in which the average value of the positive and negative 𝑏1 =
3+𝜓
moments of inertia 𝐼1 and 𝐼2 will be used as the moment
where 𝑏e is the effective width of the web determined
of inertia I.4.14 This simplified method and other approaches
in accordance with Eqs. (3.32) through (3.36) with 𝑓1
have been used in Refs. 4.6 and 4.7 for a nonlinear analysis
substitued for f and 𝑘 = 24.0 as follows:
of continuous beams.

= 54.48
Example 4.15 Determine the moment of inertia of the 𝑡

I-Section (Fig. 4.12) to be used for deflection calculation 1.052 26.03
when the I-section is loaded to the allowable moment as 𝜆 = √ (54.48) = 0.348 < 0.673
24 29,500
determined in Example 4.4 for the ASD method. The Effec-
tive Width Method is used. 𝜌 = 1.0
172 4 FLEXURAL MEMBERS

𝑏e = ℎ = 7.355 in. The same stress of 𝑓 = 24.37 ksi will be assumed in the
𝑏e 7.355 calculation of the effective design width for deflection deter-
𝑏1 = = = 1.839 in. mination.
3+𝜓 3+1
Using Eqs. (3.32)–(3.36) and the same procedure
Since 𝜓 > 0.236, employed in Example 4.6, the effective width 𝑏d of the
1 stiffened compression flange is computed as follows:
𝑏2 = 𝑏e = 3.6775 in.
2
𝑤 = 14.415 in.
𝑏1 + 𝑏2 = 1.839 + 3.6775 = 5.5165 in. 𝑤
= 137.29
Since 𝑏1 + 𝑏2 is greater than the compression portion of 𝑡
the web of 3.6775 in., the web is fully effective as assumed. 𝑘 = 4.0
Because both the compression flange and the web are fully
𝑓d = 24.37 ksi
effective, the moment of inertia 𝐼𝑥 of the full section can be √
computed as follows: 1.052 24.37
𝜆 = √ (137.29)
4 29,500
Distance = 2.076 > 0.673
from 1 − 0.22∕2.076
Area Middepth 𝜌= = 0.431
2.076
Element A (in.2 ) y (in.) Ay2 (in.4 )
𝑏d = 𝜌w = 0.431(14.415) = 6.213 in.
Flanges 4(1.6775)(0.135) = 0.9059 3.9325 14.0093 By using the effective width of the compression flange and
Corners 4(0.05407) = 0.2163 3.8436 3.1955 assuming the web is fully effective, the moment of inertia
Webs 2(7.355)(0.135) = 1.9859 0 0 can be computed from the line elements shown in Fig. 4.20
Total 3.1081 17.2048 as follows:
1
2𝐼web = 2 × 12 (0.135)
× (7.355)3 = 8.9522
Distance
𝐼𝑥 = 26.1570 in.4 from
Effective Top
Length Fiber
𝑀𝑥 𝑦cg 178.0(4.0)
𝑓= = = 27.22 ksi Element L (in.) y (in.) Ly (in.2 ) Ly2 (in.3 )
𝐼𝑥 26.1570
1 2(1.0475) = 2.0950 9.9476 20.8400 207.3059
In view of the fact that the computed stress of 27.22 ksi
is less than the assumed value of 28.31 ksi, the moment of 2 2(0.3768) = 0.7536 9.8604 7.4308 73.2707
inertia 𝐼𝑥 computed on the basis of the full section can be 3 2(9.415) = 18.8300 5.0000 94.1500 470.7500
used for deflection calculation without additional iteration. 4 2(0.3768) = 0.7536 0.1396 0.1052 0.0147
5 6.2130 0.0525 0.3262 0.0171
Total 28.6452 122.8522 751.3584
Example 4.16 Compute the moment of inertia of the hat 122.8522
section (Fig. 4.19) to be used for deflection calculation when 𝑦cg = = 4.289 in.
28.6452
the hat section is loaded to the allowable moment as deter-
mined in Example 4.6 for the ASD method. The Effective
Width Method is used. The total 𝐼𝑥 is determined as follows:
( )
1
2𝐼3′ = 2 (9.415)3 = −139.0944
SOLUTION 12

(Ly2 ) = −751.3584 − 890.4528
1. First Approximation. From Example 4.6, the (∑ )
allowable moment is 193.59 in.-kips. The estimated − 𝐿 (𝑦2cg ) = −28.6452(4.289)2
compressive stress in the top flange under the allowable
moment is = −526.9434

𝑀𝑥 𝑦cg 𝐼𝑥′ = −363.5094 in.3


193.59(4.487)
𝑓= = = 24.37 ksi
𝐼𝑥 35.646 𝐼𝑥 = 𝐼𝑥′ 𝑡 = 363.5094(0.105) = 0 − 38.168 in.4
BENDING STRENGTH AND DEFLECTION 173
The compressive stress in the top fiber is roofs or panels provide more or less lateral support and
𝑀𝑥 𝑦cg rotation restraints to the connected beam and to the flange,
193.59(4.289)
𝑓= = and increase the beam bending capacity. Two approaches are
𝐼𝑥 38.168
used to account such restraints in determining the member
= 21.75 ksi < the assumed value (no good) strength:

2. Second Approximation. Assuming 𝑓d = 21.00 ksi and 1. Use the empirical equations developed based on exper-
using the same values of w/t and k, imental study. This approach simplifies the design but
√ is limited to certain beam and panel sizes, and beam
1.052 21.00
𝜆 = √ (137.29) = 1.927 > 0.673 span lengths that are tested, and
4 29,500 2. Use the Direct Strength Method in which the buckling
𝜌 = 0.460 moments are determined considering the effects of roof
or panel covering and span continuity. This method is
𝑏d = 𝜌w = 6.631 in. essentially applicable to any cross-section types and
span lengths. The complexity of this approach is how
Distance to numerically model the system so that the roof or
from panel restraints and the structural system connectivity
Effective Top are realistically modeled.
Length Fiber
Element L (in.) y (in.) Ly (in.2 ) Ly2 (in.3 ) Both approaches are discussed in the subsections.

1 to 4 22.4322 122.5260 751.3413 4.2.8.1 Strengths for Members with General Cross-
5 6.6310 0.0525 0.3481 0.0183 Sections and Connectivity The Direct Strength Method
Total 29.0632 122.8741 751.3596 provides a means to directly determine the strength of the
ycg = 4.289 in. members that are connected to sheathings through-fastened
or with standing seam clips. The method requires that the
member buckling moment considers the restraints from
The total 𝐼𝑥 is the connected sheathing or panels. Research work4.297–4.303
has been conducted to model these restraints with springs
2𝐼3′ = −139.0944
∑ which posessess translational and rotational stiffnesses.
(Ly2 ) = −751.3596 − 890.4540 Methods on how to determine the stiffness based on the
(∑ ) deck or sheathing configuration and connectivities are
− 𝐿 (𝑦2cg ) = −29.0632(4.228)2 provided in the study. 4.305 A finite element method such
as MASTAN24.312 was then used to determine the member
= −519.5333
(global) buckling moment. Using this numerical analysis
𝐼𝑥′ = −370.9207 in.3 in metal building wall and roof systems were documented
for bare deck through-fastened to members,4.398,4.303 and
𝐼𝑥 = 𝐼𝑥′ 𝑡 = −38.947 in.4
for through-fastened and standing seam insulated metal
𝑀𝑥 𝑦cg 193.59(4.228) panels.4.301
𝑓= = = 21.01 ksi
𝐼𝑥 38.947 In 2016, the following design provisions were introduced
into the North American Specification:
Since the computed value of f is close to the assumed value
of 21.00 ksi, the moment of inertia for deflection calculation
I6.1.2 Flexural Member Design
under the allowable moment is 38.947 in.4 It is of interest to
The nominal flexural strength [resistance], Mn , shall be the
note that the difference between the I values computed from minimum of 𝑀ne , 𝑀𝑛𝓁 , and 𝑀nd as given in Specifica-
the first and second approximations is only about 2%. tion Sections I6.1.2.1 to I6.1.2.3. For members meeting the
geometric and material limits of Specification Section B4, the
4.2.8 Beams in Metal Roof and Wall Systems safety and resistance factors shall be as follows:

Beams in metal roof and wall systems usually have one of Ωb = 1.67 (ASD)
the flanges attached to metal roofs or wall panels through 𝜙b = 0.90 (LRFD)
through-fastened connections or sliding clips. These metal = 0.85 (LSD)
174 4 FLEXURAL MEMBERS

For all other members, the safety and resistance factors in In the past, the bending capacity of flexural members
Specification Section A1.2(c) shall apply. The available strength having the tension flange through-fastened to deck or
[factored resistance] shall be determined in accordance with the sheathing has been studied by a large number of investiga-
applicable method in Specification Section B3.2.1, B3.2.2, or tors in various countries.4.30–4.40 Based on the results of these
B3.2.3. studies, reduction factors for the effective yield moment have
been developed for simple- and continuous-span conditions.
I6.1.2.1 Lateral–Torsional Buckling
These factors are given in Section I6.2.1 of the 2016 edition
The nominal flexural strength [resistance], Mne , for
lateral–torsional buckling shall be calculated in accordance of the North American Specification. For the convenience of
with Specification Section F2, except 𝐹cre or 𝑀cre shall be readers, the following excerpts are adapted from the North
determined including lateral, rotational, and composite stiffness American Specification:
provided by the deck or sheathing, bridging and bracing, and
span continuity.
I6.2.1 Flexural Members Having One Flange
I6.1.2.2 Local Buckling
Through-Fastened to Deck or Sheathing
This section shall not apply to a continuous beam for the
The nominal flexural strength [resistance], 𝑀𝑛𝓁 , for local buck-
region between inflection points adjacent to a support or to a
ling shall be calculated in accordance with Specification Section
cantilever beam.
F3, except 𝐹cr𝓁 or 𝑀cr𝓁 shall be determined including lateral,
The nominal flexural strength [resistance], 𝑀n , of a C- or
rotational, and composite stiffness provided by the deck or
Z-section loaded in a plane parallel to the web, with the tension
sheathing.
flange attached to deck or sheathing and with the compres-
sion flange laterally unbraced, shall be calculated in accordance
I6.1.2.3 Distortional Buckling
with Eq. 4.143. Consideration of distortional buckling in accor-
The nominal flexural strength [resistance], 𝑀nd , for distortional
dance with Specification Section F4 shall be excluded. The
buckling of girts and purlins shall be calculated in accordance
safety factor and resistance factors given in this section shall be
with Specification Section F4, except 𝑀crd shall be determined
used to determine the allowable flexural strength or design flex-
including lateral, rotational, and composite stiffness provided by
ural strength [factored resistance] in accordance with the appli-
the deck or sheathing.
cable design method in Specification Section B3.2.1, B3.2.2, or
B3.2.3.
I6.1.3 Member Design for Combined Flexure and Torsion
The nominal flexural strength [resistance], 𝑀n , for members in 𝑀n = R M𝑛𝓁𝑜 (4.143)
combined flexure and torsion shall be reduced by applying the
Ωb = 1.67 (ASD)
reduction factor, R, determined in accordance with Specification
Eq. H4-1. 𝜙b = 0.90 (LRFD)
= 0.90 (LSD)
A design example that illustrates Specification Section I6.1
is provided in the 2017 edition of the AISI Cold-Formed Steel where R = A value obtained from Specification Table I6.2.1-1
Design Manual.1.428 for C- or Z-sections
𝑀𝑛𝓁𝑜 = Nominal flexural strength with consideration of
local buckling only, as determined from
4.2.8.2 Strengths for Members with Specific Cross-
Specification Section F3 with 𝐹n = 𝐹y or
Sections and Connectivity The design methods provided 𝑀ne = 𝑀y
in the following subsections are applicable within the given
limitations. Specification Table I6.2.1-1C- or Z-Section R Values

4.2.8.2.1 Beams Having One Flange Through Fastened Simple Span


to Deck or Sheathing When roof purlins or wall girts are Member Depth Range, in. (mm) Profile R
subject to the suction force due to wind load, the compression d ≤ 6.5 (165) C or Z 0.70
flange of the member is laterally unbraced, but the tension 6.5 (165) < d ≤ 8.5 (216) C or Z 0.65
flange is supported by the deck or sheathing. The bending 8.5 (216) < d ≤ 12 (305) Z 0.50
8.5 (216) < d ≤ 12 (305) C 0.40
capacity of this type of flexural member is less than the fully
Continuous Span
braced member but is greater than the laterally unbraced
condition because of the rotational restraint provided by the Profile R
panel-to-purlin (or girt) connection. The rotational stiffness C 0.60
has been found to be a function of the member thickness, Z 0.70
sheet thickness, fastener type, and fastener location.
BENDING STRENGTH AND DEFLECTION 175
The reduction factor, R, shall be limited to roof and wall 4.2.8.2.2 Flexural Members Having One Flange
systems meeting the following conditions: Fastened to a Standing Seam Roof System Standing
seam roofs were first introduced in the 1930s.4.171 Because
(a) Member depth ≤ 12 in. (305 mm).
standing seam roof panels are attached to supporting purlins
(b) Member flanges with edge stiffeners. with a clip that is concealed in the seam, this type of roof
(c) 60 ≤ depth/thickness ≤ 170. system has proved to be a cost-effective roof membrane
(d) 2.8 ≤ depth/flange width ≤ 5.5. due to its superior weather tightness, its ability to provide
(e) Flange width ≥ 2.125 in. (54.0 mm). consistent thermal performance, its low maintenance
(f) 16 ≤ flat width/thickness of flange ≤ 43. requirements, as well as its ability to adjust to thermal
(g) For continuous span systems, the lap length at each interior expansion and contraction.4.172
support in each direction (distance from center of support to For C- or Z-purlins supporting a standing seam roof
end of lap) is not less than 1.5d. system, the bending capacity is greater than the bending
(h) Member span length is not greater than 33 feet (10 m). strength of an unbraced member and may be equal to the
(i) Both flanges are prevented from moving laterally at the bending strength of a fully braced member. The bending
supports.
capacity is governed by the nature of the loading, gravity or
(j) Roof or wall panels are steel sheets with 50 ksi (340 MPa uplift, and the nature of the particular standing seam roof
or 3520 kg/cm2 ) minimum yield stress, and a minimum
system. Due to the availability of numerous types of standing
of 0.018 in. (0.46 mm) base metal thickness, having a
minimum rib depth of 1-1/8 in. (29 mm), spaced at seam roof systems, the method to determine the member
a maximum of 12 in. (305 mm) on centers and attached in strength through experiments was developed, and was first
a manner to effectively inhibit relative movement between added in the 1996 edition of the AISI Specification and is
the panel and member flange. retained in Appendix A of the 2016 edition of the North
(k) Insulation is glass fiber blanket 0 to 6 in. (152 mm) thick, American Specification as Section I6.2.2 for beams having
compressed between the member and panel in a manner one flange fastened to a standing seam roof system. In this
consistent with the fastener being used. section, it is specified that the available flexural strength of a
(l) Fastener type is, at minimum, No. 12 self-drilling or C- or Z-section loaded in a plane parallel to the web with the
self-tapping sheet metal screws or 3/16 in. (4.76 mm) rivets, top flange supporting a standing seam roof system shall be
having washers with 1/2 in. (12.7 mm) diameter.
determined using discrete point bracing and the provisions
(m) Fasteners are not standoff type screws. of Chapter F of the Specification or shall be calculated as
(n) Fasteners are spaced not greater than 12 in. (305 mm) on follows:
centers and placed near the center of the member flange, and
adjacent to the panel high rib.
𝑀n = 𝑅𝑀𝑛𝓁𝑜 (4.145)
(o) The ratio of tensile strength to design yield stress shall not
be less than 1.08. Ωb = 1.67 (ASD)

If variables fall outside any of the above-stated limits, the 𝜙b = 0.90 (LRFD)
user shall perform full-scale tests in accordance with Section
K2.1 of the Specification or apply a rational engineering anal- where R is the reduction factor determined in accordance
ysis procedure. For continuous purlin and girt systems in which with AISI S908 and Mn𝓁o is the nominal flexural strength
adjacent bay span lengths vary by more than 20%, the R values with consideration of local buckling only, as determined
for the adjacent bays shall be taken from the simple-span values from Section 4.2.3 with 𝐹n = 𝐹y or 𝑀ne = 𝑀y . For addi-
in Specification Table I6.2.1-1. The user is permitted to perform
tional design information, see Ref. 4.172, which includes
tests in accordance with Specification Section K2.1 as an alter-
native to the procedure described in this section.
detailed discussion and design examples using standing seam
For simple-span members, R shall be reduced for the effects roof systems. The major advantage of the base test is that
of compressed insulation between the sheeting and the member. a simple span test may be used to predict the performance
The reduction shall be calculated by multiplying R from Speci- of continuous-span systems for reducing experimental costs.
fication Table I6.2.1-1 by the following correction factor, r: The concepts for the base test was developed by T. M. Murray
𝑟 = 1.00 − 0.01𝑡i when 𝑡i is in inches (4.144) and his associates at Virginia Polytechnic Institute & State
University.
𝑟 = 1.00 − 0.0004𝑡i when ti is in millimeters (4.145)
In Canada, this type of member is designed in accor-
where 𝑡i = Thickness of uncompressed glass fiber blanket dance with Specification Chapter F based on discrete bracing
insulation. provided.
176 4 FLEXURAL MEMBERS

4.2.9 Strength of Standing Seam Roof Panel System distance from the web, as shown in Fig. 4.47a for a box-type
beam and an I-section. This phenomenon is known as
Under gravity loading, the nominal strength of standing seam
shear lag.
roof panel systems can be determined according to Chapter
Analytical and experimental investigations of the problem
F of the Specification because the load-carrying capacity of
on shear lag have previously been conducted by Hildebrand
usual panels can be calculated accurately. The strength of this
and Reissner,4.41 Winter,4.42 Miller,4.43 and Tate.4.44,4.45 This
type of panel system can also be determined by the AISI
subject has been investigated by Malcolm and Redwood,4.46
S906, Test Standard for Determining the Load-Carrying
Parr and Maggard,4.47 Van Dalen and Narasimham,4.48 and
Strength of Panels and Anchor-to-Panel Attachments for
Lamas and Dowling4.49 and in Refs. 4.50–4.55.
Roof or Siding Systems Tested in Accordance with ASTM
In their paper, Hildebrand and Reissner concluded that
E1592, in accordance with Section I6.3.1 of the North Amer-
the amount of shear lag depends not only on the method of
ican Specification.1.345
loading and support and the ratio of span to flange width
For uplift loading, the nominal strength of standing seam
but also on the ratio of G/E and the ratio 𝑚 = (3𝐼w + 𝐼s )∕
roof panels and their attachments or anchors cannot be calcu-
(𝐼w + 𝐼s ), where 𝐼w and 𝐼s are the moments of inertia of webs
lated with accuracy; therefore, it can only be determined by
and of cover plates, respectively, about the neutral axis of the
tests using the AISI S906 with the requirements and excep-
beam.
tions prescribed in Section I6.3.1 of the Specification on the
Based on the theory of plane stress, Winter analyzed the
use of Factory Mutual FM4471, Corps of Engineers CEGS
shear lag problem and developed tabular and graphic data,4.42
07416, and ASTM E1592. The load combinations including
from which the effective width of any given beam section
wind uplift are provided in Section I6.3.1a of Appendix A of
can be obtained directly for use in design. The ratios of the
the Specification.
maximum and minimum bending stresses in beam flanges
The evaluation of test results should follow the AISI S906.
were computed and verified by the results of 11 I-beam
When three or more assemblies are tested, safety factors
tests. It was indicated that shear lag is important for beams
(not less than 1.67) and resistance factors (not greater than
0.9) shall be determined in accordance with the procedure
of Specification Section K2.1.1 (c) with the target relia-
bility index and statistical data provided in Section I6.3.1 of
the Specification. The justifications for these variables are
discussed in the AISI Commentary.1.431 When the number
of physical test assemblies is less than 3, a safety factor of
2.0 and resistance factors of 0.8 (LRFD) and 0.7 (LSD) shall
be used.

4.2.10 Unusually Wide Beam Flanges and Unusually


Short Span Beams Figure 4.47 (a) Stress distribution in both compression and
tension flanges of beams due to shear lag.
When beam flanges are unusually wide, special consider-
ation should be given to the possible effects of shear lag
and flange curling, even if the beam flanges, such as tension
flanges, do not buckle. Shear lag depends on the type of
loading and the span-to-width ratio and is independent of the
thickness. Flange curling is independent of span length but
depends on the thickness and width of the flange, the depth
of the section, and the bending stresses in both flanges.

4.2.10.1 Shear Lag For conventional structural members


with ordinary dimensions, the effect of shear deformation on
flange stress distribution is negligible. However, if the flange
of a beam is unusually wide relative to its span length, the
effect of shear deformation on bending stress is pronounced.
As a result, the bending stresses in both compression and Figure 4.47 (b) Analytical curve for determining effective width
tension flanges are nonuniform and decrease with increasing of flange of short-span beams.4.42
BENDING STRENGTH AND DEFLECTION 177
Table 4.5 Ratio of Effective Design Width to Actual Width for Wide Flanges
L/wf
Loading Condition Investigator 6 8 10 12 16 20 30
Hildebrand and Reissner 0.830 0.870 0.895 0.913 0.934 0.946

Hildebrand and Reissner 0.724 0.780 0.815 0.842 0.876 0.899

Hildebrand and Reissner 0.650 0.710 0.751 0.784 0.826 0.858

Hildebrand and Reissner 0.686 0.757 0.801 0.830 0.870 0.895 0.936
Winter 0.550 0.670 0.732 0.779 0.850 0.894 0.945
Miller — — 0.750
Hildebrand and Reisser 0.610 0.686 0.740 0.778 0.826 0.855 0.910

Hiildebrand and Reisser 0.830 0.897 0.936 0.957 0.977 0.985 0.991
Winter 0.850 0.896 0.928 0.950 0.974 0.984 0.995
Miller — — 0.875

with wide flanges subjected to concentrated loads on fairly compression flanges shall be limited to the value given in
short spans; the smaller the span-to-width ratio, the larger Table 4.6 in accordance with the 𝐿∕𝑤f ratio.
the effect. For beams supporting uniform loads, shear lag is In the application of Table 4.6 the effective span length
usually negligible except that the L/wf ratio is less than about of the beam is the full span for simple-span beams, the
10 as shown in Fig. 4.47b. Winter also concluded that for a distance between inflection points for continuous beams, or
given span-to-width ratio the effect of shear lag is practically twice the length of cantilever beams. The symbol 𝑤f indi-
the same for box beams, I-beams, T-beams, and U-shaped cates the width of the flange projection beyond the web for
beams. I-beams and similar sections or half the distance between
Table 4.5 is a summary of the ratios of effective design webs for multiple-web sections, including box or U-type
width to actual width based on the results obtained by several sections (Fig. 4.47a). When I-beams and similar sections are
investigators.4.45
In Table 4.5, 𝑤f is the width of the flange projection Table 4.6 Maximum Allowable Ratio of Effective
beyond the web for I-beams and half the distance between Design Width to Actual Width for Short-Span, Wide
webs for multiple-web sections (Fig. 4.47a); L is the span Flanges
length. It should be noted that the values obtained by Hilde-
brand and Reissner were for 𝐺∕𝐸 = 0.375 and 𝑚 = 2. L/wf Effective Design Width (b) /Actual Width (w)
As far as the design criteria are concerned, the “effective 30 1.00
width” concept used in the design of compression elements 25 0.96
(Section 3.3) can also be applied to the design of beams 20 0.91
whenever the shear lag problem is critical. 18 0.89
Based on the results of Winter’s investigation,4.42 design 16 0.86
provisions for shear lag have been developed as included 14 0.82
in Section B4.3 of the North American Specification.1.417 It 12 0.78
is specified that when the effective span L of the beam is 10 0.73
less than 30𝑤f and when it carries one concentrated load 8 0.67
or several loads spaced farther apart than 2𝑤f , the ratio of 6 0.55
effective design width to actual width of the tension and
178 4 FLEXURAL MEMBERS

stiffened by lips at outer edges, wf shall be taken as the sum of Example 4.17 Compute the nominal moment for the beam
the flange projection beyond the web plus the depth of the lip. section shown in Fig. 4.48a if it is used to support a concen-
The tabulated ratios in Table 4.6 are also plotted in trated load on a simple span of 2 ft. Assume that the minimum
Fig. 4.47b for comparison with the analytical values. The yield stress of steel is 40 ksi.
AISI design values are slightly larger than the analytical
results when 𝐿∕𝑤f ratios exceed about 16. SOLUTION
Although the above-discussed provision relative to shear From Fig. 4.48a,
lag is applicable to tension and compression flanges, local
buckling in compression as discussed in Section 3.3 may 1
𝑤f = (3.25) − 0.135 = 1.490 in.
be a critical factor and should also be investigated sepa- 2
rately. The shear lag problem is of particular importance in 30𝑤f = 44.70 in.
the analysis and design of aircraft and naval structures. In
𝐿 = 2 ft = 24 in.
cold-formed steel building construction, however, it is infre-
quent that beams are so wide that they would require consid- Since 𝐿 < 30𝑤f and the beam is subject to a concentrated
erable reduction of flange widths. load, shear lag is an important factor.
For members designed by the Direct Strength Method, Using Table 4.6 for 𝐿∕𝑤f = 16.1, the ratio of effective
Commentary on the Specification1.431 recommended that the design width to actual width is 0.86. The effective design
ratio of effective width (b) to the actual width (w) be replaced widths for both compression and tension flanges are
by the corresponding ratio of 𝑀n ∕𝑀y . In building construc-
tion, in the cases of short spans under concentrated loads, 𝑏′ = 0.86 × 1.49 = 1.28 in.
web crippling is typically controlling the limit state. There-
fore, web crippling (Section 4.3.6) must be checked. See Fig. 4.48b.

(a)

(b)

Figure 4.48 Example 4.17.


BENDING STRENGTH AND DEFLECTION 179
To check if the web is fully effective according to 𝑤 = 12 (3.25) − (0.1875 + 0.135) = 1.3025 in.,
Section 3.3.1.2,
( ) 𝑤 1.3025
1 = = 9.62
𝑓1 = 40 × 2.6775 = 35.7 ksi (compression) 𝑡 0.135
3 ( ) √
𝑓2 = 35.7 ksi (tension) 1.052 40
λ= √ (9.62) = 0.568 < 0.673
|𝑓 | 0.43 29,500
Ψ = || 2 || = 1.0
| 𝑓1 | 𝑏 = 𝑤 = 1.3025 in.
𝑘 = 4 + 2[1 + 1]3 + 2[1 + 1] = 24 The nominal moment is
From Fig 4.48a, 𝑀n = 𝑆𝑥 (full section)𝐹𝑦
ℎ0 = out-to-out depth of web In view of the fact that the nominal moment determined
= 6.00 in. above for local buckling consideration is larger than that
determined for shear lag, the nominal moment of 136 in.-kips
𝑏0 = out-to-out width of the compression flange of will govern the design.
each channel
Example 4.18 For the tubular section shown in Fig. 4.49,
= 3.25∕2 determine the nominal moment if the member is to be used
= 1.625 in. as a simply supported beam to carry a concentrated load at
midspan. Assume that the span length is 5 ft and Fy = 50 ksi.
Since ℎ0 ∕𝑏0 = 3.69 < 4, then use Eq. (3.47a),
𝑏e SOLUTION
𝑏1 =
3+𝜓
1. Nominal Moment Based on Effective Width of
where 𝑏e is the effective width of the web determined in Compression Flange. For the compression flange,
accordance with Eqs. (3.32–3.35) with 𝑓1 substistuted for f
and 𝑘 = 24 as follows: 𝑤 8 − 2(3∕32 + 0.06) 7.693
√ = =
1.052(5.355∕0.135) 35.7∕29,500 𝑡 0.06 0.06
𝜆= √ = 0.296
24 = 128.2
Since λ < 0.673, 𝜌 = 1.0, Based on Eqs. (3.32)–(3.35),
( )
𝑏e = 5.355 in. ( )√𝑓
1.052 𝑤
𝜆= √
𝑏e 5.355 𝑘 𝑡 𝐸
𝑏1 = = = 1.339 in. ( )
3+𝜓 3+1 √
1.052 50
For 𝜓 = 1, which is larger than 0.236, = √ (128.2) = 2.776 > 0.673
4 29,500
1 1
𝑏2 = 𝑏e = × 5.355 = 2.678 in.
2 2
𝑏1 + 𝑏2 = 1.339 + 2.678 = 4.017 in.
Since 𝑏1 + 𝑏2 is larger than the compression portion of the
web of 2.6775 in., the web is fully effective.
Based on the method discussed previously, the effective
section modulus is
𝑆e = 3.4 in.3
The nominal moment is
𝑀n = 𝑆e 𝐹y = 3.4 × 40 = 136 in. − kips
The nominal moment determined above for shear
lag should be checked for local buckling. Since Figure 4.49 Example 4.18.
180 4 FLEXURAL MEMBERS

1 − 0.22∕𝜆 1 − 0.22∕2.776 𝑏e = 0.871(4.692) = 4.087 in.


𝜌= = = 0.332
𝜆 2.776 𝑏e 4.087
𝑏 = 𝜌w = 0.332(7.693) = 2.554 in. 𝑏1 = = = 1.141 in.
3+𝜓 3 + 0.581
See Fig. 4.50a. Since 𝜓 > 0.236, b2 = be ∕2 = 4.087∕2 = 2.0435 in., and
Assume that the web is fully effective. The distance ycg b1 + b2 = 1.141 + 2.0435 = 3.1845 > 2.968 in. (compres-
can be determined as follows: sion portion of the web). The web is fully effective. The total
𝐼𝑥 is determined as

Distance (Ay2 ) = 15.4872
from
Top 1
Fiber
𝐼webs = 2 × (0.06)(4.6925)3 = 1.0333
2
Element Area 𝐴 (in.2 ) y (in.) Ay (in.3 ) Ay2 (in.4 ) (∑ )
− 𝐴 (𝑦2cg ) = −1.2245(3.122)2 = −11.9351
1 2.554 × 0.06 = 0.1532 0.030 0.0046 0.00014
2 2 × 0.01166 = 0.0233 0.073 0.0017 0.0013
𝐼𝑥 = 4.5850 in.4
3 2 × 4.6925 × 0.06 = 0.5631 2.500 1.4078 3.5194
4 2 × 0.01166 = 0.0233 4.927 0.1148 0.5656 The section modulus is
5 7.6925 × 0.06 = 0.4616 4.970 2.2942 11.4019 𝐼 4.5850
Total 1.2245 3.8231 15.48717 𝑆𝑥 = 𝑥 = = 1.469 in.3
∑ 𝑦cg 3.122
(Ay)
𝑦cg = ∑ = 3.122 in. The nominal moment about the x axis is
𝐴
𝑀n = 1.469(50) = 73.45 in.-kips
1. To check if the web is fully effective (see Fig. 4.50b),
2. Nominal Moment Based on Shear Lag Considera-
( ) tion. According to Figs. 4.47 and 4.49,
2.968
𝑓1 = 50 = 47.53 ksi (compression) 8 − 2(0.06)
3.122 𝑤f = = 3.94 in.
( ) 2
1.724
𝑓2 = 50 = 27.61 ksi (tension) 𝐿 5 × 12
3.122 = = 15.23 < 30
| 𝑓 | 27.61 𝑤f 3.94
𝜓 = || 2 || = = 0.581 Because the 𝐿∕wf ratio is less than 30 and the member
| 𝑓1 | 47.53
carries a concentrated load, additional consideration for
𝑘 = 4 + 2[1 + 0.581]3 + 2[1 + 0.581] = 15.066 shear lag is needed. Using Table 4.6,
From Fig. 4.49, Effective design width
= 0.845
Actual width
ℎ0 = out-to-out depth of web = 5.00 in.
Therefore the effective design widths of compression and
𝑏0 = out-to-out width of the compression flange tension flanges between webs are (Fig. 4.51)
= 8.00 in. 0.845[8 − 2(0.06)] = 6.6586 in.
Since ℎ0 ∕b0 = 0.625 < 4, use Eq. (3.47a), Using the full areas of webs, the moment of inertia about
the x axis is
𝑏e
𝑏1 = 𝐼𝑥 = 4[3.22356(0.06)(2.5 − 0.03)2
3+𝜓
where be is the effective width of the web determined in 1
+ 0.01166(2.5 − 0.073)2 ] + 2 (0.06)(4.6925)3
accordance with Eqs. (3.32)–(3.35) with 𝑓1 substituted for f 12
and 𝑘 = 15.066 as follows: = 6.046 in.4
( )
( )√ and the section modulus is
1.052 4.692 47.53
𝜆= √
0.06 29,500 𝐼
15.066 𝑆𝑥 = 𝑥 = 2.418 in.3
2.5
= 0.851 > 0.673
The nominal moment is
1 − 0.22∕𝜆 1 − 0.22∕0.851
𝜌= = = 0.871 𝑀n = 2.418(50) = 120.9 in.-kips
𝜆 0.851
BENDING STRENGTH AND DEFLECTION 181

(a)

(b)

Figure 4.50 (a) Effective width of compression flange for postbuckling strength. (b) Webs order
stress gradient.

4.2.10.2 Flange Curling When a beam with unusually


wide and thin flanges is subject to bending, the portion of the
flange most remote from the web tends to deflect toward the
neutral axis. This is due to the effect of longitudinal curvature
of the beam and bending stresses in both flanges. This subject
was studied by Winter in 1940.4.42
Let us consider an I-beam which is subject to pure bending
as shown in Fig. 4.52. The transverse component q of the
flange force 𝑓av 𝑡 per unit width can be determined by
𝑓av 𝑡 dϕ 𝑓av 𝑡 𝑓 𝑡 2𝑓 2 𝑡
𝑞= = = av = av (4.146)
Figure 4.51 Effective design widths of compression and tension dl 𝑟b EI∕𝑀 Ed
flanges for shear lag. where 𝑓av = average bending stress in flanges
𝑡 = flange thickness
d𝜙, dl, 𝑟b = as shown in Fig.4.52
3. Nominal Moment for Design. From the above calcu- 𝐸 = modulus of elasticity
lation, the nominal moment for design is 73.45 in.-kips. 𝐼 = moment of inertia of beam
Shear lag does not govern the design. 𝑑 = depth of beam
182 4 FLEXURAL MEMBERS

The above approximate treatment for I-beams can also be


applied to the design of box and U-type beams, except that
for the latter the flanges of the closed box beams may be
regarded as simple plates freely supported at webs and that wf
is to be measured as half of the distance between webs. Using
the same analogy, one can determine the flange curling cf for
closed box sections as follows:
( ) ( )
5 𝑞(2𝑤f )4 𝑓av 2 𝑤4f
𝑐f = × =5 (1 − 𝜇2 )
384 𝐷 𝐸 𝑡2 𝑑
Figure 4.52 Flange curling of I-beam subject to bending.4.42 (4.149)
A comparison of Eqs. (4.147) and (4.149) indicates that
the use of Eq. (4.148), which is derived on the basis of
If the value of q is considered to be a uniformly distributed I-beams, to determine wf for box beams, may √ result in
load applied on the flange, the deflection or curling at the a possible error of 13%. This is because 4 5∕3 = 1.13.
outer edge of the flange can be computed by the conventional However, this discrepancy can be reduced if the restraint of
method for a cantilever plate, namely, webs and the variable values of the transverse component q
( ) ( ) are taken into consideration.
qw4f 𝑓av 2 𝑤4f No specific values are given by the North American Spec-
𝑐f = =3 (1 − 𝜇2 ) (4.147)
8𝐷 𝐸 𝑡2 𝑑 ification for the maximum permissible amount of curling.
where 𝑐f = deflection at outer edge However, it is stated in the AISI Commentary that the amount
𝑤f = projection of flange beyond web of curling that can be tolerated will vary with different kinds
of sections and must be established by the designer. An
𝐷 = flexural rigidity of plate, = Et3 ∕12(1 − μ2 )
amount of curling on the order of 5% of the depth of the
By substituting 𝐸 = 29.5 × 103 ksi and 𝜇 = 0.3 in section is usually not considered excessive.
Eq. (4.147), the following formula for the maximum width Assuming 𝑐f ∕𝑑 = 0.05, Eq. (4.148) can be simplified as
of an unusually wide stiffened or unstiffened flange in √
tdE
tension and compression can be obtained: 𝑤f = 0.37
√ √ 𝑓av
1800td 4 100𝑐f In general, the problem of flange curling is not a critical
𝑤f = ×
𝑓av 𝑑 factor to limit the flange width. However, when the appear-
√ √ ance of the section is important, the out-of-plane distortion
= 0.061𝑡 dE∕𝑓av 4 100𝑐f ∕𝑑 (4.148) should be closely controlled in practice.
In general, the problem of flange curling is not a critical
where 𝑐f = permissible amount of curling, in. factor to limit the flange width. However, when the appear-
𝑓av = average stress in full unreduced flange width, ance of the section is important, the out-of-plane distortion
ksi (𝑤f , t, and 𝑑 were defined previously) should be closely controlled in practice.
When members are designed by the effective design width Example 4.19 Determine the amount of curling for the
procedure, the average stress equals the maximum stress compression flange of the hat section used in Example 4.6
times the ratio of the effective design width to the actual when the section is subjected to the allowable moment. Use
width. Equation (4.148) is included in Section L3 of the the ASD method.
North American Specification to limit the width of unusually
wide flanges. SOLUTION
The above formula for determining wf is derived on the The curling of the compression flange of the hat section can
basis of a constant transverse component q. As soon as flange be computed by Eq. (4.148). In the calculation,
curling develops, the distance from the flange to the neutral
1
axis becomes smaller at the outer edge of the flange. This 𝑤f = (15.0 − 2 × 0.105) = 7.395 in.
results in the reduction of bending stresses. Therefore the 2
values of q vary along the flange as shown in Fig. 4.52. Since 𝑡 = 0.105 in.
the amount of cf is usually limited to a small percentage 𝑑 = 10.0 in.
of the depth, the error in the determination of wf by using 40.69 4.934
Eq. (4.148) is negligible and on the conservative side. 𝑓av = × = 8.34 ksi
1.67 14.415
DESIGN OF BEAM WEBS 183
Using Eq. (4.148), under concentrated loads can be designed as compression
√ √ members. The nominal strength, 𝑃n , for bearing stiffeners
0.061 × 0.105 × 10 × 29,500 4 100𝑐f
7.395 = × is the smaller of the values determined by 1 and 2 as
8.34 10 follows:

= 26.77 4 𝑐f 1. 𝑃n = 𝐹wy 𝐴c (4.150)
2. 𝑃n = Nominal axial load evaluated according to Spec-
𝑐f = 0.0058 in. ification Section E3.1 with 𝐴e replaced by 𝐴b . Speci-
fication Section E3.1 is provided in Section 5.6 of this
book.
4.3 DESIGN OF BEAM WEBS
4.3.1 Introduction Ωc = 2.00 (ASD)
{
Not only should thin-walled cold-formed steel flexural
0.85 (LRFD)
members be designed for bending strength and deflection as 𝜙c =
discussed in Section 4.2 but also the webs of beams should 0.80 (LSD)
be designed for shear, bending, combined bending and shear, where
web crippling, and combined bending and web crippling.
⎧18𝑡 + 𝐴s for bearing stiffeners at interior
2
In addition, the depth of the web should not exceed the ⎪ support or under concentrated load (4.151)
maximum value permitted by Section B4.1 of the North 𝐴c = ⎨ 2
American Specification. ⎪10𝑡 + 𝐴s for bearing stiffeners
⎩ at end support (4.152)
When the Effective Width Method is used, the maximum
allowable depth-to-thickness ratio h/t for unreinforced webs 𝐴s = cross-sectional area of bearing stiffeners
is limited to 200, in which h is the depth of the flat portion
of the web measured along the plane of the web and t 𝐹wy = lower value of 𝐹y for beam web,
is the thickness of the web. When bearing stiffeners are or 𝐹ys for stiffener section
provided only at supports and under concentrated loads,
the maximum depth-to-thickness ratio may be increased to ⎧
260. When bearing stiffeners and intermediate shear stiff- ⎪𝑏1 𝑡 + 𝐴s for bearing stiffeners at
eners are used simultaneously, the maximum h/t ratio is 300. ⎪ interior support or under

These limitations for h/t ratios are established on the basis of 𝐴𝑏 = ⎨ concentrated load (4.153)
the studies reported in Refs. 3.60 and 4.56–4.60. When the ⎪𝑏 𝑡 + 𝐴 for bearing stiffeners
Direct Strength Method is used, the h/t is limited to 300 for ⎪ 2 s
⎪ at end support (4.154)
any of the above conditions. If a web consists of two or more ⎩ [ ( ) ]
sheets, the h/t ratios of the individual sheets shall not exceed 𝐿st
𝑏1 = 25𝑡 0.0024 + 0.72 ≤ 25𝑡 (4.155)
the maximum allowable ratios mentioned above. 𝑡
The following discussions deal with the minimum require- [ ( ) ]
𝐿st
ments for bearing and shear stiffeners, the design strength 𝑏2 = 12𝑡 0.0044 + 0.83 ≤ 12𝑡 (4.156)
for shear and bending in beam webs, the load or reaction to 𝑡
prevent web crippling, and combinations of various types of 𝐿st = length of bearing stiffener
strengths.
𝑡 = base steel thickness of beam web

4.3.2 Stiffener Requirements In addition, the specification stipulates that w/ts ratios

for the stiffened and unstiffened elements of cold-formed
Section F5 of the 2016 edition of the North American steel bearing stiffeners should not exceed 1.28 𝐸∕𝐹ys
Specification, provides the following design requirements √
for attached bearing stiffeners and shear stiffeners. When and 0.42 𝐸∕𝐹ys , respectively. In the above expressions,
the bearing stiffeners do not meet these requirements, the 𝐹ys is the yield stress and 𝑡s is the thickness of the stiffener
load-carrying capacity for the design of such members can steel.
be determined by tests. It should be noted that Eq. (4.150) is used to prevent end
crushing of bearing stiffeners, while the second 𝑃n is used
a. Bearing Stiffeners. For beams having large h/t ratios, to prevent column buckling of the combined web stiff-
bearing stiffeners attached to beam webs at supports or ener section. The equations for computing the effective
184 4 FLEXURAL MEMBERS

areas 𝐴b and 𝐴c and the effective widths 𝑏1 and 𝑏2 1. Full bearing of the stiffener is required. If the bearing
are adopted from Nguyen and Yu.4.59 Figures 4.53 and width is narrower than the stiffener such that one of the
4.54 show the effective areas 𝐴c and 𝐴b of the bearing stiffener flanges is unsupported, 𝑃n shall be reduced by
stiffeners. 50%.
b. Bearing Stiffeners in C-Section Flexural Members. For 2. Stiffeners are C-section stud or track members with a
two-flange loading (Figs. 4.68c and 4.68d) of C-section minimum web depth of 3 12 in. (89 mm) and minimum
flexural members with bearing stiffeners that do not meet base steel thickness of 0.0329 in. (0.84 mm).
the above requirements of Section 4.3.2a, the nominal 3. The stiffener is attached to the flexural member web
strength 𝑃n should be determined as follows: with at least three fasteners (screws or bolts).
𝑃n = 0.7(𝑃wc + 𝐴e 𝐹y ) ≥ 𝑃wc (4.157) 4. The distance from the flexural member flanges to the
first fastener(s) is not less than d/8, where d is the
Ωc = 1.70 (ASD) overall depth of the flexural member.
{ 5. The length of the stiffener is not less than the depth of
0.90 (LRFD)
𝜙c = the flexural member minus 38 in. (9.53 mm).
0.80 (LSD)
6. The bearing width is not less than 1 12 in. (38.1 mm).
where 𝑃wc is the nominal web crippling strength for c. Shear Stiffeners. All shear stiffeners shall be designed to
C-section flexural members calculated in accordance with satisfy the following requirements for spacing, moment of
Eq. (4.197) for single web members, at end or inte- inertia, and gross area:
rior locations; 𝐴e is the effective area of the bearing 1. Spacing a between Stiffeners. When shear stiffeners
stiffener subjected to uniform compressive stress, calcu- are required, the spacing shall be based on the nominal
lated at yield stress; and 𝐹y is the yield stress of the shear strength 𝑉n permitted by Section 4.3.3.2f and the
bearing stiffener steel. Equation (4.157) is based on the following limits:
research conducted by Fox and Schuster at the Univer- ( )2
sity of Waterloo.1.299,4.229–4.231 The program investigated 260
𝑎≤ ℎ (4.158)
the behavior of 263 stud- and truck-type bearing stiff- ℎ∕𝑡
eners in cold-formed steel C-section flexural members 𝑎 ≤ 3ℎ (4.159)
subjected to two-flange loading at both interior and end
locations. This equation is applicable within the following 2. Moment of Inertia Is of Shear Stiffeners. With refer-
limits: ence to an axis in the plane of the web, the moment

Figure 4.53 Effective area Ac of bearing stiffener: (a) at end support; (b) at interior support and
under concentrated load.

Figure 4.54 Effective area Ab of bearing stiffener: (a) at end support; (b) at interior support and
under concentrated load.
DESIGN OF BEAM WEBS 185
of inertia of a pair of attached stiffeners or of a single where 𝑓v = actual shear stress
stiffener shall satisfy the following requirements: 𝑉 = total external shear force at a section
( ) 𝑄 = static moment of area between the extreme
ℎ 0.7𝑎
𝐼s ≥ 5ht3 − (4.160) fiber and the particular location at which the
𝑎 ℎ
( )4 shear stress is desired, taken about neutral axis
ℎ 𝐼 = moment of inertia of entire cross-sectional
𝐼s ≥ (4.161)
50 area about neutral axis
3. Gross area As of Shear Stiffeners. The area 𝐴s shall 𝑡 = width of section where shear stress is desired
satisfy the requirement of Eq. (4.162):
Even though Eq. (4.167) gives the exact value at any loca-
[ ]
1 − 𝐶𝑣 𝑎 (𝑎∕ℎ)2 tion, it has been a general practice to use the average value
𝐴s ≥ − √ YDht in the gross area of the web as the shear stress for design
2 ℎ (𝑎∕ℎ) + 1 + (𝑎∕ℎ)2
purposes. This average shear stress can be computed by using
(4.162)
the following equation:
where
𝑉
⎧ 1.53Ekv 𝑓v = (4.168)
⎪ when 𝐶v ≤ 0.8 (4.163) htw
⎪ 𝐹y (ℎ∕𝑡)
2
𝐶v = ⎨ √ where ℎ = depth of the flat portion of the web measured
Ekv
⎪ 1.11 when 𝐶v > 0.8 (4.164) along the plane of the web
⎪ ℎ∕𝑡 𝐹y 𝑡w = thickness of the web

⎧4.00 + 5.34 when 𝑎∕ℎ ≤ 1.0 (4.165) The use of Eqs. (4.167) and (4.168) is illustrated in
⎪ (𝑎∕ℎ)2 Example 4.20.
𝑘v = ⎨
4.00
⎪5.34 + when 𝑎∕ℎ > 1.0 (4.166)
⎩ (𝑎∕ℎ)2 Example 4.20 Determine the shear stress distribution at
and 𝑎 = distance between shear stiffeners the end supports of the uniformly loaded channel shown in
𝑌 = yield stress of web steel∕yield stress of Fig. 4.55. Assume that the load is applied through the shear
stiffener steel center of the cross section so that torsion is not involved.∗
𝐷 = 1.0 for stiffeners furnished in pairs See Appendix B for a discussion of the shear center.
= 1.8 for single-angle stiffeners
= 2.4 for single-plate stiffeners SOLUTION
Most of the above requirements for shear stiffeners 1. Exact Shear Stress Distribution Using Eq. (4.167).
are based on the AISC Specification1.148 and the study For simplicity, use square corners and midline dimen-
reported in Ref. 4.59. sions as shown in Fig. 4.56 for computing the exact
d. Nonconforming Stiffeners. According to Section F5.3 shear stresses at various locations of the section.
of the North American specification, the available a. Shear stress at points 1 and 4:
strength of members with stiffeners that do not meet the
requirements of Section 4.3.2a, 4.3.2b, or 4.3.2c, such 𝑉 = 𝑅A = 1.5 kips
as stamped or rolled-in stiffeners, shall be determined 𝑄1 = 𝑄4 = 0
by tests in accordance with Section K2 of the Specifica-
VQ1
tion or rational engineering analysis in accordance with =0
(𝑓v )1,4 =
Section A1.2(c) of the Specification. It
b. Shear stress at points 2 and 3:
( )
1
𝑄2 = 𝑄3 = 1.4325(0.135) × 6.865
4.3.3 Shear 2
= 0.664 in.3
4.3.3.1 Shear Stress In the design of beams, the actual
VQ2 1.5(0.664)
shear stress developed in the cross section of the beam can be (𝑓v )2,3 = = = 0.941 ksi
calculated by using the following well-known equation4.61 : It 7.84(0.135)

VQ
𝑓v = (4.167) ∗ When the load does not pass through the shear center, see Appendix B.
It
186 4 FLEXURAL MEMBERS

Figure 4.55 Example 4.20.

the beam is probably governed by shear yielding. The


maximum shear stress at the neutral axis can be computed
by Eq. (4.169):
𝐹y
𝜏y = √ (4.169)
3
in which τy is the yield stress in shear and 𝐹y is the yield
stress in tension.The nominal shear strength for yielding
can be determined by the shear stress given in Eq. (4.169)
and the web area, ht, as follows:
( )
𝐹y
𝑉n = √ (ht) ≅ 0.60𝐹y ht (4.170)
3
where 𝑉n is the nominal shear strength, h is the depth of
the flat portion of the web, and t is the thickness of the web.
Figure 4.56 Shear stress distribution.
b. Elastic Shear Buckling. For webs with large h/t ratios,
the shear capacity of the web is governed by shear
c. Shear stress at point 5: buckling. Figure 4.57 shows a typical pattern of shear
( )( )
1 1 failure.4.56 Based on studies by Southwell and Skan
𝑄5 = 𝑄2 + 0.135 × 6.865 × 6.865
2 4 on shear buckling of an infinitely long plate, the
= 1.46 in. 3 plate develops a series of inclined waves, as shown
VQ5 1.5(1.46) in Fig. 4.58.4.62,4.63 The elastic critical shear buckling
(𝑓v )5 = = = 2.07 ksi stress can be computed by Eq. (4.171)∗ :
It 7.84(0.135)
2. Average Shear Stress on Beam Web by Using 𝑘v 𝜋 2 𝐸
𝜏cr = (4.171)
Eq. (4.168) 12(1 − 𝜇2 )(ℎ∕𝑡)2
𝑉 1.5
𝑓v = = = 1.65 ksi where 𝑘v = shear buckling coefficient
htw 6.73(0.135) 𝐸 = modulus of elasticity of steel
From the above calculation it can be seen that for the 𝜇 = Poisson’s ratio
channel section used in this example the average shear stress ℎ = depth of plate
of 1.65 ksi is 25% below the maximum value of 2.07 ksi. 𝑡 = thickness of plate

4.3.3.2 Shear Strength of Beam Webs without Holes ∗ The problem of shear buckling of plane plates has also been studied by
a. Shear Yielding. When a beam web with a relatively small Timoshenko and other investigators. For additional information see Refs.
h/t ratio is subject to shear stress, the shear capacity of 3.1 and 4.63.
DESIGN OF BEAM WEBS 187

Figure 4.57 Typical shear failure pattern (h/t = 125).4.56

Figure 4.58 Shear buckling of infinitely long plate4.63 : (a.) simply supported edges; (b.) fixed edges.

In Eq. (4.171), the value of 𝑘v varies with the Substituting 𝜇 = 0.3 in Eq. (4.171),
supporting conditions and the aspect ratio a/h (Fig. 4.59), 0.904𝑘v 𝐸
𝜏cr = (4.172)
in which a is the length of the plate. For a long plate the (ℎ∕𝑡)2
value of 𝑘v was found to be 5.34 for simple supports and Thus if the computed theoretical value of τcr is less than
8.98 for fixed supports, as listed in Table 3.2. the proportional limit in shear, the nominal shear strength
188 4 FLEXURAL MEMBERS

By substituting the values of τpr and τcri into Eq. (4.174),


one can obtain Eq. (4.176) for the shear buckling stress in
the inelastic range, that is,

0.64 𝑘v 𝐹y 𝐸
𝜏cr = (4.176)
ℎ∕𝑡
Consequently, the nominal shear strength in the
inelastic range can be obtained from Eq. (4.177):

0.64 𝑘v 𝐹y 𝐸 √
𝑉n = (ht) = 0.64𝑡2 𝐾v 𝐹y 𝐸 (4.177)
ℎ∕𝑡
d. Direct Strength Method. The research by Pham and
Hancock4.309,4.310 showed that considerable tension field
action is available for local buckling if a web is fully
restrained at the loading and support points over its full
depth by bolted connections. This post-buckling strength
can be considered using the direct strength method
Figure 4.59 Shear buckling stress coefficient of plates versus that Pham and Hancock has developed. In addition, for
aspect ratio a/h.3.1 members with transvers stiffeners, the stiffness contri-
bution can also be included in web strength through the
for elastic buckling can be obtained from Eq. (4.173): Direct Strength Method as long as the critical shear buck-
ling force, 𝑉cr , used in the method includes the transverse
0.904𝑘v 𝐸 0.904𝑘v Et3 stiffener effects. Since it is difficult to develop analytical
𝑉cr = (ℎ∕𝑡) = (4.173)
(ℎ∕𝑡)2 ℎ solutions for shear buckling of webs with transverse stiff-
The above elastic critical buckling stress (τcr ) and the eners, numerical methods are employed. The methods
resultant (𝑉cr ) are based on the web alone ignoring the such as: semi-analytical finite strip method4.307 or spline
interaction from the flanges. Aswegan and Moen4.306 finite strip method4.308 can be used to determine the shear
developed analytical expressions which enable to take buckling force with stiffness contribution.
into consideration of the interactions of the connected The Direct Strength Method design for determining
elements. These equations are provided in Section 2.3.5 shear strength was first adopted into Appendix 1 of the
of the North American Specification.1.417 The shear 2012 edition of the North American Specification, and
buckling can also be determined numerically using the was incorporated into the main body of the 2016 edition
semi-analytical finite strip method4.307 or spline finite of the Specification as Section G2.2.
strip method4.308 developed by Hancock and Pham. e. Safety Factors. Prior to 1996, the AISI ASD Specifica-
c. Shear Buckling in Inelastic Range. For webs having tion employed three different safety factors (i.e., 1.44 for
moderate h/t ratios, the computed theoretical value of τcr yielding, 1.67 for inelastic buckling, and 1.71 for elastic
may exceed the proportional limit in shear. The theo- buckling) for determining the allowable shear stresses
retical value of the critical shear buckling stress should in order to use the same allowable values for the AISI
be reduced according to the change in the modulus of and AISC Specifications. To simplify the design of shear
elasticity. Considering the influence of strain hardening elements for using the allowable stress design method,
observed in the investigation of the strength of plate the safety factor for both elastic and inelastic shear buck-
girders in shear,4.64 Basler indicated that Eq. (4.174) can ling was taken as 1.67 in the 1996 edition of the AISI
be used as the reduction formula: Specification. The safety factor of 1.50 was used for
√ shear yielding to provide the allowable shear stress of
𝜏cr = 𝜏pr 𝜏cri (4.174)
√ 0.40Fy , which has been used in steel design for a long
where τpr = 0.8τy = 0.8(𝐹y ∕ 3) is the proportional limit time. The use of such a smaller safety factor of 1.50 for
in shear and the initial critical shear buckling stress is shear yielding was justified by long-standing use and by
given as the minor consequences of incipient yielding in shear as
𝑘v 𝜋 2 𝐸 compared with those associated with yielding in tension
𝜏cri = (4.175) and compression.
12(1 − 𝜇2 )(ℎ∕𝑡)2
DESIGN OF BEAM WEBS 189
In the 2001 and 2007 editions of the North American where
Specification, the constant used in Eq. (4.176) for deter- 𝐴w = Area of web element
mining the inelastic shear buckling stress was reduced
slightly from 0.64 to 0.60 on the basis of Craig’s calibra- = ht (4.182b)
tion of the test data of LaBoube and Yu.1.346,4.56,4.232,4.233
where h = Depth of flat portion of web measured along plane
For the purpose of simplicity, a single safety factor of 1.60
of web
was used in the 2007 edition the Specification for shear
t = Web thickness
yielding, elastic and inelastic shear buckling for the ASD 𝐹y = Design yield stress as determined in accordance
method with a corresponding resistance factor of 0.95 with Specification Section A3.3.1
for LRFD and 0.80 for LSD. These safety and resistance 𝑉cr = Elastic shear buckling force as defined in
factors are retained in Section G2 of the 2016 edition of Specification Section G2.3 for flat web alone, or
the Specification.1.417 With these minor revisions of safety determined in accordance with Specification
factor and design equations, for ASD, the North Amer- Appendix 2 for full cross-section of prequalified
ican Specification reduces 6% of the allowable inelastic (Table 3.1) members
shear buckling strength as compared with the AISI 1996 E = Modulus of elasticity of steel
ASD Specification. For beam webs with large h/t ratios, 𝑘v = Shear buckling coefficient, determined in
accordance with Specification Section G2.3
the North American Specification allows 4% increase of
the allowable strength for the elastic shear buckling.
f. North American Design Criteria for Shear Strength of G2.2 Flexural Members With Transverse Web
Webs without Holes. Based on the foregoing discussion Stiffeners
of the shear strength of beam webs, the 2016 edition of For a reinforced web with transverse web stiffeners meeting the
the North American Specification includes the following criteria of Specification Section G4, and spacing not exceeding
design provisions in Section G2 for the ASD, LRFD, and twice the web depth, this section is permitted to be used to
LSD methods.1.417 In the 2016 edition of the Specifica- determine the nominal shear strength [resistance], 𝑉n , in lieu of
tion, the equation format has been revised from shear Specification Section G2.1.
For λv ≤ 0.776,
stress to shear resultants by multiplying the web shear
𝑉n = 𝑉y (4.183)
area.
For λv > 0.776,
[ ( )0.4 ] ( )0.4
G2.1 Flexural Members Without Transverse Web 𝑉 𝑉cr
𝑉𝑛 = 1 − 0.15 cr 𝑉𝑦 (4.184)
Stiffeners 𝑉𝑦 𝑉𝑦
The nominal shear strength [resistance], 𝑉n , of flexural members
where 𝑉cr = Elastic shear buckling force as defined in
without transverse web stiffeners shall be calculated as follows:
Specification Section G2.3 for flat web alone, or
For 𝜆𝑣 ≤ 0.815,
determined in accordance with Specification
𝑉 n = 𝑉y (4.178)
Appendix 2 for full cross-section of prequalified
For 0.815 < λv ≤ 1.227 (Table 3.1) members

𝑉n = 0.815 𝑉cr 𝑉y (4.179a) Other variables are defined in Specification Section G2.1.

= 0.60𝑡2 𝐸𝑘𝑣 𝐹y (4.179b) G2.3 Web Elastic Critical Shear Buckling Force, V cr
The shear buckling force, 𝑉cr , of a web is permitted to be deter-
For λv > 1.227 mined in accordance with this section:
𝑉n = 𝑉cr (4.180a) 𝑉cr = 𝐴w 𝐹cr (4.185)
= 0.904Ekv 𝑡 ∕h
3
(4.180b) where 𝐴w = Web area as given in Eq. (4.182b)
where √ 𝐹cr = Elastic shear buckling stress
𝑉𝑦
𝜆𝑣 = (4.181) 𝜋 2 Ek𝑣
𝑉cr = (4.186)
12(1 − 𝜇 2 )(ℎ∕𝑡)2

𝑉y = Yield shear force of cross-section where 𝐸 = Modulus of elasticity of steel


𝑘v = Shear buckling coefficient calculated in
= 0.6 𝐴w 𝐹y (4.182a) accordance with (a) or (b) as follows:
190 4 FLEXURAL MEMBERS

(a) For unreinforced webs, 𝑘v = 5.34


(b) For webs with transverse stiffeners satisfying the require-
ments of Specification Section G4
when
𝑎∕ℎ ≤ 1.0
5.34
𝑘𝑣 = 4.00 + (4.187a)
(𝑎∕ℎ)2
when
𝑎∕ℎ > 1.0
4.00
𝑘𝑣 = 5.34 + (4.187b)
(𝑎∕ℎ)2

where 𝑎 = Shear panel length of unreinforced web element


= Clear distance between transverse stiffeners of
reinforced web elements
Other variables are defined in Specification Section G2.1.
For the ASD method, the allowable shear stresses in webs
are shown in Fig. 4.60 by using 𝐸 = 29,500 ksi (203 GPa,
or 2.07 × 106 kg∕cm2 ). Table 4.7 gives the allowable shear
stresses for 𝐹y = 33 and 50 ksi (228 and 345 MPa, or 2320
and 3515 × 106 kg∕cm2 ). Examples 4.21–4.23 illustrate the
applications of the shear design provisions.

Example 4.21 Use the ASD and LRFD methods to deter-


mine the available shear strength for the I-section used in
Example 4.4. Use 𝐹y = 50 ksi.
Figure 4.60 Allowable shear stress in webs for the ASD method.

SOLUTION
The allowable shear strength for the I-section having two
A. ASD Method
The depth-to-thickness ratio of each individual web element webs is
is 2𝑉 2(29.79)
ℎ 8 − 2(0.135 + 0.1875) 7.355 𝑉a = n = = 37.24 kips
= = = 54.48 Ωv 1.60
𝑡 0.135 0.135
Based on the North American design criteria, the value B. LRFD Method
of 𝑘v for unreinforced webs is 5.34. Therefore, according to Using the same nominal shear strength computed in item
Section G2.3 of the North American Specification, A, the design shear strength for the I-section having double
webs is
𝜋 2 Ek𝑣 𝜋 2 (29,500)(5.34)
𝐹cr = = 2𝜙v 𝑉n = 2(0.95)(29.79) = 56.60 kips
12(1 − 𝜇2 )(ℎ∕𝑡)2 12(1 − 0.32 )(54.48)2
= 47.97 ksi
𝑉cr = 𝐴w 𝐹cr = (7.355)(0.135)(47.97) = 47.63 kips Example 4.22 Use the ASD and LRFD methods to deter-
mine the available shear strength for the channel section used
From Specification Section G2.1: in Example 4.5. Use 𝐹y = 50 ksi.
𝑉y = 0.6Aw Fy = 0.6(7.355)(0.135)(50) = 29.79 kips
√ √ SOLUTION
𝑉𝑦 29.79 A. ASD Method
𝜆= = = 0.791
𝑉cr 47.63 For the given channel section, the depth-to-thickness ratio of
the web is
Since λ < 0.815
ℎ 10 − 2(0.075 + 0.09375) 9.6625
𝑉n = 𝑉y = 29.97 kips = = = 128.83
𝑡 0.075 0.075
DESIGN OF BEAM WEBS 191
Table 4.7 Allowable Shear Stresses for ASD Method, ksi
𝐹y = 33 ksi 𝐹y = 50 ksi
a/h a/h
h/t 0.5 1.0 2.0 3.0 >3 0.5 1.0 2.0 3.0 >3
50 12.4 12.4 12.4 12.4 12.4 18.8 18.8 18.8 18.8 18.8
60 12.4 12.4 12.4 12.4 12.4 18.8 18.8 18.8 18.3 17.5
70 12.4 12.4 12.4 12.4 12.2 18.8 18.8 16.4 15.6 15.0
80 12.4 12.4 11.6 11.1 10.7 18.8 17.4 14.3 13.7 13.2
90 12.4 12.4 10.4 9.9 9.5 18.8 15.5 12.7 11.9 11.0
100 12.4 11.3 9.3 8.9 8.6 18.8 13.9 10.6 9.6 8.9
110 12.4 10.3 8.5 8.0 7.4 18.8 12.7 8.7 8.0 7.4
120 12.4 9.4 7.3 6.7 6.2 18.8 10.8 7.3 6.7 6.2
130 12.4 8.7 6.3 5.7 5.3 17.6 9.2 6.3 5.7 5.3
140 12.4 7.9 5.4 4.9 4.5 16.4 7.9 5.4 4.9 4.5
150 12.4 6.9 4.7 4.3 4.0 15.3 6.9 4.7 4.3 4.0
160 11.6 6.1 4.1 3.8 3.5 14.3 6.1 4.1 3.8 3.5
170 11.0 5.4 3.7 3.3 3.1 13.5 5.4 3.7 3.3 3.1
180 10.4 4.8 3.3 3.0 2.7 12.7 4.8 3.3 3.0 2.7
190 9.8 4.3 2.9 2.7 2.5 11.7 4.3 2.9 2.7 2.5
200 9.3 3.9 2.6 2.4 2.2 10.6 3.9 2.6 2.4 2.2
220 8.5 3.2 2.2 2.0 1.8 8.7 3.2 2.2 2.0 1.8
240 7.3 2.7 1.8 1.7 1.5 7.3 2.7 1.8 1.7 1.5
260 6.3 2.3 1.6 1.4 1.3 6.3 2.3 1.6 1.4 1.3
280 5.4 2.0 1.3 1.2 1.1 5.4 2.0 1.3 1.2 1.1
300 4.7 1.7 1.2 1.1 1.0 4.7 1.7 1.2 1.1 1.0
Notes:
𝑎
Values above the single underlines are based on Eq. (4.178); values between the single underlines and double underlines are based on
Eq. (4.179); and values below the double underlines are based on Eq. (4.180). For the case of a/h > 3, 𝑘v = 5.34.
𝑏
1 ksi = 6.9 MPa or 70.3 kg/cm2 .

Based on 𝑘v = 5.34 and 𝐹y = 50 ksi, The allowable shear strength using the ASD method for
the design of the given channel section is
𝜋 2 Ek𝑣 𝜋 2 (29,500)(5.34)
𝐹cr = = 𝑉n 6.22
12(1 − 𝜇 )(ℎ∕𝑡)
2 2
12(1 − 0.32 )(128.8)2 𝑉a = = = 3.89 kips
Ωv 1.60
= 8.578 ksi
B. LRFD Method
Based on the same nominal shear strength computed in item
𝑉cr = 𝐴w 𝐹cr = (9.6625)(0.075)(8.578) = 6.217 kips A, the design shear strength using the LRFD method for the
From Specification Section G2.1: design of the channel section is

𝑉y = 0.6𝐴w 𝐹y = 0.6(9.6625)(0.075)(50) = 36.23 kips 𝜙v 𝑉n = (0.95)(6.22) = 5.91 kips


√ √
𝑉y 26.23
𝜆= = = 2.4 > 1.227
𝑉cr 6.217
Example 4.23 Use the ASD and LRFD methods to deter-
Therefore, mine the available shear strength for the hat section used in
𝑉n = 𝑉cr = 6.217 kips Example 4.6. Use 𝐹y = 50 ksi.
192 4 FLEXURAL MEMBERS

SOLUTION in Eq. (4.188a) or (4.188b), under the same limits given in


A. ASD Method Section 3.4.3:
The depth-to-thickness ratio of the web is ⎧ 𝑐
ℎ 10 − 2(0.105 + 0.1875) 9.415 ⎪1.0 when 𝑡 ≥ 54 (4.188𝑎)
= = = 89.67 𝑞s = ⎨ 𝑐 𝑐
𝑡 0.105 0.105 ⎪ 54𝑡 when 5 ≤ 𝑡 < 54 (4.188𝑏)

Based on 𝑘v = 5.34 and 𝐹y = 50 ksi,
⎧ ℎ − 𝑑h for circular holes (4.189𝑎)
𝜋 2 Ek𝑣 𝜋 2 (29, 500)(5.34) ⎪
𝐹cr = = 𝑐 = ⎨ 2 2.83
12(1 − 𝜇2 )(ℎ∕𝑡)2 12(1 − 0.32 )(89.67)2 ⎪ ℎ − 𝑑h for noncircular holes (4.189𝑏)
⎩2 2
= 17.71 ksi
where 𝑑h = depth of web hole
𝑉cr = 𝐴w 𝐹cr = (9.415)(0.105)(17.71) = 17.505 kips ℎ = depth of flat portion of web measured along
plane of the web
From Specification Section G2.1: 𝑡 = web thickness
𝑉y = 0.6𝐴w 𝐹y = 0.6(9.415)(0.105)(50) = 49.429 kips Similar to Section 1.1.1 of the Specification (Section 3.4.3
√ √ in this volume), the above design provisions for circular and
𝑉y 49.429 noncircular holes apply to any hole pattern that fits within an
λ= = = 1.680 > 1.227
𝑉cr 17.505 equivalent virtual hole, as shown in Figs. 3.65 and 3.66.
Therefore, 𝑉n_single = 𝑉cr . For two webs, the total nominal
shear strength is 4.3.4 Bending
Webs of beams can buckle not only in shear but also due to
𝑉n = 2𝑉cr = 2(17.505) = 35.01 kips
the compressive stress caused by bending, for example, at the
𝑉n 35.01 location of a maximum moment. Figure 3.27 shows a typical
𝑉𝑎 = = = 21.88
Ωv 1.60 pattern of bending failure of beam webs.
The web buckling stress due to bending and the post-
B. LRFD Method buckling strength of flat beam webs are discussed in
Based on the same nominal shear strength computed in item Section 3.3.1.2. The same section also discusses the AISI
A, the design shear strength using the LRFD method for the design equations for computing the effective design depth
design of the hat section is of beam webs.
For beam webs having relatively large depth-to-thickness
𝜙v 𝑉n = (0.95)(35.01) = 33.26 kips
ratios, the buckling of web elements becomes more impor-
tant. The structural efficiency of such beam webs can be
4.3.3.3 Shear Strength of C-Section Webs with Holes improved by adding longitudinal stiffeners in the compres-
When holes are present in beam webs, the effect of web sion portion of the web, as shown in Fig. 4.61. References
perforation on the reduction of shear strength of C-sections 4.60, 4.65, and 4.66 present the studies made by Nguyen
was investigated in the 1990s by Shan et al.,3.184,3.197 and Yu on the structural behavior of longitudinally reinforced
Schuster et al.,3.187 and Eiler.3.192 In these studies, three beam webs.
hole geometries (rectangular hole with corner fillets,
circular hole, and diamond-shaped hole) were considered
in the test programs. Based on the results of research
findings, the design provisions were added in the supple-
ment to the 1996 edition of the AISI Specification in
19991.333 as Section C3.2.31.336,1345 and is retained in
the 2016 edition of the North American Specification as
Section G3. 1.417
Based on the Specification provisions, for C-section webs
with holes, the shear strength shall be calculated in accor-
dance with Specification Section G2 for section without Figure 4.61 Typical sections for longitudinally reinforced beam
holes and multiplied by the reduction factor 𝑞s , as defined specimens.
DESIGN OF BEAM WEBS 193
In Europe, the design methods for profiled sheeting and beam webs subjected to the combined bending and shear
sections with stiffeners in the flanges and webs are provided stresses:

in Refs. 1.209 and 3.56. ( ) ( )2
The Direct Strength Method provided in Section G2.2 of 𝑓bw 2 𝑓v
+ ≤1 (4.191)
the North American Specification can handle the available 𝐹bw 𝐹v
bending strength of beams having longitudinal web stiff-
eners. where 𝑓bw = actual compression stress at junction of flange
and web
4.3.5 Combined Bending and Shear 𝐹bw = allowable compressive stress
𝑓v = actual average shear stress
When high bending and high shear act simultaneously, as 𝐹v = allowable shear stress,
in cantilever beams and at interior supports of continuous
For additional information on simply supported plates
beams, the webs of beams may buckle at a lower stress than
under combined shear and uniformly distributed longitudinal
if only one stress were present without the other. Such webs
stresses, see Ref. 3.3.
must be safeguarded against buckling due to this combined
In the past, the structural strength of beam webs subjected
bending and shear.
The critical combination of bending and shear stresses to a combination of bending and shear has been studied by
in disjointed flat rectangular plates has been studied by LaBoube and Yu.4.57,4.67 The results of 25 beam tests indi-
Timoshenko.3.2 Figure 4.62 shows the interaction between cated that Eq. (4.190a), originally developed for a disjointed
𝑓b ∕𝑓cr and τ∕τcr , in which 𝑓b is the actual computed bending individual sheet, would be conservative for beam webs with
stress, 𝑓cr is the theoretical buckling stress in pure bending, adequate transverse stiffeners, for which a diagonal tension
𝜏 is the actual computed shear stress, and τcr is the theo- field action can be developed. Based on the test results shown
retical buckling stress in pure shear. It can be seen from in Fig. 4.63, the following interaction equation was devel-
Fig. 4.62 that for a/h ratios ranging from 0.5 to 1.0 the rela- oped for beam webs with transverse stiffeners satisfying the
tionship between 𝑓b ∕𝑓cr and τ∕τcr can be approximated by requirements of Section 4.3.2:
Eq. (4.190), which is a part of the unit circle: 𝑓b 𝜏
( )2 ( )2 0.6 + = 1.3 (4.192)
𝑓b 𝜏 𝑓b,max 𝜏max
+ =1 (4.190a)
𝑓cr 𝜏cr
or √
( )2 ( )2
𝑓b 𝜏
+ ≤1 (4.190b)
𝑓cr 𝜏cr
By using proper safety factors, the following interac-
tion formula can be used for the allowable stress design of Eq. (4.192)

Eq. (4.190a)

Figure 4.62 Interaction relationship between 𝑓𝑏 ∕𝑓cr and


T∕Tcr .4.63 Figure 4.63 Interaction diagram for T∕Tmax and fb ∕fb,max .

You might also like