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 DEIGN

CONTENTS

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

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

-9
t

se, w~thoutthe pnor wntten permissron of the publ~sher.

2.

&.@!

'67890

DODO 89876543210

Preface

Chapter 1 General Design Considerations

Thi, '..do'# was set in Times Roman by Science Typographers, Inc.

Tfie ..iilurs were Julienne V. Brown and Madelaine Eichberg;
thd cover was designed by Anne Canevari Green;
the ?rs.,duction supenisor was Dominick Petrellese.
'The ?..:wings were done by J & R Services, Inc.
K.'8. Donnelley & Sons Company was printer and binder.

'Uh.--*<
of Congress Cataloging in Publication Data
Bor ' s. 'oseph E
rdral steel des~gn

.>graphy:

p.

udes mdex.
Bulldmg, Iron and steel. 2. Steel,
Str~
3. Structures, Theory of. I. T ~ t l e
'TA6!' "478
624'.1821
79-18155
1SBW r 37-006765-1
!

1-1
1-2
1-3
1-4
1-5
1-6
1-7
1-8
1-9
1-10
1-1 1
1-12
1-13
1-14
1-15

Types of Structures
Design Procedures
Steel as a Structural Matenai
Steel Products
Steel Strength
Temperature Effects o n Steel
Structural Design Codes
Building Loads
Highway a n d Railroad Bridge Loads
Impact Loads
Earthquake Loads
Fatigue
Steel Structures
Accuracy of Computations and Electronic Calculators
Structural Engineering Computations in SI

Chapter 2 Elements of Frame, Truss, and Bridge

Design
2-1
2-2
2-3
2-4
2-5
2-6

Methods of A n a l y s ~ s
Beam Analysis
Determinate Structures
Truss Analysis
h a d Frame Analys~s
Bndge Analysls

2-7
2-8
2-9
2- I0
2- 1 1

The Computer Program Furnished in the Appendlx

The P Matrix
Load Conditions
Checking Computer Output
Design Examples

Chapter 3 Elastic, Plastic, and Buckling Behavior

of Structural Steel
Introduction
Elastic versus Plastic Design Theory
Safety Factors in Elastic and Plastic Design
Elastic versus Plastic Design Deflections
Length of Plastic Hinge
Elastic versus Plastic Design
Load Resistance Factor Design
Local Buckling of Plates
Post-Buckling Strength of Plates

Chapter 4 Design of Beams for Bending

4-8
4-9
4- 10
4- 1 1
4-12
4- 13
4-14

General Considerations
Design of Beams by the Elastic Method
Design of Continuous Beams
Web Buckling and Crippling
Shear Criteria
Strong versus W e a k - h i s Bending
Deflections
Biaxial Bending and Bending on Unsymmetrical
Sections
Shear Center of Open Sections
Design of Laterally Unsupported Beams
Beams with Nonparallel Flanges
Design of Bridge Stringers and Floor Beams
Composite Beams
Beam Design Using Load Resistance Factor Design
(LRFD)

Chapter 5 Design of Tension Members

5- 1
5-2
5-3
5-4
5-5
5-6
5-7
5-8
5-9
5-10

Types of Tenslon Members

Allowable Tension Stresses
General Deslgn Charactenstlcs
Stresses Due to h a 1 Load on the Net Sectlon
Des~gnof AISC Tenslon Rods
Net Sectlons
Deslgn of Tenslon Members
Design of Bndge Tenslon Members
Cable Deslgn
Deslgn of Tenslon Members Usmg LliFD

Axially Loaded Columns and Struts

~ntroduction
The Euler Column Formula
Columns a l t h End Condltlons
Allowable Stresses in Steel Columns
Deslgn of Bu~lt-upCompression Members
Column Base Plates
Lateral Brac~ngof Columns
Column and Strut Design Us~ngLRFD

Beam-Column Design
Introduction
General Considerations of Axial Load with Bending
Effective Lengths of Columns in Building Frames Developing the Beam-Column Design Formulas
Determination of the Interaction Reduction
Coefficient C,,,
AASHTO and AREA Beam-Column Design Formulas
Beam-Column Design Using Interaction Equations
Stepped Columns and Columns with Intermediate Axial
Load
Control of Sidesway
>
Beam-Column Design Using LRFD

Bolted and Riveted Connections

Introduction
Rivets and Riveted Connections
High-Strength Bolts
Factors Affecting Joint Design
Rivets and Bolts Subjected to Eccentric Loading
Beam Framing Connections
Fasteners Subjected to Tension
Connections Subjected to Combined Shear and Tension
Moment (Type 1) Connections
Load Resistance Factor Design (LRFD) for Connections

Welded Connections
General Conslderat~ons
Weldmg Electrodes
Types of Joints and Welds
Lamella Teanng
Onentation of Welds
Welded Connectlons
Eccentrically Loaded Welded Connections
Welded Column Base Plates
Welded End Plate Connections
Welded Comer Connectlons
Fillet Weld Design Using LRFD

Chapter 10 Plate Girders

General
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
10- 1
10-2

Appendix
A-1
A-2

A-3

Selected Computer Programs

Frame Analysis Program
Load Matrix Gcnerator for AASI-IT0 Truck Loading
on a Truss Bridge
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

PXEFA

'

at least in American practice. Specialized problems are not generally addressed

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 studentno 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
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

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 structural 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
in the later chapters. The use of simple examples gives the reader a quick grasp
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
"/
\
,'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:
i'l

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.
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 structure 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 specificaand 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 contributions 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 Department, Washington Univers~ty,St. Louis, who revlewed the load resistance factor
design material.

-1 TYPES OF STRUCTURES
e structural engineer wlll be concerned wlth the design of a v a n
d the follo\wng:
structures including, but not necessarily I ~ m ~ t eto,

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
_'.way bndge completed in 1874 represents one of the first uses of steel (and high-strength F, = 50
: ? 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.

cable-stayed, and cantilevered. Numerous lateral bracing schemes, incI

trussed, staggered trussed, and rigid central core, may be considered or
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.

GENERALDESIGN COP.SIDER%TI

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 considerattons 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/flnancia1 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
In structural englneerlng practice the designer will have available for p
k,
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

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
n as 29 000 ksl or 200 000 MPa-

The value for deslg

2. Shear modulus, G.

E
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.
as
3. coefficlenf of expnnsron, cu T h e coefficient of expansion may be
G =

//

a = 11

AL

25 x

a(T, -

per " C

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

y j o9m Q
-w
0 0

3$2&Q

m
0

k?

-8

4
3

g 3

;aa
.- 9 9
ggmz

5)

a
rd

"0

o
+

..

*9' 0

.O

$ 2

k;"

.
2

*,

.
2

m m

+ ? Y r :
w - 0 0

"
"
%
3
Q)

88

$898

I
m
m

o m - o

Q W W t -

S: , 8 2 3
N

N N m m

%Q%Z

.z
ba

<.ga

.V1
L

Z
m

m
.L

er properties o/ some interest. These properties include the mass de

(1 t = 1OOO kg); or in terms o
76.975 kN/m3. The specif
conversion of fps units
its of kN/m and kg/m is accomplished as follows.
Given: lb/ft and required to convert to:

8%3$

Note that lb mass and lb weight or force have been used interc
ably in the fps system because the acceleration-producing force
of gravity. This cannot be done in the SI system, since the ne
derived unit that defines the force necessary to accelerate a
'1 m/s2. The acceleration due to gravity is approximately 9
xample: Given: A rolled structural shape weighs 300
Required: mass/m and weight/m.
Solution: mass/m = kg/m = 1.488164(300) = 446.
wei&t/m = kN/m = 0.0145939(3

k 4-44

B
E

V1

$i nh

M
I

aJ
Z;

E"

g3
g 7;
-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 consisting 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 dimension (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.

---r
--- -

-.

hd

'V shapes

S \hapes

C shape,

L shape

:;gc

Anierrian Stdndard
bean1 (I-beatri)

Chdnnel

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
guaranteed values converge.

UReitdllgular

"_,;id
L-shdpe

7 slldpe

1 1 ~ q u dleg
l dtlgle

Strut tirrel Tec

i l l ! Iron1 \V.stidpc

8 Square
0Rounds
Ihrt

Plate

k + $ i s 1-2 ktmctural shapes as drrectly produced by Be steel producen.

t--:,.j

L Shapes

i i . e , ~shapes are either equal or unequal leg angles. All angles have parallel

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

1960 with application to weld~ng.All three of these steels have a yield

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

S"
WI.:1

~ r a tees
l are structural members obtatned by splitting W (for WT), S (for

.. 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

me1 design takes into consideration the yield strength of the mate
yleld strength of several grades of steel available for design is given in Table 1yield strength is that minimum value guaranteed by the steel producers an
All

Figure lJa Typical stress-stram

structural steel.

curves for

E;lgure 1-36 Erllargement of lmti

stram curve for two grades of
Note that the plasbc r a g e u

.- ...---

Since about 1964, specifications for several other high-strength (low-alloy)

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
High-"irength steel beams may deflect excessively, owing to reduced
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.

G E N E R U DESIGN CONSIDERATIO

"C x 100

14 Effect of elevated temperatures on either y e l d or ult~matetensile strength ucpresse

f strength at room temperature of approximately 70F

TEMPERATURE EFFECTS ON STEEL

1-6.1 High-Temperature Effects

Steel is not a flammable material; however, the strength is heavily temperaturedependent, as illustrated in Fig. 1-4. Both the yield and tensile strength at
1000F is about 60 to 70 percent of that at room (about 70F) temperature. The
drop in strength 1s rather marked at higher temperatures, as shown on the figure,
where the strength at 1600F is only about 15 percent of that at room temperature.
Steel frames enclos~ngrnater~alsthat are flammable will require fire protection to control the ternperature of the metal for a sufficient time for the
occupants to seek safely or for the fire to either consume the flammables or be
extir~guishedbefore the building collapses. In many cases the building does not

t See footnote a of Table 1-1 In J. E. Bowles, Structura/ Steel Desrgn Data Manual, McGraw-Hd,
New York, 1980

1-6.2 IAV-TemperatureEffects
Brittle fracture is a failure often associated with low temperatures. Essentially,
brittle fracture is failure that takes place without material yielding. The
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
percent elongation is specified for steel in the ASTM standard tensile test.
Inlplicit in steel design is the resultant deformation (yieldixig) of the material

Fireproofing materials
Cinder concrete
Gypsum board
plaster, cclnetit and sand
Expanded shale concrete
Vertnicul~te
Perlite

Unit w e ~ g h t
Pcf
l 10
30-40
100

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.

k~/rn~
17.3
4.7-6.3
15.7

,.,

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

GENERXL. DESIGN CONSIDERATI

any abrupt change in cross-sectional area-to

tension situation.

inhibit lateral contraction

brittle fracture. This may initiate as a crack that propagates to a member f

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.

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

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

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,
ntity, frost depth, and average wind velocities.
list gives several design codes and/or specifications
may have occasion to use:

Lo.~,d.lbuilding departments almost always require structural de

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.
and Steel Institute (AISI), 1000 Sixteenth Street,
. 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

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-

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 pretnrctzlrnl 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

carry the dead 1

GENLRAL DESIGN CONS1

fps: R
SI:

0.0008 x area

R = 0.0086 x area

(when area
(area

>

150 ft2)

> 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.
From this list it is evident that any part of the building w
nstalled contributes to the total dead load. Dead loads can

(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.

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
:lot known.

Ceiling, metal deck, steel bar joists

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

= 12 psf

GENERAL DESIGN COKSID

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
El-1, 18 X 22 (area ABCD):
R = 0.0008(18 X 22) = 0.32

< 0.60

R = ----- = 6o 8o - 0.40 < 0.60

4.33 L
4.33(80)
Use the smaller value of R computed, 0.32. The reduced live loa
L' = (1 - 0.32)80 = 54.4
say 55 psf
Compute R for the column; the contributory area is centered on the column
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.
+

R = ------60 80 - 0.40 < 0.60

4.33(80)
Using the smaller value of R, 0.40, the reduced live load on
L' = (1 - 0.4) x 80 = 48 psf
We note that O.$0 is the maximum R for the column and is
uppermost floor level.
+

+ 8)/2 X 91 = 0.774 > 0.40 (and also 0.

D +L
= ----- = 3.703 + 5.00 = 0.402 < 0.60

0.0086[(12
4.33 L

4.33(5.00)

Since the problem statement limlts live-load reduction to not mor

0.40, the reduced l~veload is
L ' = ( 1 - 0.4) x 500 = 300kPa

d loads have been extensively studied in recent years, particularly for larger
-rise structures. Generally, for tall structures ~wnd-tunnelstudies should bz
or smaller regular-shapsd
30 m, the wind pressure
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.
Weight of concrete: 0.125 x 23.5

o allowance is commonly made for the shielding effect of adjacent structure

r from ground cover.
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.

Snow and other live loads

lilllllllllllllllII1111111111'

Figure 1-8 Snow and other roof live and dead loads.

In addition to the types of pressure or area loads noted, building codes may
stipulate checking for a concentrated load of some magnitude which may be
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
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 adequacy 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
I6 kips 7 3 kS

Bridge span

---

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)

. !

,. 3

.-.' 1

- ..

23 712.0~
35 118.0b
48 800.0~

17 990.0
27 154.0
38 246.0

479.6

282.0

128.1

522.0
626.4
729.3

306.8
367.3
426.4

197.9

1225.3

.U1 values shown are for one rail (one-half track load). Axle loads shown in diagram. Obtain
for other E loads by proportion. ,,
.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.

. . :s

,;

I.

f ::-cd

on the locomotive weight, the Cooper load is designated as E-40, E-50,

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
s.+c.;;rs at selected locations for girder bridges, with values given for a single rail
fbading (based on one-half the axle load shown in Fig. 1-1I).
Where multitracks or road lanes are carried by the bridge, the live load is as
ici!ows:

total uniform loading, however, must not be less than the following:

2. AREA wind requirements:

Pressure, force/area
Unloaded span
fps, psf

SI, kPa

Loaded span
fps. psf

SI, kPa

Percent of live load

Lane or track

AASHTO

AREA

100
90
75
75

100
2X100+1X50
2~100+1X50+1X25
As specified by designer

4
XI.xe than 4

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

live load (without impact).

Other loadings that may require consideration include differential temperatures between top and bottom flanges or chords, ice and snow loads, possibIe
overloads, and for continuous bridges, support (pier) settlements.

.-

<..*,,.,

6. ..4

- -..'i

.- - . llr,

....,-.., -

G E N E W . DESIGN CONSIDERATIONS

~ L O I U I Y

nt; must be divided by 100 to use in de

Railroad bridges make a distinction between those bridges which consi

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
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!?

1-10 IMPACT LOADS

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

impact load as

Item
Elevator loads
Macbnery and other moving loads

1.00

ilroad bridge consists in two trusses spa

sses are made up of seven panels at 27.60 ft/
at is the impact factor?

> 0.25

r,unoN Since L > 80 ft,

L = 7(27.60) = 193.2
= 25.6 percent
600
193.2 - 30
The AASHTO impact requirement is
SI

f PS

I,

=.

2
< 0.30
I,+ 125 -

I,

-l5

i0.30

L+38-

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:

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
trusses at panel points.
SOLUTION
S = 27.6 ft, L = 17 f t

< 80.

3(17)*
+ 40 - = 43.1 percent
1600

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 earthcitation 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

,,.,.'.b..*

"

..

-.

.,.

G E N E W DESIGN CONSID
9

Any floor:

Partitions in 40 X 30 apartment at 8-ft height and

cross-walls at 20 psf
gives: (40 x 2

+ 30 x 2)(8)(20)/(40 x 30)

,I

.4

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

number of stories

0.1(10) = 1.0 s

The author will average the two values of T to obtain T = 0.895 s.

1
1
C = ----=
= 0.0705
1 5 r ~ 15Substitution of this accumulation of factors/weight into Eq. (1-4)
F = 0.75(1.0)(0.80)(0.0705)(1.5)(3446) = 218.6 kips
The roof value is
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
follows: Z W,hn = 337(90) + 337(80) + 337(60) + . .
151 650 f t . kips

(F- F

lop

)-=Wn A,
2 W,hn

(218.6 - 13.7)

For tenth floor

F,, = 0.001351(337

90) = 40.98 kips

F, = 0.001351(337 X 80) = 36.42 kips

'

F8 = 0.00135 l(337 X 70) = 3 1.87 kips

F,

0.001351(337

10) = 4.55 kips

(shown in Fig. El-7b)

Summing the horizontal floor loads and including the top value of 13.7
218.59 versus 218.6 kips as a check. These lateral floor loads
further prorated to the several bays in the E-W direction for the fr
analysis load condition(s), which includes earthquake forces.

12 FATIGUE
failures which have been attributed t
ue. Fatigue failure is a material fracture caused by a sufficiently Iarg
ulsating stresses, or stress reversals. Th
ture of the material at a location where
oscopic in size) exists. A crack form
nding on stress level, rapidly or slowly (sometimes so slowly that the
re) progresses to failure of the part
Most metals tested under repeated or cyclic loadings display stress r
s qualitatively illustrated in Fig. Irly, these curves were commonly displayed as stress level versus cycIes.
t, the stress range is used as the parameter of interest. The stress range

GENERAL DESIGN CONSIDERATIONS

4(8 STRUCTURAL STEEL DESIGN

to

Base metal with rolled surfaces

41

100 to 500 500 to 2OCO Over 2

!@I

T or R e p 60 (415)

36 (250)

24 (165)

24 (165)

T o r Rev

32 (221)

19 (131)

13 (90)

lob(

T or Rev

2 1 (145)

12.5 (86)

8 (55)

T o r Rev

45 (310)

27.5 (190)

18 (124)

16

TorRev

27(186)

16(110)

10(69)

7(48

T or Rev

45 (310)

27.5 (1%)

18 (124)

16 (110

T or Rev

45 (3 10)

27.5 (190)

18 (124)

16 (110

T or Rev

21 (145)

12.5 (86)

8 (55)

plates and shapes with full or

of girder webs or flanges adjacent to welded transverse stiff-

Figure 1-13 Qualitative plot of stress range F,, versus number of cycles to failure.
. .-",., ... ',"....-,. -

Base metal at end pr partial

3.

anically fastened connections

can be defined as

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
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
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,
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
usually not considered for routine building design, since 10 load cycles/day over

_ . . _.

. . - " .,.....

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

se 12 ksi or 83 MPa for girder webs.

___-________

._
_.
---

-.--.--------_I-

.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
stresses.

Tal)crcd CUILIIIIII .lnJ r ~ k1r

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
.: zpt in multistory apartments.
.

'L-13.1 Industrial Buildings

T ~ L I ~ S -soli~rnn
on

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
leng:;. uf the diagonal members.

1-13.2 Bridges

reduced to reduce member Iengors.

OF COMPUTATIONS AND ELECTROMC

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
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
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

produce the end results 62.4 and 9.807 are not shown.
Several other useful conversion factors are as follows:

To convert

to

kilogram (kg)
:mund (lb)
pound
';ips (1000 Ib)
ib/ft
: ips/ft
!b/ft
psi
psi
ksi

kilonewton (kN)
kg
kN
kN
kN/m
kN/m
kg/m
MPa
kPa
MPa
kPa
kPa
kPa
kN . m
kN m

.ASI.

psf
isf
Lip . ft (moment)
kip . in

What are the wind forces at points 1 through 9 of the roof shown in Fig. Pi-7?
Answer: ~ ( 5 =
) - 18.07 kips.

Multiply by

0.009806650
0.4535924
,
0.004448222
4.448222
0.014593727
14.593727
1.488 16404
0.006894757
6.8947577
6.8947577
6894.7577
0.04788026
47.88026
1.35584
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.

1-3 What is the increase in weight/ft of a W16 x 40 beam with 2 i

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?

Answer: M = 7865 kN . m; I

40.1 percent.

-12 What are the story shears in the N-S direction of the building of Example 1-7?

ELEMENTS OF FRAME, TRUSS,

AND BRIDGE DESIGN

'

. of

ben
Ib )

Figure 11.1 Two of world's tallest buildings using structural steel frameworks. (
(442-4 Sears Tower building in Chicago, Illinois-c-ently
the
tallest buil

337-111John Hancock Center building, also in Chicago, Illinois. View from the sears

ANA
I

.*

requires dev eloping frame member forces (axial,

based on the dead and cntical live-load combina-

The method of analysis depends oh. the complexlty of the structure and
whether it is rigid (indeterminate) or simply framed. An analysis may consider
the structure as either two- or three-dimensional. Simple framed structures are
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:
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
rapidly, one may initialize a problem using relative va
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,
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.
primarily to obtain the effects of lateral (wind) forces on a frame.
The portal method (refer to Fig. 2-1) makes the following assumptions:

'

,q = 4 + 6 + 3 = I3

'

'

1. Point of contraflexure occurs at midheight of all columns.

2. Sum of wind (lateral) load is distributed as shears in proportion to ba
3. Beam or girder moment is zero at midspan.
4. Interior columns
carry..
na axial loads.
~--.
... .

--

With these four assumptions and application of statics,

joint moments can be obtained at any location.

~ I ,

21 portal

of approximate frame andysis [br a typical three-ba~bent in

The cantilever method (Fig. 2-2) is an alternative method of appro

beams and columns.

. Axial loa_d_s_isin~-al columns and are pro~ortionalto the

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
known, a value of A = 1 unit area is usually used). The equivalentmo
inertia of the vertical cantilever beam is computed as
I =C

The column loads are computed as

'

Mc
v.=I

:'

STRUCTURAL STEEL DESIGN

ELEMENTS OF FRAVE, TRUSS, AND BRJDGE DESIG

The concentrated loads produce a moment of

M

= -3-P- -L= -

2 2

PL
4

PL
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
." 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.
'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.:

one in which much of the roof load is carried by the perime

+,C ,

- wx2
EIy" = M = moment = -+ c 1 x
2

+ c2

- wx3
c1x2
EZy' = slope = -+ 6
2

+ C2x + C,

TERMINATE STRUCTURES

'.

successively:
EIy"' = V = shear = - wx

Load-brar~rigwail -,

- 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
approximate deflections should be computed before a de
The reader should note that highly refined deflectio
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
fix?? by the moment of inertia rather than by the bending stress.

ical sense almost all beam-column connections can, and do, transmi

2j

--

n = number.of bar members

e-rea&Iy- _
optimized for least
-- - weigh~,-since--hes_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

s-ndeterminate
trusses are commonly used wh
some members, unsymmetrical loading conditions
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
(or any frame) js that which"ba1ances weight, fabrication, and
against the total client cost. Minimum weight alone does not sati
of "optimization."
-ate
trusses are readily solved by hand using the m
equilibrium or the alternative of
a cut section. Both determinate and indeterminat
using one of the large number of computer progr
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,
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

V Z ) The
~ . nodal equilibrium in matrix notation is

P=AF.,

member forces can be written as

F = S e = SATX

combining Eqs. (c) and (a), we obtain

',

'

n ( d ) can be inverted to obtain

x = ( A S A ~ ) -P'

mpute member forces:

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

Truss member

PI + F I cos cx = 0
ZFv = O

gives the element stiffness matrix as follows:

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 incorrect) 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
..

,," ..,.,.....,.

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
xamples displaying computer output..The reader should parti
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
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" :
FEM =

12(0.707 11)
6
(-)'
0.707 11
12

For the horizontal member:

RL=--4(72)
6

- 48.0

,= 50.91

kN - m

70 STRUCTURAL

STEEL DESIGN

Now consider the following sketch for nodes 1 , 2, and 3:

P, = 0
p2 = 64 - 50.91 = 13.09 (note +)

P4 =

- (50.91 + 48.0) = - 98.91 kN

e several load conditions for a bridge would include

Dead
Dead

+ live, including impact

+ live + wind

.",,.,,

4,

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

7 mPP1)

-(does

not matter)

several load conditions.

Computer programs are included in the Appendix to deveIop
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

(ASA~)-'{P}

night include:

NLC

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

..

1
2
3

+
+

the end of the coding sequence is checked and both satisfy statics,
problem has been correctly solved for that irlpltt data.
Problem checks can be performed in several ways, as illustrated in

+ wind

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
Advantage can often be taken of symmetry of structure and lo
two office building examples that follow (without wind). Sometimes t
matrix can be used to advantage, as illustrated in the column check of th
building.
Truss checking is similar to that for rigid frames. Always take
truss geometry by checking members in which the internal force is
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
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
either and the problem needs to be reprogrammed.

2-11 DESIGN EXAMPLES

The Sallawmg several design examples will further illustrate frame coding and
computer input/output for the analysis computer program in
computer output from these examples will be used in many
examples in later chapters toillustrate the problems in structural (
a manner somewhat like that which would take place in actual design.
interest of saving text space and maintaining reader interest, the examp
considerably edited from actual design problems.

7'-6"

1
Z

2 1'1"
7
3

(2)

(5)

(3)
16)

(9)

110)

(12)

(13)

10 ( 1 6 )

I I

(19)

(20)

14123)

I j

(75)

(26)

117)

124)

,.s.v
4
(;

17 45.47.47

---.\'P = 4 b
.YPPI = 47

18 47

r.2.i
Elevation o f typical interior bent

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 Ebay 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
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 minimum requirements-the designer can always use larger values.

- -

Plan view

~-

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
following computer program control data information:

NP = 46
NBAND = 11
The next step is to develop the beam and column loadings for this
typical interior bent (refer to Fig. E2-3b, which displays loads after comcputations).
Roof:
Dead load: estimate 5 in of concrete on a metal deck supported by steel
bar joists.
NM

Dead load

0.080(18.5)

Live load office

26(members)

Live load corridor

0.080(0.70)(18.5)= 1.036 kips/ft

=

O.lO(l8.5)

= 1.85 kips/ft

These loads are shown in Fig. E2-36.

.02(6)18.5 = ? . 2 ? k

Concrete: (5/12)144
Metal deck and joists (estimated)

Ceiling, ductwork, electrical, etc.

Total

5 psf

= 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

"

(100) = 43.3 percent

R = - D+L(lOO)=4.33 L
4.33(80)

or

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

The office space will have a 30 percent reduction in 1

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

12(18.5)(0.045) = 9.99 kips

STRUCTURAL STEEL DESIGN

The longitudinal walls defining the corridor and lo

Ute interior column loads:
L

Tile wall at 8 in = 35 psf,

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)

L'

~ & l e c any
t beam weight e tributing shear to the roof line of exteri
columns, since the values are tw small to be either reliable or to affect t
design.
Now make some preliminary member size estimates with the followin
practical considerations:

r
rO
T rr 0C0 0 0
0.0. .. -0 e
0 0CC 0
*r
a ~0O 0
oo
- O
~ .O
O*
o C c ~ - ~ O

M,

I
u

"I

;f
the moment is M =
lue is somewhere
- - - in
--A

.a

0.925

-9-0000---0000-*-000O_t_O
rr*
* D *
0-0

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
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
00
00
00
00
00
00
00
00
C0
00
00
00
00
00
00
00
0

--ldddd222dddd222;ddd22fdd
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
* ~ * * . * * n D r * * *
i

= 142.7'ft , kips

N----N

* r O r * * ~ * ~ w ~ ~ ~
N----N
*----.+
L--

6
-.

1 I.OU

in-

1, = 3 I 6.u in'

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

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

For columns:
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,,

........................

12!0(12)
= 60 in3
24

n =

O O o C O O O O O O O O O o O O O O O O O O o o

dl,d;dd;ddddld;,hdb:dD'O'd::
----*.-----*<-----*.-wm~=*

*rw----nrn----ccn----.oQ*--

000000000000000000000000
000000000000000000000000
00000000000000000000O0O0

dddddddddddddddddddddddd
000000000000000000000000
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

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

.=.

O O O o 0 O O O O O o O O C C O O o o o O o o o

-u.

-- ,--a

................._...._._
- - - *---=*
- --

The required section modulus based on an allowable

24 ksi (Chap. 7) is

ox

-,8

0 0 0 - - - - C 3 0 - - - - C 0 0 - - - - ~ 0 0

Sx =

~ * 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

8
(142.7)- == 95.1 ft . kips
12
Ma, = 118.9, say 120 ft . kips
M,

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 0
0 0
0 0 0 C
0 0 C
0 0 0 c
0 0 0 0 0 0
~ 0 0
C C
0 0 0 0 ~
0

+ 1.48)(21'3)2

.
L

1. Use continuous column (no splices) for full 36 ft of hei

interior columns (maybe) at fi st floor level.]
2. Use constant-depth beams acr ss the bent.
For beams:
If the roof beam is simply su
fully fixed, the moment is M =
between, so taking an average:
-

.
c
.
.
0-0

.-.

;l;dddd:;:dddd-.;:dddd2;*

dddddddddddddddddddddddd
000000000000000000000000
000000000000000000000000
000000000000000000000000

dddddddddddddddddddddddd

00000000000000000000000000
00000000000000000000000000

'

ddd;;;;ddd;;;;ddd;;;:ddddd

'"I.

' Z

4
W
.

'4-

\ l I T F I I I tN"IE
U
h/D

YO
YP
ND
N>
ND
NO
NP
YO
yo
NO
Yo
LIP
qo
YP
YD
q P
NP
LIP
ND

1
7

10

=
=
=
=
=
=
=

YP =
"
I =
VP =
NO =
No =
NO

VP =
YP =

YO =
LIP =
ND =
No =
YO

.,D
YO
NP
NP

=
=
=
=

NO

yo =
NO =
NP =
NP =
YD T
NO =

=
No =
NO

OANJ h I O T H =

13

AND U NX =
NX =
NX =
NX =
WX =
NX =
VX =
NX
NX =
NX =
NX =
NX =
MX =
NX
NX=
YX =
NX =
NX =
NX =
NX =
NX =
YX =
NX =
NX =
NX =
NX =
NX =
NX =
NX =
NX =
NX =
NX =
NX =
NX =
NX =
FIX =
NX =
NX =
NX =
NX =
NX =
NX =
YX =
YX =
NX =
NX =

=
=
=

690

11
I 2

114
15
16
17
18
19
20
21
22
23
24
25
26
27
2R
20
30
31
32
33
34

35
36
37
38
30
40
41
42
4'
44
45
46

1
2
4
5

6
7

8
0

10
11
12
13
14
19
16
17
18
l a
20
71
22
23
24
25
76
77
ZP
29
10
31
32
31
34
3<
36
77

38
39
40
41
42
43
44
4T
46

0.00394
O.OO4Rl
-0.09541
-0.00169
0.00033
-0.19343
0.00169
-0.00033
-0.19343
-0.00394
-0.00481
-0.09941
0.00272
-0.00081
-0.08302
-0.00121
-0.00006
-0.17721
0.00121
0.00006
-0.17721
-0.00272
0.00082
-0.OR302
0,00275
0.00061
-0.05121
-0.00125
0.00007
-0.14018
0.00125
-0.00007
-0.14018
-0.00275
-0.00061
-0.05121
0.00353
-0.00136
0.00023
-0.08232
0.00136
-0.00023
-0.Oq232
-0.00353
0.00068
-0.00068

Sketch lo ~denttfyP - X codlng

XAL F3RCE. K

3
4
5

6
(I

10
11
12
13
14
15
16
17

-6.07
-22.77
-38.21
-38.21

-82.42
34.19
-20.14
20.1+

38.11
3R.70
-22.49
22-49

1.01
0.*6
1.01
-58.49
-81.26
-81.21
-58.64
-0.72
-0.54
-0.12

-64.52
-49.19
-87.27
30.45
-18.14
19.14
-30.45
-63.13
-50.34
-81.36

87-21
49.19
64.52
30.1.3
-11.44
11.94
-30.33
81-36
50.34
63.13

DESIGN EN0 *O*tNTS CORRECTED

F n R F E I A N 0 WIN0 lNE.49 EN0 FIRSTI.
-26.18-

64-31

I-F1

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

For the two basement columns, use W8

Check CF, for t o p story.

110 in"

I,

48:

C/',l,,,8d,, = ? . 3 3 1 ( 2 1 . 3 ) ( 2 ) + 2 . 5 1 6 ( 7 . 5 ) + 0.925(2) = 120.02 kips

A = 14.10 in2

From computer o u t p u t for values shown which have been reduced 25 percent I'or wind
5o.that o u t p u t is directly comparable t o D + L o u t p u t :

T F , = 120.01 kips O.K.

. .-.

7.

..l

1,

... ...,...*.,_

101.55
Actual F = ------ = ]35,4k
0.75

..

R3= 12.09~

'Mom have been

reduced for wlnd

V1= 8 965

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

Figure E23f Output checking for NLC = 2 (with wind).

Use W8

40:

A = 11.70 in2

Ix = 146 in4

I, = 184.0 inJ

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

loading cards (interspersed as appropriate in member data), and a P-m
data set. Note the P-matrix entries read (PR(i, j)) are zero for NLC = 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.
e reader should note that when wind is considered, the allowable stre
increased by one-third. So that the design is all on the same b
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 2units. Note that the dimensions are slightly different than when
oft conversion. The same general design considerations/parameters
pply as in that example except that the brick veneer wall will be take
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.
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 )
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.

Total

= 0.263

kPa
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:
=

21.1 k N / m

Live load in 6.5-m span: 2.8(5.9)

16.5 k N / m

Live load in 2.3-m span: 4(5.9)

23.6 kN/m

Dead load
Elevat~ono f t y p ~ c a lInterlor bent

59

,!
1

59

59
1

L
1

5.9

,
1

59

3.581(5.9)

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

11590'0
',5100'0ZLZOO'O
010b9'1OhlCh',
. .
01200'0fbO1L.kEUOUb'f
EVIOU-0
0kEII.EbV8bb.b
~0100'0U0015'19b9L0.4
bLkOO.0
ObULP'ZO99Z5'8
8?Z00'0siSOE'ZEZEIS.8
E~IOO'O
bLLL0.bIEIZS'Y
blIOO'0SZBfb'ZVEZSS'8
Zli00'0

l'd

01100'0
LLEOL'I91LOO'O
OLlOO'OOCZOO'O
IlbO9'1ZfU20.0ULZOO'Ob
b
9OZOO.OSETOO'O
ElZIZ'EZIb00'0
5E100'0bObO9'1LZOEO'O
bLZOO.0
ULEUS'ZILPZO'O
18200'01Ulbl'tbuZOO'O
LE100'0
bZ1bl-295000'0EiIOO'OOEEYS'ZIECZO-00UZ00'0

12
09
bk
dC
~r

VZ

= XN
= XN
XN
= Xh
- Xh
XN
Xh
xh
XN
Xh
= Xh
XN
Xh

52

?C

VL
><
ZE
rL

ZL
le
Or
bZ

VZ
LZ

iZ
22
1Z
UZ
01
11
LI
VI
51
?I
LI

Ol.bCb511
sb'51Z00'0
OI'bZbS110u'Z~ObLl
UL'LIZ-

KN

= XN
= XN

=
=
=
=

OI'bZb$ll
56'5120 0 '0
Ol'bZb'YI1OU'ZLOI,CI
Ok'iIZ-

Of
br
Hk

LC

9 LC

6ZZEl'O
ZZfOO'O

Z
1

XN
XN
XN
XN

XN
= XN

OU'ZlObLI5b.GIZ00'0
01-bzb511
11'91Z00'0
01'501911-

1 1 - 1 00-0
01-SO19I1

11'01Z00'0
01'G01011

so3

NIS

?I 01

(11

1:

lfr
;I

6 B L

P \ P

PSI;.

19L.b

XI

S6SXOItM

1
f ; I

131113W
Z

= XN
= XN

-..
.

..

OZ'OI
OE'EBEZEI

00-0
0'r@EZ~1
MY-W

~ H L

--

- o h

=
=
=

dN
oh
dh

AllJIlSVl3 OOY
Od3Z-YON ON
92
= S138M3Y

dh

OOb
3J

ON

U
L
9
5

b
2
I

= cN
= dh
oh
ON
dN
ON

=
=
=
=
=
=

dN

= dk

S3181N3

tIl3jllS

030151 3 3 l ~ l iL N
31N U N i l Oi dlh G V
9C
- dN
dN

dh

= dk

'XIdllW-d

P ~ H
L ~ +N d ~E ~ NZ ~ NI ~ .N3 8 h z r

21
I1
01

NW

ONV

000'58
OOZ'LZ.
OOZ'LZ
000-0
000'0

O.OOOOOL
11 31I1JM
E
= llONJ3 O V J l 30 ON

62
HC

= XN

OOZ.LZ
OOZ'LZ
000'0
000-0
OOZ'LZ
002-LZ
000'58
000'58
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*..

CC
Zi
II C
0k

=
-

= XN
= XN
= XN
= XN

= H l O l H ONVB

1~33

dN

ah
dk
ON
dN

5k

21

= XN

S N V i O V M MO Yvl ' X I d l V W - X

2 ~ 3 3

=
=

00'0

Ob'ZlOhCI56*$1Z00'0
O I ' ~ L ~ ~ I
1I'VIZ00'0
01~b01011-

XI%

. -. .

OCZSZ'OI
OESOO'O

= dk
= dk

01

00'0

OZ'ZZIOZ'ZZI00-0
00'0
O L ~ E B E 2 E I -O C - E B E Z E I 00'uLI00.8~100'0
00'0
OO'uL9211
00'8L9ZII
00'8L100'8LI00-0
00'0
OO'LfL9ZII- 00'8L9ZII~Aluo)

= dN
= dh
= dN

3H1

ELEMENTS OF FRXLIE, TRUSS, AND BRIDGE DESIGN

89

e reactions at the basement wall are computed:

+ axial load,, + R i + axial load,, + axial load?, + axial load,

387.29 + 402.51 + 626.26 + 665.79
= 111.95 + 119.28 +
u,..-/>

= RI

-2081.85

0.75
Now consider C F,,

53.64

- 3007.03 (vs. 3006.36 kN)

O.K.

0:
17.60

14.92

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

dd:dd'::;;;'dd;';:;;;;f;;r:

0 0 f f " m N * O * C N " ~

"IOtCN"

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
~IDP~N-C*QONV
I

I , , ,

m n n o m *

. .. . .

*l,-"...a~

n m r p t

""ti 'i"

'(Mem.

22)

'(24)

Forces on two basement reactions, including effects from horizontal beam

members 22 and 24:
CI;;t(applied
externally)

= 10.2

, ,,,=

CFh

15.3(2) = 40.8 k N

-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

Total

= 5 psf

=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

wt = 11.6 Ib/ft

3 x

i:

.A = 3.38 in2

The approximate truss weight is computed as follows:

Length of top

+ bottom chord (approx.) main truss = 2 x

Verticals: av. height x 1 1

(15

+ 27)(11)/2

Diagonals: av. length x 10 = (27.3

+ 19.2)(10)/2

120

=2

92

STRUCTURAL STEEL DESIGN

The average truss weight:

W(120) = 240(0.0206)

w=-

120

+ 0.0116(231 + 232) = 10.31 kips

= 0.09 kip/ft

use 0.10 kip/ft

Truss weight for side sheds is as follows:

+ bottom chord (approx.) = 2 X 72

Verticals: (3 + 15)5/2
Diagonals: (12.4 + 19.2)(5)/2

Length of top

= 144 ft

= 45 ft
= 80ft

The average truss weight:

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

W(72)
I

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)
main truss + 1.5 snow + dead on right shed

+ dead on

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.
Shed dead loads: note that both shed and main roof truss have 12-ft
panels.

Interior nodes: (0.07 0.50)(12)

Exterior nodes: 6.84/2

= 6.84 kips
=

3.42 kips

% STRUC'IVRAL STCCI. DESIGPJ

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

Interior nodes: 0.025(25)(12)

Exterior nodes: 7.50/2

4812

= 7.50 kips
=

3.75 kips

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

0 = tdn

-=

Main roof truss

7.50'
8 = tan-'
'1

Py

18
72

- = 14 04'

3 87'

12

- = 12 37'

Los e
= 0 025(25)(12 3 7 ) = 7 73'

r':, =

7.73 cos B = 7.50'

Ph = 7.73 sin 0 = 1 .88'

Figure E t S c

Shed trusc (left side)

ELEMENTS OF FRA!!IE,

% STRUCTURAL STEEL DESIGN

~ 1 1 0 1 9c ~
rhotTlo*

? N O I T I O N NO

YQ

STIFFIII

THF P - Y A T R I X ,

ENTRIES

AND

IN-K

954

IL
FORCE. K
&XI

D E S I G N EN0 VOlENlS
CO99EClE0 GO. F E N bNC Y l Y O
INEL9 EN0 F I R S T I . FT-X

----

8.15

----

--..
-..-

YO r
qp =
'$0 =
YP =
YO =
YP
YO =
YD =
YP =
NO =
N O =
NO z
NP =

10
11
12
13

14
15

16
17

1R

14
20
21
22
23

24

25

=
=

26

=
YD
NP
NO
ND
No
NP
NO
ND
N'P
YP
NO
NP
NP
YO
NP
NP
NO
ND
Yo
YD
YP
NP
l o
rlD

;
.

27
28

24

t . PO'-

=
=

31
32

=
=
=

79
34
35
96
37

=
=
5

38
30

;
.

40
41
47
41

44

45

;:

$2

.NO

yb
NP
ND
NP
ND
NP
yo
YO
NP
NO

=
r

49
50
51

=
=
=
=

52
53

=
=
=

56

54
55
57
5R

Figure E2-5e

TRUSS, A N D BRIDGE DESIGN

-----..
----------------------------

----

---.

-----.----

----------------------

0.00

*---

.
.
.

----

-------

----

--.-

-------------

...-

----

----

rjrsFs

NO

rrlu
FO~CE.I

OESIG* EN0 .C*E*TS

COPPECIEO FOP
U W
# * E L M ENO F I S S ~ I . FT

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

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

30.77 x lo-' m'

I,

2151.9

76 X

low6m

Two interior columns (members 24, 25, 27, 28, 67, 68. 69, 70): W690

The computation of frame loads now follows.

Roof?
Miscellaneous, including purllns
Total

= 0.72

kPa

= 0.24

kPa
= 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:
=72 m
Length of top and bottom chord = 2(36)
=70.7 m
Approx. length of diagonals = lO(5.84 + 8.3)/2
= 70.4 m
Approx. length of verticals = 1 l(4.6 + 8.2)/2

Take as horizontal pro~ect~on,

as truss slope

IS

very small and values are estimates.

'1@ .. .KUCTURAL

STEEL DESIGN

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

~u-. J I K U L ~UKAL

SIEEL DESIGN

KN &NO

YP

ND ;
N
=
VD =
VP =
LiO

3
4
5
6
7

NP

qu

NO

NO

vu =
rrr =
=

<D

8
9
10
11
12

17

UO

'40

\ Y

16

NP

17

19

Nu
NP

Y O :
'40

,do

NO
L(P

..
2

=
YP =

Ye =

v;
NU

*iP
40

14
15

1R
70
71

77
73
24

7r,
26
27

%.

2R
2s

30

31

U I

=
=
-

YO

34
35

NP

YP

=
UP =
ND =
UD

NO

S'
N %,

.1R

=
=
=

'40

VP

=
=
No =
LID =
YO

NO

37

3,
36
37
38
39
40
41
42
43
44
45

16
47

Vo

"R

yo

4Q

'40

50

No

U'

NP

V?

hi"

=
=

:Z
53
54
55

YO

56

YO

YP

=
NP =
NO
YO

','o i
LIP =

No =
YP =
Yo =
ND =
NV ;
NO =
hi0 ;
NO

.,D

=
=

VP

VO

YO =
=

Urn
ND

ND

=
LID =
YO

ii,.,..

57
58
59
60
61

62
61
64
55
66
67
be
69

70
71

'2
73
74
75
Ih
77
78
79
80

K-MR

!X,

MU 01 R A D I A N S

CONOI'IOH

YO

AXIAL
FORCE, K N

106 STRUCTURAL STEEL DESIGN

present maximum horizontal deflections are:

X, = 101.7 mm
X,, = 105.1 mm
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

gles with long legs back to back for both top and bottom chords, find the an

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
solution in either fps or SI units, as assigned by the instructor.
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.

w=

4 kipsift or 60 k N / m

4 p a n t i s '0 20' = 80'

( 4 paneis '.i 7 rn = 78 rn)

sok
2.4

200 k N

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
c
d
e

30 rt 3.0
29 t 0.5
13.75 r 0.5
7.5 r 0.5
20.0 t 2.0
3.25
6.0 -C 1.5

f
g
h

9.1 r 1.0
8.8 r 0.20
4.20 r 0.15
2.3 r 0.2
6.1 r 0.6
1
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:
25 r 3 ft

3.

Members
Bottom chord (1,4, . . ,21,25)
Top chord (6, 10, 15, 19, 23)
Vertxxils (3, 7, . . . , 20, 24)

W8 x 48
2 C12 x 30
W8 X 40

SI:
7.5

+Im

ELASTIC, PLASTIC, AND BUCKLING BEHAVIO

OF STRUCTURAL STEE

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

NTRODUCTION

-.

,.

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

.

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

\

From Eqs. (c) and (b),

el

h we obtaiq
60P3 = 60%

e, =

BEHAVIOR OF SlRUC?UIWL SEEL

S o ~ u n oWe
~ must apply a factored load of sufficient magnitude t

0.8 P3(48)
- P3(60)
0.64(29 000)
1.0(29 000)

velop f,

which checks displacements = constant

Fy in the three bars. At this time, the bar forces are simply
P I = A, Fy
P2 = A2F,
P3 = A3F,
P I + P2 + P, = P,,,,,,,
P,,= 36(0.64) + 36(0.75)

. :jlso, f r o b Eq. (c):

or simply P,

+ 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
0.8 P,

1. Plastic analysis is much simpler.

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?
3. The elastic analysis (Example 3-1) indicates bar 2 yields first. Wh
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
they reach F, in turn. When Fy is reached in bar 2, the load in the bar i
P2 = 0.75(36) = 27 kips.

+ 1.25P3 + P, = 30 kips
P, = 30/3.05

= 9.836 kips

P2 = 1.25(9.836)= 12.295 &ips

PI = 0.8(9.836) =
Total
:!ack substitution for e in each of Eq. ( c ) gives e
r ~ a d e should
r
verify).

$ 0.02035 in (which the

By proportion from Example 3-1 the load at this point is

///

p = - 27'0 (30) = 65.88 kips

12.295

l 'ow let us reconsider Example 3-1 using "plastic analysis" in the following
exnrple.

amble 3-2 For

the sketch shown in Big. E3-2 (same as Example 3-I),

n hat are the bar forces when all three bars have yielded?

Beam behavior based on a plastic analysis is similar to the bar problem.

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
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.

Within material homogeneity and rolling tolerances, as well as practical measuring limitations.

120

ELASTIC, PLASTIC, kWD BUCKLING BEHAVIOR OF STRUCKJR.U

STRUCTURAL STEEL DESIGN

The plastic section modulus is obtained as the statical moment

about the neutral axis, which div~desthe area equally. Note that
necessary in order to satisfy statlcs on the section of CF, = 0.

proportions are such that the section can become fullv ~ l a s t i cbefore the nnce
strain hardening (i.e., .depth/web thickness and flange width/flange' thicknes
not too large).
moment
in t
W e ' will now investigate
- .- --- --- ---- rnnrent
--*--vy.
- in detail the Plastic
following several paragraphs. Referring to Fig. 3-1, the moment at initial yield i
A

A..

.LLW

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.
~ ~----C F C P Fshnwn
textbook. The plastic moment M...Y , bv ins~ectionof
- - the
---- S---..".. nn t*. ~ e
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
avo3
--- -- with
---- rliqtnnre /17 t* nv u.vu
centroid from neutral axis, is

..

- --

The value of A7 is called the plastic modulus, Z , so that we may rewrite the
moment as
3 ,

Mp

--

= Lr,,

'lle,.ratia.of Z / S is termed the shape factor, f.

Example 3-3 What is the section modulus S,, plastic modulus Z , and th
shape factor f for the rectangular shape shown in Fig. E3-3?

The shape factor is computed as

- 2
8

'

..

* 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
made of the tables for T shapes, as illi~stratedin the following ex,mple.

Example 3-4 Compute the plastic section modulus and shape factor for
W610 x 241.1 rolled shape.
SOLUTION
The value of A/2 is readily obtained from the WT table CVT30
X 120.5), since this T shape is made from splitting a W610 shape. The
value in the table also locates the center of the area of the T but is with
respect to,the flange.
From Table V-18 of SSDD, obtain

I I ,

I'

I
c-

0
0
xr
I1
iY

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 =-

Figure E3-3

SOLUTIONThe elastic section modulus is computed us

materials equations:

ll

=I - =bh3= . - bh2
c
12(h/2)
6

m3
2
m3 and
The value given in' Table V-3 for 2, = 7.659 x
discrepancy is due to the extra digits used by the computer in computing
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
S,

6.78 x

m3

and the shape factor f can be directly computed as

S =

200(0.4)~

= 5.333

m3

122 STRUCTLTRAL STEEL DESIGN

ELASTIC,

PLASTIC, AND BUCKLING

BEHAVIOR OF STRUCTURAL

portional to strain as defined b

Z bli2/6
/=I50

Z=

h( b/12 - b,/iZ 1

/ ' = L.SI1

( bh2 - -

bill,?

(bli'

b,/i:

Z = b/iZ/3
f = 2.00

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
safety factor F = 1. A,.safety factor of 1 is unacceptble, as it allows for no future changes in structural use/occupancy, or for
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.

z = 4r3/3

/=I70
J

= + ( r i -- r j )

ri

32 --=ri + r j

= from tables

f = Z / S = 1.10 t o 1.18
Average = 1.1 4

Modal value = 1.12

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

The basic safety factor for the steel members in building construction is
obtained as follows. Let the computed strength of the member be defined as the
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
Fig. 1-3b) with no increase in load. Elastic design allows for this unique
behavior of steel but limits the working stresses to the elastic region of the

This value of F is taken as the basic value of F for use in the elastic design
ethod for structural steel for structures other than bridges for highways and
railroads. Railroad and highway bridges are generally subjected to a more

ELASTIC, PLASTIC, AND BUCKLING BEHAVIOR

124 STRUCTURAL STEEL DESIGN

overloading, so the un
hostile environment and a greater possibility
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
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
termed compact shapes. The geometry criteria for these shapes will be considered
in Chap. 4.
For A-36 steel the basic allowable stresses using the previously defined
safety factor becomes
Fa

s obtained based on section properties from these ultimate loa

red to the yield stress Fy and adjustments made until computedf;,

EFLECI-IONS
uld a plastic hinge develop at a point along a beam or column, a very
ection would result. This deflection would, however, have no meaning
ould result in a structure collapse. No structure is

0.6< = 0.6(36) = 21.6 ksi

(the AISC specification allows use of 22 ksi for this case)

Fa = 0.6(250) = 150 MPa
(in SI units)

OF STRUCTURAL

r v

For the AREA and AASHTO specifications, we have

under actual working load conditions will be elastic values.

Since plastic deflections result in a structure co
ignificance, so for this reason deflections
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)
Fa = 0.55(250)

137.5 MPa

We note that the optional rounding of 21.6 ksi to 22 ksi (as should be done using
desk calculations) can create a slight computational discrepancy if a digital
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,
since it is allowed and A-36 steel is the most common grade used. It is not
recommended (at this time) to round 137.5 MPa to 140 MPa, since some
rounding up has already taken place to obtain 250 MPa from 36 ksi.

LENGTH OF PLASTIC HINGE

of the plastic hinge for a W shape ca
d-end beam (refer to Fig. 3-4) as foll
e horizontal at midspan, the offset to t
yo = IM, - ( - 1tfp)

2hlP a

The equation of a parabola with the origin of s as shown is

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
working stress fb for either design method, we have
MY- = -=
1.52s

MP
F,S

l l l i l l l l l l l l l l l l l l l l

fMY

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
which the working or design loads are multiplied to obtain the "ultimate'

beam.

ELASTIC, PLASTIC,AND BUCKLING BEHAVIOR OF STRUCIIIRAL

NOWwe need to find x such that M,
have

My = S,F,. Since y = Mp -

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:
1. Widespread availability of computer programs, which can rapidly sol
simple and complicated structures using elastic methods.
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.

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

Example 3-5 Derive an expression for hfp for the fixed-end beam s
and select a W shape with adequate Z for P,,, = 120 kN.
P, = 170 k N

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
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
val'ue be Mp, and it is evident that Mp will form first at the fixed-end
locations, since the elastic moment is largest at those points. Further
increases in moment increases the elastic moment into the plastic range. It is
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
where the elastic moment is large enough that increases in Pw to P, will
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)

The total Z, required is

Z, = 0.612

+ 0.0058 = 0.6178 < 0.6566 x

m3 furnished

O.K.

Use a W360 x 38.7 beam.

It is still necessary to check bracing requirements. For an elastic design
using Fa = 0.6Fy (commonly used allowable stress), the beam would be
M = - =PL
8- = 9 0120(6)
k 8N . m
The required section modulus S is

Use a W410 x 38.7 section.

from which
O.K.
Sx(reqd,
= 0.60 + 0.0076 = 0.6076 < 0.629 X l o u 3 furnished
By coincidence we have found a section that has exactly the same mass
per meter; in most cases sections obtained by plastic design methodswe
somewhat lighter than those obtained using elastic design, at least when the
///
beam is indeterminate.

!-'or A-36 steel, Fy = 250 MPa.

Pu = Pw X load factor

= 120(1.7) = 204 kN

,,q

The required plastic section modulus is

From Table VI-2 of SSDD, select

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
W360 X 38.7

For the W360 x 38.7, the weight/m

0.38 kN/m (Table V-3).

Example 3-6 Given the propped cantilever beam shown in Fig. U-6, it is
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.

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

2- R,x

0
Taking moments of beam segment AC about A ( Z M , = 0) gives
M~ +

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

x 2 + 2 x L - L* = 0
Solve this by completing the square, we obtain

Figure E3-7

Right span:

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

M, = 0.08579(45

1 . 7 ) ( 6 )=
~ 236.26 kN . m

z,=--236'26 - 0.9451 x
250

The maximum value of M, from either span is used for design (since beam
runs across both spans using a constant section).

Mp = 15(1.7)(18) = 459 f t . kips

low3m3
M~=20(1.7)(15)=510ft.kips

usethis

beam weight is

AZ,

0.5 1
45

-(0.945 1 )

Zx.,,,d, = 0.9451 + 0.0107

0.0107
=

0.9558

!< 1.0861 x

m3

O.K.

bracing requirements).

From Table 11-2, select a W24 x 68 w ~ t hZ,= 176.4 in3. Check tde beam
weight effect as approx~mately (if a borderline case is found, one may be
justified in the additional work for an exact analys~s):

AM'
w, L~
AM; + 2
2 = -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

values of Mp to accomplish this are:
Left
Lett span:

and the required AZ, is

AZ,
= 170.0
Total Zx(reqd)

2.17(12)
36

-= 0.72 in3

+ 0.72 = 170.72 < 176.4 furnished

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

O.K.

?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
~ ~ ~ fully accepted by
elastic design method. It is expected that L R F D W become
the AISC within the useful life of this textbook. This forecast is based on the
facts that this procedure (at least the essentials, called limit states design) is
.,!;eady accepted in Canada and several other countries outside the United
States. The current AASHTO bridge specifications (12th edition) provide an
.,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.
.--!li the LRFD designs are very similar to each other and to the strengt
,jrocedure used in reinforced concrete design. In LRFD, as in rei
concrete, +,factors are used to reflect uncertainties in the material (in t h s
t-h% specified steel strength, F,). These factors are under current study with the
.&rent suggestions as in Table 3- 1.
LRFD uses an equation of the general form

case

+R = $(FdD

+R 2 l . l ( l . 1 0

+ 1 . 4 1 + 1.6PVm,,)

+R

l.l(l.10

l.l(l.10

+ 1.4L)

+R

>

1.5Sm,,)

The general objective with LRFD is to assess each item that influences
design of a structure rather than "lumping" several effects together, as,
example, simply adding the dead and live loads to obtain the composite lo
Larger factors are used with those items that carry more uncertainty, such
snow and wind loads (live-load factors of 1.5 and 1.6 versus the dead-load fa

+ 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

2%ble 3-1 Current recommendations for

equations for several loadings including wind and snow mi&t re

LOCAL BUCKLING OF PLATES

factors

Stress condition

Suggesteda

Canada

AASHTO

~ & s i o nmembers
Yielding (4)
Fracture (F,)

0.88
0.74

0.90
0.90

1.O

sending
Rolled sections and plate girders

0.86

0.90

1.o

0.90

1.o

1.0

Columnsb
TJ 2 0.16
0.16 < 7
1.0
TJ > 1.0
:!hear
Webs of beams and girders

0.86

0.90

1.o

Cvnnections
BoltsC
'Welds

0.70- 1.00
0.80

0.90
0.90

0.86
0.90-0.25~~
0.65

a1 stresses, perfectly plane, homogeneous, and isotropic that is subjecte

o m compressive load along opposite edges. Under this stress the ~ l a t
press uniformly until the buckling stress is reached, When the buc
stress is reached, the plate will deflect in a single wave or a series of wav
depending on the edge (boundary) conditions and length to width ( a / b ) ra
with a resultant redistribution of the compressive stresses until, with the additio
of load, the entire plate is buckled.
From the theory of plates as proposed by several authorities,j the critic
elastic buckling stress F,, is

" 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).
See Sec. 8-10.

Metal Structures (New York: McGraw-Hill Book Company). or Johnston, Guide ro Srabiliry Deri2n
Criteria for Metal Structures, 3rd ed. (New York: John Wiley 8i Sons. Inc.).

,
7

136

STRUCTURAL STEEL DESIGN

where

ELASTIC, PLASTIC,

AND BUCKLING BEIMVIOR OF STRUCTURAL STEEL

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

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
rious boundary conditions that are possible. This has been done by several
7lthorities, but as a convenience the author has further combined the effect of
?/[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.
Rewrite Eq. (3-6) in terms of k:, to obtain

:p
9

u7

h'[

i--- h ---I
i =O;Y

k c = I 1 5 ,re111
FIdnze ,dnle >I.

it

,h.111=

k; = 0 63 l l ~ n g e \
k: = 4 5 for web

J,

$,

Figure 3-7 Compress~oncharactens~csfor rolled shapes shown. Note that 111s generafly nec
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.

Fcr = X E ~ ; ( ; )

13

dew divide through by

2
Fcr A -~k:(i)
Ye will temporanly hold this equation. Now using Eq. (3-7) for X 9 t h
F,, = 0.755, we can, with some rearranging, obtain

Fcr

-=

0.1875

c2

4 - Fcr

Compression charactenstics (kc and wldth b in compress~on)of three rolled

shapes are g~venin Fig. 3-7. The kc values shown have been found to agree
reasonably with tests. Adjustments In kc are necessary because very few plates
are free of imperfections and residual stresses
If we use the value of k: = 0.63 shown in Flg 3-7 for the flange of a W
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
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
'9

For kc

= 11 3

=pC

65
say -

vz

Fcr / A
3.615, X = 1, and Fc, = 0.754, the limiting b / t ratio for A-36 steel is

",/@=Em
t
0.75 x 36

= 62.3

The current AISC specification allows a b / t (uses b,/21/) ratio of 6 5 / 6 . Note

also that if we consider the web, k: = 4.9, we obtaln
190
say -

I t is often useful in using Eq. (3-9) to set up a table of A vs. F, with values from
FCr = 0.754 to F,. For A-36 steel, typical values are as follows:
/

i;,,ksl
/F

27.,
28.0;

2ii 0
33.0
36

E,, ksl

1.00
0.922
0.741
0.407
0.0

29 M)O
26 738
21 489
11 803
0

say 11

which is also in AISC.

Values of k, = k, are also shown in Table 3-2 for the cr~ticalstress to
produce shear buckling. The critical buckling stress for shear can be derived in a
similar manner to that for compression, with the substitution of an appropriate
buckling coefficient k, to obta~n,from Eq. (3-6),
F C ~=S

k,n2XE
12(1 - p 2 ) ( b / l ) 2

138 s d u m

ELASTIC, PLASTIC, AND BUCKLING

STEEL DESIGN

BEHAVIOR OF STRUCTURAL STEEL

139

I is usual to assume that the four plate edges are simply supported in shear and

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
. . ."
C
r7
C
crs
l y
- 'y
Fs =
SFX*
1.67xfl
2.89
Mo~t~practlcal
steel design problems consider either buckling in compression 6r b'uckling in shear. w h e r e both stresses act simultaneously, the reader
should consult books such as those by Bleich, Johnston, and Timoshenko and
Goodier cited in an earlier footnote.
a

where all terms have been previously identified except be, which is shown in Fig.
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
is not often directly evaluated. It is used more often in a more indirect manner;
for example, AISC allows use indirectly via Appendix C3, which states: "When
the width-thickness ratio of a uniformly compressed stiffened element exceeds
the applicable limit given in Sec. 1-9.2.2, a reduced effective width, b,, shall be
used in computing the . . . ."

PROBLEMS

3-9 POST-BUCKLING STRENGTH OF PLATES

Experimental evidence shows that a buckled plate does not result in immediate
failure. Rather, there is a considerable strength reserve attributed to the effect of
the adjacent plate material, which restrains the buckling and allows transfer of
any post-buckling load increase to the unbuckled zones. This situation is
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

3.1 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?

3-3 What is the plastic moment capacity (kN . m) of a W 9 2 0 x 200.9 section for Fy
Answer: 2865.2 kN . m.
34 Verify 2, in Table V - 3 for a W 6 1 0 x 241.1 rolled shape.

345 MPa?

3-5 What is Z, for the geometrical shape shown in Fig. P3-5 if it is used for a beam?
*

3-6 Select the lightest W section to satisfy

F, = 36 ksi and plastic design.

bending for the span and loading shown in Fig. P3-6. Use

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?

Answer: S = R = $ .

0.5Fy? What are two

>

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.
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.

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).

----

.-ii--ii--

#!fI

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.
Cloxup k$s.*

(n

'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

--------

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)

Q = statlcal moment of area above point where shear stress is being
determined = A7 (refer also to Flg. 4-2)
t = thickness of beam at point of shear stress investigation
V= critical shear force from shear diagram or direct computations

e same time one with an adequate S..

ISC) have the sections commonly used
r beams ranked in descending order of S and wth respect to the X-X (strong)
weight, so that considering a group as
defined by extra llne spacings, we have

The section is checked to ke sure that it is adequate to carry its own weight,
and finally the working load deflestions are checked. Sometimes deflections

2. The hghtest section in the group is at the top and the heavlest is at the

1. The largest S In any group

is

at the top of that group

radang allows rapid detemnatlon of the most economical roIIed

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

Chechng the beam weight, we obtain

ST,,, = 138 + 3.24 = 141.2 < 143 in3

Lb = 0 < LC
rind use Fb = 0.665 = f x 36 = 24 ksi.
Rearranging Eq. (4-l), the required section modulus with strodg axis
bending is

We must find a section with a somewhat greater value of S so that the beam.
,
weight can be carried. From Table 11-1, select

furnished

Use a W14 x 90 beam.

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
for the SI units and selection of section sizes. Use F, = 250 MPa.

<-

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
depth is limited to 16 in?
Sreqd is approximately 145 in3, since the weight will have to be
SOLUTION
larger than 68 lb/ft of the most economical section selected in the example
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 ,

SOLUTION
First, draw shear and moment diagrams.
Second, obtain M,,, = 436 kN . m. Assuming full lateral support, w
obtain
Fb = 0.67 X 250 = 167 MPa
The required section modulus is

Try

W16 X 89:

S = 155.0 in3 d = 16.75 in

W 14 x 90:

S = 143.0 in3 d = 14.02 in

W12 x 106: S

145.0 in3 d

12.89 in

> 16

N.G.

(next try)

Note that 167 MPa x lo3 = 167 x l d kPa, which is consistent with 436
kN . m. Check Table VI-I for Sx somewhat larger. Find:

154

STRUCTURAL STEEL DESIGN

b

Check AS, for beam weight = 1.11 kN/m:

Col.

-r

S,,,

= S,

+ AS,

= 2.611

+ 0.0532 = 2.6642 < 2.8841

Col
>,-

:!.,.

..

cur
.,,.-,

..

O.K.

m3

Use a W610 x 113.1 beam.

major difference
Continuous beams are designed similar to simple beams.
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
now be illustrated.
Example 4-4 Use the computer output of Example 2-2 and check and/or
redesign members 22, 23, and 24 (refer to Fig. E4-4 for the gravity load
moment (which inspection of Fig. E2-4d indicates are the critical values for
design). Use F, = 250 MPa.
*

+ M for design is
+ M = 1i1.3 + 57.86 +2 140.6 (0.1) = 121.2 < 126.5kN - rn

Since the largest moment is - 126.5 kN . m, this is used to select the sectionNote that the compression flange is part on the top and part on the bottom,
with inflection points as shown on the moment diagram of Fig. E4-4.
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
in Fig. E4-4.
Since we are using the same section for all spans, the largest design
( k m o m e n t is in span 1 or 3, by inspection. Note the moment in span 2 is all
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.

r,,

,The

From Table VI- I , select a

W410 x 46.1 : S = 0.7735 x lo-' m3
LC = 1.78 m > 1.20
O.K. for no bracing
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 compression 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

midspan brace to the bottom flange for the center span, to produce

& = -2'3
- - 1.13 rn < 1.78

2
&Check the beam for self-weight (weight = 0.45 kN/m). By proportion
( s i ~ c ethe beam loading is uniform)
t

S,,,, = S

DESIGN OF B W L I S FOR B M N G

"I

+ A S = 0.7575 + 0.0086 = 0.7611 x

l o v 3m3 < 0.7735 furnished

Use a W410 x 46.1 beam.

The reader should verify that the method used to obtain A S is both
correct and the most practical means available. gince a W410 x 59.5 beam
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
'.i?a later chapter.
///

4-4 WEB BUCKLING AND CRIPPLING

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
dtfier.the.d/tw ratio or the stress that can be used for the given d/tw ratio. This
1s allowed for in the several specifications. Web buckling is illustrated in Fig.
4-5b.
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

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
reaction length + an additional distance using a 1 : 1 (45") slope through the k
distance of the section. The section property tables tabulate k for the several
rolled sections. The k distance is measured from the outer flange face to the top
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
A, = (A' + k ) t ,
The allowable stress at this location is taken by AISC (see SSDD Sec. 1-10.10.1
to be
F,
At a reaction with j

0.75Fy

R I A , , we obtain

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

where

N = reaction width; a basic value of 3 f in or 89 mm (width of standard

brick) is often assumed; iV = width of column or load-deliverin
element for interior loads
R =.reaction or other concentrated load, kips or kN

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
The shear stress distribution across any section subjected to bending c
computed using the equation presented earlier:

When large uniform loads are carried through the flange to the we
be necessary to check the compression stress& and limit the valuedto

&

VQ

5 0.75Fy

f, = It

Example 4-5 What is the allowable reaction for a W16 x 40 using the basic
value of N = 3; in, with A-36 steel? What column load can be transmitted
using a W 8 X 31?

A plot of shear stress using this equation is illustrated in Fig. 4-7. We note that
the average shear stress based on

f" = dlw

SOLUTIONFrom Table 1-3, obtain for a W16 x 40:

k
t,
mr.

=
=

1.03 in
0.305 in

R = (N

+ k)tW(0.75<) = (3.5 + 1.03)(0.305)(0.75 X 36) = 37.3 kips.

in = N.

differs somewhat from the maximum value shown in Fig. 4-7 (in this case about
23 percent) but is considerably easier to compute. AISC allows use of Eq. (4.T)
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,).
The allowable shear stress F, for rolled sections is computed as:

+ 2k)t(0.75F,)
= [8.W + 2(1.03)](0.305)(0.75 X 36) = 82.8 kips

P = (N

F, = 0.405

~ x a m ~ 4-6
l e What is the allowable reactidn for a W460 x 74.4 .section
using the basic value of N = 89 mm and F, = 345 MPa? What column load
can be transmitted using a W200 x 46.1?

F,

(AISC)

(AASHTO)

0.335

F, = 0.35Fy

(AREA)

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

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.

P = [203 + 2(27.8)](0.009)(0.75 X 345) = 602.2 kN

\
X: 9 ~

C114h0 X 1 0 5 . 7

///

Bridge reactions are generally supplied by special fabricated bearings, and

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

100 \I PJ

1 ' 1 1 ' 1 ' 1

I:,< T r ~ b i ero gcr T. 1

i= <I 2

107 4

Y rzc

\!I'd

-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

shown Fig. E4-7 such that either F, or Fb will control? m a t length
required for the reaction?
II

4-6 STRONG- VERSUS WEAK-AXIS BENDING

k~p\/ft

. -- ,L-

Figure E4-7

SOLUTION
Data for a W24 x 94:
dz24.31in
tw = 0.515 in

Depth limitations or other considerations may require onentation so

bending is produced about the Y-Y m s . The allowable bending stress in
AISC specification for compact rolled sections bent about the Y- Y axis and
solid round, square, or rectangular bars may be taken as

k=1.53ln
S,

222.0 in3
where B= 0.005 for fps units
= 0.0019 for SI units

wL
f,=-- 0.44
2twd
W L ~

f -8Sx = 0.665
From Eq. (a),

0.4(36)(2)(24.31)(0.515)
L
Substltutlng this w into Eq. (b), we obtain
=

360.6
L
/

=-

4-7 DEFLECTIONS

-360'6L - 0.66(36)
L

8%
24(8) (222)
360.6 x 12

Back substltutlon gives w = 360.6/9.85

tain
fb =

= 2(24.3 1)(0.515)

(N
.

indows to fail to function properly. Cracks can be produced in

36.61 hps/ft. Checking, we ob-

14.4 ksi

The reaction length N based on R

0.75Fy = 27 ksi is
1

Deflect~onestimates under working load are often required to ensure that floo

O.K.

36.61(9.85)2(12) = 24.0 ka
8(222.0)
36.61(9.85)

fv

9.85 ft

N = - -R

about both the X and Y axis takes place.

0.K.

= WL/~=

180.4 kips and F, =

+ k)twF, = R

a failure of the roof drainage system by pluggng from accumulat

flections for simple beams can be computed using superposition of

t w Fa

180.3
- 1.53 = 11.44 in
0.515(27)

can be computed using the general differentlal equation

EIylV = - w

L STEEL DESION

For L
and

=.

6.5 m, we obtain C, = - 173.98 (computer output = 0.0043 rad

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

12.6 mm (approximately

in)

//

-.-8 BIAXIAL BENDING AND BENDING ON UNSYMMETRICAL

SECTIONS

use superposition to flnd the stresses at cntical locations, using

.,

r! .

and for shear

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
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
This is a very complex stress state that involves biaxial bending and torsion.
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%

FbXSX

+ -=

MY

1.0

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.
The latter equation for Ixy is a product of the inertia term produced when at
least one of the X and Y axes is not an axis of symmetry. Principal axes a
produced as mutually perpendicular axes through the centroid of area such th
the moments of inertia are a maximum with r e s ~ e c tto one of the axes and
minimum with respect to the other. The principal axes are axes of symmetry
symmetrical sections but can be found for unsymmetrical sections. If
product of inertia Ixy is not zero, the axes of interest are not principal axes.

4-8.1 Symmetrical Sections with Biaxial Bending

,the X and Y axes are principal axes but th
to the principal axes (but the resultant
ikin ipf axes), biaxial bending is produced. For this
.PQ
'"
* cessary
to determine the components of load perpend
I

Example 4-10 Deslgn a roof purlin using a channel section for the si
sheds and for the most cntical loadlng cond~tionof Example 2-5. Space the
purlins 6 ft horizontally as shown in Fig. E4-10 and use sag rods at the
midbay spacing, giving an unsupported length of 12.5 ft for bendin
moments about the Y axis of the purlin.
SOLUTION
Additional data from Example 2-5:
Dead load
Live load

= 20.0

snow load

25 cos 14.04

25 cos B
re

= 24.25

'Total
.

.?

Vertical load/ft

psi

= 44.25

psi
psf

= 6.1 8(0.04425)
= 0.273

kip/ft

>?
.!&

15

DESIGN OF B M M S FOR BENmING

STRUCTURAL STEEL DESIGN

i.
{

values, we obtain

-*hk
tJ

fb = - 2046x - 1220y

(if set to 0, gives location of neutqal

t point A , x = - 0.0419, y = + 0.1357 m.

fb = - 2046( - 0.0419) - 1220(0.1357)
= - 79.8

b t point B, x = - 0.0419, y
fb

MPa (compression)

0.0673.
= 167.8 MPa (

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
be as large as 14 mm but has a good probability of being much less than
this.

4-9 SHEAR CENTER OF OPEN SECTIONS

= -

+ = 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
bt
J= b/t 4 10 (section webs and some flanges)
3

At point C, x

0.1 10, y
fb =

- 0.0673.

- 142.9 MPa ( -

compression)

If we use fb = M c / I :
At point A :

14'21(0'1357) (lo3) = 57.3 MPa (compression)

33.63
At points B and C:
fb =

14'21(0.0673) (lo3) = 28.4 MPa (tension)

33.63
These values are considerably different'from those computed using
metrical bending.
Check the deflections approximately using Table IV-5 for a
supported beam:
fb =

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
shapes. Fortunately, most sections have the shear center at a convenient Iocation
(see Fig. 4-10), for example:

1. If a section contains an axis of symmetry, the shear center is on the axis.

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).
3. For all sections consisting of two intersecting plate elements (angles, tee
etc.), the shear center is at the plate intersection.
The shear center E, (using symbols as in Table 1-6) for channels is rea
derived to be
bl
E, =
2b,$ + h1t,/3
b, = b, - t,/2
where
h'= d - t/ = average section depth
9, t,= flange and web thickness, respectively

rr

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:
L

x=--

f,

Shear enter l o ~ a t ~ o n

E,

- 4-10 Shear center location for several rolled shapes.

v = o

82
20.7

-= 3.96 ksi << 0.6FV(22ksi)

At x ,= L / 4

3.13 ft:

M,

w(L - x2) - 0.35(12.5 - 3.13')(12)

2
2

j.7 in kips

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
of ?,~terials.
1' I-.: bending stress for conditions where torsion is developed by the applied
lo22 * ,t passing through the shear center is approximately

whe .e M c / I = usual bending stress computation

V = shear at a distance x along beam from origin
by= width of flange being stressed
x = distance from origin of axis to V and where fb is desired
If= moment of "inertia of flange being stressed; approximately Iy/2
for a channel

Example 4-12 What is the approximate bending and torsion stress in a

C10 X 30
channel loaded as shown in Fig. E4-12?
.,",
4

"

_-.
,,

= 0.3

+ 31.6 = 31.9 ksi >> 0.65,

Note that if the channel beam were designed without considering torsio
and shear center location, the apparent bending stress at rnidspan indicates
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
author would assume that the section is considerably underdesigned and use
a different section, as there is too much design risk involved for the small
amount of savings obtainable from using a lighter section.
The torsion shear stress can be evaluated as
Ve' t,

L=,

J = contribution of web 4 two flanges

B r ~ i kv e n e e r

I+-

Figure E4-12

I 2 4-5'

SOLUTIONFrom Table 1-6, the section properties of a C10 X 30 are:

S, = 20.70 in3
I, = 3.94

= 21,

Eo = 0.705 in
b, = 3.033 in

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
the increased stiffness at the flange-to-web intersect~on.We w11 use the
table value of J = 1.22.
e' = Eo + 7 = 0.75 1 + 0.649 = 1.40 in
f, = 1'092(1'40)(0.436)
1.22

0.546 ksi which appears satisfactory

///

174 STRUCTURAL STEEL DESIGN .

4-10 DESIGN OF LATERALLY UNSUPPORTED B

DESIGN OF B E A M FOR BENDING

= 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
bracing."'&>,v6ry large lateral restraint is not required, as evidenced by the
. 'i
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 involving 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
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
,

ll",,. ,

'

hen*

or, alternatively,

where

J=c,t:orsion
constant previously defined and given in the table of
ib
:'section properties in SSDD
C, = warping constant =$ t,bf3(d - t,)2

~h~ AISC specifications can now be developed using the preceding several
ations and some additional simplification and rounding of numbsrs. Since
ut 1946, the AISC specifications have used the following formula based On
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

Fb =
Fb =

12 OOOC,
Ld/A j
82 700Cb
MIAj

stead of 0.41d, to obtain

[AISC Eq. (1.5-7)]

2 0.6.

(SI units)

wher: ,I,= area of compression flange = bjtf

d=depth of section; note that the ratio d/A, is computed and tabulated
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
sign. Use Cb = 1.0 when the moment in the interior of the unbraced
length is larger than either of the end moments and regardless of
sign.

The minimum L/rT ratio for using Eq. (4-26) is found by equating F,
.6Fy to obtain (with slight rounding):

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

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

+ [Eq. (1.5-7)12
2

Eq. (1.5-6b)

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
for a W36 x 300, obtain LC = 17.6 f t and L, = 35.3 ft. Since L, > I,
(46 > 35.3), we immediately note that Fb < 0.6Fk.

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

1. Initially assume that Fb = 0.6Fy and make a tentative section selection using
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
design factor. Using the tentative section or one somewhat larger, compute Fb
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
determineckby L/r,.

+ 952.2 in . kips

From Table 1-3, obtain the section properties for a W36 x 300:

Alternative computation of d / A,:

b,
d = 36.74 in

16.655 in

r,

1.680 in

Alternative computation of 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 computedCb = 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

Estimate the beam weight as approximately 0.1 x load:

Also use AISC Eq. 1.5-7 [Eq. (4-23)]:

12 oOo
Fb = 12 oOo(1) Ld/A,
46(12)(1.3 1)

(0.03 x 35 + 30)(0.1)
35

16.6 ksi

0.09 kips/ft

Use , f b = 16.6 ksi (largest value):

t.

M
S
952.2

204P

M due to beam

Fb + M = SFb

f b C - =

85.6 ki

'"he preceding example was easy to check, since the beam size has been
s e l e c ~ ~and
d it is only necessary to determine the allowable bending stress.
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
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
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
so!:~!ion (the reader should verify why this is true).
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.
trolley travels on a 90-lb railroad rail fastened to the top of the flange
(90 lb = 90 lb/yd = 30 lb/ft).

-=

= 237 ft

M,,,,,

16.6(1110)

P = l7 473'8
204

. kips
,.
..

.-

.'.

With a large unsupp

somehow estimate the bea
selected sections with moments givkn up to
ft is only slightly more, use this table as a guide and select the fol
tentative sections:
~

1 x 296:
14 x 90:

W16

S,

131.0

L,

39.9

S, = 143.0

L,,

34

L,

28

100: S,

W18 x 97:

175.0

S, = 188.0

(Table 11-4 for W12)

L, = 24.1

ry a W14 x 90 and check AISC Eq. (1.5-7). Take Cb = 1.0.

12000
Fb = (35)(12)(1.36) =21ksi
M , = F bS

.r

21(143)
=-------=250ft,kips>237
12

O.K.

Check the deflections:

I, = 999.0 in4

From Fig. 4-8, obtain the needed deflections at forward load:

~ . .,

,,

,,,

.., .,..,..,-.

= 3ok

Figure E4-14

S ~ L U T IFind
~ N the maximum moment. Write an equation for M in terms of
and take d M / d x = 0.
[15x + 15(x + 6)][L - ( x f 6 ) ]
M =
L
dM - 0 = -60x - 270 + 30L
x = 13 ft from left end
dx
M = (30 x 13 + 90)(35 - 13 - 6 ) = 219.4 f t . kips
35

Aload

15(19)'(1612
3 EIL
1

= -(47.80

3 IL
-=--

15(13)(16) ( L Z6EIL

1-31 -

162)

+ 43.03)(1728) = 1.5 in

'('2)-l.2<l.jin
360 3 360
Use a larger section; a side computation indicates that a W18 x 97 does not

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
3
4 or more

100
90
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
1 lane

Bending moment and shear (lateral distribution

on steel interio? beams)
For:

concrete floor
steel grid deck
steel grid deck

+ lanes

( ) = SI value

5.5 (1.676)
4.0 (1219)
5.0 (I 524)

7.0 (2.134)
4.5 ( 1.372)
6.0 (1.828)

<'4 in (102 mm)

> 4 in (102 mm)

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
With steel grid < 4 ~n (102 mm)
> 4 in (102 mm)

6.0 (1.828)
4.5 (1.372)
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
P, = rear wheel load of H 20, H 15, H 10, etc. (H 20 orSHS
+ 0.61
20 = 32/2 = IQkips = 72 !d)

.P, LN
M = - -S + 2 P , f t . ~ p ~ / f t M=----9.74 r,
32
<f
Use M ' = 0.8M if three or more stnngers.
>, am*
I

*..*

""

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

Step 3. Find the beam section.

The required sectlon modulus is based on F,
pression flange is laterally supported.

Ballast plate: include 10 percent for corrosion protection

0.015(77 k~/m~)(0.76)(1.10)
Estimate cross-beam weight

S =-=-=
F,
137.5
591

Miscellaneous maintenance, storage of ties, material

Total =9.01 kN/m

the dead-load moment is

- 9.01(5.79)2
Y*

8
The dead-load shear is

4'44/2(5.79 2

= 37.8 + 4.8 = 42.6 kN

.m

1.15PwD
s
1.15(110/80)(80
x 4.448)(0.76) = 280.6 kN
1.524
with P / 2 placed on each rail. The live-load shear is
=

Pw -_ -280'6
=
2

140.5 kN

The impact factor is

3 L2
I = -30'5 + 40 - S
150

where S = beam spacing, L

3(5 79)2
30.5
= 79.5 percent
+
40 0.76
150
The design live-load shear is
140.5 x 1.795 = 252.2 kN
The design live-load moment is
=-

ML

280.6(2.177 x 1.795)
2

The total design moment is

M,,,,,
= M, + ML = 42.6
&vt

548 kN

c,,
1'

+ 548 = 591 kN . m

The total design shear is

Vdeslgn

'd

= 28.3
a

"

&

'L

+ 252.2 = 280.5 kN

4.30~10-~m'

m3
d

t , = 13.8

> 8.50 mm

758 mm

We note that the weight is 0.08 kN/m larger than assumed, but the sec
modulus is more than adequate for this small difference. Check the sh
f

Vd = 9.01(5.79) + 4.44 - 28.3 kN

2
2
Step 2. Find the live load on the beam. (D = 0.76 m, s = 5 ft8=1.524 m)
P

t, = 19.3 mm

0.555, since th

ould use a W760 X 147.3 section; however, we will arbitrarily go

ost economical section, W760 x 160.7/ 1.58:
Sx = 4.8997

WL2 + track
Md = 8

FO

"

- V dtw

280.5
0.758(13.8)

26.8 << 0.35 F,

O.K.

(The end connections to the plate grder wlll be designed in Example 8-9
Use a W760 x 160.7/ 1.58 rolled section.

4-13 COMPOSITE BEAMS

A composite beam is one whose strength depends upon the mechanical interac
tion between two or more matenals. Reinforced concrete beams are actualIy
composite members but are not generally designated as such. Most often the
term "composite beam" in building and bndge construction is applied to a ste
section on which a concrete floor or bndge deck has been cast. The concrete
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
illustrate shear connectors in position to bond the concrete to the beams so
composite action is obtained. The shear connectors shown are called shears
and are welded to the beam (and through the deck pan in Fig. 4-16). Other
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
eck, relative slip occurs at the interface of the two materials and the result 1
omposite section. In some cases we may design for a specified amoun
site action. Actually, there wlll always be some small slip due to
nequal deformations in the shear studs, concrete, and steel
ractical purposes it can be neglected In composite design. The brief introd
composite design made here will only consider full composite action.
e method of construction and code specifications are significant desl
meters in design of composite sections. Figure 4-18 illustrates the effe
h of concrete to use in a transformed section b' using the modular
n = E,/E, as in several mechan~cs-of-matenalstexts.
The modular ratio depends on the 28-day concrete design strength 1,
shown in Table 4-2.

~i

190

a*",:
STRUCTURAL STEEL DESIGN

Figure 4-16 Composite building construction. (a) Shear studs in

construction and metal deck for form. ( b ) View of inetal deck
encased by concrete floor. (c) Bottom vlew of metal deck resting on floor beams and girders. Metal
deck is welded to floor beams wlth shear studs.

DESIGN OF B U \ S FOR B

4-17 Composite bridge deck (two lanes

interstate hlghway uslng six stringers. (a
site bridge deck nearly ready for concrete wit reinforcing and shear studs in place. Side vertical
is for parapet. ( b ) Closeup wew of shear studs (c) Undernew showng stringers and unshored
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
the concrete slab is added. This assembly is then shored (braced or propped)
so that no (or relative small amounts of) deflection can occur, and the
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.
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
concrete deck (refer to Fig. 4-17). The concrete is poured and at this time the
steel beam carries the dead load of steel, formwork (as used), and the
concrete. After the concrete hardens, any formwork is r
at this stage of the construction that the steel beam h
stressed with the weight of the steel beam plqs a proporti
weight of the concrete deck.

II

.0

(,'I

i =

*pJl#

se srn.~llestvalue ol b
b = LI-I

b=b,+i

AASHTO ~ = I s + ~ I 2, I
h=b,+ 16i

-\-\SIITO h = I:[

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

ny differences ldentrfled. ( a ) Edge. (b) Intenor.

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
2500
3000
3500
4000
5000
6000

13.8
17.2
20.7
24.1
27.6
34.5
41.4

2550 @SI)
2850
3100
3400

3600
4000

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
is

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

= 264.6

fq I = -2
4

= 33.07
8

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 actingV = 30.96 klps

V
j--=-=
dt,

30.96
= 4.9 ksi << 0 . 4 5
1g(0.355)

O.K.

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.

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

L
<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

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

bridge decks AASHTO Art. 1.6.16):
.

40 rnm haunch

A: - 0.02148AS = - 1.0140 x loT5

Solve for A, by completing the square:
I

v
Cross frames

. . . ..

.'I..,

77.200
4 (? 5550

The maximum ratio of A,/A, = 0.0214, based on f: and f, and taking 75

percent of the value to ensure initial steel yield. The actual steeI percentage
is
0'0004829
= 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
Increase 10 percent for roadway debris, snow, etc.:
6.37 X 1.10 = 7.0 kPa
The dead-load moment is computed as

The impact factor for the slab is (using L = s = 2.033 m)

z2= l5 = 0.375 > 0.30

use 1 = 0.30

L 38
Since the s / L = 2.033/22.4 = 0.091 is so small, one-way principal slab
reinforcement perpendicular to the traffic flow will be used in the slab
design. For this case and five spans providing slab continuity across the

JY

Use A, = 0.0033(0.19 x 1.0) = 0.000627 m2/m.

Step 2. Design the steel stringers.
Note it was necessary to design the slab so that the dead load carried by
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
Haunch: 0.040(0.42)(23.6)

= 0.40

kN/m

Rolled beam assumed

=3.60 kN/m

2~iscellaneous,
including debris, formwork, etc.
Total
(I I!

= 1.00 kN/m
= 18.00 kN/m

A4J.Z STRUCTURAL STeEL DESIGN

DESIGN OF BWki.5 FOR BENDWG' 21)3

The maximum dead-load moment at the center of span is

The maximum live-load moment, including impact, is based on the

AASHT3 truck and the stringer spacing.
I

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 haunch.

15
= 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
Factor

-= 2.033 = 1.213 wheels

1.676
1.676
Slnce 1.213 wheels = 1.213/2 = 0.6065 axle, the adjusted (for deck1 [slab
action) stringer bending moment is

It is usually sufficient to select a steel beam one or more sizes Iess than
that required for noncomposite action. The section required for noncomposite action is approximately (and flange laterally supported)
s x = - = MT
0.555
Tentatively, try a W920

The maximum shear occurs when one of the rearmost truck wheel's [144-kN
(32-kip) axle] load is at the beam end:

1129 + 1090.5
0.55(250)

16.14 x 10-3

m3

.*

364.6/3.58:

Ix=6701.3x10-6m4

S,=14.67X10-3m3

t, = 34.3 mm

This value must also be adjusted for plate action and impact:

b, = 419 mrn
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
= 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*.

i..)

sgF-

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

or, preferably, have the stringers cambered at the mill with a
(and being sure to place the cambered side up on the job
d'faster would occur). The remainder of the beam properties
ek.) are within the assumptions, so we can continue.
Compute I of the composite section (refer to Fig. E4-17c).

6'

'

snnL DnsmN

AY = ZM*.,
A = 46.52 b'd'
= 46.52 + 0.254(215) = 101.13 x

Final:

I top '%=-!!??!-=26.1Mpa
Stop

41.77

26.1
8

3.26 MPa << 0.41

f,=-=

O.K.

m2

The section stress profiles are shown in Fig. E4-17d. Check 100 percen
overload stresses against 150 percent of allowable stresses:

f, = ----2(1091) -

f,,,,,,, =

106.2. MPa
20.54
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

= 16 121.7

m4

The section modulus values are

""P

= I31

SbOt

l6 121a7
+ 40 215 = 41.77 x 10-3

l6 121'7
= 20.54 x 10-3
458 327

m3

The stresses are:

Initial:

f;=f,=-=
M"

S,

m3

Step 4. Design the shear connectors.

Use three 20-mm studs, as shown in Fig. E4-17e, with L = 150 mm
Thus the L / d = 150/20 > 4 is 0.K:
The studs must resist the smaller of
0.85f:bdf
vh~=
-7
(neglect haunch)

-I 129

14.67

- 77.0 MPa

5130 kN

A F
v , = x =
2

controls (after computing V,)

46.52(250)
2

5815 kN

@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

R,

182 kN

The ultimate strength of the shear stud is

S,,

0 . 4 d 2 m

V = 182(0.6065)(1.25) = 138 k N

E=4 7 4 0 e

= 0.4(0.020)~(27.6 x 24 900); (w/SI adjustments

N, =

5130
= 38.7
132.6

4
use 39 (multiples of 3)
I

With studs in groups of three, there will be 13 groups. Now set,the

spacing based on fatigue.
.
The maximum range of live-load shear is 0 to 214.6 kN.
Zr = c d 2 = 0.0541(20)~= 21.6 kN

At X = 3.5 m from supports, the live-load shear plus impact is (refer to

E4- 17f)

0.5022 mm

0- .---K~

Using the spacings above, the first 3.44 m requires 1 1 groups of studs,
next 3.43 m requires 9 groups, and the next 4.33 m requires 9 zrou~s.for
total of 29 groups at 3 studs each, for a total of 87 studs vs. th
for ultimate shear considerations.
Step 5. Design the diaphragms.
Lateral load:
Wind at 24 kPa X exposed surface area and, referring to Fig
E4-17a: H = 4 6 0 + 225 + 270+ 916mm = 1.871 m
Curb load at 500 lb/ft for side opposite
Wind

nz. = ---301.6) = 0+323

Pitchp = S,
200.6

0.323(%)

< 0.6 1 max. s ~ a c i "n zallowed

= 132.6 kN

. The minimum number of studs required from the end to the middbn
tance of L = 22.4/2 = 11.2 m) is

24(1.871)
Total

Note: there is no wind load from the truck because the deck acts as
diaphragm.
With diaphragms spaced at 5.55 m,
Pdiaph
= 52.2(5.55)

289.7 kN

(interior diaphragm)

- 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<

W530
$rk@yproportion since Q / I = constant,

65.5/0.64:

d = 525 mm
1, =

8.9 rnrn

A = 8.39 x
r, = 32 mm

Y to
> --j-

> 8.0 mm

O.K.

O.K.

m'
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.

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

L
22400
-_
- 22.4 mm > 17.6 mm O.K.
1000
1000
If the designer assumes that the actual moment diagram resulting from thre
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

bv, =

+(FdD

+ FLL)

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
statistical terms, we obtain

3
-

--I

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

4-14 BEAM DESIGN USING LOAD RESISTANCE FACTOR

-,-.,-.--=-

,w

---.

wUL'
4.7 1 (25)2
Mu=-= 367.97 f t - kips
8
8
(4 = 0.86 from Table 3-1)
Mu = OFyZ

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
Fb = +F,

+ = 0.86

Mu = +F,Z

vu=

'

+( v3 )

dt,"

and Mu is computed using factored dead and live loads. This is illustrated in the
following example.

Example 4-18 Given the floor system shown in Fig. E4-18 for an off
and live
80 psf; use LRFD design and A-36 steel.
Make a preliminary design selection of a W shape with depth no
factor.

From Table 11-2 of SSDD, obtain a W21 x 62 with Z = 144.4

20.99 in, and t, = 0.0400 in.
Check the beam weight:
Awu

1.1(1.1)(0.062) = 0.075 kip/ft

By proportion,
Aw Z
0.075(142.6)
A Z = U -= 2.27 in3
4.7 1
Wu

O.K.

Check the shear:

: building using simple framing. Take the loads as dead = 75 psf

+- F~

Use a W 2 1

v3

dtw = 0.86- 36 (20.99)(0.40) = 180.3 >> 58.9 kips

62 section.

v3

O.K.

,
,

212

STRUCTURAL STEEL DESIGN

.$r"

-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

ction under live load to L/800. Use j; = 4 ksi.

no twisting occurs due to the eccentricity of the brick. Use A-36 steel.

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
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 59.8 for Al.

Answer: W410 x 46.1.

4-14 Design the floor beams spanning between columns assuming supports and
= 345 ~
p ~
Answer: W410 x 46.1.
4-15 Do Prob. 4-8 using unshored composite construction. U s e x = 3000 psi.
4-16 Do h o b . 4-12 using composite unshored construction but use F, = 250 MPa steel (instead of

F,

,7 +(I

= 345 MPa steel; why?). U s e x = 21 MPa.

4-17 Redo Example 4-16 using shored construction instead of the "unshored" construction of he
example.
46 and see if a workable solution can be obtained.
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

4-18 Redo Example 4-16 using a W18

an ,HS 20 truck and A-36 or&

y6

P=

&
T
*

[II

30 5 '
i .; :. 111

t>OZ

- 'lo

L\

--

B J . ~ pi~r:,

il

'1

i
/i
iI

250 MPa steel.

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.

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.
27 A

11000 - 14000 rnm

Figme P4-19

4-29 Redo Example 4-18 if @ = 0.90 and FL = 1.5.

4-30 Redo Example 4-18 if Fy = 50 ksl.
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.
6.75 m. Use
Answer: W530 X 92.3.

,~

' ,

',

, ,

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

I
I
Squdre bar
J

Fldt bar o r plate

%;

I
,

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
s, however, specifications will require a minimum amount. of
ss for esthetic and safety reasons.
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,

W shape

0.6 F,

(gross section area)

F, = 0.5FU

(net section area)

(5- 1)

e AASHTO and AREA allowable tension stress is somewhat more conservaF,

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

Figure 5-1 Tenston members. See Fig. 5-2 for cables used as tension members. (a) Structural shapes
used for tens!?? members. (b) Upset bar. ( c ) Threaded bar and use of a turnbuckle to adjust bar
lei&. Applicable for square and round bars.

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
reductions occur along the member.
The effective net area at approximately the root of the thread of thread
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-

I,.

= A*

1%.

'>
., .

4.

5 81

,.. .,...,.

1.33Ab S A , 5 1.5

r 2 0.5" (12.5 rnm)

4
i-

..

1.;

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
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
n~ectsof 7 to 61 or more single wires are wrapped around a cen
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 850F is poured into the wire matrix. After the zinc cools, and the
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.

using r, of element to compute L ' / r between stitch bolts.

d stretching it so that the component parts are fitted together.

The modulus of elasticity of bridge strand and rope may be taken

67 (and larger)

23 000

158

CCO

DESIGN OF TENSION ~ ( E . M B ~

Table 5-2 Selected cable design data

rl

Welght

in

Ib/ft

mm

kN/m

Area
in2

pu

m2 ( X lo-')

bps

kN

project and design uncertainties.

B dge Strand (single strand with multiple wires)

I

13
16
18

0.52
0.82
0.99

0.008
0.012
0.014

19
22

1.18
1.61

0.017
0.023

;
;
f

:
1

10
13
16
119
"22
25

0.15
0.234

0.0968
0.1510

29.0
46.6

129
207

0.284
0.338
0.459

0.1832
0.2181
0.2961

56.2
66.0
89.2

250
294
397

Bndge Rope [6 X 7 (6 strands of 7 wues/strand)]

0.24
0.004
0.065
0.0419
042
0.006
0.119
0.0768
0.65
0.009
0.182
0.1 174
0.95
0.014
0.268
0.1729
1.28
0.019
0.361
0.2329
1.67
:0.024
0.471
0.3039

13.0

58

23.0
36.0
52.0
70.0
91.4

102
160
231
31 1
407

286.0
372.0
576.0

1272
1655
2562

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

15.1
0.220
4.25
2.7419
824.0
21.0
0.306
5.83
3.7613
1110.0
27.0
0.394
7.56
4.8774
1460.0

3665
4938

-3.2 Shear Flow and Maximum Effective Cross-sectional.Areas

3. For all other shapes, including built-up shapes, with at least three fasteners
a line (Fig. 5-5 has five fasteners in line), A, = 0.85A,.
4. Any tension members with only two fasteners in a line, A, = 0.75A,.

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

1;
2
2f

44
50
64

3
3f
4

75
90
1,00

--

5.24
6.85
10.60

0.076
0.100
0.155

1.47
1.92
2.97

0.9484
1.2387
1.9161

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
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. '

6494

. -

. -

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

,-.,,a,

..4

(0)

T ~ p i i . l ilr u , ,

\.ur<

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
welded. (a) Single lacing. (b) Double lacing. (c) Batten plates. (d) Perforated cover plates.

(f)
Figure 5-7 Cross sections of several built-up sections. General cross-section configuration is limited
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
.
( d ) Two channels with lacing. ( e ) Two channels with flanges reversed from (d) and both plate and
lacing. (A Four plates welded to form box section.
$,

'.i

4.

not occur. It is also necessary that the member(s) be constructed so that painting
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
being obtained on those sides by use of lacing bars or by use of perforated cover
plates. Either of these configurations allows maintenance of the interior as well
as the exterior of the member, as shown in Fig. 5.8.

iccrlon

i b ~ l i i t n gIS prescntiy used

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

parts, making the connection sufficiently strong that no relative movement can
take place across the hole. Tension tests using plastic and photoelastic techniques indicate stress concentrations at the edges of the bolt hole which
have a maximum value on the order of two to three times the average s

f, =

5,it is reasonable to assume a uniform stress distribution across

indeterminate, but should the working load be inaccurately estimated so h a t

redistributed until the load is
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" .y;,

a<g
. , .. .

,,

it.

'

t,i

!, .". ,),<..
,;;*,.,i
>\ ,
,

,?,$

"

:.

',

14
P= 1

.03~

Vol. 89, 1922, pp. 847-848) and is almost universally used. The meth
Figure E5-1b

Notes:

';y;;,~'~~,,

, , .v

...,
(

'1 .";Tie rod size could be reduced in the lower-half group of purlins for a rod
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
effect considerably, but the designer will have to decide how much.
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.
\

one hole width for each hole encountered.

For each change in direction from one hole to the next hole, add ba

me possible path as just cited, which

vior along irregular failure lines. The

'

5-6 NET SECTIONS

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
of holes, the designer must determine a failure section that will yield the
minimumiarea. The path across the section producing the minimum area is t
critic&l:ine;tsection.
-4 li$@ig. 5-9b, which is a portion of a repeating bolt or rivet pattern @mi&
be btained in a large plate ass
~ .d t # & ~ r e )the
, critical section is
;
:..
.- .\: ;. ... .......... .

where s= pitch, or longitudinal distance between adjacent holes

g= gage distance between adjacent holes across the-width

"hole" on the diagonal, thus canceling that hole reduction on the net
The effective hole diameter (except for slotted holes) is taken as
AISC specifications
AASHTO and AREA specifications

The hole is always larger than the fastener by at least

$ in (or

1.5

,t,

""

, I ,

for possible metal damage around the hol

iameter as D + in or'D + 3 rnm.

\[-I/

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)

fy,

(a)

Mgure 5-9 Critical net section for tension members with holes. (a) One or more rows of holes w
the net section is obvious. (b) Hole pattern where net section must be found by trial.

=:+$=

+ 3 + 2(2)

I.OOin

t width along path ABDF':

w = 10 - 2(1.W) = 8.00 in

DESIGN OF TEXSION Sf33

Net width along path ABCDFG:

rv, = 13.25 - 4(0.875)
Net width along path ABDEFG:

+ 0.333 + 0.333 =

10.42 in

>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)

Pallow= A,F,
P

5-10 Net section of angle used as tension member with staggered holes in both legs.

A,F,

7.78 in2

13.25(0.75)(0.65)

7.78(0.5Fu)

whichever is smaller.

*
2d"

5-7 DESIGN OF AISC TENSION MEiMBERS

Example 5-4 What is the critical section of the 8 x 6

angle with

s computations to determine the

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

F4W-e E5-4

SOLUTION

'

D = 0.75
Net width along path ABDFG:

+ 0.125 = 0.875 in

,&,w = 8 + 6 - 0.75 - 3(0.875) = 10.625 in

Net width til&gpath ABCDEFG:

= 13.25 -' 4.375

= 10.83 in

+ 0.667 + 0.667 + 0.222 + 0.400

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,
section for all the diagonals, and similarly, top and bottom chords d s s i p e
using constant section sizes. Only the vertical web members will be d e s i g e
in this example. Use F, = 250 MPa and the AISC specifications.

9 3 STRUCTURAL STEEL DESIGN

!'

which
S o ~ u n o NFrom inspection of the computer output (Pa*
shown in Example 2-6), the following values are obtained from the tw
conditions used:
Left side

Right side
Member
4
8
12
16'
20

LC-1

@pq

LC-2 (kpq

0.0
31.18
62.36
93.54
124.72

Member
74
78
82
86
90

0.0
48.62
97.26
145.88
194.50

LC-1 @W
124.73
93.56
62.37
31.18
0.0

157.57
118.18
78.80
39.40
0.0

I n the table above,

Since the signs are (+), all the vertical members have only tension Iorces Ior
either load condition. Note that the computer program automatically
output with wind by the factor 0.75 so that all loads are on the same design
basis. Member 20 has the largest axial tension force with wind from the left;
wind from the right would produce this design value in member 74. Member
20 is the longest vertical of the side shed members and has L = 4.5 m. If One
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

1
1
b*

3i

4.5(1000) = 18.75 mm
240
A preliminaly side computation indicates that the bolt Pattern !&Own
to
~ 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
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

'mm

i:

't

i1

faa
I"
i

m2

must be
Based on F, = 0 . 5 ~ "(use Table 5-1 for F,), the effective net area
194 5
A, =
= 0.9725 x 10-3 m2
200
m i s effective net area A, must be obtained from a gross section of least

0.9725
0.85(0.90)

1.2712 x

2Dt

2(25.0)(0.0063) = 0.315 x 10-3 m2

An,, = A, - Aholes2

0.9725
= 1.081 x
0.90

m'

Here the radius of gyrauon controls the selectlon of he section

than the area requirements. Tentatively use: L 152 x 89 6.3 (as ae
lightest). We must select a section that can be used in hejoint as shownCheck if it is possible to get two 22-mm bolts side by side in the long leg of
the angle. Use Table V-13 and obtain:
'Onsidered

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
~
1.16.4 SSDD) = 28 rnm (or 1; in)

A, = ---194.5 = 1.2967 x lo-'

0.6(250)

A,

Ahole,

> 2.670
0 . ~ .
Actual hole diameter = 22 + 1.5 = 23.5 mm
Edge distance = 152 - 57.2 - 63.j = 3 1.3 > 28
g2 = 63.5

Use an L152

89 x 6.3 section.

O.K.

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-

LC-1 OrN)

LC-2 GN)

L, m

39

- 63.65

47

+70.18

+ 20.39
- 37.60

6.76
8 2 (longest of verticals)

Member

m2 < 1.2967

P,, = 70.18

- (-37.60) = 107 78 L N

' 2&

STRUCTURAL STEEL DESIGN

Member 47 is critical for tension design.

Minimum r
* '

8.2(1000)
= 34.17 mm
240

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
only 70.18 kN, which gives A, = 0.90An.

DESIGN OF TEU'SION LIMBERS

Example 5-7 A portion of a highway bridge truss is shown in Fig. E5-7.

required to select the lightest W12 section for member 9 using A-36
and the AASHTO specifications. Assume 2 X lo6 load cycles for the scrwce
life of the structure.
I,:

SOLUTION
From the computer output, we obtaln the following (includes
impact in live loads):
=

+ 80.3 kips

Maximum live load =

+ 60.2 kips

Dead load
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
table of double angles using Tables V-10 and V-11 of SSDD [A, = smaller
of (A, or 0.85Ag) X 0.901:

1,

mm

6.3
7.9

'

Al
0.353
0.442

m2 Anet, m2 Ac, m2 Section

2.307
3.468

2.03
2.991

2L127 X 8 9 X 6.3
2L127X 127 X7.9

r,

mm ,A
,

37.5
39.9

(X

P,,
P,,,

= 80.3

+ 60.2 = 140.5 kips

= 80.3 - 22.8 = 57.5 kips

The force range (analogous to stress) is

P,, = 140.5 - 57.5

2.660
3.910

"

p:'
,

A,

83 klps

We will assume four holes in the flange at any net section, as shown in the
insert of Fig. E5-7. Use i-in-diameter bolts, so that the effective hole
diameter D = + = 1.0 in.
Aholcs

There is nothing lighter, than these two double angles. Note that minimum
%v
L / r ' controls the design. Now checking stress range and using A,, we have
f = -psr
=

mZ

107.78
= 54.0 MPa
2.03

This value of 63.6 is much smaller than any value of F,, in Table 1-4 up to
2000 X lo3 stress cycles for "base metal" at mechanically fastened joints.
Use two L127 X 89 X 6.3 sections.
///

5-8 DESIGN OF BRIDGE TENSION MEMBERS

The design of bridge tension members is similar to that using the AISC
specifications except that fatigue will have to be considered, as outlined in Secs.
5-4 and 1-9. This will be illustrated by the following example.

= 4(l)tj

High-strength bolts will be used to connect the member to the joint,

producing stress range conditions for "base metal at friction fasteners." The
data given in Sec. 1-12 are based on an allowable stress range when a
catastrophic failure does not immediately occur when the member fails. In
this case it is not likely the truss would collapse if member 9 fails. With this
consideration we obtain the allowable stress range:

F,,

16 ksi

(first column and checking footnote)

We will use lateral and sway bracing across the top and spanning
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.
(AASHTO requires the effective depth of this bracing to be at least 5 ft or
1.8 m.) Joint fabrication requires orienting the X axis in the plane of the
truss for rolled sections. Note that commonly the transverse floor beams are

242

STRUCTURAL STEEL DESIGN

{ t

,*
;"
r

somewhat below the point of intersection of the tmss web me

bottom chord, but this should not affect the unbraced length L
Y.
The AASHTO limitation for L/r = 200 for main tension
rers and
no stress reversals (Sec. 1-7.5), and for a P = 140.5 kips, this would seem
deem ttn
qualify. The net area, assuming that Fa = 0.55F,, is
-- 140e5
=
140e5
Anet = Fa
20

7.025
7.025 in2
in2

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

A >-- 025 - 8.26 in2

- 0.85
By tnal set up the following table, where Ahole.= 4(1.0)$; AWqd=
- - reqa

- Aholcs

* $4

Figure 5-11 Cable geometry ior derelopmg dei~pnequauobs. N@tc

Bat m%rs conf~gurauonthe T at
:'
the o r i p is larger than the T at the other end iif $e cable.

Y d ,

Section
-

+, Ln

Aholcs,

W12x53
W12~50

0.575
0.640

2.30
2.56

ln2

Arcqd,

13.30
12.14

ln2

Arm, m2

15.60 > 8.26

14.70

2.48 O.K.

In general, the cable sag y at any point IS

4h x
y = - ~ ( u- L )
L
Differentiat~ng,we obtain

The W12 x 53 sectlon is the lightest W12 that is satisfactory for both
area and L / r requirements. This section 1s also selected so that the connection can be more easily fabricated when the vert~calmember (No. 7) is
designed in Chap. 6.
Check the stress range for the W12 x 53 section:
Psr
A , = - -PSr
= - =-

A,

83.0
13.30

- 6.24 ksi << 16 ksi

O.K.

tan8

+
=8 h x- 4-Lh + t a n ~
.

(6)

~2

Since

dr = [ I

+ ($)2]1'2

H
and

T.= cos 0 -

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
"x-bie geometry as shown in Fig. 5-1 1 gives

Noting that all the terms under the square root are relatively imi,onif icant except
the first, third, and fourth, we may simplify to obtain

b t :

H z -wL2
8h
xhere = mdspan sag, as shown in Fig. 5-1 1
w = uniform cable loading/unit of length;
be
.
.
.
-- there will always
load caused by the cable weight
L = span length (the cable length is always somewhat longer)

43

-J-

When the ends of the cable are at greatly dlffenng elevations, Eq. ( 5 - 7 ) rather
than Eq. (5-8) should be used, because the tension at the upper end of h e cable
will be considerably different (it is "caqing" the weight of the cable + any
additional cable loading). For horizontal cables T has the same value at b o b

STRUCTURAL

.,

DESIGN OF TMSION bEEJ

STEEL DESIGN

ends and is directly computed as

L
. The
. , , .. cable
..,..... ,.. ... length

T=~

[ + l16($)]

1/2

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 pretension 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
In round
to
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
of a vibrating system and if reliably introduced, eliminates resonance amplitudes
of vibration. The natural frequency of a cable system (the same as any other
type of natural frequency computation) is

i).

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
change in T produces a new fn

A convenient way to dampen a cable to control vibration is to attach it to a

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).

C.ible JrTJnpelllcnI m.1) 'cc r ~ d i . i l i n g f r o m

rov.<r or in o n e d ! r < i r ~ o n

Figure 5-12 Several conflguratlons of cables In building construction (u) Cabies spaced across span
for a rectangular buildlng plan ( 6 ) S~nglecable using a central tenalon nng/;md for clrculr b u i g
(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.

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

U=2(D + L + WorE)
U = 2.0 X erection loads

3, the anchorage resistance would be 3 x 154.7 = 464

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

S ~ L U T For
~ ~ the
N 100-m level, using Eq. (5-6), we obtain
-wL2
=

8h

8(110)
-w ---h
lOd

llOkN

Roofing dead load, including weight of cables

(horizontal projection)
Live load at 40 psf:

40 x 4

- 0.088

Also, the cable tension is (Eq. (5-7):

(:r

T = H I + 16 ,

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

+tan20-8

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

19 = 45" +tan O = 1.0:

Trial

h = w/0.088

T"

1
2

0 (assumed)
0.272 m

155.6
154.7

389
386

Taking 1/2 the roof load as w,

0.024

Close enough to initial

DESIGN OF TENSION hD31BEXS

.M8 S - I U C W R A L STEEL DESIGN

The bottom cable will carry the uniform load, producing the 23.3-ft sag
of the upper cable plus the total roof load:

249

e used. Once this is done and the ultimate axial tension load P, is obtained,
k

1 1 Q V V

p,,
The required cable areas using F = 3.0:
Top cable: P,,= 3(74.8) = 224.4 kips -+ use I f-in-diameter bridge
strand
From Table 5-2, w = 4.73 Ib/ft.
Bottom cable: P, = 3(224.4) = 673.2 -+ use 2f-in-diameter bridge
strand
From Table 5-2, w = 12.8 Ib/ft.
The struts wlll carry a compression load based on q producing an h
value in the upper cable of 23.3 ft. This gives an equivalent uniform strut
"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:

A,+Fy

(+ = 0.68, Table 3-1)

Example 5-10 Given a roof truss member with a length of 11.5 ft in a

building frame with a dead load of 18.52 kips and a live load of 22.54 kips
(snow load). Use the LRFD method, A-36 steel, and i-in-diameter A-325
high-strength bolts. Design the member using the lightest C shape possible.
SOLUTION
We will use a bolt pattern as shown In F-lg. Ej-10 so that the
AISC shear lag reduction factor will be 0.85 (at leasr three fasteners in the
line of stress).

= 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(;

For n = 1, 2, and 3, we obtain f , = 0.96, 1.92, and 2.88 Hz.

For the bottom cable, only the bottom cable weight will be used, since
only the struts make continuity with the top cable:

+ ;)lw

From Sec. 3-7, obtain ( 9 factors from Table 3-1)

Pu = l . l ( l . 1 0 4- 1.55')
=

1.1[1.1(18.52) t 1.5(22.54)] = 59.6 hps

P U
59.6 - 1.88 in'
A >-..-=--.----- 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
different, no resonance 1s likely to occur. When one cable is at resonance,
the other is at a different frequency, which acts to dampen the resonance
vibrations so that the total vibration amplitude is kept small.
///

A >

"

59.6 - = ----p u

- dF

1.39 in?

0.74(58)

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
the dead and live loads must be identified so that the appropriate load factor

Try a C7
* x 9.80:
A, = 2.87 in2

3'

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
Answer: 62 mm.

*"5-10.
g, = 2.50 in
ih

4,:+
,

rl

eel is used and the plate

2g, = 2(2.50) = 5.00 in

This leaves a bolt hole spacing of

7.00 - 5.00 = 2.00

,':,
,

A > ,

and

< 2.67(7/8)

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
thickness is 12 mm?
Answer: 423 kN.
5-9 What is the allowable tensile load for the plate shoan in Fig. P5-9 using F, = 345
20-mm-diameter bolts, and the AISC specifications?

N.G.

Try a C8 x 1 1.50:
A = 3.38 in2

g , = 2.50 in

r, = 0.625 in > 0.575

O.K.

t, = 0.220 in

This leaves a center-to-center bolt spacing of

8-5=3in>3D
Check the net area.: with two holes out:
A, = 3.38 - 2

Check the shear lag:

O.K.

(f+ ;)

- 0.22

A, = 2.940(0.85) = 2.499
Joint efficiency and shear lag:
X

2.940 in2

> 1.88 in2

A , = 3.38(0.85)(0.85) = 2.44

Use a C8

>

1.88

Flgum PS-9

O.K.

5-10 What pitch is necessary in Prob. 5-9 so that only three bolt holes are deducted from
to produce the net section?
Answer: 32 mm.
5-11 Select the lightest single angle for a tension load of 50 kips. The length is 6 ft and I 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 .

O.K.

11.50 member.

PROBLEMS
.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
MPa and the AISC specifications. Use t 2 15 mm and w in multiples of 3 mm.
5-3 What is the net width of the plate shown in Fig. P5-3 using the given fps units?
Answer: 6.75 in.

v, +
v,
rl

75

3 0
40

50

2"

1,s"

= 250

Figure P5-1 1
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.
Answer: L102 X 76 X 6.3.

3"

40

F,

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..'
Answer: P = 154 kips.

5-15 Select the lightest single angle for a tension load of 68 kips, assuming one
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!

..,. .

,.,

.,,..,.

..

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.
5-24 Do Prob. 5-22 using the AREA specifications. Do not consider fatigue.
Answer: 747.4 kN.

Live load = 45 psf (horizontal project)

Use A-36 steel and the AISC specifications.
Answer: Diameter = 1$ in.

Answer: L3;

Figure P5.20

,,

7/16.

5-26 Redo Example 5-10 if the truss member length is 14.5 f t instead of 11.5 ft.
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 highstrength bolts, and A-36 steel. Panel length = 25 ft, as shown in Fig. Ej-7.

45 lb/ft2

..&&+ha.%

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.

construction. Joints are field-fabricated using high-strength bolts.

Material

Limiting L/r (approximately)

Steel
Aluminum
Wood

60
30
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.
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

<I I

-------------1_____

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

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 mathematician. The formula is readily derived (refer to Fig. 6-1) as follows. Writing the
fundamental bending equation for moment,
,. , . .

x = A

sin hy

ection of the equation, so the solution must b

sin ky = 0. This is only possible for values of k L as follows:
kL

= n,

Zn, 3n, . . . , n7;

..,,..,...

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

dZx
EI-=
(iy2

-Px

Using Eqs. (e) and (c) to obtain P, we obtain

'

&

Ql, b
I

This equation for the critical column buckling load P is generally called
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 ~

.
I
1

Fcr= (L/rI2
we obtain from Eq. (6-2),
. I

4n2E
Fc, = -- .
( ~ / 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 load
g (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
bending [Eq. (a)], however, is stress being proportional to strain. Thus the uvv
limit of validity is the proportional limit, which is often taken as Fcr-+ Fy

&,

Taking the derivative dx/& for the slope, we obtain

&
Since dx/dy

5-3 COLUMNS WITH END CONDITIONS

Rotation of the ends of columns in buildings is restrained by the beams that
frame into them. The ends of truss members are similarly restrained. In both
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 dpnvgtion 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
the fixed-end column shown in Fig. 6-2a. The differential equation for bending
now becomes
J

cos hy - k sin ky
P

sin k L

0 at y = 0, we have
k M o 1 - cos k L
P
sin k L

The smallest root of this equatlon glves k L

have

2 7 and with k

(P/EZ)'/', we

or the effective column length KL is L / 2 = 0.5 L and K = 0.5. We have now

introduced one of the first attempts to adjust the column Iength for end
conditions. The effective length K L or some equivalent is used in nearly aiI
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
1

Mo
+ B cos ky - P

and in a similar manner obtain

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 .
sin ky
P
sin k L

x=-(

+ cos I+ - 1

or K = 0.7. Several of the more common cases of column end f i t y are s h o ~ i l

in Fig. 6-3.

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

F,,

T ~ E
=

(KLI~)~

differentiating obtain
Recotnineniled

d(F,,)

d(KL/r)
Rotation fixed

conditions

Rotation free

.,,.

ce the maximum value of the Euler stress or any critical buckling stress is
.to
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):

5,

Translation free

Flgm 6-3 Theoretical and design values of K for columns with end conditions shown.
..

(a>

(KL/~))

Translation fixed
Translation fixed

-277'E

.,...,...

Strictly, according to the derivation of these equations for critical buckling

load, the tangent modulus of elasticity E, should replace the elastic modulus in
tangent modulus concept was introduced in Sec. 3-7.

F,,

CY- "j7)

KL

(6)

uation can be differentiated to obtain

d(Fcr)

-,+)

KL

P-'

d(KLlt.1
ake the slope d ( F , , ) / d ( K L / r ) = 0 at K L / r
will arbitrarily define a parameter

(c>
=

0.Also, at some point

KL

-=

6-4 ALLOWABLE STRESSES IN STEEL COLUMN

The column stress obtained with the Euler equation of

r
Cc
lopes of Eqs. ( a ) and (c) will be equal (i.e., the two curves defined by these
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=-

account for eccentricity, residual stresses, and the several other factors that
complicate the theoi'y. It would also be appropriate to use a variable safety

T*E

(dl

c,*
rom rearranging Eq. (b), inserting K L / r = C,,p

2, and using the Euler

.iAnwLIUn,u

.d

bltjEL DESIGN

AXIALLY LOADED C0LU;CNS

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
If

rLW STRUTS 2.63

value of m, p, and Cc are placed in Eq. (b), we obtain

the critical buckling stress becomes 0 . 5 q at K L / r = Cc under our

acsumptions. In general, we have the buckling stress from Eq. (b):
711:!;

\k

Equations (6-5) to (6-7) are somewhat awkward for routine computations

(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.

6-4.2 AASHTO Axially Loaded Colunln Design Stresses

sv

the allowable stress is obtained for K L / r 5 Cc from Eq. (

j
)
:

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
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
for all values of K L / r I Cc. For K L / r > Cc, use a constant value of safety
factor based on using K L / r = Cc in the equation above to give

When K L / r > Cc, the Euler equation with SF = 23/12 is used to obtain
:he allowable column stress as

> Cc:

4'

F',=

n * ~

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.
*

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:

..

fps:

Fa =

149 000

fps, ksi

ksi

(KL,'~)~
b

F',

1.03

lo6

Mpa
(KL/~)~
Eqx~tlons(6-5) and (6-7) are used for main compression members. Secondary

f 6- 10)

SF(KL/~)'

SI, MPa

c,

264

STRUCTURAL STEEL DESIGN

643 AREA Axially Loaded Column Des;ign Stress

6-45 Column Design

~h~ AREA allowable column design stress formulas are somewhat similar to the

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. niusual 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-

Allowable stress

Limitation

-,
-

3. Make a tentative section selection (sometimes use can be made of bb!rs such

I>.

1U7 < % _ < =

-

': fi

F , = O . ~ F , - (1k.2)

#,A,

.7

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

(6- 12)

ssnn

which

O~VPP

fnr r ~ ~ r n r oVT
l

--+:--\

when Po = P,,,,,.
In the AREA specifications,
F, = ksi
k = 0.75 for riveted, bolted, or welded compression

Two additional factors should be given consideration:

.*

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 adequate 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.

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 correcponding 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-

< C,

ularly when KL/r

KL/r.

60 lhlit! I I

+ 30 iblft'

rooi X 10 = I 8 k 1 0 (1

since both stress and safety factor now depend on

. ..
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 departstore 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).

':.merit

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)
Y'

136.6

> Cc

[use Eq. (6-7)]

,?&

1.23
LI.Yb

Pa = 5.26(7.99) = 42.0

> 36 kips

Fa = -- 11.5 k s ~(vs. 11.54 in Table 11-5)

1.915
O.K.
Pa = 1 lS(3.17) = 36.45 > 36 kips

Fa = 149 OoO = 7.99 ksi

(136.6)~

O.K.

Use a W8 X 18 section (tentatively); the Table 11-4 value of 41 b p s is due

to computer roundoff.
Let us also investigate a pipe section (see Table 1-14). Now that we have
some "experien~e" from just checking a W shape, let us investigate a
4-in-diameter pipe:

Use steel pipe 4 in x 10.79 lb/ft.

Note that pipe and any other sections where r , / r ,
most economical shapes to use when KL, = KA,.

KL,

KL, =
From Eq. (6'-5),

1 are geneidly <n.e

///

Example 6-2 Design a W section for the conditions shown in Fig. E6-2. Use
F = 50 ksi steel and the AISC specifications.

1
I

16 f t
8 ft

11-4 to obtain an inltial estimate of column r:.:r.

able 11-4 is dased on Fy = 36 ksi: use a ratio as
36
PCquiv= (750) 30 = 540 kips

Since

l ji
I

AXIALLY LOADED COLUILCiS .LSD STRLTS

%3

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.
LC-1 (kN)

Member

LC-2 (kN)

L, m

The section selected for tension considsrarlons was

2L127 x 89 x 6.3: cl = 2.66 x

SOLUTION
Take K

&

rn'

r,,,

= 37.5 mm

1. For member 39:

Figure E6-2

--

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

1.6 in the column
sizes approaching P = 540 kips; therefore, look for L = 1.6 X 8 = 13 ft
column length and 540 kips.
Try W14 X 99:

For member 47:

N.G.
Revise the section for minimum r:

Compute
HI

r, = 1.66ry = 1.66(3.71)
-=-=
KL

rx

16(12)
6.16

3 1.2

6.16 in

controls since it is largest

Note that a 12-mm gusset plate is used between angles at the connection,
Possible angles are:
2L152 x 152 x 7.9: r,,, = 48.0
(required for tension)
A = 4.71 > 0.5504
2 L152 x 102 x 6.3:

r,,,

C, = 107

(Table 6- 1)

41.5 mm

3 15 >0.5504

We will check the pair of angles 152 z: 102 x 6.3. slncs they are the !Ijl:est.
Member 47:
8'2(1W0) = 197.6 -+ F, = 26.3 \(Pa
41.5
Member 39:
-KL
=

F,=--47'9 - 26.9 ksi

..

..

(= 27.0 in Table 11-6)

1.78
Po = 26.9(29.10) = 782.8 > 750 kips
O.K.

P,, = 38.8(3.15) = 122.2 kN > 63.65

P4,=26.3(3.15)=82.8kN>37.6

(Table VI-5)

0.K
O.K.

,
XYIALLY LOADED

STRUCTURAL STEEL DESIGN

COLmfiS .<4ns n m s nl

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

Sectlon

A ~ , , ~in2
.

ry, I D

Fa. k s ~

W12X53
W12x58

15.60
17.00

2.48
2.5 1

9.35

P,,,,

=A

F*,Iups

N G , since ry too small

159.0> 99.3 O.K.

The r, value for the W12 x 58 IS 5.28 > 2.5, so the section is satisfactory. 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.
According to the AISC specifications, this would require a radius of a ~ a tion r of at least

Maximum column force = 56.75 42.56 = 99.3 kips (compression)

Minimum column force = 56.75 - 16.11 = 40.6 kips (compression)
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 deadand live-load stresses. The minimum r is

Try a W12 section, since we have used a W12 x 53 in Example 5-7 for,
amember 9, w p h frames into the joint on one end of this member. A gusset$
plate can covkr both members with a minimum of filler material if all the
web members have the same nominal depth.
We further note that Eq. (6-9) always determines the column stress for
members using A-36 steel since K L / r I 120 and C, = 126.1 > 120.

Commonly, plpe struts are used with diameters rdnging from 4 to 6 ic. H e i ~
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
A

For KL/r

46.6(12)/2.88

F,
P
'

12.8 in'

194.2,
(Table 11-5)

3.95 ksi

AF, = 12.8(3.95) = 50.6>>>2.3kips

Use 8-in-diameter extra-strong pipe. I t m a be advantageous to cse

272 STRUCTURAL STEEL DESIGN

,,

smaller section in the outer one-fourth of the span, where the value
L = 2h is less than 46.6 ft.

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
the sdqdard rolled W shapes, where the ratio of r / r y is often 1.5 to 5 or more,
unless bracing is provided with respect to the Y axis.
Table 6-2 Approximate radius of gyration for several built-UDsham

..

-up sections are very commonly'uskd for bridge trusses and..in co.
r towers. Antennas are essentially built-up colu mns, although not
onsidered ,as sudh. In any case, where compre:jsion (and tension)
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.
.
.
.
. -.. .
._
3. .The resulting computed values of I,, I.,, r,, and ry and K L / r,,,, ^whicil
produce the allowable compression stress.
4. Producing an acceptable section area based on the allowcable stress from s:tp
3, which is not known until the area has already been established.
i

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.

r..

rb7

= same as

r , = 0.4211

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,
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

t-7

---

----]l---I

I.oi2,L-s

/ r

7 7

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

""

"A:-..

,--A

".-A

-1

""

,-,

---I..

-..A

-1

....

,I.,

-r~.,.L

!iri:lr:d

;;,:,n!ctr)
5-

"-2

?
'

.L--.

.~:1:>
:?j
/.\

"-

.I
of using lacing, single and double batten, and perforated
plates (called, collectively, cover plates) are usually used. Where the steel is located inside a building,
the cover plates may be solid and their use could reduce fabrication costs. In
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
without the careful fabrication required to completely seal the interior. Presently, it appears that the economics of fabrication favors perforated cover plates
to lacing, since automatic gas-cutting methods allow rapid cutting of the plate
openings in a length of cover plate.
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
batten plate when the compression member "buckles," as shown in Fig. 6-5.

If we assume equal end moments, as shown in Fig. 6-5c, and use the
differential equation for bending as used to develop the Euler column equation,
and allow for boundary conditions, we obtain

The derivative at y

0 is
d.~

kL
2

- = ke tan -

Now referring to Fig. 6-5h, we obtain for the $hear in the lacing,

where k

P / E I as in the Euler equation. AISC simply takes

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

fps

(u)

(b)

(c)

Figure 6-5 Shear development for a laced (or battened) compression member. (a) Laced column. ( b )
Identification of laced column shear V. For lacing on both through faces, divide V equally on both
lacing bars; for four-side lacing obtain 90" as above. (c) Basic concept of shear in lacing of built-up
section.

SI

ARW

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
batten plaie should be so designed.
The spacing of lacing and batten plates must be such that the L / r of the
main elements between fasteners is not greater than K L / r of the entire member;
otherwise, local buckling might develop, particularly where angles are used, m d
L / r , between fastener points may be critical. Ths XXSH-TO and =IIRE,I

End

AISC
C':

L,,
=

< (',

.:I-

[?j.

i..

\/F'J i i p s i

C',

2 if,50

5.;: Sl

.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.
3, Dec~deon lacing, batten, or perforated cover plates (Fig. 5-7) or if sect~onis
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
,

#.

IS

I,

=o(J) + Z A , ~ ? , )

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

compressive stress
based on the largest KL/r.

7. Check P
= AF, > P
,,,
and iterate as necessarq.
8. Design lacing, perforated cover plates, batten plates. and/or tie (or SIZ:~)
plates.
This procedure will be illustrated by the following ex~mples.

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:

Assuming that F,
mately

Total design load = 240.77 94.13

= 334.9 kips
14 ksi the area required in the section will be approxi-

--

I,,

Both

t/ and

t,

+ 211,d2 + i l p d 2

2(162) 2(8.82)(1.66)'
= 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
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
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
the W12's have depths greater than 12 in.
C12 X 30 data:

2I,,

+ 6.375(4.59)'

= 507.0 in4
=

4.59 in

About the Y axis: Iyy = 2 1,

controls (after computing r, below)

+ 211,d2 + I,

> 0.23 in (Sec. 1-7.7)

334'9 22.5
=14.9 -

< 24.02 in2 furnished

O.K.

280

STRUCTURAL STEEL DESIGN

LSD STXbTS

.%XIALLY LOADED COLL-NS

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
welding to attach it to the rolled sections. The only bolting will be the field
conn~ecti-onsof the joints.
The distance between flange holes (and the approximate center-tocenter weld distance) is shown in Fig. E6-6c.

L' = 12 - 2(1.75) = 8.5

<

15 in

Try t

-&

3l

in (0.313 in also rmnirnum AASHTO):

L =
r

0.288(0.3 13)

109

O.K.

Compute the lacing bar force. The bar force component psrpzndicu!~
to the member axis is computed using Eq. (6-14):

(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

pd

8.3
---- 9.6 kips

cos 30'
36[1 - 0.5(109/126.1)']
Fa =
2.12
btFa = 9.6 kips
d/"

10.64 k s ~
q

Figure E6-6~

Use single k i n g bars at an angle of 60' to the member axis, as shown in

the figure (refer also to Fig. 6-5a):

L=-'

5 - 9.81 in
cos8 300

L' = 2(8.5)sin 30' = 8.5 in

L'

[ -1"'

r = ,$if)
t=

use 1 :-in plate

The final lacing bar dlmenslons wlll be taken as ( L

9.81 i- l.b9 in)

- 8.5 - 11.1 << 40

---

0.763
Also, 11.1 is less than 0.67 X L / r ( = 41.9). Limit the L / r of the lacing to
130. The radius of gyration of a flat bar is
ry(of channel)

Let us revise t to 7/ 16 in; Fa = 13.78 ksi:

9.6
= 1.58 in
b =
0.44(13.78)

Design the end tie plates (AASHTO calls these stay plates). The
AASHTO requirements are:

= 0.2881

9.8 1
= 0.262 in
130(0.288)

t > - -L= - - 981 - 0.245 in

40
40

Use a minimum of three fasteners (or equivalent weld) each side. Transllited
into design (Fig. E6-6d):

t = - l2
=
50

0.24

use $-in, to match lacing bars

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~~nbie

stress 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
m'.
approximately 1350/100 = 13.5 x
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

r
Pal,,,

and from Tab,? Y 1-5 obtzin F ,= 8 1 \fPr

. 224

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

SOLUTION
The minimum K L / r will be taken as 200 for main members.
Therefore,

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

w = 2r = 150.mm a
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

AXIALLY LOADED COLUMNS AUD STRUTS

STRUCTURAL STEEL DESIGN

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.
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,
present practice tends to the use of a base plate that is shopwelde. to the
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 construction. The overs~zedholes allow some misall~amentof thz pnchor bolts
without redrilllng the base plate holes or taking out and resetting the anchors.
The general method of fastening the column to the foundation IS illustrated in
Fig. 6-7.
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 dimsnsions must be adjusted so that

Figure S 7 c

"r

Lb = 200 mm the length of 23.8 m gives

48 holes
)I

48(0.30)

= 14.4

47 spaces = 47(0.20)= 9.4 m

Total
=23.8 m

Find the cover plate thickness (refer to Fig. 6-6):

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

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

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

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

Allowdble AISC stre.s<s

Conract marerill

287

F-P
ips: kipsiin2

S t . >L?:l

7.8

04
ii.25

Sandstone & l~rne.;tone

Brick In cement InJtriu
Full dred of conire[<
ioiindarion

1.75

0.357;

L o s t h ~ nitill Jrru on
iound~rion

dr

0.j5fc'

5 0.71;

lq.r
n oi rn

,Lf =

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

m
q(m)Z

or

1 unit wide X

r thick, we have
?

!\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.
anchor bolts being used to attach column to footing. Diagonal member is a
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
Iirectly. Note again the use of shop-welded base plates.

=
where

6M 'I2

3 X q X ( r n 2 0 r n 2 ) 'I2
=[

Fb

q = actual contact pressure

Fb= 0.755 (AISC Sec. 1-5.1.4.3, based on bending on r e c t a r i p i ~ r

section)

-STRUCTURAL STEEL DESIGN

M A L L Y LOADED C O L L J S S AND STXLTS

Example 6-8 Design the base plate for a column as shown in Fig. E6-8. Use
= 250 MPa, f,' = 20.7 MPa, and the AISC specifications.

F,

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
The pedestal floor line dimensions will be the same as the bas
plate. Thus
F, = 0.35f,' = 7.245 MPa (Fig. 6-8b)

Let us make m n:
From Table V-3 obtain d = 360 mrn; bf = 256 mm

(0.205

m = 0.106 m
B = 205

C = 342

+ 2(106) = 417mm

+ 2(106) = 554mm

B X C = 0.417 X 0.554
The actual contact pressure q is
q=--

231.0

After several side computations, let us assume that C

+ 2m)(0.342 + 2m) = 0.2305 shifting decimal

m2 + 0.2735m = 0.04009

Check:

0.2310 m2 > 0.2305

O.K.

- 7.229 MPa

..

= 50.9 mm

Use a column base plate 417

554

52 mm thick.

SOLUTION
We will design the column-to-base plate weld in ExampIe 9-7.
F, = 7.245 MPa 5 q

say 52 mm

750 rnm:

= 656 mm
Figure E6-9b illustrates data so far including q = 7.245 and tained from using B and C in the preceding equatlon for q.
t
Along line x - x :
q = 7.248 - 10 22x

ob-

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.

6-7 LATERAL BRACING OF COLUrvfNS

70.6 mm
say 75 mm
.*-om the other direction (line y-y) at point A :
=

q = 7.248

- 10.22(0.219) = 5.01 MPa

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
P, = 0 . 0 2 P

(average for a unit width)

L-,e the largest thickness of 75 mm. The final column base plate dimensions
nre 780 x 656 x 75 mm thick.
For anchor bolts as shown in Fig. E6-9c, assume that the bolts will
carry the full moment even though the axial force will reduce the effect of
the moment considerably. This assumption provides some reserve capacity
of the anchoring system to resist a considerable lateral force (colu.mn
shifting laterally).
I

1 :se

F,

= 345 for anchor bolts.

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.
A series of tests at Cornell University by Winter (see "Lateral Bracing of
Columns and Beams," Transacfions, A S C E . Vol. 125. 1960) inclicates that very
little lateral bracing is required to allow the compression element to develop Pie
allowable design stress. This restraint could generally be developed by t5s
weight of the floor system onto beams when full-length contact with the
compression flange is made. Because of the variable nature of flooring (metal
deckfine. concrete-to-beam. wood-to-beam. and so on). i t is suooested
.,- that the 2
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
" Z

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,.
F,

Areqd=----

l5

3.73

-j,02in'

rru

a i n u b i U K A I . blEEL DESIGN

AXIALLY

Simply search Tables 1-6 and 1-7 for this combination of A and r, and find
MClO

28.5: A = 8.37 in2

ry = 1.17in

Using Example 6-1 as a guide, try a 4 in

A = 3.17 in2
r = 1.51 in

LO.ADED COLUhNS .WD STRUTS

10.79 Ib/ft pipe column:

If bending or other requirements are also satisfied, this section can be used
for the girj (or spandrel).
!

'

///

6-8 COLUMN AND STRUT DESIGN USING LRFD

The use of LRFDrequires the separation of dead and live loads on the column!?
+t*.
Once thls is done, obtain the ultimate column load in the following form:
'4
.

P,,=$(FdD

+ FLL. . . )

Areqd

P" 54'12
= 3.79 in' > 3.17
OFc, - 0.65(21.94)

N.G.

Try a 5 in x 14.62 Ib/ft pipe:

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
note that the value of F,, depends on the value of 11 as follows:

Fc, = 36[1 - 0.25(1.002)']

26.96 ksi

+ = 0.65
54'12
= 3.09 in2 < 4.30
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:

Example 6-11 Given the columns of Example 6-1 spaced on 20 X 20

centers, dead load = 30 psf, live load = 60 psf (snow), column length = 14
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.

SOLUTION
P,, = l . l ( l . 1 0

+ 1.5S)A

= 1.1 [ l.l(O.030)

+ 1.5(0.060)] (20 X 20) = 54.12 kips (vs. 36 kips P,)

O.K.

///

PROBLEMS
6 -1 Determine the allowable load that can be carried by a W14
steel and the AISC specifications if:
(a) KL = 16 ft.
(b) KL = 42 ft.
(c) KL, = 68 ft and KL, = 4 4 ft.

21 1 column using

I;, = 50 hi

Make appropriate ,somments.

Answer: (6) 602.7 kips. (c) 549.1 kips.

6 -2 Determine the allowable load for a W360

AISC specifications if:
(a) KL = 5.1 m.
( b ) K L = 13.2 m.
(c) KL, = 20.5 m and K L , = 13.6 m.

3 4 . 7 rolled ssc~lonusing

F, = 415 Sip2

md

Make appropriate comments.

Answer: (c) 2669 kN.

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.

6 -12 Design the lacing for the allowable load found in Prob. 6-10, allowing a 20 percent i n w e m

('310 X 44 6

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.
6 -14 Redo Example 6-1 if the contributory column arc3 u 30 X 20 instead ol 20 X 20 but dl the

5.380 X 74 4

other data are the same.

Answer. 5 In at 14.6 Ib/ft.

Figure P67

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

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?

"

4-L's 6" X 6" X

I - p l ~ t eI 2"

$"

x L"

"

.,i:;"
;: '>.

li

"

Figure P6-8

i -9 Referring to Fig. P6-7, place a second channel C310 X 44.64 on the bottom of the S shape to

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.
5 -10 What is the allowable column load for the built-up section shown in Fig. P6-10 if the member
is of A-36 steel and the length is 14.5 ft? Use the AASHTO specifications.

>q

- 422 kN (wthout impact). Use F, = 250 hfPa and the AriSHTO spsdicatioos.
live load
Answer: Try two C380 X 50.4, 300 x 15 mm cover plate.
6 -16 Redesign the truss end post of Example 6-6 using a perforated cover plate for both sides of the
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..

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
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
column directly to the footmg, wh~chhas a total depth of 21 In.
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
AISC specifications. The column is interfaced to a concrete pedestal.
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,

l6"-4

4
Figure P6-10

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.

the maximum live load, including impact = 109.4 kips, and the minimum Live load = 0.0 kip.
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.

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,
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.

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 )

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.

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
8 th'e &oi@length ~ivlth,
respect to the
joint bending can be minimized by kk'ipin~
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!$:, Ion, member with smal!
moment of inertia), the moment g;ad?@@,,isih&$
. .. .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 computational 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:
Q

:'

"'

or

Number of equations: 100 X 2

200

20d

40,000 words

CPU requirements:

n g ~ d :100 x 3

3 0

300'

90,020 words

Since each number (or word) requires approximately 4 bytes the requirements
become:
:b

Pinned:

40,000 x 4

160.000 = 160K bytes

90,000 x 4 = 360,000 = 360K bytes

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.

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-offreedom 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
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
-f,+ - < fb
I
F,
Fb will be safe. This safety is partially obtained by neglecting the effects of P - A on
fb, which could be computed (to be strictly correct) as
Mc
c
fb = +-+
P,AI
I
When Pb is at midspan, we have (with the limitation A 2 0, see Fig. 7-2)
pbL3
P,~L'
A=+--(a>
48 EI
8 EI
and the resulting bending stress is
(b)

The actual bending stress can only be obtained by iteration of Eq. (a) until the A
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
hand computations, in only a few design situations (where the number of
members is limited ) is this approach economically justified. Neglecting the P-A
effect in Eq. (b) is an error, but on the conservative side. Inspection of Eq. (a)
indicates that the tension stress reduces the deflection and also reduces the
compression stress due to bending. With the allowable tension stress F, a
constant value and recalling that Fb may depend on the unbraced length of the
compression flange (and possibly reducing the allowable compressive stress), we
see that use of
-fb+ - - f<, I
Fb
-

FI

provides a satisfactory solution.

Pb

i=ll801n
\ r = 21

Figure EY-1

L)O~rl
#

Example 7-1 Glven the portlon of* b#hway bndge truss w i t h 102:s
an2
members as in ~ l g E7-1,
.
what is t h i ' ~ & u m tension stress in the bwer
*?"
,
chord?
.SOLUTION
(neglect the P-A effect)
0.0.10(25~)(
12)
8(5l 9)

WL?

jb=-=

8sr
P

J = -AF

0 72 ksl

165
fb -- - +101782-k s 1

f, = 13.98 + 0.72 = 14.70 k s ~

(mau~rnum)

f, = 13.98 - 0.72 = 13.26 ksl

(m~n~mum)

///

When a compressive axla1 load acts together bb~th a bending morncrt. tks
deflection is ampl~fiedand the compress~vestresb incre~seb S~ncrthe aJ!ow\.ab!e
compression stresses take Into account poss~ble buckl~ng(l~teraldefIecuons),
member design IS more senslt~veto t h ~ sloaciTnc rp&e than to one producing
tension stresses.
Referring to Flg. 7-26, we note that rhe ax131 load (absumed to be applrcd
last as belng easler to vlsuallze) Increases the deflect~on.The order of Iorid
application does not affect the outcome, honever. as long 3s jrelding is not
produced. The Important concern IS that there IS an Increase In the detlection
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
A

P ~ L P ~A L '
--- '+
48 E l
8E l

and the resulting bendlng stresses w ~ t hP, =

(-)

(c)

are

(4
Figure 7-2 P-A effects on tension and compression members.

A critical evaluat~onof Eq. (c) indicates that an lterat~vesolution IS required as

for the tenslon mode, but that the deflection "feeds" upon ltszll (dcGtctlcrr

- 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
I-l..,kling failure can occur.
The P-A effect can also develop in tall buildings, as qualitatively shown in
Fig. 7-3, when wind or earthquake forces or unsymmetrical loading produces
lateral displacements of the upper floors with respect to the lower building
eiencnts. A computer analysis can be made to analyze the P-A effect but
requires iteration. The steps include:
!. 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

Mi = ~ I + l ( X I +l Xi)
arid a shear as indicated in Fig. 7-36 using

which is applied at the top and bottom of the story with due regard to signs.

Pi + = axial force in column between i and i

e. .

+ 1 floor levels

4. Compute new X, and compare to previous values used in step 3. Iterate until
satisfactory convergence is obtained, such as, say, 0.02 ft or 0.006 m (app;oximately
in).

r?ik P<A effects are also called "secondary effects" and have been largely
ignored until more recently. Some designers have arbitrarily increased the design

A. Another
stresses (from loads) by a factor such as 10 percent to a110
eloped from
fwtor that tends to mitigate the P-h effect IS that it is;u'&
stresses that
wind or earthquake analysis where the designer can use a
are increased by one-third. The P-A effect would only exce& this in rare cases.
The analysis computer program in the Appendix h
a
gk
! eeily modified to
automatically scan the deflection matns for the appropriate X values, recompute
the P matrix using these values, then storlng them for comparison with the new
values from the current cycle until convergence and exit.

7-3 EFFECTIVE LENGTHS OF COLUiLlNS I N BUILDfNG

FRAMES
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)
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.
We shall find that K in multistory buildings, which can translate. m ~ be
y
considerably more than 2.0, as illustrated in Fig. 7-40. For :he portion of the
elastic frame shown in Fig. 7-46 and using Eq. (6-1):
x = A sin hy

(6-1)

With k = ( P / E I ) ' / ~ and using the effective .+ength L'

x = A

sin-

K L , we obtain

';Y

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:

.' = ,I,

v, = y ,

?'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

(a) and (6):

Flg-le 7-3 P-A effect for tall bulldings. ( a ) Structure with lateral loads. ( b ) ith story with deflections
gt j exaggerated.

8,

=-

"TI

IS

d d t / l $ , and we obtain. from Eqs.

cos -cos - + sin -s!n KL

K
KL
A'
11

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),
A G sin-T I
K a
7 KL

8,

- - -G

(:)iLc (

"Yl
sm-cos-

7i

KL

1
. 77
cos-s~nKL
K

At the column-beam intersection of rigid frames, the rotztion of the c o l u m

equals the rotation of the beam, so equating the two values of 0,. we obtain
nG,

TY~

..

K 6
KL
Similarly equating Eqs. ( h ) and (d), we obtain
Lu"

- ?I ( t a n 2
K 6
KL

tan

K
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 '

where the summation ( c ) is taken because load and moment are coming from
both directions. From the earlier derivation of the Euler equation and summing

"YI
+ tan -tan KL
K
77

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

h,

A plot of Ga and Gb vs. K can be made as shown in Fig. 7-5a. is nomogiaph

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
G,

~b

($1'

Gb

tan s/ K

) + -_v2 tan
K I
T/

(7-3)

% STRUCTURAL STEEL DESIGN

(i0

1000-50 0 30.0:

I1

0:

'5.8
8'0 -

K
70.0
1 00
5.0 -4 o--

3 0 --

1.0 6.0 -

so-

2.0 --

- CC

100.050 0
30.0 20.0 -

10.0 9.0 8.0 7.0 6.0 5.0

4.0

3.0 -

3.0:

1.5 --

2.0

1.0 -

1.0-

Go

Gh

0 -

5 0 . 0 3 ~
10.0~
5.0 4.0 3.02.0 -

10-

0.9 --

Gh

~ 0 . 0 ~
1005.0 4.0
3.0

--

0.7 --

0.6

0.5 -

0.3 -

0.6 --

963. Since smaller K values had been used in structures that had an adequate
ervice history, a new look was taken of the derivation for G,, Gb and the
resulting K. Yura (see "The Effective Length of Columns in Unbraced Frames,"
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,
should be used in Eqs. (7-4) and (7-5). The use of E, is equivalent to

0.2 -

0.1 -

0.1-

o-

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

0.3 -

a. The far end is hinged: multiply the E $ / L , ratio x 1.5.

b. The far end is fixed: multiply the EI,/L, ratio x 7.0.
4. If beams are simply framed to columns, use Fig. 6-3 for K.

1.00.9 0.8 0.7 -

0.4 -

0.5 -

a beam or girder is used with adequate attachment to the coiurnn of

0.4 -

0.2 -

. When

2.0 0.8 --

I .O 0.9 0.8 0.7 0.6 0.5 -

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

true pinned connection.
2. When the column is (rigidly) fixed to an infinitely rigid base, the EIb/L,
ratio -+ co and the corresponding G -. 0. For this situation it is suggested to
use G = 1.0 since there is some difficulty in producins a tmIy rigid connec-

o-

- E,
Gine~asilc- E Ge,,st,c
( a ) S ~ d e s w a ypermitted

( b ) No sidesway

Figure 7-5 Nomographs for the effective length of columns in continuous frames for lateral restraint
conditions indicated.

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

2 EIC / LC
2 EIb / L b
- 2 EIc/ LC
G
- 2EIb/Lb

Go =
and

(at far end of column)

(75)

ince El = AE and A I 1, it follows that use of G,, ,,,,, ic gives a Kkchric <
nce El is somewhat awkward to obtain and recalling in the derivation of &e
ISC equations for Fa in the inelastic region,

-9;.
..

we used essentially a SF on FCr where

(7-4)
and in the elastic region

(at near end of column)

(7-5)

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 interchangeably).
When E = constant it may cancel from Eqs. (7-4) and (7-5); however, when
inelastic buckling is developed, E, should be used for E in the E I c / L c ratio.
Other considerations include:
1. When the column is pinned to the base, the E I b / L b ratio is zero, since the
theoretical value of I --+ 0 that results is G + co. For this situation it is

Fa

SF(KL/~)'
From these equations it follows (using i

inelastic. r

FO'
G,= G,

elastic) that

(7-7)
Fa,
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
e range of K L / r = 40 to 60, the variation in SF is essentially negligible. The
se of Eq. (7-7) requires values of Fa, for the same K L / r value as F,,, so i t is

3U8

BE.L\I-COLUbN

STRUCTURAL STEEL DESIGN

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
F,' =

1 2 - 7 ~ .' ~
23(~~/r)~

Note thgtothis value is independent of 5. The values shown in the table of

F,' >
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.

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,
1973) proposed that the iterations for K could be eliminated by using fa = P / A
instead of F, to obtain

Figure E7-2
, ' ,l

Find K:use W14 x 99 through three floors:

3.94

Reduce for inelastic effects:

In lieu of using f,, which might reduce Ge excessively, one may elect to use
F, = 0 . 6 4 . Again carefully note that the reduced G is used only when KL/r <
C,. It should be self-evident that values of K 5 1.0 are not to be adjusted.
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
K,, does not have to be considered (e.g., close contact with interior masonry
walls) and F, = 36. Using AISC specifications, design columns CD.
SOLUTION
Assume that KL
section W14 x 99:

fo

From Table 11-5 for

compute
Fd =

fa =

712E

23/12 x 5 j 2

570

= -z--= 17.86 ksi

29.1

17.9 ksi, obtain K L / r

49.4 ksi

5 5 . For K L / r = 55

(calculated also in Table 11-7)

The revised

= 14 ft and use Table 11-4 to obtain a tentative

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

attached to the base so that

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

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

F, = 178.2 MPa. For Fa = 178.2 MPa, the tentative column area is

Fa = 19.37 ksi

=29.1(19.37)=563.6>520kips

DESIGN 311

O.K.

Si:lce this value is about 40 kips larger than needed, try a W14

90:
Try W250

114.6:

520
KL = 34 (Table 11-5)
- 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):
fa=--

Fd

129.18 ksi

F'

1 as before, and from Fig. 7-5a obtain K

Fa = 19.64 ksi
Pa

1.25.

> 520 kips

r2(200 OoO) = 366.6 MPa

23/12~53~

(also Table V1-7).

--- /

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
Left column:

(Table 11-5)

26.50(19.64) = 520.5

Use a W 14 x 90 for column 1.

O.K.

189.4/4.6
0
Obtain from Fig. 7-5a:

Gb =

///

Example 7-3 Given the frame shown in Fig. E7-3, with sidesway possible,
use the AISC specifications and Fy = 345 MPa to find the required column
size. Use the same section for both.

co

1.2

use 10.0 (AISC recommendation)

for right column

K = 1.75
for left column
Check the left column first, since K is much larger than the right column:
1.75(4.6 x 1000) = 70,6
r
114.05
Fa = 143.7 MPa (Table VI-6)

KL

-=

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.;

A solution: use W250 x 114.6 sections for the columns

O.K.

Example 7-4 Redo Example 7-3 with sidesway somehow restricted.

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
the right column and in both cases, using the "recommended" design values.
Alternatively, use Fig. 7-5b.
Since the K must be less than 1.0, let us use the "experience" of the
previous problem to check a somewhat lighter section. Try a W250 x 101.2:

SHTO, and AREA) interaction equation with bending about both ~ x e A,,
s
as well as axial load as
fa

fbx

-fby

Fa

Fb

Fb

(7- 10)

-+----I--<

rior to 1963, the value of F,, -. F,, = Fb. Currently, we recall that F, depends
Witeria,
a d
n several factors, including unbraced length and compact section
.C
'6
"
general AISC allows
..>
"$

Fbx = 0.66Fy

or

F,,

or

F,, = values from AISC Eq. (1.5-6a),( 1 3-6b), or ( 1.5-7)

0.60Fy

Fby = 0 . 7 5 5 for W shapes due to having solid rectangular flanges

F,' = 1032 MPa (interpolating Table VI-7)

Gb = 10

for right column

K = 0.72
K = 0.65
Check the right column:

KL

'4

for left column

Gb = 1.0
Using Fig. 7-56, we obtain

-=

Currently, Eq. (7-10) is used only in certain limited stress conditions. For t5z
remaining stress cases, ther more complicated formulas based on reszarch,
plastic design,..,and elastic stability concepts are uszd. These will be psr-tidiy
developed in the next several paragraphs to indicare some of the limirations so
that the practitioner will have some idea of how to follow through should tbs
design vary from routine.
Refer to Fig. 7-6 for a short ( L / r + 0) rectangular section of dimensions
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

for left cblumn

for right column

0.65(4.6 X 1000)
= 26.5
112.78

rx
Pa = 12.9(190.3) = 2455

Fa

> 2400 kN

190.3 MPa
O.K.

Check the left column:

-KL- - 0.75(4.6 X

1000)
= 30.6
Fa = 186.8 MPa
112.78
rx
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
f, + fb

' -I,= b X

:I.,,

Fallow

Dividing this equation by Fa,,,, = Fa, one obtained the widely used (AISC,

Figure 7-6 Plastic hnge formation in a very short membsr subjected to both an 1G2! force .n\f
moment.

314

STRUCTURAL STEEL DESIGN

.i'

'* "

~uk$i?&in~ for yo (shown on Fig. 7-6), webobtain

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

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:

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 adjstment in bending moment, we may rewrlte Eq (e) to 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

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

Figure 7-7 Plot of interaction

equations as shown. P, and Ma =
allowable values.

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.

F*

7-8 Influence of K L / r ar

BF.A\i-COLW

316 STRUCTURAL STEEL DESIGN

2
hlo

The coefficients are KL/r-dependent, and several values given by Galambos

and Ketter are as follows:
r

Figure 7-10 Plot of c o l u m Inr:raction for end moment on o a s end

and with efiscr of using C,=
(dashed lines) ior comp~fison.

Figure 7-9 Plot of K L / r for 40

and 120 for a column with moment on one end.

and for unequal end moments a linear equation of the form

DESIGN 317

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

%,

KL/r

0
20
40
80
120

0.42
0.70
0.99
1.81
3.16

0.77
0.46
0.17
- 0.72
-2.51

1.13
1.14
1.16
1.19
1.25

1.11
1.18
1.23
1.52
2.53

is in agreement with the definition for C, (i.e., single curvature bending is

critical for buckling instability than reversed cufvature, and simiIarIy the
The effect of using C, is to generally decrease the effect of the amplification
factor and is illustrated in Fig. 7-10. The use of C, produces a n intersection of

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
/ M , ratio. To the right of the intersection of the two curves, Eq. (d) cor:;:i.!i.
is requires use of two equations in design and using the most critical .of

researchers are as follows:

Cm= 0.60 -

cm=

0.4Ml
2 0.4
M2

1.75 - 1.05M,/M2

+ o.~(M,/M,)~

We obtain for bending about the X axis the f o l l o ~ n g

'
it

by

'llis is because it is easier to solve for the most critical value by using the two
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
sf PC,for Py and M,,,,, for M, and with section area and section modulus to
ob , stresses. This gives

a,

+ C , ~ B ~ ~ a,
~ , (P(KL)?
~ )

ax= 0.149~i-2x lo6 ksi

where

PdliOW

(fps unlts)

P(KL)* = ksl
#

'

By analogy, Eq. (7-13) becomes

Fb = allowable bending stress

fb

,#@
E ~r

= -;;
3

For Eq. (7-14), we obtain in a similar manner:

CG iining terms and extending the case to biaxial bending, we obtain

Fa
Fa
Fa
P0 . 6 5 + B, M,yFb.r + B,. M, - I Pal,,W

AISC Eq. 1.6- 1a

-fa+
Fa

'1

cmxfb~
f

'myfby

(l-fa/F&)Fby

< 1.0
-

(7- 13)

When j,/Fa (or P I P a )

(7-140)

*84

. Fby

< 0.15, we can obtain

and
AISC Eq. (1.6-1b)

fa + - + - - < I
fbx

fby

A careful analysis of these equations shoxvs that F, is based on P,,,which

depends on the critical oallie of KL/r. Bending resistance and any ampiification/reduction of moment effects is with respect to the bending axis ~~iitfi
nrnh)pmc+ h P n\\nwahlP
2uial
stress F_will be
...
-l. ,...,
;*
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
t/l
.. --.-.
.-.
based on K L / r , , but the P(KL)' term will depend on the axis resisting bending
Since the X axis is usually used, it is the P(KL,)~product that would usualIy be
required.
The modified form of these interaction equations is generally considered to
can be tabulsted for a number
be easier to use, since the right-hand side
L---:-b:--

IIIUIIJ

Referring agaln to Fig. 7-7, we note no reduction in moment capacity until

P/ P, > 0.18, so rounding 0.18 to 0.15 for convenience and to be conservative,
we obtain:
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

but ,$A
the actual column load P and AF, = maximum allowable column
*(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,
niult~ply/./F: by A/A, and take FiA = P, = 0.149 X IO~/(KL/~)'ksi (in fps).

u u r r ~ r t rr

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

COEFFICIENT C,
When a column in a structural frame is restrained against lateral translation
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

C,,

1.0 + -7
Fe
+la

(7-15)

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,

(l-fo/~;;)~bx

CmJby
( 1 --I;/F;)F~~

' "

+ -b - - + - fh
---I1
Fb, Fb

fo

0.472F;
F"

_< 1

(7-17)

(7- 15)

n ' ~
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,

fbx

Figure 7-11 C, reduction factor for beam-column interaction equations. (a) No sidesway: C
, = 0.6
- 0 . 4 M , / M 2 . (b) Sidesway: C
, = 0.85. (c) Column with transverse loading: C, = 1 + +fa/F,'. ( d )
Several cases of transverse loading and factors shown.

1 -a

/F

.b

rb."
( 1 - .fa / 1";') Fbv

<

V'E

F;" =
is computed using Eq. (7-12a):

1.431(~L/r)'
lso, at points braced in the plane of bending:

fa
with attention to signs (single curvature = - M,/M2). Note also that M, is
defined as the smaller of the two end moment values. When sidesway (Flg.
7-1 16) is possible, AISC specification allows:

o.55Fy

fbx

jby

Fbx

Fb,y

+-+---I1

fo

fbx

fby

Fo

fby

Fg,

-+-+
1
---I

Cm = 0.85

(sidesway present)

> 0.15:
(7- 19)

32.'; ... :UCTURAL STEEL DESIGN

BEAU-COLL3c.F

7-'? 9EAM-COLUMN DESIGN USING INTERACTION

EQUATIONS

value of KL' to

The design of beam columns using the interaction equations is essentially a trial
(i!.-i.;:tive) process. A section is tentatively selected and analyzed and if the
i. se;.:ien is too small, a new section is selected and the process repeated until a
sz:i:i'actory section (both strength and weight) is obtained. The steps may be
ou: ,-.edas follows:

,.

Fb = 0.66Fy

gA
1.

nr

C,BM-

AP

BM-Fa
Fb

Fa
,B
Fb

[using a part of Eq. (7-13a)]

[using a part of Eq. (7-14a)]

'::,pection of Table 11-4 (or VI-4) indicates that

C, Bx-Fa /3

0.2 to 0.3

(6.5 to 9.0 in SI)

Bx- Fa
Fbx

0.1 to 0.2

(3.3 to 6.5 in SI)

,B

estimation for

ax
ax - P(KL,)~

...

, '
,

:,

..

+ A P = P + (0.2 to 0.3)M
= P + (6.5 to 9.O)M

: *6

(M in in . kips)

( M in kN m and SI)

ihk value of P, enter a Table such as 11-4 or YI-4 (or in AISC Manual
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
P.',/K; > rx/ry, the KLx/rx ratio controls and one must use an adjusted
r'

check for Ilghte~s ctlon

use a 1 a ~ ~ e r s e c t i 6to4sa:~sfqdesign

With care and some preliminary scrutlny ofv Table 11-4 (or Y I J ) , an
adequate section can be obtained In one to three inals. Thls IS possible. since the
tables show little change in r,/r, and B for the two or three sect~onson either
side of the selected section.

&mate can be made similarly. An estimate for AP can be made in a

s,.rnewhat similar manner for using the AASHTO and AREA equations if
desired.
4. Compute the equivalent allowable column load as

':!

L,nb,acedI L,
computed from Eq. (4-23), (4-26), or (4-27)
if

+ P,, << P,,,,,

P + Peq > P,,,,

~ h t values
:
for bending about the Y axis are considerably larger, but a AP

+w,(;&

Fb

0.6Fy

--,

P,, = P

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:

fbx

;~ntl

Fb

if LUnbrace,
I LC

i f L ,,r,,
> L,
The unbraced length IS the actual and not the KL value and IS taken with
respect to the bending 2x1s.
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
is adequate. If fa/Fa > 0.15, ~t will be necessaryfto :heck barh Eqs. (7-130)
and (7-140). In this case compute:
P ( K L ) ~ use KL with respect to the bendlng a.uls. ~ h l c hmay be different
from the cntical K L / r used to compute the allonable avlai stress
Fa
8. Check both Eqs. (7-13a) and (7-14a) to sat~sfq

AP

" #*?

e tables, obtained as
KL
KL' = ---I
rr/ rv

It is necessary to use r,/r,, slnce the ratlo IS fixed for a section but is cor
known until a section is tentatively selected.
6. Record Ptable,A 1 ry, r,/ry, LC,L,, B,, and a,. L, and L, are needed, so a rspid
determination of Fb can be made:

,-(erminethe axial. force and column moments. We note that this step is also
1 yitive,, since indeterminate frame values are not found until a tentative
.,.!~tlon is used in the analysis.
''2.::>&ipute K to obtain KL. It may be necessary to determine both Kx and Y,
!':bending on column end conditions and lateral bracing.
2 . .'?.,timate the moment contribution (and we will use bending about a single
axis for illustration) as an equivalent axial load AP:

.*

'

'

Example 7-5 Given column and bending moments s h o \ n in Fig. E7-5 as

part of building frame in which sidesway is possible, we will limit the
column section to not over a W12. Use the AISC specifications and A-36
steel. Select a tentative section and use the basic equations (7-13), (7-i4), or
(7-15) as applicable.
SOLUTION
TO keep the deslgn of a beam column using the ~nterzctian
equations In perspective, we wlll assume that K, = 1.3 ~ n dX; = 1.0 2r,d

3.44

JIKUGIURAL STEEL DESIGN

BE.4'4-COLL5B

DESIGN

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
Figure E7-5

7.48
22

Try a smaller section-try

A

>

1 1.80 1n2

LC = 8.5

<

L,
.

12

L,,

19.6 > 12 ;
.
Fb = 0.6% = 22 ksi

16

&=z=-=

12'ft -+ Fb

F, = 15.98 k s ~

1.92(36.2)'

0.6F,

22 ksi

42.5( 12)
jb=----- 9.83 ksi
5 1.9

.112~

'"

1.92(36.62)

1 1 I .3 ksl

9.32
I I 1.3

= 1 - ---- - 0.92

By Eq. (7-14):
0.58

a 2 ~

74.2

15.98

7.48 ksi

1.92(~~,/r,)'

KL

-+ - =

-f, -- - =9.32
0.58>0.15
F,

We must use both Eqs. (7-13) and (7-14):

Fel =

>

F'

lo
14.70

9.32 Lsl

S, = 51.9 in3

Sx = 64.7 in3 (Table 1-3)

l lo
1 1.80

= ----- =

Pa,,, = 236 kips

Fa = 16.06 ksi (Table 11-5)

Y'

ju

'7

x'
= 2.64

x 40.

W 12

ry = 1.94 In

A = 14.70 in2 (use Table 11-4)

7.88
22

-+ - = 0.70 << 1 .O

that the K values do not change with section size.

Use C, = 0.85 with sidesway. After some study of Table 11-4 (with P
1 lo), let us try W12 X 50:

Fb,

= 113.8 ksi

+ 0.85(9.83)
= 0.991 <
0.92(22)

O.K.

1.0

By Eq. (7-14):

ru

0.6%

932
+ -9'83
-22
22
Use a W12 x 40 column.

+-<

F, -

1.0

- 0.87 < 1 0

O.K.

3 3

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

Check Eq. (7-14a):

List. Cm = 0.95 (AISC actually allows Cm= 0.85 if desired).

Tentatively:

+ pequv

pgwen

. 445.. + 7(41)

[estimate factor as 7 (between 6.5 and 9)]

= 872 kN
Sc2.1 Table VI-4 and select W310

x 59.5:

P,,,,,

A = 7.61

l o v 3 m2

rx

-=

114.1
I 14.1 = 753
(445) -+ 8.94(61) --_
150
130
Using Eqs. (7- 13) and (7- 14) would give:
Eq. (7-13):

< 868.3 kY

860.5 kN
2.64

'Y

1.25
>I

O.K.
Use as tentative section, W310 x 59.5.

Fa = 1 14.1 MPa

LC = 2.58 m
Fb = 0.6%

4.87 m
I~O'MP~

F ( x L ~ )=~445(1.25
fa

Actual

7.61
=

(Table VI-5)
L,

> 3.45

3.45)2 = 8.27 X lo3 kN . mi

58.5 MPa

AFa

- --- 58'5
Fa

114.1

(in same form as ax)

-0.51>0.15

P + G,,,,BxMx-

\*@

Fa
Fbx

445 +,0.95(8.94)(613-3

///

Example 7-7 The top chord member (No. 6) of the truss used in Examples
6-4 and 6-5 will be designed to include the member w e i a t and a temporary
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 temporarily closed to traffic if the maintenance load is too large to be carried
safely with traffic (live loading). Other data:
Dead load = - 283.75 kips
Live load = - 109.44 kips
P = - 393.19 kips

7.61 (1 14.1) = 868.3 (vs. 860.6 interpolated)

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.
E7-76):

Using Eq. (7-13a):

vB '

O.K.

ax
< 'allow
ax - P(KLX) -

132'8
150 132.8 - 8.27

865.3

< 868.3 kN

A = 14.70 in2

O.K.

I,

11.00 in'

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
i&

Crane colurnns in an industrial building. ( a ) Column w t h bracket. ( b ) Stepped column.

4 ~ stepped
p
column.

';62

(c) 1 .

bo.,. l:;e X and Y axes. Now if we take the X axis of the main column member
' %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:

reasonable member sizes and deflections are obtained. Different member

.sizes can be used for any member. This was not done initially in the two
examples, as the work is still highly preliminarq..
2. Based on the computer output. begin to redesisn the rnc.rnb.srs. Increesc or
decrease sizes, depending on forces and deflections.
. Reprogram and check output for forces and deflections.
. Repeat as necessary.

-8.1 Modification of K for Stepped Columns

1. 5 x e d or pinned against rotation at the roof truss level. The roof truss of the
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
excessive, knee bracing may be required from the column to the roof truss.
3. There wiN likely be moments in the column at the roof truss. There will possibly
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.

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. IrrrgeIf we use the differential equation

and allow for boundary conditions of different loads in the lower se-ment, no
lateral displacement at the column ends, change in I for the lower segment and a
common slope at the junction of the upper and lower sezgnents. we obtain

'

n*

Y Axis
i

**

$1

d&iyi4!yss
:. :
I

" 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.
i . &!&Iintermediate bracing. This is necessary to reduce K L / r , of the main
!::4ccr.;mn and of the column segment above the crane girder point. Girts may

where J = 1 + Po/ P, - a
a = L,/(Lo+ L , )
This is a solution as given by Sandhu, "Effective Length of Columns nich
Intermediate and Axial Load," A ISC Engirieeririg Journal, October 1977. Ths ki

32

STRUCTURAL STEEL DESIGN

BE.L.f-COLL?Ci

DESIGS

33

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.

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

This value of Ke is for a pin-ended column, and it is necessary to multiply this

by the AISC value of K as obtained from Fig. 7-5 or 6-3, where the actual end
conditions are taken into account. For general design office use, Eq. (7-22)
should be programmed on a computer, so a plot of a' = Po/Ph versus K, can be
made for selected ratios of a.

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 displacements. From the computer output (and for a tentative initial built-up section
iteration) for LC = 1, the axial loads and moments (for left column) are:
Member 26 (uppermost):

= 318.74 kN

moment

120.59 k3

c'

'4 STRUCTURAL STEEL DESIGN

Member 25 (intermediate):
P = 387.1 1 kN
Member 24 (lower section):

moment

338.99 kN . m

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.

considered stable by adequate use of girts and siding). Use the largest force
in the upper column, 387.1 1 kN.

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 progammable calculator to obtain k L = 3.525.
Figure E7-86

I'

W360

314 data:

PI

I, = 1107.2 x

m4

A = 40.0 X lo-' m2

W360

Zy = 428.7 X
=

399 mm

bi

m4

PI =

+ 1262.11 = 1805.1 kN (actual)

12.426(1 + a') EI =- 15.82EI

543.01

= 401 mm

L~

Also,

101.2 data:
1, = 300.9 X l o v 6 m4

A = 12.9 X

m2

ly = 50.4 X

d = 357 mm

m4
bf = 255 mm

Compute r and ry. Use EMx, = 0 to locate the new Y- Y axis.

(40

+ 1 2 . 9 ) ~= 910(12.90)
X

910(12.9)
= 22 1.9 mm
52.9

(use 222 rnm)

The actual K
1.2.

Ke

Kc,,

,,,,,,,, .,
K

From Fig. 6-3c obtain k',,,

c,,,,,,,

0.79(1.2) = 0.95

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

#
U.

13'6(0.8 16) = 27.7 MPa

;#
(note decimal shifted for 1 )
, &'= I
9.235
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

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 calculatlons to a minimum and maxlmize use of the computer.

///

in

zcl<crsd bay.

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!ambos ("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:

7 9 CONTROL OF SIDESWAY

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
a ~ obtained (see Fig. 7-14) via:
most efficient columns. ~ h i s ' m be

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.

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 13bracing member is so small and flexible that it will buckle (with stresses we:!
below
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.

6)

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.

BE.k\f-COLb\fX

DESIGN

339

4%

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)

P,

AFy
4

(7-30)

~i\'>

(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,
Mu,
P,,

1.5 (author's cholcc).

l.l[1.1(120) + l.5(\80)]

1 . 1 [ 1.1(400)

+ Ii(6Cn'l:

42.2 L
S
i - rn

= Id7-!

'K"

BMU-COLUMX DESIGN

STRUCTURAL STEEL 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

show that K

+ B cosky + Cy + D

x = Oaty = Oandy =

M=Oaty = O

Z, = 2.687 X

138.9

r,

-0aty
-"
Q-

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:

rx =

=d-&

0.7 for the column of Fig. 6-,3b. Note the boundary conditions:
,

A = 20.13 X
ry = 79.0 mm

and

controls

m3

from which
qic =

Figure W-2

0.90 - 0.25(0.534) = 0.77

Pu = AF,[ 1 - 0.25(0.534)~]= 20.13(250)(0.9287) = 4674 kN

1

7-3 Determine the effective length coefficients K for the columns Ln the frame shown in Fig

M, = ZF, = 2.687(250) = 672 kN m

Substituting values into Eq. (7-28), we obtain
1474
442
N.G.
0.86(4674)
1.18(0.86)(672) = 0.367 + 0.648 = 1.015 > 1.0
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

Annuer: BC = 0.65; EF = 0.63.

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
from the first term of Eq. (7.28). Use a W310 X 178.6 section.

///

Figure W-3

ill

t:

Figure P7-5

Figure W-8

I!I

344 STRUCTURAL STEEL DESIGN

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.
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 materiIs, 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-

FIgure Vm-1 High-strength bolted joints. ( a ) Splicing smaller-to-larger column using filler plates.
(b) Splicing same-size column. ( c ) Diagonal bracing.

tions using bolts are further classified as bearing orjricrion-type cop. nsctions.
2. Connection rigidity, which may be simple, rigid (as produced by an indeterminate structure analysis), or of intermediate rigidity. The AISC, in See.
1.2 of the specifications, classifies joints based on connection rigidity as:
Type 1: rigid connections that develop the full moment capacity of tilt
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 requires 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
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.

r'

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 consideror bearing connections, since some small amount of relative moveeen 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

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.

connections is not developed as the shear resistance of the connectors;

rather, it is developed as the product of the clamping force produced by
tightening the bolts (or driving the rivets) and the coefficient of friction
between the clamped parts. It is expected that in load factor resistance
design this will be directly used as the design parameter, producing an
equation of the general form used in several other design codes (outside the
United States) as

where

+ = performance factor (0.67 to 0.70)

p= coefficient of friction x number of slip surfaces
= total developed clamping force as the sum from all the bolts used in

2 A, F,

the connection

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
mporarily. &vets were heated in a furnace (portabie for onsitz use) to a cherry
ed color (approximately 980C) and inserted into the aligned hole through tLe
everal parts to be connected. One member of the riveting crew then applied a
ng bar with a head die to the manufactured rivet head to hold the rivet 1I1
and to shape. Another crew member used a pneumatic driver with a h a 3
to forge the protruding rivet shank to produce the other head. The forg.ing
ation simultaneously reworked the rivet metal and caused a shank enluget to very nearly fill the oversized hole. This reworking and shank enlargeent, together with the shrinking of the hot rivet, produced a substantial joint
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
ed plates making up the joint. The riveted joint has had a long history of
cess under fatigue stresses as in railroad bridges. Only recently has ,ARE.-:
llowed use of high-strength bolts and welds in joints for railroad bridgts.

- 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 coxosion. A-325 bolts may also be obtained with a galvanizing coating,
&LAG.?LS.
When high-strength bolts were first introduced into structurd applic-':
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

A-325"
A-490b

Figure 8-4 (a) Several sues of undnven structural nvets. ( 6 ) k v e t s in structural apphcations

Turn-of-nut

Specified torque

No
Yes

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.
Use two washers when the fastened material has
F, < 40 ksi.

-F',.-'i;?,

BOLTED &\Q R

w C O h > % ~ O ~ 359
S
,J

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)

i.

p
4.

eration of Eq. (8-1) we can readily see that regardless of +e t_Vcof

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
t3.s
may now see the rationale for P, = t(,K,

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-

kN (Table 8-2). The allowable

ess using AISC specifications (Table 8-1) is 120 hfPa with threads.
ne. From Eq. (8-1):
LunoN The bolt proof load is 141

Psli, = rn,uNT

iTU=

ultimate tensile strength (see Table 8-1).

k11(141)

since it is only necessary to consider one bolt and we will consider

slip plane, as in a lap joint. The safety factor is always defined as

SF =

one
Y

Prrslsting
Pallowable

~dequacy.
A bolt tension of approximately 0.7Fu gives adequate reserve strengt
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
relati' ~ositil~n...
This clamping effect also tends to hold the joint against nut
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
3/4 turn of nut) they may be reused one or more times. Tests on reuse
indicate that A-490 bolts should not be reused in any situations.

Pallow

= F,Ab(nornina~)

Combining, we obtain the safety factor:

SF

0.35(141)
120(0.7854 X 0.0'0@)I @

1.31

The reader should note that this safety factor is against slip and is not
safety factor of the joint, which is on the order of 1.67 (for a tension
f//
t and may depend on tension on net section).

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

assuming reasonable bolt spacing on the order of 3 x diameter

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

0.85 - C , ( L - C2)

fps: C, = 0.007

SI:

L 2 Cz

C, = 0.00275
C, = 406 mm

C, = 16 in

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 stresslues 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 Figbe 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
p

,Is

.'

4
y

Strain distribution (approx) in

bottom plate ofjoint above

m 8-8 Load and strain distribution in bolted joints.

.

. . ,,

, ., ,,,
Oi*

,,. ....

..

distance is required by AISC in Sec. 1-16.4 and using an edge distance

he bolt is in double shear.
thing holes, it is necessary to have an adequate edge and end
avoid warping damage to the material. The AISC specifications (Seee 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

BOLTED

I'

&\ID RIVETED COWEC~~OE~S

363'

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
,,s,

2.67 x diameter (3 x diameter preferred)

AASHTO (Sec 1-7.22C): ,s,

3 x diameter

AISC (Sec 1-16.4):

. .

-,

AREA (Sec. 1-9.3):

smin= 3

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-diaizster 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
produced.

:-

F b#:-. Several modes of joint failure. ( a ) Several joints. (6) Bolt shear failure. (c) Tension on net
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:

Sheared edge:

D, = 1.75 X diameter

Rolled edge:

De = 1.25 X diameter

?*A
8-42 Bolt Distribution and Gage Distances
It is necessary to ensure a reasonably compact joint and one where the
cted material is in reasonably good contact so that the developed friction
is uniform between the parts. If the bolts are too close together,
is obtained, since the maximum coefficient of friction is p 0.35,

--

AISC specifications require that all connections. except those in rrusses, c2rrying
calculated stresses be designed for the design load but not less thrin 6 kips (or 77
kN).
AISC specifications require that truss joints in either tension or compression
be designed for the design load but not less than 50 percent of the effective
strength of the member based on the type of desi,on stress.
The AASHTO specifications require that connections be designed on the
basis of the average of the design load and the effective strength of the member
but not less than 75 percent of the effective strength of the member. This is
because many of the structural members in U S H T O design are controlled by
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.

-Pi

8-4.5 Shear Lag

Long Joints are undesirable from the standpoi& of reduced efficiency (below 85
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

used for strut

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

where

& I P ( ~m)In eXCe,S

of

L = 450
Same as M S H T O

x= distance from gravity axis f 0 fastener plane (= d / 2 for w shapes and

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

the connection load. For a joint with constant-size fastener~in a s ~ m e t r i c z I

'fastener

'"%re 8-10 Shear 1% in a W shape connected to a pair of gusset plates.

Ptotai
- number of fasteners

is assumpllon made for ellher fastener shear o i bcanns. BeannZ i s ccriided In some speclflcatlons as beans: of fastener on ha$ nlelal 37J G~ C a t
+* *
,"cur

STRUCTURAL STEEL DESIGN

BOLTED .LXD RIbEXED COh3ZCTIOE;S

Table 8-3 Table to determine required bolt length.Fased on grip and thread
length (use table to determine if threads will fall idshear
plane; see example below)
L,,,
gnp + L; round Lrcqdto next larger f in or 6 rnm

Bolt sue, m
I
I
'
i

Thread L

L, m

Bolt sue, rnm

Thread La

I1
16

12.5

25

If
1;

7
8

15
20

30
35

2 proximate and based on soft conversion.

r,xample Gwen the connection shown m sketch and uslng a 2-~n-&meter bolt. Are
tt : J ~ S In the shear plane (defmed by the plate junction)?

Grip=;+
Fiwi. the table, L

%:
tqr

L,,,

& + ;=

1&1n

1 In:

* R ~ u n to
d nearest larger

= I&

+ 1= 2 1
16

ble 8-4 Allowable bolt and rivet stresses for AASHTO specific~fions
stress mm normnal wit area except l u r r \ - ~ v iwiu 211
(use glven
b!
on)B

r ~ i d w a b ~shear
e stress. F-

Data from AASHTO specificat~ons,12th ed , Secs 1-7 22 and 1-7 41

Fb = allowable beanng stress on nvet or bol: from fastened material.
F n c t ~ o nconnectlons are required for connectlons subject ro stress reversal.
Bolts In beanng-type connections must be X-type (thresh excluded from s h a r
nes). Beanng-type connections are to be used,@
condary members. Reduce h~gh-strengthbolt values 2
an 24 In and F, < 42 k s ~
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

Distance for two plates = + $ m = 0.8125 In

Therefore, threads are out of the shear plane.

ineta1 on fastener. The shear area is obtained using the nominal fastener
diameter. The bearing area is the projected fastener diameter x plate thickness:
the fastener load Pi is

b.

P, = A , x F,
(shear)
Pb = D X t X Fb (bearing)
No allowance is made for the hole size being
in (1.5 mm) larger than the hnl
shank in bearing.
4
$
n the fastener group is unsymmetrical or the load does not pass through
oid of the fastener group, the fasteners are not equally
- * stressed. This
ii?,ibn is considered in detail in the next section.
.We note that the allowable shear stress F, depends on the design assu
,,tie: (:.at the joint is either a friction or a bearing type, and whether the
- 1 0

$?

ds are In the shear plane for

h
to determrne ~f the threads are In the S
av
be
used
to
obtain
the
alIowabl
-,ecificat~ons,respect~vely. Table 8-6 may be
earing stress for base metal-to-bolt or
his table are based on the ultlmate

- --

bection. Table 5-3 W!i be

!.XASHTO, and A E X
$0 obta~nthe dIott-ahis
XISC values s h o i t ~in
,
material, compcisd

No
made for fasteners in either sin&
. distinction
- --- - -- - .. is
.
- .
-

or doubIr shrzi. T i e

SHTO bearing stress based on metal to fastener is

F, = 1.225
nd will generally be limited by fastener-to-metal stresses. as shown in Tab12 8-4
der the column headed F,.
The AREA bearing values are stipulated for the fastener type except th2t
ring is not considered. in- the
for connections using his!-strength WE.
. . design
- ,
ese values are shown In 1 able 8-6.

BOLTED hSD RIV-

Table 8-5 AREA fastener stresses (all connections are "friction"type)

F,
Fastener
&vetsa
Hand-driven
Power-driven
~olts~
A-325
A490

MPa

ksi

11.0
13.5

248
248

20
27

36
36

ber 1965) give average values of ultimate shear in terms of ul

rength F, (see Table 8-1) of

Fo

ksi

MPa

76
93
138
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.
Need not cons~derbeanng on bolts In frict~onconnections.

Table 8-6 Bearing values for rivets and bolts by several specifications
(top part of table is metal-on-fastener,bottom part is fastener-to-metala
Material
ksi

AISC

FU

MPa

ksi

MPa

ksi

MPa

AASHTO

ksi

MPa

AREA

ksi

MPa

F=

0.62(825 MPa)
120

250
Not required

Top part: F, = 1.5F,,; AASHTO: Fb = 1.22Fy. Bottom part: depends on fastener;

fastener bearing generally controls for bridge design.

4.26

int geometry reduces this value of F to something on the order of 3.3 rot
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
ction for A-36 steel of

F = - - F" - 1.72
0.58 F,

A-307 bolts
&vets A-501 and A-502
Power dnven slngle shear
double shear
A-325 and A-490 bolts

CO

(net section after slip)

---F" - 400 - 2.67

0.6Fy

150

(gross section before slip)

servations of joints in long service indicate that a safety factor of F 2 2.0 lor
fasteners gives satisfactory service.

-4.10 Splices in Beams

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,

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
= 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.

x 100/20

370

BOLTED

STRUCTURAL. STE

A-325 bolts a r i required for the tension sp

= 250 MPa a
be friction type, or bearing type.

. E8-2a if the metal is F,

-mm bolts, so the hole diameter = 22 3.0 =

must be sheared twice for the beam web to
splice plates, the load per bolt is

P,,,, = 2

I /

11 ,'
For a frictidn-type connection:

X A,

x F,

P,,,,= 2(0/7854 x 0.022~)(120)= 91.2 kN/bolt

(double shear)
he number of bolts required is N = 442.5/91.2 = 4.85. Use five bolts.
the bolt pattern shown in Fig. E8-26, .so that the splice plate wihth will
maximum but the maximum net section is obtained for the channel.

.OmRIVETED COh?iF,mOt;S 371

r a three bolt line and sheared edges, the minimum edge dis
use 40 mm
1.7(22) = 37.4-mm
Minimum bolt spacing = 3 0 = 3(22) = 66 mm
250 - 2(40)
O.K.
Use spacing =
2
For the forward two bolts:
85
Center bolts as 40 + - = 82.5 nim fronI edges
2
Spacing to edge is less than maximum allowed of 12t = 12 x 10 = 120 m a .
Use the distance from the front bolt to the edge of splice plate at 40
(1.75D for sheared edges) and similarly from the back bolts to the edge of
the W410. Set the gage distance so that-only two holes are deducted from
the critical net section.

C310 X 44.6

s =\
/
2
1
z
= 46.1 mm

use s

50 mm

(arbitrary choice)

Check the bearing:

On splice plates:

F, = 5(0.01)(27)(1.50 X 400)
660 > 442.5
O.K.

5(0.013)(22)(1.50 x 400)

Ph = iV x ,+Ihx

On web of C3 10: Ph

O.K.
858 > 442.5 kY
Use five bolts for the friction connection.
For a bearing-ope corlrrecrlorr
Grip = 2 x 5 x 13.0 = 23.0mm
L,,,,=23.0+28=51
iise55rnrn
=

Figure W2a

Figure E8-2b

Check the net section of the splice plates. With three bolts out, the are
requirements are:
Gross:

442.5
A, =
= 1.475 X
2 x 0.6Fy

m2

Net:

Ap =

442'5
= 1.106 X
OSF,

m2

A >-=
"lo6
- 0.85

1.301

13 = 18.0 mrn

say 4 bolts

m2

> 1.106

Distance from bolt head to end of thread = 55 - 38 = I7 mm < 18 ma

This computation shows that the threads are in the shear plane, so F, = 13
MPa (instead of 205) and P,,,, is

Try two plates 250 x 5 mm with three holes in the critical section.
A,,, = [250 - 3(25.0)]0.01 = 1.75 X

L,,,,,, = 38 mm (Table 8-3)

Width of one splice plate + beam,web = 5

O.K.

.=

2.6

(also use 4)

Use two columns of two bolts for bearing-type splice

STRUCTLTRAL STEEL DESIGN

Example 8-3 Design the connection for t

(NO. 7) for the highway truss of Example
E6-6 of Example 6-6 and to Fig. E8-3 (ref
8-46). The previous examples were used
members 7 and 9. Let us use the same cro
used for the end post (all in compression
use the same sections as for 7 and 9 (
stress). In any case, let US design the conn
(No. 7), which is in compression. Data f

II

BOLTED .kKD RIYETE# CObYECTIOhS

.-3

6-4, the allowable axla1 compressiv; stress was 9.35 ksi.

TODchord

AF, = 17.00(9.35) = 159.0 kips

159.0 -t 99.3
= 129.2 kips
2
The 75 percent member strength criteria give
pa" =

or the top chord as was

ng web members may
controlled rather than
e vertical web member

718 bolts as req'd.

.

P,,,

, and 6-6. Refer to Fi

Use A-36 steel and the AASHTO specificatiJns. Use A-325 high-strength
bolts. P,,, = - 99.3 kips (P,, = - 40.6 kips). The gusset plate t = 5/16
in (minimum t allowed by AASHTO for a plate).

373

Memno

I'

Po,,

159.0(0.75) = 119.3 kips

< 129.2

From Table 8-4, the allowable bolt shear stress F, = 13.5 ksi. Thz nuxSsr
of bolts required in the connection to transfer 129.2 kips is
129.2
P
N=---?=
= 15.9 bolts
AbF,
0.7854 x 0.875"
13.5
Use N = 16 bolts for symmetry and since 0.9 bolt is not possible. Tne use of
16 bolts requires four rows. Use a bolt spacing of 3D:

Use a minimum edge distance in l~neof stress,,=

1; in (1.750 for cut ecZ):
,.
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

Figure J33-3

SOLUTION
The fastener design will be based on a friction-type connection.
AASHTO does not allow a bearing-type connection in a main membe
Stress range does not have to be considered for connection design.
We will try four 7/8-in-diameter bolts at a section as shown in Fi
E8-3, the same as assumed in the design of member 9 for tension
Example 5-7. A deduction for net area does not have to be made f
compression members unless a complete stress reversal occurs (this does not
occur here).
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

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
Check the bolt beanng on the gusset plate. slnce I, = 0.313 < 0.640 of
flange of a W12.
O.K.
x 0.3 13)(40) = 175.3 kips > 129.2
ade sufficiently wide that tension on a net section
li be some f~llerplates needed between the gusset
nce the W 12 sections are deeper than I2 in.
///
'

.Jf\

Example 8-4'Design,.thg connection for the vertical members of the mSn

roof truss of Example 2-6 that were designed in Exampls 5-6. F, = 250
MPa. Other data include:
P = 70.18 kN (tension)

."a

STRUCTURAL STEEL DESIGN

'

,t*'

BOLTED .CYD WYETED. ~ O ~. &..T I O K S

This design was based on using a 12-mrn gusset plate and 25high-strength bdlts.

With the'threads in the shear plane, F, = 150 MPa

P,,,,

SOLUTIONAISC requires that connections for truss members be designed

for either the design load or 50 percent of the effective strength of the
td
member.
L

P,, = 0.5(0.6<)~,

= 0.5(150)(2.66) =

199.5 kN

controls

2t0.7854 x 0.025')(150) l d

375

able 8-1):
147 kN

199.5
= 1.36
147
Use two bolts for shear, since a fraction is not possible.
Check the beanng:
N

=-

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
illustrated in Fig. E8-4b. Assume that a bearing-type connection (slip can
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.

0.E';On angles: Pb = 2(2 x 0.0063)(25)(600) = 375.0 > 175.9 kN

On gusset plate: P, = 2(0.012 x 25)(600) = 360.0 > 175.9 !c3

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

15-mrn bolts
1 2 7 X 89 X 6 . 3
I

:I -,.I

V\-FJ/v
I

_
I
-

Figure E846

Bolt grip

6.3 x 2 -t 12 = 24.6 mm

24.6 + 30 = 54.6
use 60 mm
(Table 8-3)
45 mm
One,aQgle + gusset = 6.3 12 = 18.3 mm
Thread runout location = 60 - 45 = 15 mm
15 mm

<

L,,,,,,

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
= PP,
through the pattern centroid and a moment with the magnitude Lti
ere e is the eccentricity of the load. This is illustrated in Fig. 8-l3a a ~ 5.d
considering that each bolt in a pattern that is centrally loaded carries its
ed share of the total load, we have, for equal-sized bolts,
P
p.=-

"

!V

where Psi is the shear force on the ith bolt with a vector to resist the a p p l i d
force P .
An additional
bolt force is develo~edbv the eccentric moment ,bf = ?r.
~
.
Assuming a group of bolts acting as an elastic unit. we have a concept similx to
that of beam resistance being developed and as related to the beam moment of
inertia. Referring to Fig. 8-14, we have a bolt pattern \%-ithan applied moment .El
which produces a resisting moment for rotational equilibrium that is equal to
-

,=n

i- l

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

the action of P through the bolt group centroid produces an additi

l'olt resistance R, which must be added to the R,,,,,, vector, it is better
obtain the H and V components of R,,,,,,.
Referring to Fig.
components of Rl are Rh and R, can be obtained by proportion as:
L::-IJ~

Rh - RL
-

Rv =
-

R,

COk>~~3
o77~8 ,s

he bolt pattern and plate adequate for the g~venload in a bearing-~?c-nnection assuming threads In the shear plane?
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
shown on sketch
Compute z ( x 2 + y 2 )

Y - - MdY - -My
R,=R-=
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
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,

<

1l s.?

Figure E & 5 b

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
f, =
0.7854 X 0.025'
Check the plate bearing:

'

1000

63'03
0.025(12)

63.03 kN

= 128.4 < 150 MPa

kips

O.K.
x = 2.75 in
I

210.1

< 1.5(400)

y, depends on the number of bolts Try 12 bolts Note t h ~ boIt

t
I is alw~ys
the most stressed for a bolt pattern such as this.

O.K.

Check the possible tension rupture of the plate along the forward bolt line:
Moment of inertia, I = 0'012(0'270)3
12

- 2(0.012 X

C(x2 + y2) = 12(2.7512+ 4(7.j2 + 4.j2 + 1.5')

= 405.75
50
R, = 12 = 4.17 kips

J
8

0.025)(0.075)~

887.5

Section modulus S = - =
C

16.308(2)
= 0.12-.. . . .,
270

Moment at forward bolt line = Pe'

R, = --(7.5) = 16.40 kips

405.75
,
,
,

= 110(0 15) = 1 6< LN

- 136.6 MPa < 0.6G

f,=s=-0.1208

. .,

Since 16.40/0.7854 = 20.88 > 17.5 ksi, the numbzr of bolts at a I-in diameter is too small without computing R.
Try 16 bolts (bypassing 14, since 12 bolts were so hi@Iy stressed):

O.K.

Check plate buckling:

- = - -150

- 12.5 < 2 5 0 / f i

887.5
405.75

R, = ---- (2.75) = 6.01 kips

C(x2 + y2) = 16(2.75)'

O.K.

t
12
The joint is adequate for bolt shear, plate bearing on bolt, and bending.

R,

+ 4(10.j2 + 7.5' + 4.5' + 1.5')

= 877.0

50
= 3.12 kips
16

//J
Example 8-6 Given a crane runway bracket that carries a load as shown in
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
bracket plate thickness.

R, = 1.012(2.75) = 2.78 kips

R =\/(2.78

+ 3.12)' + 10.62'

12.15 kips

Use sixteen 1-in diameter A-325 bolts.

< 0.7Sj-l( 17.5)

O.K.

'Wh6

'w

IS

Find the plate thickness for the bearing (assume that the column flang
adequate):

~q 8-7 for the fastener eccenrnc:ty shov*n

tp(1)(1.5 X 58) = 12.15 kibs

tp=-=

12.15
87(1)

l + 5
0 625 In
2
ccentnclty (for 101nt performance). it is

e,,,

0.139 ~n

h=7x3+3=24in
tph3
12

M=50(15)=750in.k1ps
2tp(lO.S2 + 7.5' C 4.52

$1

fi

Me

Fb=22ksl

+ 1 S 2 ) = 7745

3.625 -

~~~~~l~

S ~ L U T I ~Referring
N
to

Fig. E8-7, try twelve I-in-diameter bolts-

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
15.8

- 0.95 in

use tp = 1.O in

Summary: use sixteen 1-in-diameter A-325 bolts and a bracket plate

In.

Figure W7

= 1.0

8-3.: AISC Reduced Eccentricity for Connections

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:
T h b

%&

'$c

eeff= e

I + 2n
- --4

llne? of fasteners with n fastenen in any line:

' 8 ,

1
f i

eeff= e

l + n
2

--

to sre

8-7 Redo Example 8-6 taking into account the reduced ecceL;:::-

'6

t=--

Fig-

----- =

750(12)
774%

f b = F b = 2 2 + - = -1

1.

,333

ctlce of neglecting the eccentnclty for most f r ~ * d

nable design procedure.

Check the plate bearing along the forward fastener line and neglect bolt
$ales 1/16 In larger than bolt:

Ip=--

BOLTED f i RI\ETED.
~
CO~+XXOP.IS

STRUCTURAL STEEL DESIGN

jO(14.25)

C ( x 2 + y 2 ) = 405.75

712.5 In . kips

(Example 5-6)

R,

*" (2.75) = 4.83 kips

405.75

R,

1.756(7.5)

R =\/(4.17

13.17 kips

+ 4.83)'

j-,=-- 15'95 - 20.3

0.7854

i13.17'

> 17.5

15.95 kips

N.G.

% STRUCTURAL STEEL

BOL I E D L\I) R I L E r t D COh3ECTIOhS

DESIGN

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)'

H"

+ 4(92 + 62 + 3')

609.9

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
O.K.
jv = l.O(O.7854) = 15.45 ksi < 17.5
Using the reduced eccentricity requires two fewer bolts in the connection. 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
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
of fasteners. Equations are given in several textbooks which may be used in an
attempt to reduce the computational effort in finding the number of fasteners
for an eccentrically loaded connection. The author suggests that with the
increased computational efficiency available to the designer with the pocket
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

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 transferred across the connection. Frame stability is.provided by use of wind bracing
pi masonry walls closely fitted to the columns as shear walls.

(dl

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

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.
der. Coping IS shown for small floor beams IJI nesr foreground (d) F r z m g for bndse Siriz,t,s.

F
=A.

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
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 conncctions should be standardized for the given bridge to reduce fabrication co,jE;
neral standardization for bridges is not as easily done.

3PtL STRUCTURAL STEEL

DESIGN

BOLTED LL?)

Zxample 8-8 Design a beam framing connection for a W 14

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.

1/16) as in Fig. ES-8-this .@'?s

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

ch side) as shown, but we may note t
angle on only one side of the web.
sides it is evident that the bolts in t
ii-ebs of both the floor 'girder and the floor beani will be in double sheay.

v = -w=L

(2.5 + 0.0')(18) = 22-77 lups

2
2
LTsc. 3/4-in
A-325 bolts in a friction connection.
,..
",_....
P b , , , = 15.46 kips double shear
(Table 11-7, SSDD)
I!! a W14 x 30:

~h~ floor beam is a W760

Total shear = 280.5 k s

160.7 and frames into a plate girder (Example
I,.

16.0 mm

t,.,

13.8 mm

~ e ust use 22-mm-diameter bolts:

p,,,, = 0.7854(0.022~)(135)
jId) = 52.3 k s

.'I....

3.

22.77
use two bolts
Nshear = -- 1.47
15.46
22.77
= 1.29
Nbrg - 1.5 X 58 X 0.75 x 0.27
In the web of a W18 x 50:
22.77(2)
Nbrg

= 1.5 X 58 X 0.75 X 0.355

use two bolts

= 1.96

requires two bolts

2t(0.75)(1.5 X 58)

22.77

2.7

use three bolts (double shear)

or bearing with bolts in double shear, Fb

280.5
= 3.7
N b r g - 0.022 x 250 x 13.8

F~~the web of a plate

value); Table 8-5)):

22.77(2)
use four bolts for symmetry
15.46 = 2.95
Use a web angle with a length of 6 in and t to be determined:
For bearing:
Nshear

For the web of a W760:

Nbrg =

Nshear =

150 hfpa (Table S-5):

use bolts as required

[bolts in single shear and Fb

280.7
0.022 x 185 x 16 = 4.3 1
280'7
-- 5.36
52.3

Place bolts as shown in Fig. E8-9.

~ 7 6 x0 160.7

use six bolts

,
*L

2t(6 - 2(0.75)(0.4~,)= 22.77

ii

tz0.176

.r
(1
'

uset=$in

an angle with a length of 6 in. Use two L4 x 4 x 1/4 x 6 in long with

,

,I:-

- 1 1 ; ~

; 1 , ; +(-1;- 1

185 hfPa (rivet

use six bolts

T , -'77

t = 0.174 in
For shear:

!!-

.. ,

![;a

STRUCTURAL STEEL DESIGN

Design the framing angle:

'

'

'

L=5

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.
X

2t(0.022)(3)(185) = 280.5
i'

,!fx;& Checking the tables, try: L127

t 2 ll.5^'mm
127 X 11.1 (Table V-9):

g = 75 mm (Table V- 13)
Effective edge distance de = 127 - 75 = 52 mm
."

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
whether required or not.

8-7 FASTENERS SUBJECTED TO TENSION

~.~
.,,,
... .,," ,

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 tLclamped 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
q npnrlv nf
and A, relatively large. On the other hand. e,-,, = A
- -T-r ,/ / A E i-constant magnitude as long as the plates remain in contact.
""ll

------

In equation form and now considering a single bolt as in Fig. 8-16d. we

*

-$4

j-

,$

- Ae,

= Aebolt

he two clhmped plates expand, and taking P' = prorated part of total P

3%

STEEL DESIGN

where terms

BOLTED hVD RIVETED

~reviouslydefined in Eqs. (8-9) and (8-10) and ~

for A-36 steel,

A,,, = 0.7854(0.875)~= 0.601 ln'

F, = 22 ksl (bearing-type connection and threads In shear plane)

r = (36 - 22)/36 = 0.3889

62.4
0.601 X 22

'a = thickness of beam flange supporting load

=
X

Check the bearing:

pb = 3(2 x 0.25

8110 Design a hanger using a WT section for a load to be

by a palr Of
The load Is
E8- I Oa.

i 8-17
~ are
.

force carried by flange on one side of beam web

=stress ratio, defined as (5 - F,)/F,;

COh%xmNS 3%

Use A-325 bolts and A-36 steel; the load is 62.4 kips.
from the bottom of a W33 x 221 beam as shown in Fig.

W33 X 201

2.36

use three bolts

x 1.50 X 58) = 131 < 62.4 kips

~~t~ if thickness of the WT web 2 2 x

land,

o.K-

bearing and shear

Step 3. Design WT.

~h~~ step simply lnvoIves studying tables after making a computation
approximate depth based on edge distances and bolt spacing in
d, 2 1.5 + 2 x 3 + 1.5 + k

= 9 + kln
a WT12

47:

tf

2L's

%dge

b, = 9.065

d = 12.15 in
=

0.875

- 9.qo-c - :5.
= - -

t, = 0.515
5
- *I

-& \

'
7

k = 1-53 In

&?\K

\ '??l+.

S>L3)

7-

6 2 4k

Required depth
3 = 9 + 1 53 = 1053 > 12.15 ln
Check the bending moment at the toe of the fillet In web.

Figure =lob

Firwe m-loo
>

*>

SoLuT1oN The hanger design will requlre design of a

WT to
attach the
for the load and to select enough bolts of the proper sire to
carry the load in tension.
Step 1. Design the angles.
Assume that L / r is not critical and choose 7/8-in bolts:

A, =

2.15in2

2.15
0.85

- 2.53 in2

A&---

= (3'38

+ 1/8)]

~t~

+ 41.92 in - kips
~(0.875)~

= 0.1276L

Fb

0.50 in2

- 0.50)0.90 = 2.60 > 2.15 required

M = )I.*(?)

'=6=
6
s =M

at least 1/4 in thick, SO that bearing is not a problem. T~ two

1/4 (long legs back to back):
= 2[1/4 X (7/8

55
1
b = -L
-= 2.6875 in
2
16
T = -62.4
- - 3 1.2 kips
2

Fb = 0.75Fy = 27 ksl

62.4
= 2.84 in2
22

A , = - 62.4
=29

0.K-

O.K,

41.92 =12.161n
useL=12.5in
0.1276(27)
~~~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:
62.4
= 2.35
N=
0.601(4)
L =

' 4 STRUCTURAL STEEL DESIOFI

side of the flange is

P = 36(1.15)'(2 x 0.3889)+(1 +

pf

P
1.70

61.1
1.7

s)
=

62.4
-= -= 35.9 kips > 7

61.1 kips

O.K.

///

The flange does not require stiffening for this load.

Figure WlOc

Use four bolts for symmetry. Also, data for a W33

b,

201:

15.745 in

tf = 1.15 in
g = 5.50 in
Step 4. Check the "prying action" using Eq. (8-9):

F = - -62.4 - 15.6 kips

4
a = dedge= 1.78 > 2(0.875)
use a = 2(0.875)

--.' ( 70(1.75)(0.875)'

Example 8-11 A stairway hanger rod is attached to the lower flange of a

W410 x 59.5 floor beam. The load to be camed by the hanger rod is
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-toflange stiffener.

SOLUTION
Design the rod:

1.75 in

+ 21(6:25)(0.875)'

By trial, obtain D = 22 mm (at 10 threads/25 mm), or

x l o T 3= 0.299 X 10-3 mi > 0.2583

=: I 1.1 kips

0.K.

The posslble moment at the bolt is

1 75
1l.l(t)=9.71<41.92in.kips
\

O.K.
--

--L--/

Since this value is only 300 lb/bolt over the allowable, take it as 0.K
WT12 x 47 with four 7/8-in-diameter A-325 bolts to the flange . Use a
of the
W33 X 201 beam.

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.
tb = 1.15 in (tnanee
- of W33 X 201)

t... = 0.715
in
- --w

Check the beam for the hanger force. Neglect the torsion produced bythis hanger rod on only one side of the beam flange, since it is located very
near the end of the beam. Take the hanger rod hole at the standard gage
distance of 88.9 mm for a W410 x 59.5 section. Also,
r, = 12.8 rnrn = t,
tw = 7.7 mm
bf = 178 mm
b=

89 - 7.7 - 1.5
2

39.9 mrn

By Eq. (8-13) and directly incorporating L F = 1.7, we obtain

= -.5 .F in

From which

250(12.8)'[0.389(1
=

+ 44.5/39.9)] + ( lo-')
1.7

19.8 kN

< 38.75

N.G.

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%

FJOL.TED AND RIVETED COhXECTI

STRUCI'UIUL STEEL DESIGN

8-8 CONNECTIONS SUBJECTEDTO

TENSION

COMBINED

SH

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)

MPa

(8-17a)

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
(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.

F:

93 - 0.221

3.

STRUCTURAL STEEL DESIGN

'?'-..,le8-7 Coefficients for allowable fastener stresses for combined shear

1

2.

'

tension cohnections
AISC
Bearing

c,

,-

a-astener

.; 502 grade 1
grade 2

c2

f ~ q

SI

30
38

207
260

26
55

180
380

c
3

fps

SI

1.3
1.3

23
29

158
200

1.8
1.8

20
44

138
303

Friction

AASHTO

cqa

c5

fps

SI

0.75
0.75

i(S

A.307
.". .325Nb

17.5

121

Set F,

FigurrF.8-12

303 MPa and find P based on shear limitations.

sin 40
Iv= lO(O.4908)
303
P

" Standard holes 1/16 in (or 1/5 mm) larger than nominal bolt diameter; see the AISC
~ ~ (Sec.
u l1-6.3) for slotted or oversize holes.

ILL

" N,'threads In shear plane.

For bolt diameters larger than If in (38 mm); F , = 90 ksi (620 MPa).
Only static tensile loads allowed.

For bearing-type connections AASHTO uses:

Check P

P=1941>>420kN
420 kN.
=

wi.~:..fv and f , are the actual computed stresses and C, is a coefficient from
Table 8-7.
The AREA currently has no provisions for fasteners in combined shear and
ter:sii:n.

Example 8-12 Given the tension-shear connection of Fig. E8-12,~whatis th

allowable load P for the W T to column connection using the AISC specif
:ations, F, = 250 MPa steel, and 25-mm-diameter A-325 bolts?
SOLUTIONAssume a bearing-type joint with threads included in shear plan
F/ = C , - C2f,I C,
From Table 8-7 (A-325N). obtain

0.13094P

380 - 1.4(0.13094P)
420 kN

Find P based on maximum tension stress.

P cos 40"
F, = 303 =
1O(0.4908)

- X, threads excluded from any shear plane.

*

85 MPa

F,' = 380 - 1.4(85) = 260

<

useP=420kN

< 303 b1Pa

0.K.
* ,.
,:*

4
fl

0.K.

150 hlPa of Table 8-1

//I

..

) CONNECTIONS

o c are commonly used building connections, with

e of Figs. 8-160 and 8-180 and b used for rigid (or AISC tlpc 1) connections
g mechanical fasteners. Figure 8-18b (see also Fig. 8 - l j a and b) is cononly used for both simple (type 2) and semirigid (type 3) connections. The
esignations "simple" and "semirigid" for this connection are determined to
me degree by the thickness of the top clip angle. For simple connections this
imited to 1/4 in (or 6.3 mm), so that the an& can
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):

0.5(0.6FV)X 1.46 = 109.5 kN . m

T = -M
= - - 109.5 - 239.6 kN
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

Distance xo = L / 2 w ~ t hsolne
XO

des~gners
= N / 2 w ~ t hsome

= 3.25

designers

F i N 8-24) Critical sections and dimensions for clip (top) and scat angle d e s i a for use in
beams to columns.

use four for symmetry

frarmng

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-diameter 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
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

SOLUTIONW460

74.4 data:
=

Assume that web angles will be used to c a q shear so that the seat
angle carries only compression due to moment and the clip angle carries
only tension.

Fb

0.755

Fb

0.75(250) = 187.5 MPa

0.01643 x

lo-'

m'

due to type of connection, location, and due to rectangular

shape of cross section

402 STRUCTURAL STEEL DESIGN

Since this 'is thickest angle in this group, go to the top of the next
since thickness controls. We would not make L > 205 mm, b
extend.outside the column flange and also because the ben&
b i ~ o m equestionable with a very long width and only two fastener
TV L203 X 102 x 25.4 mm.
a = 50 - 25.4 = 24.6 mm
0.0246
M = (239.6)= 2.947 kN . m
2
S = 200(0.0254)2 = 0.02 15 10- 3 m3
6
2.947
f = 137 < 187 MPa
0.K. (and no angle th
-0.0215
A routine check for tension shows the section to be adequate. ~~t us
see if it necessary to use a web angle for shear, since this angl
thick. Refer to Fig. E8-13d for critical dimensions and other da
analysis.
R = 275 k~

ole cames one-half of

an=

This is not :atisfactory even if we assumed the

shear; therefore, web shear angles are required.

///"

~h~ moment connection shown in Fig. 8- 1 8 may

~ be treated in one of two

1. Assuming initial tension in the bolts (which is alivays developed with

strength bolt connections).
2. Assuming no initial tension.
.."

hi@~h~~~ two assumptions are illustrated in Figs. 8-21 and 8-22.

strength bolts the assumption of initial bolt tension allows the connection to
bending
as an elastic unit and the stresses can be calculated using the
moment equation, fb = M ~ / Iwhich
, is valid up until the cOnnection plates
separate
int is never designed for a moment large eno"@
desi@
d shown in Fig. 8-21 is adequately conservative
ry to design or a n a l ~ z ea j o i n t

1.4
3.

,,"

.'....,, .k
.I/

j; = -

-- 0 ~'I;I:\

lL

---

,.'p,lrl

LIL!
j;~)t
[t

Figure E8-13d

With 12-mm standard beam end clearance, R =

ity of

v acts

203 - 12
= 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
in the seat angle. Assume that the reaction is concen
from the end of the beam for computiilg the moment.
e =

(N+k)t,Fb=R

k = 27.8 mm
(N
N = 135.2 mm
e = - =135.2 67.6 mm
2

Re' = 275(0.0415)

ection with initial bolt tension assumed.

[Eq.(4-5)]
tw = 9.0

+ 27.8)(0.009)(187.5)

= 275
(.OITI~!II:

1:

.I[,
or bolt> t: = --

col:~p!l[e I L

('he.h I n r < r l ~ r l o ne ~ l u ~ [ i o n

e ' = e + 12 - kangle=67.6
=

--

+ 12 - 38.1 =41.5mm

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
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.

too two bolts:

Fi

8 2 5k
+I

&

( "f) i

17.5 1 - -

17.5 1 -

= 11.08

);I

<

13.7 ksi

N.G.

It will be necessary to redesign the connection using either larger bolts or

more bolts.
///

1.?0"+

* . .

1.5"

-10 LOAD RESIS'I'ANCF: FAC'I'OK DESI(;N (1,KFD) FOR

ONNECTIONS

V)

%'
Q'

1.5"

Figure ES-14a

SOLUTION
Assume a friction-type joint. For 7/8-in-diameter bolts, T, = 39
kips/bolt (Table 8-2).

--

'

2-32 ksi

15(11.20)
82.5

= 13.72 ksi

= lO(0.601)

11.20(15)~
= 420.0 in3
6

S = -bh2
=
6

M = 82.5(1.47

+ 3.95) = 447.15 in . kips

f , = f =-=--

447'15
420

- 1.06 ksi

The resulting stress diagram is shown in Fig. E8-146. The hatched part pf
the M / S diagram is tension due to moment, which must be carried by the

a current form of the LRFD equation for connection design using A-325 and
-490 high-strength bolts (and with size limited to diameter 2 1.5 in) is
R, = 1.1(1.1D + 1.4L)
is value is compared to the fastener resistance or plate-to-bolt bearing as
Bearing-type connection:
R, = 0.625A,FU
Friction-type connection:
R, = 0.7rny A,Fu
Combined shear and tension:
2

0.70

+ = 1.00

(pdhear))+ (0.6~u[iens,uni)' +(0.6"b~")'

Plate bearing on bolt:
Rn=3tdF,,,, ,,,,
+=0.&

here Ab=
A,=
Fu=
m=
p=

0.75

nominal area of bolt

bolt tension area (in thread zone) [see Eq. (5-4)]
ultimate tensile strength (nominal) of bolt
number of slip surfaces ( = 1 for single shear lap joints)
coefficient of friction (commonly 0.35 for clean mill scale)

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
bolts using a friction-type connection, and a bearing-type connection.

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

ased on the final decision as to connection type, "friction" or '

+ 1.4L)
= 1.1[1.1(45) + 1.4(75)] = 170kN

R, = I.l(l.10

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
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'
surface).
A, = 0.7854(0 -

9743)'
(fps)
n

Figure P 8-1

m i n e the number of 22-mm-diameter A 4 9 0 bolts needed to develop the full effscti3-eT

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.

[Eq. (5-4)]

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):

use two bolts

Check the plate bearing using two bolts:
R, = 3tdF,

qRn
I

= 0.64

F, = 400 MPa

= 0.64(3)(0-012 x 20)(400)(2) = 368.6 kN

\
L as rzq'd.

+R, = +(0.625AbF,) N = 170 kN

>> 170 kN

O.K.

Figure P8-2

Redo Rob. 8-2 using F, = 250 MPa and the AASHTO spzcificarions for the full
pacity of the angles. Use A-325 bolts and a friction connection.
Answer: 18 bolts, L = 610 rnm just under L for no reduction.
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.
Answer: Eight 7/8-in-diameter bolts.

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

&

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.
I

are adequate for tearing but that the bearing should be checked as appropriate. .UI boIt hes
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
8 x 50 section.

the

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
and F, = 250 MPa for the T section.
the AISC specif~cat~ons
Answer. 10 bolts.

Figure P8-7

Answer: P,

217 kips, N

six f -in bolts.

7 Do Prob. 8-2 using LRFD.

18 Do Prob. 8-5 using LRFD. Assume that the 225-k?J load is mads up of D
= 125 kN. Check only for number of bolts.
Answer: 16 bolts with I percent overstress.

ICO kN acd

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 simuItaeously 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

material at the joint when cooling takes place. If the qua

electrode 'is small relative to the thickness of the joined parts, the proce
to be unreliable (i.e., insufficient melting of the base metal occurs so
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
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
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
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

Electrode covering
Weld f ~ l l e rrneral
Shielding atmosphere

Molten weld pool

Ih

s:Fcctj
develop the gas shield and to obtain any
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 Figrial 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 platesthe order of 20 to 450 mm can be welded by this process in one pass-

9-3.1 Allowable Weld Stresses

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 9allowable 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

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

cj

E E
I

computed as

T = D X cos 45" =

I
D ~"0.70711

where T= throat dimension

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.

%
8

:32 u3
.L

0 .2

$*,a%g %

m m m

-.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

9-5 AISC

for plug and slot welds.

(0)

Slot

(b)

welds.

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

here

Part

( 5 1/4
f>1/4

SI, mm'

AISC and AASHTO

t i 6
t>6

Use D = thickness of part r

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
b

D 5 3/8 in
D > 3/8 in

x 0.7071 1 (in)
use throat = D x 0.7071 1 + 0.1 1 (in)
use throat = D

If this is contemplated, care must be taken to differentiate between the shop

[,,,I(

Pz F ~ )

p, =0.3, 0.27, etc. from the appropriate code for

p,= 0.33, 0.4, etc. from appropriate code for the

""'
rod

---a

-.-"<<,A,

1v 5 8"
or betweet,

A I S ( ' S c c . I . I 7.4
Flat bar ro gusset plate

Improieil

Weld both

loin[ p l l l l

llc)l

'i17rl11~l

I O rollins .I\!,

Ii.1

at a welded joint. Also shown 1s one solution lo

of a lamella tear occufing. ( a ) Lamella tear in base metal. (6) Joint %eomstV

the
reduce

A[SC Set. 1. I 7 . 6
Lap joints

a 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 ~ ~ ~ butt weld tends to increase the possibility of tearing (i.e.. from using an
with A-36 base metal). A thin, stiffened
~f

Figure 96

Stnrcturai

AISC welded connection specifications. sine

Code, they should be generally followed for

Other AISC specifications conc

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

must be large relative strains in the base metal (lamella

not
does
in the weld metal). These straihs occur where large
localked
stre:& occur.
. 2:: Lohc:;
is generally perpendicular to the mill rolling
direction that produced
the '"lnber being welded. Beams welded (0
flanges
produce this type
loading in the column flanges p u t not in
the beam flange).
3.
:nust be strain restraint in the base metal,

to the column flange produce restraint.

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 Enineering Journal, vol. 3, 1973) discusses lamella tearing in some detail and gives
a number joint alternatives that may be used to reduce lamella

5 ORIENTATION OF WELDS
aboratory tests on small to medium-size joints show that butt welds
limit joint capacity where the electrode has been "matched" to the base
~h~ orientation of the applied stresses does not have a si~dificanteffect
butt joint strength.
~h~ orientation of stresses for fillet welds is a significant factor
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'

Figure E9- 1a

Joint efficiency = 100 percent

9-6 - WELDED CONNECTIONS

0.015(150)(0.6 F")= 337.5 kN

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 FigSet. 1-17*7satisfies 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

of f b is compared to the allowable weld shear stress.

Figure 9-8 Welded Imment connections. ( a ) Butt-welded moment come

connection.

side'?

a 3/gin
ample 9-2 ~~~i~~ the welds for connecting an L4 3i
20. Use the AISC specifications7 'I8
r static loading and dbnamic loading-

A,,@, = 1.8 1 in'

p = 1.81(22) = 39.8 kips

'-

- ----

"CI'U-L.

Ufi,.,,UN

.,,UO,.

Substituting L,

,.

"A,.-.,.

..

0.707 1 1 0(0.3 Fu)5 to( 0 . 4 ~ )

D 5

0.25(0.4 x 36) - 3.6

-0.70711X21
14.84

Use D = 3/16 in.

39.8
(0.707 1 1 X 0.1875)(21)
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

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

heck: 6.62

10.5 - L , into Eq. ( a ) yields

+ 3.89 + 4 = 14.5 in.

.1 Rigid Beam-Column Connections

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:
dc

<

100r2\/1;

(fps)

Pbf
7

dc

10.73t:\jF,

<

(SO

Pb/

(9- 1m)

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
yield stresk of kglur~fiSee 9: sl. or MPa

c=

*
,Q

tTpTi:r

ner is required oppofite the"co+ression flange if the actual column

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.

as

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
'A,, =

The earlier AISC specifications rounded 183 to 180. Th

substitution of values

;+

,"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

Pb, - ~,,t,(tb/ + 5kc)

(9-2)

F,,,

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
t excessive deformation. A yield line analysis his given the

,
I

"

i-'

~ U C I T J R N . STEEI;
DESIGN

wELUE;

'.a""r,'.',glq:
'
.
.

.W

j ~. ;. , .: ..,'*"
.

CONNECI~&$~
(
:.<

. . .

. '

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

"

, . 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.
6 siiffeners for tension requirements must be welded to the column fla

*a,,)!

"

dfficient to carry A,tF,, (i.e., use full penetration butt welds).

:ompression stzfeners must be welded or accurately fitted to the
opposite the beam flange delivering the compression load.

compute the plate length. Note that the plate will have to be long
enough to allow placing 5 in of weld at D = 0.5 in. I t will also have
the
some length between the end of the beam welds and the butt
column in order to develop adequate strain. BY proportion:

% 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.
.,

, , ..

.,

1 :

. ,.. ...

b E9-3a

b o ~ u m oDesigning
~
for full moment capacity:
M = FbS, = 24
I

64.7 = 1552.8 in . laps

T = C=-=

'd

-=

16.01

" ..

',Se a fop plate that is 8 in wide at the column weld and tapered to 3 in,
zhown in Fig. E9-36.

'P

Figure E9-3c

'

97

= --- = 0.73
22 X 6

26.4

use 3/ 16-in plate

= 0.1 j
12x0.4FY
26.4
use3/16in
Weld D = 0.70711 X 12 X 21 =0.14gin
check if the column needs reinforcing (AISC Set. '-I5.'):
t =
p

use 3/4in plate

9 -,TakeD for fillet welds at 1/2 in.

P,

= 0.7071 1(0.5)(21) = 7.42 kips/in

& = - =97

7.42

13 in and distribute as shown

97( 5/3) - 36(0.295)[0.505 + 5 ( 1.2j)] - 161.7 - 71-74 = 2.49 b

36

Therefore, a pair of stiffener plates are required opposite the tension flangeu 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-

Use a f -in

e,

(16.91 X 2)(0.70711 x 21)

on both top and bottom of each stiffener plate.

flange for
der web), the 3-in leg should be welded to the beam and bolts in the 44x3 Q-sJthat there will be adequate erecti~nclearance and edge distanceangle is welded to the beam web and field-bolted to the colu-

% = (R:oment + R : ~ ~ ~ ~ ) ~ ' ~
where
:.:\

a:

...

.i

= lorceor stress/unit df length (or area). The value of 4 is lifited

-< Rallowable

to

in
~~~~~l~ 9-4 Design the web ande connections for a W18 x 55 as
framed
beam-to-column
flange
connection
cqi
~ i ~ g~- 4 .afor a
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.

o ~ u n o NFor the beam web weld, locate neutral aGs as

(12

+2

2.5
x 2.5)s = 2 C 2 . 5 ) ~

.X = 0.367 in

4-

- L K U C T U WSTEEL DESIGN

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

7 he weld shear resistance

.2 Welded Beam Seat Angles

27.5

, r , - Y = 1.617 kips/in
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

4. = ((0.465 +

5&&,

:ompu t.ing the

odulus S = b

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
t
9 mm) to obtain

e.1n111al
.. =2
b

R,=-=-=
R
L 27'5
12

2.29 kips/in

+ 0.5 - ( t + 0.375)

einlt~al
. . . = - 2+ 1 2 - ( t + 9 )

in
rnrn

(and neglecting end returns of 2 0 )

- .

3.38
D =
0.70711 x 21 = 0.227 in

la1 bending str

the required

use 3/ 16-in weld

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??)
54Pe,
Rh =
(from Fig. 9-1 1, refer also to Fig E9-46)
25 L~

2.46
0.7071 1 X 0.3 X 70 = 0.165 in
Check the web shear capacity:

O.K.

e beam seat angle shown in Fig. 9-12 must be designed for bending stability
must be of sufficient thickness and leg length that an adequate fillet weld
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

1.617)~ 1.308~)"~
= (6.b45)'I2 = 2.46 kips/in

D=

=D

+*

0.257 in

Use L3 x 3 x 5/16 x 12 in long.

& = -R-

I 23 + 2(2.5)(6)2 + 12(0.367)' + 2(0.367~+ 2.133~)

LP = Zx + 1
, =12
3
= 144 + 180 + 1.62 + 6.49 = 332.1 in4/width of weld

use 1/4 in weld

actual required angle thickness t is very sensitive to the k value, so the

ative angle should always be accurately checked.

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
of the WT at least as large as f+ the

Use the
Also,

- < - = 335
21.2'
t - m
an inspection of Table V- 18 (S
b, = 395.4 mm

r/

Figure E 9 d b

Shear Stress due to load:

= 30.2 mm

Find the maximum weld size D for s

on each side.
loot,
=
= 9.19
205.8
Check the bending stress in the WT

fb=--

6M
twd2

-.

.,,.

,..

6(15.16)
= 132.9 MPa
18.9(0.19022)

870

0.842

> 8 as minimum for r/

-= 0.3 x 485 = 145.5

< 150

...~.,.
y=--=
50 040

15.16

L=~=o.lsoo,=~,

57.5 mm

0.~.

Use D = 9 mm.

/li

WELDED COLUMN BASE PLATES

= 2.4733 D, x 10-6 m4

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

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.

STRUCTURAL STEEL DESIGN

Beam: W610 X 241.1

d = 635 m m , b, = 329 m m
l / = 3 1 mm,r,= 17.9mm

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
column compression flange) only has to be one-half of h a m depth, since
the load is only on one side.
33.3

tc/

Is,,"

> T = -7(0.67

Column: W360 X 261.9

SOLUTION
Step I. Design vertical stiffeners 1 and 2.
For stiffener I, opposite the tension flange of the column:
610
:
f b ' - - M --- 132.6 MPa
S,
4.60
AfC = 6,$ = 0.398 X 33.3 = 13.2534 x 10-

G c tcw tbf

17 m m

329)
2

use tJt = 20 m m

+ 17.9 = 1 19.2 rnrn

use b,,

120 mm

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

Since the furnished t,,

diagonal stiffener.
8

Pw = ; f b ~ =
/ j(132.6 X 13.2534) = 2929 kN
'bf

bst,ff =

d Z 3 8 7 mm, b f = 3 9 8 mm

As, 2

WELDED C

17.9 < 31.6 mm required, it is necessary to use a

635
-- tan-' = 58.64'
387

cos 8

0.52042

= k,

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
2929 - 250 X 0.0179(33.3
250
= 10.146 X
m2

As, =

- 2.5 x 48.4)

6.2615 x l o p 3m2

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)

Step 3. Design the welds.
The weld for plate 1:

t, = 28mm
A , = 0.028 X 398 = 1 1.14 x lo-' m2

0.K.
For stiffener 2, opposite the compression flange of the column:
db(,+,.s, = d - 2k = 635 - 2(48.4) = 538.2 mm
A stiffener is required if bb > dbw:

, , ,..

.,.,..,...

dbw

>
>

F,

0.3 Fu x 1.7 = 247 MPa (plastic design)

Using inside fillet welds (see Fig. E9-9b), we obtain D,, = 8 mm, since
= 25 > 20 mrn. Check the effective D of the beam web for shear:

t,

10,73t & f i
pbf

PbJ = 2929 kN

10.73(17.9)~a
= 332.2 mm
2929

So only 7.4 mm of the weld on the beam web is effective, since the shszr of
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:

pWe,
= 2(600)(0.0074)(0.70711)(247) = 1550.9 kN >> 600

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
web and

for plate 2: with maximum

weld D = 7.40 nM in
required for 20-mm stiffener plates and four lengths

= 250(0.0179)[33.3
Pweld

+ 2(20 + 48.4)]

= 761.2 kN

"- P b f - PWeb
= 2929 - 761.2 = 2167.8 kN

4L x 0.0074 x 0.707 11 X 247 = 2 167.8 kN

Weld for diagonal stiffener:

allowable for the base metal.

r: p = 8 1 kips, L, = 12 in, using D = 9 / 16.

.
WELDED C O W E

9 ,

,,

W410

..

84.8 beams frame into a W530 x 123.5 girder as in Fig. P9-19 in

trcdes, Fy = 250
and

'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
1:

--

hear due to uniform load is

steel, A-325 bolts, and E70
A

9":

'

YO specifications.
.~.moer:1, = 5/16; D = 1/4 in.
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

;. design sketch. Use E70 electr
fram. ::-;a W12 x 53 column.
9-17 J-ai@ a stiffened b e ~ mseat using a WT fo

. neat

':

.: .

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

.
PLATE GIRDERS

4g
,,

."+
i

,,

..,

( C)

epth D and clear distance hatio h,rb is a significant parameter in plate girder design. (a) Welded plate girder- ( b )
in drsim).,( Girder elevation (side view) illusirating other
plate

the reduction7 or
and larger clear web

rn.ical for spans UP

15 f t ( 5 m) Or morePan (D/Q ratio, Ge
of
to
Larger

A-

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 confisrationough 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-

460

STRUCTURAL STEEL DESIGN

However, the current AREA specifications do not

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.
An inspection of Figs. 10-1 to 10-3
sel-vice readily indicates that there is n
problem. Within the constraints of D / L
possibilities exist for cross-sectional area. If we define optimization o f a pl
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
and fabrication cost to produce as economical (
effort, which is also a cost factor) a solution as possible.
20-2 LOADS
plate girder loads are obtained similarly to those for be
building design the loads are often moving, as for crane
industrial buildings or warehouses. AISC specif
factor for these types of loadings.
Plate girder design for highway bridges involves a
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
lpading to find 'the maximum moment and
specifications allow use of only one truck per lane for simple spans, in general,
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
\

s [ ~ i r ~ Jt
. lr
(t
Jl c
o ri I J I I C

I o ~ I J\$kith
I~~

I O J ~or

L~~~ = .![her

j r ~ n d ~ truck
r d

because of truck length factors and positioning of wheels. When the 'pan is
a line
long, the lane loading with single concentrated load
tvine
to find lhc
enough
and
is
more
convenient
to
use
than
of trucks
number of sucks and their positioning for maximum stress

computer program.
me
ln all girder design the standard Iane/tmck loads are incrcased
,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
lin2'
girders. The designer must make a similar type of 3 n n l ~ for
~i~
using more than two girders.
Example 10-1 what are the design moment and shear values for the *d'r!
of a highway bridge as shown in Fig. 10-5?
HS 20 loading, s = 19 it, w = 28 ft, span = 110 f t - D a d load due "
deck, sidewalk, and so on = 3.1 kips/ft.
S o L u n o ~nelane factor L, is obtained by 9"1 about the left

X+ Y
Lf = S

where X = S / 2 - 5 = 19/2 - 5 = 4.5 f t

y= x + ~ / =2 4.5 + 2812 = 18.5 f t

aderFig

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
ine with apex at the centerline for a larger number of wheels on the span.
given in Table 1-2 are the end reactions (shear) and the maximum reactio
(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
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;
any other loading X is X/80 x table value.

The impact factor is computed (see Sec. 1-10) as

50

I =

L
The moments are

+ 125

50
110

+ 125 = 0.21 < 0.30

O.K.

Use the equations given in Sec. 1-9 for L = 110 < 145.6 ft.

10-3
FLANGES AND WEBS OF GIRDERS
- - - PROPORTIONING
- BUILT-UP SECTIONS

Substituting values yields

,

. -..

40
110
The shear values are

= ~ ( 5 9 1 6 . 8- 1232.0) = 1703.6 ft

The flange cover plate of a built-up section usmg a rolled section as the bas
may be proportioned
as follows (refer to Fig. 10-6). Assurmng that the maxim
- bending stress fb is the allowable value of Fb,the average flange value is abo
0.95Fb. At the junction of the cover plate and the rolled section flange the stres
s about 0.90Fb. Therefore, let
M, =
= O.90Fb.Sr

kips

Vd = W L = 3.1(110) = 170.5 kips

2
2
'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

We note that Md is at the centerline of the span, whereas ML is 2.33 ft

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,,,,,

Substitution of Eqs. (a) and ( b ) into (c) and taking (d + C) = (d + q) definin'g

the flange area A , = bit;, and taking the ratio 0.90/0.95 as approximately 1, we

+ impact + Md
= 1703.6(1.21)(1 + 0.21) + 4689 = 7183 ft . kips
= ML(Lf)

The design shear is

+ Vd
= 65.9(1.21)(1.21) + 170.5 = 267 kips

Vdeslgn
= VL(Lf)+ impact

///

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

(c)

Figure 1 0 4 Bending stress distribution on a built-up beam cross sccrlon

PLATE GIRDERS

Mtota~
- sx
( d + C)0.95Fb d + C
..
17 5" )u!d be evident that Eq. (10-1) does not give a uniqr.. r b ovluLlun a n a the final
f!$P.:;- cover plates must be investigated using f, = M ,/ , q.
A
reasonabl
m
e$\iliit' of C = 1 in or 25 -.
Af

ZX

--'--a:-

' , + ~ 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~ U I
irle conditions shown in
: EIO-20. Use the AISC specifications.

465

We could use flange cover plates 25 X 548 mm; however, let us arbitrarily
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
b; = 418 + 52 = 470 mm

t;=-=13'82
0.47

29.4 rnm

use 30 mm

&'-

Now check if cover plates 30 x 470 mm are adequate:

I, = I,

r:;t

5..

Data for a W920 x 342.3:

c'JTION

wt

3.36 kN/in

d = 912 rnm

12 499.4 x
The section modulus is
=

M L - . . .wL2
-=
8
Md "

72.6(20)'
8

.,

f b = 2 5 .3855
719=

Mdes~gn

=3800 kN . m

3800
0.025)(0.95

x 0.6F,)

- "'I. =

rrFh,

20

- 1

IS

ill

---I

<

150

O.K.

13.72
0.912 + 0.025

"

by- 418

t;

,'

726 KC 1

'

w- 1)J

149.9 MPa

?'he cover plates will be welded as shown in Fig. E10-2b. This requires
checking by/$ for acompliance with AISC Sec. 1-9.2.2 (stiffened edge
element) :

72.6 kN11n

i\

- 1

ruqina
Fn i ~ 1n1 1 :.
------a
-y,
(1

(0.912

.m

= 17OkN.b

.... ...

A, =

i-

= 3630 kN

3.36(20)~
8

Te~t~-:l"ely,the area of each cover plate

, ,..

m4

The additional moment due to the flange cover plates is proportional to the
flange area:

,'noments are:

is?

+2~d'

30

13.93 << 42.4

It is not necessary to check b/29 of beam since

plates 30 x 470 mm.

Figure E10-26

O.K.

9 = 32.0

mm. Use cover

///

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.

I
1

v$ =

vQ
I

(kips/in or kN/m)

where us = force/length of developed shear

Q = statical moment of cover plate with respect to neutral axis
I = moment of inertia of section, including cover plates
AlSC Sei 1 1 0 4

AASIlTO Set 1 7 12

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
Q, I= terms previously defined

F,

I*
Example 10-3 What 1s the theoretical cutoff dlstance for the cover-plat
beam of Example 10-2? Also, design the weldlng to fasten the ~ l a t e sto the
beam using E60 electrodes

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,,,,

fps 3 d b + 3 ft
SI 3 d b + 0 9 1 ~n

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

Figure E10-3a

a*

Since the moment diagram is a ~ a f a b o l awith slope

equation for moment IS
,+fx

M, - k,~'

0 at midspan,

" .
ddS STRUCTURAL.
S'I'EBL DESlG ii

,'
x =
;i :

:$
9

'8

?.

4% x 2
M, - L
With M, = M, = 2058 kN . m at the distance x from midspan where
beam capacity is adequate without cover
the theoretical cut
is
M,

g
:
.

L/2.wehave
=

,, :,,, v :,

(3855 - 2058)
[
The theoretical length of the cover plate
x =

,:

= 6.828 m
]'I2
= 2x = 2(6.865) = 13.73 m.
470 rnm; I, = 12 499.4 x

3855
L2

l ' o t ~ lengrh
i
CP= I i . l i

Design the welds (cover plate 30 x

at the cutoff point x' = 10 - 6.83 = 3.17 m.
Shear V

771 - 77.1(3.17)

= AJ = 0.030(0.470)

nl-

1,=:0~,--

r:=;O;iirn

530 k ~

(Y

- + 15

/I/
= 6.41

m3

10-5 GENERAL PROPOR1[1IONS OF PLATE GIRDERS

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

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

follows. The moment camed by the flange is

The moment camed by the web is
t,h2

-Pwm-- 2(0.008 X 0.7071 1)(124.5) = 1.4085

Tentatively, use a 40-mm length of weld (40 mm = 5D
carry

Pi

iVf, =

> 2D),

which will

flSrw= f;7

The total moment is

M

IMJ

.\/,

= 40(1.4085) = 56.34 kN

This corresponds to a distance along the beam of

56.34 X lo3
= 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
S =

'

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-

hd,
..

472 STRUC.IUK~.L y

i b c ~
L)L!SII,~J

Figure 10-10 illustrates the assumed girder web plate loading on a strip
dx(h). From Chap. 3 the critical buckling stress was found to be
$!
.,-,,"
~,T~E
Fcr =
12(1 - p2)(h/t)2
..l
3
I

,,&

e AISC rounds this to obtain the limiting h / t ratio as

14 000
h
fps - I
[~,(~~+16.j)]'/'

\' '

SI:

where the ratiogh/t has been substituted for b / t as given in Eq. (3-5)
The force exerted on the web of the web segment shown in Fig. 10-lq$
A f f sin @

r A,fb

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
f =er+ey=-

&

t and the angle @:is

Fr +
E

97 100

-5

[F,(F,

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

F,
.

.
I

+=

d~
&-

The maximuq h / t ratios for several grades of steel are as follows:

2 ( F r + F , ) dx

fi/2

+ 114)11/*

At web yield the applied force is

..
2

, ..

.,. .. , .. .
.

lor small angles and

A d tan

AJffl

but

+ in radians, giving
Afffl

MPa

dx
2F,(F, + F')Eh

h
t

<

1.5

259
248

kcT2E

( L ) 2 ( t )dx
12(1 - p2) h

Equating Eqs. ( a ) and (b) and solving for h / t , we obtain

-=[

a/h

333
284

This value'should not be larger than the critical web buckling force
FCrtdx=

[by Eq. (10-4)]

kCnZE A,
24(1 - p2)

F,(< + F,)

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.-ently being',used for all values o f . steel F,. It I is! also',assumed 1
.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 ) ] ' / ~

Experimental studies on full-size girders have shown that the web

the compression zone deflects laterally by small amounts at very early
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
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 expressed in terms of A,, A,, h / t , and Fy. A possible equation in terms of ultimate
moment is

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,
"Strength of Plate Girders in Bending," Journal of Structural Division, ASCE,
August 19611, where F,, is the critical web buckling stress. If we replace
F,. = 1.65Fb,we obtain (with slight rounding):

T
l ii I

i bl

Inspection of Eq. (10-5) indicates that if

girder to Pratt truss. ( a ) R a t [ truss ( - )

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

vu =

Vb +

(b)

the van ~Misescriteria is

v"j

(4

e obtain A , and substitution into ( b ) ,

Vb =

fi (V,)Fr,

. g ~ , , , the value of FA is, from Eq. (3-5)-

( e>

here

(3-5)
~ q (10-6),
.
the critical stress is taken as the mean of
q. (10-6) and the value of 0.8Fy,, to obtain
(10-7)

v,

The beam shear contribution is

cornpression

FYs = 5

10-6.3 Shear and Stiffener Requirements

Where the l i l t ratio is sufficiently small, web buckling will not occur under
shear before shear yielding occurs. Actually, a beam web as part of a flexural
member subjected to a bending moment is canying the shear in a "tension field"
mode. Some researches have likened this to,a Pratt truss, where the stiffeners (if
uszd) are the compresbion members and the web segment between a Pair of
stiffeners is the tension element, as illustrated in Fig. 10-11a. Actual girder tests
carried to yield so that mill scale forms stress lines displays a stress pattern
similar to that idealized in Fig. 10-11b.
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
stresses are reached; then the additional shear is carried by tension field action.
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

d where as previously defined (see Table 3-21;

!
V; = Fc,(htw) = FcrAw
ultimate shear force capacity of a girder for web plastification (plastic
capacity) is

., y,

v-

= -ysAw

k = 4.00

+(10-8)

P L i T E GLRl

-t:;d

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:
For C, 2 0.8:

The AISC uses

''

10-6.4 Combined Bending and Shear (Interaction Check)

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
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. (10ust 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

Substitution for f,, $, and A,, = ~',/F,II,, gives

) .

19'A
i*";,:

<'..
i

are always required.

2. Intermediate transverse Stiffeners (AISC does not have a specification for
longitudinal stiffmx'S) are not required if bottf~of the following
are
met:

Intermediate stiffeners are required for any other shear stress condition.
When stiffeners are required, the spacing is limited to
:J

but a / h I 3.0.

[ I + (o/h)']'/'

YDht,

(10.14)

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

D = 1.0 for stiffener plates on both sides of web plate

D = 1.8 for angle used as stiffener on one side of web
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
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

$4

:.
.'*

inei:.~cr ust be at least4 ,$

AlSC a n d A R E A

.A.iStlTO

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

P, =
r 0.015~,h~<
"LJsiui ?, = F,/ E, a safety factor of 1.67, and assuming that the stiffener force is
debe1;qxd in the distance h/3 from the compression flange, we obtain
.B;

ha&u
tips

1.67

D o r fillet

h some additi*
obtained as

.I

hfr

C o m p u t e L'lr
L 1 = O75h
r = [11'4~"~

Cornpure L r

h r

~ ~ ( 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;

1096.f

in or m.
The allowable bearing stress F,,, = 0.90FY (AISC Sec. 1-5.1.j.1) based on the
lesser F, if the flange and stiffener are of different y~eldgrades of steel.

w i n g Stiffeners
b

Beanr,, tdfeners are always required in pairs over the reactions. Bearing
stiffece-s mgy be required beneath concentrated loads carried by plate girders.
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
as to extend approximately to the outer edges of the flange or angles.
Bear'ng stiffeners are designed as columns with an area that includes the
stiff n ~ r sand a central weblarea of 12twfor end and 25tw for interior bearing
?
stiffeners (see Fig. 10-13). This area is used for computing the radius of gyration
andrfo. checking the coludn stresses. The effective length of the stiffener may
be taken as 0.75h because of being securely connected to the web.
The effective bearing area A: is taken as the area outside the flange angle
'ffiet oi ihe flange-to-web
&~ds.
design requires co
ing L r / r to find the allowable column stress Fa

$$@$$:eking

<

PY
$4

fif

P = AF, I applied load or reaction

Al5a.check using the effective bearing area A, = bits (Fig. 10-13) to obtain
6

= AbFbrg I applied load or reaction

:i r

10-6.7 Web Crippling

Webs of plate girders are required to be proportioned so that web crippling
(same phenomena as for rolled beams) does not occur. This 1s accounted for at
reactions by bearing stiffeners. Where the compression flange supports a uniform load or concentrate
ds for which bearing stiffeners may not be
requirgd2
compressive
delivered by the flange to the girder web must
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
Basler, '.'New Provisions for Plate Glrder Design,'' Proceedings 1961 AISC
National Conference, AISC,, New York), SF = 2 65. E = 19 000. and some
rounding, to obtain the allowable web compressive stress as
10 OOOk,
fps: F =
(hllw12

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 cequivalent 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
f l n n a ~st, the ends and at the
will assume lateral bracing" of the comnressinn
- --r-------I- - concentrated loads. Other design data: E60 electrodes,
.
---- ,n,..~.,,,r,,i,.f i r tai n n c
ATSC
s,
A-36 steel, and girder depth limited to 84 in.

Figure ElO4b

90k

FO;

h/r

For

t, = 0.5 in, h/t = 78/0.5 = 156.

Step 2. Make a preliminary flange design.

162, obtain t

78/162
h
Maximum - =
=

0.48 in. Try tentative t,,,

14 000
= 322

0.5 in.

105'

--

21 ksi and an average distance to the center of the flange

(Assume that fb
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

2R

Step 3 . Compute the actual moment of inertia and section modulus of

the trial section and revise the dimensions as required.

= 343.4 kips

bh3
I,, = 12 + 2,4d2

Now draw shear and moment diagrams as shown in Fig. E10-4b.

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
h 760
--=-=

$8

760

-,It

162 (approximately)

(neglect I, of flange plates about cenlioid axis

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

ilnce the moment due to other loads is slightly off center) is

7-

0.24(53)~
= 84.3 ft - kips
8
The total bending moment = 2722 84.3 = 2806.3 ft . kips.
- M 2806.3f12'1
JbC-=:
19.53 k s ~
S
1724.5

M=

-/

Since this is considerably less than 21 or 22 ksi,

let us tentatively revise the
web thickness 2, to $ in.

,
I

-w

-.

Check the web shear SO that the plate is not too thin (neglect beam weight at
j'lls pmnt):
1)
1 O(U.3

-. .-

13)

v.n.

step 4 R e c o ~ n ~ uIt eand&,

I

= 14 829.8

+ 48

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:
L = 18 f t (largest value), and m , = 0 and ibf, = 2664 f t . klps

992 = 6 822 in4

For the interior panel: C,

ft . kips):

1.0 (since Interior moment of 2722

> 2664

s, = 63-822

1600 ln3
39.875
0.490
New girder weight =
(3 1.5 29.3) = 0.207 kip/ft
144
'be approximate additional moment due to girder weight is

M=

For this L / r , ratio, F, = 0 . 6 5 = 22 ksi for both the end and interior
panels.
Now check AISC Eq. 1.10-5, since h / t w = 208 > 7 6 0 / a :

0.207(53)~

MtOw = 2722

= 72.7

+ 72.7 = 2795 ft . kips

2795(12)
"4, = 1600 = 20.96 ksi
Continue the design. Use two flange plates x 18 in and a web plate
3
y X 78.

==

2 1.05

> 10.96 ksi

O.K.

- At this point the bending stress and slider proportions are adequaie
unless a later interaction check requlres a revlsion of girder section.
Step 6. Compute the stiffener requirements.
-AISC specifications require bearing stiffeners under reactions and the
two concentrated column loads. In the end panel the actual shear stress at
the reaction is
-

h = 208
tw

= 79.75 m

SX=1600in3

A, = 29.3 in2
Af=15.75in2

f =-=-

A,

173.25 + 0.207(53/2)
- = 6.11 ksi
0.375(78)

According to AISC Sec. 1-10.5.3, intermediate stiffeners (other than the end

486 STRUCTURAL STEEL DESIGN

beanng stiffener) are not required if

-,E;.,,C.,"
h
- I L6U
and
j, f, FF,= t
2.89
We will provide bearing stiffeners under the column loads, which leaves
clear distance in the 18-ft end panel of 18 x 12 = 216 in.

been obtained from Table 11-8 of SSDD but requires double i

Step 7. Check the interaction at concentrated loads [Eq. (10-13)
At dx to the right of the right column, the shear V is

V = 178.4 - 18(2.80 + 0.207) = 124.6 kips (including girder wei

For V = 124.6 kips, the web shear stress is
124.6
j" =
= 4.25 ksi
29.3

We have just found F, = 7.67 ksi in step 6.

I]

4'25 36
0.825 - 0.375 7.67

Assuming that C, will compute

C" =

45 OOOk - 45
q ( h / tJ2
36(20gL)
is limited

36(0.169)
= 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):

F" =

a
h

108
78

--- =

1.38 < 3

22.2

> 22 ksi

O.K.

Since the allowable stress in interaction is 22 ksi (maximum) and the a c d

and allowable values based on flange stability are less, the lesser values
control.
_Tb = 20.96 k s ~
F, = 21.05 ksi

< 0.8, we will try Eq. (10-1 1):

The allowable shear stress if no intermediate stiffeners are used

Step 8. Check the stiffener requirements of the 17.5-ft interior span

between columns.
The maximum shear is obtained coming from the left:

(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:

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,
With intermediate stiffeners and C,
Eq. (10-9):

< 1, we

can use AISC Eq. (1.10-2) or

Since 7.67 ksi is greater than the actual shear stress at the bearing stiffener,
6.1 1 ksi, and the shear is less at the interior points, it is not necessary to
check the shear stress further for stiffener analysis. Note that Fv could have

O.K.

no st~ffenzrs
required

Step 9. Check the web cnppllng under the compression tlange due ti
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
stress is

2.8

jc = 0.375(12)

= 0.622 ksi

STRUcrCnW, SIEliL DfiSICb*

The allowable compressive stress (checking at the location where a / h is

critical is

F = IO"*[5.5
( h j tWl2

1.4

> 0.622 h i

O.K.
208'
2.69'
Note that if F <f,, we would have to either add stiffeners to decr~asea
'or'inctease t,.
Step 10. Design bearing and intermediate stiffeners.
For bearing stiffeners at girder ends (we will use the same size
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
- - 8= 1 6 ~ 95
O.K.
t
0.5
=

0.215 and a / h = 1.38 from step 6.

[Eqk (10- 17) combined with kc]

- &(5.5

'

for web and stiffener); C,

In terms of A,, x 100/A, this value could have been obtained from TabIe
11-8 by using double interpolation. Try two bars 3/8 x 6 in:
A

2 ( : x 6)

4.5 1n2 > 3.02

O.K.

-16
O.K.
0.375
The minimum moment of inertia of the stiffeners is
-=--

fi

oS(16.35)~
= 182.9 in4
12
The "effective" column area for the radius of gyration (see Fig. 10-13):
I =

A = 16 x 0.5

+ 12(0.375)(0.375) = 9.69 in2

=dF
182.9

9.69

76.5 in

Step 11. Design welding to fasten the stiffeners to the web.

jU= h[(&y]l/'

- 18.45 > 20.98 ksi

O.K.

Check the bearing stress; assume a 5/16-in weld, so that the effective
bearing area = (8 - 0.3125)(0.5)(2) = 7.7 in2:

F,,, = 0.9G = 32.4 ksi

P, = 32.4(7.7)

kips/in

[ Eq. (10-16)]

= 4.34 in

L = 0.75h = 0.75(78) = 58.5 in

L
58.5
---- 13.5
r
4.34
From Table 11-5, the allowable column stress is Fa = 20.98 ksi. The
column stress is
fa=-----178'74

Use a stiffener plate length

h - 4 t , = 78 - 4(0.375)

= 249.5 >> 178.74 kips

O.K.
For intermediate stiffeners, with only one intermediate stiffener use two
plates.

For a plate on both sides of the beam web, D

1.0 and Y

1 (A-36 steel

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.
F,,, = 0.1875(0.70711)(0.3 X 60) = 2.39 kips/in
Use a 3/16-in fillet weld continuous for both bearing and intermediate
stiffeners.
Rationale: Few stiffeners and the weld can be made in one pass. I t 2
too difficult to measure and set up alternating weld distances and gaps.
Step 12. Design welding to fasten the flange plate to the web.
Check the end for maximum shear:

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 savingsularly if the weld can be made in one pass, as here.
Figure E10-4c illustrates the design summary for the girder.

490

STRUCTURAL STEEL DESIGN

9ok

105~
17 5'

AREA:
When cross-ties rest directly on the girder flange:

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

AND AREA

Z1

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
the girder will be subjected. In general, however, the same general considerations apply:
1. The girder is proportioned by the moment-of-inertia method.
2. No uniq'ue solution is possible.
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)

,,

In general, the basic allowable bending stress is

Fb 1 0.555
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
specification.
The b/2$ ratio for the flanges is also limited:
AASHTO:

AASHTO also uses Eq. (3-5):

and solving for h/tw, substituting k c

obtain

23.9, E = 29 000, and F ,

1.19Fb, we

'Xlis gives h/tw = 163 for A-36 steel.

The AREA limiting h/t, ratio is obtained from Eq. (3-5) by substitution of
kc = 23.9, E = 29 000, and Fcr = 0 . 6 5 to obtain (with slight rounding)
h
1030
2700
SI: - - fps: - = lw

fi

lw

For A-36 steel, the limiting h/t, = 171. When the value of fb
compression flange, AREA allows an increase in the h/t, ratio:

< Fb in

the

with the ratio (F,/ fb)'I2 I 2.0.

AASHTO and AREA limits the web thickness of fabricated plate girders to:
WSHTO:

Ths limiting 6/29

Alsg

= 12 for A-36 steel when fb =

ARE AREA, but for continuous girders.

Fb = 0.55Fy.

tw 2 5/ 16 in (8.0 mm)

AREA:

t, 2 0.335 in (8.5 rnm)

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
is based on Eq. (3-5) using kc = 129 (since a more substantial edge fixity is
?

."a

492 S T R U C ~ U I L A I:,. 1 ,XL u~.zl,.;:-r

8

..

PWTE GI&LRS

4 ~ 3

obtained) and F,,

'

1.64, to obtain

fps:

v3

'W

i 3 e limiting h / t w for A-36 steel; is 327.

fi

:he AASHTO requirement for stiffeners is also computed using E ~ (3-5)

.
and
the k factors given in Eq. (10-8) and with varying safety factors. If the
such that a / h -+
(no stiffeners), kc = 5.34 from ~ q (10-8),
.
and taking
F,,-2.5 f,, we obtain
h

234

h
618
-5tw
lw
The limiting h / t w = 68 for A-36 steel when f, = F, = 0 . 3 3 ~ ~ .
AREA requirements are slightly more conservative:
h
360
h
950'
fps: - 2 SI: - <lw
The limiting h / t w = 60 for A-36 steel.
Intermediate stiffeners are not required by either specificati& if the h / t w
ratio is less than that given by Eq. (10-20) or (10-21) and may not be required
for larger h / t w ratios as given in the following section.
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
-

<-

vx

S1:

ips:

332t

a = - (in)

fi

875r

SI: a=-----

fl

(mm)

In these equations t has units of in or mm. The stiffener spacing is limited to not
more than 72 in or 1800 mm when using Eq. (10-22).

fi

fi

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
from the bearing stiffener. Currently, the XASHTO specifications for intermediate stiffeners is somewhat more indirect and is given in Sec. 10-7.4 foI10wing under "Intermediate Stiffeners."
The AREA specifications reduce the 348 factor slightly, to obtain

10-73 Shear and Stiffener Requirements

fps:

.I_

From an approximate aqeraging of fi we obtain 348, which if used to back-conpute the SF gives 1.37 at a / h = 0.5 to 2.02 at a / h . 1.0. In earlier AASHTO
specifications the stiffener spacing could be computed as
348 t
914t
fps: a = - (in)
SI: a = - (mm)

h <3840
1455 ' S I : -5 -

"

>

10-7.4 Stiffener Design

.

Longitudinal stiffenem The AASHTO value for longitudinal s$feners in terms

of moment bf inertia (see Erickson and Eenam, "Application and Development
of AASHTO Specifications to Bridge Design," Journal of Structural Dit'inbn
ASCE July 1957) is
(10-23

IJ = 11t1.[24(%)2- 0.131

with a plate thickness limited to

b
71
fps: -5---

from which
=

kT ~ E
12(1-p2)SF

--

- 0-

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
a/h

0.5
0.6
0.7
0.8
0.9
1 .o

25.36
18.83
14.90
12.30
10.58
9.34

332.8
344.4
357.2
370.9
387.2
404.0

SI:

b
-

188

5 -----

<i

(1 0-24

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
l
negative-moment regions and the specifications are exactly the same as fo
AASHTO for I, and b/t.
Bearing stiffeners are always required for both AASHTO
Bearing
AREA girders. They are designed as columns using an area as shown in Fig
10-13. The allowable column stress depends on the L / r ratio computed a
shown on Fig. 10-13. The bearing stiffeners must be checked for bearing as we!
as acting as columns. The bearing area is as shown on Fig. 10-13
allowable bearing stress is
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
tw

and

f,5Fo

where
I

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

ti

& [ I+ ( h / a ) 2 ]
where C= -I I
~,(h/t,)~
'
B = 222 000 in fps units
= 1 520 000 in SI units

It is often convenient to plot the limits of these stresses as in Fig. 10-14.

When ~ m s v e r s estiffeners are required, the first stiffener is placed sue
a / h 5 0.5 and the actual shearing stress, f, 5 Fi, is obtained from the followi
equation:

where B' = 70 000 in fps units

= 483 000 in SI units
The U S H T O specifications require a minimum moment of inertia of inter-

ere
I

p=0.3 for steel

a = actual stiffener spacing
tw= thickness of girder web plate
J = 25(h/a)l - 20 but J 2. 5.0

AREA requires intermediate stiffeners if the h / t , ratio is Wate

values defined by ~ q (10-21).
.
~f stiffeners are required the spacing is

but a I 72 in (or 1.829 m),

where t,= girder web thickness, in or mm
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:

AREA:

+ D/30 and width 2 bf/4

2 51 mm + D/30 and width 2 bf/4

width 2 2 in

thickness 2 width/ 16
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 fiansOn 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

:IR+

16

71

Figure 10-17 Top flange bracing using web plate and bracket for the through-girder ballasted-deck
rallroad bridge shown m Fig. 10-4.

designed as a column for an axial force based on the horkontal corn

2; percent of the compression flange force
=

H/ V

The approximation of the axial force using the sin defined by H and
than the actual length is sufficiently accurate for design.
----

use 28 percent

(Note that tlphimpact factor is not the same as used in Example 4-15
design momecf

..

M, = 9984(1.28)

12 779.5 kN . m

The design shear

V,

1680(1.28)

2 150.4 kN

We will neglect the live load on the pedestrian/semce walkway. With rhis
assumption the dead-load estlmate is

0.025&Af
Pkb

+ 27.5 - 9 = 27.6

v rather

Example 10-5 Design a welded railroad bridge girder for a 27.5-m-span

single-track ballasted-through deck bridge. Use a Cooper E-110 loading,
A-36 steel, E70 electrodes, and the AREA specifications. The general bridge
configuration is as shown in Fig. E10-5a.
S ~ L ~ TFrom
I ~ NTable 1-2, the design live load for one-half the track (one
rail load to each girder) and adjusted for E-110 loading is:

Girder weight (including stiffeners and weld)

= 10.0 kll

Ballast, including ties at 350 rnrn depth and a reduced

effective width: 120(9.807/62.4)(0.3j0)(4.42/2)

= 14.6

Steel deck and cross beams and factor of 1.10

From Example 4-15, the beam is a W760 X 160.7/ 1.58
Deck: 0.015(77.0)(6.49>(1.10)/2.0

x 4.45 = 1680 k~

=4.1 kN

kN

= 6.2

Track at 200 lb/ft: 0.200 x 14.59 (AREA spzc~f~catlon)

=2.9 kN

kN
42.2 kN/n

= 4.4

0.300 x I4 59
Total

= 377.4 kips

kN

Cross beams: 37(1.58)(5.79)/(27.5 x 2 )

Walkway: estimate 300 Ib/ft

V = 274.5(%)

> 25 m as

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,
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

M = - =wL2

42'2(27'5)2 = 3990 kN
8
8
Tne dead load shear (at reaction) is

v = -WL
=

The section modulus is

The corresponding maximum bending stress is

M
16770
fb = - =
= 135.4 MPa < 137.5
Sx
123.81

40.2Q7.5) = 553 kN
2

The total design moment is

+ Mdead= 12 780 + 3990 = 16 770 kN . m

VdeSi,,= V L + Vdead= 2150.4 + 580 = 2730.4 kN

Step 3. Check the girder weight.

MdeSi,, = M L

Step 1. Find the girder proportions.

Take h 2134 - 2(50) 2034 mm. Also the web must be at least

--

--

"

A,,,,, = 2(0.070)(0.815) + 0.016(1.994) = 0.1460 m2

'

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
along the girder, so we will take t,,, = 16 mm.
~

- =-=
tw

Try one web plate 16 x 1994 mm and two flange plates 70

815 rnrn.

2730.4
2.034(0.35 x 250) = 15.3 mm

N-L1:

1:

(O.K. so far)

127 << 170

2034
16

O.K.

Step 4. Compute the allowable bending stress F,.

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 ) .
radius of the gyration of the compression flange is

'fie tentative web area

A, = 0.016(2034) = 32.54 X lop3 m2
Wsing Eq. (10-2), we find the tentative flange area as
, r'

'

16 580
32.54
--- 2034 X 0 . 5 c
6

A -

Flange area

- 65.21 - 5.42 = 59.79 x

m2

0.070 x 0.815 = 0.05705 m2

A w - 1.994 x 0.016
-

0.0052 m~

total

= 0.06237

m2

i i y two flange plates 70 X 815 mrn ( t available only in 10-mm increments):

h = 2134 - 2(70) = 1994 rnm
---815 - 5.82 << 190
O.K.
25 2 x 7 0

-b -

fi

- = - =1994

16

tw

125 << 170

O.K.

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

bftf3
Ix = Iweb+ 2 ~ # ~
12

720)' +

- 106(0'016)(1'994)3+ 2(0.070 x 0.8 15) 997 + -

12
= 10 571 + 121 510

+ 24 = 132 114 x

lOW6m4

815(0,070)~10~
12

Step 5. Design the bearing stiffeners.

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,

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=-=
ht

'

2730.4
= 85.6 MPa
1994(0.0 16)

< 0.355

O.K.

This value of j, results in an allowable spacing of

a =

0.872(16)

1.5 1 m

<

1.828

O.K.

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

Total

290 kN

= 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

= 58.88

+-

;"I

m3

j-

u = - =V Q

2730.4(58.88 X
I
0.132114
, The shear resistance is limited to

I Field bolt

= 1216.6 kN/m

~ I O O IO V J ~ ~r I

ou

A I O U . I I 1.30

s h o p weid
Framing angle, designed ln Example 8-9

F, = 0.35F,ASh,, = 0.35(250)(0.016>(1rn)
---/ x lo3
-"
- / \ -

-~-twnlu/rn~1~10.0

"

Figure E10-5d

0.K.

Step 8. Check thc:deflection.

. that
. the
. live-load moment is due to an equivalent uniform
nssumlng
loa~ d we
, obtain
WL =

8M

L~

and the total equivalent uniform load we is

we = wD

+ wL = 42.2 + 8(12 780) = 177.3 kN/m

27.5'

.The.dsflection is approximately

A = - - =5wL4
384EI

force

5(0' 1773)(27'5)4
= 0.04996 m (say 50 -)
384(200 000)(0. 132114)

27.5(1000)
= 43 mm = 50 mm
(take as O.K.)
Amax =
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 22high-strength bolts at each end. Take the horizontal component of stiffener

The number of 22-mm A-325 bolts in double shear required

horizontal component of stiffener force to attach stiffener to floor
193
= 1.83
N =
(0.7854 x 0.022*)(138 x 1d)(2)
Use three fasteners (AREA minimum number of fasteners in a co
For the stiffener try a WT 385 x 73.7
A = 9.39 X
m2
t, = 17.02 m m
d = 376.6 mm

Pa

t,

13.21 m m

AFa = 9.39 x 133 = 1249 kN >> 579

O.K.

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)

Design the channel-to-beam weld for the girder of Prob. 10-3.
Design the channel-to-beam weld for the girder of Frob. 10-4.

An additional concentrated force of one-fourth of the heav

applied at each critical point;. this gives a force of

&
d!,S

1 10(4.45)(0.25) = 122.4 kN

End panel:

Interior panels:

122.4 +

90
= 143.01 k~
2

122.4 + 9.0(4.58) = 163.6 kN

ave a separate set of design data of length depth and load. Be sure to w th

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.
If no specific data are assigned, take P = 2 15 kips, L = 54 ft, and D = 80 in.)

-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
@der depth to 2.15 m. Use W/4 = 110 kN, P/2 = 950 k~

*dtr)
bC
(,vt,

D ~ S ~a W
welded plate girder for a single-track open-deck railr
,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.
'

SELECTED COMPUTER PROG

. ,.,.

A-1 FRAME ANALYSIS PROGRAlM

The frame may be
This computer program will analyze any plane
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.
1Determine NP. The computer program computes NPPl = NP

,, \$.510 STRUCTURAL STEEL DESIGN

S E U C I ED

5. Determine NLC that uses wind the first time as.NLW.

6. Determine NLC that uses D + L, which is to be later comb
NLC that uses wind.
7. Determine the NLC that has wind on any beams or columns. Call this
values, which should be only wind.
8. Find NBAND using the method outlined in the text.
9. Make sure that ISTIFF is dimensioned greater or at

1 = NP value of the entry

This inputs the NLC nodal forces at
1 , ~ L) ~
= 1,NLC
for NLC load conditions using 8FlO.
Note that this requires at least two cards
NNZP.
5. Add your system control cards.
Note that you may (but not recommended) read nodal moments
hove. If you do read nodal moments, the output moments must be Correc
and to get the final design moments as
Ffinal = F ~ ~ t p +
~ tMrrad

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.

2. Punch member data cards according to the FORMA
MAT(7 14, 4G 10.4, 12)
The G format allows use of either F10.4 or xxxE-3 or xxx
NPE(7) uses the I2 format and goes in column 70.

f input units are as,identified here, the output will be

X = displacements = in, mm, or rad

F = member forces
Truss members:
kips or kN
Beam members:

Use NPE(7) = 1 for all members without transverse loads.

For any members with transverse loads:
NPE(7) = 0 for normal gravity beam loadings.
NPE(7) = - 1 for wind loads on beam or transverse loa
columns.

kips or kN for axial for

f t . kips or kN

m for mom

entification of Program Variables

of gravity
X 1, X2 = horizontal distance from origin end of member of P1 a
or m
Vl, V2 = column shears from alternate direction for a space frame, V1 =
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.

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
UNIT4
UNIT2

FT or M (start in column 1 of card)

I N or MM (start in column 5 )
K-IN or K-MM (start in column 11-unit

used for

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

~LW;
NNZP

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)
first load condition with wind
number of nonzero P-matrix entries to
0 for fps problems = 1 for SI problems
1 for all truss members; = 0 if any
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
1 to write X (deflection) matrix
0 to obtain output with few members
1 to obtain output when large number of NLC a
modulus of elasticity, ksi or MPa
member number assigned to element during cod
NP values for the element; truss has 1 through
through 6
horizontal distance to far end of member including si
or m
vertical distance to far end of member
m
cross-sectional area of member, in2 or m2
moment of inertia of member, in4 or m4 (not need
truss members)
control switch for members with trans
0 for members with transverse loading
1 for members without transverse loading
MENT cards in computer program)
BE SURE TO PUT BLANK CARD AT
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

IWINRE
IOLPL

UNIT^
UNIT,,

..

..

COUNTEL! TO ADO WINO LOAD OF&r'$E>ll

C O U N T E R F O R D E A D + LTVE.C?~O~-LI(:

-- UNITZ . IN-K OR UN-PM .AU U I T 3 $ 1 4 011 UN11M -- * * * NOTE C ~ E F U L L YU N I T S OF U M I T ~ I IFOR I N P U

J

"

-Y9

K OR K N

I N OR

* ,

lrPMIN(I15I
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 ~ ~ ~

READ CARD
2 R E A D ( ~ , ~ ~ O O ) L Z Z , U N I T I ~ U H I T ~ ~ U ~ ~ ~ Z ~ U ~ ~ T ~
000 F O R M A T ( Z O A ~ , I , A Z S Z X S A Z ~ ~ X ~ ~ ~ A ~ S ~ X ~ )
IF(EOF(lll750~3
3 WRITE ( 3 r Z 0 0 0 ) ZZZ
2000 F O R M A T ( ~ ~ ' J ~ ~ S ~ X ~ Z O A ~ ~ ~ ~
REAO CARO

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
1 0 0 1 FORMAT(lbI5l
L++*+*
REAO MODULUS OF E L A S T I C I T Y - K S I
READ(lr1002lE

1 0 0 2 FORMAT(BF1O.C)

- *

D E F I N E U N I T CONVER$ION FACTORS
IUNIT
1
IF(IMET.GT.O)IUNIT
2
0 0 4 0
1 . 1 ~ 6
FU(I)
FFU( I U M I T s I I
4 0 CONTINUE

NPP1

NP

OR B P I

***********

I
I

5 STRUCTURAL STEEL DESIGN

FASAT(Ze3) = -EASATII.Z)
FASAT12s4) = -EASAT(2,2)
EASAT(391) = -EASATf1.1)
E d S A T f 3 r 2 ) = -EASAT(1,2)
EASAT(3131 * EASATf1.1)
EASAT(3r4) = EASATf1.2)
EASAT(4rl)
-EASAT(I.Z)
FASATf4.21
= -EASATf?,Z)
EASAT(4.31
EASAT(1.2)
E A S A T ( 4 * 4 1 = EASAT(2.7)
0 0 1 7 9
111.4
IF(NPE(I).GE.NPP~IGo
TO 1 7 9
N t l = (NPE(I )-~)*NRANO
00 178 J = l r 4
I F ~ N P E ( J I . G E . N P P I ) G o TP 1 7 9
~F(NPE(J).LT.NPE(I))G~
rn 178
NSZ
NPE(J)
NPE(1) t 1
STTFF(NSl
NS2)
S T I F F I N S I t N S Z ) + EASAT(I,J)
7 8 CONTINIJF
1.9
CONTINLJF
G n TO 5 5 0 0
1 0 5 EAOL = E * A / ~ L
E I O L = (E*xIIxL)*Fu(?)
S I N O L = XSINIXL
COSOL
XCOSlXL
CA(1.l)
0.
EAlloZ)
1.
EA(lv3) = 01
FAf2rl)
-XCOS
FA(2rZI
SINOL
Eb(2r31
SINOL
EA(3.1)
-XSIN
EA(3.21
-COSOL
E A ( 3 r 3 ) = -COSOL
E A ( 4 . 1 1 = 0.
EAf4.2)
0.
EAf413)
1.
EA(5.1
= XCOS
FA(5.2)
= -SIYOL
EA(5.31
-SINOL
EA(6.1)
= XSIN
E b ( b t 2 l = COSOL
E A t 6 r 3 ) = COZOL
ES(lr1)
EAOL
E S ( 1 ~ 2 1 = 0.
ES11.3)
= 0.
ET(2.1)
0.
ES12.2)
4.*EIOL
ES(2.3)
= Z.*EIOL
ESf3rl)
0.
ES(3.2)
= ES(2.31
ES(3.3)
= ES(2.21
DO 2 0 2
I = 1.3
0 0 202
J
1.6
E S A T ~ I P J )= 0.
0 0 1 8 6 K 1 1.3
ESAT(I*J) * ESATfIpJl + E S f I ~ Y ) * E A f J s ~ )
1 - - CONTINUF
to.? CONTINUE

SELECTED CObIP
2 0 4 CONTINUE
DO 2 0 6
1 = l r 6
IFfUPE(I).GE.NPPl)GO

'

-.

----

--

NS2 = N P E f J l
NPE(II + 1
S T I F F I N S 1 4 NSZ) * STIFF(NSI+NSZ)
2 0 5 CONTINUE

c
c

EASAT(1,J)

5 5 0 0 GO TO 1 0 6
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
WRITE BAN0 R A T Q I X I F L I S T 4 > 0

72R IF(I1.GT.O)GO
C
H2
C

IN

ISTIFF

IN

COPE

TO 6 0 1

NSANO*NP

IF(LIST8.LE.O)GO
TO A 9 R 9
WRITE(3r2009)
2 0 0 9 FORMAT(
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
ni = 1
R2
N8ANO
MCOU
NP*NBAHO
On 3 0 5
I = 1.NP
W R I T E ( 3 r Z O l O ) I r ( S T I F F 1 II?.;~.MI,MZ
1
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
1FIMZ.GE.NCOUIGO
TO 8 8 4 9
11 = M2 t 1
M2 = M2 t NBANO
r
3 0 5 CONTINUE
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
C
C
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
C
DO NOT READ P - M A T Q I X E N T R I E S FOR FEM--INOUT
SO COUPUTEQ
C
COHPUTES F E * 5 0 F I N A L MOMEMTS A Q E COQDECTEO FOP FEM * * * * * a *

--

TO 4 2 6
IF(NNZP.EO.OIGO
0 0 9_91 NN = 1,NNZP

READ( l r 1 0 1 0 l t ! l
C
1 0 1 0 FORMATf1615)
C

READ CARD******

READ(1,1011)(PRfLI~L*l~NLC)
C

IF(II.GT.09GO
TO 6 0 5
203 0 0 204
I = l,b
0 0 204
J = 1.6
EASAT(1.J)
0.
DO 1 8 7 K 9 1 . 3
E A S A T f I # J ) * E A S A T f I p J ) + EAfI,K)*ESATfK.JI
1 8 7 CONTINUE

TO 2 0 6

1011 F O R M A T ( 8 F 1 0 . 4 1
OD 9 9 0 L = 1,NLC
P(n1.L)
= Pfn1.L) + P R ( L )
9 9 0 CONTINUE
9 9 1 CONTINUE
4 2 6 CONTINUE
DO 9 9 2
NS1 9 l r N P
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.

,.

************~*****;~******T~***lOtOOO~OOO.**

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,
' r A 2 r 1 AN0 ' , A 4 r / )
NS1 = 1
NS2 = 1 0
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

i 0 1 4 F O R M A T l T 5 r 'MP
',13~1X,lOFll.2)
IF(NS2.EQ.NLC)GO
TO 4 2 8
N S l = NSZ + 1
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

C'

SUBROUTINE TO REOUCE THE RAND MATRIX

CO N 1
1'
0080
N-1pNP
I - N
DO 7 0
L = ZtNBANO
NL
(N-l)*NBAND + L
I = I+1
IFISTIFF(NLI.EQ.O.IG0
TO 7 0
8
STIFF(NL)lSTIFF(Nl)
J - 0
00 68 K
L,NBANO
J = J + 1
I J = 11-l)*NBAND 4 J
NK
(N-lI*NBAND + K

67
70

66
80

C
C

= P(NtM)

IFfSTIFF(NK).NE.O.)P(NrM)

ENO OF MATRIX

REDUCTION

c
C
C

F(Z)-F(Zl*3~/4.
F(3l*F(31*3.14.
IF(JJ.NE.IDLPLIG0
Tn 973
C
PUT 0 t L I N T F - M A T Q I X
00 970
I
193
9 7 0 TF(YEMNO.11
* F(I)

--

SOLUTION

IS

IN

P(I.JI

1.58 CONTINUE
I F ( 1TRUSS.EQ.O)GO

TO LATER A D O TO Y I V O L O I D COHOIT.

TO 1 8 9

'

4 3 1 CONTINUE

I F AOOEO TO WINO

STIFF(NK)*P(L*KpM)

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
503 U R I T E ( ~ ~ ~ O ~ ~ ) I ~ ( P ( I ~ J ) P J * ~ ( S ~ , N S Z )
L O 1 8 FORMAT(6X,'NX
9
*rI3~1XrlUFll.5)
IF(NSZ.EQ.NLC)GO
TO 4 3 1
N S 1 = NSZ + 1
NS2 = NSZ 4 1 0
~ ~ 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

REDUCE MOMENTS FOP WINO

ALSO NECESSARY TO REDUCE 0 t L MOMENTS B Y 0 . 7 5

9 6 CONTINUE
GO TO 8 5
\fi CONTINUE
C

NOTE I F YOU PEA0 D-HATQIY ENTRIES FOR FE* THEN OUTPUT *UST 5 E
AOJUSTEO BY HAHO TO ACCOUNT FOP F I X E D END MOMENTS DM 9 E A M ENDS

6 0 7 CONTINUE
Fl2l-(F(ZI-FEMlI/FU(l)

INCLUDE LOAD MATRIX I N REDUCTION

00 67 H
1,NLC
P(IsM) = P(IPM)
B*P(NrM)
CONTINUE
M = 1,NLC
0 0 66
P(H,M)
= P(NrM)lSTIFF(Nl)
CONTINUE
N 1 = N 1 4 NBAND
COMPLETE SOLUTION BY BACK S U B S T I T U T I O N
IF(N.LE.O)GO
TO 9 0
L = N - 1
0 0 06
K = 2, NBANO
NK * ( N - l ) * N B A N D 4 K
0 0 86
H
1pNLC

602 IF(ITRUSS.EO.O.OR.IWPITP.E~~~)WPITE(~~~O~OIJJPUNIT~~UNIT~
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
5Xr *MEMBERit 4 Y p ' A X I P L
1 37X. 'DESIGN END MOMENTS C0RRECTEO'r/,
7FOWCFs ' r A29 8Yr'FOR FFM AN0 WINO (NEAR FNO F I P S T l r ' , A L P I ~ ,
RFWINO 5
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

INCREMENT COUNTERS I 1 AND J J - - J J

COUNTS NL

IF(NLC.EO.1)GO

TO 1 9 5

IF(F(1I.GT.PMAX(ME*NOIIDMA~(MFRNO)

IF(F(1I.LT.PMIN(MEHNOIIPMIN~~EMNOI

* F(11

TF(HEMNO.JJ)
F(1)
I F ~ I W R I T P . L E . O ) G O TO 4 4 1
WRITE 9AR FORCES FOR NLC

F(11

1 AND r V Q I T P

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

5U)

. .
STRUCTURAL STEEL DESIGN

SELECfED COMPUTEX

A-2 LOAD MATRIX GENERATOR FOR AASHTO TRUC

LOADING ON A TRUSS BRIDGE

JJ = JJ 4 1
GO T O 1 8 9
195 JJ
1

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
IF(NLC.EO.l)GO
TO 1 9 3
REWIND I
C
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
0 FP
O R M A X AND IN
B A R F O R C E S ONLY---USE
I W R I T P = -1 F O R C O ~ P L E T E L I S T I N G
B A R F O R C E S WHEN N L C IS L A R G E R THAN 1

C
VRIT~(3,2033)
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

z',/T

0 0 447

.
7

lrNn

~ R I T ~ ( 3 r 2 0 3 5 ) I ~ P n ~ ~ ( ~ ) . ~ n ~ ~ ( ~ ) , ~ ~ ( ~ y ~ ~ ~ )

FORHAT(6X*T ~ ~ X ~ F ~ O . ~ Y ~ X , F ~ O . Z ~ ~ X , F ~ O . ~ ]
4 4 2 CONTINUE
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

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,

OF

This program steps a movlng load consisting in either two or thre

loads along a truss. At each step the panel loads are computed as
with concentrated loads. These panel loads are the P-matrix entries f
weightless truss to use in the frame analys~sprogram.
The loads move from left to right. If the larger wheel loads ar
truck is "backed" across the truss. The llmtatlons are:
1. The user must use at least two and not more than three loads2. The "step" must be an integer multiple of the panel length. Since this
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

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
NS1-NSZ41
NSE=NSZ+~O
IF(NS2.GT.NLC)NSZ.NLC
URITE(3r2032)NSl
GO T O 4 4 3
1 9 3 GO T O 2
C
7 5 0 STOP
END

Output is automatically punched onto cards m the correct format for

input as the P matrix in the frame analysis program. The user may wish to
an optional punch control to inspect the output prior to having it punche

Variable Identification
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
increment of wheel movement, f t or m
STEP
0 for fps; = 1 for SI bndges. This is used to proper1
MET
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.
cumulative wheel spaclng as 14. 28. for using 14 ft between each set of
X(1)
wheels. This spacing or the SI equivalent wdl usually be the most
critical spacing.
NP numbers in order from left to nght where the panel loads are
Ml(1)
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
r

S T R U C T U R A L STEEL DESIGN

NP IS THE NUMBER OF LOADS THAT A R E WANTED ON THE TRUSS

SL IS THE
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
CONDITION
P t I ) I S THE WEIGHT OF EACH LOAD I N THE S E R I E S
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
MET=O.O FBS ~ 1 . 0 M E T R I C
D I M E N S I O N T~MPt20~70)rREACTtZO170),Pt5lvXt5Is M l ( 2 0 1

C
(

r-

'

1 FORMATt215r2Fl0.4,15)

i,

WRITEt3v5)NPrSL,NSvSTEP

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

lPANEL'vFB.2
v//v'
NUMBER O F PANELS I N TRllSS'v2X113. I/.'
LENGTH OF

2MOVEMENT OF THE LOA0S1,F8.2

DO 2 0 I = l . Z O
DO 2 0 J = 1 . 7 0
TEMP( I,J)=O.O
70 REACTtI.Jl=O.O
DO 1 0 0 1 - 1 1 5
Xt Il=O.O
L')
P t Il=O.O
NLS=SL/STEP
MX=NP-1
NSPl=NS+1
NSMl=NS-1
FORMATt1OFf3.21

WRITE(~~~)~IIP~I)~I=~~NP)
b FORMATt///.*

P'.Ilr'

=lrF8.3)

WRITE~~.~)~JIX(J)~J=I~NXI

'

REA0(1.81(MltI)rI*lrNSPl

r//)

R E A O ~ ~ ~ ~ ) ~ P ~ I ) ~ I = ~ ~ N P ) ~ ~ X ~ J ) I J = ~ ~ N X )
'

2 0 1 CONTINUE
I F (XIMPAT.GT.0.30)
XIMPAT-0.30
00 1 0 3 I-1.MZ
W 1 0 3 J-1,NSPl
103 REACT~J~II=REACTIJ,II*~~+XIMPAT)
WRITEt3r4)
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 0 6 W R I T E t 3 . 3 1 Jt ~ R E A C T ~ I I J ) ~ I - ~ , N S P ~ ) "
*. t
3
FORMATtT5,13qlOt5XvF6.1))
PUNCH M l FROM L E F T TO R I G H T ACROSS fk T R U S S "
C
PUNCH
FORMAT
I
S
COMPATIBLE
W
I
7
H
A'NALYSIS
PROGRAM FOR DIRE$$
C

7 FORMATl///r9
X 1 1 I 1 v ' =',F8.31
SK=O.O
SK2=0.0
DO 1 1 0 M = l v N P
DO 1 0 9 I = l r N S
IPl=I+l
DO 1 0 8 J = l , N L S
RR=P(M)*tJ*STEP-SK-SK2)/SL
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
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

FORMAT( 2 0 1 4 I
00 1 1 2 I=Z.NS
WRITE(2v9)MltII
FORMATt15)

8
9

112 W R I T E ~ ~ ~ ~ O ) I R E A C T I I ~ J I ~ J J ~ ~ M ~ I
10
FORMATtBF10.41
4000

STOP
EN0

****

FOLLOWING 3 CARDS ARE SET FOR FPS OUTPUT

5.
0
7
25.
14.
28.
32.
32.
8.
26
2
6
1 0 1 4 1 8 22
26
F O L C M I N G 3 CARDS ARE SET FOR S I OUTPUT
C
1.5
1
3
7
7.5
36.0
142.0
142.0
4.25
8.5
26
2
6
1 0 1 4 1 8 22
26

****

I*

INPUT

5k-4smUclVbU.. STEEL DESIGN

"ft

SELECTED COLLPUTER. PROG

,$i

A-3 LOAD MATRIX GENERATOR FOR

LOADING ON A TRUSS BRIDGE
n l s program steps Cooper's E-80 train lo
(DX) the panel loads are computed as a
loads are the P-matrix entries for a weight1
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/

1 FaR M E T R I
ISUIT
NO OF SPANS OR GIRDER SEGMENTS:
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
TRUSS SPAN OR GIRDER SEGMENT:
F A C = 0.75
FOR E ~ W - 0 . 9 FOR E72--1.00
FOR E B O - - L . ~ ~F~~
~
~ 1 1 0 NOSPAN

c
c
c

= 0 FOR FPS:

WIOH

IWRIT

CENTER T O CENTER SPACING OF T R U S S E S

SELECTED COMPUTATIONS FOR DEBUGGING

= 1 TO WRITE

'

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
ISWIT
0 for fps; = 1 for SI prob
spacing to SI and to
IPUNCH
0 for not punching O U ~ P U
= 1 to punch output in f
frame analysis program
for no impact factor;
to write selected inter
am; = 0 when pro
crement of wheel movement left to right, f t or m
ISPAN
panel length, ft or m
WIDH
width between two trusses making up bridge an
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

i$
%

"r,Tt,

~ E ~ 0 1 1 ~ 1 0 0 2 1SPAN,UIOHHFAC
0 X ~

'

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

NSPN = NOSPAN
1
REAO(lr100611NPSIJI .J=lqNSpNI
0 6 FORHATl1615)
I F 1 1SWIT.GT.OIDIS
= 25.
8 = NOSPAN
TOTSPN = B*SPAN
SET WHEEL LOAOS AND WHEEL SPACING
SUHXL = 0.
0 0 5
I-1.18
X I 1 I = FFXI I 1
IF(ISWIT.GT.OIXII
I = x111*0.3048
P I 1 I = FFPI II*FAC
F
I( ISWIT.GT.OIPII
I = PlI)*4.44822
SUMXL = SUHXL + XI T I
5 CONTINUE

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

2000

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
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//)

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

WRI T E ( 3 r 2 0 0 6 )
2906 FOPMAT(lOX.'***

4
I.,

6F9.1.//)

LOAOS M I 0 IMPACT--USE

FOR FLOOR BEAM D E S I G N AN0 CH

F I N O NO OF LOAOS IN. CURRENT SPAN

.TOTSPNIXILI
IFlL.EQ.NP+1.WO.AOXH.GT-TOTSPNIXIL)

IFIL.GT.NP+l.OR.XlLI-GTTSPANIXIL)

'

j4$!JZSTOP

= TOTIP. + SUMXL + o x
0 0 8 3 K=1,20
DO 7 3 L = 1 . 2 1 0
7 ' PR1K.L)
= 0,
~ ' iCONTINUE

DO 9 5 L * NCIZOO
IFIL .EQ.NP+I.ANO.AOXH.LE

= Z*SPAN
SUMXL
= O I F F + Z*SPAN
SPAN

SWXL

IFlXfL)..LT.O.tGO
TO 1 4 8
IF(XIL).NE.SPANIGO
TO 2 2
0 1 = SPAN
DZ = 0.
2 2 sun1 = sun1 + XILI
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
9 5 CONTINUE
96 C O N T I N U E
SUM3 = SUM3 + SUM1
TOTSUM = SUM3 + SUM2
N C = L * l
02 = 01
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.
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 ,
4 3 x 1 'SUM3 = ',F7.3,3X,'SUMZ
=
.F7.3r3X1'01FF~',F7.3,3Xv
'TOT
5'rF6.lr2Xv'JJ~',IZ1.
/r5X1 'XI9 ~ ' ~ F 7 . 3 ~ 3 X ~ e X 2 0 = 1 ~ F 7 . 3 3 3 X ~ * X 2 1 = * 7
6F7.3v3XvQXL='.F7.31II
CONTINUE
ACCUMULATE EFFECTS OF LOAOS I N U I Y SPAN
0 0 9 9 L L = NN1L
IFILL.LE.NP)GO
TO 9 7
P I L L ) = WLOAO*XlLLl
IFILL.EQ.NP+l.OR.NN.EO.L)OZ
= 02
X1LL)IZ.
9 7 RR = O Z * P f L L ) / S P A N
RL = P I L L 1
RR
PTIKC) = PTIKCI
RR
PTIKC-21 = PTIKC-21 - RL
02 = 02
XILLI
SUMP = SUMP + P I L L )
9 9 CONTINUE
SUMPT = SUMPT + P T I K C )
1FlJJ.EO.IM)SUMPT
= SUUPT + P T I K C - 2 )
Z = 1M
JJ
OZ = Z*SPAN + TOTSUM
AOXH
NN = L L
SUM1 = 0.
1 0 2 CONTINUE
SO CAN
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.)
U S E P R I N T E D OUTPUT TO OESIGN TRANSVERSE FLOOR BEAMS
BASED ON LARGEST PANEL LOAD * FLOOR BEAM REACTIONS
AN0 DEPENDING ON NUCBER OF TRACKS ON BRIOGE
OOES M T
I N C L U D E ANY IMPACT FACTOR *a*+****+
WRITEl3rZOO8)I~lPTIJl~J,Z~NY~2~
2 0 0 8 FORMAT(//.2Xv15v12F10.2v/,
5X,12F10.2,/1
WRITEI~~~OO~ISUMP~SU*PT
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 = * ,
1F10.2,/)
0 0 1 3 8 J = 1,NY
PRIJIII = PTIJI
1 3 8 CONTINUE
1 4 0 CONTINUE
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

DO 1 4 0
A

I = 1.210

M A I N DO LOOP FOR INCREMENTING LOADS

D I F F = 0.
ADX = A*OX
AOXH
AOX
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
GO TO 1 4 1
I A O X = TOTSPN
SUM = 0.
IFI AOX.GE.TOTSP

F I N O NO OF SPANS TO USE
, 5 SPUSO = A O X l S P A N

ZERO
DO 8 6

CONTINUE

MATRICES
ARBITRARY
K = 19.200

AMOUNTS

X I K I = 0.
1

i
i

I
b

4..

0 0 8 8 K = 1,50
T I K I = 0.
CONTINUE
I M = SPUSO
RM = I M
IFI8M.LT.SPUSOIIM
= IM + 1
BM = I M
1FIAOXH.GT.TOTSPN)OIFF
= AOXH
;UM2 = 0.
? F I A B S ( O I F F ~ . L E . O . O O ~ ) G ~ TO 1 5

DO 1 2 MM = l v N P
>OM2 = SUMZ + x t n n )
:~lSUM2.GE.OIFFlGO
TO 1 5
1 2 CONTINUE
t K NC = tin + 1
fin = I M
6 7 = BM*SPAN
AOXH + SUM2
'WR = NC
LUM1 = 0.
5UM3 = 0.
cUHP = 0.
\UMPT = 0.
I~~~~.EO.NP.ANO.OZ.LE.O.)OZ

i!

TOTSPN

3 1 = SPAN
OZ
= 2*lIM-JJ)
7 = JJ

KC

+ 4

**

= 0.

FOR NUMBER OF SPANS U S E 0 FOR CURRENT LOAD P O S I T I O N S

J J LOOP
= 1.TM

DO 1 0 2

C
C

In1 = I- 1

528 STRUCTURAL STEEL DESIGN

C

142
143
25

2014
145
149
148
2016
150

**

C
...,,,..

**

J E
9
1.05
4

***I,

FOLLOWIHG 4 CARDS T Y P I C A L SET O F F P S DATA FOR 9 PANEL TRUSS

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
0'
1
0
27.60
17.00
1.0
8
12
16
20
24
28
32
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
1
1
1
0
8 -40
5.20
8
12 i l l
20
2 4 lo&
32

J E BOYLES
3 . ,...
9 8 .....
o...

3.45
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
PR(KK.J)
= PR(KK,J!*PAC
b,
CONTINUE
KK = 2
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
FORMAT(BF10.4)
CONTINUE
GO TO 5 0 0 0
WRITE(3r2016)
FORMAT(II/v5X*****
PROBLEM T E R M I N A T E D - - X ( L )
I S NEGATIVE
STOP
EN0

/*

.
---- -"....-'---C,C.L~B:~$,.;<'~-'M,~DRAS

-...-*-..--.------

--

'

II

<
. c$,.;-.!

i
:,

.:. P!=3.

I ,

(-,>t,:

-- ...

3x0.

.-,

-,.. ,.---.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

tearing of welds, 420

.' F tamella
,

Lateral bracing:
.
for beams, 150, 155, 174
for columns, 291
for girders, 496
knee bracing, 497
Live loads, 18
reduction for, 19
Load conditions, 70
Load resistance factor design
(LRFD), 132
table of,@factors, 132
Loads, 18
bridge: AASHTO, 26, 27
AREA, 28
equations for shear and
moment, 26
impact, 32
tabulated shear and moment,
AREA, 29
dead, 18
earthquake, 33
hi:
eccentric, on fasteners, 375
impact, 32
live, 18, 19
ponding, 25
snow, 24
map for, 25
, wind,21,31
Low-temperature effects, 13
Llr ratio:
compression members, 258, 260,
277
tension members, 231, 237, 238,
242
LRFD, 132
beam column desig*, 338
beam design, 208
column design, 292
connections, 405
tension member design, 248
Lug angles, 3,65..
Mass density, steel, 7

tdius of gyration., 175, 237, 255,

MUNSE, W . ,

364

embers, tables, 272

ompression flange of beams, 179
Euler formula used in, 260
lacing for, 276
alh, 473, 478, 494
blt, 150, 465, 47 I , 490, 493

Net area:
column, 264
in tension, 219, 225, 232
effective hole diameter for;233
at thre'dd root, 219
use of s2/4g, 233
N E W L I N , D. E., 427
NEWMARK, N. M., 185

hlt, 456, 48 1 , 49 1
Llr, 263, 27 1 , 275, 307
Reaction distance for beams, 157
bracing for, 158
Reduced eccentricity of fasteners.

Plastic section modul

stresses: bendin

AISC specifications fo

, computation: for beams, 146,

joint length as factor, 360

Shear stress, allowable: in beams,

S1 conversion factors. 39
Sidesway, control of, 337
Slot weld, 414, 419
AISC specifications for, 4
Spacing of bolts, 363
Stepped columns, 329
K factor for, 33 1
Stiffeners, bearing, 158, 37
493
column flange, 427
web. 4'76

diagonal, for corner

connections, St6
glrder, requlred for, 456
~ntermediate,479, 493,
long~tudinal,AASHTO,
moment of Inertia of, 480,
spacing for, 493,495
seat angks for, 436
St~ffnebsmethod of analysis, 54
Stress, allowable: base plate des
287
bending, 149, 174, 469, 471 4
compression: AASHTO, 2
AISC, 261
AREA, 26-1.
fasteners, tables, 353, 3
366-368
interaction, 3 12
shear, 146, 159
in girder webs, 478, 492
tension, 219
table of values, 221
weld, 416
Stress range:
defined, 40
table of values, 41
Stress-strain curves, 1 1 , 113
Structural codes, list of, 17
Structural shape data, table,
Structural shapes, 7-10

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

$.%
$& ', @j "able

($.,&"!;I
&,

d
,

*'

Ult~rnatestrength. 7
of values, 6
Unhraced length, I75
LL'.1-51, 323
L", 151
Unlt welght of steel, 7

. VALLAERT,
J . , 368

Warping constant C, , 174