AISC Engineering Journal Q1 2010

Engineering
Journal
American Institute of Steel Construction
First Quarter 2010 Volume 47, No. 1
1 Critical Evaluation of Equivalent Moment Factor

Procedures for Laterally Unsupported Beams
Edgar Wong and Robert G. Driver
21 Impact of Diaphragm Behavior
on the Seismic Design of Low-Rise
Steel Buildings
Colin A. Rogers and Robert Tremblay
37 Notes on the Impact of Hole Reduction
on the Flexural Strength of Rolled Beams
Louis F. Geschwindner
41 A Case for a Single Stiffness Reduction
Factor in the 2010 AISC Specification
Louis F. Geschwindner
47 Bolt Shear Design Considerations
Raymond H.R. Tide
65 Current Steel Structures Research
Reidar Bjorhovde
www.aisc.org
ENGINEERING
JOURNAL
AMERICAN INSTITUTE OF STEEL CONSTRUCTION
Dedicated to the development and improvement of steel construction,

through the interchange of ideas, experiences and data.
Editorial Staff
Editor: Keith Grubb, P.E., S.E.
Research Editor: Reidar Bjorhovde, Ph.D.
Production Editor: Areti Carter
Officers
David Harwell, Chairman
Central Texas Iron Wroks, Inc., Waco, TX
William B. Bourne, III, Vice Chairman
Universal Steel, Inc., Atlanta, GA
Stephen E. Porter, Treasurer
Indiana Steel Fabricating, Inc., Indianapolis, IN
Roger E. Ferch, P.E., President
American Institute of Steel Construction, Chicago
David B. Ratterman, Secretary & General Counsel
Charles J. Carter, S.E., P.E., Ph.D., Vice President and
Chief Structural Engineer
John P. Cross, P.E., Vice President
Louis F. Geschwindner, P.E., Ph.D., Vice President, Special Projects
American Institute of Steel Construction, University Park, PA
Scott L. Melnick, Vice President
The articles contained herein are not intended to represent official attitudes,
recommendations or policies of the Institute. The Institute is not responsible for any
statements made or opinions expressed by contributors to this Journal.
The opinions of the authors herein do not represent an official position of the
Institute, and in every case the officially adopted publications of the Institute
will control and supersede any suggestions or modifications contained in any
articles herein.
The information presented herein is based on recognized engineering principles
and is for general information only. While it is believed to be accurate, this
information should not be applied to any specific application without competent
professional examination and verification by a licensed professional engineer.
Anyone making use of this information assumes all liability arising from such use.
Manuscripts are welcomed, but publication cannot be guaranteed. All manuscripts
should be submitted in duplicate. Authors do not receive a remuneration. A “Guide
for Authors” is printed on the inside back cover.
ENGINEERING JOURNAL (ISSN 0013-8029) is published quarterly.
Subscriptions: Members: one subscription, $20 per year, included in dues;
Additional Member Subscriptions: $15 per year. Non-Members U.S., Canada, and Subscribe to Engineering Journal by visiting our web site
Mexico: $40 per year, $110 for three years, single copy $15. International Members www.aisc.org/ej or by calling 312.670.5444.
and Non-Members: $90 per year; $250 for three years; single copy $25. Published
Copies of current and past Engineering Journal articles
by the American Institute of Steel Construction at One East Wacker Drive, Suite
are available free to members online at www.aisc.org/ej.
700, Chicago, IL 60601.
Periodicals postage paid at Chicago, IL and additional mailing offices. Non-members may purchase Engineering Journal article
Postmaster: Send address changes to ENGINEERING JOURNAL in care of downloads at the AISC Bookstore at www.aisc.org/ej for
the American Institute of Steel Construction, One East Wacker Drive, Suite 700, $10 each.
Chicago, IL 60601.
Copyright 2010 by the American Institute of Steel Construction. All rights
reserved. No part of this publication may be reproduced without written permission.
The AISC logo is a registered trademark of AISC.
Critical Evaluation of Equivalent Moment Factor
Procedures for Laterally Unsupported Beams
EDGAR WONG and ROBERT G. DRIVER
ABSTRACT
This paper compares the numerous approaches to determining equivalent moment factors used in evaluating the elastic critical moment of later-
ally unsupported beams for a wide variety of moment distributions. The investigation revealed that the procedure used currently in the Canadian
design standard produces unacceptable results for the majority of the common bending moment distributions considered. Large abrupt changes
in Cb values with only slight changes in the shape of the moment diagram were observed in 6 out of the 12 moment distribution comparisons,
which contributes to the overall poor performance of the procedure.
The study also revealed drawbacks inherent in other methods. Overall, the quarter-point moment equations developed for general moment
distributions capture the trends of the numerical data reasonably well. However, for example, the evaluations show that the 2005 AISC equation
produces non-conservative results in some situations, while the British equation, although generally conservative, produces comparatively less
accurate results. Other equations examined capture the trends of the numerical data more consistently by implementing a square root format
in the quarter-point moment method. However, they produce results that exceed the numerical data in several cases, implying that they are too
aggressive for design purposes.
To capture the best features of the various methods investigated, yet improve the overall suitability for general design purposes, a modified
quarter-point moment equation using the square root format is proposed. Not only does it simulate the trends of the numerical solutions closely,
but it also produces reasonable and conservative equivalent moment factors, even in cases where other methods do not. Like all quarter-point
moment methods, the proposed equation does not produce good results in some situations where concentrated moments are applied. Never-
theless, it is believed to be appropriate for the vast majority of typical design cases.
Keywords: lateral support, equivalent moment factors, Cb, beams.
INTRODUCTION 1. The internal moment distribution between brace

points;
The elastic lateral-torsional buckling moment capacity of a
doubly-symmetric steel beam is governed primarily by the 2. The elevation of the applied load with respect to the
member’s weak-axis moment of inertia, Iy , and the torsion shear center;
parameter. The latter factor can be expressed as
3. The degree of lateral, rotational, and warping restraint
ECw at the brace points; and
L GJ
4. The potential for less critical adjacent unbraced seg-
where L is the laterally (and torsionally) unbraced length of ments to restrain buckling (i.e., interaction buckling).
the beam, E is the elastic modulus, G is the shear modulus,
Cw is the warping constant, and J is the torsional constant. Although methods that consider all of these factors in the
Previous research has shown that the following factors can computation of the elastic critical moment are available,
also influence the critical moment capacity significantly most steel design specifications simplify the analytical pro-
(Clark and Hill, 1960; Nethercot and Rockey, 1972; Nether- cess by accounting only for the moment distribution effect
cot and Trahair, 1976): among the four factors. That is, loads are assumed to be ap-
plied at the shear center (unless, perhaps, they are applied
significantly above the shear center (for downward loads) by
a means that does not also serve as a brace), lateral braces
Edgar Wong, M.Eng., P.Eng., structural engineer, Walters Chambers and are assumed to prevent both lateral displacement and twist
Associates Ltd., 501, 10709 Jasper Ave., Edmonton, AB, T5J 3N3, Canada.
E-mail: [email protected]
of the beam’s cross section, while restraining neither weak-
axis rotation nor warping, and the potential for interaction
Robert G. Driver, Ph.D., P.Eng., Professor, Department of Civil and Environ-
mental Engineering, University of Alberta, Edmonton, AB, T6G 2W2, Canada buckling is neglected. For these reasons, this paper addresses
(corresponding author). E-mail: [email protected] the effect of the moment distribution only.
ENGINEERING JOURNAL / FIRST QUARTER / 2010 / 1

For a doubly-symmetric beam subject to a uniform (con- member delivers its reaction load to the primary beam away
stant) moment about the strong axis along its length, the from its compression flange without providing significant
critical lateral-torsional buckling capacity, Mcr , can be ex- rotational restraint to the beam and the designer deems this
pressed as: to be inadequate as a bracing mechanism. Other typical ex-
amples of loads that are not associated with the provision
E 2 of effective bracing include suspended loads, supported col-
Mcr EI y GJ I y Cw (1)
L L umn reaction loads and loads where the connection of the
tributary beam to the primary beam is bolted and employs
The boundary conditions assumed in this equation are such horizontally slotted holes. Considering the multitude of con-
that both ends of the unbraced segment are restrained as ditions that a structural designer may face, group 1, 2 and 3
described in the previous paragraph. The value determined moment distributions must all be included in any evaluation
from Equation 1 is commonly referred to as the “basic” of equivalent moment factors.
lateral-torsional buckling moment capacity, not only because There are many equations and methods published in the
it is the simplest to derive theoretically, but, more impor- general literature and design specifications for determining
tantly, it gives the lowest possible capacity of a beam seg- equivalent moment factors. In this paper, comparisons of
ment between properly designed brace points when loads are equivalent moment factors determined using various meth-
applied at the shear center (Kirby and Nethercot, 1979). It ods for 12 different moment distribution types, described in
is widely accepted that the effect of a non-uniform moment Table 1, are presented. In order to generalize the moment
distribution can be approximated simply by multiplying distribution types, three factors are introduced in Table 1:
Equation 1 by an equivalent (uniform) moment factor, Cb. (1) for Type 1, the factor L is the ratio of the absolute value
Since a non-uniform moment distribution is less severe than of the smaller to larger end moment of the unbraced seg-
a uniform one, the value of this factor is always greater than ment, and it is taken as positive for double curvature bending
or equal to 1.0. and negative for single curvature; (2) for Types 2 to 5, 8, 9,
In general, non-uniform moment distributions between 11 and 12, the factor C is the ratio of the actual end moment
brace points can conveniently be categorized into three to the fixed end moment; and (3) for Types 6, 7 and 10, the
groups: (1) linear moment distributions arising when there factor a is the distance from a concentrated load to the near-
are no loads or moments applied between brace points; est vertical support (see Table 1). The moment distribution
(2) non-linear moment distributions with multiple constant types selected are believed to be common enough to cor-
moment gradient regions; and (3) non-linear moment distri- respond with typical design loading cases and cover a broad
butions with continuously varying moment gradients. The enough range to lead to general conclusions.
primary difference between the last two groups is that beams The original objective of this study was to examine the
within group 2 are not subjected to any distributed load and adequacy of the equivalent moment factor specified in Cana-
their moment distributions can be transformed into group 1 dian standard CAN/CSA-S16-01. During the course of this
distributions by adding braces at points where the moment examination, a critical evaluation of other published meth-
gradient changes. It is important to realize that some existing ods was also conducted. Not only are the shortcomings of
equivalent moment factor equations have been derived for the CAN/CSA-S16-01 procedure clarified, but a broad col-
group 1 moment distributions only, whereas others purport lection of solutions determined by other methods is also pre-
to be applicable for all groups. Misusing the equations may sented herein to illustrate their performance and limitations.
lead to significant errors in critical moment predictions. Although relevant physical test data are scarce, numerical
It should be noted for clarity that in many cases the means data are included as reference values where available. Fi-
of delivering loads to a beam will also provide effective lat- nally, a new equivalent moment factor equation is proposed
eral bracing to that beam, apparently making group 1 mo- based on the findings of the investigation that incorporates
ment distributions the only case that will occur in practice. the best features observed in the various existing methods. It
However, circumstances where loads are applied to a beam is demonstrated that the proposed equation effectively rec-
with little, or perhaps uncertain, resulting bracing effective- tifies current CAN/CSA-S16-01 deficiencies and produces
ness are relatively common. One example of this is where accurate, yet conservative, approximations to the numerical
two parallel primary beams have loads delivered to them by solutions over a wide range of moment distribution types.
simply supported transverse secondary framing members
(i.e., the two primary beams “lean on” each other with re- PROCEDURES PUBLISHED IN THE
spect to the intermediate lateral support points at the ends of GENERAL LITERATURE
the secondary members). If the two primary beams become
unstable at a similar time in the loading regime, they can- The fundamental aspects that characterize the non-uniform
not be considered to support each other laterally (Galam- moment effect are the rate of change of the moment along
bos, 1998). Another common example is where a secondary the beam length, the number of curvature reversals between
2 / ENGINEERING JOURNAL / FIRST QUARTER / 2010

Table 1. Types of Moment Distributions Considered in the Study *
Type 1—End Moments Only
Type 2—Uniformly Distributed Load with Equal End Moments
Type 3—Uniformly Distributed Load with One End Moment
Type 4—Mid-span Concentrated Load with Equal End Moments
Type 5—Mid-span Concentrated Load with One End Moment
Type 6—Concentrated Load with Pinned Ends
* The ends of the members depicted are brace points, as well as points denoted by the symbol x.

Table 1 (cont.). Types of Moment Distributions Considered in the Study *
Type 7—Two Equal Concentrated Loads Symmetrically Placed with Pinned End
Type 8—Uniformly Distributed Load with Equal End Moments, Brace at Mid-span
Type 9—Uniformly Distributed Load with One End Moment, Brace at Mid-span
Type 10—Two Equal Concentrated Loads Symmetrically Placed, Brace at Mid-span
Type 11—Two Equal Concentrated Loads at Third Points with Equal End Moments
Type 12—Two Equal Concentrated Loads at Third Points with One End Moment
* The ends of the members depicted are brace points, as well as points denoted by the symbol x.

brace points, and the distance between the maximum moment original publication, it has been indicated in numerous sub-
and the brace locations. Due to the challenge of developing a sequent publications that using the absolute values of these
simple procedure for determining equivalent moment factors moments in the equation is appropriate. Unlike Equations 2
that can capture all three aspects concurrently for all kinds and 3, the quarter-point moment methods are independent of
of moment distributions, not all published methods are ap- the magnitudes of the end moments, unless one or both are
plicable to all moment distribution types. Available methods also the maximum moment in the unbraced segment.
for determining equivalent moment factors can be divided Another quarter-point moment method was developed by
into three categories: Serna et al. (2006) by curve fitting their numerical analysis
results that account not only for the effect of a non-uniform
1. Methods developed for unequal end moments only
moment distribution, but also the lateral, rotational and warp-
(e.g., Salvadori, 1955; Austin, 1961).
ing restraints at the brace points. Since the latter influences are
2. Methods developed for a general moment distribution not within the scope of this paper, the equation is written in a
(e.g., Kirby and Nethercot, 1979; Serna et al., 2006) simplified form for laterally and torsionally simple end condi-
tions as Equation 5 in Table 2. The main distinction of this
3. Methods developed to address numerous specific mo- equation as compared to Equation 4 is that the individual mo-
ment distributions individually (e.g., Trahair, 1993; ment terms are squared and a square root format is assigned.
Clark and Hill, 1960; Suryoatmono and Ho, 2002).
Methods Developed for Specific Moment Distributions
Methods Developed for Unequal End Moments Only
Clark and Hill (1960) and Nethercot and Trahair (1976) each
Two equations that are commonly used to approximate the published a list of equivalent moment factor values based on
effect of a constant moment gradient between brace points numerical analyses for specific non-uniform moment distri-
on the critical elastic moment are Equations 2 and 3 shown butions, as shown in Table 3. Although not applicable to all
in Table 2. The parameter L quantifies the influence of the typical design loading cases, they provide a good database
flange force variation between the two ends. That is, if a from which designers can approximate equivalent moment
beam flange is subjected to a bending-induced compression factors for other distributions. The two sets are nearly identi-
that varies between lateral supports, the degree of varia- cal, except for the value in the Type 2 distribution when C
tion dictates the tendency of the beam to buckle elastically equals 1.0. For this case, the value of 1.30 from Clark and
(Zuraski, 1992). Furthermore, if the flange force varies be- Hill (1960) appears to be incorrect, and if it is recalculated
tween tension and compression (i.e., when the unbraced seg- using the original source of data, a value of 2.52 is obtained,
ment of the beam is in double curvature), the beam is even as reflected in Table 3.
less susceptible to lateral-torsional buckling. Equation 2 Instead of discrete values, Trahair (1993) published indi-
represents a lower bound to the original solutions developed vidual equations (see Table 3) for several moment distribution
by Salvadori (1955) using the Rayleigh-Ritz method, and types based on curve fitting of numerical data. These equations
Equation 3 is from the work of Austin (1961) for in-plane apply to a much wider range of moment distributions than do
bending of beam-columns. Equation 3 is considered inap- the lists of discrete values because the designer can adjust the
propriate for assessing out-of-plane buckling due to flexure point load location along a beam or the magnitude of the end
alone because it is derived for members subjected to both moments. Nethercot and Rockey (1972) also proposed a Cb
axial load and bending simultaneously (AISC, 2005b), and, equation for moment Type 7 that is a function of the distance
as such, it is not considered further. between the point load and the closer support. Analogous
equations presented by Suryoatmono and Ho (2002) are rel-
Methods Developed for a General Moment atively complex as compared to those of Trahair (1993), and
Distribution—Quarter-Point Moment Methods they address moment Types 1 through 3 only; therefore, they
are not included in the comparisons in this paper.
Equation 4, shown in Table 2, was developed to be applicable
to all types of moment distributions (Kirby and Nethercot,
PROCEDURES IN DESIGN SPECIFICATIONS
1979). It utilizes the magnitudes of the bending moments
at four specific locations along the unbraced segment: the Because the methodologies discussed in the previous section
quarter point, Ma, centerline, Mb, third-quarter point, Mc, have various degrees of practicality, accuracy, consistency
and maximum, Mmax, moments. Equations with this format and computational complexity, different steel design speci-
are referred to as the “quarter-point moment methods” in fications have adopted different procedures for determining
this paper. The main function of these four moments is to the equivalent moment factor to be used for the design of
describe the degree of non-uniformity of the moment along laterally unsupported beams. Nevertheless, the majority use
the unbraced length, thus approximating its influence on a single method that is intended to be applicable to all types
the critical moment. Although not specified explicitly in the of moment distributions.

Table 2. Equivalent Moment Factor Equations Evaluated in This Study
Equation a Publication Equivalent Moment Factor Equation
2 Salvadori (1955) b, c Cb 1.75 1.05 0.3 2 2.3
3 Austin (1961) Cb (0.6 0.4 ) 1 2.5
12Mmax
4 Kirby and Nethercot (1979) Cb =
2Mmax + 3Ma + 4Mb + 3Mc
2
35Mmax
5 Serna et al. (2006) Cb = 2
M max + 9Ma2 + 16Mb2 + 9Mc2
12.5Mmax
6 AISC Specification Cb 3.0
2.5Mmax 3Ma 4Mb 3Mc
Mmax
7 British Standard BS 5950-1 Cb 2.273
0.2Mmax 0.15Ma 0.5Mb 0.15Mc
1.7Mmax
8 Australian Standard AS4100 Cb 2.5
M Mb2 Mc2
2
a
4Mmax
9 Proposed Equation Cb 2.5
2
M max 4Ma2 7Mb2 4Mc2
a
Equation 1, specified in the body text, is the general lateral-torsional buckling equation to which these Cb equations are applied.
b
Adopted by CAN/CSA-S16-01 and CAN/CSA-S6-06, but with an upper limit of 2.5.
c
Adopted by AASHTO with additional requirements.
CAN/CSA-S16-01—Limit States Design of Steel only (without any transverse loads), then it is silent on how
Structures (CSA, 2001) to account for non-uniform moment distributions that are not
captured by this additional clause.
This Canadian design standard specifies Equation 2 in Table
2 for determining the equivalent moment factor for unbraced
beam segments subjected to end moments, except that an up- CAN/CSA-S6-06—Canadian Highway Bridge Design
per limit of 2.5 is used instead of 2.3. It is not clearly stated Code (CSA, 2006)
in the standard whether or not this is intended to apply to This standard has adopted the same procedure as CAN/
beams that are also subjected to transverse loads within this CSA-S16-01; however, the commentary to the standard re-
length. Although the sign of L is assigned as described previ- fers users to the procedures of Clark and Hill (1960) as an
ously, no indication is given of how to account for triple cur- alternative approach.
vature (e.g., Moment Type 2 in Table 1). The standard also
specifies that for non-linear moment distributions where “the
ANSI/AISC 360-05—Specification for Structural Steel
bending moment at any point within the unbraced length is
Buildings (AISC, 2005a)
larger than the larger end moment” (CSA, 2001), the equiva-
lent moment factor be taken as 1.0. This additional clause ef- The American steel design specification for buildings stipu-
fectively requires the designer to ignore the beneficial effect lates that Equation 6 in Table 2 be used to determine the
of a non-uniform moment distribution under this common equivalent moment factor. As shown in the table, the only
circumstance. Moreover, if the standard is interpreted to differences between Equations 4 and 6 are the coefficients
mean that Equation 2 applies just for cases of end moments for the terms Mmax. This adjustment was made in an attempt

Table 3. Discrete Equivalent Moment Factors and Equations from the Literature a
Moment Clark & Hill Nethercot & Trahair Eurocode 3

Type (1960) Trahair (1976) (1993) (ECS, 1992)
L −0.5 Cb 1.31 L0 Cb 1.75 Equation 2 (Table 2) L −0.75 Cb 1.141
L0 Cb 1.77 L 1.0 Cb 2.56 L −0.5 Cb 1.323
L 0.5 Cb 2.33 L −0.25 Cb 1.563
L 1.0 Cb 2.56 L0 Cb 1.879
1
L 0.25 Cb 2.281
L 0.5 Cb 2.7
L 0.75 Cb 2.927
L 1.0 Cb 2.752
C0 Cb 1.13 C0 Cb 1.13 Numerical result: C0 Cb 1.13
C 1.0 Cb 1.30 C 1.0 Cb 2.58 C0 Cb 1.09 C 1.0 Cb 1.285
2b C 1.0 Cb 2.52 Equations: C 1.0 Cb 2.52
For 0 c C < 0.75, Cb 1.13 0.12C
For 0.75 c C c 1.0, Cb −2.38 4.8C
same as Type 2 for same as Type 2 for Numerical result: same as Type 2 for C 0
C0 C0 C0 Cb 1.09
3 Equations:
For 0 c C < 0.7, Cb 1.13 0.1C
For 0.7 c C c 1.0, Cb −1.25 3.5C
C0 Cb 1.35 C0 Cb 1.35 Numerical result: C0 Cb 1.365
C 1.0 Cb 1.70 C 1.0 Cb 1.70 C0 Cb 1.31 C 1.0 Cb 1.565
4
Equation:
For 0 c C c 1.0, Cb 1.35 0.36C
same as Type 4 for same as Type 4 for Numerical result: same as Type 4 for C 0
C0 C0 C0 Cb 1.31
5 Equations:
For 0 c C < 0.89, Cb 1.35 0.15C
For 0.89 cC c 1.0, Cb −1.2 3.0C
a L /2 same as a L /4 Cb 1.44 Numerical result: a L /2 same as Type
Type 4 for C 0 a L /2 Cb 1.35 a L/2 Cb 1.31 4 for C 0
6 Equation:
For 0 c a/L c 0.5, Cb 1.35
0.4(1−2a/L)2
a L /4 Cb 1.04 a L /4 Cb 1.04 Numerical result: a L/4 Cb 1.046
a L /2 same as a L /2 same as a L/2 Cb 1.31 a L/2 same as Type
7c Type 4 for C 0 Type 4 for C 0 Equation: 4 for C 0
For 0 c a/L c 0.5, Cb 1.0
0.35(2a/L)2
a
No Cb values or equations were published for moment Types 8 to12.
b
Strikethrough indicates error in original reference; refer to text for clarification.
c
For moment Type 7 only, Nethercot and Rockey (1972) propose Cb 1.0 (a/L)2.
to give better results for cases of fixed end supports (AISC, AASHTO—LRFD Bridge Design Specifications
2005b). For design purposes, this specification sets an upper (AASHTO, 2007)
limit to the equivalent moment factor of 3.0, which is the
Similar to the current CAN/CSA-S16-01 procedure, this
highest among all specifications discussed here. The com-
specification uses Equation 2 as the primary equivalent
mentary to the specification indicates that Equation 2 is also
moment factor equation and also specifies that the value
appropriate for cases where the moment distribution is linear
be taken as 1.0 whenever the larger end moment is not the
between brace points.
largest moment throughout the unbraced segment. However,

the main difference between the two specifications is that in ASSESSMENT APPROACH
order to avoid the non-conservative results obtained when
Equation 2 is used for certain non-uniform moment distribu- Equivalent moment factors for 12 diverse selections of bend-
tions, AASHTO (2007) introduces an equivalent linear mo- ing moment distribution types have been determined from
ment distribution. The larger end moment and the mid-span the various equations discussed in this paper. These solutions
moment are projected back linearly to obtain an imaginary are compared against each other, as well as with numerical
smaller end moment, and then the larger value of the actual results presented in the literature. Although White and Kim
and imaginary smaller end moment is used to determine L (2008) included the results of hundreds of physical test re-
in Equation 2. Figure 1 illustrates an example for which a sults collected from numerous sources in a comprehensive
more appropriate solution is obtained if the magnitude of the statistical study on the flexural resistance of beams, the ma-
mid-span moment is taken into consideration. The AASHTO jority of these do not fall within the scope of the current study
(2007) procedure requires the calculation of two different because the experiments either involved transverse loading
equivalent moment factors—one for each of the top and bot- applied above or below the shear center, were conducted on
tom flanges—if both flanges experience compression due to beams with mono-symmetric or hybrid cross-sections, were
reversing curvatures. influenced by interaction buckling, or resulted in inelastic
global or local buckling. Nonetheless, they included four
BS 5950-1—Structural Use of Steelwork in Building: experimental results from simply supported beams tested
Code of Practice for Design (BSI, 2000) with a mid-span concentrated load applied through the shear
center (i.e., moment Type 6, a/L 0.5) that failed by elastic
The British standard specifies Equation 7 in Table 2 for lateral-torsional buckling. Since there are so few suitable
determining the equivalent moment factor. Among all speci- test results available, and because this particular moment
fications discussed in this study, BS 5950-1 has the lowest distribution is associated with a relatively well-established
upper limit (2.273). equivalent moment factor, their inclusion would add little to
the discussion presented herein. Therefore, these four tests
Eurocode 3 EN-1993-1-1—Design of Steel Structures are excluded from the comparisons in this paper.
(ECS, 1992)
In Annex F of the European design code, tabulated discrete Representative Moment Distributions
equivalent moment factors are provided for moment Types 1,
2, 4 and 7 (see Table 1). These values are similar to those As shown in Table 1, the bending moment types considered
published by Clark and Hill (1960), as shown in Table 3. in this study have been selected to cover a broad variety of
As in the original publication, the value for moment Type 2 potential situations. Moreover, each moment type envelopes a
when C equals 1.0 appears to be in error, and it has been wide range of moment diagrams by varying either the magni-
corrected accordingly in Table 3. tude of the end moments or the concentrated load locations.
The value of L for Type 1 (linear) moment distributions
AS 4100—Australian Standard: Steel Structures reflects the ratio of the end moments and can therefore vary
(SAA, 1998) only from 1.0 to 1.0. The variable C, used for moment
Types 2 to 5, 8, 9, 11 and 12, was assigned to alter the mag-
The Australian design standard specifies Equation 8 in Table nitude of the end moments. When C is set to 0, it represents
2 for determining the equivalent moment factor. Similar to a pinned (in plane) boundary condition, whereas when it is
Equation 5, it employs a square root format. equal to 1.0, it represents a fixed boundary condition. The
Fig. 1. Simplified moment diagrams according to CAN/CSA-S16-01 and AASHTO.

factor C was considered over a range from negative to a large (Type 2) and the other half are from models subjected to
enough positive value so that the scenarios of single, double a mid-span point load (Type 4). The solutions selected are
and triple flexural curvature, combined with either a maxi- based on cross-sectional properties similar to the two dis-
mum moment at the end or away from the braces, were all tinct numerical models used in the analyses of Serna et al.
covered and examined. Moment diagrams with negative val- (2006).
ues of C or values greater than 1.0 provide useful insight for Suryoatmono and Ho (2002) published a suite of fi-
evaluating continuous span structures. The variable a, used nite difference solutions for a 10-m-long (32.8 ft) doubly-
for moment Types 6, 7 and 10, was assigned to provide a symmetric wide-flange section with several different mo-
means for altering the point load location along an unbraced ment types: Type 1 with L varying from 1.0 to 1.0; Types 2
beam segment. As the location of the concentrated load coin- and 3 with C varying from 0 to 2.0; and Type 6 with a L/2
cides with the maximum moment for these moment distribu- (same as Types 4 and 5 with C 0). A total of 38 data points
tion types, this variation helps to develop trends of solutions are used in the comparisons.
that depict the influence of a varying distance between the Serna et al. (2006) published an extensive set of equiva-
point of maximum moment and the brace points. lent moment factors based on numerical results for moment
Moment Types 2, 4 and 11 were selected, in part, because Types 1 to 5, with various end support conditions. Only data
they include moment distributions that correspond to triple associated with no end lateral rotational restraint and no
curvature. Due to the ambiguous instruction provided in warping restraint are used in the evaluation. These research-
standard CAN/CSA-S16-01 regarding the sign of L for seg- ers analyzed two models with different flexural slenderness
ments in triple curvature, solutions calculated by setting L to values to ensure that the effect of the flexural slenderness
both negative and positive are compared. on the equivalent moment factors was observed. Only the
Moment Types 8, 9 and 10 are transformations of Types 2, lower value of Cb from the two models is utilized for each
3 and 7, respectively, that simulate situations in which an ex- loading condition in this study. As such, a total of 67 data
tra brace is placed in the middle of the original unbraced seg- points are used.
ment. Moment Types 8 and 10 are included in the analysis to Other numerical results used in the comparisons were
create circumstances where the moment is relatively uniform published by Clark and Hill (1960), Nethercot and Trahair
over much of the length, but no moment within the unbraced (1976), and Trahair (1993). They are summarized in Table 3.
segment exceeds the larger end moment. As noted previously,
this situation is not covered explicitly in CAN/CSA-S16-01.
Unlike Types 8 and 10, the moment distributions of the left RESULTS AND IMPORTANT OBSERVATIONS
and right segments for moment Type 9 are different; thus, two
sets of equivalent moment factor results are evaluated. Comparisons Among Methods
The equivalent moment factor values determined by the
Numerical Simulation Data
methods discussed previously for all 12 moment distribu-
Numerical analysis results from six different research pro- tion types are graphically presented alongside available
grams are included in the comparisons as reference solu- numerical results in Figures 2 to 14. The purposes of these
tions. No numerical simulations were found in the literature comparisons are to identify deficiencies and strengths of the
for moment Types 8 to 12. various methods and to propose a method that optimizes the
A total of 1500 critical bending stresses calculated using trade-off between computational effort and accuracy over a
numerical analysis were tabulated by Austin et al., (1955). broad range of moment distribution types.
This extensive set of results was created by determining Due to the large quantity of data assembled, for clarity
the critical moment of wide-flange sections with 10 dif- of the graphs in Figures 2 through 14 not all results from
ferent levels of flexural slenderness. Two loading cases—a the various methods and equations could be included. There-
uniformly distributed load and a mid-span point load—and fore, methods that are deemed not to provide any particular
25 discrete levels of in-plane and out-of-plane end rotational insight are sometimes omitted. To further alleviate difficul-
restraint combinations were used. Moreover, three discrete ties in interpreting the graphs due to congestion of the data,
levels of load application (i.e., load applied at the top flange, all numerical results use filled symbols so as to distinguish
shear center, and bottom flange) were evaluated for each them from the open and unfilled symbols used for design
combination of the loading and boundary conditions. Be- specifications and other published equations. Where the so-
cause the effects of the height of load application and the lutions for Equations 6 through 9 are controlled by the pre-
out-of-plane rotational end restraint are not considered in the scribed upper limit in the relevant design specification, the
current study, only 10 out of the 1,500 solutions are used in curves above the limit are shown dashed to reflect the ac-
the comparisons. Half of these solutions are selected from curacy of the equations in the event that the limit should be
numerical models subjected to a uniformly distributed load modified or eliminated.

Figure 2 demonstrates that all methods provide satisfac- equation should be positive or negative for such a case, both
tory approximations to the numerical results for Type 1 (lin- positive and negative values were used to develop two differ-
ear) moment distributions for L values up to about 0.5. As ent sets of solutions for comparison. Nonetheless, both sets
expected, results calculated using the CAN/CSA-S16-01 fail to follow the trend of the numerical solutions. As illus-
equation closely match the numerical results over the full trated in Figure 3, all other methods produce reasonable ap-
range. Among all the quarter-point moment equations, the proximations to the numerical data, with several utilizing a
AISC equation gives the most conservative results within the maximum permissible value to prevent the use of very large
region 0.5 L 1.0, with differences up to 18% compared to values in design. One exception is that the value suggested
the numerical results. by Eurocode 3 for the fixed end moment case (C 1.0) is
For moment Type 2, the beam segment is under triple very low. Although it provides an excellent representation of
curvature bending when C is greater than 0. Since CAN/ general trends, the equation proposed by Serna et al. (2006)
CSA-S16-01 does not specify whether the sign of L in the Cb appears to be too aggressive for design purposes in the
Fig. 2. Cb Results for Moment Type 1.

region 1.0 C 1.8 because the values obtained exceed distribution type over the majority of the range of common
many of their own numerical results. C values. If the upper limits of the design equations are not
The CAN/CSA-S16-01 results also fail to follow the trend considered, AS 4100 and Serna et al. (2006) seem to ap-
of the numerical data for moment Type 3. Since there is no proximate the upper and lower bounds, respectively, of the
moment at the left end of the unbraced segment, the Cb equa- numerical data in the upper range of C, while the AISC equa-
tion in this standard always gives results equal to 1.75 unless tion is the most conservative method in the same region.
the opposite end moment is not the maximum moment in the As seen in Figure 5 for moment Type 4, results obtained
segment, in which case the value is 1.0. The abrupt transition from the Kirby and Nethercot (1979) and AISC formulae
between these two cases is at C 0.69. As shown in Fig- differ significantly from the numerical results on the non-
ure 4, non-conservative results exist where 0.69 C 0.85. conservative side in the region of 0.6 C 1.1 (restraint
Most other methods perform relatively well for this moment approaching a fixed end condition). Although the BS 5950-1
Fig. 5. Cb Results for Moment Type 4

equation is conservative in this region, it fails to capture taken positive for triple curvature. Otherwise, grossly con-
the relatively abrupt change observed in the numerical data servative values are obtained throughout.
trends at about C 1.0. Conversely, the AS 4100 and Serna et The Cb values obtained using CAN/CSA-S16-01 for mo-
al. (2006) equations, which both use the square root format ment Type 5 change abruptly from 1.0 to 1.75 when C 0.89.
in the Cb equation, produce accurate approximations for this Figure 6 shows that these results are dissimilar to all other
region and also capture the abrupt change in slope. The equa- methods. Other methods generally give reasonable results,
tion of Serna et al. (2006) is the more conservative of the two with the BS5950-1 equation being the most conservative.
in this region. Similar to moment Type 2, two sets of solu- Since there are no end moments in moment Type 6, CAN/
tions calculated using the CAN/CSA-S16-01 procedures are CSA-S16-01 sets Cb equal to 1.0, regardless of the location
plotted for this moment type. Significantly non-conservative of the point load. Figure 7 shows that this solution is highly
values are observed within the region 1.0 C 1.4 if L is conservative in all situations. Although numerical results are

unavailable over the entire range of a /L, the actual trend of quarter-point moment methods, BS5950-1 gives the most
the solution can be reasonably predicted. The correct Cb so- conservative results in this region.
lutions should decrease gradually from about 1.75 to 1.35 as Similar to moment Type 6, the CAN/CSA-S16-01 equa-
the point load moves from one end of the unbraced segment tion gives Cb values equal to 1.0 for the full range of a/L for
toward mid-span. It is believed that the Trahair (1993) equa- moment Type 7 because no end moment is present. In this
tion provides the closest approximation to the true solutions, case, it is apparent that the correct solutions should increase
although it is intended for use with this load case only. The gradually from 1.0 to about 1.35 as the point loads move
local maxima in the curves predicted by most quarter-point from the ends of the unbraced segment toward mid-span, as
moment methods at about a /L 0.35 appear unreasonable, obtained from the equations by Trahair (1993) and Nether-
and in the case of the AS 4100 equation the Cb values around cot and Rockey (1972) that were derived for this load case
the peak are likely significantly non-conservative. Of all the only. Figure 8 shows that all quarter-point moment methods

fail to reflect the effect of the non-uniform moment distribu- for the ranges of 1.0 C 0.85 and 0 C 0.75, respec-
tion when the point loads are between the end and the first tively, and are considered to be highly non-conservative over
quarter-point of the unbraced segment, although all methods most of these ranges. On the other hand, Figure 10 (Type 9,
appear to provide conservative solutions over the full range left unbraced segment) shows that it produces highly con-
with the exception of AS 4100. servative results as compared to other methods. Solutions by
No numerical results are available for moment Types 8 to other methods are, in general, consistent and appear to be
12. However, evaluation of the performance of the various reasonable approximations to the true solutions. The equa-
methods can be based on judgment and the knowledge ob- tion in BS5950-1 tends to be the most conservative of the
tained from the results observed for moment Types 1 to 7. quarter-point moment methods over much of the ranges, and
Figure 9 (Type 8) and Figure 11 (Type 9, right unbraced especially when the maximum value is invoked.
segment) show that the Cb results obtained by CAN/CSA- As illustrated in Figure 12, the CAN/CSA-S16-01 equa-
S16-01 are much higher than the results of other methods tion gives a solution of 1.75 for moment Type 10, regardless
Fig. 10. Cb Results for Moment Type 9, left unbraced segment.
Fig. 11. Cb Results for Moment Type 9, right unbraced segment.

of the point load location. However, solutions determined segment, as described previously, are in better agreement
by all other methods are much lower than 1.75 for the great with other methods. Again, BS5950-1 provides the most
majority of the range. It is evident in this case that the cor- conservative solutions among the quarter-point moment
rect solutions should increase gradually from 1.0 to about methods, although all such methods provide similar results
1.75 as the point loads move from the ends of the beam to- over the full range of a/L.
ward the braced mid-span. Although there are no numerical Solutions developed for moment Types 11 and 12 are il-
results to verify the correct solutions directly, it is clear that lustrated in Figure 13 and Figure 14, respectively. Findings
the solutions obtained using CAN/CSA-S16-01 are highly and observations are similar to those discussed previously
non-conservative. Conversely, the solutions of the AASHTO for moment Types 2 and 3.
procedure, which uses the same equation as CAN/CSA- Although a broad investigation is presented herein that in-
S16-01 but is based on an imaginary smaller end moment cludes many procedures from the literature and design speci-
determined using the moment at the center of the unbraced fications, it is instructive to clarify the deficiencies of the

CAN/CSA-S16-01 method for determining equivalent mo- distribution effect. Finally, CAN/CSA-S16-01 is ambiguous
ment factors. This design standard gives inconsistent results in some common design circumstances because it does not
for all moment types evaluated in this study except for the clearly state whether or not its provisions are applicable to
linear moment distribution (Type 1), for which the procedure an unbraced segment that is subjected to end moments in
was originally derived. Driver and Wong (2007) summarize combination with other loading, or whether the sign of L
the ranges where this method produces acceptable results for should be positive or negative for the case of triple curva-
each moment type presented here and conclude that the pro- ture. The latter ambiguity in some cases creates a situation
cedure is unsatisfactory over the full range in two out of the where the choice of sign results in either a highly conserva-
12 types (Types 6 and 10) and over a significant part of the tive or a highly non-conservative solution. Driver and Wong
range in nine others. In all, they identified four general types (2007) provide a more detailed discussion of the CAN/
of deficiencies in the CAN/CSA-S16-01 approach. First, its CSA-S16-01 provisions.
provisions tend to produce highly conservative results for
simply supported beams that are unbraced between their Important Observations Concerning Quarter-Point
ends because Cb always defaults to 1.0. Figure 7 indicates Moment Methods
that Cb can be underestimated by more than 40% for mo-
ment Type 6, for example. Significantly conservative results As shown in all 12 comparisons, the quarter-point moment
can also occur in transversely loaded unbraced segments that methods, which are purported to be applicable for any mo-
experience either zero moment at one end of the segment ment distribution, tend to give reasonable results for dif-
(Cb 1.75) or equal and opposite end moments (Cb 1.0). ferent moment types even though their levels of accuracy
Second, this method potentially overestimates Cb when a and conservatism vary. The coefficients for the four discrete
transversely loaded segment experiences a maximum mo- moments used in these equations are selected deliberately to
ment at either end. For moment Type 10 (Figure 12), for weight the influence of each quarter-point moment magni-
example, the overestimation can be as high as 75%. Third, tude relative to the maximum moment, and the coefficients
in 10 out of the 12 moment types discussed in this paper, selected are highly influential to the accuracy of the results.
Cb either remains unchanged over the entire range of a/L or A few common characteristics of these coefficients are
C or it experiences abrupt changes at particular C values, observed by examining Equations 4 to 8. For example, the
whereas for a gradually transforming moment distribution sum of all coefficients in the denominator is always equal
a gradually changing Cb function would appear more ap- to the coefficient in the numerator. This condition ensures
propriate. This suggests that the CAN/CSA-S16-01 provi- that Cb 1.0 for a uniform moment distribution (i.e., when
sions, although not always producing non-conservative re- Ma Mb Mc Mmax). Also, the coefficient of Ma is identical
sults, inconsistently accounts for the non-uniform moment to the coefficient of Mc to ensure that the Cb value is the same

for any two mirrored moment distributions about the center- of using the actual quarter-point values, Equations 4 to 8
line of the unbraced segment. The last common characteristic produce the result Cb 1.0, which may be highly conserva-
is that the coefficient of Mb is always at least equal to that of tive, depending on the actual locations of the concentrated
Ma and Mc. This represents the fact that beams with moment moments. One way to address this deficiency would be to in-
distributions where the point of maximum moment is close crease the number of moment parameters in the equation to
to the centerline of the unbraced segment (i.e., Mb ≈ Mmax) are better represent the actual moment distribution, but it would
more prone to lateral-torsional buckling than those where it also increase the complexity of the equation as well as the
is close to the quarter points. concomitant computational effort required for design. Due
All quarter-point moment methods may fail to provide to the relative rarity of these cases, this increase in complex-
conservative approximations of the actual equivalent mo- ity is likely unwarranted if designers are simply aware of
ment factor under the presence of abrupt changes in the mo- cases where the quarter-point moment equations should be
ment diagram, i.e., for segments loaded with concentrated applied with due caution.
moments. Arguably, this condition is rare in typical design Another concern with the quarter-point moment equations
problems, but it can occur, for example, when a vertical can- arises because the resulting equivalent moment factor is in-
tilever post affixed to the beam flange is loaded parallel to dependent of the sign of the internal moments. It is unclear
the beam axis. Figure 15 demonstrates one situation where how these equations can account for the effect of an abrupt
the accuracy of the quarter-point moment equations is ques- reversal of curvature in a beam, such as the one illustrated in
tionable because they fail to capture the uniformity of the Figure 16, Case 2. Although it is apparent that Case 2 load-
moment distribution between the quarter points. Using any ing should result in a more favorable equivalent moment fac-
of Equations 4 through 8 for the two different moment distri- tor than Case 1 due to the presence of double curvature, all
butions shown in this example results in the same Cb value, quarter-point moment equations incorrectly give the same
although one case is clearly more critical than the other. If a Cb value for both diagrams because the absolute values of
designer were simply to set Ma and Mc equal to Mmax instead the moment parameters fail to distinguish between the two
Fig. 15. Inaccuracy of quarter-point moment methods for case of abrupt change in moment.
Fig. 16. Inaccuracy of quarter-point moment methods for case of abrupt curvature reversal.

cases. As a result, the solution for Case 2 is highly conserva- ues of all methods considered. Similar drawbacks exist for
tive. Again, refining the equations to rectify this shortcoming Equation 5, although many of the cases where the numerical
is likely unnecessary because this scenario is also relatively data are exceeded are for the larger values of Cb that could
uncommon, but designers need to be aware of the limitations be eliminated by using an upper bound on the permissible
of the procedure. values. In light of the rather sparse set of corroborating nu-
merical and experimental data available, both equations are
PROPOSED EQUATION judged to be too aggressive for design use. Therefore, a mod-
ified quarter-point moment equation utilizing the superior
It is evident from the foregoing discussion that modifica-
square root format is proposed for design in order to provide
tions are required to improve the accuracy of the CAN/CSA-
accurate Cb values and properly represent the data trends,
S16-01 equivalent moment factor procedures. All other
while at the same time minimizing the chance of obtaining
methods considered in this investigation are more versatile;
non-conservative beam capacities. This equation takes the
however, these methods have their own drawbacks. The
following form:
method proposed by Trahair (1993) tends to provide very
good results, but it relies on several independent equations,
4M max
each having a limited scope of application, and therefore be- Cb = ≤ 2.5 (9)
2
comes somewhat cumbersome for general design purposes. M max + 4M a2 + 7Mb2 + 4M c2
The use of a table of individual Cb values for specific cases,
similar to those of Clark and Hill (1960), Nethercot and where the moment parameters are defined in the same way
Trahair (1976) and Eurocode 3, is considered undesirable as for the other quarter-point moment equations discussed
for design specifications due to the innumerable common herein. The upper limit of 2.5 is selected to prohibit the use
cases for which no guidance would be provided. Although of very high Cb values in design, although a different upper
the AASHTO procedure effectively eliminates many of the limit could be selected and justified based on reliability con-
non-conservative results obtained from Equation 2 by using siderations. Nevertheless, it must be noted that the proposed
an equivalent linear moment diagram, it still gives highly equation appears to produce good results even without this
conservative results for simply supported beams braced only limit. Therefore, if the limit were to be increased or removed,
at the ends and subjected to transverse loading. Despite their the better performance of the proposed equation, as com-
shortcomings for certain rare cases, as discussed in the prepared to the other quarter-point moment methods that use the
vious section, the quarter-point moment approach shows the square root format, becomes even more important.
most promise of wide applicability, combined with simplic- As shown in Figures 2 through 14, the proposed equivalent
ity, and for the most part these equations capture the trends moment factor equation (Equation 9) provides far better ap-
observed in the numerical data well in the cases considered. proximations to the numerical data (and to estimated correct
Their accuracy, however, depends largely on the coefficients solutions where no such data exist) than does the equation
of the moment terms. The equation in the British standard in CAN/CSA-S16-01, and it also addresses shortcomings of
tends to give very conservative results for several moment the other methods, while producing appropriately conserva-
types. The Kirby and Nethercot (1979) and AISC equations tive Cb values for design. For instance, it effectively avoids
are nearly equivalent and generally give good results. How- producing the non-conservative results obtained by the AISC
ever, they are unable to capture the trends of the numeri- equation for moment Type 4 in the range 0.6 C 1.1 (the
cal data for the common case of moment Type 4 and give AISC and proposed equation curves are plotted with a heavi-
non-conservative results (up to about 32%) in the region of er line weight to facilitate comparison). The proposed equa-
0.6 C 1.1. tion also gives very good results for loading that produces a
Equations 5 and 8 (Table 2), by Serna et al. (2006) and linear moment distribution between brace points (Type 1), as
specified in the Australian design standard, respectively, shown in Figure 2.
generally capture the Cb trends very well. Their ability to
provide better results for moment Type 4 than the Kirby SUMMARY AND CONCLUSIONS
and Nethercot (1979) and AISC equations is attributed to
the square root format that makes Equations 5 and 8 unique Numerous published methods and equations for determin-
among the quarter-point moment equations considered in ing equivalent moment factors used in evaluating the elastic
this investigation. However, both Equations 5 and 8 produce critical moment of laterally unsupported beams have been
Cb values in some situations that exceed the numerical data compared for a wide variety of moment distribution types.
significantly, thereby producing non-conservative results. The investigation revealed that the procedure used currently
Equation 8 gives results that exceed the numerical data for in the Canadian design standard produces unacceptable
moment Types 2, 4 and 6, and in cases where no numeri- results for the majority of the common bending moment
cal data are available, it often produces the highest Cb val- distributions considered. Not only does this method give

grossly conservative results for many common cases, it also Austin, W.J., Yegian, S. and Tung, T.P. (1955), “Lateral
frequently gives unconservative results. Large abrupt chang- Buckling of Elastically End-Restrained I-Beams,” Pro-
es in Cb values with only slight changes in the shape of the ceedings of the American Society of Civil Engineers,
moment diagram were observed in 6 out of the 12 moment Vol. 81, pp. 673-1–673-25.
distribution comparisons, which contributes to the overall BSI (2000), Structural Use of Steelwork in Building: Code
poor performance of the procedure. Moreover, it does not of Practice for Design, Rolled and Welded Sections,
give clear direction as to the sign of L when the beam is BS5950-1, British Standards Institution, London, United
under triple curvature. Kingdom.
The study also revealed drawbacks inherent in other meth-
Clark, J.W. and Hill, H.N. (1960), “Lateral Buckling of
ods. Overall, the quarter-point moment equations developed
Beams,” Journal of the Structural Division, American So-
for general moment distributions capture the trends of the
ciety of Civil Engineers, Vol. 86, No. ST7, pp. 175–196.
numerical data reasonably well. However, the evaluations
show that the Kirby and Nethercot (1979) and AISC (2005) CSA (2001), Limit States Design of Steel Structures, CAN/
equations produce non-conservative results in some situa- CSA-S16-01, Canadian Standards Association, Missis-
tions, while the BS 5950-1 (BSI, 2000) equation, although sauga, ON.
generally conservative, produces comparatively less accurate CSA (2006), Canadian Highway Bridge Design Code,
results. The Serna et al. (2006) and AS 4100 (SAA, 1998) CAN/CSA-S6-06, Canadian Standards Association, Mis-
equations capture the trends of the numerical data more con- sissauga, ON.
sistently by implementing a square root format in the quarter- Driver, R.G. and Wong, E. (2007), “Critical Evaluation of
point moment method. However, they produce results that the CSA-S16-01 Equivalent Moment Factor for Later-
exceed the numerical data in several cases, implying that ally Unsupported Beams,” 2007 Annual General Meeting
both equations are too aggressive for design purposes. and Conference, Canadian Society for Civil Engineering,
To capture the best features of the various methods inves- Paper GC-189.
tigated, yet improve the overall suitability for general design
ECS (1992), Eurocode 3: Design of Steel Structures, Gener-
purposes, a modified quarter-point moment equation using
al Rules and Rules for Building, EN-1993-1-1, European
the square root format (Equation 9) is proposed. Not only
Committee for Standardization, Brussels, Belgium.
does it simulate the trends of the numerical solutions closely,
but it also produces reasonable and conservative equivalent Galambos, T.V., Ed. (1998), Guide to Stability Design Crite-
moment factors, even in cases where other methods do not. ria for Metal Structures, Chapter 12 “Bracing,” Structural
Moreover, it is simple and well-suited to design applica- Stability Research Council, John Wiley and Sons, Inc.,
tions. Like all quarter-point moment methods, the proposed New York, NY.
equation does not produce good results in some situations Kirby, P.A. and Nethercot, D.A. (1979), Design for Struc-
where concentrated moments are applied. Nevertheless, it is tural Stability, Halsted Press, New York, NY.
believed to be appropriate for the vast majority of typical Nethercot, D.A. and Rockey, K.C. (1972), “A Unified
design cases. Approach to the Elastic Lateral Buckling of Beams,”
Engineering Journal, American Institute of Steel Con-
REFERENCES struction, July, pp. 96–107.
AASHTO (2007), LRFD Bridge Design Specifications, Nethercot, D.A. and Trahair, N.S. (1976), “Lateral Buckling
4th edition, American Association of State Highway and Approximations for Elastic Beams,” The Structural Engi-
Transportation Officials, Washington, DC. neer, Vol. 54, No. 6, pp. 197–204.
AISC (2005a), Specification for Structural Steel Buildings, SAA (1998), Australian Standard—Steel Structures, AS4100,
ANSI/AISC 360-05, American Institution of Steel Con- Standards Australia, Homebush, NSW, Australia.
struction, Chicago, IL. Salvadori, M.G. (1955), “Lateral Buckling of I-Beams,”
AISC (2005b), Commentary on the Specification for Struc- ASCE Transactions, American Society of Civil Engineers,
tural Steel Buildings, American Institution of Steel Con- Vol. 120, pp. 1165–1177.
struction, Chicago, IL. Serna, M.A., López, A., Puente, I. and Yong, D.J. (2006),
Austin, W.J. (1961), “Strength and Design of Metal Beam- “Equivalent Uniform Moment Factors for Lateral-
Columns,” Journal of the Structural Division, Ameri- Torsional Buckling of Steel Members,” Journal of Con-
can Society of Civil Engineers, Vol. 87, No. ST4, April, structional Steel Research, Vol. 62, pp. 566–580.
pp. 1–32.

Suryoatmono, B. and Ho, D. (2002), “The Moment–
Gradient Factor in Lateral–Torsional Buckling on Wide
Flange Steel Sections,” Journal of Constructional Steel
Research, Vol. 58, pp. 1247–1264.
Trahair, N.S. (1993), Flexural–Torsional Buckling of Struc-
tures, CRC Press, Boca Raton, FL.
White, D.W. and Kim, Y.D. (2008), “Unified Flexural Re-
sistance Equations for Stability Design of Steel I-section
Members: Moment Gradient Tests,” Journal of Struc-
tural Engineering, American Society of Civil Engineers,
Vol. 134, No. 9, pp. 1471–1486.
Zuraski, P.D. (1992), “The Significance and Application of
Cb in Beam Design,” Engineering Journal, American In-
stitute of Steel Construction, Vol. 39, No. 1, pp. 20–25.

Impact of Diaphragm Behavior
on the Seismic Design of Low-Rise
Steel Buildings
COLIN A. ROGERS and ROBERT TREMBLAY
ABSTRACT
Modern building codes allow engineers to use reduced seismic loads in design provided that the seismic load resisting system (SLRS) of the
structure is adequately designed and detailed to withstand strong ground shaking through ductile inelastic response. This approach has been
adopted by the North American model codes, which typically include special provisions to achieve satisfactory inelastic seismic performance.
Single-story buildings often incorporate a steel roof deck diaphragm that is relied on to transfer lateral loads to the vertical bracing bents. The
vertical braces are usually selected as the energy dissipating fuse element, while the diaphragm and other elements in the SLRS should be
designed such that their capacity exceeds the nominal resistance of the braces. Steel bracing members designed for compression inherently
possess significant reserve strength when loaded in tension, which means that large brace tension loads must be considered in the design of
the surrounding protected structural components. Capacity design seismic provisions have led to the need for much thicker roof deck panels
and more closely spaced diaphragm connection patterns compared with past practice in Canada. This paper describes the current U.S. seismic
design approach and provides examples as it is applied to single-story buildings and their diaphragms. An overview of the related aspects of an
ongoing research project on the flexibility and ductility of the roof diaphragm in low-rise steel buildings is also included.
Keywords: diaphragms, seismic performance, low-rise steel buildings.
INTRODUCTION Diaphragms may also contribute to the overall dynamic

properties and response of a building due to their in-plane
Single-story buildings often incorporate a steel roof deck di-
flexural and shear flexibility.
aphragm that is relied on to transfer lateral wind and seismic
North American model building codes (ASCE, 2005;
loads to the vertical bracing bents. Roof deck diaphragms
NRCC, 2005) and steel design specifications (AISC, 2005a,
in North America are commonly constructed of corrugated
2005b; CSA, 2005) allow engineers to use reduced seis-
cold-formed steel panels that are connected to the under-
mic loads in design provided that the seismic load resisting
lying structure and to one another at side-laps. Standing
system (SLRS) of the structure is adequately designed and
seam roofs (SSRs) also incorporate a form of steel deck,
detailed to withstand strong ground shaking through ductile
although it is not rigidly attached to the supporting structure;
inelastic response. Building codes and standards include
therefore, SSRs do not provide the necessary diaphragm ac-
special provisions to achieve satisfactory inelastic seismic
tion for the purposes of this discussion. Design of roof deck
performance for the various SLRSs used in steel building
diaphragms for in-plane shear forces can be carried out us-
construction (Tremblay, 2005). In particular, the design of
ing the SDI Diaphragm Design Manual (Luttrell, 2004). The
the vertical structural system must be carried out with strict
flexural capacity of the diaphragm can be developed through
compliance to capacity design principles, i.e., the fuse ele-
the use of continuous chord members (Figure 1a). Transfer
ments of the SLRS are sized and detailed to dissipate seis-
of the horizontal forces to the vertical bracing bents relies on
mic input energy through cyclic inelastic response, whereas
the action of the diaphragm collector elements (Figure 1a).
the remaining elements should be provided with sufficient
capacity to carry the maximum forces that are anticipated
along the lateral load path.
The vertical braces of steel buildings are usually selected
as the energy dissipating fuse element in the seismic load-
Colin A. Rogers, Associate Professor, Department of Civil Engineering and resisting system, while the other elements in the SLRS
Applied Mechanics, McGill University, Montreal QC, H3A 2K6, Canada (cor- are designed to have a capacity that is equal to or exceeds
responding author). E-mail: [email protected]
the expected strength of the braces. Figure 1b depicts the
Robert Tremblay, Professor, Group for Research in Structural Engineering, hierarchy of inelastic behavior in the elements located in
École Polytechnique, Montreal, QC, H3C 3A7, Canada. E-mail: robert.trem-
[email protected] the SLRS. When tension-compression bracing is used the
steel bracing members designed for compression inherently

possess significant reserve strength when loaded in tension, ture is located in Los Angeles, CA. In addition, the paper
which means that large brace tension loads must be consid- includes the interim findings of a study currently under way
ered in the design of the surrounding protected structural for which the objective is to develop seismic design strate-
components. In Canada the SLRS of single-story buildings gies that account for the flexibility and ductility of the roof
includes the roof diaphragm as well as the other compo- diaphragm in low-rise steel buildings. The scope of research
nents in the vertical structural system. This design objec- includes quasi-static diaphragm shear tests (Tremblay et al.,
tive is clearly stated in the 2005 National Building Code of 2004; Essa et al., 2003), large-scale dynamic diaphragm
Canada (NBCC) (NRCC, 2005): “Diaphragms and their tests (in progress), ambient vibration building measurements
connections shall be designed so as not to yield.” (Article (Paultre et al., 2004; Lamarche, 2005; Tremblay et al. 2008a,
4.1.8.15.1) and in CSA-S16 seismic provisions: “In capacity 2008b), as well as dynamic analyses of representative build-
design … diaphragms and collector elements are capable of ings (in progress). At project end the aim is to make design
transmitting the loads developed at each level to the vertical recommendations on the following aspects: diaphragm stiff-
seismic-force-resisting system” (Clause 27.1.2). These seis- ness under seismic loading, period of vibration for the build-
mic provisions have led to the need for much thicker roof ing, seismic response modification factors, ductile detailing
deck panels and more closely spaced diaphragm connection requirements and inelastic performance levels.
patterns compared with past practice, especially in areas of
high seismicity. Complying with these newly introduced de- SEISMIC DESIGN OF LOW-RISE BUILDINGS
sign requirements has impacted significantly on the cost of ACCORDING TO U.S. PROVISIONS
steel building structures in Canada, making this system less
attractive economically than in past years (Tremblay and Seismic Design Provisions
Rogers, 2005). In contrast, no specific guidance is given by
ASCE 7-05 provides the minimum seismic design loads for
AISC (2005a, 2006) to prevent yielding or failure of roof
building structures in the U.S. Except for buildings with
diaphragms or beams acting as collectors or chords, and the
horizontal torsional irregularity, the equivalent lateral force
designer must refer to ASCE 7 (2005) for the design forces.
procedure can be used for single-story steel buildings. This
A capacity design requirement for the diaphragm to meet
procedure comprises the application of an equivalent lateral
the expected yield strength of the braces in an ordinary or
seismic force that varies as function of the seismicity at the
special concentrically braced frame (OCBF or SCBF) with
site, the soil type, the period of the buildings and the type of
R > 3, for example, does not exist.
seismic load resisting system. The minimum lateral load, or
This paper contains a description of the U.S. seismic de-
seismic base shear, V, is given by:
sign provisions for low-rise steel buildings, as well as a de-
sign example of a single-story building located in Boston, V CsW (1)
MA. The design is also performed assuming that the struc-
(a) (b)
Bracing
Roof Members Anchor
Diaphragm (Inelastic) Rods
V V
Chord (typ.)
Steel Deck Collector Bracing Foundations

Units (typ.) Elements Connections
Roof
V Diaphragm Bracing Anchor
(Inelastic) Members Rods
V V
Vertical
X Bracing
(typ.)
Collector Collector Bracing Foundations
(typ.) Elements Connections
Fig. 1. Single-story buildings with capacity-based design concepts for SLRS.

where SCBFs, which allow this framing system to qualify for an R
S DS factor of 6.0. An OCBF may be designed with less restrictive
Cs
(R I ) provisions; however, the seismic loads must be computed
with R 3.25. The third system, for which special ductil-
S D1
for T TL ity detailing requirements need not be considered, must be
T R I( ) designed with R 3.0 according to the AISC specification
S D1 (AISC, 2005b). The importance factor varies from 1.0 to 1.5,
for T TL
T 2
(R I ) depending upon the occupancy category.
Also of key importance in seismic design is the Seismic
0.044 S DS I Design Category of the building. This parameter depends on
0.01 the occupancy category and the spectral acceleration values
0.5S1 at the site. Seismic Design Categories A and B typically
if S 1 0.6 g apply to buildings that are located in low seismic areas or
(
R I ) represent a low hazard to human life in the event of failure.
Braced steel frames designed with an R 3.0 are only per-
In these equations, W is the effective seismic weight, SDS and
mitted for seismic design categories A through C. Structures
SD1 are, respectively, the short-period and the one-second
built in moderately or highly active seismic regions and/or
design spectral acceleration parameters, R is the response
buildings that represent a high hazard to human life or that
modification factor, I is the importance factor, T is the fun-
are used for essential facilities generally are assigned to the
damental period of the building, and TL is the long-period
more severe Seismic Design Category D. Where S1 > 0.75,
transition period at the site. For a single-story steel building,
the structure is assigned to Seismic Design Category E, ex-
W includes the roof dead load, half the weight of the exterior
cept that essential facilities are assigned to Seismic Design
walls, and 20% of the roof snow load when the snow load
Category F. The OCBF system is limited to a height of 35 ft
exceeds 30 psf. The spectral acceleration parameters are
for Seismic Design Category D or E and is not permitted for
obtained from:
Seismic Design Category F. The height limits for SCBFs—
S DS = 2
Fa Ss ; SD1 = 2
Fv S1 (2) 160 ft for Seismic Design Category D or E and 100 ft for
3 3
Seismic Design Category F—typically would not apply
where Fa and Fv are, respectively, the short-period and the to most single-story building applications. In ASCE 7, the
long-period site coefficients that depend on the site class, seismic loads must be amplified by a redundancy factor,
and SS and S1 are, respectively, the mapped spectral accelera- R 1.3, for Seismic Design Category D, E or F. For braced
tions at short-period and one second corresponding to the steel frames, however, this factor can be taken equal to 1.0
maximum credible earthquake (MCE) level. Values of Fa and if removal of one brace does not result in more than a 33%
Fv are specified in ASCE 7 for different site classes, while reduction in lateral strength nor result in an extreme tor-
values of SS and S1 and TL can be obtained from maps that sional irregularity condition. The redundancy factor can also
are also included in ASCE 7. The period T for concentrically be ignored for rectangular buildings that are regular in plan
braced steel frames can be taken equal to the approximate provided that at least two bracing bays are constructed on
period Ta 0.02h0.75, where h (ft) is the height of the build- each of the perimeter walls.
ing. Alternatively, the period obtained from dynamic analy- In view of their higher R factor, SCBFs are expected to
sis may be used, although the so-computed period cannot develop significant inelastic response under the design earth-
exceed the limit given by T CuTa, where Cu is a coefficient quake. The aim of the AISC seismic provisions is to limit, for
that varies from 1.4 for high seismic zones to 1.7 for low the most part, the inelastic demand to the bracing members
seismic zones. The R factor depends on the type of lateral so that the integrity of the gravity supporting system formed
framing system. Single-story steel buildings typically rely by the beams and columns will remain intact during a strong
on steel bracing for lateral resistance. Three categories of earthquake. Tension/compression bracing must be used for
concentrically braced steel frames are described in ASCE 7: SCBFs. In addition, limits are imposed on the brace overall
special concentrically braced steel frames (SCBFs); ordinary slenderness and width-to-thickness ratios to ensure ductile
concentrically braced frames (OCBFs); and braced frames brace response and minimum energy dissipation without
not specifically designed for seismic resistance. The main premature fracture under inelastic reversed cyclic loading.
difference between the three systems is their expected in- Brace connections must be designed to resist loads corre-
elastic deformation capacity under seismic ground motions. sponding to the expected brace axial strength in tension,
The AISC seismic design provisions (AISC, 2005a) provide Texp ARyFy, and compression, Pexp 1.1Ry Pn, where A is the
detailing rules to ensure ductile inelastic response for the first brace cross-section, Ry is the ratio of the expected yield stress
two systems. More stringent requirements are prescribed for to the nominal yield stress, Fy, and Pn is the nominal brace

compressive strength. When the stress ratio, Pu /GcPn, under lation, the deformations from elastic analysis, Exe, must be
seismic load combinations exceeds 0.40, columns in bracing multiplied by Cd /I to obtain the design story drift reflecting
bents must be designed for the axial load obtained using the inelastic response, $. The factor Cd is, respectively, equal to
seismic load combinations, including system overstrength. 5.0 and 3.25 for SCBFs and OCBFs. For braced steel frames
Seismic load effects are amplified by the overstrength factor, designed without ductile detailing, Cd 3.0. When checking
7 0, to approximate the maximum seismic induced force the drift limits, it is noted that ASCE 7 allows the use of deflec-
columns will experience during strong ground shaking. In tions due to seismic loads based on the building fundamen-
ASCE 7, 7 0 2.0 for braced steel frames. These amplified tal period obtained from dynamic analysis, without applying
seismic loads need not exceed the forces arising from 1.1Ry the upper limit CuTa.
times the nominal strengths of the connected braces nor the
forces producing uplift of the foundation. Building Design Example (Boston)
OCBFs are expected to undergo limited inelastic defor-
mations under a design earthquake; thus fewer, less stringent The simple rectangular building located in Boston, MA,
ductility requirements apply. Both tension/compression and shown in Figure 2a is used to illustrate the seismic design
tension/only bracing designs are permitted in this category. provisions for single-story buildings with lateral seismic
The braces must still meet limits on their slenderness and loads resisted by steel braced frames acting together with
width-to-thickness ratios. Brace connections must have a a metal roof deck diaphragm. The roof structure is made of
tensile strength equal to or greater than the expected brace open web steel joists supported on steel trusses spanning
yield tensile strength, Texp ARy Fy, but need not exceed the across the entire width of the building. Single-bay X-bracing
load combination effects based upon the amplified seis- is used on each of the four exterior walls. Only the design of
mic loads. the seismic load resisting system in the direction parallel to
No specific guidance is given in AISC to prevent yield- the short walls is considered in this example. In addition, the
ing or failure of roof diaphragms or beams acting as col- calculations are performed assuming that an SCBF system
lectors or chords, and the designer must refer to ASCE 7 with R 6.0 and Cd 5.0 is adopted for the bracing bents.
for the design forces for these components. For single-story The main differences between this and an OCBF design are
structures, roof diaphragms are designed for the lateral force discussed at the end of the example.
V, but this force must not be less than 0.2SDS IW and need The dead load of the roof and walls are given in the figure
not exceed 0.4SDS IW. The redundancy factor, R, must be the together with the roof snow load. The seismic weight, W, is
same as that used for the vertical bracing bents. Collector equal to 593 kips. An Occupancy Category II is assumed for
beams must be capable of transferring the forces used in the the building, the site class is D, and the importance factor
design of the diaphragm to the supporting framework. For I 1.0. The seismic data SS, S1 and TL for the chosen location
Seismic Design Category C, D, E or F, the collector elements are given in Figure 2a. For this site, Fa 1.56 and Fv 2.4,
must resist the load combinations including seismic loads which gives design spectral accelerations SDS 0.31g and
amplified for overstrength. No specific requirement is given SD1 0.11g. The building is 22 feet tall and the period
for beams acting as diaphragm chords and it is assumed that Ta 0.02(22)0.75 0.20 s. For this site (SD1 0.11g), the fac-
forces consistent with diaphragm design can be used. tor Cu 1.6 and the amplified period CuTa 0.32 s. This
In the analysis of a building’s structure, the minimum ac- period estimate is used for the design and, hence, will need
cidental eccentricity corresponding to 5% of the dimension to be checked at the end of the design process. Using these
perpendicular to the loading direction must be considered parameters in Equation 1, it is found that Cs 0.052 and
if the diaphragm is anticipated to act as a rigid element. V 30.8 kips. Following ASCE 7 procedures, it is deter-
Resistance to the induced in-plane torsional moments can mined that the building can be assigned to Seismic Design
be assumed to be provided by all bracing bents if the roof Category B. Therefore, the redundancy factor, R, can be tak-
diaphragm has sufficient in-plane stiffness to efficiently dis- en equal to 1.0.
tribute the loads to the vertical system. If the diaphragm is The bracing bents are designed first. At this point, one
flexible, the load becomes essentially resisted only by the cannot determine whether the roof diaphragm will be classi-
bracing bents acting in the direction parallel to the applied fied as flexible; as such it is conservatively assumed that the
load. Single-story buildings are said to have a flexible dia- diaphragm is rigid and that in-plane torsional effects must
phragm when the maximum in-plane deformation of the be accounted for. The structure is symmetric and acciden-
roof diaphragm is more than twice the average of the build- tal torsion is included by moving the center of mass (CM)
ing deflections computed along the two end walls parallel away from the center of rigidity (CR) by 5% of the length
to the direction under consideration. ASCE 7 requires that of the building (10 ft), as prescribed in ASCE 7 and illus-
in-plane deformations of the roof diaphragm be included in trated in Figure 2b. Assuming that all four bracing bents are
the determination of the building story drift. In this calcu- of equal stiffness, the load on the bracing bent on gridline A

is equal to 54% V 16.6 kips. This load is resisted equally The roof diaphragm is formed of 1'-in.-deep, wide-rib
by the tension and compression acting braces (Figure 2c). (WR) Canam P3606 steel deck sheets having a trapezoidal
Once the braces and columns are designed, analysis of the cross-section. The sheets are 36 in. wide, with flutes spaced
braced frame will be performed to assess gravity load ef- 6 in. on center, as illustrated in Figure 3a. The sheets are 25 ft
fects on the braces. At this step, a first trial is made using the long and span over four equal spans between the roof joists
compression force of 11.0 kips; square tubing HSS 3×3×m (see Figure 2a). No. 10 self-tapping screws are used for the
conforming to ASTM A500 Grade C (Fy 50 ksi) is se- side-lap connections whereas Hilti X-ENP-19 L15 pins are
lected for the braces. The factored resistance of these braces chosen to connect the steel deck to the supporting structure.
GPn 13.9 kips, assuming a brace effective length KL (0.5) The diaphragm design is performed according to the SDI
(400 in.) 200 in. The braces have a design thickness method (Luttrell, 2004) together with the 2004 supplement
td 0.174 in., a cross sectional area A 1.89 in.2, and to the 2001 specification for the design of cold-formed steel
they meet the AISC limits for overall slenderness (KL/r members (AISI, 2004). The resistance factor associated with
175 < 200) and width-to-thickness ratio (b/t 14.2 < 15.4). fastener failure modes, Gd, is equal to 0.65. In the direction
For steel tubing, Ry 1.4 in the AISC seismic provisions studied the maximum shear flow in the diaphragm arises
and the expected brace capacities can be determined: along gridline A with Su 16.6 kips/100 ft 0.166 kip/ft. A
Texp 132 kips and Pexp 23.8 kips. light diaphragm design consisting of 0.0295-in.-thick steel
W-shapes made of ASTM A992 steel (Fy 50 ksi) are (22 ga.) panels is found adequate with two side-lap screws
used for the columns. Axial compression due to gravity roof per joist span and pins installed on a 36/3 pattern (18 in. on
dead and snow loads are, respectively, equal to PD 5.25 kips center). Such a diaphragm has a factored shear resistance
and PS 8.75 kips. Since the brace KL/r 175 exceeds Gd Sn 0.281 kip/ft and a shear stiffness, G b, of 12.9 kip/in.
4(E/Fy)0.5 96 (taking E 29,000 ksi), the AISC Seismic The edge beams along the 200-ft-long walls act as the
Provisions require that the columns be designed to carry chord members resisting the axial loads induced by the
gravity load effects plus the brace force Texp transferred to the in-plane bending moment, which is produced by the seis-
column; PE 132[sin(41.3o)] 87.1 kips. A W8×40 shape mic load of 30.8 kips assumed to be uniformly distributed
is found adequate to withstand the various load combina- over the length L 200 ft, wE 0.154 kip/ft. The maxi-
tion effects and meet the minimum width-to-thickness lim- mum axial loads develop at the diaphragm mid-span,
its prescribed for columns in the AISC Seismic Provisions. Pu (±0.154)(2002/8)/100 ±7.7 kips. This situation is il-
The analysis of the braced frame under combined gravity lustrated in Figure 3b; W8×10 beams with a cross-sectional
and seismic loads shows that the brace compression force area, A 2.96 in.2, are selected. It is noted that seismic loads
is increased to 11.8 kips, still lower than the brace factored acting parallel to the long walls also induce axial loads in the
resistance (13.9 kips). The elastic deflection, EB, of the brac- same beams and the worst scenario must be considered for the
ing bent computed under half the seismic load (15.4 kips), beam design. As illustrated in Figure 3c, when considering
is 0.11 in.
a) b) 0.02 V
1.5'' steel deck X-Bracing
(sheets 25'-0" long) (typ.) Joists @ 75''o/c K
V = 30.8
0.54 V
5
= 16.6 K 0.46 V
CM CR
4 @ 25'-0" = 100'-0"
0.02 V
10' 100'
c)
K
-1
1. K
0 K
.0
1
11
Truss (typ.) 22'
10 @ 20'-0" = 200'-0"
A K 41.3O
Roof dead load = 21 psf Site Class D
Weight of walls = 5 psf Ss = 0.30 g ; S1 = 0.07g 25'
Roof snow load = 35 psf TL = 6 s
Fig. 2. (a) Plan view of the building studied, (b) in-plane torsion effects, and (c) bracing bent studied.

accidental eccentricity, a maximum force of 15.7 kips is deflection to the bracing bent deflection (Figure 4a) is equal
transferred through the roof diaphragm to the bracing bent lo- to 0.69/0.11 6.3, which is much greater than 2.0, indicating
cated along the long walls. The edge beams then act as collector that a flexible roof diaphragm could have been considered
elements transferring the shear flow from the diaphragm to in design. The design should then be redone assuming that
the bracing bent (Figure 3d). In this particular case, the max- the diaphragm acts as a simply supported beam spanning
imum compression axial load in the beams reaches 7.9 kips, between the bracing bents parallel to the load, thus neglect-
which is more critical than the load induced when the same ing the contribution of the bracing bents perpendicular to the
beams act as chord elements. It is noted that maximum load in the resistance to the in-plane torsion due to the acci-
forces in collector beams along braced column lines will be dental eccentricity of the seismic load. For regular rectangu-
minimized if the bracing bents are located half-way along lar buildings such as the one studied herein, a seismic force
these grid lines. The designer must also provide proper con- equal to 0.55V would then need to be considered along each
nections between the steel deck and the perimeter members of the perimeter bracing bents. This is slightly larger than
to allow the transfer of the shear flow from the diaphragm to the values obtained assuming in-plane torsional resistance
the perimeter beams. In addition, attention must be paid to provided by four equally stiff bracing bents on the perimeter:
ensure the transfer of the computed beam axial loads through 0.54V and 0.51V, in the short and long directions, respec-
the beam-column joints. Once the diaphragm is designed and tively. For simplicity, however, the design obtained herein is
the chord members are selected, in-plane elastic deforma- kept unchanged in the example.
tions of the diaphragm due to flexure, EF, and shear, EW, can The design loads for earthquakes as calculated using
be determined. For this simple case, these two values can be ASCE 7 are based largely on the fundamental period of vi-
calculated using: bration of the vertical structure. It has, however, been shown
through analytical means that the period of vibration of a
5 wE L4 wE L2 (3) single-story building with a flexible roof diaphragm may be
D F W
384 EI d 8 Gb longer than that predicted based on the stiffness of the verti-
cal SLRS (Tremblay and Stiemer, 1996; Medhekar, 1997;
In this expression, Id is the moment of inertia of the dia- Tremblay et al., 2000). In the determination of the equivalent-
phragm in the direction considered (Id 2.13 t 106 in.4). static lateral loads for single story-buildings ASCE 41 (2006)
The deflections EF and EW are, respectively, equal to 0.09 in. allows for the introduction of the flexibility of the roof dia-
and 0.60 in., giving ED 0.69 in. The ratio of the diaphragm phragm to estimate the fundamental period of vibration.
a) c)
Deck Sidelap
Joist Sheet
(typ.) Fastener
(typ.) CR
CM K
Frame V = 30.8
Fastener
(typ.)
15.7K / 200' = 0.0785 kip/ft

PLAN
b) d) 6.3 K
7.7K
-1.6 K
0.0785 kip/ft -7.9 K
- 7.7 K
22'
20’
30.8 K / 200' = 0.154 kip/ft
(typ.)
PLAN ELEVATION (LONG WALL)
Fig. 3. (a) Steel deck panels, (b) axial loads in beams acting as diaphragm chord members,
(c) force transfer from the diaphragm to the edge beam under seismic loads in the long direction, and
(d) edge beams acting as collector elements under seismic loads acting in the long direction.

Similarly, an expression for the period was proposed by Equation 1 with this longer period, as permitted in ASCE
Medhekar (1997) and validated by shake table testing by 7-05. Figure 4b illustrates the significant reduction in design
Tremblay and Bérair (1999) and Tremblay et al. (2000). A loads with V(T 1.12 s) 0.0167W 9.9 kips, which is 32%
longer building period can often provide for much lower of the seismic load determined using the period T CuTa.
seismic design forces based on the uniform hazard spectrum Under this reduced lateral load, the elastic deflections EB and
that is now required for design. Tremblay and Rogers (2005) ED are respectively equal to 0.035 in. and 0.22 in. The de-
illustrated that the use of this extended period of vibration sign story drift can then be determined as $ (5.0)(0.035 +
can lead to significant savings in terms of the cost of the 0.22)/1.0 1.28 in., giving an inter-story drift of 0.49%,
lateral load carrying system, mainly because of the lower which is less than the limit of 2% typically applicable to this
design loads. Studies by Naman and Goodno (1986), Dubina type of building. One can check that for this structure P-delta
et al. (1997), Tena-Colunga and Abrams (1996), Tremblay effects are small and can be considered negligible.
and Stiemer (1996) and Tremblay et al. (2002), among oth- In the structure as designed, the strength of the diaphragm
ers, also showed that the calculated seismic forces can be re- is not related to the actual capacity of the vertical bracing
duced by incorporating the diaphragm flexibility. Using the system and there is no guarantee that the system will behave
elastic deflections of the bracing bents and the diaphragm, EB as intended under the design earthquake, i.e. with inelastic
and ED in Figure 4a, one can estimate the fundamental period response developing in the bracing members that have been
of the structure with (ASCE 41, 2006): specially sized and detailed to undergo significant inelastic
response without fracture. For instance, Figure 4c shows the
T (W V )( 0.10 B 0.078 D ) (4) diaphragm shear flow along the perimeter beams on the short
exterior walls on grid lines A and F that was considered for
where EB and ED are in inches. When compared to the original the design of the roof diaphragm. The corresponding axial
equation found in ASCE 41, the ratio W/V has been incorpo- loads in the edge beams acting as collectors are also given
rated in Equation 4 because the expression requires the use in the figure (maximum 8.30 kips). In Figure 4d, the same
of deflections due to a horizontal load equal to the effective shear flow and beam axial loads are given when the braces
seismic weight, W. For the design presented herein, the com- reach their expected axial compression and tension strength
puted fundamental period from Equation 4 is 1.12 s, which Pexp and Texp as determined earlier. The second set of forces is
is much longer than the value assumed in design (0.32 s). much greater, more than 7.0 times the forces used in design.
For drift calculations, one uses the load obtained from However, considering that an R factor of 6.0 was used in the
a) δD c) Beam Axial Load 4.15K

δB
- 4.15 K
16.6K/100' = 0.166 kip/ft - 8.30 K
22'
25' (typ.)
29.3 K
b) T = CuTa
d) Beam Axial Load
0.06
K - 29.3 K
117K/100' = 1.17 kip/ft - 58.5
0.04
Computed T
V/W
23.8K
132 K 22'
0.02
0.00 25' (typ.)

0.0 0.5 1.0 1.5 2.0
Period, T (s) ELEVATION (END WALL)
Fig. 4. (a) Bracing bent and roof diaphragm deformations, (b) variation of the seismic load with the period,
(c) axial loads in edge beams acting as collector elements along the short walls under V = 30.8 kips, and
(d) axial loads in edge beams acting as collector elements along the short walls upon brace yielding.

Table 1. Summary of the Design Parameters for the Building Examples
Boston area Los Angeles area

Ss 0.30, S1 0.07, TL 6.0 s Ss 1.70, S1 0.60, TL 12.0 s
Parameters
SDC B SDC D
W 593 kips W 453 kips
SLRS SCBF OCBF SCBF OCBF
R 6.0 3.25 6.0 3.25
CuTa (s) 0.32 0.32 0.28 0.28
V (kips) 30.8 56.9 85.6 158
R 1.0 1.0 1.0 1.3
Brace design T/C T/O T/C T/O
Brace section HSS 3×3×m L2½×2×X HSS 4×4×) 2L3×3×k
Texp (kips) 132 71.3 236 262
Pexp (kips) 23.8 — 75.1 37.5
Su (kip/ft) 0.166 0.307 0.462 1.11
Deck sheets 22 ga. 22 ga. 22 ga. 16 ga.
Frame fasteners 36/3 36/4 36/4 36/4
Screws/joist spacing 2 3 2 11
G b (kips/in.) 12.4 (F) 22.8 (R) 25.1 (F) 102 (R)
T from Eq. 4 (s) 1.12 1.11 0.71 0.56
E/hs (%) 0.49 0.53 1.67 1.43
Note: (F) flexible diaphragm, (R) rigid diaphragm
calculation of the design seismic load V, it is unlikely that along the side-lap with eight screws per joist spacing and
such a high force demand will develop during the design pins arranged in a 36/5 pattern.
earthquake. The resistance of the foundation to overturning The example building could have been designed using
uplift can also limit the forces delivered to the bracing bent. tension-only (T/O) bracing of the OCBF category. The main
AISC (2006) provides an example illustrating how founda- differences between the SCBF and OCBF design solutions
tion uplift can be included in this calculation. Nevertheless, are summarized in Table 1. An R factor of 3.25 would have
it is highly probable that forces in excess of the capacity of been considered for the OCBF design, leading to a base shear
the perimeter beams and the diaphragm as designed will force V 56.9 kips. For this system, L2×2'×X single-
be reached in future earthquakes, which may cause severe angle braces made of ASTM A36 steel (A 1.32 in.2,
damage and, possibly, failure of the diaphragm structure and Fy 36 ksi) would have represented an acceptable solu-
collapse of the roof gravity system that it laterally supports. tion [GTn 42.8 kips > 40.9 kips (0.54)(56.9/cos(41.3o)].
Caution should therefore be exercised by designers in the The required shear resistance for the diaphragm, without
selection of the diaphragm and its chords and collectors to consideration of capacity design, would have been equal to
ensure that proper response will be achieved. 0.307 kip/ft [(0.54)(56.9)/100]. Deck panels 0.0295-in.-
As mentioned, the 2005 NBCC in Canada states that dia- thick (22 ga.) are found adequate with three side-lap screws
phragms must be designed not to yield. They must therefore per joist span and pins installed on a 36/4 pattern (12 in. on
be provided with sufficient strength to match the expected center). Such a diaphragm has a factored shear resistance
strength (actual capacity) of the vertical system. The design GdSn 0.315 kip/ft and a shear stiffness, Gb, of 22.8 kip/in.
forces need not exceed, however, the forces corresponding For angles, Ry 1.5 in AISC Seismic Provisions and the ex-
to elastic response, i.e., forces determined with the seismic pected brace tensile strength Texp 71.3 kips. Had capacity
response modification coefficients equal to 1.0. Had this design principles been adopted, the design force for the dia-
concept been applied to the example building, the required phragm would have been equal to 0.535 kip/ft, as governed
shear strength for the roof diaphragm would have been by tension yielding of the braces. This value is nearly half
Su (6.0)(16.6 kips)/100 ft 1.0 kip/ft. In this case, the that required for the more ductile SCBF system. This differ-
diaphragm design would call for a much stronger configura- ence is attributed to the fact that tension/compression bracing
tion such as 0.0474-in.-thick (18 ga.) deck panels connected is required for SCBFs. The size of the braces is governed by

the compression capacity requirement; since the braces are considered in capacity design (Pexp 37.5 kips with KL/r 221).
generally long and slender due to the height of single-story The expected shear flow when the braces reach their tensile
structures they possess significant overstrength resulting and compressive strengths is therefore equal to 2.25 kips/ft,
from the large difference between tension and compression approximately twice the design value according to the
capacities. When applying capacity design principles, this current seismic provisions. Although less pronounced than
large overstrength impacts on the forces that will be deliv- the building designs for the Boston area, both the SCBF and
ered to the components in the SLRS, which need to remain OCBF systems in Los Angeles have computed fundamental
essentially elastic. periods longer than the values assumed in design.
Building Design Example (Los Angeles) EFFECT OF DIAPHRAGM FLEXIBILITY ON

BUILDING PERIOD OF VIBRATION
The example building is designed for the Los Angeles area
assuming the same site class (D), Occupancy Category (II) In theory, accounting for the in-plane flexibility of the roof
and importance factor (I 1.0). Table 1 gives the key design diaphragm may lead to more economical design solutions
parameters for the SCBF and OCBF designs. At this site, for single-story steel buildings, as was illustrated in the de-
Fa 1.0 and Fv 1.5, which result in design spectral accel- sign example. However, recent ambient vibration studies on
erations SDS 1.13 g and SD1 0.60 g. In view of these higher buildings of this type at the University of British Columbia
design spectral accelerations, the building must be assigned and the University of Sherbrooke have shown that the period
to a more severe Seismic Design Category D. The Cu factor of vibration may be shorter than that predicted by analyti-
for the period is also limited to 1.4, giving a design period cal means (Paultre et al., 2004; Lamarche, 2005). The pos-
CuTa 0.28 s. There is no roof snow load in Los Angeles, thus sible stiffening effect of non-structural roofing components
the seismic weight W reduces to 453 kips. The base shear diminishes to some extent the period lengthening effect of
for the SCBF and OCBF categories are, respectively, equal the roof deck diaphragm (Yang, 2003; Mastrogiuseppe et
to 85.6 kips and 158 kips. For the SCBF system, tension/ al., 2008), but this stiffening effect was found to be limited
compression brace design is selected and a redundancy fac- and not large enough to explain the differences between field
tor R 1.0 can be used despite the more critical SDC because measurements and analytical predictions. One drawback to
removal of one brace results in less than a 33% reduction in the previous ambient vibration tests is that the building pe-
lateral strength, and an extreme torsional irregularity con- riods were obtained from the measurements of small build-
dition does not exist. This is not the case when a tension- ing movement caused by relatively calm wind conditions.
only bracing system is adopted for the OCBF example; It is believed that roof diaphragms exhibit a stiffer response
S 1.3 must be used for this design. under such low amplitude loading because of the inherent
Rigid diaphragm behavior is assumed for the distribution friction resistance of the deck connections and the partial
of the lateral loads, and 0.54 V is considered to be resisted prevention of shear deformations from warping of the deck
along the bracing bent. The selected braces and brace expect- sheets at their ends due to the overlapping of the adjoining
ed tensile and compressive strengths are given in Table 1. sheets. This represents a stiffer condition compared to the
The design shear flow for the diaphragm and the properties single sheet case that was considered in the development of
of the selected roof deck system are also given in Table 1. the Steel Deck Institute (SDI) stiffness equations (Luttrell,
For the SCBF system (R 6.0), the Su value is 0.456 kip/ft 2004), which could also contribute to the observed differ-
[(0.54)(85.6) kips/100 ft] and a 22 ga. (0.0295 in.) deck design ences between field tests and predictions (Figure 5). In this
with 36/4 frame fastener pattern and eight side-lap screws per figure, the periods computed using Equation 4 are compared
joist span is selected: Gd Sn 0.489 kip/ft and G b 25.1 kips/ to those obtained from an empirical expression proposed by
in. For this structure, a shear flow of 2.34 kips/ft is associ- Lamarche (2005) based on field test data. The values 0.05hn
ated with the braces reaching their expected tensile and com- and 0.025hn (where hn building height in m) are the predic-
pressive strengths. This is 5.1 times the design value, indi- tor equations for the period of vibration of a concentrically
cating that the system may not perform as intended when the braced frame based on the 2005 NBCC.
building is subjected to strong seismic ground shaking. For A more recent investigation that compared the results of am-
the OCBF system (R 3.25), Su 1.11 kips/ft [(1.3)(0.54) bient vibration measurements of a 74,100 ft2 (approximately
(158) kips/100 ft]. This higher force demand requires much 23 ft in height, 300 ft by 234 ft in plan) single-story commer-
heavier diaphragm design: 16 ga. (0.0598 in.) deck sheets cial building located in Magog, Quebec, with a 3D SAP 2000
with closely spaced side-lap fasteners. In this OCBF de- building model (Figure 6a) showed that to obtain the mea-
sign, back-to-back angles were selected for the braces. Even sured periods (Figure 6b) a rigidly connected frame structure
if tension-only braces are assumed in design, this type of with infinitely stiff braces and continuous diaphragm panels
brace still possesses compressive strength that should be would need to be assumed instead of the more common pin

connections and standard three- to four-span deck panels as mass. The frame was mounted on rockers and two 220-kip
would normally be used (Tremblay et al., 2008a, 2008a b). high-performance dynamic actuators were used to apply the
Analysis of the building accounting for the in-plane flexibility load along both edges of the supporting steel frame, thus
of the diaphragm and using assumptions commonly made in representing the ground motion forces being transmitted to
practice regarding member end fixity, brace stiffness and the roof by the vertical braces or the walls at the diaphragm
panel lengths led to a period of vibration of 1.11 s, signifi- ends. The intent was to vibrate the diaphragm at increasing
cantly longer than the 0.39 s that was measured. In com- amplitudes to identify whether a decrease of the in-plane
parison, the NBCC would require that the period Ta be a shear stiffness would occur and if so by what extent. Infor-
maximum of 0.35 s, i.e. Ta c 0.05 hn, which is in line with the mation of this nature could be used to determine whether the
ambient vibration result. A question remains as to whether period expressions defined in model building codes could
this NBCC defined period should be allowed to be increased be modified to account for diaphragm behavior under earth-
based on dynamic analyses of building models. quake induced vibrations. Furthermore, the impact of end
Preliminary large-scale dynamic tests were carried out panel overlaps on the in-plane stiffness was investigated by
at École Polytechnique of Montreal in the summer of 2007 testing diaphragm specimens both with and without end-
on three diaphragm test specimens approximately 24 ft by laps. A variety of dynamic tests ranging from low amplitude
69 ft in plan (Figure 7). The test specimens were construct- basic white noise vibrations and harmonic signals to inelas-
ed using the popular 0.0295-in.-thick (22 ga.) 1'-in.-deep tic loading signals were used. In the tests, the influence of
wide-rib deck profile with flutes spaced 6 in. on center. Nail the loading amplitude on the period was assessed through
frame connections and screw side-lap connections were used all of the above dynamic loading protocols (Tremblay et al.
throughout. The test specimens represent a large portion 2008b). The inelastic response of the different diaphragm
of a building’s roof, including the roof structure and roof designs was also examined.
Fig. 5. Analytically computed periods and periods predicted using the empirical expression
based on field test measurements (Tremblay and Rogers, 2005).
Fig. 6. (a) 3-D model of the structure and (b) measured fundamental mode and natural frequency (Tremblay et al., 2008a).

Figure 8 shows the variation of the fundamental period frequency decreased) as larger amplitude cycles were ap-
of vibration of a steel diaphragm specimen with the ampli- plied. The results of the two loading protocols illustrate that
tude of dynamic excitation in terms of absolute acceleration the stiffness and natural frequency of a diaphragm are de-
at mid-span. Several white noise (random vibration signal pendent on the level of deformation demand.
with flat power spectral density used to obtain the frequency
response) tests for which the displacement amplitude of the DUCTILE DIAPHRAGM DESIGN
cycles was systematically increased were used to produce AND BUILDING ANALYSES
this figure. The plot shows that the period rapidly length-
It may also be possible to rely on the inelastic behavior
ens (flexibility increases) as the acceleration level exceeds
of the diaphragm in design; that is, specify the roof deck
that observed in field ambient vibration tests (typically less
diaphragm to be the fuse element in the SLRS instead of
than 0.002 g). The higher initial stiffness is attributed to the
the vertical braces (Figure 1b). The shear capacity of the
lack of slip at the side-lap and deck-to-frame connections
diaphragm can be adjusted by changing the connector spac-
under low amplitude loading. Similarly, specimens submit-
ing and panel thickness, thus leading to a capacity force that
ted to a sine sweep protocol at increasing amplitudes showed
is only marginally higher than the code calculated seismic
that the resonant frequency of the diaphragm did not remain
force. This could lead to a less expensive seismic load
constant; rather the specimen became less stiff (resonant
Fig. 7. Large-scale dynamic diaphragm test setup: (a) plan view and (b) during construction.
Fig. 8. Change in diaphragm period with white noise loading amplitude (adapted from Tremblay et al., 2008b).

resisting system and simpler detailing requirements for the of forces and inelastic deformations in metal roof deck dia-
braces and their connections because they would be ex- phragms under actual seismic conditions, which lead to the
pected to remain in the elastic range (Tremblay and Rogers, tests illustrated in Figure 7.
2005). Experimental and analytical studies of the inelastic The impact of capacity design provisions and period
performance of diaphragms are summarized in the work of limitations on the seismic design of low-rise steel buildings
Tremblay et al. (2004) and Essa et al. (2003) (Figures 9 and was investigated (Tremblay and Rogers, 2005). Several
10). It was shown by means of testing that steel deck made of design strategies were examined, including design without a
0.0295-in.-thick (22 ga.) and 0.0358-in.-thick (20 ga.) sheets capacity-based approach, capacity design with ductile brac-
and connected with mechanical fasteners could undergo ing components, and capacity design assuming the cold-
some limited cyclic inelastic deformations. It was noted, formed steel roof diaphragm acts as the main energy dis-
however, that relying on this inelastic behavior in the design sipation element in the SLRS. The effects of relaxing the
of actual roof diaphragms could result in concentrations of period limitations and the capacity design forces for the roof
large amplitude inelastic deformations in localized regions diaphragm were also considered. A parametric study was
of the roofs, i.e., along braced wall lines, which could lead carried out to evaluate the impact of the different strategies
to undesirable diaphragm failures. This aspect could not be on the cost of the seismic load resisting systems (Figure 11).
addressed in the past test programs due to the limited size The building geometry, the seismic hazard level, the brac-
of the diaphragm specimens (12 ft by 20 ft) and the type of ing configuration and the level of ductility were varied in
displacement controlled loading that was used. Tests under this study. The results show that capacity design provisions
dynamically applied loading on larger diaphragms were have a significant impact on the structure, especially when
needed to properly assess the shear stiffness, distribution tension-compression bracing is used. Substantial savings
1500 kN Actuator Pin

Specimen Frame (typ.)
3.658 m
Hor. Reaction Joist

Vert. Reaction (typ.)
(typ.)
6 096 m
Fig. 9. Quasi-static diaphragm test specimen setup (Essa et al., 2003; Tremblay et al., 2004).
8.0
D3A - 0.6 qu
4.0
q (kN/m)
0.0
-4.0
-8.0
-30.0 -20.0 -10.0 0.0 10.0 20.0 30.0
γ (mrad)
Fig. 10. Weak diaphragm design—concentration of inelastic design

(adapted from Tremblay et al., 2004; Tremblay and Rogers, 2005).

could be realized with the use of a period obtained from Northridge earthquake scaled to match the design spectrum
methods of mechanics that incorporate diaphragm flexibility. for Vancouver, Canada.
Additional solutions for decreasing the cost of the structure The time history response of the drift due to brace de-
also include relaxation of the capacity design provisions by formation (EB) and total building deformation (EB + ED) for
reducing the upper limit on diaphragm forces or selecting the three design scenarios is provided in Figure 13: (a) protected
diaphragm as the main energy dissipating system. Nonlinear diaphragm (brace fuse), where Ta 0.05 hn; (b) weak
dynamic analyses of a limited number of these structures (fuse) diaphragm, where Ta 0.05 hn; and (c) weak (fuse)
(Figure 12) were carried out using the RUAUMOKO (Carr, diaphragm, where Ta T of the building accounting for
2004) computer program. The roof diaphragm was modeled the full diaphragm flexibility. Note: a protected diaphragm
as a deep horizontal plane truss. A Stewart hysteretic model is selected so that its shear and flexural strength exceed
was selected for the diagonal roof truss members in order the forces that correspond to the expected strength of the
to reproduce the cyclic inelastic response measured for the brace (fuse) elements; a weak diaphragm is selected to act
screwed-nailed diaphragm system as described by Tremblay as the inelastic fuse and thus has a strength that only needs
et al. (2004). The software and hysteretic model did not al- to meet the building code seismic force. In all cases, the
low for the simulation of the strength degradation, which building experienced a maximum roof deformation that is
was observed during testing. The response of the example below the prescribed limit of 2.5% hs. A large portion of the
building was examined under one record from the 1994 overall story drift occurred in the bracing members due to
Fig. 11. SLRS cost ratios for (a) Vancouver and (b) Montreal (adapted from Tremblay & Rogers, 2005).
Ground
Motion
20.0
Test 28
(Tremblay et al. 2004)
10.0 Model
q (kN/m)
0.0
-10.0
-20.0
-15.0 -10.0 -5.0 0.0 5.0 10.0
γ (mrad)
L/2
Hysteretic model
Model (half-building)
Fig. 12. Building model and Stewart hysteretic element (Tremblay and Rogers, 2005).

yielding; the in-plane diaphragm displacements were much results show that allowing the inelastic response of the structure
less because this structural element was protected from entering to take place in the roof diaphragm made of thin steel sheets can
into the inelastic range through implementation of the capacity result in an acceptable overall seismic performance. However,
design approach. After the strong motion segment of the record variation in strength and localized demand may result in exces-
which ended at 14 s, significant deformations still developed sive plastic deformations of the diaphragm, and further studies
in the bracing bents because the bracing members had been are needed before this design approach can be adopted.
permanently elongated and were not able to offer any lateral
resistance near the zero deformation position. As also expected CONCLUSIONS
from the design assumptions, plastic deformation was not ob-
Seismic provisions of modern building codes now rely on
served in the roof diaphragm. In contrast, for the buildings in
capacity design procedures to provide a desired hierarchy
which the diaphragm was designed to be the fuse element, the
of material yielding in the SLRS and better control of the
peak roof displacement remained nearly the same but the in-
inelastic response of a structure. For single-story steel
elastic demand switched from the bracing members to the roof
buildings with concentrically braced steel frames, inelastic
diaphragm. The peak plastic demand in the roof, Gp, is in ac-
response is typically concentrated in the diagonal bracing
cordance with the recommended permissible value of 10 mrad
members of the braced bays. Other components along the
( 10 × 10-3 rad) of shear deformation for nailed-screwed decks
lateral load path, such as the roof diaphragm, including its
(Essa et al., 2003).
chords and collectors, must be designed to resist the forces
It must be realized that this is a single example building sub-
that will develop upon yielding in the vertical components of
jected to only one ground motion and that the performance can
the seismic load resisting system. Current seismic provisions
vary significantly with ground motions and building dimen-
in the U.S. for buildings with R > 3 do not result in entirely
sions. In particular, inelastic demand can be very sensitive to
consistent design between the steel framing and the roof
design and modeling assumptions, as well as loading condi-
diaphragm. If full capacity design principles were required,
tions. In addition, this study was limited to uniform rectangular
much higher design forces would need to be applied to the
structures and it is expected that higher ductility demand can
diaphragm. For simple metal roof deck design, an SCBF
be induced in structures with irregularities or a non-symmetric
example studied herein for the Boston area showed that the
footprint, as often encountered in practice. Nonetheless, the
roof deck would need to be increased from a thickness of
0.0295 in. to 0.0474 in. (22 ga. to 18 ga.) with a more closely
spaced fastener arrangement. Similar results were obtained
for the other cases that were studied. Alternative design ap-
proaches that reduce the force demand on the diaphragm are
being evaluated. The designer could possibly take advantage
of the in-plane flexibility of the roof diaphragm, as this is
currently permitted in ASCE 41 for the seismic retrofit of
existing structures. Parametric studies performed for Cana-
dian seismic conditions have shown that there is a significant
potential for savings if the period from dynamic analysis is
used in design. However, field ambient vibration test data
seem not to support this approach and caution must be ex-
ercised before using the period prediction that accounts for
roof diaphragm flexibility in seismic design. It may also be
possible to allow for inelastic deformation in the roof dia-
phragm, instead of the diagonal bracing members. For thin
deck sheets, these deformations can develop in the form of
bearing or tearing in the vicinity of the deck fasteners. De-
formation capacity is however limited and means must be
taken to ensure that it will be properly distributed over the
diaphragm area so that no concentration will develop that
can lead to failure of the diaphragm system, i.e., a loss in
the ability to transfer lateral forces to the bracing bents and
a possible decrease in the effectiveness of the deck panels
to laterally brace the supporting joist and beam structure.
Fig. 13. Time histories of the story drifts for various design
Research projects have been undertaken to examine the po-
scenarios under a site-representative earthquake ground motion
tential use of these two alternative design strategies.
(adapted from Tremblay and Rogers, 2005).

ACKNOWLEDGMENTS Essa, H.S., Tremblay, R. and Rogers, C.A. (2003), “Behavior
of Roof Deck Diaphragms under Quasistatic Cyclic Load-
The authors would like to acknowledge the support provided ing,” Journal of Structural Engineering, ASCE, Vol. 129,
by the Canada Foundation for Innovation, the Canadian No. 12, pp. 1658–1666.
Sheet Steel Building Institute, the Canam Group Inc., WSB
Lamarche, C.-P. (2005), “Étude expérimentale du comporte-
Consulting Structural Engineers, RJC Consulting Engineers,
ment dynamique des bâtiments de faible hauteur en acier,”
the Vancouver Steel Deck Diaphragm Committee, the Steel
M.Sc.A. Thesis, Department of Civil Engineering, Uni-
Deck Institute and Hilti Limited. Additional funding for this
versity of Sherbrooke, Sherbrooke, QC, Canada.
research was provided by the Natural Sciences and Engineer-
ing Research Council of Canada and the Funds for Research Luttrell, L. (2004), Diaphragm Design Manual, 3rd ed., in-
in Nature and Technologies of the Province of Quebec. The cluding Appendix VI Addendum November 2006, Pub-
authors thank research assistants Camelia Nedisan, Charles- lication No. DDMO3, Steel Deck Institute, Fox River
Philippe Lamarche, John Franquet and Robert Massarelli, as Grove, IL
well as Patrice Belanger, technician at École Polytechnique, Mastrogiuseppe, S., Rogers, C.A., Nedisan, C.D. and Trem-
for their valuable assistance during the large-scale diaphragm blay, R. (2008), “Influence of Non-structural Components
test program. on Roof Diaphragm Stiffness and Fundamental Periods of
Single-Storey Steel Buildings,” Journal of Constructional
REFERENCES Steel Research, Vol. 64, No. 2, pp. 214–227.
AISC (2005a), Seismic Provisions for Structural Steel Build- Medhekar, M.S. (1997), “Seismic Evaluation of Steel Build-
ings, ANSI/AISC 341-05, American Institute of Steel ings with Concentrically Braced Frames,” Ph.D. Thesis,
Construction Inc., Chicago, IL. Department of Civil and Environmental Engineering,
University of Alberta, Edmonton, AB, Canada.
AISC (2005b), Specification for Structural Steel Buildings,
ANSI/AISC 360-05, American Institute of Steel Con- Naman, S.K. and Goodno, B.J. (1986), “Seismic Evaluation
struction Inc., Chicago, IL. of a Low-Rise Steel Building,” Engineering Structures,
Vol. 8, No. 1, pp. 9–16.
AISC (2006), Seismic Design Manual, American Institute of
Steel Construction Inc. and The Structural Steel Educa- NRCC (2005), National Building Code of Canada, 12th
tional Council, Chicago, IL. ed., National Research Council of Canada, Ottawa, ON,
Canada.
ASCE 7 (2005), Minimum Design Loads for Buildings and
other Structures, including Supplements Nos. 1 and 2, Paultre, P., Proulx, J., Ventura, C., Tremblay, R., Rogers, C.,
ASCE/SEI 7-05, American Society of Civil Engineers, Lamarche, C.-P. and Turek, M. (2004), “Experimental
Reston, VA. Investigation of Low-Rise Steel Buildings for Efficient
Seismic Design,” Proceedings of the 13th World Confer-
ASCE 41 (2006), Seismic Rehabilitation of Existing Build-
ence on Earthquake Engineering, Vancouver, BC, Canada,
ings, American Society of Civil Engineers, Reston, VA.
Paper No. 2919.
AISI (2004), Supplement 2004 to the North American Speci-
Tena-Colunga, A. and Abrams, D.P. (1996), “Seismic Be-
fication for the Design of Cold-Formed Steel Structural
havior of Structures with Flexible Diaphragms,” Journal
Members, 2001 ed., American Iron and Steel Institute,
of Strucural Engineering, ASCE, Vol. 122, pp. 439–445.
Washington, DC.
Tremblay, R. (2005), “Canadian Seismic Design Provisions
Carr, A.J. (2004), RUAUMOKO, Inelastic Dynamic Anal-
for Steel Structures,” Proceedings of the 2005 North
ysis Program, Dept. of Civil Engineering, University of
American Steel Construction Conference, Montreal, QC,
Canterbury, New Zealand.
Canada, AISC, Chicago, IL, 12 pp.
CSA (2005), Supplement No. 1 to CAN/CSA-S16-01 Limit
Tremblay, R. and Stiemer, S.F. (1996), “Seismic Behav-
States Design of Steel Structures, CSA-S16s1-05, Cana-
ior of Single-Storey Steel Structures with Flexible Dia-
dian Standard Association, Toronto, ON, Canada.
phragm,” Canadian Journal of Civil Engineering, Vol. 23,
Dubina, D., Zaharia, R. and Georgescu, M. (1997), “Stress pp. 49–62.
Skin Design of Single-Story Steel Buildings in Seismic
Tremblay, R. and Bérair, T. (1999), “Shake Table Testing
Areas,” Proceedings of the STESSA 97 Conference on the
of Low-Rise Steel Buildings with Flexible Roof Dia-
Behaviour of Steel Structures in Seismic Areas, Kyoto,
phragms,” Proceedings of the 8th Canadian Conference
Japan, 1997, pp. 426–433.
on Earthquake Engineering, Vancouver, BC, Canada,
pp. 585–590.

Tremblay, R., Bérair, T. and Filiatrault, A. (2000), “Experi- Tremblay, R., Nedisan, C., Lamarche, C.-P. and Rogers, C.A.
mental Behaviour of Low-Rise Steel Buildings with Flex- (2008a), “Periods of Vibration of a Low-Rise Building
ible Roof Diaphragms,” Proceedings of the 12th World with a Flexible Steel Roof Deck Diaphragm,” 5th Inter-
Conference on Earthquake Engineering, Auckland, New national Conference on Thin-Walled Structures, Brisbane,
Zealand, Paper No. 2567. Australia, pp. 615–622.
Tremblay, R. and Rogers, C.A. (2005), “Impact of Capacity Tremblay, R., Rogers, C.A., Lamarche, C.-P., Nedisan, C.D.,
Design Provisions and Period Limitations on the Seismic Franquet, J., Massarelli, R. and Shrestha, K.M. (2008b),
Design of Low-Rise Steel Buildings,” International Jour- “Dynamic Seismic Testing of Large Size Steel Deck
nal of Steel Structures, Vol. 5, No. 1, pp. 1–22. Diaphragms for Low-Rise Building Applications,” 14th
Tremblay, R., Rogers, C.A. and Nedisan, C. (2002), “Use of World Conference on Earthquake Engineering, Beijing,
Uniform Hazard Spectrum and Computed Period in the China, Paper No. 05-05-0066.
Seismic Design of Single-Storey Steel Structures,” 7th Yang, W. (2003), “Inelastic Seismic Response of Steel Roof
U.S. National Conference on Earthquake Engineering, Deck Diaphragms Including Effects of Non-Structural
Boston, MA, Paper No. 195. Components and End Laps,” M.A.Sc. Thesis, Department
Tremblay, R., Rogers, C., Martin, É. and Yang, W. (2004), of Civil, Geological and Mining Engineering, École Poly-
“Analysis, Testing and Design of Steel Roof Deck Dia- technique, Montreal, QC, Canada.
phragms for Ductile Earthquake Resistance,” Journal of
Earthquake Engineering, Vol. 8, No. 5, pp. 775–816.

Notes on the Impact of Hole Reduction
on the Flexural Strength of Rolled Beams
LOUIS F. GESCHWINDNER
ABSTRACT
The use of G 0.9 and 8 1.67 with the provisions in Section F13.1 of AISC 360-05 (AISC, 2005) to account for the reduction in flexural strength
for a beam with holes in the tension flange has been questioned several times since the publication of the Specification for Structural Steel Build-
ings in 2005. The intent of this paper is to review and provide justification for the use of the resistance/safety factors within the 2005 Specification
provisions for the impact on flexural strength of holes in the tension flange.
Keywords: bolt holes, tension flange, resistance factors, safety factors.
INTRODUCTION Fu A fn
Fcr Fy (2)
The use of G 0.9 and 8 1.67 with the provisions in Sec- Yt Fy A fg
tion F13.1 of AISC 360-05 (AISC, 2005) to account for the
reduction in flexural strength for a beam with holes in the Because ASTM A36 and A992 steels meet the limit given
tension flange has been questioned several times since the in Section F13 for Yt 1.0, and the term
publication of the Specification for Structural Steel Build-
ings in 2005. The research basis for the 2005 provisions Fu A fn

originated from a report by Dexter et al. (2002). Their report Fy A fg
includes a proposed formulation for the limits on when the
impact of the holes must be considered and how the strength in Equation 2 is always less than 1.0 if this check is appli-
should be determined in those cases; however, the report cable, the critical stress is always less than the yield stress.
does not address the appropriate resistance or safety factors Therefore, use of the resistance/safety factors associated with
to be used. The intent of this paper is to review and provide yielding (i.e., G 0.9, 8 1.67) appears to be warranted.
justification for the use of the resistance/safety factors within To examine this interpretation more closely, three models
the 2005 Specification provisions for the impact on flexural are developed for determining flexural strength when holes
strength of holes in the tension flange. are present in the tension flange of W-shapes.
THEORY Model 1
AISC 360-05 Equation F13-1 is intended to present a simple For ease of calculation, the W-shape is modeled with holes in
yet reasonable approach to account for holes in the tension both the tension and compression flanges. The flange forces
flange of beams. Although not presented this way in the are taken as the rupture force and the web is assumed to be
Specification, Equation F13-1 can be rewritten in terms of yielding throughout. Thus,
critical stress, always less than Fy, times the full elastic sec-
tion modulus as illustrated by the following two equations: ( )
(
M n Fu A fn d t f Fy Z x A fg d − t f

) (3)
M n = Fcr S x (1) If the ratio of flange-rupture strength to flange-yield

strength, which is always less than 1.0, is taken as
Fu A fn (4)

Fy A fg
Equation 3 can be stated as

Louis F. Geschwindner, Ph.D., P.E., Vice President, American Institute of
Steel Construction, 1 E. Wacker Drive, Suite 700, Chicago, IL, 60601. E-mail:
[email protected]
( )( )(
M n Fy Z x 1 Fy A fg d t f ) (5)

Since this equation represents a rupture failure mode, the flange and the web contributions. This multiple factor ap-
resistance factor, G 0.75, and safety factor, 8 2.00, are proach is similar to that used in connection design. For ex-
used to determine the available strength. ample, a bolted flange plate moment connection could have
When there is no reduction for the presence of holes, the tension flange plate controlled by rupture and the com-
: 1.0, Equation 5 reduces to Mn Fy Zx. Thus, the de- pression flange plate controlled by yielding. Thus, different
sign strength is GMn 0.75Fy Zx and the allowable strength resistance/safety factors would be applied in the design of
is Mn /8 0.5Fy Zx. Similarly, the provisions of Sec- each element yet they both participate in resisting the same
tion F13, as represented by Equations 1 and 2, reduce to connection moment. In the application here, using Equation
Mn Fy Sx. Thus, the design strength, using G 0.9, becomes 6, this approach yields directly the design strength as
GMn 0.9Fy Sx and the allowable strength, using 8 1.67,
d tf
becomes Mn /8 0.6Fy Sx. In all cases where the shape fac-
tor, Zx /Sx, is greater than or equal to 1.2 (0.9/0.75 1.2 or
(
M n 0.9 Fy Z x 0.9 0.75 Fy A fg )( ) 2

( ) (7)
2
2.00/1.67 1.2), the Specification approach gives a lower or 0.9 1 Fy A fg
equal available strength when compared to this model. But,
4 Fy t w
if the shape factor is less than 1.2, this model, which was
initially thought to be conservative, gives a lower value than
or the allowable strength as
the Specification approach.
d tf
)( )
Mn
Model 2

(
0.6 Fy Z x 0.6 0.5 Fy A fg
2
( ) (8)
2
A second model is investigated to see if this underprediction 0.6 1 Fy A fg
can be reversed by eliminating the holes at the compression
4 Fy t w
flange which were included for convenience only.
For this model, only the holes in the tension flange are
For this model, as : approaches 1.0, the contribution of the
accounted for and the compression flange is not reduced. It
tension flange is not fully restored to its yield strength since
takes a bit more calculation effort to determine the nominal
its contribution is always modified by the rupture resistance/
strength with this approach, but it is expected to yield a more
safety factor when : 1. This amounts to a 15% reduction in
accurate representation of the true behavior. In this case, pro-
the contribution of the tension flange to the design strength.
vided the plastic neutral axis remains in the web,
For this model, the design strength and the allowable strength
d t f (1 ) Fy A fg (6)
2

for all values of the ratio of flange-rupture strength to flange-
M n Fy Z x (1 )( Fy A fg ) yield strength results in available strengths greater than that
2 4Fy t w obtained using Equation F13-1 from AISC 360-05.
and the resistance/safety factor is again taken as G 0.75 or SUMMARY

8 2.00 since the strength calculation considered rupture
Figure 1 illustrates the LRFD results for the three models
of the tension flange. This model yields higher available
discussed earlier compared to the Specification equation
flexural strength for most of the range of the ratio of flange-
for a W8×24. This particular shape was chosen because it
rupture strength to flange-yield strength. However, as the re-
is compact and has a shape factor close to the lowest of all
duction for holes gets smaller—that is, as : approaches 1.0,
W-shapes, 1.105. It can be seen that Model 3 predicts design
the nominal strength approaches Fy Zx and the same problem
strengths greater than those predicted by Specification Equa-
occurs as for Model 1, where the available strength predicted
tion F13-1. Identical comparisons would result if ASD had
by this model is lower than that predicted by the Specifica-
been used for the figure.
tion approach for W-shapes with a shape factor below 1.2.
The intent of this study was to confirm that Equation F13-1
with G 0.9 or 8 1.67 provides a prediction of flexural
Model 3
strength that is conservative. Models 1 and 2 show that for
A third approach is developed with the goal of increasing all W-shapes with a shape factor of 1.2 or greater, the predic-
the design strength for those cases where the reduction for tion by the Specification equation is conservative. However,
holes is small. a more accurate model was needed for W-shapes with a shape
Since the flange in tension is controlled by tension rupture factor less than 1.2. Model 3 is a reasonable analytical ap-
and the remainder of the shape is controlled by yielding in proach that can be considered conservative. Since Model 3 al-
Model 2, Model 3 simply applies two different resistance/ ways provides an available flexural strength greater than that
safety factors, G 0.75 or 8 2.00, for the tension flange obtained using the Specification provisions, it is considered
contribution and G 0.9 or 8 1.67, for the compression acceptable to use Equation F13-1 with G 0.9 or 8 1.67.

NOTATION REFERENCES
The notation used in this paper is consistent with that used in AISC (2005), Specification for Structural Steel Buildings,
ANSI/AISC 360-05 with one symbol added, :. ANSI/AISC 360, American Institute of Steel Construc-
tion, Chicago, IL.
Afg gross flange area, in.2 (mm2)
Dexter, R.J., Alttstadt, S.A. and Gardner, C.A. (2002),
Afn net flange area, in.2 (mm2) Strength and Ductility of HPS70W Tension Members and
Fcr critical stress, ksi (MPa) Tension Flanges with Holes, University of Minnesota,
Minneapolis, MN, March 2002.
Fu specified minimum tensile strength, ksi (MPa)
Fy specified minimum yield stress, ksi (MPa) APPENDIX
Mn nominal flexural strength, kip-in. (N-mm) Derivations for the equations presented in this paper follow.
Sx elastic section modulus about the x-axis, in.3 (mm3)
Model 1
Yt hole reduction coefficient
In this case, both flanges are assumed to have the same re-
Zx plastic section modulus about the x-axis, in.3 (mm3) duction for the presence of holes. The nominal plastic mo-
ment strength is given by Fy Zx when no holes are present.
d depth of section, in. (mm)
To account for the reduced strength of the flanges, the yield
tf thickness of flange, in. (mm) contribution of both flanges is deducted and replaced by the
tension rupture contribution. This is clearly a conservative
tw thickness of web, in. (mm)
approach for determining the nominal flexural strength since
G resistance factor it ignores the actual contribution of the compression flange.
Thus,
8 safety factor
: flange rupture to yield strength ratio M n Fy Z x 2 Fy A fg d t f 2F A d t f (a)
2 2 u fn
2 2
W8x24 with Holes in Tension Flange
90
80
70
Design Strength, Phi(Mn), (kip-ft)
60
50
40
30
20
10
0
0 02 0.4 06 0.8 1
Ratio of Flange Rupture Force to Flange Yield Force
Model 1, phi=0.75 Model 2, phi=0.75 Model 3, phi varies AISC 360-05 Eq. F13-1, phi=0.9
Fig. 1. Design strength for a W8×24 with holes in the tension flange.

Combining terms and multiplying the second term by Substituting for x yields
Fy Ag yields ⎛ d tf ⎞ ⎛ d tf ⎞
Fy Ag ( )
M n 3 = Fy Z x − 2 Fy A fg ⎜ − ⎟ + Fu A fn ⎜ − ⎟
⎝2 2⎠ ⎝2 2⎠
M n = Fy Z x +
(F A u fn − Fy A fg ) ( d − t )( F A ) (b) ⎛ d tf ⎞ (
Fy A fg − Fu A fn )
2 (g)
Fy A fg
f y fg + Fy A fg ⎜ − ⎟ −
⎝2 2⎠ 4 Fy t w
Fu A fn Fu A fn
Defining and substituting into Equation b gives Combining terms and substituting yields
Fy A fg Fy A fg
( )( )( )
2
M n Fy Z x 1 Fy A fg d t f (5)
d t f 1 Fy A fg

M n Fy Z x 1 Fy A fg
4Fy t w
(6)
2
Model 2
Model 3
For this model, only the holes in the tension flange are
considered. First, the contribution of the web is determined The only difference between Model 3 and Model 2 is the
by deducting the flanges from the nominal plastic moment application of the yielding and rupture resistance/safety
strength of the W-shape. factors. Based on Equation g, for LRFD the design strength
becomes
d tf
M n1 = Fy Z x − 2 Fy A fg −
2 2
(c)

d tf

d tf
M n y Fy Z x y 2 Fy A fg r Fu A fn
2 2
2 2
(h)

2
Then the tension flange rupture and compression flange yield d t f y Fy A fg Fu A fn
contributions are added. y Fy A fg
2 2
4 Fy t w
d tf d tf
2 2

M n 2 = Fy Z x − 2 Fy Afg − + Fu Afn −
2 2
where Gy is the resistance factor for yielding and Gr is the
resistance factor for rupture.
(d) Fu A fn
d tf Combining terms and substituting , Gy 0.9, and
+ Fy Afg − Fy A fg
2 2 Gr 0.75 yields
⎛ d tf ⎞
Finally, the last factor to consider is the impact of the shift in (
φM n 0.9 Fy Z x 0.9 0.75 Fy A fg ⎜ )(⎟ )
the plastic neutral axis into the compression zone of the web, ⎝ 2 ⎠
( )
2
defined as distance x from the centroid of the gross area. 0.9 ⎡⎣ 1 Fy A fg ⎤⎦ (7)
This results in a moment reduction based on the removal of
4 Fy t w
some compression force and the addition of an equal ten-
sion force, captured through the multiplication by 2 in the For ASD the safety factors are applied to Equation g, yielding
last term. These forces are half of the difference between the
flange yield force and flange rupture force. Thus, M n Fy Z x
2

Fy A fg d t f Fu A fn d tf

y y 2 2 r 2 2
d tf d tf (i)

M n 3 Fy Z x 2 Fy A fg Fu A fn
2 2 2 2 Fy A fg d t f
Fy A fg Fu A fn
2
(e)
d tf x2
y 2 2 y 4 Fy t w
Fy A fg 2 F t
y w
2 2 2 where 8y is the safety factor for yielding and 8r is the safety
factor for rupture.
and the distance that the plastic neutral axis moves up into Fu A fn
the compression zone, x, is Combining terms and substituting , 8y 1.67, and
8r 2.00 yields Fy A fg
Fy A fg − Fu A fn (f) ⎛ d tf ⎞
x=
)( )
Mn
2 Fy t w

(
0.6 Fy Z x 0.6 0.5 Fy A fg ⎜ ⎟
⎝ 2 ⎠ (8)
( )
2
0.6 ⎡⎣ 1 Fy A fg ⎤⎦

4 Fy t w

A Case for a Single Stiffness Reduction
Factor in the 2010 AISC Specification
LOUIS F. GESCHWINDNER
ABSTRACT
The 2005 AISC Specification for Structural Steel Buildings includes a stiffness reduction factor, Ub, in Appendix 7 to be used in the direct analysis
method to account for the presence of residual stresses and their influence on the second-order effects of frame behavior. The 2005 Commentary
includes a stiffness reduction factor, Ua, to be used along with the effective length nomograph to account for the influence of column inelasticity
due to residual stresses on effective length. These two stiffness reduction factors are intended to account for the same effect yet they are differ-
ent. This paper provides the background for these two factors, and it will demonstrate that Ub is the more correct stiffness reduction factor. The
2010 AISC Specification will recommend its use with both the direct analysis method and the effective length nomograph.
Keywords: stiffness reduction factor, direct analysis method, effective length nomograph, column inelasticity, residual stresses.
INTRODUCTION derivation of the equations upon which the nomographs

are based was that all members in the frame behave elasti-
Appendix 7 of the Specification for Structural Buildings
cally. Since it was known at the time that columns behaved
(AISC, 2005) includes a stiffness reduction factor, Ub, to be
inelastically if they had sufficiently low slenderness ratios,
used in the direct analysis method to account for the pres-
there was a need to address the influence of inelastic column
ence of residual stresses and their influence on the second-
behavior on K. This was accomplished by using a stiffness
order effects of frame behavior. The 2005 Commentary in-
reduction factor.
cludes a stiffness reduction factor, Ua, to be used along with
The first introduction of a stiffness reduction factor to be
the effective length nomograph to account for the influence
used in conjunction with the nomograph is found in the 1971
of column inelasticity due to residual stresses on effective
paper by Yura (1971). He shows that the joint stiffness ratio,
length. These two stiffness reduction factors are intended to
G, used with the nomograph included the modulus of elastic-
account for the same effect, yet they are different. This paper
ity, E, in the numerator and denominator so that
provides the background for these two factors and the first
stiffness reduction factor introduced in the literature in the
early 1970s. It will demonstrate that Ub is the more correct G=
Column stiffness EI L col
=

stiffness reduction factor and it should be used with both the Beam stiffness EI L beam (1)
direct analysis method and the effective length nomograph.
I L
= = Gelastic
col
STIFFNESS REDUCTION FACTOR FOR I L beam

DETERMINATION OF INELASTIC K-FACTOR
However, if the columns behaved inelastically, E for the col-
The requirement to use an effective length factor, K, in
umns should be replaced by the tangent modulus, ET, and the
the determination of column strength was introduced with
joint stiffness ratio became
the 1963 AISC Specification. At about the same time, two
nomographs were published to assist in determining these
effective length factors. One of the assumptions used in the Ginelastic
(
ET I L ) col

ET
Gelastic (2)
( EI L ) beam
E
Yura pointed out that the difficulty would be in determining

ET /E. He went on to say that this could be accomplished in a
reasonably accurate manner by noting that for a given Kl/r,
ET Fcr (inelastic )
Louis F. Geschwindner, Ph.D., P.E., Vice President, American Institute of Steel = (3)
Construction, 1 E. Wacker Drive, Suite 700, Chicago, IL, 60601. E-mail: ge- E Fe
[email protected]

where Fcr(inelastic) is the critical stress in the inelastic region 1969 AISC SPECIFICATION
and
From the 1969 AISC Specification, column allowable stress-
2E
Fe 2
es are given:
Kl
r for Kl Cc 2 2 E (inelastic behavior)
r Fy
By using the stress equations from the AISC Specification,
2
an approximate relationship could be established as Kl r
1 − Fy
ET Fcr (inelastic ) Fa 2Cc2
(4) Fa = (5)
E Fe Fe' 3
5 3 Kl r
+
−
Kl r
where Fa is the allowable stress given by the Specification 3 8Cc 8Cc3
and
12 2 E 2 2 E (elastic behavior)
Fe 2
and for Kl Cc
Kl r Fy
23
r
12 2 E
Fa (6)
Yura also noted that Fa and Feb use different factors of safety,
( )
2
23 Kl r
but that this could be ignored as a minor factor in the devel-
opment. while
Equation 4 is the basis for the stiffness reduction factors
12 2 E
used with the Allowable Stress Design (ASD) specifications Fe (7)
( )
2
from 1969 through 1978, while Equation 3 is the basis for 23 Kl r
Ua presented in the 2005 AISC Specification. Although these
stiffness reduction factors are all based on the same rela- Equations 5, 6, and 7 include the effects of both residual
tionship, Equation 3, the actual value of the stiffness reduc- stresses and initial out of straightness (Johnson, 1966). Thus,
tion factor has gone through some variation as the column the ratio of Fa /Feb is not the ratio of ET /E but an approxima-
strength equations have changed through subsequent edition, as already indicated by Yura.
tions of the specification. Substituting Equations 5 and 7 into Equation 4 and defin-
ing the stiffness reduction factor as SRF1969 yields
USE OF ACTUAL STRESS RATHER
THAN ALLOWABLE STRESS 23 Fy Fy
1
Fa 12 4 Fe Fe
The use of the stiffness reduction factor as originally pre- SRF1969 (8)
sented by Yura (1971) was an iterative process. First, the Fe
3

5 3 Kl r Kl r
stiffness reduction factor was assumed to be 1.0 and K was
3 8Cc 8Cc3
determined. With this K, the allowable stress was determined
and a new stiffness reduction factor was found. This stiffness
reduction factor lead to a new, reduced K, which led to a new The factor 23/12 in the numerator is the safety factor for
allowable stress and another new reduced K. This process elastic buckling while the three-term denominator is the
eventually converged and, in the example presented by Yura, safety factor for inelastic buckling. In the inelastic region,
the column actually ended up having K 1.0. for Kl /r 0 the safety factor is 5/3 while for Kl /r Cc, the
In order to eliminate the need to iterate in the determi- safety factor is 23/12. For design according to the 1969 AISC
nation of the stiffness reduction factor, Disque (1973) rec- Specification, the safety factors account for the influence of
ommended using the actual stress rather than the allowable initial out of straightness while inelasticity is accounted for
stress. He also suggested that an even more conservative through the basic equations.
approach would be to use the maximum possible allowable It should be noted that in the elastic buckling region, when
stress, 0.6Fy, and he provided tables in his paper for that ap- Kl/r > Cc, the ratio of Fa /Feb is 1.0 and there is no inelastic
proach. The 8th edition Steel Construction Manual included stiffness reduction factor, as would be expected.
a stiffness reduction factor table based on the actual stress.

1986 AISC LRFD SPECIFICATION Where Pn FcrAs and Fcr is determined through Equation 9.
Equations 11, 12 and 13 provide the same results.
With the introduction of the 1986 AISC LRFD (Load and
A new stiffness reduction factor, Ub, to be used in the
Resistance Factor Design) Specification, column behavior
Direct Analysis Method was introduced in the 2005 AISC
was defined at the critical stress level, without the use of
Specification. It is based on column strength curves pro-
factors of safety. In addition, the column strength equation
posed by the Column Research Council in 1960 (Johnson,
in the inelastic region was changed. Thus,
1960; Galambos, 1998), where
Kl Fy
for c 1.5 (inelastic behavior) KL Fy
r E for 2 (inelastic behavior)
r E
(
Fcr 0.658
c2
)F
y
(9)
Fcr 2 (14)
1
and for Mc > 1.5 (elastic behavior) Fy 4
0.877 and for 2 (elastic behavior)

Fcr 2 Fy (10)
c Fcr 1
= 2 (15)
These equations also include the influence of initial out of Fy
straightness and residual stresses. In the elastic buckling The slenderness parameter, M, is the same as Mc introduced
region, Equation 10, the primary factor influencing column in the 1986 AISC LRFD Specification. Equations 14 and 15
strength is the initial out of straightness. Thus, the 0.877 fac- include the effects of residual stresses but do not include
tor can be thought of as accounting for initial out of straight- the effects of out of straightness (Johnson, 1960). With the
ness. In the inelastic region, Equation 9, the separate influ- substitution of M, Equation 15 reduces to the Euler buckling
ences of residual stresses and initial out of straightness can stress equation,
not be clearly distinguished. This is the same situation that
2E
was seen when using the 1969 ASD equations. Fcr Fe 2
(16)
The stiffness reduction factor based on Eq. 3 and the Kl
LRFD column strength equations then becomes, for col- r
umns with Mc c 1.5
and Equation 14 reduces to
SRF1986 =
ET Fcr (inelastic )
= =
λ2
0.658 c Fy (11) Fy
E Fcr (elastic ) 0.877 Fcr = 1 − Fy (17)
4 Fe
2 Fy
λc
The 2005 Commentary says that Ub is similar to the inelastic
stiffness reduction factor. For use with the nomograph, the
2005 AISC SPECIFICATION stiffness reduction factor defined by Equation 3, using Equa-
tions 16 and 17, yields
The column strength equations found in the 2005 AISC
Specification are essentially the same as those from the 1986 ET Fy Fy
AISC LRFD Specification. The only change is how the divi- b = = 1 − (18)
E 4 Fe Fe
sion between elastic and inelastic behavior is defined and
the format of the actual terms in the equations. For the 2005
Specification, Eq. 11 becomes COMPARISON OF STIFFNESS REDUCTION
Fy
FACTORS BASED ON SLENDERNESS RATIOS
0.658 Fe Fy
The three stiffness reduction factors given in Equations 8,
E
a T (12) 12 and 18 are each a function of Kl/r, directly and through
E 0.877 Fe Fe. Figure 1 presents these three equations as a function of
M, which itself is a function of Kl/r, for Fy 50 ksi. Table 1
The 2005 Commentary gives Ua in the format developed by gives values for Equations 8, 12, and 18 for M 0 to 1.5
the ASCE Task Committee on Effective Length (ASCE, (Kl/r 0 to 113.5). It is seen from these results that the
1997) as three equations for determining the stiffness reduction factor
Pn Pn provide essentially the same results when presented as a
a 2.724 ln (13)
function of slenderness. This confirms the claim of Yura
Py Py

(1971) that the variation in safety factor was a minor factor. Table 1. Stiffness Reduction Factors
It also shows that the use of strength equations that include Based on Slenderness
initial out of straightness and residual stresses (Equations
8 and 12), rather than just residual stresses (Equation 18) M SRF1969 Ua Ub
is also a minor factor in determining the stiffness reduction Eq. 8 Eq. 12 Eq. 18
factor. Since the stiffness reduction factor was initially de- 0 0.000 0.000 0.000
veloped to assist in the proper use of the nomographs, which 0.05 0.003 0.003 0.002
are of course based on Kl/r, it is clear that any of these three 0.10 0.011 0.011 0.010
approaches would be reasonable for this purpose.
0.15 0.025 0.025 0.022
0.20 0.044 0.045 0.040
STIFFNESS REDUCTION FACTOR FOR USE
0.25 0.068 0.069 0.062
IN THE DIRECT ANALYSIS METHOD
0.30 0.097 0.099 0.088
The stiffness reduction factor to be used with the direct 0.35 0.129 0.133 0.119
analysis method is defined as Ub as discussed above and is
0.40 0.166 0.171 0.154
given directly as an unnumbered equation in Appendix 7 of
the 2005 AISC Specification. 0.45 0.207 0.212 0.192
0.50 0.250 0.257 0.234
For Pr /Py ≤ 0.5
0.55 0.297 0.304 0.280
b 1.0 (19)
0.60 0.346 0.353 0.328
and for Pr /Py > 0.5 0.65 0.396 0.404 0.378
0.70 0.449 0.455 0.430
P Pr
b 4 r 1 (20) 0.75 0.502 0.507 0.483
Py Py

0.80 0.555 0.558 0.538
0.85 0.608 0.609 0.592
In Equation 20, B 1.0 for LRFD and B 1.6 for ASD. This
factor is used to insure that the calculation is performed at an 0.90 0.661 0.658 0.646
ultimate strength level. 0.95 0.712 0.705 0.699
This relationship was developed from the CRC column 1.00 0.762 0.750 0.750
strength equations presented above as Equations 14 and 15. 1.05 0.808 0.792 0.799
1.10 0.852 0.831 0.844
1.15 0.891 0.867 0.885
1.200
1.20 0.926 0.899 0.922
Ĳa (T 1.25 0.955 0.926 0.952
1.000
1.30 0.977 0.950 0.976
0.800
1.35 0.993 0.969 0.992
SRF1969 (T 1.40 1.000 0.984 1.000
Factor
0.600 1.45 1.000 0.994 1.000

Ĳb (T 1.50 1.000 1.000 1.000
0.400
0.200
If the inelastic critical stress is taken as Ub times the elastic

0.000 critical stress, then from Equation 15
0 0.5 1 1.5 2 2.5
Slenderness
1 (21)
Fcr b 2 Fy

Fig. 1. Comparison of the three approaches to determining the
stiffness reduction factor as a function of slenderness parameter, M, This is simply a restatement of the Equation 3 and applies
for use with the nomograph. for both elastic and inelastic behavior when using Equations

19 and 20 for Ub. If Equation 21 is solved for M2 and the result
substituted into Equation 14, Table 2. Stiffness Reduction Factors Based on Pn /Py
Pn /Py SRF1969 Ua Ub
Fcr 1 b Fy
1 (22)
4 Fcr
Eq.8 Eq. 13 Eq. 24
Fy 1.000 0.000 0.000 0.000
0.991 0.003 0.023 0.034
Equation 22 can then be solved for Ub, which yields
0.982 0.011 0.049 0.071
Fcr Fcr
b 4 0.971 0.025 0.077 0.112

1 (23)
Fy Fy
0.960 0.044 0.108 0.155
0.947 0.068 0.140 0.200
which can be written in terms of strength as 0.934 0.097 0.175 0.248
Pn 0.919 0.129 0.211 0.297
Pn
b 4 1 (24) 0.904 0.166 0.249 0.347

Py Py
0.888 0.207 0.287 0.398
Equations 20 and 24 are the same except that one is written 0.871 0.250 0.327 0.449
in terms of BPr, the amplified required strength, and one in 0.853 0.297 0.368 0.500
terms of Pn, the nominal strength. This difference is simi- 0.835 0.346 0.410 0.551
lar to the two approaches presented by Disque (1973) and 0.816 0.396 0.452 0.601
Yura (1971).
0.796 0.449 0.495 0.649
To allow for additional comparisons, the stiffness reduc-
tion factor using the 1969 AISC Specification, SRF1969, can 0.776 0.502 0.537 0.696
be linked to nominal strength. To accomplish this, Fa /Fy is 0.754 0.555 0.579 0.741
increased by 5/3, removing a uniform factor of safety and in- 0.732 0.608 0.622 0.784
creasing the allowable stress at Kl/r 0 to Fy and the stresses 0.710 0.661 0.663 0.824
at other values of Kl/r to a comparable critical stress level.
0.686 0.712 0.704 0.861
This links Equation 8 to Fcr /Fy or Pn /Py.
Figure 2 shows the three stiffness reduction factors as a 0.662 0.762 0.744 0.895
function of Pn /Py and Table 2 gives values for Pn /Py from 1 0.637 0.808 0.782 0.924
to 0.39. It is clear from these results that the three approach- 0.612 0.852 0.819 0.950
es no longer can be viewed as giving essentially the same 0.586 0.891 0.853 0.971
0.559 0.926 0.886 0.986
0.531 0.955 0.915 0.996
0.503 0.977 0.942 1.000
0.474 0.993 0.964 1.000
1.200
0.443 1.000 0.982 1.000
1.000
Ĳb (T
0.435 1.000 0.986 1.000
0.800 0.414 1.000 0.995 1.000
Factor
0.600
Ĳa (T 0.390 1.000 1.000 1.000
SRF1969 (T
0.400
0.200
0.000 answer. The differences between the factors can be attributed

0.000 0.200 0.400 0.600 0.800 1.000 1.200 to the inelastic column strength equations used. A detailed
comparison of Equations 5 and 14 would show that the 1969
PnPy
ASD equation is the CRC equation with a variable safety
Fig. 2. Comparison of three stiffness reduction factors as a func- factor. The LRFD strength equation, Equation 9, is quite dif-
tion of the nominal strength to yield strength ratio. ferent when compared to both of the others.

A SINGLE STIFFNESS REDUCTION FACTOR REFERENCES
There are three stiffness reduction factors available for use AISC (2005), Specification for Structural Steel Buildings,
with the nomograph as discussed earlier: Equations 8, 12 ANSI/AISC 360, American Institute of Steel Construc-
and 18. If the intent is only to use the stiffness reduction tion, Chicago, IL.
factor to modify the column end stiffness ratios, G, for use ASCE Task Committee on Effective Length (1997), Effec-
with the nomograph, it has been shown that any of these ap- tive Length and Notional Load Approaches for Assessing
proaches will provide satisfactory results. Frame Stability: Implications for American Steel Design,
The SRF1969 and Ua are based on strength equations that American Society of Civil Engineers, New York, NY.
include initial out of straightness and residual stresses while
Disque, R. O. (1973), “Inelastic K-factor for Column
Ub is based on column strength equations that include only
Design,” Engineering Journal, AISC, Vol. 10, No. 2,
the effects of residual stresses. A comparison of these three
pp. 33–35.
approaches, Equations 8, 13 and 24, as a function of Pn /Py,
shows that they do not give similar results. Since the intent of Galambos, T.V. (ed.) (1998), Guide to Stability Design Cri-
the stiffness reduction factor, in all cases, is to include only teria for Metal Structures, Structural Stability Research
the influence of inelastic behavior due to residual stresses, Council, 5th edition, John Wiley and Sons, New York,
clearly Ub should be used. Thus, the stiffness reduction fac- NY.
tor for both the nomograph and the direct analysis method Johnson, B.G., (ed.) (1960), Guide to Design Criteria for
should be taken as Ub. Metal Compression Members, Column Research Council,
Equations 19 and 20 for Ub will be included in the 2010 1st edition, John Wiley and Sons, New York, NY.
AISC Specification as Equations C2-2a and C2-2b for use Johnson, B.G., (ed.) (1966), Guide to Design Criteria for
with the direct analysis method. The effective length meth- Metal Compression Members, Column Research Council,
od is covered in Appendix 7 of the 2010 Specification, and 2nd edition, John Wiley and Sons, New York, NY.
Commentary Section A7.2 indicates that the same equations
Yura, J. A. (1971), “The Effective Length of Columns in
used for stiffness reduction with the direct analysis method
Unbraced Frames,” Engineering Journal, AISC, Vol. 8,
should be used with the effective length method.
No. 2, pp. 37–42.

Bolt Shear Design Considerations
RAYMOND H.R. TIDE
Abstract
In this paper, bolt shear capacities are reviewed using the Load and Resistance Factor Design (LRFD) philosophy. Only bolt-shear limit states
are addressed, although one aspect of slip critical limit states is addressed incidentally. This paper does not consider bolt bearing limit states.
Test data used to justify the adoption of ASTM A325 and A490 high-strength bolts was obtained from previous research programs. The data also
included various types of rivets and Huck bolts for general comparison. First, the test data are used to evaluate the current American Institute
of Steel Construction (AISC, 2005) and Research Council on Structural Connections (RCSC, 2004) bolt shear provisions and to determine the
current reliability, C, which is found to be conservative when based on a resistance factor, G, of 0.75. The appropriateness of the G-factor for bolt
shear is addressed. Canadian (CSA S16-01) and Eurocode (EN 1993) provisions are also evaluated and shown not to be compatible with the test
results. Two design equations are developed—one linear, one a step function—that result in a C value slightly greater than 3.0, appropriate for a
manufactured product. The single-step function (with a step at 38 in.) is recommended for inclusion in updated design specifications. This design
provision increases the design strength by 12.5% for short connections and by 17.2% for long connections. The test data indicate that there is no
need for a bolt strength reduction due to the length of the connection, provided that the connection material gross and net section areas exceed
certain ratios. That ratio is a function of the connection material yield and tensile strength, the total bolt shear area and the bolt tensile strength.
Keywords: bolt shear, reliability, resistance factor, connection length factor.
BACKGROUND The design values are based on an extensive research pro-

gram conducted by the steel industry at the Fritz Engineer-
The current shear strength of a high strength bolt may be
ing Laboratory at Lehigh University from the 1950s through
expressed by the following equation:
the early 1970s. As was the custom at the time, the high-
Pn Pu Ab R1 R2 R3 R4 (1) strength bolts were fully pretensioned. Bolt threads were ex-
cluded from the shear plane. In addition, an earlier research
where
investigation at the University of Illinois and at University
Pu ultimate tensile strength of bolt, ksi
of California by Davis et al. (1940) was reviewed concern-
R1 0.625, shear-to-tension ratio
ing riveted connections for the San Francisco–Oakland Bay
R2 0.80, connection length reduction factor for
Bridge. All of the data was summarized in the Guide to De-
L ≤ 50 in.
sign Criteria of Bolted and Riveted Joints (the Guide) by
R3 1.00 if threads are excluded from the shear plane
Kulak et al. (1987). The roles of 12 basic variable groups
0.80 if threads are included in the shear plane
resulted in approximately 45 test variables that are described
R4 0.80, additional connection length reduction factor
in the Guide and will not be repeated in this paper. However,
for L > 50 in.
three of the basic variable groups will be subsequently ex-
L connection length, in.
amined and used to develop a proposed design procedure.
Ab nominal unthreaded body area of bolt, in.2
The types of connections tested were the basic lap splice,
The design shear values for ASTM A325 and A490 bolts the butt splice, the open shingle splice, and the closed shin-
are given in RCSC Specification Table 5.1 (RCSC, 2004). gle splice, as shown in Figure 1. A review of the literature re-
The design values, for other fasteners, such as ASTM A307 vealed that the data from each test series were not uniformly
bolts and threaded material, are given in AISC Specification reported. As a result, the original research reports were used
for Structural Steel Buildings (hereafter AISC Specification, to augment the background data. The connection length for
2005) Table J3.2. In Load Resistance and Factor Design a lap splice is the distance between the centerlines of the
(LRFD( terms, the design shear strength of a bolt is GRn, extreme end bolts. The connection length for a butt splice is
with G 0.75 and Rn Pn. the distance from the centerline of the bolt at one end of the
connection to the centerline of the bolt closest to the overall
connection centerline. Fortunately, large quantities of bolts
were obtained from production lots, so essentially identical
bolts were used in several test programs. Each lot of bolts
Raymond H.R. Tide, D.Sc, S.E., P.E., Principal, Wiss, Janney, Elstner Associ-
ates, Inc., 330 Pfingsten Rd., Northbrook, IL 60062. E-mail: [email protected].
was tested to determine both the tensile and shear strength.

A literature search identified 11 papers and reports in ad- TEST DATA
dition to several supporting reports that resulted in data from
As previously indicated, test data were obtained from 11 pa-
119 tests. Because of incomplete background information,
pers and reports: Bendigo et al. (1963), Davis et al. (1940),
the 40 tests from Davis et al. (1940) were not used other
Fisher et al. (1963), Fisher and Kulak (1968), Fisher and
than to document its test results. Of the remaining 79 tests,
Yoshida (1970), Foreman and Rumpf (1961), Kulak and
the breakdown is as follows: 54 used ASTM A325 bolts,
Fisher (1968), Power and Fisher (1972), Rivera and Fisher
18 used ASTM A490 bolts, 5 used rivets, and 2 used Huck
(1970), and Sterling and Fisher (1965, 1966). Because of
bolts.
the various reporting formats and test parameters, it was not
The CSA S16-01 standard, as supplemented in 2005,
possible to directly compare the results. The published test
adopted the 2003 Eurocode EN 1993, Clause 3.8, which is
ultimate shear strength of each connection was reduced to
shown as Equation 2, converted to U.S. customary units:
an average ultimate shear strength, PTEST, of a single connec-
VR 0.60GFu Ab (2) tor, bolt or rivet, loaded on two shear planes (double shear).
The predicted ultimate shear strength of the same connector
where
was computed using the reported appropriate single shear
G 0.80, resistance factor
connector test data times two, PPRED, for each lot of bolts or
Fu nominal bolt tensile stress, ksi
rivets.
Ab bolt area, in.2
The ratio PTEST /PPRED was then computed to compare the
The R1 factor of 0.60 represents the shear-to-tension ratio results, with connection length as the only independent vari-
used in the CSA S16 document. When the threads are in- able. All of the reconfigured test data are given in Table 1
cluded in the shear plane, an R3 factor of 0.70 is used. Of and plotted in Figure 2. The solid line represents the current
greater significance, VR is valid up to a connection length, step function for the length reduction factor. The dotted line
L, of 15d, where d is the bolt diameter in inches. When the represents the same equation multiplied by the current AISC
connection length exceeds 15d, VR is reduced by the factor G for bolts, 0.75. The plotted data are in a non-dimensional
(1.075 – 0.005L/d) but is not taken as less than 0.75 times the form, eliminating the independent variables of bolt diameter,
original value given in Equation 2. material type and connection configuration. Although this
Lap Splice
Butt Splice
Closed Shingle Splice
Open Shingle Splice
Fig. 1. Four types of connections.

Table 1. Connection Test and Computation Data
Bolts 0.90 As Fv 0.90 As Fv
Dia. in L PTEST PPRED PTEST Fyp Fup Fv As Ag Fyp An Fup (1) (2) Ref.
No. ID (in.) Type Line (in.) (kip) (kip) PPRED (ksi) (ksi) (ksi) (in.2) An/As (in.2) (in.2) (in.2) (in.2) R2 Mat. No.
1 F42b 1- A325 4 10.5 131.5 131.4 1.001 114.0 121.4 66.1 15.9 0.51 13.0 8.3 8.07 7.8 0.90C A514 4
2 F42c 1- A325 4 10.5 133.0 131.4 1.012 114.0 121.4 66.1 15.9 0.56 13.8 8.3 8.9 7.8 0.90C A514 4
3 F42d 1- A325 4 10.5 132.0 131.4 1.005 114.0 121.4 66.1 15.9 0.61 14.5 8.3 9.66 7.8 0.90C A514 4
4 F42e 1- A325 4 10.5 132.8 131.4 1.01 114.0 121.4 66.1 15.9 0.66 15.4 8.3 10.5 7.8 0.90C A514 4
5 F42g 1- A325 4 10.5 134.3 131.4 1.022 114.0 121.4 66.1 15.9 0.72 16.3 8.3 11.4 7.8 0.90C A514 4
6 F111 1- A325 11 35.0 167.8 131.4 1.036 114.0 119.8 81.5 21.8 0.74 18.9 14.0 15.5 13.3 0.90C A514 4
7a HJ131 Y A325 13 42.0 85.7 113.0 0.758 36.5 66.1 94.0 15.6 1.13 21.1 36.2 17.6 20.0 0.75S A36M 4
7b 44.1 76.3 94.0 15.6 0.91 16.9 29.9 14.1 17.3 0.75S A440L 4
8A HJ132 Y A325 13 42.0 69.8 113.0 0.618 36.5 66.1 94.0 15.6 0.89 16.3 36.2 13.9 20.0 0.75S A36M 4
8b 44.1 76.3 94.0 15.6 0.71 13.0 29.9 11.1 17.3 0.75S A440L 4
9a HJ133 Y A490 13 42.0 116.8 115.3 1.013 44.1 76.3 95.9 15.6 1.92 33.6 30.5 29.8 17.6 0.90C A440M 4
9b 114.0 121.4 95.9 15.6 0.90 15.8 11.8 14.0 11.1 0.90C A514L 4
10a HJ135 Y A325 13 42.0 95.7 96.9 0.988 44.1 76.3 80.6 15.6 1.68 29.8 25.7 26.1 14.8 0.90C A440M 4
10b 114.0 121.4 80.6 15.6 0.79 14.1 9.9 12.3 9.3 0.90C A514L 4
11 J42a 1 A490 4 10.5 154.8 151.7 1.020 114.0 121.4 96.6 12.6 0.76 13.9 9.6 9.58 9.0 0.90C A514 7
12 J42b 1 A490 4 10.5 153.5 151.7 1.012 114.0 121.4 96.6 12.6 0.82 14.6 9.6 10.3 9.0 0.90C A514 7
13 J42c 1 A490 4 10.5 150.8 151.7 0.994 114.0 121.4 96.6 12.6 0.86 15.2 9.6 10.9 9.0 0.90C A514 7
14 J42d 1 A490 4 10.5 152.5 151.7 1.006 114.0 121.4 96.6 12.6 0.92 16.0 9.6 11.6 9.0 0.90C A514 7
15 J072 Y A490 7 21.0 121.4 116.6 1.041 101.6 111.9 97.0 8.41 0.91 9.58 7.2 7.66 6.6 0.90C A514 7
16 J132 1- A490 13 42.0 201.2 191.8 1.049 101.6 111.9 96.5 25.8 0.91 28.6 22.1 23.7 20.0 0.90C A514 7
17 J172 Y A490 17 56.0 118.5 119.8 1.016 101.6 111.9 97.0 20.4 0.90 20.4 17.5 18.5 15.9 0.90C A514 7
18 J251 Y A490 25 84.0 109.4 119.8 0.913 101.6 111.9 99.7 30.1 0.82 28.4 26.6 24.6 24.1 0.90C A514 7
19 J252 Y A490 25 84.0 124.0 119.8 1.035 101.6 111.9 99.7 30.1 1.12 37.6 26.6 33.7 24.1 0.90C A514 7
20 D71 Y A325 7 21.0 80.4 102.5 0.784 28.2 60.0 85.3 16.8 1.08 21.9 45.7 18.2 21.5 0.75S A7 1
21 D81 Y A325 8 24.5 80.1 102.5 0.782 28.2 60.0 85.3 19.2 1.09 24.7 52.3 21.0 24.6 0.75S A7 1
22 D91 Y A325 9 28.0 75.4 102.5 0.736 28.2 60.0 85.3 21.6 1.10 27.5 58.8 23.7 27.6 0.75S A7 1
23 D101 Y A325 10 31.5 75.3 102.5 0.734 28.2 60.0 85.3 24.0 1.08 29.8 65.3 26.1 30.7 0.75S A7 1
24 D701 Y A325 7 21.0 86.6 112.0 0.773 33.6 64.3 93.2 16.8 1.14 23.8 41.9 19.3 21.9 0.75S A7 1
Notes:
(1) 0.90 indicates data represented by a circle (C) in Figure 3;
0.75 indicates data represented by a square (S) in Figure 3;
0.75* indicates data represented by a triangle (T) in Figure 3.
(2) L and M refer to lap plates and main member plates in hybrid connections.

Table 1. Connection Test and Computation Data (Cont.)
25 D801 Y A325 8 24.5 82.1 112.0 0.733 33.6 64.3 93.2 19.2 1.10 26.0 47.9 21.1 25.0 0.75S A7 1
26 D901 Y A325 9 28.0 83.2 112.0 0.742 33.6 64.3 93.2 21.6 1.12 30.0 53.9 24.3 28.2 0.75S A7 1
27 D1001 Y A325 10 31.5 83.4 112.0 0.744 33.6 64.3 93.2 24.0 1.12 33.1 59.9 26.8 31.3 0.75S A7 1
28 D10 Y A325 10 31.5 77.2 109.0 0.708 28.2 60.0 90.7 24.0 1.09 33.7 69.5 26.2 32.7 0.75S A7 1
29 D13A Y A325 13 31.5 76.5 109.0 0.701 28.2 60.0 90.7 31.3 1.09 41.7 90.6 34.2 42.6 0.75S A7 1
30 D13 Y A325 13 42.0 71.3 109.0 0.654 28.2 60.0 90.7 31.3 1.09 41.7 90.6 34.2 42.6 0.75S A7 1
31 D16 Y A325 16 52.5 65.2 109.0 0.598 28.2 60.0 90.7 38.5 1.09 49.4 111.4 41.9 52.4 0.75S A7 1
32 L2 Y A325 2 3.5 98.5 100.2 0.983 28.2 60.0 83.4 2.40 1.65 5.76 6.4 3.96 3.0 0.75*T A7 1
33 L5 Y A325 5 14.0 89.2 100.2 0.89 28.2 60.0 83.4 6.01 1.32 9.63 16.0 7.92 7.5 0.75*T A7 1
34 L7 Y A325 7 21.0 91.4 100.2 0.912 28.2 60.0 83.4 8.41 1.57 14.9 22.4 13.2 10.5 0.75*T A7 1
35 L10 Y A325 10 31.5 74.8 100.2 0.746 28.2 60.0 83.4 12.0 1.10 14.9 31.9 13.2 15.0 0.75S A7 1

36 DR71 Y A141R 7 21.0 52.7 66.5 0.793 28.2 60.0 55.3 16.8 0.785 16.9 29.7 13.2 13.9 0.75S A7 1
37 DR101 Y A141R 10 31.5 47.1 66.5 0.709 28.2 60.0 55.3 24.0 0.769 22.2 42.4 18.5 19.9 0.75S A7 1
38 DR131 Y A141R 13 42.0 46.8 66.5 0.704 28.2 60.0 55.3 31.3 0.762 27.5 55.2 23.8 26.0 0.75S A7 1
39 B3 Y A325 4 10.5 87.5 96.2 0.910 36.6 65.5 80.0 24.0 1.11 36.0 47.2 26.6 26.4 0.75*T A7 6
40 B6 Y A325 3 7.0 86.1 96.2 0.896 36.6 65.5 80.0 21.6 1.15 36.0 42.5 24.8 23.7 0.75*T A7 6
41 BR2 Y A141R 5 14.0 52.0 60.0 0.867 36.6 65.5 49.9 30.0 0.887 36.0 36.8 26.6 20.6 0.75*T A7 6
42 A3 1 A325 4 12.0 113.8 125.6 0.906 36.6 65.5 80.0 25.1 1.10 36.0 49.4 27.5 27.6 0.75S A7 6
43 G1 1- A325 3 8.0 149.8 167.0 0.897 36.6 65.5 84.0 23.9 1.11 36.0 49.4 26.5 27.6 0.75S A7 6
44 B5 1- A325 5 14.0 84.0 96.2 0.874 36.6 65.5 80.0 24.0 1.11 36.0 47.2 26.6 26.4 0.75*T A7 6
45 K42a Y A490 4 10.5 122.5 124.4 0.985 43.0 76.0 103.5 9.62 1.22 15.5 20.8 11.7 11.8 0.75S A440 10
46 K42b Y A490 4 10.5 122.5 124.4 0.985 43.0 76.0 103.5 9.62 1.27 16.0 20.8 12.2 11.8 0.75*T A440 10
47 K42c Y A490 4 10.5 124.5 124.4 1.001 43.0 76.0 103.5 9.62 1.31 16.4 20.8 12.6 11.8 0.75*T A440 10
48 K42d Y A490 4 10.5 125.5 124.4 1.009 43.0 76.0 103.5 9.62 1.37 17.0 20.8 13.2 11.8 0.75*T A440 10
49 K131 Y A490 13 31.6 109.6 121.5 0.902 43.0 76.0 101.1 15.6 1.30 24.2 33.0 20.4 18.7 0.75*T A440 10
50 K132 Y A490 13 63.0 100.9 121.5 0.830 43.0 76.0 101.1 15.6 1.30 24.1 33.0 20.3 18.7 0.75*T A440 10
51 K133 Y A490 13 63.0 127.7 121.5 1.051 43.0 76.0 101.1 15.6 1.92 33.8 33.0 30.0 18.7 0.90C A440 10
52 K191 Y A490 19 63.0 94.4 121.5 0.777 43.0 76.0 101.1 22.8 1.30 33.4 48.2 29.7 27.3 0.75*T A440 10
Notes:
Table 1. Connection Test and Computation Data (Cont.)
53 E41b Y A325 4 10.5 94.3 101.4 0.929 45.3 75.8 84.4 9.62 0.947 12.9 16.1 9.11 9.6 0.75S A440 3
54 E41c Y A325 4 10.5 96.3 101.4 0.949 45.3 75.8 84.4 9.62 0.981 13.3 16.1 9.49 9.6 0.75S A440 3
55 E41e Y A325 4 10.5 97.8 101.4 0.964 45.3 75.8 84.4 9.62 1.11 14.4 16.1 10.7 9.6 0.75*T A440 3
56 E41f Y A325 4 10.5 90.9 92.4 0.983 45.3 75.8 76.9 9.62 0.996 13.3 14.7 9.58 8.8 0.75*T A440 3
57 E41g Y A325 4 10.5 95.9 98.6 0.973 45.3 75.8 82.0 9.62 1.00 13.4 15.7 9.66 9.4 0.75*T A440 3
58 E41 Y A325 4 10.5 91.0 92.4 0.985 45.3 75.8 76.9 9.62 1.01 13.5 14.7 9.70 8.8 0.75*T A440 3
59 E71 Y A325 7 21.0 84.9 92.4 0.918 45.3 75.8 76.9 16.8 0.999 20.6 25.7 16.8 15.3 0.75*T A440 3
60 E101 Y A325 10 31.5 80.5 92.4 0.871 45.3 75.8 76.9 24.0 1.00 27.8 36.7 24.0 21.9 0.75*T A440 3
61 E131 Y A325 13 42.0 81.7 95.2 0.859 45.3 75.8 79.2 31.3 0.994 38.5 49.3 31.1 29.4 0.75*T A440 3
62 E161 Y A325 16 52.5 79.5 95.2 0.835 45.3 75.8 79.2 38.5 0.994 45.7 60.6 38.2 36.2 0.75*T A440 3
63 E46 Y A325 4 10.5 90.8 92.4 0.983 45.3 75.8 76.9 28.9 1.01 40.6 44.2 29.2 26.4 0.75*T A440 3
64 E74 Y A325 7 21.0 86.1 92.4 0.931 45.3 75.8 76.9 33.7 0.995 41.0 51.5 33.5 30.8 0.75*T A440 3
65 E741 Y A325 7 21.0 80.4 92.4 0.869 45.3 75.8 76.9 33.7 1.00 41.2 51.5 33.7 30.8 0.75*T A440 3
A502
66 Rivet Y 32 94.0 43.8 54.1 0.809 59.0 88.0 45.0 76.9 0.660 57.5 52.8 50.8 35.4 0.90C A572 9
Gr 1
67 Bolt Y A325 32 94.0 55.5 91.4 0.607 59.0 88.0 76.0 76.9 0.660 57.5 89.2 50.8 59.8 0.75S A572 9
68 1 Y A325 18 52.0 67.2 105.9 0.635 49.0 79.0 88.1 14.4 0.750 12.8 23.3 10.9 14.5 0.75S A572 8
69 2 Y Huck 12 34.0 81.8 110.0 0.744 49.0 79.0 91.5 10.8 1.00 12.8 18.2 10.9 11.3 0.75S A572 8
70 3A Y A325 12 33.5 85.8 105.9 0.810 49.0 79.0 88.1 10.8 1.00 12.8 17.5 10.9 10.8 0.75*T A572 8
71 3B Y Huck 12 33.5 87.0 110.0 0.791 49.0 79.0 91.5 10.8 1.00 12.8 18.2 10.9 11.3 0.75S A572 8
72 5 Y A325 18 52.0 63.4 105.9 0.599 49.0 79.0 88.1 16.8 0.646 12.8 27.2 10.9 16.9 0.75S A572 8
73 6 Y A325 12 33.0 95.8 107.6 0.891 49.0 79.0 89.5 14.4 1.13 20.0 23.7 16.3 14.7 0.75*T A572 8
74 7 Y A325 32 94.0 30.1 47.0 0.640 49.0 79.0 78.2 25.2 0.322 10.0 36.2 8.13 22.5 0.75S A572 8
75 8 Y A325 21 61.0 44.4 47.0 0.944 49.0 79.0 78.2 18.6 0.436 10.0 26.7 8.13 16.6 0.75S A572 8
76 E721 Y A325 7 21.0 76.4 92.6 0.826 45.3 76.0 77.0 16.8 0.805 34.4 25.7 13.6 15.3 0.75*T A440 11
77 E722 Y A325 7 21.0 90.7 92.6 0.980 45.3 76.0 77.0 16.8 1.21 48.1 25.7 20.4 15.3 0.90C A440 11
78 E163 Y A325 16 52.5 68.1 95.0 0.717 45.3 76.0 79.0 38.5 0.805 76.8 60.4 31.0 36.0 0.75*T A440 11
79 E164 Y A325 16 52.5 87.0 95.0 0.917 45.3 76.0 79.0 38.5 1.20 107.4 60.4 46.2 36.0 0.90C A440 11
Notes:

Table 2. Test Data for San Francisco–Oakland Bay Bridge Connections
Dia. Rivets L PTEST PPRED PTEST Fv Ref.
No. ID (in.) Type* in Line (in.) (kip) (kip) PPRED (ksi) An /As No.
80 CCC 7-1 1 R(c) 8 6 89.0 89.2 0.998 56.8 1.19 2
81 CCC 7-2 1 R(c) 8 6 90.0 89.2 1.009 56.8 1.19 2
82 DCC 7-1 1 R(c) 14 4 89.1 89.2 0.999 56.8 0.771 2
83 DCC 7-2 1 R(c) 14 4 88.9 89.2 0.997 56.8 0.771 2
84 ACM 12-1 1 R(m) 12 15.75 121.7 119.0 1.023 75.8 1.58 2
85 ACM 12-2 1 R(m) 12 15.75 116.5 119.0 0.979 75.8 1.58 2
86 ASM 12-1 1 R(m) 12 15.75 113.3 119.0 0.952 75.8 1.22 2
87 ASM 12-2 1 R(m) 12 15.75 113.8 119.0 0.956 75.8 1.22 2
88 ACC 18-1 1 R(c) 18 24.50 87.0 89.2 0.975 56.8 1.06 2
89 ACC 18-2 1 R(c) 18 24.50 87.8 89.2 0.984 56.8 1.06 2
90 ACC 36-1 1 R(c) 36 24.50 79.5 89.2 0.891 56.8 0.528 2
91 ACC 36-2 1 R(c) 36 24.50 81.1 89.2 0.909 56.8 0.528 2
92 ACC 54-1 1 R(c) 54 40.75 77.1 89.2 0.864 56.8 0.352 2
93 ACC 54-2 1 R(c) 54 40.75 79.9 89.2 0.896 56.8 0.352 2
94 ASC 18-1 1 R(m) 18 19.5 86.7 119.0 0.728 75.8 0.812 2
95 ASC 18-2 1 R(m) 18 19.5 85.4 119.0 0.718 75.8 0.812 2
96 ASC 36-1 1 R(m) 36 24.5 84.9 119.0 0.714 75.8 0.406 2
97 ASC 36-2 1 R(m) 36 24.5 84.4 119.0 0.710 75.8 0.406 2
98 ASC 54-1 1 R(m) 54 31.25 80.3 119.0 0.675 75.8 0.271 2
99 ASC 54-2 1 R(m) 54 31.25 80.1 119.0 0.673 75.8 0.271 2
100 ACM 24-1 1 R(m) 24 35 119.8 119.0 1.006 75.8 0.791 2
101 ACM 24-2 1 R(m) 24 35 123.6 119.0 1.039 75.8 0.791 2
102 ACM 36-1 1 R(m) 36 54.25 108.2 119.0 0.909 75.8 0.528 2
103 ACM 36-2 1 R(m) 36 54.25 115.8 119.0 0.973 75.8 0.528 2
104 ASM 24-1 1 R(m) 24 35 118.1 119.0 0.992 75.8 0.609 2
105 ASM 24-2 1 R(m) 24 35 126.8 119.0 1.066 75.8 0.609 2
106 ASM 36-1 1 R(m) 36 54.25 115.3 119.0 0.969 75.8 0.406 2
107 ASM 36-2 1 R(m) 36 54.25 112.9 119.0 0.949 75.8 0.406 2
108 ANM 12-1 1 R(m) 12 15.75 118.8 119.0 0.999 75.8 1.053 2
109 ANM 12-2 1 R(m) 12 15.75 115.8 119.0 0.973 75.8 1.053 2
110 ANM 24-1 1 R(m) 24 35 124.8 119.0 1.049 75.8 0.526 2
111 ANM 24-2 1 R(m) 24 35 119.2 119.0 1.001 75.8 0.526 2
112 ANM 36-1 1 R(m) 36 54.25 118.1 119.0 0.992 75.8 0.351 2
113 ANM 36-2 1 R(m) 36 54.25 119.7 119.0 1.006 75.8 0.351 2
114 BCC 20a-1 1 R(c) 22 15 86.6 89.2 0.971 56.8 1.031 2
115 BCC 20a-2 1 R(c) 22 15 86.6 89.2 0.971 56.8 1.031 2
116 BCC 20b-1 1 R(c) 22 22.5 91.2 89.2 1.022 56.8 1.031 2
117 BCC 20b-2 1 R(c) 22 22.5 93.0 89.2 1.043 56.8 1.031 2
118 BCC 20c-1 1 R(c) 22 30 90.6 89.2 1.016 56.8 1.031 2
119 BCC 20c-2 1 R(c) 22 30 89.0 89.2 0.998 56.8 1.031 2
Notes:
* R(c) Carbon rivet; R(m) Manganese rivet

did not completely eliminate the bending effect of the lap- An /As in some form was reported in each document. The test
and open-shingle splice, the quantity of these tests is small results, as clearly reported in the Guide, demonstrated that
compared to the quantity of butt-splice connection tests, so as An /As increased, the connection performance PTEST /PPRED
that their overall effect is very limited. The available data for also improved.
the 40 tests conducted by Davis et al. (1940) are in Table 2. Recently, Moore et al. (2008) reported results that in-
As expected, there are a few data points that are randomly cluded tests on 1,533 high-strength bolts. The program in-
scattered throughout the plot. Because 22 data points are cluded ASTM A325, A490, F1852 and F2280 bolts. The
concentrated at 10.5 in., they have been distributed to 9.5 in., latter two are nominally referred to as tension control bolts
10.5 in. and 11.5 in. for clarity. The bottom of the vertical and are comparable to A325 and A490 bolts, respectively.
scale is also truncated to spread out the data. The program reported on both threads included in the shear
Figure 3 shows a plot of the test data, where the test data plane as well as threads excluded from the shear plane. Ten-
are identified relative to the connection’s strength and quasi- sion tests were also performed to calibrate the various lots
stiffness characteristics (developed later in this paper). A of bolts. Compared to the earlier Lehigh tests (e.g., Fisher
review of Figure 3 indicates that the test results fall into et al., 1963), these bolts were tested in the snug-tight con-
groupings that suggest different design criteria for differ- dition and not fully pretensioned. The results indicate that
ent connection lengths—as indicated by the earlier AISC manufactured bolts have reasonably uniform properties as
and RCSC step function, which is shown in Figure 2. It ap- compared to an assembled connection and therefore warrant
pears that there is a band of data above the 0.90PTEST /PPRED consideration of a lower target reliability, C.
level that extends across the full range of connection lengths. Based on the data for bolt shear with threads excluded, a C
However, there is another group of data that slopes down- of 4.0 was obtained by Moore et al. (2008) for a live to dead
wards between approximately 15 in. and approximately load ratio of 3.0. A resistance factor, G, of 0.85 was obtained
40 in. After 40 in., the boundary line is a minimum of ap- for the same condition. Thus, the current AISC/RCSC G of
proximately 0.60PTEST /PPRED. 0.75 appears to be conservative. This observation was con-
The earlier research identified the connection net section, sidered, along with other factors, when proposed revisions to
An, as a significant variable. Similarly, the total area, As, of the AISC and RCSC provisions were developed.
all the bolt shear planes was also found significant. A ratio of
Fig. 2. Results of 119 connection tests with current design criteria superimposed.

DESIGN CRITERIA Equation 3 represents, in non-dimensional form, the net
section of a connection, and Equation 4 represents, in non-
The test data that are plotted above the 0.90PTEST /PPRED level, dimensional form, the gross section of the connection. Equa-
in Figure 3, indicate that under certain conditions there is tion 3 can be considered to represent a strength relationship.
no reduction in connection capacity regardless of the con- Similarly, Equation 4 can be considered to represent a quasi-
nection length. As previously noted, early research showed stiffness concept, because as the ratio increases, the plates
that as the ratio An /As increased, connection capacity also essentially remain elastic as the ultimate shear strength of the
increased. A review of the test data indicated that there was a bolt is reached. These are not unfamiliar concepts, because
better correlation when the following ratios were compared checking the net and gross sections of a connection has been
to the PTEST /PPRED ratio: part of AISC specifications for years. The numerical values
An Fup of N1 and N2 must be chosen to satisfy both the test data and
= N1 (3)
As Fv Chapter D of the AISC Specification.
For design purposes, it is more appropriate to rearrange
and Equations 3 and 4, substituting nominal values for ultimate
Ag Fyp values, e.g., replacing Fv with R1Fu , taking R1 as 0.625. The
= N2 (4)
As Fv procedure is shown in Appendix A, which evaluates the net
and gross section requirements of Chapter D of the AISC
where Specification, taking the length reduction factor, R2, as 0.90.
Ag gross area of connection material, in.2 The computed values for N1 and N2 are determined to be 0.56
An net area of connection material, in.2 and 0.47, respectively. Coincidentally, the 0.56 value for N1
As total bolt area in shear plane, in.2 is equivalent to a PTEST /PPRED ratio of 0.90 (shown by the
Fup nominal tensile strength of connection material, horizontal line in Figure 3).
ksi The general forms of the design equations in Appendix
Fyp yield stress of connection material, ksi A follow as Equations 5 and 6. Because of the uncertain-
Fv ultimate shear strength of the bolt, ksi ties associated with bolt installation (pretensioned versus
N1, N2 target ratios, selected considering test data or snug-tight), second-order effects, and the dictated resistance
specification criteria
Fig. 3. Test results (79) identified by strength and quasi-stiffness criteria.

factors, N2 was set equal to N1 at a common value of 0.56. DESIGN EQUATIONS
Solving for the net and gross areas so that the predicted
A practical approach must be chosen to satisfy the needs of
PTEST /PPRED ratio will exceed 0.90 yields:
design office, detailer and fabrication requirements. For shear
An ≥ 0.56AsFu /Fup (5) connections with lengths less than approximately 15.5 in.,
a basic reduction factor R2 of 0.90 is recommended to ac-
and count for variability in connection behavior. This is an in-
Ag ≥ 0.56AsFu /Fyp (6) crease from the current basic reduction factor value of 0.80,
resulting in a 12.5% increase in bolt capacity from current
where Fu is the nominal tensile strength of the bolt in ksi. methods for “short” connections. The resistance factor, G, of
The 0.56 factor results from the product of 0.625 and 0.90 0.75 is still appropriate.
(R1 × R2) rounded to two significant figures. The increase in R2 is considered appropriate because all
The values for An and Ag, as well as the computed values of the tests were uni-axial, whereas actual connections typi-
for Equations 5 and 6, are shown in Table 1 for the test data. cally have a nominal bi-axial contribution. Finite element
Because the shear strength of the bolts was established in studies have demonstrated this effect for shear and bending
the research reports, there was no need to convert the bolt at the end of both simply-supported and fixed-end beams.
tensile strength to the bolt shear strength. When the data Similarly, the connections at the ends of diagonals in long
were tabulated, three conditions were identified. The first span trusses, although designed with pin ends, actually have
condition was when both net and gross area, An and Ag, ex- some bending due to transverse differential displacements at
ceeded the respective inequalities shown in Equations 5 or their ends as the truss deflects under load.
6, respectively. These PTEST /PPRED data are shown as circles For connection lengths greater than 15.5 in. but less than
in Figure 3. The second condition was when only one of the 28.8 in., R2 could be taken as a function of connection length
two inequalities was exceeded. Typically it was the net area, as follows:
An, and these are shown with triangles in Figure 3. The third R2 1.075 – 0.0113L (7)
condition was when neither inequality was satisfied, and
these are shown as squares in Figure 3. The letters C, T and R2 is limited to a minimum of 0.75. Beyond approximately
S are used in to R2 column and footnote of Table 1 to identify 28.8 in. there is a constant strength reduction, R2, of 0.75.
the shape of the data point in Figure 3. With the application of G 0.75, the overall bolt design value
With the exception of the one test at 94 in., all of the test is less than all of the test data considered, whether bolt, rivet
results out of the 21 shown as circles satisfied both inequali- or Huck connector. Connection lengths greater than 28.8 in.
ties and had PTEST /PPRED ratios greater than 0.90. This one result in a nominal bolt strength increase of 17.2% compared
test, No. 66, with a PTEST /PPRED ratio of 0.809, was for an to current practice, because the length reduction factor would
ASTM A502 Grade 1 rivet with ASTM A572 connection increase from 0.64 to 0.75.
plates. The other 20 tests had PTEST /PPRED ratios that varied With the foregoing design criteria in mind, and observing
from 0.913 to 1.051. the distribution of the test data in Figures 2 and 3, a simpli-
The next group of bolts consists of 28 test results where fied design criterion was chosen. An initial straight line with
only one inequality was satisfied, shown as triangles. Only a constant R2 of 0.90 extending to 38 in. was chosen. Next
26 test data show up in Figure 3 because in two cases, both is a step function that drops to 0.75. Thereafter, R2 remains
at a connection length of 10.5 in., the test results were essen- constant at 0.75 for connection lengths greater than 38 in.
tially identical. Twenty-three of these data had connection These design equations are shown graphically in Figure 4,
lengths less than 38 in. and 14 were less than 14 in. For the where they are superimposed on the data of Figure 2. In both
short connection lengths the test data fall above or very close cases G remains at 0.75. The proposed design criteria are
to the 0.90PTEST /PPRED line. compatible with the theoretical results shown in Figure 5.18
There are 30 test results shown as a square in Figure 3, of the Guide (Kulak et al., 1987).
where neither inequality was satisfied. Only 29 squares are The 38-in. length was arbitrarily chosen because it was
evident because there is a duplicate at 31.5 in. The one test not a multiple of any of the standard bolt gage spacings and
data at 61 in. for an enclosed shingle connection with a PTEST / was less than 42 in. The 42-in. length represents the begin-
PPRED ratio of 0.944 is an anomaly. A review of the original ning of the lower plateau for test results. Once again the
research data did not identify any obvious inconsistency. The resulting final design criterion is less than the least of the
square and triangular data indicate that between approxi- test data. Using the proposed criteria, a revision to the bolt
mately 21 in. and 42 in. there is a transition in connection shear portion of AISC Specification Table J3.2, is presented
behavior depending on the material properties and plate area in Table 3. Similar revisions to RCSC Table 5.1 for ASTM
(Ag, An) proportions relative to the total bolt shear area, As. A325 and ASTM A490 bolts are appropriate.

Table 3. Proposed Bolt Shear Revisions to AISC Table J3.2
Nominal Shear Stress in Bearing-Type Connection, Fnv (ksi)
Bolt Type Less than 38 in. 38 in. and greater
A307 27 23
A325 threads included 54 45
A325 threads excluded 68 56
A490 threads included 68 56
A490 threads excluded 84 70
Threaded rods threads included 0.45Fu 0.375Fu
Threaded rods threads excluded 0.563Fu 0.469Fu
The historical tests were performed on fully tightened average value of 1.43 and a standard deviation of 0.24.
high-strength bolts with hardened washers using the turn-of- Similar results for the less critical gross area represented by
nut method. A high degree of slip resistance (friction) was Equation 6 yield an average value of 1.32 and a standard
achieved. The effect of pretensioned bolts is demonstrated deviation of 0.26. With a connection frictional component of
by examining Figure 3. The eight test data identified by a approximately 30%, these relationships would indicate that
circle with connection length greater than 38 in. and above all the bolt shear strength would be fully engaged at the con-
the 0.90 horizontal line indicate that the shear strength of nection’s ultimate load.
all bolts was reached. In Table 4 the ratio of connection With the use of snug-tight bolts there is effectively no fric-
net area divided by Equation 5 is greater than 1.0 with an tional component to the connection capacity. Regardless, the
Fig. 4. Proposed design criteria superimposed on 119 connection test results.

Table 4. Ratio of Ag and An to Proposed Design Criteria
Test Length PTEST Ag Eq. 6 Ag An Eq. 5 An
No. (in.) PPRED (in.) (in.2) Eq. 6 (in.2) (in.2) Eq. 5
9 42 1.013 33.6 30.5 1.10 29.8 17.6 1.69
10 42 0.988 29.8 25.7 1.16 26.1 14.8 1.76
16 42 1.049 28.6 19.7 1.45 23.7 18.5 1.28
79 52.5 0.917 107.4 60.4 1.78 46.2 36.0 1.28
17 56 0.989 20.4 16.1 1.27 18.5 15.1 1.23
51 63 1.051 33.8 33.0 1.02 30.0 18.7 1.60
18 84 0.913 28.4 23.7 1.20 24.6 22.2 1.11
19 84 1.035 37.6 23.7 1.59 33.7 22.2 1.52
Average 1.32 1.43
Standard Deviation 0.26 0.24
test results can still be used for snug-tight bolted connec- Bolt Shear Design Sequence
tions because the frictional component is offset by using a
1. Determine design load, P.
reduced G of 0.75 from 0.85 (Moore et al., 2008), a reduc-
tion of 13%; and by limiting the connection length reduction 2. Initially assume maximum bolt capacity and L c 38 in.
factor, R2, to 0.90 from 1.0 (the single-bolt connection case), Select ASTM A325 or A490 bolts, bolt diameter,
a reduction of 11%. In addition, the coefficient N2 in Equa- thread condition (included or excluded), and single or
tion 6 was increased from 0.47 to 0.56 (19%), reducing the double shear to obtain Vn.
stress on the plates and more uniformly distributing the force
to the bolts. As reported in the literature (Fisher and Kulak, 3. Determine number of bolts by dividing P by Vn.
1968), the bolts were ordered and supplied near the low end 4. Calculate As, considering thread condition (included
of the applicable ASTM standard. In comparison, it is likely or excluded) and single or double shear.
that the average production bolt will have a slightly higher
ultimate strength. All of these factors would justify not hav- 5. Choose a bolt pattern and determine the connection
ing a length reduction factor less than 0.90 for connections length, L.
exceeding 38 in. The coefficients could be fine tuned by per- 6. If L c 38 in., the design is complete for bolt shear. If
forming a limited number of tests. However, the proposed L 38 in., continue.
step function is conservative.
The proposed design criteria do not require any apprecia- 7. Compute Ag and An.
ble difference in design methodology from current methods.
8. Check Equations 5 and 6 (L > 38 in.):
The only new item is that the total bolt shear area, As, has
Equation 5: An ≥ 0.56AsFu /Fup
to be computed. Because As reflects the number of bolts in
Equation 6: Ag ≥ 0.56AsFu /Fyp
the connection times the shear area (a function of the bolts
being in single or double shear, including or excluding the 9. If Ag and An criteria are not satisfied, revise bolt capac-
threads), it is a number that the connection designer already ity for L > 38 in. criteria and recompute the number of
has available. The design equation is a modification of Equa- bolts.
tion 1 as follows:
10. Size splice plates to satisfy main member require-
Pn Pu Ab R1 R2 R3 (8) ments.
The value of R2 is either 0.90 or 0.75 depending on the con-
nection’s strength and quasi-stiffness as well as whether the
connection has a length greater than 38 in.

Table 5. Reliability ( ) and Resistance (G) Values for Current AISC/RCSC Design Criteria*
Reliability (C) Resistance (G)
Connection No. of Mean
R2 Std. Dev. Live/Dead Load Ratio (Ln /Dn) Live/Dead Load Ratio (Ln /Dn)
Length (in.) Tests Value
2 3 4 5 2 3 4 5
10.5 0.80 21 0.988 0.0293 5.5 5.1 5.0 4.8 0.933 0.949 0.958 0.963
21.0 0.80 9 0.893 0.0889 4.4 4.2 4.1 4.1 0.807 0.818 0.825 0.829
31.5 0.80 7 0.772 0.0804 3.7 3.6 3.5 3.4 0.731 0.738 0.743 0.746
42.0 0.80 7 0.848 0.176 3.2 3.1 3.1 3.1 0.714 0.718 0.721 0.723
52.3 0.64 7 0.756 0.159 3.5 3.5 3.4 3.4 0.757 0.762 0.766 0.768
62.0 0.64 4 0.901 0.122 5.1 4.9 4.8 4.7 0.897 0.912 0.921 0.928
94.0 0.64 2 0.624 0.0233 4.2 4.0 3.9 3.8 0.781 0.790 0.795 0.798
Notes:
* Rivet tests not included in these values
Gaverage 0.814
RELIABILITY The critical issue was the importance of connection

strength and quasi-stiffness as the connection became longer.
With the bolt shear strength design criteria established, it is
Once the connection strength and quasi-stiffness exceeded a
now possible to evaluate the results in terms of LRFD con-
predetermined amount, length was no longer a variable in
cepts. The reliability, C, is determined using the equation:
the performance of the connection. As a result, the test data
R was examined for both cases: first, by examining the data
ln
Q from the test results above the 0.90PTEST /PPRED ratio, and then
(9)
by examining the data excluding the test results above the
VR2 VQ2 0.90PTEST /PPRED ratio.
The current design criteria were examined for all the ap-
and the corresponding resistance factor, G, is determined us-
plicable bolt test data for nominal live load (Ln) to dead load
ing the equation:
(Dn) ratios of 2.0 to 5.0. The results for C and G are given in
Rm 0.55VR Table 5. Because of the scatter in test data, C is quite variable
e (10)
and ranges from 3.1 to 5.5. It is not surprising that short
Rn
connections have a high C value because connection length
In Equation 10, G is dependent on knowing C. Similarly, is not really a variable, although a 0.75 reduction factor is
when the step by-step-procedures are followed in Equation mandated. A similar spread in G was also obtained, ranging
9, G is required to solve for C. This dilemma is resolved by from 0.714 to 0.963, with a high average value of 0.814.
using the current G 0.75 from the AISC Specification. A This would suggest that the bolt shear design criteria could
step-by-step solution for these equations and explanation of be increased.
terms is given in Appendix B using criteria established by A review of reliability, C, for the CSA S16 (CSA, 2001,
Fisher et al. (1978). 2005) and Eurocode EN 1933 (CEN, 2003) criteria shown in
There are two procedures that can be followed to deter- Table 6 will reinforce the understanding that increasing bolt
mine C. One approach would be to establish a least-square strength with bolt diameter is not justified. For 1-in.-diameter
determination of the PTEST /PPRED relationships relative to the bolts, the C values drop to approximately 2.0 for a live to
overall connection length and solve for one C for the total dead load ratio of 5.0. When the C values are computed for
database. The second approach would determine C at dis- 1'-in.-diameter bolts, the value drops to 1.9, an unaccept-
crete connection lengths that have adequate test results. The ably low value. Although 1'-in.-diameter bolts were not
first procedure has the advantage of using more test data tested, the computed C values are sufficiently accurate be-
in one computation; however, large amounts of data at one cause the test data have been converted to a non-dimensional
length can disproportionately mask other issues. The second format. The Y-in., 1-in. and 1--in. bolt data indicate that the
procedure was used in this study to try to identify significant non-dimensional concept appears reasonable. The CSA S16
variables from the multitude that were identified in the test- document uses a G of 0.80.
ing programs.

Table 6. CSA-S16 and Eurocode Reliability ( )*
Connection Live/Dead Load Ratio
Bolt Dia.
Length Avg.
(in.) 2 3 4 5
(in.)
10.5 3.8 3.6 3.4 3.4 0.82
21.0 3.1 2.9 2.9 2.8 0.74
31.5 2.6 2.5 2.5 2.4 0.70
1 42.0 2.6 2.5 2.5 2.5 0.72
52.3 2.3 2.3 2.3 2.3 0.70
62.0 3.9 3.7 3.6 3.6 0.85
94.0 2.9 2.7 2.6 2.6 0.72
10.5 3.8 3.6 3.4 3.4 0.82
21.0 2.9 2.8 2.7 2.7 0.73
31.5 2.4 2.3 2.2 2.2 0.67
1' 42.0 2.3 2.2 2.2 2.2 0.69
52.3 2.0 2.0 2.0 1.9 0.66
62.0 3.3 3.2 3.1 3.1 0.78
94.0 2.8 2.6 2.6 2.5 0.71
* G 0.80 in CSA-S16
Table 7. Reliability ( ) and Resistance (G) Values for Proposed AISC/RCSC Design Criteria*
Reliability ( ) Resistance ( )
Connection No. of Mean Std.
R2 Live/Dead Load Ratio (Ln /Dn) Live/Dead Load Ratio (Ln /Dn)
Length (in.) Tests Value Dev.
2 3 4 5 2 3 4 5
10.5 0.90 21 0.988 0.0293 4.8 4.6 4.4 4.3 0.856 0.868 0.875 0.879
21.0 0.90 9 0.893 0.0889 3.9 3.7 3.6 3.6 0.747 0.755 0.760 0.763
31.5 0.90 7 0.772 0.0804 3.2 3.1 3.0 3.0 0.677 0.682 0.685 0.687
42.0 0.75 7 0.848 0.176 3.4 3.4 3.3 3.3 0.740 0.745 0.748 0.751
52.3 0.75 7 0.756 0.159 3.0 2.9 2.9 2.9 0.693 0.696 0.698 0.699
62.0 0.75 4 0.901 0.122 4.4 4.2 4.1 4.1 0.813 0.824 0.831 0.835
94.0 0.75 2 0.624 0.0233 3.4 3.2 3.1 3.1 0.696 0.701 0.704 0.707
Notes:
Gaverage 0.754
The proposed AISC/RCSC design criteria, all the reported value because of the test results. The magnitude of the values
bolt test data, and the computed C and G values are given in is centered, average 0.754, on the starting value of 0.75. This
Table 7. The range in C values has been reduced to 2.9 to indicates that appropriate adjustments have been made to the
4.8. The test data ratios of PTEST /PPRED that are still above current design criteria.
0.90 result in a large coefficient of variation resulting in the The final set of computations included only the test data that
low values for the 52.3 in. connection length. As previously exhibited a change in performance with connection length.
mentioned, high-strength bolts are a manufactured product, The test data with a PTEST /PPRED ratio above 0.90 were ex-
which suggests that a C of approximately 3.0 would be ac- cluded. These C and G results are shown in Table 8. Once the
ceptable. The resistance factor, G, has a similar variation in high PTEST /PPRED data are removed from the calculations, the

Table 8. Reliability ( ) and Resistance ( ) Values with Limited PTEST/PPRED Data (< 0.90)*
Reliability ( ) Resistance ( )
Connection No. of Mean Std.
R2 Live/Dead Load Ratio (Ln /Dn) Live/Dead Load Ratio (Ln /Dn)
Length (in.) Tests Value Dev.
2 3 4 5 2 3 4 5
10.5 0.90 0 — — — — — — — — — —
21.0 0.90 4 0.813 0.0438 3.7 3.6 3.5 3.4 0.731 0.738 0.742 0.745
31.5 0.90 6 0.751 0.0618 3.2 3.1 3.0 2.9 0.676 0.681 0.684 0.686
42.0 0.75 4 0.722 0.108 3.3 3.2 3.2 3.1 0.701 0.706 0.710 0.712
52.3 0.75 5 0.677 0.101 3.0 3.0 2.9 2.9 0.675 0.679 0.682 0.684
62.0 0.75 2 0.804 0.0385 4.6 4.4 4.3 4.2 0.830 0.841 0.848 0.852
94.0 0.75 2 0.624 0.0233 3.4 3.2 3.1 3.1 0.696 0.701 0.704 0.707
Notes:
Gaverage 0.726
C values ranged from 2.9 to 4.6. The C of 2.9 does not change The CSA S16 provision was identical to and transferred
because at 52.3 in. there was no data above 0.90. As previously from the Eurocode document. The two key issues in these
explained, for a manufactured product, a C of 2.9 is accept- provisions are variable and decreasing bolt shear strength
ably close to the target value of 3.0. This change is reflected with increasing connection length and increasing bolt shear
in the resistance factor, G, that on average (0.726) is below the strength with increasing diameter. The reviewed test data
starting value of 0.75. The difference is not significant. indicate that the first issue is justified, although the benefit
gained by having a sliding scale is probably not justified
SUMMARY AND CONCLUSIONS relative to the complexity. The second issue is the increasing
bolt shear strength with increasing diameter, which is not
A review of the historic research test data was made to deter- justified by the test data and at large bolt diameters results in
mine bolt shear strength in terms of LRFD principles. A total unacceptably low reliability, C.
of 119 connection tests were identified. Of these, 40 tests The current LRFD principles have a target reliability, C
were with rivets associated with the design and construction of approximately 4.0 for connections, which include slip-
of the San Francisco–Oakland Bay Bridge. Unfortunately, critical connections and bolt-bearing connections. In com-
insufficient information was reported to allow full utilization parison, the target C for main members—a manufactured
of the test data. Of the remaining 79 connection tests, the product—typically have C of approximately 3.0, or slightly
connector distribution was 54 with ASTM A325 bolts, 18 lower. Because the bolt itself is a manufactured product,
with ASTM A490 bolts, 5 with rivets, and 2 with Huck bolts. there is some leeway as to what C is acceptable for bolts.
The statistical analysis was performed using the ASTM A325 As a practical consideration, it is reasonable to use a com-
and A490 bolts. Subsequently, it was possible to show that mon resistance factor, G, value of 0.75 for slip critical con-
the rivet and Huck bolt test data were compatible with the nections, bolt bearing connections, and for this study of bolt
recommended design criteria. shear strength.
Because of the many connection variables, the test data The current AISC/RCSC design criteria result in variable
were reduced to a non-dimensional form to limit the signifi- C from 3.1, and in some cases, to a conservatively high value
cance of all the variables. As a result, the connection length of 5.5. In comparison, the proposed design criteria C range
remained as the desired and predominate independent vari- from 2.9 to 4.8. When the PTEST /PPRED test data above 0.90
able. Recent tests sponsored by RCSC also indicated that are excluded, the range for C becomes 2.9 to 4.6. The short
the reliability, C of the shear strength of bolts was similar connection values for C are going to be high because the test
to plates and shapes reported in earlier literature. Based on results are for axially loaded specimens and do not include
other anecdotal information there does not appear to be any the secondary forces associated with biaxial beam end re-
justification to change the current resistance factor, G. actions or adjacent truss-panel-point relative displacement.
In addition to the AISC/RCSC design criteria, the equiva- The effect of pretensioned bolts versus snug tight bolts has
lent Canadian CSA and Eurocode provisions were examined. not been directly evaluated.

The proposed criteria have a step function at 38 in. to suggestions; D.D. Crampton, who performed all the spread-
permit an initial length reduction factor, R2, increase from sheet iterations to locate the preferred design criteria; and fi-
0.80 to 0.90. This represents a 12.5% increase in bolt shear nally, W.F. Baker, R.B. Johnson, M.V. Holland, L.S. Muir, and
strength. Beyond 38 in., the length reduction factor is in- A. Vercellino, who provided examples of large connections.
creased from 0.64 to 0.75, a 17.2% increase. As a result, the
proposed design procedure is identical to the current require- REFERENCES
ments except with an increased bolt shear design value and a
AISC (2005), Specification for Structural Steel Buildings,
step function at 38 in. instead of 50 in.
American Institute of Steel Construction, Chicago, IL.
An unexpected result of the study was the realization that
under circumstances of sufficient connection strength, repre- Bendigo, R. A., Hansen, R.M. and Rumpf, J.L. (1963),
sented by the net area, An, and in conjunction with sufficient “Long Bolted Joints,” Journal of the Structural Division,
connection quasi-stiffness, represented by the connection ASCE, Vol. 89, No. ST6, Proc. Paper 3727, December.
gross area, Ag, in comparison to the total bolt shear area, As, (Reference 1 in Figures 2 and 4.)
there would be no need for a connection strength reduction CEN (2003), Eurocode 3: Design of Steel Structures, EN
less than 0.90 with increasing length. This condition exists 1993, European Committee for Standardization, Brussels,
when the inequalities expressed in Equations 5 and 6 are sat- Belgium.
isfied. Equation 6 is not exactly a stiffness criterion, but it CSA (2001), Limit States Design of Steel Structures, CSA-
indicates that the connection plates remain essentially elastic S16-01, Canadian Standards Association, Rexdale, On-
as the bolt ultimate shear strength is reached. tario, Canada.
Because the historical tests were performed on fully tight-
CSA (2005), Supplement 1 to CAN/CSA-S16-01, Limit
ened high strength bolts, the use of Equations 5 and 6 when
States Design of Steel Structures, CSA S16s1-05, Cana-
snug-tight bolts are used has not been experimentally con-
dian Standards Association, Rexdale, Ontario, Canada.
firmed. However, the strength component attributed to fric-
tion has been offset by the reduced G of 0.75 (13%), limiting Davis, R. E., Woodruff, G.B. and Davis, H.E. (1940), “Ten-
the connection capacity ratio to 0.90 (11%) and increasing sion Tests of Large Riveted Joints,” Transactions, ASCE,
the gross area coefficient requirement in Equation 6 from Vol. 105, p. 1193. (Reference 2 in Figures 2 and 4.)
0.47 to 0.56 (19%). In addition, the statistical bolt strength Fisher, J.W., Galambos, T.V., Kulak, G.L. and Ravindra,
will be somewhat higher than the research programs inten- M.K. (1978), “Load and Resistance Factor Design Cri-
tional use of low end bolt strength. Performing a few tests teria for Connectors,” Journal of the Structural Division,
would quantify and refine the N1 and N2 coefficients. ASCE, Vol. 104, No. ST9, September.
The statistical study was based on ASTM A325 and A490 Fisher, J.W., Ramseier, P.O. and Beedle, L.S. (1963),
bolts; however, when the limited rivet and Huck bolt data “Strength of A440 Steel Joints Fastened with A325 Bolts.”
were compared with the bolt results, no inconsistency was Publications, IABSE, Vol. 23. (Reference 3 in Figures 2
found. Similarly, the connection plate material varied from and 4.)
relatively low-strength ASTM A7 steel to high-strength
Fisher, J.W. and Kulak, G.L. (1968), “Tests of Bolted Butt
ASTM A514 steel with intermediate-strength ASTM A440
Splices,” Journal of the Structural Division, ASCE, Vol
and A572 steel in between, again with no inconsistencies.
94, ST11, November. (Reference 4 in Figures 2 and 4.)
This would indicate that the data in a non-dimensional for-
mat did not have any apparent bias and indicates that the Fisher, J.W. and Yoshida, N. (1970), “Large Bolted and
procedure is acceptable for all current grades of connectors, Riveted Shingle Splices,” Journal of the Structural Divi-
plates and shapes. In conclusion, the proposed design crite- sion, ASCE, Vol 96, No. ST9, Proc. Paper 7534, Sept., pp.
ria are essentially identical to the current provisions except 1903-1918. (Reference 5 in Figures 2 and 4.)
the bolt strengths are adjusted slightly upwards, resulting Foreman, R.T. and Rumpf, J.L. (1961), “Static Tension Tests
in a more uniform reliability, C closer to the professionally of Compact Bolted Joints,” Transactions ASCE, Vol. 126,
accepted values. In addition, Equations 5 and 6 provide a p. 228. (Reference 6 in Figures 2 and 4.)
means of proportioning a connection to gain optimum bolt Kulak, G.L. and Fisher, J.W. (1968), “A514 Steel Joints Fas-
shear strength. tened by A490 Bolts,” Journal of the Structural Division,
ASCE, Vol. 94, ST10, October. (Reference 7 in Figures 2
ACKNOWLEDGMENTS and 4.)
The author wishes to acknowledge several colleagues who Kulak, G.L., Fisher, J.W. and Struik, J.H.A. (1987), Guide
assisted him in this bolted connection study: T.V. Galam- Design Criteria for Bolted and Riveted Joints, 2nd edi-
bos, who set up the reliability procedure that was used tion, John Wiley and Sons, New York, NY.
to calibrate C; J.A. Yura, who made several meaningful

Moore, A.M., Rassati, G.A. and Swanson, J.A. (2008), Rivera, U. and Fisher, J.W. (1970), “Load Partition and Ul-
“Evaluation of the Current Resistance Factors for High timate Strength of Shingle Joints,” Fritz Laboratory Re-
Strength Bolts,” Final report submitted to the Research port 340.6, Lehigh University. (Reference 9 in Figures 2
Council on Structural Connections. and 4.)
Power, E.H. and Fisher, J.W. (1972), “Behavior and Design Sterling, G.H. and Fisher, J.W. (1965) “Tests of Long A440
of Shingle Joints,” Journal of the Structural Division, Steel Bolted Butt Joints,” Fritz Engineering Laboratory
ASCE, Vol. 98, No. ST9, September. (Reference 8 in Fig- Report No. 288.26, Lehigh University. (Reference 11 in
ures 2 and 4.) Figures 2 and 4.)
RCSC (2004), Specification for Structural Joints Using Sterling, G.H. and Fisher, J.W. (1966), “A440 Steel Joints
ASTM A325 or A490 Bolts, Research Council on Struc- Connected by A490 Bolts,” Journal of the Structural Di-
tural Connections, Chicago, IL. vision, ASCE, Vol. 92, No. ST3, Proc. Paper 4845, June.
(Reference 10 in Figures 2 and 4.)
APPENDIX A
AISC SPECIFICATION SECTION D2 Plate Fracture

Use Equation D2-2 from the 2005 AISC Specification:
Plate Yielding
Pn Fu An
Use Equation D2-1 from the 2005 AISC Specification:
Pn Fy Ag The design rupture strength of the plate is GtPn and Gt 0.75.
Thus,
The design tensile strength of the plate is GtPn and Gt 0.90. GtPn GtFu An
Thus,
GtPn GtFy Ag Equating the bolt shear strength to the plate rupture strength
(with Fup the rupture stress of the plate):
For bolts, the design shear strength is Gb Pn, with Gb 0.75.
Gb(Fu As)R1R2 GtFup An
Also, Pn Pu R1R2, assuming threads excluded from the shear
plane. We also know that Pu Fu As. Substituting, we have:
Solving for An and substituting as before, we obtain:
Gb Pn Gb(Pu R1R2) Gb(Fu As)R1R2 0.563Fu As
An
R1 has been established at 0.625. Take R2 as 0.90 by assum- Fup
ing L 38 in., avoiding reducing R2 from 0.90 to 0.75 for Bolt shear will control over tensile rupture as long as this
L 38 in. inequality is satisfied.
Equating the bolt shear strength to the plate yield strength
(with Fyp the yield stress of the plate): Notes:
Gb(Fu As)R1R2 GtFyp Ag 1. As and Fu are bolt properties.
Solving for Ag, 2. For design purposes, use a coefficient of 0.56 for both
calculations (i.e., for N1 and N2) until further research
b Fu As R1 R2 0.75( Fu As )(0.625)(0.90) 0.469 Fu As
Ag quantifies pretensioning and second-order effects.
t Fyp 0.90 Fyp Fyp
Bolt shear will control as long as:
0.469 Fu As
Ag
Fyp

APPENDIX B
SAMPLE CALCULATION FOR BOLT SHEAR ( DVD )2 ( LVL )2

VQ =
R Q
ln
Q 2
(1.05 DN )(0.10) (3.00D
DN )(0.25)
2

VR2 VQ2 =
4.05 DN
where = 0.1870
R mean resistance
Q mean load effect 3. Nominal Design Strength
VR coefficient of variation
In non-dimensional form, RN R1R2FUN, where R2 is a de-
VQ coefficient of variation
sign level criteria. The connection length is 42 in., which is
greater than 38 in. Therefore, R2 is taken as 0.75.
R f {R1, FUN, RP, RM, RF}
RN (0.625)(0.75)FUN
where
0.4688FUN
FUN nominal tensile strength of bolt
R1 ratio of shear to tensile strength 0.625, Set GRN equal to QN, with G 0.75, and solve for DN:
VR1 0.05
RP mean test value, VRP ratio of standard deviation 0.75(0.4688FUN) 6.0DN
to mean test value DN 0.0586FUN
RM fabricating factor 1.20, VRM 0.07
Now, calculate Q:
RF fabricating factor 1.00, VRF 0.02
Q 4.05DN
These data are from Fisher et al. (1978).
4.05(0.0586FUN)
0.2373FUN
EXAMPLE
L 42 in., 7 tests The reliability, C, may now be calculated:
RP 0.848, standard deviation 0.1761
0.6360 FUN
ln
1. Mean Resistance 0.2373FUN

R R1 RP RM RF FUN (0.2257)2 (0.1870)2
R (0.625)(0.848)(1.20)(1.00)FUN 0.6360FUN 0.9859

VRP 0.1761/0.848 0.2077 0.2931
3.36
VR = V R21 + VR2P + VR2M + VR2F
VR = 0.052 + 0.20772 + 0.072 + 0.022 = 0.2257 4. Resistance Factor,

RM 0.55VR
2. Mean Load Effect e
RN
DN, LN nominal dead and live loads, respectively RM mean test value (RP)from C calculations
D 1.05DN , VD 0.10 RN proposed design criteria, R2
L LN, VL 0.25 C from previous calculations (Step 3)
For this example, assume LN /DN 3.0 VR coefficent of variation
QN nominal load effect 0.848
( 0.55)(3.36)( 0.2257 )
1.2DN 1.6LN DN [1.2 1.6 (LN /DN)] 6.0DN e
0.75
Q mean load effect
DL
1.1307 e 0.417

1.05DN LN 1.05DN 3.0DN 4.05DN (1.1307)(0.659)
0.745

Current Steel Structures Research
REIDAR BJORHOVDE
INTRODUCTION locales, the research contract budgets do not need to cover
A
the 25% of the professor’s annual salary, which means that
mong the thorniest problems for researchers in gener-
the overall cost is reduced significantly. At the same time,
al—and maybe for American structural engineering
the sheer size of the total American academic research un-
researchers in particular—is the subject of money, that is,
dertaking dwarfs that of all other countries. So the rest of
research funding. This is especially acute for academic re-
the world is working hard to catch up with the American
searchers under the current economic conditions, what with
research enterprise, which is still the most effective and
limited governmental and private research grants funding
powerful in the world. But the efforts within the European
being further reduced or in many cases entirely eliminated.
Union, for example, are notable and very impressive, as is
What many government leaders and business executives and
the support offered in countries like China, Japan, Singapore
accountants do not appear to understand is that when mar-
and Australia. Therefore, unless American federal and state
kets “turn south,” investment in research and marketing is
governments and industry are prepared to address the fund-
more important than ever, for that is when the foundation is
ing issue aggressively, the leading role of U.S. universities
placed for rapid advances when business picks up again.
will be eroded—significantly. That will not be helpful to
American university research and researchers are espe-
anyone, anywhere.
cially vulnerable, considering that funding must be obtained
Many of the research studies presented here have been
not just for graduate student support (commonly $20,000
conducted in Europe, with financial support from various
or more per year, to cover tuition and stipend), test speci-
instances within the European Union. Some of these are
mens, equipment, and other legitimate expenses, but also to
typical of many current research projects in Europe, as they
cover typically three months of income for the professor(s)
reflect collaboration between universities in different coun-
as well as university overhead of 40% to 60% of the con-
tries. For example, a study of the axial and flexural capaci-
tract amounts. How many Americans know that university
ties of welded built-up box cross sections is a joint effort
professors (like public school teachers) are paid only for
between institutions in Germany and Slovenia, and an ex-
nine months of the year and that the remainder of the annual
amination of the buckling capacity of slender composite tu-
income has to come from research funds or other sources?
bular columns has been a joint effort between schools in the
And if the researcher wants to advance in his or her aca-
U.K. and Lithuania. Certain bridge evaluations have been
demic career, research funding must be obtained to ensure
conducted in Denmark and the U.K., and an interesting as-
that results, publications and recognition will be obtained for
sessment of structural strengthening using adhesives has
scientific achievement. In brief, an academic career without
been done jointly by researchers in Germany and Poland.
research and attendant evidence of scholarly success equals
The potential performance and use of girders with corrugat-
failure in most institutions, because teaching alone is simply
ed webs in building structures is conducted in Germany; this
not sufficient. You must demonstrate scholarly success, and
is an application that until now has only been seen in bridge
that comes from successful research work.
construction. Finally, composite construction continues to
The modus operandi is different for researchers and in-
be the subject of aggressive research work in Australia, as
stitutions in most other parts of the world, although there
reported here, as is a major investigation of the steel stor-
are some signs that universities in the U.K., for example,
age racks that are so important for many industrial and other
are looking at the American model as a means of reducing
manufacturing operations. By all indications, these studies
the demands on university budgets. In most of these other
are typical of research institutions that conduct high qual-
ity research on subjects that are likely to attract attention in
many engineering locales.
References are provided throughout the paper, whenever
such are available in the public domain. However, much
of the work is still in progress, and in many cases reports
or publications have not yet been prepared for public
Reidar Bjorhovde, Ph.D., P.E., Research Editor for Engineering Journal, dissemination.
5880 E. Territory Ave., Suite 202, Tucson, AZ, 85750-1803. E-mail:
[email protected]

PERFORMANCE OF STRUCTURAL MEMBERS Plate Girders with Corrugated Webs for Building Appli-
cations: This project has been conducted at the Brandenburg
Slender Thin-Walled Box Columns: This study has been
University of Technology in Cottbus, Germany, in partner-
conducted at the University of Ljubljana in Ljubljana, Slo-
ship with the Bochum University of Applied Science in
venia, and the University of Stuttgart in Stuttgart, Germany.
Bochum, Germany. The project director has been Professor
Professors Darko Beg (Ljubljana) and Ulrike Kuhlmann
Hartmut Pasternak.
(Stuttgart) have been the directors of the project.
Plate girders with corrugated webs were originally devel-
Focusing on the use of thin-walled box cross section col-
oped for use in bridge structures. The depth and spans of
umns, the members were fabricated using hot-rolled plates
the typical plate girders with flat-plate webs often required
as well as cold-formed elements. The usual material property
extensive transverse and even longitudinal stiffeners, with
tests were conducted, as were complete evaluations and mea-
attendant stability problems as well as long-term cyclic load
surements of residual stresses, member out-of-straightnesses
performance and difficulties in the form of fatigue cracking.
and out-of-plumbs. Eight full-scale column tests were con-
The flat-plate web thickness also could be substantial, result-
ducted, using concentric and eccentric axial loads, and the
ing in heavy girders.
correlation between the analytical and the experimental re-
The corrugated webs of the girders that have been studied
sults was found to be very good (Pavlovcic et al., 2009).
by Professor Pasternak benefit from increased web buckling
Figure 1 shows two of the column specimens during testing.
strength due to the shape of the web and its innate stabil-
The residual stress magnitudes and distributions were de-
ity. As a result, these girders are now being used in building
termined using the well-known method of sectioning, as de-
structures, due to their reduced self weight. Plate thicknesses
veloped and used extensively at Lehigh University (Tebedge
in these girders are typically 1.5 to 3 mm (o in. to - in.),
et al., 1973). There was very good agreement with the data
and the corrugations are sinusoidally shaped (Pasternak et
provided by other investigations, for welded built-up as well
al., 2009a). The sinusoidal fabrication process has been au-
as cold-formed members (Galambos, 1998).
tomated, with the result that corrugated web thicknesses up
The analytical computations were performed using the
to 12 mm (' in.) can be used. The girders have seen applica-
ABAQUS software. It predicted correct ultimate limit states
tion in storage warehouses, as shown in Figure 2, and now
and accurate failure loads for local as well as overall buck-
also in short span bridges. Design criteria for the girders are
ling limit states, with an accuracy of q6 percent. This is
provided by Eurocode 3, Section 1-5, Annex D, although the
comparable to what was found in the classical column study
current requirements limit the web thickness to no more than
of Bjorhovde and others, although the earlier investigations
3 mm (- in.) (CEN, 2005a).
did not have access to such advanced software as ABAQUS
(Bjorhovde, 1972; Galambos, 1998).
Fig. 1. Slender box column specimen during different loading stages

(Photos courtesy of Professors Darko Beg and Ulrike Kuhlmann).

The researchers are now developing formulations that can The project at the University of New South Wales devel-
be used for larger web thicknesses, taking into account non- oped a push-test to replace the push-out test that has been
linear response characteristics, lateral-torsional buckling the traditional one to use to assess stud shear capacity, etc.
and patch loading. This effort is a joint project between the (Ranzi et al., 2009). The push-out tests in many cases led to
Brandenburg and Braunschweig technological universities premature failures of a type that never occurred in full-scale
in Germany (Pasternak et al., 2009a). beam tests, hence the need for the improved modeling and
test as provided by the UNSW group. Figure 3 shows the
STRENGTH AND BEHAVIOR OF COMPOSITE push-test laboratory assembly, and Figure 4 shows the very
ELEMENTS AND STRUCTURES favorable ductile load-slip response of the studs.
Two full-scale beam tests were run at the University of
Behavior of Composite Beams with Deep Trapezoidal
Sydney to verify and validate the results of the push-test that
Slabs Containing Headed Stud Shear Connectors: This
had been developed. With beam spans of 8 m (26 ft 8 in.),
is a major project that has been conducted at the Centre for
the beams performed very well, with ductile behavior as pre-
Infrastructure Engineering and Safety of the University of
dicted by the push-test. The correlation with the analytical
New South Wales in Sydney, Australia, with Professor Mark
results was excellent. Figure 5 shows the test setup for the
Bradford as the director. Funding and other support have
full-scale beams.
been provided by the Australian Research Council and the
Additional work focuses on the development of service-
Australian Steel Institute, as well as the Australian steel deck
ability criteria for these types of beams, including the warp-
producer BlueScope Lysaght.
ing effect of slab shrinkage and its interaction with the par-
Composite beams using corrugated steel deck and headed
tial shear connection between the slab and the steel beam.
stud shear connectors have been very common since an early
research project was conducted at Lehigh University in the
late 1960s (Fisher, 1970; Grant et al., 1977). For a number
of years the types of steel deck exhibited various forms of
trapezoidal and other corrugations, and the deck depths were
mostly 1' in., 2 in., and 3 in. These systems were eminently
suitable for floor beams and girders in commercial and resi-
dential low- and high-rise building frames. The span of the
concrete slab on top of the steel deck was commonly 8 ft to
12 ft, due to the limitations imposed by moment capacity of
the slab, the shear stud length and capacity, and the deflec-
tions of the slab. Various deck types and geometric configu-
rations have since been developed, with depths up to 6 in. to
8 in. (150 mm to 200 mm), but the shear connectors and the
overall design criteria have not evolved accordingly. How-
ever, longer spans clearly would be very useful for many Fig. 3. Push-test for stud shear connector performance
applications. (Photo courtesy of Professor Mark Bradford).
Fig. 2. Storage warehouse with corrugated web girders Fig. 4. Ductile load-slip performance of push-tested shear connector
(Photo courtesy of Professor Hartmut Pasternak). (Figure courtesy of Professor Mark Bradford).

Buckling of Slender Composite Concrete-Filled Steel Col- full-scale tests that have been conducted demonstrate the re-
umns: This has been an international collaborative research sponse of the structures under gravity and lateral loads. One
project between the Vilnius Gediminas University in Vilnius, such test is illustrated in Figures 6 and 7, which show the
Lithuania and the University of Manchester in Manchester, rack before the test and the condition after a typical pallet
England. Professors Audronis Kvedaras (Vilnius) and Cam- loading case, as described earlier.
eron Goode (Manchester) have been the project directors.
The project has effected a major data collection and evalu-
ation of test results for square and round tubular columns,
to assess the design criteria that are specified in Eurocode
4 (Goode, 2008; CEN, 2005b). It has been found that Euro-
code 4 tends to be conservative for circular tubular columns,
and the criteria can be readily extended up to concrete cyl-
inder strengths of 100 MPa (14 ksi), or significantly higher
than the current limitation of 75 MPa (approximately 11
ksi). This does not extend to rectangular columns, for which
the current limit is 50 MPa (7 ksi).
INDUSTRIAL STRUCTURES
Strength and Behavior of Steel Storage Racks: This has
been a major, multi-year project at the University of Syd-
ney in Sydney, Australia. Professor Kim Rasmussen is the
project director for a study that has wide governmental and
industrial support.
The study has focused on the performance of the storage
racks when they are subjected to impact forces due to fork-
lifts. This is a very common and sometimes disastrous oc- Fig. 6. Full-scale storage rack before testing
currence. In addition to detailed three-dimensional strength (Photo courtesy of Professor Kim Rasmussen).
and stability analyses of the racks, full-scale tests have also
been conducted to ensure as realistic conditions as possible.
Among other considerations, the tests include actual colli-
sions of forklifts with rack columns (“uprights”), evaluating
the dynamic response of the structure (Gilbert and Rasmus-
sen, 2009). Progressive collapse scenarios are also assessed.
The pallet loading of the racks is a central issue, including
what takes place when a pallet falls through several “stories”
of a rack that may have experienced upright damage. The
Fig. 7. Pallet-loading-induced failure of storage rack (notice the

Fig. 5. Full-scale composite beam test with deep trapezoidal steel deck damaged upright on the right-hand side of the photograph)
(Photo courtesy of Professor Mark Bradford). (Photo courtesy of Professor Kim Rasmussen).

The project work includes assessments of component tests Figure 10 shows the variation of the effective length of
and other aspects of various international storage rack cri- the girders, as a function of the spacing of the U-frames and
teria, such as the American code by RMI (Rack Manufac- the buckling mode. For instance, mode 1 (top curve) is the
turers Institute) and the relevant European code. Additional usual single wave form, reflecting the lowest buckling load.
component tests are being developed, including one for the Increasing the U-frame spacing from 2 m to 3 m increases
stiffness of base plates subjected to uplift. the effective length by 20%, meaning that the buckling load
is reduced by (1.2)2 or 44%. This is as should be expected.
BRIDGE STRUCTURES
STRUCTURAL REPAIR AND RETROFIT
Stability of Girders Using U-Frames: This study was
conducted at City University of London, England, while the Use of Adhesives to Strengthen Steel Structures: This
project director, Kuldeep S. Virdi, was a faculty member at project has been a joint international effort between the
that university. He is now at Aarhus University in Aarhus, Brandenburg University of Technology in Cottbus, Ger-
Denmark. many, and the Krakow University of Technology in Krakow,
Lateral-torsional buckling is a major consideration for Poland. The project directors are Professors H. Pasternak
beams and girders, and maybe especially for the long-span (Brandenburg) and M. Piekarczyk (Krakow). The funding
members often used in bridge construction. While actual sources have not been identified, but they are likely to be
bracing is the most common approach to increase the stabil- major European Union programs.
ity and strength of the girders, there are a number of cases Adhesives have actually been used in a number of con-
where such is not suitable, especially for through-type bridg- struction projects in Germany, starting with several box gird-
es and where overhead bracing members are not practical. er bridges in the 1960s. These structures have continued to
The so-called U-frame provides girder stability through con- function properly, now after nearly 50 years of service. In the
nections between the floor beams and the girders, as shown
in Figure 8, and the floor beams also support the roadway
slab.
The study examined the influence of a range of girder and
floor beam spacings to determine the buckling strength of
the system (Virdi and Azzi, 2009). An example of the buck-
led shape for one of the U-frame configurations is shown in
Figure 9.
A parametric analysis of the U-frame system was per-
formed, using a bridge span of 18 m (60 ft), a bridge
deck width of 6.3 m (21 ft), 1.8-m-deep (6 ft) girders, and
0.8-m-deep (30 in.) floor beams. The bridge deck was de-
Fig. 9. One of the buckling shapes of a typical structure with U-frames
signed as composite with the floor beams. The U-frame
(Figure courtesy of Professor Kuldeep Virdi).
spacing was varied from 2 m to 3 m (6 ft 8 in. to 10 ft),
giving structures with eight to five U-frames.
Buckling Mode 1
Buckling Mode 5
Fig. 8. Girders and floor beams constitute the U-frames of the bridge Fig. 10. Variation of girder effective length as a function of U-frame spacing
(Figure courtesy of Professor Kuldeep Virdi). (Figure courtesy of Professor Kuldeep Virdi).

current project, a box girder was tested at Krakow and knee Pasternak, H., Kubieniec, G. and Robra, J. (2009a), “Plate
joints were tested at Brandenburg, and ABAQUS analyses Girders with Corrugated Webs,” in Stability and Ductility
were conducted for all of the test specimens. There was good of Structures, Audronis Kvedaras, ed., International Col-
correlation between the tests and the analytical results. loquium, Vilnius Gediminas Technical University, Vilni-
For the knee joints it was found that it was better to use us, Lithuania.
epoxy-glued plates on both sides of the joint; a single sided Pasternak, H., Kubieniec, G. and Piekarczyk, M. (2009b),
plate was sufficient when polyurethane glue was used. Simi- “Adhesives in Strengthening of Steel Structures,” in Sta-
lar results were found for the box girders. The increases in bility and Ductility of Structures, Audronis Kvedaras, ed.,
load-carrying capacity were substantial, including the fact International Colloquium, Vilnius Gediminas Technical
that local buckling did not affect the ultimate limit states. It University, Vilnius, Lithuania.
is clear that the use of various forms of adhesives can play an
Pavlovcic, L., Froschmeier, B., Kuhlmann, U. and Beg, D.
important strengthening role—it is still a question whether
(2009), “Slender Thin-Walled Box Columns Subjected to
the organic stability of the glued joints and surfaces will suf-
Compression and Bending,” in Stability and Ductility of
fice. However, the performance of the older German bridges
Structures, Audronis Kvedaras, ed., International Collo-
speaks well for the method and the materials.
quium, Vilnius Gediminas Technical University, Vilnius,
Lithuania.
REFERENCES
Ranzi, G., Bradford, M.A., Ansourian, P., Filonov, A., Ras-
Bjorhovde, Reidar (1972), “Deterministic and Probabilistic mussen, K.J.R., Hogan, T.J. and Uy, B. (2009), “Full-
Approaches to the Strength of Steel Columns,” Ph.D. Dis- Scale Tests on Composite Steel-Concrete Beams with
sertation, Lehigh University, Bethlehem, PA. Trapezoidal Steel Decking,” Journal of Constructional
CEN (2005a), Eurocode 3—Design of Steel Structures—EN Steel Research, Vol. 65, No. 7, pp. 1490–1506.
1993-1, Comité Européen de Normalisation, Brussels, Tebedge, N., Alpsten, G.A. and Tall, L. (1973), “Residual
Belgium. Stress Measurement by the Sectioning Method,” Experi-
CEN (2005b), Eurocode 4—Design of Composite Steel and mental Mechanics, Vol. 13, pp. 88–96.
Concrete Structures—EN 1994-1, Comité Européen de Virdi, K. S. and Azzi, W. (2009), “Stability of Girders Us-
Normalisation, Brussels, Belgium. ing U-Frames,” in Stability and Ductility of Structures,
Fisher, J.W. (1970), “Design of Composite Beams with Audronis Kvedaras, ed., International Colloquium, Vil-
Formed Metal Deck,” Engineering Journal, AISC, Vol. 7, nius Gediminas Technical University, Vilnius, Lithuania.
No. 3, pp. 88–96.
Galambos, T.V., ed. (1998), Guide to Stability Design Crite- ACKNOWLEDGMENTS
ria for Metal Structures, 5th edition, John Wiley & Sons,
Special thanks are due the following members of the Inter-
New York, NY.
national Structural Steel Research Advisors (ISSRA) who
Gilbert, B.P. and Rasmussen, K.J.R. (2009), “Impact Tests provided input to this paper:
and Parametric Impact Studies on Drive-In Steel Storage
Racks,” International Conference on Advances in Steel r %BSLP #FH 6OJWFSTJUZ PG -KVCMKBOB -KVCMKBOB
Structures (ICASS’09), Hong Kong, December 16–18. Slovenia
Goode, C.D. (2008), “Composite Columns—1819 Tests r .BSL #SBEGPSE 6OJWFSTJUZ PG /FX 4PVUI 8BMFT
on Concrete-Filled Steel Tube Columns Compared with Sydney, Australia
Eurocode 4,” The Structural Engineer. Vol. 86, No. 16,
r 6MSJLF ,VIMNBOO 6OJWFSTJUZ PG 4UVUUHBSU 4UVUUHBSU
pp. 33–38.
Germany
Goode, C.D., Kuranovas, A. and Kvedaras, A.K. (2009),
“Buckling of Slender Composite Concrete-Filled Steel Additional assistance has been provided by Audronis Kve-
Columns,” in Stability and Ductility of Structures, Audro- daras, Vilnius Gediminas Technical University, Vilnius,
nis Kvedaras, ed., International Colloquium, Vilnius Ge- Lithuania; Hartmut Pasternak, Brandenburg University of
diminas Technical University, Vilnius, Lithuania. Technology, Cottbus, Germany; Luka Pavlovcic, University
Grant, J.A., Fisher, J.W. and Slutter, R.G. (1977), “Compos- of Ljubljana, Ljubljana, Slovenia; Kim Rasmussen, Univer-
ite Beams with Formed Steel Deck,” Engineering Jour- sity of Sydney, Sydney, Australia; and Kuldeep Virdi, Aarhus
nal, AISC, Vol. 14, No. 1, pp. 24–43. University, Aarhus, Denmark.

GUIDE FOR AUTHORS
SCOPE: The ENGINEERING JOURNAL is dedicated to the improvement and
advancement of steel construction. Its pages are open to all who wish to
report on new developments or techniques in steel design, research, the
design and/or construction of new projects, steel fabrication methods, or
new products of significance to the uses of steel in construction. Only
original papers should be submitted.
GENERAL: Papers intended for publication may be submitted by mail to the Editor,
Keith Grubb, ENGINEERING JOURNAL, AMERICAN INSTITUTE OF
STEEL CONSTRUCTION, One East Wacker Drive, Suite 700, Chicago,
IL, 60601, or by email to [email protected].
The articles published in the Engineering Journal undergo peer review
before publication for (1) originality of contribution; (2) technical value to
the steel construction community; (3) proper credit to others working in
the same area; (4) prior publication of the material; and (5) justification of
the conclusion based on the report.
All papers within the scope outlined above will be reviewed by engineers
selected from among AISC, industry, design firms, and universities.
The standard review process includes outside review by an average of
three reviewers, who are experts in their respective technical area, and
volunteers in the program. Papers not accepted will not be returned to the
author. Published papers become the property of the American Institute of
Steel Construction and are protected by appropriate copyrights. No proofs
will be sent to authors. Each author receives three copies of the issue in
which his contribution appears.
MANUSCRIPT PREPARATION: Manuscripts must be provided in Microsoft Word 2003 format. A laser-
quality proof or high quality PDF must accompany your submittal.
Download our complete author guidelines at www.aisc.org/ej.
There’s always a solution in steel.
ENGINEERING JOURNAL
American Institute of Steel Construction
One East Wacker Drive, Suite 700
Chicago, IL 60601

AISC Engineering Journal Q1 2010

Uploaded by

Copyright:

Available Formats

AISC Engineering Journal Q1 2010

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

AISC Engineering Journal Q1 2010

Uploaded by

Copyright:

Available Formats

What are some of the topics covered in the journal?

What are some of the topics covered in the journal?

What guidelines must authors follow when submitting papers to the journal?

What guidelines must authors follow when submitting papers to the journal?

Engineering

First Quarter 2010 Volume 47, No. 1

1 Critical Evaluation of Equivalent Moment Factor

Dedicated to the development and improvement of steel construction,

Keywords: lateral support, equivalent moment factors, Cb, beams.

INTRODUCTION 1. The internal moment distribution between brace

ENGINEERING JOURNAL / FIRST QUARTER / 2010 / 1

2 / ENGINEERING JOURNAL / FIRST QUARTER / 2010

Type 1—End Moments Only

Type 2—Uniformly Distributed Load with Equal End Moments

Type 3—Uniformly Distributed Load with One End Moment

Type 4—Mid-span Concentrated Load with Equal End Moments

Type 5—Mid-span Concentrated Load with One End Moment

Type 6—Concentrated Load with Pinned Ends

ENGINEERING JOURNAL / FIRST QUARTER / 2010 / 3

Type 10—Two Equal Concentrated Loads Symmetrically Placed, Brace at Mid-span

4 / ENGINEERING JOURNAL / FIRST QUARTER / 2010

ENGINEERING JOURNAL / FIRST QUARTER / 2010 / 5

Equation a Publication Equivalent Moment Factor Equation

2 Salvadori (1955) b, c Cb 1.75 1.05 0.3 2 2.3

3 Austin (1961) Cb (0.6 0.4 ) 1 2.5

6 / ENGINEERING JOURNAL / FIRST QUARTER / 2010

Moment Clark & Hill Nethercot & Trahair Eurocode 3

ENGINEERING JOURNAL / FIRST QUARTER / 2010 / 7

Fig. 1. Simplified moment diagrams according to CAN/CSA-S16-01 and AASHTO.

8 / ENGINEERING JOURNAL / FIRST QUARTER / 2010

ENGINEERING JOURNAL / FIRST QUARTER / 2010 / 9

Fig. 2. Cb Results for Moment Type 1.

Fig. 3. Cb Results for Moment Type 2.

10 / ENGINEERING JOURNAL / FIRST QUARTER / 2010

Fig. 4. Cb Results for Moment Type 3.

Fig. 5. Cb Results for Moment Type 4

ENGINEERING JOURNAL / FIRST QUARTER / 2010 / 11

Fig. 6. Cb Results for Moment Type 5.

Fig. 7. Cb Results for Moment Type 6.

12 / ENGINEERING JOURNAL / FIRST QUARTER / 2010

Fig. 8. Cb Results for Moment Type 7.

Fig. 9. Cb Results for Moment Type 8.

ENGINEERING JOURNAL / FIRST QUARTER / 2010 / 13

Fig. 10. Cb Results for Moment Type 9, left unbraced segment.

Fig. 11. Cb Results for Moment Type 9, right unbraced segment.

14 / ENGINEERING JOURNAL / FIRST QUARTER / 2010

Fig. 12. Cb Results for Moment Type 10.

Fig. 13. Cb Results for Moment Type 11.

ENGINEERING JOURNAL / FIRST QUARTER / 2010 / 15

Fig. 14. Cb Results for Moment Type 12.

16 / ENGINEERING JOURNAL / FIRST QUARTER / 2010

ENGINEERING JOURNAL / FIRST QUARTER / 2010 / 17

18 / ENGINEERING JOURNAL / FIRST QUARTER / 2010

ENGINEERING JOURNAL / FIRST QUARTER / 2010 / 19

20 / ENGINEERING JOURNAL / FIRST QUARTER / 2010

Keywords: diaphragms, seismic performance, low-rise steel buildings.

INTRODUCTION Diaphragms may also contribute to the overall dynamic

ENGINEERING JOURNAL / FIRST QUARTER / 2010 / 21

Steel Deck Collector Bracing Foundations

Fig. 1. Single-story buildings with capacity-based design concepts for SLRS.

and the resistance/safety factor is again taken as G 0.75 or SUMMARY

b 4 0.971 0.025 0.077 0.112

b 4 1 (24) 0.904 0.166 0.249 0.347