STRUCT

STEEL
DESIGN

Joseph E. Bowles
Professor of Ciod Engineering

McGraw-Hill Bmk Company

New York St. Louis San Francisco Auckland Bog& Hambur
Johannesburg London Madrid Mexico hiontreal New&
Panama Paris Sao Paulo Singapore Sydney Tokyo Toron
 S k;lrrLTURALSTEEL DE§IGN

Cv-.:- r;ht 0 1980 by McGrdw-H111, Inc. All nghts reserved CONTENTS

Pr: t.2 m the Unlted States of Amenca. No part of this publ~catlon
n t y ~ereproduced, stored In a retneval system, or transm~tted,In any
I'm or by any means, electronic, mechanical, photocopy~ng,recording, or

-9t I 2. se, w~thoutthe pnor wntten permissron of the publ~sher.

&.@! '67890 DODO 89876543210

Preface
Chapter 1 General Design Considerations
1-1 Types of Structures
1-2 Design Procedures
1-3 Steel as a Structural Matenai
1-4 Steel Products
1-5 Steel Strength
1-6 Temperature Effects o n Steel
1-7 Structural Design Codes
Thi, '..do'# was set in Times Roman by Science Typographers, Inc. 1-8 Building Loads
Tfie ..iilurs were Julienne V. Brown and Madelaine Eichberg; 1-9 Highway a n d Railroad Bridge Loads
thd cover was designed by Anne Canevari Green; 1-10 Impact Loads
the ?rs.,duction supenisor was Dominick Petrellese. 1-1 1 Earthquake Loads
'The ?..:wings were done by J & R Services, Inc. 1-12 Fatigue
K.'8. Donnelley & Sons Company was printer and binder. 1-13 Steel Structures
1-14 Accuracy of Computations and Electronic Calculators
1-15 Structural Engineering Computations in SI

'Uh.--*<
of Congress Cataloging in Publication Data
Chapter 2 Elements of Frame, Truss, and Bridge
Bor ' s. 'oseph E Design
rdral steel des~gn
2-1 Methods of A n a l y s ~ s
^ .>graphy: p. 2-2 Beam Analysis
.
udes mdex. 2-3 Determinate Structures
Bulldmg, Iron and steel. 2. Steel,
!
2-4 Truss Analysis
Str~ 3. Structures, Theory of. I. T ~ t l e
L I ~ .

'TA6!' "478 624'.1821 79-18155 2-5 h a d Frame Analys~s

1SBW r 37-006765-1 2-6 Bndge Analysls
 2-7 The Computer Program Furnished in the Appendlx Axially Loaded Columns and Struts
2-8 The P Matrix
2-9 Load Conditions ~ntroduction
2- I0 Checking Computer Output The Euler Column Formula
2- 1 1 Design Examples Columns a l t h End Condltlons
Allowable Stresses in Steel Columns
Deslgn of Bu~lt-upCompression Members
Chapter 3 Elastic, Plastic, and Buckling Behavior Column Base Plates
of Structural Steel Lateral Brac~ngof Columns
Column and Strut Design Us~ngLRFD
Introduction
Elastic versus Plastic Design Theory
Safety Factors in Elastic and Plastic Design Beam-Column Design
Elastic versus Plastic Design Deflections Introduction
Length of Plastic Hinge General Considerations of Axial Load with Bending
Elastic versus Plastic Design Effective Lengths of Columns in Building Frames -
Load Resistance Factor Design Developing the Beam-Column Design Formulas
Local Buckling of Plates Determination of the Interaction Reduction
Post-Buckling Strength of Plates Coefficient C,,,
AASHTO and AREA Beam-Column Design Formulas
Beam-Column Design Using Interaction Equations
Chapter 4 Design of Beams for Bending Stepped Columns and Columns with Intermediate Axial
General Considerations Load
Design of Beams by the Elastic Method Control of Sidesway >
Design of Continuous Beams Beam-Column Design Using LRFD
Web Buckling and Crippling
Shear Criteria Bolted and Riveted Connections
Strong versus W e a k - h i s Bending
Deflections Introduction
4-8 Biaxial Bending and Bending on Unsymmetrical Rivets and Riveted Connections
Sections High-Strength Bolts
4-9 Shear Center of Open Sections Factors Affecting Joint Design
4- 10 Design of Laterally Unsupported Beams Rivets and Bolts Subjected to Eccentric Loading
4- 1 1 Beams with Nonparallel Flanges Beam Framing Connections
4-12 Design of Bridge Stringers and Floor Beams Fasteners Subjected to Tension
4- 13 Composite Beams Connections Subjected to Combined Shear and Tension
4-14 Beam Design Using Load Resistance Factor Design Moment (Type 1) Connections
(LRFD) Load Resistance Factor Design (LRFD) for Connections

Welded Connections
Chapter 5 Design of Tension Members
General Conslderat~ons
5- 1 Types of Tenslon Members Weldmg Electrodes
5-2 Allowable Tension Stresses
Types of Joints and Welds
5-3 General Deslgn Charactenstlcs
Lamella Teanng
5-4 Stresses Due to h a 1 Load on the Net Sectlon Onentation of Welds
5-5 Des~gnof AISC Tenslon Rods Welded Connectlons
5-6 Net Sectlons Eccentrically Loaded Welded Connections
5-7 Deslgn of Tenslon Members
Welded Column Base Plates
5-8 Design of Bndge Tenslon Members Welded End Plate Connections
5-9 Cable Deslgn
Welded Comer Connectlons
5-10 Deslgn of Tenslon Members Usmg LliFD Fillet Weld Design Using LRFD
Chapter 10 Plate Girders
10- 1General
10-2 Loads
10-3 Proportioning Flanges and Webs of Girders and Built-up
Sections
Partial-Length Cover Plates
General Proportions of Plate Girders
Plate Girder Design Theory-AISC
Plate Girder Design Theory-AASHTO and AREA

Appendix Selected Computer Programs

A-1 Frame Analysis Program
A-2 Load Matrix Gcnerator for AASI-IT0 Truck Loading
on a Truss Bridge
A-3 Load Matrix Generator for AREA Cooper's E-80
Loading on a Truss Bridge

Index

The primary purpose of this textbook is to provide the basic material for th
course in structural steel design. The text contains elements of both buil
and bridge design for use in the structural engineering sequence of c i d
gineering programs. If the instructor wishes to emphasize building frames,
text is also suitable for an introduction to structural steel design in arc
programs.
Approximately equal emphasis is given to fps and SI units. In the dis
material both systems of units are used; the examples and homework pro
are in either fps or SI. This format was arrived at throu$ discussions wi
nurnber of interested faculty members and people in industry. The conse
was that the text discussion should continue to use both systems of units beca
transition to metric is not occurring as rapidly in the construction indust
other areas'of engineering. Dual usage seems necessary to provide both
and instructor with a feeling for what is a reasonable member size (num
deflection, or other design parameter in both systems of units.
Practical SI instruction requires use of design data, a n d since none
readily available, I have assembled a set of computer-generated rolled
section data tables as a supplement to the text. These tables are in ge
agreement with the AISC and ASTM A-6 specifications. This bound set of
also includes edited material from the AISC, AASHTO, and AREA
tions. It is intended that the textbook, together with the supplemental Stnict
Steel Design Data (SSDD) manual, will provide adequate material for a
design course without the need for any other reference material. T b e
material should be sufficient to enable students to design routine (an
not-so-routine) structural members in either fps or SI units and by any
the three steel design specifications which are most likely to control the deslgn

at least in American practice. Specialized problems are not generally addressed
1
in a classroom environment, and for these (as well as for design office practice
and other nonacademic work) the reader should obtain a copy of the latest
specifications from the appropriate agency.
I use the digital computer as a design aid in a somewhat interactive mode
(via batch processing) for the design portion of the steel design course. I have
found that the use of the computer in the steel design course is one of the better
academic experiences for students, because it helps them rapidly gain experience
in structural behavior. This may be by acci'dent (from mispunching data on the
modulus of elasticity, cross-sectional area, or moment of inertia of a member) or
by iteration of a design problem in which member sizes are changed as indicated
by the computer output. In either case, students readily see the effects of
member section properties on structural behavior. Using the computer programs
pe:;~ii'fs' thisb.with only a modest amount of work on the part of the student-
no program writing.
Several computer programs are listed in the Appendix to the text for those
o are not already using the computer as a design aid. These programs are
relatively simple, but efficient, and can easily be punched on cards for use on a
local computer system. The band matrix reduction method is used so that
computer cpu requirements are minimal. I can furnish these programs on tape at
the cost of tape, reproduction, and mailing.for anyone who is using the text in
the classroom.
I have not attempted to cite, or promote the use of, desktop programmable
calculators for simple tasks such as beam or column designs because of the
variety of devices available (e.g., HP, TI, Sharp, Casio, etc.), each requiring a
different programming method, and because of continuing rapid change in the
state of the art. Listing of the multiple programs necessary for use of the various
calculators would take too much text space, at the expense of more important
topics.
The text attempts to strike a balance between theory and "how to." The
topical treatment is not so exhaustive as to obscure the fundamentals but is of
sufficient depth that the reader is aware of the source of the design equations in
the various specifications. A number of the equations are partially to completely
! derived so that the reader can be aware of the limitations. A reasonably detailed
' in the later chapters. The use of simple examples gives the reader a quick grasp
explanation is given of the basic design problems; and the illustrative examples
are essentially step by step. With this format students should be able to cope
with the more complex design problems on the professional level, and to obtain
design solutions for the assigned home problems.
Appropriate references are cited directly in the text for topics for which
coverage is limited but which are sufficiently important that the reader may wish
to study the subject in greater depth. The inclusion of references will generally
be of more use to those in professional practice than to student users. My
expcr-icnce in teaching steel design for a number of years is that most students in
the ,first design course are primarily interested in learning how to design the
various types of structural members they will be assigned for home or laboratory
work. At this point in their professional development they are not overly
PXEFA
terested in the theoretical considerations and the extensive laboratory work
researchers and theoreticians that has produced the current design equations.
The complexity of semitheoretical and empirical design equations, couple
with the nature of structural design and its intimate association with desim
specifications and codes, makes i t necessary to take a strong '-how to" approa
in teaching steel design. I t is essential to present the user with a set
hypothetical (or real) data and by illustration produce a design. Students
presumed to have a sufficient background in the basic engineering and
sequence to appreciate what has been illustrated and are taught how to dup
the steps with a similar problem to gain confidence, and, based on the illus
tive problems, to extrapolate to a problem where the desim parameters
considerably different, with a minimum of super-vision.
Fabrication and practical considerations are introduced in the exam
problems as appropriate. Fastener spacing, edge distances, erection clearanc
standard gage distances, thread runout. and maintenance are considered
various sections. This should give the user an appreciation of fabricati
problems and other practical considerations. In conjunction with this, the te
has a large number of photographs, supplemented with line drawings of struc-
tural elements and connections, which should be of particular aid to the no
The reader should supplement these illustrations by observing steel frames un
construction. The photographs were all taken especially for this text, to disp
individual structural features as appropriate to the development of the disc
, sion.
Plastic design is introduced briefly in Chapter 3 together with the basics o
plate theory. This is done so that the design equations with origins in
design or.plate theory can be efficiently referenced back to Chapter 3, there
saving text space. Plastic design methods are not emphasized, for two basi
reasons: there is not enough time in a first course to adequately treat the subje
and elastic design seems to be preferred in professional practice.
I have deviated from the current textbook, trend to reflect the fonna
ibcorporated in some of the steel texts published in the 1950s. This
ibcludes the use of simple illustrative examples where the design data are
stated as well as more realistic design examples. These examples are anaI
Chapter 2 using the computer, and selected members are subsequently designed
of the general objectives of the discussion. More detailed design examples are
used to generate a sense of realism and to clearly indicate that steel d e s i g is not
just a matter of manipulating numbers. The examples are accompanied with a
reasonable amount of discussion of the analysis provided,
Within the framework of classroom time restraints, a steel design course
should be as realistic as possible. For this reason the user is encouraged to c a n y
iiny structural design problems assigned in Chapter 2 through succeeding
chapters, redesigning members as necessary and recycling the problem one or
more times for member sizing before the connections are desiqed in Chapters 8
and 9. A false sense of security regarding the actual complexity of structural
design, and even how the design loads are finally arrived at, can be developed if
 fl PttZFACE
$~$he>,seris simply given the loads for each design problem. Admttedly, the more
reall, IC design problenls require more physical and mental effort on the part of
"/
,
\
i'l
,'r'lcs;udent and more grading effort on the part of the instructor. This extra
~".ff~:t can be offset somewhat by assigning fewer total problems, but including
t~i;;s In which loads are glven, to bu~ldconfidence, and some with design
prc i !i31:is, to bulld des~gnsk~ll.
7 iie following text sequence might be appropriate in the semester system:
L1
, dmester hour:, Rapid coverage of Chapters 1 and 3, with Chapter 2
assigned for reading. Reasonable coverage of Chapters 4
to 10, Probably two wceks each on Chapters 4, 7, and 10.
I semester hours Rapid coverage of Chapters 1 and 3. Two weeks on
Chapters 2, 4, 7, and 10, followed by actual design of a
building frame and highway bndge truss, or industrial
building, based on the analysls in Chapter 2. One struc-
ture should be done In fps, the other in SI. A design
notebook should be kept, showing computations and
computer input/output. It is also suggested that this
work be done in groups, each wlth no more than four
students.

AC KNQWLEDGMENTS
Several persons and organizations have provlded considerable encouragement
arid assistance in produclng this textbook. First, I should llke to express my
,incere appreciation to Dr. Peter Z. Bulkeley, Dean of Englneenng and Technol-
jgy, Bradley University, who provided me with released teacbng time.
I would'also like to thank Mr. Andrew Lally and Mr. Frank Stockwell, Jr.,
of AISC, who provided me with a prelimmary copy of the new AISC specifica-
and took the time to go over the major changes with me. Mr. Lally also
;..-ovided useful ~nformationon maklng the SI conversions. Mr. Robert Lorenz
sf the Chicago Reg~onalOfflce, AISC, was also helpful in providing me with
.ast-minute corrections to the preliminary specification changes.
Both Bethlehem and US Steel corporations were most helpful ln providing
copres of their new steel section profiles, nearly a year in advance of their
becomlng official. This allowed work to proceed early on computer generation
,.; the Structural Steel Deslgn Data Manual tables. Particular appreciation is due
to Mr. Roland Graham of US Steel, who carefully revlewed selected portions of
he manuscript and the entire steel data manual and made some very useful
sl:ggestions.
Grateful acknowledgment is also made of the very considerable contribu-
tions of Dr. Eugene Chesson, Civil Engineering Department, University of
Delaware, who carefully reviewed both the preliminary and flnal text
manuscripts. Thanks are due Dr. T. V. Galambos, Civil Engineering Depart-
ment, Washington Univers~ty,St. Louis, who revlewed the load resistance factor (

design material.
a
 -1 TYPES OF STRUCTURES

e structural engineer wlll be concerned wlth the design of a v a n

structures including, but not necessarily I ~ m ~ t eto,
d the follo\wng:

Bridges: for railroads, highways, and pedestrians.

Buildings: including rigid framed, simple connected frames, load-beari
Fi ,?:re 1-1 The Eads bndge across the Mississipp~&ver at St. Louis, Missouri. This rallroad and cable-stayed, and cantilevered. Numerous lateral bracing schemes, incI
_'.way bndge completed in 1874 represents one of the first uses of steel (and high-strength F, = 50 trussed, staggered trussed, and rigid central core, may be considered or
: ? S ksl steel) in the United States for a major structure. The 192-m (630-ft) hlgh St. Louis
'It teway" arch, wlth an extenor s k ~ nof stalnless steel, can be seen In the background.
Buildings may be further classified as to occupancy o r height as
industrial, mill, high-rise, and so on.
Other structures: including power transmission towers. towers for radar an
installations, telephone relay towers. water supply facilities, a n d trans
tion terminal facilities, including railroad, trucking. aviation, and mari

In addit~on to the foregoing structures, the structural e n p e e r is

engaged in the design of ships, a~rplanes,parts of vanous machines a n d
mechanical equipment, automob~les,and dams and other hydraulic struc
including water supply and waste d~sposal.
This text will focus pnmanly on structural d e s ~ g nusmg metal, an
particular standard structural shapes as produced directly by the several
producers or in a few cases use of members that are built u p from steel p
and shapes and fabncated either by the steel producers or in local st
fabrication shops.

4,
1-2 DESIGN PROCEDURES
Structural design lnvolves application of engineering judgment to produce a
structural system that will adequately satlsfy the client/owner's needs. Next, this
system IS incorporated ~ n t oa mathematical model to obtain the member forces.
Since the mathematical model never accurately represents the real structure,
engineenng judgment is agaln required to assess the validity of the analysis so
that adequate a ~ ~ o w a n can
c e be made for uncertainties in both deformations and
statlcs.
Based on material properties, structural function, environmental considera-
ttons and esthetics, geometrical modlf~catlonsin the analysls model are made
and the solutlon process Iterated untll a solutlon is obta~nedthat produces a
satisfactory balance among material selection, economics, client desires/flnan-
cia1 ability, and various architectural cons~derat~ons. Seldom, except possibly in
the most elementary structure, will a unique solution be obtained-unique in the
sensc that two structural engineering firms would obtain exactly the same
82 @Qlutlon.
,# 'r
k, In structural englneerlng practice the designer will have available for p
w e use numerous structural materials, including steel, concrete, wood,
' posslbly plastics and/or other metals, such as alumlnu
"
'occupancy/use, type of structure, location, or other design parameter
dlctait: the structural material. In thls text we will assume that the design
proceeded to the point where the structural form has been decided (i.e., as trus
g~rder,frame, dome, etc.) and the several possible alternative structural materia
have all been eliminated in favor of using steel. We will then proceed with an
.?ddA~lonalanalysis required, and make the member selection and connecti
deslg~lappropriate to the topic being studied.
Textbook space and classroom t ~ m elimitations will of necessity reduce
the bare essentials the complexity of the design presentations. The reader shou
be aware that real design 1s considerably more comp
than the simplifications presented in the following chapters.
%) Safety as a design concern takes precedence over all other design consid
t q8, The "safety" of any structure, of course, de
?#&np. Since the structure is always loaded after it is built and not always i
t ~ e ~ k o dore manner used in the design, the selection of design loads is
pioblem in statlstlcs and probability. This part of t
rather subjective and produce extremely divergent designs had not bu
codes been developed (and in some form or another, almost universally
which place minimum required/suggested bounds f
is an important factor.
1-3 STEEL A S A STRUCTURAL MATERIAL
S i ~ e is
l one of the most important structural matenals. Properties of particular
*-ce In structural usage are h ~ g hstrength, co
BP%:t:rlal, and ductil~ty.Ducabiy 1s the ability
GENERALDESIGN COP.SIDER%TI
compression before fallure Other Important cons~deratl
e use of steel include widespread ava~labiiit:,~ n durability,
d particularly wt
odest amount of weather protection
Steel 1s produced by refining iron ore and scrap metals together
g agents, coke (for carbon), and oxygen in hlgh-tempe
aces to produce large masses of lron called "pigs" or "pig iron."
is further refined to remove excess carbon and other lm~uritiesand
r metals, such as copper, nickel, chromum, man
molybdenum, phosphorus, sillcon, sulfur, titanlum, columbium and vanadi
to produce the desired strength, ductility, welding, and corrosion-resis
characteristics
The steel lngots obtained from this process are pasaed between two roU
pposlte directions to produce a semiflnis
called either a slab, bloom, or billet, depen
nal area. From thls point the product 1s sent to other 10
ills to produce the final sect~ongeometry, Including structural shapes as
plates, and pipes. The process of rolling, in addi
red shape, tends to lmprove the m a t e d properties of to
malleability. From these rollln:: mils the structural sh
are shipped to steel fabricators or warehouses on order.
The steel fabricator works from the englneerlng or architectural dra
produce shop detail drawings from which the requlred dlmenslons are
to shear, saw, or gas-cut the shapes to sue and to accurately locate
drilling or punching. The origlnal draw~ngsalso indicate the necessary
finishes to cuts. In many cases the parts are assembled in the shop to det
if a proper f i t has been obtained. The pieces are marked for ease of
identification and shipped rn pleces or subassemblies to the ~ o site b for erec
general contractor.
Some of the most important structural properties of steel are the follo
1. ~ ~ of e[astlcrp,
d E~ The typlcal
l ~rdnge for
~ all steels (relatively l n d e ~ e
The value for deslg n as 29 000 ksl or 200 000 MPa-
2. Shear modulus, G.
E
G =
2(1 + P )
where p = Poisson's ratlo taken as 0.3 for steel. Using ,u = 0.3 9ves G =
11 000 ksl or 77 000 MPa. //
3. coefficlenf of expnnsron, cu T h e coefficient of expansion may be as
a = 11 25 x per " C
AL = a(T, - T,)
L (ft or m depending on length L )
 GENERAL DESIGN CONSID

in degrees Celsius. To con

eit to Celsius, use
C = ; ( F - 32)
and ultimate strength. Table 1- 1 gives the yield poin
es of steel of interest to the structural designer that are

mmd;z
er properties o/ some interest. These properties include the mass de
y j o9m Q
0 0 -w (1 t = 1OOO kg); or in terms o
3$2&Q 76.975 kN/m3. The specif
conversion of fps units
k? m
0 N its of kN/m and kg/m is accomplished as follows.
;aa
-8 .- 9 9
Given: lb/ft and required to convert to:
4
3
g 3 ggmz
5)
a
rd
"0
-
?
*9' 0
o
+ 0 m m
+ ? Y r :
2 .O m w - 0 0 m
.. Z
m
$ 2
k;"
5 2 88 $898
Note that lb mass and lb weight or force have been used interc
. m
I
o m - o ably in the fps system because the acceleration-producing force
2 m Q W W t -
of gravity. This cannot be done in the SI system, since the ne
2 S: , 8 2 3 derived unit that defines the force necessary to accelerate a
*, P N N N m m
.
d
'1 m/s2. The acceleration due to gravity is approximately 9
2 8 %Q%Z xample: Given: A rolled structural shape weighs 300
m
u
"
.L

a
Q)

"
p 2.z
ba
Required: mass/m and weight/m.
Solution: mass/m = kg/m = 1.488164(300) = 446.
% <.ga wei&t/m = kN/m = 0.0145939(3
8%3$
3
-
.V1
L
k 4-44

B
E
V1
M
$i nh
I
3
aJ
g3
g 7;
E"
Z; -E 6 s
8 5 re rolled into plates of varying
round, square, and rect
Most of the rolling is done on hot steel, with the p
steel." Sometimes the thinner plates are further rolled
' steel products. Several of
following sections.
 b S l RUCTURAL STEEL Ukbiut4

1 4 . 1 W Shapes

X,e most commonly used structural shape is the wide-flange or W shape. This is
a doubly symrnetncal (symmetrical about both the x and y axes) shape consist-
ing i;f two rectangular-shaped flanges connected by a rectangular web plate. The
flange faces are essentially parallel with the inner flange distance for most of the
graiips, with a constant dimension.? There is some variation due to roll wear
and other factors, but the distance is held constant within ASTM tolerances.
h e shape is produced as illustrated in Fig. 1-1.

Khe dssignation: W16 X 40 means a nominal overall section depth of 16 in with

a weight of 40 Ib/ft.
, The des~~narron:W410 X 59.5 is the same W 16 as above with a nominal depth in
1 rr~m(based on the approximate average depths of all the W16 sections and
rounded to the nearest 5 mm) and with a mass of 59.5 kg/m.

&or to 1978, at least ope W section in a group designation was "exactly" the
t:)$kpnal depth given (i.c., one W16 was 16.00 in deep; one W18 was 18.00 in
"';,$@p).
, Now the closest W16 is the W16 x 40, with a designated depth of 16.01
$0: There can be substantial deviations between the nominal and actual depth
.
. .(<.g., the W21 ranges from 20.66 to 22.06 in). For the W14 the SI equivalent is
W360, but the actual range is 349 to 570 mm (in this case the "average" was too
. fr*
'
'from the nominal value and W360 was somewhat arbitrarily used).
It should be noted that the rolled product will contract on cooling and at a
vat~ablerate depending on the thickness at any point on the cross section. The
rolls used to produce the shapes will undergo wear, and coupled with the
enoinlous forces involved in the rolling process, only shapes of nominal dimen-
sion (varying from theoretical or design values) can be produced. American
Soclety for Testing and Materials (ASTM) specification A-6 in Part 4 gives
a:lowabie rolling tolerances, including amount of flange and web warping and
deviation of web depth permitted for the section to be satisfactory. Generally,
the maxipum permissible variation in depth as measured in the plane of the web
is 5 in or 3 mm. Note, however, that the permissible difference in depth
between two rolled beams with a theoretical depth of 16.01 can produce extreme
depth:; of 15.885 to 16.135 in or a difference of 4 in or 6 mm. These variations
should be' kepr rn mind, particularly when converting to SI dimensions for
detailing, clearances, and mating of parts.

1-4.2 S Shapes
These are doubly symmetric shapes produced in accordance with dimensions
adopted in 1896 and were formerly called 1 beams and American Standard
%

t The several sections with a constant nominal depth. Where a group consists of a large number
of S~CIICJI~S,
a second inner flange distance may be used.
hd
i
C
---r
--- -

-.
-
'V shapes
:;gc
S \hapes
Anierrian Stdndard
[
C shape,
Chdnnel
L shape
kqiral leg angle
to 330 MPa and refer to Figs 1-30 and 1-36. Similarly, A-44
point of 345 MPa, will have a yield strength on the order of
bean1 (I-beatri)
guaranteed values converge.
i UReitdllgular

I
"_,;id
L-shdpe 7 slldpe
8 Square
0Rounds
C 1 1 ~ q u dleg
l dtlgle Strut tirrel Tec
i l l ! Iron1 \V.stidpc Ihrt

1
Plate

k + $ i s 1-2 ktmctural shapes as drrectly produced by Be steel producen. ed and designated A-272 (described ~n ASTM specification A-272). Sp
on ASTM A-440 was wntten in 1959 for another h ~ g hstrength steel
t--:,.j L Shapes 1960 with application to weld~ng.All three of these steels have a yield
i i . e , ~shapes are either equal or unequal leg angles. All angles have parallel at is dependent on the thickness of the metal, as shown In Table 1-1.
flal~gefaces. Angle leg dimens~onscan vary on the order of +- 1 mm in width.
An L6 x 6 X IS an equal leg angle wtth nominal dimens~onsof 6 m and a
2thLE~nessof $ In.
An L89 X 76 x 12.7 IS an unequal leg angle w ~ t hleg dimensions of 89 and
76 k:m, respectively, and a leg thickness of 12.7 mm (L3f x 3 x +).
t
1-4.'; T Shapes
St: LC l are structural members obtatned by splitting W (for WT), S (for
~ r a tees
S"
WI.:1
.. M (for MT) shapes. Generally, the spltttlng is such to produce
one-half the area of the parent section, but offset splitting may be
d .: -r tze section is required. Published tables of T shapes are based
slA.$~netncal splitting. No allowance is made for material loss from splitting
p21 t .:t shape by sawing or flame cuttmg.
4 WT205 x 29.8 is a structural tee with a nominal depth of 205 mm
m: c*c of 29.8 kg/m and is obtained by splitting the W410 x 59.5 section (f
R'IG x 40).
Exera1 rolled structural shapes are Illustrated in Fig 1-2.

1-3 STEEL STRENG??-I

All me1 design takes into consideration the yield strength of the mate Figure lJa Typical stress-stram curves for E;lgure 1-36 Erllargement of lmti
structural steel. stram curve for two grades of
yleld strength of several grades of steel available for design is given in Table 1- Note that the plasbc r a g e u
yield strength is that minimum value guaranteed by the steel producers an
 .- ...--- . G E N E R U DESIGN CONSIDERATIO

Since about 1964, specifications for several other high-strength (low-alloy) "C x 100
steels-hzve been- incorporated into ASTM specifications as A-572 and A-588.
Table 1-1 shows that the steel covered by the A-572 specification covers several
$yield strengths, termed grades, such as grades 42, 45, 50, 55, 60, and 65 for the
corresponding guaranteed minimum yield stress in ksl. Generally, the yield
{strengths of these newer steels are also thickness-dependent, as shown in the
table under the heading "plates and bar thickness." The steel producers have
designated the several W shapes into five groups, depending on flange thickness
(and as shown in Tables 1-1, 1-2, V-1, and V-2)t, compatible with the steel grade.
The designer merely has to check these tables to see if the shape is available in
the raquired/deslred yleld-stress grade. For example, in the 450-MPa grade,
only shapes In group 1 qualify from flange thickness. W18 shapes are available
in group 1 only from 35 to 60 Ib/ft inclusive (the five smallest sections and with
a maximum part thickness of 0.695 in).
Specification ASTM A-588 allows F, = 345 MPa for a high-strength low-
@lo$ steeltwhich may be up to 100 rnm (4 in) thick. The steel covered in this
@$hPecifl,rttion is primarily for welding and is corrosion-resistant.
in terms of cost/unit of mass, the A-36 steel is most economical. High
stre2g:h steels have principal application where the stresses are primarily 14 Effect of elevated temperatures on either y e l d or ult~matetensile strength ucpresse
High-"irength steel beams may deflect excessively, owing to reduced f strength at room temperature of approximately 70°F
rnodul:.~. The high-strength steel columns may be less economical than
steel tf the slenderness ratio ( K L / r ) is large. Hybrid girders that use
strength steel in the flanges, or built-up columns using high strength steels
provide better solutions where member sizes are restricted. In a given cas
necessary to perform an economic and availability analysis to determi
suitability of using high-strength steel.

/ TEMPERATURE EFFECTS ON STEEL

1-6.1 High-Temperature Effects

Steel is not a flammable material; however, the strength is heavily temperature-
dependent, as illustrated in Fig. 1-4. Both the yield and tensile strength at
1000°F is about 60 to 70 percent of that at room (about 70°F) temperature. The
drop in strength 1s rather marked at higher temperatures, as shown on the figure,
where the strength at 1600°F is only about 15 percent of that at room tempera-
ture. 1-6.2 IAV-TemperatureEffects
Steel frames enclos~ngrnater~alsthat are flammable will require fire protec-
tion to control the ternperature of the metal for a sufficient time for the Brittle fracture is a failure often associated with low temperatures. Essentially,
occupants to seek safely or for the fire to either consume the flammables or be brittle fracture is failure that takes place without material yielding. The
extir~guishedbefore the building collapses. In many cases the building does not stress-strain curves of Fig. 1-3 indicate that m the usual failure of a tensile
specimen, considerable elongation takes place. As a matter of fact, a minimum
t See footnote a of Table 1-1 In J. E. Bowles, Structura/ Steel Desrgn Data Manual, McGraw-Hd, percent elongation is specified for steel in the ASTM standard tensile test.
New York, 1980 Inlplicit in steel design is the resultant deformation (yieldixig) of the material
 Fireproofing materials I Pcf
Unit w e ~ g h t
k~/rn~
Cinder concrete l 10 17.3
Gypsum board 30-40 4.7-6.3
plaster, cclnetit and sand 100 15.7
Expanded shale concrete
Vertnicul~te
Perlite ,.,
I I
Usually use 35* nim o f flreprotection for 2-h fire rating;
obtain specific thickness values from either tests o r from
the producers o f gypsuln, perlite, etc.

rigure 1-5 Methads of producing fireproofing of structural steel members. ( a ) Sprayed fiber. (6)
Lath and plaster. (c) Lightweight concrete (formed). ( d ) Gypsum board-use boards to build
'hick;us. ( e ) Corner detail of sprayed on fireproofing. Thickness is built in several sprayings. (fl
Bear. dzwl of sprayed fireproofing.
16 STRUCTURAL STEEL DESIGN
any abrupt change in cross-sectional area-to inhibit lateral contraction
tension situation.
brittle fracture. This may initiate as a crack that propagates to a member f
ntity, frost depth, and average wind velocities.
occurring is of little aid in settling the resulting damage claims that are sure
follow. Brittle fracture can be controlled in several ways:
1. Detail niembers and their connections to minimize stress concentrations.
2. Specify the fabrication and assembly sequence to minimize residual ten
stresses.
3. Use steels that are especially alloyed for low-temperature environments
tures are encountered. . .
5 , If possible, machine (or grind) the notch into a sm
6. P-educe the rate of tensile strain application.
1-7 STRUCTURAL DESIGN CODES . and Steel Institute (AISI), 1000 Sixteenth Street,
Lo.~,d.lbuilding departments almost always require structural de
The various state departments of transportation
generally use the specifications put forth by AASHTO,
several railroads generally use specifications put forth by
The structural designer doing highway or railroad work
closely the design specifications of these publications, partic
government is involved with any of the financing.
the owner/client may require a more stringent design than the building co
c,piteria. Only in rare cases can the designer get a variance from the local
governing body to deviate in a less conservative manner from the code. Vari-
GENERXL. DESIGN CONSIDERATI
tructural design. On the one hand, it sometimes takes
ials and methods; on the other hand, th
fast." If the local .building code is care
minimum design requirements met, or exceeded, and a catastro
of exists that good engineering practice has been followed.
ng codes are supposed to reflect that part of the structural
unique for that locale, such as temperatures, earthquakes,
list gives several design codes and/or specifications
may have occasion to use:
Code, published by, and available from, the
iation, 85 John Street, New York, N.Y. 10038.
Building Code by International Conference of Building
Workman Mill Road, Whittier, California 90601.
lding Code, Building Officials and Code Administrators Inte
East 60th Street, Chicago, Illinois 60637 (formerly BOCA).
itute (ANSI), Minimum Design
ngs and Other Strucrures, ANSI 58-1, 1430 Broadway, New Yo
titGte of Steel Construction (Specifications), Steel
th ed. (1979), 101 Park Avenue, New Y0rk;N.Y. 1001
lding Society (AWS), Structural Welding Code, 2501
iami, Florida 33125.
. Publishes various specifications for using iron and ste
Association of State Highway and Transportation Of
l.ITO), Specifications for Highway Bridges, 341 National Press Bu
Railway Engineering Association (AREA), Speci/ica
ay Bridges, 59 East Van Buren Street, Chicago, Illinois 6060
national and city codes use specification standards as applic
zations, such as AWS, AISC, and AIS
cies for the other construction materials.
nd AISC, as well as the steel producer
tables of structural shape design data as well as data on other steel
, wire, and bolts. Certain of these data together
AISC, AASHTO, arid AREA, have been pre-
tnrctzlrnl Steel Design Data (SSDD) man
by McGraw-Hill Book Company and used as a supplement with
manual was developed so that both fps and SI units would
18 STRUCTURAL ST GENLRAL DESIGN CONS1

carry the dead 1 fps: R = 0.0008 x area (when area > 150 ft2)
SI: R = 0.0086 x area (area > 11.2 m2)
include:

Ceiling materials, including duct work for environmental co
supplies.
-xterior walls supported by the frame, including windows, doors, and ba
Interior walls Qat are permanently placed.
'iechanical eqmpment (heating, air conditioning, ventilati
' (such as elevators, including cage, cables, motors).
r .reproofing.
Beams, girders, and columns, including the footings making
frame.
(some codes limit R < 0.40 for horizontal membe
here R= reduction factor used as ( 1 - R ) x L,,,,,
D = dead load, psf or kPa (kllonewtons/m)
L = live load, psf or kPa, but L 1s limlted to not over 100 psf or
generally, values larger than this are not reduced
lic assembly (such as auditoriums), garages, and roofs.

IFrom this list it is evident that any part of the building w

nstalled contributes to the total dead load. Dead loads can
Example 1-1 A port~onof an office (multistory) floor plan is shown
El-1. The floor is 4-111concrete on a metal deck over steel bar joists. W
the reduced live load for the floor beams and for an exterior column
floors down from the top floor?

prescribed by building codes based on occupancy and 1

anow, and earthquake loads are considered. In addition to
!oads include:

People, as in auditoriums, assembly halls, and classrooms.

Movable room partitions.
Office equipment and production machines if they are m
Warehouse products.
Furniture.

Building code values of live loads tend to be based on

SOLUTIONEstimate the dead load on the contributory (centered

ber) floor area as:
Concrete floor and finish: 4 x 144/ 12 = 48 psf
computational convenience and because the actual buildin
Ceiling, metal deck, steel bar joists = 12 psf
:lot known.

f As p e n m the several natlonal b u l l l n g codes c ~ t e dearher

 GENERAL DESIGN COKSID

El-1, 18 X 22 (area ABCD):
R = 0.0008(18 X 22) = 0.32
+
From Table IV-4 of SSDD, the live load = 5.00 kPa. The r
r for a grder based on a contnbutory area as shown is
R = + 8)/2 X 91 = 0.774 > 0.40 (and also 0.
0.0086[(12
< 0.60
R D +L
= ----- = 3.703 + 5.00 = 0.402 < 0.60

4.33 L
-
R = ----- = 6o 8o -
+

4.33(80)
- 0.40 < 0.60 4.33 L 4.33(5.00)

Use the smaller value of R computed, 0.32. The reduced live loa Since the problem statement limlts live-load reduction to not mor
0.40, the reduced l~veload is
L' = (1 - 0.32)80 = 54.4 say 55 psf
Compute R for the column; the contributory area is centered on the column L ' = ( 1 - 0.4) x 500 = 300kPa
of 9 X 22 (AEFD); but for the accumulation of three stories, we have
R = 0.0008(3 x 9 x 22) = 0.475 < 0.60 O.K.
60 80 - 0.40 < 0.60
R = -------
+

4.33(80) d loads have been extensively studied in recent years, particularly for larger
Using the smaller value of R, 0.40, the reduced live load on -rise structures. Generally, for tall structures ~wnd-tunnelstudies should bz
L' = (1 - 0.4) x 80 = 48 psf or smaller regular-shapsd
We note that O.$0 is the maximum R for the column and is 30 m, the wind pressure
uppermost floor level. ory to use. The Nation

Example 1-2 A meeting/banquet room in a hotel has

22 X 27 m. The floor is 125 mm of concrete with a tile surf

15 0.75

d this pressure times

SOLUTION Note that a public room is not the same as

,whece the loading is pnmarily seating in fixed or movable
definition we may use a live-load reduction factor.
First, estimate the dead load using Table IV-3 of S
weight of concrete = 23.5 kN/m3. o allowance is commonly made for the shielding effect of adjacent structure
r from ground cover.
Weight of concrete: 0.125 x 23.5
ge ground level at th

The wind pressure is commonly computed between floor levels and prorate
he adjacent floors using simple beam theory if the vertical distance compare
 The sever31 wind values are shown in Fig. El-36. The data display is
convenient for computer programming for frame stresses using the com-
~uter-gr,ogramdiscussed in Chap. 2.
///
Wind pressures can be approximately computed as

where V is mi/h or km/h. This equation is readily derived as q = f mu2, where

-
the mass density of air is approximately 0.00238 lb s2/ft4.
Since wind is a transient load, the building codes usually allow a one-third
increase in the allowable design stresses with wind included as a part of the load
condition as long as the required section is not less than required in the load
condition of dead + live loads alone. For example, if a stress of 20 ksi is
allowed, then with the wind load condition a stress of 20 X 1.33 = 26.6 ksi could
be used.

1-8.2 Snow Loads

Snow loads are live loads acting on roofs. Snow and any other live loads are
taken with respect to the horizontal projection of the roof, as illustrated in Fig.
1-8. Figure 1-9 is a map illustrating snow loads that may be used in the absence
of spzcific load building code requirements. Even in areas where snow loads are
minir~~al, a minimum roof live load should be.used. The NBC stipulates the
larger of the snow load or 20 psf or 1.0 kPa. Since 10 in of snow approximates 1
in of water, a 20-psf snow load corresponds to a roof snow depth of nearly 40 in
-easily obtained where snow drifting occurs. When rain later falls on snow,
however, the saturated snow weighs considerably more and the unit weight can
approach that of water.

In addition to the types of pressure or area loads noted, building codes may
Snow and other live loads stipulate checking for a concentrated load of some magnitude which may be
lilllllllllllllllII1111111111' placed anywhere on the floor or roof. Where roofs are used as recreational areas
or sun decks, the live loads must be adjusted to values based on occupancy in
addition to considering snow and/or wind.
Po~idingis a special roof load that may require investigation. Ponding is a
condition where water collects on a flat roof which has deflected locally
(possibly due to an overload, poor construction, foundation settlement, or
plugged roof drain), causing a concentration of water which in turn increases the
load and deflection, causing a further concentration of water. Noting that a
water depth of 1 in results in a live-load pressure of 5.2 psf, loads are readily
eveloped which can locally fail roof members. Through progressive failure, the
roof may collapse. Ponding design is considered in some detail by Marho
Figure 1-8 Snow and other roof live and dead loads.
July 1966 AISC Engineering Journal.
 Erection loads are not directly considered in biiildiiig c o d ~ s 'rhese
. loads
may control the design of certain members, particularly very high rise buildings,
cantilevered bridges, or cable-supported structures. The engineei responsible for
esy phase of the erection may be held legally responsible for damages or loss of
life resulting frorn a structural failure during erection. Most structural failures
(at least that are reported) tend to occur during erection rather than later.
'
Erection methods andequipment tend to vary from project to project; thus ii is
'

not practical in textb~oksto do more than point out this very important design
area. The engineer musf+d$termine what equipment will be used, where it is
placed, loads to be lifted; quantities of material, and the storage locations,so
tLat the affected individual steel structural members can be checked for ade-
quacy using princip
chapters on design.

1-9 HIGHWAY AND RAILROAD BRIDGE LOADS

'The American Association of Highway and Transportation Officials (AASHTO

stacdard truck loadi

General lucarion of rndxrrnurn momsnr

is oLtained with eith
ti,c i l S truck and span lengths, shears are as follows:
, ,,, ,*,. ,. . .
),
31 kips IJJ kN
-. - 24 k ~ p s108 k N
Bridge span I6 kips 7 3 kS
---
33.8 to 145.6 ft M,., = :[(0.9~ + 4.206)(0.5L + 2.33) - 1l 2 L ] It - kips

26 ?a :27.5 ft

2.53 to 38.86 m

Over 127.5 ft

Over 38.86 m

a W = 40, 30, and 20 kips or equivalent in kN (and is the basic truck load, not the

total).

&re 1-10 Standard AASHTO truck loadings for bridges.

 GENERAL DESIGN COWIDEIW

Moments, shears, and floor-beam reactions for Cooper's E-81) loa

al axle lodd
nd of train:
sfS'RUCTURAL STEEL DESIGN
~ a b 1-2
~ d (Continued) . ,

479.6 282.0 128.1

. ! 23 712.0~ 17 990.0 522.0 306.8
,. 3 35 118.0b 27 154.0 626.4 367.3
.-.' 1 48 800.0~ 38 246.0 729.3 426.4 197.9 1225.3
- .. . -
.U1 values shown are for one rail (one-half track load). Axle loads shown in diagram. Obtain total uniform loading, however, must not be less than the following:
,; . . :s for other E loads by proportion. ,,
I. .4t center of span; other moment values are usually close to center of span, so one may obtain
, !utal moment as the sum of w ~ ' / 8 for dead load + live load value shown in the table.

f ::-cd on the locomotive weight, the Cooper load is designated as E-40, E-50, 2. AREA wind requirements:
E.50, E-75, E-80, or E-110 and is directly proportional (i.e., E-60 = $XE-80). The
cl:rrent AREA design criteria are based on the E-80 (sometimes E-110) loading
s.:dwn in Fig. 1-11. Table 1-2 can be used to obtain the bending moments and Pressure, force/area
s.+c.;;rs at selected locations for girder bridges, with values given for a single rail Unloaded span Loaded span
fbading (based on one-half the axle load shown in Fig. 1-1I). fps, psf SI, kPa fps. psf SI, kPa
Where multitracks or road lanes are carried by the bridge, the live load is as
ici!ows:

-
Percent of live load

Lane or track AASHTO AREA

2 100 100
3 90 2X100+1X50
4 75 2~100+1X50+1X25
XI.xe than 4 75 As specified by designer

Other bridge loadings that must be considered include impact, wind, and
longitudinal forces. Impact and longitudinal forces allow for dynamic effects
from rolling equipment going across as well as for starts and stops made on the
bridge. Impact will be considered in the next section. The wind force is
>elf-explanatory and in the case,of a loaded railroad bridge, the wind against the
train may be'a substantial load.
AREA: 0.15 X live load (without impact).
Other loadings that may require consideration include differential tempera-
tures between top and bottom flanges or chords, ice and snow loads, possibIe
overloads, and for continuous bridges, support (pier) settlements.
 .- - -- .- - . llr, G E N E W . DESIGN CONSIDERATIONS
....,-.., -
~ L O I U I Y
<..*,,., 6. ..4 ..'i

Railroad bridges make a distinction between those bridges which consi nt; must be divided by 100 to use in de

een centers of single or groups of longitudinal s

which frame into transverse floor beams or girders, or between
trusses or girders, ft or m
= length between transverse floor beams or between supports as app
cable, ft or m
beneath the ballast. e 90 percent of I, computed above for ballasted-deck bridges.

Example 1-4 Given a highway truss bridge with HS 20 loading. The truss
1-10 IMPACT LOADS anels are 7.5 m, and the distance between reactions is 37.5 m. T h e distance
etween trusses (width) is 14.1 m. What is the impact factor, I!?

OLUTION The impact factor will vary for the floor beams, stringers, and
impact load as russ, depending on their lengths. For the stringers the impact factor is
l5 = 0.330 > 0.30 therefore. use 1/ = 0.30 //

Item
ilroad bridge consists in two trusses spa
sses are made up of seven panels at 27.60 ft/
Elevator loads 1.00 at is the impact factor?
Macbnery and other moving loads > 0.25
r,unoN Since L > 80 ft,
L = 7(27.60) = 193.2
600 = 25.6 percent
193.2 - 30

The AASHTO impact requirement is Example 1-6 What is the impact for the floor beams of the AREA truss of
Example 1-5? Floor beams are transverse members connecting the two
f PS SI trusses at panel points.

I,
-
=. - 2< 0.30
I,+ 125 - I, -- l5
L+38-
i0.30
S = 27.6 ft, L = 17 f t
SOLUTION < 80.
3(17)*
+ 40 - -1600
= 43.1 percent
where L is the length of span or portion of span that is loaded, in ft or m. The
AREA impact specifications depend on the rolling equipment. For diesel and
electric locomotives and tenders:
HQUAKE LOADS
general trends. One is to attempt to model tfie
asses and springs and use a digital computer to
s assumed earthquake accelerations- The
the earthquake accelerations based on earth-
citation based on building geometry, and apply
 GENERAL DESIGN

roof

Elevation
Figure 1-12 Earthquake zone map for the Umted States. (After Unrform Bu~ldngCode, 19

Example 1-7 A 10-story apartment building with basement is as shown

Fig. El-7a. The exterior is insulated curtain walls and thennopane windo
with an estimated weight of 15 psf. The interior walls are generally stud
partitions plastered on both sides with insulation between apartments. Use
4-in concrete floors (tiled or carpeted) on corrugated metal pan supports
carried by open web steel bar joists. The building site is in Memphis,
Tennessee. Estimate the earthquake force and corresponding story load.

SOLUTIONEstimate the roof and floor dead loads as follows:

... Alloaf....
Wood sheathing = 3 psf
5-ply felt roofing = 7 psf
Ceiling and bar joists = I1 psf
 . .. -. .,. G E N E W DESIGN CONSID
,,.,.'.b..* " L 9
Any floor:
Partitions in 40 X 30 apartment at 8-ft height and
cross-walls at 20 psf .
,I
.4
gives: (40 x 2 + 30 x 2)(8)(20)/(40 x 30) /

4
Floor: -(144) (concrete)
12
Ceiling (estimated)
Bar joists and metal pan
Exterior wall at 10-ft height
(2 x 40 + 2 x 90)(15)(10)/(40 x 90)
Total = 93.5 psf
Total floor weight = 0.0935(40 x 90) = 336.6 kips
These weights are illustrated in Fig. El-7a. For easi
weights of 76 and 337 kips, respectively, for remaining wo
Fig 1-12, the Z factor is 0.75. Take I = 1.00; take K = 0.
Table 1-3. The total building weight = 76 + lO(337) = 3446
The earthquake force in the E-W direction is computed
D=40ft
T = : 0'05(100) = 0,7906 Figure El-76

Since the frame is of steel, the alternative computation for period is

T = 0.1 x number of stories = 0.1(10) = 1.0 s
The author will average the two values of T to obtain T = 0.895 s. Summing the horizontal floor loads and including the top value of 13.7
1
C = ----- =
1 218.59 versus 218.6 kips as a check. These lateral floor loads
= 0.0705
1 5 r ~ 15- further prorated to the several bays in the E-W direction for the fr
Substitution of this accumulation of factors/weight into Eq. (1-4) analysis load condition(s), which includes earthquake forces.
F = 0.75(1.0)(0.80)(0.0705)(1.5)(3446) = 218.6 kips
The roof value is 12 FATIGUE
F,,, = 0.07TF = 0.07(0.895)(218.6) = 13.7 kips ( T > 0.7 s)
The story loads are found using hn = distance grou failures which have been attributed t
follows: Z W,hn = 337(90) + 337(80) + 337(60) + . . ue. Fatigue failure is a material fracture caused by a sufficiently Iarg
151 650 f t . kips ulsating stresses, or stress reversals. Th
ture of the material at a location where
F = (F- F )-=Wn A, (218.6 - 13.7)
lop 2 W,hn oscopic in size) exists. A crack form
For tenth floor nding on stress level, rapidly or slowly (sometimes so slowly that the
re) progresses to failure of the part
F,, = 0.001351(337 X 90) = 40.98 kips (shown in Fig. El-7b)
F, = 0.001351(337 X 80) = 36.42 kips Most metals tested under repeated or cyclic loadings display stress r
' F8 = 0.00135 l(337 X 70) = 3 1.87 kips s qualitatively illustrated in Fig. I-
rly, these curves were commonly displayed as stress level versus cycIes.
F, = 0.001351(337 X 10) = 4.55 kips
t, the stress range is used as the parameter of interest. The stress range
4(8 STRUCTURAL STEEL DESIGN GENERAL DESIGN CONSIDERATIONS 41

to !@I 100 to 500 500 to 2OCO Over 2

Base metal with rolled surfaces T or R e p 60 (415) 36 (250) 24 (165) 24 (165)

plates and shapes with full or

of girder webs or flanges adja-

cent to welded transverse stiff-
T o r Rev 32 (221) 19 (131) 13 (90) lob(
Figure 1-13 Qualitative plot of stress range F,, versus number of cycles to failure. Base metal at end pr partial
. .-",., ... ',"....-,. -
3.

T or Rev 2 1 (145) 12.5 (86) 8 (55)

anically fastened connections
can be defined as

T o r Rev 45 (310) 27.5 (190) 18 (124) 16

TorRev 27(186) 16(110) 10(69) 7(48

The AISC, AASHTO, and AREA specifications are very nearly .identical
(where rnembcr failure is not catastrophic) in specifying the stress range and
number of stress cycles. These specifications are based on a large number of
fatigue tests performed (see Fisher, "Fatigue Strength of Steel Members with T or Rev 45 (310) 27.5 (1%) 18 (124) 16 (110
Welded Details," AISC Engineering Journal, No. 4, 1977). Typical stress range
values which may be used for all three specifications are given in Table 1-4. The T or Rev 45 (3 10) 27.5 (190) 18 (124) 16 (110
reader should consult the Structural Welding Code, Sec. 9-14, which is the ori
of the material used in the three specifications or the appropriate des
specification for a more complete presentation of fatigue cases and F,.
The largest value in each cycle category in Table 1-4 is generally applicable
to buildings. Lesser values than shown are necessary for reduced sections, T or Rev 21 (145) 12.5 (86) 8 (55)
certain types of joints, type of joining material, and for certain members in
industrial buildings. The AISC (Appendix B) manual, AWS Structural Welding
Code, the AASHTO specifications (Sec. 1-7.2) or AREA should be consulted for
those situations where fatigue must be considered. Note that fatigue is not
usually considered with wind or earthquake loadings on buildings. Fatigue is
se 12 ksi or 83 MPa for girder webs.
usually not considered for routine building design, since 10 load cycles/day over
_ . . _.
. . - " .,..... ___-________ .__.
--- -.--.--------_I-

N = 10 x 365 x 2 0 = 73 OM ,,
This is seldom enough cycles of whatever thk;,itress.rang6 to require a reduction
.A STRUCTURAL STEEL DESIGN

Example 1-8 A rolled section will undergo an expected 1 X 106 cycles of

stress over the design period of the structure. The stress analysis gives
f,, = P,,,/A = 16 ksi; f, = - P,,/A = - 12 ksi. The basic stress'for
this member is Fa = 22 ksi (tension) and 16.5 ksi (compression). 'The
structural configuration limits F, = 24 ksi in the base metal. Is the section
satisfactory?
S l u l r ~ p l sb ~ bent
y
SOLUTION
f,,=16-(-12)=28ksi>24ksi N.G.
The section is inadequate for this number of cycles; increase the section so
that f,, < 24 ksi. Note that the section is adequate for "allowable" static
Tal)crcd CUILIIIIII .lnJ r ~ k1r
stresses. i n a r i g ~ di r ~ n hcnt
~ r

1-13 STEEL STRUCTURES

Slructures of steel include bridges, buildings, trans

sign supports, and even art objects. The primary focus of this
asid buildings, since these are the most common projects invol
Buildings are commonly classified according
:,ti?ry buildings. Little use is made at present of s T ~ L I ~ S -soli~rnn
on
.: zpt in multistory apartments. .

'L-13.1 Industrial Buildings 1-14 Several bents used in steel building frames.

Iildustrial. buildirigs are commonly one- or two-story structures

fnr industrial (such as manufacturing, storage, or retail/wholesal
and institutional (including schools, hospitals,
Other structures may include gymnasiums, aren
tr: rlsportation terminals (land, sea, and air). Thes
frame, as illustrated in Fig. 1-14, or have a roo
resting on load-bearing walls (see Fig. 2-4). The
rnay"8e~"?lgidor pinned; may be a two or th
t;uss-on-column system. The truss may be rig
frames under construction are illustrated in Fig. 4-1.
A building frame is a three-dimensional skeleton b
rigid in only one plane. Some buildings are rigid in both t
but this: type of frame will not be considered
resulting' from considering only the principal frame
termed a bent and may be one or more stories in
iiiusirate; terms defined here and later). The ter
vlhether'rigid, truss-on-column, rafters-on-colu
to span between columns in the principal pla
the third dimension is the bay spacing. Span
1-15 Additional terms used to identify structural members in industriai buildings.
 GENERAL DESIGN CONSIDEIUTIONS

h type, where traffic passes between the trusses. The deck-type trus
preferred if clearance beneath the truss is not a factor, because pr
Many truss bridges combine both types (see Fig. 1-19) o
of a truss for the longer spansand girders for the short
s a common practice. This latter scheme is illustrated I
with bridge trusses are shown in Fig. 1-20 (see also Fi
n bridge design is to use girder structures, which req
sses. In all cases as much welding is used as
lions either welded or fabricated using high-strength

ACCURACY OF COMPUTATIONS AND ELECTROMC

leng:;. uf the diagonal members. .

the 10-in slide rule was the principal computational tool in the structur
er's office, the computations rarely exceeded three significant dipits. This
satisfactory, for the reasons presented earlier in this
y and as implied in the example computations.
Presently, the electronic calculator and/or the digital computer are almost
nivetsally used for structural computations because of both the greater corn
lexity of structural configurations and the greater speed of performing calcul
ally setting the decimal. Now it is almost mandato
ethical and economical reasons) to provide several iterations on a
1-13.2 Bridges he design. This step almost always requires use of

These calculating devices can give a rather large number of digits to any
blem is how to treat this increased computing capacity.
ot any better than the input, but with a large number of
arently significant digits it looks very impressive. In nearly all design offices,
design computations are checked by a second person as a design precaution,
ner carries a large number of string calculations on an
ronic calculator, the results will differ from those .obtained where the
ker truncates intermediate steps, then reenters these values and continues
computations. Where the discrepancies are not large, the question arises of
ether the problem has been "checked" or whether one (or both) of the
persons has made a design omission.
For these reasons it is suggested that regardless of the initial input data
accuracy, computations should be camed to as many decimal places as is
aps) to obtain good checking convergence. The extra
alculation effort is minimal. Any intermediate steps should be written to the
me precision as they are used in subsequent calculations (e.g., do not write
reduced to reduce member Iengors. 06.1 and then use 106.153 in the following computations).
 50 STRUCTURAL S T B ~ LDESIGN

The reader should note that the several intermkdiate What are the wind forces at points 1 through 9 of the roof shown in Fig. Pi-7?
produce the end results 62.4 and 9.807 are not shown. ) - 18.07 kips.
Answer: ~ ( 5 =
Several other useful conversion factors are as follows:

To convert to Multiply by

kilogram (kg) kilonewton (kN) 0.009806650

:mund (lb) kg 0.4535924
pound kN , 0.004448222
';ips (1000 Ib) kN 4.448222
ib/ft kN/m 0.014593727
: ips/ft kN/m 14.593727
!b/ft kg/m 1.488 16404
psi MPa 0.006894757
psi kPa 6.8947577
ksi MPa
.ASI. kPa
6.8947577
6894.7577
psf kPa 0.04788026
isf kPa 47.88026
Lip . ft (moment) kN . m 1.35584
kip . in kN m . 0.1 129862
cago Building Code requires 1.25 kPa?
What is the maximum shear and bending moment in a bridge span of 92 f t for an HS 15
at is the impact factor?
Answer: V = 48.5 kips, I = 0.23.
PROBLEMS

1-1 What is E, for a steel with F, = 40 ksi?

1-2 What is E, for a steel with F, = 365 MPa?
Answer: 0.00182.
Answer: M = 7865 kN . m; I = 40.1 percent.
1-3 What is the increase in weight/ft of a W16 x 40 beam with 2 i
-12 What are the story shears in the N-S direction of the building of Example 1-7?
1-5) fireproofing as in Fig. PI-3? Note that the method of applica
exact geometric section; therefore, the coinputations should not be
, Answer: Aw 30 Ib/ft. -

1- rep roofing

Figure PI-3
. .,,. c, .,. ,.,,., ".,..,..

1.1What is the increase in weight of a W410 X 46.1 rolled section with 55 mm of vermiculite
fi:.eproofing as in Fig. P1-31 (See the comment in Prob. 1-3 before starting this problem.)
i-5 What is the R factor for the interior column at the top floor and three floors down of Ex
1-I?
1-6 What is the R factor for beam B1 and column B2 of Example 1-2?
 ELEMENTS OF FRAME, TRUSS,
AND BRIDGE DESIGN

' ANA .*

. of I requires dev eloping frame member forces (axial,

ben based on the dead and cntical live-load combina-

Ib ) The method of analysis depends oh. the complexlty of the structure and
whether it is rigid (indeterminate) or simply framed. An analysis may consider
Figure 11.1 Two of world's tallest buildings using structural steel frameworks. (
(442-4 Sears Tower building in Chicago, Illinois-c-ently the
the structure as either two- or three-dimensional. Simple framed structures are
tallest buil
337-111John Hancock Center building, also in Chicago, Illinois. View from the sears generally determinate; that is, the three equations of statics (2FA,2 F,, Cbf =
are sufficient to obtain the internal member forces. In any case, with simpEe
ming the ends of the members are assumed to have no moment resistance
ransfer to adjacent members. The internal member forces of determinate
tructures are readily obtained by hand calculations and with considerabIe
efficiency using pocket calculators.
Rigid,framed structures are generally indeterminate since the member ends
transfer shear forces and moments to the adjacent members. Indeterminate
ructures require d e f o m o n compatibility to supplement the equations of
atics to determine3e internal member forces. DigitaEomputers are used to
obtain solutions for all but the simplest indeterminate structures. Continuous
beams and certain simple rigid frame structures have solutions that can be
 ELEMENTSOF FXAME,TRUSS, M
.D

obtained from handbooks (or readily derived using mechanics-of-m

methods). A few beam solutions are presented in Part I
manual, and in most engineering handbooks.
A fundamental part of structural design is to es
.f.rai,c,.is to be-rigid or simple. A rigid frame gener
moments but does not necessarily result in a more economical design: This
because: '
,q = 4 + 6 + 3 = I3

1. Practical considerations of quality control/design o

connections.
2. Tendency to use the same depth beams across a bay even though certain
spans are shorter or carry less load.
3. The use of a constant column size through at least two and' often three or
more floors to reduce splicing. Since the colum
it is over designed in the upper floor(s).
?
' . The stiffness (formerly called slope deflection) method of analysis is most
commonly used for.indeterminate structural analysis u
This method is particularly adapted for'computer use,. as
sparse and symmetrical. Advantage can be taken of
inversion effort so that rapid solutions of very large
' economically. .
All indeterminate structure solutions are iterative a
in that the output depends on*the input; or, stated di
have member properties (area, A ; moment o
displacement compatibility. Since digital comp
~ I ,

rapidly, one may initialize a problem using relative va 21 portal of approximate frame andysis [br a typical three-ba~bent in
this output to select preliminary members and iterate as required: Wh
input is considerable, with frames containing lar
be better to use some approximate methods of
member sizes. Approximate methods include simply
columns and beams (e.g., A = constant, Ib = 1.21, The cantilever method (Fig. 2-2) is an alternative method of appro
Moment distribution (but limited to not more than
used together with relative values of A and I for pr
The-portal and cantilever methods of approxim
for building frames of one or more story heights. beams and columns.
primarily to obtain the effects of lateral (wind) forces on a frame. . Axial loa_d_s_isin~-al columns and are pro~ortionalto the
The portal method (refer to Fig. 2-1) makes the following assumptions: neutral axis of the bent treated as a vertical cantilever beam. The location
the neutral axis is based on column areas (since these are not g
1. Point of contraflexure occurs at midheight of all columns. known, a value of A = 1 unit area is usually used). The equivalentmo
2. Sum of wind (lateral) load is distributed as shears in proportion to ba inertia of the vertical cantilever beam is computed as
3. Beam or girder moment is zero at midspan.
4. Interior columns carry..na axial loads. I =C A ~ ~
... . -- . ~---
The column loads are computed as
With these four assumptions and application of statics, Mc
joint moments can be obtained at any location.
' v.=-
I
 :' STRUCTURAL STEEL DESIGN
ELEMENTS OF FRAVE, TRUSS, AND BRJDGE DESIG

The concentrated loads produce a moment of

M = -3-P- -L= - PL PL
2 2 4 2
'The error is L(0.5 - 0.375)/0.5](100) = 25 percent (too small). The error ra
: decreases and reverses sign as the number of con
(i.~~ting that the situation is one where the total beam load is a
! illcreasing with the number of loads). For example, if the tota
Load-brar~rigwail -,
." 1" and placed at five equal spaces, the M, value is still 3 p L /
9 .';'L/25.
Beams may be simply supported, overhanging, cantilevered, fixed, or con-
. t;iious, as illustrated in Fig. 2-3. The designer must always m
~ ~ o i involving
is point loads as well as how realistic the fixed, cantilever,
continuous-beam models will, compare to the actual member geometry. It
seldom possible and never desirable to have actual point application
l;nds/reactions, although this is a common assumption made in all analy
,rfie.:.ods. TERMINATE STRUCTURES
'The proportioning of beams can be done as soon as the s
di.1 :rams can be obtained. The general differential equation for a beam is
EZylV=-w '.

2.1.: successively: one in which much of the roof load is carried by the perime
EIy"' = V = shear = - wx +,C ,
- wx2
EIy" = M = moment = -+ c 1 x
2
+ c2
- wx3 c1x2
EZy' = slope = -+ -
6 2
+ C2x + C,
ical sense almost all beam-column connections can, and do, transmi
- wx4
EIy = deflection = -
24
Tile general equation for the elastic curve for a beam produ~cesfour
integration. Constants C2 and C4 = 0 for simple beams, and the w
for beams loaded with concentrated loads.
The general beam equation is sometimes useful in that the appro
i n f ~ i n a t i o nmay be approximately obtained by replacing a ser
ce~tratedloads with the equivalent uniform load value.
,@,earnde.fl,c.ctionsare often limited for both buildings and bridges. A va
of L/360 to as low as L/1000 can be fo
; occupancy and/or interior finish. Where' deflection n = 2j -- R
approximate deflections should be computed before a de
The reader should note that highly refined deflectio n = number.of bar members
possible because of the uncertainties in loading. It should also b
deflection is heavily depende,nCton the moment of inertia_an
constant for steel. Therefore, where a rigid deflection criterion i
use of high-strength steel ma'y not be economical, since the sectlo e-rea&Iy- _
optimized for least
-- - weigh~,-since--he-
fix?? by the moment of inertia rather than by the bending stress. s_-
geometry. Indeterminate trusses are not SO readily
.
60 snivcl.rfRA~STEEL DESIGN ELEMENTS OF F W I E , TRUSS, Ah?) BRIDGE D

optimized, since the bar forces will depend on both V Z ) The

~ . nodal equilibrium in matrix notation is
s-ndeterminate trusses are commonly used wh
some members, unsymmetrical loading conditions P=AF.,
extra members are used to reduce the L / r ratio
members. ~ontrolling.'theL / r ratio is likely 'to dictate the geometry
trusses. Where the L / r ratio controls the design, op
very difficult. The reader should bear in mind, however, that the member forces can be written as
(or any frame) js that which"ba1ances weight, fabrication, and F = S e = SATX
against the total client cost. Minimum weight alone does not sati
of "optimization." combining Eqs. (c) and (a), we obtain ', '

-ate trusses are readily solved by hand using the m

equilibrium or the alternative of n ( d ) can be inverted to obtain
a cut section. Both determinate and indeterminat
using one of the large number of computer progr x = ( A S A ~ ) -P'
computer program in the Appendix can be used t
indeterminate trusses. This program uses the stiffness method of matrix analysis
as outlined below. The user is expected to have taken or be taking a course in
advanced structural analysis, so only a brief outline of the stiffness method as
used in the computer program is given.
Referring to Fig. 2-5, the coding of a typical truss element is as shown, mpute member forces:
where the P-X code refers to nodal effects in a truss system of which only ,P,
through P, are shown as affecting the ith member. The internal member force' of
e general A-matrix' entries for any truss member are the direction cosines:
- COS a

p4 - x4

rforming matrix multiplication, we can obtain the S A as

Truss member connecting 2 nodes

e product of A x SA gives the element stiffness matrix as follows:

Truss member

PI + F I cos cx = 0
ZFv = O
66 STRUCTURAL STEEL DESIGN
determine the number of P-X entries (NP) and those definin
(NP + 1 = NPPI), where the X values (displacements) are zero.
members (NM) is also determined from this step. The ho
distances from one end to the other of each member is obtained from t
building geometry. The program computes the actual member length, using'
and V from this step as input data.
If the corrert member cross-sectional area A and moment of inertia I values
are input, the displacements (X matrix) are actual values. If relative (or incor-
rect) values are used, the displacements are not the true values (also the member
,;orces may be incorrect in many cases). The member forces of a determinate
.,ass can be obtained from the use of any area (use area = 1.0 so program does
i;ot divide by zero), but again the displacements will only be correct if the
correct member area is used. The moment of inertia is not used in determinate
or indeterminate pinned truss computations. The member forces of inde
rerminate truss configurations are dependent on the member cross-sectiona
nreas. The member forces in rigid frames depend on b
area and moment of inertia; thus this is always an iterative
use some estimated intial values, obtain output, revise
required, and make additional solutions as required.
If i t is desired to obtain floor beam deflections (or deflections along
highway bridge), it is only necessary to add nodes at those locations
illustrated in Fig. 2-8. While adding nodes increases the size of the matrix to
inverted, the routine (a banded reduction method) used in th
gram here only uses a part of the global matrix, so that very la
be very efficiently solved: To take full advantage of this reduction meth
coding should be such that the difference between the P, on the near end
(for trusses) or P, at the far end is as small as possible, sin
[NBAND = NPE(4 or 6) - NPE(1) + I] sets the size of the
reduced. Note that NBAND is only 4 for the beam coding
This means that a very large number of beam segments
alternatively, a continuous bridge with five or six spans with
span can be easily solved using a very small amount of computer core, since
value of STIFF(1) is NBAND X NP = 4 X NP; with NP = 100, there are o
400 STIFF(1) entries versus 100 X 100 = 10,000 entries for
inversion. The solution time is down to about that of the 100 X 100 mat
.. , , ,," ..,.,.....,. ~ xamples displaying computer output..The reader should parti
2-8 THE P MATRIX
'The P matrix is developed from the structure loads c
element a s either a truss element that has just node forces
team element based on using statics and the external for
member, such as wind, roof loads, and wheel loads. It
wheel loads between truss panels are prorated to adjacent
Seam analysis, as if that truss member is a beam with the re
ELEMENTS OF FWLIE, TRUSS, AND BRIDGE DES
elements making up a beam are considered as a seri
s; thus the P-matrix entries will include the fixed-end mome
ed for the several beam elements meeting at a joint in additio
the shears as illustrated in Fig. 2-2.
In both beam and truss analyses, the node forces (equal and opposite t
those acting on the connecting elements) are resolved into components par
to the translation Pi's.
Where beam and beam-columns (axial force and bending) are analyze
is convenient to have the computer program compute the FEMs and shears
the several beam elements using the beam loading, and make the summ
process as each element contribution to a node is found and then buiId tho
matrix. The program should also be able to read in selected additional P-ma
entries, which can be used and/or added to the value(s) already in the P mat
The computer program in the Appendix allows this procedure.
The user should note in using the computer program to develop th
P-matrix entries that all dead and live loads are applied along the length of
beam element in the direction of gravity using a (+) sign. This alIows
program to correctly compute the P-matrix entries (and signs) for sloping be
Wind loads applied to sloping, flat, and vertical beam members are aIwa
applied along and normal to the member axis. If the slope is 0 _< 0 _< 30°, t
sign will usually be ( - ) because of the aerodynamic (uplift) effect. This
the program to develop the P matrix for the wind NLC. Since gravity an
loads along a member cannot be combined because of a horizontal compone
of wind on a sloping member, it is convenient to store the D + L (or D o
nd treat the wind case totally separately, then combine wind with
L analysis to obtain the design case for D + L + I.V. Specificatio
the allowable steel stresses to be increased one-third in afTy stress co
ing wind (as long as the resulting member is not smaller than
rately). This stress increase is equivalent to reducing the membe
including wind (axial forces and moments), by 25 percent. The comput
m also does this, so that the designer merely obtains the maximum ax^
r moment from any of the NLC for design of that element. T o
antage of the program ability to do this, it is necessary to order the win
g NLW) after the D + L load NLC. This procedure is illustrate
lrly inspect the input/output for Example 2-7, which uses wind pressure on
loping beam element.
The coding of NP and noting NP + 1 = NPPl for specifying zero displac
nts (including but not limited to) for reactions excludes those global matr
ries from consideration. Those P-matrix entries developed by the comput
re not used in the analysis. The reader'should note that thes
directly into the reaction and therefore do not cause intern
mber forces (see the output check in Fig. E2-4d).
One of the very early decisions the structural designer must make is wheth
he superstructure (columns in particular) is pinned (allowing rotation) or rigdl
 P S X E N T S OP PPAUE, TRUSS, AND BIUDGE

p6

VP = 6
I
I
1VlII = 2
NPPI = 7

SOLUTIONIt will be necessary to develop the necessary equations for

for sloping members. For this refer to Fig. E2-2b and note that loads no
to member cause FEM. Now, returning to the initial problem, fo
sloping member 0 = 45" :
12(0.707 11) 6
FEM = (-)' ,= 50.91 kN - m
12 0.707 11

For the horizontal member:

RL=--4(72) - 48.0
6
70 STRUCTURAL STEEL DESIGN

Now consider the following sketch for nodes 1 , 2, and 3: e several load conditions for a bridge would include
P, = 0
p2 = 64 - 50.91 = 13.09 (note +) Dead + live, including impact
I
Dead + live + wind
P4 = - (50.91 + 48.0) = - 98.91 kN

.",,.,, 4, .,. ,,.,,.... ,"..-"

several load conditions.

Computer programs are included in the Appendix to deveIop
7 mPP1) -(does not matter)
trix via punched cards for analysis of a highway or railroad truss.

0 CHECKING COMPUTER OUTPUT

2-9 LOAD CONDITIONS

of two basic steps:

condition (NLC in computer program), The load conditions

!;)ads making.up the P matrix in the equation
X = (ASA~)-'{P}

night include:

b a d combination
--
~ l e a d+ live
Dead + live in alternate bays
Dead + live snow
Dead live f snow+
Dead + live + snow one side
Dead live wind +
..
NLC the end of the coding sequence is checked and both satisfy statics,
problem has been correctly solved for that irlpltt data.
1
2 Problem checks can be performed in several ways, as illustrated in
+ 3
+ 4
+ wind 5
+ 6
of computational effort-often by inspection. Summing forces above
72. STRUCTURAL STEEL DESIGN

a story are also often convenient, as the column axial forces are direct1 L 7'-6" 2 1'1"
1 7
Advantage can often be taken of symmetry of structure and lo Z 3
two office building examples that follow (without wind). Sometimes t (2) (3)
matrix can be used to advantage, as illustrated in the column check of th
building. (5) 16)
Truss checking is similar to that for rigid frames. Always take
truss geometry by checking members in which the internal force is 6 (9) 7 110)
zero. If the computer output does not give zero and the load
there is something surely wrong with the input data, such as mispunching an NP (12) (13)
or a mismatch of H and V (either length or signs). Perform any additional
statics checks near each end of the truss at joints where a minimum of bars
connect. If these joints do not satisfy statics, the more 10 ( 1 6 ) I I 117)
either and the problem needs to be reprogrammed.
(19) (20)

2-11 DESIGN EXAMPLES 14123) I j 124)

The Sallawmg several design examples will further illustrate frame coding and ,-
.s.v
(75) (26)
computer input/output for the analysis computer program in 4
(; ---.\'P = 4 b
computer output from these examples will be used in many .YPPI = 47
17 45.47.47 18 47
examples in later chapters toillustrate the problems in structural (
a manner somewhat like that which would take place in actual design. r.2.i
interest of saving text space and maintaining reader interest, the examp Elevation o f typical interior bent
considerably edited from actual design problems. ,

5 @ 18'-6"
Example 2-3 A small three-story office building is as shown in plan
elevation on Fig. E2-3a. Wind bracing will be used in the E-
bay and locations inarked wb) together 'with a simple fra
rigid frame (with resulting member end moments) will be used in the N-S
direction. A brick veneer exterior will be carried by lintel beams to the
exterior columns and at each floor. The framing method allows a corridor
between the interior columns with a clear space for the office areas. Rental
space will be in blocks: a, b, c, and so on. This space will contain
miscellaneous office furniture, partitions, and so on, based
and desires. With three stories, a flight of stairs at each
elevator on the west end for large equipment will be the to - - ~-
allowance. Air conditioning and heating will utilize a h Plan view
auxiliary shed (not shown). The basement will contain the remainder of the
environmental equipment and provide. extra storage. Use the NBC for
general design.
We will use a flat roof which may be used for worker exercising during
the day. For this additional activity, the roof will be designed for a live load
of 80 psf (as opposed to 20 to 30 psf for usual live and/or snow). The reader
"should note that building codes and material specifications stipulate mini-
mum requirements-the designer can always use larger values.

- -
STRUCTURAL STEEL DESIGN

SOLUTION The general computer coding is partially shown .in Fig. E2-3a
and also on Fig. E2-3d (computer output sheet). This coding gives the
Dead load = 0.080(18.5)
following computer program control data information:
NM = NP = 46
26(members) NBAND = 11 Live load office = 0.080(0.70)(18.5)= 1.036 kips/ft
The next step is to develop the beam and column loadings for this Live load corridor = O.lO(l8.5) = 1.85 kips/ft
typical interior bent (refer to Fig. E2-3b, which displays loads after com-
These loads are shown in Fig. E2-36.
cputations).
Roof:
Dead load: estimate 5 in of concrete on a metal deck supported by steel
bar joists.
.02(6)18.5 = ? . 2 ? k

Concrete: (5/12)144
Metal deck and joists (estimated) , = 5 psf
Ceiling, ductwork, electrical, etc.
Total
-
= 5 psf
=70 psf

Live load: check reduction in 21.3 x 18.5 spa

live load is not reduced; however, we are not using a stand
loading.
R = 21.3(18.5)(0.08) = 31.5 percent
" +

or R = - D+L(lOO)=- (100) = 43.3 percent

4.33 L 4.33(80)
Note that the L = 80 psf value is from Table IV-4. Use a 30 perce
reduction.
Live load = 8OC1.00- 0.30) = 56 psf
In a 7.5-ft span, R = 7.5(18.5) = 138.8 ft2 < 150 (no reduction).
The equivalent beam loads for the roof are:
Dead load = 0.070(18.5) = 1.295 kips/ft
.. . ... ..,..,.... . . Live load in 18.5-ft span = 0.056(18.5) = 1.036 kips/ft
Live load in 7.5-ft span = 0.080(18.5) = L.48 kips/ft
Other floors:
Estimate dead load with increase for floor surfacing
and ceiling finishing The exterior brick veneer wall will contribute column loads for' t

Corridors (Table IV-4 and conservative use)

Office space terior wall finish, fixtures, etc., of 5 psf gives a total ='45 psf.. .."

Column load = 12(18.5)(0.045) = 9.99 kips

The office space will have a 30 percent reduction in 1
STRUCTURAL STEEL DESIGN

The longitudinal walls defining the corridor and lo

Ute interior column loads:
L
i
Tile wall at 8 in = 35 psf,
L' Beam at 50 lb/ft (est.) = 50 plf
Column shear = 12(18.5)(0.035) + 18.5(0.050) = 8.695 kip
, For roof = 0.050(18.5) i=0.925 kip (interior columns)

i
t beam weight e tributing shear to the roof line of exteri
~ & l e c any
columns, since the values are tw small to be either reliable or to affect t
design. rO
r T rr 0C0 0 0
0.0. .. -0 e
0 0CC 0
*ra ~0O 0
oo- O
~ .OO*
o C c ~ - ~ O

Now make some preliminary member size estimates with the followin
practical considerations:
*
.
L
-
x
8
.
-
-
c.
.
0-0
8 -
,
.-.
;l;dddd:;:dddd-.;:dddd2;*
-,-
8 8
-- ,---
.=.
a , I ,

1
1. Use continuous column (no splices) for full 36 ft of hei
................._...._._
~ * r 0 0 0 0 = - C o O O C O - ~ 0 0 O O O O e o
* C r 0 0 ' 0 0 . ~ . 0 0 0 0 * - * 0 0 0 0 * C * 0

- - - *----
-9-0000---0000-*-000O_t_O
=*
rr*
- -- * D *

interior columns (maybe) at fi st floor level.] I

u
0-0

2. Use constant-depth beams acr ss the bent. O O o C O O O O O O O O O o O O O O O O O O o o

O O O o 0 O O O O O o O O C C O O o o o O o o o
0 0
0 0
0 0 ........................
0 0 0 0 0 0 0 0 0 0 0 0 0 C 0 0 0 0 0 C 0 0 0 0
0 0 0 0 C0 0 C
0 0 0 c
0 0 0 0 0 0~ 0 0
C C
0 0 0 0 ~
0

For beams: 0 0 0 - - - - C 3 0 - - - - C 0 0 - - - - ~ 0 0

"-
I 0 0 O O O 0 C c 0 O C C 0 0 0 3 0 0 0 0 0 0 9 0 0 0
If the roof beam is simply su the moment is M =A .a U J ;f
~
l 0 0 0 ~ 0 C 3 0 0 1 0 0 0 0 0 0 0 0 5 0 0 0 0 0 0
0
0 0
C 0
0 0
0 0
C 0
C 0
C 0
0 0
0 0
0000000000000000C0000000000000000
C D D o C o O O O O o O 0 E O O O O o c O o o o o O
fully fixed, the moment is M = lue is somewhere
- - - in
--- --ldddd222dddd222;ddd22fdd
between, so taking an average:
-
c 0000000000000000000000000

Q O C O C o o ~ O 9 O O O O I O e O O O O * o - ~
;d;;;;;;d;;;;;;dJ;;:;n.dddO
+ 1.48)(21'3)2 i
* ~ * * . * * n D r * * *
h N----N
* r O r * * ~ * ~ w ~ ~ ~
N----N *----.+ L--
M, = = 142.7'ft , kips
8 6
-.

M, =
8
(142.7)- == 95.1 ft . kips
- dl,d;dd;ddddld;,hdb:dD'O'd::
----*.-----*<-----*.-wm~=*
*rw----nrn----ccn----.oQ*--
12
Ma, = 118.9, say 120 ft . kips

The required section modulus based on an allowable 000000000000000000000000

000000000000000000000000
00000000000000000000000000
00000000000000000000000000
00000000000000000000O0O0 ddd;;;;ddd;;;;ddd;;;:ddddd
24 ksi (Chap. 7) is dddddddddddddddddddddddd ' . : '"I. ' Z

12!0(12) 000000000000000000000000
Sx = = 60 in3 O O O O o O O O o O O O O O O O o O O O O O O O
000000000000000000000000
24 dddddddddddddddddddddddd

0 0 0 0 ~ 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 ~ 0 0
0000000000000000000000~0
000000000000000000000000

ox -u. n = in- 1, = 3 I 6.u in'

dddddddddddddddddddddddd
1
I
1 I.OU

I
000000000000000000000000
For columns: 000000000000000000000000
000000000000000000000000
dddddddddddddddddddddddd
Estimate a length coefficient (Chap. 7) K = 1.2 -+KL = 12 x
14.4 ft. Estimate F, = 16 ksi (som thing less than 22 ksi).

Papp,, = 0.925 +

This computation
utions + the uniform beam loads from roof to the first flo
'c01umn~
Z

A -
 - -

4
W
. '4- \ l I T F I I I tN"IE 690 OANJ h I O T H = 13

U AND U -

h/D ; 1 NX = 1 0.00394
YO = 7 NX = 2 O.OO4Rl
YP - 3 NX = -0.09541
ND = 4 NX = 4 -0.00169
N> r 5 WX = 5 0.00033
ND = 4 NX = 6 -0.19343
-
NO
NP
YO
=
-
7
P
9
VX =
NX
NX =
- 7
8
0
0.00169
-0.00033
-0.19343
yo = 10 NX = 10 -0.00394
NO = 11 NX = 11 -0.00481
Yo = I 2 NX = 12 -0.09941
LIP
qo
YP
=
=
=
1-
14
15
MX =
NX
NX=
- 13
14
19
0.00272
-0.00081
-0.08302
YD = 16 YX = 16 -0.00121
q P = 17 NX = 17 -0.00006
NP = 18 NX = 18 -0.17721
LIP = 19 NX = l a 0.00121
ND = 20 NX = 20 0.00006
YP = 21 NX = 71 -0.17721
"I = 22 YX = 22 -0.00272
VP = 23 NX = 23 0.00082
NO = 24 NX = 24 -0.OR302
No = 25 NX = 25 0,00275
NO I 26 NX = 76 0.00061
VP = 27 NX = 77 -0.05121
YP = 2R NX = ZP -0.00125
YO = 20 NX = 29 0.00007
LIP = 30 NX = 10 -0.14018
ND = 31 NX = 31 0.00125
No = 32 NX = 32 -0.00007
YO = 33 NX = 31 -0.14018
.,D = 34 NX = 34 -0.00275
YO = 35 NX = 3< -0.00061
NP = 36 FIX = 36 -0.05121
NP = 37 NX = 77 0.00353
NO = 38 NX = 38 -0.00136
yo = 30 NX = 39 0.00023
NO = 40 NX = 40 -0.08232
NP = 41 NX = 41 0.00136
NP = 42 NX = 42 -0.00023
YD T 4' YX = 43 -0.Oq232
NO = 44 YX = 44 -0.00353
NO = 45 NX = 4T 0.00068
No = 46 NX = 46 -0.00068

Sketch lo ~denttfyP - X codlng

XAL F3RCE. K

3 -6.07 -82.42 38.11

4 -22.77 34.19 3R.70
5 -38.21 -20.14 -22.49
6 -38.21 20.1+ 22-49

(I 1.01 -64.52 87-21

9 0.*6 -49.19 49.19
10 1.01 -87.27 64.52
11 -58.49 30.45 30.1.3
12 -81.26 -18.14 -11.44
13 -81.21 19.14 11.94
14 -58.64 -30.45 -30.33
15 -0.72 -63.13 81-36
16 -0.54 -50.34 50.34
17 -0.12 -81.36 63.13

DESIGN EN0 *O*tNTS CORRECTED

F n R F E I A N 0 WIN0 lNE.49 EN0 FIRSTI. I-F1

-26.18- 64-31
 STRUCTURAL STEEL DESIGN
ELWlE?4TS OF FRAME,TRUSS, AND BRIDGE DESIG

check if O.K.: K L / r , = 14.4(12)/3.53 = 49. The allowable col

ess from Table 11-5 .= 18.5 ksi.
We will try this value for the first design iteration, since there will
moments on the columns of an unknown amount that will effective
increase P. We will use a smaller section for exterior columns. Use WS

A = 9.13 in2 I, = 110 in"

Check CF, for t o p story. For the two basement columns, use W8 X 48:
C/',l,,,8d,, = ? . 3 3 1 ( 2 1 . 3 ) ( 2 ) + 2 . 5 1 6 ( 7 . 5 ) + 0.925(2) = 120.02 kips
From computer o u t p u t for values shown which have been reduced 25 percent I'or wind A = 14.10 in2 I, = 184.0 inJ
5o.that o u t p u t is directly comparable t o D + L o u t p u t :

These data are used to make up a set of member data cards, b

T F , = 120.01 kips O.K. 101.55
Actual F = ------ = ]35,4k loading cards (interspersed as appropriate in member data), and a P-m
0.75
data set. Note the P-matrix entries read (PR(i, j)) are zero for NLC = 1
. .-. ..l
7.
1, ... ...,...*.,_ .. ake the values shown in Fig. E2-36 for NLC = 2. The remainder o
trix is built by the computer program using beam loading data and th
umn shears separately computed and also shown on Fig. E2-36. Th
ut is written back as a part of the output for designer checking (Fig.
3c). The remainder of the output is shown on Figs. E2-3d and
ether with some output checks. Figure E2-3f further illustrates
eck of NLC = 2.
R3= 12.09~
'Mom have been
e reader should note that when wind is considered, the allowable stre
V1= 8 965 increased by one-third. So that the design is all on the same b
reduced for wlnd
gram multiplies all member forces by 0.75 for NLC that have wind forces as
ut. With this adjustment, the designer merely has to scan all the output for
maximum moment or axial force. The largest value in any load condition i
hen used with the allowable stress to determine if the member is adequate. Not
in checking the output (as illustrated in Fig. E2-lj-) the fact that this 0
or has been used must be considered so that statics is satisfied.

ample 2-4 We will design the small office building of Example 2-

units. Note that the dimensions are slightly different than when
C h e ~ k l n gcomputer o u t p u t for column (member 25) uslng computer o u t p u t and the statics o f node 1 4 oft conversion. The same general design considerations/parameters
pply as in that example except that the brick veneer wall will be take
Figure E23f Output checking for NLC = 2 (with wind). 00 mm (approximately 2 x thickness of Example 2-3). Figure E2-4a
lays the general building layout and SI dimensions.

LUTION The computer program will solve either SI or fps probIems. I

cessary to include a "UNITS" card with fps or SI identification for use in
the FORMAT statements to identify the units of the output.
Use W8 X 40: Since'the same building design criteria are being used, we may pi%
directly to coding the joints (refer back to Fig. E2-3d since frame is s ~ i l x )
A = 11.70 in2 Ix = 146 in4 eveloping the frame loads.
STRUCTURAL STEEL DESIGN
ELEhlENTS OF F I U M E , TRUSS, A N D BRI

oof loads:
Dead load: estimate 130 mm of concrete on a metal deck
steel bar joists.
Concrete: 0.13(23.5 k ~ / m ' ) = 3.055 kPa
Metal deck and joists (estimated) =0.263 kPa
Ceiling, finishing, electrical, etc. = 0.263 kPa
Total 3.581 kPa

Live load: a live-load reduction can be used in the 6.5 x 5.9 m area
Basic live load = 4.0 kPa (Table IV-4).

Use a 30 percent live-load reduction. The reduced live load = 4.q1.0

0.30) = 2.8 kPa. D o not use a live-load reduction across the 2.3-m span
The roof beam loads are:
Dead load = 3.581(5.9) = 21.1 k N / m
Elevat~ono f t y p ~ c a lInterlor bent Live load in 6.5-m span: 2.8(5.9) = 16.5 k N / m

59 ,!
1
59
1
59 L
1
5.9 ,
1
59
N
Live load in 2.3-m span: 4(5.9) = 23.6 kN/m
Frame loads for other floors:
Use 30 percent reduction for office area live loads but none in co

Plan vrew

Figure E2-4a

3.77 kPa (brick) + f~nish:use 4.0 kPa

 58100'0 01100'0 12 = KN Ol.bCb511 OI'bZb$ll = dN
I - LLEOL'I- 09 = XN sb'51Z- 56'512- Of = dh
11590'0 91LOO'O bk = XN 00'0 0 0 '0 br = dN
',5100'0- OLlOO'O- dC = XN OI'bZbS11- Ol'bZb'YI1- Hk = dk
ZLZOO'O
010b9'1-
OhlCh',
. .
OCZOO'O
IlbO9'1-
ZfU20.0-
~r
VL
><
-
= XN
XN
= Xh
0u'Z~ObLl
UL'LIZ-
00'0
OU'ZLOI,CI
Ok'iIZ-
00'0
LC
01
5k
= dk
=
=
dN
ah
01200'0- ULZOO'O- ZE - Xh OU'ZlObLI- Ob'ZlOhCI- 21 - dk
fbO1L.k- b b rL ‘ XN 5b.GIZ- 56*$1Z- CC = ON
EUOUb'f 9OZOO.O- ZL - Xh 00'0 00'0 Zi - dN
EVIOU-0 SETOO'O le - xh 01-bzb511 O I ' ~ L ~ ~ I II C - o h
0kEII.E- ElZIZ'E- Or - XN 11'91Z- 1I'VIZ- 0k = dN
bV8bb.b ZIb00'0 bZ - Xh 00'0 00'0 62 = oh
~0100'0- 5E100'0- VZ = Xh 01'501911- 01~b01011- HC = dh
U0015'1- bObO9'1- LZ - XN
9b9L0.4 LZOEO'O VZ - Xh
bLkOO.0 bLZOO.0 52 = XI%
ObULP'Z- ULEUS'Z- ?C ; XN
A O99Z5'8 ILPZO'O iZ = XN
8?Z00'0- 18200'0- 22 = XN
l'd siSOE'Z- 1Ulbl't- 1Z = XN 1 1 - 1 - 11'01Z-
EZEIS.8 buZOO'O UZ = XN 00-0 00'0
E~IOO'O LE100'0 01 = XN 01-SO19I1 01'G01011
bLLL0.b- bZ1bl-2- 11 = XN
IEIZS'Y 95000'0- LI = XN
blIOO'0- EiIOO'O- VI = XN
SZBfb'Z- OEEYS'Z- 51 = XN
VEZSS'8 IECZO-0- ?I = XN
Zli00'0 0UZ00'0 LI = XN
OZ'ZZI- OZ'ZZI- 21 = dh
00'0 00-0 I1 = cN
O L ~ E B E 2 E I -O C - E B E Z E I - 01 = dh
00.8~1- 00'uLI- 0 = oh
00'0 00'0 U = ON
00'8L9ZII OO'uL9211 L = dN
00'8LI- 00'8L1- 9 = ON
00'0 00-0 5 = dh
OO'LfL9ZII- 00'8L9ZII- b = dN
. -. . -... ..
~Aluo) OCZSZ'OI 6ZZEl'O Z = XN OZ'OI 00-0 2 = dk
OESOO'O ZZfOO'O 1 = XN OE'EBEZEI 0€'r@EZ~1 I = dk

9 LC S N V i O V M MO Yvl ' X I d l V W - X ~ H L MY-W ONV NW 'XIdllW-d 3H1

= H l O l H ONVB S3181N3 tIl3jllS

2 ~ 3 3 1~33 NIS so3 1 I v A 4 L ~ +N d ~E ~ NZ ~ NI ~ .N3 8 h z r

P ~ H

OOZ.LZ
OOZ'LZ
000'0
000-0
OOZ'LZ
002-LZ 000'58
000'58 OOZ'LZ.
000'58 OOZ'LZ
000'0 000-0
000'0 000'0
OOZ'LZ
OOZ'LZ
000.5u
OOO'b8
000.0
000'0
OOZ'LZ
OOZ'LZ
000-0
000'0
000-0
000'0
009-t
OO*..

lfr 1: (11
PSI;.
19L.b
XI
1 - 131113W - 11 31I1JM
O.OOOOOL
E -- d
AllJIlSVl3 OOY
Od3Z-YON ON Z
OOb
=
- 030151 3 3 l ~ l iL N
31N U N i l Oi dlh G V
Z = llONJ3 O V J l 30 ON 92 = S138M3Y 3J ON 9C - dN dN
;I ?I 01 6 B L P \ P S6SXOItM f ; I
 .
w
ELEMENTS OF FRXLIE, TRUSS, AND BRIDGE DESIGN 89

e reactions at the basement wall are computed:

= RI + axial load,, + R i + axial load,, + axial load?, + axial load,
387.29 + 402.51 + 626.26 + 665.79
= 111.95 + 119.28 +
u,..--
/>
--
2081.85 - 3007.03 (vs. 3006.36 kN) O.K.
0.75
Now consider C F,, = 0:

53.64 14.92 17.60

J1

LJ
47.18

Member forces

m
! .'
~ c * ~m c~c * ~c o ,~ r c ~o c u
. .
o m ~c o ~
m o ~- o -~* D ~O m - o w o m ~ ~ a e c c ~ ~

dd:dd'::;;;'dd;';:;;;;f;;r:
0 0 f f " m N * O * C N " ~ "IOtCN" *l,-"...a~

I 4 l - l l - l l -

'(Mem. 22) '(24)

f . .. . .
m m m R - C I O n O m O m N n o m C ~ m O * C C O O o
m F C E m N N m ~ O C * m n * 0 ~ 0 0 3 e 0 3 0 0 0
d;:;;;;;d'd';;;d:;:d;;d;;d Forces on two basement reactions, including effects from horizontal beam
~IDP~N-C*QONV m n n o m * n m r p t
I l l I I , , , I members 22 and 24:
""ti 'i"
CI;;t(applied
externally) = 10.2 + 15.3(2) = 40.8 k N
CFh , ,,,= -22.52+ 15.15 + 3 7 . 3 3 + 11.80+4.32-5.11
= 40.93 kN O.K.

Example 2-5 This is a partial design of an industrial building (fo

manufacturing process) with a general elevation view shown in Fig. ,I$-
together with a reduced size plan view. With the large unsupporied' cl
span and height required, roof beams are not deemed pr~ctical,so roo
trusses with purlins will be used. These trusses will both reduce roof
deflections and increase the overall frame rigidity, particularly by making
e columns continuous to the top of the main roof truss. Making the
lumn continuous will reduce column rotation and translation at the roof
 ELEMENTS OF F R h X E ,

line. The purlins (beams across bays to support th

spaced at about 6-ft centers; thus one will fall bet
will produce bending in the top chord members in addition to the axial
from the truss analysis.
It will be necessary to make an initial estimate of member siz
program these data, inspect the computer output. and iterate as required.
first iteration (not shown) used W14 x 68 columns. The most severe Ioadi
condition (with wind) produced lateral displacements on the order of 36
which would never be acceptable. The columns were resized using W27
178 for the second iteration shown here.
The building plan shows 22 bays at 25-ft sp
carry wind bracing (not shown here or designed in
be longitudinal beams between columns at app
spacings to carry the exterior walls. These memb
load to the column and also provide lateral braci
the weak axis. The wall loads are not considered in
The wall contribution can be readily accounted for by
shears directly to the axial force on the computer output when it
determined what the exterior siding will be (sheet metal, metal and woo
etc.).
General design parameters for a typical interior bent (or frame) inclu
Roof + purlins, estimated at
Miscellaneous additional details = 5 psf
-
Total =20 psi
Load/ft of roof = 25(0.02) = 0.50 kip/ft

SOLUTIONThe dead load of the truss is obtained after a n approximate

analysis (not indeterminate truss) has been made to estimate possibIe
forces. This analysis is very approximate and not shown here, as there
very crude assumptions and there are several which
make to obtain similar results. From this initial estimate the following
member sizes are selected:
T o p and bottom chords of shed and main roof tr
wt/ft = 20.6 Ib/ft A = 6.06 in2
Vertical and diagonal members: two L4 x 3 x i:
wt = 11.6 Ib/ft .A = 3.38 in2
The approximate truss weight is computed as follows:
Length of top + bottom chord (approx.) main truss = 2 x 120 =2
Verticals: av. height x 1 1 = (15 + 27)(11)/2
Diagonals: av. length x 10 = (27.3 + 19.2)(10)/2
92 STRUCTURAL STEEL DESIGN

The average truss weight:

W(120) = 240(0.0206) + 0.0116(231 + 232) = 10.31 kips
w=- 120
= 0.09 kip/ft use 0.10 kip/ft

Truss weight for side sheds is as follows:

Length of top + bottom chord (approx.) = 2 X 72 = 144 ft

Verticals: (3 + 15)5/2 = 45 ft
Diagonals: (12.4 + 19.2)(5)/2 = 80ft

The average truss weight:

W(72) 144(0.0206) + 0.01 16(80 + 45) = 4.42 kips
=
4.42
w=-- - 0.06 kip/ft use 0.07 kip/ft
I
72
t

The weights of the trusses have been rounded up to account for connections.
The truss will be analvzed for two load conditions.

NLC = 1: dead load + snow

NLC = 2: wind + snow + dead on left shed + wind (suction) + dead on
main truss + 1.5 snow + dead on right shed

The truss ,will be symmetrically designed, since wind can blow

either direction. The rationale for this combination of loads is as follo

1. Wind from the right will not blow snow off the left shed. Also, the
vertical wall will create some stagnation, so the direction of the wind is
normal downward.
2. On the main truss, the slope is such that aerodynamic action will result
(in addition to blowing snow off) in a suction.
3. The right shed should be protected from the wind, but the snow from the
roof can accumulate; to account for this, the snow is increased to

We can now begin to compute the node forces.

For dead load: Apply truss weight (computed as kips/ft of span) to the
top node along with the roof dead load since the small total weight causes
too much additional work for the increase in computation precision.
p
Shed dead loads: note that both shed and main roof truss have 12-ft
panels.
B
+
Interior nodes: (0.07 0.50)(12) = 6.84 kips
Exterior nodes: 6.84/2 = 3.42 kips
% STRUC'IVRAL STCCI. DESIGPJ

Main truss:
Interior nodes: (0.10 0.50)(12) += 7.20 kips
d .,... Exterior nodes: 7.20/2 = 3.60 kips
Snow load at 25 psf of horizontal (span) projection:
Interior nodes: 0.025(25)(12) = 7.50 kips
Exterior nodes: 7.50/2 = 3.75 kips

4812 = 5 74
12
0 = tdn -= 1 1 31'
60
sln 0 = 0 19612
LOS 0 = 0 98058
12
L=- =1224'
cos 0

Main roof truss

7.50'
18
8 = tan-' - = 14 04'
72 3 87'
12
'1 - = 12 37'
Los e
Py = 0 025(25)(12 3 7 ) = 7 73'
r':, = 7.73 cos B = 7.50'
Ph = 7.73 sin 0 = 1 .88' Shed trusc (left side)

Figure E t S c
 ELEMENTS OF FRA!!IE, TRUSS, A N D BRIDGE DESIGN
% STRUCTURAL STEEL DESIGN

YQ STIFFIII ENTRIES = 954

? N O I T I O N NO
D E S I G N EN0 VOlENlS
~ 1 1 0 1 9c ~
rhotTlo* NO - 2
OESIG* EN0 .C*E*TS
IL
&XI CO99EClE0 GO. F E N bNC Y l Y O rjrsFs rrlu COPPECIEO FOP U W
FORCE. K INEL9 EN0 F I R S T I . FT-X FO~CE.I # * E L M ENO F I S S ~ I . FT

8.15 0.00
---- ----
THF P - Y A T R I X , Y AND IN-K ---- ----
--.. ----
-..- *---
---- ----
--.. ----
---- ----
---- ----
---- -...
---- ----
---- ----
---- ----
---- ----
YO r R ---- --.-
---- ----
qp =
'$0 =
q
10
---- ----
----
YP = 11 ---. ----
----
YO =
YP
YO =
. 12
13
14
--.-
----
...-
----
----
YD = 15
YP = 16
NO = 17
N O = 1R
NO z 14
NP = 20
= 21
YD 5 22
NP = 23
NO = 24
ND L 25
No = 26
NP = 27
NO ;
. 28
ND = 24
N'P t . PO'-
YP = 31
NO = 32
NP = 79
NP = 34
YO = 35
NP r 96
NP = 37
NO = 38
ND 5 30
Yo = 40
YD r 41
YP = 47
NP ;
. 41
l o = 44

;:
rlD = 45

$2
.NO
yb
NP
.
=

=
49
50
ND r 51
NP = 52
ND = 53
NP = 54
yo = 55
YO = 56
NP = 57
NO = 5R

Figure E2-5e
 ELEMENTS OF FRXME, TRUSS, A N D BRIDGE D a 1

chapters to design (or redesign the members). Note that the maxim
lateral displacement is now on the order of 4.7 in at the base of the sh
truss and only 4.0 in at the top of the main truss, indicating some additio
bending in the interior columns. Note that this displacement is occunin
the wind NLC, as one would reasonably expect.

Example 2-6 Deslgn the ~ n d u s t n a lwarehouse bullding In SI units with the

general dimens~onsshown in Fig E2-6a Use the same assumptions and
problem parameters as used for Example 2-5

SOLUTION The initial member sizes are selcc~ed as follows (and bein
considerably guided by the computer output of Example 2-5, which
professional terminology is using "experience"):
Top and bottom chord members of both main and shed truss: two
L152 x 102 x 7.9 mm:

I, = not needed for truss-type members

Intermediate truss members (verticals and diagonals): two L102 X 76 X
6.3 mm:
wt = 0.169 k N / m A = 2.18 x mZ
Shed columns (member numbers 1 and 93): W610 X 241.1:
wt = 0.72 k N / m A = 30.77 x lo-' m' I, = 2151.9 X low6m
Two interior columns (members 24, 25, 27, 28, 67, 68. 69, 70): W690

The computation of frame loads now follows.

Roof? = 0.72 kPa
Miscellaneous, including purllns = 0.24 kPa
Total = 0.96 kPa

Load/m on bent = 0.96(7.6) = 7.3 k N / m

Dead load of truss: note that this is computed directly as a horiz
projection.
Main truss:
Length of top and bottom chord = 2(36) =72 m
Approx. length of diagonals = lO(5.84 + 8.3)/2 =70.7 m
Approx. length of verticals = 1 l(4.6 + 8.2)/2 = 70.4 m

t Take as horizontal pro~ect~on,

as truss slope IS very small and values are estimates.
'1@ .. .KUCTURAL STEEL DESIGN

I Note (-1 s~gnfor P m a t r ~ x

~u-. J I K U L ~UKAL SIEEL DESIGN

KN &NO K-MR !X, MU 01 R A D I A N S CONOI'IOH YO I

YP = 1
AXIAL
ND ; 2 FORCE, K N
N = 3
VD = 4
VP = 5
LiO ; 6
NP = 7
qu ; 8
NO , 9
NO = 10
vu = 11
rrr = 12
<D = 17
UO ;
14
'40 : 15
\ Y = 16
NP
Nu
NP 5
.
r 17
1R
19
Y O :
'40
,do
.. 70
71
77
NO 2 73
L(P = 24
YP = 7r,
Ye = 26
v; 7 27
NU = 2R
*iP %. 2s
40 = 30
NP = 31
YP = 37
U I - 3,
YO = 34
UD = 35
UP = 36
ND = 37
NO ; 38
S' - 39
N %, = 40
VP = 41
.1R = 42
'40 = 43
YO = 44
NO = 45
No = 16
LID = 47
Vo = "R
yo ; 4Q
'40 : 50
No - ii,.,..
U' 7 :Z
NP = 53
V? = 54
hi" = 55
YO = 56
YO . 57
YP : 58
NO = 59
NP = 60
YO = 61
','o i 62
LIP = 61
No = 64
YP = 55
Yo = 66
ND = 67
NV ; be
NO = 69
hi0 ;
NO
VO
-
=
70
71
'2
.,D = 73
VP : 74
YO = 75
Urn= Ih
ND = 77
ND = 78
YO = 79
LID = 80
106 STRUCTURAL STEEL DESIGN

present maximum horizontal deflections are:

X, = 101.7 mm X,, = 105.1 mm gles with long legs back to back for both top and bottom chords, find the an
X6 = 106.4 mm X,, = 90.4 mm
XI, = 106.4 mm X,, = 88.5 mm
The difference
X,, - X,, = 90.4 - 85.8 = 4.6 mrn

tension and compression members and connections.

PROBLEMS

The following problems are of two 'types. The first four problems are for the student to obtain
familiarity with using the computer program given in the Appendix to solve structural problems
(other computer programs, such as STRESS or STRUDL, may also be used, but the program in the
Appendix is likely to be considerably faster to run). Use E = 200 000 MPa or 29 000 ksi. Obtain the 4 p a n t i s '0 20' = 80'
solution in either fps or SI units, as assigned by the instructor. ( 4 paneis '.i 7 rn = 78 rn)
One or more of the last three problems should be used for the design projects to be carried
along with other later chapter problems.
2-1 Given the following beam. Obtain a W section that limits the deflection at point A to 1 in or 25
mm. Note that all NPE(2) can be made NPPI.
Answer: MA = 171.6 ft kips or 236.2 kN . m.

sok
* 2.4 200 k N
w= 4 kipsift or 60 k N / m

Wind = NBC value (take height as 9 rn to points B and E for wind in eith

Vertical load from a hoist at point D = 20 kips or 90 kN

Bay spacing = 25 ft or 8 m
Try to limit iterations to five.
Figwe PZ1

2-2 Given the following academic beam/frame. Find a W section that limits the'deflection at point
A to 0.5 in or 12.5 rnm.

Note. Joints B. E. and Fare "iixed."

Joints .-I. C. a n d D arc "pinned."
110 STRUCTURAL STEEL DESIGN ELEMENTS OF FRAME, TRUSS, BRlDGE DESX

dunensions for your analysis:

Dunensions

Part a'
b 30 rt 3.0 9.1 r 1.0
c 29 t 0.5 8.8 r 0.20
d 13.75 r 0.5 4.20 r 0.15
e 7.5 r 0.5 2.3 r 0.2
f 20.0 t 2.0 6.1 r 0.6
g 3.25 1
h 6.0 -C 1.5 2.0 +- 0.5 for the analysls program for the wheel loads at selected pos~tlonsalong the span. It is sugg
use wheel d~stance~ncrementsof 5 f t or 1.8 m.
Make t h s dunension consistent w t h d~mensione.

Notes ' - -
1. The knee brace is plnned to the column, but the column is contmuous to the truss.
2. The lateral crane lmpact load wdl be applied to the column at the attachment of the runway
guder to the column, as shown m the figure.
3. An optional vehcal member is used to reduce the K L / r of the bottom chord. If you use h s
member, it should be pinned to the bottom chord, but the bottom chord is continuous across this
connecuon (thus wll have bendng).
4. Sketch where you would place w n d bracmg, but do not design.
2-7 Given the hghway bndge truss shown in Fig. P2-7, make a computer analysis for the s
truck loadmg ass~gnedby the mstructor. Use the followmg dimensions and mtial data:

Dimensions/section, fps: SI:

3. 25 r 3 ft 7.5 +Im
Members
.
Bottom chord (1,4, . . ,21,25) W8 x 48
Top chord (6, 10, 15, 19, 23) 2 C12 x 30
Vertxxils (3, 7, . . . , 20, 24) W8 X 40
 ELASTIC, PLASTIC, AND BUCKLING BEHAVIO
OF STRUCTURAL STEE

m-1The Chicago "Picasso". A massive sculpture using corrosion-resistant steel designed NTRODUCTION

-.
9

. .

,. .
. ... ,,",,>, .... ,,,,..,,.,.. ,. *

. .

tendency of unsupported structural elements to buckle under

gned using equations that have been developed as a combination

 l i e STRUCW STEEL BESIGN
ELASTIC, PLASTIC, AND BUCKLING \ BEHAVIOR OF SlRUC?UIWL SEEL

From Eqs. (c) and (b),

S o ~ u n oWe
~ must apply a factored load of sufficient magnitude t
0.8 P3(48) velop f,
el = e, = - P3(60) = Fy in the three bars. At this time, the bar forces are simply
0.64(29 000) 1.0(29 000) P I = A, Fy P2 = A2F, P3 = A3F,
h we obtaiq
60P3 = 60% which checks displacements = constant
P I + P2 + P, = P,,,,,,, or simply P,
. :jlso, f r o b Eq. (c):
P,,= 36(0.64) + 36(0.75) + 36(1.0) = 84.04 kips
Since the actual load is only 30 kips, the load factor is P , / P :
86 04
L F = -= 2.87
30
Several comments are in order:
' ~ wsubstituting into Eq. (a), we obtain 1. Plastic analysis is much simpler.
0.8 P, + 1.25P3 + P, = 30 kips 2. The rigid bar ABC will rotate under the applied load P, = 84.04 ki
which was not the case in the elastic analysis. Why?
P, = 30/3.05 = 9.836 kips 3. The elastic analysis (Example 3-1) indicates bar 2 yields first. Wh
P2 = 1.25(9.836)= 12.295 &ips When F, = 36 ksi in bar 2, it carries no additional load but mer
elongates with any additional load being carried by adjacent bars un
PI = 0.8(9.836) = they reach F, in turn. When Fy is reached in bar 2, the load in the bar i
Total P2 = 0.75(36) = 27 kips.

:!ack substitution for e in each of Eq. ( c ) gives e $ 0.02035 in (which the By proportion from Example 3-1 the load at this point is
r ~ a d e should
r verify).
/// p = - 27'0 (30) = 65.88 kips
12.295
l 'ow let us reconsider Example 3-1 using "plastic analysis" in the following
exnrple.
Beam behavior based on a plastic analysis is similar to the bar problem.
amble 3-2 For the sketch shown in Big. E3-2 (same as Example 3-I), Consider the beam shown in Fig. 3-1. If we apply a bending moment to the
ection, the moment-rotation (M-+) curve is linear? to !tfy. From the point at
n hat are the bar forces when all three bars have yielded?
ch the most stressed beam fiber is at F, (producing the yield moment itl,) to
point at which all of the beam fibers are at F, (either tension or compression
epending on which side of the neutral axis we are inspecting) and producing
plastic moment M,, the curve is nonlinear. When iCfp is reached, the beam
ply rotates at this point, with no further increase in moment capacity (or
tress) and we say that a plastic "hinge" has formed. There is some small
itional increase in moment capacity when some of the beam fibers most
ant from the neutral axis reach strains into the strain-hardening region. This
ect depends on beam cross-sectional geometry of the flanges and web and, of
he beam span and boundary conditions.. If the beam is loaded with a
greater than My (but not M,) and unloaded, the curve branch BE is
with a permanent amount of residual beam rotation OE.

t Within material homogeneity and rolling tolerances, as well as practical measuring limitations.
120 STRUCTURAL STEEL DESIGN ELASTIC, PLASTIC, kWD BUCKLING BEHAVIOR OF STRUCKJR.U

proportions are such that the section can become fullv ~ l a s t i cbefore the nnce The plastic section modulus is obtained as the statical moment
strain hardening (i.e., .depth/web thickness and flange width/flange' thicknes about the neutral axis, which div~desthe area equally. Note that
not too large). necessary in order to satisfy statlcs on the section of CF, = 0.
- in detail the Plastic
W e ' will now investigate - .- ---
A - --- ---- rnnrent
moment --*--vy. in t A.. .LLW

following several paragraphs. Referring to Fig. 3-1, the moment at initial yield i
My = S,F,
where S, is the section modulus, I / c . The moment of inertia I and the distanc
from the neutral axis to the extreme fiber c are as in any mechanics-of-material.
textbook. The plastic moment M...Y , bv ins~ectionof - - the ~ ~-----
---- S---
- -- ..
-..".. nn t*. ~ e
C F C P Fshnwn

cross section in Fig. 3-1 with a fully plastic section, M = Mp and noting that the
neutral axis at this point divides the area in two- narts ---- rliqtnnre /17 t* nv u.vu
--- -- with avo3
The shape factor is computed as
centroid from neutral axis, is
- 2 8

The value of A7 is called the plastic modulus, Z , so that we may rewrite the
moment as
'
..
* a
The plastic modulus and shape factor for a W shape can be computed
manner similar to the rectangular shape of Example 3-3. Here convenient use
3 ,
Mp
--
= Lr,, made of the tables for T shapes, as illi~stratedin the following ex,mple.
'lle,.ratia.of Z / S is termed the shape factor, f.
Example 3-4 Compute the plastic section modulus and shape factor for
Example 3-3 What is the section modulus S,, plastic modulus Z , and th W610 x 241.1 rolled shape.
shape factor f for the rectangular shape shown in Fig. E3-3?
I I , SOLUTION The value of A/2 is readily obtained from the WT table CVT30
I' X 120.5), since this T shape is made from splitting a W610 shape. The
I value in the table also locates the center of the area of the T but is with
respect to,the flange.
-
c-

E From Table V-18 of SSDD, obtain

0
0
xr
I1

Y
i-

+ X
From Table V-3, the depth of a W610 x 241.1 is 635 mm. The total
area = 30.77 X lo-' m2.
&, = d - 2& = 635 - 2(68.6) = 497.8 rnrn = 8.498 m
AJ = 15.39 x lo-' x 0.498 = 7.664 x
Z =- m3
I Figure E3-3
2
The value given in' Table V-3 for 2, = 7.659 x m3 and
SOLUTIONThe elastic section modulus is computed us discrepancy is due to the extra digits used by the computer in computing
materials equations: directly as opposed to rounding for Table V-18 and the use of 0.497
0.498 above.
From Table V-3, the section modulus of a W6 10 x 241.1 is
J ll S =I - =bh3= . - bh2
c 12(h/2) 6 S, = 6.78 x m3
and the shape factor f can be directly computed as
200(0.4)~
S = = 5.333 x m3
6
 , 8 122 STRUCTLTRAL STEEL DESIGN ELASTIC, PLASTIC, AND BUCKLING BEHAVIOR OF STRUCTURAL

portional to strain as defined b

esign method. The current elastic design procedures as found in the several
esign specifications is based on the linear stress-strain response tb the elastic
mit, but there is implicit recognition of the steel behavior beyond the elastic
limit. Elastic design as commonly used places the limiting steel stress as the yield
that used the limiting stress of F, in the elastic design
-
Z bli2/6
/=I50
Z= h( b/12 - b,/iZ 1
( bh2 - - bill,?
Z = b/iZ/3
f = 2.00
safety factor F = 1. A,.safety factor of 1 is unaccept-
ble, as it allows for no future changes in structural use/occupancy, or for
/ ' = L.SI1
(bli' b,/i:

material properties (flaws, under dimensions of sections, and minor mztallurgj-

value of F > 1 is required.

Ideally, every element of a steel structure should have the same factor of
fety. In practice, this is not the case. Flexural member response tends to be the
ost reliable to predict and those members have a minimum value of F.

results in a structure collapse, rightfully have the largest values of F.

z = 4r3/3 z = + ( r i -- r j ) z = from tables
The basic safety factor for the steel members in building construction is
/=I70 f = Z / S = 1.10 t o 1.18
J
32 ---
=- ri Average = 1.1 4
obtained as follows. Let the computed strength of the member be defined as the
ri + r j Modal value = 1.12 d be defined as R. The safety factor can

Figure 3-3 Plastic section modulus and shape factor for selected cross sections.
computed service load
The shape factor is a measure of the increase in plastic moment capacity
over the value of yield moment My, and we have
M, = ZF, = f(SF,)
For F = 1, this new ratio becomes

The plastic section modulus and resulting shape factor for several cross
.sectisn.s.isshown in Fig. 3-3. Now if we take A S / S = AR/R = 0.25 and noting that S / R = F, we obtain

3-3 SAFETY FACTORS IN ELASTIC AND PLASTIC DESIGN

Solving for F, we obtain
I -
Steel design may be based on the yield strength (termed plastic, or limit states,
design) or on an elastic design. In limit states design the analysis proceeds based
on assumed plastic behavior-the member continues to strain from E, to e,, (see
This value of F is taken as the basic value of F for use in the elastic design
Fig. 1-3b) with no increase in load. Elastic design allows for this unique
ethod for structural steel for structures other than bridges for highways and
behavior of steel but limits the working stresses to the elastic region of the
railroads. Railroad and highway bridges are generally subjected to a more
 ELASTIC, PLASTIC, AND BUCKLING BEHAVIOR OF STRUCTURAL
124 STRUCTURAL STEEL DESIGN

s obtained based on section properties from these ultimate loa

hostile environment and a greater possibility overloading, so the un
red to the yield stress Fy and adjustments made until computedf;,
factor A S = AR is taken as 0.29, which gives F = 1/0.55 = 1.82.
The value of F = 1/0.6 ,is modified to 1/0.66 when the cross-section
geometry is such that a plastic hinge can fully develop at the most high
EFLECI-IONS
stressed point. Rolled shapes whose section geometry is such that the plast
hinge can fully develop so that the basic value of F = 1/0.66 can be used are
uld a plastic hinge develop at a point along a beam or column, a very
termed compact shapes. The geometry criteria for these shapes will be considered
ection would result. This deflection would, however, have no meaning
in Chap. 4.
For A-36 steel the basic allowable stresses using the previously defined ould result in a structure collapse. No structure is
safety factor becomes
Fa = 0.6< = 0.6(36) = 21.6 ksi
under actual working load conditions will be elastic values.
(the AISC specification allows use of 22 ksi for this case)
Since plastic deflections result in a structure co
r v Fa = 0.6(250) = 150 MPa (in SI units) ignificance, so for this reason deflections
,
For the AREA and AASHTO specifications, we have stic analytical procedures for both general metho
Fa = 0.55(36) = 19.8 ksi
(these specifications allow use of 20 ksi for this single case) LENGTH OF PLASTIC HINGE
Fa = 0.55(250) = 137.5 MPa
of the plastic hinge for a W shape ca
We note that the optional rounding of 21.6 ksi to 22 ksi (as should be done using d-end beam (refer to Fig. 3-4) as foll
desk calculations) can create a slight computational discrepancy if a digital e horizontal at midspan, the offset to t
computer is used, unless a rounding procedure for this grade of steel is set in the
computer program. The author suggests that the rounding to 22 ksi be done, L
since it is allowed and A-36 steel is the most common grade used. It is not yo = IM, - ( - 1tfp) = 2hlP a
recommended (at this time) to round 137.5 MPa to 140 MPa, since some
The equation of a parabola with the origin of s as shown is
rounding up has already taken place to obtain 250 MPa from 36 ksi.

3-3.1 Factor of Safety for Current Plastic Design

The factor of safety used (called load factor) for plastic design according to t
present AISC procedure is obtained by using the average shape factor f defin
in Sec. 3-2 and illustrated in the computations for a typical rectangular shape
Example 3-3. In elastic design with compact sections, the value of F = 1/0.66 =
1.52. The value of plastic moment Mp = fMy, where the shape factor = 1.12 as
the modal value for all the rolled W shapes. Now using the same l l l i l l l l l l l l l l l l l l l l
working stress fb for either design method, we have
MY- = -
-= MP fMY
1.52s F,S F,S
Canceling the section modulus S, we obtain
F, = 1,.52f = 1.52(1.12) = 1.70 (as used in Part 2 of the AISC specifi
This value of F is used in plastic (or limit states) design as a load factor
beam.
which the working or design loads are multiplied to obtain the "ultimate'
 ELASTIC, PLASTIC,AND BUCKLING BEHAVIOR OF STRUCIIIRAL

NOWwe need to find x such that M, = My = S,F,. Since y = Mp -

have
Y = F,(Z, - S,) = F,(fS, - S,)
Substituting for y and Mp in Eq. (3-l), we obtain
1
8(fF, s,)($)I = f ~ ( 1- 7)s,
"Caricdi'i and F, and rearranging, we obtain for the hinge length defined by
the length x (which is half the rnidspan hinge length),
Figure 3-5 Plastic hinge formation for several beams loaded as shown. (a) Simply s u p p
hinge for failure. (b) Propped cantilever two hinges for failure. ( c ) Fixed-cnd, three hinges fo

When f =, 1.12, the length 2 x of the plastic hinge in the center of the b

beams with concentrated loads, and load. combinations, can be obtained in

when the collapse mechanism has been determined. These two

3-6 ELASTIC VERSUS PLASTIC DESIGN

There are several advantages in using plastic design fo
small one- or two-story structures:
s to produce a collapse mechanism fop
1. The rapidity of obtaining design moments.

sary for the beams in the following e

Offsetting these advantages are several disadvantages:
Example 3-5 Derive an expression for hfp for the fixed-end beam s
1. Widespread availability of computer programs, which can rapidly sol and select a W shape with adequate Z for P,,, = 120 kN.
simple and complicated structures using elastic methods.
P, = 170 k N
2. Most designers have more familiarity with elastic design methods.
3. Difficulty of obtaining the collapse mode 'if the structure is reasonably
complicated.
4. There is little savings in column design (and sometimes for other members
depending on the fabrication methods).
5. Difficult to design for fatigue.

I
In plastic design it is necessary to determine the location of the plastic
hinges that form at locations where M, develops. It is necessary that enough
SIRUCIWRAL STEEL DESIGN

SOLUTION Three hinges are necessary to produce a collapse mechanism

Note that the beam is indeterminate to second degree (no horizontal load
so Z F, = 0 has no significance). From symmetry the three hinges necess
to form the mechanism must be as shown in Fin. - E3-5. The effect of The total Z, required is
fixed-end moments is to reduce the simple beam moment diagram as shown
byi,the dashed lines. For hinges to form it is necessary that the moment Z, = 0.612 + 0.0058 = 0.6178 < 0.6566 x m3 furnished O.K.
val'ue be Mp, and it is evident that Mp will form first at the fixed-end Use a W360 x 38.7 beam.
locations, since the elastic moment is largest at those points. Further It is still necessary to check bracing requirements. For an elastic design
increases in moment increases the elastic moment into the plastic range. It is using Fa = 0.6Fy (commonly used allowable stress), the beam would be
also evident that the only other possible location for M, is under the
concentrated load, since the moment at this point will be the next location 8- = 9 0120(6)
M = - =PL k 8N . m
where the elastic moment is large enough that increases in Pw to P, will
The required section modulus S is
orce the moment into the plastic range. When this hinge forms, the
. !ructure cbllapses (theoretically) and no further increase in load is possible.
With this consideration, we have (again referring to Fig. E3-5)
Use a W410 x 38.7 section.

from which
= 0.60 + 0.0076 = 0.6076 < 0.629 X l o u 3 furnished
Sx(reqd, O.K.
By coincidence we have found a section that has exactly the same mass
!-'or A-36 steel, Fy = 250 MPa. per meter; in most cases sections obtained by plastic design methodswe
Pu = Pw X load factor = 120(1.7) = 204 kN somewhat lighter than those obtained using elastic design, at least when the
beam is indeterminate. ///
Example 3-6 Given the propped cantilever beam shown in Fig. U-6, it is
The required plastic section modulus is ,,q required to obtain a general expression for itl, and design the beam if
ww = 5 kN/m and F, = 250 MPa. Also derive a general expression for the
location of M, in the span.
From Table VI-2 of SSDD, select
W360 X 38.7 Zx = 0.6566 x m3
The beam must cany its own weight, so for self-weight the simple beam
moment is M = w ~ ~ /For 8 .plastic analysis use the same concept as for the
concentrated load, which gives

For the W360 x 38.7, the weight/m = 0.38 kN/m (Table V-3).
LJU STRUCTURAL STEEL DESIGN ELASTIC, PLASTIC, AND BUCKLING BEHAVIOR OF STRUCTLRU STEEL

SOLUTIONThe collapse mechanism will consist of two hinges located as

shown. From statics the moment Mp at B is a maxjmum. From mechanics of
materials, V = 0, where M = maximum. This gives
Rc = wux
Also, using statics Z M B = 0 for segment BC, which gives
wux2
M~ + 2- R,x
0 =
Taking moments of beam segment AC about A ( Z M , = 0) gives

Mp + R,L-;=Ow L2
2
Substituting Eq. ( a ) for Rc in Eq. (c), then substituting Eq. (c) into Eq. (b)
for Mp, we obtain Figure E3-7
x 2 + 2 x L - L* = 0
Right span:
Solve this by completing the square, we obtain
x = 1.414L - L = 0.414L
Now a general expression for Mp can be obtained from Eqs. ( a ) and (b):
Mp = 0.08579~~~~
Using Eq. (3-4) with the given beam length and loading, the value of Mp is The maximum value of M, from either span is used for design (since beam
runs across both spans using a constant section).
M, = 0.08579(45 X ~ 236.26 kN . m
1 . 7 ) ( 6 )=
Mp = 15(1.7)(18) = 459 f t . kips
z,=--236'26 - 0.9451 x low3m3
250 M~=20(1.7)(15)=510ft.kips usethis

beam weight is
0.5 1
AZ, = -(0.945 1 ) = 0.0107
45 From Table 11-2, select a W24 x 68 w ~ t hZ,= 176.4 in3. Check tde beam
Zx.,,,d, = 0.9451 + 0.0107 = 0.9558 !< 1.0861 x weight effect as approx~mately (if a borderline case is found, one may be
m3 O.K.
justified in the additional work for an exact analys~s):
bracing requirements). AM' w, L~
AM; + 22 = -8-

Example 3-7 Given the two-span continuous beam shown in Fig. E3-7,
select an economical W section using plastic design and A-36 steel.
I .

Sofiyno~Two hinges are necessary to collapse at least one span. The and the required AZ, is
values of Mp to accomplish this are:
2.17(12)
Left
Lett span: AZ, = -= 0.72 in3
36
Total Zx(reqd)
= 170.0 + 0.72 = 170.72 < 176.4 furnished O.K.

Use a W24 x 68 section with Z, = 176.4 in3.

 ?32 STRUCTURAL STEEL DESIGN

3-7 LOAD RESISTANCE FACTOR DESIGN occupancy; other values are also used (e.g., 1.5 for maximum sn

Load resistance factor design (or LRFD) is a recent proposal which

undergoing some development as an alternative approach to the cumently used equations for several loadings including wind and snow mi&t re
elastic design method. It is expected that L R F D W become
~ ~ ~ fully accepted by
the AISC within the useful life of this textbook. This forecast is based on the +R 2 l . l ( l . 1 0 + 1 . 4 1 + 1.6PVm,,)
facts that this procedure (at least the essentials, called limit states design) is
.,!;eady accepted in Canada and several other countries outside the United
+R l.l(l.10 + 1.5Sm,,)
States. The current AASHTO bridge specifications (12th edition) provide an +R > l.l(l.10 + 1.4L)
.,hernative design method in steel termed load factor design for simple and
;3~tin~,o.!!s*beams and girders of moderate length which use compact sections. The general objective with LRFD is to assess each item that influences
.--!li the LRFD designs are very similar to each other and to the strengt design of a structure rather than "lumping" several effects together, as,
,jrocedure used in reinforced concrete design. In LRFD, as in rei example, simply adding the dead and live loads to obtain the composite lo
concrete, +,factors are used to reflect uncertainties in the material (in t h s case Larger factors are used with those items that carry more uncertainty, such
t-h% specified steel strength, F,). These factors are under current study with the snow and wind loads (live-load factors of 1.5 and 1.6 versus the dead-load fa
.&rent suggestions as in Table 3- 1.
LRFD uses an equation of the general form
+R = $(FdD + FLL)
where $= analysis factor (also termed importance factor); value cu
suggested, 1.1
Fd = uncertainty factor for dead load with a value of 1.1 suggeste
FL = uncertainty factor for live load with a value of 1.4 sugges LOCAL BUCKLING OF PLATES

2%ble 3-1 Current recommendations for factors

Stress condition Suggesteda Canada AASHTO

~ & s i o nmembers
Yielding (4) 0.88 0.90 1.O
Fracture (F,) 0.74 0.90 1.0 a1 stresses, perfectly plane, homogeneous, and isotropic that is subjecte
sending o m compressive load along opposite edges. Under this stress the ~ l a t
Rolled sections and plate girders 0.86 0.90 1.o press uniformly until the buckling stress is reached, When the buc
Columnsb stress is reached, the plate will deflect in a single wave or a series of wav
0.90 1.o
TJ 2 0.16 0.86 depending on the edge (boundary) conditions and length to width ( a / b ) ra
0.16 < 7 1.0 0.90-0.25~~ with a resultant redistribution of the compressive stresses until, with the additio
TJ > 1.0 0.65 of load, the entire plate is buckled.
:!hear From the theory of plates as proposed by several authorities,j the critic
Webs of beams and girders 0.86 0.90 1.o elastic buckling stress F,, is
Cvnnections
BoltsC 0.70- 1.00 0.90 -
'Welds 0.80 0.90 -

" See Journal of Structural Division, ASCE ST9, September 1978 (contains eight
Fapers on LRFD).
7 = ( K L / ~ ~ ) V F (~K/ =E length factor as given in Chap. 6). Metal Structures (New York: McGraw-Hill Book Company). or Johnston, Guide ro Srabiliry Deri2n
See Sec. 8-10. Criteria for Metal Structures, 3rd ed. (New York: John Wiley 8i Sons. Inc.).
 ,
7

136 STRUCTURAL STEEL DESIGN

ELASTIC, PLASTIC, AND BUCKLING BEIMVIOR OF STRUCTURAL STEEL u7

where F,,= steel stress at the proportional limit (a value of F,

0.70 to 0.75FY may be used)

+,
F, = yield stress of steel
Fcr= critical buckling stress of Eq. (3-6).
rC
If we attempt to solve Eq. (3-6) for the critical buckling stress, s h'[

problems develop, particularly if A < 1. First, we must determine kc. Wh

general expression for kc has been given, it is necessary to adjust this f i--- h ---I

rious boundary conditions that are possible. This has been done by several k c = I 1 5 ,re111
i =O;Y
7lthorities, but as a convenience the author has further combined the effect of FIdnze ,dnle >I. it ,h.111=
?/[12(1 - p2)] = 0.9038 to give values shown in Table 3-2 for k; [i.e.,
a(0.9038) = 3.6151. If the X term is less than 1, it is necessary to iterate to Fcr.
B' is is illustrated as follows. k; = 0 63 l l ~ n g e \
Rewrite Eq. (3-6) in terms of k:, to obtain k: = 4 5 for web
:p
a
9

2 Figure 3-7 Compress~oncharactens~csfor rolled shapes shown. Note that 111s generafly nec
8 13 Fcr = X E ~ ; ( ; ) to mvesbgate the cntical b / t ratlo, which may be as shown for a W shapc w t h a cover plate
, welded or bolted.
.
J,

$, dew divide through by

2
Fcr -
A -~k:(i) Compression charactenstics (kc and wldth b in compress~on)of three rolled
Ye will temporanly hold this equation. Now using Eq. (3-7) for X 9 t h shapes are g~venin Fig. 3-7. The kc values shown have been found to agree
F,, = 0.755, we can, with some rearranging, obtain reasonably with tests. Adjustments In kc are necessary because very few plates
are free of imperfections and residual stresses
Fcr
-=
0.1875 c2 If we use the value of k: = 0.63 shown in Flg 3-7 for the flange of a W
4 - Fcr section and a SF = 2.00 and Fc, = F,, we obtain
Since Fc, is on both sides of Eq. (3-9), we must solve for Fcr/X by trial. Once the
= 11 3 say 11
value is obtained, this can be used in Eq. (3-8a) for the ratio of b / t , which is
~sually,the item of interest. From Eq. (3-8a) the b / t ratio is

Lt =pC
Fcr / A
65
say -
vz
'9 For kc = 3.615, X = 1, and Fc, = 0.754, the limiting b / t ratio for A-36 steel is
The current AISC specification allows a b / t (uses b,/21/) ratio of 6 5 / 6 . Note
",/@=Em
t 0.75 x 36
= 62.3 also that if we consider the web, k: = 4.9, we obtaln

I t is often useful in using Eq. (3-9) to set up a table of A vs. F, with values from 190
say -
FCr = 0.754 to F,. For A-36 steel, typical values are as follows:
6
/
which is also in AISC.
i;,,ksl A E,, ksl Values of k, = k, are also shown in Table 3-2 for the cr~ticalstress to
/F produce shear buckling. The critical buckling stress for shear can be derived in a
27., 1.00 29 M)O similar manner to that for compression, with the substitution of an appropriate
28.0; 0.922 26 738
2ii 0 0.741 21 489 buckling coefficient k, to obta~n,from Eq. (3-6),
33.0 0.407 11 803 k,n2XE
36 0.0 0 F C ~=S
- 12(1 - p 2 ) ( b / l ) 2
I 138 s d u m STEEL DESIGN ELASTIC, PLASTIC, AND BUCKLING BEHAVIOR OF STRUCTURAL STEEL 139

. I C

I is usual to assume that the four plate edges are simply supported in shear and where all terms have been previously identified except be, which is shown in Fig.
i
t e sha'ar stress F,, = </\/5. This value must be combined with the safety
factor of 1/0.6 so that the design shear stress becomes
3-8. For a very long, thin, simply supported plate, it appears that the theoretical
value of k: is 3.615 for this equation. The use of post-buckling strength of plates
a . . ." C r7 C
is not often directly evaluated. It is used more often in a more indirect manner;
Fs = crs -
- l y - 'y for example, AISC allows use indirectly via Appendix C3, which states: "When
SFX* 1.67xfl 2.89 the width-thickness ratio of a uniformly compressed stiffened element exceeds
Mo~t~practlcal steel design problems consider either buckling in compres- the applicable limit given in Sec. 1-9.2.2, a reduced effective width, b,, shall be
sion 6r b'uckling in shear. w h e r e both stresses act simultaneously, the reader used in computing the . . . ."
should consult books such as those by Bleich, Johnston, and Timoshenko and
Goodier cited in an earlier footnote.
PROBLEMS

3-9 POST-BUCKLING STRENGTH OF PLATES 3.1 What is the allowable b,/2t, for any W section using a steel with F;. = 50 ksi?
Answer: 9.2.
3.2 What is the allowable b,/2!, for any W section with F, = 250 MPa?
Experimental evidence shows that a buckled plate does not result in immediate
3-3 What is the plastic moment capacity (kN . m) of a W 9 2 0 x 200.9 section for Fy = 345 MPa?
failure. Rather, there is a considerable strength reserve attributed to the effect of Answer: 2865.2 kN . m.
the adjacent plate material, which restrains the buckling and allows transfer of 34 Verify 2, in Table V - 3 for a W 6 1 0 x 241.1 rolled shape.
any post-buckling load increase to the unbuckled zones. This situation is 3-5 What is Z, for the geometrical shape shown in Fig. P3-5 if it is used for a beam?
idealized in Fig. 3-8, which illustrates the central buckled zone in the loaded
width b. On either side are strips that confine the buckling and are loaded to a *
lesser effective load fe. The concept of effective width be is applied as the sum of
the two strip widths on each side of the buckled zone. When the effective stress
on these two edge strips (stress on width of be) reaches a value such that
deformation is constant with no further increase in load, the full load capacity of
the plate has been reached. The difference between the initial buckling load and
this new value is the post-buckling strength of the plate.
The stress fe may be evaluated using Eq. (3-8), to obtain
3-6 Select the lightest W section to satisfy bending for the span and loading shown in Fig. P3-6. Use
F, = 36 ksi and plastic design.

4 kipa't't

3-7 Select the lightest W section to sorisfy bending for the span and loading shown in Fig. P3-7. Use
Fy = 250 MPa and plastic design.
Answer: W 6 1 0 X 101.2.

340 k;\'

50 kN1m

Ngum 3-8 Effective width for post-buckling plate capacity. +-

7 -- in 8 IT]--1 FIgwp P3.7
3-8 Select the lightest W section for bending for the beam shown in Fig. P3-8. Use F, = 345 MPa
and plastic design.

3-9 What is the uncertainty factor on both S and R to produce F,

reasonable alternate nonequal values to produce the same effect?
- 0.5Fy? What are two

Answer: S = R = $ .
i

>
144 STRUCTURAL STEEL. DESIGN
.

of our attention will.be'

.
directed to cases where
,

1. Has no end moments (simple beam), or

2. Has moments at the ends of each span (continuous or
I .

Beam loadings will consist of both dead and live loads.

in 'terms of whether a series of concentr
roduced depends on the general framing pl
rhe beam dead load is its own weight. Wher
self-weight may be a significant part of the total beam load. where th
small and/or the external loads are also sma
small, and here the strength/weight ratio of
any case, the section should always be chec
loads, including the beam weight.
Beams may be classified as:

1. Girders: main load-carrying members into which floor beams

as shown in Fig. 4-1. Chapter 10 consider
for bridges. ----
.-ii--ii--

2. Joists: members used to carry roofing and floors of buildings.

3. Linfels: beam members used to carry wall loads over wall openings.
3. Spandre/s:exterior beams at the floor level in building construction used to

compressive stresses in the masonry. #!fI

5. Stringers: members used in bridges parall
slab and commonly frames into transverse (
6. Floor beams: secondary members of a floor
members in bridge construction into which stringers f
4- 14).

In most cases, particularly for maximizing economy, a

loaded so that the bending is about the strong a
section property tables in SSDD). Occasionally, the bending
:he weak (Y- Y) axis and in some instances there is simultane
both the X and Y axes. In nearly all of the appli
bending, the load is considered to be applied th
or S shapes. The shear center for these shapes
' .,ad position produces simple bending about ei

When the load does not pass through the

channels, angles, and some built-up sections, u
a torsional moment is produced along with the
taken into account to avoid overstressing the member.
The design of beams requires an analytical iterative a n
the shear and moment diagrams based on the se
ms simple framed to columns. Wind is resisted in both directions using cross bracing at fou
( e ) Girder (left to right) and floor beam system in an office building. (n
Cloxup k$s.*
'framing of floor beams and joists. ( n ) Roof beams and spandrel for f l a t roof buildin?. (hi
- .-. "-...,.--
I... U l L L b ULOlUIY

Y -------- -

inoment together with the loads can be used to back-compute the critical shear.

on mechanics of materials:
For bending:

For extreme fiber stresses:

For shear: N OF BEAMS BY THE ELASTIC METHOD

The W shapes will usually be used for beams. On occasion, S shapes, M shapz
channels are used, depending on location, mode of connection to

desired (see Fig. 4-2)

c = distance from neutral axis to extreme fiber
I = moment of inertia of cross section
S= I / c = section modulus of section (both I and S are tabulated in
section property tables such as Tables 1-4 and V-4 of SSDD) e same time one with an adequate S..
Q = statlcal moment of area above point where shear stress is being ISC) have the sections commonly used
determined = A7 (refer also to Flg. 4-2) r beams ranked in descending order of S and wth respect to the X-X (strong)
t = thickness of beam at point of shear stress investigation
weight, so that considering a group as
V= critical shear force from shear diagram or direct computations defined by extra llne spacings, we have

1. The largest S In any group is at the top of that group

The section is checked to ke sure that it is adequate to carry its own weight, 2. The hghtest section in the group is at the top and the heavlest is at the
and finally the working load deflestions are checked. Sometimes deflections

radang allows rapid detemnatlon of the most economical roIIed

 . .-
I
"4.

STRUCTLTRAL STEEL DESIGN /DESIGN OF B

SOLUTIONDraw shear and moment diagrams as shown in Fig. E4-1 Next,

obtain M,,, = 276 ft - kips.
With a uniform loading on the beam, the top flange is in contact with
I Chechng the beam weight, we obtain

something; thus we will assume full lateral support, so that

Lb = 0 < LC ST,,, = 138 + 3.24 = 141.2 < 143 in3 furnished
rind use Fb = 0.665 = f x 36 = 24 ksi. Use a W14 x 90 beam.
Rearranging Eq. (4-l), the required section modulus with strodg axis
bending is Example 4-3 Given the beam span and loading shown in Fig. E4-3
approximately the beam system of Example 4-1. Most of the SI
will be approximately the same as the fps problems (but not
conversion of units), so the reader may be asslsted in developing a feeling
We must find a section with a somewhat greater value of S so that the beam. for the SI units and selection of section sizes. Use F, = 250 MPa.
, <-
weight can be carried. From Table 11-1, select

Checking the beam weight, the additional S required is

S,,,, = 138 + 2.45 = 140.45 > 140 furnished

If we use this beam, the actual fb= 24(140.45)/140 = 24.077 ksi, or this
beam will be overstressed 77 psi based on the given load conditions. We
may:

1. Use engineering judgment that this small overstress is acceptable, or

2. Use the next lightest section, which in this case is a
W24 x 68: Sx = 154.0 > 140.45

With Sx so much larger and beam weight the same as in the earlier
section, it is not necessary to again check the beam weight.
The reader should observe that from a deflection stand~ointthe W24 is
the better selection. ///.
Example 4-2 What size of beam should be used in Example 4-1 if the beam SOLUTION First, draw shear and moment diagrams.
depth is limited to 16 in? Second, obtain M,,, = 436 kN . m. Assuming full lateral support, w
obtain
SOLUTION Sreqd is approximately 145 in3, since the weight will have to be Fb = 0.67 X 250 = 167 MPa
larger than 68 lb/ft of the most economical section selected in the example
The required section modulus is
Always use the maximum depth possible when the beam is laterally sup
:;ported. Therefore, by inspection of the W16's in Table 1-3, select , ,

W16 X 89: S = 155.0 in3 d = 16.75 in > 16 N.G.

Note that 167 MPa x lo3 = 167 x l d kPa, which is consistent with 436
Try W 14 x 90: S = 143.0 in3 d = 14.02 in (next try) kN . m. Check Table VI-I for Sx somewhat larger. Find:
W12 x 106: S = 145.0 in3 d = 12.89 in
 154 STRUCTURAL STEEL DESIGN

b
Check AS, for beam weight = 1.11 kN/m: Col. Col cur
-r >,- :!.,. .. .,,.-, ..
? I

S,,, = S, + AS, = 2.611 + 0.0532 = 2.6642 < 2.8841 m3 O.K.

Use a W610 x 113.1 beam.

Continuous beams are designed similar to simple beams. major difference

when
" ,
using
. .-.. AISC specifications is that if the section is compact and is not a
cantilever beam, the section may be designed on the basis of using either:

1. 0.9 X largest negative moment in span, or

2. Positive moment based on maximum positive moment from moment diagram
+
0.1 x average of the negative span moments,

chever is larger. These moments are based on gravity (D + L, D L S, + +

etc., and not wind) load moments. Where the beam or girder is rigidly framed
into a column, the design moment value for the column at this point can also be
reduced 10 percent. This procedure is based on recognition of the method of
plastic hinge formation and resultant transfer of moment from the neg+tive zone
to the pos~tivezone until a hinge finally forms at that point. This procedure will
,The + M for design is
now be illustrated. + M = 1i1.3 + 57.86 +2 140.6 (0.1) = 121.2 < 126.5kN - rn
Example 4-4 Use the computer output of Example 2-2 and check and/or Since the largest moment is - 126.5 kN . m, this is used to select the section-
redesign members 22, 23, and 24 (refer to Fig. E4-4 for the gravity load Note that the compression flange is part on the top and part on the bottom,
moment (which inspection of Fig. E2-4d indicates are the critical values for with inflection points as shown on the moment diagram of Fig. E4-4.
design). Use F, = 250 MPa. Tentatively assume that LC will be larger than the values for the bottom
flange; if so, no lateral bracing will be required and we can use Fb = 0.665,
* SOLUTION
From the computer, output obtain beam end moment values of
- 57.86 + 140.64 - 86.86 +
86.88 - 140.64 +57.86 kN m .
These values are needed to complete the shear and moment diagrams shown From Table VI- I , select a
in Fig. E4-4.
Since we are using the same section for all spans, the largest design W410 x 46.1 : S = 0.7735 x lo-' m3
( k m o m e n t is in span 1 or 3, by inspection. Note the moment in span 2 is all LC = 1.78 m > 1.20 O.K. for no bracing

r,, t n ~ a t i v e ,requiring possible lateral bracing on the bottom (compression)

.?/hange if & = 2.3 > LC.We must check this possibility when the section is
selected.
L, = 2.16 m
The point of inflection in either outside span produces 0.87 m of
unsupported compression flange on bottom of beam. This will be deemed
adequate compression flange bracing, since the remainder of the compres-
sion flange (on the top) is braced by the floor. We will have to use a
156 W U C l V R A L STEEL DESIGN
1 , DESIGN OF B W L I S FOR B M N G
"I

midspan brace to the bottom flange for the center span, to produce distance or column base plate width in the following way. The needed reactio
distance is obtained by considering an area in web compression defined by the
- - 1.13 rn < 1.78
& = -2'3 reaction length + an additional distance using a 1 : 1 (45") slope through the k
2
t distance of the section. The section property tables tabulate k for the several
&Check the beam for self-weight (weight = 0.45 kN/m). By proportion rolled sections. The k distance is measured from the outer flange face to the top
( s i ~ c ethe beam loading is uniform) of the fillet transitioning',the web-to-flange interface. At this location the
resulting web area in compression is nearly (if not exactly) a minimum. At a
reaction the area in web compression is
S,,,, = S + A S = 0.7575 + 0.0086 = 0.7611 x l o v 3m3 < 0.7735 furnished A, = (A' + k ) t ,
Use a W410 x 46.1 beam. The allowable stress at this location is taken by AISC (see SSDD Sec. 1-10.10.1
The reader should verify that the method used to obtain A S is both to be
correct and the most practical means available. gince a W410 x 59.5 beam F, = 0.75Fy
was used in the initial computer analysis, it appears that the problem will
have to be reprogrammed after the column design/revision has been made At a reaction with j = R I A , , we obtain
'.i?a later chapter. ///

At a concentrated load in the span, the distance k can develop on both sides
4-4 WEB BUCKLING AND CRIPPLING of the load as illustrated in Fig. 4-6. For thls condition, we obta~n

Web buckling is an out-of-plane web distortion resulting from a combination of

iarge d/tw ratio and bending stress. The unbraced length of compression flange
,nay also contribute to web buckling. Web buckling is controlled by limiting where N = reaction width; a basic value of 3 f in or 89 mm (width of standard
dtfier.the.d/tw ratio or the stress that can be used for the given d/tw ratio. This brick) is often assumed; iV = width of column or load-deliverin
1s allowed for in the several specifications. Web buckling is illustrated in Fig. element for interior loads
4-5b.
R =.reaction or other concentrated load, kips or kN
Web crippling can occur if the web in-plane compressive stresses are
5ufficiently large. This can occur if reaction distances or load-bearing plates
used to deliver column loads to the beam flange are too narrow. Web crippling
can also occur if a uniform load on the flange is too large for the web thickness.
Web crippling control will be obtained by determining the required reaction

Figure 4-5 Web failures to avoid in design. (a) Web crippling. (6) Web buckling.
Figure 4-6 Bearing length for concentrated loads on beams accord~ngto AISC spcnf~cahons.
 DESIGN OF BE4bf.5 FOR B

-5 SHEAR CRITERZA

When large uniform loads are carried through the flange to the we The shear stress distribution across any section subjected to bending c
computed using the equation presented earlier:
be necessary to check the compression stress& and limit the valuedto
& 5 0.75Fy VQ
f, = -
It
Example 4-5 What is the allowable reaction for a W16 x 40 using the basic A plot of shear stress using this equation is illustrated in Fig. 4-7. We note that
value of N = 3; in, with A-36 steel? What column load can be transmitted the average shear stress based on
using a W 8 X 31?
v
SOLUTIONFrom Table 1-3, obtain for a W16 x 40: f" = -
dlw
k = 1.03 in differs somewhat from the maximum value shown in Fig. 4-7 (in this case about
t, = 0.305 in 23 percent) but is considerably easier to compute. AISC allows use of Eq. (4.T)
mr.
R = (N + k)tW(0.75<) = (3.5 + 1.03)(0.305)(0.75 X 36) = 37.3 kips. for either rolled or fabricated (plate girders) sections.
The USHTO and AREA specifications simply allow computation of f,
based on the area of the "gross" section. This can be interpreted as in the AISC
specifications (i.e., f, = V/dt,).
in = N. The allowable shear stress F, for rolled sections is computed as:
P = (N + 2k)t(0.75F,)
= [8.W + 2(1.03)](0.305)(0.75 X 36) = 82.8 kips F, = 0.405 (AISC)

F, = 0.335 (AASHTO)
l e What is the allowable reactidn for a W460 x 74.4 .section
~ x a m ~ 4-6
using the basic value of N = 89 mm and F, = 345 MPa? What column load F, = 0.35Fy (AREA)
can be transmitted using a W200 x 46.1?
Shear stresses seldom govern in building construction unless the section
both very short and heavily loaded, as illustrated in the following example-
k = 27.8 mrn
tw = 9.0 mm
50 100 \I PJ
1 ' 1 1 ' 1 ' 1
Note that MPa X mm2 x = kN (low3not shown). The column lo
(assuming base plate same size as column depth) is: for a ~ 2 0 dx 46.1, the \
depth, d = N = 203 mm. C114h0 X 1 0 5 . 7
X: 9 ~ 1 1 ~ ~
I:,< T r ~ b i ero gcr T. 1

P = [203 + 2(27.8)](0.009)(0.75 X 345) = 602.2 kN /// i= <I 2 Y rzc

107 4 \!I'd

Bridge reactions are generally supplied by special fabricated bearings, and

-'~A;4'SHfO-~pecificationsrequire that web stiffeners will be used when the web
shear at the bearing is f, > 0.75F0. Column loads are usually not carried by -1.7
/ 71.6

4-7 Theoretical and average shear stress distribution on section for conditions shown-
 160 STRUCTURAL STEEL DESIGN

4-7 What is the span length and load/ft for a beam of A 4-6 STRONG- VERSUS WEAK-AXIS BENDING
shown Fig. E4-7 such that either F, or Fb will control? m a t length
required for the reaction?

II k~p\/ft

Depth limitations or other considerations may require onentation so

bending is produced about the Y-Y m s . The allowable bending stress in
. -- ,L- Figure E4-7 AISC specification for compact rolled sections bent about the Y- Y axis and
Data for a W24 x 94:
SOLUTION solid round, square, or rectangular bars may be taken as
dz24.31in k=1.53ln
tw = 0.515 in S, = 222.0 in3
wL where B= 0.005 for fps units
f,=--
2twd
- 0.44 = 0.0019 for SI units

W L ~
f -
--8Sx = 0.665
From Eq. (a),
0.4(36)(2)(24.31)(0.515) 360.6 about both the X and Y axis takes place.
W = =-
L L /
Substltutlng this w into Eq. (b), we obtain
4-7 DEFLECTIONS
--
360'6L - 0.66(36)
8% Deflect~onestimates under working load are often required to ensure that floo
24(8) (222)
L = = 9.85 ft * indows to fail to function properly. Cracks can be produced in
360.6 x 12

Back substltutlon gives w = 360.6/9.85 = 36.61 hps/ft. Checking, we ob-

tain

36.61(9.85)2(12) = 24.0 ka O.K.

fb = 8(222.0)

36.61(9.85)
= 14.4 ksi 0.K.
fv = 2(24.3 1)(0.515)
The reaction length N based on R = WL/~= 180.4 kips and F, =
0.75Fy = 27 ksi is a failure of the roof drainage system by pluggng from accumulat

(N + k)twF, = R flections for simple beams can be computed using superposition of

1 . 1
N = - -R k
t w Fa

- 180.3 can be computed using the general differentlal equation

- 1.53 = 11.44 in
0.515(27) EIylV = - w
 L STEEL DESION

For L
and
=. 6.5 m, we obtain C, = - 173.98 (computer output = 0.0043 rad use superposition to flnd the stresses at cntical locations, using .,
r! .

and for shear

It appears that Ax occurs at + M = M,,, at x = 0.87 + 2.37 = 3.24 m from
the left end. For this value of x in the preceding equation we compute

Equation (4-9) is commonly used for spandrel beams and roof purli
where there are both vertical and horizontal force components. When the load is
= 12.6 mm (approximately in) // applied to the top flange of the member (common for purlins on sloping roofs)
and is separated into components perpend~cularand pafallel to the X and Y
axes, the force parallel to the Y axis does not pass through the origin of
-.-8 BIAXIAL BENDING AND BENDING ON UNSYMMETRICAL This is a very complex stress state that involves biaxial bending and torsion.
SECTIONS approximate solution is obtained as
I6
1. :

The preceding sections have considered bending about either the X or the Y axi
of W, M, or S shapes. In all these cases the moment produced by the load
zpplied perpendicular to the X or Y axis with a line of action through the o
of axes. This produced the type of bending stress that can be computed usi
Eq. (4-1). The W, M, and S sections are symmetrical with respect to
,lane containing the X and Y axes and thus produce principal axes.
xres are such that
that is, using one-half the section modulus of the section for the tangential
bending component and the (+) sign used to obtain the stress.
An alternative form of Eq. (4-9) is widely used for biaxial bending and
similar form for combining bending and axial stresses. To obtain this form,
setf, = Fb and then divide both sides of the equation by the allowable stress
(noting that this is done even if Fbx# Fb), delete the (-), and reorder to ob
1%
+ -=
MY
1.0
FbXSX F b s,
When this is done, a section with any combination of moments and s
modulus producing 1.0, or less, is satisfactory; one should try to o b t a u v
as nearly 1.0 as possible. Any value larger than 1.0 represents an overstress.

Example 4-10 Deslgn a roof purlin using a channel section for the si
The latter equation for Ixy is a product of the inertia term produced when at sheds and for the most cntical loadlng cond~tionof Example 2-5. Space the
least one of the X and Y axes is not an axis of symmetry. Principal axes a purlins 6 ft horizontally as shown in Fig. E4-10 and use sag rods at the
produced as mutually perpendicular axes through the centroid of area such th midbay spacing, giving an unsupported length of 12.5 ft for bendin
the moments of inertia are a maximum with r e s ~ e c tto one of the axes and moments about the Y axis of the purlin.
minimum with respect to the other. The principal axes are axes of symmetry
symmetrical sections but can be found for unsymmetrical sections. If SOLUTION
Additional data from Example 2-5:
product of inertia Ixy is not zero, the axes of interest are not principal axes.
Dead load = 20.0 psi
Live load = snow load = 25 cos B
4-8.1 Symmetrical Sections with Biaxial Bending re

= 25 cos 14.04 = 24.25 psi

,the X and Y axes are principal axes but th 9
to the principal axes (but the resultant .?'Total
. = 44.25 psf
ikin ipf axes), biaxial bending is produced. For this Vertical load/ft = 6.1 8(0.04425)
.PQ '"
* cessary to determine the components of load perpend s = 0.273 kip/ft
I
1
 .!&
>?
15
P
STRUCTURAL STEEL DESIGN DESIGN OF B M M S FOR BENmING 1
i.
{ values, we obtain veneer to crack if it actually occurred. Eng~neeringjudgment would h
be applied to decide if this angle is satisfactory with a deflection that coul
-*hk
- tJ
fb = - 2046x - 1220y (if set to 0, gives location of neutqal
t point A , x = - 0.0419, y = + 0.1357 m.
be as large as 14 mm but has a good probability of being much less than
this.
fb = - 2046( - 0.0419) - 1220(0.1357)
I

= - 79.8 MPa (compression) 4-9 SHEAR CENTER OF OPEN SECTIONS

b t point B, x = - 0.0419, y = -0.0673.
fb = 167.8 MPa ( + = tension) The shear center locates the point with respect to a cross section to apply
flexural load so that no twisting (or torsion) occurs when shear stresses due t
bending act on the plane through the point. Thus. if the loading passes through
the shear center, the section may be analyzed for simple bending and shear
using Eqs. (4-9) and (4-9a). I f the beam loading does not pass through the she
center, a torsion moment is developed that produces torsional s
stresses of
Ve' t

where Ve'= shear and shear eccentricity with respect to the shear center
t = thickness of element where shear stress is desired
J = torsional constant of section, for a thin rectangle
J= -
bt
b/t 4 10 (section webs and some flanges)
3
bt
J = - - 0.21t4 b/r 4 (stubby flanges as for channeis)
3
The computation of the shear center is complicated for all but the sim
At point C, x = 0.1 10, y = - 0.0673. shapes. Fortunately, most sections have the shear center at a convenient Iocation
fb = - 142.9 MPa ( - = compression) (see Fig. 4-10), for example:

If we use fb = M c / I : 1. If a section contains an axis of symmetry, the shear center is on the axis.
At point A : 2. From (1) it follows that the shear center of all sections with symmetry abo
both axes is on the intersection of the two axes (all W, M, and S shapes).
14'21(0'1357) (lo3) = 57.3 MPa (compression) 3. For all sections consisting of two intersecting plate elements (angles, tee
fb =
33.63 etc.), the shear center is at the plate intersection.
At points B and C:
The shear center E, (using symbols as in Table 1-6) for channels is rea
14'21(0.0673) (lo3) = 28.4 MPa (tension)
fb =
33.63 derived to be
These values are considerably different'from those computed using E, =
bl rr
2b,$ + h1t,/3
metrical bending.
Check the deflections approximately using Table IV-5 for a where b, = b, - t,/2
supported beam: h'= d - t/ = average section depth
9, t,= flange and web thickness, respectively
Using these values for a C10 x 30 with d = 10.00 in. b, = 3.033 in, t f =
0.436 in, and t, = 0.673 in, we obtain E, = 0.705 in, as in AISC and Table 1-6
This deflection is computed at 14.2 mm (over f in), which could cause the
1':l STRUCTURAL STEEL DESIGN /$ DESIGN OF BEAAIS FOR BENDCjG

At midspan:
x=--
L v = o

82
Shear enter l o ~ a t ~ o n f, = -= 3.96 ksi << 0.6FV(22ksi)
20.7
At x ,= L / 4 = 3.13 ft:

E, - 4-10 Shear center location for several rolled shapes. M, = - 0.35(12.5 - 3.13')(12)
w(L - x2) - = j.7 in kips
2 2

Qe reader should note that there are several equations used in various
dp,;g-~ tables which give Eo values for channels slightly different from those of
d ,,!-14), depending on the assumptions made for shear flow in the flanges.
The computation for shear center of more complicated shapes is beyond the _-.
,,

s c c ~ ~ofe this text, and the reader is referred to any text on advanced mechanics
= 0.3 + 31.6 = 31.9 ksi >> 0.65,
of ?,~terials. Note that if the channel beam were designed without considering torsio
1' I-.: bending stress for conditions where torsion is developed by the applied and shear center location, the apparent bending stress at rnidspan indicates
lo22 * ,t passing through the shear center is approximately that the section is considerably overdesigned. When considering the sh
center effects, the section is not adequate unless it can be assumed that
analysis is too approximate and that the end connections are such as to
restrain rotation so that the larger bending stresses do not develop. The
whe .e M c / I = usual bending stress computation author would assume that the section is considerably underdesigned and use
V = shear at a distance x along beam from origin a different section, as there is too much design risk involved for the small
by= width of flange being stressed amount of savings obtainable from using a lighter section.
x = distance from origin of axis to V and where fb is desired The torsion shear stress can be evaluated as
If= moment of "inertia of flange being stressed; approximately Iy/2 Ve' t,
for a channel L=,
Example 4-12 What is the approximate bending and torsion stress in a J = contribution of web 4 two flanges
C10 X 30
4
channel loaded as shown in Fig. E4-12?
.,", " A

U
B r ~ i kv e n e e r

= 0.927 + 0.152 = 1 08 1n4(vs 1 22 In SSDD)

The discrepancy between J as computed above and In the SSDD tables is
due to using a more refined computatlon in the tables, which accounts for
I+- I 2 4-5' Figure E4-12
the increased stiffness at the flange-to-web intersect~on.We w11 use the
SOLUTIONFrom Table 1-6, the section properties of a C10 X 30 are: table value of J = 1.22.
e' = Eo + 7 = 0.75 1 + 0.649 = 1.40 in
S, = 20.70 in3 Eo = 0.705 in
f, = 1'092(1'40)(0.436)
1.22
I, = 3.94 = 21, b, = 3.033 in
= 0.546 ksi which appears satisfactory ///
 174 STRUCTURAL STEEL DESIGN . DESIGN OF B E A M FOR BENDING 1

4-10 DESIGN OF LATERALLY UNSUPPORTED B

= shear modulus = = I 1 150ksifory=0.3
I,.;: moment of inertia about the Y axis
s,= section modulus about the X axis
L= unsupported length
floor or roof. Cases where the floor system consists of metal decking placed onto
the bea 5 ay not provide adequate lateral flange support if the resistance is
$$4 ~olely4 . f s i c t i . n . In these situations, carefuily placed welds to take, say; 2 to
, . 3 ~erc$$., ::%he flange force (0.02 to 0.03A, f,) will provide adequate lateral
ll",,. ,

. 'i
bracing."'&>,v6ry large lateral restraint is not required, as evidenced by the
applicatidrrq~only hand force to restrain reasonably sized beams against lateral
f
buckling in.laboratory tests.
The lateral bending and warping of laterally unsupported beams was
illustrated.,in Figs. 4-2 and 4-3. There are several situations, particularly involv-
ing vertical flexural members and beams used to c a m column loads across large
'"Pen area4 crane runway girders, and continuous beams in large spans, where
the compression flange is not in contact with decking and others, where it is not
practical to provide lateral bracing except at the ends and/or at only a few
interior span points. For conditions where the compression flange is laterally
unsupported for some distance, the resulting column-type action may result in
flange buckling, a warping (partial-to-full section rotation), and lateral bending.
Just the superposition effect of vertical compressive bending stresses adding to
the compressive stresses due to lateral bending as the section deflects out of
plane may produce yield or buckling stresses in one side of the compression
flange. In any case, where this situation is possible, the allowable bending
stresses are reduced. This reduction is critical when the section is light and deep,
since with lateral bending the warping resistance of the flanges and web is small.
'If lfo,true for long unsupported lengths, since, similar to buckling of a long
~'~?~&h~f~t'rakes a smaller load (or stress) to cause the compression flange to
hen* laterally out of plane and producing a tendency for warping.
'A 'theoretical combination of torsion and lateral bending resistance can be
made to obtain the critical compression flange buckling stress as

or, alternatively,

~h~ AISC specifications can now be developed using the preceding several .
where J=c,t:orsion constant previously defined and given in the table of ations and some additional simplification and rounding of numbsrs. Since
ib
:'section properties in SSDD ut 1946, the AISC specifications have used the following formula based On
C, = warping constant =$ t,bf3(d - t,)2 g a safety factor of 1.67 and a factor Cb to
 where rT= radius of gyration of the compression flange + 1/3 of the
sion web area taken about the Y axis; these values are also given in

a c - r ~ n tfor moment gradient (the latest modification) into Eq. (4-21), to obtain stead of 0.41d, to obtain
12 OOOC,
Fb = 2 0.6. [AISC Eq. (1.5-7)]
Ld/A j
82 700Cb
Fb = (SI units)
MIAj
wher: ,I,= area of compression flange = bjtf
d=depth of section; note that the ratio d/A, is computed and tabulated
I
in tables of section properties
Cb = 1.75 -I- 1.05(Ml/M2) + 0.3(M,/ M,)' 5 2.3. In this equation M, is
always the smaller and M, the larger moment at the end of the'
'*:
unbraced length. M l / M 2 = (+) when moments are of same sign
(producing reversed curvature) and Ml/M2 = (-) when of opposite The minimum L/rT ratio for using Eq. (4-26) is found by equating F,
sign. Use Cb = 1.0 when the moment in the interior of the unbraced .6Fy to obtain (with slight rounding):
length is larger than either of the end moments and regardless of
sign.
T!I-: Column Research Council proposed as a n alternative to Eq. (4-20) the
basic ,. Iumn formula

T a k i ~:SF = 1.67, this equation becomes

SC Eq. 1.5-7) when ( L / r , is less thdn that obtained from the equation abo
e minimum L / r , ratio at which Eq. (4-27) applies is found by equating E
(4-26) and (4-27) to obtain (again slightly rounded):
where the term Cc= slenderness ratio for which residual stresses cause inelastic

-
buckliilg or the transition L / r from inelastic to Euler (or elastic) buckling:

Substituting this value and using 0.667 instead of 0.6 for the first term (1.0) and
approximately 0.75 for the second term in Eq. (4-24), we obtain

I F, [AISC Eq. (1.5-6a)] (4-26)

 1/U STRUCTURAL STEEL DESIGN

m*y be used according to AISC, which means that we may use but will be neglected.

SOLUTIONUsing the given load geometry, complete Fig. E4-13 by drawing

the shear and moment diagrams also shown. Next, from Table 11-1 of SSDD
+ [Eq. (1.5-7)12 for a W36 x 300, obtain LC = 17.6 f t and L, = 35.3 ft. Since L, > I,
Eq. (1.5-6b) (46 > 35.3), we immediately note that Fb < 0.6Fk.
2
acEording to Eq. (4.22) and as illustrated in Fig. 4-1 1.
The design of laterally unsupported beams may be summarized as follows:
* . M = 204P + 952.2 in . kips
1. Initially assume that Fb = 0.6Fy and make a tentative section selection using From Table 1-3, obtain the section properties for a W36 x 300:
a table such as Table 11-1 of SSDD, which also gives LCand L,. If the actual
unbraced length Lb I LC or L,, a direct solution can be obtained since the
unbraced length will not be a factor.
2. If the tentative section indicates L, < Lb, the unbraced length.may be a Alternative computation of d / A,:
design factor. Using the tentative section or one somewhat larger, compute Fb r, 1.680 in
d = 36.74 in b, = 16.655 in =
using Eq. (4-23). If this equation gives a value of Fb that satisfies bending, a
so:ution (but possibly not the best) is obtained.
3. If Eq. (4-23) does not supply a (satisfactory) solution, the designer must use
either Eq. (4-26) or (4-27), depending on the L / r , ratio. Use the largest F,
from either Eq. (4-23) or from the controlling equation (426) or (4-27) as Alternative computation of r,:
determineckby L/r,.

A table of unbraced lengths versus allowable bending moments such as

Table 11-3 or VI-3 may be used to obtain a direct design or to give an indication
of sections that may possibly prove to be adequate.
The following examples illustrate the use of the equations for laterally
unsupported beams.

Example 4-13 Given a girder using a W36 x 300 supporting two columns as
shown in Fig. E4-13, what is the maximum column load using the AISC
specifications and A-36 steel? Assume that the girder is restrained against
rotation only at the ends. The columns may provide some lateral restraint The slight discrepancy is due to the somewhat approximate computations
used; however, this method is satisfactory where r, must be computed-
Cb = 1.0, since the moment diagram shows that the end moments are 0
and the interior span moment is larger.

Check to see if AISC Eq. 1.5-66 [Eq. (4-27)] applies:

Use AISC Eq. 1.5-66 [Eq. (4-27)]:

 -m
" rli

I&.; :TRUCTLTRALSTEEL DESIGN

Also use AISC Eq. 1.5-7 [Eq. (4-23)]: Estimate the beam weight as approximately 0.1 x load:

Fb = 12 oOo(1) - 12 oOo (0.03 x 35 + 30)(0.1)

= 16.6 ksi = 0.09 kips/ft
Ld/A, 46(12)(1.3 1) 35
Use , f b = 16.6 ksi (largest value):
t. M M due to beam = -=
f b C - =
Fb + M = SFb
S
204P + 952.2 = 16.6(1110) M,,,,, = 237 ft . kips .-
,. .. .'.
P = l7 473'8 = 85.6 ki With a large unsupp
204 somehow estimate the bea
selected sections with moments givkn up to
'"he preceding example was easy to check, since the beam size has been ft is only slightly more, use this table as a guide and select the fol
s e l e c ~ ~and
d it is only necessary to determine the allowable bending stress. tentative sections:
most design situations the problem is more of an iterative process, in that wh
th? loads.~w,igiven,we do not know what section wi1l:be ~ 1 x 296: S, = 131.0 L, = 39.9 (Table 11-4 for W12)
011c;:casion: one may use charts such as those in AISC; which give the allbwa

mom:nt fbr several unbraced lengths or, alternatipely, computer-genera 14 x 90: S, = 143.0 L,, = 34
ta&s such as Table 11-3 or VI-3 of SSDD, which giye'the a1
se!ccted shapes for several unbraced lengths Lb. We\should also observe that
Eq. (4-23) controls the design, the use of A-36 steel is the 'most 'economic W16 x 100: S, = 175.0 L, = 28
so!:~!ion (the reader should verify why this is true).
W18 x 97: S, = 188.0 L, = 24.1
Example 4-14 Given the laterally unsupported girder
for a crane runway in an industrial warehouse. Select t
that also limits deflection to L/360. Use A-36 steel. ry a W14 x 90 and check AISC Eq. (1.5-7). Take Cb = 1.0.
trolley travels on a 90-lb railroad rail fastened to the top of the flange
(90 lb = 90 lb/yd = 30 lb/ft). 12000
Fb = (35)(12)(1.36) =21ksi

M , = F bS .r
21(143)
=-------=250ft,kips>237 O.K.
12
Check the deflections:
I, = 999.0 in4
From Fig. 4-8, obtain the needed deflections at forward load:
~ . ., ,, .., .,..,..,-. P = 3ok Figure E4-14
,,,
15(19)'(1612 + 15(13)(16) ( L Z- 1-31 - 162)
Aload = 3 EIL 6EIL
~ N the maximum moment. Write an equation for M in terms of
S ~ L U T IFind
x and take d M / d x = 0.
1
[15x + 15(x + 6)][L - ( x f 6 ) ] = -(47.80 + 43.03)(1728) = 1.5 in
M = 3 IL
L
dM - 0 = -60x - 270 + 30L
- - x = 13 ft from left end -=--
dx 360 3 360
'('2)-l.2<l.jin
M = (30 x 13 + 90)(35 - 13 - 6 ) = 219.4 f t . kips Use a larger section; a side computation indicates that a W18 x 97 does not
35
 DESIGN OF BEAMS FOR B .

where & = 10 ft or 3.05 m

W = roadway width (including any shoulder allowance)
The number of lanes N is an integer; thus a lane fraction should be rounded
to the next integer. A reduction in load intensity is allowed when the number
lanes N > 2 as follows:

N Percent of live load

2 100
3 90
4 or more 75

The distribution of wheel loads to stringers (or longitudinal beams) as

as transverse floor beams is based on a paper by Newmark ("Design of I-B
Bridges," Transactions, ASCE, Vol. 114, 1949) and is given in terms of span
divided by a coefficient as given in Table 4- 1.
The computation of shear and bending moments for deck stringers is bas
on a'simple beam analysis for the critical load (either truck or lane). After thes
values are computed and adjusted for impact, the distribution factors from
Table 4-1 are used to obtain the design effect on any of the interior stringers.

Table 4-1 Wheel load distribution coefficients for AASHTO bridge design
used as S/coefficient where S = stringer (beam) spacing
Bending moment and shear (lateral distribution 1 lane 2 + lanes
on steel interio? beams) ( ) = SI value

For: concrete floor 7.0 (2.134) 5.5 (1.676)

steel grid deck <'4 in (102 mm) 4.5 ( 1.372) 4.0 (1219)
steel grid deck > 4 in (102 mm) 6.0 (1.828) 5.0 (I 524)
Note: For exterior stringers use statics to o b m n effectwe wheel loads causmg
bending and shear (refer to Fig. 4- 15).
Bending moment in transverse floor beams w~thoutstnngers:
With concrete deck 6.0 (1.828)
With steel grid < 4 ~n (102 mm) 4.5 (1.372)
> 4 in (102 mm) 6.0 (1.828)
Bending in transverse floor beams w t h stnngers: see Fig. 4-15.
Bending moments in concrete deck slabs (mam relnforcement transverse):
S = effective span length = clear flange edge-(+flange edge dlstance + b,/2
+ 0.61
P, = rear wheel load of H 20, H 15, H 10, etc. (H 20 orSHS 20 = 32/2 = IQkips = 72 !d)
9
M = - -S + 2 P , f t . ~ p ~ / f t M=-----
32 9.74 r,
.P, LN, m
Use M ' = 0.8M if three or more stnngers.
<-
I

. f *-
""
>, am* ..*
a Exterior stnngers to be at least as large as mtekor.stnngers. to allow for future bndge w d z n i n g
488 STRUCTURU STEEL QESIGN DESIGN OF B%W FO

Ballast plate: include 10 percent for corrosion protection Step 3. Find the beam section.
0.015(77 k~/m~)(0.76)(1.10) The required sectlon modulus is based on F, = 0.555, since th
pression flange is laterally supported.
Estimate cross-beam weight
S =-=-=
Miscellaneous maintenance, storage of ties, material F, 591
137.5 4.30~10-~m'
Total =9.01 kN/m ould use a W760 X 147.3 section; however, we will arbitrarily go
the dead-load moment is ost economical section, W760 x 160.7/ 1.58:
Sx = 4.8997 x m3 t , = 13.8 > 8.50 mm
WL2 + track
Md = -
8 t, = 19.3 mm d = 758 mm
-
- 9.01(5.79)2 + 4'44/2(5.79 - = 37.8 + 4.8 = 42.6 kN .m We note that the weight is 0.08 kN/m larger than assumed, but the sec
Y* 8 2 modulus is more than adequate for this small difference. Check the sh
The dead-load shear is - V - 280.5
f
= 26.8 << 0.35 F, O.K.
" dtw 0.758(13.8)
Vd = 9.01(5.79) + 4.44 - 28.3 kN
2 2 (The end connections to the plate grder wlll be designed in Example 8-9
Step 2. Find the live load on the beam. (D = 0.76 m, s = 5 ft8=1.524 m) Use a W760 x 160.7/ 1.58 rolled section.
1.15PwD
P =
s 4-13 COMPOSITE BEAMS
- 1.15(110/80)(80 x 4.448)(0.76) = 280.6 kN
1.524 A composite beam is one whose strength depends upon the mechanical interac
tion between two or more matenals. Reinforced concrete beams are actualIy
with P / 2 placed on each rail. The live-load shear is composite members but are not generally designated as such. Most often the
-
Pw -_ -280'6
=
140.5 kN term "composite beam" in building and bndge construction is applied to a ste
2 2 section on which a concrete floor or bndge deck has been cast. The concrete
The impact factor is securely bonded to the steel section via carefully designed shear connectors so
that the concrete and steel act together as a tee beam. Figures 4-16 and 4-17
3 L2
I = -30'5 + 40 - - where S = beam spacing, L illustrate shear connectors in position to bond the concrete to the beams so
S 150 composite action is obtained. The shear connectors shown are called shears
30.5 3(5 79)2 and are welded to the beam (and through the deck pan in Fig. 4-16). Other
0.76
+ =-
40 -
150
= 79.5 percent
of shear connectors can be used, but shear studs are most common.
When there is no particular effort to bond the steel beam and concrete flo
The design live-load shear is
eck, relative slip occurs at the interface of the two materials and the result 1
140.5 x 1.795 = 252.2 kN omposite section. In some cases we may design for a specified amoun
The design live-load moment is site action. Actually, there wlll always be some small slip due to

ML =
280.6(2.177 x 1.795) = 548 kN . nequal deformations in the shear studs, concrete, and steel
2 ractical purposes it can be neglected In composite design. The brief introd
composite design made here will only consider full composite action.
c,, The total design moment is
e method of construction and code specifications are significant desl
1' &vt M,,,,, = M, + ML = 42.6 + 548 = 591 kN . m meters in design of composite sections. Figure 4-18 illustrates the effe
The total design shear is h of concrete to use in a transformed section b' using the modular
n = E,/E, as in several mechan~cs-of-matenalstexts.
Vdeslgn = 'd + 'L
The modular ratio depends on the 28-day concrete design strength 1,
= 28.3 + 252.2 = 280.5 kN shown in Table 4-2.
a . " &
~i a*",:
190 STRUCTURAL STEEL DESIGN DESIGN OF B U \ S FOR B

Figure 4-16 Composite building construction. (a) Shear studs in 4-17 Composite bridge deck (two lanes interstate hlghway uslng six stringers. (a
construction and metal deck for form. ( b ) View of inetal deck site bridge deck nearly ready for concrete wit reinforcing and shear studs in place. Side vertical
encased by concrete floor. (c) Bottom vlew of metal deck resting on floor beams and girders. Metal is for parapet. ( b ) Closeup wew of shear studs (c) Undernew showng stringers and unshored
deck is welded to floor beams wlth shear studs. work for deck.

The two basic construction methods for producing composite beams are:

1. Shored construction. The steel beams are put in place and the formwork for II

the concrete slab is added. This assembly is then shored (braced or propped) 9
so that no (or relative small amounts of) deflection can occur, and the * .0

concrete is poured. After the concrete has hardened for'about 7 days (about
70 to 75 percent off,' obtained), the shoring is removed. At this point the
stresses in the composite beam are due to the dead weight of the steel beam
plus a proportionate share of the concrete deck. (,'I
2. Unshored construction. The steel beams are placed and formwork (metal
decking may be the necessary formwork as in Fig. 4-16) supplied for the i = *pJl#
se srn.~llestvalue ol b
concrete deck (refer to Fig. 4-17). The concrete is poured and at this time the b = LI-I
steel beam carries the dead load of steel, formwork (as used), and the b=b,+i
concrete. After the concrete hardens, any formwork is r AASHTO ~ = I s + ~ I 2, I

at this stage of the construction that the steel beam h h=b,+ 16i -\-\SIITO h = I:[

stressed with the weight of the steel beam plqs a proporti 4-18 Effectwe flange wdth of composite sections by both AlSC and M H T O s w c a t i o n s
weight of the concrete deck. ny differences ldentrfled. ( a ) Edge. (b) Intenor.
w Y

192 STRUCTURAL STEEL DESIGN

use an overst~e~C@torof 135 percent (1.35 to account for &e
te capacity), tKk" stress on the bottom flange of a steel composite

2000 13.8 2550 @SI)

2500 17.2 2850
'" 3000 20.7 3100
3500 24.1 3400
4000 27.6 3600
5000 34.5 4000
6000 41.4 4400

The actual bending stresses in a composite beam are limited by the materi

ccvnstruchon methods is illustrated in Fig. 4-19.

The ultimate load of a shored composite beam has been found to be on th

S, = section modulus of steel beam referred to tension flange

S, = section modulus of composite beam referred to tension flang

Shored

!----- b -----4

igure 4-20 Stress dstnbutlon at ultmate load m compoate beam

 DESIGN OF BE4!!S FOR

Use $ connectors with a capaclty of 8 hps/connector; the nu

is
= 264.6
fq I = -2
4
-
8
= 33.07 use 34 connectors
Use two connectors at each section at a spacing w = 4d =
center to center. The longitudinal connector spacing is

S 30(12) = 10.6 in > 6 0 (minimum)

= ------
2(34)/2
< 81 (maximum)
Step 7. Check the shear with both dead and live loads acting-
V = 30.96 klps
V
j--=-= 30.96
= 4.9 ksi << 0 . 4 5 O.K.
dt, 1g(0.355)
Step 8. Check the deflection under live load. Note that th
deflection has been built out of the floor via flow of the wet con
placing.

A = L! 5(0.13X8)(3@)(12~)
=~ W ~
= 0.32 in
L
<-
384EI 384(29 000)(2056.1) 360

Example 4-17 Design a composite bridge section for a two-Ian

12-ft lanes + two 8-ft shoulders and pedestrian walkways) as
E4-17a. Other data: HS 20 loading; /; = 27.6 MPa (class A
ksi);f, = 415 MPa (reinforcing steel); A-36 steel for the stnngen;
specifications; 2 X lo6 loading cycles.

SOLUTIONThe two parapets and rallmgs,.ulll bg pl ed after the

placed. We will assume that this load~n$ is &%ed by the two
stringers, which are not deslgned here(ddjsd.t& interior s t k g e r checke
see if it is adequate as an extenor stnnger, since the exterior stringer mur
at least as large as the interior stnngers). We wlll also assume that a fu
40 mm of wearing surface will be added as the initial surface deteriora
Step 1. Design the concrete deck slab.
The steel beam spacing is selected at 2033 rnm. as shown in Fig.
Estimate the stringer flange width b, = 420 mm (a W920 se
compute the effective stnnger spaclng (AASHTO Sec. 1-3.2) of the interior
beams as

s = sac< - b, b/ = 2033 - 210 =

+- 1823 mm
2
The slab dead load, including future wearing surface based on D = 230
 m

STRUCTURAL STEEL DESIGN

roadway, the live-load moment is

The design moment for the deck slab using strength design is

M,,, = M, ~ M =L 2.33 + ------

+- 5(18.7) = 33.5 kN - m
3 3
Take d = D - 40 (allowing 25 rnm clear cover for bottom reinforcement

. • I 40 rnm haunch I
.
bridge decks AASHTO Art. 1.6.16):

A: - 0.02148AS = - 1.0140 x loT5

Solve for A, by completing the square:
I I I
I
I
v The maximum ratio of A,/A, = 0.0214, based on f: and f, and taking 75
Cross frames
. . . .. percent of the value to ensure initial steel yield. The actual steeI percentage
is
77.200 .'I.., 0'0004829
4 (? 5550 = = 0.0025 << 0.0214 O.K.
0.19(1)
The minimum percentage is
of concrete, is
(0.230 + 0.040)(23.6 kN/m3) = 6.37 kPa JY

Increase 10 percent for roadway debris, snow, etc.: Use A, = 0.0033(0.19 x 1.0) = 0.000627 m2/m.
Step 2. Design the steel stringers.
6.37 X 1.10 = 7.0 kPa Note it was necessary to design the slab so that the dead load carried by
The dead-load moment is computed as the stringers could be computed. Assume unshored construction-the
stringers must carry the dead weight of the deck slab until the concrete
hardens.
Dead load of slab: (0.230 + 0.040)(23.6)(2.032) = 13.00 kN/m
The impact factor for the slab is (using L = s = 2.033 m)
Haunch: 0.040(0.42)(23.6) = 0.40 kN/m
z2= -
l5 = 0.375 > 0.30
L 38 + use 1 = 0.30
Rolled beam assumed =3.60 kN/m
Since the s / L = 2.033/22.4 = 0.091 is so small, one-way principal slab I

2~iscellaneous,
including debris, formwork, etc. = 1.00 kN/m
reinforcement perpendicular to the traffic flow will be used in the slab
design. For this case and five spans providing slab continuity across the (I I! Total = 18.00 kN/m
A4J.Z STRUCTURAL STeEL DESIGN
r $>
DESIGN OF BWki.5 FOR BENDWG' 21)3 :
The maximum dead-load moment at the center of span is conservative, so that we may check stresses at the center of span under
and add the composite effect of M L (unshored construction sequznce).
Step 3. Find the properties of the composite section (refer to
E4- 17b).
Assume that 15 mm of concrete is not usable in composite
because of wear and surface deterioration. Neglect the area of con
The maximum live-load moment, including impact, is based on the the haunch.
AASHT3 truck and the stringer spacing.
15
I = 0.25
=
+
22.4 38
From Sec. 1-9, the live-load moment due to total truck load on span of this
length is

b = 0.42 + 1.613 = 2.033 rn

b = 0.42 + 161 = 3.86 m
Use b = 2.033 m; for f,' = 27.6 MPa, n = 8.

The distribution factor for live load to a stringer of a bridge with two or
more lanes is obtained from Table 4-1 as It is usually sufficient to select a steel beam one or more sizes Iess than
that required for noncomposite action. The section required for noncom-
Factor = -= -
2.033 = 1.213 wheels posite action is approximately (and flange laterally supported)
1.676 1.676
Slnce 1.213 wheels = 1.213/2 = 0.6065 axle, the adjusted (for deck1 [slab s x = - = MT 1129 + 1090.5 = 16.14 x 10-3 m3
0.555 0.55(250)
action) stringer bending moment is .*
Tentatively, try a W920 X 364.6/3.58:
The maximum shear occurs when one of the rearmost truck wheel's [144-kN Ix=6701.3x10-6m4 S,=14.67X10-3m3
(32-kip) axle] load is at the beam end: I

t, = 34.3 mm b, = 419 mrn

This value must also be adjusted for plate action and impact: A = 46.52 x m2
Check the shear:
We note that M, is at the center of the stringer span, whereas ML is slightly
off center. The error of summing MD + ML for design is negligible and
F
= 26.1 MPa << 2
3 O.K.

Check the dead-load deflect~on:

= 0.04402 m (44.02 mm)

This much deflection in a 22.4-m span would tend to aggravate during
placing the concrete to a level top surface (requires an additional 44 mm of
 i..) sgF- snnL DnsmN

i k c r e t e at the center of the span. This will require a temporary Final:

or, preferably, have the stringers cambered at the mill with a
(and being sure to place the cambered side up on the job I top '%=-!!??!-=26.1Mpa
' I*. b d'faster would occur). The remainder of the beam properties Stop 41.77
ek.) are within the assumptions, so we can continue. 26.1
f,=-= 3.26 MPa << 0.41 O.K.
@
s
,
6' Compute I of the composite section (refer to Fig. E4-17c). 8
' ,
AY = ZM*.,
A = 46.52 b'd' +
= 46.52 + 0.254(215) = 101.13 x m2
The section stress profiles are shown in Fig. E4-17d. Check 100 percen
overload stresses against 150 percent of allowable stresses:

2(1091) -
f, = ----- 106.2. MPa
20.54
f,,,,,,, = 77.0 + 106.2 = 183.2 MPa < 1 5(0 55F,) also O.K.

Figure E4-17c

= 6701.3 + 46.52(0.327)~x Id + Dedd load Lnc. l o ~ d

Figure E4-17d

= 6701.3 + 4974.3 + 210.4 + 4235.7 Step 4. Design the shear connectors.

= 16 121.7 x m4 Use three 20-mm studs, as shown in Fig. E4-17e, with L = 150 mm
The section modulus values are Thus the L / d = 150/20 > 4 is 0.K:The studs must resist the smaller of
l6 121a7 0.85f:bdf
""P = I31 +
+ 40 215 = 41.77 x 10-3 m3 vh~= -7 (neglect haunch)
l6 121'7
= 20.54 x 10-3
SbOt =
+
458 327
m3

I The stresses are: = 5130 kN controls (after computing V,)

A F 46.52(250)
Initial: M"
f;=f,=-= --
I 129 - 77.0 MPa v , = x = = 5815 kN
S, 14.67 2 2
. .
@6 STRUCTURAL STEEL DESIGN
DESIGN OF B W M S FOR B

1. -.v 95
1.
Ka
R .A Figure W17f

At X = 7 m from supports, the live-load shear plus impact is

Figure W 1 7 e I' [ 22.4Ra = 4078.8
The ultimate strength of the shear stud is R, = 182 kN
V = 182(0.6065)(1.25) = 138 k N
S,, = 0 . 4 d 2 m E=4 7 4 0 e
= 0.4(0.020)~(27.6 x 24 900); (w/SI adjustments p = 0.323(%) = 0.5022 mm
= 132.6 kN < 0.6 1 max. s ~ a c i "n zallowed 0- .----
K~
. The minimum number of studs required from the end to the middbn Using the spacings above, the first 3.44 m requires 1 1 groups of studs,
tance of L = 22.4/2 = 11.2 m) is 4 next 3.43 m requires 9 groups, and the next 4.33 m requires 9 zrou~s.for
5130 total of 29 groups at 3 studs each, for a total of 87 studs vs. th
N, = -
132.6
= 38.7 use 39 (multiples of 3)
I
for ultimate shear considerations.
Step 5. Design the diaphragms.
With studs in groups of three, there will be 13 groups. Now set,the
, Lateral load:
spacing based on fatigue.
The maximum range of live-load shear is 0 to 214.6 kN. . Wind at 24 kPa X exposed surface area and, referring to Fig
E4-17a: H = 4 6 0 + 225 + 270+ 916mm = 1.871 m
Zr = c d 2 = 0.0541(20)~= 21.6 kN
Curb load at 500 lb/ft for side opposite
Wind = 24(1.871)
Total

Note: there is no wind load from the truck because the deck acts as
diaphragm.
nz. = ----
Pitchp = - 301.6) = 0+323 With diaphragms spaced at 5.55 m,
S, 200.6
Pdiaph
= 52.2(5.55) = 289.7 kN (interior diaphragm)
At X = 3.5 m from supports, the live-load shear plus impact is (refer to
E4- 17f) - 289.7- 144.9 kN (exterior diaphragms)
2
Use all same-size diaphragms; also, the diaphragms must be at Ieast one
third and preferably one-half of the grder depth. Arbitrarily select
nr<
Y to
W530 65.5/0.64: d = 525 mm > --j- O.K.
X
-
$rk@yproportion since Q / I = constant, 1, = 8.9 rnrn > 8.0 mm O.K.
A = 8.39 x m'
r, = 32 mm L = 2.033 m
 STRUC'IURAL STEEL DESIGN
S
A check (not shown) as a column indicates P = 994 kN >> 289.7 k ~ so,
section is amply adequate. Use the same section for both ends ahd inten
points.
Step 6. Check the live-load deflection.
Convert the live-load moment to an equivalent uniform load as a first
approximation.

I I I
--I I 1- Flgum FX-18

SOLUTION Since the girder is symmetrical on the floor plan and

are closely spaced, assume that the loads will give a uniform girder I
Taking the maximum allowable deflection as
bv, = +(FdD + FLL)
L 22400 -_
-3--

1000 1000
- 22.4 mm > 17.6 mm O.K. Using load factors glven in Sec. 3-7 (and Table 3-1) and noting th
no live load reduction for floor area since the load factors are the
If the designer assumes that the actual moment diagram resulting from thre
statistical terms, we obtain
I wheel loads is sufficiently close to a uniform load diagram, the deflection a
computed is adequate; otherwise, a more exact analvsis should be m a d e w, = 1.1 [ 1.1(22 x 0.075) + 1.4(22 x 0.080)]
= l.l(l.815 + 2.464) = 4.71 kips/ft
The design moment is
wUL' 4.7 1 (25)2
4-14 BEAM DESIGN USING LOAD RESISTANCE FACTOR
-,-.,-.--=- ,w ---. I Mu=--
8
-
8
= 367.97 f t - kips

Mu = OFyZ (4 = 0.86 from Table 3-1)

The design of a beam for bending using LRFD is relatively straightforward. It
necessary to have separate values for dead and live loads. The allowable bendin
stress is taken as
From Table 11-2 of SSDD, obtain a W21 x 62 with Z = 144.4
Fb = +F, + = 0.86 20.99 in, and t, = 0.0400 in.
Mu = +F,Z Check the beam weight:

vu= '
+( v3 ) dt," By proportion,
Aw Z
Awu

A Z = U --
= 1.1(1.1)(0.062) = 0.075 kip/ft

0.075(142.6)
= 2.27 in3 O.K.
Wu 4.7 1
and Mu is computed using factored dead and live loads. This is illustrated in the
following example. Check the shear:

Example 4-18 Given the floor system shown in Fig. E4-18 for an off
: building using simple framing. Take the loads as dead = 75 psf
and live
80 psf; use LRFD design and A-36 steel.
Make a preliminary design selection of a W shape with depth no
= +- F~ dtw = 0.86- 36 (20.99)(0.40) = 180.3 >> 58.9 kips O.K.
v3 v3
factor. Use a W 2 1 X 62 section.
 , .$r"
, ,
212 STRUCTURAL STEEL DESIGN
i
-be
7'-
DESIGN OF'BEAMSFO

gn the floor stringers, noting full lateral support for the compressi
= 250 MPa. Use the AASHTO specifications.
interior floor stringers of Fig. P4-19 using composite daigo
no twisting occurs due to the eccentricity of the brick. Use A-36 steel. ction under live load to L/800. Use j; = 4 ksi.
Design the interior floor stringers of Fig. P4-19 using composite desip and unshored cons
Limit live-load deflections to L/800. Use f; = 28 MPa.
~ollowingare miscellaneous beam problems for laterally unsupported spans and
Answer: W410 x 59.8 for Al.
considerations.
4-a A column load of 160 kips is carried across an open work area as in Fig. P4-23. The
unsupported length is 39.5 ft. Select the lightest W shape with the deflection limited to L
any grade of steel if A-36 is not adequate. Note that you must assume your end conditio
Answer: W36 X 300. (Simply supported.)
Answer: W410 x 46.1.

--
4-14 Design the floor beams spanning between columns assuming supports and = 345 ~ p ~ .
Answer: W410 x 46.1. i t>OZ
4-15 Do Prob. 4-8 using unshored composite construction. U s e x = 3000 psi. ,7 - P= - 'lo L\

&
T
*
4-16 Do h o b . 4-12 using composite unshored construction but use F, = 250 MPa steel (instead of +(I y6 [II

F, = 345 MPa steel; why?). U s e x = 21 MPa. B J . ~ pi~r:,

4-17 Redo Example 4-16 using shored construction instead of the "unshored" construction of he
example.
4-18 Redo Example 4-16 using a W18 X 46 and see if a workable solution can be obtained.
'1
For Probs. 4-19 to 4-22, refer to the cross section shown in Fig. PC$-19. The bridge span will be
assigned by the instructor (36 to 46 ft or 11 to 14 m). If not assigned, use 40 ft or 12.5 m for the span.
For exterior beams the maximum possible load due to truck is one-half of the truck load. The dead
30 5 '
i .; :. 111
il t

i
/i
an ,HS 20 truck and A-36 or& 250 MPa steel. iI
I ._.I'
,
steel. Use the AASHTO specifications.
A m e r : W33 x 130.

24 ~~d~ h o b . 4-23 with p = 790 kN and the metric dunensions shown

g. 4-6 is required for the beam selected in h o b . 4-23?

Anrwer: N = 8.50 in.
4% ~f the beam selected in Prob. 4-24 supports the column load of 790 kN carried
200 x 86.3 column section, what length N of Fig. 4-6 is required?
Answer: N = 222 m.
27 A w section is to be used in a span of 10 f t to carry a midspan concentrated load of 650
lect the section, check for shear stresses, and compute the values for both the reaction and
ncentrated load, using A-36 steel and AISC specifications.
28 A W section is to be used in a span of 3.5 m to cany a midspan concentrated load of 2ICO
lect the section, check for shear stresses, and compute the N values for both the reaction and
ncentrated load using Fy = 250 MPa steel and the AISC specifications.
Answer: W840 x 299.1, N = 260 mm for reaction.
4-29 Redo Example 4-18 if @ = 0.90 and FL = 1.5.
4-30 Redo Example 4-18 if Fy = 50 ksl.
11000 - 14000 rnm Annuer: W21 X 50.
4-31 ~~d~ Example 4-18 if D = 3.75 kpa, L = 4.0 k ~ a L, = 7.75 m (bay). and the joist l e W h =
= 250 MPa and LRFD with factors suggested in k c . 3-7.
Figme P4-19 6.75 m. Use
Answer: W530 X 92.3.
 I , ' ,
,~ ', , ,

214 STRUCTURAL STEEL DESIGN

4-32 Redo Example 4-10 for the lightest W or M section.

Answer: W10 X 22 or MI0 X 22.9.
4-33 Design the roof purlins for Prob. 5-20 (of the next chapter) using the lightest rolled W or C
section and with a sag rod at midspan. Use the AISC specifications and A-36 steel.
Answer: W14 X 43.
434 Obtain the lightest W section for a beam span of 45 ft with two 20-kip concentrated loads at 15
ft from each end. Lateral support is available only at the ends and concentrated loads. Use A-36
steel and the AISC specifications.
Answer: W27 X 84.
4-35 Obtain the lightest W section for a 30-kips/ft uniform load on a 15-ft span. Also find the
reaction distance N. Use A-36 steel and the AISC specifications.
218 STRUCTURAL STEEL DESIGN

Round bar n all these uses the tensile' strength of the steel is used. In 'thi
uration plate buckling or warping is not a consideration. In some
I
I Squdre bar s, however, specifications will require a minimum amount. of
I J Fldt bar o r plate ss for esthetic and safety reasons.
I,
%; enerally, tension members may be categorized as rods and bars,
tural shapes, built-up members, and wires or cables. Several of
bers are illustrated in Figs. 5-1 and following.

ALLOWABLE T E N S I O N S T R E S S E S

e AISC allowable tension stress of members, except eyebars, is limited to

F, = 0.6 F, (gross section area)

1 W shape
F, = 0.5FU (net section area)
e AASHTO and AREA allowable tension stress is somewhat more conserva-
(5- 1)

F, = 0.555 (5-2)
cification further limits this basic stress to the lesser value of

ut the net section is used for both these equations. For steel with not over 80
i the basic tensile steel stress is governed by Eq. (5-2) for AASHTO design.
On the net section across the pin hole of an eyebar (see Fig. 5-3), th
owable AISC stress is

Allowable tensile stresses for several steel grades are shown in Table 5-1,
ere the reader should note that established practice allows rounding of the
ues for A-36 steel to the values shown for both AISC and AASHTO/AREA
In all cases, except eyebars, the tension stresses must be computed based o
both the gross and net cross-sectional area when using AISC specifications. Ody
the net area is required for AASHTO and AREA specifications. The net area is
the gross (total) area where welded connections are used. The net area is the
least effective cross-sectional area for all other cases as where bolt or rivet holes
are used for mechanical fasteners at the ends or where holes and/or
Figure 5-1 Tenston members. See Fig. 5-2 for cables used as tension members. (a) Structural shapes reductions occur along the member.
used for tens!?? members. (b) Upset bar. ( c ) Threaded bar and use of a turnbuckle to adjust bar The effective net area at approximately the root of the thread of thread
lei&. Applicable for square and round bars. ension members using the AISC specification is
$+'7
1

where D = nominal outslde diameter of threads

n= number of threads/ln (or the SI equivalent of n/25.4)
 DESIGN OF TENSIOS & E M B E R S UP

le 5-1 Allowable tensile stresses for specifications and equations shown

Anchor rod

On gross section not at pinholes.

On gross section at pinholes.

For A-572 steel.

(strand or rope)

-3 GENERAL DESIGN CONSIDERATIONS

principal factor is how to affect a connection of the tension member to the

tension member.

5-3. Next in simplicity would be some kind of threaded bar or cable. Here
, 1
2.'- STRUClVRAL STEEL DESIGN

,$4:;
ii.
.

r-
t .
1 . I,. = A* 1%. 5 81
!
'>
., . 4. ,.. .,...,.
I
1.33Ab S A , 5 1.5

r 2 0.5" (12.5 rnm)

4
i-
..
I
I /
I
I
I I
t
1.; 2 5-3 AISC eyebar dimensions.

several problems develop, including the area in tension for the

and fitting the member into the structure. This problem is usually solved b
of a turnbuckle or by having an extra threaded distance on one end to ta using r, of element to compute L ' / r between stitch bolts.
the slack. Bars and cables used as tension members are generally giLen a
initial tension when installed to eliminate any tendency to s
rattling when the structure vibrates under service loads. The
sometimes useful in "tightening" up the remainder of the structure., .
Cables may be strands of 'wire rope, with the terms "bridge stra
"bridge rope" being used to specify the structural quality of the cables.
wires are not used in structural applications; rather, strands that are a d stretching it so that the component parts are fitted together.
n~ectsof 7 to 61 or more single wires are wrapped around a cen The modulus of elasticity of bridge strand and rope may be taken
produce a symmetrical section. Wire rope is produced by laying several
helically around a wire core. Commonly wire rope.consists of 6 to 37
T.:bie 5-2 gives design data for selected sizes of both bridge
r0y.e.
Cable connections commonly take the form of one of the configurations
i1lus::ated in Fig. 5-2. In the connections shown, the end of the cable is carefully
c!:~ned and then fed through the opening in the connector. The cable end is
then broomed (strands separated somewhat), carefully cleaned, and molten zinc
at about 850°F is poured into the wire matrix. After the zinc cools, and the 67 (and larger) 23 000 158 CCO
connection is again cleaned and assembled, the member is ready to be installed.
Tiiis method of attaching the end connector produces a joint at least as strong as
the cable., Occasionally (but not shown), the cable can be inserted into a longer
c~.:r?ectorwhich is squeezed (termed swaging) to produce a friction connection.
 DESIGN OF TENSION ~ ( E . M B ~

b
Table 5-2 Selected cable design data
rl

! 4 Welght Area pu

in mm Ib/ft kN/m in2 m2 ( X lo-') bps kN

I project and design uncertainties.
B dge Strand (single strand with multiple wires)
I
13 0.52 0.008 0.15 0.0968 29.0 129 -3.2 Shear Flow and Maximum Effective Cross-sectional.Areas
16 0.82 0.012 0.234 0.1510 46.6 207
18 0.99 0.014 0.284 0.1832 56.2 250
f 19 1.18 0.017 0.338 0.2181 66.0 294
22 1.61 0.023 0.459 0.2961 89.2 397

3. For all other shapes, including built-up shapes, with at least three fasteners
Bndge Rope [6 X 7 (6 strands of 7 wues/strand)] a line (Fig. 5-5 has five fasteners in line), A, = 0.85A,.
; 10 0.24 0.004 0.065 0.0419 13.0 58 4. Any tension members with only two fasteners in a line, A, = 0.75A,.
13 042 0.006 0.119 0.0768 23.0 102
; 16 0.65 0.009 0.182 0.1 174 36.0 160
f 119 . 0.95 0.014 0.268 0.1729 52.0 231
: "22 1.28 ,
0.019 0.361 0.2329 70.0 31 1
1 25 1.67 :0.024 0.471 0.3039 91.4 407

Bndge Rope (6 X 19 (6 strands of 19 wires/strand)]

1; 44 5.24 0.076 1.47 0.9484 286.0 1272
2 50 6.85 0.100 1.92 1.2387 372.0 1655
2f 64 10.60 0.155 2.97 1.9161 576.0 2562

Bridge rope [6 X 37 (6 strands of 37 wires/strand)]

3 75 15.1 0.220 4.25 2.7419 824.0 3665
3f 90 21.0 0.306 5.83 3.7613 1110.0 4938 At point I : Uniform >tress d ~ s t r i b i i t ~ u~i is s u r n e dI I I a n s l y s ~ i
2 . S o ~ r i eload traiisferred to gusset p b t e s I e ~ v l n pP' Load
4 1,00 27.0 0.394 7.56 4.8774 1460.0 6494 transfer is at flanges w h ~ c hresults In qurllitati~estresses
-- across sectlon sliown. Interior web stresscs w1II be lower
than P1.4, due to shear lag. Thus. In long j o ~ n t sthe web
may tear d u e t o large diifercntial s t r ~ i n sresult~ngIn I!
progressive tension failure at [tie tbrwdrd end ot'joint
3: Additional load trsnsferred to gusset plates I e ~ v i n gY .
4: A11 load transferred t o gusset plates and tension member
stresses are zzro. '

. - , . -
228 S p U
,-.,,a,
C l U R A L STEEL DESIGN
..',...--,..
A,
..4

(0)

T ~ p i i . l ilr u , , iccrlon

\.ur< i b ~ l i i t n gIS prescntiy used

I r i i i i i l o i rivering.

(f) Figure 5-8 Lacing and other means of producing a built-up member with access to interior.
fabrication shop practice gives preference to use of batten or perforated cover plates which are
Figure 5-7 Cross sections of several built-up sections. General cross-section configuration is limited welded. (a) Single lacing. (b) Double lacing. (c) Batten plates. (d) Perforated cover plates.
only by designer's need and ingenuity and may include W and S shapes with'lacing and/or plates.
Where additional area is needed, more plates may be added to any of the above sections. ( a ) Four ;,
angles with lacing. ( 6 ) Four angles with both plates and lacing. (c) Two angles and one or two $,.
, parts, making the connection sufficiently strong that no relative movement can
( d ) Two channels with lacing. ( e ) Two channels with flanges reversed from (d) and both plate and take place across the hole. Tension tests using plastic and photoelastic tech-
lacing. (A Four plates welded to form box section.
'.i niques indicate stress concentrations at the edges of the bolt hole which
4. $
have a maximum value on the order of two to three times the average s

not occur. It is also necessary that the member(s) be constructed so that painting f, = 5,it is reasonable to assume a uniform stress distribution across
of the complete member can be affected. This requirement generally precludes
use of fully enclosed box or other built-up sections with .enclosed cavities.
Instead, the built-up sections are open on one or more sides, with continuity indeterminate, but should the working load be inaccurately estimated so h a t
being obtained on those sides by use of lacing bars or by use of perforated cover redistributed until the load is
plates. Either of these configurations allows maintenance of the interior as well
as the exterior of the member, as shown in Fig. 5.8. ber of stress reversals or to
range to the member. The
id a fatigue failure if the
er of stress cycles and/or the stress range is large.
is usual to assume that fatigue can be neglected in normal building
load cycles/day, the total
es over a period of 20 years is only
N,,,,,, = 20(365)(20) = 146 000
 232 STRUC-IVRAL STEEL DESIGN DESIGN OF TENSIOK
?

'

+
L:
0.4 195 kipsift

0.066 kiplft

P= 1
14 .03~
me possible path as just cited, which
vior along irregular failure lines. The

Vol. 89, 1922, pp. 847-848) and is almost universally used. The meth
Figure E5-1b
a<g
it. p" .y;,
. , .. .
!, .". ,),<..
' t,i
,,
Notes:
,;;*,.,i ';y;;,~'~~,,
, ,?,$
>\ ,
) "
, . ',
,

, , .v
. . one hole width for each hole encountered.
:. : . '1 .";Tie rod size could be reduced in the lower-half group of purlins for a rod For each change in direction from one hole to the next hole, add ba
...,( force of 6 X 1.03 = 6.18 kips.
,

; 2. A rafter, substantial girt, or other member spanning the 25-ft bays will be
, '
required to carry the concentrated rnidspan load of 12.18 kips without
excessive outward deflection. The roofing will undoubtedly reduce this where s= pitch, or longitudinal distance between adjacent holes
effect considerably, but the designer will have to decide how much. g= gage distance between adjacent holes across the-width
3. Computations neglect the benefits of the roof in contact with the top
flange of the purlins, and may reduce the sag rod force 50 percent or
more. \

5-6 NET SECTIONS "hole" on the diagonal, thus canceling that hole reduction on the net
The effective hole diameter (except for slotted holes) is taken as
When the cross section contains a row of holes, the critical section will occur
through one of the holes, as shown in Fig. 5-9. When there is more than one row AISC specifications
of holes, the designer must determine a failure section that will yield the AASHTO and AREA specifications
minimumiarea. The path across the section producing the minimum area is t
critic&l:ine;tsection. The hole is always larger than the fastener by at least $ in (or 1.5
-4 li$@ig. 5-9b, which is a portion of a repeating bolt or rivet pattern @mi&
,t,

be btained in a large plate ass

"" ~ .d t # & ~ r e )the
; :..
, I , , critical section is
for possible metal damage around the hol
.- .\: ;. ... .......... .
iameter as D + in or'D + 3 rnm.

\[-I/ r 0 0
Example 5-2 What is the critical net section of the hole pattern shown
Fig. E5-2? Take holes for i-in-diameter bolts. Note that this pattern is rno
academic than practical, but is used to illustrate the method of using s2/4
Use the AISC specifications for the hole diameter.

Total width w = 2(1.5) + 3 + 2(2)

(a) D =:+$= I.OOin
fy,
Mgure 5-9 Critical net section for tension members with holes. (a) One or more rows of holes w t width along path ABDF':
the net section is obvious. (b) Hole pattern where net section must be found by trial. w = 10 - 2(1.W) = 8.00 in
 DESIGN OF TEXSION Sf33

Net width along path ABCDFG:

rv, = 13.25 - 4(0.875) + 0.333 + 0.333 = 10.42 in
Net width along path ABDEFG:
>vn = 13.25 - 4(0.875) + 0.222 + 0.400 = 10.37 in
Check the 85 percent requirement:
~ v ,= 13.25(0.85) = 1 1.26 in

Based on the critical net width of 10.37 in (note that bo

connected, so a shear lag reduction is not required) and c = 0.7
A, = 10.37(0.75) = 7.78 in2
Pallow= A,F, = 13.25(0.75)(0.65)

5-10 Net section of angle used as tension member with staggered holes in both legs.
P , = A,F, = 7.78(0.5Fu)
whichever is smaller.
J
*
2d"

5-7 DESIGN OF AISC TENSION MEiMBERS

Example 5-4 What is the critical section of the 8 x 6
s computations to determine the
X angle with
tional computations to determine the critical net section where
for mechanical fasteners. Member selection proceeds based o n
ncluding the 85 percent maxim
are in the member, and any reduction in net area from

be met in all the specificatio

and AREA) have minimum
mum connection requirement is

Example 5-5 Select the lightest single angle section for the vertical
of the side shed truss of the industrial building of Example
desirable to use the same size angle section for all the verticals,
F4W-e E5-4
section for all the diagonals, and similarly, top and bottom chords d s s i p e
SOLUTION . ' using constant section sizes. Only the vertical web members will be d e s i g e
D = 0.75 + 0.125 = 0.875 in in this example. Use F, = 250 MPa and the AISC specifications.
Net width along path ABDFG:
,&,w = 8 + 6 - 0.75 - 3(0.875) = 10.625 in
Net width til&gpath ABCDEFG:

= 13.25 -' 4.375 + 0.667 + 0.667 + 0.222 + 0.400

= 10.83 in
 1

9 3 STRUCTURAL STEEL DESIGN

S o ~ u n o NFrom inspection of the computer output (Pa* which

shown in Example 2-6), the following values are obtained from the tw

!' conditions used:

Right side Left side

Member LC-1 @pq LC-2 (kpq Member LC-1 @W 0

4 0.0 0.0 74 124.73 157.57
8 31.18 48.62 78 93.56 118.18
12 62.36 97.26 82 62.37 78.80
16' 93.54 145.88 86 31.18 39.40 I n the table above,
20 124.72 194.50 90 0.0 0.0
d Ahole, = 2Dt = 2(25.0)(0.0063) = 0.315 x 10-3 m2
0.9725
Since the signs are (+), all the vertical members have only tension Iorces Ior An,, = A, - Aholes2 -0.90
= 1.081 x m'
either load condition. Note that the computer program automatically
I output with wind by the factor 0.75 so that all loads are on the same design
I basis. Member 20 has the largest axial tension force with wind from the left; 'Onsidered Here the radius of gyrauon controls the selectlon of he section
1 wind from the right would produce this design value in member 74. Member than the area requirements. Tentatively use: L 152 x 89 6.3 (as ae
1
20 is the longest vertical of the side shed members and has L = 4.5 m. If One lightest). We must select a section that can be used in hejoint as shown-

b*
3i
of the other members had been longer, that member length
the ~ / , . - i ~this case both the controlling length and the maximum
force are in the same member. Design P = 194.5 kN.
L/,. = 240 (per AISC, Set. 1-8.4, and assuming that this is a main member
with such a large axial force). The minimum radius of gyration is
the angle. Use Table V-13 and obtain:
Standard gage distances: g, = 57.2 mm g, = 63.5
AISC specifications: center to center of hole = 2,670 ( 3 ~
specifications: center of hole to edge of angle leg ~
-
Check if it is possible to get two 22-mm bolts side by side in the long leg of

~ b l ~
4.5(1000) = 18.75 mm 1.16.4 SSDD) = 28 rnm (or 1; in)
'mm 240
=

I
A preliminaly side computation indicates that the bolt Pattern !&Own
~ 5 - 5can probably be used, as L / r rather than stress is
m i s pattern may be able to use standard gage distances and
economical. For two bolts at the section and using 22-mm A-325
D = 22 + 3.0 = 25.0 mm
to
> 2.670
g2 = 63.5 0 . ~ .
Actual hole diameter = 22 + 1.5 = 23.5 mm
Edge distance = 152 - 57.2 - 63.j = 3 1.3 > 28
Use an L152 X 89 x 6.3 section.
- O.K.

't
Ii:
L
me effectiveangle area using AISC criteria for shear 1% (see Set. 5-3.2
A, = 0.90An
~~~~don F, = 0.6FY,the gross angle section must be at least
i1 A, = ----
194.5 = 1.2967 x lo-'
0.6(250)
m2
designed for both tension and compression (and p o ~ ~ l b lfor
y fatigue). F~~~
the 'Omputer output, members 39 and 47 will be crltica] for axial forms-
faa Based on F, = 0 . 5 ~ "(use Table 5-1 for F,), the effective net area must be
I"
194 5 LC-2 GN)
i
I A, =
200
-= 0.9725 x 10-3 m2
Member
39
LC-1 OrN)
- 63.65
L, m

t + 20.39 6.76
m i s effective net area A, must be obtained from a gross section of least 47 +70.18 - 37.60 8 2 (longest of verticals)

A, = 0.9725 = 1.2712 x m2 < 1.2967

0.85(0.90) P,, = 70.18 - (-37.60) = 107 78 L N
 ' 2& STRUCTURAL STEEL DESIGN 1 DESIGN OF TEU'SION LIMBERS

Member 47 is critical for tension design. Example 5-7 A portion of a highway bridge truss is shown in Fig. E5-7.
8.2(1000) required to select the lightest W12 section for member 9 using A-36
Minimum r = = 34.17 mm and the AASHTO specifications. Assume 2 X lo6 load cycles for the scrwce
240
life of the structure.
* ' +
Use 25-mm bolts: D = 25 3.0 = 28.0 mm. Use two L sections with a
12-mm gusset plate as in Fig. E5-6. Assume two lines of holes, since P is I,: ,
only 70.18 kN, which gives A, = 0.90An.

SOLUTION From the computer output, we obtaln the following (includes

impact in live loads):
Dead load = + 80.3 kips
Maximum live load = + 60.2 kips
For A, = 0.75An, the gross angle area is at least Minimum live load = - 22.8 kips

From which
Use long legs of unequal leg angles back to back. (Why?). Again set up a P,, = 80.3 + 60.2 = 140.5 kips
table of double angles using Tables V-10 and V-11 of SSDD [A, = smaller P,,, = 80.3 - 22.8 = 57.5 kips
of (A, or 0.85Ag) X 0.901:
The force range (analogous to stress) is
P,, = 140.5 - 57.5 = 83 klps
1, mm Al m2 Anet, m2 Ac, m2 Section r, mm ,A
, (X mZ
We will assume four holes in the flange at any net section, as shown in the
6.3 ' 0.353 2.307 2.03 2L127 X 8 9 X 6.3 37.5 2.660 insert of Fig. E5-7. Use i-in-diameter bolts, so that the effective hole
7.9 0.442 3.468 2.991 2L127X 127 X7.9 39.9 3.910
diameter D = + = 1.0 in.
Aholcs = 4(l)tj
There is nothing lighter, than these two double angles. Note that minimum
%v High-strength bolts will be used to connect the member to the joint,
L / r ' controls the design. Now checking stress range and using A,, we have
producing stress range conditions for "base metal at friction fasteners." The
f = -psr
= 107.78
- = 54.0 MPa
data given in Sec. 1-12 are based on an allowable stress range when a
" A, 2.03 catastrophic failure does not immediately occur when the member fails. In
This value of 63.6 is much smaller than any value of F,, in Table 1-4 up to this case it is not likely the truss would collapse if member 9 fails. With this
2000 X lo3 stress cycles for "base metal" at mechanically fastened joints. consideration we obtain the allowable stress range:
p:' Use two L127 X 89 X 6.3 sections. /// F,, = 16 ksi (first column and checking footnote)
We will use lateral and sway bracing across the top and spanning
, 5-8 DESIGN OF BRIDGE TENSION MEMBERS between the two trusses to satisfy stability, but this wdl not reduce the
$
unbraced length of the several web members with respect to the Y axis.
The design of bridge tension members is similar to that using the AISC (AASHTO requires the effective depth of this bracing to be at least 5 ft or
specifications except that fatigue will have to be considered, as outlined in Secs. 1.8 m.) Joint fabrication requires orienting the X axis in the plane of the
5-4 and 1-9. This will be illustrated by the following example. truss for rolled sections. Note that commonly the transverse floor beams are
 242 STRUCTURAL STEEL DESIGN

somewhat below the point of intersection of the tmss web me

bottom chord, but this should not affect the unbraced length L
The AASHTO limitation for L/r = 200 for main tension Y.
rers and
no stress reversals (Sec. 1-7.5), and for a P = 140.5 kips, this would seem
deem ttn

,*
;" { t qualify. The net area, assuming that Fa = 0.55F,, is

r ,
Anet = -
Fa
-- 140e5
=
140e5
20
7.025
7.025 in2
in2

25(1.414)(12)
rmn
rmn = = 2.12 in (for r,,)
200
We will arbitrarily use the AISC 85 percent reqmrement, so A, is at least

A >-- 025 - 8.26 in2 I * $4

- 0.85
Figure 5-11 Cable geometry ior derelopmg dei~pnequauobs. N@tc
:' Bat m%rs conf~gurauonthe T at
By tnal set up the following table, where Ahole.= 4(1.0)$; AWqd= the o r i p is larger than the T at the other end iif $e cable.
- - reqa
- Aholcs
Y d , a
Section +, Ln ln2 ln2 Arm, m2 In general, the cable sag y at any point IS
- Aholcs, Arcqd,
4h x
y = - ~ ( u- L ) + x tan8
W12x53 0.575 2.30 13.30 15.60 > 8.26 2.48 O.K.
W12~50 0.640 2.56 12.14 14.70
L
Differentiat~ng,we obtain
The W12 x 53 sectlon is the lightest W12 that is satisfactory for both +
-
.
=8 h x- 4-Lh + t a n ~
~2
(6)
area and L / r requirements. This section 1s also selected so that the connec-
tion can be more easily fabricated when the vert~calmember (No. 7) is Since
designed in Chap. 6. H
Check the stress range for the W12 x 53 section: dr = [ I + ($)2]1'2 and T.= -
cos 0 -
A , = - -PSr
Psr
= - =- 83.0
A, 13.30
- 6.24 ksi << 16 ksi O.K. we obtain a general equation for the tension force in the cable at any point as
Although fatigue is not a controlling parameter for this member, all the truss
the truss
members should be checked similarly. Use a W12 x 53 section.
///
5-9 'CABLE DESIGN
* '
H moment summation about a convenient location with respect to the parabolic Noting that all the terms under the square root are relatively imi,onif icant except
"x-bie geometry as shown in Fig. 5-1 1 gives the first, third, and fourth, we may simplify to obtain
b t :
H z -wL2
43 8h
xhere = mdspan sag, as shown in Fig. 5-1 1
When the ends of the cable are at greatly dlffenng elevations, Eq. ( 5 - 7 ) rather
I
w = uniform cable loading/unit of length;
- there will always
..
.
be
-- -J- - than Eq. (5-8) should be used, because the tension at the upper end of h e cable
load caused by the cable weight will be considerably different (it is "caqing" the weight of the cable + any
L = span length (the cable length is always somewhat longer)
additional cable loading). For horizontal cables T has the same value at b o b
 STRUCTURAL STEEL DESIGN DESIGN OF TMSION bEEJ

1 .,
ends and is directly computed as

L 1/2
T=~ [ + l16($)]

, . The
. , , .. cable
..,..... ,.. ... length is approximately given by
S I T ~ I2nd
I > I u * r r L ~ b l pruduic
c
,riiii.ning 6ileir r o ,~s[e111
Id)

Cables have been used to support large-span roof structures as well as

bridges and guys for towers. In buildings the cable roof is constructed by
stringing cables across the open space at sufficiently close spacing and preten-
sion force T to produce the desired sag based on Eq. (5-9). The sag in buildings
is on the order of to $ (bridges may be as large as to i).
In round
structures the cables may be attached to a large compression ring at the building
perimeter and terminate in the center in a tension ring. This configuration is the
most desirable, since the cable tension must be carried by some kind of
, anchorage. The compression ring is the most desirable, since large compressive
stresses can be used if the nng is made of steel. The use of prestressed concrete
roofing "plank" directly on the cables produces the necessary roof and at the
same time tends to reduce the vibrations, since the concrete planking is rather
heavy and develops a large system mass.
Where the use of concrete planking is not sufficient to restrict damping,
other means must be employed. Damping tends to modify the natural amplitude C.ible JrTJnpelllcnI m.1) 'cc r ~ d i . i l i n g f r o m
of a vibrating system and if reliably introduced, eliminates resonance amplitudes rov.<r or in o n e d ! r < i r ~ o n

of vibration. The natural frequency of a cable system (the same as any other
type of natural frequency computation) is

whe?e rn = w / g = mass of the entire cable (including any attachments, such as

a roof)
n = any integer, such as 1, 2, or 3, used to obtain the fundamental
modes; the value at n = 1 is of primary interest, but we may need
values for n = 2, 3, and perhaps 4
+
T= cable tension; can be written as T AT, so that it is evident that a
Figure 5-12 Several conflguratlons of cables In building construction (u) Cabies spaced across span
change in T produces a new fn for a rectangular buildlng plan ( 6 ) S~nglecable using a central tenalon nng/;md for clrculr b u i g
A convenient way to dampen a cable to control vibration is to attach it to a (c) Double-cable system for increased stiffness and mbration control For round buridmg p h - ( 6 )
Cables used In roof support system If cables rad~are.the towers may be esamnally self-supprtug.
second cable, as illustrated in Fig. 5-12, which has a different natural frequency.
This can be accomplished by use of Eq. (5-1 l), which indicates that changes in T
different f , values. It will then be necessary to install the cables using
ruts to a tension sufficient to satisfy design forces and at T values such
values at any of the fundamental modes does not match, which would
uce resonance (very large vibration amplitudes that would likely cause a
tructural collapse).
 I

STRUCTURAL STEEL DESIGN

The safety factors when using cables have been suggeste

k i n g from 2.5 to 3.0; compare the factored ultimate tension
tables such as Table 5-2. The factored load may be
U = 1.5 dead + 3 live
U = 2.5(D + L)
With SF = 3, the anchorage resistance would be 3 x 154.7 = 464
U=2(D + L + WorE)
U = 2.0 X erection loads
xample 5-8 Design the guy cables at
The horizontal forces for design are shown in Fig. E5-8.
used, but each is to be designed for the force shown.
20-ft centers and will be in compression. The design of the
considered in Chap. 6.

-200 m+- 120 k N

-100 m--L I10 k N

Figure E5-8 Roofing dead load, including weight of cables

(horizontal projection)
S ~ L U T For
~ ~ the
N 100-m level, using Eq. (5-6), we obtain
Live load at 40 psf: 40 x 4
8(110)
-wL2
= llOkN -w ---- - 0.088
8h h lOd
Also, the cable tension is (Eq. (5-7):

[
T = H I + 16 -
, (:r +tan20-8

obtained with slight simplification of terms. By Ulal and using SF = 2.5,

shown, and this tension will be [combine Eqs. (5-6) and (5-8) to obtain th

19 = 45" +tan O = 1.0:

Trial h = w/0.088 T T"

Taking 1/2 the roof load as w,

1 0 (assumed) 155.6 389 0.024

2 0.272 m 154.7 386 Close enough to initial
.M8 S - I U C W R A L STEEL DESIGN DESIGN OF TENSION hD31BEXS 249

The bottom cable will carry the uniform load, producing the 23.3-ft sag e used. Once this is done and the ultimate axial tension load P, is obtained,
of the upper cable plus the total roof load: k 1 1 Q V V

p,, = A,+Fy (+ = 0.68, Table 3-1)

The required cable areas using F = 3.0:

Top cable: P,,= 3(74.8) = 224.4 kips -+ use I f-in-diameter bridge Example 5-10 Given a roof truss member with a length of 11.5 ft in a
strand building frame with a dead load of 18.52 kips and a live load of 22.54 kips
From Table 5-2, w = 4.73 Ib/ft. (snow load). Use the LRFD method, A-36 steel, and i-in-diameter A-325
Bottom cable: P, = 3(224.4) = 673.2 -+ use 2f-in-diameter bridge high-strength bolts. Design the member using the lightest C shape possible.
strand
From Table 5-2, w = 12.8 Ib/ft. SOLUTION We will use a bolt pattern as shown In F-lg. Ej-10 so that the
The struts wlll carry a compression load based on q producing an h AISC shear lag reduction factor will be 0.85 (at leasr three fasteners in the
value in the upper cable of 23.3 ft. This gives an equivalent uniform strut line of stress).
"or diaphragm" load of 0.220/2 = 0.1 10 kip/ft. At a 20-ft spacing, the strut
load is 20(0.110) = 2.2 kips.
Check the natural frequency of the top and bottom cables: First find
the cable length:

= 354.13 ft
With dead and live loads in contact with the top cable and using 350 ft
as the contribution span for these loads, the mass is

With two holes out at the critical section, the net area is
Anet= A, - 2(; + ;)lw
For n = 1, 2, and 3, we obtain f , = 0.96, 1.92, and 2.88 Hz. From Sec. 3-7, obtain ( 9 factors from Table 3-1)
For the bottom cable, only the bottom cable weight will be used, since
only the struts make continuity with the top cable:
Pu = l . l ( l . 1 0 4- 1.55')
= 1.1[1.1(18.52) t 1.5(22.54)] = 59.6 hps
59.6 - 1.88 in'
A >-..-=--.-----
P U

- OFy 0.88(36)
For n = 1, 2, and 3, we obtainf, = 6.78, 13.6, and20.3 Hz.
Slnce the natural frequencies f , of the two cables are considerably - = -----
59.6 -
- 1.39 in?
A >
p u

different, no resonance 1s likely to occur. When one cable is at resonance, " - dF 0.74(58)
the other is at a different frequency, which acts to dampen the resonance
vibrations so that the total vibration amplitude is kept small. Using the largest A,, the gross section using the AISC efficiency factor and
/// the shear lag factor gives the gross area as at least

5-10 DESIGN OF TENSION MEMBERS USING LRFD

The design of tension members using LRFD is relatively straightforward. Again L * x 9.80:
Try a C7 3'

the dead and live loads must be identified so that the appropriate load factor A, = 2.87 in2 ry = 0.625 in > 0.575 O.K.
 250 STRUCTURAL STEEL DESIGN
DESIGN OF m & I O N L

NOWcheck if we can get two bolts in the web, as illustrated in What is the bolt pitch in Fig. P5-3 so that the critical net section is at [east 205
*"5-10. Answer: 62 mm.

g, = 2.50 in and eel is used and the plate

2g, = 2(2.50) = 5.00 in
ih
This leaves a bolt hole spacing of
4,:+ 7.00 - 5.00 = 2.00 < 2.67(7/8) N.G.
Answer: 97.9 kips.
5-8 What is the allowable plate capacity of Fig. P5-3 ~f F, = 250 &(Pa steel is used and the
, ,':,
, rl thickness is 12 mm?
A > , : d Try a C8 x 1 1.50: Answer: 423 kN.
5-9 What is the allowable tensile load for the plate shoan in Fig. P5-9 using F, = 345
A = 3.38 in2 r, = 0.625 in > 0.575 O.K. 20-mm-diameter bolts, and the AISC specifications?
g , = 2.50 in t, = 0.220 in
This leaves a center-to-center bolt spacing of
8-5=3in>3D O.K.
Check the net area.: with two holes out:

Check the shear lag:

A, = 3.38 - 2
(f+ ;) - 0.22 = 2.940 in2

Flgum PS-9
A, = 2.940(0.85) = 2.499 > 1.88 in2 O.K.
Joint efficiency and shear lag:
5-10 What pitch is necessary in Prob. 5-9 so that only three bolt holes are deducted from
A , = 3.38(0.85)(0.85) = 2.44 > 1.88 O.K. to produce the net section?
Use a C8 X 11.50 member. Answer: 32 mm.
5-11 Select the lightest single angle for a tension load of 50 kips. The length is 6 ft and I -
PROBLEMS bolts wiU be used as shown in the pattern on Fig. P5- 11. Use A-36 steel and the AISC s
Answer: L6 X 4 X f .

.For all problems, assume adequate fastener strength so that only the net/gross section requirement$, *
control.
5-1 Design an eyebar to cany a tension load of 40 kips using a I-in-diameter pin. Use A-36 steel
and the AISC specifications. Use t 2 in and w in multiples of in. , .
5-2 Design an eyebar to carry a tension load of 200 kN using a 25-mm-diameter pin. Use F, = 250
MPa and the AISC specifications. Use t 2 15 mm and w in multiples of 3 mm. Figure P5-1 1
5-3 What is the net width of the plate shown in Fig. P5-3 using the given fps units?
Answer: 6.75 in. 210 kN. The length is 1-9 m a d
shown in Fig. P5-11. Use 5 =. E O

5-13 Select the lightest pair of angles back to back to c a m a tensile load of 433 kN. Ux
22---diameter bolts, a 12-mm gusset plate, and F, = 3 4 5 &{Pa steel. The member lenglh is 3.2 EL
Use the bolt pattern of Fig. P5-13 and the AISC specifications.
v, + 75 3" Answer: L102 X 76 X 6.3.
v,
rl 3 0
50 2"
40 40 1,s" Flsurep5-3

5-4 What is the net width of the plate shown in Fig. P5-3 using the given SI units?
Answer: 176.3 mm.
5-5 What is the bolt pitch in Fig. P5-3 so that the critical net section is at least 8 in?
Flgure F5-13
 ' ~ 2STRUCTURAL
: STEEL DESIGN

5-14 Select the lightest pair of angles back to back to cany a tensile load of 92.5 kips.
i-in-diameter bolts, a f-in gusset plate, and F, = 50 ksi steel. The member is 8.375 ft long.
bolt pattern of Fig. P5-13 and the AISC specifications..'
5-15 Select the lightest single angle for a tension load of 68 kips, assuming one Answer: P = 154 kips.
the critical section. The member is 7.5 ft long. Use A-36 steel, the bolt pattern
and both the AISC and AASHTO specifications. Assume no stress reversals fo
Answer: By AASHTO, L7 x 4 x A.

F!
..,. . 4 ,., .,,..,. ..
P

Figure P5-15

5-16 Select the lightest single angle for a tension load of 220 kN assuming one 20-mm-diameter bolt
at the critical section. The member is 4.3 m long. Use F, = 345 MPa steel, the bolt pattern shown in
Fig. P5-15, and both the AISC and AASHTO specifications. Assume no stress reversals for
AASHTO.
5-17 Design the bottom chord members to satisfy tension for the side shed truss of Example 2-5
using WT (structural tee), i-in-diameter bolts, A-36 steel, and the AISC specifications. Assume two
bolts at the critical section is the web of the tee. The tee is continuous across the critical joint.
Answer: WT9 X 27.5.
5-18 Design member 5 of Fig. E5-7 using a W12 section if the computer output (including impact) is
Dead load = 160.5 1 kips
Live load = 77.17 kips (maximum)
Live load = - 5.8 1 kips (minimum)
Use the AASHTO specifications and A-36 steel.
A m e r : W12 x 53.
5-19 Design the guy cable for the 200-111 level of the TV antenna of Example 5-8.
5-20 Design sag rods for the purlins shown in Fig. P5-20. The purlin span is 28 ft and spaced on 8-ft
centers. The sag rod is at the midspan of the purlin. The roof slope is as shown.
Dead load = 25 psf of ioof surface 5.23 Do Prob. 5-21 using the AASHTO specifications. Do not consider fatigue.
4
Live load = 45 psf (horizontal project) 5-24 Do Prob. 5-22 using the AREA specifications. Do not consider fatigue.
Use A-36 steel and the AISC specifications. Answer: 747.4 kN.
Answer: Diameter = 1$ in.
Answer: L3; X 3 X 7/16.
5-26 Redo Example 5-10 if the truss member length is 14.5 f t instead of 11.5 ft.
45 lb/ft2 5-27 Given the bottom chord of a truss using a pair of C2CO X 17.1 1 back to back with a 15-mnr
gusset plate between them. Using two 25-mm-diameter A-325 bolts at the critical section and for a
dead-load bar force of 120 kN, what is the maximum live-load bar force that is d o w e d = k g
FL = 1.67? Use F, = 250 MPa and a panel length of 5.1 rn. Assume at least thrrt fastrnrrs i n , ~ e
line of stress and the AISC specifications as applicable.
Answer: LL = 365 kN.
5-28 Design the bottom chord member (No. 12) of the highway bridge truss oi E w p l e 5-7 (refer to
Fig. E5-7) using a built-up section. Use Examples 6-7 and 8-3 as a guide in selecting the rolled
sections to make up the cross section. Loads: dead = 336.9 hps (tension); live mimimum == 13.8
kips (tension); live minimum = 0.0 kip. Use the AASHTO splcifications, i-in-diameter high-
Figure P5.20 strength bolts, and A-36 steel. Panel length = 25 ft, as shown in Fig. Ej-7.

..&&+ha.% ,, 8 , 4,%W '

 AXIALLY LOADED COLUMNS AND STRUTS

6-1 INTRODUCTION

The vertical compression members in a structure are commonly identified as

columns (sometimes stanchions in fore~gnliterature). Sometimes verti
pression members are called posts. The diagonal compression members
ing the top chord of bridge approaches are end posts. The diagonals of a truss
members used in wind bracing may be called stnlts. Short compression m
at the junction of columns and roof trusses or beams may be called knee
In all cases, however, the member under consideration is carrying a compress1
load.
A structural member carrying a compression load is termed a column if
length is sufficiently great. For lesser lengths the member may be called
compression block. The length which divides these two classifications is such
it affects the maximum compressive stress which can be developed under
load. The length is seldom used alone in describing column behavior. Rather,
a n offshoot of the Euler column fonnula developed in the next section, the r
of column length to radius of gyration ( L l r ) is used.

Material Limiting L/r (approximately)

Steel 60
Aluminum 30
construction. Joints are field-fabricated using high-strength bolts. Wood 10
 S T R U C STEEL
~ DESIGN AXIALLY LOAD
COLUSD~S

reasons, which include the following:

1. The difficulty 'of determining the exact point of demarcation between co

pression blocks and columns.
2. Columns, although appearing straight and homogeneous, may have sm
imperfections and always have residual stresses from mill operations, such
rolling, cooling, and so on. Any small imperfections will result in a
eccentricity about one (or both) of the axes and produce lateral deflectio
,(buckling) due to the bending moment that is produced as the product
;load x eccentricity.
<I I -------------1_____
3. It is often difficult to apply a load through the center of area (i.e., appl

From these several considerations it is evident that if an ideal, isotropic x = A sin hy + B cos Xy

member.
undary conditions of x = 0 at y = 0, we obtain B = 0 an

6-2 THE EULER COLUMN FORMULA x = A sin hy

One of the most popular column formulas ever proposed (and there are a ve
large number) was derived in about 1759 by Leonhard Euler, a Swiss mathema- ection of the equation, so the solution must b
tician. The formula is readily derived (refer to Fig. 6-1) as follows. Writing the sin ky = 0. This is only possible for values of k L as follows:
fundamental bending equation for moment, kL = n, Zn, 3n, . . . , n7;

,. , . . ..,,..,...

we have for the Euler case that the moment is P x , so

Using Eqs. (e) and (c) to obtain P, we obtain
dZx
EI-= -Px
(iy2
' & Ql, b
This equation for the critical column buckling load P is generally called t
I
Euler .equation (and the load, the Euler load; the stress, the Euler stress).
es of Eq. (6-2) by the column area A, noting that the ra
ction r = , Fcr= P,,/A, and that n = 1 gves
(or Fcr), we obtain
7 r 2 ~
Fcr= -
(L/rI2
we obtain from Eq. (6-2),
. . I

I t
1 I
Fc, = -4n2E
-- .
. a ( ~ / r ) ~
which is equivalent to the column containing two sine waves in the length L.
~ G i is
s called the second buckling mode, n = 1 is the first buckling mode (single
sine wave), and from Eq. (6-2) it becomes evident that the minim m crit'
. ,
b ~ c q loadg (or stress) isiobtained in the first buckling mode. &,
lnispection of:,Eq. (6-3) indicates that very large values of Fc;,'kfn b
obtai'ried by using L / r + 0. Implicit in writing the differential, .equatibn Taking the derivative dx/& for the slope, we obtain
bending [Eq. (a)], however, is stress being proportional to strain. Thus the uvv
cos hy - k sin ky
limit of validity is the proportional limit, which is often taken as Fcr-+ Fy & P sin k L
Since dx/dy = 0 at y = 0, we have
5-3 COLUMNS WITH END CONDITIONS k M o 1 - cos k L
= 0
P sin k L
Rotation of the ends of columns in buildings is restrained by the beams that The smallest root of this equatlon glves k L = 2 7 and with k = (P/EZ)'/', we
frame into them. The ends of truss members are similarly restrained. In both have
cases the design may be made on the basis of pinned ends. Note that the Euler
equation derivation was for a perfectly straight, pin-ended column T h e dpnvg-
tion for the critical buckling load for columns with various end restraints can be
done in a manner similar to that for the Euler case. This will be illustrated for or the effective column length KL is L / 2 = 0.5 L and K = 0.5. We have now
the fixed-end column shown in Fig. 6-2a. The differential equation for bending introduced one of the first attempts to adjust the column Iength for end
now becomes conditions. The effective length K L or some equivalent is used in nearly aiI
J 2
design formulas in all bullding specifications.
When the top of the column :s f~xedagalnst translation and the base fixed
against both translation and rotation as in Fig. 6-2b. we can rewrite Eq. ( f ) u
Using the standard method of solution as before and noting that we have simply
added a constant, we obtain

x = A sin ky + B cos ky - -
Mo
P
and in a similar manner obtain
1
where k = (P/EI)'/~, as before. With boundary conditions of x = 0 at bot
y L'6 and L, we odtain the constants A and B :
A,
MO 1 - cos kL . or K = 0.7. Several of the more common cases of column end f i t y are s h o ~ i l
x=-( P sin k L
sin ky + cos I+ - 1 in Fig. 6-3.
a
 b l K U L 1 UKAL STEEL DESIGN .AXIALLY LOADED COLmNS .L\iD S&LTS 261

at accounts for a lessening effect of some of these variables as the

(or K L / r ) decreases. For example, a small eccentricity is not a3 critical
short column as for a long slender one, and so on. With these considria-
, let us investigate the AISC column equation for axially loaded columns.

.1 AISC Axially Loaded Column Design Stresses

develop the AISC design stresses by taking the Euler equation for critical
ss, and with n = 1 we have
T ~ E
F,, =
(KLI~)~
differentiating obtain
Recotnineniled d(F,,) - -277'E
(a>
d(KL/r) (KL/~))
Rotation fixed Translation fixed
ce the maximum value of the Euler stress or any critical buckling stress is
conditions
Translation fixed .to5, let us also define the critical stress in any re,mion where the Edzr
tress is not valid (such as smail K L / r values):

7 KL
CY- "j7)
Rotation free Translation free
F,, = (6)
Flgm 6-3 Theoretical and design values of K for columns with end conditions shown.
uation can be differentiated to obtain
.. .,,. .,...,...
Strictly, according to the derivation of these equations for critical buckling d(Fcr) - -,+) KL P-'
(c> , ,
load, the tangent modulus of elasticity E, should replace the elastic modulus in d(KLlt.1
ake the slope d ( F , , ) / d ( K L / r ) = 0 at K L / r = 0.Also, at some point
tangent modulus concept was introduced in Sec. 3-7. will arbitrarily define a parameter
KL
-=
r Cc
6-4 ALLOWABLE STRESSES IN STEEL COLUMN
lopes of Eqs. ( a ) and (c) will be equal (i.e., the two curves defined by these
The column stress obtained with the Euler equation of equations will have a common tangent). Also, experimental column test
ta indicate that takingp = 2 is adequate. Now equating slopes at K L / r = C,
d for p = 2, we obtain
- 271'~
- 2 r n ( ~ , )=~- 7
( C,)
m which obtain m as

m=-
T*E
(dl
account for eccentricity, residual stresses, and the several other factors that c,*
complicate the theoi'y. It would also be appropriate to use a variable safety rom rearranging Eq. (b), inserting K L / r = C,,p = 2, and using the Euler
 .iAnwLIUn,u bltjEL DESIGN AXIALLY LOADED C0LU;CNS rLW STRUTS 2.63
.d k
i ~ ~ . for
i c Fcr, we obtain compression members may be designed for an allowable stress based on the
following amplification factor using L / r ( K = 1) when L / r exceeds 120:

arid solving, we obtain for Cc,

We obtain the allowable deslgn stress uslng e~therEq. (6-5) or (6-7):
F; = F, x \k
Equations (6-5) to (6-7) are somewhat awkward for routine computations
If value of m, p, and Cc are placed in Eq. (b), we obtain (even with the programmable desktop calculators), and ~t IS convenient to write
a computer program to produce a table of Fa vs K L / r for various values of F,
shown In Tables 11-5, 11-6, and VI-5. VI-6 The AISC manual has morz comp'ere
tables using the several commonly used grades of I., , ~ncludlngthe amplrficairon
factor In Eqs. (6-5) and (6-7) for secondarq members uhen L / r > 120.
the critical buckling stress becomes 0 . 5 q at K L / r = Cc under our
711:!;
acsumptions. In general, we have the buckling stress from Eq. (b):
6-4.2 AASHTO Axially Loaded Colunln Design Stresses
The AASHTO formulas for axlally loaded columns are derived similarly to the
AISC values, but the SF tends to be somewhat more conservative, since the
sv - the allowable stress is obtained for K L / r 5 Cc from Eq. (
j): members of the bridge structure are, in general, in a more hostile environment
than building members. The AASHTO formulas are as follows:
For K L / r I Cc:

The AISC kas used the following variable safety factor since 1963:

For K L / r > Cc: 4'

n * ~
F',= f 6- 10)
for all values of K L / r I Cc. For K L / r > Cc, use a constant value of safety SF(KL/~)'
factor based on using K L / r = Cc in the equation above to give
The SF = 2.12 for the AASHTO specifications and the values for the lens&
factor K and Cc are computed the same as for the AISC spzc~fications.Tab!e 6-1
gives Cc for the several values of F, commonly used for columns.
When K L / r > Cc, the Euler equation with SF = 23/12 is used to obtain *
:he allowable column stress as Table 6-1 Values of Cc according to the AlSC
and AASHTO specifications for several
values of F,; C, = ( 2 n 2 ~F,)'/~
/

With standard values of n 2 ~we

, obtain:
. .. 149 000 fps, ksi SI, MPa c,
fps: Fa = ksi
(KL,'~)~

b 1 F', =
1.03 x lo6
Mpa
(KL/~)~
Eqx~tlons(6-5) and (6-7) are used for main compression members. Secondary
264 STRUCTURAL STEEL DESIGN
643 AREA Axially Loaded Column Des;ign Stress
~h~ AREA allowable column design stress formulas are somewhat similar to the
4 B
Limitation Allowable stress
1U7 < % _ < = #,A,
- F , = O . ~ F , - (1k.2) .7 (6- 12)
f i
I>.
. ': fi
In the AREA specifications,
F, = ksi
k = 0.75 for riveted, bolted, or welded compression
member end connections (and k , not K )
= 0.875 for pinned-end members
#.4 Net versus Gross Column Cross-sectional Area
q e net area (= gross area - loss for holes) is used in tension member design.
In any connection design using mechanical fasteners (rivets or bolts), it is
assumed that the fastener completely fills the hole. This assumption is very
nearly met in riveted work, where the head fabrication enlarges the rivet shank, . .
and very nearly occurs in bolted work since the hole is only about 1.5 rnrn larger
than the nominal bolt diameter. Under axial compression, although there are
stress concentrations at the hole, it can be safely assumed that no loss qf net
area occurs when a mechanical fastener fills the hole. When an intermediate
open hole is in a compression member (as for utilities, erection, etc.), the
designer must exercise judgment as to whether to use the net or the gross area.
The AISC specifications give the allowable stresses as "On the gross 'section of
.
axially loaded compression members . . ." Undoubtedly, .there will be ade-
quate arching to transmit the load around $he hole if there is only a small
an'iount of area lost at any section due to the holes. The most conservative
procedure would be to use the gross section when the hole has a mechanical
fastener that fills, or nearly fills, it and the net section for all other cases.
6-45 Column Design
It is necessary to use an iterative process in the desiw of compression members
using any of the AISC, U S H T O , or A3EA allowable stress equztioas. ni-
usual design problem involves the following steps:
1. Detennlne column loads (unless the ~roblemonlv ~ n v n l v e <r h ~ r k i n o
2 CAC~;R-
- -,
-
3. Make a tentative section selection (sometimes use can be made of bb!rs such
as Table 11-4 or Table VI-4 of SSDD).
4. Compute KL/r for the section selecied (now that r is known) and uit the
appropriate stress equation to compute F, (or use tables such as Table 11-5 of
ssnn which O~VPP F fnr r ~ ~ r n r oVT
l w --+:--\
when Po = P,,,,,.
Two additional factors should be given consideration: c
.*
1. Commonly, only the,W8, W10, W12, and W14 sections and rectangulrrr tube
and round pipe sections are used for columns, since the critical radiu of
gyration is with respect to the Y axis. These sections have the best 5 vrtIues
(and corresponding r , / r , ratios). In building design. practical consider-1-
+' IORS
often necessitate use of a given nominal column size throughout the buildng.
It is usual in building construction to run a single column t h ? o u ~ at I z a r
two, and often three or more floors to avoid fleld column $ f i w . v ~ n c Iabor
savings more than offset the Increased we~ghtof steel.
2. When KL/r > C, the AISC speclflcat~onreq6lres use of Eq. (6-7), the
AASHTO specif~cation requlres Eq. (6-10). and ARE,-\ has a somewhat
similar requirement for Eq. (6-13). In all these equations F, is independent of
F,. Therefore, In column design one should use A-36 steel for all cases where
K L / r exceeds C, or the AREA lirmtation, and even if F, > 36 ksi is being
used for some of the other members. For example. if we are using F, = 60 ksi
and a section is found where KL/r > 97.7 (refer to Table GI), we should try
to specify A-36 steel instead of the more expenslve 60-ksi steel. The sligqtly
larger safety factor in Eq. (6-7) allows for the transition between K L / r = C,
= 126.1 (for A-36 steel) and the lower values of C, for the higher-strength
steels.
Advantage should be taken of any ava~lable design aids in mdcilting a
preliminary column selection. Tables such as Table 11-4 or VI-4 of SSDD c a
often be used to make the final design. Tables 11-5 and 11-6 and the correcpond-
ing SI tables (Tables VI-5 and VI-6) can be used to advantage for any o h s r
266 STRUCTURAL STEEL DESIGN
i
{md including W) shapes to quickly obtain Fa when KL/r is computed-partic- 60 lhlit! I I + 30 iblft' rooi X 10 = I 8 k 1 0 (1

ularly when KL/r < C, since both stress and safety factor now depend on
KL/r.
. ..
6-4.6 Design Examples
The design of simple axially loaded columns and struts will be illustrated in the
following examples.

Example 6-1 Design a column to be used in a one-story discount depart-

':.merit store building. Columns are spaced 20 ft on center both ways. The
roof load is taken as 30 psf dead load and 60 psf snow. This gives a column
load of 20(20)(0.090) = 36 kips. Use A-36 steel (C, = 126.1).

SOLUTIONFrom inspection of Fig. E6- 1, take K = 1. By using Table 11-4 as

a guide, the lightest W section (W8 X 18) for KL = 14 ft can carry 41 kips.
This could be a iolution but is not very economical. We will first check this
W8 section and then compare to using a tube column.
Check the a W8 x 18 section:

-
KL- --=14(12) 136.6 > Cc [use Eq. (6-7)] ,?&
Y' 1.23
LI.Yb -
Fa = 149 OoO = 7.99 ksi Fa = -- 11.5 k s ~(vs. 11.54 in Table 11-5)
1.915
(136.6)~ O.K.
Pa = 1 lS(3.17) = 36.45 > 36 kips
Pa = 5.26(7.99) = 42.0 > 36 kips O.K. Use steel pipe 4 in x 10.79 lb/ft.
Note that pipe and any other sections where r , / r , 4 1 are geneidly <n.e
Use a W8 X 18 section (tentatively); the Table 11-4 value of 41 b p s is due ///
most economical shapes to use when KL, = KA,.
to computer roundoff.
Let us also investigate a pipe section (see Table 1-14). Now that we have
Example 6-2 Design a W section for the conditions shown in Fig. E6-2. Use
some "experien~e" from just checking a W shape, let us investigate a
F = 50 ksi steel and the AISC specifications.
4-in-diameter pipe:

1 KL, = 16 f t
KL, = 8 ft
I
11-4 to obtain an inltial estimate of column r:.:r. Since
From Eq. (6'-5), able 11-4 is dased on Fy = 36 ksi: use a ratio as
36
PCquiv= (750) 30 = 540 kips
 l ji AXIALLY LOADED COLUILCiS .LSD STRLTS %3
I I

This appears to be the lightest rolled section possible for this lozdisg
situation. // /
Example 6-3 Check the pair of angles selected in Ertamplz 5-6 to be usid 23
vertical members in the main roof truss of Example 2-6. Use A-36 stet1 a d
the AISC specifications.

Member LC-1 (kN) LC-2 (kN) L, m

The section selected for tension considsrarlons was

2L127 x 89 x 6.3: cl = 2.66 x rn' r,,, = 37.5 mm

SOLUTION
Take K = 1. For member 39:
& Figure E6-2

Since the table is based on KL/ry, we qote that rJr, --

1.6 in the column For member 47:
sizes approaching P = 540 kips; therefore, look for L = 1.6 X 8 = 13 ft
column length and 540 kips. N.G.
Try W14 X 99:
Revise the section for minimum r:

Compute Note that a 12-mm gusset plate is used between angles at the connection,
HI

Possible angles are:

r, = 1.66ry = 1.66(3.71) = 6.16 in
2L152 x 152 x 7.9: r,,, = 48.0
KL
-=-= 16(12) 3 1.2 controls since it is largest A = 4.71 > 0.5504 (required for tension)
rx 6.16
2 L152 x 102 x 6.3: r,,, = 41.5 mm
A = 3 15 >0.5504
We will check the pair of angles 152 z: 102 x 6.3. slncs they are the !Ijl:est.
C, = 107 (Table 6- 1) Member 47:
-KL
= 8'2(1W0) = 197.6 -+ F, = 26.3 \(Pa (Table VI-5)
r 41.5
Member 39:

F,=--47'9 - 26.9 ksi (= 27.0 in Table 11-6)

1.78 P,, = 38.8(3.15) = 122.2 kN > 63.65 0.K
, .. .. Po = 26.9(29.10) = 782.8 > 750 kips O.K.
P4,=26.3(3.15)=82.8kN>37.6 O.K.
 ,
STRUCTURAL STEEL DESIGN XYIALLY LOADED COLmfiS .<4ns n m s nl

By trial [and programming Eq. (6-9) on a pocket programmable ci!cuIa-

tor]:

Sectlon A ~ , , ~in2
. ry, I D Fa. k s ~ P,,,, =A F*,Iups

W12X53 15.60 2.48 N G , since ry too small

W12x58 17.00 2.5 1 9.35 159.0> 99.3 O.K.

The r, value for the W12 x 58 IS 5.28 > 2.5, so the section is satisfac-
tory. Also, this section is the same size as member 9. so the joint ullI be easy
to fabricate. Use a W12 x 58 section. ///
Example 6-5 Design the struts for the cable-supported roof of Examp!r 5-9.
The strut load is 2.2 kips based on cables at 4 f t on centers and struts as
shown in Fig. E6-5. The maximum strut length IS 2('33.3) = 46.6 ft.

Figure E65

SOLUTION The maximum K L / r for compression rnembsrs is limited to 2W.

+
Maximum column force = 56.75 42.56 = 99.3 kips (compression) According to the AISC specifications, this would require a radius of a ~ a -
1
Minimum column force = 56.75 - 16.11 = 40.6 kips (compression) tion r of at least
P,, = P,,, - P,,, = 99.3 - ( + 40.6) = 58.7 kips (stress range)
The AASHTO specifications (Sec. 1-7.5) limits KL/r for compression
nembers to KL/r = 120 for main members or members with both dead- Commonly, plpe struts are used with diameters rdnging from 4 to 6 ic. H e i ~
and live-load stresses. The minimum r is we have a rather large unbraced length. qo euarnindtlon of Table 1-14
(SSDD) indicates that we can use.
Extra strong plpe: 8-in diameter: r = 3.88 > 2.80 in
Try a W12 section, since we have used a W12 x 53 in Example 5-7 for, A = 12.8 in'
amember 9, w p h frames into the joint on one end of this member. A gusset$ For KL/r = 46.6(12)/2.88 = 194.2,
plate can covkr both members with a minimum of filler material if all the F, = 3.95 ksi (Table 11-5)
web members have the same nominal depth.
We further note that Eq. (6-9) always determines the column stress for P , = AF, = 12.8(3.95) = 50.6>>>2.3kips
members using A-36 steel since K L / r I 120 and C, = 126.1 > 120. '
Use 8-in-diameter extra-strong pipe. I t m a be advantageous to cse 2
 272 STRUCTURAL STEEL DESIGN
4 * ,, ..

smaller section in the outer one-fourth of the span, where the value
L = 2h is less than 46.6 ft.
/ onsidered ,as sudh. In any case, where compre:jsion (and tension)
6-5 DESIGN OF BUILT-UP COMPRESSION MEMBERS
,
A built-up section is a more practical design than using a rolled shape in many
situations. This is particularly true when there is a very long unsupported
column length involved such that to meet the L / r requirements would require
one of the heavier rolled shapes. Another factor of primary importance is that
the- radius oPgyration of built-up members can be controlled (see Table 6-2 for
selected examples so that the value of rx can be made more nearly equal to r,, to
produ<cemaximum section efficiency. This efficiency cannot be obtained using
-up sections are very commonly'uskd for bridge trusses and..in co.
r towers. Antennas are essentially built-up colu mns, although not
embers are used in large unsupported spans, a built-up m ember may need to
e considered. Any of the sections considered in Chap. 5 may be used (refer to
igs. 5-4, 5-7, and 5-8) in addition to any other section configuration, which
ay, or could, be made appropriate for the design problem.
It is somewhat more difficult to produce an optimum, or least-weish~,
built-up section since there are several design parameters to satisfy, including:
. Types of members to use, including rolled angles; cha
shapes, as well as plate segments.
Arrangement of the basic members, including any size limitations for overz!I
section dimensions. i

the sdqdard rolled W shapes, where the ratio of r / r y is often 1.5 to 5 or more, . . . . -.. . ._
3. .The resulting computed values of I,, I.,, r,, and ry and K L / r,,,, ^whicil
unless bracing is provided with respect to the Y axis. produce the allowable compression stress.
4. Producing an acceptable section area based on the allowcable stress from s:tp
Table 6-2 Approximate radius of gyration for several built-UDsham
3, which is not known until the area has already been established.

One or more iterations are usually required to f~nallydevelop a satisfactory

section. The number of iterations obviously wlll depend on
1. Engineering design versus material cost.
2. Number of sections to be fabricated; if 100 sections are to be fabricated, the
material savings can be considerable, whereas in the fabrication of only focr
or five sections, the cost of producing a refined design might cost more ti223
the material saved.
3. Reliability of load data a,nd intended use of the built-up members.
In general, a 5 to 15 percent "overdesign" of a built-up member will be
satisfactory.
Built-up sections may be built using rolled shapes as in Fig. 6-4, but more
commonly are constructed using lacing, perforated cover plates, or batten plates,
I I r.. = same as 1 - 1 as shown in Fig. 5-7. A larger radius of gyration and more efficient use of steel is
obtained by separating the load-carrying parts as much as practical (and as done
with the flanges of W shapes). Where this is done, it is necessary to interconnect
the parts so that these several parts act as a load-carrying unit. Several methods
rb7 r , = 0.4211
t-7
---

----]l---I I.oi2,L-s
- U / r i 7 7 y

. 6-4 Built-up shapes using cornbinatlons of rolli.,! sh-psi. Sc.iti,,n

Figure ;;,:,n!ctr) !iri:lr:d :?j
.~:1:>
.
"" "A:-.. ,--A ".-A -1 "" ,-, ---I.. -..A -1 .... ,I., -r~.,-
.L 5- "-2 ?
' .L--. /.\ "-
 .I
of using lacing, single and double batten, and perforated plates (called, collec- If we assume equal end moments, as shown in Fig. 6-5c, and use the
tively, cover plates) are usually used. Where the steel is located inside a building, differential equation for bending as used to develop the Euler column equation,
the cover plates may be solid and their use could reduce fabrication costs. In and allow for boundary conditions, we obtain
exterior environments, where corrosion is always a problem, it is necessary to
have access to the interior of the section for maintenance and inspection;
otherwise, the interior must be completely sealed. The "open" cover plates and
lacing allow access to the interior of the section for cleaning and painting The derivative at y = 0 is
without the careful fabrication required to completely seal the interior. Pres- d.~ kL
ently, it appears that the economics of fabrication favors perforated cover plates - = ke tan -
4 2
to lacing, since automatic gas-cutting methods allow rapid cutting of the plate
openings in a length of cover plate. Now referring to Fig. 6-5h, we obtain for the $hear in the lacing,
The design of lacing and batten plates, in particular, requires attention to
several details: Lack of proper attention to lacing design was believed to have
caused the first' Quebec Bridge in Canada to fail in 1907. It is standard practice
to allocate a portion of the axial load as the shear developed in the lacing or where k = P / E I as in the Euler equation. AISC simply takes
batten plate when the compression member "buckles," as shown in Fig. 6-5. kL
ke tan - = 0.02
2
The AASHTO and AREA specifications make the assumption that the end
eccentricity e shown in Fig. 6-5c is equal and opposite on the two ends of the
column (shown equal and with the same sign in Fig. 6-5c). With some a d d i c o d
simplification of the preceding equation for V, we obtain

where F,= steel y~eldstress, k s ~or lCf Pa

L / r = value for entlre member wlrh rebpect to an axls perpendicular to
the plane of lacing or perforated co\er plate as follo\~s:

AASHTO ARW
fps SI fps SI

AASHTO and A ~ E Aspecifications also require that V be increased for any

additional shear on the member, such as section weight or other tr, P-sverse
loadings. Wind on bridge trusses would also contribute an increase in V ilndzr
this interpretation. The value of V obtained for either the . U S H T O or XZSC
computation may be either a tension or a compression value, and the l x i n g or
(u) (b) (c) batten plaie should be so designed.
Figure 6-5 Shear development for a laced (or battened) compression member. (a) Laced column. ( b ) The spacing of lacing and batten plates must be such that the L / r of the
Identification of laced column shear V. For lacing on both through faces, divide V equally on both main elements between fasteners is not greater than K L / r of the entire member;
lacing bars; for four-side lacing obtain 90" as above. (c) Basic concept of shear in lacing of built-up otherwise, local buckling might develop, particularly where angles are used, m d
section. L / r , between fastener points may be critical. Ths XXSH-TO and =IIRE,I
 End

AISC L,, I < (', \/F'J i i p s i

C': = .:I- [?j. C', = 5.;: Sl

i.. 2 if,50 .oillnrs~~t~~ri

;? L i 1 :
- -

.I I ,<> ZL,L
:," 2 ------
- 110.:ai.-, IML,,I(L: iI

Figure 6-6 Design of perforated cover plates

where Iq,,=moment of inertia of the jth pdrt aith respect to the pciriicl r.ui
I and through the centrold of the j t h part
A,= cross-sectional area of the jth parr
d,(,,= perpendicular d~stancefrom the centroid of the jth area to rile i
axis
5. Compute the radlus of gyration wlth respect to both axes.
1. General outside dimensions and load to be carried.
2. Estimate the tentative compression area based on assuming an F, between 15
and 20 ksi or 100 to 140 MPa (based on F, = 36 ksi). This allows a modest
- , reduction in the allowable stress from 0.6F, due to the K L / r of the built-up
#.

shape. 6. Compute KL/r, and KL/r, and obtaln the allo~~able

compressive stress
3, Dec~deon lacing, batten, or perforated cover plates (Fig. 5-7) or if sect~onis based on the largest KL/r.
to be someth~nglike that of Fig. 6-4.
4. combute area, I,, I,, r,, and r,. The moment of inertla of the built-up section
7. Check P ,
= AF, > P ,,, and iterate as necessarq.
8. Design lacing, perforated cover plates, batten plates. and/or tie (or SIZ:~)

IS I I, = =o(J) + Z A , ~ ? , )
t plates.

This procedure will be illustrated by the following ex~mples.

 4

AXIALLY LOADED COLLMXS .W STRL.?S z9

Exampie 6-6 Design a laced section for the end post of the highway bridge
truss of Example 6-4, which has seven panels at 25 ft each (see Fig. E6-6a).
The unsupported length 4, = Lx is 24.04 ft. The computer output (with an
impact factor of 0.17 included) is as follows:

Member LC-I I, kN Dead load, kips

Compute the radius of gyration about the X and Y axes:

About the X axis: locate a new X axls:

+
Total design load = 240.77 94.13
= 334.9 kips
Assuming that F, -- 14 ksi the area required in the section will be approxi-
mately I,, = 2I,, + 211,d2 + i l p d 2
= +
2(162) 2(8.82)(1.66)' + 6.375(4.59)'
= 324 + 48.6 + 134.4
Let us try two channels with a solid cover plate aqd lacing as shown in Fig.
E6-6b. This configuration, with solid cover plate up, will provide some = 507.0 in4
protection to the interior of the built-up section, and lacing will allow access
for painting and cleaning. The spacing and configuration will be such that a = 4.59 in controls (after computing r, below)
reasonably easy framing of the W12 web sections can be made, as shown,
using a pair of gusset plates. We note that filler plates will be required, since About the Y axis: Iyy = 2 1, + 211,d2 + I,
the W12's have depths greater than 12 in.
C12 X 30 data:

Both t/ and t, > 0.23 in (Sec. 1-7.7)

334'9 -
=-14.9 -
22.5 < 24.02 in2 furnished O.K.
280 STRUCTURAL STEEL DESIGN
.%XIALLY LOADED COLL-NS LSD STXbTS 3l

Current practice is to use perforated cover plates welded to the rolled

sections rather than lacing. We will design lacing for this example and use Try t -& in (0.313 in also rmnirnum AASHTO):
welding to attach it to the rolled sections. The only bolting will be the field L =
- = 109 O.K.
conn~ecti-onsof the joints. r 0.288(0.3 13)
The distance between flange holes (and the approximate center-to-
Compute the lacing bar force. The bar force component psrpzndicu!~
center weld distance) is shown in Fig. E6-6c.
to the member axis is computed using Eq. (6-14):
L' = 12 - 2(1.75) = 8.5 < 15 in (also AISC)

100
62.8 + 10 6;ig)]
+ --------- = 3.349(2.06)
'?
= 6.9 kips

We will increase this value 20 percent (author's decision) to allow for

member weight, wind, and any other factors; thus Vdalgn is
V, = 1.2(6.9) = 8.3 kips
The axial force is
8.3
pd =
cos 30'
---- - 9.6 kips
36[1 - 0.5(109/126.1)']
Fa = = 10.64 k s ~
2.12
btFa = 9.6 kips q

d/" Figure E6-6~

Use single k i n g bars at an angle of 60' to the member axis, as shown in Let us revise t to 7/ 16 in; Fa = 13.78 ksi:
the figure (refer also to Fig. 6-5a): 9.6
b = = 1.58 in use 1 :-in plate
L=-' cos8 300
5 - 9.81 in 0.44(13.78)

L' = 2(8.5)sin 30' = 8.5 in The final lacing bar dlmenslons wlll be taken as ( L = 9.81 i- l.b9 in)
L' - 8.5 - 11.1 << 40
---
ry(of channel) 0.763
Design the end tie plates (AASHTO calls these stay plates). The
Also, 11.1 is less than 0.67 X L / r ( = 41.9). Limit the L / r of the lacing to AASHTO requirements are:
130. The radius of gyration of a flat bar is

[ -1"'
r = ,$if) = 0.2881
Use a minimum of three fasteners (or equivalent weld) each side. Transllited
9.8 1 into design (Fig. E6-6d):
t= = 0.262 in
130(0.288)
t > - -L= - - 981 - 0.245 in t = - l2
= 0.24 use $-in, to match lacing bars
40 40 50
STRUCTURAL STEEL DESIGN AXIALLY LOADED COLUMXS h
1\3 STRUTS

Weld
\

Flgure E6-6d
Example 6-7 Design a built-up section using perforated cover plates for use
as a..column in a water tower (see Fig. E6-7a). The unsupported column
length is 24.7 m and the axial load for design is 1350 kN. Use F, = 250 MPa
and the AISC specifications (noting that a water tower is not a "building"
and may be located where a collapse is more of an expensive nuisance than
a hazard to people, so it may not be necessary to use any specifications),
since adherence to these specifications, although not necessary, will ensure a
safe design. There is usually some wind bracing in water towers, but we will
assume that the bracing is not sufficient to develop restraint against column
buckling.
For this L / r value of 200, the allo~~nbiestress F, = 25.7 !vfPa. Tnerefoie
we must use an r greater than this or the steel area WILL be excessive. For ar
L / r = 100, the value of F, = 90 MPa and the area of steel will bc
approximately 1350/100 = 13.5 x m'.
Try 4 L152 x 152 x 14.3 (somewhat arbitrary choice). The data are:
-
,x = ). = 43.4 mrn
Place the angles in a symmetrical section with spacing as shown in Fig
E6-7b. Compute I, = I, and r, = r, using section data:

- = 24'7(1000) = 1 10.26 and from Tab,? Y 1-5 obtzin F ,= 8 1 \fPr

r . 224
Pal,,, = 81(4.184 x 4) = 1356 > 1350 kN required O.K.
This cross section will be considered adequate and we will proceed to dzsis
the perforated cover plates.
The cover plate design is not so much "design'. as satisfying seIectec
criteria and producing a hole spacing that will f i t the column Ien&
Referring to Figs. E6-7c and 6-6, we obtain:
Figure E67a

r =.75 mrn (arb~trarqselection of hole rzd~us)

SOLUTION
The minimum K L / r will be taken as 200 for main members. w = 2r = 150.mm a
Therefore,
Use L, 2 1.25Lo + 0.45 m. Le,ngth of cover plate = 21.7 - 2(0.45) = 23.
m. This choice 1s made so th5t bqstR' a hole length of L, = 3CYI rnm
 STRUCTURAL STEEL DESIGN AXIALLY LOADED COLUMNS AUD STRUTS %

composite building construction, but more commonly the column is ferminated

on a footing, pedestal, or pilaster. A pedestal is used to support the metal column
above the ground to prevent corrosion when the footing is below ground level. A
pilaster is an enlarge'd sect1)n of the basement wall used to transrmt the column
load through the wall zone:to the footing. Sometimes, but not commonly, the
column terminates6direc;tly o h the footing.
d
A base plate 1s pzcessary when a steel column termmates on 'a 4. type of
masonry to spread'tbe high intenslty of stress in the steel to a vaJu.s"Vlat can be
safely carried by th? mqonry matenal. Masonry is, here deflned as concrete,
concrete block, briGE<,,'a~:,tile; concrete IS most often used and will be the only
\
material considered>&$ '
The base plate arjd the mating end of the column may be planed to affect
load transfer by direct beanng. The base plate is seated on the foundation using
a cement grout (thick sand-cement rmxture, often with an expansive a,vent to
produce an intimate contact, since cement paste tends to shrink on dryins)-
Grout can take out up to about 1 in (25 mm) of footing-to-column mismatch as
long as it is a "fill-in" discrepancy.
Angles may be used to bolt or weld the base plate to the column. However,
Figure S 7 c present practice tends to the use of a base plate that is shopwelde. to the
"r
column. Slightly oversize anchor bolt holes are drilled in the base p18 e to fit
over anchor bolts placed in the foundation/footlng element during field con-
Lb = 200 mm the length of 23.8 m gives
struction. The overs~zedholes allow some misall~amentof thz pnchor bolts
48 holes = 48(0.30) = 14.4 m
without redrilllng the base plate holes or taking out and resetting the anchors.
47 spaces = 47(0.20)= 9.4 m The general method of fastening the column to the foundation IS illustrated in
)I

Total =23.8 m Fig. 6-7.

Find the cover plate thickness (refer to Fig. 6-6):
The requ~redbase plate dimensions are based on the allouable unit contact
pressure of the footing. The base plate thickness IS based on the base plitz
contact pressure, producing bending on the cntical section uith dimensions as
shown in Fig. 6-8c. When the column base resists a moment, the plate dimsn-
sions must be adjusted so that

Use 6 rnrn ( 1 / 4 in). Use welding and locate the plate as shown. Welding
may be intermittent. If the welding is adequate to allow the cover plate and
angles to act as a unit, AISC allows a contribution of the cover plate based
on
( L - 2r)tp = [500 - 2(75)]0.006 = 2.1 x m2
of area to be used to increase the axial capacity of the section. ///
6-6 COLUMN BASE PLATES
Steel columns are placed on some type of supporting member to interface the
column and support. The supporting member may be a concrete column in
where S is the sectlon modulus of the rectangular base plate kith respect to the
moment axis. The plate th~cknessfor this case is also shown in Fig. 6-Sc.
From Fig. 6-8a the area of the column base plate is
A number of combinations of side dimensions B and C can be obtained, bzt
--
that combination should be used which produces r n n . Fabrication practice
favors 3 and C in integers.
The thickness of the base plate is obtalned by considering bznding on a
critical section the distance m or n from the corresponding free edge (Fig. 6 - 3 ~ ) .
%.t . AXUCTURALS n E L DESIGN
AXIALLY LOADED COLU;L[NS AND SIRUTS 287

Allowdble AISC stre.s<s

Conract marerill F-P
ips: kipsiin2 S t . >L?:l

Sandstone & l~rne.;tone 04 7.8

Brick In cement InJtriu ii.25 1.75
Full dred of conire[<
ioiindarion 0.357;
L o s t h ~ nitill Jrru on
iound~rion 0.j5fc' 5 0.71;

dr A

,Lf = lq.r
n oi rn
d.v

With rnornenr
(ci

S!
Figure 6-8 General base plate dunensions and other desigo criteria. ( a ) Base plate dimensions. (b)
Allowable stresses F,. (c) Base plate moment to compute base plate thickness.

For a uniform pressure and a strip tn or n x 1 unit wide X r thick, we have

?
m n
M = q(m)Z or !\I = q ( n ) 2-

Using the largest value of M (and noting that if m = n they are cqual), we have
M
fb -- Fb -- - = *
(for a strip one unit wide)
t2
Solving for t , we obtain
I'tgue 6-7 Column to foundation interfacing using base plates. (a) Use of angles to attach column
I foundation. Method not widely used at present due to extra fabrication (cutting two angles and

&"fully p l p n g column end and base plate). ( b ) Widely used shop-welded column-to-base plate
pod f ~ attachihg
r .columns to foundations [see also field photographs in (c), (d), and (e)]. (c)
using shop-welded base plate. Gap for grouting to final grade can be easily seen. where
=
q = actual contact pressure
6M 'I2
( =[
3 X q X ( r n 2 0 r n 2 ) 'I2
Fb 1
anchor bolts being used to attach column to footing. Diagonal member is a Fb= 0.755 (AISC Sec. 1-5.1.4.3, based on bending on r e c t a r i p i ~ r
bracing element. (d) Column also using shopwelded base plate. Grouting space *
The cable is being used to align frame. (e) Senes of interior columns fastened to section)
Iirectly. Note again the use of shop-welded base plates.

STRUCTURAL STEEL DESIGN
Example 6-8 Design the base plate for a column as shown in Fig. E6-8. Use
F,= 250 MPa, f,' = 20.7 MPa, and the AISC specifications.
--
M A L L Y LOADED C O L L J S S AND STXLTS 7s
Example 6-9 Redesign the column base plate of Example 6-8 to resist a
bending moment of 265 kN . m in addition to the axial load (Fig. E6-9a).

Figure E6-8

SOLUTION
plate. Thus
The pedestal floor line dimensions will be the same as the bas

Let us make m n: -
F, = 0.35f,' = 7.245 MPa (Fig. 6-8b) I SOLUTION
We will design the column-to-base plate weld in ExampIe 9-7.
F, = 7.245 MPa 5 q

After several side computations, let us assume that C = 750 rnm:

From Table V-3 obtain d = 360 mrn; bf = 256 mm

= 656 mm
(0.205 + 2m)(0.342 + 2m) = 0.2305 shifting decimal Figure E6-9b illustrates data so far including q = 7.245 and - ob-
m2 + 0.2735m = 0.04009 tained from using B and C in the preceding equatlon for q.
m = 0.106 m Along line x - x : t

Check: q = 7.248 - 10 22x

B = 205 + 2(106) = 417mm
C = 342 + 2(106) = 554mm
B X C = 0.417 X 0.554 = 0.2310 m2 > 0.2305 O.K.
The actual contact pressure q is

q=-- - 7.229 MPa

231.0

4 . ..
= 50.9 mm say 52 mm
Use a column base plate 417 X 554 X 52 mm thick.
STRUCTURAL STEEL DESIGN AXIALLY LOADED COLLIMFS At?,

Usine three anchor rods for each side, the diameter is

The corresponding thickness is

Use three 32-mm-diameter anchor rods.

70.6 mm
= say 75 mm 6-7 LATERAL BRACING OF COLUrvfNS
.*-om the other direction (line y-y) at point A :
A common and widely used empirical rule for lateral bracing for both
compression flanges of beams and columns is to provide a bracing (z!so z
compression member) element to carry a lateral brace force of
q = 7.248 - 10.22(0.219) = 5.01 MPa (average for a unit width) P, = 0 . 0 2 P
where P is the axial force in the compression member being braced; that is,
P = A& for bending members, where A, = area of compression flange and
f, = average (or maximum) bending compressive stress. For columns, use P =
average axial force in column. This recommendation is given by the Structurzl
Stability Research Council in Guide to Stabilip Design Criteria for hferal
Structures, 3rd ed., edited by Johnston.
L-,e the largest thickness of 75 mm. The final column base plate dimensions A series of tests at Cornell University by Winter (see "Lateral Bracing of
nre 780 x 656 x 75 mm thick. Columns and Beams," Transacfions, A S C E . Vol. 125. 1960) inclicates that very
For anchor bolts as shown in Fig. E6-9c, assume that the bolts will little lateral bracing is required to allow the compression element to develop Pie
carry the full moment even though the axial force will reduce the effect of allowable design stress. This restraint could generally be developed by t5s
the moment considerably. This assumption provides some reserve capacity weight of the floor system onto beams when full-length contact with the
of the anchoring system to resist a considerable lateral force (colu.mn compression flange is made. Because of the variable nature of flooring (metal
shifting laterally). I " Z .,- that the 2
deckfine. concrete-to-beam. wood-to-beam. and so on). i t is suooested
percent criterion be used. Winter also de'nved an analytical expression for the
bracing requirement based on both. reitraint and the relative stiffness of the
column and brace. If a SF of appro xi mat el^ 2.5 is used with this derived
expression, the empirical rule of 2 percefft can be obtained
1 :se F, = 345 for anchor bolts. P

Example 6-10 Determine the minimum-size spandrel (or girt) to brric;. tke
W section with respect to the Y axis for the 730-kip a..tial load of Example
6-2. The distance between columns may be taken as IS ft. Fv for column =
50 ksi.

SOLUTIONThe axial force in the channel section used for the brace is
.- P, = 0.02 P = 0.02(750) = 15 kips
The maximum L / r (AISC) = 200 (compression member). The minimum
radius of gyration, r, = 18(12)/200 = 1.08 in. For an L / r ratio = ZCO, Ci::
allowable axial stress F, = 3.73 ksi (Table 11-5. SSDD), applicable for all F,.
- l5 -j,02in'
Areqd=---- F, 3.73
rru a i n u b i U K A I . blEEL DESIGN AXIALLY LO.ADED COLUhNS .WD STRUTS
Simply search Tables 1-6 and 1-7 for this combination of A and r, and find
Using Example 6-1 as a guide, try a 4 in X 10.79 Ib/ft pipe column:
MClO X 28.5: A = 8.37 in2 A = 3.17 in2
ry = 1.17in r = 1.51 in
If bending or other requirements are also satisfied, this section can be used
for the girj (or spandrel).
! 1 1 * ///
' 1

6-8 COLUMN AND STRUT DESIGN USING LRFD

P" - 54'12 = 3.79 in' > 3.17 N.G.
The use of LRFDrequires the separation of dead and live loads on the column!? Areqd = -
OFc, - 0.65(21.94)
Once thls is done, obtain the ultimate column load in the following form: +t*.
Try a 5 in x 14.62 Ib/ft pipe:
.
'4
+ i
P,,=$(FdD + FLL. . . )
Also,

Here we need Table 3-1, since the value of @ ranges from 0.86 to 0.65 depending
on q , which in turn depends on KL/r of the column as well as 5.We may also Fc, = 36[1 - 0.25(1.002)'] = 26.96 ksi
note that the value of F,, depends on the value of 11 as follows:
+ = 0.65
54'12 = 3.09 in2 < 4.30 O.K.
0.65(26.96)
=
Use a pipe column 5 in X 14.62 lb/ft. ///

where 7 = value given in Table 3-1 and is repeated here: PROBLEMS

6 -1 Determine the allowable load that can be carried by a W14 X 21 1 column using I;, = 50 hi
steel and the AISC specifications if:
(a) KL = 16 ft.
(b) KL = 42 ft.
Example 6-11 Given the columns of Example 6-1 spaced on 20 X 20
(c) KL, = 68 ft and KL, = 4 4 ft.
centers, dead load = 30 psf, live load = 60 psf (snow), column length = 14 Make appropriate ,somments.
ft, a d K = 1 (as in Example 6-1 and Fig. 6-3). Use A-36 steel and the
%
LRF method. (A4 in x 10.79 lb/ft pipe column was selected in Example
6- 1 .) Redesign the column using LRFD as given in Sec. 3-7.
Answer: (6) 602.7 kips. (c) 549.1 kips.
6 -2 Determine the allowable load for a W360 X 3 4 . 7 rolled ssc~lonusing F, = 415 Sip2 md
AISC specifications if:
(a) KL = 5.1 m.
( b ) K L = 13.2 m.
SOLUTION (c) KL, = 20.5 m and K L , = 13.6 m.
P,, = l . l ( l . 1 0 + 1.5S)A Make appropriate comments.
Answer: (c) 2669 kN.
= 1.1 [ l.l(O.030) + 1.5(0.060)] (20 X 20) = 54.12 kips (vs. 36 kips P,) 6 -3 What is the lightest square tube section (see Table 1-15, SSDD) for a column loading of 121
kips and an unsupported length of 12.4 ft? Use the AISC specificarions and '2-36 steel.
Answer: 6 X 6 x 0.375.
6 4 What is the allowable column load using the AISC specifications and iF, = 315 %(Pa for a
rectangular tube section 300 X 200 X 9.52 mrn wall (Table V-16. SSDD) for an u n s u ~ p o n dl-.cgr;l
of 4.8 m?
6 -5 What is the allowable load using the AREA spssificaiions lor s W l 4 X 145 cset L; a
 % STRUCTURAL STEEL DESIGN
ALLY LOADED C O L W S AND STRUTS

compression member in a bolted end connection for a bridge truss? The member is 15.5 ft long and
cises A-36 steel.
Answer: 771.6 kips.
6 -6 What is the allowable load using the AREA specifications for a W310 X 178.6 rolled section
used as a compression member w t h a bolted end connection for a bndge truss? The member is 4.75
n? long and uses F, = 345 MPa steel.
Answer: 3 163 kN.
Q -7 What is the allowable load for the built-up section shown in Fig. P6-7 using the AREA
speclfications and A-36 steel? Assume a bolted end connection and a length of 5.25 m.

i
('310 X 44 6 6 -12 Design the lacing for the allowable load found in Prob. 6-10, allowing a 20 percent i n w e m
V for wind, member weight, etc.
6 -U Design the perforated cover plates for the allowable load and wcuon w d in Prob. 6-1 1.
5.380 X 74 4
6 -14 Redo Example 6-1 if the contributory column arc3 u 30 X 20 instead ol 20 X 20 but dl the
other data are the same.
Figure P67 Answer. 5 In at 14.6 Ib/ft.

6 -8 Whatlis the allowable load for the built-up member shown in Fig. P6-8, using the AASHTO
Z
speclfications, F, = 50 ksi, and an unsupported length of 18.7 ft?
live load
-
6 -15 Design a laced column section (refer to Example 6-5) for a highway bndge truss end post The
length is 7 . 2 ~ 10.18 m. The truss span 1s 50.4 m and the loads are dead load = - 1 I20 kY a d
- 422 kN (wthout impact). Use F, = 250 hfPa and the AriSHTO spsdicatioos.
Answer: Try two C380 X 50.4, 300 x 15 mm cover plate. t
" $"
4-L's 6" X 6" X
J 6 -16 Redesign the truss end post of Example 6-6 using a perforated cover plate for both sides of the

i a "
- "
I - p l ~ t eI 2" x L"
. .,i:;"
li
;: '>.
channels. Note that AASHTO allows use of the net area of the perforated cover plate in computing
the total section area and column capacity.
Answer: 2C12 X 25, A,,, = 23.7, r,, = 4.58 m, ~ncludestwo 12 m X f plates with 3-in holes..

L 6 -17 Design a column base plate for the m a m u m capaclty of a W12 x 170 column with an
unbraced length of 12.0 ft. Assume that K, = 5 = 1.0. Use F, = 50 k s ~ /,,' = 4 !GI, and Pis AISC
W
Figure P6-8 speclficatlons. The column 1s Interfaced to a concrete psdesral
6 -18 Design the column base plate for a W14 X 120 secuon that carnes an m a 1 load of 5 0 ) kips
and a base moment of 200 ft . lups. Use A-36 steel. /; = 3 ksi, and the base plate lntsriacrs 'Je
i -9 Referring to Fig. P6-7, place a second channel C310 X 44.64 on the bottom of the S shape to column directly to the footmg, wh~chhas a total depth of 21 In.

-
zaake it symmetrical. What is the allowable column load using the AASHTO specifications if
L 6.6 m and using A-36 steel?
Answer: 1962 kN.
A w e r : 24: X 22; X 2;.
6 -19 Design a column base plate for the maximum capacity of a W3 10 X 117.6 rolled section wirh
an unbraced length of 4.1 m. Assume that K, = K, = 1 .O. Use Fy = 3-45 ;LIP&f: = 28 blPa, and the
5 -10 What is the allowable column load for the built-up section shown in Fig. P6-10 if the member AISC specifications. The column is interfaced to a concrete pedestal.
is of A-36 steel and the length is 14.5 ft? Use the AASHTO specifications. 6 -20 Redo Example 6-1 1 for the lightest available W8 section.
6 -21 Redo Example 6-1 1 if the loads are as follows: dead load = 35 psf; live load = 75 psf.

-
Answer: 5-in pipe at 14.6 Ib/ft.
6 -22 Redo Example 6-11 using the following data: dead load = 1.75 kPa; hve load = 3.75 LPa;
column contributory area = 6.1 x 7.1 rn, column length = 4.98 m. and K, = K, = 1.0. Use eikrr a
round pipe or a square structural tube for the column and Fy = 250 MPa steel.
6 -23 Design member 6 of Example 6-6 (refer to Fig. E6-60) if the dead load bar force = 283.9 Lips,
the maximum live load, including impact = 109.4 kips, and the minimum Live load = 0.0 kip.

1 >q J
w
,
-
l6"-4
4
Figure P6-10
-
Compare the section to that obtained in Example 6-6. Take P, = +(pd DL + pL LL). wkere
4 1.3, 8, = 1.0, and BL = 1.67 (latest AASHTO). Also, P, = 0.85A Fc,, where F<, = Fa from Eq*
(6-9) or Eq. (6-lo), without using the SF = 2.12.
Answer: Use the same section as in Rob. 6-16.
6 -11 What is the allowable column load for the built-up section shown in Fig. P6-1 I? The length is
5.3 m. Use the AISC specifications and A-36 steel. Neglect the contnbution of the perforated cover
plate.

Figure W-1 Beam-columns and beams making up an industrial frame. Note alternate orientation
of strong axis of beam-columns along sides of frame. A closeup of selected joints is shown in Fig.
9-15.
BEAM-COLUhm DESIGN
7-1 INTRODUCTION
When a structural member 1s !oaded In a manner to produce more t#n oce
stress mode, some adjustments must be made in the allowable stres$s- Vvnnere
the stresses are produced as a combination of bending about the X,and Y axes
as in Sec. 4-8, the final stresses used for deslgn are obtained by superposirioa,
Since F,, may not be equal to F4, (part~cularlyin the case of FV shapes from
flange geometry) the beam deslgn In Chap. 4 was obtalned by iteration.
Accumulation of compressive (or tens~on)stresses at one edge of one of h e
flanges was used In the following form of Eq ( a )
i
T /is equat~onwas obtained by d ~ v ~ d i nEq.
g (a) by F,.
A problem similar to this 1s often involved when the structural member is
1 , ded in a combination of bending and axlal load. These situtlons are always
p oduced in rigid frame building construction- (i.e., the columns carry the
building load axially as well as end moments from the girders that frame into
them). In industnal buildings column brackets may be used to car:;, ::i~.e
runway girders and, ultimately, the crane load. The resu!ting bracket eccsnti-xiry
298 STRUCTURAL STEEL DESIGN
4
produces a bending moment in addition to the axial loads in the column. In this
case the column moment is not at the column ends. Similarly, wind pressuregon
long vertical members may produce bending moments, since a large distahce
between floors (or ground to roof) may disallow the concept of wind being
carried, analogous to one-way slab action. In Examples 2-5 and 2-6 the framing
of the side sheds to the columns of the main .bay produces large column
moments which must be allowed for in their design (to be considered in a later
section).
Other design conditions produce bending in addition to axial forces. For
example, the top chords of roof and bridge trusses are normally "pin-ended"
compression members, but the weight of the member will produce bending as
yell. Purlins placed between panel joints of a roof truss or rafter as a means of
redu* both the purlin member size and roofing span will produce bending in
\he chofd or rafter.
In general, compression members are loaded with axial forces and moments.
'The moment(s) may be at the ends of the member, as in rigid framed buildings,
or developed at an interior point from a bracket, local beam, cable attachment,
or other loading. When the moment effect produces single curvature (see Fig.
7-1) a much more critical design condition is developed than when the
moment(s) produce reverse curvature.
Bending may also be produced in tension members such as the bottom
:herds of bridge trusses where floor beams may frame into them. Bottom chords
of building trusses may be used to attach hoisting devices; other temporary
loads attached to the bottom chords of building trusses will produce bending in
addition to the axial load present.
Figure 7-1 Column loading curvature resulting. The single curvature of ( b ) is often the most critical.
(a) Reversed curvature in building frame. ( b ) Single curvature interior loading. (c) Reversed
cwature interior loading.
In many of these situations, particularly with truss members, the bendrng
stresses are neglected. This may be a reasonable procedure where the bending
stress results from the member weight, or even from purlins if they are relatively
small (and light weight), and the resulting effects are possibly less than about 10
percent of the analyzed stresses. There are undoubtedly small bending stresses in
the truss chord elements due to continuity across one or more panel points (a
technique used to reduce fabrication costs). and additionally there is usually
some overdesign, since it is a common practice to use a constant-sue top (and
bottom) chord member (again to reduce fabrication costs). Exceptions to this
may be obtained when it is safe to use only one bolt or rivet at the ends of the
truss members, so that rotation-is less inhibited-it is. of course, not practical to
fabricate "pinned" joints for the usual b.rfd~,e~o~~roof truss. The effects of actual
joint bending can be minimized by kk'ipin~ 8 th'e &oi@length ~ivlth, respect to the
member 1engths.into the joint .is"short as" practikal&l, coupled w i t h the fact
that truss rnenibers usually have small E l / L rqJi&.,li!$:, Ion, member with smal!
Q
moment of inertia), the moment g;ad?@@,,isih&$ . .. .SF hnd much of the member
:'
length is essentially moment-free. ;."',; "'
The moments that are prodixed-ig t;<sses due to end fixity (and commonly
ignored.in analysis by -using "pinned'" joints) ire collectively termed secondary
effects. It is not necessary to ignore secondary effects. since modem computa-
tional methods (such as the analysis computer program in the Appendix) can
analyze a "rigid" (three-degrees-of-freedom) truss almost as easily as a pinned
(two-degrees-of-freedom) truss. The early computers had only limited cpu/core
storage, and before efficient methods of matrix solution. the "pinned" tNSS
geometry was necessitated, since the difference in matrix size for a truss with ICO
odes or joints was:
Number of equations: 100 X 2 = 200 or n g ~ d :100 x 3 = 3 0
CPU requirements: 20d = 40,000 words 300' = 90,020 words
Since each number (or word) requires approximately 4 bytes the requirements
become:
:b
40,000 x 4
Pinned: = 160.000 = 160K bytes
90,000 x 4 = 360,000 = 360K bytes
.
A capacity of 360K + program requirements would tax the capacity of some of
the larger computers currently available had not efficient methods of matrix
solution been developed so that with clever coding (reduce the bandwidth of the
matrix to a minimum) and/or use of boundary conditions, the three-drser-of-
freedom truss might only require some 30 to 5OK. which is well urithin the
capacity of all but the smaller desktop minicomputers.
 300 STRUCTURAL STEEL DESIGN BEAM-COLLhl?i DESIGN 331

7-2 GENERAL CONSIDERATIONS OF AXLAL LOAD WITH

B~NDING

When tension axial load and bending occur simultaneously, the principal of +

superposition may be safely applied. This is because (see Fig. 7-2a) the tension
2
load tends to reduce the bending effects, so that the value of A is reduced
somewhat, and, of course, the actual stress is also somewhat less so that the
maximum allowable stress conditions computed as
i=ll801n
-f,+ - < fb
I \ r = 21 L)O~rl Figure EY-1
F, Fb -
#
will be safe. This safety is partially obtained by neglecting the effects of P - A on
Example 7-1 Glven the portlon of* b#hway bndge truss w i t h 102:s an2
fb, which could be computed (to be strictly correct) as members as in ~ l g E7-1,
. what is t h i ' ~ & u m tension stress in the bwer
Mc c ,
fb = +-+ P,A- chord? .- *?"

I I
When Pb is at midspan, we have (with the limitation A 2 0, see Fig. 7-2) SOLUTION
(neglect the P-A effect)
pbL3 P,~L'
A=+--- WL? 0.0.10(25~)(12)
48 EI 8 EI (a> jb=-= = 0 72 ksl
8sr 8(5l 9)
and the resulting bending stress is
P 165
J = -AF fb -- - +101782-k s 1
(b)
The actual bending stress can only be obtained by iteration of Eq. (a) until the A f, = 13.98 + 0.72 = 14.70 k s ~ (mau~rnum)
value used on the right side is sufficiently close to the A obtained on the left side.
This iterative solution may be reasonably practical on a computer, but with f, = 13.98 - 0.72 = 13.26 ksl (m~n~mum) ///
hand computations, in only a few design situations (where the number of
members is limited ) is this approach economically justified. Neglecting the P-A When a compressive axla1 load acts together bb~th a bending morncrt. tks
effect in Eq. (b) is an error, but on the conservative side. Inspection of Eq. (a) deflection is ampl~fiedand the compress~vestresb incre~seb S~ncrthe aJ!ow\.ab!e
indicates that the tension stress reduces the deflection and also reduces the compression stresses take Into account poss~ble buckl~ng(l~teraldefIecuons),
compression stress due to bending. With the allowable tension stress F, a member design IS more senslt~veto t h ~ sloaciTnc rp&e than to one producing
constant value and recalling that Fb may depend on the unbraced length of the
tension stresses.
compression flange (and possibly reducing the allowable compressive stress), we Referring to Flg. 7-26, we note that rhe ax131 load (absumed to be applrcd
see that use of
last as belng easler to vlsuallze) Increases the deflect~on.The order of Iorid
-fb+ - - f<, I application does not affect the outcome, honever. as long 3s jrelding is not
Fb FI - produced. The Important concern IS that there IS an Increase In the detlection
provides a satisfactory solution. due to the P-A effect, wlth d correspond~ngIncrease In the bending stresses. T:
value of the deflect~on1s (w~thPb at m~dspan)for t h ~ scase
P ~ L P ~A L '
Pb
A = --- '+ (c)
48 E l 8E l
and the resulting bendlng stresses w ~ t hP, = (-) are

(4
A critical evaluat~onof Eq. (c) indicates that an lterat~vesolution IS required as
Figure 7-2 P-A effects on tension and compression members. for the tenslon mode, but that the deflection "feeds" upon ltszll (dcGtctlcrr
 - e. .
2 - 1 STRUCTURAL STEEL DESIGN B U M - C O L L W DESIGN

ch.ljes more deflection), and for members with an I too small or an L too large a stresses (from loads) by a factor such as 10 percent to a110 A. Another
I-l..,kling failure can occur. fwtor that tends to mitigate the P-h effect IS that it is;u'& eloped from
The P-A effect can also develop in tall buildings, as qualitatively shown in wind or earthquake analysis where the designer can use a stresses that
Fig. 7-3, when wind or earthquake forces or unsymmetrical loading produces are increased by one-third. The P-A effect would only exce& this in rare cases.
lateral displacements of the upper floors with respect to the lower building The analysis computer program in the Appendix h agk ! eeily modified to
eiencnts. A computer analysis can be made to analyze the P-A effect but automatically scan the deflection matns for the appropriate X values, recompute
requires iteration. The steps include: the P matrix using these values, then storlng them for comparison with the new
values from the current cycle until convergence and exit.
!. Vake a conventional computer analysis using the lateral forces.
2. Cjbtain ith-story lateral displacements Xi and for XI+, (next story above).
3. Compute an additional P-matrix moment as 7-3 EFFECTIVE LENGTHS OF COLUiLlNS I N BUILDfNG
FRAMES
Mi = ~ I + l ( X I +-
l Xi)
arid a shear as indicated in Fig. 7-36 using The concept of effective column length L' = K L was introduced in Sec- 6-3 rind
a value of K was obtained for several common cases. I t was observed that when
the column ends were laterally restrained so that the P-A effect (as in Fig. 7-3)
which is applied at the top and bottom of the story with due regard to signs. could not develop, the K factor was K <_ 1.0. With the "flagpole" of Fig. 6-3e or
the pinned base of Fig. 6-3f, the K factor was 2.0 or more.
,
Pi + = axial force in column between i and i + 1 floor levels We shall find that K in multistory buildings, which can translate. m ~ be y
4. Compute new X, and compare to previous values used in step 3. Iterate until considerably more than 2.0, as illustrated in Fig. 7-40. For :he portion of the
satisfactory convergence is obtained, such as, say, 0.02 ft or 0.006 m (ap- elastic frame shown in Fig. 7-46 and using Eq. (6-1):
p;oximately 4
in). x = A sin hy (6-1)
With k = ( P / E I ) ' / ~ and using the effective .+ength L' = K L , we obtain
r?ik P<A effects are also called "secondary effects" and have been largely
';Y
ignored until more recently. Some designers have arbitrarily increased the design x = A sin-
KL
Taking the origin of coordinates at a point of inflection as shown in Fig. 7 4 b ,
we have at the top of the column:
v, = y , .' = ,I,
?'I
s , = A sin-
KL
At the bottom of the column:

Noting that sin (a - P ) = sin a cos ,O - cos a sin ,B,

At the top and base of the column, the slope IS d d t / l $ , and we obtain. from Eqs.
(a) and (6):

Flg-le 7-3 P-A effect for tall bulldings. ( a ) Structure with lateral loads. ( b ) ith story with deflections "TI -
cos -cos - + sin -s!n -
11

gt j exaggerated.
8, =-
( KL K KL A'
304 STRUCTURAL STEEL DESIGN

Substituting,
- we obtain
2
8, = C o x , where Ga = 2 El, / LC
2 EIb / L,
Similarly,

Substituting for x , and x , [using Eqs. ( a ) and (b)], we obtain

8, = (z),
K a
A G sin-T I
7 KL
8, =
(:)iLc (
- - -G "Yl
sm-cos-
KL
7i

K
-
1
cos-s~n-
KL
At the column-beam intersection of rigid frames, the rotztion of the c o l u m
. 77
K

equals the rotation of the beam, so equating the two values of 0,. we obtain
nG, TY~ ..
-
K 6 KLLu"

Similarly equating Eqs. ( h ) and (d), we obtain

K 6
5
- ?I ( t a n 2
KL
- tan
K
= 1
"YI
+ tan --
KL
tan -
K
77

Substituting Eq. ( i ) for tan v , / K L , we obtain

Figure 7-4 Elastic frame for derivation of G, and Gb terms to obtain effective column length KL. (a)
Part of an elastic frame. KL defined as distance between points of inflection. ( b ) Column element
isolated from (a) with terms used in derivation identified. (c) Conjugate beam and moment variation
(assumed).
From the bending moment diagram for the assumed moment distribution along
the beams linearly varying as shown in Fig. 7-4c, the slope of the beam at the
juncture with the column (and using conjugate beam principles) gives '
-
f, ' -A,/
6(Ga + G,) tan s / ~
We may program Eq. (7-2) for increments of Ga and G, and Iterate untd a \ d u e
of F = n / K 1s obtalned to sat~sfythe equality. The value of 7 / K thus obtalned
is used to obtain K as
r
A plot of Ga and Gb vs. K can be made as shown in Fig. 7-5a. is nomogiaph h,
was first developed by Julian and Lawrence in unpublished lecture notes and as
cited by several references, including the Structural Stability Research Couaci:.
Guide to Stability Design Criteria for Meral Srnrctures, 3rd ed., edited by
Johnston and published by John Wiley & Sons, Inc., New York.
In a somewhat similar manner, since the boundary conditions are different.
one may develop equations for Ga, Gb, and K for frames restrained ~ g a i m t
lateral translation as

where the summation ( c ) is taken because load and moment are coming from G, ~b ($1' + +

2
Gb -
tan s/ K
) + -_v-
2 tan K I
T/ = (7-3)
both directions. From the earlier derivation of the Euler equation and summing
, % STRUCTURAL STEEL DESIGN

suggested that one use G = 10 since there is some difficulty in prodlrcin~a

(i0 K Gh Go K Gh true pinned connection.
- CC 5 0 . 0 3 ~ 10- ~ 0 . 0 ~ 2. When the column is (rigidly) fixed to an infinitely rigid base, the EIb/L,
j 1000--
50 0 -
70.0
1 00 100.0-
50 0 10.0~ 100- ratio -+ co and the corresponding G -. 0. For this situation it is suggested to
I 30.0: 5.0 -- 30.0 - 5.0 - 5.0 -
4.0 - use G = 1.0 since there is some difficulty in producins a tmIy rigid connec-

I1
4.0 - 0.9 --
20.0 - 3.0- 3.0
4 o-- --
2.0 - 2.0 -
. When a beam or girder is used with adequate attachment to the coiurnn of
0°: -
1 '5.8
8'0 -
3 0 -- 10.0 -
9.0 - 0.8 --
1.0 -
8.0 -
7.0 - I .O - 1.0- a. The far end is hinged: multiply the E $ / L , ratio x 1.5.
6.0 - 6.0 - 0.9 - 0.9 - b. The far end is fixed: multiply the EI,/L, ratio x 7.0.
0.8 - 0.8 -
i so- 5.0
-
- 0.7 - 0.7 - 4. If beams are simply framed to columns, use Fig. 6-3 for K.
i 2.0 -- 4.0 - 0.6 - 0.7 -- 0.6 -

1 3.0: 3.0 -
-
0.5 -
0.4 -
-
0.5 -
0.4 -
-
-
The 'use of K factors obtained in the manner just described has been
required by AISC since 1963 and by A A S H T O since 1974. The K factors tend to
2.0 - 0.3 - 0.3 -
1.5 -- -
-
0.6 --
- 963. Since smaller K values had been used in structures that had an adequate
0.2 - 0.2 - ervice history, a new look was taken of the derivation for G,, Gb and the
1.0 - - -
-
resulting K. Yura (see "The Effective Length of Columns in Unbraced Frames,"
0.1- 0.1 -
- A I S C Engineering Journal, April 1971) correctly pointed out that where the
-
- K L / r ratio was less than C,, inelastic buckling should be considered and E,
1.0- 0 - o- 0.5 - o- should be used in Eqs. (7-4) and (7-5). The use of E, is equivalent to
i
~ - E,
Gine~asilc- -
E Ge,,st,c (75)
( a ) S ~ d e s w a ypermitted ( b ) No sidesway
ince El = AE and A I 1, it follows that use of G,, ,,,,, ic gives a Kkchric <
Figure 7-5 Nomographs for the effective length of columns in continuous frames for lateral restraint nce El is somewhat awkward to obtain and recalling in the derivation of &e
conditions indicated.
ISC equations for Fa in the inelastic region,
This equation may also be programmed for values of Go and Gb and to find the
corresponding value of F = a / K to satisfy the equality. A plot of this is shown
in the +-iornograph in Fig. 7-5b. we used essentially a SF on FCr where -9;.
..
\

The use of both nornographs shown in Fig. 7-5 involves computing values of

Go =
2 EIC / LC (at far end of column) (7-4)
2 EIb / L b
and in the elastic region
and G
- 2 EIc/ LC (at near end of column) (7-5)
- 2EIb/Lb Fa =
SF(KL/~)'
From the derivation involving Go and G,, it is evident that if we call one of the
values Go, the other end produces Gb (i.e., the values can be used interchange- From these equations it follows (using i = inelastic. r = elastic) that
ably).
FO'
When E = constant it may cancel from Eqs. (7-4) and (7-5); however, when G,= - G, (7-7)
Fa,
inelastic buckling is developed, E, should be used for E in the E I c / L c ratio.
Other considerations include: is computation neglects the variable SF. which ranges from 1.67 at K L / r = 0
23/12 (use 1.92 for hand computations) at K L / r = C'..For most columns in
1. When the column is pinned to the base, the E I b / L b ratio is zero, since the e range of K L / r = 40 to 60, the variation in SF is essentially negligible. The
theoretical value of I --+ 0 that results is G + co. For this situation it is se of Eq. (7-7) requires values of Fa, for the same K L / r value as F,,, so i t is
3U8 STRUCTURAL STEEL DESIGN BE.L\I-COLUbN DESIGN 3@

necessary to compute a table of values such as Table 11-7 or VI-7 (SSDD) and
using Foe= Fd to correspond to the AISC Specifications
1 2 - 7 ~ .' ~
F,' =
23(~~/r)~
Note thgtothis value is independent of 5. The values shown in the table of
F,' > 6 are for the purpose of using Eq. (7-7) and are not intended to be real
stresses that might be used in a design.
The use of Eq. (7-7) in an actual design requires iteration; that is, assume a
column; compute G,, G,, and find K; compute KL/r and Fa,Fi;revise G,, G,;
and find a new K for

1. Several cycles, or
2. To convergence, or *
3. To an arbitrary limiting K, such as 1.2 or 1.5. Figure E7-2
, ' ,l

Of course, if the section chosen is not adequate, a new section must be selected
and the computations repeated. Disque (see AISC Engineering Journal, No. 2, Find K:use W14 x 99 through three floors:
1973) proposed that the iterations for K could be eliminated by using fa = P / A
instead of F, to obtain
= 3.94
Reduce for inelastic effects:
In lieu of using f,, which might reduce Ge excessively, one may elect to use 570
fo = -z--= 17.86 ksi
F, = 0 . 6 4 . Again carefully note that the reduced G is used only when KL/r < 29.1
C,. It should be self-evident that values of K 5 1.0 are not to be adjusted. From Table 11-5 for fa = 17.9 ksi, obtain K L / r = 5 5 . For K L / r = 55
compute
Example 7-2 Given a portion of a high-rise frame shown in Fig. E7-2 with
sidesway permitted, assume adequate bracing perpendicular to bent so that (calculated also in Table 11-7)
K,, does not have to be considered (e.g., close contact with interior masonry Fd = 712E = 49.4 ksi
23/12 x 5 j 2
walls) and F, = 36. Using AISC specifications, design columns CD.
The revised
SOLUTION Assume that KL = 14 ft and use Table 11-4 to obtain a tentative
section W14 x 99:

For G, we note that the column is "dly attached to the base so that
I, -+ KI:

This value is not reduced for ~nelastlcbehav~or.From Fig. 7 - j a , we obtain

K = 1.38. With K = 1.38,
 BE.L'.l-COLL3W DESIGN 311

From Table 11-5 with this value of K L / r , we obtain SOLUTIONMake an initial column-size estimate that KL/r = 40, for xxEc5
Fa = 19.37 ksi F, = 178.2 MPa. For Fa = 178.2 MPa, the tentative column area is
P , =29.1(19.37)=563.6>520kips O.K.
Si:lce this value is about 40 kips larger than needed, try a W14 x 90:
Try W250 X 114.6:

520 KL = 34 (Table 11-5)

fa=-- - 19.62 ksi -
26.50 r
i c2m Table 11-7 obtain (note that this table is computer-generated and uses
S F = 23/12):
Fd = 129.18 ksi
F' =
r2(200 OoO) = 366.6 MPa (also Table V1-7). !
--- /
23/12~53~
7 = 1 as before, and from Fig. 7-5a obtain K = 1.25. The adjusted G, = 0.673(164.6/366.6) = 0.302
Right column:
G'b = 189.4/4.6 = 0 . use 1.0 (AISC rzcommendaticn)
co
Fa = 19.64 ksi (Table 11-5) Left column:
Pa = 26.50(19.64) = 520.5 > 520 kips O.K. 189.4/4.6
Gb = = co use 10.0 (AISC recommendation)
Use a W 14 x 90 for column 1. 0
///
Obtain from Fig. 7-5a:
Example 7-3 Given the frame shown in Fig. E7-3, with sidesway possible, K = 1.2 for right column
use the AISC specifications and Fy = 345 MPa to find the required column
size. Use the same section for both. K = 1.75 for left column
Check the left column first, since K is much larger than the right column:
KL
-=
1.75(4.6 x 1000) = 70,6
r 114.05
Fa = 143.7 MPa (Table VI-6)
Pa = AFa = 14.58 x 143.7 = 2095 kN > 2OOO 0.K.
Now check the right column:

F, = 169.7 MPa
Pa = 14.58(169.7) = 2474.2 > 2400 ki.; O.K.
A solution: use W250 x 114.6 sections for the columns
 Example 7-4 Redo Example 7-3 with sidesway somehow restricted. SHTO, and AREA) interaction equation with bending about both ~ x e A,,
s
as well as axial load as
SOLUTION One solution is to use K = 0.80 for the left column aqcording to
Fig. 6-36; correspondingly, one would obtain K = 0.65 using Fig. 6-3a for fa fbx
-+----I--<
-fby 1 (7- 10)
the right column and in both cases, using the "recommended" design values. Fa Fb Fb -
Alternatively, use Fig. 7-5b. rior to 1963, the value of F,, -. F,, = Fb. Currently, we recall that F, depends
Since the K must be less than 1.0, let us use the "experience" of the n several factors, including unbraced length and compact section.C -
Witeria, a d
previous problem to check a somewhat lighter section. Try a W250 x 101.2: general AISC allows "
'6
..>
"$

Fbx = 0.66Fy
or F,, = 0.60Fy
or F,, = values from AISC Eq. (1.5-6a),( 1 3-6b), or ( 1.5-7)
Fby = 0 . 7 5 5 for W shapes due to having solid rectangular flanges
F,' = 1032 MPa (interpolating Table VI-7)
Currently, Eq. (7-10) is used only in certain limited stress conditions. For t5z
remaining stress cases, ther more complicated formulas based on reszarch,
'4
plastic design,..,and elastic stability concepts are uszd. These will be psr-tidiy
Gb = 10 for left column
developed in the next several paragraphs to indicare some of the limirations so
Gb = 1.0 for right column that the practitioner will have some idea of how to follow through should tbs
Using Fig. 7-56, we obtain design vary from routine.
K = 0.72 for left cblumn Refer to Fig. 7-6 for a short ( L / r + 0) rectangular section of dimensions
K = 0.65 for right column X d that is stressed with both an axial force and a moment sufficient to
evelop a plastic hinge. The plastic moment in the presence of a compressive
Check the right column:
KL
-=
0.65(4.6 X 1000)
= 26.5 Fa = 190.3 MPa
rx 112.78
Pa = 12.9(190.3) = 2455 > 2400 kN O.K.
Check the left column:
-KL- - 0.75(4.6 X
1000)
= 30.6 Fa = 186.8 MPa
rx 112.78
Pa = 12.9 x 186.8 = 2410 > 2000 kN O.K.
Use a W250 X 101.2. ///

7-4 DEVELOPING THE BEAM-COLUMN DESIGN FORMULAS

Pnor to the sixth edition of the AISC Manual of Steel Construction in 1963, the
des~gnof compression members subjected to bending was obtained as
' -I,= b X :I.,,
f, + fb Fallow
Figure 7-6 Plastic hnge formation in a very short membsr subjected to both an 1G2! force .n\f
Dividing this equation by Fa,,,, = Fa, one obtained the widely used (AISC, moment.
314 STRUCTURAL STEEL DESIGN
.i'
'* "
~uk$i?&in~ for yo (shown on Fig. 7-6), webobtain
effect as Eq. (b) of Sec. 7-2. However, i t can be shown [see, for e:c~mp!e,
Timoshenko and Gere, Theory of Elastlc Stablliry, 2nd ed. (New Yock:
McGraw-Hill Book Company), Sec. 1-1 I ] that i t is sufficiently accurate (a
1 to 2 percent error) to amplify the moment for P-1:
Multiplying the P ratio by d 2 / d 2and noting that b2d2u.f= P.,', we obtain

However, from Sec. 3. Example 3-3, M, = u,bd2/4; thus we obtain
Although the development above has been made for a rectangular cross section,
it is also valid for all (including W, S, and M) shapes. A plot of Eq. (7-1 1) is
shown in Fig. 7-7. Also shown is the plot of a linear equation of the form
This factor may be called an amplfl~cationjocioi. since its effect is !o aF$if-j or
increase M, T h ~ svalue has been uced In the curvec shoun I n Flgs. 7-8 tb3 7-10.
The P I P , ratio is the ratlo of the actual column load to the Euler column load
and f,/F: is simply dlviding both loads by the column area. W ~ t hthis adjst-
ment in bending moment, we may rewrlte Eq (e) to obtain

If one were to plot
for k L / r = 40 and for K L / r = 120, the straight lines shown also on Fig. 7-7
would be obtained. These curves will be somewhat in error, since the P-A effect
$as been neglected. We could use an iterative approach to include the P-A
Shown in Flg. 7-8 is a plot of the loading situation u here tf, = .t12 = Jf
and in Fig. 7-9 the case where M I = itt and itlz = 0. These t-o plots represent
the extreme range of cases where a column is loaded w t h end moments, ma
building frame. The curves shown in F ~ g s7-8 and 7-9 have been made using a
modification of Eq. (e) proposed by Galambos and Ketter (Transacf~om,ASCE.
Vol. 126, pp 1-25, 1961), whlch gives, for equal end moments.

Figure 7-7 Plot of interaction

equations as shown. P, and Ma = F* 7-8 Influence of K L / r ar
allowable values.
 BF.A\i-COLW DESIGN 317
316 STRUCTURAL STEEL DESIGN

Figure 7-10 Plot of c o l u m Inr:r-

2 Figure 7-9 Plot of K L / r for 40
action for end moment on o a s end
and 120 for a column with mo-
and with efiscr of using C,=
hlo ment on one end.
(dashed lines) ior comp~fison.

and for unequal end moments a linear equation of the form may note that the reciprocal of Eq. (7-12b) is used as the C, amplification
or for laterally unbraced beams. Equation (7-12a) is used as C, in the AISC
s when using t : i ~

! I I ~ / , I I=~ ( - ) Ill \!! = 1-1

The coefficients are KL/r-dependent, and several values given by Galambos
and Ketter are as follows:

r %,

KL/r C D F G is in agreement with the definition for C, (i.e., single curvature bending is
0 0.42 0.77 1.13 1.11 critical for buckling instability than reversed cufvature, and simiIarIy the
20 0.70 0.46 1.14 1.18
40 0.99 0.17 1.16 1.23 The effect of using C, is to generally decrease the effect of the amplification
80 1.81 - 0.72 1.19 1.52 factor and is illustrated in Fig. 7-10. The use of C, produces a n intersection of
120 3.16 -2.51 1.25 2.53

The plots using Eqs. ( h ) and ( i ) are reasonably satisfactory for all of the cases
with equal end moment, but it is rather conservative for those cases of unequal

values other than at M , / M , = 1.0 and with a rapidity determined by the

researchers are as follows: / M , ratio. To the right of the intersection of the two curves, Eq. (d) cor:;:i.!i.
0.4Ml
Cm= 0.60 - - M2
2 0.4
is requires use of two equations in design and using the most critical .of

cm=
1.75 - 1.05M,/M2 + o.~(M,/M,)~
 We obtain for bending about the X axis the f o l l o ~ n g
a,
'llis is because it is easier to solve for the most critical value by using the two
P + C , ~ B ~ ~ a,
~ -, (P(KL)?
~ ) PdliOW -
equstlons than to make a plot and locate the intersection and then use the
goverli~ngequation. These equations are adjusted for design use by substitution where ax= 0.149~i-2x lo6 ksi (fps unlts)
' sf PC,for Py and M,,,,, for M, and with section area and section modulus to
P(KL)* = ksl
it by ob , stresses. This gives
' #

By analogy, Eq. (7-13) becomes

Fb = allowable bending stress

M
fb = -;;
3
,#@ For Eq. (7-14), we obtain in a similar manner:
E ~r CG iining terms and extending the case to biaxial bending, we obtain
Fa Fa Fa (7-140) *84
P-
0 . 6 5 + B, M,y-
Fb.r + B,. M, - I Pal,,W
AISC Eq. 1.6- 1a . . Fby

-fa+ cmxfb~ + 'myfby < 1.0 When j,/Fa (or P I P a ) < 0.15, we can obtain
Fa f a F b x (l-fa/F&)Fby - (7- 13)
'1 i
and
AISC Eq. (1.6-1b) A careful analysis of these equations shoxvs that F, is based on P,,,which
fa + - + - - < I
fbx fby
depends on the critical oallie of KL/r. Bending resistance and any ampii-
fication/reduction of moment effects is with respect to the bending axis ~~iitfi
... -l. ,..., ;* nrnh)pmc+ h P n\\nwahlP stress F_will be
2uial
.. --.-
L---:-b:--
S U v S C I 1 P L I I I ~> I I U W I I . 1 1 1 1 1 3 111
IIIUIIJ t/l
u u r r ~ r t rr ~ r r -. .-.
Referring agaln to Fig. 7-7, we note no reduction in moment capacity until based on K L / r , , but the P(KL)' term will depend on the axis resisting bending
P/ P, > 0.18, so rounding 0.18 to 0.15 for convenience and to be conservative, Since the X axis is usually used, it is the P(KL,)~product that would usualIy be
we obtain: required.
A

Limitation of f,/Fa 5 0.15: AISC Eq. (1.6-2)
Since Eqs. (7-13) to (7-15) are somewhat awkward to use except on a
cqmputer, let us multiply through by AFa to [using Eq. (7-13) without fby for a
particular illustration] obtain
The modified form of these interaction equations is generally considered to
be easier to use, since the right-hand side can be tabulsted for a number
of column sections and for several assumed column lengths based on kTL/ry.
Similarly, the terms ax, a,, B,, and By can be computed and tabulated. These a x
shown in Tables 11-4 ( 5 = 36 ksi) and VI-4 (Fy = 250 hipa) for the W shspcs
commonly used as columns. The AISC manual also has these tabulations for
F. = 50 ksi steel and includes use of S shapes and tube and pipe sections.
-Y

7-5 DETERMINATION OF THE INTERACTION REDUCTION

but ,$A the actual column load P and AF, = maximum allowable column COEFFICIENT C,
*(bad ${anbis not the same value shown on Figs. 7-8 and 7-9 and used to
-devcin:; ,.the curves shown). Also noting that /, = M/S, define A / S = B,
When a column in a structural frame is restrained against lateral translation
niult~ply/./F: by A/A, and take FiA = P, = 0.149 X IO~/(KL/~)'ksi (in fps).
with end moments, as illustrated in Figs. 7-8 to 7-10 and 7-1 la, the value of C,
320 STRUCTURAL STEEL DESIGN BEh?f-COLL'4N DESIGN 331

en the column has transverse loadings as in Fig. 7-1 1 c, use

1.0 + -7
+la
C,, = (7-15)
Fe
here f, = actual column stress
Fb= Euler stress as previously defined (including SF = 23/17)
$= factor determined from Fig. 7-1 ld, which depends on end restraint
and transverse loading
."?*

'7-6 A A S M O AND AREA REAM-COLUMN DESIGN F$&V~C'L~S

,,* ,,*$

AASHTO working stress design uses' essentially~thzsame equations as .A.,ISC,

with some additional adjustment to the amplification factor. The C
, factor is
defined similar to AISC.

'"Ib, + CmJby
_< 1 ' " (7-17)
(l-fo/~;;)~bx ( 1 --I;/F;)F~~
fo + -b - - + - fh
---I1
0.472F; Fb, Fb (7- 15)
n ' ~
F" F

2. I ~ ( K L / ~ ) '

Generally use C,,, = 0.85 for end conditions of Fig. 7-1 1 b and c; use C, = 1.0
hen the interior moment is greater than end values or with an interior moment

The AREA equations are similar to the AISC equations. For,f;/F, > 0.15:
rb."
Figure 7-11 C, reduction factor for beam-column interaction equations. (a) No sidesway: C , = 0.6
fbx
+ < 1 (7- 19)
, = 0.85. (c) Column with transverse loading: C, = 1 + +fa/F,'. ( d )
- 0 . 4 M , / M 2 . (b) Sidesway: C
1 -a /F .b . ( 1 - .fa / 1";') Fbv
+
Several cases of transverse loading and factors shown.

F;" = V'E
is computed using Eq. (7-12a): 1.431(~L/r)'

lso, at points braced in the plane of bending:

fa fbx
+-+---I1
jby
with attention to signs (single curvature = - M,/M2). Note also that M, is o.55Fy Fbx Fb,y
defined as the smaller of the two end moment values. When sidesway (Flg.
7-1 16) is possible, AISC specification allows:
fo 1
-+-+ ---I
fbx fby

Cm = 0.85 (sidesway present) Fo fby Fg,

 32.'; ... :UCTURAL STEEL DESIGN BEAU-COLL3c.F

7-'? 9EAM-COLUMN DESIGN USING INTERACTION value of KL' to e tables, obtained as

EQUATIONS '
.* KL
i , " #*? KL' = ---I
The design of beam columns using the interaction equations is essentially a trial rr/ rv
(i!.-i.;:tive) process. A section is tentatively selected and analyzed and if the It is necessary to use r,/r,, slnce the ratlo IS fixed for a section but is cor
i. se;.:ien is too small, a new section is selected and the process repeated until a known until a section is tentatively selected.
sz:i:i'actory section (both strength and weight) is obtained. The steps may be 6. Record Ptable,A 1 ry, r,/ry, LC,L,, B,, and a,. L, and L, are needed, so a rspid
ou: ,-.edas follows: determination of Fb can be made:
Fb = 0.66Fy I LC
if LUnbrace,
,.
gA
,-(erminethe axial. force and column moments. We note that this step is also
1.
1 yitive,, since indeterminate frame values are not found until a tentative
.,.!~tlon is used in the analysis.
Fb
Fb
= 0.6Fy L,nb,acedI L,
if

computed from Eq. (4-23), (4-26), or (4-27)

,

''2.::>&ipute K to obtain KL. It may be necessary to determine both Kx and Y,

= i f L ,,r,, > L,
!':bending on column end conditions and lateral bracing. The unbraced length IS the actual and not the KL value and IS taken with
2 . .'?.,timate the moment contribution (and we will use bending about a single respect to the bending 2x1s.
axis for illustration) as an equivalent axial load AP: 7. Compute K L / r cntical, obtaln or compute F,, and compute Ju = P / A .
Compare fa/ Fa I 0.15. If the ratio IS less. use Eq. (7-15) to see ~f the section
Fa is adequate. If fa/Fa > 0.15, ~t will be necessaryfto :heck barh Eqs. (7-130)
AP = C,BM- ,B [using a part of Eq. (7-13a)]
and (7-140). In this case compute:
Fb
P ( K L ) ~ use KL with respect to the bendlng a.uls. ~ h l c hmay be different
nr AP = BM-Fa [using a part of Eq. (7-14a)] from the cntical K L / r used to compute the allonable avlai stress
Fb Fa
'::,pection of Table 11-4 (or VI-4) indicates that 8. Check both Eqs. (7-13a) and (7-14a) to sat~sfq

C, Bx-Fa /3 = 0.2 to 0.3 (6.5 to 9.0 in SI)

fbx One should attempt to achieve an equality as clossly as possible for one of
the equations (it cannot be done simultaneously for both equations) and if:
;~ntl Bx- Fa = 0.1 to 0.2 (3.3 to 6.5 in SI)
Fbx + P,, << P,,,,, check for Ilghte~s ctlon
ax
P
P + Peq > P,,,,
4
use a 1 a ~ ~ e r s e c t i 6to4sa:~sfqdesign
,B = estimation for
ax - P(KL,)~
--, With care and some preliminary scrutlny ofv Table 11-4 (or Y I J ) , an
~ h t values
: for bending about the Y axis are considerably larger, but a AP
adequate section can be obtained In one to three inals. Thls IS possible. since the
&mate can be made similarly. An estimate for AP can be made in a tables show little change in r,/r, and B for the two or three sect~onson either
s,.rnewhat similar manner for using the AASHTO and AREA equations if
side of the selected section.
desired.
4. Compute the equivalent allowable column load as
Example 7-5 Given column and bending moments s h o \ n in Fig. E7-5 as
P,, = P + A P = P + (0.2 to 0.3)M (M in in . kips) part of building frame in which sidesway is possible, we will limit the
= P + (6.5 to 9.O)M - column section to not over a W12. Use the AISC specifications and A-36

+w,(;&
( M in kN m and SI)
steel. Select a tentative section and use the basic equations (7-13), (7-i4), or
ihk value of P, enter a Table such as 11-4 or YI-4 (or in AISC Manual ' (7-15) as applicable.
...
* ':! o i Yables which have been prepared for design use using a computer) with the
, ' K4, value and obtain P,,,,. If Kx/K; I rx/ry, the KL/r,, values control; if SOLUTION TO keep the deslgn of a beam column using the ~nterzctian
,
:,
.. P.',/K; > rx/ry, the KLx/rx ratio controls and one must use an adjusted equations In perspective, we wlll assume that K, = 1.3 ~ n dX; = 1.0 2r,d
@

: *6 r'
3.44 JIKUGIURAL STEEL DESIGN
BE.4'4-COLL5B DESIGN 3 3

By Eq. (7-13):
fu
- + ----
' d b r
5 1
Fu PFb,

0.47+ O 85(7'88) = 0,796 << I 0

0 934(22)
By Eq. (7-14):
i;+f- L <
1
0.65 Fb,
Figure E7-5
7.48 7.88
-+ - = 0.70 << 1 .O
that the K values do not change with section size. 22 22
:
Try a smaller section-try W 12 x 40.
l lo
Use C, = 0.85 with sidesway. After some study of Table 11-4 (with P > A = 1 1.80 1n2 ju = ----- = 9.32 Lsl
1 1.80
1 lo), let us try W12 X 50: KL
ry = 1.94 In -+ - = 74.2 F, = 15.98 k s ~
'7
A = 14.70 in2 (use Table 11-4) Pa,,, = 236 kips

5 L, = 16 > 12'ft -+ Fb = 0.6F, = 22 ksi

x'
- = 2.64 Fa = 16.06 ksi (Table 11-5)
Y'
. , S, = 51.9 in3

LC = 8.5 < 12 L,, = 19.6 > 12 ;

.Fb = 0.6% = 22 ksi -f, -- - =9.32
0.58>0.15
42.5( 12)
jb=----- - 9.83 ksi
F, 15.98 5 1.9
Sx = 64.7 in3 (Table 1-3)
F' =
.112~ = 1 1 I .3 ksl
P '" 1.92(36.62)
&=z=-= lo
14.70 7.48 ksi
9.32
p = 1 - ---- - 0.92
I I 1.3

By Eq. (7-14):
We must use both Eqs. (7-13) and (7-14):
0.58 + 0.85(9.83)
0.92(22)
= 0.991 < 1.0 O.K.
a 2 ~
Fel = - = 113.8 ksi By Eq. (7-14):
1.92(~~,/r,)' 1.92(36.2)'
+-<ru 1.0 1

0.6% F, -

932
22
+ -9'83
-22 - 0.87 < 1 0 O.K.
Use a W12 x 40 column.
326. 5 bTURAL STEEL DESIGN

E;.awple 7-6 Given the column and bending moments shown for a building
frame, with sidesway for K, restricted by use of bracing and shear walls, use ,
the AISC specifications and F, = 250 MPa steel to select a tentative column
section.
SOLUTION
Refer to Fig. E7-6 and assume that
1

Kx=1.25 k;=1.0

Figure E7-6

List. Cm = 0.95 (AISC actually allows Cm= 0.85 if desired). Check Eq. (7-14a):
Tentatively:

4 pgwen + pequv 4
114.1
E . 445.. + 7(41) [estimate factor as 7 (between 6.5 and 9)] (445) -+ 8.94(61) --_
150
I 14.1 = 753
130
< 868.3 kY O.K.
= 872 kN Using Eqs. (7- 13) and (7- 14) would give:
Eq. (7-13):
Sc2.1 Table VI-4 and select W310 x 59.5:
P,,,,, = 860.5 kN
1.25
A = 7.61 x l o v 3 m2 rx
-= 2.64 >- I
O.K.
'Y

Use as tentative section, W310 x 59.5. ///

Fa = 1 14.1 MPa (Table VI-5) Example 7-7 The top chord member (No. 6) of the truss used in Examples
LC = 2.58 m 4.87 m
L, = > 3.45 6-4 and 6-5 will be designed to include the member w e i a t and a temporary
Fb = 0.6% = I~O'MP~ concentrated force of 2.2 kips that will be applied to the center of the chord
during maintenance operations (see Fig. E7-7a). The bridge may be tempo-
rarily closed to traffic if the maintenance load is too large to be carried
safely with traffic (live loading). Other data:
F ( x L ~ )=~445(1.25 X 3.45)2 = 8.27 X lo3 kN . mi (in same form as ax)
Dead load = - 283.75 kips
fa =
7.61
= 58.5 MPa - ---
- 58'5 -0.51>0.15 Live load = - 109.44 kips
Fa 114.1
P = - 393.19 kips
Actual = AFa = 7.61 (1 14.1) = 868.3 (vs. 860.6 interpolated)
Using Eq. (7-13a): SOLUTION Since a pair of C15 x 40 channels was used in Example 6-4 for a
somewhat smaller load, we will try a pair of Cl5 x 50 channels (Fig.
Fa
P + G,,,,BxMx- ax < 'allow E7-76):
vB ' Fbx ax - P(KLX) -
A = 14.70 in2 I, = 11.00 in'
\*@
445 +,0.95(8.94)(613--
3
132'8
150 132.8 - 8.27
= 865.3 < 868.3 kN O.K.
 . 3% .A ~ C T U R A ISTEEL
. DESIGN

e used to support the siding and for lateral bracing. Caution is nec&sary,
owever, that girts should be continuous and, should building repair require
emoval of a girt, that it be done in only one bay at a time so that the Isteraf.
support is not lost.
3. There will be a column moment at the crane n l n w q l a e l due to the IongirudinoI
thrust of the crane starting or stopping suddenly. This will also produce a
moment at the base plate even if the analysis is made in such a manner as to
ignore the moment at the crane level. This force will also produce a column
shear that must be resisted by the anchor bolts at the base.

The general design of a column of this type proceeds as follows:

. As in Examples 2-5 and 2-6, tentatively analyze the structure and revise until
reasonable member sizes and deflections are obtained. Different member
.sizes can be used for any member. This was not done initially in the two
';62 Crane colurnns in an industrial building. ( a ) Column w t h bracket. ( b ) Stepped column.
examples, as the work is still highly preliminarq..
i&
(c) 1 . 4 ~ stepped
p column. 2. Based on the computer output. begin to redesisn the rnc.rnb.srs. Increesc or
decrease sizes, depending on forces and deflections.
I
. Reprogram and check output for forces and deflections.
bo.,. l:;e X and Y axes. Now if we take the X axis of the main column member . Repeat as necessary.
' %Pic.r~~~d for bending in the plane of the bent and the Y axis for bending out of
)iac,: (with respect to building length), we have the following considerations:
-8.1 Modification of K for Stepped Columns
pin-ended column free to buckle and with a load Poand an interior Ioad Pi a
in Fig. 7-13 is in a state of unstable equilibrium if the loads are sufficientlq. Irrrge-
1. 5 x e d or pinned against rotation at the roof truss level. The roof truss of the If we use the differential equation
ii~dustrialbents of Examples 2-5 and 2-6 provides rotation fixity (at least
nearly so), but translation may take place.
2. Sidesway control. The side sheds in Examples 2-5 and 2-6 plus any bracing in
the plane of the first and last bents will act to control sidesway. If it is still and allow for boundary conditions of different loads in the lower se-ment, no
excessive, knee bracing may be required from the column to the roof truss. lateral displacement at the column ends, change in I for the lower segment and a
3. There wiN likely be moments in the column at the roof truss. There will possibly common slope at the junction of the upper and lower sezgnents. we obtain
be moments in the column at the crane girder level due to lateral thrust of the
crane trolley against the rails making the track. There will be a column
moment at the base due to assumed base fixity. These moments produce
beam-column interaction for which Eqs. (7-13) and (7-14) must be used.
n* '
Y Axis
i $1

f
** d&iyi4!yss
:. :
" E I X ~or plnned at roof level. If we put some cross bracing in the plane of the
chord and vertical cross bracing in one or more of the bays on
'op;osite s<des;'the major amount of sidesway can be controlled.
where J = 1 + Po/ P, - a
a = L,/(Lo+ L , )
i . &!&Iintermediate bracing. This is necessary to reduce K L / r , of the main
I

This is a solution as given by Sandhu, "Effective Length of Columns nich

!::4ccr.;mn and of the column segment above the crane girder point. Girts may Intermediate and Axial Load," A ISC Engirieeririg Journal, October 1977. Ths ki
 32 STRUCTURAL STEEL DESIGN
BE.L.f-COLL?Ci DESIGS 33
I
Example 7-8 Make a tentative design for the crane column for the industrid
building in Example 2-6. Refer to Fig. E7-8a and assume the folio-kg:

I. Adequate bracing with respect to the Y axls of main column (don2

bays).
2. Sidesway in plane of bent with no top rotation and base fixed.
3, For initial tentative design, use only dead and live loads from computer
output + crane loads.
4. Crane girder will be placed at a level with the bottom of the side sksd
truss.
5. Use a stepped (built-up) section.

Figure 7-13 Figure for derivation of effective length coefficient for a stepped column.
i

value obtained as the solution of Eq. (7-22) is as used in Chap. 6:

Rearranging yields
'1 I'

but

and equating this load to the Euler load, we obtain

Solving for the equivalent length factor K, yields SOLUTION A single W690 x 264.9 was used to obtain prclimin~ryou:put
shown in Example 2-6. This size section (area and moment of ~ n e r t ~ ais)
necessary to reduce the lateral deflections at maln truss roof level acd at
side shed-to-column intersection to tolerable values. The built-up section
will require this or larger values to provide satisfactory lateral displace-
This value of Ke is for a pin-ended column, and it is necessary to multiply this ments. From the computer output (and for a tentative initial built-up section
by the AISC value of K as obtained from Fig. 7-5 or 6-3, where the actual end iteration) for LC = 1, the axial loads and moments (for left column) are:
conditions are taken into account. For general design office use, Eq. (7-22) Member 26 (uppermost):
should be programmed on a computer, so a plot of a' = Po/Ph versus K, can be
made for selected ratios of a. P = 318.74 kN moment = 120.59 k3 c' m
'4 STRUCTURAL STEEL DESIGN

Member 25 (intermediate): considered stable by adequate use of girts and siding). Use the largest force
P = 387.1 1 kN moment = 338.99 kN . m in the upper column, 387.1 1 kN.
Member 24 (lower section):
P=543.01kN moment=-313.6kN-m
base moment = - 190.1 kN . m
Let us somewhat arbitrarily try a section made up of one W360 x 314 main
. column and one W360 X 101.2 crane column, as shown in Fig. E7-86.
, . A.
Substituting values into Eq. (7-22) yields

Noting that the term to be evaluated as cotangent must be identifix! i$

radians, the problem can be quickly programmed on a pocket progamma-
ble calculator to obtain k L = 3.525.
I' Figure E7-86
W360 X 314 data: PI = 543.01+ 1262.11 = 1805.1 kN (actual)
I, = 1107.2 x m4 Zy = 428.7 X m4 12.426(1 + a') EI =- 15.82EI
PI =
A = 40.0 X lo-' m2 L L~
d = 399 mm bi = 401 mm
W360 X 101.2 data: Also,

1, = 300.9 X l o v 6 m4 ly = 50.4 X m4
A = 12.9 X m2 d = 357 mm bf = 255 mm
Compute r and ry. Use EMx, = 0 to locate the new Y- Y axis.
(40 + 1 2 . 9 ) ~= 910(12.90) -
The actual K = Ke X Kc,, ,,,,,,,, ., From Fig. 6-3c obtain k',,, c,,,,,,, -
910(12.9) 1.2.
X = = 22 1.9 mm (use 222 rnm)
52.9
K = 0.79(1.2) = 0.95

From Table VI-5, obtain F, = 134.6 MPa From Table VI-7, obtain F,'=
732.8 MPa. Check the Interaction equation [Eq (7-13)]. 9
A

Check allowable stress Fa based on K L / 5 (since the other direction is

 736 STRUCTURAL STEEL DESIGN BE.+(-COLWW DESIGN 337

' Use ,C = 0.85. (We will not check for bending about the X axis at this
time, since computations are very preliminary. After the next computer run,
if this section is still satisfactory, we would check bending about both axes.)
I
13'6(0.8 16) = 27.7 MPa
# ;#
, &'= I
=
9.235 (note decimal shifted for 1 )

U. Assume that Fb = 0 . 6 5 . [We will have to compute this later using Eqs. 1.5-6
and 1.5-7 (AISC numbers) as appropriate.]

Figure 7-14 Sldesway control. ( a ) Masonry shear wall ( b ) D~agonalbracing in zcl<crsd bay.

By inspection, Eq. (7.14) will be satisfied for the lower column segment.
The upper column segment should be checked for interaction to make
sure that the W360 X 314 is adequate as a column for the full height of the
bent.
We can now reprogram this example with the new column sections and
any revised sectlons for the truss members and see if the lateral deflections
far the several load conditions is satisfactory and that the bending moments
and axlal forces are compatible with the section being analyzed. This should
be done prior to refining the final design, to keep the engineering calcula-
tlons to a minimum and maxlmize use of the computer.
///
7 9 CONTROL OF SIDESWAY
1
It is evident that the most efficient column design results when the frame is
adequately braced against sidesway. With no sidesway:
1. C,,, can be less than 0.85 (but may be more in selected cases).
2. The effective length factor K is not greater than 1.0.
A rigidly framed structure can translate laterally sufficiently to undergo
"~idesway."Note also that the development of the equations for the G factors is
based on a common slope at a joint that can only be obtained for a "rigid" joint.
It is therefore necessary to provide specific resistance to sidesway to obtain the
most efficient columns. ~ h i s ' m be
a ~ obtained (see Fig. 7-14) via:
1. Shear walls (use rigid vertical walls of brick, tile, or concrete block to contain
the lateral movement). If masonry walls are used, a close contact with the
column should be provided so that the column cannot translate in the
construction void between materials.
2. Diagonal bracing (essentially produce a vertically oriented truss) in one or
more bays.
It may not be necessary to use n g ~ djoints wlth shear walls or diagonzr
bracing. However, some materlal economy may be obtalned via use of rigid
joints as well as providing some addltlonal room for uncsrtalnties.
0 criter;2
Diagonal braclng requirements are usually small The controlhn,
may be the L / r ratio rather than the cross-sectional area requirements. Gs!am-
bos ("Lateral Support for Tier Bullding Frames." A ISC Engtneering Jozirml.
January 1964), using essentially the same method proposed by Winter for Iatsrzi
bracing of columns and beams, develops an expression for the area of tF,c
diagonal brace member:
where A,= area of bracing, in2 or m'
C P = sum of all column loads at a story level In plane of bracing
a = horizontal run of diagonal brace/length of column
E= modulus of elasticity assuming column and brzce of s a ~ e
material, ksi or MPa
Diagonal bracing is usually designed only for tension. It is assumed that 13-
bracing member is so small and flexible that it will buckle (with stresses we:!
below 6)under a very small compression load. For opposite-direction loading,
the buckled member straightens with no damage due to small buckling stresses
and prevents sidesway from occurring by carrying the necessary tension load.
Example 7-9 Given the story of a tiered building shown in Flg. E7-9, design
the diagonal bracing for the intenor bay. Assume that this wll be placed in
alternate bents in the out-of-plane direction.
 4%
BE.k\f-COLb\fX DESIGN 339

Additionally, we must always satisfy the following

P,, 5 0'.pu
When there is bending about only one axis we have [he following:

and

+
I n these equations = 0.86 (for beams) and o = value given in Table 3-! i ~ i
columns a n d noting +,
varies from 0.65 to 0.86 d e p s n d ~ n gon k Z / r . The C,
terms are as in AISC. The values of P, are
Pu=AFy(l-0.25r12) ~i\'> (7-30)

AFy
P, = -4 (but use appropriate bending axis)

T h e design forces a n d moments are as computed in Chaps. 6 and 4.

T h e use of these LRFD equations will be tllustrated by the folIouing
example.

Example 7-10 Given the beam column and loading shown in Fig. E7-IC.
select the lightest W310 shape using F, = 350 bIPa and LRFD.

Figure E7-10

SOLUTIONUse F, = 1.5 (author's cholcc).

Mu, = l.l[1.1(120) + l.5(\80)] = 42.2 L
Si - rn
P,, = 1 . 1 [ 1.1(400) + Ii(6Cn'l: = Id7-! 'K"
 U

STRUCTURAL STEEL DESIGN BMU-COLUMX DESIGN

/
With bending only about the X axis, Eq. (7-28) applies, so that PROBLEMS

4-1 Using the general solution for column buckling,

-
Inspection of this equation indicates that x = A rin!iy + B cosky + Cy + D and X =d-&
show that K = 0.7 for the column of Fig. 6-,3b. Note the boundary conditions:
, x = Oaty = Oandy = L

M=Oaty = O -"
Q-
-0aty = L

7-2 Determine the effective length coeific~entsK lor the colurnns In the frame shown In Fig.
Note that the far end of the girder of column A is pinned.
Answer: AB: 1.75; FG: 0.78.
Using these values of A and Z , try a W310 x 157.7 section:
A = 20.13 X
ry = 79.0 mm rx = 138.9 r, controls
Z, = 2.687 X m3

from which
Figure W-2
qic = 0.90 - 0.25(0.534) = 0.77
Pu = AF,[ 1 - 0.25(0.534)~]= 20.13(250)(0.9287) = 4674 kN
7-3 Determine the effective length coefficients K for the columns Ln the frame shown in Fig
1
M, = ZF, = 2.687(250) = 672 kN m - Annuer: BC = 0.65; EF = 0.63.
Substituting values into Eq. (7-28), we obtain
1474 442
0.86(4674) 1.18(0.86)(672) = 0.367 + 0.648 = 1.015 > 1.0 N.G.
+

- I ill
t:
Check:
+ 3 u 2 p,

0.77(4674) = 3599 >> 1474 O.K.

Since this column is just over, the next largest W310 section should work.
Tabulating data so that a comparison can be made, we obtain

W310 x 178.6: A = 22.77 X m2

ry = 79.5 mm (and controls, as before)
Zx = 3.055 X m3
* By inspection of this data it is evident that the section is adequate. It is not
necessary to check that Pud i +=PU,since this ratio is almost that obtained Figure W-3
from the first term of Eq. (7.28). Use a W310 X 178.6 section.
///
Figure P7-5

Figure W-8
 344 STRUCTURAL STEEL DESIGN I!I
I ,

7-9 Design the top chord of a railroad truss with a panel length of 27.583 ft for the flollowmg
conditions:
Live load = 867.2 h p s (E-80 loading)
Impact = 473.6 kips
Dead load = 402 kips (estimated)
.@
use a bullt-up section somewhat as shown in Fig. P7-9. Use the AREA specdications and A-36
steel. The chord member ends will be either riveted or bolted.
t
A m e r : Two S24 X 100 and one S24 X 79.9 with a 30 x cover plate.

7-10 Check the section of the column above the crane runway girder of Example 7-8 and redesign as
required.
7-11 Using Example 7-8 as a guide, make a tentative redesign of the main column of Example 2-5.
Use A-36 steel and the AISC specifications.
7-12 Redo Example 7-2 with the column sue limited to W12.
A m e r : W 12 X 96.
7-13 Design the diagonal bracing for the bent shown in Fig. E7-3 to inhibit sidesway. Use the
lightest pair of angles with a 12-mm gusset plate.
Annoer: Two L63 X 51 X 4.8.
7-14 Check the exterior columns of Example 2-3 using the computer output and resize the columns
if necessary. Note that both exterior columns are to be the same sue. Use a single column (no
sphc*for full building height. Use A-36 steel and the AISC specifications.
7-15 CHeck the interior columns of Example 2-3 using the computer output and resize the columns
as necessary. Note that basement columns are not necessarily the same size as the upper column
wbch is to be used for full building height. Use A-36 steel, the AISC specifications, and not aver
W 10 columns.
Answer: W8 X 48 bottom; W8 X 40 upper.
7-16 Use the computer output of Example 2-4 for the exterior columns as outlined in Prob. 7-14.
Limit column size to W250.
A m e r : W250 X 67.0.
7-17 Use the computer output of Example 2-4 for the interior columns as outlined in Prob. 7-15.
Limit column size to W250.
7-18 Venfy w t h computahons that the W310 X 178.6 column section of Example 7-10 is adequate.
.
$f@ 7-19 R & ~ OExample 7-10 d ML = 210 kN m and D = 425 kN. AU other data are the same.
Answer: W310 X 178.6.
7-20 Redo Example 7-10 if My moments are present: M,, = 50 kN . m; MLy = 75 kN . m.
Answer: W310 X 282.8.
 BOLTED A N D RIVETED CONNECTIOhTS

8-1 INTRODUCTION

A steel structure is produced as an assemblase of the structura1 members

making up the framework. Connections are required where the various member
ends must be attached to other members sufficiently to allow the load to
continue an orderly flow to the foundation. Since the connection serve5 ru carry
a d from or to adjoining members, i t must be adequately designeti. c :
;
I
onnection design involves producing a joint that is safe, economical of materi-
Is, and capable of being built (it must be practical). The more practical
onnections are usually more economical. since fabrication costs greatly affec:
economy of both connections (or joints) and the members thsmsc.!v:s, 2.i
ustrated earlier, particularly concerning built-up tension and compression
embers. Several structural connections are illustrated in Fig. 8-1.
Connections (or structural joints) may he classified according to:

. Method of fastening, such as rivets (hardly ever), bolts, or welding. Connec-

tions using bolts are further classified as bearing orjricrion-type cop. nsctions.
2. Connection rigidity, which may be simple, rigid (as produced by an inde-
terminate structure analysis), or of intermediate rigidity. The AISC, in See.
1.2 of the specifications, classifies joints based on connection rigidity as:
FIgure Vm-1 High-strength bolted joints. ( a ) Splicing smaller-to-larger column using filler plates. Type 1: rigid connections that develop the full moment capacity of tilt
(b) Splicing same-size column. ( c ) Diagonal bracing. connecting members and retain a constant relative ang!: between tI.5
connected parts under any joint rotation.
 P
(el

used as tension hanger.

Type 2: simple framing with no
connected parts. Actually, a small amount of moment will be developed,
but it is ignored in the design. Any joint eccentricity less than about 2; in
(63 rnrn) is neglected.
Type 3: semirigid connections with less than the full moment capacity of the
connected members being transferred. Design of these connections re-
quires assuming (with adequate documentation) an a;bitrary amount of
moment capacity (e.g., 20, 30,
3. Type of forces transferred across the structural connection:
a. Shear forces: common for floor beams and joists.
b. Moment: either bending or torsion.
c. Shear and moment: as in type 1 or 3 connections.
. . , d. Tension or compression: as for column splices and for "pinned" truss
members.
e. Tension or compression with shear: as for diagonal bracing.
4. Connection geometry:
a. Framing angles used to connect floor joists and stringers to beams and
columns.
and a web angle for shear. Web angle is optional.
b. Welded connections using plates and angles.
c. End plates on beams or rafters.
d. Plates or angles used on one side of a floor joist or beam.
e. Seat angles with or without stiffeners.
Several of these connections are illustrated in Fig. 8-2.
5. Fabrication location:
a. Shop connections: produced in the fabrication shop.
b. Field connections: joint parts fabricated in the shop but assembled on 2-s
, .
a. Friction connections. Connections designed as friction connections ha~:s
 connections. Connections where the joint resistance is taken as
ation of connector shear resistance and bearing of the c o m s c

r'
1 against the connector. This mode of beh , r #evelo
sufficient slip occurs to bring the connected matePlal id contact wi
back projection of the connector pgar the w6king or 'ddrsipfocd.
connector shear is a portion of the resistance in beanag$onnecfio
analysis, the reduced shear .area available for threaded connectors w'cz
the threads are in any slip plane requires 3, reduction in the design load. In
actual practice, threads in the shear plane.result in a lower allow3S:s
design shear stress for the fastener.

The design of both friction and bearing corinections involves use of an

wable shear stress. The value is much lower for friction connections, siccr
nt'any joint slippage under working loads. The value is consider-
or bearing connections, since some small amount of relative move-
een the parts making up the joint can be t o l e ~ a t ~ dBoth
. types of
dition to being designed for "shear," are rout~nelychecked for
on the net section and for bearing of the connected material against the

fabrication practice tends to use of slotted over-sized holes

connections. The slot allows easier field erection since more a l i m e n t
e is available for temporary erection bolts.

2 RIVETS AND RIVETED CONNECTIONS

For many years rivets were the sole practical means of producing safe 2nd
iceable metal connections. The process required piinching or drilling ho!:s
roximately 1/16 in (1.5 mrn) oversize, assembling the parts using drift pins to
ign the holes, and using one or more bolts to hold the parts togetbet
Figure 8-3 Several modes of joint resistance. (a) Bolt shear. (b) Plate shear or tear-out. (c) Bolt
bearing. (d) Plate bearing. (e) Bolt tension failure. (A Tension on net section.
mporarily. &vets were heated in a furnace (portabie for onsitz use) to a cherry
ed color (approximately 980°C) and inserted into the aligned hole through tLe
everal parts to be connected. One member of the riveting crew then applied a
connections is not developed as the shear resistance of the connectors; ng bar with a head die to the manufactured rivet head to hold the rivet 1I1
rather, it is developed as the product of the clamping force produced by and to shape. Another crew member used a pneumatic driver with a h a 3
tightening the bolts (or driving the rivets) and the coefficient of friction to forge the protruding rivet shank to produce the other head. The forg.ing
between the clamped parts. It is expected that in load factor resistance ation simultaneously reworked the rivet metal and caused a shank enluge-
design this will be directly used as the design parameter, producing an t to very nearly fill the oversized hole. This reworking and shank enlarge-
equation of the general form used in several other design codes (outside the ent, together with the shrinking of the hot rivet, produced a substantial joint
United States) as of the time. The rivet contraction during cooling is resisted by the joint
rial and develops tension in the rivet so that a riveted joint is intermediate
een a friction- and a bearing-type connection (a bearing type is commonly
ed). This joint transmits the design load primarily by friction between the
where + = performance factor (0.67 to 0.70) ed plates making up the joint. The riveted joint has had a long history of
p= coefficient of friction x number of slip surfaces cess under fatigue stresses as in railroad bridges. Only recently has ,ARE.-:
2 A, F, = total developed clamping force as the sum from all the bolts used in
llowed use of high-strength bolts and welds in joints for railroad bridgts.
the connection
# - 3 STRUCTURAL STEEL DESIGN BOLTED &\TRIVETED COhC+TCnO?iS 355
41
t the procedure for the design of a riveted connection is exactly the sao;,e a
for a bolted connection. Figure 8-4 illustrates several sizes of undriwn rivers 2nd
structural applications using rivets.

8-3 HIGH-STRENGTH BOLTS

There are two general classes of bolts used in structural appIications. These artre
general-use A-307 (ASTM designation), sometimes called unfinished baits.
These bolts have a somewhat rough shank and bearing surfaces, since not as
at care is taken in their manufacture. The A-307 bolts are made of steel ~ i t I . ~
ultimate tensile strength F, on the order of 60 (grade A) to 100 ( g a d s B)'hi
15 to 690 MPa) and available from $ in (6 mm) to 4 in (102 mm) in diarnerer
in lengths from 1 to 8 in, in increments of 1/4 in. and over 9 ir.. in
ements of 1/2 in. A-307 bolts are available with several head a n d auf
configurations, but the hexagonal and square head are most cornmody used.
Several sizes of A-307 bolts are illustrated in Fig. 8-5.
A-307 bolts are cheaper than A-325 and A-490 bolts and sbouid be use2 in
tatic load structural applications whenever possible. Applications incIuds csc in
mall structures, locations where the bolt installation is visibIe for i ~ s d ~ i
serviceability checks, and in service loads which are relatively smaII.
High-strength bolts are available in the ASTM classifications, sizes, sad
ultimate tensile strength shown in the lower inset of Table 8-1.
The general length, head, a n d nut configurations are the same IS for 11-307
olts except that larger diameters may not be available. The A-325 bolts can be
btained with metallurgy for special purposes, such as high resistance to coxo-
sion. A-325 bolts may also be obtained with a galvanizing coating,
When high-strength bolts were first introduced into structurd applic-': &LAG.?LS.
washers were required to spread the bolt load to a larger area of the sofie:. r;,ttir!
of the fastened parts. This requirement was partially caused by the nut a n d head
to dig (called galling) into the A-33 and A-7 (F,= 33 ksi) steel avrri!?ds
time. Current high-strength bolting application; require that a hardraed
asher be used under the turned element as follows:

Method of tightening

Turn-of-nut Specified torque

A-325" No Yes
A-490b Yes Yes

a Washers are required when uslng oversued bolt

holes. Washers are requlred when flange slops is

greater than 1 : 20 (S and C shapes both have a
slope of inner flange face of approximately 1 : 6).
Use washers on sloplug flange faces as above.
Figure 8-4 (a) Several sues of undnven structural nvets. ( 6 ) k v e t s in structural apphcations Use two washers when the fastened material has
F, < 40 ksi.
 i
i;?,
-F',.-'-

BOLTED &\Q R I w C O h > % ~ O ~ 359

S
,J
;':'i. p
. 4.

r proper installation re p= Slip coefficient (usually can use 0.35 for clean mill scale; most 0~~~~
surfaces are less than this value. and i t may be necessctry to deternine
the value by test)
m = number of slip surfaces
N = number of fasteners
T = proof load of each fastener (as in Table 5-2)
eration of Eq. (8-1) we can readily see that regardless of +e t_Vcof
L
d (friction, or no slip, or bearing with some slip acce~&bIehthe
eed P,,,, before either bolt shear or bolt (or m a t i
7 may now see the rationale for P, = t(,K, t3.s

joint actually develops resistance as n cornbinstion of boir shear

e 8-1 What is the nominal safety factor against relative slip in a

-type joint using 20-mm A-325 bolts? Take ,U = 0.35-

LunoN The bolt proof load is 141 kN (Table 8-2). The allowable
ess using AISC specifications (Table 8-1) is 120 hfPa with threads.
ne. From Eq. (8-1):
Psli, = rn,uNT = k11(141)
since it is only necessary to consider one bolt and we will consider Q ~ one
Y

iTU= ultimate tensile strength (see Table 8-1). slip plane, as in a lap joint. The safety factor is always defined as
Prrslsting
SF =
Pallowable

~dequacy. Pallow = F,Ab(nornina~)

A bolt tension of approximately 0.7Fu gives adequate reserve strengt
Combining, we obtain the safety factor:
the bolt be somewhat overstressed (say, 3/4 turn instead of 1/2 turn).
The belt tension acts as a mawive spring in tension to hold the fastened parts in 0.35(141)
SF = = 1.31
relati' ~ositil~n...
This clamping effect also tends to hold the joint against nut 120(0.7854 X 0.0'0@)I @
loosening in fatigue load situations, SO that most of the time a locking nut is not
If A-325 bolts have not been excessively overstressed (not more than The reader should note that this safety factor is against slip and is not
3/4 turn of nut) they may be reused one or more times. Tests on reuse safety factor of the joint, which is on the order of 1.67 (for a tension
indicate that A-490 bolts should not be reused in any situations. t and may depend on tension on net section). f//

FACTORS AFFECTING JOINT DESIGN

has considered several factors involved in c @ n ~ e c : i ~ n
pertains to fasteners. We *ill now consider ~evc;a1
 S ! STRUCTURAL STEEL DESIGN
8-4.1 Joint Length
One factor of considerable importance is joint size. Obviously, smaller joints
economical of material. However, since an assumption is made that eac
fastener in a joint carries a prorated share (equal for constant-size fasteners),
problem arises for long joints. Referring to Fig. 8-8, we see that the distribution
of strain is unequal from the frontmost bolt to the rear bolt. If the joint is too
long (with "too" not being specifically defined here), it is evident that the first
bolt will carry more than P / N of the load and the last bolt will carry nothing to
almost nothing. With the base metal or plate designed to be adequate for tension
in the net section, the plate does not pull apart but does stretch based.on
P L / A E , so the forward bolts (or rivets) will either undergo compatible shear
strains or will shear off if the strain and resulting forward displacement of the
hole is too great. The loss of the forward bolt will transfer the load to the next
bolt(s) in line, and the next bolt may shear, and so on-producing a progressive
joint failure (a process called "unbuttoning"). Note, however, that with the large
loads involved, this process is very nearly instantaneous. If the joint is short
enough that all the bolts carry load, the first bolt strains with the plate. When
strains corresponding to yield stress (shear) develop, the bolts continue to strain
with no increase in load and the next bolt(s) in line will pick up the transfemed
load. The ultimate joint load is reached when all the bolts have yielded. Strain
compatibility analyses are seldom made, since the factors of safety used.together
with the property of steel ductility are such that except for long joints, only the
first bolts (if any) in a connection are yielded or are close to yield.
We should note that the safety factor for the connection (particularly he
fasteners) shocld be higher than for the members being connected. This is so
+that a'member failure will always occur before a joint failure. A joint failure will
generally be catastrophic, whereas a member fai,lure is likely to allow time for
safety measures to be undertaken.
Recalling that no joint (with holes and in tension) is more than 85 percent
efficient, and based on the work of Bendigo, Hansen, and Rumpf ("Long Bolted
Joints," Proceedings, ASCE, Vol. 89, ST6, December 1963), the efficiency of a
I I I I
p i {
P he bolt is in double shear.
,Is
.'
4 Strain distribution (approx) in
bottom plate ofjoint above
y m 8-8 Load and strain distribution in bolted joints.
. . . ,, , ., ,,, ,,. .... ..
Oi*
assuming reasonable bolt spacing on the order of 3 x diameter
E = 0.85 - C , ( L - C2) L 2 Cz
fps: C, = 0.007 SI: C, = 0.00275
C, = 16 in C, = 406 mm
is equation indicates that connections with joint lengths up to 4%
efficiency of 85 percent (i.e., no reduction in connection capacity for j
greater lengths there is nearly a linear 10s~o f Joint
ching a capacity of about 60 percent of the short joint whe~lthe 1sng:h is Qa
order of 50 in (1250 mm).
The cufient AISC specifications indirectly ailow for long joints by use of
percent efficiency factor and adjusting the allowable fastener stress-
lues for fastener stress are considered valid (for bearing connections) up to a
nt length of 50 in (1250 mm). Above this length the allowable shear stress is (0
reduced 20 percent. Recalling that the specifications provide minimum
uirements, the structural designer has the option of using Eq. (8-2) for
termediate joint lengths, between 16 and 50 in.
-4.2 Edge Distance
lifie of stress are located too close to the i.dgr. i t may be possible
plate as shown in Fig. 8-3b and in the actual joi%t shown in Fig-
be avoided by using an edge distance obtained by equating s h c a
es using F, = constant for both bolt and base metal, to obtain
A,F, = dtF,
h the edge distance d is
A
d =4
t
distance is required by AISC in Sec. 1-16.4 and using an edge distance
thing holes, it is necessary to have an adequate edge and end
avoid warping damage to the material. The AISC specifications (See-
e distances for this based on the nominal bolt diameter. For bolt
5 1: in (30 rnm) nominal dimension measured from center of
Rolled edge: D, = 1.4 >( diameter (rounded to n?ar?St in Or 3 mm)
 30- .UC?URAL STEEL DESIGN

I'
BOLTED 363'
&\ID RIVETED COWEC~~OE~S

and superposition of effects eventually produces a l~mitingfnction resistance

(i.e., yN does not increase w~thoutbound w ~ t hlncreaslng N). A spacing that is
too close can cause difficulty In ~nstallingthe fasteners, slnce the wench hezd
requires a mlnimum working space Theae problems are rcsol~zdby usmg the
mlnimum spaclng requirements
AISC (Sec 1-16.4): ,,s, = 2.67 x diameter (3 x diameter preferred)
AASHTO (Sec 1-7.22C): ,s, = 3 x diameter
. .
-, AREA (Sec. 1-9.3): smin= 3 X diameter
Maximum spacing of a single line of fasteners in the direction of sris:is
should generally be limited to 12r, where t = thickness of thinnest part being
clamped. This spacing can be used for AISC, AASHTO. and AREA.
Fabrication practice coupled with wide usage of the :-
through I-in-diaizs-
ter fasteners has led to certain standard gage distances. These values are shavn
in the tables for rolled shapes (see Tables 1-3 to 1-7 and V-3 to V-7 of SSDD)
and depend on width of angle' leg for angles. as in Tables 1-13 and V-13 and in
the AISC manual.
The user must also check the minimum edge distance for angles so that the
fastener size does not result in a hole too close to the edge to satisfy the -

specifications.
The rolled shapes, which have very wide flanges. may have a second gagz
line. Values for the very yide flanges (e.g.. W14 larger than 142 lb/ft) are shown
in the AISC manual, as well as in tables available from the steel producers, but
are-not shown 'in the SSDD tables.
The reader should note that the "standard" gage distances will generally
result in a more economical fabrication cost, but these distances are not the only
ones that can be used. The designer should, however. consult the fabricator if
other than standard gages are contemplated so that an economical joint is
F b#:-. Several modes of joint failure. ( a ) Several joints. (6) Bolt shear failure. (c) Tension on net produced.
section iL.:-re. ( d ) Tear-out failure due to bolts being too close to end in the direction of the stress.

8-4.4 Minimum Joint Design

For bo!: diameters larger than I f in (or 30 mm) use:
AISC specifications require that all connections. except those in rrusses, c2rrying
Sheared edge: D, = 1.75 X diameter calculated stresses be designed for the design load but not less thrin 6 kips (or 77
kN).
Rolled edge: De = 1.25 X diameter AISC specifications require that truss joints in either tension or compression
?*A be designed for the design load but not less than 50 percent of the effective
8-42 Bolt Distribution and Gage Distances strength of the member based on the type of desi,on stress.
The AASHTO specifications require that connections be designed on the
It is necessary to ensure a reasonably compact joint and one where the basis of the average of the design load and the effective strength of the member
cted material is in reasonably good contact so that the developed friction but not less than 75 percent of the effective strength of the member. This is
is uniform between the parts. If the bolts are too close together, because many of the structural members in U S H T O design are controlled by
is obtained, since the maximum coefficient of friction is p 0.35, -- factors other than stress, such as L / r . At least two fasteners are :equired in any
AASHTO-designed connection.
 364 S~RUCRRAL STEEL DESIGN BOLTED A\?) W ~ E CDO ~ X X C ~ O ~ S
.c

tion.

I
8-4.5 Shear Lag
-Pi
Long Joints are undesirable from the standpoi& of reduced efficiency (below 85 used for strut
percent when L > 406 mm), but in cases where W, S, or C shapes are used with
gusset plates on the flanges (see Fig. 8-10), ~t is necessary to produce a jo
sufflclentl~long that the stress in the section at A-A can be transferred to a

resulting progressive failure across the section. A measure of shear lag

efflclenc~is based on the distance from the gravity axis of the member to the
fastener (or gusset plate) plane. Munse and lChesson ("Riveted and Bolte
Joints: Net Section Design," Proceedings, ASCE, STI, February 1963) give an
equation for shear lag efficiency:
L = SD
1 percent f o r each & I P ( ~m)In eXCe,S of
L = 450
where x= distance from gravity axis f 0 fastener plane (= d / 2 for w shapes and Same as M S H T O
simply 2 given in tables for angles)
L = joint length
For most well-designed sections, Eq. (8-3) should give a value of e , from 85
A to 90 Percent. The AISC reductions for shear lag were presented in Set. 5-3.2.

I n ~ t ~ transfer
dl of some flange

basic premise In connection design is that each fastener c a r 7 a prorated share

--- f the connection load. For a joint with constant-size fastener~in a s ~ m e t r i c z I

- Ptotai
'fastener - number of fasteners
is assumpllon made for ellher fastener shear o i bcanns. BeannZ i s ccriid-
'"%re 8-10 Shear 1% in a W shape connected to a pair of gusset plates. ed In some speclflcatlons as beans: of fastener on ha$ nlelal 37J G~ C a t
? +* *
,"cur
 - t STRUCTURAL STEEL DESIGN 1
BOLTED .LXD RIbEXED COh3ZCTIOE;S

Table 8-3 Table to determine required bolt length.Fased on grip and thread ble 8-4 Allowable bolt and rivet stresses for AASHTO specific~fions
length (use table to determine if threads will fall idshear (use glven
b! stress mm normnal wit area except l u r r \ - ~ v iwiu 211
plane; see example below)
L,,, -
gnp + L; round Lrcqdto next larger f in or 6 rnm
on)B
/ e stress. F-
r ~ i d w a b ~shear
Bolt sue, m Thread L L, m Bolt sue, rnm Thread La
-
I I1
I 1 -
16 12.5 25
7
'
i If 8 15 30
1; 1 20 35

2 proximate and based on soft conversion.

r,xample Gwen the connection shown m sketch and uslng a 2-~n-&meter bolt. Are Data from AASHTO specificat~ons,12th ed , Secs 1-7 22 and 1-7 41
tt : J ~ S In the shear plane (defmed by the plate junction)?
Fb = allowable beanng stress on nvet or bol: from fastened material.
Grip=;+ & + ;= 1&1n F n c t ~ o nconnectlons are required for connectlons subject ro stress reversal.
Fiwi. the table, L = 1 In: Bolts In beanng-type connections must be X-type (thresh excluded from s h a r
nes). Beanng-type connections are to be used,@
%:
L,,, = I& + 1= 2 1 condary members. Reduce h~gh-strengthbolt values 2
16
an 24 In and F, < 42 k s ~
tqr + * R ~ u n to
d nearest larger t N, threads m shear plane. X, threads exciudbd
Lreqd= 2f In
T h e a d length = 1; (from table): *"\% .
2.25 - 1 375 = 0 875 In ds are In the shear plane for bection. Table 5-3 W!i be
to determrne ~f the threads are In the S
h
Distance for two plates = + $ m = 0.8125 In !.XASHTO, and A E X
av
-,-be used to obtain the alIowabl
Therefore, threads are out of the shear plane.
ecificat~ons,respect~vely. Table 8-6 may be $0 obta~nthe dIott-ahis
earing stress for base metal-to-bolt or XISC values s h o i t ~in
,
ineta1 on fastener. The shear area is obtained using the nominal fastener his table are based on the ultlmate material, compcisd
b. diameter. The bearing area is the projected fastener diameter x plate thickness:
the fastener load Pi is
P, = A , x F, (shear) - --
- --- - -- - .. is
. distinction
No . made for fasteners in either sin&
- .
- or doubIr shrzi. T i e
Pb = D X t X Fb (bearing) SHTO bearing stress based on metal to fastener is
No allowance is made for the hole size being in (1.5 mm) larger than the hnl
shank in bearing. 4
- 1 0 F, = 1.225
$
n the fastener group is unsymmetrical or the load does not pass through nd will generally be limited by fastener-to-metal stresses. as shown in Tab12 8-4
oid of the fastener group, the fasteners are not equally
- * stressed. This der the column headed F,.
ii?,ibn is considered in detail in the next section. The AREA bearing values are stipulated for the fastener type except th2t
.We note that the allowable shear stress F, depends on the design assu
$? ring is not considered. in- the
. . design
- ,
for connections using his!-strength WE.
,,tie: (:.at the joint is either a friction or a bearing type, and whether the ese values are shown In 1 able 8-6.
1
 BOLTED hSD RIV- CO

Table 8-5 AREA fastener stresses (all connections are "friction"type) ber 1965) give average values of ultimate shear in terms of ul
rength F, (see Table 8-1) of
F, Fo
Fastener ksi MPa ksi MPa

&vetsa
Hand-driven - - 11.0 76
Power-driven - - 13.5 93
~olts~
A-325 36 , 248 20 138
A490 36 248 27 186

Use beanng stress on n y t s : single shear, 27 ksl or 185 MPa; double shear, 36 k s ~or
i 2 5 0 MPa. e
Need not cons~derbeanng on bolts In frict~onconnections. . I

0.62(825 MPa)
F= = 4.26
Table 8-6 Bearing values for rivets and bolts by several specifications 120
(top part of table is metal-on-fastener,bottom part is fastener-to-metala int geometry reduces this value of F to something on the order of 3.3 rot
Material FU AISC AASHTO AREA mpact joints (in tests) and to around 2.0 for joints whose length is in excess of
70 mm (50 in). This value of F compares to the tension value on the cross
ksi MPa I
I
ksi MPa I
I

ksi MPa I ksi MPa I ksi MPa ction for A-36 steel of

F = - - F" - 1.72 (net section after slip)

0.58 F,

F = F" - 400 - 2.67

A-307 bolts
&vets A-501 and A-502
Power dnven slngle shear
double shear
A-325 and A-490 bolts
a Top part: F, = 1.5F,,; AASHTO: Fb = 1.22Fy. Bottom part: depends on fastener;
fastener bearing generally controls for bridge design.
8-4.9 Nominal Factor of Safety of Fasteners and Connections
The nominal safety factor against slip for bolted joints is on the order of 1.25 to
1.30 (as computed in Example 8-1). One might ask what the nominal safety
factor,d,the..jointand of the mechanical fasteners is against failure.
The safety factor of the mechanical fasteners can be readily estimated based
on the ultimate bolt tension and shear stress values divided by the allowable
values given in Tables 8-1, 8-4, and 8-5.
Shear strength tests on high-strength bolts (see Wallaert and Fisher, "Shear
Strength of High Strength Bolts," Journal of Structural Division, ASCE, ST5,
---- (gross section before slip)
0.6Fy 150
servations of joints in long service indicate that a safety factor of F 2 2.0 lor
250
fasteners gives satisfactory service.
Not required
-4.10 Splices in Beams
ms are often spliced to produce continuous spans. Splices are usualIy p!aceiI
lose to the location of zero shear in the span. Any use of mechanical fast*Pnzrs
the tension flange will reduce the effective area somewhat. Based on otes:s, r~ss
ISC and recent AASHTO specifications allow ths designer to use the ,
lange area for stress calculations as long as th? holes (in the tension flange) are
s than 15 percent of the tension flange area. When the area of holes e x c r d s
percent, the flange area is reduced by that portion of holes in excess oi 15
For example, if A, (of tension flange) = 20, A,,,,, = 4. the percent holes = 4
x 100/20 = 20 percent. The excess hole area = 20 - 15 = 5 percent and, the
ension flange area is 20(1 - 0.05) = 19, not 16, as would be obtained by t-.l.ring
t all the hole area (20 - 4 = 16).
These computations should also be used for the flnngz spiics plates.
370 STRUCTURAL. STE BOLTED .OmRIVETED COh?iF,mOt;S 371

A-325 bolts a r i required for the tension sp r a three bolt line and sheared edges, the minimum edge dis
. E8-2a if the metal is F, = 250 MPa a 1.7(22) = 37.4-mm use 40 mm
be friction type, or bearing type.
Minimum bolt spacing = 3 0 = 3(22) = 66 mm
+
-mm bolts, so the hole diameter = 22 3.0 = Use spacing =
250 - 2(40)
- O.K.
must be sheared twice for the beam web to 2
splice plates, the load per bolt is For the forward two bolts:
85
I / P,,,, = 2 X A, x F, Center bolts as 40 + - = 82.5 nim fronI edges
11 ,' 2
For a frictidn-type connection:
I Spacing to edge is less than maximum allowed of 12t = 12 x 10 = 120 m a .
P,,,,= 2(0/7854 x 0.022~)(120)= 91.2 kN/bolt (double shear) Use the distance from the front bolt to the edge of splice plate at 40
he number of bolts required is N = 442.5/91.2 = 4.85. Use five bolts. (1.75D for sheared edges) and similarly from the back bolts to the edge of
the bolt pattern shown in Fig. E8-26, .so that the splice plate wihth will the W410. Set the gage distance so that-only two holes are deducted from
maximum but the maximum net section is obtained for the channel. the critical net section.

C310 X 44.6

s =\/21z
= 46.1 mm use s = 50 mm (arbitrary choice)
Check the bearing:
On splice plates: Ph = iV x ,+Ihx F, = 5(0.01)(27)(1.50 X 400)
= 660 > 442.5 O.K.
On web of C3 10: Ph = 5(0.013)(22)(1.50 x 400)
= 858 > 442.5 kY O.K.
Use five bolts for the friction connection.
For a bearing-ope corlrrecrlorr
Figure W2a Figure E8-2b Grip = 2 x 5 x 13.0 = 23.0mm
L,,,,=23.0+28=51 iise55rnrn
Check the net section of the splice plates. With three bolts out, the are L,,,,,, = 38 mm (Table 8-3)
requirements are:
Width of one splice plate + beam,web = 5 + 13 = 18.0 mrn
442.5 Distance from bolt head to end of thread = 55 - 38 = I7 mm < 18 ma
Gross: A, = = 1.475 X m2
2 x 0.6Fy This computation shows that the threads are in the shear plane, so F, = 13
442'5 MPa (instead of 205) and P,,,, is
Net: Ap = = 1.106 X m2
2 X OSF,

say 4 bolts
A >-=
"lo6 1.301 x m2
- 0.85

Try two plates 250 x 5 mm with three holes in the critical section. .= 2.6 (also use 4)

A,,, = [250 - 3(25.0)]0.01 = 1.75 X > 1.106 O.K. Use two columns of two bolts for bearing-type splice
 373
STRUCTLTRAL STEEL DESIGN
II BOLTED .kKD RIYETE# CObYECTIOhS
.-3
Example 8-3 Design the connection for t 6-4, the allowable axla1 compressiv; stress was 9.35 ksi.
(NO. 7) for the highway truss of Example , and 6-6. Refer to Fi P,,, = AF, = 17.00(9.35) = 159.0 kips
E6-6 of Example 6-6 and to Fig. E8-3 (ref 159.0 -t 99.3
8-46). The previous examples were used pa" = = 129.2 kips
2
members 7 and 9. Let us use the same cro or the top chord as was
used for the end post (all in compression The 75 percent member strength criteria give
ng web members may
use the same sections as for 7 and 9 ( controlled rather than Po,, = 159.0(0.75) = 119.3 kips < 129.2
stress). In any case, let US design the conn e vertical web member From Table 8-4, the allowable bolt shear stress F, = 13.5 ksi. Thz nuxSsr
(No. 7), which is in compression. Data f of bolts required in the connection to transfer 129.2 kips is
P 129.2
N=---?= = 15.9 bolts
AbF, 0.7854 x 0.875" 13.5
Use A-36 steel and the AASHTO specificatiJns. Use A-325 high-strength
Use N = 16 bolts for symmetry and since 0.9 bolt is not possible. Tne use of
bolts. P,,, = - 99.3 kips (P,, = - 40.6 kips). The gusset plate t = 5/16
16 bolts requires four rows. Use a bolt spacing of 3D:
in (minimum t allowed by AASHTO for a plate).

718 bolts as req'd. Memno

Use a minimum edge distance in l~neof stress,,=
,.
1; in (1.750 for cut ecZ):
TODchord / I'
1 .
Total length of joint = 4(2 625) + Z(1.5) = 13.5 in
W ~ t ha joint length of 13; in (nominal), shear lag IS not
minimum required transverse edge distance = 1.25D rounded
larger 1/8 in or 3 mm for flanges of beams and channels (but
and other elements). Thls glves
4 = 1 2 j ( 8I ) = l $ in (1.125 in)
The distance furnished and based on the standard gags distance = 5.5 in
(see Table 1-3) is computed as
b, - 5.5 10.01 - 5.5
dlurn= ------- = = 2.26 ~n > I . 125 O.K.
2 2
I Check the bolt beanng on the gusset plate. slnce I, = 0.313 < 0.640 of
flange of a W12.
Figure J33-3
x 0.3 13)(40) = 175.3 kips > 129.2 O.K.
SOLUTION The fastener design will be based on a friction-type connection. ade sufficiently wide that tension on a net section
AASHTO does not allow a bearing-type connection in a main membe li be some f~llerplates needed between the gusset
Stress range does not have to be considered for connection design. nce the W 12 sections are deeper than I2 in. ///
We will try four 7/8-in-diameter bolts at a section as shown in Fi
' .Jf\
E8-3, the same as assumed in the design of member 9 for tension Example 8-4'Design,.thg connection for the vertical members of the mSn
Example 5-7. A deduction for net area does not have to be made f roof truss of Example 2-6 that were designed in Exampls 5-6. F, = 250
compression members unless a complete stress reversal occurs (this does not MPa. Other data include:
occur here). P = 70.18 kN (tension)
AASHTO (Sec. 1-7.16) requires that a connection be designed for th
average of the design load and the full effective member strength but not
less than 75 percent of the effective strength of the member. From Example
 ."a ' ,t*'
STRUCTURAL STEEL DESIGN 375
BOLTED .CYD WYETED. ~ O ~. &..T I O K S

This design was based on using a 12-mrn gusset plate and 25- With the'threads in the shear plane, F, = 150 MPa able 8-1):
high-strength bdlts.
P,,,, = 2t0.7854 x 0.025')(150) l d = 147 kN
SOLUTIONAISC requires that connections for truss members be designed 199.5
for either the design load or 50 percent of the effective strength of the N =- = 1.36
147
member. td
L Use two bolts for shear, since a fraction is not possible.
P,, = 0.5(0.6<)~, = 0.5(150)(2.66) = 199.5 kN controls Check the beanng:
Fb = 1.5F, = 1 j(400) = 600 b1Pa ( Fu from Table 8-6)

The bolts through the angles and gusset plate will be in double shear, as On angles: Pb = 2(2 x 0.0063)(25)(600) = 375.0 > 175.9 kN 0.E';-
illustrated in Fig. E8-4b. Assume that a bearing-type connection (slip can On gusset plate: P, = 2(0.012 x 25)(600) = 360.0 > 175.9 !c3
tolerated) is satisfactory (a designer's prerogative). Check the bolt len
using Table 8-3 to see if the threads are in the shear plane. Use two 25-mm A-325 bolts and a 12-mm gusset plate.

-5 RIVETS AND BOLTS S U B J E m D TO E C C E h i C LOXDlN

nerally, when the eccentricity of the load on a bolt group is less than sboct
in (60 mm), it is neglected. Joints such as the simple frame connection of Fig
a, which is widely used, are in this category. The bracket connectiori of
g. 8-12a is loaded with an eccentricity that is obviously too l a r ~ eto be
glected. The framed beam connection may be large enough that the resuf5ng
ccentricity,,isalsd too large to neglect. One may note that the standard frzrnstf
connection angles in the AISC design manual neglect the eccentricity for vduls
to about 3.7 in (one of: the standard framing angle connections with nearly
value of maximum eccentricity is shown in Fig. 8-12b).
A load to be resisted by a bolt group that is eccentric with respect to the
id of the group pattern can be replaced with a force that has a'lineof
through the pattern centroid and a moment with the magnitude Lti = PP,
ere e is the eccentricity of the load. This is illustrated in Fig. 8-l3a a ~ 5.d
15-mrn bolts considering that each bolt in a pattern that is centrally loaded carries its
1 2 7 X 89 X 6 . 3
ed share of the total load, we have, for equal-sized bolts,
:I -,.I
I V\-FJ/v P
I
I p.=-
_I-
I
( " !V
I
where Psi is the shear force on the ith bolt with a vector to resist the a p p l i d
Figure E846
force P .
An additional
~ . -- bolt force is develo~edbv the eccentric moment ,bf = ?r.
Bolt grip = 6.3 x 2 -t 12 = 24.6 mm Assuming a group of bolts acting as an elastic unit. we have a concept similx to
24.6 + 30 = 54.6
L = use 60 mm that of beam resistance being developed and as related to the beam moment of
L,,,,,,
45 mm =(Table 8-3) inertia. Referring to Fig. 8-14, we have a bolt pattern \%-ithan applied moment .El
One,aQgle + gusset = 6.3 12 = 18.3 mm + which produces a resisting moment for rotational equilibrium that is equal to
,=n
Thread runout location = 60 - 45 = 15 mm
i- l
15 mm < 18.3 mm. ' .threads in the shear plane (see Fig. E8-4a)
f vbe assume that the value of R, is proportional to the distilnce f r o n the
 1I : Z ,
1.
3 73 STRUCTURAL STEEL DESIGN BOLTED L\D RWEED COk>~~3
o77~8 ,s
\
L::-IJ~ the action of P through the bolt group centroid produces an additi he bolt pattern and plate adequate for the g~venload in a bearing-~?c--
l'olt resistance R, which must be added to the R,,,,,, vector, it is better nnection assuming threads In the shear plane?
obtain the H and V components of R,,,,,,. Referring to Fig.
components of Rl are Rh and R, can be obtained by proportion as: SOLUTIONSince the bolt pattern 1s symrnetncal (as in most ~ractlclrpeob-
]ems), the centroid of the pattern is readllq located and marked 3s c-g-, 3s
Rh - -
- RL Rv =
- R, shown on sketch
Compute z ( x 2 + y 2 )
R,=R-=Y - - MdY - -My
I dl C.d:dl

S,rnllarly, from Fig. 8-14b, it is evident that

erld can be interpreted as the polar moment of inertia of a group of unit are
:.ate that if we use the area A with the denominator of either Eq. (A or ( g ) ,
arLJR, are obtained as stresses. Multiplying the numerator by the area A of
zth !~olt produces the force R, or Rh. With A in both the numerator
denominator, it cancels, giving R as a force. For general design, the equat
are

f 18.3 1l s.? Figure E & 5 b

Example 8-5 Given the bracket connection shown in Fig. E8-5a and th
<
the fasteners are 25-rnm A-325 bolts and the plate is = 250 MPa steel,
Com~ute

P = l I0 k N

shown on Fig. E8-56 to iesist applied P.

compute e = 150 + 125/2 = 212.5 mm:

R,,, = 0.5088~

Set up a table as follows and omit the signs of I and y (use veciors
previously drawn on Fig. E8-5b to determine the direction of the ir afid c
 vectors):

Placing these values on Fig. E8-5b, it is easy to see that bolts 1 and 5 are th
most highly stressed (critical). Bolt 3 is loaded the least amount (13.47 kP.,
The resisting force on bolt 1 is computed as

R =\/(31.8 + 18.33)' + 38.2' = = 63.03 kN

63.03
f, = = 128.4 < 150 MPa O.K. kips
0.7854 X 0.025' X 1000
x = 2.75 in
Check the plate bearing: I

' =
63'03
0.025(12)
= 210.1 < 1.5(400) O.K.
y, depends on the number of bolts Try 12 bolts Note t h ~ boIt
the most stressed for a bolt pattern such as this.
t

C(x2 + y2) = 12(2.7512+ 4(7.j2 + 4.j2 + 1.5')

I is alw~ys

Check the possible tension rupture of the plate along the forward bolt line: 8
J
= 405.75
Moment of inertia, I = 0'012(0'270)3 - 2(0.012 X 0.025)(0.075)~ 50
12 R, = -
12 = 4.17 kips

887.5
R, = ---
405.75
(7.5) = 16.40 kips
I 16.308(2)
Section modulus S = - = = 0.12-.. . . ., ,
,,
C 270 887.5
R, = ---- (2.75) = 6.01 kips
4 405.75
Moment at forward bolt line = Pe' = 110(0 15) = 1 6< LN . ., Q
Since 16.40/0.7854 = 20.88 > 17.5 ksi, the numbzr of bolts at a I-in diame-
f,=s=--
0.1208
- 136.6 MPa < 0.6G O.K. ter is too small without computing R.
Check plate buckling:
Try 16 bolts (bypassing 14, since 12 bolts were so hi@Iy stressed):
C(x2 + y2) = 16(2.75)' + 4(10.j2 + 7.5' + 4.5' + 1.5') = 877.0
- = - -150 - 12.5 < 2 5 0 / f i O.K.
t 12 50
The joint is adequate for bolt shear, plate bearing on bolt, and bending.
R, = -
16
= 3.12 kips

//J
Example 8-6 Given a crane runway bracket that carries a load as shown in
R, = 1.012(2.75) = 2.78 kips
Fig. E8-6. Use A-36 steel and either 7/8- or 1-in-diameter A-325 bolts.
Assume a friction-type connection. Find the number of bolts and the R =\/(2.78 + 3.12)' + 10.62' = 12.15 kips < 0.7Sj-l( 17.5) O.K.
bracket plate thickness.
Use sixteen 1-in diameter A-325 bolts.
 'Wh6 I - 1.
'w STRUCTURAL STEEL DESIGN BOLTED f i RI\ETED.
~ CO~+XXOP.IS ,333

Find the plate thickness for the bearing (assume that the column flang ~q 8-7 for the fastener eccenrnc:ty shov*n Fig-
IS adequate):
tp(1)(1.5 X 58) = 12.15 kibs l + 5
e,,, = 3.625 - 0 625 In
----- =
2
12.15
tp=-= 0.139 ~n ccentnclty (for 101nt performance). it is to sre
87(1)
ctlce of neglecting the eccentnclty for most f r ~ * d
Check the plate bearing along the forward fastener line and neglect bolt nable design procedure.
$ales 1/16 In larger than bolt:
h=7x3+3=24in M=50(15)=750in.k1ps Fb=22ksl ~~~~~l~ 8-7 Redo Example 8-6 taking into account the reduced ecceL;:::-

tph3
Ip=--
12
2tp(lO.S2 + 7.5' C 4.52 + 1 S 2 ) = 7745 S ~ L U T I ~Referring
N to Fig. E8-7, try twelve I-in-diameter bolts-
$1
Me 750(12)
fi f b = F b = 2 2 + - = -1 774%
'6

750(12) = 0.53 in
t = ------
774(22)

AIsC, Set. 1-9.1.2, for unstiffened elements under compres

-b< - - 95 - 15.8
t - v36

15.0
t=--
15.8
- 0.95 in use tp = 1.O in
Figure W7

Summary: use sixteen 1-in-diameter A-325 bolts and a bracket plate = 1.0
In.

M = jO(14.25) = 712.5 In . kips

8-3.: AISC Reduced Eccentricity for Connections
C ( x 2 + y 2 ) = 405.75 (Example 5-6)
T h b preceding method of joint analy~lswith eccentricity is widely used. Based
on a series of tests which displayed that this method of joint analysis is
ccrslderabl~~ ~ n s e ~ a t i AISC
v e , uses (in tables for design of eccentrically
lo . ? L J fastener groups in the design manual) a reduced eccentricity which is
cc-.?~uted,for a single line of n fasteners:
R, =
*" (2.75) = 4.83 kips
405.75
I + 2n
%& eeff= e - --- R, = 1.756(7.5) = 13.17 kips
'$c 4
llne? of fasteners with n fastenen in any line: R =\/(4.17 + 4.83)' i13.17' = 15.95 kips
l + n
--
' 8 ,
. eeff= e
2 j-,=-- 15'95 - 20.3
0.7854
> 17.5 N.G.

1
f i
% STRUCTURAL STEEL DESIGN
BOL I E D L\I) R I L E r t D COh3ECTIOhS x!?

Try fourteen I-in-diameter bolts:

8
e,,, = 17.75 - - = 13.75 in
2
M = 50(13.75) = 687.5 in . kips
C ( x 2 + y 2 ) = 14(2.75)' + 4(92 + 62 + 3') = 609.9

H" 6873
--(2.75) = 3.10 kips
=
609.9
R, = 1.127(9) = 10.14 kips

R =\/(3.57 + 3.10)~+ 10.14' = 12.14 kips

12.14
jv = l.O(O.7854) = 15.45 ksi < 17.5 O.K.

Using the reduced eccentricity requires two fewer bolts in the connec-
tion. The plate thickness of 1.0 in is still necessary to satisfy b / t , and
beardg IS not a problem.
i/.t
The AISC Design Manual gives tables based on one, two, and four vertical
columns of fasteners which may be used to design eccentric connections. By
assumlng the number of fasteners in a row, a computation for e,, is made, and (dl
with n and e,,, we obtain a coefficient that is multiplied by the allowable
fastener load to give the total group eccentric load. One may readily derive an
equation for the allowable eccentric load for a single vertical (or horizontal) line
-Q-15 Frarmng connecttons. (a) Moment-reslstlng connecbon using top and s a t m$a bckd
o column flange. Web angles c a n y shear. View l w k m g down (b) M o m c n r - r e ~ t m gc o c n c c ~ ~~rz, : o
of fasteners. Equations are given in several textbooks which may be used in an olumn web. Note that top and bottom plates are welded lnto column web a d flrrc&eto ~ ; i
oiumn snlrclici>.(c) F r a m ~ n gfloor system In pgwer statlm Note >hear zufirnen 113 %eb of r ~1.
attempt to reduce the computational effort in finding the number of fasteners der. Coping IS shown for small floor beams IJI nesr foreground (d) F r z m g for bndse Siriz,t,s.
for an eccentrically loaded connection. The author suggests that with the F
increased computational efficiency available to the designer with the pocket =A.

calculators, it is as easy to "punch it out'' as to try to use an equation developed
by someone else. This is so particularly because no simple equation exists and
most equations require some iteration anyway.
8-6 BEAM FRAMING CONNECTIONS
8
'
A number of the joints shown in Fig. 8-20 are ..~tandardizcd'~ as to b9uir
ttern and with an angle length that depends on the beam size and as given in
AISC design manual. The angle is selected based on bolt bearing and wirh
ee dimensions that depend on producing adequate edge distance and without
I I L L G L L G L G L L L G"etween bolts and wrench during installation if the boIt hokes are

Figures 8-2a and 8-15 illustrates the most common methods of framing steel
structures for small buildings of five or fewer stories where the connections are
AISC Type 2 (simple shear connections). The eccentricity of the beam shear is
neglected with these connections, and Ao moments are assumed to be trans-
ferred across the connection. Frame stability is.provided by use of wind bracing
pi masonry walls closely fitted to the columns as shear walls.
3ligned both vertically and horizontally. Use of these mbles often prodcccr u
:onnection that is overdesigned. However, the cost of ovcrdesip is gsncraiiy
more than offset by reduced fabrication costs from using standard dimensions.
The simple framing tonnection is used to connecr strin,oers to floor beams
~ n dfloor beams to -girders in bridge fabrication. In bridge design the conncc-
tions should be standardized for the given bridge to reduce fabrication co,jE;
neral standardization for bridges is not as easily done.
 b
3PtL STRUCTURAL STEEL DESIGN BOLTED LL?)

Zxample 8-8 Design a beam framing connection for a W 14 + 1/16) as in Fig. ES-8-this .@'?s
to frame into a W18 X 50 girder. The W14 x 3 0 carries
(live + dead) = 2.5 kips j f t and the span is L = 18.0 ft.
; specifications, A-36 steel, and A-325 high-strength bolts.
designed in Examples 4-1
rength bolts. F, = 138 hfPa
ction-type connections.
SOLUTION
From Example 4- 15, obtain
V,,,, = 78.3 kN
v , , ~=,140.5 kX
Figure W 8 V,,, = 178.0 x 0.795 = 1 1 1.7 k N
0 should be coped as shown in Fig. E8 Total shear = 280.5 k s
ch side) as shown, but we may note t ~h~ floor beam is a W760 X 160.7 and frames into a plate girder (Example
angle on only one side of the web.
sides it is evident that the bolts in t I,. = 16.0 mm
. ii-ebs of both the floor 'girder and the floor beani will be in double sheay.
t,., = 13.8 mm
v = -w=L
(2.5 + 0.0')(18) = 22-77 lups
2 2 ~ e ust use 22-mm-diameter bolts:
LTsc. 3/4-in A-325 bolts in a friction connection. jId) = 52.3 k s
p,,,, = 0.7854(0.022~)(135)
,.. ",_....
3. .'I....

P b , , , = 15.46 kips double shear (Table 11-7, SSDD) For the web of a W760:
I!! a W14 x 30: = 2.7 use three bolts (double shear)
22.77
Nshear = -- 1.47 use two bolts or bearing with bolts in double shear, Fb = 150 hfpa (Table S-5):
15.46
- 280.5 use bolts as required
22.77 = 3.7
- N b r g - 0.022 x 250 x 13.8
= 1.29 use two bolts
Nbrg - 1.5 X 58 X 0.75 x 0.27
F~~the web of a plate [bolts in single shear and Fb = 185 hfPa (rivet
In the web of a W18 x 50:
value); Table 8-5)):
22.77(2) 280.7
= 1.96 requires two bolts Nbrg = 0.022 x 185 x 16 = 4.3 1 use six bolts
Nbrg = 1.5 X 58 X 0.75 X 0.355
22.77(2)
15.46 = 2.95
Nshear = use four bolts for symmetry Nshear = - --
280'7
52.3 - 5.36 use six bolts
Use a web angle with a length of 6 in and t to be determined: Place bolts as shown in Fig. E8-9. ,
For bearing:
2t(0.75)(1.5 X 58) = 22.77 ~ 7 6 x0 160.7 *L

t = 0.174 in T , -'77 ,I:- # !!- ![;a

For shear: - - 1 1 ; ~ 1 .. ,
-
; 1 , ; +(--
- 1

ii
! 2t(6 - 2(0.75)(0.4~,)= 22.77 1;-
(1
.r tz0.176 uset=$in
'
an angle with a length of 6 in. Use two L4 x 4 x 1/4 x 6 in long with
, I
 STRUCTURAL STEEL DESIGN

4 ' Design the framing angle:

'
'
L=5 X 30 +2X 1.750 = 18.50 = 18.5(22) = 407rnrn
use410
The angle thickness will be controlled by the bearing of the three
in the outstanding leg (o.s.1.) fastening the beam to the girder; therefore,
t 2 16 mm.
2t(0.022)(3)(185) = 280.5
i' t 2 ll.5^'mm
,!fx;& Checking the tables, try: L127 X 127 X 11.1 (Table V-9):
g = 75 mm (Table V- 13)
Effective edge distance de = 127 - 75 = 52 mm
." 4 d,,,,,,, 2 1.5D = 33 mm
A L102 X 102 X 11.1 will not give sufficient clearance between the --
-- - W7hO
bolts and the bolts through the girder web in the o.s.1. The small amount of
overdesign is rather negligible in any case. The design is summarized in Fig.
E8-9. Note that AREA does not require a "bearing" check using A-325
bolts. This check established the approximate angle t and is recommended
I
whether required or not.

8-7 FASTENERS SUBJECTED TO TENSION

~.~
,.,,, ... .,," -
1

Figure 8-16 illustrates the usual conditions for bolts in tension. When the bolt ;c
tightened to develop the proof load, the shank elongates. Simultaneously tL-
clamped plates are compressed. When we apply a load to the connection, we
have the free body of Fig. 8-16c, which gives

Obviously, as P becomes larger and larger, two events occur simultaneouslv:

1. The bolt tension T in9reases slightly, producing a slight shank. elongation.

2. Shank elongation reduces the plate clamping pressurk u, since the plate
compression e, = 2T$,/ApE is small and caused by r, being relatively small
and A, relatively large. On the other hand. e,-,, = A - -T-r ,/ / A E i--
""ll ------
q npnrlv nf

constant magnitude as long as the plates remain in contact.

In equation form and now considering a single bolt as in Fig. 8-16d. we

* -$4
- Ae, = Aebolt
j- ,$
he two clhmped plates expand, and taking P' = prorated part of total P
3% STEEL DESIGN BOLTED hVD RIVETED COh%xmNS 3% *
where terms ~reviouslydefined in Eqs. (8-9) and (8-10) and ~ i 8-17
~ are
.
force carried by flange on one side of beam web A,,, = 0.7854(0.875)~= 0.601 ln'
=stress ratio, defined as (5 - F,)/F,; for A-36 steel, F, = 22 ksl (bearing-type connection and threads In shear plane)
r = (36 - 22)/36 = 0.3889 62.4 = 2.36 use three bolts
0.601 X 22 X 2
'a = thickness of beam flange supporting load
Check the bearing:
pb = 3(2 x 0.25 x 1.50 X 58) = 131 < 62.4 kips o.K-
8110 Design a hanger using a WT section for a load to be a
by a palr Of ~~t~ if thickness of the WT web 2 2 x land, bearing and shear
Use A-325 bolts and A-36 steel; the load is 62.4 kips.
The load Is from the bottom of a W33 x 221 beam as shown in Fig.
E8- I Oa. Step 3. Design WT.
~h~~ step simply lnvoIves studying tables after making a computation
W33 X 201 approximate depth based on edge distances and bolt spacing in
d, 2 1.5 + 2 x 3 + 1.5 + k
= 9 + kln
a WT12 X 47:
d = 12.15 in b, = 9.065
tf = 0.875 t, = 0.515 k = 1-53 In
- 2L's -
%dge
- 9.qo-c - :5.
-
= - -
- *I
5 -& \ '
7 ; &?\K \ '??l+. S>L3)
7-
I
3 = 9 + 1 53 = 1053 > 12.15 ln
Required depth 0.K-
6 2 4k
Figure =lob Check the bending moment at the toe of the fillet In web.
Firwe m-loo
55 1
> *> . b = -L -- = 2.6875 in
SoLuT1oN The hanger design will requlre design of a
2 16
WT to
attach the for the load and to select enough bolts of the proper sire to T = -62.4
- - 3 1.2 kips
carry the load in tension. 2
Step 1. Design the angles. M = )I.*(?) + 41.92 in - kips
Assume that L / r is not critical and choose 7/8-in bolts:
Fb = 0.75Fy = 27 ksl
A, = -
62.4
22
= 2.84 in2
~t~ ~(0.875)~
= 0.1276L
A , = - 62.4
=29 2.15in2 '=6= 6

2.15 s =-
M
A&---
0.85
- 2.53 in2 Fb
41.92 =12.161n useL=12.5in
L =
at least 1/4 in thick, SO that bearing is not a problem. T~ two 0.1276(27)
1/4 (long legs back to back): ~~~i~~~~the number of bolts to can7 the hanger force. Use 7/8-in-diameter
bolts for the hanger to connect the WT to the beam:
= 2[1/4 X (7/8 + 1/8)] r 0.50 in2
62.4
N= = 2.35
= (3'38 - 0.50)0.90 = 2.60 > 2.15 required O.K, 0.601(4)
' 4 STRUCTURAL STEEL DESIOFI

side of the flange is

P = 36(1.15)'(2 x 0.3889)+(1 +
P 61.1
s) = 61.1 kips

pf -= -= 35.9 kips > 7

62.4 O.K.
=
1.70 1.7 -
The flange does not require stiffening for this load. ///
Example 8-11 A stairway hanger rod is attached to the lower flange of a
Figure WlOc 4 W410 x 59.5 floor beam. The load to be camed by the hanger rod is
Use four bolts for symmetry. Also, data for a W33 X 201: estimated to be 38.75 kN. F, = 250 MPa for both rod and beam. Design a
hanger rod and check if the beam can carry thls load without a web-to-
b, = 15.745 intf = 1.15 in g = 5.50 in flange stiffener.
Step 4. Check the "prying action" using Eq. (8-9):
SOLUTION
Design the rod:
F = - -62.4 - 15.6 kips
4
a = dedge= 1.78 > 2(0.875) use a = 2(0.875) = 1.75 in

--.' ( 70(1.75)(0.875)' + 21(6:25)(0.875)' By trial, obtain D = 22 mm (at 10 threads/25 mm), or

=: I 1.1 kips x l o T 3= 0.299 X 10-3 mi > 0.2583 0.K.
The posslble moment at the bolt is
Check the beam for the hanger force. Neglect the torsion produced by-
1 75 this hanger rod on only one side of the beam flange, since it is located very
1l.l(t)=9.71<41.92in.kips O.K.
near the end of the beam. Take the hanger rod hole at the standard gage
\ I \ -- . --L--/
distance of 88.9 mm for a W410 x 59.5 section. Also,
Since this value is only 300 lb/bolt over the allowable, take it as 0.K bf = 178 mm tw = 7.7 mm r, = 12.8 rnrn = t,
WT12 x 47 with four 7/8-in-diameter A-325 bolts to the flange . Use a
W33 X 201 beam. of the 89 - 7.7 - 1.5
b= = 39.9 mrn
2
L = 12.5 in
2L4 X 3 X 1/4 with long legs back to back
Step 5. Check the beam flange for adequacy without using stiffeners.
By Eq. (8-13) and directly incorporating L F = 1.7, we obtain
tb = 1.15 in (tnanee
- of W33 X 201) t... = 0.715
w in
- --- = -.-
5 .F in .
From which

250(12.8)'[0.389(1 + 44.5/39.9)] + ( lo-') = 19.8 kN < 38.75 N.G.

P =
1.7
Since 19.8 < 38.75 kN, the flange of the W410 x 59.5 is too thin and we
must either use a section with a thicker flange or reinforce the flange of this
section. It will be about as economical to use a section with a thicker flange.
///
 3% STRUCI'UIUL STEEL DESIGN FJOL.TED AND RIVETED COhXECTI

8-8 CONNECTIONS SUBJECTEDTO COMBINED SH

TENSION

Figure 8-18 illustrates several cases of connections subjected to

sion is adequately conservative if designed as either a friction or a beari

connection. It is evident that the compressive force on the connectiqn increa
the slip resistance, whereas a tension force tends to decrease it.

of the nominal fastener area are

f, + f, r fa,,,,
A better fit is obtained using a quadrant of a stress ellipse, since F,,,,, has
separate values for tension and shear F,, whi& give

(8-ljj

ere T,= bolt proof load as obtained from Table 8-2

J A , = nominal bolt tension force (f,= T / A d
Tfie ~ S H T O
equation for the combined stress cases are the same as AISC
some additional conservatism and simplification, giving, for friction-
' (c)
(d) ,, F; 5 13.5 - 0.221 ksi (8- 17)
Fm Connections where fasteners are subjected to c0mbio.d & a r and t e b .
Moment-resistin~ c o ~ e c t i o nusing WT and column stjffeners. (b) Type 2 ( k p l e ) or,tm 3 F: 93 - 0.221 MPa (8-17a)
(scm%d) wnnection If angle is used with clip angles, assvme &at it cames
of shear. (c)
connection using a Pair of angles back to back or a WT ( d ) Co-0~ type
for wind bracing. One may use a cable instead of the angle showd.
 3. STRUCTURAL STEEL DESIGN

'?'-..,le8-7 Coefficients for allowable fastener stresses for combined shear 1

2. ' tension cohnections

AISC
Bearing Friction AASHTO
c, c2 c
3 cqa c5
,-
a-astener f ~ q SI fps SI fps SI
.; 502 grade 1 30 207 1.3 23 158 - 0.75
grade 2 38 260 1.3 29 200 -
0.75
. i(S
A.307 26 180 1.8 20 138 - -
.". .325Nb 55 380 1.8 44 303 17.5 121 - P FigurrF.8-12

Set F, = 303 MPa and find P based on shear limitations.

sin 40
Iv= lO(O.4908) = 0.13094P

303 = 380 - 1.4(0.13094P)

" Standard holes 1/16 in (or 1/5 mm) larger than nominal bolt diameter; see the AISC
ILL ~ ~ (Sec.
u l1-6.3) for slotted or oversize holes. P = 420 kN
" N,'threads In shear plane. Find P based on maximum tension stress.
- X, threads excluded from any shear plane. P cos 40"
* For bolt diameters larger than If in (38 mm); F , = 90 ksi (620 MPa). F, = 303 =
Only static tensile loads allowed. 1O(0.4908)
P=1941>>420kN useP=420kN
For bearing-type connections AASHTO uses: Check P = 420 kN.
= 85 MPa < 150 hlPa of Table 8-1 0.K.
4
fl
wi.~:..fv and f , are the actual computed stresses and C, is a coefficient from F,' = 380 - 1.4(85) = 260 < 303 b1Pa 0.K. //I *
Table 8-7. * ,. .
The AREA currently has no provisions for fasteners in combined shear and ,:* ..
ter:sii:n. ) CONNECTIONS

Example 8-12 Given the tension-shear connection of Fig. E8-12,~whatis th o c are commonly used building connections, with
allowable load P for the W T to column connection using the AISC specif e of Figs. 8-160 and 8-180 and b used for rigid (or AISC tlpc 1) connections
:ations, F, = 250 MPa steel, and 25-mm-diameter A-325 bolts? g mechanical fasteners. Figure 8-18b (see also Fig. 8 - l j a and b) is con-
only used for both simple (type 2) and semirigid (type 3) connections. The
SOLUTIONAssume a bearing-type joint with threads included in shear plan esignations "simple" and "semirigid" for this connection are determined to
me degree by the thickness of the top clip angle. For simple connections this
F/ = C , - C2f,I C, imited to 1/4 in (or 6.3 mm), so that the an& can
From Table 8-7 (A-325N). obtain t a moment is not developed. The L V T for the
ay be designed similar to Example 8-10. The clip
angle for the connection of Fig. 8-186 requires a design for bending that is
somewhat similar to the W T of Fig. 8-180. Critical sections and assumptions for
the clip and seat angle design are shown in Fig. 8-20.
 Find the tension force to be resisted by the clip a n d e (a
= 0.6Fy = 150 MPa):
M = 0.5(0.6FV)X 1.46 = 109.5 kN . m

= - - 109.5 - 239.6 kN
T = -M
d 0.457
number of bolts in tension in the clip angle is based on F,= 3
A, = 0.4908 x lo-' m2.

= 1.6 use two bolts

We note that it is good to not have to use more than two bolts, since

I a
Distance xo = L / 2 w ~ t hsolne
des~gners
XO = N / 2 w ~ t hsome = 3.25 use four for symmetry
designers
F i N 8-24) Critical sections and dimensions for clip (top) and scat angle d e s i a for use in frarmng
beams to columns.

Example 8-13 Design a semirigid (type 3) connection to resist one-half the

capacity of a W460 X 74.4 section, as shown in Fig. E8-130. Use 25-mm-di-
ameter A-325 bolts, 5 = 250 MPa steel, and the AISC specifications.

Figure B 1 3 c

205 mm use 200 mm

long leg outstanding. The long Icg will
have to be long enough to place two bolts at 3 D + edge distance + vaIuc of

-
SOLUTIONW460 X 74.4 data:
g, from Table 1-13 of SSDD.
L > 28 + 75 + g, = 28 + 75 + 65 = 168 mm
Try L178 X 102 X 22.2 (refer to Fig. E8- 13c):
a = 50 - 22.2 = 27.8 mm .P

= 0.01643 x lo-' m'

Fb = 0.755 due to type of connection, location, and due to rectangular
Assume that web angles will be used to c a q shear so that the seat shape of cross section
angle carries only compression due to moment and the clip angle carries Fb = 0.75(250) = 187.5 MPa
only tension.
402 STRUCTURAL STEEL DESIGN

Since this 'is thickest angle in this group, go to the top of the next
ole cames one-half of
since thickness controls. We would not make L > 205 mm, b This is not :atisfactory even if we assumed the an=
extend.outside the column flange and also because the ben& shear; therefore, web shear angles are required. ///"
b i ~ o m equestionable with a very long width and only two fastener
TV L203 X 102 x 25.4 mm. ~h~ moment connection shown in Fig. 8- 1 8 may
~ be treated in one of two
a = 50 - 25.4 = 24.6 mm
0.0246
M = (239.6)- = 2.947 kN . m 1. Assuming initial tension in the bolts (which is alivays developed with
2
strength bolt connections).
S = 200(0.0254)2 = 0.02 15 10- 3 m3 2. Assuming no initial tension.
6 .."
2.947
-- = 137 < 187 MPa
f - 0.K. (and no angle th ~h~~~ two assumptions are illustrated in Figs. 8-21 and 8-22. hi@-
0.0215
A routine check for tension shows the section to be adequate. ~~t us strength bolts the assumption of initial bolt tension allows the connection to
as an elastic unit and the stresses can be calculated using the bending
see if it necessary to use a web angle for shear, since this angl
moment equation, fb = M ~ / Iwhich , is valid up until the cOnnection plates
thick. Refer to Fig. E8-13d for critical dimensions and other da int is never designed for a moment large eno"@ separate
analysis. desi@
d shown in Fig. 8-21 is adequately conservative
R = 275 k~ ry to design or a n a l ~ z ea j o i n t
.
1.4
3. ,," .'....,, .k
lL-- 0 ~'I;I:\ LIL!
j;~)t
.I/
j; = -
S ---
,.'p,lrl [t

Figure E8-13d

With 12-mm standard beam end clearance, R = v acts

ity of
203 - 12
e == 95.5 mm
2 --
Since this value is so large and because we are using the long leg out
ing to provide adequate length for two rows of bolts, let us
distance required for the beam and use that value to &termin ection with initial bolt tension assumed.
in the seat angle. Assume that the reaction is concen

-
from the end of the beam for computiilg the moment.
(N+k)t,Fb=R [Eq.(4-5)]
k = 27.8 mm tw = 9.0
(N + 27.8)(0.009)(187.5) = 275 .I[,
(.OITI~!II: 1: or bolt> t: = --
N = 135.2 mm f
col:~p!l[e I L
e = - =135.2 67.6 mm ('he.h I n r < r l ~ r l o ne ~ l u ~ [ i o n
2
e ' = e + 12 - kangle=67.6 + 12 - 38.1 =41.5mm
= Re' = 275(0.0415) = 11.412 >> 2.947 kN . m as used for top angle
 STRUCTURAL STEEL DESIGN BOLTED .&\TI RIVETED COhXECnONS &

Example 8-14 The bracket connection shown in Fig. E8-14a uses a piece o
too two bolts:
WT and two pieces of angle to make a stiffened beam seat. The fasteners
are A-325 high-strength bolts and A-36 steel. Is the connection adequate for
the shear and moment to be resisted? Use the AISC specifications.
8 2 5k Fi =
( "f) i
17.5 1 - - = 17.5 1 - -
);I
= 11.08 < 13.7 ksi N.G.

+I 1.?0"+ It will be necessary to redesign the connection using either larger bolts or * . .

more bolts. ///

1.5"

- -10 LOAD RESIS'I'ANCF: FAC'I'OK DESI(;N (1,KFD) FOR

V) ONNECTIONS
%'
Q'

a current form of the LRFD equation for connection design using A-325 and
1.5" -490 high-strength bolts (and with size limited to diameter 2 1.5 in) is
Figure ES-14a
R, = 1.1(1.1D + 1.4L)

SOLUTION Assume a friction-type joint. For 7/8-in-diameter bolts, T, = 39 is value is compared to the fastener resistance or plate-to-bolt bearing as
kips/bolt (Table 8-2).
Bearing-type connection:
R, = 0.625A,FU o = 0.70
-- - = 2-32 ksi Friction-type connection:
A 15(11.20) R, = 0.7rny A,Fu + = 1.00

' 82.5
= lO(0.601)
= 13.72 ksi
Combined shear and tension:
2
(pdhear))+ (0.6~u[iens,uni)' +(0.6"b~")'
Plate bearing on bolt:
6 = 0.75

Rn=3tdF,,,, ,,,, +=0.&

S = -bh2
= 11.20(15)~
= 420.0 in3
6 6 here Ab= nominal area of bolt
A,= bolt tension area (in thread zone) [see Eq. (5-4)]
M = 82.5(1.47 + 3.95) = 447.15 in . kips Fu= ultimate tensile strength (nominal) of bolt
m= number of slip surfaces ( = 1 for single shear lap joints)
p= coefficient of friction (commonly 0.35 for clean mill scale)
f , = f =-=-- 447'15 - 1.06 ksi
S 420 Example 8-15 Given the connection shown in Fig. E8-15, determine if the
plate thickness shown is adequate and find the number of 20-mm A-325
The resulting stress diagram is shown in Fig. E8-146. The hatched part pf bolts using a friction-type connection, and a bearing-type connection.
the M / S diagram is tension due to moment, which must be carried by the

Figure B 1 S
406 STRUCTURAL STEEL DESIGN BOLTED AND RIVETED COhTcZCTIOSS
I1
SOLUTION
e connection is "safe" with either two or four bolts. Choose th
I R, = I.l(l.10 + 1.4L) ased on the final decision as to connection type, "friction" or '
= 1.1[1.1(45) + 1.4(75)] = 170kN
Checking the plate dimensions.
A,+F, = 170 (see Sec. 5- 10)
shown in Fig. P8-1 wing 3/4-in
bolts. Set the gage so that the critical net section is a minimum. Also determine the
ess of cover plates to the nearest multiple of 1/16 in. Use the AISC spxifications an
eel. Show a neat final sketch with all critical items required. Use a bearing connection.
Answer: P = 159 kips.

does not
The A, furnished is

Find the number of bolts required in a friction-type connection (n P

tolerated, but note that after plates slip, !the bolts must still shear for
connection failure). Assume that y = 0.35 (clean mill scale on fay' Figure P 8-1
surface).
m i n e the number of 22-mm-diameter A 4 9 0 bolts needed to develop the full effscti3-eT
A, = 0.7854(0 - 9743)'
n
(fps) [Eq. (5-4)] of the tension connection shown using a pair of angles and the usset plare shown in Fig
a bolt spacing of at least 3D. Make a near sketch of the finj$ainr design rhoviag the
ired, length, and any other critical design information. Usi:the AISC specifications,
= 345 MPa steel, and a friction connection.

Assume 10 threads/25.4 mm for the 20-mm-diameter bolt.

+R, = 0.7mp4,Fu = 170 kN
m = 1 for single shear
0.70 x 1 x 0.35 X 0.24122 x 725 x N = 170 kN
N = 3.97 use four bolts
Flnd the number of bolts required in a bearing-type connection (slip

\
tolerated and joint failure by bolt shear):
L as rzq'd. -
+R, = +(0.625AbF,) N = 170 kN
Figure P8-2

Redo Rob. 8-2 using F, = 250 MPa and the AASHTO spzcificarions for the full
use two bolts pacity of the angles. Use A-325 bolts and a friction connection.
Answer: 18 bolts, L = 610 rnm just under L for no reduction.
Check the plate bearing using two bolts: DO rob. 8-2 for the fps equivalent of the pair of angles and a 3/5-in _wset plate for an axid
oad of 240 kips, A-36 steel, and A-325 bolts in a bearing connection.
R, = 3tdF, 4 = 0.64 F, = 400 MPa Answer: Eight 7/8-in-diameter bolts.

qRn = 0.64(3)(0-012 x 20)(400)(2) = 368.6 kN >> 170 kN . O.K. 5 Given the beam splice shown in Fig. P8-5, use the AASHTO specifications to (a) W i m b o l ~
d w v e r plates for the full moment capacity of the beam. @) D e s i p b l o : and web plzta for the
I I I
 & m

3Ol.TED .\.\TIRIVETED COh3ECTIOh.S Jpjg

iull web shear capacity of the beam. Use 7/8-in-diameter A-325 bolts for all splice parts, and A-3
steel.
Parr~alanswer. M = 1950 In . laps, T = 107 laps. are adequate for tearing but that the bearing should be checked as appropriate. .UI boIt hes
I angles are on standard gage distances, as in Table 1- 13 of SSDD.

ikfj+~f*\
Figure P8-5

8-6 Design the eccentrically loaded bracket connection for the load shown in Fig. P8
25-mm-diameter A-325 bolts in a friction connection. Determine the plate thickness for both
and tear along the forward row of bolts.

P = 225 kN
9 Redo Example 8-14 for twelve 7/8-in-diameter ,A-325 bolts.

1 Redesign the connection of Example 8-8 if the floor beam on the left docs not frame into the
8 x 50 section.

Answer: Yes.

8-7 Determine the number and placlng of 22-mm-d~ameterA-325 bolts for a cable connection to a
W360 x 314 column as shown In Fig. P8-7. Assume any needed T d ~ m e n s ~ oton produce L that is
'~va~lable.
Check the tenslon in the stem of the T so that a large enough sectlon is used. Assume that
the hole for the cable attachment ulll be reinforced so that capacity 1s not limted at that point. Use Answer: P, = 217 kips, N = six f -in bolts.
the AISC specif~cat~ons and F, = 250 MPa for the T section. 7 Do Prob. 8-2 using LRFD.
Answer. 10 bolts. 18 Do Prob. 8-5 using LRFD. Assume that the 225-k?J load is mads up of D = ICO kN acd
= 125 kN. Check only for number of bolts.
Answer: 16 bolts with I percent overstress.

Figure P8-7
approach sawed cuts in smoothness.
As stated earlier, most welding uses an electric current. The current is used
o heat a n electrode to a liquid state, which is then deposited as a filler along b e
nterface of the two or more pieces of metal being joined. The process simuIta-
eously melts a portion of the base metal (metal being joined) at the interface so
hat the electrode intermixes with the base metal and develops continuity of
411
 412 STRUCTURAL STEEL DESIGN

Electrode covering
material at the joint when cooling takes place. If the qua
electrode 'is small relative to the thickness of the joined parts, the proce Weld f ~ l l e rrneral

to be unreliable (i.e., insufficient melting of the base metal occurs so Shielding atmosphere -
weld may pop off:or not fully join). This event can be
preheating the base metal or limiting the minimum size of the weld.
welding operation takes place in a very cold environment, it may be ne
preheat the parts, particularly where the parts are thick so
temperature differential does not develop in a short dista
resulting thermal stresses are so high that the weld zone fails.
Electric welding involves passing either dc or ac current through an
:rode. By holding the electrode 'a very short distance fro
is connected to one side of the circuit, an arc forms as the
"shorted." With this "shorting" of thencircuit, a very large current flow t
place which melts the electrode tip (at the arc) and the base metal in the vici
of the arc. The electron flow making the circuit "carries"
metal to the base metal to build up the joint. Careful con
and current are necessary to produce a .quality weld wi Molten weld pool
define an adequate melt zone and while keeping electrode s
mum.
The electrode may be either the anode (+) side of the circ
( - ) side. Most commonly, the electrode is the anode and t
is conducted using "reversed polarity." When the weld electrode is the cath
(-), the circuit uses straight polarity. Most welding is done using dc current
ac is used as the power supply, it is first transformed to dc.
Of the numerous welding processes available, the foll
to be used for structural applications:
1. Shielded metal arc welding (SMA W). This is the most common welding
method using stick electrodes. The electrodes are available in leng!hs of 9 to 18
Ih )
in and are coated with a material that produces an inert gas and slag when the
welding current melts the metal. This gas surrounds the weld zone to prevent
,jxidation (see Fig. 9-la), which is a critical factor if more than one weld pass is
Iiecessary to build the weld to the required size. The slag, being lighter than the
metal, floats to the top of the weld and can be brushed away. On subsequent develop the gas shield and to obtain any s:Fcctj

passes it is necessary to brush the earlier passes to remove any slag, dirt, or other
foreign material whose presence might cause a flaw in t
welding is the most commonly used field method using
The maximum size of weld produced in one pass is about 5 / 16 in or 8
2. Gas-shielded metal arc welding (GMA W). This method of.weldi
often used in shop welding, where uncoated electrodes ar
welding unit. The unit controls electrode spacing and we
inert gas source to shield the weld against the surrounding atmosphere.
3. Submerged arc welding (SAW). This method of welding is also used in
fabrication shops. The joint is aligned and covered with a blanket of granular
fusible material containing alloying and fluxing agents as well as inert gas,
producers. The electrode is inserted into the granular material, the arc produced,
and the melting of electrode and base metal takes place. The heat fuses the
n e slag is later brushed away to expose the
g.This welding process is vew similar to the submerged
an electrocondu~ti~e slag which is held in position between
etal to be joined by water-cooled retaining plates (see Fig-
rial is melted and current passed throu& it to maintain the
and filler metal. The filler is obtained from the
ed into the slag. The process is generally done in
the filler melts the retaining plates are slowly
rnpleted, partly cooled weld. which has a thin 'la=u cover in^
must be brushed away.
~ l welding is~ used to shop-weld
~ thick
~ plates ~together* I~t ha\ been~
er o p u l a r in bridge work to weld girder plates and floor plates-
the order of 20 to 450 mm can be welded by this process in one pass-
9-3.1 Allowable Weld Stresses I

II
The butt joint is the only joint likely to be ill direct tension. The a110
tension or compression stresses for weld metal is given in Table 9-
allowable stresses in tension or com~ressionfor the weld metnl mnv a

into this by further limiting the F, of the base metal to 42 ksi for E60 elec
and'to 55 ksi for E70 electrodes in structural grade steel.
The allowable shear stress for fillet welds is limited to
Fo = Oa3Fu(electrode)

in the AISC specifications, but it is always necessary to check that th

sufficient base metal to resist the same shear stresses. Tn penern1 t h e ~ P I A

checked against the specif;cations being used.

The max~mumshear stress in the weld metal is the limiting v
:onnectlons where the weld is subjected to combined shear and

'1-3.2Fillet Welds I
The fillet weld shown in Fig. 9-4 is approximately triangular in crdss section.
Care must be taken that the throat dimension shown in Fig. 9-4c & built ou
adequately. In most cases the legs of the weld D are made eoual hiid t h i ~i s nnt
* ..-2
,A
m cj
E E
computed as
I
%
8
T = D X cos 45" =
I
D ~"0.70711
L
0 -
L
0 .2

:32 u3
where T= throat dimension
$*,a%g %
.-
+ m m m
D = nominal leg dimension
The leg dimension for fillet welds D should be taken to the nearest 1/16 in or 1

Figure 9-4 Critical shear area for fillet welds. (a) Fillet weld for tee joint. (b) Fillet weld for lap joint.
!'c) Throat dimension for minimum shear area.
 -.xu alnuLr UKAL. >[EEL DESIGN

9-33 Plug and Slot Welds

is because the quantity of molten weld metal is much smaller and the resultin < ,'
shrinkage is much less. >
t ~ ;~ r 7

In1

9-3.4 Minimum Weld Size

(b) 0 welds.
for plug and slot welds. (0) Slot
9-5 AISC

9-3.5 Maximum Weld Size

The maximum size of fillet welds along the edges of connected parts is:
0.707 1 1D(0,
F,) 5 Pz F ~ )
Part here
[,,,I(

p, =0.3, 0.27, etc. from the appropriate code for ""' rod

SI, mm' AISC and AASHTO

p,= 0.33, 0.4, etc. from appropriate code for the

( 5 1/4 t i 6 Use D = thickness of part r

f>1/4 t>6 Use D = t - 1/16 in or r - 1 ma

Unless the throat dimension is specifically built out to use the

full value of r.
~ ., ,,, .,.. ,. ..
It is possible in the fabrication shop when using the submerged arc welding
process to specifically produce a fillet throat thickness by rounding the wela out
so that for
D 5 3/8 in use throat = D x 0.7071 1 (in)
b
D > 3/8 in use throat = D x 0.7071 1 + 0.1 1 (in)

If this is contemplated, care must be taken to differentiate between the shop

 ---a -.-"<<,A,

1v 5 8"
or betweet,

A I S ( ' S c c . I . I 7.4
Flat bar ro gusset plate

Improieil loin[ p l l l l llc)l 'i17rl11~l

Weld both
I O rollins .I\!,

Ii.1

at a welded joint. Also shown 1s one solution lo the

reduce
of a lamella tear occufing. ( a ) Lamella tear in base metal. (6) Joint %eomstV

A[SC Set. 1. I 7 . 6
Lap joints
~fa large weld (or welds from both sides) is imposed on a thick piece b2se
a lamella tear can occur. The tear can occur because the shrinkage strains
om [he welding operation will be large and restrained. T h e restraint may be
he far side or from the member rhickness Or a
tch of electrode and base metal in a f u l l - ~ e n e ~ ~ ~ -
Figure 96 AISC welded connection specifications. sine
Stnrcturai butt weld tends to increase the possibility of tearing (i.e.. from using an
Code, they should be generally followed for with A-36 base metal). A thin, stiffened

Other AISC specifications conc to the column flange produce restraint.

building consfru~tionare shown in ~i~ 9-6.
9-4 L M E L L A TEARING
'eaMg tsee Fig. 9-7) is a phenomenon that may occur in certain welded
Joi'ts It is not a common condition because it involves several
use of fillet welds, a joint design that allows strain relief or loading
g direction, and the order of welding to minimize
methods used to avoid lamella tearing- The AISC
ly Restrained Welded Connections" (AISC En-
ineering Journal, vol. 3, 1973) discusses lamella tearing in some detail and gives
a number joint alternatives that may be used to reduce lamella

must be large relative strains in the base metal (lamella 5 ORIENTATION OF WELDS
not does
in the weld metal). These straihs occur where large
stre:& occur. localked aboratory tests on small to medium-size joints show that butt welds
. 2:: Lohc:;
is generally perpendicular to the mill rolling limit joint capacity where the electrode has been "matched" to the base
the '"lnber being welded. Beams welded (0 direction that produced ~h~ orientation of the applied stresses does not have a si~dificanteffect
loading in the column flanges p u t not in flanges produce this type
the beam flange). butt joint strength.
3. ~h~ orientation of stresses for fillet welds is a significant factor
:nust be strain restraint in the base metal,
ultimate joint strength. Tests (see Butler and Kulak. "Strength of Filiet
 ,422 STRUCWRAL STEEL.DESIGN wioU, C O ~ E C X ' I O N S

joint shown in Fig- E9-l a wing

Pa, and the AlSC specifications-

(?
. .. jr'
d

Figure E9- 1a

Joint efficiency = 100 percent

p = 0.015(150)(0.6 F")= 337.5 kN
9-6 - WELDED CONNECTIONS
Use D = 15 - 2 . 0 13
~ mm.
F, = 0.3 x F, = 0.3(415) = 124.5 MPa

se a 150-mm weld on each side of joint as s h o r n in Fig-

Set. 1-17*7-
satisfies AISC set. 1-17.4). Use 26-mm end returns Per

allowable shear strength of the weld.

It be noted that a butt weld used to resist a moment develops stresse
Figure E9- 1b

eld and 75 mm along side'?

ample 9-2 ~~~i~~ the welds for connecting an L4 3i a 3/gin

of f b is compared to the allowable weld shear stress. 20. Use the AISC specifications7 'I8
r static loading and dbnamic loading-

A,,@, = 1.8 1 in'

Figure 9-8 Welded Imment connections. ( a ) Butt-welded moment come
connection. p = 1.81(22) = 39.8 kips
'- - ---- "CI'U-L. .,,UO,. Ufi,.,,UN

Substituting L, = 10.5 - L , into Eq. ( a ) yields

,. "A,.-.,. .. 0.707 1 1 0(0.3 Fu)5 to( 0 . 4 ~ ) heck: 6.62 + 3.89 + 4 = 14.5 in.
0.25(0.4 x 36) - 3.6
D 5 --
0.70711X21 14.84 .1 Rigid Beam-Column Connections

Use D = 3/16 in.

= 39.8
(0.707 1 1 X 0.1875)(21)
I
Use the weld shown in Fig. ~ 9 - 2 b .

Figure E9-26
Figure E9-2c
For loading it is necessary to balance the weld
neutral axis of the angle (AISC, s
Referring to Fig. E 9 - 2 ~and placing the weld across
reduce the joint length, we have

L1 + L2 + 4 = 14.5
Ll + L2 = 10.5
L l = 10.5 - L,

Take the sum of moments about t

eliminated; also,
Pw = 0.707 1 1 X 0.1875 x 21 = 2.78 kips/in

L2(2.78)(4 - 1.16) + 4(2.78

Canceling Pw, we obtain
2.84L2 - 1.16Ll = 3.36
Substituting L l = 10.5 - L2, we obtain
2.84L2 - 1.16(10.5 - L2) = 3-36
L2 = 3.89 in
 ewlin, "Column Web Strength in Beam-to-Column Connections,"
0 ves
tructural Division, ASCE, ST9, September 1973) 3
4100t;fi
PC, = J
,
"c

we equate PC, to the beam flange compressive force (Pbf = A,&) and Iook
he dimension d,, we obtain the current AISC equation [Eq. (I.15-2)l:

4 100r2\/1;
dc < (fps)
Pbf
7

10.73t:\jF,
dc < (SO (9- 1m)
Pb/
ere Pb,= beam flange force (compressive) x F
F= 5/3 for dead and live and 4/3 for dead + live + wind, kips or kN
t,= column web thickn~shan 0,
d, = required colum~'.w,&'#h'i~$?@~s as d - 2k, in or mm
c= tTpTi:r
yield stresk of kglur~fiSee 9: sl. or MPa
. *
,Q
ner is required oppofite the"co+ression flange if the actual column as
reviously defined is greater than that given by the right side of Eq. (9-1).
Stiffeners (for the column web) are required opposite the beam
ge (see Figs. 9-lob and 9- 15a) as follows:
be concerned with determination of whether these column web and fl
stiffeners are required. this inequality is satisfied as shown (right side equal or larger), no c o l u m
is required. A more convenient method to determine the stiffener
ents is to equate
p,, + pm = b'f
ith P,, = A,,F ,, we obtain
Pb, - ~,,t,(tb/ + 5kc)
'A,, = (9-2)
F,,,
The earlier AISC specifications rounded 183 to 180. Th AISC equation for column stiffeners opposite the beam
that only a positive stiffener area is valid.
The column flange must be of sufficient thickness to resist the beam flange
substitution of values t excessive deformation. A yield line analysis his given the

;+ ,"here the additional term (~,/36)"' is used to adjust for other grades of steel.
A forther adjustment to the factor 33 400 to incorporate test results (see Chen
 ,
I
, #

i-'
"
.
'.a""r,'.',glq:
. .
'
.W
~ U C I T J R N . STEEI;
DESIGN ~ wELUE; CONNECI~&$~
( :.< j ~. ;. , .: ..,'*"
.
. . . . '

I' the w1umn flange thickness tcf is less than that given on the right sides of
eqyations above, flange stiffeners are required. neMSC sp&fications req
column web or flange stiffeners meet the following criteria:
I
*a,,)! "
, . S f . - Eq. (9-1) if this equation is applicable.

2. Width of both siiffeners + t, 2 0.67%.

>'/+A,

"
. -' Stiffener thickness 1, 2 $3/2 (also, the b/t ratio must be satisfied).

pzr beams on one side of the column, the stiffeners may extend onlyone-h
::i the column depth.
." &'hew e l d j o c n g the stiffeners to the column web must be sked to carry
i-~balancedmoments on each side of the celumn. compute the plate length. Note that the plate will have to be long
6 siiffeners for tension requirements must be welded to the column fla enough to allow placing 5 in of weld at D = 0.5 in. I t will also have
some length between the end of the beam welds and the butt the
dfficient to carry A,tF,, (i.e., use full penetration butt welds).
:ompression stzfeners must be welded or accurately fitted to the ,
column in order to develop adequate strain. BY proportion:
opposite the beam flange delivering the compression load.

% m ~ l e9-3 Design the moment connection shown in Fig. ~ 9 - j usi

u
the AISC specifications, and A-36 steel.
e a full-penetration butt weld using the

-3c), use a plate 12 in deep on one side.

.,

, , , .. ., . ,.. ... F
1 :
b E9-3a
b o ~ u m oDesigning
~ for full moment capacity:

M = FbS, = 24 X 64.7 = 1552.8 in . laps Figure E9-3c

I '

26.4 = 0.1 j use 3/ 16-in plate

T = C=-= -= t =
'd 16.01 p 12x0.4FY
" .. 26.4
',Se a fop plate that is 8 in wide at the column weld and tapered to 3 in, Weld D = 0.70711 X 12 X 21 =0.14gin use3/16in
zhown in Fig. E9-36.
check if the column needs reinforcing (AISC Set. '-I5.'):
97
'P = --- = 0.73 use 3/4in plate
22 X 6
9 -,TakeD for fillet welds at 1/2 in.
97( 5/3) - 36(0.295)[0.505 + 5 ( 1.2j)] - 161.7 - 71-74 = 2.49 b 2
P, = 0.7071 1(0.5)(21) = 7.42 kips/in 36
Therefore, a pair of stiffener plates are required opposite the tension flange-
& = - =97 13 in and distribute as shown
7.42 u s e plates 3 in wide (2 x 3 > 0.67 x 8 o.K.).
' t430 STRUCTURAL4 STEEL DESIGN

Neglect ttus zone

Check the compression flange:

1 <,r. line = L ; 3 t ~ b o u bast)

t
/ o r line = [.' I l iccntrord)

9-11Eccentrically loaded beam framing an!&s also Fig. 9-IZd).

Figure E93d

to carv97 kips and across the ends to the flange adjacent

The stiffener may be only one-half of the column depth, since

the eccentricit4- e,

(16.91 X 2)(0.70711 x 21)

1 Use a f -in on both top and bottom of each stiffener plate.

angle is welded to the beam web and field-bolted to the colu- flange for
der web), the 3-in leg should be welded to the beam and bolts in the 44x3 Q-sJ-
that there will be adequate erecti~nclearance and edge distance-

~~~~~l~ 9-4 Design the web ande connections for a W18 x 55 as in

~ i ~ g~- 4 .afor a framed beam-to-column flange connection cqi
% = (R:oment + R : ~ ~ ~ ~ ) ~ ' ~ a 55-kip end reaction. u s e two ~3 x 3 x t in x 12 in long. Use the
specifications, E70 electrodes, and A-36 steel.
where
4 , a:
= lorceor stress/unit df length (or area). The value of 4 is lifited to o ~ u n o NFor the beam web weld, locate neutral aGs as
. ...
-< Rallowable
:.:\ 8 .i
2.5
(12 +2 -
x 2.5)s = 2 C 2 . 5 ) ~
-
.X = 0.367 in
 4- - L K U C T U WSTEEL DESIGN

4 Figure E9-40

'~nsideronly one side, so

R = - - = -55=
27.5 kips
2 2 /'
/~ / 3 ) :
: 'le polar moment of inertia is (I of line about base = ' ~ 3.11
' a = - - 14.4
- 0.257 in =D O.K.
LP = Zx + 1
, =- I 23 + 2(2.5)(6)2 + 12(0.367)' + 2(0.367~+ 2.133~)
12 3 Use L3 x 3 x 5/16 x 12 in long. f
= 144 + 180 + 1.62 + 6.49 = 332.1 in4/width of weld
C
7 he weld shear resistance .2 Welded Beam Seat Angles
& = -R- 27.5
e beam seat angle shown in Fig. 9-12 must be designed for bending stability
, r , - Y = 1.617 kips/in must be of sufficient thickness and leg length that an adequate fillet weld
ioint 1 of the weld is critical by inspection (or drawing rays from c.g.)
M = Re = 27.5(3.0 - 0.367) = 72.4 in . kips be placed along the vertical legs to carry the shear and.moment due to the
ccentricity of the reaction. The angle is checked for bsnding at the fillet runout
k distance from tables) as in Fig. 9-12a. The allowable bending stress is taken
+* 5&&, ,

:ompu t.ing the la1 bending str

4. = ((0.465 + +
1.617)~ 1.308~)"~
= (6.b45)'I2 = 2.46 kips/in
odulus S = b the required
D= 2.46
0.7071 1 X 0.3 X 70 = 0.165 in use 3/ 16-in weld
Check the web shear capacity:
2(3/16 X 0.7071 1)(21) 5 0.39(0.4 X 36)
For the column flange weld (assume 1/4 5 r, 5 3/4: why??) ere b = seat angle width. The value of r is obtained by trial. We may initialIy
timate the eccentricity (assuming the k distance for any angle is t + 0.375 in
Rh = -54Pe,
25 L~
(from Fig. 9-1 1, refer also to Fig E9-46) t+ 9 mm) to obtain
N
.. =-
e.1n111al
2
+ 0.5 - ( t + 0.375) in

I
b N
, , R,=-=-=
R . . . = - 2+ 1 2 - ( t + 9 )
einlt~al rnrn
L 27'5
12
2.29 kips/in (and neglecting end returns of 2 0 )
actual required angle thickness t is very sensitive to the k value, so the
ative angle should always be accurately checked.
- .
D = 3.38
0.70711 x 21 = 0.227 in use 1/4 in weld Example 9-5 Design a beam seat angle and weld for the conditions shown
in Fig. E9-5a. Use F, = 250 MPa and E70 electrodes. Data for a W4lO X
 STRUCTURAL STEEL
WELD W COWECTI

Use N = 120 mm:

-N= - -120 - 6 0 m m
2 2
Use the of the WT at least as large as f+ the
Also,

- < - = 335 21.2'

Figure E 9 d b
t - m
an inspection of Table V- 18 (S Shear Stress due to load:
b, = 395.4 mm r/ = 30.2 mm

Find the maximum weld size D for s M 15.16 0.842

on each side. L=~=o.lsoo,=~,

loot,
= = 9.19D
205.8 > 8 as minimum for r/
Check the bending stress in the WT

6M
-= 0.3 x 485 = 145.5
fb=-- - 6(15.16)
= 132.9 MPa < 150
twd2 18.9(0.19022) 0.~.

-.
.,,. 1 ,.. ...~.,.
/li
Use D = 9 mm.
y=--=
50 040 57.5 mm
870
WELDED COLUMN BASE PLATES

ob site. Several situations using CO~UIIIII end plates are sho"n in Fig* '-I4.
base plate may be either butt- or fillet-welded to the column- The decision is
= 2.4733 D, x 10-6 m4

quires additional fabrication. The general base plate d e s i 9 for dimensions

width x length x thickness) is as outlined in Set. 6-6.
 -
r I

W STRUCTURAL STEEL DESIGN

WELDED C

Since d,, furnished = 538.2 > 332.2 required, a web stiffener is rtq&&
~hi'stiffener(a pair with one on -each side of the web op&site the
Beam: W610 X 241.1
column compression flange) only has to be one-half of h a m depth, since
d = 635 m m , b, = 329 m m
l / = 3 1 mm,r,= 17.9mm the load is only on one side.
tc/ 33.3
Is,," > T = -7- = 17 m m use tJt = 20 m m

Column: W360 X 261.9 (0.67 X 329) + 17.9 = 1 19.2 rnrn use b,, = 120 mm
d Z 3 8 7 mm, b f = 3 9 8 mm bst,ff = 2

Step 2. Check i f a d~agonalst~ffener1s required.

Use plastic design [Eq. (9-7)]:
M, = 1.7 x IM = 1.7 x 610 = 1037 kN . m
SOLUTION
Step I. Design vertical stiffeners 1 and 2.
For stiffener I, opposite the tension flange of the column:
: f b ' - - M ---
610
- 132.6 MPa
S, 4.60 Since the furnished t,, = 17.9 < 31.6 mm required, it is necessary to use a
diagonal stiffener.
AfC = 6,$ = 0.398 X 33.3 = 13.2534 x 10- 635
/ j(132.6 X 13.2534) = 2929 kN
Pw = ; f b ~ =
8 -- tan-' -
387
= 58.64' cos 8 = 0.52042

'bf - (
G c tcw tbf = k,
As, 2
F,,
Use the beam for the "column" dimension and 2.5 instead of 5k, sin
column tension flange is at the end and not centered on the 5k zone

As, =
2929 - 250 X 0.0179(33.3 - 2.5 x 48.4) = 6.2615 x l o p 3m2
250
= 10.146 X m2 Use two plates 20 mm x 160 mm wide:
Use end plate:
A, = 6.4 x low3m 9 . K .
b, = 398 mm (width same as column)
t, = 28mm Step 3. Design the welds.
The weld for plate 1:
A , = 0.028 X 398 = 1 1.14 x lo-' m2
0.K. PbJ = 2929 kN F, = 0.3 Fu x 1.7 = 247 MPa (plastic design)
For stiffener 2, opposite the compression flange of the column:
Using inside fillet welds (see Fig. E9-9b), we obtain D,, = 8 mm, since
, , ,..
db(,+,.s, = d - 2k = 635 - 2(48.4) = 538.2 mm
.,.,..,...
t,= 25 > 20 mrn. Check the effective D of the beam web for shear:
A stiffener is required if bb > dbw:

10,73t & f i
dbw >
pbf . .
10.73(17.9)~a
i > = 332.2 mm
So only 7.4 mm of the weld on the beam web is effective, since the shszr of
2929 the base metal controls. Use D along the inside flange as required and an
 WELDED C O h W C

Ssuming that 600 mm is effective, check P:

= 2(600)(0.0074)(0.70711)(247) = 1550.9 kN >> 600
pWe,

Use flange D = 14 mm.

esign using LRFD is similar to load resistance factor- D
ctions. ~t is necessary to compute the.factored load
R, = l.l(l.10 + 1.4L)

R, = 00.6 FEmA,

+= 0.80 [currently (197811

I\'(J/~.
All plates but I ore 20 mm thick. ~ l fillet
l
are D = 8 lnln except b e a m - t o - ~ o e]n~d ~plate,
~

Figure E9-96

The for plate 2: with maximum weld D = 7.40 nM in

web and =
required for 20-mm stiffener plates and four lengths

= 250(0.0179)[33.3 + 2(20 + 48.4)] = 761.2 kN

Pweld "- P b f - PWeb
= 2929 - 761.2 = 2167.8 kN , allowable for the base metal.
r: p = 8 1 kips, L, = 12 in, using D = 9 / 16.
4L x 0.0074 x 0.707 11 X 247 = 2 167.8 kN

Weld for diagonal stiffener:

 , .
WELDED C O W E
.
i 9 ,
,,

W410 84.8 beams frame into a W530 x 123.5 girder as in Fig. P9-19 in
trcdes, Fy = 250 and

..
9
1:
'sign the welds and gusset plate for a pair of L152 x 102 x 19 - for
ctive angle capacity in static tension. Use F, = 250 MPa, Ej'0 electr
pi.,. . : , : a t i ~ t ~
Keep
. the joint length to a minimum,
9. - 30Prob. 9-9 for a dynamic load. . ,
9. --Design a welded framed simple bea+ connection to ca
22 ).: 50 beam to a W12 x 53 column. Use
ai.2 Li X 3 X t angles with a length of 12 in.
dnnuer:tP5/16in;Dwcb=3/16;Dcol=1/4. ,

9-.: : Design the framing angles and weld for

hear due to uniform load is

steel, A-325 bolts, and E70
A ' , YO specifications.
.~.moer:1, = 5/16; D = 1/4 in.
9 , 9": Design the welded end connec

9 - f . ~Design a framed beam c o ~ e c t i o nto carry a beam shear of 152 k~ from a

U s e a pair of angles 76 X 76 x t with L
and E70 electrodes.
9-15 Design the weld and plate for
Answer: D = 3/16 in; r = 1/2 in.

20

Figure P9-15

simple :Pan carrying a uniform load of 1

(. . neat ':
;. design sketch. Use E70 electr
fram. ::-;a W12 x 53 column.
9-17 J-ai@ a stiffened b e ~ mseat using a WT fo

.: . ,
 , Built-up rolled sections may be used where the overall depth is limited in

section is to use two lighter rolled shapes in parallel. Even where the to& weight

r the built-up section.

g 10-10 is by far the most common use of girders in building coqtmction.

Plate girders are commonly fabricated by welding two flmge plates to a we1
e as illustrated in line details in Fig. 102. The most common materids i r l
 ."+
. i

PLATE GIRDERS 4g ,,

,, ..,

( C)

epth D and clear distance h-

atio h,rb is a significant parameter in plate girder design. (a) Welded plate girder- ( b )
).,( Girder elevation (side view) illusirating other in drsim-
plate

the reduction7 or
and larger clear web

rn.ical for spans UP

15 f t ( 5 m) Or more-
Pan (D/Q ratio, Ge
of A to A-Larger

n exceeds 50 to 00 ft
for spans up to 300
00 m) or more. Roued beams are generally more economical for bridge
less than 50 to 60 ft and are used in a deck stringer confisration-
ough girders are generally more economical than trusses, the latter are sm
in many situations for esthetics, particularly where additiond lanes
ired and the existing facility is a truss-
 I
I
460 STRUCTURAL STEEL DESIGN

However, the current AREA specifications do not s [ ~ i r ~ Jt

. lr
(t o ri I J I I C
Jl c I o ~ I J\$kith
I~~
L~~~ = .![her I O J ~or j r ~ n d ~ truck
r d
rather, the computed deflection of the girdei under
including impact, is limited to A/span 5 1/640.
;Girders are fabricated' in segments limited by the
ere4tion/transport equipment and rolling mill capaci
plate girders are almost exclusively used in American
the reduced fabrication costs and the fact that the we1
the web can be almost fully automated.
A-36 steel is most commonly used for plate gird
for webs and stiffeners. .Continuous gird
sections in the positive moment region and d
moments regions over reactions. An increa
esthetically pleasing; however, fabrication costs are increased, since
web must be cut and additional stiffeners are usually required,for the $1
rati9:prod~ced.~ An alterriative to an incre
strength steels in the zones of increased moment.
for the flange either throughout the span or in
produces a hybrid plate girder. because of truck length factors and positioning of wheels. When the 'pan is
An inspection of Figs. 10-1 to 10-3 long, the lane loading with single concentrated load a line
sel-vice readily indicates that there is n
of trucks enough and is more convenient to use than tvine to find lhc
problem. Within the constraints of D / L
possibilities exist for cross-sectional area. If we define optimization o f a pl number of sucks and their positioning for maximum stress
girder design as that producing the least overal
ilsing a computer program, to "optimize" the gir
necessarily (and probably will not) mean least ov
considerations. What one should attempt to obt
computer program. me
and fabrication cost to produce as economical ( ln all girder design the standard Iane/tmck loads are incrcased
effort, which is also a cost factor) a solution as possible. ,impact factor, which is a function of span length.
Figure 10-5 illustrates the placing of the standard truck loading On
two-lane bridge to obtain the contribution of truck loads to either of ths
20-2 LOADS girders. The designer must make a similar type of 3 n n l ~ for
~i~ lin2'
using more than two girders.
plate girder loads are obtained similarly to those for be
building design the loads are often moving, as for crane Example 10-1 what are the design moment and shear values for the *d'r!
industrial buildings or warehouses. AISC specif of a highway bridge as shown in Fig. 10-5?
factor for these types of loadings. \ , HS 20 loading, s = 19 it, w = 28 ft, span = 110 f t - D a d load due "
Plate girder design for highway bridges involves a deck, sidewalk, and so on = 3.1 kips/ft.
girdqrs are to be used per lane, using the AAS
amount o f truck or lane loading to the s
influence-line type analysis using either stan
S o L u n o ~nelane factor L, is obtained by 9"1 about the left aderFig
lpading to find 'the maximum moment and
X+ Y
specifications allow use of only one truck per lane for simple spans, in general, Lf = -
S
although as shown in Fig. 10-5, more than one truck may be on: the bridge at the
critical location at a given instant. When more than one truck is on a lane, the where X = S / 2 - 5 = 19/2 - 5 = 4.5 f t
y= x + ~ / =2 4.5 + 2812 = 18.5 f t
, rWrL STRUCTURAL STEEL DESIGN
PLATE G

Substituting values, we obtain

ue to the moving train either at the centerline or very close to it.
made based on a combination of use of d M / d x = 0 for short sp& I
where only three or four wheels are simultaneously on the span and an influ
The impact factor is computed (see Sec. 1-10) as ine with apex at the centerline for a larger number of wheels on the span.
50 given in Table 1-2 are the end reactions (shear) and the maximum reactio
- 50
L
I =
+ 125 110 + 125 = 0.21 < 0.30 O.K. (shear) developed in a floor beam into which two stringers of lengths s h o \ v ~i ~
the first column frame. [For example, for a floor beam betwee
The moments are stringers, the maximum reaction is obtained when one 10-kip wheel load is on
the floor beam and the other two wheels are 5 it away, giving R = 10 i
2(2/7)10 = 15.7 kips; similarly, for I! = 10 ft, thc reaction R = 10 + 2(5/10)IO
= 20.0 kips; etc.] The values in the table are shown for Cooper's E-80 loading;
Use the equations given in Sec. 1-9 for L = 110 < 145.6 ft.
any other loading X is X/80 x table value.

Substituting values yields - - - PROPORTIONING

10-3 - - -
- FLANGES AND WEBS OF GIRDERS
, . -.. BUILT-UP SECTIONS

The flange cover plate of a built-up section usmg a rolled section as the bas
40 may be proportioned as follows (refer to Fig. 10-6). Assurmng that the maxim
= ~ ( 5 9 1 6 . 8- 1232.0) = 1703.6 ft kips - -
110 bending stress fb is the allowable value of Fb,the average flange value is abo
The shear values are 0.95Fb. At the junction of the cover plate and the rolled section flange the stres
s about 0.90Fb. Therefore, let
2
-
Vd = W L = 3.1(110) = 170.5 kips
2 M, = = O.90Fb.Sr
'i'he live-load shear (one 32 k p wheel on reaction) at reaction is M , = afAC = (d + t;)(t;bj x 0.95Fb)
The required section capacity is
lvftou, = lbf, 4- lbf2 (c)
Substitution of Eqs. (a) and ( b ) into (c) and taking (d + C) = (d + q) definin'g
We note that Md is at the centerline of the span, whereas ML is 2.33 ft the flange area A , = bit;, and taking the ratio 0.90/0.95 as approximately 1, we
to the left of the centerline (truck moving left to right). The difference is so
small that we will simply add the two values as if they both occur at the
centerline to obtain a slightly conservative design M :
M,,,,, = ML(Lf) + impact + Md
= 1703.6(1.21)(1 + 0.21) + 4689 = 7183 ft . kips
The design shear is
= VL(Lf)+ impact
Vdeslgn + Vd
= 65.9(1.21)(1.21) + 170.5 = 267 kips
///
Railroad girders are more difficult to design, since there are 18 wheels in the
Cooper E-series standard loading. Design aids are available, such as those
shown in Table 1-2, which gives the approximate (to exact) maximum moments
Figure 1 0 4 Bending stress distribution on a built-up beam cross sccrlon
 PLATE GIRDERS 465

We could use flange cover plates 25 X 548 mm; however, let us arbitrarily
Mtota~ - sx
( d + C)0.95Fb d + C
Af ZX - use a plate width 52 mm larger than the bf of the beam, which allows a
.. 26-mm overhang on each side for welding the plate to the flange. This gives
17 5" )u!d be evident that Eq. (10-1) does not give a uniqr- .. - --'--a:-
b; = 418 + 52 = 470 mm
r b ovluLlun a n a the final
f!$P.:;- cover plates must be investigated using f, = M ,/ , q. m A reasonabl ,
e$\iliit' of C = 1 in or 25 -.
t;=-=13'82 29.4 rnm use 30 mm
0.47
' , + ~ m p l e10-2 Design the cover plates for a W92O x 34" ^

,Ii ,-250 MPa for a moment of 3630 kN . m L- &'-

Now check if cover plates 30 x 470 mm are adequate:
irle conditions shown in
: EIO-20. Use the AISC specifications.
r:;t
~ U I
I, = I, +2~d'
5.. c'JTION Data for a W920 x 342.3:
wt = 3.36 kN/in d = 912 rnm
12 499.4 x
= m4
The section modulus is

The additional moment due to the flange cover plates is proportional to the
,'noments are: flange area:

M L - . . .wL2
-= 72.6(20)'
8 8 = 3630 kN .m
<
Md "
3.36(20)~ f b = 2 5 .3855
719= 149.9 MPa 150 O.K.

8 = 17OkN.b
?'he cover plates will be welded as shown in Fig. E10-2b. This requires
Mdes~gn =3800 kN . m checking by/$ for acompliance with AISC Sec. 1-9.2.2 (stiffened edge
Te~t~-:l"ely,the area of each cover plate ruqina Fn i ~ 1n1 1 :.
------a
-y, (1 w- 1)J IS element) :
, ,.. ., .... ...
A, = 3800 13.72
-
(0.912 + 0.025)(0.95 x 0.6F,) 0.912 + 0.025

- - ' " -
by- 418 - 13.93 << 42.4 O.K.
i-
"'I. = 72.6 kN11n t; 30
i\ 1 It is not necessary to check b/29 of beam since 9 = 32.0 mm. Use cover
is? ,' 7
rrFh,
20 plates 30 x 470 mm.
726 KC 1
- 1

- 1
ill
---I

Figure E10-26 ///

 466 STRUCTURAL STEEL DESIGN

10-4 PARTIAL-LENGTH COVER PLATES acing of bolts or welds can be found by equatlng the weld
or bolt capacity and the spacing s as
resistance capaclty
S =
us

The weld capacity at the theoretical cover plate cutoff 1s obtalned by e

the moment of the total weld capacity to the moment at the theoretic

lance.
F, =- -
lLfQ (hips or kY)
I
where F,= total force to be carried by the weld In the cutoff length a
M = bending moment at the theoretical cutoff point
I
1 v$ = -
vQ
I
(kips/in or kN/m) Q, I= terms previously defined

where us = force/length of developed shear

I*
Q = statical moment of cover plate with respect to neutral axis Example 10-3 What 1s the theoretical cutoff dlstance for the cover-plat
I = moment of inertia of section, including cover plates beam of Example 10-2? Also, design the weldlng to fasten the ~ l a t e sto the
beam using E60 electrodes
AlSC Sei 1 1 0 4 AASIlTO Set 1 7 12
S~LTJTION Obtain the cutoff points (Fig E10-3a) The capacity of the beam
without cover plates is
M, = SxF, = 13.72(150) = 2058 kN . m

-b 20 -----------+I

AASHTO rnlnlmurn Lover plate length L,,,,

Figure E10-3a
fps 3 d b + 3 ft
SI 3 d b + 0 9 1 ~n a*
Since the moment diagram is a ~ a f a b o l awith slope = 0 at midspan,
equation for moment IS
,+fx = M, - k,~'
Figure 10-7 Code requirements for theoretical cutoff d~stancesfor cover plates. Requireme
same for a v e r plates narrower (shown above) or wder than next underlying plate or beam
 +
" .
ddS STRUCTURAL.
S'I'EBL DESlG ii

,' x = L/2.wehave
f ;i :
M, - -
M, 4% x 2
=
L
g With M, = M, = 2058 kN . m at the distance x from midspan where
9
:$ beam capacity is adequate without cover the theoretical cut
.
:
'8
is
,, :,,, v :,

,
,:
,
[
x = (3855 - 2058)

The theoretical length of the cover plate

3855
L2 ]'I2
= 6.828 m

= 2x = 2(6.865) = 13.73 m. l ' o t ~ lengrh

i CP= I i . l i nl-

Design the welds (cover plate 30 x 470 rnm; I, = 12 499.4 x 1,=:0~,-- r:=;O;iirn

at the cutoff point x' = 10 - 6.83 = 3.17 m.

Shear V = 771 - 77.1(3.17) = 530 k ~
/I/
Q = AJ = 0.030(0.470)
(Y - + 15
) = 6.41 x m3
10-5 GENERAL PROPOR1[1IONS OF PLATE GIRDERS hd,
..

us = 530(6'41) = 281.6 kN/m

12499.4
(Noting that I is and Q = low3,we obtain 281.6 k N / h and not the
direct calculator reading of 0.28 16.)
The allowable stress using E60 electrodes:
Fw = 0.3(415) = 124.5 MPa follows. The moment camed by the flange is
We will arbitrarily use D = 8 mm for the weld (tf > 20 mm per AISC Sec.
1-17.2) so that the theoretical spacing at the cutoff point to the interior of
the span is The moment camed by the web is
t,h2
-Pwm-- 2(0.008 X 0.7071 1)(124.5) = 1.4085 iVf, = flSrw= f;-7
Tentatively, use a 40-mm length of weld (40 mm = 5D > 2D), which will The total moment is
carry M = IMJ + .\/,
Pi = 40(1.4085) = 56.34 kN
This corresponds to a distance along the beam of
?. 56.34 X lo3
S = = 200 mm O.K.
281.6
Use 40-nun lengths of intermittent weld alternating on each side for each
2d0-mm increment of length.
The weld in the distance a (a = 1.5b, since D < 0.75t;) is continuous.
nd the effective moment Me is
'
Me = [0.705(2) + 0.418](1.4085)[912 + 2(30)]
= 2503 kN m > 2058 O.K.
If Me had been less than the moment at the theoretical cutoff point, it would
have been necessary to either increase D or the distance a (or both). de
final sketch is as shown in Fig. E10-3b. Figure 10-8 ~ i section
~ to dobtain~ an approximate
~ expression for the flan~
area-
 472 STRUC.IUK~.L y i b c ~
L)L!SII,~J
4

Figure 10-10 illustrates the assumed girder web plate loading on a strip e AISC rounds this to obtain the limiting h / t ratio as
dx(h). From Chap. 3 the critical buckling stress was found to be
h 14 000
$!
.,-,,"
..l
3

~,T~E fps - I
[~,(~~+16.j)]'/'
I

Fcr =
,,&

12(1 - p2)(h/t)2
\' ' h 97 100
where the ratiogh/t has been substituted for b / t as given in Eq. (3-5) SI: -5
The force exerted on the web of the web segment shown in Fig. 10-lq$
[F,(F, + 114)11/*
A f f sin @ r A,fb The AISC specification allows a somewhat larger h / t ratio if transvers
feners are used at a spacing ratio of a / h I 1.5 of
+
since (in radians) is a very small angle. The stress f/ must be of suffi
magnitude to overcome any residual stress Fr in the web; thus the web strai
units of F L - 2 ) at yield is
Fr + F,
f =er+ey=-
&
E . .

I t and the angle @:is I I

v d~ 2 ( F r + F , ) dx The maximuq h / t ratios for several grades of steel are as follows:

+= &- = -
fi/2 E h
At web yield the applied force is

2 , ..
..
.,. .. , .. .
.
A d tan + = AJffl
but
lor small angles and + in radians, giving MPa [by Eq. (10-4)] a/h < 1.5

dx 333
Afffl = 2F,(F, + F')-
Eh 284
259
This value'should not be larger than the critical web buckling force 248

kcT2E
FCrtdx= ( L ) 2 ( t )dx
12(1 - p2) h Experimental studies on full-size girders have shown that the web
Equating Eqs. ( a ) and (b) and solving for h / t , we obtain
the compression zone deflects laterally by small amounts at very early
1
bending, with the resulting transfer of stresses from the compressio? web l r
to the compression flange. This results in an increase in the flange strebs.'over
h
-=[
t
kCnZE A,
24(1 - p2)
When A-7 steel was used, it was assumed that the residual stress Fr could be
~dequatelyapproximated as Fr = f , / 2 = 33/2 = 16.5 ksi. This value !is cur-
1
F,(< + F,)
I that amount indicated by theory as shown in Fig. 10-10c. This increme in
compression flange stress requires a reduction of the allowable compressive
stress so that the stress actually developed does not cause a flange failure. The
experimental studies indicated that this flange stress reduction could be ex-
pressed in terms of A,, A,, h / t , and Fy. A possible equation in terms of ultimate
.-ently being',used for all values o f . steel F,. It I is! also',assumed 1
moment is
.4,/A, = ,I Poisson's ratio for steel p = 0.3, E = 29 000 ksi, and for
'2"Yjtain
-h - 13 784
t [ F ~ ( F+, 1 6 . 5 ) ] ' / ~
Since the ratio of M u / M v = Fu/Fv = F,'/F,, we may rewrite this equation ih
 474 STRUCTURAL STEEL DESIGN
P U T E GIRDERS '47

terms of stresses as long as the compression flange is sufficiently stable with

respect to L / r , and d / A P This is done by limiting the maximum allowable
stress io F, as defined by AISC Eqs. (1.5-6) and/or (1.5-7). The value of 8,
depends on the flange restraint effects on the web, and if we assume partial
restraint,' ,& may be taken as 5 . 7 v m (see Basler and Thurlimann, T
"Strength of Plate Girders in Bending," Journal of Structural Division, ASCE,
l ii I
August 19611, where F,, is the critical web buckling stress. If we replace
F,. = 1.65Fb,we obtain (with slight rounding):

i bl

Inspection of Eq. (10-5) indicates that if girder to Pratt truss. ( a ) R a t [ truss ( - ) = cornpression (b)
ate girder with tension fields.

or the SI equivalent, no flange stress reduction is necessary (i.e., F; = Fb).

t the yield shear stress F, according to the van ~Misescriteria is

10-6.3 Shear and Stiffener Requirements 5

FYs = - (4
Where the l i l t ratio is sufficiently small, web buckling will not occur under v"j
shear before shear yielding occurs. Actually, a beam web as part of a flexural e obtain A , and substitution into ( b ) ,
member subjected to a bending moment is canying the shear in a "tension field"
fi (V,)Fr,
mode. Some researches have likened this to,a Pratt truss, where the stiffeners (if Vb = ( e>
uszd) are the compresbion members and the web segment between a Pair of 5
stiffeners is the tension element, as illustrated in Fig. 10-11a. Actual girder tests . g ~ , , , the value of FA is, from Eq. (3-5)- here
carried to yield so that mill scale forms stress lines displays a stress pattern
similar to that idealized in Fig. 10-11b. (3-5)
Girder webs are designed by AISC specifications assuming that the shear is,
carried by beam action shear-as for rolled shapes-until shear web buckling ~ q (10-6),
. the critical stress is taken as the mean of
stresses are reached; then the additional shear is carried by tension field action. q. (10-6) and the value of 0.8Fy,, to obtain
The ultimate,shear resistance in a web panel (space between two stiffeners) is the
sum of the two shear (beam, Vb, and tension field, V,) contributions, or
(10-7)
vu = Vb + v,
The beam shear contribution is d where as previously defined (see Table 3-21;
., y, ! V; = Fc,(htw) = FcrAw k = 4.00 +-
ultimate shear force capacity of a girder for web plastification (plastic
capacity) is (10-8)
v- = -ysAw
 -t:;d P L i T E GLRl

4. In girders designed on the basis of tensional;! action, the spacing of the

interior stiffener from the end, bea,ri&;ifi'ffener, or any stiffgners adj
large holes, will be based on limiting h e shear stress to AISC Eq. (1.

appropriate to obtain: 10-6.4 Combined Bending and Shear (Interaction Check)

For C, 2 0.8:
When a girder is simultaneously subjected to a large value of shear and bzn
simultaneously, the allowable shear stresses may have to be reduced.
interaction solution was developed by Basler (see "Strength of Plate Girde
Under Combined Bending and Shear," Journal of Stndcrziral Dioision, ASCE
The AISC uses '' October 1961) based on girder tests at Lehigh University. AISC has slight1
rounded the values to obtain
0.375f,
5, 2 0.6Fy

eneral, no interaction check is required if

(a) f, < 0.6F" and f, 2 0 . 6 5
(b) f, = F, butf, 5 0.75Fb.
en either one of these conditions is not met, the interaction using Eq. (10-
ust be checked for possibly reducing the allowable flange bending stress F,
e of this equation.

-6.5 Intermediate Stiffeners

ffener design is. based on obtaining the vertical stiffener
= 0 from Fig. 10-12c to obtain
P, = f,t,a sin + sin + ) .

19'A
i*";,:

Substitution for f,, $, and A,, = ~',/F,II,, gives <'..

i

are always required. YDht, (10.14)

2. Intermediate transverse Stiffeners (AISC does not have a specification for [ I + (o/h)']'/'
longitudinal stiff€mx'S) are not required if bottf~of the following are
P

met: which is AISC Eq. 1.10-3. Where Y = F,,/F,,, to account for ~ossibleusciof a
different yield-grade steel for the stiffener'than ior the web, D = factor to
account for reduced efficiency for a stiffener plate on one side of the web or

Intermediate stiffeners are required for any other shear stress condition. D = 1.0 for stiffener plates on both sides of web plate
When stiffeners are required, the spacing is limited to
D = 1.8 for angle used as stiffener on one side of web
:J
D = 2.4 for stiffener plate on one side of web plate
The required area of stiffener is often very small. To ensure a stiffener
but a / h I 3.0. ufficiently large and stiff to maintain the girder web shape, the moment of
 *&" .JCTURAL STEEL I)i,%k<N' P U T E GI1U)EY.S

inei:.~cr ust be at least4 ,$

:. $4 AlSC a n d A R E A .A.iStlTO
1 .'*
6 h
4, 2 (%) (10-15)
= in4 or m4. Since the stiffener is a compression
,.:.the minimum io should be maintained.
he stiffener must be fastened to the web to carry some vertical force.
! i'ifferentiation of the equation for P, with respect to a / h gives
{ *":I 2
, P, =
.B; r 0.015~,h~<
"LJsiui ?, = F,/ E, a safety factor of 1.67, and assuming that the stiffener force is
Idebe1;qxd in the distance h/3 from the compression flange, we obtain

a 6 h 1.67 N D o r fillet

ha&u .I h some additi*

tips hfr obtained as Cornpure L r = h r
C o m p u t e L'lr
L 1 = O75h
r = [11'4~"~

~ ~ ( f i Check
: as column based o n cross-hatched area shown above
Check bearing: use on!y area of plates as .Ah = birr.
Check b,/r, ratlos.

Figure 10-13 Bearing stiffener design requirements for speclficatlons ~nlcated.

Use t; = in or m.
The allowable bearing stress F,,, = 0.90FY (AISC Sec. 1-5.1.j.1) based on the
1096.f + w i n g Stiffeners lesser F, if the flange and stiffener are of different y~eldgrades of steel.
b

Beanr,, tdfeners are always required in pairs over the reactions. Bearing
stiffece-s mgy be required beneath concentrated loads carried by plate girders. 10-6.7 Web Crippling
These s:,:feriers must extend the full flange to flange distance and have a close
bearir~,:against the flange delivering the load. The stiffener width must be such Webs of plate girders are required to be proportioned so that web crippling
as to extend approximately to the outer edges of the flange or angles. (same phenomena as for rolled beams) does not occur. This 1s accounted for at
Bear'ng stiffeners are designed as columns with an area that includes the reactions by bearing stiffeners. Where the compression flange supports a uni-
stiff n ~ r sand a central weblarea of 12twfor end and 25tw for interior bearing form load or concentrate ds for which bearing stiffeners may not be
? e
stiffeners (see Fig. 10-13). This area is used for computing the radius of gyration requirgd2 compressive delivered by the flange to the girder web must
. andrfo. checking the coludn stresses. The effective length of the stiffener may be iuff.iciehtly low that c (Fig. 10-96 illustrates buckling) does not occur.
This is an cage-loaded plate 'stabihty problem, and agaln Eq. (3-5) IS used (see
be taken as 0.75h because of being securely connected to the web.
Basler, '.'New Provisions for Plate Glrder Design,'' Proceedings 1961 AISC
The effective bearing area A: is taken as the area outside the flange angle
'ffiet oi ihe flange-to-web &~ds. National Conference, AISC,, New York), SF = 2 65. E = 19 000. and some
-
design requires co ing L r / r to find the allowable column stress Fa rounding, to obtain the allowable web compressive stress as
$$@$$:eking
fif
< PY
$4
fps: F =
10 OOOk,

6
P = AF, I applied load or reaction (hllw12
Al5a.check using the effective bearing area A, = bits (Fig. 10-13) to obtain
P = AbFbrg I applied load or reaction
:i r
and for a flange delivering compressive load restrained against rotation.

In the AISC specifications, Eq. (10-17) is directly combined with the appropriate
kc term for displaying the design equations.
Ir
Example 10-4 Design a welded plate girder to support two columns span
ning an auditorium space in a high-rise building. Floor loads deliver c- Figure ElO4b
equivalent uniform load to the top flange of 2.8 kins/ft- - (not i..
,---- n.. r. l.i i d i n a t&he
"
girder weight): General span and loading is a's shown in Fig. E10-4a. We
FO; h/r = 162, obtain t78/162
= = 0.48 in. Try tentative t,,, = 0.5 in.
will assume lateral bracing" of the comnressinn f l n n a ~st, the ends and at the
- --r--------
I- - -
concentrated loads. Other design data: E60 electrodes, .ATSC
-
---- -
s,,n,..~.,,,r,,i,.f i r tai n n c h
Maximum - =
14 000 = 322
A-36 steel, and girder depth limited to 84 in.

90k 105'

For t, = 0.5 in, h/t = 78/0.5 = 156.

Step 2. Make a preliminary flange design.

(Assume that fb --
21 ksi and an average distance to the center of the flange
area of 79 in.) Try a flange plate 7/8 X 18 in.
- b = 18 9j
= 10.3 < ---- (AISC Sec. 1-9.1.7)
2rj 2(0.875) fi
Step 3 . Compute the actual moment of inertia and section modulus of
2R = 343.4 kips the trial section and revise the dimensions as required.
Now draw shear and moment diagrams as shown in Fig. E10-4b. bh3
I,, = -12 + 2,4d2 (neglect I, of flange plates about cenlioid axis
Step 1. Make a preliminary web plate design.
Assume that web depth = 78 in. The limiting h / t w for no reduction in
flange bending stress is
$8
h -
- 760
--=-= -,It
760 162 (approximately)
m
 31 8L'C 1 U~U'L b 4 LLi. UIjblbiu

Gxnpute the weight of the girder:

Check the allowable stress F,.
te L / r , with r, computed using I, of the compressi
., f gyration is obtained using A, + A,/6:
The approximate additional 'bending moment (conservatively compute 7-
I ilnce the moment due to other loads is slightly off center) is

0.24(53)~
M= = 84.3 ft - kips
8
+
The total bending moment = 2722 84.3 = 2806.3 ft . kips.
I - M 2806.3f12'1
-/
JbC-=:
S 1724.5 19.53 k s ~ \
I

Since this is considerably less than 21 or 22 ksi,

let us tentatively revise the
web thickness 2, to $ in.
We must investigate both th; end panel gnd the Intenor (between two
- -. columns) panel smce C, is different for each locat~on.For the end panel:

,
-w

I Check the web shear SO that the plate is not too thin (neglect beam weight at L = 18 f t (largest value), and m , = 0 and ibf, = 2664 f t . klps
j'lls pmnt): 1)

1 O(U.3 13)
-. .- v.n.

step 4 R e c o ~ n ~ uIt eand&, For the interior panel: C, = 1.0 (since Interior moment of 2722 > 2664
ft . kips):
I
= 14 829.8 + 48 992 = 6 822 in4
I s, = -
63-822
1600 ln3
39.875
For this L / r , ratio, F, = 0 . 6 5 = 22 ksi for both the end and interior
New girder weight =
0.490
144
-
(3 1.5 29.3) = 0.207 kip/ft + panels.
Now check AISC Eq. 1.10-5, since h / t w = 208 > 7 6 0 / a :
'be approximate additional moment due to girder weight is
0.207(53)~
M= = 72.7

MtOw = 2722 + 72.7 = 2795 ft . kips == 2 1.05 > 10.96 ksi O.K.
2795(12)
"4, = 1600 = 20.96 ksi - At this point the bending stress and slider proportions are adequaie
Continue the design. Use two flange plates x 18 in and a web plate unless a later interaction check requlres a revlsion of girder section.
3
y X 78. Step 6. Compute the stiffener requirements.
AISC specifications require bearing stiffeners under reactions and the
-- -

h = 208
- two concentrated column loads. In the end panel the actual shear stress at
tw the reaction is
D = 79.75 m A, = 29.3 in2 f =-=-
V 173.25 + 0.207(53/2)
- = 6.11 ksi
A, 0.375(78)
SX=1600in3 Af=15.75in2
According to AISC Sec. 1-10.5.3, intermediate stiffeners (other than the end
486 STRUCTURAL STEEL DESIGN

beanng stiffener) are not required if

been obtained from Table 11-8 of SSDD but requires double i
h -,- E;.,,C.," Step 7. Check the interaction at concentrated loads [Eq. (10-13)
- I L6U and j, f, FF,= -
At dx to the right of the right column, the shear V is
t 2.89
We will provide bearing stiffeners under the column loads, which leaves V = 178.4 - 18(2.80 + 0.207) = 124.6 kips (including girder wei
clear distance in the 18-ft end panel of 18 x 12 = 216 in.
For V = 124.6 kips, the web shear stress is
124.6
j" = -
29.3
= 4.25 ksi

We have just found F, = 7.67 ksi in step 6.

(
0.825 - 0.375 -
7.67 I]
4'25 36 = 22.2 > 22 ksi O.K.

Since the allowable stress in interaction is 22 ksi (maximum) and the a c d

Assuming that C, will compute < 0.8, we will try Eq. (10-1 1): and allowable values based on flange stability are less, the lesser values
control.
45 OOOk - -45
C" = F, = 21.05 ksi _Tb = 20.96 k s ~
q ( h / tJ2 36(20gL)
Step 8. Check the stiffener requirements of the 17.5-ft interior span
The allowable shear stress if no intermediate stiffeners are used is limited t
between columns.
36(0.169) The maximum shear is obtained coming from the left:
F" = = 2.1 1 ksi << 6.1 1 N.G. stiffeners required
2.89
Try one stiffener at half the distance (noting that if this works for the 18-ft
panel it will also work on other end in 17.5-ft ane el):
(Y1
V = 170.15 + 0.207 - - 17.5(2.8 + 0.207) - 90 = 33.01 kips

j,=--33'01 - 1.13 ksi

29.3
Try no stiffeners:
a
-- = 108
-- 1.38 < 3
h 78 k
O.K. (less tha maximum a / h allowed)

= 1.56 > 1.38 O.K.

F, = 36(0'170)
2.89 = 2.12 > 1.13 ks, O.K. no st~ffenzrs
required
With intermediate stiffeners and C, < 1, we can use AISC Eq. (1.10-2) or Step 9. Check the web cnppllng under the compression tlange due ti
Eq. (10-9): uniform load.
Assume that the flange is restrained against rotation (since it carria
uniform load). The load carried in compression to the web is 2.8 kipsift -t
weight of top flange. We will neglect flange weight, so that the compressiv
Since 7.67 ksi is greater than the actual shear stress at the bearing stiffener, stress is
6.1 1 ksi, and the shear is less at the interior points, it is not necessary to
2.8 = 0.622 ksi
check the shear stress further for stiffener analysis. Note that Fv could have jc = 0.375(12)
 I
STRUcrCnW, SIEliL DfiSICb*

The allowable compressive stress (checking at the location where a / h is for web and stiffener); C, = 0.215 and a / h = 1.38 from step 6.
critical is
4
F = IO"*[5.5 + [Eqk (10- 17) combined with kc]
1 ( h j tWl2
In terms of A,, x 100/A, this value could have been obtained from TabIe
- &(5.5 + = 1.4 > 0.622 h i
O.K. 11-8 by using double interpolation. Try two bars 3/8 x 6 in:
208' 2.69'
A = 2 ( : x 6) = 4.5 1n2 > 3.02 O.K.
Note that if F <f,, we would have to either add stiffeners to decr~asea
' 'or'inctease t,. -=-- -16 O.K.
Step 10. Design bearing and intermediate stiffeners. 0.375
t
For bearing stiffeners at girder ends (we will use the same size The minimum moment of inertia of the stiffeners is
column loads) try two 8 X 1/2 in bars, for a width (2 X 8 0.375 = +
in) approaching the width of the flange plate of 18 in.
b
I
- - - 8= 1 6 ~ 95 - O.K.
t 0.5
fi
oS(16.35)~ Use a stiffener plate length
I = = 182.9 in4
12 h - 4 t , = 78 - 4(0.375) = 76.5 in
The "effective" column area for the radius of gyration (see Fig. 10-13): Step 11. Design welding to fasten the stiffeners to the web.
A = 16 x 0.5 + 12(0.375)(0.375) = 9.69 in2
r =dF182.9
= 4.34 in
jU= h[(&y]l/' kips/in [ Eq. (10-16)]

L = 0.75h = 0.75(78) = 58.5 in

L
-- 58.5
---
r 4.34
- 13.5
From Table 11-5, the allowable column stress is Fa = 20.98 ksi. The
For a pair of stiffeners, we have 2.7 kips/in; for each side this becorn&;
2.7/2 = 1.35 kips/in.
For 1/2-in bearing stiffeners, use a 3/ 16-in weld and E60 electrodes.
I
column stress is
F,,, = 0.1875(0.70711)(0.3 X 60) = 2.39 kips/in
178'74
fa=------
9.69
- 18.45 > 20.98 ksi O.K. Use a 3/16-in fillet weld continuous for both bearing and intermediate
stiffeners.
Check the bearing stress; assume a 5/16-in weld, so that the effective
Rationale: Few stiffeners and the weld can be made in one pass. I t 2
bearing area = (8 - 0.3125)(0.5)(2) = 7.7 in2:
too difficult to measure and set up alternating weld distances and gaps.
F,,, = 0.9G = 32.4 ksi Step 12. Design welding to fasten the flange plate to the web.
P, = 32.4(7.7) = 249.5 >> 178.74 kips
Check the end for maximum shear:
O.K. .

For intermediate stiffeners, with only one intermediate stiffener use two
plates.
Use a 5/16-in (t, > 3/4 in) continuous weld on both sides at F, = 392
kips/in. Note that the weld 1s considerably overdesigned, but for an im
portant girder the use of intermittent welds is not worth the savings-
ularly if the weld can be made in one pass, as here.
For a plate on both sides of the beam web, D = 1.0 and Y = 1 (A-36 steel Figure E10-4c illustrates the design summary for the girder.
490 STRUCTURAL STEEL DESIGN

9ok 105~ AREA:

When cross-ties rest directly on the girder flange:
17 5' 18'

16
? - 8 X j t 2 - 6 x i 8 ' t Web X 78" 1 2-8X+"e 1-2 - 6X :"t
X 18" The limiting b/2$ = 10 for A-36 steel
When cross-ties do not rest directly on the girder flange:

The limiting b/29 = 12 for A-36 steel.

10-7 BL'km GIRDER DESIGN THEORY-AASHTO
Z1
AND AREA
8 d 10-7.2 Web Plate Design
Plate girder design using AASHTO and AREA specifications is very similar but
more conservative than AISC because of the more hostile environment to which AASHTO also uses Eq. (3-5):
the girder will be subjected. In general, however, the same general considera-
tions apply:

1. The girder is proportioned by the moment-of-inertia method.

and solving for h/tw, substituting k c = 23.9, E = 29 000, and F , = 1.19Fb, we
2. No uniq'ue solution is possible.
obtain
3. Shear and stiffener requirements are more rigid. Bearing stiffeners are always
required, but AASHTO allows use of longitudinal stiffeners.?

10-7.1 'Firder Flanges (AASHTO and AREA) ,, 'Xlis gives h/tw = 163 for A-36 steel.
The AREA limiting h/t, ratio is obtained from Eq. (3-5) by substitution of
In general, the basic allowable bending stress is kc = 23.9, E = 29 000, and Fcr = 0 . 6 5 to obtain (with slight rounding)
Fb 1 0.555 h 1030 2700
fps: - = - SI: - - -
However, if the compression flange is laterally unsupported in a length L, the
stresses must be reduced as in Part I11 of SSDD and in the appropriate
lw fi lw 6
specification. For A-36 steel, the limiting h/t, = 171. When the value of fb < Fb in the
The b/2$ ratio for the flanges is also limited: compression flange, AREA allows an increase in the h/t, ratio:
AASHTO:

with the ratio (F,/ fb)'I2 I 2.0.

AASHTO and AREA limits the web thickness of fabricated plate girders to:
WSHTO: tw 2 5/ 16 in (8.0 mm) AREA: t, 2 0.335 in (8.5 rnm)
Ths limiting 6/29 = 12 for A-36 steel when fb = Fb = 0.55Fy. AASHTO allows the limiting h / t , to be increased with a longitud
stiffener at 1/5 of the clear web depth from the compression flange. This value
t Alsg ARE AREA, but for continuous girders. is based on Eq. (3-5) using kc = 129 (since a more substantial edge fixity is

? C *
."a
 492 S T R U C ~ U I L A I:,. 1 ,XL u~.zl,.;:-r 5 . ..
I PWTE GI&LRS 4 ~ 3
8 .

'

obtained) and F,, 1.64, to obtain

.I_

= From an approximate aqeraging of fi we obtain 348, which if used to back-con-

h <-3840 pute the SF gives 1.37 at a / h = 0.5 to 2.02 at a / h . 1.0. In earlier AASHTO
fps: 1455 ' S I : -
-5 - -
specifications the stiffener spacing could be computed as
" v3 'W 6 348 t 914t
i 3 e limiting h / t w for A-36 steel; is 327. fps: a = - (in) SI: a = - (mm)
fi fi
10-73 Shear and Stiffener Requirements where t has units of in or mm. The previous two editions of the AfSC
specifications included this requirement for spacing the first interior stiffener
:he AASHTO requirement for stiffeners is also computed using E ~ (3-5) . and from the bearing stiffener. Currently, the XASHTO specifications for inter-
the k factors given in Eq. (10-8) and with varying safety factors. If the mediate stiffeners is somewhat more indirect and is given in Sec. 10-7.4 foI10w-
such that a / h -+ (no stiffeners), kc = 5.34 from ~ q (10-8),
. and taking ing under "Intermediate Stiffeners."
F,,-2.5 f,, we obtain The AREA specifications reduce the 348 factor slightly, to obtain
h 234
h 618 332t 875r
fps: <-
-
tw
-
vx
-5-
lw
S1:
w ips: a = - (in)
fi
SI: a=-----
fl
(mm)

> The limiting h / t w = 68 for A-36 steel when f, = F, = 0 . 3 3 ~ ~ . In these equations t has units of in or mm. The stiffener spacing is limited to not
AREA requirements are slightly more conservative: more than 72 in or 1800 mm when using Eq. (10-22).
h 360 h 950'
fps: - 2 - SI: - - <-
lw 6 fi 10-7.4 Stiffener Design
The limiting h / t w = 60 for A-36 steel. Longitudinal stiffenem The AASHTO value for longitudinal s$feners in terms
Intermediate stiffeners are not required by either specificati& if the h / t w . of moment bf inertia (see Erickson and Eenam, "Application and Development
ratio is less than that given by Eq. (10-20) or (10-21) and may not be required of AASHTO Specifications to Bridge Design," Journal of Structural Dit'inbn
for larger h / t w ratios as given in the following section. ASCE July 1957) is
M~'hen h / t w is larger than given by the preceding equations, the stiffener
: '1 spacing is obtained from Eq. (3-5) by obtain IJ = 11t1.[24(%)2- 0.131 (10-23
I
with a plate thickness limited to
b 71 b 188 (1 0-24
from which fps: -5---
a SI: - 5 -----
<i
=
a
-
kT ~ E
12(1-p2)SF I --
tw

'JsiM,S.F..= 1.5, several a / h ratios from 0.5 to 1.0, and computing kc fr

(10-8), we typically obtain
- 0-
t where f, = calculated bending stress in the compression flange. The stiffener
may be placed on one side of the web at the 1/5 point from the compressio~
flange. The AREA allows longitudinal stiffeners for continuous girders in d
negative-moment regions and the specifications are exactly the same as fo
l

AASHTO for I, and b/t.

a/h k P Bearing Bearing stiffeners are always required for both AASHTO
AREA girders. They are designed as columns using an area as shown in Fig
0.5 25.36 332.8
10-13. The allowable column stress depends on the L / r ratio computed a
0.6 18.83 344.4
0.7 14.90 357.2 shown on Fig. 10-13. The bearing stiffeners must be checked for bearing as we!
0.8 12.30 370.9 as acting as columns. The bearing area is as shown on Fig. 10-13
0.9 10.58 387.2 allowable bearing stress is
1 .o 9.34 404.0
USHT~: Fbrg = 0.80Fy AREA: Fbrg= 0.835
 494 STRUCTURAL STEEL DESIGN

AREA requirements for the b,/t, ratio of bearing

for any other compression member.

feners if
h
-<-I50 and f,5Fo
tw
where

but!,F, I 5 / 3 . This criterion gives the limiting h / t w = 68 for F, = ~ , / 3for

A-36 steel. Thus any h / t , ratio that is less than 68 does not require transverse
stiffeners if fo < 5 / 3 . Any h / t w < 150 does not require stiffeners if f , < F, as
givdn in Eq. (10-26).
When intermediate stiffeners are required or when j, > F~of E ~ (10-2 .
the maximum spacing is limited to a 2 1.5h and the allowable shearing
limited to

lnediate stiffeners(including other than reaction bearing stiffeners) of

& [ I+ ( h / a ) 2 ]
where C= -- I I
ti ~,(h/t,)~
'
B = 222 000 in fps units
ere p=0.3 for steel
= 1 520 000 in SI units
a = actual stiffener spacing
It is often convenient to plot the limits of these stresses as in Fig. 10-14. I
tw= thickness of girder web plate
When ~ m s v e r s estiffeners are required, the first stiffener is placed sue J = 25(h/a)l - 20 but J 2. 5.0
a / h 5 0.5 and the actual shearing stress, f, 5 Fi, is obtained from the followi AREA requires intermediate stiffeners if the h / t , ratio is Wate
equation:
values defined by ~ q (10-21).
. ~f stiffeners are required the spacing is

where B' = 70 000 in fps units

= 483 000 in SI units but a I 72 in (or 1.829 m),
where t,= girder web thickness, in or mm
The U S H T O specifications require a minimum moment of inertia of inter- f v = actual web shear stress = V/dr,, ksi or blPa
 SSTRUCTUKRLb 1LLL. UEa~crr.r

The stiffener dimensions must be at least as follows:

AASHTO: width 2 2 in + D/30 and width 2 bf/4
2 51 mm + D/30 and width 2 bf/4
t
thickness 2 width/ 16
AREA: same as for AASHTO

10-7.5 Interaction
.\ASHTO specifications include a bending stress reduction if the shear
f, I 0.6Fv according to

id-7.6 Lateral Bracing and Diaphragms

Figures 10-15 and 10-16 illustrate bracing and diaphragms in current practice.
.AASHTO specifications require either cross frames or diaphragms at the ends
;-nd at intervals along the span not to exceed 25' ft (7.62 m). Diaphragms (cross
beams) must be at least one-third and preferably one-half or more of the gird
depth.
When spans exceed 125 ft (38.1 m) for concrete deck bridges, an additional
,:ystem of lateral bracing will be used parallel to and as close as practical to the
bottom flange and in at least one-third of the bays (when multiple girders are
ysed). The smallest angle permitted by AASHTO for this is the 3 X 2; (76 X 64
,mm) fastened with at least two bolts or the kquivalent weld in each end
konnection of the angles.
Fi;rve 10-15 Bridge bracing using diaphragms. (a) Diaphragms used with five plate girders. (b) E
6i:.phragm. Must be sufficiently set back from adjacent diaphragm or abutment so that it can
i*.?ected and periodically painted.
Figure 10-16 Longitudinal and cross bracing. Note use of verrical and longitudinal stiffeners. Bridge
in background of (c) is older and uses riveted construction where bridge in foreground is alI-welded
The three girders carry a four-lane roadway plus shoulders and walkway for both bridges. ( a ) Cross
bracing. (b) Fabrication details. Note that the end fastener plates ire welded to both web and fians-
On continuous bridges these plates provide flange brace points. (c) Use of both cross bracing and
longitudinal truss bracing. Note longitudinal stiffener along compression flange, where h / is ~
excessive.
AREA requires the following:
1. For beams or girders less than 42 in (1066 mrn), use of I-shaped diaphragms
(cross beams) as deep as practical and asin:, double-angle beam framing
connections.
2. For girders deeper than 42 in and spaced more than 48 in (1220 mm) on
centers, a system of cross-frame bracing in which the angle of the cross-frame
-
diagonals with the vertical is not greater than 60".
3. Spacing of diaphragms or cross frames is
18 ft (5.5 m) for open-deck construction
12 ft (3.65 m) for ballasted-deck bridges
4. Where ballasted decking is carrled on Gross beams without stringers (as in
Fig. 10-5), at least one line of longitudinal d ~ a p h r a p s1s to be used, as shown
in Fig. 10-3d.
5. Knee bracing is required to support the compression flange for through deck
girders. This bracing (see Fig. 10-17) is usually placed at selected intermediate
stiffeners for ease of attachment and ranges in slope from about 3 V : l H to
somewhat less. The maximum spacing is limited to 12 f t (3.66 m). It may be
 Cross b a r n s a t 2 . 3 f r
760 min

Figure E10-5a

The impact factor is computed (see Fig. E10-5a) based on L > 25 m as

:IR+ 71
16 + 27.5 - 9 = 27.6 use 28 percent

(Note that tlphimpact factor is not the same as used in Example 4-15
Figure 10-17 Top flange bracing using web plate and bracket for the through-girder ballasted-deck design momecf
rallroad bridge shown m Fig. 10-4.
..
M, = 9984(1.28) = 12 779.5 kN . m
The design shear
designed as a column for an axial force based on the horkontal corn
V, = 1680(1.28) = 2 150.4 kN
2; percent of the compression flange force
We will neglect the live load on the pedestrian/semce walkway. With rhis
0.025&Af
Pkb = assumption the dead-load estlmate is
H/ V
Girder weight (including stiffeners and weld) = 10.0 kll
The approximation of the axial force using the sin defined by H and v rather
than the actual length is sufficiently accurate for design. Ballast, including ties at 350 rnrn depth and a reduced
---- effective width: 120(9.807/62.4)(0.3j0)(4.42/2) = 14.6 kN
Example 10-5 Design a welded railroad bridge girder for a 27.5-m-span
Steel deck and cross beams and factor of 1.10
single-track ballasted-through deck bridge. Use a Cooper E-110 loading, From Example 4-15, the beam is a W760 X 160.7/ 1.58
A-36 steel, E70 electrodes, and the AREA specifications. The general bridge Deck: 0.015(77.0)(6.49>(1.10)/2.0 =4.1 kN
configuration is as shown in Fig. E10-5a.
Cross beams: 37(1.58)(5.79)/(27.5 x 2 ) = 6.2 kN
S ~ L ~ TFromI ~ NTable 1-2, the design live load for one-half the track (one Track at 200 lb/ft: 0.200 x 14.59 (AREA spzc~f~catlon) =2.9 kN
rail load to each girder) and adjusted for E-110 loading is:
Walkway: estimate 300 Ib/ft = 0.300 x I4 59 = 4.4 kN
Total = 42.2 kN/n

Note that some approximation 1s used for the steel deck to allow for
forming the trough to hold the ballast to g v e an effective length of 6.49 m,
V = 274.5(%) = 377.4 kips x 4.45 = 1680 k~
Note also that an estimat~ngfactor of 10 percent is applled for

<
 S T R U C T n k L Ki'W:L3 1 )I: , I C ~ ? ;
P'LATE G~RDERS

mcertainties. The dead load moment is

The section modulus is
M = - =wL2
42'2(27'5)2 = 3990 kN .
8 8
Tne dead load shear (at reaction) is
The corresponding maximum bending stress is
v = -WL
= 40.2Q7.5) = 553 kN
M 16770
2 2 fb = - =
Sx
-
123.81
= 135.4 MPa < 137.5 (O.K. so far)
The total design moment is
MdeSi,, = M L + Mdead= 12 780 + 3990 = 16 770 kN . m Step 3. Check the girder weight.
VdeSi,,= V L + Vdead= 2150.4 + 580 = 2730.4 kN A,,,,, = 2(0.070)(0.815) + 0.016(1.994) = 0.1460 m2
Step 1. Find the girder proportions.
-- --
Take h 2134 - 2(50) 2034 mm. Also the web must be at least '
Try one web plate 16 x 1994 mm and two flange plates 70 X 815 rnrn.
t
v 2730.4
2.034(0.35 x 250) = 15.3 mm
N-L1:
" hF,
Ihe,maximum
,.,. , , . ..,...
~ shear is at the reaction, but with a moving load contributing
.he major effects (with impact) will not change much for some, distance Step 4. Compute the allowable bending stress F,.
along the girder, so we will take t,,, = 16 mm. The compression flange will be laterally braced using knee brackets
in Fig. 10-17 at every fourth cross-beam supporting the deck. These intern
- =-= are approximately 4(0.76) = 3.04 m < 3.66 (maximum by A m ) .
1: tw 2034
16 127 << 170 O.K.
radius of the gyration of the compression flange is
'fie tentative web area
, r'
I ,
A, = 0.016(2034) = 32.54 X lop3 m2
Wsing Eq. (10-2), we find the tentative flange area as
Flange area = 0.070 x 0.815 = 0.05705 m2
16 580 32.54
'
I
A - --- - 65.21 - 5.42 = 59.79 x m2 A w - 1.994 x 0.016 = 0.0052 m~
- 2034 X 0 . 5 c 6 - - total = 0.06237 m2
6 6
i i y two flange plates 70 X 815 mrn ( t available only in 10-mm increments):
h = 2134 - 2(70) = 1994 rnm
-b -
---815 - 5.82 << - 190 O.K.
25 2 x 7 0
fi
- = - =1994 125 << 170 O.K.
tw 16
Step 2. Compute the moment of inertia of girder and determine the
a-tual bending stress fb so that we can checkf, I Fb.
The allowable bending stress is
Ix = Iweb+ 2 ~ # -
bftf3
~
12
+
- 106(0'016)(1'994)3+ 2(0.070 x 0.8 15) 997 + -
-
12
( 720)' + 815(0,070)~10~
12 Step 5. Design the bearing stiffeners.
= 10 571 + 121 510 + 24 = 132 114 x lOW6m4 We will wrap the top flange around the end of the girder web to provide
additional tension field resistance for the erid post as shown in both Fig,
 P

The,clear distance a is limited to

0.8721,
a = -------- 5 1.828 m (AREA 72-in limitation on spacing
fI
At the bearink stiffener '

'
jv=-=
v 2730.4
= 85.6 MPa < 0.355 O.K.
ht 1994(0.0 16)
This value of j, results in an allowable spacing of
0.872(16)
a = = 1.5 1 m < 1.828 O.K.
m
We can now do one of two things:

1. Space all stiffeners (but so that they come out as an integer) at

mately this spacing.
2. Use this spacing for each end part of the bridge and a Iarg
the interior part, since the live-load shear (without impac
(Table 1-2). Similarly, the dead-load shear is only 1/2 of

Let us investigate the spacing for the approximate center h

span:
Live load = 764 x 1.28 = 1234 kN
Dead load = 580/2 = 290 kN
Total = 1524 kN

and
1524
jv = 1994(O.Q16) = 47.8 MPa

We will use the spacing as shown in Fig. ElO-5c.

1

Figure E10-5c

9
a SmuC?UiUU S'TYEL DESiON
Use a pair of stiffener plates for each intermediate stiffener.
D 2134
6, 2 50 + - = 50 + -- = 121 mm use b, = 125 mm
30 30

Step 7. Design the flange-to-web weld.

The AREA specifications allow either full-penetration groove welds o
fillet welds; either must be continuous. We will arbitrarily use full-penetra
tion groove welds. For this type of weld we will only have to check the she
produced and compare it to the maximum value allowed for E70 electrod

Q = A,-- = 0.070(0.8 15) 997

( +- ;"I
= 58.88

u = - =V Q
x m3
2730.4(58.88 X
i j- s h o p weid
~ I O O IO V J ~ ~r I ou A I O U . I I 1.30

I 0.132114
= 1216.6 kN/m I Field bolt 1 1 Framing angle, designed ln Example 8-9

, The shear resistance is limited to

F, = 0.35F,ASh,, = 0.35(250)(0.016>(1rn)
---/ x lo3
-"
- / \ - "

-~-twnlu/rn~1~10.0 0.K. Figure E10-5d

Step 8. Check thc:deflection. force

. that
nssumlng . the . live-load moment is due to an equivalent uniform
loa~ d we
, obtain
8M
WL = -
L~ The number of 22-mm A-325 bolts in double shear required
I
and the total equivalent uniform load we is horizontal component of stiffener force to attach stiffener to floor
193 = 1.83
we = wD + wL = 42.2 + 8(12 780) = 177.3 kN/m N =
(0.7854 x 0.022*)(138 x 1d)(2)
27.5' @

.The.dsflection is approximately Use three fasteners (AREA minimum number of fasteners in a co

For the stiffener try a WT 385 x 73.7
A = - - =5wL4 5(0' 1773)(27'5)4
= 0.04996 m (say 50 -) A = 9.39 X m2 t, = 17.02 m m
384EI 384(200 000)(0. 132114)
d = 376.6 mm t, = 13.21 m m
27.5(1000)
Amax = = 43 mm = 50 mm (take as O.K.)
640
Step 9. Design the upper flange knee bracing.
We shall arbitrarily place a brace on every fourth floor beam f
spacing of q0.76) = 3.04 m. Where this coincides with or is "close eno
to a stiffener, we will increase the stiffener thickness from 8.5 mm to 1
For any bracing plate alone, we will use a 12-mm plate. Shop-we
stiffener or brace plate to girder web and compression flange as shown i
Fig. E10-5d. Field-weld the bottom of the stiffener brace plate to the c
beam. Shop-weld the angle for the bracket and punch for three 22- O.K.
high-strength bolts at each end. Take the horizontal component of stiffener Pa = AFa = 9.39 x 133 = 1249 kN >> 579
 -
Annvec 1638 kN m.

given in Sec. 1-9.

I
Wind on girder: .1.5(0.030 ksf)(47.88)(2.134 rn dept
, Wind on train: 0.300 klf(14.59)

i Design the channel-to-beam weld for the girder of Prob. 10-3.

An additional concentrated force of one-fourth of the heav Design the channel-to-beam weld for the girder of Frob. 10-4.
applied at each critical point;. this gives a force of
& ;
d!,S ,
1 10(4.45)(0.25) = 122.4 kN ave a separate set of design data of length depth and load. Be sure to w th

I End panel: 122.4 + 90 = 143.01 k~

P - 175 to 250 kips, L = 40 to 60 ft, and D = 60 to 96 in, as assigned. Use A-36 stre
electrodes, and the AISC specifications. Assume lateral support at the ends and load points.
2 If no specific data are assigned, take P = 2 15 kips, L = 54 ft, and D = 80 in.)
) Interior panels: 122.4 + 9.0(4.58) = 163.6 kN

-L-

welded plate,girder for a crane runway for the conditions (-urn loads o n
in Fig. P10-9. Use the AISC specifications, A-36 steel, and E7O electrodes. Limit
 STRUc'nRAL STEIiL DESIGN

gipder depth to 84 in. Lateral support is only at girder ends. Use W /

L aOft, and S = 15 ft.
19-10 Desi~na welded plate girder for a crane runway for the condit
shown in Fig. P10-9. Use the AISC specifications, A-36 steel, a
*dtr)
bC
(,vt, @der depth to 2.15 m. Use W/4 = 110 kN, P/2 = 950 k~
D ~ S ~a W
' welded plate girder for a single-track open-deck railr SELECTED COMPUTER PROG
,rrack contributes 0.25 kip/ft loading, including rail, ties, and so on.
loading and span from 40 to 110 ft as assigned. Use A-36 stee
s~--.zications.Limit the overall girder depth to L/10 2 D 2 ~ / 1 5 ,
:t. Take the girder spacing S 2 L/15.Design sway bracing and/or diaphragms in

-
stby flanges, stiffeners, and welds. (Note: If specific problem data are not assigne
E-1 i0 loading, L 90 ft, D = 96 in, and s = 6.5 ft.)
&-12 Denim a welded plate girder for a singletrack open-deck railroad bridge. A
tr7;k contributes 3.65 kN/m of loading, including rail, ties, and so on. ~~~i~~for
C-l live loading and span from 14 to 34 m, as assigned: Use the

-
a-d depth D but limit D to between 1.22and 2.45 m. If specific problem
CooPr E-80 loading, L 23 m, D = 2.3 m, and S = 1.6 m.

. ,.,. .

A-1 FRAME ANALYSIS PROGRAlM

This computer program will analyze any plane The frame may be
pinned, or a combination. It must have a constant modulus of elasticity, E.

ated loads are on the member. The program computes fixed-end

r a uniform beam loading (and orientation from horizontal to vertic
so computes fixed-end moments for up to two concentrated loads on

The following steps are required to run the program, with tenninol
matching the computer program listing.

1, Code the structure according to rigid frame, truss, or combination- T

account any hinges, as in chap. 2.
Determine NP. The computer program computes NPPl = NP 1-
 ,, \$.510 STRUCTURAL STEEL DESIGN S E U C I ED
i I

1 = NP value of the entry

5. Determine NLC that uses wind the first time as.NLW. 1 , ~ L) ~
= 1,NLC This inputs the NLC nodal forces at
6. Determine NLC that uses D + L, which is to be later comb for NLC load conditions using 8FlO.
Note that this requires at least two cards
NNZP.
NLC that uses wind. 5. Add your system control cards.
7. Determine the NLC that has wind on any beams or columns. Call this
Note that you may (but not recommended) read nodal moments
hove. If you do read nodal moments, the output moments must be Correc
values, which should be only wind. and to get the final design moments as
8. Find NBAND using the method outlined in the text.
Ffinal = F ~ ~ t p +
~ tMrrad
9. Make sure that ISTIFF is dimensioned greater or at
careful attention to signs. The output signs are interpreted using Fig
e program computes the fixed end moments, the output is automati
corrected for the FEMs to obtain the design values. Member forces do not
to be corrected for any node forces read in other than moments-

1. Punch cards for lines 4, 8, and 10 of computer listing. f input units are as,identified here, the output will be
' 2. Punch member data cards according to the FORMA X = displacements = in, mm, or rad
MAT(7 14, 4G 10.4, 12)
The G format allows use of either F10.4 or xxxE-3 or xxx F = member forces
NPE(7) uses the I2 format and goes in column 70. Truss members: kips or kN
Use NPE(7) = 1 for all members without transverse loads. Beam members: kips or kN for axial for
For any members with transverse loads: f t . kips or kN m for mom
NPE(7) = 0 for normal gravity beam loadings.
NPE(7) = - 1 for wind loads on beam or transverse loa
columns.
entification of Program Variables
any alphanumeric data using not more than SO c
Recommend using your name, problem number, o r
, UNIT3, U N I T 4 units identification for fps an
of gravity
X 1, X2 = horizontal distance from origin end of member of P1 a FT or M (start in column 1 of card)
or m UNIT4 I N or MM (start in column 5 )
Vl, V2 = column shears from alternate direction for a space frame, V1 = UNIT2 K-IN or K-MM (start in column 11-unit used for
near end shear; use kips or kN and sign same as P1, P2
3. Put a blank card at the end of the member data. K-FT or K N . M (start in column 21)
number of P's coded = 2 x number of nodes - number
reactions
 JllAL STEEL DESIGN SELECTED C O W

number of members (NM = N P for determinate trusses

number of load conditions
1 to list band matrix (part of total AS
finding X's)

..
~LW; first load condition with wind
IWINRE COUNTEL! TO ADO WINO LOAD OF&r'$E>ll
NNZP number of nonzero P-matrix entries to IOLPL C O U N T E R F O R D E A D + LTVE.C?~O~-LI(:
0 for fps problems = 1 for SI problems
1 for all truss members; = 0 if any
UNIT^
UNIT,,
.. K OR K N
I N OR
-- UNITZ . IN-K OR UN-PM .AU U I T 3 $ 1 4 011 UN-
J " , -Y9
9

11M -- * * * NOTE C ~ E F U L L YU N I T S OF U M I T ~ I IFOR I N P U

present * ,

band matrix and must be dimensioned NP x NBAND at

least
band width used. Is difference betwe
NP on ends of any member? Valu
needed but not larger than NP. If valu
numbers too small, tends to give a
sometimes-other times gives overflow messages.
NLC with D + L which is saved on disk file to add to win
* NLC
+
NLC which uses D L from disk file with current force
lrPMIN(I15I
1 to write X (deflection) matrix
0 to obtain output with few members DATA F F U 1 1 2 ~ r 1 0 0 0 . r 1 . ~ 1 0 0 0 . ~ 1 ~ ~ 1 0 0 0 0 0 0 0 1 1 2 ~ t ~ ~ ~

1 to obtain output when large number of NLC a READ CARD

modulus of elasticity, ksi or MPa 2 R E A D ( ~ , ~ ~ O O ) L Z Z , U N I T I ~ U H I T ~ ~ U ~ ~ ~ Z ~ U ~ ~ T ~

member number assigned to element during cod 000 F O R M A T ( Z O A ~ , I , A Z S Z X S A Z ~ ~ X ~ ~ ~ A ~ S ~ X ~ )

NP values for the element; truss has 1 through IF(EOF(lll750~3
3 WRITE ( 3 r Z 0 0 0 ) ZZZ
through 6 2000 F O R M A T ( ~ ~ ' J ~ ~ S ~ X ~ Z O A ~ ~ ~ ~
horizontal distance to far end of member including si REAO CARO
or m R E A O ( ~ . ~ ~ ~ ~ ) N P J N ~ ~ ~ N L C P L I S T R , N L Y ~ N N Z P J I ~ E ~ #
~IDLPL,IYINBE~IWRITX~IW~ITP
vertical distance to far end of member
1 0 0 1 FORMAT(lbI5l
m
cross-sectional area of member, in2 or m2 L++*+* REAO MODULUS OF E L A S T I C I T Y - K S I
READ(lr1002lE
OR B P I

moment of inertia of member, in4 or m4 (not need

1 0 0 2 FORMAT(BF1O.C)
truss members) - *
control switch for members with trans ***********
-
D E F I N E U N I T CONVER$ION FACTORS
IUNIT 1
0 for members with transverse loading
1 for members without transverse loading
MENT cards in computer program)
FU(I)
4 0 CONTINUE
-
IF(IMET.GT.O)IUNIT
0 0 4 0 1 . 1 ~ 6
FFU( I U M I T s I I
2

BE SURE TO PUT BLANK CARD AT NPP1 NP 1

BEK DATA
NP value of P matrix entry where a con
moment) is located
value of corresponding P-matrix entry. The
are in a DO loop, so use as many pai
1
I
I
 5 STRUCTURAL STEEL DESIGN
SELECTED CObIP

FASAT(Ze3) = -EASATII.Z)
FASAT12s4) = -EASAT(2,2) 2 0 4 CONTINUE
EASAT(391) = -EASATf1.1) DO 2 0 6 1 = l r 6
E d S A T f 3 r 2 ) = -EASAT(1,2) IFfUPE(I).GE.NPPl)GO TO 2 0 6
EASAT(3131 * EASATf1.1)

-
EASAT(3r4) = EASATf1.2)

-
EASAT(4rl) -EASAT(I.Z)
FASATf4.21 = -EASATf?,Z)
EASAT(4.31 EASAT(1.2) NS2 = N P E f J l -
NPE(II + 1
S T I F F I N S 1 4 NSZ) * STIFF(NSI+NSZ) t EASAT(1,J)
E A S A T ( 4 * 4 1 = EASAT(2.7)
0 0 1 7 9 111.4 2 0 5 CONTINUE
IF(NPE(I).GE.NPP~IGo
N t l = (NPE(I )-~)*NRANO
TO 1 7 9
5 5 0 0 GO TO 1 0 6 . I

00 178 J = l r 4 c
I F ~ N P E ( J I . G E . N P P I ) G o TP 1 7 9 c THE B A N D M A T R I X IS N O W F O R M E D F O R R F O U C T I O N IN ISTIFF IN COPE
~F(NPE(J).LT.NPE(I))G~ rn 178 C WRITE BAN0 R A T Q I X I F L I S T 4 > 0
- c
NSZ
STTFF(NSl
NPE(J)
NS2) .
NPE(1) t 1
S T I F F I N S I t N S Z ) + EASAT(I,J) 72R IF(I1.GT.O)GO TO 6 0 1
'
1.9
7 8 CONTINIJF
CONTINLJF
G n TO 5 5 0 0
C
C
H2 . NSANO*NP

1 0 5 EAOL = E * A / ~ L IF(LIST8.LE.O)GO TO A 9 R 9
E I O L = (E*xIIxL)*Fu(?) WRITE(3r2009)

COSOL -.
S I N O L = XSINIXL
XCOSlXL
2 0 0 9 FORMAT(
ni = 1
--
1 1 0 1 ' T H E RAND " A T Q I X W I T H 1 0 0 0 F A C T O R E D ' n I l

CA(1.l)
EAlloZ)
EA(lv3) = 01
. 0.
1.
R2
MCOU
On 3 0 5
N8ANO
NP*NBAHO
I = 1.NP
FAf2rl) -XCOS W R I T E ( 3 r Z O l O ) I r ( S T I F F 1 II?.;~.MI,MZ 1

---
FA(2rZI SINOL 2 0 1 0 F O R M A T ( ~ X I I ~ , ~ X I - ~ P ( ~ F 11 ~ " ~5 .X ~~ )- 3~ P ( 9 F l Z . 2 ) ~ / , 5 X 1 - 3 P ( 9 F 1 2 . 2 ) 1
Eb(2r31 SINOL 1FIMZ.GE.NCOUIGO TO 8 8 4 9
EA(3.1) -XSIN 11 = M2 t 1
EA(3.21 -COSOL r M2 = M2 t NBANO
E A ( 3 r 3 ) = -COSOL 3 0 5 CONTINUE

EAf4.2)
EAf413)
--
E A ( 4 . 1 1 = 0.
0.
1. C
C
8 8 8 9 WRITE(3.20111MZ~NRANO
2 0 1 1 F O R M A T ( I / ~ T ~ I ' M O S T I F F ( 1 I ENTQIES . ' ~ I ~ . ~ O X P ' S A NY TOD T H . ' r I 4 t l / l

EA(5.1 = XCOS
FA(5.2) = -SIYOL C NOTE--DO NOT READ P - U A T P I X E N T Q I E S COW UNFOQM LOADS OM 3 E I N S
EA(5.31 -SINOL C DO NOT READ P - M A T Q I X E N T R I E S FOR FEM--INOUT SO COUPUTEQ
EA(6.1) = XSIN C COHPUTES F E * 5 0 F I N A L MOMEMTS A Q E COQDECTEO FOP FEM * * * * * a *
E b ( b t 2 l = COSOL c
ES(lr1) -
E A t 6 r 3 ) = COZOL

E S ( 1 ~ 2 1 = 0.
EAOL
c
IF(NNZP.EO.OIGO TO 4 2 6
0 0 9_91 NN = 1,NNZP

ES11.3)
ET(2.1)
ES12.2)
--= 0.
0.
4.*EIOL
C
READ( l r 1 0 1 0 l t ! l

1 0 1 0 FORMATf1615)
ES(2.3)
ESf3rl)
ES(3.2)
- = Z.*EIOL
0.
= ES(2.31
C

C
READ(1,1011)(PRfLI~L*l~NLC)
READ CARD******

ES(3.3) = ES(2.21 1011 F O R M A T ( 8 F 1 0 . 4 1

0 0 202
DO 2 0 2
J
E S A T ~ I P J )= 0.
. I = 1.3
1.6
OD 9 9 0 L = 1,NLC
P(n1.L)
9 9 0 CONTINUE
= Pfn1.L) + P R ( L )

0 0 1 8 6 K 1 1.3 9 9 1 CONTINUE
ESAT(I*J) * ESATfIpJl + E S f I ~ Y ) * E A f J s ~ ) 4 2 6 CONTINUE
1 - - CONTINUF DO 9 9 2 NS1 9 l r N P
to.? CONTINUE DO 9 9 2 N S 2 9 l r N L C
9 9 2 T F f N S l r N S Z l = 0.
IF(II.GT.09GO TO 6 0 5 c ************~*****;~******T~***lOtOOO~OOO.**
,.
203 0 0 204 I = l,b L

0 0 204
EASAT(1.J)
J = 1.6
-
0.
DO 1 8 7 K 9 1 . 3
4 0 6 URITE(3,2012)UNITl~UNIT2
2 0 1 2 F O R M A T ( / * Z X s ' T H E P-MATPIX,
NS1 = 1
' r A 2 r 1 AN0 ' , A 4 r / )

E A S A T f I # J ) * E A S A T f I p J ) + EAfI,K)*ESATfK.JI NS2 = 1 0
1 8 7 CONTINUE 4 2 7 IF(NSZ.GT.NLCINS2 = NLC
DO 4 0 8 I = ltNP
408 V P I T E ( ~ ~ ~ O ~ + ~ I ~ ( P ( I ~ J I ~ J J N S ~ ~ N S Z ~
5 1%,STRUCTURAL STEEL DESIGN

IF(NS2.EQ.NLC)GO
N S l = NSZ + 1
-
i 0 1 4 F O R M A T l T 5 r 'MP ',13~1X,lOFll.2)
TO 4 2 8
I
NS2 9 NS2 + 1 0
URITE13r2016)NSl
2 0 2 6 FORHbT(11,5Xr1THE P-MATRIX CONTINUED BEGINNING U I T H NLC
GO TO 4 2 7
6 2 8 CONTINUE 602 IF(ITRUSS.EO.O.OR.IWPITP.E~~~)WPITE(~~~O~OIJJPUNIT~~UNIT~
c
C'
CO N 1
0080
1'.
SUBROUTINE TO REOUCE THE RAND MATRIX

N-1pNP
2 0 2 0 F O R M A T ( I P 4 x 9 'LOADING CONOITION NO 9 ' r I 3 r l r
1 37X. 'DESIGN END MOMENTS C0RRECTEO'r/, 5Xr *MEMBERit 4 Y p ' A X I P L
7FOWCFs ' r A29 8Yr'FOR FFM AN0 WINO (NEAR FNO F I P S T l r ' , A L P I ~ ,
RFWINO 5

-
I - N
DO 7 0
NL
I = I+1
L = ZtNBANO
(N-l)*NBAND + L
GO T O 5 5 5
6 0 5 P ( N P P 1 r JJ).O.
FEU1 = F M I I J J I .
-
IFISTIFF(NLI.EQ.O.IG0 TO 7 0
C NOTE I F YOU PEA0 D-HATQIY ENTRIES FOR FE* THEN OUTPUT *UST 5 E

-
8 STIFF(NL)lSTIFF(Nl) C AOJUSTEO BY HAHO TO ACCOUNT FOP F I X E D END MOMENTS DM 9 E A M ENDS
J - 0
00 68 K L,NBANO
J = J + 1

-
I J = 11-l)*NBAND 4 J
NK (N-lI*NBAND + K

67
70
00 67 H

CONTINUE
-
INCLUDE LOAD MATRIX I N REDUCTION
1,NLC
P(IsM) = P(IPM) -
B*P(NrM)
c
6 0 7 CONTINUE
Fl2l-(F(ZI-FEMlI/FU(l)

C REDUCE MOMENTS FOP WINO

0 0 66 M = 1,NLC C ALSO NECESSARY TO REDUCE 0 t L MOMENTS B Y 0 . 7 5 I F AOOEO TO WINO
P(H,M) = P(NrM)lSTIFF(Nl)
66 CONTINUE
80 N 1 = N 1 4 NBAND
C COMPLETE SOLUTION BY BACK S U B S T I T U T I O N
F(Z)-F(Zl*3~/4.
F(3l*F(31*3.14.
IF(N.LE.O)GO TO 9 0
IF(JJ.NE.IDLPLIG0 Tn 973
L = N - 1
0 0 06

-
K = 2, NBANO
NK * ( N - l ) * N B A N D 4 K
C
00 970 I
9 7 0 TF(YEMNO.11
-
PUT 0 t L I N T F - M A T Q I X
193
* F(I)
TO LATER A D O TO Y I V O L O I D COHOIT.

0 0 86 H 1pNLC
IFfSTIFF(NK).NE.O.)P(NrM)
9 6 CONTINUE
= P(NtM) - STIFF(NK)*P(L*KpM)

GO TO 8 5
\fi CONTINUE
C
t ENO OF MATRIX REDUCTION -- SOLUTION IS IN P(I.JI 1.58 CONTINUE
I F ( 1TRUSS.EQ.O)GO TO 1 8 9

I C 1 7 F O R M A T ( I I r 5 X r ' T H E X-HATRIXp ',A29' OR RADIANS',/)

NS1 = 1
NS2 .
10
4 2 9 IF(NSZ.GT.NLCbNS2 = NLC
DO 5 0 3 1 = lrNP IF(NLC.EO.1)GO TO 1 9 5
503 U R I T E ( ~ ~ ~ O ~ ~ ) I ~ ( P ( I ~ J ) P J * ~ ( S ~ , N S Z )
L O 1 8 FORMAT(6X,'NX
IF(NSZ.EQ.NLC)GO
9 *rI3~1XrlUFll.5)
TO 4 3 1
IF(F(1I.GT.PMAX(ME*NOIIDMA~(MFRNO)
IF(F(1I.LT.PMIN(MEHNOIIPMIN~~EMNOI
TF(HEMNO.JJ) F(1)
-
* F(11
F(11
N S 1 = NSZ + 1
NS2 = NSZ 4 1 0 c
I F ~ I W R I T P . L E . O ) G O TO 4 4 1
WRITE 9AR FORCES FOR NLC 9 1 AND r V Q I T P . 1
~ ~ 1 T E 1 3 r 2 0 1 9 ~ ~ S l
2 0 1 9 F O R M A T ( l l r 5 X r ' T H E X-MATRIX CON1 D BEGINNING WITH NLC
IF(MEHNO.EO.~IURITE~~PZOZ~~JJ
FORCES [ K I P $ O Q K N I FOR NLC
2 0 2 5 FORHAT(5X,'BAR . 'rI3*11
4 3 1 CONTINUE
C INCREMENT COUNTERS I 1 AND J J - - J J COUNTS NL
 . .
5U) STRUCTURAL STEEL DESIGN SELECfED COMPUTEX

JJ = JJ 4 1 A-2 LOAD MATRIX GENERATOR FOR AASHTO TRUC

C
195 JJ 1 -
GO T O 1 8 9
LOADING ON A TRUSS BRIDGE
1 9 1 GO T n 5 5 5
1 9 2 I F ( I T R U S S . E ~ . O I G O TO 1 9 3 This program steps a movlng load consisting in either two or thre
IF(NLC.EO.l)GO TO 1 9 3
loads along a truss. At each step the panel loads are computed as
4 C
REWIND I
I T R U S S = 1 AND N L C I S L A R G E U S E I W R I T
B A R F O R C E S ONLY---USE
.
O R M A X AND IN
0 FP
I W R I T P = -1 F O R C O ~ P L E T E L I S T I N G OF
with concentrated loads. These panel loads are the P-matrix entries f
B A R F O R C E S WHEN N L C IS L A R G E R THAN 1 weightless truss to use in the frame analys~sprogram.
The loads move from left to right. If the larger wheel loads ar
C
VRIT~(3,2033)
truck is "backed" across the truss. The llmtatlons are:
2 0 3 3 F 0 R M A T ( 1 1 ' 8 X * ' T ~ ~n A Y I n U n L I V E L O A D B A R F O R C E S & N O D E A D L O A D v A L U E
1S"lr5Y'qnEf4 N O ' P ~ X Y ' M A X L L 1 , 9 X y * M f N L ~ q , f j x ~ rL O~ A~D ~ORo L ~ S TN L C
1. The user must use at least two and not more than three loads-
z',/T
0 0 447 I . lrNn
~ R I T ~ ( 3 r 2 0 3 5 ) I ~ P n ~ ~ ( ~ ) . ~ n ~ ~ ( ~ ) , ~ ~ ( ~ y ~ ~ ~ ) 2. The "step" must be an integer multiple of the panel length. Since this

c
4 4 2 CONTINUE 7
FORHAT(6X*T ~ ~ X ~ F ~ O . ~ Y ~ X , F ~ O . Z ~ ~ X , F ~ O . ~ ]

W R I T E B A R F O P C F n A T a I v IF I W R I T P -I
IF( I W R I T P . N E . - ~ ) G O T O 1 9 3
.
is computed, the step should be input suffic~entlyaccurately and
smaller so that in computer truncation the correct integer is obtained.
3. All panels must be of the same length.

An impact factor is automatically computed and the panel 1

WRITE ( 3 1 z 0 3 z ) ~ ~ 1
2 0 3 2 F O R n A T ( I I * 5 X ~ ' T B~ A~R F O R C E M A T R I X
4 4 3 DO 4 4 4 1 . 1 , y n
S T A R T I N G WITH NLC . qy13y11, impact factor in the output. MET is used to compute the impac
444 ~ R I T E ( ~ ~ ~ ~ ~ ~ ) I I ( T F ( I ~ J I ~ J ~ N ~ ~ ~ N ~ ~ )
2028 F O R ~ A T ( ~ X ~ I ~ , ~ X Y ~ D F ~ ~ . ~ )
IF(NSZ.EO.NLC) GO TO 1 9 3 Output is automatically punched onto cards m the correct format for
NS1-NSZ41
NSE=NSZ+~O input as the P matrix in the frame analysis program. The user may wish to
IF(NS2.GT.NLC)NSZ.NLC
URITE(3r2032)NSl
an optional punch control to inspect the output prior to having it punche
GO T O 4 4 3
1 9 3 GO T O 2
C
7 5 0 STOP
Variable Identification
END
A sample set of data cards are llsted ulth thls program for both an fps and an
bridge truss using three wheel loads, seven truss panels (25 ft or 7 5 m),
heels 5 ft or 1.5 m. The coding is such that the panel NP
are 26,2, 6, . . . and NPPI = 26 here and is not punched
number of wheel loads
number of truss panels
step length, ft or m
STEP increment of wheel movement, f t or m
MET 0 for fps; = 1 for SI bndges. This is used to proper1
impact factor.
wheel loads. Note the order of using loads on the sample cards. U s h g
8. 32. 32. runs the truck forward on the bndge.
X(1) cumulative wheel spaclng as 14. 28. for using 14 ft between each set of
wheels. This spacing or the SI equivalent wdl usually be the most
critical spacing.
Ml(1) NP numbers in order from left to nght where the panel loads are
placed. M1 has the same meanlng here as in the frame analysis
program. Note that NPPl is not used. For a through deck truss, this
section of the program requlres a sllght modiflcatlon to get loads onto
the truss where NPPl may be, since the current method of use omits
the flrst and last Ml(1) values (they are input for ~dentdication).
 922 S T R U C T U R A L STEEL DESIGN

r NP IS THE NUMBER OF LOADS THAT A R E WANTED ON THE TRUSS SL IS THE 2 0 1 CONTINUE

I F (XIMPAT.GT.0.30) XIMPAT-0.30
C LENGTH OF EACH PANEL I N THE TRUSS NS I S T H E NUMBER OF PANELS I N THE TRUS
( STEP I S THE MOVEMENT THAT I S WANTED FOR THE LOADS BETWEEN EACH LOAD 00 1 0 3 I-1.MZ
r- CONDITION P t I ) I S THE WEIGHT OF EACH LOAD I N THE S E R I E S W 1 0 3 J-1,NSPl
X t I I I S THE O f S1 ANCE THE LOAO I S B E H I N D THE F I R S T L O A D 103 REACT~J~II=REACTIJ,II*~~+XIMPAT)
' MET=O.O FBS ~ 1 . 0 M E T R I C WRITEt3r4)
D I M E N S I O N T~MPt20~70)rREACTtZO170),Pt5lvXt5Is M l ( 2 0 1 4 FORMAT( I / / / / *LOAOm,/* ' CONOITIOW9.TZO.' THE R E A C T I O N S ' )
0 0 1 0 6 J=11M2 a
1 FORMATt215r2Fl0.4,15) 1 0 6 W R I T E t 3 . 3 1 Jt ~ R E A C T ~ I I J ) ~ I - ~ , N S P ~ ) "
i, WRITEt3v5)NPrSL,NSvSTEP 3 FORMATtT5,13qlOt5XvF6.1)) *. t ,
C PUNCH M l FROM L E F T TO R I G H T ACROSS fk T R U S S "
5 F O R M A T t ' l N U M B E R OF LOADS ON T R U S S 1 ~ 2 X , 1 3 ~ / / ~ ' SPAN LENGTH O F EACH PUNCH FORMAT I S COMPATIBLE W I 7 H A'NALYSIS PROGRAM FOR DIRE$$ INPUT
lPANEL'vFB.2 v//v' NUMBER O F PANELS I N TRllSS'v2X113. I/.' LENGTH OF C
2MOVEMENT OF THE LOA0S1,F8.2 r//)
REA0(1.81(MltI)rI*lrNSPl ) ,
DO 2 0 I = l . Z O 8 FORMAT( 2 0 1 4 I
DO 2 0 J = 1 . 7 0 00 1 1 2 I=Z.NS
TEMP( I,J)=O.O WRITE(2v9)MltII
70 REACTtI.Jl=O.O 9 FORMATt15)
DO 1 0 0 1 - 1 1 5 112 W R I T E ~ ~ ~ ~ O ) I R E A C T I I ~ J I ~ J J ~ ~ M ~ I
Xt Il=O.O 10 FORMATtBF10.41
I L') P t Il=O.O
NLS=SL/STEP 4000 STOP
MX=NP-1 EN0
NSPl=NS+1
NSMl=NS-1

'
R E A O ~ ~ ~ ~ ) ~ P ~ I ) ~ I = ~ ~ N P ) ~ ~ X ~ J ) I J = ~ ~ N X )
FORMATt1OFf3.21 C **** FOLLOWING 3 CARDS ARE SET FOR FPS OUTPUT
3 7 25. 5. 0
WRITE(~~~)~IIP~I)~I=~~NP) 14. 28.
b FORMATt///.* P'.Ilr' =lrF8.3) 8. 32. 32.
26 2 6 1 0 1 4 1 8 22 26
WRITE~~.~)~JIX(J)~J=I~NXI
7 FORMATl///r9 X 1 1 I 1 v ' =',F8.31 C **** F O L C M I N G 3 CARDS ARE SET FOR S I OUTPUT
3 7 7.5 1.5 1
SK=O.O
SK2=0.0 36.0 142.0 142.0 4.25 8.5
DO 1 1 0 M = l v N P 26 2 6 1 0 1 4 1 8 22 26
DO 1 0 9 I = l r N S I*
IPl=I+l
DO 1 0 8 J = l , N L S
RR=P(M)*tJ*STEP-SK-SK2)/SL
I ' ALR=PtMI-RR
I F .LJ*SJEP+STEP .GT. SLI SK-SL-J*STEP
JZ= tSL/STEPI*tI-11
J3=J+J2
TEMPERARY STORAGE OF R E A C T I O N VALUES
TEMPtIvJ3)=A1R
TEMPfIPlrJ3)-RR
1 0 8 CONTINUE
1 0 9 CONTINUE
L PLACE CONSECUTIVE LOADS I N T H E I R PROPER P O S I T I O N
IX=XtM)/STEP
IST=IX*STEP
SK2=Xt M I - I S T
INT=O
I F tM.EQ.1) GOT0 1 0 5
DO 1 0 7 I=lrNSP1
DO 1 0 7 J = l r J 3
n n i = n-1
INT=XfMMl)/STEP
M2=J+INT
C ADO TEMPERARY STORAGE TO PERMANENT STORAGE
107 REACTII~MZ)=REACT~IIM~)-TEMP~I~J~
GOT0 1 1 0
105 00 104 I=ltNSPl
I
DO 1 0 4 J = l r J 3
1 0 4 REACTtI.J)=-TEMP(1.J)
1 1 0 CONTINUE
IFtMET.EQ.1IGOTO 200
XIMPAT=5O./tNS*SL*125.1
GOT0 2 0 1
700 X I M P A T = ~ ~ . ~ ~ / ~ N S * S L + ~ ~ . ~ ~ I
 5k-4smUclVbU.. STEEL DESIGN SELECTED COLLPUTER. PROG
"ft ,$i
A-3 LOAD MATRIX GENERATOR FOR
LOADING ON A TRUSS BRIDGE NOSPAN =
NO OF SPANS OR GIRDER SEGMENTS: ISUIT 1 FaR M E T R I
c = 0 FOR FPS: OX = INCR. OF LOAD MOVEMENT. F T OR M; SPAN LEN
FAC = L O I D R A T I O I F F P = € 8 0 LOAD
c TRUSS SPAN OR GIRDER SEGMENT:
n l s program steps Cooper's E-80 train lo c F A C = 0.75 FOR E ~ W - 0 . 9 FOR E72--1.00 FOR E B O - - L . ~ ~F~~ ~ ~ 1 1 0 -

(DX) the panel loads are computed as a WIOH CENTER T O CENTER SPACING OF T R U S S E S
loads are the P-matrix entries for a weight1 IWRIT = 1 TO WRITE SELECTED COMPUTATIONS FOR DEBUGGING '

Program. Note that the program does not

bl;idge,pnd only the uniform load is on.
The only limitation to this program is
The Cooper's E-80 loads and wheel sp
rated into the program as DATA FFP/
n

Variable Identification
(see sample data cards listed at end of program listing)
TITLE up to 80 columns of alphanumeric data for probl
NOSPAN number of panels ~ E ~ 0 1 1 ~ 1 0 0 2 1SPAN,UIOHHFAC
0 X ~ '
ISWIT 0 for fps; = 1 for SI prob 1 0 0 2 FORMAT(BF10.41 N P S ( J ) = NP N U M B E R S F O R NON-ZERO *-MATRIX ENTRIES

spacing to SI and to NSPN = NOSPAN 1 -

REAO(lr100611NPSIJI .J=lqNSpNI
IPUNCH 0 for not punching O U ~ P U 0 6 FORHATl1615)
= 1 to punch output in f
frame analysis program I F 1 1SWIT.GT.OIDIS = 25.
8 = NOSPAN
for no impact factor; TOTSPN = B*SPAN
SET WHEEL LOAOS AND WHEEL SPACING
to write selected inter SUHXL = 0.
am; = 0 when pro 0 0 5 I-1.18
X I 1 I = FFXI I 1
crement of wheel movement left to right, f t or m IF(ISWIT.GT.OIXII I = x111*0.3048
P I 1 I = FFPI II*FAC
ISPAN panel length, ft or m F
I( ISWIT.GT.OIPII I = PlI)*4.44822
WIDH width between two trusses making up bridge an SUMXL = SUHXL + XI T I
5 CONTINUE
computing the impact factor, ft or m
FAC factor to convert E-80 to E-110, E-60, etc.; use F
E-80 loading
NPs(J) NP numbers (same a
established (note that these entries do not inc
cation may be required for a through deck b

w. @i 2000
7 PAC = 1. + PAC
~ ~ ~ ~ ~ ~ ~ , ~ ~ ~ ~ I T I T L E . N P , H O ~ P A N , O X , S P A N ~ ~ L O A O ~ S U ~ X L ~ P A C ~ F A
F O R ~ A T ( / / / ~ ~ X , Z O A ~ , I , ~ X ~I ' N* PV I ~ V ~ X * ' N O S P = ~ ' N* 1 5 * 3 X *
1 2 x 1 I D X = o r ~ 7 . ~ l / r 5 ~ , ' S P A N LENGTH = ' . F 8 . 3 r 3 X v ' U N I F LOAO/LENGTH
i$ 2~,~6.l, /, ~ X , * S U H OF X ( L 1 OF CONC. LOADS a ' t F ~ . Z I ~ X V ' I M P & C T
~ A C T D R= * , ~ 6 . 3 , / , 5 ~ , IFACTOR FOR E-LOADS--1.0 FOR E - 8 0 = ' *
465.3, ~ x , S O I S T C-TO-C OF TRUSSES 'rF7-31//)
%

"r,Tt,
 5Ad S T R U C T U R A L STEEL DESIGN

URITE13r2005~1P(Il,I=1,NP)
= 'rl2F9.l~lrl2X.
,145 F O R M A T I S X v ' P l I i 6F9.1.//)
F I N O NO OF LOAOS IN. CURRENT SPAN
* WRI T E ( 3 r 2 0 0 6 )
2906 FOPMAT(lOX.'*** LOAOS M I 0 IMPACT--USE FOR FLOOR BEAM D E S I G N AN0 CH
DO 9 5 L * NCIZOO
IFIL .EQ.NP+I.ANO.AOXH.LE .TOTSPNIXILI = Z*SPAN -
SUMXL
E N * ,
IFlL.EQ.NP+1.WO.AOXH.GT-TOTSPNIXIL) = O I F F + Z*SPAN - SWXL
r
IFIL.GT.NP+l.OR.XlLI-GTTSPANIXIL) SPAN
4 j4$!JZSTOP = TOTIP. + SUMXL + o x
0 0 8 3 K=1,20
'
IFlXfL)..LT.O.tGO
IF(XIL).NE.SPANIGO
0 1 = SPAN
TO 1 4 8
TO 2 2
DO 7 3 L = 1 . 2 1 0
7 ' PR1K.L) = 0, DZ = 0.
I., i 2 2 sun1 = sun1 + XILI
~ ' iCONTINUE
IFlSUMl.GE.01-0.0001)GO TO 9 6
IFIL.EO.NP+lIZZX -
= SPAN (OZ + 0 . 0 0 5 )
IFIL-€O.NP+1~ANO.SUMl.GE.ZZXlGO TO 9 6
****** M A I N DO LOOP FOR INCREMENTING LOADS 9 5 CONTINUE
DO 1 4 0 I = 1.210
A = l 96 C O N T I N U E
SUM3 = SUM3 + SUM1
D I F F = 0.
ADX = A*OX TOTSUM = SUM3 + SUM2
N C = L * l

1
AOXH AOX
02 = 01
C
TO STOP COMP T A T I O N S WHEN ONLY UNIFORM LOAD ON B R I D G E
IFIA0X.GT.ZSTOP
IFI AOX.GE.TOTSP
SUM = 0.
GO TO 1 4 1
I A O X = TOTSPN
RR = 0.
R L = 0.
I F 1 IWRIT.LE.OIG0
. TO 5 5 0
u R I T E I ~ ~ ~ ~ ~ ~ ~ T O T S P N ~I MNr C 0~2 A 1 O O lXt HS~U M l t O Z t NNv L.SPAN.SUH
1 7 SUM21 O I F F t T O T S U M v J J v X 1 1 9 1 ~ X 1 2 0 ) ~ X 1 2 1 1 t X f L ~
3 0 0 5 FORMATl3X.'TOTSPN = 1 , F 8 . 1 , 3 X ~ e N C = ' 1 1 5 r 3 X , 1 A 0 X X H = *,F7.3,
2 3 X V 1 1 M = l.15, 5 X v e O Z = 1.F7.312X.'01 = '~F7.3,' SUM1 *@,F7.3.
Y 3/r5X* 'OZ a ' ~ F 7 . 3 r 3 X ~ ' N N - ' ~ 1 5 . 2 X , v L ~ 1 , 1 5 t 2 X T * S P A N = ~ , F 7 . 3 ,
F I N O NO OF SPANS TO USE
, 5 SPUSO = A O X l S P A N 4 3 x 1 'SUM3 = ',F7.3,3X,'SUMZ = .F7.3r3X1'01FF~',F7.3,3Xv 'TOT
C
c ZERO MATRICES
I
ARBITRARY AMOUNTS
5'rF6.lr2Xv'JJ~',IZ1.
6F7.3v3XvQXL='.F7.31II
/r5X1 'XI9 ~ ' ~ F 7 . 3 ~ 3 X ~ e X 2 0 = 1 ~ F 7 . 3 3 3 X ~ * X 2 1 = * 7

B DO 8 6 K = 19.200 CONTINUE
X I K I = 0. ACCUMULATE EFFECTS OF LOAOS I N U I Y SPAN
0 0 9 9 L L = NN1L
b CONTINUE IFILL.LE.NP)GO TO 9 7
1 L
P I L L ) = WLOAO*XlLLl
i 0 0 8 8 K = 1,50 IFILL.EQ.NP+l.OR.NN.EO.L)OZ -
= 02 X1LL)IZ.
T I K I = 0.
9 7 RR = O Z * P f L L ) / S P A N
CONTINUE
i
RL = P I L L 1 - RR
I M = SPUSO
RM = I M
PTIKC) = PTIKCI - RR
PTIKC-21 = PTIKC-21 - RL
IFI8M.LT.SPUSOIIM = IM + 1
BM = I M 02 = 02 - XILLI
SUMP = SUMP + P I L L )
I 1FIAOXH.GT.TOTSPN)OIFF = AOXH - TOTSPN 9 9 CONTINUE
;UM2 = 0.
b ? F I A B S ( O I F F ~ . L E . O . O O ~ ) G ~ TO 1 5 SUMPT = SUMPT + P T I K C )
1FlJJ.EO.IM)SUMPT = SUUPT + P T I K C - 2 )
4.. Z = 1M - JJ
DO 1 2 MM = l v N P
>OM2 = SUMZ + x t n n )
OZ = Z*SPAN + TOTSUM - AOXH
NN = L L
:~lSUM2.GE.OIFFlGO TO 1 5
SUM1 = 0.
1 2 CONTINUE
1 0 2 CONTINUE
t K NC = tin + 1
fin = I M WRITE ONLY V E R T I C A L VALUES AT PANEL P O I N T S I H O R I Z = 0.) SO CAN
6 7 = BM*SPAN -
AOXH + SUM2 U S E P R I N T E D OUTPUT TO OESIGN TRANSVERSE FLOOR BEAMS
BASED ON LARGEST PANEL LOAD * FLOOR BEAM REACTIONS
'WR = NC
LUM1 = 0. AN0 DEPENDING ON NUCBER OF TRACKS ON BRIOGE ** OOES M T
5UM3 = 0. I N C L U D E ANY IMPACT FACTOR *a*+****+
cUHP = 0. WRITEl3rZOO8)I~lPTIJl~J,Z~NY~2~
\UMPT = 0. 2 0 0 8 FORMAT(//.2Xv15v12F10.2v/, 5X,12F10.2,/1
I~~~~.EO.NP.ANO.OZ.LE.O.)OZ = 0. WRITEI~~~OO~ISUMP~SU*PT
C 2 0 0 9 F O R M A T f 5 0 X , ' S U M TRUSS LOAOS = ' , F 1 0 . 2 . 3 X , 'SUM NODE LOAOS = * ,
C 1F10.2,/)
i! DO 1 0 2 J J LOOP FOR NUMBER OF SPANS U S E 0 FOR CURRENT LOAD P O S I T I O N S
= 1.TM 0 0 1 3 8 J = 1,NY
PRIJIII = PTIJI
3 1 = SPAN OZ - 1 3 8 CONTINUE
1 4 0 CONTINUE
KC = 2*lIM-JJ) + 4
7 = JJ 141 U R I T E I 3 ~ Z O l l ) I
2 0 1 1 FORMATI/I~SXI 'ONLY UNIFORM LOAO ON BRIDGE. I = ' , I 3 , / / )
IFlIPUNCH.LE.OIG0 TO 1 4 9
In1 = I- 1
 528 STRUCTURAL STEEL DESIGN

C A P P L Y I M P A C T FACTOR TO LOADS I F I M P > o

IF(IMP.LE.O)GO TO 2 5
DO 1 4 3 J = 1 v I M l
DO 1 4 2 K K = Z v N Y v 2
142 PR(KK.J) = PR(KK,J!*PAC b,
143
25
CONTINUE
KK = 2
4
DO 1 4 5 J = l r N S P N
KK = K K + 2
JJ = N P S ( J )
WRITE12vlOOb)JJ
WRITE(2v2014)(PR(KKvM)vM=l,W
2014 FORMAT(BF10.4)
145 CONTINUE
149 GO TO 5 0 0 0
148 WRITE(3r2016)
2016 FORMAT(II/v5X***** PROBLEM T E R M I N A T E D - - X ( L ) I S NEGATIVE ***I,
150 STOP
EN0

C ** FOLLOWIHG 4 CARDS T Y P I C A L SET O F F P S DATA FOR 9 PANEL TRUSS

J E BOYLES AREA B R I O G E FOR T E X T U S I N G COOPER E - 0 0 L O A D I N G AN0 I
...,,,.. 9 8 .....
3 . ,... o... 0' 1 0
3.45 27.60 17.00 1.0
4 8 12 16 20 24 28 32
C
J E
** F O L L O W I N G 4 CARDS T Y P I C A L S E T O F S 1 0 A T A . F O R 9 P A N
BOULES AREA B R I D G E FOR T E X T U S I N G COOPER'Ei80 LOAOIN
9 1 1 1 0
1.05 8 -40 5.20
4 8 12 i l l 20 2 4 lo& 32
/*

-...-*-..--.------ ---- -"....-'---- . '

I -- C,C.L~B:~$,.;<'~-'M,~DRAS

II <
. c$,.;-.!
I -
.:. P!=3.
I - I ,

i (-,>t,: 3x0. I

:, -- ... .
.-, -,.. ,.---.F. ---. . . .-r.-7-., -
.,...-.*7.".,-~,-
 AASHTO column formulas, 263
with bending, 32 1
' AASHTO minimum thickness of
metal, 491
AISC column f o h u l a , 262
with bending, 3 18
Angle tension members, 777
Approximate analysis of frames, 54
cantilever method, 55
portal method, 54
AREA column formulas, 264 .
with bending, 32 1
A d E A minimum thickness of
plates, column, 285
bending stresses in, 287
design criteria for, 287
BASLER, K., 474, 479, 48 1
Beam analysis, 57
biaxial bending, 164
Beam analysis:
differential equation, 58
for unsymmetrical bending. 167
Beam columns, 297
design formulas: XXSHTOand
AREA, 311
AISC, 313, 318-319
design methods, 322
effective length of, 303, 306
G factors for, 306
inelastic eRect/reduction, 307
K factor chart, 306
Beam framing connections, 385
Beam stresses, 146
allowable: Ar\SHTO and AREA,
15 1
AISC, 149, 174
laterally unsupported, 174
bending, 146
biaxial bending, 164
elastic design, 148
shear, 146, 159
Beams:
compact section criteria, 149
deflections, 146
laterallv unsu~norted.174
.' F tamella
, tearing of welds, 420
tdius of gyration., 175, 237, 255, S1 conversion factors. 39
Lateral bracing: .
for beams, 150, 155, 174 Sidesway, control of, 337
for columns, 291 embers, tables, 272 Slot weld, 414, 419
for girders, 496 ompression flange of beams, 179 AISC specifications for, 4
knee bracing, 497 Euler formula used in, 260 Spacing of bolts, 363
Live loads, 18 lacing for, 276 Stepped columns, 329
MUNSE, W . , 364
reduction for, 19 K factor for, 33 1
Load conditions, 70 alh, 473, 478, 494 Stiffeners, bearing, 158, 37
Load resistance factor design blt, 150, 465, 47 I , 490, 493 493
(LRFD), 132 column flange, 427
Net area: hlt, 456, 48 1 , 49 1
table of,@factors, 132 web. 4'76
column, 264 Llr, 263, 27 1 , 275, 307
Loads, 18 diagonal, for corner
in tension, 219, 225, 232 Reaction distance for beams, 157 connections, St6
bridge: AASHTO, 26, 27 effective hole diameter for;233
AREA, 28 bracing for, 158 glrder, requlred for, 456
at thre'dd root, 219 Reduced eccentricity of fasteners. ~ntermediate,479, 493,
equations for shear and use of s2/4g, 233
moment, 26 long~tudinal,AASHTO,
N E W L I N , D. E., 427
impact, 32 moment of Inertia of, 480,
NEWMARK, N. M., 185
tabulated shear and moment, spacing for, 493,495
AREA, 29 seat angks for, 436
dead, 18 St~ffnebsmethod of analysis, 54
earthquake, 33 Stress, allowable: base plate des
hi:
eccentric, on fasteners, 375 287
impact, 32 Plastic section modul bending, 149, 174, 469, 471 4
live, 18, 19 compression: AASHTO, 2
ponding, 25 AISC, 261
snow, 24 AREA, 26-1.
map for, 25 fasteners, tables, 353, 3
, wind,21,31 366-368
Low-temperature effects, 13 interaction, 3 12
Llr ratio: shear, 146, 159
stresses: bendin
compression members, 258, 260, , computation: for beams, 146, in girder webs, 478, 492
277 tension, 219
tension members, 231, 237, 238, table of values, 221
242 weld, 416
LRFD, 132 Stress range:
beam column desig*, 338 defined, 40
beam design, 208 table of values, 41
column design, 292 Stress-strain curves, 1 1 , 113
connections, 405 Structural codes, list of, 17
tension member design, 248 joint length as factor, 360 Structural shape data, table,
Shear stress, allowable: in beams, Structural shapes, 7-10
Lug angles, 3,65.. AISC specifications fo

Mass density, steel, 7

 536 INDEX
*
Tapered flange beam, 182
Temperature coefficient, 5
Temperature e r s u s strength, 13
Tension stress&, 219
allowable values, 219
effective net area for, 225
Threaded area in tension, 2 19
table of design data for, 230
THURLIMANN, B., 474
rIMOSHENK0 and GOODIER, 133
Torsio&pnstant, J, 171, 174
rruss analysis, 59

$.% Ult~rnatestrength. 7
$& ', @j "able of values, 6
($.,&"!;I Unhraced length, I75
LL'.1-51, 323
&, *' 3 L", 151
d Unlt welght of steel, 7

.
, J . , 368
. VALLAERT,
Warping constant C, , 174