Steel Frame Design Examples

STEEL FRAME DESIGN EXAMPLES
By the same author

MORE STEEL FRAME DESIGN EXAMPLES
STEEL FRAME
DESIGN EXAMPLES
lan Robb
M.Sc., C.Eng., M.I.Struct.E.
Lecturer in the Department 0/ Civil Engineering
University 0/ Sa/ford
SI EDITION
THIRD EDITION
Palgrave Macmillan
ISBN 978-1-349-81764-1 ISBN 978-1-349-81762-7 (eBook)
DOI 10.1007/978-1-349-81762-7
© lan Robb 1961, 1965, 1972

Softcover reprint ofthe hardcover 1st edition 1972 978-0-333-02412-6
All rights reserved. No part of this publication may be

reproduced or transmitted, in any form or by
any means, without permission
First published 1961

Second edition 1965
Third edition (SI) 1972
Published by
THE MACMILLAN PRESS LTD
London and Basingstoke
Associated companies in New York Toronto
Dublin Melbourne Johannesburg and Madras
ISBN 978-0-333-13404-7
Preface to Third Edition
This book is primarily intended for students of structural steelwork
design up to intermediate standard, and as an aid to young engineers
gaining experience whilst engaged in professional practice.
It is assumed that the reader will have carried out some preliminary
study in theory of structures and strength of materials wh ich will enable
hirn to appreciate the behaviour of simple structures, and steel as a
material of construction.
The examples which foIlow are intended to give a broad outline of
structural design using the medium of steeI. This is accomplished in
two ways, firstly by a thorough treatment of the design of structural
elements, and secondly by demonstrating the arrangement of structural
elements required to produce a safe and economical structure. Neither
aspect can be viewed in isolation if the reader is to become a competent
design engineer.
The diagrams and drawings illustrating the text should be viewed as
typical details rather than as fuIly dimensioned working drawings.
Students with no industrial experience may find it beneficial to make
their own detailed drawings from the sketches given as a preliminary
to the transition from the study of theory to the practice of design,
a development that is not always easy to achieve. Drawings are an
essential language of communication for the engineer and their impor-
tance should not be underestimated.
As far as is practicable, the provisions of BS 449 have been in-
corporated in the text. Students should not regard this specification
in any other light than as aseries of recommendations which form the
basis for commercial design standards as weIl as professional practice.
The introduction of SI units into engineering is sufficient reason for
a thorough revision of the text, but modifications in BS specifications
regarding both design procedure and materials are equally important
reasons for undertaking the task. In the ten years since the first edition
was published the popularity of the triangulated framed structure has
vi PREFACE TO THIRD EDITION
diminished and the popularity of the portal type frame has increased
and for this reason the section of the text dealing with triangulated
framed structures has been reduced. (For details of portal type framing,
see 'More Steel Frame Design Examples' by the same author.) On the
other hand, more detailed consideration has now been given to con-
nections between structural members.
The author is indebted to the many students, coIleagues and cor-
respondents whose constructive comments have resulted in additional
material as weIl as improved presentation being incorporated in the
third edition.
I. ROBB
Contents
List of Abbreviations ix
1 Preliminary Considerations of Design

1 The role of the engineer as a designer. 2 The role of the speci-
fication in design. 3 Design procedure. 4 Choice of sections
when designing. 5 Relative costs of fabrication. 6 Relation of
strength to weight in selecting sections. 7 Importance of load
assessment. 8 Use of tabulated information.
2 Elementary Design Definitions 9

9 Relationship between load, stress, and strain. 10 Relationship
between shear and bending. 11 Relationship between bending
moment and moment of resistance. 12 Relationship between
ultimate stress (or rupture stress), yield stress, and working stress.
13 Factor of safety. 14 Fatigue failure of material. 15 BrittIe
fracture. 16 Deflection of structures. 17 Geometrical properties of
sections.
3 Elements of Basic Design 15

18 Beam design. 19 Struts. 20 Triangulated frameworks. 21
Design of tension members.
4 Design of Multi-Storey Shop Premises 24

22 Particulars of scheme. 23 Layout of steelwork. 24 Design of
floor beams. 25 Effect of wind on building. 26 Design of columns
(or stanchions). 27 Design of welded plate girder. 28 Design
of welded connections. 29 Design of mass concrete foundation
block. 30 Design of reinforced concrete foundation. 31 More
about connections. 32 Typical strength calculations for connectors.
33 Typical strength calculations for connections.
viii CONTENTS
5 Design of 18 m Span Shed With Ridge-Type Roof Trosses 81
34 Particulars of scheme. 35 Layout of steelwork. 36 Effect of
wind on building. 37 Roof purlins. 38 Design of roof trusses.
39 Design of side columns. 40 Design of mass concrete founda-
tion to resist overturning. 41 Design of side and gable sheeting
rails. 42 Design of gable steelwork. 43 Sienderness ratios of
struts. 44 Note regarding design loads and choice of sections.
45 Design of rainwater gutters and pipes.
6 Design of Braced Tower Supporting Water Tank 122

46 Particulars of scheme. 47 Layout of steelwork. 48 Investiga-
tion of wind pressure. 49 Design of tank-supporting beams. 50
Design of tower members. 51 Design of foundation to resist
uplift. 52 Check on estimated data. 53 Design of connections.
54 Provision of access to the tank. 55 Maintenance of the struc-
ture.
7 oe
Some Aspects Economical Design 137
56 Costs of fabrication. 57 Effect of layout upon design.
Index 145
List of Abbreviations
The following abbreviations have been used in the text and these
follow the recommendations of BS 449, CP3 and CPl14.
A Cross-sectional area
BS British Standard as issued by the British Standards
Institution
Bor b Breadth
CP Code of Practice for buildings as issued by British
Standards Institution
D ord Depth
dia Diameter
E Young's modulus of elasticity (taken as 210 kNjmm 2 for
structural steel)
Lever arm
Stress, in a general sense
Calculated stress in axial compression
Calculated stress in axial tension
Calculated compressive stress induced by bending
Calculated tensile stress induced by bending
Horizontal reaction
Moment of inertia (second moment of area)
Moment of inertia measured about x or y axes
Kilogramme
Overalliength of member
Effective length of member
Effective length of member measured on the x or y axis of
the member
M Bending moment
m Metre; m 2 square metre
mm Millimetre; mm 2 square millimetre; mjs meter/second
N Newton; MN meganewton; kN kilonewton; kNm
kilonewton metre
x LIST OF ABBREVIATIONS
p Load in a general sense (used where W would cause
confusion)
Allowable stress in axial compression
Allowable stress in axial tension
Allowable compressive stress induced by bending
Allowable tensile stress induced by bending
Unit wind pressure
Dynamic wind pressure
Reaction to a system of loads
Radius of gyration
Radius of gyration measured about x, y, v and u axes
respectively
SI, S2, S3 Wind speed factors
T Thickness of compression flange of a section
t Thickness
UB Universal beam section
UC Universal column section
V, V. Basic wind speed and design wind speed respectively
W Load or force
x Major axis of a section
y Minor axis of a section; also distance from an axis to a
plane of investigation
z Section modulus
1. Preliminary Considerations of
Design
1 The Role of the Engineer as a Designer
All design work should result in a structure that fulfils its intended
function at an economic price. Simplicity does not necessarily mean a
lack of effort and yet the most successful structures achieve their
function with an impression of ease which is the result of a combination
of complex skills. Each part of the structure will fit naturally into the
pattern of the whole giving an overall sense of unity free from the
distractions of unnecessary parts or a poor arangement of the various
members.
Consider the normal procedure in commercial design when a steel-
framed building is required. The customer, or dient, will have some
idea of the size, purpose, cost, and use to which the building will be
put. He will consult a plant engineer, if industrial plant is required,
and an ar~hitect if aesthetic appearance or space planning is required.
These experts will then confer with the customer, and an initial con-
ception of the building will be sketched out. It is at this stage that
the designer of the building frame should be consulted, so that his
knowledge of construction materials and economics of design can be
added to the discussion. The result should be aseries of final layout
drawings which will enable the customer to see whether or not the
building meets his requirements. On these drawings should also be
marked the position of all structural members, together with all
Hoor and plant loadings, so that the designer will be able to begin his
detailed design.
By adopting this procedure all the interested parties will have taken
their share in the planning stage of the project, and the final result
should be a complete set of layout drawings enabling each party to
know the requirements of the others.
Frequent exchanges of information should take place throughout
2 STEEL FRAME DESIGN EXAMPLES
the planning stage of the project, and as a result of this the designer
may undertake several incomplete preliminary designs for a variety of
reasons, e.g. to investigate probable intensities of loading, to compare
the appearance of alternative schemes, or to compare relative costs.
During the design of the structural members the designer will call
upon his knowledge of applied mechanics and mathematics. The use
of these two aids will enable him to estimate the probable effect of load
upon the structure. Having assessed the relative importance of each of
his findings, he will proceed to design a framework using the necessary
material in the appropriate positions to resist the loading upon the
frame members. These frame members will, in turn, transmit all
loads to the foundations of the building. In the ideal structure, each
member would be as strong as any other member, although this is
never achieved in practice.
To many students, the application of applied mechanics or mathe-
matics presents a lesser difficulty than the choice of frame layouts to
suit a particular problem, the assessment of loadings, or the choice of
sections from the many standard shapes available. The experienced
designer would describe this as 'judgement', 'experience' or 'general
engineering knowledge'. It is on the development of this knowledge
that the student must concentrate if he is to become a successful
design engineer. Tbis knowledge may be developed by practical
experience in designing, by observation of actual structures, reading
trade journals, and by discussion with fellow students or mature
engineers. Trade journals carry many advertisements with excellent
photographs, and these may be collected to form a 'type catalogue'
which may prove useful in appreciating the practical application of each
type of framework.
Because some students feel themselves lacking in this engineering
knowledge there is a tendency for them to take refuge in their knowledge
of mechanics or mathematics, which they understand more readily.
This is dangerous. The young designer should realise that his role is
primarily that of an engineer, relying on his knowledge of mathematics
and mechanics as useful tools to be used in his profession. This text
endeavours to explain the basic elements of steel-frame design. As in
any form of design in any industry or profession, the results from
different designers may vary, even though all received an identical
design brief. Tbe student must make up his own mind, and not be
led into thinking that there is only one solution to any particular
problem in design.
In the theoretical analysis of problems the student will be familiar
with the format of classroom problem in which the geometry of the
structure, intensity of loading, relative stiffness of members are specified
as pre-requisites of the investigation. Only some, and sometimes none,
PRELIMINARY CONSIDERATIONS OF DESIGN 3
of these pre-requisites are available to the designer who is required
to produce a safe and economical structure. Inevitably this must mean
some degree of trial and error investigation, usually on the basis of
making assumptions and proving them by analytical means. The
student should realise that although both theoretical and practical
problems are structural engineering, the approach to practical design
is different from the technique employed in solving theoretical class-
room problems. As experience of each particular problem progresses,
so will the degree of error in the assumptions diminish.
The adoption of SI units has presented the engineering industry
with one of its biggest challenges. Because of the enormous amount of
work involved in a change in basic units of design, the large variety of
industries involved, and the difficulties of transition, it will be some
time before final uniformity is achieved.
BS 449: 1959: Part 2, expresses allowable stresses in units ofnewton/
mm 2 (N/mm 2 ) although there is an opinion amongst some engineers
that the basic unit should be meganewtons/m2 (MN/m 2) in line with
the acceptance of the metre as a unit. F ortunately the two are indentical
numerically and no difficulty should exist. The author follows BS 449
practice and states stresses in N/mm 2 •
For simplicity the kg mass is ignored and all loads are expressed as
forces in newton. If gravitational aeceleration is taken as 10 m/s 2
(exact value = 9·81 m/s 2 ) then
I kgf = 10 newtons
and it is eonvenient to make this transition to units of newton before

ealculations are eommenced.
Unless stated otherwise the diagrams are noted in rnillimetre without
the unit being stated, e.g. 969 means a length of 969 mm. On layout
drawings where longer lengths or distanees may be speeified, then the
metre is used but in this ease the unit is speeified, e.g. 5 m means a
length of 5 metres.
It is likely that BS 4, 'Specifieation for Structural Steel Seetions',
will eontinue to speeify seetion modulus in units of em 3 notwithstanding
that the eentimetre is not a reeommended unit. However, this need not
eause eonfusion in simple stress ealculations for the following reason.
Bending stress = M
z
If M is ealculated in kNm then
M(kNm) x 103
stress (N/mm 2 or MN/m 2) = (3)
Z em
This procedure is adopted throughout the text without further co m-
ment.
For some considerable time the sizes quoted for structural sections
will be metric equivalents of imperial sizes and a certain amount of
'rounding off' is inevitable.
2 The Role of the Specification in Design

Before beginning to study design the student should equip hirnself with
the necessary guidance regarding working stresses, quality of material,
and workmanship. This may be summed up as 'good commercial
practice', and is contained in the following relevant British Standard
Specifications as published by the British Standards Institution.
BS4 Dimensions and Properties of Structural Sections
BS 449 The Use of Structural Steel in Building
BS 648 Schedule of Weights of Building Materials
BS 916 Black Bolts, Screws, and Nuts
BS 1856 General Requirements for the Arc-Welding of Mild Steel
BS 3294 The Use of High Strength Friction Grip Bolts
BS 4360 Weldable Structural Steels
The customer usually stipulates that the design must be carried out in
accordance with a certain specification. In this country it is usual to
specify BS 449. This requirement is not mandatory but its acceptance
provides common ground on which interested parties can meet. A
specification must not be regarded as a text book in design. Additional
or alternative clauses may be specified by the engineer or customer.
Examples of additional items are:
(a) Minimum size and thickness 0/ material. Where special conditions of

corrosive atmosphere exist (e.g. at the sea-coast or adjacent to certain
chernical processes) it may be necessary to specifya minimum thickness
of material, say 10 mm. It mayaiso be desirable to specify a minimum
size of angle section, particularly in regard to frameworks such as
roof trusses. In secondary members a small section such as a 31·8 X
31·8 X 6·25 angle may be suitable. This would result in a light but
extremely flexible structure in which there may possibly be difficulty in
making a riveted or bolted connection. For general work a minimum
size of angle such as 62 mm X 62 mm X 6 mm should be adopted and a
50 mm X 50 mm would be suitable for short subsidiary members. In
50 mm legs of angle sections the maximum size of bolt that can be
accommodated is 16 mm Ci in.) and the smallest leg size that will
accommodate a 20 mm dia rivet or bolt is 62 mm. If welding is pre-
ferred as a means of connection, and flexibility is of no importance,
then smaller sizes of angle are quite practicable. Other minimum
recommended sizes are:
gusset or connection plates, 8 mm thick;
base plates to main stanchions, 12 mm thick;
web plates to plate girders, lO mm thick;
bolts and rivets to main connections, 20 mm dia;
fillet welds in main connections, 6 mm leg size
(see Fig. 4.23 for definition of leg size).
(b) Size 0/ jinished pieces prior to delivery /rom workshop to site.

Many engineers prefer as much fabrication as possible to take
place in the workshop, where good conditions exist, rather than on the
site, where conditions are generally against good workmanship. The
size of fabricated pieces is limited by the capacity of the fabrication
shop in regard to space, machine capacity, or maximum loading of
available lifting apparatus.
Due to pressure of work in the fabrication shop it may be more
convenient for the work to be despatched in small pieces and assembled
at the site.
The transport of large pieces tends to be expensive, and limits of
size may be imposed by the transport authorities, especially rail
transport. Available access to site mayaiso decide the permitted size of
fabrications.
Consideration must be given, at an early stage in the design, to the
question of size of fabricated pieces as it will have an important bearing
on the cost of the structure.
(c) Other clauses may specijy: (i) conditions for payment and pro-
cedure in the event of faulty material and workmanship, (ii) limitations
to access to site, due to many building operations taking place at the
same time, (iii) statutory requirements on accident prevention and
employees' welfare.
One of the problems the engineer will have to solve before he can
design the foundations to the structure is the safe ground press ure.
This information is not generally available in standard specifications,
and the local authority which controls the area in which the building is
to be built should be consulted. Apart from the need to obtain their
approval for all proposals regarding building layout and design, they
will be in a position to advise on the subject of safe bearing pressures in
their particular locality. If the problem is complicated, or the nature of
the ground uncertain, then an expert in soil mechanics should be
consulted.
3 Design Procedure
Before detailed design is undertaken it may be necessary to consider
several preliminary design arrangements. Detail design procedure is, in
fact, the reverse ofthat which takes place during the construction ofthe
building at site and it is not possible to design the foundations until
all the loads are known. However the foundation conditions may
control the layout and spacing ofthe columns and should be considered
at an earlier stage in the programme. Detail design procedure should
start at the rooflevel and end at the foundations. Assuming the building
to be multi-storeyed, the floor beams should be designed, and their
size, together with a reference mark, entered on a layout drawing.
Subsidiary beams should be designed before main beams and in general
it is easier to work from the simple to the complex. All loadings re-
quired for column or stanchion design should be entered on the layout
drawing and the column design can then be commenced. Finally, the
foundations will receive attention.
4 Choice of Sections when Designing

A large selection of standard sections (refer to BS 4) is available to
meet designers' requirements in a variety of grades of steel. Certain
makers also list special sizes which are available for particular re-
quirements. The most economical design to fabricate and erect is not
always the one in which each member has been designed individually
for maximum economy. For example, a floor may consist of 50 beams
each having slight variations in loading. To consider each as an in-
dividual design would result in 50 different sections being chosen. The
work entailed in obtaining material in such variety, possibly from many
sources, the large amount of drawing-office work involved, and the
fabrication of many pieces all different from one another, would
certainly not result in an economical building. Common sense dictates
that the variety of sections be kept to a reasonable level.
5 Relative Costs of Fabrication

If it becomes necessary to design a compound section built up from
smaller units the cost of fabrication should be considered. The use of
complicated forms of latticed bars, or sections which are not easily
obtainable commercially, may result in an expensive form of construc-
ti on even though the proposed design suggests that the minimum
amount of material required to resist the load has been used. Economi-
cal, speedy fabrication and erection will be achieved by using sections
which are readily obtainable, by simplicity of design and choice of
section, which, in turn, gives simplicity of connection between the
members.
Hot-rolled universal beam and column seetions provide a wide
choice of standard sizes for the designer as weIl as ranges of angle, tee,
channel and tube sections. Cold-rolled sections made from relatively
thin steel plate (3 mm or 5 mm thick) of which a 'z' form is typical
are also available. These shapes are particularly useful for light struc-
tural members such as roof purlins where deflection is a criterion
rather than flexural stress due to imposed loading.
BS 4360 'Weldable Structural Steels' specifies the various types of
steel available to the design engineer. The grade of steel most com-
monly used is grade 43, corresponding to what the designer used to
specify in the past as 'Steel to BS 15'. Approximately 80-90 per cent
of steel used in building is grade 43 and it is appropriate that the
student should gain experience in its use. All the examples given in the
text are based on this grade of steel. Higher grades of steel (e. g. grade 50,
commonly called 'high-tensile steel') are available which permit
higher working stresses with a resulting economy in tln: weight of steel
required. This saving must be offset by the higher cost of grade 50
steel as compared with grade 43 steel and by consideration being
given to the deflection and flexibility which results from lighter sections
being used. Careful consideration of all the factors involved must be
made before a high grade steel is specified.
6 Relation of Strength to Weight in Selecting SectioDS

The standard sections available to the designer have been chosen to
suit a variety of requirements and each section is available in a number
of weights. When selecting a section it is necessary to try and achieve
the maximum strength for the minimum weight of material. This may
be illustrated by a typical example of beam design. Calculation has
established that the required section modulus is 750 cm3 • Possible
sections available to meet this requirement are:
311 X 167 X 54 kg z = 752cm3

356 X 171·5 X 51 kg z = 794cm3
381 X 152 X 52 kg z = 842 cm3
From an examination of the sections listed it will be noted that the
section which possesses the modulus nearest to the required figure is
not the most economical when weight is considered.
7 Importance of Load Assessment
Assessment of the magnitude and type of load is an essential first stage
in design. For instance, a load which is comparatively static, such as a
floor load, will have a different effect on the structure than that caused
by the movement of an overhead travelling crane in a workshop.
Much care and thought must be given to the correct assessment
of loading because the economy of the structure will be greatly affected
by it.
Dead load comprises all permanent loads such as self weight of the
structure and its covering, weight of floors, and permanent plant.
Live load or superimposed load comprises all occasionalloads such as
wind pressure, stored goods, occupants, or snow. Dynamic loading
is caused by moving loads such as cranes, road trafik, or vibrating
machinery.
The designer decides the combination of loads likely to be en-
countered by the building during its expected life. This does not
necessarily mean the summation of all possible loading conditions and
it may be that certain alternative combinations of loading conditions
require investigation.
8 Use oe Tabulated Information

Many items of information are repeatedly used by the designer in the
course of his work. The properties of structural sections, the loading
capacity of beams or columns,and safe loads on bolts and welds are
typical examples. Repeated calculations for these items would result
in time-wasting and tedious operations and it is convenient for all
such items of information to be tabulated in the form of a handbook
reference. The student is advised to obtain one of these volumes
before embarking on serious design.
In commercial practice the tabulation of information is frequendy
taken a stage further. During preliminary design work it is sometimes
necessary to compare relative costs of one type of structure with
another before deciding which is to be adopted for the project in hand.
These items of information are usually graphed or tabulated and are
prepared from cost records of previous projects. This information is
naturally confidential and is not usually released to the public.
2. Elementary Design Definitions
9 Relationship between Load, Stress, and Strain
If a structural member is subjected to an axial puB the member is said
to be in tension, and a tensile stress equal to the applied load divided by
the cross-sectional area of the section will be present. Conversely, an
axial thrust will produce a compressive load in the member. Stress is
load per unit of area. Alteration in length will also take place, and the
amount of extension or contraction divided by the original length will
give the strain in the member. The relationship of stress divided by
strain gives the value of Young's modulus of eIasticity for the material,
provided that the relationship is considered while the material is
within the elastic state.
10 Relationship between Shear and Bending

In a simply supported beam subjected to vertical loading, the action
of bending the beam will produce compressive stresses which will be a
maximum in the extreme top fibres, and a tensile stress which will be a
maximum in the extreme bottom fibres. In a beam of symmetrical
shape these stresses will be equal but opposite in character, and at
the mid-depth of the section, the stress will be zero. This line of zero
bending stress is termed the neutral plane if the beam is viewed in side
elevation, and the neutral axis if viewed transversely in section. The
variation in stress which occurs in the beam results in shear stress caused
by the tendency of the particles of material to sIide past each other.
If the beam is unsymmetrical in shape, then the neutral axis will
not be at the mid-depth of the section but will pass through the cen-
troid of the cross-sectional area. The stresses in the top and bottom
extreme fibres will not be equal as in the case of the symmetrical beam
section because the extreme fibres lie at unequal distances from the
neutral axis.
11 Relationship between Bending Moment and Moment of Resistance

In a member subjeeted to bending, the effeet of bending upon the
beam is obtained by seetioning the beam at the required point and
ealculating the moment required to maintain equilibrium at that point.
This ealculated moment is referred to as the bending moment eaused by
external forees aeting on the beam. Internally the beam must resist the
external moment and the resistanee is ealled the moment of resistance.
This is summarized in the design formula: M = z x f(see §17).
12 Relationship between Ultimate Stress (or Rupture Stress), Yield

Stress, and Working Stress
This may be explained by eonsidering the standard tensile test for a
mild steel speeimen in its simplest form (see Fig. 2.1). As load is
applied, elastie extension takes plaee in the material until the elastie
MAX LOAD
POINT
FRAC.,.URE
POIN,.
El~STlC
p..I\NGE
1/
STRAIN
FIG. 2.1. Stress/strain diagram.
limit or yield point is reaehed. If loads are applied within the elastie
range no permanent deformation will oeeur when the load is removed.
If loading is eontinued beyond the yield point some permanent deforma-
tion will oeeur when the load is removed. Conventional design, based
on the elastic theory, is eoneerned with the maximum stress that may
be plaeed on the material before yielding and permanent deformation
takes plaee. In mild struetural steel the yield point for grade 43A steel
ELEMENTARY DESIGN DEFINITIONS 11
is approx. 257 N/mm 2 and for grade 50B steel is approx. 355 N/mm2 •
To design a structure to this stress would be unwise, as no account
would be taken of such factors as:
(i) imperfections of design material, or workmanship;
(ii) loss of strength due to rusting;
(iii) accidental increases in load.
These unmeasurable factors are taken into account by adopting a
working stress less than the yield stress. The basic flexural working
stress for mild steel manufactured to BS 4360, as laid down by
BS 449: 1969, is 165 N/mm 2 •
STRESS = f
~~=~] ~ ~
~ J~
- I~
l_-: -- =-
--
-
--
-
-
---
--
--
--
--
--
-
--
--
-
- ---
[bl [c]
FIG. 2.2. Stress distribution in simple beam.
(a) Elastic distribution
(b) Partially plastic
(c) Fully plastic
This text is concerned with designs based on the elastic theory.

It is appropriate to mention an advanced method of design based
upon the failure load of the structure rather than upon an arbitrarily
decided working stress which takes no account of the mode of ultimate
failure. This method is referred to as the plastic theory or collapse
method of design.
Consider the test specimen previously mentioned, and loaded so
that the strain is increased beyond the elastic limit. Plastic strain
occurs, and it will be noted that an increase in strain is not accompanied
by an increase in stress. In a simple beam the stress distribution will be
as shown in Fig. 2.2.
In Fig. 2.2(a) the stress distribution lies within the elastic range of
loading. In Fig. 2.2(b) the stress distribution is shown as the yield point
of the material is passed. (No increase in stress takes place and the
section becomes partially plastic.)
In Fig. 2.2(c), as the loading is increased the whole section becomes
plastic, a plastic hinge is formed, and the beam fails.
The collapse load is divided by a load factor to give the working load.
Derivation of collapse loads and load factors requires a different
technique from that employed in conventional or elastic design. A
further consideration of plastic design is given in the author's 'More
Steel Frame Design Examples'.
13 Factor of Safety
The factor of safety is usually expressed as the ratio of the ultimate
stress divided by the working stress. If the ultimate stress is 435/
500 N/mm 2 and the working stress is 165 N/mm 2 then the factor of
safety will be 2·6/3'0 approx. Some designers contend that the factor of
safety should be related to the yield stress rather than the ultimate
stress giving a ratio of 1·5 approx. It is not possible to quantify the
factor of safety with any degree of accuracy because it is bound up
with such factors as the judgement of the designer and the quality of
materials as weIl as workmanship.
Experienced designers of mature judgement can work to reduced
factors of safety with a consequent economy in the structure without
endangering safety. Such practices are not for the beginner and the
examples given in the text are not meant to be irreducible standards of
design as might be appropriate in certain circumstances in commercial
practice.
14 Fatigue Failure of Materials

Fatigue failure occurs when material has been subjected to a large
number of stress reversals. In buildings, any variation in stress is not
likely to result in stress reversal and is likely to be gradual. No account
of this mode of failure is taken in normal commercial practice, with the
exception of those parts subjected to dynamic loading (e.g. crane
structures).
15 Brittle Fracture
When certain types of steel are subjected to stress at low temperature,
failure may take place by brittle fracture at stresses in mild steel as low
as 30/60 N/mm 2 rather than at normal yield stresses (250 N/mm 2).
It is most likely to occur in large structures exposed to extremes of
temperature in winter (e.g. bridges). Most building frames are protected
to some degree and it is not the usual practice to allow for this mode of
failure in the design of buildings. This type of failure is avoided by
using grades of steel not subject to this disadvantage (see BS 4360).
ELEMENTARY DESIGN DEFINITIONS 13
16 Deflection of Structures
Structural steel is an elastic material and as such will deform under
load. Deftections of members may require limiting if they interfere
with the efficient use of the building. Excessive deftection of beams
may affect the behaviour of supported plant or result in the cracking
of ceilings. In the case of columns it is possible for excessive deftection
in a horizontal plane to break brittle cladding materials or open up
joints allowing the ingress of the weather.
The use of high grade steels (e.g. grade 50 or 55) will result in smaller
sections being used with an increase in deftection as compared with a
design using a lower grade steel (e.g. grade 43). This fact may in
certain circumstances preclude the use of high grade steels.
Deftections which offend the eye are generally unacceptable as weIl as
those which result in discomfort to the occupants even though the
building structure may possess an adequate safety margin.
17 Geometrical Properties of Seetions

The geometrical properties of sections are frequently used in structural
design, and areminder of the most commonly used definitions is given
be1ow.
rb "I _ . - y
ci ci
"2 "2
x-r--- x cl x 0 x
d.
d "2
2. D
l.N E R1IA A80UT)( Ax 15

= bd
Tz
3
= I"c. x x
I.NERTIA ABOUT X- X AXIS
=I,,+ bxdxD 2
FIG. 2.3. Moment of inertia (second moment of area).
Moment ofinertia (or second moment of area). Refer to Fig. 2.3. The
moment of inertia of the basic rectangular section is b X d 3 /12 about
the x axis and d x b3 /12 about the y axis. If a moment of inertia is
required about an axis other than the x or y axes, then the following
expression is used
Ix = Ir + b X d X D2
A Iess frequently used value is the polar moment of inertia about the
Z axis which passes through 0 and is perpendicular to the other two
axes
Section modulus. The section modulus about the x axis = Zn and

about the y axis, ZY.
Also Zx = lx/yx
Zy = ly/yy
And
I
Hence M=fX-=fxz
Y
3. Elements of Basic Design
18 Beam Design
(a) In structural design the term beam is conventionally applied to a
horizontal member subjected to bending which occurs most frequently
in a vertical plane (e.g. a floor-supporting beam). Bending in a horizon-
tal plane may exist separately, or in conjunction with bending in a
vertical plane (e.g. a girder supporting a travelling crane).
(b) Simply supported beams are those in which the end supports are
assumed to be knife edges and no restraint is placed upon the bent
shape of the beam at the supports. When a beam is continuous over
several spans or a single-span beam is rigidly held at the ends then the
continuity or end restraint affects the be nt shape of the beam. These
types ofbeams are referred to as continuous orfixedbeams respectively.
(c) When bending takes place in a simply supported beam the top
fibres will tend to compress and the lowest fibres tend to elongate. Or,
the top portion of the beam will tend to act as a strut and the bottom
portion will tend to act as a tie. Unless the top flange is adequately
restrained laterally it will buckle in a similar manner to a strut. The
layout of structural members is usually arranged to give this restraint
but if this is not possible then the working stress in the compression
flange must be reduced to a level below that at wh ich buckling takes
place.
(d) The strength of beams must also be investigated when it is necessary

to guard against possible failure by buckling, shearing, or crushing of
the web caused by heavy point loads applied to the flange. This condi-
tion will frequently require attention at the be am support.
(e) Three basic methods are used to provide the strength required in a
beam section:
(i) using standard rolled sections, usually of I shape, although
channel, angle, tee, or zed shaped sections may be used if found con-
venient for a particular purpose,
(ii) using standard roIIed sections to whieh has been added to the
flanges a flat plate to inerease the seetion modulus of the seetion,
(iii) using a girder completely fabrieated into an I shape from flat
plates to give a seetion modulus mueh larger than either of the methods
(i) and (ii).
Figure 3.1 iIIustrates the three types.
ßASIC
! SECTION
BASIC SECTION W\TH
FL,o\NGE PLATE5 ADDE D
FABRICAiED
PLATE 61ROER
FIG. 3.1. Basic beam and girder sections.
(f) The span and spaeing offloor beams are deeided by the following:
(i) Type of flooring (e.g. reinforced conerete, pre-cast units, steel
plating). Eaeh type of flooring has an eeonomieal span suited to that
partieular form of eonstruetion.
(ii) Architectural requirements. Features sueh as the size and spaeing
of doors and windows frequently control the spacing of floor beams
and the supporting eolumns. As a general rule it may be said that
short-span beams are more eeonomieal than long spans, even when
taking into aeeount the extra eolumns required. Demands by the
eustomer for a certain area of uninterrupted floor spaee may deeide
eolumn spacings, and these in turn affeet the spans of the beams.
(Hi) General struetural arrangement. A building may require a
variety ofmembers to make up a framework (e.g. floor beams, eolumns,
roof trusses and purIins). Eaeh member must be designed bearing in
mind the eeonomic neeessities of the other members with the result
being a eompromise.
Reeent years have seen riveting as a means of eonneetion largely
replaeed by welding or boiting. Large sizes of roIIed seetions have also
beeome available making the addition of plates to the flanges of beams
a mueh rarer praetiee than was onee the ease. Automatie fabrieating
plant for weiding plate girder seetions together is also a modern
development.
ELEMENTS OF BASIC DESIGN 17
19 Struts (including Columns or Stanchions)

(a) 'Strut' is a general term applied to a compression member, usually
one in a framework. 'Stanchion' or 'column' is a term applied to a
vertical compression member.
(b) Normal failure of steel struts takes place by buckling. Vertical
loading produces a compression stress which results in a tendency for
the column to bend or buckle laterally. Consequently the compressive
stress must be kept below the level at wh ich buckling takes place. The
limit of permissible stress is calculated by means of a column formula
based partlyon mathematical analysis and partlyon experimental
work. An example of a column formula may be seen in BS 449: 1969,
appendix B.
The height over wh ich the column will buckle is called the effective
or buckling length, and it is over this length that the stress in the column
must be limited. The length for design purposes is the effective length.
(c) The amount of fixity at the ends of a strut has an important effect
upon the effective length of the strut. Degrees of restraint vary con-
siderably and the designer must take care in correctly assessing the
effect of end restraint on the column. This task is simplified by the
typical examples given in BS 449: 1969, appendix D. Basically there
are two types of end restraint.
(i) If a column were supported by a fixed pin at each end, the pin
would be unable to provide any restraint to the bent shape of the
column. The column would be free to take any deflected form that a
load might induce. This type of fixity would be termed jixity in position,
but not fixity in direction.
(ii) If a column were continuous beyond the supported ends, as in an
intermediate length of a multi-storey column, and the floor beam
connections were reasonably substantial, then these two factors would
tend to prevent the column ends from taking up a deflected form.
The ends of the column would be restrained (or jixed in direction) as
weil as jixed in position. In some cases the degree of end restraint may
be a combination of each of these types.
(d) A further important point is illustrated in Fig. 3.2. In this case the
column will buckle over the length LI about the x axis, and over the
length L 2 about the y axis, provided that the tie at mid-height is suf-
ficient to secure this point of the column in position. This device may
be effectively used in practice to gain the maximum economy in design.
(e) The column formula is most conveniently applied by first calculating
the slenderness ratio of the column. This ratio is equal to the effective
length of the column divided by the radius of gyration (r) of the
proposed column seetion about the appropriate axis (I/r). Permissible,
or safe working stresses corresponding to each value of I/r can be
calculated using the column formula, and tabulated (see BS 449:
1969, Tables 17a, 17b, and 17c).
COLUrv1N
/ L'Z
IY
"TI E\
-1---.
1" I E\
LI
-xf-~
Iy
SEeiloN A-A
L2
r ~
A , A
HG. 3.2. BucklingJength of columns.
20 Triangulated Frameworks
(i) Rooftrusses. Examples of steel roof trusses supporting roof cover-
ings are shown in Fig. 3.3. The angle of roof slope is decided by:
(a) Economy of design. A ridge height of between one-quarter and one-

fifth of the span is suitable for this condition.
(b) Architectural requirements. The appearance of the building may

depend upon a certain emphasis given to the roof proportions by the
architect.
(e) Type ofroof eovering. The cheapest, and therefore the commonest,
type of roof covering is asbestos-cement c1adding. Steel, aluminium,
and plastic take a smaller share of the market. There are so many
variations of each type available that it is not possible to generalise
as to the pitch of roof slope for each type of sheet. The recommenda-
tions issued by the manufacturers as to roof pitch and fixing should be
\
~ \ S .... SPAN
~ \ 10"", SPAN
DOTTEO L.INE.S REPRESENTS CENTR.E

L.INE OF RAF'TER
RAFTER DIVIDED TO GIVE. PANeL.
LENGTHS OF APPRoX 2..,.,oR TO
~"<,o, W'T" PUR"" Po,,"o",
\14msp~
17 ..,., :SPAN
23m SP .... N
FIG. 3.3. Roof truss frameworks for various spans.

closely followed particularly if the site is exposed to harsh c1imatic
conditions or the framework subject to deflection.
Internal framing of roof trusses is usually decided by the division of
the rafter into economicallengths of 1·75 m to 2·5 m. Once the type of
covering is known and a suitable spacing for the purlins supporting
the covering has been decided, a trial setting out should be made on
paper to determine the most economical layout of the sheets and
purlins. Maximum economy will be gained if the rafter panel (or
(~)
~ (bJ
1. 1. 1. 1.
~ 111. 1.
(e. )
HG. 3.4. Lattice girder frameworks.

(a) Basic 'N' type framing
(b) Basic type framing with sub-framing added
(c) Superimposition of two basic 'N' type framings
node) points coincide with the purlin spacings. A position may arise,
where owing to the required purlin spacing, it is not possible to locate
the purlins over the panel points. If this condition is unavoidable,
care must be taken to ensure that the rafter is designed to resist bending
caused by purlin loading in addition to the axial loads in the rafter
member.
In the examples of roof trusses shown in Fig. 3.3 the members are
arranged in such a way as to give symmetry about the centre line of the
rafter as weIl as the centre line of the whole frame. If this rule is followed
a pleasing appearance should be obtained.
(ü) Lattice girders. Examples of lattice,girder frameworks are shown in

Figs. 3.4 and 3.5. Lattice girders are large beams in which the web
members take the form of open lattice work. As in the case of solid
lV\t\ZV\
~ I :[
(C)
HG. 3.5. Lattice girder frameworks.

(a) Basic 'Warren' type framing
(b) Basic type framing with verticals added
(c) Superimposition of two basic 'Warren' type framings
web construction, it is assumed that the flanges resist the bending

moment and the shearing forces are resisted by the web members.
Large span girders may take either the form of fabricated plate girders
(see Fig. 3.1) or a lattice framework.
Both forms of girder are usually associated with large spans, the
plate girder being used for spans above which a rolled section is
available. An upper limit of span of about 30 m is usual for plate
girders, with lattice girders being used above this limit, although it is
difficult to be precise because common ground exists between the
two types of girders especially at medium spans. Parallel flanges have
been shown in the examples given but curved top members for very
large spans may prove economical (e.g. in bridge construction).
Two basic types of lattice girders are commonly used:
(a) The N-type framing shown in Fig. 3.4(a). Economy is achieved

because the long diagonals are tension members.
(b) The Warren-type framing shown in Fig. 3.5(a). In this framing the
diagonals usually slope at 45° or 60° and are alternatively struts and
ties beginning with a strut at each support. As a result of this type of
framing the Warren girder is much less susceptible to deflection than
the N-framing provided that the inclined members make a reasonably
large angle with the horizontal. This type of framing is frequendy seen
in bridge work.
Additional members, or sub-framing, may be added to the basic
pattern (see Figs. 3.4(b), 3.5(b)) when the loading points and panel
points do not coincide, and this arrangement is likely to prove more
economical than designing long members for flexural as weIl as axial
loading.
Figures 3.4(c) and 3.5(c) show an arrangement in which a second
girder is superimposed upon the oudine of the original girder as an
alternative means of supporting a group of flange loads. These some-
what fussy arrangements were popular towards the end of the 19th
and early 20th centuries, particularly in America where a number
of different types of frames were developed following the general
acceptance of steel as a suitable construction material for large trian-
gulated frameworks. Modern trends prefer the simpler arrangements
shown in Figs. 3.4(b) and 3.5(b).
The span-depth ratio for lattice girders is usually 10 to 15 although
this figure may be as low as 8 or as high as 20 depending upon loading
conditions.
Latticed arrangements in structural work tend to be associated with
environments in which labour is relatively cheap in relation to material.
If the reverse holds good then methods of fabrication using less labour
but a larger amount of material tend to develop.
21 Design of Tension Members

The amount of cross-sectional area that may be loaded in a tension
member is less than the total or gross area because reductions in
area must be made to allow for holes provided for connection pur-
poses. Further, in single angle ties it is usual to connect only one leg
of the angle resulting in eccentricity of connection, and a reduction in
area is also made for this fact.
BS 449: 1969, cIause 42, defines the net area to be used in design
caIculations. If double angle ties are used and no eccentricity of con-
nection exists (e.g. one angle positioned on each side of a connecting or
gusset plate) then no reduction of the area of the unconnected legs is
required.
4. Design of Multi-storey
Shop Premises
22 Particulars of Scheme
A steel-framed, multi-storey building is required for the shopping
centre of a large town. An architect was employed by the company
requiring the new premises and architectural drawings were made
showing the plan and elevations of the proposed building. Close
consultation took place between the customer, architect, and structural
engineer. Approval of the proposals was obtained from the local
authority and from the fire prevention officer during the planning stage
and before building work commenced on the site.
The main features ofthe building are as follows (see Fig. 4.1):
(a) The ground floor display window is to be free from all obstructions
for the full width of the building which means that no stanchions can
be positioned in the front of the building at ground-floor level except
adjacent to the side walls. To support the stanchions in the front of the
building above first floor level a plate girder is introduced, the casing
of which is to be used for advertisement purposes.
(b) Vertical curtain walling and glazing, divided into four vertical
spaces, is used as a feature on the front elevation. Support for the
curtain walling is provided by the beams at each floor level.
(c) Side and rear elevations are of 300 mm cavity brick walls supported
on steel beams and no window openings are required because of
adjacent property.
(d) Floor and flat roof construction is of pre-stressed concrete plank

units with 40 mm thick finishes to the floors and screed and asphalt
finishes to the roof.
ROOF 207M 134-
ua
3·7 5.
4 TH
i HOOR
kI5~"cIS3
207M 134
1'::=
8± I I 1- CU R"T .... IN
WA LLlt-IG
e=±
UB (ON CRE"TE
_LLI FE ATURE
3'75
"" 3"'D ....
(OL.~PL\C!
L- 0207,. 134
0 I
'FTOOii lJB
~
16'l" 154 _LLI
3·75 .. I VC
2 ND 207)1, 134
e±
HOOR
ue. 0
4·5 . 210'>< 20S _LLI
'" o (OL.5PL\(E ./ UG
."
- I8
IST
FlOOR NA ME
m
PL"'TE GIRDER.
0A 11.0
.<l·S m
/~53,,3IS
uC DISPLAY, I--" EH TfI,ANCE
GRO~ ND FLOOR
(0.)
zS,"x 14'" 4
ua
. /
CA SONG
--
'l5\'oxI4("
UII
BRIC.K
PANELS
25",,14("4
--
lJe.
,
•
25(,x 146,~ ,
UB f==:
12%)(14{"4
r---
. r---
0
, VB ,
(b)
FrG. 4.1. Multi-storey shop premises.

(a) Front elevation
(b) Side elevation
(e) The customer requires the maximum amount of floor space free
from column obstruction for display purposes. A figure of 7·5 m to
9·0 m is suitable for column spacings in this type of building and is a
compromise between structural economy and the requirement of large
floor areas. In this case a central column was introduced giving a
spacing of 8·4 m by 8·0 m.
(f) Fire protection of the steelwork is mandatory. The usual methods

are casing the steelwork in a fireproof board casing, or concrete casing
carried out wholly in-situ or partly off-site with areas adjacent to site
connections cased in-situ. Sprayed asbestos is also feasible but is
unlikely to be used in this context.
23 Layout oe Steelwork
Layout arrangements of beams and columns are given in Figs. 4.2(a)
to (d). Column flanges have been positioned in direction so as to
receive the heavy loadings from floors in preference to the webs (e.g.
refer to column SI). In a column section the bulk of the material is
concentrated in the flanges and this concentration is best able to
receive the loading. Additional columns are introduced in the side
walls at 4 m centres as this should result in greater economy than a
spacing of 8 m. At each floor level a tie is inc1uded to this column to
provide restraint across the x axis of the column.
Pre-stressed concrete floor units spanning 4·2 mare used. The par-
ticular type ofunit used in this instance consists of a pre-stressed concrete
plank unit and an in-situ topping of I : Il: 3 mix concrete. Ducting for
electrical services may be accommodated within the depth of the
concrete topping. A pre-stressed concrete floor unit is a convenient
way of obtaining a lightweight fireproof floor.
The rolled sections used in this example are 'universal beam and
column' sections (VB and VC) and Grade 43 steel (BS 4360) is used
throughout. Welding will be adopted as a means of making shop
connections, and site connections will be made with black structural
bolts (see BS 916).
24 Design oe Floor Beams

(a) Normal commercial practice assumes that the floor beams in a
building of this type are simply supported. This means that no account
is taken of the restraint provided by the beam end connections to the
column. The end connections should have suflicient strength to transfer
all the loading from the beam to the column, to restrain the top flange
of the beam from lateral movement, and to restrain the column.
4 .....
<IJ
r-
- ,.16
w jü"-----H
~cO
-'I":) lI\
LU 4M
(.ol 55
[107] H,(
1/'1
W
uJ ----~
.:.
\Il:!
W;:::.cO 4 .....
lI\
1'.:; t"~
4·2rn 4·2", 4,2"" 4· '2 "" .[ SI'AI-l OF Floofl,.

1 1 (a.) OF. RooF UNITS
SHOWN THUS
I LIFT, STAIRS ETCI D4
~ - - - ; - - -:I: '207>< 1'3 4 1._--,1>=-4-'--_1-:-1
~ US ';t
g g,
FIG. 4.2. Multi-storey shop premises.

(a) Roof plan
(b) 4th floor plan (3rd floor plan simiIar)
1/1
eil 4..,
4·2..., -4.'2"., 4·2..., 4·2""

CG) SPAN O~ FLOOR
I LIFT, SiA1RS ET'I UNITS·RHOWN
H----
I ------ I
I
A4 THliS
1: --I'2.0~~'3~ A4
-
«'" «'"
TIE
H"2ö7Xi34 «
-
« -
(----
"TIE
....
us
'"..:
Oll
«
A.2 c.olS," c: OL 55"
/
~~
",(1751 "IOx 30S" A.7..
~ U& l" <t
01>
«
.::... D751 ,A
<.><
11>
TIE
0
... ilE """VI
H---- ~- « '"
.:(----H
-,.,..• <~
«'" U>
Ö '"«
~
PlATE
~12'-) .::. GIfl.DER) (112'-]
H 10 • "1. H
t COL SI \:'OlS2./ BEAM END REAC"TlONS
(0() ·SHOWN "THU5 [ 1
FIo. 4.2 (cont.)

(c) 2nd floor plan
(d) 1st floor plan
DESIGN OF MULTI-STOREY SHOP PREMISES 29
(b) A typical form of calculation sheet suitable for beam design is
shown in Fig. 4.3. Standardisation makes for simplicity and reduces the
chances of error. Procedure is as folIows:
(i) A sketch of the loading on the beam is made in the calculation
column of the design sheet. The span of the beam between centre lines
of supports is given in brackets, and the effective span of the beam
after deductions at supports is indicated below it. It is an advantage to
have tbis information recorded because these deductions are estimated
and may not be obvious to another person who may wish to refer to
them at some future date. I t is necessary to make all calculations of such
a standard that another person has no difficulty in following them
(e.g. the local authority of the area in which the building operations
take place is likely to require calculations to be submitted).
(ii) The designer then carries out the necessary calculations and
decides upon a suitable section for the beam. Availability of material
will usually be a consideration in the choice of section adopted. As an
aid to later work the calculated reactions are entered in the two right-
hand columns of the calculation sheet and also on to the ftoor plans
adjacent to the appropriate column.
Further points of interest are:
(iii) An estimate for the self weight of the beam and its casing must
be made before the total load on the beam can be calculated. Experience
enables the designer to estimate tbis weight with a high degree of
accuracy. The student, who cannot be expected to estimate this factor
accurately, should make acheck calculation after a provisional beam
size has been obtained (see calculation to beam EI).
(iv) The beams to the front elevation at roof level support only the
weight of casing in addition to their own weight. At the ftoor levels
the front elevation beams support the curtain walling, their own
weight, and the weight of the casing. These relatively small loadings
over a short span of 4·2 m would theoretically result in a beam ofvery
smallsection.Forpurposesofgeneralstabilitya 177·8 x 101·6 x 21·56kg
joist may be adopted. If a UB is preferred a 207 x 134 x 30 kg
may be used. The term 'general stability' in tbis context means that
the design engineer uses his judgement to decide what is the minimum
acceptable size of section wbich is consistent with bis views of what
constitutes a safe and stable structure.
(v) All beams have been designed to a bending stress of 165 N/mm 2
on the assumption that adequate lateral support is given by the detail
of ftoor construction which has been adopted (see Fig. 4.2). It would
be uneconomical to adopt a detail in which no lateral support was
given to the top ftange as it would be necessary to adopt a lower
working stress (see section 18(c». As an example ofinadequate restraint
a working stress will be calculated for beam Al assuming no restraint
"'~ qf C AlCULATION
Po\. Rp, oq.. ~
kN kN <t",r:; ot-'t-
... GALCULATION
Re
kN
RF>.
kN
~",,,, <t-'t'
IB..., SPANI 18 ..... s p,o, NI
4TH D1 ZND 1'1 611
3RO Cl ~~~:M7~~ ~:' 151" Al
C O L r T I n -o
-loO~t
LOAD" LOAD =
7·9" 4 ·'2" 5-4" 179 7·'h; 4 · 2)1. 7-4=245
5tL F W T A"lD CA$ING
SELF WT AND CASIN (Esti",,,.tcd) '" 35
(EStIMQtl1.d) " '29 (k N) 280 140 140
(kN) '20S 104 104
M ~ '208x 7.9 M '" WL -;:, 'Z~O" 7 ' 9
e
M '"
8 8 8
:. 20G kNm ~ 276 kN",
Z=M '" 206,,10

3
Z"" M "" '2.'1b~IOl
f ~ T ~
=1'Z50~",3 ::. I"'SO <:,,,,,3
USE: 457" 1<:10,5)(.74 USE: 533 ~ 209 ,,92
uB UB
184.., 5PANI
62
Wbb
2Nl)
4TH D2
3P.D CZ
15.,. A'2. SELF ZSOkN
1'3-4 SPAN)
2.08kN CO\.
~
'Ol
B·'2,
SE~
-10 - 100
WT)
LOAD (1.
BEAMS Al)
(Ol COl 280kN
:0
-100 8·'2, - 100 WL. == 1.60)<6'2
M "" 4" ~
LOAD(2 BEAMS Cl)
== '208 kN == 574 kNm
M " WL :: 208 X 'O·z 5 ELF WT AN!) CASINr,
4 -4-- (Est.rnG\t<1!d)== 70kN
= 42(, kNm 1'29 129 M = WL. :: 70 x S'L 175 175
S S
SELF WT AND
"" 72 kNm
CA 51NG (,,-s~)= 50kN
M '" WL ::. 50" 8'2 TOTAL M ="4'-k N..,
8 e z==- !::1 =. <.4" ><103
'" 51 kNm f ~
TOl'AL M '" 417kNM
c.m 3
Z. =M =.477,.103 '" 3~20
T I"S
USE.: 1010.305 X \49
'" 2900 "".,3
USE :012 X 229)( 125 UB
VB
FIG. 4.3. Calculation sheet.

Floor beams
1 k"'I't-
K-"; CALCULATION
Re R~
kN Id.J
~t-
1~,,<fJ" .t" CALCUL.AiION
RL
kN kN
RR.
14 ·2,.., SPAN) 14,.,., SPAN)

4TH D4
3R.D (4 COl~C.Ol
-o ~ -O
4TH DS"
3 ... 0 (5 :oo~=t~o~
2><0 B4
LOAD ( BP,IC.KWOII, K)
2ND es LOA D (BFI,ICKWO~K)
l'Öy /'..4 I~i A5
:. 4,,4·25 x 4 · & " ~. 8 x 4· IS x 4 · "
=731<"1
'" 78 i<N
LOAD (FLOOR)
SE L F WT A!<.JD
CA51N6 :: 81<,N -= 4 ·0 x :2·1. 7· 4
43 43 "'62kN 72 72.
86 kN
-- SHF WTA!<.JI)
(.A5IN(; =. BI<.N
M =. WL = 86 x 4·2
-8--
"'8 ~
:. 45 kN.., M~ WL :: 1 43~ 4
= M 4S" 10 1 "8 ~
Z
T
=:
165 = 721<.N..,
=
2. 74 cm~ Z = M =. 72 >< 10'
f 1t<.5
USE: 207" 134" 30 ,: 43(; c. ... J
UB
USE:25",,14{'·4J1 ~7
ROO E4 FoFl. STA8 1L1 T"l' u6
LJs E: '203x I?><l. '25 RooF Es AS foP. AS, 65, Eie
UB
FIG. 4.3 (cont.)

",,,,~ c+- R~ RR, ... RL RII.
~\,; ~~~ CALCUL.a.TION I<N kN <-"cP· I.!-.-'" CALCUL.a."TION kN kN
(4-2 ... 5P ... N) (4 ... SPAN)

4TH D4
:O~~O~
COL~C.O\. 4TH OS"
311.D C4 3 ... 0
-o~-o C5
84 2NO 85
"'4
2>lD
I foT LOAD(8RIC.KWORK)
=- 4x4'25 )(4·'"
IST
"5 LO'" D (BRICKWORK j
.. 3-8 )(4-15 x 4-"
=73kN
-=78 kN
LOAD (FLOOR)
SELF WT AI.JO
CASIN6 =
81<.N
43 43
= 4 -0 )( 2- 1• 7· 4
::. &2kN 72 7Z
- -
86kN
M =WL = 86 )(4·2
5HF WT ANO
CASIN r. =- 8kN
-8--
'8 ~
=- 45 kN M M=~", 143~ 4
2 == M =. 45" 10i 8 -e-
T 165 = 72kN...,
== 274 c...,i Z = .t::1 =. 72)( 10 1
f ~
USE: 207)(134)( 30
UB == 43"- G",3
USE:2S('xI4"'4x n
RoO E4 FoP. STABILITY v6
US E: 203)( 13~)( 25 ROOF ES AS FoP. AS, 85, Ei'
UB
FIO. 4.3 (cont.)

to buckling is provided to the top flange. This will be done by the
method given in BS 449: 1969, dause 19a. Two factors enter into this
problem, the first being the ratio of depth of section divided by flange
thickness (D/T).
I 7900
;;; = 4·34 x 10 = 183
D = 533 = 34
T 15·6
Allowable value of Pbc = 78 N/mm 2 approx.

(BS 449: 1969, Table 3a.)
This is a very low figure and illustrates the importance of adequate

restraint.
(vi) Provided that grade 43 steel is used it is unlikely that deflection
of the beams will require investigation. If higher grades of steel are
used then smaller sections will be possible but the decrease in moment of
inertia will result in increased deflections.
lS Effect of Wind on Building

The proportions of the building are approximately cubic in shape and
together with the stiff flooring and walls, render it unnecessary to
consider the effect ofwind in this instance (BS 449: 1969, clause 10).
26 Design of Columns or StanchioDS

(a) Typical calculation sheets for the columns are given in Figs. 4.4 to
4.8 and the procedure adopted is as folIows. Each floor-to-floor
portion of a stack of stanchions is treated as aseparate length for the
purposes of design. A line diagram is drawn indicating floor heights,
floor loadings, splices, and degree of end fixity. End fixities should
follow the recommendations ofBS 449: 1969, dause 31.
(b) An estimate of the section required to resist the loading may now
be made and followed with detailed calculations to justify the pre-
liminary choice of section. A first estimate of section can be obtained
by assuming that for the majority of cases the ratio of load/cross-
sectional area of the section will be approximately 90 N/mm2 •
LOf'lDING FLOOR TOTAL CALCULATION
DETAILS kN LOADS lOAD
87 87
PIN ROOF 87 41' .... TO ROOF
!.22..~1QL 107
107 kN 5EC1"ION AS FOR 3 R0 1"0 4' ....
I
ß7
lI' 12
3"'0 1"0 4T.... A~ 91·10 ....1.
?r15 ..., 400 400
l" TRY :2.11Q,,20,G,,7Ik, r'f'" 5·2Scm
O-B5L (w)
ee== 0 '7"3750= 'Zc,'Z'jm",
'2.&25 I • 51+4'1 [8
7')= 51..S Pe."'1'33 N "'", T",b\~17'"
)
104 f c :: ~H,,,IO :: ~7'4N/"''''z
1
fiXE 0
~HI w 10'
~+.!l2.
104
Fe. _ ~7-4 _ 0.73
104
129 1% - -m - rlESS THf'l N
l UNIT'"
129
104 5EC.TION IS SATISFACTORY
lf-.l.Q,
Q. '" Q'7 L 4SG '6~G
(w) 'l NO TO 3 RO
SEC1"10N AS FOR IST To 2 ND
104
FiXE D IST TO 2 NO A= 20\·2c.J
1'2') ~ 1'29
104
104 TRY 327)(~\I x158 k<; U.C ry~7·'oc",
I 129 ~:: 0'7)( 4500 :: 3150
!. =. ;150 = 40 Pe = 13' Nimmt
* 20
1'2.9
104
3-75 m i", O·7l r'f 7~ ' 0 3 [651+4'1, Table 17"J
~ 1372, fc. '" 7.022" 10 :: 100 N/m.,.,..z
(W) 7.0'1'2 >< 10 3
rLESS THA:N)
fe. = 100 :::: 0·72 L Ur-ll1''I'
140 Pe. 1'39
140 5ECTION 15 Sf>.TISFAC.TOR.'1"
17<; ~...Q1. 140
175
4·5 m
I
140 ~
175' C,ROUND TO I sr A~ 30S·be.....2
TR'f 353 X 3Uh 240 I<~ UC r)' =8'13c",
2 :=. O'7 l "so f. =0'85)( 5000 =. 41.5'0
t:o ~ = 53 ~c. = 131 Nfr.,..,t
IPPLICE 'f"j 81.:1 [ßSIt't'I Ta.blt. '7-.1
IST
fe. = 2(082" 10
3
= gg
'
N/", ... t
FI)(ED 140
140
f
30·5&_\0)
rt.. l:SS 'THAI-!
1
17':> ~...!..2E. 140
175 ...!.. '" !!. = 0 ·(07 l Uf,jI1'Y
I
140 )f
175
30
pe 131
SECLION IS SAiISF .... CTORY
e-O'SSL " .. 0 '2&82.
(\111 ) ~ REPRESEr-ITS SElF VVT
(,ROUN~ AND CASING
PIN 0"
)~'"
FlG. 4.4. Calculation sheet.

Column S6
LOADING FLOOR ,OTAL
DET AI LS kN LOADS LOAD
C.ALC.UL,6.TION
"H IS COLUMI,J IS RES,RAINEO ON /S"I'TO Z"o

'\-110 Y AXIS 6Y 8RICKWALLS ANO ,R'!' 254 X 254)( 73k'3 UC
CANNOT FAlL S'I' I3UCKLlN(, ON A '" 9'Z·':lcm~ r" '" l!'05c"" r y = {"55'",
THIS AXIS. CALCULATIONS ARE D " 254,.,,,, ' " 14·'2",,,, D/i" 18
GIVEN 1'0R LOWER TWO LENGTH5 e O'7~ 4500" 3150
1
'0
RooF 122 22 e/r~ '" 3150/110.5 '" 29
-t
22 Pe = 143N/ ",..,~ fB5449: 1%9
PIN 107 107 kN
Pb< '" I"';>N/",m~ 1,I\!lLES 3a, 17<>.
~·75 m ':O.ß5L \"3
*'2 /<03
ej<= 1>/2+100= 11.7+100= 7.7.7
L 22- - DIVIDE M AT FLOO"'- EaUALL Y
8ETWEEN UPPER AND LOWER
5(; LEN GTI-IS 0'" C.OLUMN
I
5b
4 TK 5G fc::: 10 14.,0 3 = 10' N/m",,'
\ 129 ",,29>< 10J
flXED -I--:-i 1 2.~
fb = Mx :: 175" 10 3 )< ZZ7
~ C)< "'Z')( '2 >< 89';>)( 10 '
3'1! m
1% 2.53 41"
tb v '" 0 DuE
TO SYMMETI~'CAL
= 22 N/m,.,~
11 SPLlCf t =O'7L G, LOADING ABOUT Y A)(15

% CHEC.K STRESS RATlOS
3 RD
I~
FIXEO
129 I fe '" 109::
Pe. 14 3
O· 7G2
{~:J6 '-75
tbc
~be
= ~
1&5 .
= O' 133
Q. a95
L ESS T~AN
UNIT'I'
3·75,..,
SECTION IS 5A'TISFACTOR'!'
2I-JO 6RoUND TO 1ST

72
I~ IN A 51MlLI\R MANNER To A6DVE
72
lRY 2.<07" 258" 107 kg lJG
F\XED 175
:f ..--.1E. D/'T = 2"~O'5 '" 14
4·')m
33':) 1014 t;, _ o· es" 5000 = 38
r" - 113
Pb<: = 1105 N/mm< Pe'" 140 N/mm<
IIIST 1
72 7Z fs = [3"3" 103 0·714
I~
175 72. P" 1'3.(,(, " 10 3 " \<10
'FIXED --1-;-1 175
fb<. _ 175 x 103 " 233 =
*~ Pb< - 2~1313 " 10 3 " 11>5
4·5 In 172
{ 'OQ·ß<';L SE<.TION 15 SATISFACToRY
GRD
)(. REPRESENTS SELF WT
0·5 "" A>JD GA51NG
PIN
FIG. 4.5. CaIculation sheet.

Column S5
FLOO~ "OTAl CALC.ULATION
DETAILS LOAOING LOADS LOAD
ROOF 3 RD To ~OOF
AS FOR 52, 53, 54
PIN
1'57" 1<;3 l( ~O k9 lJC
IST TO ZNO
3-75'"
47. TR'< 1~'2 X 154 X 37k9 UC
A'= 47-4c"'Z r)C:o «.-63c", fy= 3 · 86..
D=- 1102"'''''. T= tt·5"'M Dj,.-= .15
t ~ 0'7>< 4500 =3150 ""'"
510 I " 5/ö

G
e/ry :. 31"0/38./ö
Pe = 101 N/... ",Z
Pb<.::' IlöS NI"""'?"
=- 82
[6544' : 1%9
TA6LI<S .3"',
1
17a.
H-- flG e)< '" Dh,+loo = IBI.,.,,,, ey", 100
3 -1~ m
R 120 [6'544': ""', c.1 34]
DIVIDE M 1>1,. FLOOtö: E.QUALLY
I !/5PLICE
BE,.WEEN UPPER "'ND LoWER
L ENE>THS OF COLUMN
fc. '"301" 10 3 :: "4 N/"""'z
3 RD "l'74~ I 03 z
56 hc '" 7" 103 " 181 = 23 NI"''''
1
H~ )C 2" 274" 10 3
202 fbc :: 72)( 103 " 100 '" 39 N/.,.m'
3·75",
'I ';l." 'I·S" 10 3
CHEC.K STRESS R"'TIOS
fc _ "4 = 0·(;34-
PZ-jOT
= o·
7~
fbc(X) '2.3 :: 139
72 PI:>" Ilö5
1
7 *~ ~(y);-. ~ = 0·2%
J-!-I-- 99 301 I'bc. I/öS I· 009
5LIGtI,.L'( IN EXC.E5S OF ulJliY
BUT IS 5ATISF",C.ToRY
FoR
Gf1..D To IST
SEE CONTINUA"ION SHEET
1\FIXED
11
* REPRESENiS
AND CASING
SELF : Wi
11
11
4.5", TO
6?'OUND
FIG. 4.6. Calculation sheet.

Column SI
FLOOR ToTAL CALCULATION
DETAILS LOADING LOADS LOAD
301 CONTINUED
IRY 353)( 318 X '2.401<.9 uc:.
A = 305·" cm1 r" '" 14 · 5G'" ry:8' 13'"
e)t= 17"+100= 276",,., lZ.y = 100",,,,
D""' 353.,,1'" T= 37·7_ J)/T'" 10
f == O· &5 .. SOOO = 4150 .......
f/r'j '" 42So/e . 13 '" 53
Pe. = 101 N/mm'" r8SA4~: 1969
Pbe.;: 1"5 NI"''''''' lTA8LE5 34, 17e..
1
DIYIDE M AT FLOO'" LEVEL IN
4·5", TO Pp,oPORTION OF UPPE RAND
2 .. 0 LOWER C.OL.UMN STIFFNESS
~'://.t (upper) '" ~"= '2·25
3150 .. 10 3
I.Y/.{ (loW"') =' 1.0"2·3,)< 10"= 47'"
4'250 " 10 3
72 1/25 M A BOYE \ ST F LOOR
112" 24/25 M BELOW IST nOOR
)l" 30 {c. ;: ~l ;: 50 NI... ",'"
120Ze 1529 30·5& .. 103
fbe;: M~ = "H." 10 3" 27")( 24
4· Sm x Z.. 3"41" 10' 25
,. 82 N/",m~
fbc.:: My ;: 72,,10',,100)C 24
Y 'Zr lOZ73 w 10' 25
~ Go N/... ","
GRO CHECK STRESS RATlOS
0·5 In fe. ;: 2.Q.. '" O· 3B'2
pe. I '?I I
fbe.,v - '02
Pt>e. v ) - ~
= 0 · 497
f..!::...(y) '" S!.. ;: O· 03"

LESS
Pb<. 1&5
0·915 THAN
UNITY
SEC.TION IS SATISFACTOR'(
FIG. 4.6 (cont.)

FLOOl'. T01' AL
DI:TAILS LOAD IN G L O~DS LOADS (ALCU LA lloN
kN kN 41'H TO ROoF 5 EC TION A S FOR

87 31\0 To 4 TH
3 RD TO 4.,.H
ROOF
, PIN I
~I~ "".22,.
S7
<I-
4
TRY 157" I S3 ,,30l<~ UC
A "3<a·'Zcrn 2 ry '" 3·81 ",., PfT" 17
e= 0·7 x 3750 ~ 2(,25
3,75 ""
.e" Q · 85L
110 110 ejy~ ~ %2%'0. 1 = 69
!>c'" 11 7 NI.,."",,'" [8S4-4-'1·. 1'1 (,'1
Pbc. '" 1(,5 NI"' ....' llQ.bleS 3~, 17..
1
e" '" 7St" 100 == 178
104 fc. -= 241 " 10 3 = «,3 NI"'''''''''
4"" 104 3·S'2~IO~
I
1 •
b fbe = M~ = 104.IO,,178=42N/Mft
C; X 2. Z" 2 ... 12 1 ~ 10'
-"-I~ 'U?- fbc " 0
375,.,., 1'31 241 i
5PLICE t '" 0'7 L CHECK STRESS RAT lOS
fe '" hl '" 0·538
pc 117
11/ 104 tbc. = 41. = o· '255 LcSS THA~
,RD 104- P~c. iZ5 0 · 793 UNIT'I'
I b 5ECTI oN rs SATI5FACTORY
~I~ ,.,. 20
~ " O· 7L
"
13~ 377
IST TO 2 ND
TR Y 210>< '205" 59 K9 UC
DIT = IS l J ~ 0 · 85" 4500 '" 74
3·75..., 'y 5 1· 8
I>c '" 111 NimM'
P~c. = 1"5 NI", ...'
2'<1" 140
140 CHECK ST RESS RATlOS
7
I 7
fc _ 55" .. 10l
(7.5 g~IO·}x\l1
'" O. (,(0 I
w..
Pe. -
2-I---2...
~ 5% fbc '" 140" 103",~ '" 0 .1 49
4·5 "" ~ " 0 ·8S L t>bc '2,.. 581"10,, 1"5 0.810
SECT ION IS SATlsFACTORY
~REPIl.ESENTS SELF WT AND

CA!'>ING
HG. 4.7. Calculation sheet.

Column S2
ROOF <1' ~ GI
'"'"- ou
<S'
Y. Y. O:t 3 -75 ,.,
"'ou <" - v
4 Tt1 ~~J ~ ,...,? ~r-:>
~x .... )(
'!"
'-" 3 ·1S ...
11 '-"
1'1
l _SPLlCE
3 RO I
<S'
'1<1' Y.
'ÖO'
",-" o ,.G'
\1\
",ci
;;:;tOv 3-7<;'"
- .... v !" <1'> v "'mv )t~:)
2"'b Xf")? x .n;) ~ r- J
~~ ~
('< " ;:;"
'"
_SP L IC.E--.
:;:Jx.
'" 4·5 m
1ST I
'"
~~v ~fu 00-"
- ou '"
"'..t" 4 ·5
m
", ... ::J ~o:>
(,RD xl'<
~)(
''"'
Z:" 0·5
-~)(
'"
MAP.I<. 2 5 Go
BASE LOAD 15 29 55(, 13(,3 2(,8'2.
(k N)
FIG. 4.8. Column schedule.
(c) The stress in the column is made up oftwo parts:

(i) Direct stress (fc) = load on section/area ofsection. This results in a
tendency for the column to buckle and the stress must be limited by
making use of a column formula (see section 19). A convenient method
of obtaining the value of!a is to make use of the tabulated information
contained in BS 449: 1969, Table 17. The ratio fe/Pe gives the stress
ratio due to direct stress.
(ü) Stresses due to bending. When floor beams connect to stanchion
flanges the stanchion receives the load at an eccentricity from the axis
of the column. The eccentricity of loading results in a bending stress
being induced in the column. Stress nduced by bending =Mz/zz
+ Mv/zy. From this it will be seen that the stanchion resists bending
in much the same way as a beam, and the permissible bending stress is
the same as for a beam (165 N/mm 2 for Grade 43 steel) provided that
adequate lateral restraint is available. The eccentricity of a beam load
upon a co lu mn flange is taken as half the column width plus 100 mm
(see BS 449: 1969, clause 34). The ratio he/Pbe or fbtlpbt gives the
stress ratio due to bending.
For stanchion equilibrium the sum of the stress ratios must not
exceed unity.
!a/Pe + he/Pbe must not exceed 1·0
(See BS 449: 1969, clause 14). If wind stresses are a factor then
BS 449: 1969, clause 13, will also apply.
(d) Four of the stanchions in the structure have been chosen as design
examples. SI is an example of eccentric loading on both axes, S2
carries eccentric loading on one axis, S5 is an example of a stanchion
stiffened by adjoining construction and S6 is a simple example of
concentric loading.
(e) The critical 'z and '1/ values for the various stanchions require
explanation. Stanchion S6 relies on deep floor beams at each floor
for restraint and the critical value is '1/' Stanchion S2 is restrained by
floor beams on the x axis but the curtain walling on the face of the
building cannot be said to give any effective restraint on the y axis
and the critical value for S2 is T1/' Provided that the exterior wall which
restrains stanchion S5 on the y axis is of first quality construction
there is no reason why this fact should not be taken into account,
making the critical value for this column Tz. Stanchion SI is not
adequately restrained on the y axis by the exterior wall and TI/ must be
used in this case.
(j) At each floor the total bending moment in the stanchion is shared
between the upper and lower lengths of the stanchion (see BS 449:
1969, c1ause 34). This fact accounts for the divisor of2 in each fraction
of M/z except for column SI where the moment must be divided in
relation to the stiffness ratio of each length above and below the floor
level.
(g) The estimation of effective lengths of columns based on the recom-
mendations of BS 449 may cause some difficulty to the student and
the following notes may be helpful. Careful answers should be con-
sidered to two questions. Firstly, is the adjacent structural framework
capable of holding the end of the column in a fixed position? Secondly,
is the adjacent framework and continuity of the column from one
floor to another capable of holding the column ends in a verticalline?
The first case iIlustrates 'fixing in position only' and the second,
'fixing in position and direction'.
(i) Columns which are held in position by a grid of floor beams and
continue through to floors above and below are treated as having
both ends fixed in position and direction, with an effective length of
O' 7 times the length between centres of connections.
(ii) If one end is fixed as above with the other end held in position
but with the column not continuing (as in the case of the top and bottom
lengths of a multi-storey column) then an effective length of 0'85 times
the length between centres of connections is appropriate.
(iii) This previous recommendation means that the base of the
columns is treated as a pinned end and no fixing restraint from the
foundation block is taken into account. The design of the foundation
block is simplified as it will be required to support only vertical load
without the necessity of estimating the size of column base or concrete
foundation required for restraint.
(iv) Reference to BS 449: 1969, Appendix D, Fig. 8, gives a recom-
mended value for the effective length of 0·85 times the appropriate
overalliength. For column SI an effective length of 0·7 the appropriate
overall length is taken because of a large concrete casing which gives
additional support. Normally an effective length coefficient of 0·85
is appropriate in this case.
(v) If the conditions of column end restraint are in doubt it is advis-
able to err on the safe side and use a larger effective length.
27 Design of Welded PIate Girder

BS 449: 1969, clauses 20, 27 and 28, contains many recommendations
regarding plate girder design and should be read in conjunction with
the design given below.
A first assumption is that the bending moment is resisted wholly by
the Banges and all shearing forces are resisted by the web, this assump-
tion being approximately 90 per cent true. The first stage is to decide
upon a preliminary layout and then to submit the proposals to a detailed
scrutiny.
1. Proposed dimensions 0/ girder (Fig. 4.9)

(a) Depth. For economy, the ratio of depth to span varies between
1: 8 and 1: 12 depending upon intensity of loading and span. This
girder may be described as heavily loaded and a depth of span/8 will
be used.
Depth of girder = 16·8/8 = 2·1 m = 2100 mm
(b) Weight. The weight of the plate girder must be estimated and in
view of the variables present in each individual ca se (e.g. grade of
steel, means of fabrication, relationship between span, load, depth,
etc.), it is better to rely on judgement rather than on an approximate
formula. In this case a self weight of 70 kN will be adopted for the
dead weight of the girder representing about 3 per cent of the total
load carried. A concrete casing accounts for another 80 kN making a
total dead load of 70 + 80 = 150 kN.
(c) Size 0/ ftanges. The compression Bange must be provided with

adequate lateral support and an adequate width must be provided for
(ONNECI10N FOR.O 1]. 4 W' Gm xc -:]
(OLUMN 5'2 - 3L_ --,!V_'_'_'_ _'_____~__<
t'AK T PLAN ON TOP FLANGE ~__ . __ PAR.T PLAN oN BOTTOM fLANGE
Al I - . . --_. - -
350,,25 PLATE/ L IJ 11 J 1
,'1 f!' l.350" 50 PLATE,
I üoor '1100(''''
50 5050 so 50 " '1 50 50 50
r 1 8 FlllEl WELD:, 1CONNECTION 2000>< 1'2.
E
f 10O r-~ c;, I
E r- FIHET I, I FOR BEAM AI WEB
100
IOOP WELO , WEB_ 0
v" ~Cl .- D JOINT 0
/ c " 11 D ("l
100
~ 900 !! 15o" 12
'-- :1
150" 1'2.! FLAT
1 350" 25 PLATE) 'aAl ilI1 ~ /350 )( 50 PLATE /
1-
4'2.00 4'2.00 10 (,L. GIRDEf!,
1)""'; 20 CAMBI R
~ DER
'''M
AI ~ ~ffiß'oo ELEVATION. ATCLG"
60 PREPARATION
r=:q,= 150,,12 C1
3S 3~3 1 8=: 40,,40.:-1--"' I .....
FLAT 1 tl2 Y 1
u: CHAMfER
5 I
rW< wEB .JOIWT WELD
/'" SEC.HON C-c
/ ' 50 x 12 FL,6.T
~;SO" '~ FLAT
)
SE.c..'ION D-D
\ :25 &~
fiiif\(SLOPE
IIN5
~ I>: ~ ~ i II I FLANGE.. JOINT WElD
SEC.TION A-A SEC.T I ON B·B U
100
SEC.'ION E-E
FIG. 4.9. Details of welded plate girder.
this purpose. Empirical rules exist which may be used as a guide. Two
such rules are:
·d h f fl span of girder
Wl t 0 ange = 40
. h f fl distance between lateral restraints
or wldt 0 ange = 20
In this case
span = 16800 = 420'mm
40 40 .
distance between lateral restraints 4200
20 = 20 = 210mm
These are only approximate guides and must be treated as such. A
flange width of 350 mm will be adopted provisionaIly.
The size of flanges will also depend upon the allowable stress and a
figure of 140 NJmm 2 will be assumed (see BS 449: 1969, Table 2).
This stress must be verified at a later date when the proposed design
will be submitted to a more rigorous scrutiny.
An approximate area offlange may be calculated using the assumption
that the moment of resistance of the girder is equal to the moment of a
force couple in which the force is that contained in one of the flanges
and the lever arm of the force couple equal to the depth of the girder.
This caIculation will require the bending moment to be known and it is
appropriate to caIculate bending moment and shearing force diagrams
at this stage in the design (Fig. 4.10).
Moment at centre of girder
= 1126·5 x 8·4 - (701 x 4·2 + 75 x 4·2)
= 6195 kNm
Moment of resistance of the girder
= flange force x girder depth
Flange force = area of flange x allowable stress in flange
= AJx PbC
Girder depth = 2100 mm = D
then M = MR = A x Pbc X D
_ M _ 6195 X 106 _ 2
or A - Pbc X D - 140 x 2100 - 21000 mm
It must be emphasised that this figure is an approximation and gives
results on the high side. Taking this fact into account a flange size of
350 mm x 50 mm will be adopted.
COl52 556
BE"M AI 140
5
• G 8 ) 70lkN 70lkN 70IkN/GlAZING
C."SIN 0 Isolc.N . 1 11 70lkN
SEI.FWT 70 \ ~~~
~F~==4''2.~M.i~.4:::::::'2~M~~4.2~""".~I4==''2~'''1'
~ 1"·8m :
112~'5 kN LOfl.D DIA6RAM 11'26·5 kN
-.----------:----==-;:-;-"..---- -- ---,
....
.... ,
....
- - - - - __ =::"l
M DUE
TO U.D.L. BEt-.JDING MOMENT DIAGRAM
,,~I...-____ _L_+_----_-_,
.5IiEAR1NG FOfl,CE. DIAGMM

c.l. GIR.DE" OF 61RDER WITIt
'-
4.510..,
350,. 50 PLATES
AREA OF GIRDER
UMI.OF WITH 3&0 x 25 Pl.ATES
350)(50-
PlAiEs
AN'\' CONVENIENT ANGLE

FL~NGE C.U"iAILMENT DIAGP.AM
FIG. 4.10. Welded plate girder. Bending moment, shearing
force, and flange curtailment diagrams.
(d) Thickness 0/ web and spacing 0/ stiffeners. The thickness of the web
will be found by applying the rule:
. . b h' k depth of girder
MInImum we t IC ness = 200
2100
=--
200
= lO·5mm
A web thickness of 12 mm will be adopted.
Various rules exist for the spacing of web stiffeners but frequently
the best layout is settled by circumstances. In this case bearing stiffeners
will be required at the girder supports and under each concentrated
load occurring at the span quarter-points. Intermediate stiffeners will
be provided equidistant between the bearing stiffeners, and this will
give a spacing between stiffeners of 2'1 m. Two empirical rules used for
stiffener spacing are:
(i) Allowable spacing = 270 X web thickness
= 270 X 12 = 3·24m;
or (ii) = 1·5 X depth of girder
= 1·5 X 2100 = 3·15 m
A width of stiffener equal to 150 mm will give good support to the
outstand of the flange plates and the thickness should be one-twelfth
of this figure, say 12 mm. (See BS 449: 1969, clause 28.)
The provisional design is now complete.
2. Verijication 0/proposed design

(a) Value 0/ compressive stress, Pbc. In designing a provisional girder
a value for Pbc was assumed and it is now necessary to check this by
comparison with the recommendations of BS 449: 1969, clause 20.
This method allows for the fact that the resistance of a beam, or
plate girder compression flange, to lateral instability depends partly
upon the resistance of the member to torsion as weIl as resistance to
bending. Before the compressive stress can be found by this method,
it is necessary to calculate the moment of inertia of the girder about the
yaxis.
Inertia of 2 - 350 mm X 50 mm flanges
= 357 X 10 6 mm i
The inertia of the web is so small that it may be ignored.
Total inertia about y axis = 357 X 106 mm'
ry = JAI= J35759 X
X
106
103 = 77·6 mm
Distance between lateral restraints

= 4200mm
An effective length of 0·85 X 4200 mm will be adopted which corres-
ponds to a flange held in position at quarter-points of span but not
restrained in position by the beam connections at these points.
0·85 X 4200 = 3570 mm
.!. = 3570 CI: 46
r ll 77·6
D 2100
T= 5<)=42
From BS 449: 1969, Tables 7 and 8, A = 1376 (approx.), C. = A =

1376 and the appropriate value of Pbc = 150 N/mm 2 (reduced to
140 N/mm 2 because the member is over 40 mm in thickness and is
part of a plate girder).
Note regarding the provisions of BS 449: 1969 for obtaining values of
Pbcfor plate girders.
The method consists of applying a formula to obtain a critical stress,
and app1ying a load factor to give a working stress (BS 449: 1969,
Tables 7 and 8). For plate girders with equal flanges:
C. (critical stress) = A (see Table 7)
If flange curtailment takes p1ace between points of lateral restraint the
ratio DIT is modified by app1ying a factor K l •
If the girder in question has unequal flanges a further factor B is
required making the basic formula:
C. = A +B
Additionally the factor B is modified by the quantity K 2 which is
dependent upon the relative inertias of the two flanges. The general
procedure is as folIows:
(i) Design a proposed girder section.
(ii) Calculate Ilr11.
(iii) Calculate value of DIT using the appropriate values of N and
Kl in assessing T.
(iv) Obtain values of A and B from Table 7.
(v) Calculate K 2 using appropriate values of M.
(vi) Obtain critical stress C. appropriate to cases (i), (ii) or (iii).
(vii) Obtain value of Pbc from Table 8.
These steps apply to all girders likely to be encountered in practice
but as the majority of girders have equal flanges the procedure is
simplified to the steps shown previously for the girder under present
consideration.
(b) Depth 0/ girder. It is now convenient to verify the depth assumed
for the girder. One formula favoured by some engineers is given below.
= 1·1 ) ' -
Economic depth
VJbCMX t
)
= 1-1 )(6195 X 106 )

140 x 12
= 2020 mm
where M = bending moment on girder
Pbc = allowable compressive flange stress
t = web thickness
This compares favourably with the previous estimate of span/8. In
practice it is not always possible to adopt a precise economic depth
because of headroom, deflection, or transport considerations. In tbis
case the estimated depth of 2100 mm is acceptable.
(c) Deflection 0/girder. The deflection of the girder may be calculated

by traditional theoretical methods Ce.g. analysis based on the work of
Mohr). Many engineers contend that this method does not give results
that are borne out in practice, since no account is taken of the 'give'
wbich take place between the parts of the girder. A simple deflection
formula acceptable in practice is:
L2
Deflection = k x D X f"c
_ 16·8 X 16·8 X 140 _ 2 .

- 800 X 2.1 - 3 5 mm
where L = span in metres
k = constant = 800
D = depth of girder in metres
!bC = stress in flanges (N/mm2)
Tbis deflection approximates to span{700 and is satisfactory bearing
in mind the situation of the girder. Excessive deflection would not be
desirable in these circumstances and it would be advisable to warn the
shop-front designers of the deflection that might be expected. Generally
a deflection of span/400 to span/500 is acceptable in plate girder
design.
Tbe effect of deflection in the girder upon the structure may be offset
in two ways:
(i) By building a camber into the girder. This would normally be
done by shaping the web plate to give a parabolic camber equal to
approximately two-thirds of the calculated deflection. This figure is
suggested because it should be remembered that only half of the loading
is due to superimposed loads caused by temporary degrees of occupa-
tion of the building. In this way the camber provided in the girder is
matched with the working load ofthe building rather than the maximum
load.
It will not be necessary to set the flange plates to the required camber
since they can easily be drawn into position during the fabrication of
the girder. A further practical point to remember is that the depth of the
web plate required for fabrication purposes will be the net depth plus
the agreed camber.
(ii) By packings placed under the bases of columns S2, or by
increasing the length of the columns to offset the deflection of the
girder. This alternative method of compensating for the effect of
deflection may be preferred to the one above on grounds of simplicity.
In order to maintain uniformity of level at the first floor it will be
necessary to position the connections for beams Al to allow for the
girder deflection.
(d) Moment 0/ inertia (second moment 0/ area). The moment of
inertia of the proposed section will be calculated and compared with
the moment of inertia required.
Inertia of 2 - 350 mm X 50 mm thick flanges = 36·66 X 10 9
Inertia of 2000 mm X 12 mm thick web = 8·00 X 10 9
Total moment of inertia of proposed section = 44·66 X 10 9 mm4
This is a welded girder and no deduction to inertia need be made for
holes in the tension flange.
· reqUlre
Inertla . d= -
Mxy
j-
6195 X 106 X 1050
= 140
= 46·4 X 10 9 mm 4
The inertia provided by the proposed design is slightly less than that
required, but the two are sufficiently in agreement for the discrepancy
to be acceptable. The maximum stress is given by:
M X Y 6195 X 106 X 1050
-/- = 44·66 X 109
= 146 NJmm 2
(e) Thickness 0/ weh and design 0/ hearing stiffeners. Calculate average
shear stress in web of girder.
Maximum shearing force on web occurs at the girder support
= 1126·5 kN
Cross-sectional area of web = 2000 x 12
= 24 000 mm2
. 1126·5 x 103
Average shear stress In web = 24 x 103
= 46·9N/mm2
The allowable average shearing stress may be caIculated as follows
(BS 449: 1969, Table 12a).
~ = 2000 = 166
t 12
Distance between stiffeners = 2100 = 1.05 d
Depth of web 2000
From Table 12a, allowable average shearing stress = 83 N/mm 2 •
This indicates that a large measure of understress occurs in the
web and that it might be possible to reduce the web thickness to say
10 mm, or to try other arrangements of web stiffeners. Whilst this is a
possible line of action it is desirable to maintain a web thickness of
1/200th of the girder depth.
/(20)< 1:)",
240 240
= -t~-::i:::::::!:!:==;f:==i'~~
3/ 150 >< 12 FLAT / "
WEB 5TIFFENER5
FITTED Ta ToP HANGE
FIG. 4.11. Horizontal section through web stiffeners at quarter points

of span showing the effective area resisting buckling. '
The bearing stiffeners under the points of concentrated load (at

each support and quarter point across the span) will be designed as
columns with a section equal to the area of the stiffener and a length of
20 times the web thickness. (See BS 449: 1969, cIause 28.) Bearing
stiffeners enable the concentrated loads to be transmitted to the web.
To enable a connection for the Boor beam to be accommodated and
at the same time to preserve symmetry, a stiffener seetion as shown in
Fig. 4.11 will be investigated.
rotalload at quarter point = 701 kN.
Inertia of stiffeners about centre line of web
= 3 X (b
-3X d3)
- = 40·5 X 108 mm4
nertia of the portion of web is negligible and is ignored.

Area of stiffener plus portion of web
= 11160mm2
Radius of gyration about centre line of web
= JA.
I
=
J( 11-16
40·5 X
X
108 )
103 = 60mm
Effective length = 0·7 X 2000 = 1400 mm

I 1400
;:= 60 =24
pe = 145 N/mm2
(BS 449: 1969, Table 17a.)
·"fti
Actua1 stress In sti ener = 70111 X16010 =
3
I
63 N mm2
The allowable stress (Pe) is greatly in excess of requirements but a

reduction in area of stiffener would save little material. It is preferable
to maintain symmetry of detail as weIl as simplicity of connection
rather than to pursue theoretical accuracy"
Because they are used as bearing stiffeners this group of flats require
to be accurately fitted to the underside of the top flange so that the
load applied to the top face of the flange will be transmitted in bearing
through the flange to the top face of the stiffener. To allow the weId
between the flange and web to pass uninterrupted, the stiffener will be
cut away as shown on the detail drawing and it is necessary to check
that sufficient bearing area is available.
Area of stiffeners = 3 X (110 X 12) = 3960 mm 2
" stress on end 0 f"fti
Beanng sti ener = 5563960
X 103
= 140N/mm2
=
Safe bearing stress 190 N/mm2
(BS 449: 1969, Table 9.)
(Tbe load of 556 kN equals tbe column load and tbe remainder of tbe
load is taken by tbe beam web connection Al.)
Tbe bearing stiffeners at tbe girder supports can be designed in a
similar manner bearing in mind wbat metbod of support is to be
150)( 12 FLAT
100 I- PLI'.TE
ERECTloN
(,IRDtR
35 3 x 318 / CLEARANCE
1 uc.
I" 111
{,
(
\ iL1\'/ 251
1
"
FILLE
WELD
h.
~ /~ 11301 2~
I
b 11"'" LE1NGTH OF W
RESISTING CRU SHING
EB
BRACKET
500 WIDE
"
37~
P 1<0
I
1""---- FI LLET
WELD
HG. 4.12. Elevation of plate girder support.
i, J , , ,. ==~f~~~
100 240
I I (20xt)'!
ISOx 12 FLAT STIFFEJ..JERS

FITTED To 60TI-I HANGES
FIG. 4.13. Horizontal section througb web stiffener at plate girder
support showing the effective area resisting buckling.
provided (see Figs. 4.12 and 4.13). An unstiffened bracket is shown
with a 30 mm thick flat welded to the column and providing sufficient
bearing area to support the girder reaction load. To obtain the length
of web resisting bearing, or crushing, a line is drawn at 30° from the
edge of the bracket (see Fig. 4.12). It will be seen that the stiffener
layout comes within tbe dispersal limits and fuH bearing load will be
assumed as being resisted by the stiffener. These bearing stiffeners
should be fitted to the girder flanges, top and bottom.
(f) Curtailment 0/ flange plates. From the bending moment diagram
(Fig. 4.10) it will be observed that the bending moment diminishes on
each side of the centre of the girder and the area of the flange plates
can be reduced as the bending moment decreases. In the span shown
one change of section would be reasonable and it is proposed to reduce
the flange plate area from 350 mm by 50 mm to 350 mm by 25 mm
at some point in the girder span.
Using the simple geometrical construction shown in Fig. 4.10 it is
possible to determine the theoretical point of curtailment. To obtain
good dispersal of stress beyond the theoretical point of curtailment the
thicker plate should extend for a further 600 mm at each end.
Moment at centre of girder = 6195 kN m
Moment at quarter-point = 1126·5 X 4·2 - 37·5 X 2·1
= 4651·3 kNm
Area of girder with 50 mm flange = 59 X 103 mm2
Area of girder with 25 mm flange = 41·5 X 103 mm 2
From Fig. 4.10, length of 50 mm thick plate required
= 9·12 m
Add 0·6 m at each end to allow for dispersal of stress to the reduced
section.
(g) Weight o/girder. The weight of the girder is as folIows:

Weight in flanges
= 2 X 0·35 X 0·05 X 10·5 X 7860 = 2890
2 X 0·35 X 0·025 X 6·0 X 7860 = 824
3714 kg
Weight in web
= 2 X 0·012 X 16·5 X 7860 = 3110
24 X 0'150 X 2·0 X 0·012 X 7860 = 678
3788 kg
Total weight in girder +

= 3714 3788
= 7502 kg
(75,02 kN)
This confirms that the original estimate of the weight of the girder as
70 kN was sufficiently accurate and no adjustment need be made to the
ca1culations.
A further point to note is the similarity of the weight of material
contained in the flanges with the weight contained in the web and
web stiffeners. For a plate girder to be economically proportioned it is
necessary for these two weights to be approximately equal.
(h) Details oJ welding required Jor connection. The amount of welding
required to transmit load from one component part of the girder to
another may be calculated as follows:
(i) Connection of flange plates to web. The horizontal shearing force
per metre run of girder at level of connection
=SXAXYXI03
I
where S = vertical shearing force on girder
A = area of flange
Y = distance from neutral axis to centroid of area A
I = moment of inertia of girder at point of con-
sideration
There is a large reduction of shearing force at each quarter-point of
span and it is desirable in the interests of economy to investigate the
shear in the end quarter and middle half of the girder independently.
Consider end quarter of girder:
. 1126·5 x 8750 x 1012 X 103
Sheanng force per metre = 25.6 X 109
= 390kN
(Note that the values of land y are not the same for the centre of the
girder where the flange plates are thicker.)
Allowable load on 8 mm fillet weId per rnetre run at a stress of
115 N/rnrn 2 on throat size
= 0·7 x 8 x !15 X 103 = 644 kN
10
(For definition of throat and leg size of welds, see Fig. 4.23.)
Length of weId required per rnetre of girder
= !: = 0·606 m = O· 303 rn each side of web
If 50 mrn long welds are placed at 100 rnrn pitch a total weId Iength
of 0·5 rnetres per rnetre Iength of girder is provided. The extra weId
allowed is desirable for the purpose offinishing and starting each weId
length.
Consider rniddle half of girder
. 388 X 17500 X 1025 X 103
Shearmg force per rnetre = 44.66 X 109
= 157 kN
Allow load on 6 rnrn weid per rnetre run

_ 0·7 X 6 X 115 X 103 = 483 kN
103
Length of weId required per rnetre of girder
157 .
= 483 = 0·33 m = 0·16 m each slde
This amount of weId is relatively srnall and some consideration must
be given as to the minimum acceptable centre of welds required for
purposes of connection. The longitudinal space between welds should
not exceed 12 times the web thickness (in this case 144 mm). If 50 mm
long welds are spaced at 150 mm a total weId length of 0·33 m is
provided.
(ii) Connection of web stiffeners to web. Sufficient welding must be
provided to transmit the whole of the load on the stiffeners into the
web.
Load on end stiffener = 1126·5 kN
. 1126·5
Length of 6 mm weId reqUlred = ~
= 2·33 m
Use 100 mm welds at 200 mm pitch (total 6 m). A similar pattern
may be adopted for the intermediate stiffeners.
(iii) Welds made within the workshop may be required to give the
size of plates required because it is improbable that the lengths neces-
sary would be available through the normal trade channels. The
preparation of the parent metal and the welding should be in accord-
ance with the recommendations of BS 1856, 'General requirements
for the metal arc welding of mild steel'. A joint in each flange and one
in the web has been indicated (Fig. 4.9) and it will be noted that the
joints do not occur in aU the members at the same point in the span.
If possible, the joints in the flanges should occur away from the centre
line of the girder where the bending moment is high, and the joint in
the web positioned towards the centre of the girder where the shear
stresses are lowest.
DESIGN OF MULTI-STOREY SHOP PREMISES SS
(i) Note on erection clearance at end 0/ plate girder. An erection
cIearance of at least 6 mm should be allowed at each end of the girder
because of its size and weight. This cIearance will prove useful during
hoisting and erection of the girder and the rernaining space can be
filled with 3 rnrn thick packings.
28 Design of Welded Connections

For convenience the design of the various connections has been grouped
together. As a general rule it may be said that if welded construction is
used for shop fabrication in preference to riveted or bolted construction
a saving in weight of approximately 10 per cent may be expected.
Site welding is frequently inconvenient and expensive and bolts are
generally preferred in this particular context unless there are special
circurnstances. In brief one usually specifies, 'shop welds, site bolts'.
(a) Stanchion or column bases (Fig. 4.14)

(i) Built-up base for stanchion SS.
Load on base plate from stanchion shaft = 1363 kN
Safe bearing pressure on mass concrete foundation = 2150 kN/m2
(1 :2:4 mix)
. 1363
Area of base plate reqUlred = 2150
= 0·635 m 2
Assurne a square base of side = YO·635
= 0·8m
(approx.)
= 800mrn
A flange gusset thickness and a base plate thickness of 16 rnrn will
be adopted. This is approxirnately two-thirds of the flange thickness
of the stanchion and should give a weIl proportioned base.
After the flange gussets have been tack welded to the shaft of the
colurnn, the base plate will be added and tack welded in place and the
rernainder of the welding can then be cornpleted. However, the method
of fabrication will affect the design procedure. If the end faces of the
colurnn shaft and the flange gussets are machined after assembly, as
weIl as the base plate, then complete bearing is assured between the
various parts. The gusset plates may then be regarded wholly as
stiffeners to the base plate, and the welding provided will be nominal
because complete bearing of sufficient area is present to transrnit
the whole ofthe load. (See BS 449: 1969, cIause 38.)
3
Where no machine facing takes place the whole of the load must be
transmitted by the welding. Some designers contend that this provision
is unduly harsh because some area of the shaft and gussets must be
bearing upon the plate and suggest that 40 per cent of the load is
passed in direct bearing and the remaining 60 per cent is passed through
the fastenings. This will result in smaller and more economical con-
nections and may be applied where the more exacting provisions of
BS 449 are not required.
Adopt provisions of BS 449: 1969 for an unmachined base detail.
Load = 1363 kN
Profile measurement of column in contact with base plate is approxi-

mately 1 m (i.e. the wh oie profile less the two flange faces where
gussets occur).
Adopt 8 mm fillet welds.
Allowable stress on weId across throat (0'7 X leg size) is 115 N/mm2 •
Allowable load per metre run of weid
8 X 0·7 X 103 X 115 _ 644 kN

- 103 -
Load capacity of profile weid 1 metre in extent

= 644kN
Remainder of load to be taken by gusset weid
= 1363 - 644 = 719 kN
Length of fillet weid required = ~: = 1·12 m
If flange gussets 380 mm in depth are provided there will be 1· 52 m

of welding available between the column shaft and the gusset, and
considerably more is available between the gusset and the base plate.
This size of gusset allows a load dispersion from the shaft of about
45 0 , and triangular fillets are included to the base to ensure rigidity
of the base plate and an even distribution of load to the concrete
foundation.
Referring to Fig. 4.14 the critical line for the flange gussets will be
along A-A at the junction of the stanchion flange and gusset plate.
The flange gusset will be investigated along this line for the effect of
bending and shearing.
Upthrust on shaded area taken as a proportion ofthe column load
total load .
= t t I X area of shaded portIon
o a area
1363
= 800 2 X (400 X 270) = 230 kN
k I
I
. I
I
ELEVATION OF BUILT-UP BASE.

A C,USSET
A
/ PLATE
1llir
[W
ßENDING MOMENT DIA6P. ..... M FOR
c,uSSE T PLA,E ",LONG LlNE A-A
AREA OF UPTHRlIST
RESISTED BY ONE
800 ',,,~. ] GUSSET
".'~ 400
~ ... -:.,
~ ;'.
1270j
PLAN
FIG. 4.14. Design of column base.
Bending moment on line A-A

230 X 135
= 103 = 31 kNm
16 X 380 2
Z of gusset = 6 = 386 X 103 mm 3
Stress due to bending = MZ

31 X 106
= 386 X 103
= 81 N/mm 2
Allowable stress = 165 N/mm 2
. load
Shear stress on line A-A = --
area
230 X 103
- 16 x 380
= 38 N/mm2
Allowable stress = 190 N/mm2
The ftange gusset may be considered satisfactory.
(ii) Slab base for stanchion S6. A slab base has two advantages
when compared with a built-up base:
(1) The ease of fabrication and consequent reduction in workmanship

costs although the weight of material may be greater, and (2) the reduc-
tion in depth required below ground-ftoor level to accommodate it.
The size of slab base which in this case is supported by a reinforced
concrete foundation will be decided by the allowable bearing pressure
on reinforced concrete which will be taken as 4290 kN/m 2 • This
bearing stress for reinforced concrete is substantially higher than the
bearing stress recommended for the mass concrete foundation for
column S5.
. 2682
Area of slab requlred = 4290 = 0·625 m 2
If a square slab measuring 800 mm square is adopted an area of

0·64 m 2 is provided.
Slab bases are designed to resist the bending caused by the upthrust
on the underside of the base which results in the slab tending to become
'dished' in shape.
The slab thickness may be obtained by using the formula contained
in BS 449: 1969, clause 38, which is based upon the slab bending in two
directions about the edges of the column.
In this case A = 241 mm; B = 223 mm; w = 4·29 N/mm 2 ; Pbct

= 185 N/mm 2 •
_J3 x
t-
4·29 (24 2
185 I
_ 223 2 )
4
= 56·3 mm (say 60 mm to allow for machining)
Alternatively, by a more approximate method the base may be
designed on the assumption that the line B-B is critical for bending
stress. The larger of the two projections of the slab beyond the edge
ofthe column will produce the greater bending moment (Fig. 4.15).
Jl Ie,
ELEVATION OF SLAB BA5E
"'ND'~"T
ALONG LINE B-B
D"GRAM
Il;
r~
800
r,;:
"
~~
' AREAOF
VPTHP..VST
RE51STED
SY SLAI!o
ON LINE B·B
,"
W
FIG. 4.15. Design of column base.
Upthrust on shaded area taken as a proportion ofthe column load

total load
- t o t a I area x area of shaded part
2682
= 8002 X (800 X 241) = 807 kN
. 807 X 120
Moment on IIne A-A = 103 = 96·9 kNm
z = bt 2 = 800 X t 2 = 133t2
6 6
M=zxj
96·9 X 106 = 133t2 X 185
t = J96'9 X 106
133 X 185 = 63 mm, say 65 mm
A slab base relies wholly on bearing for transmission of load, and

the end face of the column shaft and the top face of the slab should be
machined to ensure good bearing. An increase in thickness of 3 mm
sbould be allowed for tbis purpose. A nominal connection is required
between tbe base and tbe column sbaft and a 6 mm fillet weid is
adequate for tbis purpose.
(h) Stancmon hracket supporting /loor heam. As stated previously,

tbe stanchion brackets supporting tbe beams bave been treated as
unstiffened brackets and it may be assumed tbat tbe verticalleg of tbe
bracket connecting to tbe column is in simple sbear. Tbe bracket sup-
porting beam A2 will be taken as an example (Fig. 4.16).
Load on bracket from beam = 175 kN
Adopt 6 rnrn fillet weid.
Allowable safe load on weid tbroat area of 0·7 X leg size
= 115 kN/mm2
Allowable safe load on weid per metre lengtb = 6 X 0·7 X 115
= 483 kN
Lengtb of weId required to develop load = !~~ = 0·36 m
A 228·6 X 101·6 X 12·57 angle cleat witb welds on eacb side will
allow 0·46 m of weid. Tbe top of tbe beam will require res training
in position and for tbis purpose it is usual to provide a connection
sufficient to carry 2·5 per cent of the flange load in tbe beam.
Flange force in beam
= maximum bending moment/depth of beam
= 646 X 103 /610 = 1060 kN
2·5 per cent of 1060 = 27 kN

Allowable single sbear value of 20 mm dia bolt
= area X allowable shear stress
= 314 X 80/103 = 25·1 kN
Number of bolts required
= 27/25·1 =2
Adopt two 88·9 X 88·9 X 9·42 angle cleats 150 rnrn long to accom-
modate 4 bolts. See Fig. 4.16 for details of tbis connection.
SPLI C.E PL AT ES
NOTES 1'2 THIC.K _
WE:LOS SHOWN )\ 900 L.O~G
4·5 m
THUS ..LLI.LL.LLli. TO 20.0 FLOOR
HOLES '2'2. DIA
DIV I SION PLATES
FOR '2.0 DIA
SITE BOL.iS ..ti.,; - /11iHICK
600
<.LEATS
SS·c:>,,88·9,,9-4 L
,,5 IST FlOOR,
~
rn I~~ I--C.L HTS
I~I'- 152-"'''76·2
A2 ,_ .~~ x 9·4-7 \...
(010"'305----
UB ' h elEAi
152·4. loH.
x 9·47L
C.L"""T /
'Z1. S·")( 101· b
)\ 1"2·57 L
4 ,5 rn
'l~ 7,. 2Sz "I 07k':)
&00 +----1_l_
I UC
ALL fILLE.T WELD

(0 m",
VIEW'X'
" HOL.ES •
I\NCHOP,
~OLTS
24 01A
GROVNO FLOOR
"lEW 'X'
FIG. 4.16. Details of column S5.

The bearing stress on the beam web at the junction of the beam
flange and web will now be checked (see Fig. 4.17). A length equal to
twice the bracket thickness will be taken as the stiff portion of the
bracket.
Stiff portion = 2 X 10 = 20 mm
3
ISS
--,~
-F
I
D~'30 o
5 UB
X. .,
53 2
JU NGTION OF
W EIS AND HANGE
;~ÖhlL
/ ~ :-~
,
~ 20 STIFF PORT ION
OF BRACKET
HG. 4.17. Beam web crushing.
From this point the bearing may be assumed to disperse at 300 to the
horizontal until it meets the junction of the web and flange.
Depth of flan ge = 39 mrn
Length of effective dispersal = 39 x cot 300
= 68mrn
Totallength of bearn web resisting crushing
= 68 + 20 - 3 = 85 rnrn
(the 3 rnrn represents erection clearance).
Allowable load on this length
= length x web thickness X allowable stress
= 85 X 11·9 X 190/103 = 192 kN
Actual load on bracket

= 175 kN
The effect of buckling on the bearn web will now be checked (see
Fig. 4.18). The stiff portion of the web resisting buckling rnay be
taken as that shown in the sketch. Frorn the edge of the stiff portion
a dispersal line at 45° is taken to cut the neutral plane of the beam to
give the length of be am web assumed to resist buckling.
Length resisting buckling = 20 + 305 - 3
= 322mm
Effective depth of web which rnay be assumed as acting as a column
resisting buckling
= O' 5 x depth of beam
= 0·5 x 610 = 305 mm
lE.N6TH OF
3 322/iWEB r-E 5I STI NG
ßUCK LI NG
D
D"IO"30~
lJB
<
2"
I
, 61 o
/ " D
,-\45' "2
/
\
I ~ ~oS
~20 SilFF PORTION

OF e>R.I'ICKET
FIG. 4.18. Beam web buckling.
Inertia of this length of web about centre li ne of web

bd 3 322 x 11.93
= 12 = 12
= 45·2 X 103 mm4
Radius of gyration = JAI= J(45'2

322 X
X
10 3 )
11 .9
= 3·43 mrn
~ = 305 ='89
r 3·43 .
pe = 92 N/mm 2
Allowable load = area x allowable stress
92
= (322 x 11'9) X 103
= 352kN
Actualload = 175 kN
The effect of shearing force on the web will now be checked.
Area of web resisting shear
= depth of section X thickness of web
=6IOxIl'9
= 7260mm2
Allowable average shearing stress
= looN/mm 2
(See BS 449: 1969, Table 11.)
. 7260 X 100
Allowable sheanng load = 103
= 726kN
Actualload = 175 kN
It will be noted that the beam web possesses adequate strength to
resist the effects of crushing, buckling, and shearing stresses. In practice
it is usual for these checks to be carried out only for beams of short
span and heavy loading. For normal cases it is only necessary to
carry out sufficient calculations to obtain the required strength for the
bottom bracket.
The bracket on stanchion SI (Fig. 4.12) at first ftoor level which
supports the plate girder is of welded and fabricated construction.
Adequate bearing area is available to carry the load from the plate
girder if a 30 mm thick ftat is provided and a 25 mm thick ftat welded
as shown provides a means of connection to the plate girder. This
25 mm thick ftat also allows adequate dispersal of load from the
30 mm thick ftat to the web stiffeners. The bracket should be designed
at the same time as the web stiffeners to the plate girder (see section 27).
Load on bracket from plate girder = 1126·5 kN
Bearing stress = 190 N/mm2
. 1126·5 X 103
Area requtred = 190 = 5920 mm2
The area of bearing provided by the 30 mm ftat = 375 X 25
= 9375 mm2 which is considerably in excess of requirements.
Adopt 16 mm fillet weid for verticalleg.
Allowable load per metre run of weid
= 16 x 0·7 x 115 = 1290kN
. 1126·5
Length of weId requtred = 1290 = 0·87 m
A bracket depth of 500 mm will be adopted giving a weId Iength of

1 metre.
(c) Beam-to-beam c01llJections. A typical exampie of beam-to-beam

connection is shown in Fig. 4.19 and shows the connection between
beam Al and A2.
~IO" 305 UB 175,. 12 il'lICK

FI.Ai
/ A'l
.
I
1/
-
WELD
'533)< 20' ,,"' ... VIEWX
UI!>
AI AI
/=~
"HOLES '22 DIA I=/=;r:.====~
FoP. 20 PI" 8LACI<. ~OLTS
VIEW'X'
FIG. 4.19. Connection between ftoor beams Al and A2.
Load to be carried = 140 kN on each side of beam A2

Total load = 2 x 140 = 280 kN
The connection consists of ftats welded to the end faces of beam AI
and these flats being bol ted through the web of beam A2. These con-
necting bolts will faH either in double shear or by bearing in the web
of beam A2, tbis bearing being treated as enclosed bearing.
Allowable load in double shear on 20 mm dia bolt
= area of shank x double shear stress
= 314 x 160 _ 50.3 kN
103 -
Allowable load in enclosed bearing

= dia of bolt x thickness of web x
allowable enclosed bearing stress
= 20 x 11'~ x 200 = 47.6 kN

10
Critical value for design = 47·6 kN
Number of bolts required = ~~.~ = 6

A thickness of end flat to beam of 12 mm will be adopted.
Allowable load per metre of 6 mm fillet weId
= 6 X 0·7 X 115 = 484 kN
Length of weId required = !:~ = 0·29 m
The length provided is weIl in excess of this figure.
(d) Stanchion splices. In stanchions of this type, the load in the shaft
is wholly compressive in nature and the only function of the splice
plates is to maintain the two parts of the shaft in position. A splice
length of 0·9 m may be regarded as fulfilling these conditions (see
Fig.4.16).
29 Design of Mass Concrete Foundation Block

Provided that the load is not too great, a mass concrete block is the
most economical method of transferring the load from the stanchion
base to the ground. Stanchion S5 will be taken as an example, and the
base load of 1363 kN is approaching the maximum for this type of base.
To ensure that no bending takes place on the block, a load dispers al
line of 45° is taken from the edge of the column base plate to cut the
vertical side face of the block and this will determine the minimum
depth required.
Load from stanchion base = 1363 kN
Assumed safe bearing pressure on ground = 215 kN/m 2
Estimated weight of block = 70 kN
Estimated weight of filling above block and brickwork in walls over
block
= lOkN
Total load bearing on ground = 1363 + 70 + 10
= 1443 kN
Area of block required = ~~:
= 6·7 m 2
Adopt a square base of side = v/6·7
= 2·6m
Say, block size to be 2·60 m square X 1·0 m deep. (See Fig. 4.20.)
It will be seen from the detailed sketch that bolt boxes are shown.
These may be made from timber, steel, or plastic tubes or polystyrene
blocks and are cast into the concrete block but are removed prior to
the erection of the steelwork. The space provided by the removal of
the bolt box together with the 25 mm grout space under the column
base provide a means of adjustment for any discrepancies in the
accuracy of position and level of the anchor bolts.
2",7" 258)( I07k<3

UC
55
.. YIEW'X:
LOAD DISPERSAL
UNE \
45°
450.20DI.o.
-'-
ANCt-\OR
100 BOLi
4 PER BASE.
~
So.UARE HEAD AND
NECK iO 80LTS
VIEW 'x'
FIG. 4.20. Mass concrete foundation block for column S5.
No uplift is present on this base, and theoretically the provision of

anchor bolts is unnecessary although they are usually provided as an
aid to column location and for steadying purposes during erection.
If they are not provided it is necessary to supply some method of
supporting the building until sufficient steelwork has been erected to
provide stability.
Procedure in erection of a steel stanchion is generally as folIows:
(i) The foundation hole is dug and the anchor bolts, anchor plates
and bolt boxes are suspended in their correct positions from a frame
firmly fixed across the top of the excavation.
(ii) The concrete is now placed in position and the bolt boxes and
suspended frame are removed before the concrete is finally set.
(iii) After the concrete has matured, and before erection of the steel-
work commences, the steelwork erector places steel packings on the
top of each block to give the exact level between the underside of the
column base and the top of the concrete block.
(iv) The steel eolumn is now ereeted, followed by the remainder of
the steelwork and after it has been plumbed and levelled the anehor
bolt holes and the grout spaee are filled with eement grout to give a
solid base.
30 Design of Reinforced Concrete Foundation

The relatively large load earried by eolumn S6 makes a mass concrete
foundation uneeonomieal and the alternatives are a steel grillage eased
in eonerete or a reinforced concrete footing. The latter is to be preferred
today, although in the past steel grillage foundations were popular
but can now be regarded as a wasteful use of material.
A steel slab base of adequate area has al ready been designed and it is
necesliary to design a reinforced concrete foundation to support the
column loading. In foundations of this type it is likely that failure will
take place by the steel base plate trying to shear its way through the
concrete rather than by a flexural failure.
Column loading 2682

Weight of foundation 250 (estimated as 7! to 10 per
Allowance for ground floor cent of column load)
loading 120
3052kN
Assume a safe ground pressure of 215 kN/m 2 •

. 3052
Area of base reqUlred = 215
= 14·2m2
Adopt a square base of side = Y14·2
= 3·8m
Say 4 metres square.
A provisiooal depth of base will now be estirnated (500 mrn) and
checked for shearing along the lioes CDFE (Fig. 4.21). Load producing
shear may be taken as upthrust 00 area of base less area CD FE
2682
= 42 x (42 - 1.8 2 ) = 2140 kN
Shearing stress along line CD FE
= load = 2140 x 103 = 0.6 N/mm2

area 4 x 1·8 x 103 x 500
~w
-r-r----------,
4000 E m!
ElEVATION
'f------
I
p
I
REINFORCEMENT
'2.9 SAftS 20 DIA.
IN EACH
DIP,ECTloN
~
p:ELEVATION
~--------~G
/~~
/ __ _
/ __ _
1'1.:1- ---
IV 01 40 MINIMUM • ___ _
L___ .J Covep, H ' , __ _
EI.
1100
.IF ',--
,-
' - - - - - -..... J
I. 4000
PLAN PLAN
FIG. 4.21 Reinforced concrete foundation
block for column S6.
This may be considered satisfactory for a 1 :2:4 concrete mix (see

CP114: 1969, Table 6).
It is not usual to provide shear reinforcement in concrete bases and a
depth of concrete which is adequate to resist shearing stress is provided.
It is assumed that bending will take place across both axes of the
base and will be produced by the upthrust on four equal areas of
which GHJ is one (Fig. 4.21).
Bending moment at centre line of base
W A W B W
= '4 x 3' -"4 x 3' = 12 (A - B)
2682
= 12 (4 - 0'8) = 716 kNm
Area of steel required

Moment
= Allowable stress in steel x lever arm of section
Allowable stress in steel
= 21ON/mm2
Effective depth of section
= 500 - 60 = 440mm
. 716 X 106
:. Area of steel reqUired = 210 X 0.875 X 440
= 8850mm 2
Twenty-nine bars equally spaced across the base each 20 mm dia give an
area of 9106 mm 2 and this arrangement of bars will be required in
both directions across the base. The lever arm coefficient of 0·875 is
adopted as being the maximum desirable in a reinforced concrete
section in wh ich the selected depth is greater than that normally
used when flexural considerations are the design criteria.
No uplift is present on this base and anchorage arrangements
similar to those adopted for stanchion S5 will be adequate.
31 More about Connections

Up to one-third of the cost of steelwork is accounted for by the con-
nections required to join the various structural members together. The
advantages of careful and economical design will be lost if insufficient
attention is given to choosing the correct type of connector. Although
students and young engineers will take much care in designing the
structural member they will frequently give insufficient consideration
to the design of the connections. The importance 01 efficient and econom-
ical design 01 connections must not be overlooked. A short description
of the types of connector most frequently used is given below.
1 Black structural bolt. This type of bolt is made from mild bolt
steel (ultimate stress 400 N/mm 2 ) and is the cheapest and the com-
monest connector used in structures. It is not manufactured to close
tolerances and the working stresses are reJatively low (see BS 449: 1969,
clause 50). These bolts are manufactured from black bars which
have not been machined, and are normally used in holes 2 mm larger
than the diameter of the bolt shank.
2 Close fitted turned bolt. In this case the bolt is of a much higher
quality and the shank is accurately turned to size and for this reason
it is sometimes referred to as a bright bolt. Essentially the bolt shank
forms an accurately finished dowel pin placed in an exact size of hole
0·25 mm larger than the bolt shank. No drifting together of connected
parts is allowed during erection. The required accuracy may be obtained
by drilling pilot holes of small size in the individual plys of material
and reamering out the holes to the correct size simultaneously through
aIl plys of material. A turned bolt is machined under the head and a
thick washer is used under the nut to ensure that the bolt can be
satisfactorily tightened without any of the screwed portion of the
bolt coming within the grip length of the bolt (i.e. the length of bolt
passing through the connected parts). To ensure easy insertion into
the hole, without damage, the screwed portion is 1·5 mm smaIler in
diameter than the shank. It will be realised that no slip can take place
using this type of connector and it is useful in certain conditions.
Expense involved in both the bolt and the hole make it an uneconomic
means of connection for widespread use.
3 Rivet (hot driven). A hot driven rivet is another type of fastening

which relies upon the dowel pin action of the shank. The shank which
is circular usually has a cup-shaped head aIthough other shapes are
available (e.g. countersunk or pan head) and is heated to a red heat
and placed in the hole. A powered hammer is then used to form the
other head of the rivet using the portion of the shank which protrudes
through the hole. One effect of this process is to make the rivet shank
fiIl the hole and for strength caIculations the diameter of the hole
should be taken in preference to the nominal shank diameter. Site
riveting as weIl as shop riveting is possible but the widespread adoption
of eIectric-arc welding has largely replaced the use of rivets both in the
workshop and on site.
4 Electric-arc welding. This process may be said to be the modern

means of making connections between structural members. Neat and
compact connections are possible and the necessity for hoIing members
can be reduced or almost eliminated. A greater accuracy of fit between
members is required than for bolted work. If it is necessary to hold
the members in position by means of cIamps or jigs this may prove
expensive and offset any saving obtained by eliminating bolts or
rivets. This is particularly true in regard to site conditions and welding
is usuaIly confined to workshops although it is used on site if occasion
demands. Site connections are usually bolted.
5 High-strength friction-grip holt. These rely on a predetermined

grip being produced between the connected parts, and this is obtained
by tightening the bolt to a specified degree of tension. A connector of
this type relies wholly upon the friction generated between the con-
nected parts and the shearing or bearing value of the bolt shank is
ignored. This is the reverse of the normal type of bolt or rivet which
reIies on the shearing or bearing value of the shank and the friction
grip is largely ignored. Several grades of bolts are available and the
user should be careful that the maker's instructions are carefully
followed as to tightening procedure. Several ingenious designs of
bolts are available which give visual indication of the degree of tightness
being achieved which enables a rapid and accurate check to be made
upon the workmanship. BS 3294 gives recommendations for ordinary
grade bolts which should prove adequate for most structural purposes.
This type of bolt is becoming increasingly popular in situations where
connections carrying heavy loads have to be designed.
32 Typical Strength CalculatiODS for Connectors

(a) Black bolts. Strength characteristics of this type of connector are
given in Fig. 4.22.
SHEAR W 5HEAf'. PLANE.
, / ' PL,o,N E
W~~==~~~~=4
,
w2:~~~~~~~::=?-.w
~ I +-W '2 SHEAR PLANE
R BOLT LOADED IN DOU8LE

BOL, LOADED IN SINGLE SHEA 5HEAR OR ENCLOSED ßEIIRING
oR SIMPLE. ßE ... RIN(. IN PLATE. IN PLATE.
AllOWABLE STRESSES
SHEA?, eo Nlm..,~
SIMPLE 8EAP.ING 1<:'0 Nj ... ",<
ENCLOSED ßEARIN6 'ZOON/"'''''
TENSIO,," 1301'1/",""
w .... -E
BOL T LOADED It-J "TENSION
ACROSS ARE" AT ROOT OF
THREAO.s
TABULATED VALUES [kN1 BOLTS GRADE 4·G
BOll 5HEARI~G BEARIN(, VALUES 1"1 PL,o,re {E] ARe ... itN51Lf
DIA AREA VALUEs UPPER VAluE - ~""CLOSED{El AT V~LVE
mm "'1\'12 f-S-IN-C;-L""'fr-DO-U-lIL-e-iLOWER VMUE -SIMPLE (5 tOR) R~~T
,.. G 8 10 '5 'Zo 'Z5 S TH~E~
1'2. 1\3 9·05 18·1 14-4 "·2 24 E 70 .4 9· 14
/1·5 15' 4 " ·2 s
'" 201 1" ' 1 32· 2
20 314 '25·1 50·3 24 32 40 ''0 &0 E 215.5" 28

".1. 2S." 32 48 "4 S
24 452 3~'2 72· 4
32·4 4301 54 81 loS E 403 52.4
27 573 45·8 91·6 25· 9 34·" ~,2 "4-8 8"'4 10& S
FIG. 4.22. Strength characteristics of black structural bolts.

EXAMPLE: Bolt dia 24 tnm.
Single shear value = Bolt area X shear stress
80
= 452 X -3 = 36·2kN
10
Double shear value = 2 X single shear

= 2 X 36·2 = n·4kN
Tension value = Area at root of thread X tensile stress
130
= 318 X -
103
= 41·4kN
(b) Rivets, close fitted turned botts. Strength characteristics of these

types of connectors can be calculated in a similar manner to that
shown above for black bolts using the appropriate stresses. The area of
bolts is taken as the shank or nominal diameter but in the case of hot-
driven rivets the area of the rivet becomes the area of the hole and this
larger area should be used in calculations.
I----- -
SQ U A
-8-p.-,,---Iq A
<:. 51NC,LE. 'v'
DOUBLE 'V.
EXAMPLES OF BUTT WELD5
THROH ,./
=0'7 LEe, SIZE
FILLET WELD
HG. 4.23. Line diagrams of typical butt and fillet welds.
(c) We/ding. Reference should be made to BS 1856: 'General require-

ments for the metal-arc welding of mild steel', and Fig. 4.23. Butt
welds are invariably 'fuH strength welds' in which the jointed plates
are given prepared edges to enable the welding process to penetrate
fuHy the thickness of the plate. Such welding is continuous, and inter-
mittent butt welds are generaHy unacceptable.
Fillet welds are specified by the leg size and for the purposes of
strength calculations it is assumed that a fillet weid fails across the
throat area (0·7 X leg size). For Grade 43 steel the allowable stress on
the throat area is 115 N/mm2 •
Leg size of weid = 6 mm
Load per metre run = Throat thickness X allowable stress X 1031
103 kN
= (6 X 0·7) X 115 = 483 kN
(d) High strength/riction grip bolt (see BS 3294).

Bolt diameter = 20 mm
Proof load of bolt (obtained from makers) = 130 kN
. h
VaI ue m slip factor b f fti . . f:
s ear = I d f: t X num er 0 e ectlve mter aces
oa ac or
X proof load
. slip factor 0·45
RatIo of load factor may be taken as H = 0·322
In double shear the number of effective interfaces
=2
Value in double shear = 0·322 X 2 X 130
= 83·6kN
Value in tension = 0·6 X proof load
= 0·6 X 130 = 78 kN
33 Typical Strength Calculations for Connections
(a) Connection using black bolts and having eccentric loading in plane
0/ bolts. Examples of this type of connection are shown in Fig. 4.24
with the analysis appropriate to a group of four 24 mm dia black
bolts assumed to be loaded in single shear and simple bearing.
Loading values of 24 mm dia bolt are:
Single shear value = 36·2 kN
Bearing value in 12 mm plate = 46·1 kN
Criterion for design is the single shear value of 36·2 kN.
The bolt carrying the maximum load will be the bolt E which lies the
greatest distance from the centroid of the bolt group.
90kt-l
(w)
w 6S
[e]
CENTROID
75
./
4 BoLTS
24 0 IA 1+-"-'-01~~
W
EXAMPLES ANALYSIS
FIG. 4.24. BoIted connection, bolts in plane of load.
Load in bolt E ean be broken down into two parts, the first of
whieh is due to the axial effeet of the load, and the seeond being the
rotating or bending effeet of the load. Once these two values have been
ealculated the resultant load ean be found whieh must not exceed the
allowable load.
Load in bolt E due to axial effeet = No. o~bolts = ~o = 22'5 kN
Load in bolt E due to bending effect

= bending stress X area of bolt
Mxy
= -/- X area of bolt
where M = bending moment on bolt group

= 90 X 65kNmm
y = distanee from centroid of group to bolt E = 70·7 mm
/ = moment of inertia of bolt group
= Area of one bolt x (Y1 2 + Y2 2 + Ya 2, ete.), where Yh Y2'
Ya, ete. are distances of bolts from eentroid of group
= 4 X 70'72 mm4
Load in bolt E due to bending effeet
90 X 65 X 70·7 X 452 = 20. k
- 4 X 70.7 2 X 452 7 N
These two loads can now be combined graphically to give a resultant
load of 40 kN.
Comparing this with the allowable value of 36·2 kN indicates that
the bolt group is overstressed by approximately ll per cent. Considera-
tion should be given to increasing the number of bolts in the connec-
tion.
(b) Connection using black bolts anti ha,ing eccentric Ioading at right
angles to the plane 0/ connectors. Reference to Fig. 4.25 indicates
the example to be considered and consists of 10 bolts of 24 mm dia.
VIEW',x'
BG. 4.25. BoIted connection: bolts at right angles

to plane of load.
In this case the turning moment produced by the load is at right angles
to the plane of the bolts indicating that the upper bolts in the group are
subjected to a combination of shear and tensile stresses whilst the
lower bolts are subjected to shear stresses, the corresponding compres-
sive stresses being taken at the interface between the column and the
bracket.
The exact behaviour of such a bolt group is difficult to evaluate.
For example, does the bracket rotate about the centroid of the bolt
group or does it rotate about the bottom edge of the bracket? Further,
it is not possible to combine tensile and shearing stresses other than
by consideration of the theory of 'principal stresses'.
A simple and practical solution to the problem is to locate the axis of
rotation through the centroid of the bolt group and assume that all
bolts above this line resist the tensile load, and all bolts lying below
this line resist the shearing load.
In consequence of this proposal it will also be assumed that the four
bolts above the axis of rotation will resist the tensile forces in propor-
tion to their distance from the axis of rotation.
If tensile load in one bolt = P kN then tensile load in top pair of
bolts = 2P with a moment about the axis of rotation = (2P x 180)/103
kN m. Similarly the next pair of bolts will produce a moment of
2 x (P/2) X (90/10 3 ) kN m.
Total moment of resistance of group
= 2P x 180 x 90 = 0.45 kN
103 + P 103 P m
Bending moment on bolt group produced by a load of 90 kN = 90 X

= 13·5 kN m
(150/103)
0·45P = 13-5
P = 30kN
Allowable tensile load (Fig. 4.22) = 41·4 kN. Load due to shearing
action on lower four bolts below axis of rotation = (90/4) = 22·5 kN.
Allowable single shear load (Fig. 4.22) = 36·2 kN.
Both the tensile and shearing loads are less than the allowable
values and the connection may be considered satisfactory provided
that the initial assumptions prove acceptable.
(e) Welded eonneetion with eeeentrie loading in plane 0/ the welds. In

this case (see Fig. 4.26) the bracket is welded on three sides only. If
HG. 4.26. Welded connection: welds in plane of load.
welding is provided on all four edges the problem is simplified in that

the weId group will be symmetrical in relation to the column and also
to its own geometry. Point A lies the greatest distance from the cen-
troid of the weId group and it is this point of weId that requires in-
vestigating. It is important that the weId be returned round point A
as shown in Fig. 4.26 so that the start or finish to the welds is away
from the point of maximum stress in the weid group. For purposes of
calculation the return weid is ignored.
Assurne throat thickness as unity (I mm). Length of weid = 450 +
2 (125) = 700 mm. The position of the y axis of the group of welds
may be found by taking moments about the verticallength of weid.
700x = 2(125 X 62) + (450 X 0)
x = 22·14mm
Eccentricity of load relative to centroid
= 230 + 40·36 = 270·36 mm
/z welds (125 mm weid) = 2 X 125 X 2252
= 12·66 X 106 mm4
4503
/z welds (450 mm weid) = 12
= 7·59 X 106 mm4
/y welds (125 mm weid) = 2(1253 /12 + 125 X 40.36 2)
= 0·732 X 106 mm4
I y welds (450mm weId) = 450 X 22.142
= 0·22 X 106 mm4
Polar moment of inertia = /z + /y
= 21·20 X 106 mm4
Distance from centroid to weid A = y(225 2 + 102.562)
= 247mm
Stress at point A due to axial effect of load
= ; = 450 :~(l25) = 0·143 kN/mm 2
Stress at point A due to bending effect of load

Mx y 100 X 270·36 X 247
="IP = 21·20 X 106
= 0·314 kN/mm 2
A graphical combination of these two stresses results in the resultant
stress ofO·415 kN/mm2 based on a throat thickness of 1 mm.
Throat thickness required if stress is not to exceed 115 NJmm 2
3·6
Leg size = 0.7 = 5·2 mm, say 6 mm weid
If 6 mm fillet welds are used the design of the bracket is satisfactory.
(d) Welded connection with eccentric loading at right angles to plane 0/

welds. It is usually assumed in this type of bracket (see Fig. 4.27)
that the maximum stress occurs along the line of the top weid as
150
t--
.
COLUMN
--
VIEW'X
.x ><. ?> 80
/
ELD
VIEW 'x'
FIG. 4.27. Welded connection: welds at right angles

to plane of load.
the rotation takes place about the x axis of the weid group and that
a simple vectorial summing of tensile and shear forces is acceptable.
Any misgivings about the theoretical analysis of welded connections
should be viewed against the relatively low stress allowed in welded
work (l15 NJmm 2 ) compared with the stress allowed in the jointed
parts (165 N/mm 2).
Assurne throat thickness of unity (I mm)
/z welds (ISO mm weid) = 2 X 150 X 1902 = 10·83 X 106 mm4

3803
/z welds (380 mm weid) =2 X 12 = 9·14 X 106 mm4
Total /z of welds = 19·97 X 106 mm4

Stress in top weid due to axial effect of load
ur 170 170
= A = 2(380 + 150) = 1060
= 0·160 kNjmm2
Stress in top weid due to bending effect of load
M urxexy
=-;-= I
170 X 255 X 190
= -1'""9--:'9=7-x---"'IO=6'---
= 0·413 kNjmm2
Resultant = v(O'160 2 + 0,413 2)
= 0·442 kNjmm2
Throat thickness required if stress is not to exceed 115 Njmm 2
-
0·442 X 103
115
= 3.85 rnrn
3·85
Leg size = 0'7 = 5·5 mm, say 6 mm
If 6 mm fillet welds are used the design of the bracket is satisfactory.

5. Design of 18 m Span Shed with
Ridge-Type Roof Trusses
34 Particulars of Scheme
(a) A workshop building 18 m span, by 40 m long, by 5·5 m to the
eaves level is required for manufacturing purposes. Details are discussed
between the customer and the structural engineer, and frequently,
when this type of industrial building is required, the services of an
architect are considered unnecessary. The customer arranges for
the building work (e.g. preparation of site, foundations, brick walls,
etc.) to be carried out by a contractor of his own choice. Fabrication
and erection of the steel frame are carried out by the steel fabricator (or
constructional engineering firm) who mayaiso supply, usually through
a sub-contractor, any glazing or corrugated sheeting required for clad-
ding the frame. This type of building tends to be very competitive.
Nowadays buildings with ridge-type roof trusses are seen much less
frequently than in the past. The portal frame type of structure (see
'More Steel Frame Design Examples') is preferred by many as being
superior in appearance and ease of maintenance as weIl as providing
more usable interior space. Modem methods of production-line
fabrication which have been applied to portal frame construction have
resulted in this form of structure becoming very competitive in this
country. Nevertheless the rooftruss and column remains a good example
of framed trusswork.
(b) The following were decided during a meeting between the customer
and the design engineer.
(i) Size 0/ building. 18 m centres of side columns, 40 m centres of
gable columns and 5i m height from ground to eaves level.
(ii) Cladding (or covering). Asbestos-cement corrugated sheeting
with 12 mm thick insulation board was suggested with the object of
combining the initial economy of the outer covering with the advan-
tages of the underlining which reduces considerably the beat losses
from within the building. Steel sheeting provides an alternative means
of cladding but the cost of protection against corrosion can be bigb.
Modern forms of steel sheeting with a coloured protective coating can
be extremely attractive but are appreciably more expensive than
asbestos-cement products. A weIl designed and protected steel sheet is
likely to have a longer life than asbestos-cement and mayaiso be
preferred for appearance.
(iii) Glazing. A good standard of daylight illumination is required
and approximately one-third to one-half of the roof plan area is
provided with continuous 'patent type glazing' (i.e. aluminium- or
lead-covered steel bars designed to give puttyless glazing). An alterna-
tive means of admitting daylight to the building would be to arrange
for a certain proportion of the roof sheets to be replaced by transparent
roof sheets wbicb match tbe profile of the remaining sbeets. These
can be arranged in a pattern to give a more even distribution of lighting
than is possible with patent glazing arranged in linear areas along the
length of the building.
(iv) Ventilation. No special provision was required for ventilation.
(v) Condition 0/ site. There was a slight slope along the length of the
site, and it was decided to introduce a dwarf brick-wall to provide a
more effective break or finish between the sloping ground and interior
of the building than could be provided with relatively brittle asbestos-
cement sbeets.
(vi) Floor construction. A 100 mm tbick floor of reinforced concrete
laid on a weIl consolidated hardcore 150 mm tbick was considered
suitable for tbe floor, bearing in mind tbe condition of the ground as
weIl as tbe floor loading.
(vii) Access doors. No large pieces of equipment were to be manu-
factured and no provision was made for large doors but double-
leaved hinged doors were provided in each gable.
(viii) No special provision was required for plant which migbt
affect either tbe spacing or loading of structural members and tbe
design engineer was free to decide bis own arrangements.
35 Layout of Steelwork
(a) Tbe layout of steelwork in tbe form of a line diagram is given in
Fig. 5.1. Tbere being no restrictions affecting the layout of tbe steel-
work, adecision must be made regarding tbe economical spacing
of columns and roof trusses. Economy in tbis instance will be a function
of botb purlin and roof truss economies and the spacing of roof trusses
and columns may vary between 3 m and 6 m. Larger span trusses may
DESIGN OF 18 m SPAN SHED 83
40 ...... CEtHR 5
rA
\5IDE/
~
(,ABLE C.01..1.I11 N
,\-·S '" WIND
(,I!>,OER
---+
-----
~
\'0 ...,
4·5",
HOOR
"Ir
1
LtVEL [------1
ROOF ROOF
/ TRlI&S TRUSS
4·5 / '- I -......
--
",
-<
/
r----<
4·5
'"
4 ... 4", 4", 4,., 4",
Li
A
PLAN AT ROOF TRUSS TIE l~VEL
ROOF Pu RLiNS '\
mpliiilllllll'- \.
I!HI I C, LAI ING \
111111 111111 111111 \..
'1111 111 111111
H (,LAZIN(,
PP.RT
-
PLAN AT
-
PUf/.I..IN
-
LEVEL
SIDE / / SflEETI ~G RA\LS

ßRAC ING
Klillh, f
11111111 1111,
PART SIDE. ElEVATIO

('I'IBLE ELEVATION
HG. 5.1. 18 m span shed.

Framing diagram.
generally be spaced at larger centres than smaller spans. In this ca se

the length of the building is 40 m wh ich divides conveniently into ten
equal spaces of 4 m each. Gable columns will be spaced at 4·5 m giving
four equal spaces in each gable frame.
This spacing of the main building frames at 4 m centres is suitable
when the purlin sections are to be fabricated from hot-rolled angle
seetions. An alternative form of purlin is available which is cold-rolled
from steel sheets 3 mm or 5 mm thick into the form of a 'Z'. Usually
galvanised steel strip is used for these lighter purlins which are larger
in depth than conventional rolled sections. Good section design results
in the section having a high moment of resistance consistent with a
minimum use of material and spans of 6 mare very economical.
If this latter pattern of purlin is preferred then the spacing of the main
building frames can be increased and a more economical building
provided.
(b) Whatever type ofroof covering is adopted it is essential to follow the

maker's recommendations as to roof slope and sheet fastenings. This
may weIl decide the pitch of the roof truss rafter slopes and in this
case a roof slope of 221° is adopted. Very shallow trusses result in
high loadings in the members and deep trusses will result in long
lengths of strut which are uneconomical. Commercial practice varies
between a ridge height of one-fifth to one-quarter of the span, cor-
responding to a rafter slope of 211° to 261°. Recommended spacings
for cladding or sheeting supports vary according to type of sheet
selected and in this case are:
roof 1·37 m; sides and gable 1·82 m.
(c) The best arrangement for the internal framing of the roof truss
would be for the rafter or node points to coincide with the purlin
positions. In this example the spacing of the purlins is 1·37 m which
does not coincide with the economical rafter panel length of 1·8 m to
2·3 m. A rafter panellength of 1·94 m is adopted with a form of roof
truss framing which does not give long lengths of struts. However,
because the purlin positions and the node or panel points do not
coincide it will be necessary to design the roof truss rafter for flexural
as weIl as axial loads. A 75 mm rise is given to the centre portion
of the roof truss, partly to offset deflection, and partly to improve
appearance.
(d) Grade 43 steel is used throughout the design.
36 Effect of Wind on Building ,

Before commencing the detailed design of the building it is necessary to
investigate the effect of wind upon it. CP3: Chapter V: Loading: Part 2,
Wind Loads, is the appropriate specification and careful consideration
should be given to it so that the wind loading on the structure may be
assessed as accurately as possible. Wind loading usually has two basic
effects:
(a) Local damage may be caused to small areas of c1adding and the
supporting members by relatively small pockets of high wind gust
pressure without endangering the stability of the structure (e.g.
isolated roof sheets or window glass may be broken).
(b) Larger pockets of wind may endanger the stability of the building
but because of their size it is likely that the average wind pressure will
be lower than in case (a) because the high local gust effect is spread
over a larger area. Class A loading refers to case (a) and Class B or C
loading refers to case (b). Tbe division between Class Band C loading
is determined by the size of the structure.
The design procedure for assessing wind loads may be carried out as
folIows.
(a) The basic wind speed (V) can be determined from the Code of
Practice and is based on the basic wind speed likely to be experienced
in the locality of the building (e.g. Manchester, V = 45 mts).
(b) Tbe design wind speed (V.) is the product of V, SI, S2, and S3, where
SI, S2, and S3 are factors taking into account the topography,
environment, and life of the building.
For this example the following values are assumed.
(i) For cladding (i.e. local damage)
V= 45 mtsec
SI = Topography factor = 1·0
GrOUnd roughness category 3]
S2 = 0·78 ( Height of building 10 m
Class A loading
S3 = 1·0
V. = 45 x 1 x 0·78 x 1 = 35·1 mts
(ii) For structure (i.e. stabiIity)

V = 45 mts
SI = 1·0
Ground roughness category 3 }
S2 - 0.74 { Height of building 10 m
- Class B loading-greatest horizontal
dimension of building is less than 50 m
V. = 45 x I x 0·74 x 1 = 33·3 mts
(e) Calculate dynamie wind pressure (q) for building. At this stage the
wind velocity is converted into apressure.
q = kV. 2
When using SI units
k = 0·613
q(cladding) = 0·613 X 35·P = 755 N/m 2
q(structure)= 0·613 X 33.3 2 = 680 N/m 2
(d) The pressure eoefficients (Cl') for the various building surfaces are
now evaluated from the Code of Practice and the wind loadings
calculated for each surface. The total wind load acting on the building
is the vectorial summation of the wind loads acting on the various
surfaces. External pressure coefficients are denoted Cl'e and internal
pressure coefficients are denoted CI'I.
Wind load normal to a building surface
= Area of surface X (sum of Cl'e and CI'I) X dynamic wind pressure
The precise assessment of internal pressures is likely to be difficult,
if not tedious, and in this case, as indeed in the majority of cases, a
simplified procedure is acceptable. Provided that no large openings
are present in the cladding surfaces the value of CI'I may be taken as
the more onerous of +0·2 or -0·3.
Internal press ure does not affect the overall conditions of building
stability but must be taken into account when individual structural
members are being considered.
In areas of cladding adjacent to eaves, ridge, and gable a special
risk occurs in that these areas are subject to high press ure as the wind
swirls and accelerates. Local values of Cl'e for the roof and sides are
-1·1 and -1·0 respectively.
The values of press ure coefficients appropriate to this example
are given in Fig. 5.2 and the following relationships were used to decide
the pressure coefficients.
. h height to eaves 5·5
Ratlo- = =-
w width of building 18
= 0·31
corresponding to hlw ~ !.
. [ length of building 40
RatIO - = =-
w width of building 18
= 2·2
corresponding to i < [lw< 4.
Roof angle = 22!O
-0,0
WIND +0'7
t
-
WIND
-0'3 i -0·4 ! +
I
-0,7 1-0'7
---,--
I I
f---t- --
-0.3 1 -004 I
-0·6 I -0·.,
-
-0.5
+0'7 I -0·25 -0·5
- ---+--- -I---t---
[ci)
~
-0·6 -0·\
PLAN PLAN
0·3
\
CRoSS SEC TIONS
FIG. 5.2. Details ofwind loading.

(a), (b), (c) Wind pressure coefficients for wind direction normal to
side of building. (d), (e), (f) Wind pressure coefficients for wind
direction normal to gable of building.
37 Roof Purlins
Roof purlins may be considered as secondary members and higher
stresses than those used for the design ofmain members are appropriate
provided that the resulting deflection does not crack brittle cladding
panels, impair the weathertightness of the roof, or give an appearance
of weakness in the structure. As previously stated it is possible to use
the normal hot-rolled sections for these members, or the use of cold
4
rolled steel purlins may be eonsidered. No standard seetions are
available for the latter type, eaeh manufaeturer marketing his own
seetion. If this type of purlin is preferred the maker's recommendations
as to loading and defleetion must be followed. For this example a
normal hot-rolled steel angle seetion is adopted.
Design 0/ sheeting pur/in

Dead load:
kN/m 2
Sheeting 0·161
Insulation 0·036
Purlin self weight 0·058
0·255
Superimposed load 0·695
0·950 kN/m 2 measured on slope
The superimposed load eorresponds to an allowanee of 0·75 kN/m 2

(see CP3: Chapter V: Loading) measured on plan area.
Area supported by one purlin = 1·37 x 4 = 5·48 m2
Load on one purlin = 5·48 x 0·95 = 5·21 kN
It is likely that some degree of eontinuity will exist over the supports
and a value for the bending moment of W x L/1O is reasonable.
However, eaeh end of the purlin should have at least two eonneeting
bolts to help eontinuity, otherwise a value of W x L/8 should be used.
M = WL/1O = 5·21 x 4/10 = 2·08 kNm
z = MI! = 2·08 X 103/165 = 12·60 em3
A 101·6 X 63·5 X 6·3 angle is suitable having a z = 15·4 em3 •
The 63·5 mm leg ofthe angle reeeives ample lateral stiffening from the
cladding and is quite satisfaetory over a span of 4 m.
BS 449: 1969, clause 45, recommends the following minimum
requirements:
Depth = span/45 = 4000/45 = 89 mm
Width = span/6O = 4000/60 = 67 mm
. _ W X L _ 5·21 X 4 X 103 _ . 3
MIO. Z - 1.8 X 103 - 1.8 X 103 - 11 6 cm
The proposed section meets these requirements apart from a slight
deficiency in width which is sufficiently small to be acceptable.
This method of purlin design is suitable for purlins located on roof
slopes of 300 or less. If the slope is greater, then the load should be
resolved into two components parallel to the legs of the angle section,
and bending stresses calculated about both axes and summated in a
similar manner to that adopted for the design of the side cladding
supporting rails.
Purlins are lightly-Ioaded members and it is possible that excessive
deftection may damage brittle c1adding materials. In cases where this
occurs it is important that some consideration be given to limiting the
deftection to, say, span/200. The true deftection will be difficult to
assess and should lie between the condition for a simply supported
purlin over one span (deftection = (5/384)(WL3 IEI» and the condition
for a two-span continuous beam (deftection = (1/185)(WL 3 IEI).
Assume deftection = span/200 = 20 mm, and since deftection =
(5/384)(WL 3 IEI) then the moment of inertia required
5 x 5·21 X 40003
- 384 x 210 x 20 = 103 cm4
A 101·6 x 63·5 x 6·3 angle has a moment of inertia = 106 cm3 •

The proposed section is satisfactory when viewed from the criterion of
deftection.
Purlins adjacent to areas of high local wind pressure should be
checked for the uplift condition in which an external pressure coefficient
of 1·1 combines with an internat pressure coefficient of 0'2, giving a
total pressure coefficient of 1·3 .
. d upI'r
Wm I' (
1 t on pur m q =
755 NI m 2) = 1,3 x 103 x 5·48
755
= 5,37 kN
Minimum vertical load acting in opposite direction
= dead load
= 0·255 x 5·48
= 1·39 kN
Net uplift = 5·37 - 1·39 = 3,98 kN
This is less than the load for which the purlin was designed and should
not produce unduly high stresses in the purlin at the underside edge
which in this loading condition will be in compression. It may be
argued that the hook bolts which secure the sheeting to the purlins
will provide some restraint to the edge of the purlin. If it is feIt to be
desirable a horizontal sag-rod can be used at the mid-span of the
purlins which occur adjacent to these areas of high local wind press ure.
An angle section 101 mm deep and spanning 4 m can be considered
adequate, indeed many designers would feel justified in using a lighter
section, say 75 mm deep.
The giazingpurlin will support a heavier load and may be designed in a
similar manner. A 101·6 X 63'5 X 7·82 angle is considered satisfactory.
An additional 50·8 X 50·8 X 6·32 angle is used in conjunction with
the glazing purlin to support the glazing bars. This angle may be in the
form of short lengths local to the glazing bar positions or if it is made
continuous in length it will prevent any broken panes of glass from
falling into the building.
38 Design of Roof Trosses
(a) Se/f weight 0/ t'IISS. From Fig. 5.3 it will be seen that a roof
truss of 18 m span weighs approximately 800 kg.
~o
·wo /
\00
/
V
/
,.......
0') /
.....
.Y.
/
t-
~
/
50 /
/
/
./
100
V
/'
o 10 20 30
SPAN (m]
FIG. 5.3. Weights of roof trosses.
Estimated weight of truss = 800 kg (8 kN).

Equivalent weight per sq. m of roof slope
weight of truss
- 2~--~~~~-----------
X rafter length X truss centres
8
- 2 X 9·72 X 4 = 0·103 kN
(b) Loading.
Dead 10ading:
Sheeting- Glazing-
Asbestos sheeting 0·161 Glazing 0·293
Insulation board 0·036 Purlins 0·122
Purlins 0·066 Self weight 0·103
Selfweight 0·103
0·366 0·518
Superimposed 10ading: 0·695 kN{m 2 (as for purlins).

Wind loading: (see Fig. 5.2(d) and (e».
Pressure coefficient = -0·7 - (+0·2) = -0·9
Dynamic wind pressure = 680 N{m 2
(c) Force diagrams (see Figs. 5.4 and 5.5). Force diagrams are the most
convenient method of obtaining the forces in the bars of the roof
truss. Strictly speaking, force diagrams can only be drawn for pin-
jointed frames but in practice the necessity for joint (or gusset) plates
between the members renders the frame statically indeterminate.
Practical necessity ignores this fact and simple force diagrams are
acceptable. A further advantage may be gained by assuming all loads
as being equal although glazing areas produce a higher intensity of
load than sheeted areas. The glazing areas are uniformly distributed
and no unacceptable error should arise because of this assumption.
Forces in bars due to the superimposed load can be obtained by
multiplying the dead load force by the ratio of superimposed load{
dead load. If conditions prevent the above assumption from being
acceptable then separate diagrams must be drawn for each loading
condition.
Length of sheeting per rafter = 6 m
Load from sheeting = 6 X 4 X 0·366 = 8·78 kN
Length of glazing per rafter = 4 m
Load from glazing = 4 X 4 X 0·518 = 8·29 kN
Total load per rafter = 17·07 kN
17·07
Dead load per panellength of rafter = -5- = 3·41 kN
Length of c1adding per rafter = 10m
FRAME DIAGRAM
s:
7 BAR 5lC-7x
DEAD LOAD FoRCE
DIAG~AM 1'\
~""~ I~ I~ ~o ~5 kN
LOAD SCALE.
FIG. 5.4. Roof tross.
Force diagram for dead load
Superimposed load per rafter =

10 X 4 X 0·695
= 27·8 kN
Superimposed load per panellength of rafter
27·8
= -5- = 5·56 kN
Super load = 5·56 = 1.6
Ratio
Dead load 3·41
Length of cIadding per rafter = 10 m
Load firom wm. d 10 X 4 x 0·9 x 680
= 103
= 24·5 kN
24·5
Wind load per panellength of rafter = -5- = 4·9 kN
~·S
~'"
L! '2·4kN
14 I M IN
FRAME
Q
DIAG.RAM 24.SkN~
o 5 10 15 ?O 25
1",,1 I I I I
lO"'o SCALE. [k.N]

FIo. 5.5. Roof truss.
Force diagram for wind load
Force diagrams can now be drawn and the forces in the bars tabulated
as shown in Fig. 5.6. It is, however, not possible to draw the force
diagrams for the truss framing as shown without making a temporary
modification. Point 5 cannot be located because it lies on a line
joining points 7 and 8 which are not known at this stage in the con-
struction of the diagram. Point 5 is 'by-passed' temporarily by inserting
the bar 5x-7x (shown dotted). This enables points 7 and 8 to be located,
leading in turn to the location of point 5. Bar 5x-7x is now deleted.
(4) Design of roof truss members. After the forces in the bars have
been tabulated the first stage in the design of the members is the
relative importance of the various combinations of the three basic
loading conditions.
FIG. 5.6. Table of forces in roof truss members.
Dead Dead
Dead Super Wind +
Super
+
Wind
Bar
C T C T C T C T C T
B-l 42·0 67 57 109 15

E-6 35·0 56 55 91 20
Q-l 38 61 52 99 14
Q-8 21 34 26 55 5
1-2 4·0 6 6 10 2
3--4 6·0 10 8 16 2
2-3 3 5 4 8 1
7-8 17 27 26 44 9
C represents compression.
T represents tension.
All loads given in kN.
(i) Dead load (permanent load)

(ii) Superimposed load (occasionalload) } Positive
(iii) Wind load (occasional load) } Negative
The following combinations of loading are possible:
Case I Condition (i).
Case 11 Condition (i) + (ii).
Case III Condition (i) + (ii) + (-iii) = (i) + (ii) - (iii).
Case IV Condition (i) + (-iii) = (i) - (iii).
From the above possible combinations it will be seen that the
maximum positive case is Case 11, and provided that load (iii) is
greater than load (i) then the maximum negative case is Case IV.
Generally speaking, a member which is a strut under Case II (e.g.
roof truss rafter) will be adequate under Case IV. However, a tie
member under Case II which becomes a strut under Case IV will
require designing as a strut. This is a case in which the tie becomes
subject to load reversal due to the effect of wind and an increased [Ir
ratio of 350 is tolerable in these circumstances.
It has already been stated that the joint (or gusset) plates used to
connect the members together induce some degree of restraint at the
connection points. As the truss deflects, and because the ends of the
truss members are restrained, some bending or 'secondary stress' is
induced. This is ignored provided that all members are in the form of
angles and the use of flat bars for tie-members is not allowed.
(i) Compression members. Rafter- Member E-6. Design load 91·0
kN; panellength 1·94 m; purlin load 6·92 kN.
Ofthe five lengths forming the rafter, the length E6 has been selected
for design as carrying the most adverse condition ofaxial load and
ftexural bending caused by the purIin load and panel or node point
not being coincident. By inspection of the preliminary layout drawing
it is possible to locate the purlin which lies nearest to the centre of any
of the panel lengths. The maximum ftexural bending due to the purlin
loading will occur in this length.
The design of the rafter is treated in a similar manner to that out-
lined previously for a multi-storied column subject to axial compression
and bending. Good continuity exists over the node points. Across the
x axis of the rafter (Le. in the depth of the truss) an effective length of
strut member equal to 0·7 times the panel length will be adopted. On
the y axis (Le. across the truss horizontaIly) the rafter is restrained by
the connection between the purlins and the roof truss rafter. The
purlin cIeats forming the connection can be assumed to give only
partial restraint to the rafter, and an effective length of 0·85 times the
purlin spacing will be adopted.
Try two 88·9 x 63·5 x 7·85 angles.
r", = 2'79 cm r = 2·70 cm A = 22·74 cm2
y
I", = 177·6 cm 4 L", = 1·94 m L = 1·95 m y
I", = 0·7 x 1·94 = 1358 mm I/r", = 1358/27'9 = 49

Iy = 0·85 x 1·95 = 1660 mm I/ru = 1660/27 = 62
pe = 124 N/mm 2 (BS 449: 1969, Table 17a)
Ic = P/A = 91 x 103 /2.274 x 103 = 40 N/mm2
The complementary compressive bending stress will occur over the
node point on the underside of the angle section (see Fig. 5.7). An
approximation may be made as to the bending moment occurring at
this point. It would be extremely difficult to achieve a true analysis of
the moments occurring along the length of the rafter.
Bending moment over support
=i x bending moment on rafter length assuming simple
supports
2 W X a X b 2 6·92 X 0·68 X 1·26
=-x L ="3 1·94
3 X
= 2·05kNm
I" =M X Y = 2·05 X 106 X 60·4 = 69.6 / 2
Jbe I 1776 X 103 N mm
Check stress ratios:
/c/p" = 40/124 = 0·323
he/be = 69·6/165 = 0·423
0·746 (Iess than unity)
Section is satisfactory.
DEFLECTED SHAPE
OF F.AFTEp, (?
~
" ..,
'2.~~ c..\~~c=T'" TENSiON
C.OMPRESSION
FIG. 5.7. Bending in roof truss rafter.
Secondary struts-Members 3-4 or 4-5. Design load 16 kN; length

1·83 m (centres of end connections). The maximum llr ratio is 180 for
members of this type (see BS 449: 1969, c1ause 33) and because the
load is relatively small it is Iikely that this limit will be approached. These
members are not continuous, as was the rafter designed above, and
provided that at least two bolts are provided at each end ofthe member,
an effective length ofO·85 times the distance between end connections is
acceptable (see BS 449: 1969, clause 30).
Try 63·5 x 50·8 x 6·25 angle.
r u = 1·07 cm A = 6·82 cm2
= 1830 mm [ = 0·85 x 1830 = 1560 mm
L
[Ir = 1560/10·7 = 146 pe = 42 N/mm 2
Je = WIA = 16 x 103 /0.682 X 103 = 23·4 N/mm2
(ru is the least radius of gyration for this section.)
The remaining angle struts carry very small loads and two criteria
should be applied to such members.
(i) Limit [Ir ratio to 180.
(ii) Decide on the minimum size of angle leg (say 50 mm) that is
suitable for connection purposes.
Minor angle struts- Members 1-2 or 6-7. Design load 10 kN;
length 860 mm.
Try 50·8 x 50·8 x 6·32 angle.
rv = 0·99cm
[ = 0·85 x 860 = 731 mm
[Ir= 731/9·9 = 74 pe = 111 N/mm 2
Ic = W/A = 10 x 103 /0.608 X 103 = 16·5 N/mm 2
It should not be necessary to check any of the strut members for load
reversal conditions caused by wind loads.
(ii) Tension members. It is convenient to use angle sections for
the tie or tension members but as only one leg is usually connected it is
necessary to make allowance for the eccentricity of the connection
by assuming that only a portion of the unconnected leg is effective
(see BS 449: 1969, clause 42). Tbe area of hole must be allowed for
and it is usual to deduct the area of one hole only from each member.
If holes are required in both legs (e.g. at a joint in the main-tie) then
the spacing of holes in one leg must be 'staggered' in relation to the
holes in the other leg. The major tie-member in the frame is bar Q-I
which is subject to load reversal under wind conditions. Tbis produces a
strut condition which is more severe than when tension is present. Tbe
length of this member is taken as 5·25 m which is measured from the
roof truss eaves to the point along the tie where angle ties running
longitudinally along the building length are provided.
Main-tie. Member Q-1. Load 99 kN (tension) 14 kN (compression).
Try 76·2 X 63·5 X 7·90 angle.
Gross area of connected leg = (76·2 - 3·9) X 7·9 = 572 mm2
Gross area of unconnected leg = (63·5 - 3·9) X 7·9 = 471 mm2
Total gross area = 1043 mm 2
(The gross area of a leg is equal to the product of the angle thickness
and the leg size less half the thickness of the angle.)
Net area of connected leg = 572 - (7·9 X 22)
= 398mm2
3al
Net area of unconnected leg = 471 X 3-:----"--
al + a2
3 X 398
= 471 X 3 X 398 + 471
= 338mm2
Total net area = 736 mm 2
Pt = 155 N/mm 2
(BS 449: 1969, clause 41)
Allowable load = 736 X 155/103 = 114 kN
The load reversal condition will now be investigated.
L = 5250 mm [ = 0·85 X 5250 = 4460 mm
[/r = 4460/13·2 = 339
(Limit of [/r ratio is 350. See BS 449: 1969, clause 44.)
pe = 10 N/mm 2 + 25 per cent
(See BS 449: 1969, clause 13.)
= 12·5N/mm2
Je = W/A = 14 X 103 /1.04 X 103 = 13·4 N/mm2
These figures indicate that the member is slightly overstressed. However,
the design length of 5·25 m ignores the restraint value produced by the
connection at mid-point for members 1-2 and 2-3, and for this reason
the proposed section is satisfactory.
Load reversal in the remaining members is not likely to prove
troublesome with the exception of the centre portion of the main tie
(Q-8) and the crown-tie (7-8). The sections shown for both these
embers exceed the [/r limit of 350 but the degree of reversal is so
small as to be capable of being ignored. If preferred the sections
mentioned can be increased as folIows:
End main-tie (Q-l) 76·2 X 76·2 X 7·85 angle
Centre main-tie (Q-8) 76·2 X 63·5 X 6·25 angle
Crown-tie (7-8) 76·2 X 63·5 X 6·25 angle
In the latter two cases the 76·2 mm leg should be outstanding from the
plane of the truss.
Minor angle fies. Following the precept used in the design of the
minor struts a 50·8 X 50·8 x 6·32 angle will be suitable.
(e) Design 0/ connections. A detailed drawing of the roof truss is given
in Fig. 5.8 and 20 mm dia black structural bolts are used throughout
in the main connections. In the smaller angles (50·8 mm size of leg)
16 mm dia bolts will be used. If convenient, it is usual to erect the truss
in halves before despatch from the workshop to the site. Each half truss
would consist of a triangular frame bounded by one rafter, a crown-tie
and the outer portion of the main-tie. The central portion of the main-
tie and the central vertical suspender would be despatched loose. Site
connections would be made with black bolts. If preferred the truss
could be despatched as a bundle of loose pieces for assembly at site.
The ultimate cost of the truss will be affected by the choice of method
employed and it may be that either will be chosen in practice depending
upon the particular conditions at the time.
In the rafter only, the connection bolts are in double shear passing
through two angles, or in bearing passing through the gusset or con-
nection plate. Usually the design of the connection is settled by the
bearing value of the plate through wbich the bolt passes. For tbis
reason it is customary to have a thicker gusset plate at points of double
shear in order that the bearing value of the bolt will be closer to the
double shear value of the bolt. Otherwise, the bearing value associated
with a thinner gusset plate will result in a larger number of bolts being
required. At the ends of the rafter 10 mm thick gusset plates will be
used and 8 mm thick gusset plates will be used elsewhere.
The tabulated load values for bolts are given in Fig. 4.22.
Load values for a 20 mm dia bolt are:
single shear 25·1 kN,
double shear 50·3 kN,
enclosed bearing (in 10 mm thick plate) 40 kN
simple bearing (in 10 mm thick plate) 32 kN.
RaJter: Load 109 kN.
Number of bolts required = 109/40 = 3 minimum.
Main-tie: Load 99 kN.
Number of bolts required = 99/25·1 = 4 minimum.
PlNGLE T IE
ASBESTOS-CEMENT SHEET5
tVV'~~A(/
~. . ) 'FDd=~ ....
r ~(n,", xG3 · S x ro .? - 41-10LIOS AT
8IH~xSS" )<7'9 AN<:>LE PUP,LlN ~ JOINT
PlIRLIN i, / 1
C.LEAT RooF TRUS5 RAFTER
PURLIN CLEAT5 V' EW f3
BOllS 20 DIA EXCEPT
\ MID- POINT OF IN 50 ·8 ANGL ES
Wt-lERE T HEY AR
{RMTEf'..
1(;; D IA. ,Ne;, ~A'O
CLADD ING ~ \
-.. C. L. TRU55
I 6~~1l~
<;Q
I'
RA FT EP-
I "2/65.9><(,,3. 5><7 .55 L
VI EW 1\ c-
GUSSETS
BARS ·c· ARE STHICK
50·6 x SO· 6)< " ' 32 L UNLESS
STATEP
RMTER A.
S LOPE
'22~D
"3 ·5" ro3 ·5" "''Z2L
[CENTRE PORTI ON OF
T I E RA ISED 75
,C.L .
9000 9000
" 371<9 UE!>
ELEVATION
FIG. 5.8. Details of roof truss.
Two bolts will be sufficient to transmit the remaining loads. No single
bolt connections should be used because the conditions of strut design
were based on an effective length of 0·85 Land an end connection of at
least two bolts.
Tbe number of bolts required in the intermediate rafter connections
should be sufficient to resist the difference in force between two ad-
joining panel members. In roof truss work the difference is small and
the necessity of providing a suitable connection is the criterion rather
than the problem of difference in force within adjoining members. It is
also necessary to ensure that the two angles forming the rafter section
act in unison with each other and do not tend to buckle apart between the
gusset plate connections. In double angle construction it is necessary
to ensure that the slenderness ratio (I/r) of the single angle between
fastenings is no greater than for the double angle over its whole design
length. To satisfy this condition an additional bolt is provided mid-way
between each gusset plate with a thick washer to act as a packing
piece between the angles.
Angle deats are used to secure the angle roof purlins. Each purlin
should have a two bolt connection at each end, making four bolts in
the vertical leg of the purlin deat where a purlin joint occurs and two
bolts elsewhere. Generally, purlins are provided with a two-bolt
connection at each end when the span exceeds 3 m. If a single bolt
connection is used below this span the effective length of the roof
truss rafter should be increased from 0·85 L to 1·0 L.
39 Design of Side Columns

(a) Design 0/ column shaft. The pressure coefficients for the wind
loads acting transversely across the building are shown in Fig. 5.2(b)
and (c). Internal wind pressure does not affect the overturning condi-
tion because it is acting equally in all directions. Wind load acts as
apressure over the whole surface under consideration, but for con-
venience of calculation it is assumed as concentrated at the centroid
of each surface.
Wind load on vertical faces
= 5·5 X 4 X (0·7 + 0·25) X 680/103 = 14·20 kN
Wind load on inclined roof surfaces
= Resultant horizontal component produced by 0·4 and 0·3
pressure coefficients
= (0·4 - 0·3) X (10 X 4 X (680/103 » X sin 22!0 = 1·03 kN
Overturning moment about column base line (see Fig. 5.9(a»
= 14·20 X 3·25 + 1·03 X 7·86 = 54·2 kN m
It will be assumed that the overturning moment is shared equally
between both columns. This assumption is based on the reasoning
that both column caps deflect an equal amount because they are
connected together by a roof truss which is capable of passing the
out-of-balance forces produced by the wind from one column to the
DIRECTION [0·3] [0-4]

OF WIND .. \~~...=.;::..::r-~!
~
"? E
[0..21..- [O'7tO' 2 51 14''20k N "
.n
~E
'"
1- 27·[ kN""
15·Z3kN
"..- 27·lkN""
3·02kN
"-INDUCED
[a.]
FIG. 5.9. Wind load sustained by side columns.
other allowing an equalisation of moment to take place. The out-of-

balance force at column cap level may be calculated as folio ws (see
Fig. 5.9(b)).
Moment at column base = 54·2/2 = 27·1 kN m
Wind load on column face
= 5·5 X 4 X 0·7 X 680/103 = 10·48 kN
Wind load at column cap level transferred from roof truss
= 1·03/2 X 7'86/6 = 0·67 kN
(Half of the roof wind load (1 ·03 kN) assumed to be applied at each cap
level and increased in relation to height of mid rafter level/co lu mn cap
level.)
Out-of-balance force (A) at column cap
= (10,48 X 3·25 + 0·67 X 6 - 27'1)/6 = 1·84 kN
This force is small and the rooftruss will be capable oftransmitting this
load without any modification being necessary to the existing design.
On large buildings it may require consideration and the roof truss
main-tie designed to accommodate this additional load.
In addition to the wind loading causing a bending moment at the
column bases it will also result in an increase in the load in the column
on the leeward side of the building and a corresponding decrease in
the load in the wind ward column as the overturning effect takes place.
Load induced in columns by wind
overturning moment 54·2
= = - = 3·02kN
span of columns 18
An assessment of maximum and minimum loading conditions in the
columns may now be made.
Max. kN Min. kN
Dead load from roof truss 17·05 17·05
Superimposed load from roof truss 27·20
Weight of side framing and sheeting 5·50 5·50
Self weight of column 3·00 3·00
52·75 25·55
Induced wind load (+) 3·02 (- )3,02
55·77 22·53
Moment at base due to wind = 27'1 kN m.

A preliminary choice of section may now be made bearing in mind
two conditions.
(i) In a column loaded as shown the effect produced by bending is
very much greater than the effect produced by axial load and for this
reason the column can be thought of as a bent vertical beam rather
than as a conventional column in wh ich the axial load effect is much
greater than the bending effect. A VB section should be chosen in
preference to a VC section.
(ii) Because of the canti lever be am effect it is likely that horizontal
deflection at the column cap level must be considered. The defiection
or displacement must not result in damage to the cIadding or lead to
unsightly deformation of the structure. It is doubtful whether an exact
analysis of this deflection could be made because it is affected by such
undefined factors as partial restraint at the eaves connection between
column and roof truss, stiffening effect of the cIadding upon the
building, floor construction, and the partial relaxation of the assumed
fixity of the foundation. For the purposes of this design the column
may be taken as behaving as a simple beam canti lever carrying a
uniformly distributed load producing a moment at the support equal
to the base moment of 27·1 kN m. Any modification to this assumption
must be made by the designer in the light ofhis experience with buildings
of this type.
Wx L3
Deflection = 8 X E X J
But M = W x L/2 and the deflection may be written as M x V/
4xExI.
Following the recommendations of BS 449: 1969, clause 31, the
deflection will be limited to 1/325 of the column height.
·
MaXlmum dfl·
e ectJOn = 6000
325 = 18 mm
27·1 X 103 X 6000 X 6000

Then 18 = 4 X 210 X I
and I = 6·45 X 107 mm 4 = 6450 cm4
A 304 X 124 X 37 kg UB has a moment of inertia = 7143 cm4 and

may be adopted as a trial section wh ich must now be checked for
conventional column behaviour.
lz = 1·5 X L = 1·5 X 6 = 9·0 m
ly = 0·75 X L = 0·75 X 6 = 4·5 m
(See BS 449: 1969, appendix D, Fig. 15.)

= 9000/123 =
I/r z 74 I/r ll = 4500/25·8 = 175
pe = 30N/mm2
(BS 449: 1969, Table 17a.)
D/T = 304/10·7 = 30
pbe = 82 N/mm 2
(BS 449: 1969, Table 3a.)
In checking the stress ratios it will be necessary to investigate two
conditions:
(i) Dead plus superimposed loadings in which the stress ratio must
notexceed unity (1.0).
(ii) Dead plus superimposed plus wind loadings in which the stress
ratio requirement may be increased by 25 per cent (1·25). The increase
in stress in case (b) when compared with case (a) is due solely to wind.
Wind forces can be classed as an occasionalload in that it is unlikely
that a building will be required to withstand maximum wind loadings
very frequently. In fact the wind pressures recommended by CP3 are
based on a maximum gale situation occurring only once in fifty years.
A 25 per cent increase in stress is justifiable in these circumstances
(see BS 449; 1969, clause 13).
Check stress ratios.
Dead plus superimposed 10ading
tc = W/A = 52·75 x 103 /4.74 x 103 = 1H N/mm2
f"e = 0
tc/Pe = 1H/30 = 0·37
f"e/Pbe =0
0·37 (less than unity)
Dead plus superimposed plus wind loadings

tc = W/A = 55·77 x 103 /4.74 x 103 = 11·8 N/mm2
!oe = M/z = 27·1 x 106 /470 x 103 = 57·7 N/mm2
tc/Pe = 11·8/30 = 0·394
fbe/Pbc = 57·7/82 = 0·705
1·099 (less than 1·25)
The proposed section may be regarded as satisfactory.
(b) Design 0/ column connections. Connections will be required on

the column for three purposes.
(i) Column cap, (ii) sheeting rails to support cladding, and (iii)
column base.
In cases (i) and (ii) the only requirement is for an adequate connec-
tion to be made between the parts and no calculations are proposed
because the loads are smalI.
Item (iii) (column base):
Maximum load = 55·77 kN
Minimum load = 22·53 kN
Bending moment = 27·1 kN m
Base shear shared equally between two columns
= 7·61 kN (see Fig. 5.9(a»
A built-up gussetted base will be adopted in which the gusset plates
will transfer the bending moment to the base plate. A suitable size
for the base may be determined by deciding the spacing of the anchor
bolts required to give a good resistance to overturning and for this
purpose an anchor bolt spacing of 450 mm is adopted. The anchor
bolt spacing across the other axis of 225 mm will result in a suitable
spacing between the bolt boxes during the casting of the concrete
foundation block.
Referring to Fig. 5.10 it is reasonable to suppose that the point of
rotation would be the line BC (point A in elevation).
55·771<.N [Ma.)C.1
.Z'Z.53kN [Min)
27"1 kNII'l
'Zoo
10
t.
~o DIA x....-"
4'50 A"ICHOP. 377
Bo LT l--..::c.:.;'-t--ook-::..;r
ALL WElDST E.LEVATION

Go FILlET
'"00
~~
1 ~~
'2'2. 5 I
]: I
:!~ "ilF
~I:: l~
450
1
c
PLAN
HG. 5.10. Column base design.
Moment occurring about point A

= 27·1 - 22·53 x 0·152 = 23·68 kNm
This moment is resisted by the tension load (T) in the anehor bolts.
23·68 x 103
Force (T) in 2 anehor bolts = 377
= 62·8 kN
or, 31·4 kN per bolt.
From Fig. 4.22 it will be seen that a 20 mm dia bolt will earry a tension
load of 28 kN. If an allowanee of 25 per cent in the allowable load is
",ROOF TRUSS
~~~~~
200 I PLATt: \0 ,HICK
H.f'. . GUT ER
: I
I·r·:~.r. ' 1150
I' .f, .u • ...1.
~ 1
;z
Ci PLAN ON CAP
A GlE R 1\ IL Cl
lOH" " 7"''2 ~ 7· 90 L -<
..J
V Sl'IAF'T Of COUJMN
304 x 124x 37k~ UB
5500
f-
Z ß.o-SE.
w
r: 450 PLATE
710 · 2 UJ ~IOTHICK
u
,
./
ANGLE CLI;. ...T
VIEW'A'
..... Vl
0 Z~S ~~300
f-
I I
81\·9.< 710·2 ~ 7·82 L
VI
ILJ
<0 15 ~OO
4 HOLES AT
JOINT
'"
..:: PLAN ON BASE
ANCHOR BOlT
20 DIA
450 LONC,\ ONCRETE
AROCORE.
ÄNCHOR
PLATf.
10 TH ICK
VIEW'A,' SECTION T\-IROUGH SIDE. OF ßUILDING
FIG. 5.11. Details of side columns.
made because the force in the anchor bolts is due to wind, then the
size of bolt proposed is adequate. The anchor bolts can be 450 mm
long with 10 mm thick anchor plates. Details of the column are given
in Fig. 5.11 and reference should be made to section 29 for a description
of procedure at site.
40 Design of Mass Concrete Foundation to Resist Overtuming
Two ca!Oes will be considered for foundation design.
(i) Dead plus superimposed plus wind loadings. (Overturning with
maximum verticalload.)
(ii) Dead plus wind loadings. (Overturning with minimum vertical
load.)
A foundation block relies for its stability upon positive pressure
acting across the fuH area of the base and the bed joint between the
ground and the concrete block cannot resist tension because there is no
jointing medium present. Because the overturning moment is large in
relation to the vertical load a block of sufficient weight should be
provided to ensure that no tension takes place at the heel of the base.
The two loading conditions are shown in Fig. 5.12.
w::: tOO'S7kN
M : ± '27· t kNrn
DETAILS OF LOADtNG
+KI I J I I I
PRESSURE DUE TO W
I:
cl. =
-IT~ 1791
PRE5SU~E~+T"i: 3 x 597 ~I•

DUE TO M ~
'!:!._t1/~!I
A Z w 1'1
COM61NED
PRESSvRE
A Z )
CASE I (ASE TI
FIG. 5.12. Foundation block design.
Case 1
Try a 2 m x I m x 1 m deep block.
Verticalload = 55'77
Weight of block = 2 x 1 x 1 x 22·4 = 44·8
100·57 kN
bd 2 1 X 22
Z of base = ""6" = - 6 - = 0·67 m3
Pressure due to verticalload
= -AW = -
100·57
2-- - 50·28 kN/m 2
Pressue due to overturning moment
= -Mz = -0·67
27·1
= 40·5 kN/m 2
Pressure at toe of block

W M
= A + Z = 50·28 + 40'5 = 90'78 kN/m:l
Pressure at heel of block
W M
= A - Z = 50·28 - 40·5 = 9·78 kN/m2
It will be seen that a size of block has been chosen with the necessary
weight to offset the overturning effect and ensure a positive pressure
over the fulllength of base.
Case 11
Verticalload = 22·53
Weight of block (as before) = 44·8
67·33 kN
A first investigation will show that the (W/A) pressure is less than the
(Mlz) pressure indicating that tension appears to be present at the
heel of the block. However, it is possible to overcome this by limiting
the length of block in the following manner.
Move the 67·33 kN load a distance e from the centre-line of the
block so that the overturning moment on the block remains the same.
This in no way affects the loading on the foundation and it is still
subject to the same verticalload and bending moment as before.
M= Wxe
M 27·1
or e = W= 67.33 = O'403m
If compression or positive pressure is present over the full base length,
the distance e must be less than 1 of the base length. In this case
1 x base length = 0·33 m indicating that if the full base length is
used, then tension will exist at the heel of the base.
Assume a restricted base of length d 1 • If zero pressure is present at
the heel of a length d 1 the position of the vertical load must coincide
with the third point next to the toe of the block. This third point lies
O' 597 m from the toe of the block.
d1 = 3 X 0·597 = 1·791 m
Apressure diagram as shown is applicable in which the area of the
pressure diagram is equal to the verticalload.
p X 1·791 X 0·5 X 1 = 67·33
P= 75 kN/m 2
A size of block was chosen to demonstrate the two conditions of
pressure and it is likely in practice that a smaller block would prove
suitable because both calculated bearing pressures are low.
41 Design of Side and Gable Sheeting Rails

On the sides of the building supporting the asbestos sheeting or cladding
are angle rails spaced vertically at 1·82 m and spanning 4 m on the
building sides and 4·5 m on the gables. The larger span will be adopted
for design purposes indicating that one size of angle will be adopted
for both locations. A bending moment equal to W X L/lO will be used
for design purposes.
Spacing of rails = 1·82 rn
span = 4·5 rn
Vertical loading: kN/rn 2
Asbestos sheets = 0·161
Insulation = 0·036
Self weight = 0·050
0·247 kN/rn 2
Load = 4'5 X 1·82 X 0·247 = 2·02 kN
M= W xL = 2·02 X 4·5 = 0'91OkNm

10 10
Try 101·6 X 76·2 X 7·90 angle (101·6 leg outstanding).
f (top edge, cornpression) = M; Y
= 0·91 X 106 X 19·0 = 26 N/ 2

668 X 103 rnrn
DESIGN OF 18 m SPAN SHED III
0'91 x 106 X 57·2
f(bottom edge, tension) = 668 X 103 = 78 N/mm2
The wind pressure coefficient is made up of two parts.
(a) Local external coefficient (Cpe ) = -1,0
(b) Internal coefficient (Cpl) = +0·2
Sum of pressure coefficients = Cpe - Cpl
= -1,0 - (+0'2) = -1,2
Dynamic wind pressure (q) for cladding = 755 N/m2 •
Horizontal loading
= Area X q = 1·82 X 4·5 X 1·2 X 755/103 = 7·41 kN
M =W xL = 7·41 X 4·5 = 3.34 kNm
10 10
3·34 X 106 X 31·6
f(front edge, tension) = 1380 X 103 = 76·5 N/mm2
. 3·34 X 106 X 70 2
f(rear edge, compresslOn) = 1380 X 103 = 170 N/mm
The disposition and summation of the stresses is given in Fig. 5.13.
'2·02. kN
+ 210
7(;"'5
- '50''5
J 31·'" 70 +26
+ 170
+1'36
-::::::::::,.. _I / -
-+!f
19·0 ,x 1~)I.tD)(7&'2 x 7·9
7·4 IkN L
':I Je
57·2 I
ALL STRESS ES
IN N/Mrtl2.
-78
-78 / STRESS DISTRIBUTION
-7&'5
DUE TO VERTICAL LOADING
--
- I 5 4 '5
,
IJnJ-r.....I
'4tt1UIIIllllil+
-710'<;;
170
STRESS DISTRIBUTION DUE
TO HORIZONTAL LOADING
FIG. 5.13. Stress distribution in sheeting rails.

Horizontal deflection assuming simple supports
5 X W X L3 5 X 7·41 X 45003
- 384 X E X 1- 384 X 210 X 1380 X 103 = 30·3 mm
30·3 I
= 4500 = 148 span
Although this deflection appears to be high it is not likely to result in
damage to the sheeting. If the angle rails are secured by two bolts at
each end the deftection will be rather less (say 20 mm) than the cal-
culated figure which is appropriate to simple supports.
The maximum calculated stress is compressive in nature, occurs at
the inside edge and is approaching the maximum of 206 N/mm2
(165 + 25 per cent). This edge at which the maximum stress occurs is
an unrestrained compression ftange but it may be assumed in the
design of sheeting rails that the stiffness of the sheeting together with
the hook bolts securing the sheets provide the necessary restraint.
Some designers contend that to prevent sag in a horizontal member
a limit of 200 should be placed on the value of [Ir for the member.
This is purely an empirical rule. Applying this idea to the present case
the l/r = 4500/22·3 = 202.
It is of interest to give further consideration to the design of the
angle rails. The span, as weil as the wind loading taken for design
purposes was the worst of several values applicable to the building
and it would be possible to adopt reduced sections where appropriate.
However, two further alternatives exist, the first of which is to reduce
the column spacings adjacent to the gable ends because it is on these
areas that the worst wind conditions apply. The effect of span reduction
would be to make the smaller section, appropriate to a lower pressure
area of the building, usable in areas of higher pressure. Or, the
size of angle rail appropriate to a lower pressure area may be adopted
and additional angles of the same section used in higher pressure areas.
Additionally it might be assumed that the vertical weight of the
sheeting was supported by the foundations, thus relieving the angle
rails of this portion of the load and resulting in a lighter section being
acceptable.
Whatever decision is made in assessing the validity ofthe above points
the design engineer will always attempt to provide an economical
solution appropriate to the particular environment in which he is
working.
42 Design of Gable Steelwork

The gable steelwork is required to resist a considerable area of wind
pressure in addition to the verticalload caused by the self weight of the
framing and cladding. It is in dealing with the wind pressure that the
main difficulty arises, and one solution which is frequently used in
practice is to support the gable columns at roof truss tie-level with a
horizontal wind girder (see Figs. 5.1, 5.14). Load reactions from the
gable wind girder at the corner columns of the buHding are transferred
to the foundation level by means of bracing in the sides of the building
(see side elevation Fig. 5.1).
(a) Design 0/ gable wind gi,de,. Two cases of wind pressure acting
on the gable frame require investigating (see Fig. 5.2).
Case J. Wind blowing on to gable. Total pressure coefficient in direction

of wind = 0·7 + 0·1 = 0·8.
An assumption will be made that the loads resulting from this
pressure may be divided equally between each gable and because the
wind may blow in either direction the members of the wind girders
must be designed for load reversal. Any out-of-balance loads between
the two wind girders will be passed along the structure, mainly by the
angle ties at roof truss tie-level, but also by the roof purlins, side rails
and cladding.
Case JJ. Wind blowing perpendicular to gable (normal to side of

building). In this case the pressure coefficient is 0·6 acting equallyon
each gable as a suction loading. Provided that the gable columns are
adequately framed and connected together, as they are in this case, by
angle ties at roof truss main-tie level, then no out-of-balance loads
exist. Provided that the angle ties are adequate to resist the tension
forces set up by this condition of wind loading it is not necessary to
consider Case 11 any further.
From the above it will be seen that Case I is the criterion for design.
Dynamic wind pressure on each of two gables
= pressure coefficient X q
0·4
= 2" X 680 = 272 N/m2
Referring to Fig. 5.1 5:
Wind load on area A = 2·25 X 3·212 X 272/103 = 1·96 kN
Wind load on area B = 4'5 X 4·60 X 272/103 = 5·65 kN
Wind load on area C = 4·5 X 5·988 X 272/103 = 7·32 kN
The force diagrams and tabulated loads in the bars of the gable wind
girder are given in Fig. 5.1 5.
11 304 x 124 x 37 kS UB C.L. OF ROO F "TRUSS
I. n j 15, C.l. OF GUS5E TS Dl
?'OOF T RUSS MAIN- TI E /"
8&'9 ~ 76-2 x 6-30L _ _ 200011'

r
GU55 ET PlATE5 8 lHleK "'/ 111710-'2.
ßOL T5 20 DIA
710-2 I
11 4000
'o0 .'bi-1o''L'' \- ~p- - _.~ 2:J!p m~
- f \1
-:>_<)1- -_\-
2000 11I
v~? =
\'
101_10 " 710-2 '" 7- 90 l ~

C 'Zo3xl:l3", 25k9 U l!> D 5ECTION D-D
B 4500 4500 4500
~Sll'
102
1 i
304xl24x37k9 UB
5Ec.rION c-c
SEC 11 0N ß-B
PLAN ON WIND GIRDER [ 5EGTION

ON FIG 5- 1 ~
A-Al
FIG. 5.14. Details of gable wind girder.
A 0 ~
- r______________~-,~----------~2rw
~ID~
3·7 ....
~~i5 :
1\ c~o-,>
WIND ,
2·75
m
,A
1
~_ , __
~::J!
<cd .....
..... "'1~oEP.
J
GA&LE
COLUMt-J
..,
2·75
FLooR ,I
4·5 4.5 AtJ

4·5 4.5 SECTION A-A
M no-. ",., m
ELEVATION oF GABL.E SHOWING
ARE.A5 RESISTED Sy WIND GIRDER
~AR FORc.E.~I<.Nl
ANGLE TIES
1 1/ V -;rROOF TROSS I
~-I
E-Z
.~~
-- - --I-H K-I 23
IM/~1~\~1F
J-3 30
1-2- 9
D '3-4 7
f
(·%kN
t f t5'6~ t
K 5.GSJ 7.32H
FRAME DIAGRAM 1'~6kN
cl.
~r-------------~~ 11·27\(101
Jr---~~~------~--~'
e~--------------~ FORCES F=OR
SIDE SP,ACING
hr---~~~------~----~
9r-------------~~
f FORc.E DIAGRAM
.5IDE Bf1..AcING
FIG. 5.15. Gable wind girder. Frame diagram, force diagram

and table of forces.
From the table of loads in the bars it will be seen that the loads are
smalI, and bars sizes are likely to be very slender. To maintain rigidity
and prevent sagging of the bars under their own weight, it will be
necessary to place a limit on the Ifr values for the individual bars.
Wind forces produce loads which only reach their full intensity on a
very few occasions, and a higher value of Ifr is permissible than for
main members of a frame carrying permanent or nearly permanent
loads. A maximum value of l/r equal to 250 will be adopted. Any
proposed section must be checked for excessive deflection as weIl as
strut behaviour.
Bar E-l. Load 25 kN.
1= 0·85 L = 0·85 X 5200 = 4410 mm
If l/r = 250 then r = 4410/250 = 17·7 mm. 2-63'5 X 50·8 X 6·25
angles bol ted together as shown in Fig. 5.14 have a r value of 19·5 mm.
Check for normal strut behaviour
~ = 4410 = 226
r 19'5
pe = 19 N/mm 2
As the loading is wholly induced by wind the above value may be
increased by 25 per cent:
pe = 19 + 25 per cent = 24 N/mm 2
Allowable load = area of section X pe
= 1364 X (24/103 ) = 32·8 kN
The proposed section is satisfactory. A depth of 110 mm (including
thickness of gusset plate) is available in this section to offset deflection
which is likely to occur in long struts carrying smallioads and which is
due to the self weight of the member. An approximate rule to allow
for this condition is to make the section depth equal to I/50 of the
span. In this case the ratio of span/50 is equal to 104 mm which is
satisfactory.
The remaining members may be designed in a similar manner. To
give the necessary strength to members K-l and J-3 it is connected to
a conveniently placed gable sheeting rail by means of batten plates
(see Fig. 5.14). Bar 1-2 is not required to carry a large axial load but
supports the dead weight of the girder and this should be borne in
mind when deciding a section for this particular member. An
88·9 X 76·2 X 6·30 angle is recommended for this section and meets the
span/50 rule previously adopted. This member must run unbroken
from the gable frame to the next adjacent truss and not be jointed at
any intermediate gussets.
The design of the bracing in the sides of the building required to
transfer the wind girder reaction load to foundation level will now be
considered. A single angle section arranged as shown in Fig. 5.16 is
satisfactory and the bracing system consists of the top sheeting rail
with the addition of the inclined angle already mentioned. To increase
the efficiency of the top sheeting rail it should be battened by bent
flats to the lowest purlin on the roof slope and if this is done it will not
be necessary to investigate this member any further. With regard to the
vertical load of 14 kN, the column is wholly adequate to carry this
small additional force. The load of 22 kN in the inclined bar may
be tensile or compressive in nature, depending upon the direction of
the wind. A 1ength of 7 m is excessive for economical strut design and
this may be reduced by fastening the bracing to the side sheeting rails
along its length. An estimated effective length for this member will be
taken as 1·25 times the distance between sheeting rail restraints,
bearing in mind the difficulty of accurately estimating the effect of
restraints in this case.
L = 2·22m 1 = 1·25 X 2220 = 2780 mm
Limit l/r to 250 then r = 2780/250 = 11·1 mm.
A 76·2 x 63·5 x 6·25 angle has a minimum r value of 13·3 mm.
Check for strut behaviour:
l/r = 2780/13·3 = 209
pe = 22 N/mm2 which may be increased by 25 per cent to 28 N/mm 2 •
28
Allowable load = 836 x 103 = 23·4 kN
(b) Design 0/ gable column. The gable columns will behave as simply
supported vertical beams spanning between the ground and the wind
girder position. This is a simplification of the situation in that the
columns will behave as a form of continuous member beyond the
wind girder level. Because of this continuity the maximum bending
moment is taken as WL/lO as compared with WL/8 for a simply
supported member.
Wind pressure coefficient for gable
= 0·7 + 0·3 = 1·0
Wind load acting on one gable column
= 4·5 X 5·5 X 1·0 X 680/103 = 16·8 kN
Bending moment = WL/lO = 16·8 X 5·5/10
= 9·25 kNm
Verticalload (gable sheeting, framing, etc.) = 14 kN
Try 203 X 133 x 25 kg UB:
l",= 0·85 L = 0·85 x 5500 = 4675 mm
l~ = 0·7 L = 0·7 x 5500 = 3850 mm
I/rz = 4675/85·3 = 55 I/r y = 3850/29·5 = 131
pe = 51 N/mm 2 (BS 449: 1969, Table 17a)
D/T = 203/7·8 = 26 pbe = 126 N/mm 2 (BS 449: 1969, Table 3a)
Check stress ratios:
_ W _ 14 X 103 _ • 2
/c - A - 3.23 X 103 - 434 N/mm
M 9·25 X 106
f"e = z= 231 X 103 = 40·1 N/mm 2
/c/Pe = 4·34/51 = 0·085

!be/Pbe = 40·1/126 = 0·319
0·404
The proposed section is satisfactory.

It should be noted that although the sum of the stress ratios is low
the I/r ratio is relatively high and any reduction in section is not
recommended.
In order to triangulate the gable frame below eaves level and to
increase its stability, a nominal bracing angle is introduced between
the corner column and the next gable column. A 76·2 X 63·5 X 6·25
angle will be satisfactory provided that it is fastened to the gable
sheeting rails in a similar manner to that employed for the side bracing.
This additional member will help to 'square-up' the structure during
erection. Such members are frequently used in practice for this purpose
although not required on the grounds of strength.
(c) Gable rafter. The loading condition for this member is similar to
the gable sheeting rails and the same section will be adopted. A small
angle (63·5 X 50·8 X 6·25) is introduced into the gable frame to
support the gable rafter at mid span points between the columns and
to complete triangulation of the frame. Details of the gable framing are
shown in Fig. 5.16.
43 Slendemess Ratios of struts (l/r)

In this chapter four limiting values of I/r have been used and to
prevent confusion they are summarized below.
(i) I/r = 180. This is the usual limit applied to columns and all
struts forming the main frame of a building.
(ii) I/r = 200. Applied to secondary members with light loading
where it is desirable to limit the flexibility of the secondary member.
'"
ALL HOLES COU!\JTERSUNK
ON SHEETING FACE
RAILS 5ECl!RED TO COlUMNS
BY 88,9" 88·9" 7-90 L C.LEATS
WElDED TO COLUMNS
GUSSET PLATES 8 TH ICK
I
101,,, x 76'1. " 7· 90L.
\J~---f ~ I 76'2 ~
f A : c =;1"""~ ...... ~~ ----.t* --..... b" ~41A
5"500 TO HOOR " ,
304 x 124)( 37kg UB 203" 133>< 25kg UB _ _ _ _ _ _
I ./
76-2" <03 - 5 >< 0 ' 25 ' , /
BRACING ........ '-11
7r,,·Z
4500 111 4500 4500
FOR 5 EC TI ON A-A
SEE FIG 5·14- 5ECT ION
PAR1 ElEVAT ION OF GABLE FRAME. B-B
FIG. 5.16. Details of gable frame.

(iii) Ifr = 250. A limit applied to members introduced into a structure
to resist wind loadings only.
(iv) Ifr = 350. Applied to roof truss members which normally act
in tension but are subject to areversal of load during maximum wind
conditions.
44 Note Regarding Design Loads and Choice of Sections

In the examples of design given in this chapter it will be noted that
many of the design loads are smaIl, and for this reason it may be
thought possible that they may be ignored. To the experienced designer
the abbreviation of calculations is only possible because of his ex-
perience. This text is primarily concerned with design from a student's
or young engineer's point of view and in this context no such abbre-
viations are possible, nor have they been made. The student should
concentrate on correct load assessment, detailed design 0/ each member,
and maintaining equilibrium in all parts 0/ the structure as the prior
considerations. From the point of view of design the condition of
equilibrium will be met if the following three conditions are satisfied.
(i) Aigebraic sum of all horizontal forces and reactions acting on the
structure should be zero.
(ii) Aigebraic sum of all vertical forces and reactions acting on the
structure should be zero.
(iii) Aigebraic sum of all bending moments taken about any point
on the structure should equal zero.
For the purposes of design a structure is broken down into smaller
parts and sometimes into a two-dimensional problem. The student
should always be prepared to think in three dimensions because it is in
this context that the building exists. When an experienced designer
decides to adopt a section which appears to be aminimum, or less
than minimum size, he is judging the issue against his experience of
three-dimensional behaviour of similar complete structures rather
than upon consideration of a particular member of the structure in
isolation. To the inexperienced onlooker this sometimes appears to
consist of arbitrary decisions, but as stated above, it is only possible
because of his considerable experience. No attempt has been made to
simulate this refinement in this text and consequently some of the
sections chosen may appear to the experienced engineer to be less
competitive than he would accept in practice.
45 Design of Rainwater Gutters and Downpipes

Although rainwater goods are not part of the steel frame it is applicable
to consider suitable sizes in order to complete the scheme. Rainwater
goods are usually designed using empirical rules and the following
may be considered suitable for rainfall conditions likely to be
encountered in this country.
Centres of down-pipes = 12 m.
Cross-sectional area of one pipe = plan area drained/14·4 X 103
(mm units).
Cross-sectional area of gutter = twice area of downpipe.
Applying these rules to this example:
Centres of down-pipes = 12 m (fastened to every third column).
. db d . 12 X 9 X 106 00
Area d rame y one own-plpe = 14.4 X 103 = 75 mm.
2
A 100 mm dia down-pipe has a cross-sectional area of 7860 mm 2 •

Cross-sectional area of gutter = 2 X 7500 = 15 X 103 mm2 •
A 200 mm half-round (Le. semi-circular) gutter has a cross-sectional
area of 15·7 X 103 mm2 •
Materials from which rainwater goods are made are cast iron,
asbestos-cement, aluminium alloy, galvanized sheet steel, or rigid
plastics. Each length of gutter should be 2 m long and supported by
two forged flat steel straps bol ted to the lowest purlin. Down-pipes
are manufactured in similar lengths but one forged steel strap is
sufficient fastening for each length when bolted to the side column.
6. Design of Braced Tower Supporting
Water Tank
46 Particulars ofScheme
An open (i.e. unsheeted) tower structure is required to support a
prefabricated steel water-storage tank. The water tank is 6 m square
by 2'5 m deep and the underside ofthe tank is to be 18 m above ground
level. Safe ground pressure may be taken as 210 kN/m2 and the struc-
ture must be capable of resisting a wind pressure of 950 N/m 2 • Grade
43 steel is to be used.
47 Layout of Steelwork
The proposed layout of steelwork is given in Fig. 6.1, and a summary
of the relevant points is given below.
(a) Tank-supporting beams. The spacing of the beams will depend

upon the size of thc units from which the tank is built. Several makers
of tanks market standard-size steel or cast iron units which can be
assembled to give any size of tank required. In this example it will be
assumed that 1·2 m square units are available and that a beam will be
required under each joint in the tank floor.
(b) Corner legs. A suitable spacing for the corner legs must be decided.
It is desirable that the legs should be spaced as far apart as possible
to give a good width of base to resist overturning. Alternativcly, it may
be considered advisable to narrow the spacing of the Iegs to allow the
tank-supporting beams to cantilever over the sides of the tower. The
bending moment induced in the cantilever portion of the tank-
supporting beams will offset the bending moment produced in the
central portion of the bcams. Approximately one third of the weight of
DESIGN OF BRACED TOWER 123
Gm
"WAHR TANK ~I
I I I I
I I I
- ~- i- - ,-""1-
I I I I
TANK ./ ~~=EA=M='=='==~~'~E
(\I
5UPPoRTING
BEAMS
J-
PROPoSED ..c
LINE OF ~#=====~=9--t~
Ace E55
LADDEF.
4.25 <l·875
"" m
PLAN TANK SUPPORTING BEAMS
HORIZONTAl. I~~~=====;i
DI/'IGONAI.
CROSS
BRACIN(i
18m
PLAN ON HORIZONTAL DIAGONAL

CROSS BRACING AT POSITIONS A-S
4·25m
Ir ' - - - - ''I
4· 5....
4·'25m
L _ _ _ _ ..J
ElEVA110N OF TOWER PLAN ON 10WER LEGS
FIG. 6.1. Water tower.

Framing diagram.
the structure is contained in these beams and it would be desirable to

effect an economy along these lines. As a compromise between these
two alternatives it is proposed to space the corner legs 4·25 m apart
giving an overhang of 0·875 m at each end of the tank-supporting
joists.
It could be argued that good economy could be obtained by giving a
generous overhang to the tank-supporting beams and sIoping the
corner legs outwards to give the required spread at the base of the
tower. This would mean that the pattern of side bracing does not
repeat, as in the case of parallel legs, and the overall economy of the
structure would be impaired. It is doubtful whether sloping legs can
be justified in any tower less than 25 m high. If the tank being supported
is smalI, then sloping legs are justifiable.
A leg spacing of 4·25 m gives a ratio of leg spacing to tower height of
4'25/18, or approximately 1 to 4. This may be considered satisfactory
as giving a good resistance to horizontal deflection due to wind press ure
as weil as giving a compromise solution to the varying economics of
tank-supporting beams, side bracings, and tower legs.
(c) Side hracings. The tower may be regarded as four vertical cantilever
lattice-girders formed together in the shape of a square box. A girder
with parallel booms has already been adopted and it now remains for
the layout of the internal bracing to be settled. This bracing must be
capable of resisting forces induced by wind pressure as weil as providing
restraints at suitable centres for the corner legs and giving three-
dimensional stability to the structure.
-
DIREC.TION I'--~
OF WIND
SUl!>
FRAMING
(GI.) (b)
BG. 6.2. Development of side bracing system.
When the wind is biowing in the direction indicated in Fig. 6.2 the
N-form of bracing shown is economical because the long diagonals
are in tension. Sub-framing is introduced to provide additional
restraints to the corner legs. A similar lattice girder is required for a
reversal of wind direction and will be 'opposite hand' in form. The
superimposition of these two girder forms gives the arrangement
of bracing shown in Fig. 6.2(c). This system of bracing is a duplicate tie
system in which the diagonals are designed as ties, it being assumed
that only one diagonal is in operation in any one panel, depending
upon the direction of the wind, and the remaining diagonal is in-
operative. Analysis is simple and this type of triangulated framing is
extremely effective in practice.
(d) Horizontal diagonal cross-bracing. In order to preserve the square

of the structure under conditions of diagonal wind, a system of bracing
is introduced at positions A-B (Fig. 6.1).
(e) Restraint 0/ compression flanges and webs 0/ tank-supporting

beams. In order that the flexural bending stress in the beam flanges
may be the maximum allowable, tie angles have been included in the
top tier of beams at the centre of the span. Web brackets are provided
at the support positions of the lower tier of beams to ensure that the
wind load occurring on the tank is adequately transmitted to the tower
portion (see Fig. 6.7).
48 Investigation of Wind Pressures

Investigation of wind on open structures raises several interesting
points.
The shape of the members will affect the intensity of pressure. A
flat plate when placed in a wind stream will exhibit different character-
istics from an angle section. It is possible for the combined effect of
the positive and negative pressures to exceed the basic unit pressure,
due to the effect of turbulence around the edges of the section. To
allow for these conditions it is desirable to adopt higher values of
unit pressure than would be considered suitable for clad or sheeted
structures.
Because the structure is open, it is possible for wind to pass through
from the wind ward side to the leeward side and the exposed surfaces
of each side are required to resist full wind pressure. The size of mem-
bers in this type of structure is usually small, and it would be unwise
to assume that the windward side affords any shelter or shielding effect
upon the leeward side. Bearing in mind the above factors the design
wind pressure of 950 N/m2 was decided.
(a) Conditions 0/ wind loading to be considered. (i) Wind blowing

normal to side of tower (see Fig. 6.3, Case I). If sides AB and CD
of the tower are exposed to wind pressure it may be assumed that this
pressure is resisted equally by the vertical lattice girders in sides AD
and BC. Local bending between A and B, and C and D, is ignored
together with effeet of wind blowing parallel to AD and Be. Overturn-
ing would take plaee about the tower eentre-line and legs A and B
would resist tension with legs C and D resisting eompression. The
bracing in sides AD and BC would beeome stressed. Overturning
would be resisted by Jour legs and two sets of side braeing.
Ul-
A.
-0- ..::
--
UJ
<- ct:
3- ..::
0-
:z- o o
UJ
z-
---
t-
~- ~--
U
UJ
3
~
I C. 0..
CASE I
CASE TI
HG. 6.3. Effect of wind upon tower.
(ii) Wind blowing aeross diagonal (see Fig. 6.3, Case 11). As in the
previous ease, loeal bending may be ignored on the sides of the tower,
and the wind load resisted by the vertical lattiee girders on all four
sides.
If side of tower = 1·0, then diagonal = 1·41. Assurne this load
resolved into two loadings within the sides of the tower.
Load E is replaeed by loads Fand G.
Load H is replaeed by loads J and K.
Eaeh of the loads, F, G, J, and K will also be equal to unity (I ·0) whieh
means that the vertieal girders are loaded equally for both Cases land
11. When overturning takes plaee aeross the diagonal, then leg A resists
tension and leg C resists eompression, whieh means that overturning
is resisted by two legs as against four legs in the previous ease. Sum-
marising this from a design point ofview, it is convenient to design the
vertieal braeing to suit Case I loading, but the tower legs should be
designed to suit Case 11 loading.
(b) Evaluation ofwind loading.

Case I (Wind normal to one face of tower):
Wind pressure = 950 N/m 2 •
Refer to Fig. 6.4.
Wind load on tank = 6 X 2·75 X 950/103 = 15·7 kN (2·75 repre-
sents depth of tank plus the depth of the tank-supporting beams).
r- ---1
[ 7·85
11)-
I I
I
'"
I
8'5 _ !E;9
2·4
10·9
2:1.1r--7\
ALL LOADS IN
E 14.~ kN
~-
~ '2'2·8 1.2
~ ~
WIND LOAD ON EQUIVALENT WIND LOAD ~..;S:-2-t-
STRUC.TURE ON STRUCTIJRE WIND LOAD 010,1
(a.) (Is ) ONE. 51DE oF
STRUCTURE
ce)
FIG. 6.4. Wind forces acting on tower.
Divide equally between windward and leeward faces = 15'7/2=

7·85 kN.
Equivalent load at top of braced portion of the structure = 7·85 X
19'25/17'75 = 8·5 kN.
In considering the wind load on the braced portion of the tower it is
necessary to estimate the amount of surface exposed to the wind. An
average figure of 20 per cent will be chosen to enable the design to
proceed.
Wind load on one face oftower = 17·75 x 4·25 x 950/103 x 0·2 =
14·3 kN.
Each tower face is exposed to wind pressure and this loading repeats
on both the windward and leeward faces (see Fig. 6.4(a». The equivalent
loading on the braced portion of the tower is shown in Fig. 6.4(b).
This loading is shared equally by two parallel faces, see Fig. 6.4(c).
Horizontal reaction at the base of tower
= 2(5·4 + 2·4 + 2·4 + 1,2) = 22·8 kN
Vertical reactions at each support
= [2(5'4) x 17·75 + 2(2'4) x 11·82 + 2(0'24) x 5'91]/4'25
= 65·2kN
The maximum load in the diagonals will occur in the bottom panel of
the frame.
Shear on this panel = 22·8 - 2(1'2)
= 20·4kN
Load in diagonal = 20·4 (diagonallength/panel width)
= 20·4(7'27/4'25)
= 35kN
Alternatively, a force diagram may be drawn for the frame.
Case II (Wind acting across diagonal):

Length of tank diagonal = y'(2 X 62 ) = 8·48 m.
Length of tower diagonal = y'(2 X 4'25 2) = 6·02 m.
Wind load on tank across diagonal on projected area
= 8·48 X 2·75 X 950/103
= 22·2kN
Some of the load will be deftected off the tank sides and the load
acting across the projected diagonal area may be reduced by some factor,
which may be taken as 0·8.
Net wind load on tank = 0·8 X 22·2
= 17·76kN
If this load is taken as acting at its centroid of application, then:
Overturning moment about base = 17·76 X 19·25 = 342 kN m.
Wind load across diagonal ofbraced portion
= 6·02 X 17·75 X 950/103
= 102kN
Exposure as before = 20 per cent.
Net wind on two faces = 2 X 102 X 0·2 = 40·8 kN.
Overturning moment about base = 40'8 X 8·875 = 362 kN m.
Total overturning about base = 342 + 362 = 704 kN m.
This overturning moment is resisted by the two diagonally opposed
tower legs acting as a force couple.
Load induced into tower legs = 704/6·02 = 117 kN.
This load may be tensile or compressive in nature, depending upon the
direction of the wind.
49 Design of Tank-supporting Beams
The details of design for these members are given in Fig. 6.5. It has
al ready been stated that the tank is composed of prefabricated steel
units. Alternatively, cast iron units may be used, but in this ca se it
F3EAt-II MI<. \ 8EAM MK. 2

1.15 AT \-2m CS~} 3 LSELF]
1 ! 1 I wT
44-\ 44-\
I
w PER
~HHHHHHHiH
UNIi LENGTH [88-21
~
1"8S-21
!
[L'10~87~. 4-'2.5 [Ll 1.0-8175 [LI)
IYI m m zn
kN
WEI(;HT OF TANK 80
WEI(,f.\T OF WATER 792-
MOMENT AT C.L. BEf>.M
:: 222 x 2·12.-44-1)(3- gg·2xl·8
1\72 - 85'Z x 0'"
== I7.S·71<:t-Jm
LOAO PER U_e>_ :. 872+ 5
=- 174-4 kN Z = M = 1"Z5"-7x 103 _ l'
7<02 G..
SELF WT '" '2 kN T 1<05-
TOTAL =- 174-4-rZ=- 17c;,-4kN
w =
17<0-4';- c;, = '2~-4o kN/m
IJSE 381>< 152 x 52 k9 Ue.
M (AT C_L):; wL7. _ wLI1.
8 '2.
.,; '2~-40 ('2-26 - 0 -38)
::: 55·3 kNm
Z. = .!::':l = 55-3>< leil = 335 CW13
f 1<05
U5E 25\ x 140x 3\ kq U ß
BENDIN(, MOMENT DIAGRAM
FIG.6.5. Calculation sheet.

Tank supporting beams.
will be necessary to limit the deflection in the tank-supporting beams
to prevent damage to the tank and the joints between tank units. The
tank fabricator will usually specify the allowable deflection and this
figure should be adopted. In other words, the criterion for design is
deflection and not flexural considerations.
50 Design of Tower Members
(a) Corner leg or upright
Loading
Dead loads: kN
Tank (empty) = 80 (data ex makers)
Self weight of tower = 70
150kN
Load per leg = 150/4 = 37·5 kN
Live loads:
Water (say 2·2 m deep) = 6 x 6 x 2·2 x 10
= 792kN
Load per leg = 792/4 = 198 kN
Induced wind = ± 117 kN per leg
(already calculated).
Maximum load (excluding wind, tankfull) = 37·5 + 198 = 235·5 kN.
Try 152·4 x 152·4 x 12·6 angle.
A = 37·03 cm2 rv = 3·01 cm
L = 2·955 m (distance between horizontals)
I = L (based on assumption that each end is held in position but
not restrained in direction)
Ilr= 2955/30'1 = 98 pe = 81 N/mm2
/c = WIA = 235·5 x 103 /3'703 x 103 = 63'7 N/mm 2
Proposed section is satisfactory.
Maximum load (including wind, tank full)
= 37·5 + 198 + 117 = 352·5 kN.
For this condition the safe stress (Pe) may be increased by 25 per cent
(BS 449: 1969, clause 13).
pe = 81 + 25 per cent = 101·25 N/mm2
/c = W/A = 352·5 x 103 /3'703 x 103 = 95·5 N/mm2
Proposed section is satisfactory.
Maximum load (including wind, tank empty)
= 37·5 - 117 = -79,5 kN (tension).
The proposed section is obviously able to resist this relatively small
tensile force and need not be checked for this loading condition.
(b) Diagonal members. Maximum load 35 kN (tension). If the load
condition is considered the size of angle required would be smaller
than is desirable because a small angle would be extremely flexible and
contribute to lack of rigidity in the structure as a whole. A limit of 250
should be placed upon the Ilr ratio for this purpose.
Length of diagonal betw6en intersections of diagonals to tower
legs = 3630 mrn, [ = 0·85 L, and if [/r = 250 then
3630 X 0·85
r= 250
= 12·3 mm = 1·23 cm
A 63·5 X 63·5 X 7·90 angle having an r v = 1·24 is satisfactory.
(c) Subsidiary horizontal members. These bars, although carrying no

theoretical loading, are required to provide restraint to the corner
legs and to the ends of the diagonal members designed in (b) above.
In the plane of the tower si des, each angle is supported at mid-span
by the gusset plate connecting the diagonals, but in the plane per-
pendicular to the tower face the angle is supported only at its connec-
tions with the tower legs. For this reason it is necessary to have the
larger leg of the angle outstanding from the face of the tower.
Try 76·2 X 63·5 X 7·90 angle with the 76·2 leg outstanding.
[= 0·85 X 2125 = 1810 mrn
I/ru = 1810/13·2 = 137
Ix = 0·85 X 4250 = 3600 mm
Ilrx = 3600/23·5 = 153
Both values are less than 180 and the proposed size of angle should be
adequate to atford the required support to the corner legs and to the
diagonal bracing.
(d) Main horizontal members. Load 20·4 kN (compression). In

addition to the axial load this member is required to provide restraint
to the corner legs. For this reason the Ilr ratio will be limited to 180
rather than the higher figure of 250 which is appropriate to members
carrying wind loads.
Try 2-63·5 X 63·5 X 7·90 angles 'starred' and battened together
as shown in Fig. 6.6. This type of double angle section is to be pre-
ferred to one in which the angles are placed back to back on each side
of a gusset, resulting in a small space between the angles which is
inaccessible for purposes of scraping and painting.
L = 4·25 m I = 0·85 x 4250 = 3600 mm
Ilr = 3600/24'1 = 150 pe = 40N/mm 2
Je = WIA = 20·4 X 103 /1.896 X 103 = 11 N/mm 2
The proposed seetion is understressed but it should be remembered
that it spans 4·25 m and is required to provide restraint to the corner
legs of the tower.
(e) Horizontal diagonal cross-bracing. This system of bracing is

important to the overall rigidity of the structure because it offsets
any tendency for twist or torque to develop in the structure due to
peculiar conditions of wind loading. It is extremely difficult to assess
the design accurately.
Try 88·9 X 76·2 X 7·82 angle (88,9 leg vertical).
Ilr = 3760/15'8 = 238
(l = 1·25 L where L is half the tower diagonal)
This is less than 250 which is acceptable for subsidiary members but it
may be argued that a deeper section is required to offset deflection
due to self weight. If depth is limited to the diagonal length/50 then
a section with a depth of 6020/50 = 121 mm is required and a 101·6 X
76'2 X 7·90 angle should be adequate bearing in mind the intersection
of the angles at mid-span will help in limiting the deflection.
51 Design oe Foundation to Resist Uplift

(a) Concrete Block
Case J. Compressive load in tower.
Corner leg = 352·5 kN.
If safe ground pressure = 210 kN/m 2 , area required = 352'5/210 =
1·68 m 2 • Weight of concrete = 24 kN/m3 •
Allow a depth of block = 0·6 m.
Weight of concrete block = 1·68 X 0·6 X 24 = 24·2 kN.
This is obviously insufficient to anchor the corner leg for Case 11
loading in which an uplift of 79·5 kN is present.
Case II. Tensile load in tower corner leg = 79·5 kN.

Allow a factor of safety against uplift of 1·5.
Weight of concrete block required = 79·5 + 50 per cent
= 119·25 kN
Let d = length of side of a square-block and d/2 = depth.
Then, d 3 /2 X 24 = 119·25
d= 2'2m
Say, 2·2 m square X 1·1 m deep.
(h) Anchorage. Uplift due to wind on each corner leg = 79·5 kN.
Allowable stress in anchor bolts on area across root of thread
= 130N/mm2 •
As the stress in the anchor bolts is induced by wind the safe stress
may be increased by 25 per cent to 152 N/mm2 •
Allow three bolts per base.
79·5 X 103
Area required per bolt = 3 X 152
= 175 mm 2
A 20 mm dia bolt has a cross-sectional area at the root of thread equal
to 215 mm 2 and is satisfactory.
To ensure that the anchor bolts are adequately secured in the concrete
block the system of anchor channels shown in Fig. 6.6 is included.
52 Check on Estimated Data

In a structure of this nature the exposed area of steelwork presented to
the wind and the self weight of the structure are important factors in
the stability of the structure. Both factors were estimated in order to
allow the design to proceed and the data used should be verified
before the proposed design is finally accepted. A check on the pro-
visional data used in this example shows that the original estimates
were substantially correct.
53 Design of Connections
Typical details of suitable connections are shown in Figs. 6.6 and 6.7.
Black bolts are used throughout and because of the size of the tower
it is unlikely that it will be erected before delivery and transported in
several large pieces. The connections such as gusset plates, base plate,
and cap material for the corner legs could be shop-bolted to the
appropriate members for despatch to the site. Gusset plates are used at
all bracing connection points to give some degree of rigidity even
though adequate width of material may be present to accommodate
the necessary connection bolts, as in the case of diagonal bracing
connected to the corner legs.
DIAGONAL
MEMBERS
CORNER LEG I G3·5 x "3·5x 7· 90 L
152·4 x 1~2-4 x 12.1<. I
ANGLE I
I 10 ,HICK
I
I
A I A
t l f
2125 BASEPLAiE
~----zr-r--=-:..c=-=----t---=------ 12 iH I C K
m'~~"~''''":'W71ik "01
\ 500 .1
PL I\N ON M5 E
C:ONCRE T E
FOUI-JDATloN
'2200 SQUARE
X 1100 D!:.E P
3 ANCHoP, BOL iS
20 DIA)<. 750 LONG
FIG. 6.6. Part elevation of tower side.
54 Provision of Access to the Tank

It is unlikely that frequent access to the tank will be required and a
system of ladders served by intermediate landings is shown in outline
on Fig. 6.1. Each ladder should be 250 mm between stringers with
rungs at 250 mm centres and provided with safety loops on the upper
IOI·"~ 76·2 X 7-90L
o DIAGONAL TII:
0
i
I
0
,....
0
- C.l. -rOWER\_
0 1
,l)
'251)( 14<0." 31 k~ U 13
r-t--- -
1
0
0 381)C 152 X 52 k5 U e, I SECTION ß· B
(SEE FIt:> 6'6)
0
0
0
~ I
I I
.. .. 251)C 14' X 31 k9 UB I
- - -
I 1 ANG LE TIE
0
0
' ~COF..NER I 76'Z x 103· :h 7·')0
~
LEG UNDER 1/
I
I -
I:
__ ,--i--- -
I 4250
'a75 ·1 "'000
PART PLAN ON TANK 5UPPORTIN& BEAMS
WATER
TANK.
251" 146 X 31 "9 UB .~.
WEB BRACKET
7<0.·7. " 7(" 1. X 7·85 L
PLAN ON CAP OF
CORNER 1
CORNER LEG
LEG --I
PART 5 1DE ELEVATION AT HE,o,D oF TowE~
HG. 6.7. Details of tank beams and head of tower.
two flights. A slope is indicated to the ladders to give ease in c1imbing.

If frequent access is required to the tank then consideration should be
given to the provision of a staircase.
55 Maintenance of the Structure

In view of the fact that the structure is exposed to the weather it is
important that some attention be given to maintenance.
(i) A minimum thickness of material should be specified for external
structures, say 8 mm. If thinner material is used, say 6 mm, then
careful maintenance is necessary and in areas of high corrosive risk a
thickness of 10 mm may be considered advisable.
(ii) The minimum acceptable finish should consist of grit blasting
all material and coating with good quality primer paint beforedelivery,
followed by a further two coats of paint at site after erection. Relatively
more expensive would be galvanizing the whole structure or spraying
with metal. Such a treatment would be likely to last 15 years and over
a long period proves less expensive than a 3 or 5 year programme of
painting.
(iii) As an alternative to mild steel a copper bearing or weathering
steel might be a feasible proposition. These steels require no protection
and over the first two years of life achieve a protective patina of rust
which prevents further deterioration. The colour of the patina depends
upon environmental conditions. These grades of steel, although more
expensive initially, are likely to prove economical over aperiod of
time and are becoming increasingly popular for external structures.
(iv) In Fig. 6.6 it will be noted that the tower corner leg is shown set
up above ground level to prevent corrosion taking place which would
occur if the steel base was positioned below ground level.
(v) No details should be adopted which behave as water-traps. All
edges of plates or angles should be carefully detailed so that they remain
exposed for maintenance purposes.
7. Some Aspects of Economical
Design
56 Costs of Fabrication
In general, it may be said that the contractor responsible for steelwork
fabrication obtains his raw material from three sources :
(a) Rolling mills (i.e. direct from the steel makers).
(b) Stockholder. These commercial undertakings hold stocks of com-

monly used sections.
(c) Own stock. The fabricator will probably hold a stock sufficient
for his immediate needs.
Different f~ctors enter into the cost of obtaining material from the
above sources.
(a) Transport. Distance of the fabrication workshop from the source
of supply will decide the cost of transport and this cost must be
reflected in the final cost of the product.
(b) Availability 0/ stock lengths. If relatively small weights of each

section are required (under, say, 5000 kg) then the material is likely to
be obtained wholly from stock, either from the material held by the
fabricator or by the stockholder. The fabricator usually avoids carrying
large stocks since he will be reluctant to layout capital on stock which
fluctuates in value or deteriorates. If stock lengths are used the required
cutting lengths will result in some degree of waste which should pre-
ferably be kept to a minimum (say 5 per cent to 7 per cent). Material
ordered from the rolling mills can usually be supplied in lengths with
a tolerance of + or - 25 mm. From the above it will be seen that
material 'ex stock' is likely to be more expensive than material 'ex mills'.
Quotations for steel frames are sometimes specified as being based on
material 'ex stock' or 'ex mills'. Average cost of plain sections may be
taken as follows:
VB and VC sections 100

Angle sections 103
Flats or plates 100
Because costs vary from time to time the material costs are based on
a unit cost of 100 units. Smaller sections require more labour to
produce than larger sections and this resuIts in price variations. The
yield point of the smaller and thinner seetions is likely to be higher
than for the heavier sections because of the probability of a certain
amount of work hardening taking place during the rolling process.
The following are typical examples of fabrication costs:
Floor beams of simple construction 285

Floor beams of compound construction 315
Simple columns 275
Bolted roof trusses 430
Angle rails and purlins 320
These costs are based on a unit material cost of 100 units. Because
material costs vary, so will fabrication costs. It is possible that fabrica-
tion costs will not vary in a constant ratio with materials because of the
conditions prevailing in the trade at any particular time: Poor trading
conditions result in keen competition amongst fabricators for the
available work and it may be necessary for them to cut their profit
margins to remain competitive even though material costs may remain
the same or even increase.
57 Effect oe Layout upon Design

The true skill of the designer is more evident in the arrangement of the
individual members of a structure rather than in mathematical skill in
analysing the forces on those members.
The arrangement of members is frequently dicta ted by circumstances
(e.g. clear floor space required, plant requirements, details of cladding,
etc.). However the designer must know something of economical
arrangements and be prepared to advise his client or to make the best
arrangement if a free choice is available. Three examples are given to
illustrate this point.
SOME ASPECTS OF ECONOMICAL DESIGN 139
EXAMPLE 1. Tbe economic spacing of roof trusses and columns
A steel-framed ridge-type shed building is required having an eaves
height of 6 m and a span of 12 m. Each main frame consists of two
columns and a roof truss and it is required to find the economic spacing
for the main frames using (a) hot-rolled angle sections, and (b) cold-
rolled z-sections for the roof purlins and side rails.
The weights of steel required for case (a) is given in Table 7.1 and for
case (b) in Table 7.2. For case (a) using conventional sections the
TABLE 7.1. Weight of steel required using hot-rolled angle

sections for roof purlins and side rails.
Centres of columns
Item (m)
3 4 5 6
Roof truss 300 330 355 370

Columns 410 490 490 530
Roof purlins 214 344 776 1315
Side raUs 143 230 520 880
Total (kg) 1067 1394 2141 3095
Weightfm2 of floor
area (kg) 2·96 2·9 3·57 4·3
TADLE 7.2. Weight of steel required using cold-rolled zed

purlins and angle rails.
Centres of columns
Item (m)
3 4 5 6
Rooftruss 300 330 355 370

Columns 410 490 490 530
Roof purlins 121 163 300 432
Side raUs 81 109 200 290
Total (kg) 912 1092 1345 1622
Weightfm2 of floor
area (kg) 2·53 2·28 2·24 2·26
economical spacing is 4 m but increases to 5 m or 6 m in case (b)
when cold-rolled zed sections are used. The figures given represent
a weight comparison rather than a cost comparison. Light sections
cost more per unit of weight than heavy sections and although the cost
-'
«
---_ CASE (bl _
oc
lLl
----------
1-'2
«
L
u.
o
I'
l-
X
<!)
UI
~
Ö -4 '5 6
5PACING Of MAIN FRAMES
[ml
FIG. 7.1. Economic spacing of roof truss and column frarnes.
difference between cases (a) and (b) will be significant it is unlikely that
it will be as great as the difference in weight suggests. Figure 7.1 shows
the information contained in Tables 7.1 and 7.2 in graphical form.
EXAMPLE 2. The economic spacing of columns supporting ßoor beams

For a given load, it is true to say that beams of short span are more
economical than beams of long span even when the cost of supplying
extra columns is taken into account. The weightJmetre of beam
required to carry a uniformly distributed load of 30 kNJm is 37 kg
when the span is 4 m but rises to 74 kg when the span is 8 m.
Consider the following example. Design a beam to support a load
of 30 kNJmetre (including dead weight of beam) assuming that (a)
the beam spans 8 m and is supported by columns 3 m high at each end,
and (b) the span is split into two equal portions supported by three
columns 3 m high.
Case (a)
Beam required = 457 x 190·5 x 74 kg
Column required = 152 x 152 x 23 kg
Weight of steel required = 74 x 8 + 23 x 3 x 2
= 730kg
Case (b)
Beam required = 304 x 124 x 37 kg
Column required = 152 x 152 x 23 kg
Weight of steel required = 37 x 8 + 23 X 3 X 3
= 503 kg
These figures show that long span beams are uneconomical.
EXAMPLE 3. Economic depth of plate girders

It is proposed to examine the weight of material required to provide a
welded plate girder of uniform section for varying depths of girder. For
the purposes of this example the following data are assumed:
Span of girder 12 m.
Uniformly distributed load inc1usive of self weight 1000 kN.
Flange width 350 mm.
Web thickness 10 mm.
Web stiffeners: spaced at a distance centre-to-centre not exceeding
the depth of the girder.
Safe compressive stress, Pbc = 140 N/mm 2 •
Bending moment = W X L/8 = 1000 X 12/8 = 1500 kN m.
Inertia required = M X Y/PbC' and when Y = d/2 = 300,
then I = 15()() X 106 X 300
140
= 3·21 X 10 9 mmi
The thickness offlange plates may be obtained by evaluating the follow-
ing expression for d (see Fig. 7.2).
BD3 (B - t)d 3
12 - 12 = Inertia required
When B = 350, D = 600, t = 10 then d = 478 and T = 60.
Table 7.3 shows details of material weights required for varying
depths of girder and the same information is presented graphically in
Fig. 7.2.
TADLE 7.3. Weight of plate girder for varying depths of girder.
Depth of girder
Item (mm)
600 900 1200 1500
Flanges 3950 2305 1319 989

Web and stiffeners 734 1131 1450 1754
Total (kg) 4684 3436 2769 2743
5000
4000
.--.
a>
~
-l
~3000
er
UJ
t-
<t
L '2000
u.
o
I-
~ 1000
w
~
600 900 1'200 1500

DEPTH 01= GI~DER [l'YIm]
FIG. 7.2. Relation of plate girder depth to weight of plate girder.
Points ofinterest which may be seen from the graph are as folIows:
(a) The weight of material required drops as the depth of section and
moment of inertia increase.
(h) Using the formula given in section 27 the economic depth for this
condition is equal to the span/lI which appears a little too shallow
if absolute economy is required. Tbe point at which the weight of
Hange material equals the weight of the web material also indicates a
point at which economy is achieved.
(c) Plate girders frequently require to be accommodated within a

certain construction depth and it may not be possible to use a truly
economic depth of girder if this results in construction costs elsewhere
in the building being increased. For example, if only one plate girder is
required, it may be more economical to reduce the girder depth than to
increase the height of the whole building, even though the weight of
material for the plate girder is increased.
(d) At the economic depth of the girder the deHection is approximately

span/500.
(e) In this example, no adjustment has been made to the value of

compressive bending stress because of possible adverse ratios of DIT
or l/r which may occur in certain cases of section geometry. If lateral
restraints are provided at third points of span, then the proposed design
is satisfactory. Should it be necessary to decrease the value of stress
in the compression Hange then the weight of material required will be
greater.
Index
Architect, 1, 16, 81 Curtailment diagram for flange
Asbestos-cement corrugated sheet- plates,52
ing, 18, 81 Customer,1
Assessment of loads, 8, 120
Dead load, 8
Beams, 7, 15-16, 26-33, 122, 129 Deflection, 13
bending, 10, 15, 16,26-33, 129 beams,33
brackets,60-5 plate girder, 47
choice of sections, 6-7, 129 Depth, plate girder, 41, 47, 141-3
connections, 65-6, 74-7 Design procedure, 1,6, 138
end restraint, 26, 60 preliminary considerations to, 1
ftange restraint, 26, 29 procedure in, 6
web, buckling, 15, 62-4 Drainage of roofs, 120-1
crushing, 15,62 Drawings,1
shearing, 15,64 Dynamic loading, 8
Bending, relationship to shear, 9
Bolts, loading values, 72-4 Economical design, 1, 16, 137-43
types, 70-1 Effective length, 17
Brittle fracture, 12 Elastic limit, 10
Buckling length, 17 range, 10
theory of design, 10, 11
Camber, in plate girder, 48 Engineer as designer, 1
Client,1 Erection procedure, 67
Collapse load, 12
Columns, 17-18, 33-41, 101-7, 140-1
base plates, 55-60, 105-7, 133-4 Fabricated units, size of, 5
buckling, 17-18 Fabrication costs, 6, 137
connections, 60,76-80 Factor of safety, 12
eccentricity of loading, 39 Fatigue,12
end fixity, 17,40-1 Flange, plate curtailment, 52
formulae, 17, 18,39 stresses, in beams, 30, 33
foundations, 66-70, 108-10, in plate girders, 43, 45-6
132-3 Floor, loads, 30
radius of gyration, 18, 40 Foundations, 66-70, 108-10, 132-3
slenderness ratio, 17, 118
splice,66 Gable and side framing, 112-19
Corrugated sheeting, 18,81 Grouting to column bases, 67-8
Cost of fabrication, 6, 137 Gyration, radius of, 18,40
146 INDEX
Holes, allowance for, in tie members, Roof, trusses (contd.)
22-3 spacing of, 82-4, 139-40
types of, 18-19
Inertia, moment of, (second moment weight of, 90
of area), 13-14 wind loading on, 91, 93-4
Ladders, access, 134-5 Safety factor, 12
Lattice girders, 21-2 Section, modulus, 14
Load, assessment, 8 properties of, 13
dead,8 Shear, in beams, 15, 64
dynamic,8 plate girders, 49
factor, 12 relationship to bending, 9
live, 8 Sheeting rails, 110-12
relationship to stress and strain, Shop premises, 24-80
9 SI units, 3
Side and gable framing, 112-19
Maintenance of exposed structure, Size of fabricated units, 5
135-6 Slendemess ratio, 17, 118
Minimum size and thickness of Specifications, 4-7, 84
material, 4, 136 Steel, grades of, 7
Moment of inertia. (second moment Strain, 10-12
of area), 13, 14 Stress, distribution in beams, 15
Moment of resistance, 10 uItimate, 10, 12
working, 11, 12
Neutral axis, 9 yieId, 10
plane, 9 Struts, 17
Plastic theory of design, 11 Tabulated information, 8
Plate girder, 41-55 Tension members, 22-3, 97-9
assumptions in design, 41 Transport oflarge fabricated pieces, 5
camber,48 Triangulated frameworks, 18
curtailment of flange plates, 52
deflection,47 Ultimate stress, 10, 12
depth, 41, 47, 141-3 Units, design, 3
flanges, 41, 43
inertia,48 Warren girder, 21-2
web, 44, 49 Water tower, 122-36
weight, 41, 52 Weight, of plate girders, 41, 52
welded connections, 53, 54 of roof trusses, 90
Pressure, wind, 33, 84-7, 125-8 Welded connections, beams, 65-6,
PurIins, 87-90 74-6
columns, 56-7, 77-80
Radius of gyration, 18, 40 plate girder, 53-5
Roof, coverings, 18,81 Wind, bracing, 113-17, 124-8
drainage, 120-1 pressure, 33, 84-7, 125-8
purIins, 87-90 Working load, 12
trusses, 18-19, 81-4, 90-101 stress, 11, 12
connections in, 99-101 Workshop design, 81-121
design of members in, 93-9
force diagrarns of, 91-4 Yield point, 10-12
framing of, 18-19,82-4 stress, 10

Steel Frame Design Examples

Загружено:

Сведения о документе

Авторское право

Доступные форматы

Поделиться этим документом

Поделиться или встроить документ

Параметры публикации

Этот документ был вам полезен?

Это неприемлемый материал?

Авторское право:

Доступные форматы

Steel Frame Design Examples

Загружено:

Авторское право:

Доступные форматы

STEEL FRAME DESIGN EXAMPLES

By the same author

© lan Robb 1961, 1965, 1972

All rights reserved. No part of this publication may be

First published 1961

1 Preliminary Considerations of Design

2 Elementary Design Definitions 9

3 Elements of Basic Design 15

4 Design of Multi-Storey Shop Premises 24

6 Design of Braced Tower Supporting Water Tank 122

and it is eonvenient to make this transition to units of newton before

2 The Role of the Specification in Design

(a) Minimum size and thickness 0/ material. Where special conditions of

(b) Size 0/ jinished pieces prior to delivery /rom workshop to site.

4 Choice of Sections when Designing

5 Relative Costs of Fabrication

6 Relation of Strength to Weight in Selecting SectioDS

311 X 167 X 54 kg z = 752cm3

8 Use oe Tabulated Information

10 Relationship between Shear and Bending

11 Relationship between Bending Moment and Moment of Resistance

12 Relationship between Ultimate Stress (or Rupture Stress), Yield

FIG. 2.1. Stress/strain diagram.

This text is concerned with designs based on the elastic theory.

14 Fatigue Failure of Materials

17 Geometrical Properties of Seetions

l.N E R1IA A80UT)( Ax 15

Section modulus. The section modulus about the x axis = Zn and

(d) The strength of beams must also be investigated when it is necessary

FIG. 3.1. Basic beam and girder sections.

19 Struts (including Columns or Stanchions)

HG. 3.2. BucklingJength of columns.

(a) Economy of design. A ridge height of between one-quarter and one-

(b) Architectural requirements. The appearance of the building may

DOTTEO L.INE.S REPRESENTS CENTR.E

FIG. 3.3. Roof truss frameworks for various spans.

HG. 3.4. Lattice girder frameworks.

(ü) Lattice girders. Examples of lattice,girder frameworks are shown in

HG. 3.5. Lattice girder frameworks.

web construction, it is assumed that the flanges resist the bending

(a) The N-type framing shown in Fig. 3.4(a). Economy is achieved

21 Design of Tension Members

(d) Floor and flat roof construction is of pre-stressed concrete plank

FrG. 4.1. Multi-storey shop premises.

(f) Fire protection of the steelwork is mandatory. The usual methods

24 Design oe Floor Beams

4·2rn 4·2", 4,2"" 4· '2 "" .[ SI'AI-l OF Floofl,.

FIG. 4.2. Multi-storey shop premises.

4·2..., -4.'2"., 4·2..., 4·2""

FIo. 4.2 (cont.)

Z=M '" 206,,10

FIG. 4.3. Calculation sheet.

14 ·2,.., SPAN) 14,.,., SPAN)

FIG. 4.3 (cont.)

(4-2 ... 5P ... N) (4 ... SPAN)

FIO. 4.3 (cont.)

Allowable value of Pbc = 78 N/mm 2 approx.

This is a very low figure and illustrates the importance of adequate

lS Effect of Wind on Building

26 Design of Columns or StanchioDS

FlG. 4.4. Calculation sheet.

"H IS COLUMI,J IS RES,RAINEO ON /S"I'TO Z"o

RooF 122 22 e/r~ '" 3150/110.5 '" 29

11 SPLlCf t =O'7L G, LOADING ABOUT Y A)(15