Steel Design 3 Connections

Behaviour of connections in steel structures and design of mechanical
fasteners and welds according to Eurocode 3
Steel Design 3
Connections J.W.B. Stark

Behaviour of connections in steel structures and design of mechanical
fasteners and welds according to Eurocode 3
Steel Design 3
Connections J.W.B. Stark
preliminary work | connections | 1

Colophon
text prof.ir. J.W.B. Stark
editing ir. C.H. van Eldik / Bouwen met Staal
graphic design Karel Ley / Fig.84-Reclamestudio
published by Bouwen met Staal

ISBN (print) 978-90-75146-05-9
ISBN (PDF) 978-3-433-61161-6
The publication of this textbook has been made possible by:
Bauforumstahl www.bauforumstahl.de
Bouwen met Staal www.bouwenmetstaal.nl
Infosteel www.infosteel.be
Stahlbau Zentrum Schweiz www.szs.ch
Tata Steel www.tatasteelconstruction.com
World Steel Association www.constructsteel.org
© Bouwen met Staal 2021
All rights reserved. No part of this publication may be reproduced, stored in an automated data-
base and/or made public – in any form or by any means, electronic, mechanical, photocopying,
recording or in any other way – without prior written permission from the publisher. The utmost
care was taken in the preparation of this publication. Nevertheless, any errors and imperfections
can not be ruled out. The publisher excludes – also for the benefit of all those who have partici-
pated in this publication – any liability for direct and indirect damage, caused by or in connection
with the application of this publication
2 | connections | preliminary work

This textbook was originally published in 2012 by Bouwen met Staal in Dutch as Verbinden by the same
author. The English translation has been prepared by prof.ir. H.H. Snijder and ir. F.M.W. van den Hove
(both Eindhoven University of Technology) and checked by dr. G. Couchman (The Steel Construction
Institute).
The text is based on the (English) EN version of the Eurocodes using default and/or recommended
values. Where a country can make a national choice – or when non-contradictory complementary in-
formation may be used – this is indicated by the following symbol: . Separate annexes contain the
national choices for Belgium, Luxembourg, The Netherlands and Switzerland. These annexes – as
well as any errata, corrections and additions to this textbook – can be downloaded free of charge
from the websites of the (national) organisations.
Illustrations
All unnamed photographs and all drawings come from the archive of Bouwen met Staal.
L = left, R = right.
Aeroview/Dick Sellenraad 1.39 MCB verbindingstechniek 2.14

Agfa-Gevaert 4.24 Nmiguel 3.1R
BCB Staalconstructie 4.1 Eppo W. Notenboom 1.1R
P. Bos 3.2 Pixabay (2427999) 3.21
Joe Cough 1.17 Tom de Rooij Vakfotografie 5.38R
C.H. van Eldik p. 1-1, 3.16 Iemke Ruige p. 3-1, 3.23
Dutch Engineering 1.26 Sarens Nederland 3.1L
A.F. Hamerlinck 5.38L Simply Creative Photography p. 5-1
Köco/KSM Holland 4.8 J.W.B. Stark 4.25
Witold Krasowski 2.9 The Steel Erection Company 1.38
Timothy Large 2.48 TU Delft/Faculteit Civiele Techniek en Geoweten
Cristian Lazzari p. 4-1 schappen/Stevin Laboratorium 2.32, 2.38
Lincoln Electric Europe 4.4, 4.6, Vercruysse & Dujardin 1.27
4.13, 4.15, 4.17 vm cover

Content
1 Connections in steel structures 1-2
1.1 The function of connections 1-2
1.2 Parts of connections 1-4
1.3 Common connections found in buildings 1-7
1.3.1 Column splice 1-8
1.3.2 Column base connection 1-8
1.3.3 Beam-to-column connection 1-11
1.3.4 Beam-to-beam connection 1-14
1.3.5 Beam splice 1-15
1.3.6 Connection of bracings 1-17
1.4 Design and detailling 1-17
1.4.1 Requirements regarding connections 1-18
1.4.2 Cost-conscious design of connections 1-19
1.5 Standards for connections 1-26
1.5.1 EN 1993-1-8 1-26
1.5.2 EN 1090-2 1-28
1.5.3 Product standards 1-28
1.6 The role of the computer 1-29
1.7 Literature 1-32
2 Bolts in clearance holes 2-2

2.1 Categories of bolted connections 2-3
2.1.1 Bolted connections loaded in shear 2-3
2.1.2 Bolted connections loaded in tension 2-5
2.2 Bolts 2-6
2.2.1 Product standards 2-7
2.2.2 Screw-thread 2-8
2.2.3 Bolt/nut combinations 2-8
2.2.4 Washers 2-9
2.2.5 Mechanical characteristics 2-10
2.2.6 Hole clearance 2-12
2.2.7 Practical bolt choice 2-12
2.3 Positioning of bolts 2-14

2.4 Resistance of elements with holes 2-15
2.4.1 Net cross-section 2-15
2.4.2 Tension resistance of a plate with holes 2-18
2.4.3 Tension resistance of an angle with holes 2-19
2.4.4 Block tearing 2-20
2.5 Resistance of bolts 2-23
2.5.1 Bolts loaded in shear 2-23
2.5.2 Bolts loaded in tension 2-28
2.5.3 Bolts loaded in shear and tension 2-29
2.5.4 Countersunk bolts 2-29
2.6 Literature 2-34
3 Slip-resistant connections, rivets and pins 3-2

3.1 Preloaded bolted connections 3-3
3.1.1 Tightening of preloaded bolts 3-3
3.1.2 Preloaded connection loaded in shear 3-6
3.1.3 Preloaded connection loaded in tension 3-12
3.1.4 Preloaded connection loaded in shear and tension 3-15
3.2 Connections with fit bolts and injection bolts 3-17
3.2.1 Fit bolts 3-17
3.2.2 Injection bolts 3-18
3.3 Rivets 3-24
3.4 Pins 3-26
3.5 Literature 3-28
4 Welds 4-2
4.1 Welding processes 4-2
4.1.1 Arc welding 4-2
4.1.2 Resistance welding 4-8
4.2 Weld shapes 4-8
4.2.1 Fillet welds 4-10
4.2.2 Full penetration butt welds 4-11
4.2.3 Partial penetration butt welds 4-14
4.2.4 Plug welds and flare groove welds 4-14

4.3 Welding procedures 4-15
4.4 Weld quality 4-16
4.4.1 Welding errors 4-17
4.4.2 Welding inspection 4-19
4.5 Weld design 4-21
4.5.1 Fillet welds 4-21
4.5.2 Butt welds 4-34
4.6 Literature 4-35
5 Design and detailing of connections 5-2

5.1 Design approach 5-2
5.2 Force distribution in connections 5-6
5.2.1 Influence of relative stiffnesses 5-6
5.2.2 Centre of rotation 5-9
5.2.3 Non-linear distribution of bolt forces 5-19
5.2.4 Force distribution in long connections 5-20
5.2.5 Combination of different types of connections 5-22
5.3 Transfer of axial compression or tension forces 5-23
5.3.1 Butt welds 5-23
5.3.2 Plates 5-24
5.3.3 Gusset plates 5-30
5.4 Transfer of shear forces 5-51
5.5 Introduction of tension forces 5-56
5.5.1 Welded connections 5-56
5.5.2 Bolted connections and prying forces 5-60
5.6 Transfer of compression forces 5-65
5.7 Transfer of shear forces 5-66
5.8 Literature 5-67

Structural basics
Steel Design1
CONTENT
1 Structural safety H.H. Snijder and H.M.G.M. Steenbergen, Structural basics.

2 Actions and deformations Analysis and design of steel structures for buildings
3 Modelling according to Eurocode 0, 1 and 3 (Steel Design 1),
4 Analysis published by Bouwen met Staal, Zoetermeer 2019,
5 Analysis methods ISBN 979-90-72830-98-2, format 23x25 cm, 272 p.
Also available as e-book at Wiley / Ernst & Sohn,
6 Assessment by code checking
at: www.ernst-und-sohn.de/en/steel-design.
7 Resistance of cross-sections
Fire
Steel Design2
CONTENT
1 Fire safety A.F. Hamerlinck, Fire safety and fire resistant design of
2 Calculation of the fire steel structures for buildings according to Eurocode 3
resistance (Steel Design 2), published by Bouwen met Staal,
3 Fire safety engineering Zoetermeer 2021, ISBN 979-90-75146-04-2,
format 23x25 cm, 164 p.
4 Design tables
Also available as e-book at Wiley / Ernst & Sohn,
at: www.ernst-und-sohn.de/en/steel-design.
Steel Design series

The textbooks in the Steel Design series are based on the (English) EN version of
the Eurocodes using default and/or recommended values. Where a country can
make a national choice – or when non-contradictory complementary information
may be used – this is indicated by a symbol (black square). Separate annexes contain
(for now) the national choices for Belgium, Luxembourg, The Netherlands and
Switzerland. These annexes can be downloaded free of charge, when available, from
the websites of the (national) organisations as well as any errata, corrections and
additions to these textbooks.

Steel Design
series
Education and high quality textbooks are crucial to developing
an interest in steel structures and their benefits for clients, archi-
tects and designers. However, despite the need to inspire the
industry’s next generation, many textbooks on steel structures
are commissioned on a low budget, resulting in material that
lacks imagination and tends to feature, at best, moderate illus
Steel Design 1
trations. These textbooks are usually intended for high school
Steel Design 2
and university level students, as well as designers who are not
yet specialised in steel and steel construction. Therefore, it is
vital that lecturers have access to up-to-date books that offer
clear and concise explanations, while inspiring readers about
the possibilities of steel through beautiful graphics and images.
Steel Design is a set of English textbooks translated from the
original Dutch that are based on the EN version of Eurocode
with differences in nationally defined parameters included in
an annex. These textbooks are intended for high-school and
university level students. The content is applicable to designers
who are not specialised in steel and steel construction.
Steel Design 3 See https://publicaties.bouwenmetstaal.nl/?p=all for more
detailed information on Structural basics, Fire and other text-
Steel Design 5 books of Bouwen met Staal.
World Steel Association

worldsteel has supported the development of study material
Steel Design 4
related to steel in construction since 2018. This allows future
architects and designers to take advantage of steel products
and their features that support designs that meet the circular
economy principles.
A separate opt-in programme has been developed called
'constructsteel.org' and is able to be joined by steel pro-
ducers and construction industry related organisations upon
application. This programme focusses on the construction
market sector exclusively to promote steel and steel products.
Please see www.worldsteel.org and www.constructsteel.org
for further details about the steel industry and specifically the
construction market.

1
Connections
Connections in steel structures
prof.ir. J.W.B. Stark

professor emeritus steel structures, Faculty of Civil Engineering and Geosciences, Delft University of Technology
connections in steel structures | connections 1 | 1

Connections in steel
structures
A steel structure cannot exist without bolted and welded connections between the elements of
the steel structure itself or with other materials or components. This chapter discusses the role
of connections in steel structures, identifies the various components, and provides examples
of the most common connections in buildings. Thereafter, important aspects of the design and
the detailing are discussed in general. An overview is given of the most important codes and
regulations for connections in steel. Finally, the role of the computer within the design of steel
structures will be discussed.
In general, the term 'connection' is used for the location at which two or more structural elements
(like plates and sections) meet and are connected (e.g. by bolting or welding). The focus is on
the transfer of the relevant internal forces and moments at the connection.
The term 'joint' is generally used for the zone where two or more structural members (like beams
and columns) are interconnected. The focus is on the structural properties (like resistance, stiffness,
and rotation capacity).
1.1 The function of connections

Steel structures usually comprise sections and plates that are processed and assembled into
structural parts in a fabrication shop, for example, a truss or a part of a bridge deck. The dimensions
of these elements are limited, because they usually have to be transported and, subsequently,
1.1 Use of large parts on the construction site.
2 | connections 1 | connections in steel structures

assembled (with a crane). The maximum dimensions depend on the capac
ity of the steel fabricator, the corrosion protection, the storage options, the
beam splice
method and route of transport to the construction site, and the capacity of
the crane. In general, it is desirable to aim for the largest possible elements. beam-to-column
This results in a reduced number of connections on site (fig. 1.1), where the connection
elements are assembled to form the steel structure. So, the steel structure column splice
contains connections that are manufactured in the fabrication shop as well
as on site. The conditions under which these two types of connections are
made strongly vary and this should be taken into account when designing the
column base
connections.
Generally the number of connections in a steel structure is rather large.
They are needed in the following situations.
– When structural elements meet from different directions. For instance,
angle web plate
this is the case for connections between beams and columns, secondary
and primary beams, and members in a truss.
– Connections on site are needed in order to limit the dimensions of
the components considering transport, storage, and assembly. As an column beam
example: the columns in a building may contain a splice every two or bolt
three storeys. When elements are galvanized thermally, the length of the
galvanizing bath must be considered. For example, the maximum length of
a galvanizing bath in a European galvanizing plant is approximately 15 m. bolt
– When the steel structure has to be connected to structural elements formed

from other materials, or connections between different types of steel
element. For example, the connection of a steel column to a concrete
foundation, the connection of a steel beam to a concrete body such as a column
bolt
lift core, or the connection of steel sheeting to steel trusses;
end-plate
– When the cross-section of a component changes, for example when the
weld
column size reduces at higher stories.
Figure 1.2 shows an example of a steel frame containing four storeys, in

which the first three situations apply.
weld
Connections are an important part of any steel structure. The structural

column base plate
properties of the connection affect the strength, stiffness, and stability of the
mortar joint
complete steel structure. The number and complexity of connections have a
decisive influence on the effort - and therefore, on the cost – that is needed
anchor with
to carry out the design calculations and produce construction drawings. Also,
screw thread
the manufacturing and assembly costs significantly depend on the chosen
connection type. concrete
foundation
1.2 Locations in a framework where

connections are necessary.

1.2 Parts of connections
Welds and bolts are most commonly used for the connections in a steel structure. For buildings,
welding is used in a workshop to create structural elements that are as large as possible. At the
construction site the connections between these welded components are often realized with
bolts. For bridges, both welded and bolted connections are used on site. For more detailed informa
tion on bolted and welded connections see Connections 2 (Bolts in clearance holes), Connections 3
(Slip-resistant connections, rivets, and pins), and Connections 4 (Welds). Other components are often
necessary to transfer the forces in a connection, such as angles, end-plates, or local stiffeners. These
components are also part of the connection.
Welds
In a welded joint, either fillet welds or butt welds may be used (fig. 1.3). For connections in a statically
loaded building structure fillet welds are often chosen, especially when cost is important. These
welds require no, or simple, preprocessing, whereas the weld edges for butt welds must be
preprocessed (beveled). Additionally, butt welds require a more accurate fit during assembly.
However, for connections between thick plates, the smaller weld content of butt welds can outweight
the higher costs of the preprocessing.
Although welding on site is technically feasible, the costs are relatively high for the following
reasons (see section 1.4.1):
– temporary facilities, such as scaffolding and welding tents, are needed to create a suitable
weld location that is safely accessible;
– a power source must be available at the location of the connection;
– bolts are often necessary in order to connect elements temporarily;
– inspections costs for the welded work;
– additional costs for repair of the conservation.
preprocessing: weld metal over

beveled edges the full thickness
1.3 Fillet weld and butt weld. double fillet weld butt weld

Bolts
Connections in building structures assembled on site are often
executed as bolted connections. Rationalizing the choice of bolt
bolts loaded
to one type (strength) is preferred, namely 8.8 bolts, in large in tension
clearance holes and hand-tight tightened, see Connections 2, bolts loaded

section 2.2. The bolts are loaded in either tension or shear (fig. 1.4), in shear
or in a combination of tension and shear (fig. 1.5), depending on

the loading, the type of connection, and the location of the bolts. A
small amount of movement will occur between the elements due to 1.4 Bolts loaded in tension
movement of the bolts in the clearance in the holes under loading. or shear.
When this movement of the bolts, together with the displacement of

the connection is not allowed – for instance, because the resulting
deformations are not acceptable – then a slip-resistant connection
with preloaded bolts, fitting bolts, or injection bolts can be chosen.
Such situations are covered in Connections 3, section 3.1 and 3.2. tension
The use of slip resistant connections should be avoided as much as shear
possible in statically loaded structures. Special treatment of the

contact surfaces, and the checks that are needed to confirm whether
the bolts are preloaded adequately, are expensive. In addition to
the bolts themselves other components are often needed to transfer
forces in a bolted connection, such as plates (stiffeners) or parts of
steel sections. Figure 1.6 shows several examples of these elements
in a beam-to-column connection.
1.5 Bolts loaded in combination
of tension and shear.
angle partial end-plate flange plate

stiffener
bolt
fin plate
full packing plate

stiffener
angles end-plate flange plates with fin plate

(bolted to the web of the (welded to the beam and (flange plates: welded to the flange of the
beam and to the flange bolted to the flange of column and bolted to the flange of the beam
of the column) the column) fin plate: welded to the flange of the column
and bolted to the web of the beam) 1.6 Parts of bolted beam-to-column connections.

Other connection means
As an alternative to welds and bolts, rivets or pins are sometimes used for connec
tions in steel structures, see Connections 3, section 3.3 and 3.4. Pins are mainly used
in hinged connections or structures that must be disassembled easily. Rivets are used
for renovation and maintenance of historic steel structures, build before 1940.
To connect cold formed sections and thin sheets, for example roof and facade
claddings, special screws, rivets and nails are used.
bending of the
Local reinforcements
column flange Sometimes it is necessary to reinforce the connected sections locally so they can
yielding of the
column web
transfer concentrated forces in the connection. Figure 1.7 shows an example of a
beam-to-column connection, with the possible local failure modes of the column.
shearing of the Depending on the dimensions and the yield strength of the column flange and
column web
column web, it might be necessary to reinforce the flange and/or web locally or
support it to prevent premature local failure of the column. Figure 1.8 shows a
local buckling of number of structural solutions that can be used to prevent local failure, for instance
the column web
horizontal and diagonal stiffeners, application of a column flange backing plate,
and column web doubler plates. Before choosing any one of these types of
reinforcement, the structural designer should check whether it may be more
1.7 Possible local failure modes in a non-stiffened beam-to-
economic to reconfigure the frame so that a smaller bending moment capacity of
column connection. the connection is acceptable, or to choose heavier section(s).
plug
fillet weld
weld
horizontal stiffeners backing plate horizontal and doubler plate welded on

diagonal stiffeners the web of the column
1.8 Possible types of stiffening of a beam-to-column connection.

1.3 Common connections found in buildings
In a steel building structure, the following types of connection can occur, see also figure 1.2:
– column splice;
– column base;
– beam-to-column connection;
– beam-to-beam connection;
– beam splice;
– connection of bracings.
In addition, connections can occur ‘within’ beams and columns, for example a beam executed
as a truss, or a build-up column.
Connections (and joints) must meet the requirements of EN 1993, which relates to the structural
characteristics of strength, stiffness and deformation capacity. In addition, when choosing and
designing a connection, the structural designer should pay attention to costs, aspects of manu
facturing and assembly, and usability, see section 1.4.
Satisfying these requirements and conditions does not lead to unique solutions for detailing.
The choice of the most appropriate type of connection also depends on the judgement and
experience of the structural designer and which requirements are most relevant for a given si
tuation. The choice of the type of connection is also influenced by the configuration of the
fabrication shop that will produce the steel structure. Therefore, the detailing should preferably
be decided by close collaboration between the designer and the steel fabricator. For these
reasons, it is impossible to present a complete overview of all possible connections: each situa
tion is different. However, due to the countless possibilities that are available when designing
in steel, the designer is always able to find a suitable detail. To give an idea of the possibili
ties, this paragraph presents the most common types of connections. Professional journals and
magazines – such as Bouwen met Staal (The Netherlands), Info_Steel (Belgium/Luxembourg),
Revue Construction Métallique (France), Stahlbau (Germany), Steel Construction (international) and
Steeldoc (Switserland) – regularly publish case studies of projects, where the chosen connections
are shown in detail. It is also instructive to look at the joints in steel structures in buildings that
are under construction, and in finished buildings.
The assumptions considered in the frame design calculations must always be reflected in the design
of the connections and their associated mechanical properties, such as the rotational stiffness. Take
for example a beam to beam connection, in a frame where it has been assumed that the secondary
beams are restrained by the primary beams against lateral torsional buckling. This may mean that
the connection between the secondary beam and the primary beam should prevent rotation along
the longitudinal axis of the beam (twist). The properties of the beam to beam connection, achieved
through its detailing, must be consistent with this assumption. When this is not the case, lateral
torsional buckling of the secondary beams must be prevented in another way, for example by
restraining the compression flange against side-ways movement using floor or roof plates.

1.3.1 Column splice
A column splice is a connection between two lengths of column, in the longitudinal direction. The
simplest splice connection is a welded splice, in which the column ends are welded directly (‘cold’)
against each other (fig. 1.9). A column splice is also possible when the two columns have a different
size, provided the web and flanges are sitting above each other (fig. 1.10). When such a connec
tion is to be used as part of a frame assembly on site, the connection will need to be temporary
fixed with, for example, two strips until the connection is welded.
weld
detail In a bolted splice connection where the column ends are placed directly (‘cold’) against each
other – depending on the preloading of the contact surfaces – the forces are either fully or partially
butt welds
transferred by cover plates. In the case of two column parts with a different cross-section – or when
there are dimensional variations (rolling tolerances) between the two parts – packing plates must
1.9 Column splice between identical sections. be applied between the flanges and cover plates (fig. 1.11). Such a connection avoids the need
for welding on site, but due to its greater size it is sometimes not desired. The connection requires
more space locally and is visually less attractive.
HEA 240 Often, a bolted column splice with welded end-plates is used (fig. 1.12). It should be noted that the
end-plates can deform during welding. In tender specifications it is sometimes stated that contact
surfaces must be milled flat after welding, although structurally this is unnecessary and in the case
of preloaded connections, deformed end-plates are even more favorable, see Connections 3
(Slip-resistant connections, rivets, and pins), section 3.1.3.
weld 1.3.2 Column base connection

detail
Generally, a column base consists of a base plate – which may be stiffened with welded plates –
butt weld
and fillet weld and an anchorage to the foundations. The dimensions of the base plate must be such that the
HEB 240
1.10 Column splice between two column parts

HEB 240
with a different cross-section.
packing plate
packing plate
welded
end-plate
cover plate
cover plate
cover plate HEM 240
1.11 Column splice with cover plates and packing plates. 1.12 Column splice with welded end-plates.

compression force in the steel column is sufficiently spread over the (less strong) concrete of the
foundation or substructure. In the case of columns that are loaded eccentrically or in tension, the
base plate must also be able to transfer the tension forces to the anchors.
A space must be left between the underside of the base plate and the top of the foundations. This
space is needed to allow for adjustments to accommodate dimensional tolerances in both the steel
structure and the concrete substructure. After the steel structure has been erected the space is filled
with a non-shrink mortar or grout. The two common methods for placing columns are:
• With adjustable nuts under the base plate. With this method, four anchors are required in order
to align the column in two directions. According to EN 1090-2, cl. 9.5.4 the adjustable nuts do not
need to be turned back unless this is explicitly specified otherwise. When preloaded anchors are
used, turn back of the adjustable nuts is necessary, because these nuts would otherwise prevent
the anchor from being preloaded over a sufficient length.
• With wedges. The wedges are removed after pouring and hardening of the grout under the 1.13 Column base connection with four anchors.
base plate.
Figure 1.13 shows a column base connection with four anchors. In addition to shear forces and
normal forces, this connection can transfer bending moments. When the anchors do not need to
transfer bending moments, one can choose a base plate that does not extend outside the column
and only uses two anchors (fig. 1.14). Since in this case alignment using the adjusting nuts is
only possible in one direction, this solution is discouraged because of execution considerations.
However, small columns and ‘architectural steel’ may be exceptions.
Shear forces can be transferred by friction and through the anchors. In the case of large shear
forces, it may be necessary to weld a shear tab (or dowel) to the bottom side of the base plate
that is then located in a recess in the concrete. The space between the shear tab and the base
plate and the concrete must be filled with a non-shrinking mortar. A shear tab can consist of a
solid steel strip or a short piece of a steel section (fig. 1.15). Figure 1.16 shows a column base
connection with welded stiffeners. This solution should be avoided. Using a thicker base plate
without stiffeners is preferable because, despite higher material costs, a thicker base plate without
stiffeners is always cheaper than a thin plate with stiffeners.
1.14 Flush base plate with two anchors.
non-shrinking
mortar
shear tab column base plate stiffener
1.15 Shear-tab to transfer large horizontal forces. 1.16 Base plate connection with stiffeners.

with a partial depth end-plate with angles with a fin plate
e e e
beam connected to the flange of the column: (top) elevation and (bottom) plan
with a partial depth end-plate with angles with a fin plate
e=0 e e
beam connected to the web of the column: (top) elevation and (bottom) plan
1.17 Connections for transferring shear and/or normal forces between beams and columns.

connection advantages disadvantages
end-plate – most strong and stiff, – high manufacturing costs due to

which is advantageous for welding the plate to the beam
assembly – less compact for transportation (beam sections
cannot be nested)
– section variations on shape and dimensions and
deviations during manufacturing and assembly
must be overcome using packing plates
– difficult to assemble
angles – automatic production of – lower strength than end-plate connections

(angle) sections – many parts (angles and bolts)
– dimensional variations can – higher assembly costs (for assembly on site)
be accommodated by the
clearance of the holes
– compact for transportation
(beam and column sections
can be nested)
fin plate – dimensional variations can – relatively low strength

be accommodated by the – fin plates can warp during welding and be
clearance of the holes damaged during transportation
– easy assembly – not compact for transportation (column sections
– no welding on the beam cannot be nested)
– asymmetrical connection of the beam to the fin
plate (decide beforehand on which side the plate 1.18 Characteristics of nominally pinned
will be connected) connections with (partial depth) end-plates,
angles, and fin plates.
1.3.3 Beam-to-column connection

EN 1993-1-1, cl. 5.1.2 distinguishes so-called ‘nominally’ pinned connections and (rigid and
semi-rigid) moment connections. A ‘nominally’ pinned connection transfers only shear forces,
possibly in combination with normal forces. A moment connection is able to transfer a bending
moment.
Nominally pinned connection

In practice, many connections between beams and columns that are used to transfer shear force
and/or normal force are formed with end-plates, angles, or fin plates (fig. 1.17). Table 1.18 shows
their most important characteristics.
Moment connections
The detailing of moment connections depends on the schematization that is assumed for the design
of the structure. When it is assumed that no rotation occurs in the connection and that the joint is
stronger than the joining sections, one speaks of a completely rigid and full strength connection.

(symmetrical) end-plate connection with horizontal stiffeners (symmetrical) end-plate connection with a haunch
1.19 Rigid moment connections between beams and columns.
1.20 Semi-rigid (flexible) moment

with a full depth end-plate with an extended end-plate connections between beams and columns.
Two details that are often used to achieve these characteristics are shown in figure 1.19. In order
to be able to meet the requirements, generally a reinforcement must be applied in the form of
stiffeners and/or haunches.
It is also possible to take the flexibility of the connection into account in the frame design and/
or make the connection less strong than the joining sections. With such approaches expensive
reinforcements can often be omitted. Figure 1.20 shows two examples of such a connection.
Other variants
For the types of connection described above the column is continuous and the beam is connected
to the column. It is also possible to have a continuous beam and connect the column to the beam
(fig. 1.21). The top of the column has an end-plate that connects to the beam. The beam transfers

beam
partial
stiffener
end-plate
column with a single continuous with welded fin plates

plate (thin walled hollow section) (thick walled holow section)
1.21 Beam-to-column connection with a continuous beam. 1.23 Connection of beams to a hollow section column.
welded bolted bolted
1.22 Connection with fitted stiffeners. 1.24 Knee connections in a portal frame.
the contact pressure by means of stiffeners. These stiffeners do not need to fit under the top flange
of the beam unless there is also a column on top of the beam (fig. 1.22). Full stiffeners must be
fabricated to exact dimensions to allow for rolling tolerances in the beam section.
Figure 1.23 shows a connection of a beam to a hollow section column. For a thick-walled hollow
section it is possible to weld the connection plates directly to the column. When calculations show
that the tube wall is not thick enough for this solution, two opposing slots can be fabricated in the
hollow section walls. A single connection plate is then inserted through the slots and welded to the
outside of the column.
Figure 1.24 shows three different variations for the connection of a column to a beam in a portal
frame (knee connection): completely welded, bolted through the flange with an additional end-
plate, and bolted with angle cut beams and columns and welded end-plates.

possible extra
stiffeners
primary secondary
beam beam
secondary primary
beam beam
1.25 Floor system with beams in two layers and simply supported secondary 1.26 Floor system with beams in one layer, in which the upper side of the
beams. secondary beams is lower than the upper side of the primary beams.
1.3.4 Beam-to-beam connection

In a floor system with beams in two (stacked) layers, the secondary beams can be simply supported
by the primary beams (fig. 1.25). The structural design should indicate whether any local stiffeners
are necessary.
In a floor system with beams in a single layer, the upper flanges of the secondary beams may be
either lower than, or at the same level as, the upper flanges of the primary beams. Flanges at different
levels may be used for composite floor slabs. A fin plate connection between the secondary beam
and the primary beam is a suitable solution (fig. 1.26). This is a low-cost connection in terms of both
fabrication and erection.
Different solutions for the upper flanges of the beams at the same level are shown in figure 1.27. The
variants in which the upper flange of the secondary beam – or both flanges – is shortened (notched)
are more expensive, but this is sometimes necessary in case of large shear forces. The characteristics
of the various beam-to-beam connections shown in this figure correspond to the beam-to-column
connections in table 1.18.
In general, non-notched connections with end-plates or angles will provide sufficient restraint against
twist of the beam end. Fin plate connections and notched connections are often not rotationally stiff
to provide full resraint against twist.

single welded fin plate double bolted angles single welded end-plate
and notched beam
single welded fin plate double bolted angles single welded end-plate
and notched beam and notched beam and notched beam
1.27 Floor system with beams in one layer, in which the upper side of the secondary beams is level with the upper
side of the primary beams.
flange plate
1.3.5 Beam splice

web plate
A beam splice is needed when the designed beam is longer than the maximum available (com
mercial) length, or when an assembly connection is necessary on site, for example due to transpor
tation limitations. In the first case, the beam splice can be welded or bolted in a workshop, whereas
a bolted connection is almost always chosen in the second case (for execution on the construction
site). A welded beam splice is in principle the same as a welded column splice (see fig. 1.9).
flange plate
When there is a moment and a shear force at the location of a bolted beam splice, the flanges are
connected with flange plates and the webs with web plates (fig. 1.28). Flange plates can be applied 1.28 Beam splice for resisting a bending moment
and shear force.
either single-sided or double-sided. When in the frame design a hinge is assumed at the location
of the splice – for example in a Gerber beam – only shear force needs to be transferred. In that
case, only web plates, or beam end-plates, are needed (figure 1.29).
welded
end-plate
1.29 Beam splice for resisting a

shear force alone.

1.30 Connection of a horizontal bracing (in the plane of the upper flange of the beams).
1.32 Wind bracing in a floor system with beams

1.31 Connection of a horizontal bracing (in the plane of the bottom flange of the secundary beam). in two layers.
1.33 Wind bracing with an extra horizontal 1.34 Connection detail of a vertical bracing. 1.35 Alternative solution for the connection detail of a
channel section. vertical bracing.

1.3.6 Connection of bracings
Bracing is used in almost all steel structures to ensure the structural stability of roofs and floors
(horizontal bracings), as well as facades and walls (vertical bracings). The structural engineer has
a lot of options to choose from, for both the choice of the bracing sections and the detailing of
the connections. A few examples are shown below:
• Figure 1.30 shows an example of a horizontal bracing using equal angles. The bracing lies in
the same plane as the upper flanges of the primary and secondary beams. The gusset plates and
the bolt heads might hinder the placement of trapezoidal sheeting used for roofs or floors. An
alternative solution is shown in figure 1.31, in which the connections with the bracing are located
in the same plane as the bottom flanges of the secondary beams. In a floor system comprising
beams in two layers, the solution shown in figure 1.32 can be chosen for connecting the bracing
members. When the floor system has beams in only one direction – and thus primary beams are
not present – an extra section is necessary to form an edge beam as part of the horizontal (stability)
truss: the channel section in figure 1.33.
• Figure 1.34 shows an example of a vertical bracing for a vertical (stability) truss in a multi-storey
building. The diagonal bracing is connected to a gusset plate that is welded to the upper flange of
the beam and the end-plate. Through this end-plate, the beam and gusset plate are connected to
the column. In the solution of figure 1.35, the diagonal and beam are both connected to the gusset
plate, which is welded to the column. The upper flange of the beam has to be notched locally.
1.4 Design and detailing

Designing and detailing a connection in a steel structure involves more than just checking
whether the connection meets the structural requirements as described in the Eurocodes. Equally
as important are the non-structural aspects related to the construction and the cost of the steel
structure (fig. 1.36). Both topics are discussed in this section.
requirements
structural requirements non-structural requirements
deformation building
strength stiffness aesthetics economy
capacity technology
– connections to – material costs
concrete, timber, etc. – fabrication
– dimensions of the – corrosion protection
structural elements – assembly
– maintenance
1.36 Requirements for the design of a steel
– recycling / re-use structure.

costs material labor total (%)
1.4.1 Requirements regarding connections
The design of a steel structure is a process in which a large number of different aspects must be
materials 40 0 40
taken into account. As a result, a wide variety of assemblies and connections exists. Each steel
manufacturing 0 20 20
structure must, of course, always meet the structural requirements with regards to safety (strength,
corrosion
4 4 8 stiffness, and stability) and usability. In addition, non-structural aspects play an important role
protection
in the design. One of the most important aspects is economy – the cost of the steel structure.
fire protection 11 11 22
Generally, in the case of an office building, 55% of the costs are material costs and 45% are labor
assembly 0 10 10
costs (table 1.37). Roughly the same ratio applies for single storey industrial buildings, except
total (%) 55 45 100 that often less or no fire protection is needed.
1.37 General distribution in percentages of direct The manufacturing costs are strongly dependent on the connections chosen in a particular case.
costs for a steel structure of an office building.
In addition, the connections influence the costs of the design and drawings, and the costs of
corrosion protection. To minimize costs it is desirable to design a steel structure with as few
connections as possible, and for connections which are simple in terms of both manufacturing
and assembly. The number of connections is, for example, influenced by the choice of truss
or plate girder, or by the choice of the grid of the main structure. It is the responsibility of the
structural engineer to take into account both structural and non-structural aspects to achieve the
best solution. It is very important to develop and plan a good method for execution of the steel
structure and its connections. The manufacturing capabilities may differ significantly between
different steelwork fabricators. A given steel construction company may prefer a certain type of
structure and a certain type of connection because of its particular capabilities. One company
may specialize in welding, while another company has the machinery and experience such that
bolted connections are preferred (fig. 1.38).
1.38 Bolted connections in a steel structure.

A significant saving in costs can be achieved by limiting the variety of types of
connections, and by standardizing the most common connections. Fabrication of
connections in the workshop is generally easier, and therefore cheaper, than at
the construction site. This is particularly true for operations such as welding and
drilling holes. On site welding is not always practical, and is even not allowed in
certain cases, such as at refineries because of explosion risks. Usually, welding on
site is too costly because:
– to access a good and safe weld location, temporary facilities must created, for
example working platforms (fig. 1.39);
– the specimen cannot be rotated in order to obtain a favorable welding position;
– additional costs arise for the manufacturing and assembly of temporary bolts;
– disadvantageous weather conditions can hinder the execution of a connection,
or make it even impossible unless extra facilities are applied (such as tents);
– additional costs can result from inspection and possible repairs.
Welding on site also can be considered an indicator that the principles of design-
for-disassembly (as often required in circular construction) will not be met.
1.39 Temporary work platforms with welding tents for welding
a structure on site.
1.4.2 Cost-conscious design of connections
The costs of a steel structure can be minimized by proper detailing. It is important
that the designer fixes the starting points for the detailing at the tender phase,
because a steel fabricator will offer a price for the steel construction and produc
tion of manufacturing drawings or models (like BIM) based on this information. A
number of principles can be applied to achieve good detailing of a steel structure:
– limit the number of different sections;
– limit the number of different joint types;
– take into account the rolling tolerances of sections;
– choose as far as is possible solutions that are suitable for automated manufac
turing;
– use fillet welds, and limit the weld volumes;
– choose one, or maximum two, types of bolt for each structure;
– avoid accumulation of dirt and water on the structure;
– ensure easy assembly of the structure.
These starting principles are briefly explained below. In addition, it is important

to use as many common steel sections as possible (in stock) and in the maximum
lengths that are available commercially. Uncommon sections are almost always
more expensive (price per kilogram) than more common sections, and a minimum
order of quantity (in tons) is also required for some uncommon sections. The en
gineer should make sure that, while detailing, the structural characteristics of the
chosen solution are aligned with the design assumptions.

1.40 Tolerances of hot-rolled sections
tf dimensions
in accordance with EN 10034 (h is the
nominal height). height tol. width tol. web tol. flange tol.
(mm) (mm) (mm) (mm) (mm) (mm) (mm) (mm)
tw h h ≤ 180 +3/–2 b ≤ 110 +4/–1 tw < 7 ±0,7 tf < 6,5 +1,5/–0,5
180 < h ≤ 400 +4/–2 110 < b ≤ 210 +4/–2 7 ≤ tw < 10 ±1,0 6,5 ≤ tf < 10 +2,0/–1,0
400 < h ≤ 700 +5/–3 210 < b ≤ 325 +4/–4 10 ≤ tw < 20 ±1,5 10 ≤ tf < 20 +2,5/–1,5
b h > 700 +5/–5 b > 325 +6/–5 20 ≤ tw < 40 ±2,0 20 ≤ tf < 30 +2,5/–2,0
40 ≤ tw < 60 ±2,5 30 ≤ tf < 40 +2,5/–2,5
tw ≥ 60 ±3,0 40 ≤ tf < 40 +3,0/–3,0
tf ≥ 60 +4,0/–4,0
b
perpendicularity k + k’ symmetry error e (for flange) straightness qxx or qyy
k’ width tol. thickness width tol. height tol.
(mm) (mm) (mm) (mm) (mm) (mm) (mm)
b ≤ 110 1,5 b ≤ 110 2,5 80 < h ≤ 180 0,30L
b > 110 0,02b ≤ 6,5 t < 40 110 < b ≤ 325 3,5 180 < h ≤ 360 0,15L
b > 325 5,0 h > 360 0,10L
k
b1 b2 110 < b ≤ 325 5,0
t ≥ 40 q
b > 325 8,0
1 L
e = (b1 – b2)
2
1.41 Preferably use

partial stiffeners based convenient less convenient notes
on manufacturing and
prevent using stiffeners
maintenance costs. as much as possible
details
general
partial stiffeners allow for rolling tolerances stiffeners are fitted between flanges
(but with reduced structural efficiency)
1.42 Example of full stiffeners.

Limit the number of different types of connection
When a lot of different connection types are used the cost of the structure increases. Each dif
ferent solution means additional drawings and calculations, and requires extra attention during
manufacturing and assembly. Adopting unique solutions is in principle permitted when they are
architecturally needed, or when the structure cannot function without this particular solution. It is
discouraged to design connections differently merely to save material, as the material cost saving
does not outweight the cost of extra labor in the manufacturing and assembly phases.
Take into account the tolerances of sections

Tables of section data only include the nominal (theoretical) dimensions of rolled sections, despite
the fact that these sections can have a rolling tolerance of a few millimeters. The designer must
keep these dimensional variations in mind when structural parts are formed from rolled sections.
EN 10034 describes the tolerances of rolled sections (fig. 1.40). Also, dimensional variations make
stiffeners over the full height of a section relatively expensive. Due to the inevitable within-
tolerance deviations, it takes a lot of labor to make the stiffeners so that they fit precisely between
the two flanges. Moreover, a structure with partial stiffeners is easier to protect and often less dirt
will accumulate (fig. 1.41; see also fig. 1.21). Sometimes, full stiffeners are necessary to transfer large
forces through a section. This may be the case for trusses consisting of multiple bays, for Vierendeel
trusses, and for supports (see fig. 1.22). Figure 1.42 shows an application of full stiffeners.
Choose solutions that lend themselves to automated manufacturing

Figure 1.43 shows schematically the routing through a steel fabrication workshop. Sawing, drilling,
cutting, shearing, and punching are carried out with a computer driven machine. In a drilling and
sawing line members with the correct length and the specified holes are produced quick and
accurate. Welding requires – depending on the welding process – a varying amount of work to
be done by hand. The manufacturing costs of structural elements produced in an automated way
are much lower than when a lot of hand work is required.
small
welding
parts
sawing,
e.g. plates intermediate corrosion intermediate
cutting, transport
assembling protection storage
and drilling
stock rust removal
plates and and adding
sections primer
cutting, 1.43 Schematic presentation
e.g. beams
setting and of the route through a steel
punching
fabrication workshop.

convenient less convenient notes
– use preferably fillet welds

– avoid preprocessing as
much as possible
– avoid fitting plates as much
as possible
welding a b c a b c
of sections
a. by welding the plates on the column the a. the plates are welded in between the flanges;
influence of weld tolerances is eliminated; every plate has to fit due to tolerances and
b. welds are not preprocessed; welds need to be preprocessed
c. welds are executed as fillet welds. b. welds are preprocessed;
c. welds are executed as butt weld
deformation – when welding, keep in mind

due to shrinkage
the effects of shrinkage
plate connected by fillet welds plate connected by a butt weld
butt weld – preferably use fillet welds

5 mm
welds
10 mm
butt weld
end-plate connected by fillet welds when using butt welds, the end-plate must
(two alternatives have been given) be preprocessed
– this mainly applies to

statically loaded structures
– in dynamic loaded structures
fillet welds are not always
possible
fillet welds butt welds
preferably use fillet welds: it is not needed to avoid the use of butt welds:
preprocess the weld preprocessing is needed
1.44 Preferably use fillet welds.

Use fillet welds and limit weld volumes
Welds can be executed as fillet welds or butt welds, see figure 1.3 and Connections 4 (Welds). An
important advantage of fillet welds is that the welding edge does not have to be pre-processed
(fig. 1.44). In the case of butt welds, the connecting parts must have beveled edges; this is time
consuming, and therefore expensive. For structural reasons, butt welds are sometimes unavoidable.
Reasons of aesthetics may also lead to the choice of butt welds.
Welding of steel structures is still often done manually, and this means that the costs of welds are
approximately proportional to the weld volume (the volume of the weld seam). The smaller the weld
volume, the lower the costs. The amount of welding work can be limited by avoiding stiffeners,
by limiting throat thickness, and – if possible – by using intermittent fillet welds (fig. 1.45). Due
to the dangers of crevice corrosion, intermittent fillet welds may only be applied for temporary
structures, and permanent structures that are exposed to a non-aggressive environment, such as
heated interior spaces and outdoor climates in corrosion classes C1 and C2, according to ISO
9223. Intermittent fillet welds can be used in galvanized structures. In that case the welds must
be positioned opposite one another and the production starts at the center of the element. When
using intermittent fillet welds, the weld volume and thus the heat input is reduced (and so the
deformation due to weld shrinkage). 1.45 Intermittent fillet welds.
Choose one or at most two types of bolt per structure

Bolted connections are commonly used in steel structures. From a structural point of view, it is
logical to use the optimal bolt diameter and bolt class for each connection. However, given cost
considerations it is actually better to use one type of bolt for the entire structure, to limit the
number of different bolt dimensions (length and diameter) and bolt classes. This makes the
assembly easier. It is best to choose bolts that are threaded over the full length of their shank,
and limit the number of different bolt lengths by increasing the lengths in 10 mm increments rather
than 5 mm. The positioning of the bolts is chosen first to satisfy assembly and manufacturing
considerations (fig. 1.46), with spacings determined as a function of the dimensions of the tools
used. After that, the structural checks follow and adjustments can be made.
1.46 Good practice for manufacturing and

convenient less convenient notes assembly of bolted connections.
– drilling equipment is
used most efficiently
when a structural
element has bolt holes
of equal diameter, and
simple bolt patterns are
drilling
used
– choosing equal edge
distances avoids
mistakes when turning
choose equal bolt diameters, and by different bolt diameters, and the over the section
drilling holes in one line mistakes holes are not drilled on one line
are less likely

convenient less convenient notes
preferably use
6 mm stiffeners and
details welded plates of plates of equal
general equal thickness thickness
8 mm
10 mm
1.47 Use of plates with equal thicknesses.
Limit the number of different plate thicknesses

Limit the number of different plate thicknesses (fig. 1.47) and choose conventional (com
mercially available) dimensions for the thickness of the plates. Use the following guidelines
for the design of plates, lips, and stiffeners:
– make plates preferably from steel strips, and align the length or the width to commercially
available widths of strip;
– form plates from steel sheet, and choose a length or width so that cutting loss is mini
mized.
Prevent accumulation of dirt and water on the structure

A steel structure that is exposed to the outside elements has to be protected to prevent
corrosion. In practice, corrosion occurs first at the connections, at locations where the
standard section
structure remains wet, and near corners and edges. On sharp corners – such as edges
of steel sections – a protective organic coating will be relatively thin. This thinning can
2 mm
be prevented by removing the sharp edges during manufacturing (fig. 1.48). However
such rounding-off of the edges is an extra treatment that is expensive, and therefore only
with rounded edges applied to structures that are exposed to an aggressive environment, such as corrosion
classes C4 and C5. Detailing should be such that dirt cannot accumulate locally, to avoid
1.48 Rounding of sharp corners of a areas where the structure remains wet.
rolled section.
Ensure that the structure is easy to assemble

A steel structure is normally so large that it must be transported in parts to the construction
site. The structure should be easy to assemble, and stabilize, during assembly (fig. 1.49).
The design of a structure can be strongly influenced by assembly considerations. This is,
for example, the case when designing a steel frame for a multi-storey building. From a
structural point of view it would be possible to reduce the cross-section size of the columns
for each storey from the bottom to the top of the building. However, given manufacturing
and assembly considerations, it is better to limit the number of different sizes. A common
compromise is to maintain a uniform column cross-section over a height of three storeys.
This limits the number of different structural components, thereby reducing the number of
assembly operations and increasing the speed of construction.

convenient economic assembly less convenient economic assembly notes
lifting of one small

structural element is
sypplying not faster than one
structural larger element
– aim for a minimum number of structural elements
elements
– trusses should be welded and delivered as large as possible – small structural elements
and with welded connection – trusses assembled on site with bolts
connection above supporting beam: connection not above supporting beam:

– erector can move safely over supporting beam – erector is not sitting safely on the beam; he has to
– (weld) activities can be carried out safely climb dangerously to connect the next beam
– no special assembly connections are needed – other special facilities are needed
position of staggered joints

the connections direction of assembly A B A B
in beams
connection
edge beam
edge beam
single span beam
beams staggered over two bays: – next main beam must provisionally be held
– cantilevered beams support the next main beam – loads on beams over three supports leads to higher loads
– uniform load of the main beam on beam A
use alternative B when

the beam does not
have to resist lateral
torsional buckling
A B
possibilities – beams are easy to assemble from above with a

for conections bolted connection; alternative B is preferred
between beams – beam length is 3 to 5 mm shorter than the theoretical – beams must be erected sideways
length due to easy assembly – temporary bolts must be removed
– right end-plate runs continuously with one additional – both beams must be erected at the same moment
bolt used to erect the right beam and temporarily fixed in order to be able to add the
– thereafter, the left beam will be erected and both bolts
1.49 Provide design
upper bolt rows are applied
for a simple assembly.

design calculations and code checking
EN 1993-1-8
Eurocode 3. Connections
execution products
references standards, group 7 references standards, group 1

EN 1090-2 weldable structural steel
references standards, group 2

dimensions, tolerances and
technical delivery conditions

structural hollow sections

bolts, nuts and washers

welding consumables
and welding

information is given in the
1.50 System of relevant standards for connections National Annex
in steel structures.
1.5 Standards for connections

Standards and regulations play an important role in the design and construction process of a
steel structure, particularly with respect to the design, manufacturing, and assembly of the con
nections. EN 1993 is used for the design of steel structures. This standard comprises several
parts, with all rules on connections included in part EN 1993-1-8.
In addition to standards for the analysis and design, there are also standards covering execution
and products. EN 1993-1-8, cl. 1.2 refers to all relevant standards. The complete European system
for connections in steel structures is schematically shown in figure 1.50.
1.5.1 EN 1993-1-8
EN 1993-1-8 provides rules for the analysis and design of connections in steel structures, subjected
to static loads and steel grades S235, S275, S355, or S460. This standard may also be used for con
nections in higher steel grades, up to and including S700, provided that the additional rules in EN
1993-1-12 are taken into account. For structures subject to fatigue loading EN 1993-1-9 applies.
Figure 1.51 gives an overview of the contents of EN 1993-1-8.
The foreword to EN 1993-1-8 contains an overview of all clauses in which a national choice is
permitted via a so-called nationally determined parameter. In most cases, recommended values
are provided in a comment. In other cases, permission is given to provide additional information
a
through a National Annex. Table 1.52 shows an overview of the nationally determined parameters.
In addition to the nationally determined parameters, the National Annex to EN 1993-1-8 may also
b contain complementary information provided that it is not in conflict with the EN text.

1 Introduction 4 Welded connections
1.1 Scope 4.1 General
1.2 Normative references 4.2 Welding consumables
1.3 Distinction between principles and application rules 4.3 Geometry and dimensions
1.4 Terms and definitions 4.4 Welds with packings
1.5 Symbols 4.5 Design resistance of a fillet weld
4.6 Design resistance of filled welds all round
2 Basis of design 4.7 Design resistance of butt welds
2.1 Assumptions 4.8 Design resistance of plug welds
2.2 General requirements 4.9 Distribution of forces
2.3 Applied forces and moments 4.10 Connections to unstiffened flanges
2.4 Resistance of joints 4.11 Long joints
2.5 Design assumptions 4.12 Eccentrically loaded single fillet or single-sided partial
2.6 Joints loaded in shear subject to impact, vibration and/or penetration butt welds
load reversal 4.13 Angles connected by one leg
2.7 Eccentricity at intersections 4.14 Welding in cold-formed zones
3 Connections made with bolts, rivets or pins 5 Analysis, classification and modelling
3.1 Bolts, nuts and washers 5.1 Global analysis
3.2 Rivets 5.2 Classification of joints
3.3 Anchor bolts 5.3 Modelling of beam to column joints
3.4 Categories of bolted connections
3.5 Positioning of holes for bolts and rivets 6 Structural joints of H or I sections
3.6 Design resistance of individual fasteners 6.1 General
3.7 Group of fasteners 6.2 Design resistance
3.8 Long joints 6.3 Rotational stiffness
3.9 Slip-resistant connections using 8.8 or 10.9 bolts 6.4 Rotation capacity
3.10 Deduction for fastener holes
3.11 Prying forces 7 Hollow section joints
3.12 Distribution of forces between fasteners at the ultimate 7.1 General
limit state 7.2 Design
3.13 Connections made with pins 7.3 Welds
7.4 Welded joints between CHS members
7.5 Welded joints between CHS or RHS brace members
and RHS chord members
and I or H section chords
and channel section chord members
1.51 Content of EN 1993-1-8.
clause parameter or item recommended
1.2.6
reference standards for rivets information may be given in the National Annex
(group 6)
partial safety factors γM2, γM3, γM3,ser, γM4, γ M2 = 1,25; γ M3 = 1,25; γ M3,ser = 1,1;
2.2(2)
γM5, γM6,ser en γM7 γ M4 = 1,0; γ M5 = 1,0; γ M6,ser = 1,0
3.1.1(3) bolt class the National Annex may exclude certain bolt classes
requirements for the level of preloading, if preload is not explicitly used in the design calculations
exclusively related to the characteristics during for slip resistances but is required for execution purpo
3.4.2(1)
manufacturing and assembly or to the quality ses, or as a quality measure (e.g. for durability) then the
requirements level of preload can be specified in the National Annex
the Nation Annex may give additional information on

additional information related to the classifica
5.2.1(2) the classification of joints by their stiffness and strength
tion of joints
to that given in cl. 5.2.2.1(2)
1.52 Overview of the clauses
additional information with regard to the use of the National Annex may give further information on the
6.2.7.2(9) in EN 1993-1-8 in which a
equation (6.26) use of equation (6.26)
national choice is permited.

1 Scope
1.5.2 EN 1090-2
2 Normative references
2.1.3 Welding consumables The design rules in the Eurocode are based on the assumption that the structure is manufactured
2.1.4 Mechanical fasteners
a
2.2 Preparation and assembled according to the principles of ‘good practice’. For steel structures this require
2.3 Welding
2.4 Testing ment is met when the rules of EN 1090-2 are followed. Figure 1.53 shows the main structure of
3 Terms and definitions this standard, with the clauses for connections indicated in blue. Table 1.54 provides an overview
4 Specifications and documentation of the content and status of the annexes of EN 1090-2, with the relevant annexes indicated
5 Constituent products in blue.
5.5 Welding consumables
5.6 Mechanical fasteners
6 Preparation and assembly
6.6.1 Dimensions of holes (table 11 covers the 1.5.3 Product standards
nominal clearances for bolts and pins in normal
round holes, oversize round holes, and long and
short slotted holes) The dimensions and mechanical properties of the products used for connections in steel
6.6.2 Tolerances on hole diameter for bolts and pins structures – such as bolts, nuts, washers, and pins – are included in the relevant product
6.6.3 Execution of holing
7 Welding standards, see Connections 2 (Bolts in clearance holes), section 2.2.1. Standards for construction
8 Mechanical fastening procedures, for example weld procedures and weld classification, are also available. The
9 Erection European standards system is based on products that are standardized in accordance with
10 Surface treatment b EN-standards, and in some cases the internationally accepted ISO (International Standard
11 Geometrical tolerances
Organization) standards. EN 1993-1-8, cl. 1.2 and EN 1090-2, chapter 2 indicate which product
12 Inspection, testing, and correction
12.4 Welding standards are applicable, see also figure 1.50. Products that are supplied in accordance with
12.5 Mechanical fastening
standards other than EN- and ISO-standards are not excluded, but it has to be proven that
those products have equivalent properties.
1.53 Content of EN 1090-2. The chapters and
sections colored in blue are those most relevant for
connections.
annex status topic
additional information, list of options, and requirements related to the

A N
execution classes
B I guidance for the determination of the execution classes
C I checklist for the content of a quality plan
D N geometrical tolerances
E I welded joints in hollow sections
F N corrosion protection
G N test to determine the slip factor
H N calibration test for preloaded bolts under site conditions
J N use of compressible washer-type direct tension indicators
K I Hexagon injection bolts
guide to flow diagram for development and use of a WPS (welding

L I
1.54 Annexes to EN 1090-2 (N = normative; procedure specification)
I = informative). The sections colored in blue are those
M N sequential method for fasteners inspection
most relevant for connections.

120 mm
1.6 The role of the computer 30 60 30
Hand calculations for the design of connections in steel structures

require a lot of repetition. So it is obvious that computer programs can
play an important role in the design. It is possible to use spreadsheets M16 a = 4 mm
40
class 8.8 80
for standard connections, but also special connection programs are
80
available in the market.
Most programs for the analysis and design of connections are limited a=4
to the type of connections that can be designed using Eurocode
rules, or with the rules for standard components covered in the 220 145
Eurocode. Therefore, standard computer programs have limitations M2 V2 V1 M1

and manual calculations may still be needed, even for conventional
IPE 220 40
connections. That said, the computer is an indispensable tool for 5
the design of both rigid (continuous) and flexible (simple, or pinned) 12
connections, simply because a reliable hand calculation of these
HEB 140
connections takes a lot of time. To find an optimal design multiple
variants have to be designed and compared, and it is nearly impossible to do this by hand. 1.55 Example of a double-sided beam-to-column
connection.
In practice numerous computer programs are available for the design of standard connections. In
many cases it is possible to couple the connection program to a frame design program, so that it
is possible to design complete frames with semi-rigid connections. Often it is possible to use the
connection geometry from the design program as input for the numerical controled machine tools.
The use of a computer program for the design of a connection is illustrated for the connection shown
in figure 1.55: a double-sided beam-to-column connection in a non-sway portal frame. The pos
sibilities and the steps needed for the design are discussed here using (parts of) screen shots of the
connection program COP [1]. The first screen shot shows that a connection is chosen from the range
of available standard configurations. In the example a moment connection is chosen using a double-
sided beam-to-column connection where the beam is connected to the column flange (fig. 1.56).
1.56 Connection program: choice of configuration,

specification of design procedures and input of
beam lengths.

1.57 Connection program: choice of the sections
and steel grade for the beams.
1.58 Connection program: choice of the sections

and steel grade for the column.
1.59 Connection program: selection of the specific details of the various components, with visualization.

The design procedure, elastic or plastic, can be specified for the global analysis
and the joint verification. Next it is necessary to insert the beam lengths.
Subsequently, the beam and column sections are chosen from the available
section files; in the example an IPE 220 is chosen for the beams and an
HEB 140 is chosen for the column (fig. 1.57 and 1.58). The steel grade must
also be chosen; in this case S235 The program automatically generates the
geometrical data from the chosen cross-sections. Hereafter all other aspects
of the connection are specified. Here an extended end-plate connection
without any stiffening is chosen (fig. 1.59).
The specific details of the various components are selected. For the end-plates
this relates to the steel grade and the dimensions, for the bolts it is the
diameter, the bolt class and the shear plane (through the thread or through the
shaft). The locations of the bolts are also determined, based on edge distances
and spacings (fig. 1.59), as well as the dimensions of the welds between the
end-plate and the beam end (not shown). The program checks whether or not
the chosen dimensions of the various elements meet the requirements of the
Eurocode. Examples include: verification of the bolt positioning (does it comply
with the limitations for edge distances), and verification of the weld sizes (do
they comply with the minimum throat thickness). When the requirements are
not met most programs display a warning and suggest adjusting one or more
parameters. The user can accept the suggestion but is also free to ignore it.
After inserting these data the geometry of the connection is fixed. Many
programs offer the opportunity to have a 3D view of the connection, from
different sides. When the connection model is deemed satisfactory the next 1.60 Connection program: input of the loads.
step is to apply the loads via an input screen (fig. 1.60). Subsequently, the
analysis and design is performed and the results can be seem directly from
the graphical presentation.
As well as the resistance, the program may also calculate the stiffness of
the connection. The result is shown in terms of a moment/rotation (M/ )
diagram. Having done this the classification of the connection can be deter
mined, based on its strength and its stiffness (fig. 1.61). The resulting M/
diagram can be implemented in the model, and after that the response
of the structure can be calculated using non-linear calculation methods. The
non-linear characteristics of the connection may influence the frame response.
As a result of this the connection may need to be rechecked for a revised
combination of forces, as part of an iterative process.
1.61 Connection program: presentation of the calculation

results (strength, stiffness, and classification), the decisive
failure mechanisms, and the M/ diagram.

The program also indicates the decisive failure mode for a given configuration. This
is useful especially when some means of improvement, for example add in stiffeners,
is needed because the specified connection does not fulfill requirements (fig. 1.62;
see also fig. 1.8). When the connection detail is finalized the results may be
presented in a calculation report, which often includes a drawing of the connection
with its dimensions and other relevant data.
The screen shots presented above were obtained from one specific connection
program. Although other suppliers offer programs with similar possibilities, the
a. horizontal stiffeners between the column flanges
on-screen presentations will obviously be different. When choosing a connection
program it is recommended to take note of the following aspects:
– the number of different types of connection available;
– the simplicity and the clarity of the data input;
– check if the dimensions of the components meet the limits given in the Eurocodes;
– the clarity of the calculation results;
– the link with the frame design program;
– translation of geometrical data to numerical data (as input for a BIM (building
information modeling) design);
b. doubler plate on the column web – bench-mark calculations to check the reliability/accuracy of the calculated results;
it is always advisable to check the computer results by hand calculations.
1.7 Literature
1. K. Weynand and J.-P. Jaspart, COP. The Connection Program. Program for
the design of joints according to EN 1993-1-8, www.fw-ing.de/en/products/cop.
2. EN 1090-2 (Execution of steel structures and aluminium structures. Part 2.
Technical requirements for steel structures), 2018.
c. haunche between beam and column 3. EN 1993-1-8 (Eurocode 3. Design of steel structures. Part 1-8. Design of
1.62 Connection program: overview of different options for

joints), 2005 (incl. AC, 2009).
stiffening the connection. 4. EN 1993-1-9 (Eurocode 3. Design of steel structures. Part 1-9. Fatigue), 2005
(incl. AC, 2009).
5. EN 1993-1-12 (Eurocode 3. Design of steel structures. Part 1-12. Additional rules
for the extension of EN 1993 up to steel grades S 700), 2007 (incl. AC, 2009).
6. EN 10034 (Structural steel I and H sections. Tolerances on shape and dimen
sions), 1994.
7. ISO 9223 (Corrosion of metals and alloys. Corrosivity of atmospheres. Clas
sification, determination and estimation), 2012.

2
Connections
Bolts in clearance holes

bolts in clearance holes | connections 2 | 1

Bolts in clearance holes
Steel structures are made from sections and plates. In the fabrication shop, normally welding
is used to assemble these components into columns, beams, trusses, and even portal frames.
These are then transported to the construction site and assembled into a complete steel structure.
Bolted connections are generally used for onsite assembly (fig. 2.1), either because the circum-
stances on site are not optimal for welding, or to facilitate demountability of the structure.
Bolts in normal size holes (see section 2.2.6) are most common for steel structures. Other types
of mechanical fasteners – in particular preloaded bolts, fitted bolts, injection bolts, rivets and
pins – are treated in Connections 3 (Slip-resistant connections, rivets, and pins).
Special bolt types also exist, for example bolts that are screwed directly into the tapped wall for
connections to hollow sections (fig. 2.2). Other connectors are used for joining thin gauge cold
formed sections such as purlins and profiled steel sheets used for floors, roofs, and facades.
Examples of these are self-tapping, thread-cutting, and self-drilling screws, and rivets and nails
(fig. 2.3). Such mechanical connectors for hollow sections, cold formed sections, and thin plates
are not discussed below.
2.1 Example of a bolted connection.
expander bolt (Lindapter) screw-thread tapped into wall (Flowdrill)
2.2 Special connectors for hollow sections.
self-tapping screw Ø 6,3 mm

with washer and 1 mm thick elastomer
– for steel plate to be connected to main steel structure
thread-cutting screw Ø 8 mm
with washer and with or without 1 mm thick elastomer
– for steel plate to be connected to main steel structure
self-tapping screw Ø 4,22-6,3 mm

– for steel plate to steel plate connections
– for fittings (without washer)
– for panels to be connected to main steel structure
shot fired nails

– for steel plate to steel structure
blind rivets Ø 4-6,4 mm

– for steel plate on thin walled steel lattice work
– for fittings
2.3 Connectors for cold-formed sections.
2 | connections 2 | bolts in clearance holes

This chapter considers bolts in clearance holes (normal holes, oversized holes, and slotted
holes), and the design rules used to check connections in steel structures using these bolts. The
following topics are discussed:
– categories of bolted connection;
– bolt characteristics;
– positioning of bolts;
– hole weakening;
– bolt resistance.
2.1 Categories of bolted connections

Based on the nature of the load, EN 1993-1-8, cl. 3.4 distinguishes between bolted connec
tions loaded in shear and connections loaded in tension. Figure 2.4 shows an example of both.
bolt loaded in shear
The necessary checks can be found in table 3.2 of EN 1993-1-8. In some cases, the bolts are
loaded in both tension and shear. An example is the beam-to-column connection shown in
figure 2.5.
2.1.1 Bolted connections loaded in shear

Bolts are often used with clearance holes in order to be able to insert the bolts easily: the holes
are drilled larger than the diameter of the bolt itself. Using clearance holes causes that fabri-
cation and erection deviations can be more easily accommodated. The oversize is normally
2 mm for an M20 bolt (see section 2.2.6). The clearance can cause that a connection loaded in
shear deforms. Usually, the nut is tightened by hand with a wrench or with a pneumatic tool.
The degree of tightening is normally based on experience, unless prescribed otherwise. The bolt loaded in tension
erector knows from experience not to excessively tighten the nut, to avoid the risk of fracture
of the bolt. 2.4 Two possible load cases for bolts.
tension
shear
2.5 Example of a connection with bolts loaded in

tension and shear.

160
Although the axial force in the bolt is not known for bolts tightened by hand, the level of this
pre-
loaded force can have a significant impact on the deformation of the connection. This is illustrated
bolt
120 in figure 2.6, in which the relationship between shear force F and the displacement of the
shear force F (kN)
hand-
tightened connected plates is shown for a small (tightened by hand) and large (preloaded) axial force
bolt
80 in the bolt respectively.
F F
The connected plates are clamped together by tightening the bolt, and contact pressure in
40 the shear planes results in friction. The frictional resistance – which is equal to the contact
pressure multiplied by the friction coefficient for each contact plane – must be exceeded
before slip occurs. When the bolt force is small, the plates move relative to each other
0 2 4 6 8 10 12 14
horizontal shift (mm) under a relatively small load. The connection will ‘lock’ once sufficient movement has taken
place for the bolt shaft to contact with the sides of the holes (fig. 2.7). As the load increases
2.6 Influence of the axial bolt force on the deformation further the sides of the holes are loaded in bearing and the bolt shaft is loaded in shear.
behavior in shear.
EN 1993-1-8, cl. 3.4.1, refers to such a connection as category A: loaded in bearing (‘loaded
in bearing and shear’ would be a better description).
In the case of predominantly statically loaded structures a minor displacement in the
connections is not usually of any concern. The displacements that occur due to clearance
holes can be taken into account during construction. In trusses with bolted (gusset plate)
connections, for example, it is possible to compensate for the effects of clearance by
increasing or applying a camber for that reason.
Slip in a bolted connection due to clearance is not permitted in the following situations, see
also EN 1993-1-8, cl. 2.6.
• It causes a different load distribution from that considered in the design.
• Stiffness requirements are not satisfied, either at the serviceability limit state or ultimate
limit state.
• There is a varying load such that both compression and tension occur: bolts can then
work loose and friction corrosion may occur between elements due to slip in the connection.
preload
bearing stress shear stress friction
shift
gap contact pressure gap
2.7 Categories of connections

loaded in shear . a. bolt in category A: bearing and shear b. bolt in category B and C: slip resistance due to friction

EN 1993-1-8, cl. 2.6(3), makes an exception for wind and stability bracings. For wind bracings
loaded in tension it is normal to shorten the length by two times the hole oversize dimension to
ensure that the bracing is tensioned after assembly, and to ensure that the bracing absorbs the
loading without slip in the connection. This is also referred to as ‘pulling’.
• A high fatigue strength is required: bolted connections loaded in bearing (and shear) have a
relatively low fatigue strength.
In all four of the situations noted above a different connection method must be chosen, rather than
bolts tightened by hand in clearance holes. The most common option is preloaded bolds, in which
force transfer does not take place through the bolt shaft but through friction between the contact
surfaces (fig. 2.7b). The bolts are tightened in such a way that a high axial preload occurs in the
bolt. This is called a slip-resistant connection: see figure 2.6 for the load-deformation behavior of
a preloaded connection. Depending on whether slip may occur at the serviceability limit state or
not, EN 1993-1-8, cl. 3.4.1, refers to these connections as category B (Slip-resistant at serviceability
limit state) or category C (Slip-resistant at ultimate limit state). In slip resistant connections controlled
tightening of the bolts is required, and the contact surfaces must be appropriately treated to ensure
the correct friction between the contact surfaces; see Connections 3 (Slip-resistant connections,
rivets, and pins), section 3.1.2.
Another way of preventing slip is to use bolts in fitted holes, or injection bolts. In both cases,
no special treatment of the contact surfaces is necessary (see Connections 3, section 3.2).
2.1.2 Bolted connections loaded in tension

A bolted connection loaded in tension – in which the bolts are tightened by hand or with a small
preload – will possess only a small fatigue resistance. The fatigue resistance can be improved by
preloading the bolts in combination with a proper design of the contacting parts, see Connections 3,
section 3.1.3. EN 1993-1-8, cl. 3.4.2 refers to these connections as category D (non-preloaded)
or category E (preloaded).
In the case of a bolted connection loaded in tension prying forces may occur, and these increase
the tension forces in the bolts. Prying forces are shown in figure 2.8 for a T-stub flange loaded in
tension. The flange of the T-stub deforms so that contact forces Q occur at the edges. The tension
force in the bolt becomes: Ft = T/2 + Q. In the case of a T-stub with a thick flange and relatively
T
small bolts, the bolts will yield before the bending resistance of the flange is reached. When the
bolts have sufficient deformation capacity, the prying forces decrease to zero as deformation
increases. Whether prying forces should be taken into account is dependent of the dimensions
of the bolts and the flange thickness. Prying forces in T-stub flanges are discussed in Connections 5
(Design of connections), section 5.5.2.
2.8 Forces in a T-stub flange loaded in tension. Q Ft Ft Q

2.2 Bolts
h
m
A bolt is a (mechanical) fastener comprising a head and a shaft
with a screw-thread on which a nut can be turned (fig. 2.9). Bolts
XY XY are made out of a circular steel bar which attains the required
e db diameter by pulling the bar through a hard metal mould.
8.8 8 Thereafter, the head is pressed in a few steps, after which the
SB SB
parts of the bolt are beveled and the screw-thread is applied.
s b
k l Nuts are made by first pressing a (hexagonal) nut with a hole,
after which the inner thread is tapped.
e outer diameter bolt head 8.8 (on bolt head) material grade: A bolt with a matching nut and, optionally, washers (see sec-
s wrench width • the first number is 1/100 of the design value of
l shaft length the tension strength fub (N/mm2) tion 2.2.4) is called a bolt set. Bolts and bolt sets are available
b thread length • the second number is 1/10 of the ratio between
in many shapes, sizes, and materials.
db nominal bolt diameter the tension strength and the yield strength or
h dikte sluitring 0,2% strain strength; this number shall be smaller Most bolts have an hexagonal head and nut and a metric screw-
m thickness nut than 10; other regular bolt grades are 4.6 and 10.9
8 (on nut) 1/100 of the minimal tension strength fyb
thread. The screw-thread is the axial displacement of the nut
k thickness bolt head
XY factory mark (N/mm2) with one complete turn.
SB bolt category:
– SB structural bolt (non-preloaded) A bolt is designated by a letter and a number, e.g.
The clamp length of a bolt is the total thickness of the plates,
– HV preloaded (low nut) M16. The M stands for ‘metric’ and the number is plus any washers, that can be connected by the bolt (fig. 2.10).
– HR preloaded (high nut) the nominal bolt diameter in millimeters.
The maximum clamp length is the bolt length l minus the height
of the nut and the minimum protruding length umin. According
to EN 1090-2, cl. 8.2.2 the protruding length must be at least
one full thread. The minimum clamp length depends on the
length of the bolt shaft that is unthreaded, including the thread
extension, plus the prescribed additional length with thread.
2.9 Bolt with nut and washers. According to EN 1090-2, cl. 8.2.2 the prescribed additional
length for non-preloaded bolts is one full thread, and for pre-
loaded bolts it is four full threads. In order to tighten the bolt,
both sides of the connection must be accessible for tightening
tools. Welding bolts or nuts is not generally allowed, because
the structure of the steel changes due to welding (applying heat),
and this reduces the strength. There are special hexagonal wel-
ding nuts (available up to M16) that allow welding. An alternative
is to use holes with a tapped thread in one of the elements, or
to weld a plate that has bolt holes with tapped threads. For
hollow sections, one of the methods in figure 2.2 can be used.
clamp length m umin

l 2.10 Clamp length of a bolted connection.

This paragraph deals with the following topics:
– product standards for bolts and nuts;
– screw-thread;
– bolt/nut combinations, in particular the fit of the nut on the bolt in case of galvanizing;
– washers;
– mechanical properties of bolts and nuts;
– hole clearance;
– practical bolt choices.
2.2.1 Product standards

The dimensions and qualities of bolts for use in steel structures are defined in European product
standards, see EN 1993-1-8, cl. 1.2.4, and – through EN 1993-1-8, cl. 2.1(1) – EN 1090-2, cl. 2.2.4.
These standards also describe the tests that the manufacturer must perform to prove the suitability
of the product (procedure tests). EN 1090-2, cl. 5.6.3 and 5.6.4 refers to the standards given in
table 2.11. Two systems are used for preloaded connections, differing mainly in the height of the
nut. Product standards for special types of bolts – such as bolts with countersunk heads – are not
listed in table 2.11.
The quality of a bolt – including its associated nut – mainly depends on the mechanical properties
and the quality of manufacturing. The mechanical properties of bolts are defined in EN ISO 898-1,
and the mechanical properties of nuts in EN ISO 898-2, see section 2.2.5. The quality of manu-
facturing concerns among other things the dimensional accuracy. EN ISO 4759-1 distinguishes
between three product classes for bolts: A, B, and C. Class A is the most accurate and class C the
least accurate.
category bolts and nuts washers procedure test
non-preloaded EN 15048-1 [1] n.v.t. [2] EN 15048-2
HR-system [3] EN 14399-3 EN 14399-5 or -6 [5] EN 14399-2

preloaded
HV-system [4] EN 14399-4 EN 14399-5 or -6 [5] EN 14399-2
1. Sets according to EN 14399-3 and EN 14399-4 may also be used.

2. EN 1090-2 does not specify a product standard for washers. EN 15048-1 provides a requirement for hardness
(see section 2.2.4).
3. High nuts: to be used for all preloading methods.
4. Low nuts: preferably to be used for preloading with the moment method.
5. Flat washer without beveling only under the nut according to EN 14399-5.
2.11 Designated product standards for bolt sets according to EN 1090-2.

2.2.2 Screw-thread
In general, the screw-threads of bolts are manufactured by rolling (cold formed) at normal
temperatures. These are known as rolled screw-threads. A favourable material structure is formed
in the core of the screw-thread due to (plastic) deformation. The pressure applied during rolling
also results in a compression preload in the bottom of the thread, which is beneficial for notch
sensivity. Usually, a rolled screw-thread is recognizable by the fact that the cross-section of the
bolt shaft reduces slightly at the transition to the screw-thread. Moreover, the end of the bolt
stem is usually concave.
A screw-thread can also be applied by cutting. A groove is cut into the material by a tap, which
cuts through the crystal structure. A rolled screw-thread is therefore stronger than a cut screw-thread.
EN 1993-1-8, cl. 3.6.1(3) specifies the same design values for bolts with a cut screw-thread as
those for bolts with a rolled screw-thread if the thread complies with EN 1090-2. This means
that the bolts comply with the product standard referenced from EN 1090-2, and so tests have
shown that the bolts have the required strength. When the cut screw-thread does not comply with
EN 1090-2, the design values for the shear and tension resistances must be reduced to allow for
greater deviations in shape, and increased notch sensivity, see also section 2.5.
Cut screw-threads are mainly used for special fasteners, such as anchor bolts and bolts with
large diameters. According to EN 14339-1, cl. 4.2.3, bolts in class 10.9 must always have a rolled
thread.
2.2.3 Bolt/nut combinations

outer middle The screw-thread of a bolt has fixed dimensions to enable a nut to be turned on the bolt shaft, with
line db
some tolerances. The fit is important, especially the corresponding tolerance between the thread of
inner tread
nut the bolt (outer thread) and the thread of the nut (inner thread) (fig. 2.12). The system of the tolerance
in EN ISO 4759-1 is based on the ISO fitting system. The fit is indicated by a number (for example
a 6), and the location of the tolerance area is indicated with a lower case letter for the bolt and an
flange upper case letter for the nut (for example a g or an H). The tolerance indicates the allowable deviation
middle line df outer tread relative to the core middle line. The required fit for preloaded bolts is specified in EN 14399-1,
bolt
cl. 4.4.2 as 6g for bolts and 6H for nuts (without coating).
core
middle line dk In general, bolts and bolt sets are supplied without any corrosion protection (black bolts) other than
a thin layer of oil. For outdoor applications this protection is insufficient, so bolts and bolt sets are
2.12 Geometry of a screw-thread. galvanized. The fit of the nut on the bolt then plays an important role, as explained below.
Galvanized bolts and nuts

The two commonly used methods for galvanizing bolts and nuts for steel structures are hot-dip
and electrolytic galvanizing. The zinc layer varies from 10-20 µm (0,01-0,02 mm) with electrolytic
galvanizing. These bolts are only suitable for indoor applications due to the small thickness of
the zinc layer. In addition, there is a real possibility of hydrogen embrittlement resulting from electro-
lytic galvanizing of higher bolt classes (10.9 or higher). With hot-dip galvanizing the zinc layer varies

between 60-200 µm, which is sufficient protection for outside applications and moreover the
galvanizing process does not lead to hydrogen embrittlement. The diameter of the screw-thread
of the bolt increases due to galvanizing and the screw-thread of the nut reduces. Therefore, the
inner thread of the nut or the outer thread of the bolt must be adjusted to compensate for the
zinc layer. Otherwise it is not possible to turn a hot-dip galvanized nut onto a hot-dip galvanized
bolt. The two main types of galvanized bolt sets are:
• ISO-fitting bolts and nuts. The outer thread of the bolt is rolled thinner before galvanizing,
and the inner thread is precisely tapped after galvanizing, according to the ISO-value.
• Oversize bolts and nuts. The bolts are normal ISO-fitting black bolts that become oversized
after galvanizing. The nuts are galvanized after they are pressed, after which the inner thread is
tapped, taking into account the maximum thickness of the zinc layer on the outer thread of the
bolt. The inner thread must be tapped larger than that of an ISO-fitted nut: thus the nuts are also
oversized. This oversize is not standardized but is typically 0,2-0,3 mm, depending on the bolt
diameter (the thicker the bolt, the thicker the zinc layer on the bolt) and the precision of tapping
the inner thread by the manufacturer. This type of galvanized bolts is most common in steel
structures.
The tolerance between the outer and inner threads is basically the same for both types of bolt
sets. It is important that hot-dip galvanized bolts are delivered complete with nuts (bolt sets) by
the manufacturer. This avoids separately delivered nuts not fitting the galvanized bolts. The design
values of the resistances of galvanized bolt/nut combinations derived from tests according to
product standards, are the same as those of equivalent black bolts.
2.2.4 Washers
In general, the application of washers is not mandatory for non-preloaded bolts in normal round
holes according to EN 1090-2, cl. 8.2.4. In a hot-dip galvanized or painted steel structure,
washers are often applied to prevent damage of the corrosion protection of the connected
elements during tightening of the nut or bolt head.
For non-preloaded connections EN 1090-2 does not specify a product standard for washers (see
table 2.11, note 2). According to EN 15048-1, cl. 4.4.2.4 the hardness must be at least 100 HV
(Vickers hardness). Sometimes a plate ring is used instead of a washer. Often, this is a square or
rectangular piece of steel plate with a drilled or punched hole. The dimensions of plate rings are
not standardized.
According to EN 1993-1-8, cl. 3.6.1(10) a washer must be used under the bolt head and nut in an 2.13 Asymmetric overlap connection with one
bolt row: washers are mandatory.
asymmetric overlap connection with a single bolt row (fig. 2.13). In such connections, the washers
for bolt classes 8.8 and 10.9 must have a hardness between 300 and 370 HV.
Washers must be applied on both sides of connections with oversize holes or long slotted holes,
and they may not be thinner than 4 mm according to EN 1090-2, cl. 8.2.4.
When the clamp length of a bolt set is too long for the total thickness of the parts to be connected,
the clamp length may be adjusted with an additional spacer or at most three washers with a total

INP UNP
tapered
plate
slope 14% slope 8%
2.14 Tapered plate for INP and UNP sections.
thickness of maximum 12 mm according to EN 1090-2, cl. 8.2.4. These washers must be applied
to the side where tightening does not take place.
According to EN 1090-2, cl. 8.2.4, tapered plates (washers) must be used when the surface of the
connected element makes an angle with the surface perpendicular to the bolt shaft larger than
1/20 (3˚) for d ≤ 20 mm and 1/30 (2˚) for d > 20 mm (d is the bolt diameter). This is the case, for
example, when connecting (older) INP or UNP sections (fig. 2.14).
To achieve correct tightening of preloaded bolts it is important that the friction between the nut
or the bolt head – depending which of the two is tightened – and the underlying material does
not show too much variation. Therefore, a hardened steel washer is used as part of the system (see
table 2.11). With preloaded 10.9 bolts a washer must be used under both the bolt head and the
nut, to prevent reduction of the preload as a result of deformations caused by high contact forces.
2.2.5 Mechanical characteristics

This paragraph covers the most important mechanical properties of bolt and nuts.
Bolts
When a bolt breaks, fracture occurs in the threaded section. The actual cross-section in the
threaded section is somewhat elliptical and not easy to determine (fig. 2.15). Therefore, for
core diameter
convenience the tension stress area As is defined as the cross-section of a smooth cylinder –
flange diameter
which has the same failure strength as the part of the bolt with the screw-thread. The value for
2.15 Actual cross-section of the threaded part. As is defined in EN ISO 898-1:

(2.1)
Where (see also figure 2.12):

db nominal bolt diameter;
dk core diameter.
In table 2.16, the actual shaft cross-section A and the tension stress area Asare given for the most
commonly used bolts. The letter M in the denotation stands for metric screw-thread and the
number is the diameter of the shaft in millimeters.

bolt
a
M12 M16 M20 M22 M24 M27 M30
shaft cross-section A 113 201 314 380 452 573 707 2.16 Cross-sections (mm2) for bolts M12 up
tension stress area A s 84,3 157 245 303 353 459 561 to M30.
bolt class
4.6 8.8 10.9
tension strength fub (N/mm2) 400 800 1000
2.17 Yield strength and tension strength of
yield strength f yb (N/mm2) 240 640 900 the three most commonly used bolt classes.
b
bolt class
EN 1993-1-8 4.6 4.8 5.6 5.8 6.8 8.8 10.9 2.18 Bolt classes for use in steel structures.
Bolts are divided into bolt classes. The format has two numbers, separated by a dot (table 2.17).
From these numbers, the guaranteed minimum tension strength fub and the guaranteed minimum
yield strength (0,2% strain limit) fyb can be derived. For example, for bolt class 8.8:
– tension strength: fub = first number x 100 = 8·100 = 800 N/mm2;
– yield strength: fyb = first number x second number x 10 = 8·8·10 = 640 N/mm2.
EN ISO 898 and the product standards listed in table 2.11 specify additional mechanical properties
such as the minimum strain before failure, and the minimum impact energy (Charpy value). The
minimum strain before failure is important for the deformation capacity (ductility) of the bolt, and
the minimum impact energy for the danger of brittle failure.
EN ISO 898 describes different bolt classes, of which seven are identified as suitable for use in steel
structures in EN 1993-1-8, cl. 3.1.1 (table 2.18). The bolt classes 4.6 and 4.8 have the same tension
strength and therefore the same resistance. However class 4.6 has the advantage of a larger margin
between fub and fyb and has a greater deformation capacity. Therefore class 4.6 should be preferred
above class 4.8. The same applies for class 5.6 and class 5.8 respectively. According to the product
standards, the bolt class and factory mark must be shown on the bolt head.
Nuts
The tension strength of nuts is not specified directly in the product standards. The nut must be
able to withstand a specified test load without failure of the screw-thread. The this test load is
the product of the tension stress area As and the minimum tension strength of the bolt fub. The
strength class of nuts is indicated by a number. This number multiplied by one hundred gives
the minimum tension strength (in N/mm2) of the bolt to which the nut is matched. The class and
factory mark must be shown on the nut.

2.19 Nominal clearance (mm) according to
bolt
EN 1090-2. type of hole
M12 M14-M22 M24 ≤ M27
normal round holes 1 2 2 3
oversize round holes 3 4 6 8
short slotted holes (on the length) 4 6 8 10
long slotted hole (on the length) 1,5d 1,5d 1,5d 1,5d
2.2.6 Hole clearance

Bolts in clearance holes – mentioned in EN 1993-1-8, cl. 3.6.1(4) – are most suitable for connections
on site. These holes are drilled a little larger than the diameter of the bolt. A normal oversize for
an M20 bolt, for example, is 2 mm.
Sometimes the erected deviations of the connecting elements are too large to allow execution
with normal bolt hole sizes. In this case, holes with a larger clearing may be used. They may be
oversize circular holes, or slotted holes with a larger gap in one direction (longitudinal direction)
and a normal gap in the other direction (transverse direction). The design must take into account
the effects of these larger holes. Table 11 of EN 1090-2 specifies the nominal clearances, depending
of the bolt diameter and the type of hole, see table 2.19.
In table 11 of EN 1090-2 the nominal clearance for a bolt M14 is 2 mm. But under the conditions
as given in EN 1993-1-8, cl. 3.6.1(5) bolts M12 and M14 may be used in holes with an oversize of
2 mm instead of 1 mm, if the shear resistance of the bolt group is larger than the bearing resistance.
2.2.7 Practical bolt choice

Although EN 1993-1-8 allows seven bolt classes, see table 2.18, it is wise to follow usual practice
and limit the bolt choice for normal structural work to class 8.8. The use of 8.8 bolts, instead of a
lower bolt class, reduces the number of bolts needed and thereby results in more compact joints
with less holes to be drilled. The price/performance ratio is optimal with this choice. Furthermore,
standardizing to one bolt class (8.8) has the following advantages:
– less stock: remember that in addition to variations in bolt diameter, different bolt lengths are
also required;
– more simple bolt lists (lists with all the bolts required for a structure);
thickness of the plate, – no risk of using bolts in a wrong class, which increases safety.
flange or web (mm) bolt
5-8 M12
Sometimes 10.9 bolts are preferred in preloaded connections, due to the large preload. This has
8-12 M16
an additional advantage; during the assembly of the steel structure the erectors can see directly
12-22 M20 if a given bolt must be preloaded or not. Also, the number of diameters and shaft lengths of
>22 M24 bolts should be limited for practical reasons. It is recommended to vary the bolt diameter as a
2.20 Practical bolt choice for connections loaded function of the thickness of the elements to be connected (table 2.20 to 2.22). For angles with
in shear, as a function of plate thickness. flanges wider than 100 mm, it is recommended to use staggered bolt positions.

edge distance
flance spacing
width w w1 p1 w
(mm) bolt (mm) (mm) (mm)
p1 p1
40 M12 22 – 40
45 M12 25 – 40
50 M16 28 – 55 d0
55 M16 30 – 55 flange width ≤ 100 mm

60 M16 30 – 55
65 M20 32,5 – 70
70 M20 35 – 70
75 M20 37,5 – 70 w1
w
80 M20 40 – 70
p1 p1
90 M20 45 – 70
100 M24 55 – 80
110 M24 50 70 80
d0
120 M24 50 80 80
flange width > 100 mm
130 M24 50 90 80
140 M24 55 100 80
150 M24 55 110 80
160 M24 60 115 80
180 M24 60 115 80
2.21 Practical bolt choices, edge distances, and
200 M24 65 150 80 spacings for angles.
section size
section
80 100 120 140 160 180 200 220 240 260 270 280 300
bolt M6 M8 M10 M12 M12 M16 M16 M16 M20 – M20 – M24
IPE w 28 34 40 44 48 54 60 60 70 – 76 – 80
hw 58 74 92 112 126 146 158 176 190 – 218 – 248
bolt – M12 M20 M20 M20 M20 M24 M24 M24 M24 – M24 M24
HEA/HEB w – 50 65 65 70 70 80 80 90 90 – 90 100
hw – 56 74 91 104 121 134 152 164 176 – 196 208
bolt – M16 M20 M20 M24 M24 M24 M24 M24 M24 – M24 M24
HEM w – 60 70 70 90 90 90 90 110 110 – 110 110
hw – 56 74 91 104 121 134 152 164 176 – 196 208
2.22 Practical bolt choice and edge distance (mm) for flanges of commonly used I-shaped rolled sections. hw is the nett web
height (between root radii).

e2
2.3 Positioning of bolts
force
direction
p2 (compression
or tension)
EN 1993-1-8, cl. 3.5 specifies minimum and maximum values for the spacing between
e2 bolts and the distance of a bolt to the edge of a structural element – both perpendicular
e1 p1 p1 e1 to and parallel with the direction of the force. Figure 2.23 gives the symbols that are
e 1: end distance p1 and e1: parallel to force direction
used for these distances for a bolt group and for slotted holes. Table 2.24 gives an
e2: edge distance p2 and e2: perpendicular to force direction overview of the requirements in EN 1993-1-8.
p1, p2: spacing
The minimum bolt distances arise from:
intermediate distances and spacings
– the required space for the bolt head, the nut, and the washer if applicable;
p1,0 – the required space for the tightening tool;
– the required material around the bolt for transferring forces to the connected structural
outer row
parts.
inner row
When the minimum bolt spacings and distances are used in a connection, the bearing
resistance must generally be reduced (see section 2.5.1). Therefore, it may be wise to adopt
p1,i
larger bolt spacings and distances. The minimum bolt spacings may be appropriate
p1,0 ≤ 14t and ≤ 200 mm
p1,i ≤ 14t and ≤ 400 mm
staggered rows in elements loaded in tension

maximum
structural steel
distance minimum
weathering steel
corrosive non-corrosive
d0 environment environment
e3 maximum of
end distance e1 1,2d 0 4t + 40 mm –
8t or 125 mm
e4 1
d
2 0 maximum of
edge distance e2 1,2d 0 4t + 40 mm –
end and edge distances for slotted holes 8t or 125 mm
distance e3 for
1,5d 0 – – –
2.23 Symbols for a particular bolt pattern. slotted hole
distance e4 for
1,5d 0 – – –
slotted hole
minimum of minimum of minimum of

spacing p1 2,2d 0
14t or 200 mm 14t or 200 mm [1] 14tmin or 200 mm [2]
minimum of
spacing outer row p1,0 2,2d 0 – –
14t or 200 mm
minimum of
spacing inner row p1,i 2,2d 0 – –
28t or 400 mm
minimum of minimum of minimum of
spacing p2 2,4d 0
14t or 200 mm 14t or 200 mm [1] 14tmin or 175 mm [2]
1. Note 1 under table 3.3 of EN 1993-1-8 indicates that no limit applies for elements in a non-corrosive
environment, except when it is necessary for elements loaded in compression to prevent local buckling.
2.24 Minimum and maximum spacing and end and edge Note 3 under table 3.3 sets the rules for calculating the maximum distance associated with local buckling.
distances ( is the thickness of the thinnest element). 2. tmin = thickness of the thinnest plate.

when there is little space available, or to prevent large gusset minimum value according to
pratical value
plates (lots of welding, thus expensive). Table 2.25 gives an bolt EN 1993-1-8
overview of minimum and practical bolt distances. e1 e2 p1 p2 e1 e2 p1 p2
Maximum bolt spacing distances refer to connections that M12 16 16 29 31 25 25 40 40
transfer no, or minor, forces, and are only used for elements M16 22 22 40 43 35 30 55 55
that are exposed to the outside environment or other corrosive M20 26 26 50 53 45 40 65 70
influences. Local buckling of plates between the connectors
M24 31 31 57 62 55 50 80 80
can be decisive for the maximum bolt spacing in structural
M30 40 40 73 79 65 60 100 100
elements loaded in compression. According to note 2 of
table 3.3 in EN 1993-1-8, the resistance to local buckling of a
plate loaded in compression must be calculated according to EN 1993-1-8, cl. 6.3.1 with a buck- 2.25 Minimum and practical bolt spacings and
distances in mm.
ling length equal to 0,6p1. It is not necessary to check for local buckling between the connectors
when p1/t < 9ε with ε = 235/ fy .
2.4 Resistance of elements with holes

In order to be able to use mechanical fasteners – such as bolts – holes must be provided in the
elements to be connected. However, the presence of a hole reduces the gross cross-sectional area
of the original section. The following subjects related to the effects of holes are discussed below.
– net cross-section;
– tension resistance of a plate with holes;
– tension resistance of an angle with holes;
– tearing out of a bolt group.
2.4.1 Net cross-section

For the design of sections or plates loaded in tension, the influence of holes must be taken
into account. The critical cross-section is referred to as the net cross-section Anet. According
to EN 1993-1-1, cl. 6.2.2.2 the net cross-section Anet is determined by reducing the gross
cross-section by the area of all holes and other openings in a failure line. For the plate shown
in figure 2.26, the net cross-section equals:
Anet = bt – d0t (2.2)

Anet A
NEd NEd
Where: b
b width of the plate;
t thickness of the plate;
d0 diameter of the hole. t
2.26 Tension bar with bolt holes. d0

When the bolt holes are staggered, like in figure 2.27, a zigzag cut (failure plane 1) can be
NEd NEd
critical in addition to a cross-section perpendicular to the bar axis (failure plane 2). For a
1 2 p2
zigzag cut, the hole reduction is according to EN 1993-1-1, cl. 6.2.2.2:
p1 p1 (2.3)
2.27 Tension bar with staggered bolt holes.

Where:
n number of holes in the zigzag cut;
p1 spacing between two adjacent holes, parallel to the bar axis (in EN 1993-1-1 denoted
with s);
p2 spacing between the same two adjacent holes, perpendicular to the bar axis (in EN
1993-1-1 denoted with p).
Figure 2.28 shows some practical examples of staggered bolt configurations. The practical
edge distances, hole diameters, and minimum spacings are given in table 2.21 and 2.22. In
a section with bolt holes in more than one plane, such as the angle of figure 2.29 in example
2.1, distances are measured along the middle of the section. According to EN 1993-1-1,
cl. 6.2.4(3) and 6.2.5(6) hole weakening does not have to be taken into account when the
cross-section is loaded in compression and all holes contain a bolt.
Example 2.1 shows that material can be saved in sections loaded in tension by staggering
of the bolts. A practical additional advantage of staggering holes is that the bolts and nuts
are easier to access with wrenches or other tightening tools.
rolled section HEA 280
double angle 100x100x10
2.28 Examples of a staggered bolt placement.

Example 2.1
• Given. An angle 100x100x10 mm with a bolt hole of diameter d0 = 26 mm in both flanges.

The spacing p1 = 70 mm (fig. 2.29).
• Question. Determine the critical net cross-section.
• Answer. There are two different possible failure cross sections: section 1 and section 2.
The net area of these sections is defined according to equation (2.2) and (2.3) respectively:
Section 2 is critical. The net area of section 2 equals that of section 1 when the spacing p1 is
chosen such that p12 / 4p 2 = d0 . This is the case for p1 = 102 mm. When the bolts are placed in a
non-staggered position, the net cross-section equals:
angle
d0 = 26 100x100x10
t
45
d0 b
55 t
p1 = 70
b–t
section 1 section 2
e2 = 45
p2 = 100 2b – t
e2 = 45 2.29 Angle with a bolt hole in

both flanges.

2.4.2 Tension resistance of a plate with holes
The design value of the tension resistance Nt,Rd of a plate with holes, according to EN 1993-1-1,
cl. 6.2.3 depends on whether the gross cross-section yields or the net cross-section fails:
Afy
yielding of gross cross-section Nt,Rd = Npl,Rd = γ (2.4)
M0

0,9A net fu
failure of net cross-section Nt,Rd = Nu,Rd = γM2 (2.5)
Where:
Npl,Rd design value of the normal force for yielding of the gross cross-section;
fy yield strength;
fu ultimate tension strength;
γM0, γM2 partial factors for the resistance of the cross-section: γM0 = 1,0 and γM2 = 1,25.
When there is a requirement for deformation capacity of the plate, the cross-section ratio Anet/A
should be such that yielding of the gross cross-section occurs before failure of the net cross-section.
The length of the stressed area in the net section is too small to allow a large elongation of the
plate. For example, according to EN 1998 requirements for deformation capacity are relevant for
steel structures in earthquake-sensitive areas. EN 1993-1-1, cl . 6.2.3(3) states that in such a situation
the design value of the resistance of the net cross-section Nu,Rd may not be smaller than the resi-
stance in yielding of the gross cross-section Npl,Rd . This is so-called capacity design:
Afy 0,9A net fu A net fu γM0

Npl,Rd = 0,9 1,0 (2.6)
γM0 γM2 A f γM2
y
Table 2.30 provides the minimum value of Anet/A for the four steel grades S235 to S460 that
follow from the requirement of equation (2.6) for a capacity design. The table shows that only for
the lower steel grades limited reduction of the cross-section by holing is allowed.
minimum value Anet /A Even when capacity design is not required, it may still be good to design connections for sufficient
steel grade
capacity deformation deformation capacity, for example for statically indeterminate structures. For these situations, a less
design capacity strict requirement than for capacity design can be used to derive the minimum values of Anet /A by
S235 0,91 0,73 assuming characteristic values instead of design values. This is based on the assumption that there
is a correlation between the yield strength and tension strength. The following then applies:
S355 0,97 0,77
S420 1 0,90 A net fu

Afy 0,9A net fu 0,9 1,0
A fy (2.7)
S460 1 0,95
The minimum values of Anet/A that follow are also given in table 2.30. The requirement of equa-
2.30 Minimum values of Anet/A to assure
deformation capacity. tion (2.7) is not included in EN 1993-1-1.

2.4.3 Tension resistance of an angle with holes median line
Nu,Rd
For angles with connections to both flanges the design value of the
tension resistance depends on the sizes of the gross and net cross- e2
sections, similar as for plates. In general, angles are connected with
only one flange (fig. 2.31). This results in a complex stress distribution hole line
e1 p1 p1 p1
in the connection, because the tension force in the total cross-section
must transfer into the connected flange alone (leg). Tests have shown
2.31 Angles connected with one flange.
that in this case a part of the section can tear out, well before the net
cross-section yields (fig. 2.32).
The design value of the tension resistance Nu,Rd depends on the
number of bolts, the spacing p1, and the edge distance e2. For a con-
nection with one bolt the following applies, according to EN 1993-1-8,
cl. 3.10.3:
(2.8)

For connections of angles with two or more bolts in one flange, a

reduction factor β is introduced. This reduction must be applied to
the design axial resistance of the net cross-section. The value of β
depends on the number of bolts and their spacing p1. The design
resistance Nu,Rd follows from:
2.32 Tearing out of a
β A f single-sided connected angle.
two bolts Nu,Rd = 2 γ net u (2.9)
M2

β3 A net fu
three or more bolts Nu,Rd = (2.10)
γM2
Equation (2.9) and (2.10) apply only when the bolts are aligned. For
other bolt patterns, EN 1993-1-8 does not provide rules. The value β
of β2 (two bolts) and β3 (three or more bolts) in relation to p1/d0 is p
β3 = 0,3 + 0,08 1
d0
given in figure 2.33. For unsymmetrical angles that are connected
0,7
via the narrower flange a value of Anet must be determined by assum
ing an equivalent symmetric (equal leg) angle. The dimensions of 0,5
the symmetric angle should be based on the narrower flange of the 0,4
asymmetrical angle. p
β2 = 0,1 + 0,12 1
d0
p1
2.33 Reduction factors β2 en β3. 2,2 2,5 5,0 d0

15 mm
3 bolts M20 equivalent
(d0 = 22 mm) angle
32,5
32,5
2.34 Connection of two asymmetrical angles 50 70 70 double angle

with a gusset plate. 100x65x9
Example 2.2
• Given. A double angle section 100x65x9 in steel grade S235 with three M20 bolts connected
to a gusset plate (fig. 2.34).
• Question. Determine the design value of the tension resistance Nu,Rd.
• Answer. The connected angles are asymmetrical. Therefore, an equivalent symmetric angle
65x65x9 must be assumed. The net cross-section of the two equivalent angles is A net = 2(A – d0t) =
2·(1098 – 22·9) = 1800 mm2. The reduction factor β = β 3 follows from figure 2.33 with p1/d0 =
70/22 = 3,18:
p1
β 3 = 0,3 + 0,08 = 0,3 + 0,08·3,18 = 0,55
d0
For three bolts, the design value of the tension resistance Nu,Rd follows from equation (2.10):
β A f 0,55·1800·360·10 –3
Nu,Rd = 3 net u = = 285 kN
γM2 1, 25
NEd
Also, the bolts must be checked for shear resistance and the angle and the gusset plate for
bearing resistance. These checks have been omitted here.
2.4.4 Block tearing

In addition to checking the tension resistance of the net-section and the gross-section a check
must also be made on the tearing out of a bolt group (block tearing) (fig. 2.35). Block tearing is
characterized by the combination of failure in shear, along the shear line of a row of holes in the
NEd force direction, accompanied by tension failure across a row of holes perpendicular to the force
2.35 Block tearing. direction. EN 1993-1-8, cl. 3.10.2 provide checks for this situation.

section II
section II
section I
section I
notch VEd VEd
a1
VRd ℓv
VRd
IIa
a2
IIb
e2 e2 2.36 Two notched beams with the tops of the upper

flanges at the same height.
For a symmetric bolt group subjected to an axial load – such as the one shown in figure 2.35 –
the design block tearing resistance Veff,1,Rd follows from:
fy A nv
f A 3
Veff,1,Rd = uγ nt + γ (2.11)
M2 M0
Where:
Ant net cross-section loaded in tension;
Anv net cross-section loaded in shear.
The first term of equation (2.11) is the same as equation (2.5) but without the factor 0,9. When
the contribution of the second term is small, the check must be carried out using both equation
(2.11) and (2.5). When the tear pattern is non-symmetrical, or when the bolt group is loaded with
an eccentric load, the part of the cross-section loaded in tension must be halved. The design
value for the tearing resistance Veff,2,Rd equals:
fy A nv
0, 5fu A nt 3
Veff,2,Rd = γM2 + γ (2.12)
M0
This situation occurs, for example, in the case of notched beam to beam connections. Removing
(or notching) part of the upper flange is common practice when connecting two beams, for which
the upper surfaces of the top flanges are required to be at the same level (fig. 2.36). When the
connecting beam has the same depth, or is even deeper, than the primary beam, then the upper
as well as the lower flange will need to be notched (fig. 2.36, right-hand side). Notching weakens
the cross-section where section I has to be checked in shear and bearing, and section II in block
tearing.

Example 2.3
• Given. The web of a notched HEA 200 beam in steel grade S235 is connected with five M24
bolts (with d0 = 26 mm) in two variations (a and b) for the positioning of the bolt group (fig. 2.37).
• Question. Determine for each variant the design value for the block tearing resistance Veff,2,Rd.
• Answer. In both variants the tearing pattern is un-symmetrical. In addition, the side of the tear
pattern loaded in tension is eccentrically loaded. The design value of the block tearing resistance
Veff,2,Rd follows from equation (2.12). For variant a:
For variant b:
This illustrates that the block tearing resistance strongly depends on the position of the bolts.
2.37 Notched beam connections with two

variants of bolt placements.

2.5 Resistance of bolts
The resistance of individual, non-preloaded bolts depends on the way the bolt is loaded, namely:
– bolts loaded in shear (section 2.5.1);
– bolts loaded in tension (section 2.5.2);
– bolts loaded in shear and tension (section 2.5.3).
Design checks are given in table 3.2 of EN 1993-1-8. For bolts with a countersunk head some of
the checks are modified (section 2.5.4).
2.5.1 Bolts loaded in shear

A connection with non-preloaded bolts loaded in shear – see section 2.1.1 and figure 2.6 – will
settle until the sides of the holes press against the bolt shafts. This settlement is reflected in the
load/deformation behavior of the connection. The force distribution is schematically shown in
figure 2.7a (category A). Different failure modes can determine the resistance of a connection,
depending on the bolt diameter, the plate thickness, and the material characteristics of the bolts
and plates (fig. 2.38 and 2.39).
In principle, each possible failure mode must be checked individually; the failure mode with the
lowest resistance is critical. However, this does not mean that it does not matter which failure mode
governs. Shearing of a bolt (option 1 in fig. 2.39) occurs with relatively small deformation, while failure
in bearing of a plate provides a larger deformation capacity. A design in which bearing is critical is
preferable. EN 1993-1-8, cl. 3.6.1 combines tearing of a plate (option 2 in fig. 2.39) and bearing of
a bolt hole (option 3 in fig. 2.39) in one failure mode. This combined failure mode is referred to as
bearing resistance.
For non-preloaded bolts loaded in shear the following resistances are therefore important:
– shear resistance (shearing of a bolt);
– bearing resistance (tearing of a plate or bearing of a bolt hole). 2.38 Different failure modes of failed test pieces.
e2 e2 small e2 large
1 2 3 4
shearing tearing of a plate bearing of rupture of net 2.39 Possible failure modes of connections
of a bolt a bolt hole cross-section loaded in shear.

In addition, the tension resistance of the net cross-section of the connecting plate or plates must
be checked (option 4 in fig. 2.39). The design values of the shear resistance and bearing resistance
in EN 1993-1-8 are determined based on a statistical evaluation of the results of a large number
of tests performed by different laboratories. The evaluation process used was in accordance with
Annex D of EN 1990.
Shear resistance
The number of shear planes is important in the case of a bolted connection loaded in shear.
A connection may have a single shear plane or double shear planes (fig. 2.40).
The shear plane can pass through the shaft or the thread, depending on the length of the
screw-thread (fig. 2.41), and this choice influences the shear resistance. For practical reasons, it is
preferable to assume that the shear plane runs through the thread, unless the length of the shaft
and the length of the thread of the bolts to be used are specified and controlled.
When the shear plane passes through the shaft (without thread) of the bolt, the shear resistance
per shear plane Fv,Rd according to table 3.4 of EN 1993-1-8 is:
0,6fub A
Fv,Rd = (2.13)
γM2
F
t1
F
t1
single shear
1
t1 F
2
F
t2
1
t1 F
2
2.40 Single and double shear connections. double shear

d
shaft
shaft
As A
> 0,1d
2.41 Possible locations of the shear plane. shear plane through thread shear plane through shaft

When the shear plane passes through the thread of the bolt, then the shear resistance per shear
plane is:
α v fub A s
Fv,Rd = γM2 (2.14)
With:
αv = 0,6 for bolt classes 4.6, 5.6, and 8.8;
αv = 0,5 for bolt classes 6.8 and 10.9.
According to EN 1993-1-8, cl. 3.6.1(4) equation (2.13) and (2.14) apply only for bolts in normal
circular holes. For bolts in oversized holes or in slotted holes, no rules are given. In this case, it is
good practice to design the connection so that bearing of all bolts in the group is critical.
The smaller value of αv for bolt classes 6.8 and 10.9 follows from a statistical evaluation, and is
caused by the reduced toughness and greater notch sensitivity of these bolts.
For shear through the thread, it is important whether the screw-thread is in accordance with
EN 1090-2. When the thread does not meet the relevant requirements, then the resistance of
each shear plane Fv,Rd in equation (2.14) must be multiplied by a reduction factor 0,85 according
to EN 1993-1-8, cl. 3.6.1(3).
Formally, on the basis of table 3.4 of EN 1993-1-8, one may calculate resistance using the shaft cross-
section – see equation (2.13) – when the beginning of the thread lays in an outer plate. The tolerances
of the bolt dimensions and plate thickness must be taken into account. However, based on tests,
equation (2.14) is advised to be applied when the thread ends a distance smaller than 0,1d from the
shear plane (fig. 2.41). The check for shear resistance according to table 3.2 of EN 1993-1-8 is:
Fv,Ed ≤ Fv,Rd (2.15)
Where Fv,Ed is the design value of the shear force per bolt per shear plane for the ultimate limit
state.
Bearing resistance
Loading tests with bolted connections have shown that high values of the mean bearing stress
in plate material can occur. However, large deformations (ovalization) occur before failure in
bearing, which results in other possible failure modes. It is therefore necessary to limit the mean
bearing value. When in the bolted connection the end distance, the edge distance or the spacing
is too small, shearing of the plate material will be decisive (see also fig. 2.39).
EN 1993-1-8, table 3.4 combines these two failure modes (bearing and shear) into one equation
for the bearing resistance Fb,Rd:
k1α b fud t
Fb,Rd = γM2 (2.16)

parallel to force direction
inner inner end bottom plate

bolts bolts bolts
end inner inner upper plate
bolts bolts bolts
perpendicular to
force direction
e2
edge bolts
p2
inner bolts
force
p2 direction
edge bolts
e2
e1 p1 p1 e1
parallel to force direction
inner end bottom plate

bolts bolts
end inners upper plate
bolts bolts
perpendicular to
force direction
e2
edge bolts
p2
force direction
edge bolts
e2
2.42 Indication of the bolts, end and edge distances, e1 p1 e1
and spacings.
The factor k1 depends on the edge distance and spacing perpendicular to the force direction
(e2 and p2) and the factor αb on the end distance and spacing parallel to the force direction (e1 and p1)
(fig. 2.42). The factor k1 is given by:
2,8e 2 1,4p 2
edge bolts k1 = smallest value of: – 1,7 or – 1,7 or 2,5 (2.17)
d0 d0
k1
2,8e2
k1 = – 1,7 1,4p 2
d0
inner bolts k1 = smallest value of: – 1,7 or 2,5 (2.18)
2,50
d0
1,66 From equation (2.17) and (2.18) follows that for e2 > 1,5d0 and p2 > 3d0 the value of k1 = 2,5,
independent of the bolt placement. Therefore, these limiting values for e2 and p2 should be taken
e2
as normal minimum values, see also section 2.3. Figure 2.43 shows the relationship between the
1,2 1,5 d0 factor k1 and e2/d0.
The factor αb is given by:
compact normal
fub
2.43 Relationship between k1 en e2 /d0. α b = smallest value of: α d or or 1,0 (2.19)
fu

The ratio between the tension strength fub of the bolt material and fu of the plate material is
αd
often larger than one, and therefore not decisive for the value of factor αd. The factor αd is αd =
e1
3d0
given by:
1,0
e1
end bolts αd = (2.20)
3d0
0,4
p1 1 shear ovalization
inner bolts αd = – (2.21) e1
3d0 4 1 2 3 d0
1,2
The distances e1 and p1 determine whether hole ovalization or shearing of the plate material will 2.44 Relationship between αd en e1/d0.
be critical. This is illustrated in figure 2.44 for e1. The design value of the bearing resistance is
maximal for e1 = 3d0 and e2 = 1,5d0 for end bolts, and for p1 = 3,75d0 and p2 = 3d0 for inner
bolts.
The check for the bearing resistance is according to table 3.2 of EN 1993-1-8:
Fv,Ed ≤ Fb,Rd (2.22)
The direction of the force on a bolt may not be parallel to the edge of the plate. According to note
3 in table 3.4 of EN 1993-1-8, the bearing resistance may then be checked separately for the force
components that are parallel and perpendicular to the plate end.
EN 1993-1-8, cl. 3.7 states that the resistance of a bolt group may be taken as the sum of the
design values of the bearing resistances of the individual bolts, provided that the shear resistance
Fv,Rd is satisfied for each bolt. If this is not the case – according to EN 1993-1-8, cl. 3.7 – the
smallest design value of the bolt group must be taken into account for all the bolts of the group.
Table 2.45 gives an overview of the design values for shear resistance Fv,Rd and bearing resistance
Fb,Rd for class 8.8 bolts. The values for Fv,Rd apply for grade S235 plate material and for the
minimum edge distances and spacings that are generally used in practice (practical values in
table 2.25).
M12 M16 M20 M22 M24 M27 M30
shear resistance Fv,Rd (kN) shaft 43,3 77,1 121 146 174 220 271
(single shear) thread 32,4 60,3 94,1 117,1 136 176 215
bearing resistance Fb,Rd (kN) 5,8t 7,7t 9,6t 10,6t 11,5t 12,9t 14,4t
single shear t≤5 t≤7 t≤9 t ≤ 10 t ≤ 11 t ≤ 13 t ≤ 14

bearing decisive for
double shear t ≤ 11 t ≤ 15 t ≤ 19 t ≤ 22 t ≤ 23 t ≤ 27 t ≤ 29
2.45 Design values for the shear resistance and bearing resistance of 8.8 bolts in S235 plate material
(t is plate thickness in mm).

2.5.2 Bolts loaded in tension
For a non-preloaded bolt loaded in tension, the following resistances are important:
– tension resistance (failure of the bolt in tension);
– punching shear resistance (pulling of the bolt head or nut through the plate).
Tension resistance
The design value of the tension resistance Ft,Rd of a bolt depends on the tension stress area As
and the tension strength of the bolt material fub, see table 2.16 and 2.17. The equation for the
design value is given in table 3.4 of EN 1993-1-8:
k 2 fub A s
Ft,Rd = (2.23)
γM2
In equation (2.23), k2 is a calibration factor determined on the basis of a statistical evaluation of

similar tests, as discussed in section 2.5.1 for the shear and bearing resistances. For normal bolts
with threads meeting the requirements of EN 1090-2: k2 = 0,9. When the thread does not comply
with EN 1090-2, k2 must be multiplied by a reduction factor 0,85 (see also section 2.2.2).
The check of the tension resistance is according to table 3.2 of EN 1993-1-8:
Ft,Ed ≤ Ft,Rd (2.24)
Table 2.46 shows the tension resistance for practical bolt classes according to equation (2.23).
Punching shear resistance

According to table 3.4 of EN 1993-1-8, the punching shear resistance of a plate Bp,Rd is:
0,6 πdm t p fu
Bp,Rd = (2.25)
γM2
Equation (2.25) assesses shear under the bolt head or nut of a cylindrical part of the plate material
with a thickness tp. The diameter dm of the cylinder is taken as the mean of the across points and
across flats dimensions (e and s in fig. 2.9) of the head or the nut , whichever is smaller.
The check for punching shear resistance is according to table 3.2 of EN 1993-1-8:
Fv,Ed ≤ Bp,Rd (2.26)

bolt
bolt
class M12 M16 M20 M22 M24 M27 M30
4.6 24,2 45,2 70,6 87,5 102 132 161
8.8 48,6 90,4 141 174 203 264 323

2.46 Design values of the tension resistance
10.9 60,4 113 176 218 254 330 404 Ft,Rd (kN).

2.5.3 Bolts loaded in shear and tension bolts loaded in
shear and tension
Connections with bolts loaded in pure tension are not very common. Usually, a combination of
shear and tension occurs (fig. 2.47). In this case the following experimentally determined condition
should be satisfied, according to table 3.4 of EN 1993-1-8:
Fv,Ed Ft,Ed bolts loaded

+ 1,0 (2.27)
in shear
Fv,Rd 1,4Ft,Rd
According to table 3.4 of EN 1993-1-8, punching does not have to be checked for bolts loaded FEd
in shear and tension.

2.47 Bolts loaded in shear and tension.
2.5.4 Countersunk bolts

Countersunk bolts are often used in situations in which the bolt head may not protrude due to
assembly or construction reasons. However, countersunk bolts are less suitable for transferring
forces from a structural point of view. For these bolts the same failure modes apply as for ‘normal’
(non-countersunk) bolts. However, the tension and bearing resistances are reduced because the
bolt head is sunk into the plate material.
The angle and the depth of a countersunk head must meet the requirements of the product
standards described in EN 1993-1-8, cl. 1.2.4. In this case, a value of k2 = 0,63 should be used
in equation (2.23) to determine the design value of the tension resistance Ft,Rd. This means that
the tension strength is only 70% of that of normal (non-countersunk) bolts: for normal bolts k2 = 0,9.
According to note 2 of table 3.4 in EN 1993-1-8, the plate thickness t cannot be fully utilized
when determining the bearing resistance Fb,Rd of a countersunk bolt. The calculation should be
done for an effective plate thickness teff equal to (fig. 2.48):
1
t eff = t – d (2.28)
2 csb
For a countersunk bolt the same checks shall be carried out as for a normal bolt.
1
d
2 csb
dcsb
t
teff
2.48 Reduction of plate thickness when

countersunk bolts are used.

Example 2.4
• Given. A tension bar of two plates 100x12 mm (steel grade S235) is connected with three 8.8
bolts M20 (d0 = 22 mm) to a gusset plate 170x20 mm (steel grade S235) (fig. 2.49). The bolts are
not preloaded (category A). The shear surface is through the thread of the bolts. The connection
is located in a corrosive environment.
• Question. Determine the design value for the resistance TRd of the connection, assuming that
the net cross-section of the gusset plate is not critical.
• Answer. The bolts are loaded in shear and bearing, and the tension resistance of the connected
plates may be governed by either the gross or net cross-section. First check if the bolt positioning
meets the requirements of table 2.24:
end distance: e1 = 35 mm ≥ 1,2d0 = 1,2·22 = 26 mm (OK)

e1 = 35 mm ≤ 4t + 40 mm = 4·12 + 40 = 88 mm (OK)
edge distance: e2,gusset plate = 85 mm ≥ 1,2d0 = 26 mm (OK)
e2,gusset platet = 85 mm ≤ 4t + 40 mm = 4·20 + 40 = 120 mm (OK)
spacing: p1 = 65 mm ≥ 2,2d0 = 2,2·22 = 48 mm (OK)
p1 = 65 mm ≤ 14t = 14·12 = 168 mm (OK)
p1 = 65 mm ≤ 200 mm (OK)
Shear
The shear surface is through the bolt thread. The shear resistance Fv,Rd per bolt and per shear
plane follows from equation (2.14) with αv = 0,6 for 8.8 bolts:
α v fub A s 0,6·800·245·10 –3
Fv,Rd = = = 94,1 kN
γM2 1,25
The design value for the total resistance of the three bolts (in double shear) is:
TRd = number of bolts x number of shear planes per bolt x Fv,Rd = 3·2·94,1 = 565 kN
200 mm
35 35 20
65 65 12 12
35
d0 = 22
50
100 TEd
50
35
inner inner end gusset plate:

bolt bolt bolt parallel to force direction
2.49 Double plate tension bar with three bolts edge edge edge gusset plate:
connected to a gusset plate. bolt bolt bolt perpendicular to force direction

Bearing
The bearing resistance Fb,Rd follows from equation (2.16) to (2.21), depending on the position of
the bolt. The gusset plate (20 mm thickness) is more critical than the two plates forming the ten-
sion bar (total thickness of 2·12 = 24 mm). In the direction of the tension force TEd, there is one
end bolt and two inner bolts. Perpendicular to the force direction there are only edge bolts (for
the gusset plate and for the two plates forming the bar); the term 1,4p2/d0 – 1,7 is not relevant
in equation (2.17). It is found that:
2,8e2,gusset plate 2,8·85

factor k1 edge bolt k1 = – 1,7 = – 1,7 = 9,1
d0 22
or k1 = 2,5 (critical)
e1 35
factor α d end bolt αd = = = 0,53
3d0 3·22
p1 1 65 1
inner bolt αd = – = – = 0,73
3d0 4 3·22 4
fub 800
end bolt α b = α d = 0,53 (critical) or α b = = = 2,22 or α b = 1,0
fu 360
k1α b fudt 2,5·0,53·360·20·20·10 –3

Fb,Rd = = = 153 kN (critical)
γM2 1,25
fub 800
inner bolt α b = α d = 0,73 (critical) or α b = = = 2,22 or α b = 1,0
fu 360
k1α b fudt 2,5·0,73·360·20·20·10 –3

Fb,Rd = = = 210 kN
γM2 1,25
The end bolt has the smallest bearing resistance.
For the (double shear) end bolt: 2Fv,Rd = 2·94,1 = 188 kN > Fb,Rd = 153 kN and for each of the
(double shear) inner bolts: 2Fv,Rd = 2·94,1 = 188 kN < Fb,Rd = 210 kN. The shear resistance of all
bolts in the group is not less than the bearing resistance. Therefore, summing up of individual
resistances is not allowed according to EN 1993-1-8, cl. 3.7. The design value of the resistance of
the bolt group is in this case equal to the product of the number of bolts and the smallest value
of bearing resistance:
Tb,Rd = 3Fb,Rd,end bolt = 3·153 = 459 kN

Tension resistance of plates forming the bar
The tension resistance Nt,Rd of the double plates follows from equation (2.4) and (2.5), depen-
ding on whether yielding of the gross cross-section or failure of the net cross-section is critical:
A = 2bt = 2·100·12 = 2400 mm2
( )
A net = 2 b – d0 t = 2· (100 – 22 ) ·12 = 1872 mm2
Afy
2400·235·10 –3
yielding of gross cross-section Nt,Rd = Npl,Rd = = = 564 kN
γM0 1,0
0,9A net fu 0,9·1872·360·10 –3

failure of net cross-section Nt,Rd = Nu,Rd = =
γM2 1,25
= 485 kN (critical)
When capacity design according to EN 1998 is required, the requirement Npl,Rd = 564 kN <
Nu,Rd = 485 kN must be met, see also equation (2.6). The connection in this example does not
meet this requirement, and therefore is not suitable for application in an earthquake-proof structure.
For the connection to contain a certain deformation capacity, the following limitation must be
satisfied according to equation (2.7):
A net fu 1872 360

0,9 = 0,9· · = 1,08 1,0 (OK)
A fy 2400 235
Finally, bearing resistance is critical for the resistance of the connection as a whole:
TRd = Tb,Rd = 459 kN.
Alternative bolt placement

A higher bearing capacity can be achieved by choosing better bolt positioning, for example
with e1 = 40 mm en p1 = 55 mm. This adapted geometry only influences the bearing resistance.
e1 40
end bolt αd = = = 0,61 α b = α d = 0,61 (critical)
3d0 3·22
k1α b fudt 2,5·0,61·360·20·20·10 –3

Fb,Rd = = = 176 kN
γM2 1,25
p1 1 55 1
inner bolt αd = – = – = 0,58 α b = α d = 0,58 (critical)
3d0 4 3·22 4
k1α b fudt 2,5·0,58·360·20·20·10 –3

Fb,Rd = = = 167 kN
γM2 1,25

For the (double shear) end bolt: 2Fv,Rd = 2·94,1 = 188 kN > Fb,Rd = 176 kN and for each of the
(double shear) inner bolts: 2Fv,Rd = 2·94,1 = 188 kN > Fb,Rd = 167 kN. The shear resistance of all
bolts in the group is larger than the bearing resistance. Therefore, summing up is allowed:
Tb,Rd = 176 + 2·167 = 510 kN
Based on this bolt positioning the tension resistance of the net cross-section is critical:
TRd = Nt,Rd = Nu,Rd = 485 kN
In this example, the alternative bolt positioning results in a higher bearing capacity. In addition,
the connection is more compact: the length of the connection is shortened from 200 mm to 190 mm.
Because bearing is decisive for all bolts, the deformation capacity is also improved.

2.6 Literature
1. H.H. Snijder and H.M.G.M. Steenbergen, Structural basics. Analysis and design of steel
structures for buildings according to Eurocode 0, 1 and 3 (Steel Design 1), Zoetermeer 2019.
2. EN 1090-2 (Execution of steel structures and aluminium structures. Part 2. Technical require-
ments for steel structures), 2018.
3. EN 1990 (Eurocode. Basis of structural design), 2002 (incl. A1, 2005 and AC, 2010).
4. EN 1993-1-1 (Eurocode 3. Design of steel structures. Part 1-1. General rules and rules for
buildings), 2005 (incl. AC, 2009 and A1, 2014).
5. EN 1993-1-8 (Eurocode 3. Design of steel structures. Part 1-8. Design of joints), 2005 (incl. AC,
2009).
6. EN 1998-1 (Design of structures for earthquake resistance. Part 1. General rules, seismic actions
and rules for buildings), 2004 (incl. A1, 2013 and AC, 2009).
7. EN 1998-3 (Design of structures for earthquake resistance. Part 3. Assessment and retrofitting
of buildings), 2005 (incl. AC, 2013).
8. EN 14399-1 (High-strength structural bolting assemblies for preloading. Part 1. General require
ments), 2015.
9. EN 14399-2 (High-strength structural bolting assemblies for preloading. Part 2. Suitability for
preloading), 2015.
10. EN 14399-3 (High-strength structural bolting assemblies for preloading. Part 3. System HR.
Hexagon bolt and nut assemblies), 2015.
11. EN 14399-4 (High-strength structural bolting assemblies for preloading. Part 4. System HV.
12. EN 14399-5 (High-strength structural bolting assemblies for preloading. Part 5. Plain washers),
2015.
13. EN 14399-6 (High-strength structural bolting assemblies for preloading. Part 6. Plain chamfered
washers), 2015.
14. EN 15048-1 (Non-preloaded structural bolting assemblies. Part 1. General requirements),
2016.
15. EN 15048-2 (Non-preloaded structural bolting assemblies. Part 2. Fitness for purpose), 2016.
16. EN ISO 898-1 (Mechanical properties of fasteners made of carbon steel and alloy steel. Part 1.
Bolts, screws and studs with specified property classes. Coarse thread and fine pitch thread),
2013 (incl. AC, 2013).
17. EN ISO 898-2 (Mechanical properties of fasteners made of carbon steel and alloy steel. Part 2.
Nuts with specified property classes. Coarse thread and fine pitch thread), 2012.
18. EN ISO 4759-1 (Tolerances for fasteners. Part 1. Bolts, screws, studs and nuts. Product grades
A, B and C), 2000.

3
Connections
Slip-resistant connections,
rivets and pins

slip-resistant connections, rivets and pins | connections 3 | 1

Slip-resistant connections,
rivets and pins
In steel structures, bolts are most commonly used in clearance holes – these are holes with a
certain oversize – see Connections 2 (Bolts in clearance holes). However, sometimes bolted
connections loaded in shear are subject to significant impact loads, or load reversal. Examples
include bolts used in bridges and crane rails. In these situations, bolts in clearance holes are not
allowed according to EN 1993-1-8, cl. 2.6, and preloaded, fit or injection bolts must be used.
Preloaded bolts increase the resistance against fatigue in structural elements that are loaded in
tension, such as elements found in chimneys and windmills (fig. 3.1). Rivets are only used when
renovating or maintaining historical steel structures that date from before 1940. Pins are mostly
3.1 Preloaded bolts are used, for example, used for hinged connections, or in structures that must be easily demountable. Sometimes archi-
in windmills.
tectural considerations play a role in the specification of pins.
This chapter describes the design rules for checking steel structures with the following mechanical
fasteners:
– preloaded bolts;
– fit bolts and injection bolts;
– rivets;
– pins.
The design rules for bolt positioning and for the influence of holes on the tension resistance of
the connected parts for preloaded, fit and injection bolts are the same as for normal bolts (in
oversize holes): see Connections 2, section 2.3 and 2.4. The product properties are discussed
in Connections 2, section 2.2.
2 | connections 3 | slip-resistant connections, rivets and pins

3.1 Preloaded bolted connections
In a preloaded bolted connection the bolts are carefully tightened using a special tool until a certain axial
tension force is obtained in the bolt. This is called preloading and the axial force obtained in the bolt
shaft is the preload. The connected plates are clamped together by the preload in the bolts. According
to EN 1993-1-8, cl. 3.4 preloaded bolts should only be used for the following bolted connection types
(see Connection 2, section 2.1):
– connections loaded in shear: category B (no slip at serviceabilty limit state) and C (no slip at ultimate
limit state);
– connections loaded in tension: category E.
For categories B and C the purpose of preloading is to prevent the connected parts slipping relative to each
other below a certain level of loading. This is called a slip-resistant connection. Preloading has a beneficial
effect on the deformation and fatigue strength of bolted connections. However, the controlled tightening
process – necessary to obtain the required level of preload – and the special treatment of contact surfaces
needed for connections loaded in shear, both increase the costs. Therefore, it is wise to consider if preloading
is really necessary or if it could be avoided by a structural reconfiguration. The quality that can be achieved
for connections with preloaded bolts mainly depends on the manufacturing method and the dimensional
accuracy. Deformed contact surfaces – for example due to shrinkage after welding – are often not permitted.
One exception to this rule are end-plate connections loaded in tension, in which deformation due to weld
shrinkage has a positive effect on the fatigue strength (see section 3.1.3). It is important to use a reliable
tightening method so that the required preload is obtained in the bolt and overloading of the bolt is
prevented. Overloading can result in damaging, even completely ‘turning-off’, the bolt.
3.1.1 Tightening of preloaded bolts

EN 1090-2, cl. 8.5 describes procedures for tightening preloaded bolts. According to EN 1090-2, cl. 8.5.1,
preloaded bolts must be tightened such that the nominal minimum preload force Fp,C is obtained:
Fp,C = 0,7fubAs (3.1)
Where:
fub nominal ultimate strength of the bolt material;
As tension stress area of the bolt.
The value of the preload force according to equation (3.1) is equal to the value specified in EN 1993-1-8,
cl. 3.9.1(2), for use in the calculation (see section 3.1.2).
EN 1090-2 discusses the following four tightening methods for preloaded bolts:
– torque method (cl. 8.5.3);
– combined method (cl. 8.5.4);
– HRC (tension control) method (cl. 8.5.5);
– DTI (direct tension indicator method) (cl. 8.5.6).

Torque method
In the torque method, the tightening tool is set to a fixed tightening torque (fig. 3.2). When the
fixed point is reached, the tool will give an audible click or will automatically turn off. The bolt
(or nut) is then tightened to the predetermined moment. A large part of the moment (tightening
torque) is needed to overcome friction between the thread of the bolt and the thread of the nut
(thread friction), and friction between the nut and the washer (face friction). Only a small part of
the moment is effectively used to induce the preloading force.
According to EN 1090-2, cl. 8.5.2 the tightening torque Mr is determined as follows:
Mr = kmdFp,C (3.2)
Where:
km coefficient, which amongst other things depends on the thread friction, the face friction and
the pitch of the thread;
d nominal bolt diameter.
The value of km is specified by the supplier of the bolt sets and must be determined in accor-
3.2 Tightening of preloaded bolts using the dance with the relevant part of EN 14399. The coefficient km can vary considerably, even with
torque method.
new bolts. EN 14399-1 defines three K-classes:
– K0: no requirements for how km is specified;
– K1: specification by the supplier of a series of individual values ki;
– K2: specification by the supplier of a mean value for km and the coefficient Vk = sk/km (whereby
sk is the standard deviation).
According to table 20 of EN 1090-2, bolt sets of k-class K2 must be applied in the torque method.
EN 14399-2 and EN 14399-4 specify for both HR bolts and HV bolts the following values for
km and Vk:
0,10 ≤ km ≤ 0,23
Vk ≤ 0,10
When the torque method is used and the thread friction and/or the face friction is larger than
expected the required level of preload in the bolt is not reached. When the thread friction and/
or the face friction is lower than expected there is a greater chance of turning off the nut.
It is important that preloaded bolts are installed with the lubrication applied by the manufacturer,
and that the complete bolt set – as tested – is used. EN 1090-2, cl. 8.5.1 states that when extra
lubrication is applied the bolt set must be tested again according to the procedure described in
Annex H of EN 1090-2.
When preloaded bolts have been loosened after preloading they may not be used again as
preloaded bolts.

In addition to variations in friction (represented by coefficient km), the tightening torque itself is
a second source of inaccuracy. EN 1090-2, cl. 8.5.1 prescribes that the maximum deviation of
the tightening tool (torque wrench) shall be limited to ± 4%. The accuracy of wrenches must be
checked at least once a week.
When the elements of the connection are not completely flat prior to preloading, preloading
of a given bolt will influence the preload of the earlier preloaded bolts. This must be taken into
account when choosing the order of tightening. Tightening must be carried moving away from
the stiffest part towards the least stiff part of the connection. According to EN 1090-2, cl. 8.5.3
adjustments must be carried out in at least two steps to achieve a uniform level of preload.
Combined method
The combined method makes use of the fact that a suitable preloaded bolt is sufficiently tough
to allow – even after the required preload is achieved – a considerable tightening of the nut
before fracture occurs. EN 14399-2 describes the tests that a manufacturer has to carry out to
demonstrate that this is the case.
The combined method is significantly more accurate than the torque method. The tightening
procedure consists of the following two steps:
• Step 1: First, the bolt is tightened with a moment equal to 75% of the moment that is required
to obtain the full preload. Generally, the resulting force in the bolt is sufficient to properly clamp
the connection parts and the chance of turning of the nut is small.
• Step 2: The position of the nut – or the head, when it is tightened – is marked with chalk or
paint and the nut is then tightened further by turning over a so-called finishing angle. Table 21
of EN 1090-2 gives values for the finishing angle, depending on the total thickness t (up to t = 10d)
of the parts to be connected, including all filling plates and washers. This angle has been deter-
mined through tests. The finishing angle varies between 60˚ and 120˚.
Figure 3.3 shows that with the combined method consistent bolt forces can be achieved, despite
the large variations one may find in clearance and stiffness of the parts to be connected.
preload
=
bolt force
plate stack:
stiff
2 3
1
3
F
4 p
middle stiff
weak 3.3 With the combined method,

α α α tightening angle consistent bolt forces can be achieved
α1 (α is the finishing angle).
α2
α3

before tightening during tightening after tightening
3.4 HRC (High Resistance Calibrated) or tension

control method.
HRC (tension control) method

The HRC (High Resistance Calibrated) method is a special form of the torque method, in which
so-called tension control (or HRC) bolts are used. The requirements are given in EN 14399-10. These
bolts are tightened via the nut, while the shaft is held in place through an extra bolt head at the
embossment
end of the shaft. The tightening moment is limited by wrenching/pulling off the extra bolt head
punched in washer
(fig. 3.4). A feature of these bolts is that they can be tightened from one side only (no access is
needed to the other side). In addition, the head can also be round.
DTI (direct tension indicator) method

The level of preload can be determined directly, for instance by using force-indicating bolts or
washers. Appendix J of EN 1090-2 deals with the use of direct tension indicators. The principle of
force-indicating washers is that embossments punched in the washer will be pressed flat under a
known force (fig. 3.5). The manufacturer ensures that the preload force Fp,C is evenly distributed
around the washer.
One point to remember when direct tension indicator washers are used is that the ‘bumps’ will
deform (flatten) permanently, so they only confirm that the required level of preload has been
present in the bolt at some stage. When the load in a given bolt reduces, for example due to
tightening of a neighbouring bolt, this cannot be seen. This means that a correct choice of
before tightening
tightening sequence and number of tightening steps is essential for a good result. Another
disadvantage is that there may be a permanent gap between the washer and the bolt (or bolt
head), where moisture can get in. Direct tension indicating washers are mainly used in the United
Kingdom.
after tightening 3.1.2 Preloaded connection loaded in shear

Tightening the nut creates a tension force in the shaft of a preloaded bolt. This tension force is
3.5 Force-indicating washer (direct tension
indicator). in equilibrium with the contact pressure on the thread and bolt head/nut. As a result, the bolt
shaft is stretched and the connection plates are compressed. In the case of a preloaded con-
nection loaded in shear (category B and C), no displacement occurs until the frictional resistance

preload force
between the contact surfaces is exceeded. Figure 3.6 illustrates the mechanism of this
contact or
force distribution. friction surface
The magnitude of the preload and the slip factor between the plates determine the 1
F
2 Ed
shear load that can be transferred. A slip- resistant connection can still be achieved
when preloaded bolts are used in oversize holes. Connections with preloaded bolts in FEd
category B should not slip at the serviceability limit state, resulting in limited deforma-
1
tion in use. Connections in category C should not slip even at the ultimate limit state. F
2 Ed
In addition to the slip-resistance of a preloaded connection loaded in shear, other criteria

must also be checked, depending on the category of the connection.
• Category B (slip-resistant at serviceability limit state). At the ultimate limit state slip
is possible, so it is necessary to check for bearing and shear according to EN 1993-1-8, 3.6 Load transfer in a preloaded connection loaded in shear.
cl. 3.6.1; see Connections 2 (Bolts in clearance holes). In addition, a check for tension
resistance (fracture of the net cross-section and yielding of the gross cross-section) of
the connected parts must be carried out.
• Category C (slip-resistant at ultimate limit state). A check on shear is not necessary,
category criterion check
because the design limit of the slip resistance is always critical. However, a check on
bearing is necessary, as well as a check on yielding of the net cross-section. The require- slip Fv,Ed,ser ≤ Fs,Rd,ser
ments for bearing resistance and for the resistance of the net cross-section follow from the B shear Fv,Ed ≤ Fv,Rd
requirement to avoid that after slip failure occurs without sufficient deformation capacity. bearing Fv,Ed ≤ Fb,Rd
When slip occurs under extreme loading, the connection is still able to absorb the load slip Fv,Ed ≤ Fs,Rd
through bearing. C bearing Fv,Ed ≤ Fb,Rd
yielding net
Fv,Ed ≤ Nnet,Rd
The various criteria to check for a preloaded bolted connection in shear are given in cross-section
table 3.2. of EN 1993-1-8 (table 3.7)
3.7 Criteria to check for a preloaded connection loaded
in shear.
Slip resistance
The resistance to slip provided by friction depends on:
– preload force Fp,C;
– slip factor µ;
– partial factor γ M;
– number of surfaces n between which friction occurs;
– reduction factor ks to take into account the use of oversize or slotted holes.
EN 1993-1-8, cl. 3.9.1(2) indicates that for 8.8 and 10.9 bolts, tightened according to
EN 1090-2, the following value for the preload force Fp,C must be assumed:
Fp,C = 0,7fubAs (= 3.1)
The slip factor µ strongly depends on the quality and treatment of the contact surfaces.
Both treated and untreated contact surfaces must always be free of oil, paint, or other

contaminants. The slip factor is a minimum for untreated contact surfaces, and
surface treatment class slip factor µ
a maximum for blasted surfaces. In connections with blasted surfaces, attention
surfaces blasted with shot or grit, must be paid to rust formation both during the period between manufacture and
with loose rust removed, and free A 0,50
assembly, and after assembly. When blasted surfaces are protected against
from pitting
corrosion, for example by spraying with aluminium and by paints based on an
surfaces blasted with shot or grit,
and then: alkali-zinc silicate, the slip factor may increase. The slip factor between contact
– spray-coated with an aluminium or surfaces with a certain finish can only be established by testing. EN 1090-2
B 0,40
zinc based product
describes in appendix G how such tests should be carried out. Some finishes
– with alkali-zinc silicate paint with a
thickness of 50-80 µm (e.g. zinc) exhibit creep, which leads to an increasing displacement under a
constant load. In such cases there is a large difference in the slip factor for short
surfaces cleaned by wire-brushing
or flame cleaning, with loose rust C 0,30 term and long term loading.
removed To be able to make a clear distinction between surface treatments, table 3.7 of
none: surfaces as rolled D 0,20 EN 1993-1-8 provides a classification system with four classes A to D, each with
an associated slip factor µ varying between 0,5 and 0,2. Table 18 of EN 1090-2
3.8 Classes of surface treatment according to EN 1090-2. gives examples of surface treatments for these four classes (table 3.8).
Hot-dipped galvanized surfaces of steel elements are less suitable for the appli-
cation of slip-resistant preloaded connections. As shown by durability test, friction
reduces due to creep of the zinc layer causing a lower value of the slip factor. In
practice, for hot-dipped galvanized surfaces a value of µ = 0,1 is often used. This
value falls outside the four classes given in table 3.8. A value of µ = 0,2 (class D)
can be achieved by coating a hot-dipped galvanized surface (layer thickness of
80-140 µm) with a layer of inorganic zinc silicate (layer thickness 50-80 µm). In
practice it is better to cover the contact surfaces during galvanizing, and then
apply one of the surface treatments from table 3.8 (after galvanizing).
The design value of the slip resistance Fs,Rd for preloaded bolts in class 8.8 or
10.9 is given in EN 1993-1-8, cl. 3.9.1(1):
k snμ
Fs,Rd = F (3.3)
γM3 p,C
Where:
ks reduction factor, which depend on the type of hole: for normal holes ks = 1
and for oversized and slotted holes ks < 1, see table 3.6 of EN 1993-1-8;
n number of friction surfaces;
γ M3 partial factor, depending on the category (limit state):
– connections in category B: γ M3 (= γ M3,ser ) = 1,10;
– connections in category C: γ M3 = 1,25.
Note that EN 1993-1-8, cl. 3.6.1(2) defines the design preload force as Fp,Cd =
0,7fubAs/γ M7. However, this value is not used in equation (3.3) for determining

the slip resistance, because in that equation the entire safety margin has already been allowed
for in the partial factor γ M3. Equation (3.3) can be applied to both the serviceability limit state
(category B) and ultimate limit state (category C). For serviceability the slip resistance is referred
to as Fs,Rd,ser and the corresponding partial factor as γ M3,ser (‘ser’ stands for: serviceability limit
state).
Shear resistance
The shear resistance Fv,Rd of a preloaded bolt is equal to that of a ‘normal’ bolt, and depends on
whether the shear plane passes through the shaft (surface A) or through the thread (surface As)
of the bolt, see Connections 2 (Bolts in clearance holes), section 2.5.1:
0,6 fub A
shear plane through shaft Fv,Rd = (3.4)
γM2
α v fub A s
shear plane through thread Fv,Rd = (3.5)
γM2

The factor αv depends on the bolt class: αv = 0,6 for class 8.8 and αv = 0,5 for class 10.9.
Bearing resistance
The bearing resistance Fb,Rd of a preloaded bolt is equal to that of a ‘normal’ bolt:
k1α b fudt
Fb,Rd = (3.6)
γM2
The factor k1 depends on the edge distance and spacing perpendicular to the direction of the
force (e2 and p2), and the factor ab on the end distance and spacing in the direction of the force
(e1 and p1); see Connections 2, section 2.5.1.
Tension resistance of cross-section

The design value of the tension resistance Nt,Rd of a member with bolt holes, depends on whether
the gross cross-section yields or the net cross-section fails; see Connections 2, section 2.4.2:
Afy
yielding of gross cross-section Nt,Rd = Npl,Rd = (3.7)
γM0

0,9A net fu
failure of net cross-section Nt,Rd = Nu,Rd = (3.8)
γM2
Yielding resistance of net cross-section

The resistance against yielding of the net cross-section Nnet,Rd follows from:
A net fy
Nnet,Rd = (3.9)
γM0

Example 3.1
200 mm
• Given. A tension bar formed from a double plate 100x12 mm (in steel grade S235) is
35 35 20
65 65 12 12
connected to a gusset plate of 170x20 mm (in steel grade S235) with three preloaded M12
35 class 10.9 bolts in normal holes (fig. 3.9). This is the same connection as example 2.4 in Con-
50
d0 = 22 TEd nections 2 (Bolts in clearance holes), except using preloaded bolts. The shear plane runs
100
through the threaded portion of the bolts. The surface treatments of the plates and the gusset
50
plate belong to class C.
35
• Question. Calculate the design value of the resistance TRd for the connection in cate
3.9 Double plated tension bar connected to a gories B and respectively C. It may be assumed that the net cross-section of the gusset plate
gusset plate with preloaded bolts. is not critical.
• Answer. The criteria to check for a preloaded connection loaded in shear are shown in
table 3.7 for both categories.
Category B (slip-resistant at serviceability limit state)

The resistances to slip, bearing, and shear must be checked for preloaded connections in
category B.
Slip
The design value of the slip resistance Fs,Rd,ser for each bolt follows from equation (3.1) and
(3.3). There are two contact surfaces for each preloaded bolt, therefore n = 2. For a surface
treatment in class C, µ = 0,3 (see table 3.8). Therefore:
Fp,C = 0,7fub A s = 0,7·1000·245·10 –3 = 172 kN
k nμ 1,0·2·0,3
Fs,Rd,ser = γ s Fp,C = ·172 = 93,8 kN
M3,ser 1,1
The design value of the resistance at the serviceability limit state is:
Ts,Rd,ser = number of bolts x Fs,Rd,ser = 3·93,8 = 281 kN
Shear
The shear resistance Fv,Rd follows from equation (3.5), since the shear plane runs through
the thread of the bolt. With α v = 0,5 for bolt class 10.9:
α v fub A s 0,5·1000·245·10 –3
Fv,Rd = = = 98,0 kN
γM2 1,25
The shear resistance of the 10.9 bolt is only slightly greater than that of an 8.8 bolt: Fv,Rd = 94,1 kN;
see example 2.4 in Connections 2. This is due to the smaller value of α v. The design value
for the total resistance Tv,Rd of the three preloaded bolts (double shear) is:
Tv,Rd = number of bolts x number of shear planes for each bolt x Fv,Rd = 3·2·98,0 = 588 kN

Bearing
The bearing resistance Fb,Rd of an M20 class 10.9 preloaded bolt is equal to the resistance of an
M20 class 8.8 bolt, as considered in example 2.4 in Connections 2. For the end bolt Fb,Rd = 153 kN,
and for the inner bolts Fb,Rd = 210 kN.
For the end bolt (double shear): 2Fv,Rd = 2·98,0 = 196 kN > Fb,Rd = 153 kN and for each of the
inner bolts (also double shear): 2Fv,Rd = 2·98,0 = 196 kN < Fb,Rd = 210 kN. The shear resistance
of the inner bolts is therefore smaller than the bearing resistance, which means that adding
resistances is not allowed according to EN 1993-1-8, cl. 3.7. The design value of the bearing
resistance of the connection is then equal to the product of the number of bolts and the smallest
design value of the bearing resistance:
Tb,Rd = 3Fb,Rd,end bolt = 3·153 = 459 kN
Conclusion: the design value of the resistance TRd at the ultimate limit state of this connection in
category B is governed by bearing: TRd = Tb,Rd = 459 kN.
Category C (slip-resistant at ultimate limit state)

The resistances against slip and bearing, and the yield resistance of the net cross-section, must
be checked for a preloaded bolt in category C.
Slip
The design value of the slip resistance Fs,Rd for each bolt is determined using equation (3.1) and
(3.3), but with a different value for γ M3 than that used for category B bolts:
k snμ 1,0·2·0,3
Fs,Rd = F = ·172 = 82,6 kN
γM3 p,C 1,25
For three bolts together: Ts,Rd = 3Fs,Rd = 3·82,6 = 248 kN.
Bearing
The bearing resistance Fb,Rd of a connection in category C is equal to Fb,RD for category B: Tb,Rd
= 3Fb,Rd,end bolt = 3·153 = 459 kN.
Yield resistance of the net cross-section

The yield resistance of the net cross-section at the location of the bolt holes Nnet.Rd is given by
equation (3.9):
A net fy
1872·235·10 –3
Nnet,Rd = = = 440 kN
γM 1,0
0
Conclusion: slip resistance governs the design value of the resistance TRd of this connection in
category C: TRd = Ts,Rd = 248 kN.

3.1.3 Preloaded connection loaded in tension
Preloading a connection that is subject solely to tension (category E) is a sensible option only
if the tension load varies considerably in size or in direction. Such variations may in fact lead to
fatigue failure of the bolts. Preloaded bolts reduce this danger provided the connection is detailed
properly. The variation of bolt force due to an external tension force applied to a preloaded
connection can be explained using a simple model (fig. 3.10).
Here, the (steel) plate stack is conceived as an elastic block and the bolts as springs. The preload
in the springs is Fp and the external load is Ft,E. Due to an external load Ft,E < Fp, the contact pressure
in the plate stack decreases and the tension force in the bolts (springs) increases. According to
Hooke’s Law, the elongation ΔLb of the bolts due to a preload force Fp is:
FpL AE ΔL b
ΔL b = Fp = (3.10)
AE L
Where L is the thickness of the plate stack and A is the diameter of the bolt shafts. For simplicity it
is assumed that the threaded part of the bolt shaft has the same cross-section as the shaft. From
figure 3.11 can then be derived using equation (3.10):
Fp
HG AE
tgα = = = (3.11)
OG ΔL b L
Ft,E
force
plate stack
N
L
spring H
Fp
1 1
F F
2 p 2 p
Ft,E
Ft,E
K
Ft,E
3.10 Schematization of a connection with
preloaded bolts loaded in tension. F
Fp’
a β
O C G D E deformation
∆Lb,t
∆Lb ∆Lpl
3.11 Force/deformation diagram of a preloaded
connection loaded in tension. f

For the plate stack, with cross-section Apl, the following may be derived:
Fp Fp A plE
HG
tgβ = = = = (3.12)
EG ΔL pl Fp L L
A plE
It is essential that the sum of ΔLb and ΔLpl does not change due to the application of an external
tension load Ft,E. The influence of such a tension force on the behavior of a connection is shown
by the force/deformation diagram figure 3.11. The force Ft,E is drawn from the origin O. The corre
sponding deformation is represented by the line OC (ΔLb,t). Line FD represents the remaining
preload force Fp’, line CD is the extension, and line DE represents the shortening of the plate stack
due to Fp’. The diagram shows that the total tension force in the bolt (line DL) increases only slightly
due to the tension force Ft,E. The stiffer the plate stack, the smaller the increase in tension force in
the bolt. With an increasing external tension force, the contact pressure eventually decreases to
zero, after which the bolt is released from the plate stack. In this (theoretical) limit case the tension
load Ft,E is equal to the bolt force (line NE).
The stiffness and variation of the bolt force under the influence of an external load Ft,E both depend
strongly on the design of the connection. Two situations are compared in figure 3.12: in the
connection on the left, the flanges are initially deformed in such a way that the contact surface is
located centrally between the webs. In the connection on the right, the deformation is such that
the flanges are in contact at their tips. Figure 3.12 also shows the associated force/deformation
diagrams (note that compared to figure 3.11 only the part on the right side of line GH is shown).
2Ft,E 2Ft,E
y y
force 2Ft,E force 2Ft,E
Ft,E ∆F
b
∆Fb
Fp Fp
Ft,E ∆Fc
∆Fc
Fp preload force
∆Fb variation in bolt force
∆Fc variation in contact force 3.12 Influence of the location of the contact
point(s) on the stiffness and on the variation
deformation (∆y) deformation (∆y) of the bolt force in a connection.

In the situation on the left, the deformation Δy is much smaller than in the situation on the
right for a given tension load Ft,E. Moreover – and certainly more important – in the situation
on the left the change in bolt force (ΔFb) is much smaller. This is of great importance for
structures subjected to fatigue loading, given that the fatigue strength of a bolt is low due to
the notching effect of its thread.
Figure 3.13 shows the results of some fatigue tests, which illustrate the importance of preload
and the correct location of the contact point(s). As long as the external load is transferred by
reducing the contact pressure, no fatigue occurs in the preloaded bolts. This means a fatigue
check of the bolts is not necessary. In an assessment of sensitivity to fatigue, according to
EN 1993-1-8, cl. 3.6.1(2) it may be assumed that Fp = Fp,Cd.
Clearly the other parts of a connection must also be checked for fatigue. End-plates deform due
to weld shrinkage in a similar way to that shown on the left in figure 3.12. This leads to a favorable
location of the contact points and thus it is beneficial to not correct such deformation.
2Ft,E
bolts M20
class 10.9 bolts M20
Fv = 176 kN class 8.8
Fv = 100 kN Fv = 100 kN
Fv = 0 kN and Fv = 105 kN
2Ft,E
160 160
140 140
120 120
100 100
Ft,E (kN)
Ft,E (kN)
80 80
60 60
40 40
20 20
0 0
104 105 106 107 104 105 106 107
number of cycles number of cycles
3.13 Influence of the preload force and the location of the contact point(s) on fatigue strength.

For designing a preloaded connection loaded in tension, according to table 3.2 of EN 1993-1-8
the same equations apply as for ‘normal’ non-preloaded bolted connections, see Connections 2
(Bolts in clearance holes), section 2.5.2.
k 2 fub A s
tension resistance Ft,Rd = Fv,Ed (3.13)
γM2

0,6 πdmt p fu
punching resistance Bp,Rd = Fv,Ed (3.14)
γM2

Where k2 is a calibration factor: k2 = 0,9 for normal bolts with a thread that meets the requirements
of EN 1090-2.
3.1.4 Preloaded connection loaded in shear and tension

When a preloaded bolt is subjected to an external tension load the contact force decreases,
and therefore also the friction force and slip resistance decrease (see fig. 3.11). The decrease
in the contact force depends on the stiffness ratio between the bolts and the plate stack (see
fig. 3.12). For simplicity, the design rules in EN 1993-1-8, cl. 3.9.2 are based on the application
of a standard reduction in the contact force equal to 80% of the external load Ft,Ed. The slip
resistance Fs,Rd then follows from:
category B Fs,Rd,ser =
(
k snμ Fp,C – 0,8Ft,Ed,ser ) (3.15)
γM3,ser
category C Fs,Rd =
(
k snμ Fp,C – 0,8Ft,Ed )
(3.16)
γM3
Equation (3.15) and (3.16) do not apply for connections in which the tension force on the bolt is
caused by an external moment. In such cases it may be assumed that the decrease of the contact
pressure in the tension zone is compensated by the increasing contact pressure in the compres-
sion zone. According to table 3.4 of EN 1993-1-8 the check for combined shear and tension is:
(3.17)
This is the same formula as used for normal, non-preloaded bolts loaded in shear and tension,
see Connections 2 (Bolts in clearance holes), section 2.5.3.

FEd,hor s = 150 mm
FEd,vert
3.14 Connection with four preloaded bolts

a = 100
loaded in tension and shear.
Example 3.2
• Given. A connection with four preloaded M22 class 10.9 bolts, loaded by a horizontal force
FEd,hor = 220 kN and a vertical force FEd,vert = 260 kN (fig. 3.14). The bolt threads are in the shear
plane and the holes have a normal clearance. The friction surfaces are treated to class A. The
connection should not slip at the ultimate limit state (category C).
• Question. Check the preloaded bolts when prying forces do not occur (see Connections 2,
section 2.1.2).
• Answer. First is checked, whether or not the connection slips. A check for combined shear and
tension follows.
Slip
The shear force per bolt is Fs,Ed = FEd,vert/4 = 260/4 = 65,0 kN and the direct tension force for
each bolt is Ft,Ed = FEd,hor/4 = 220/4 = 55,0 kN. The eccentricity of the vertical force FEd,vert
causes extra tension forces in the upper two preloaded bolts. This eccentricity results in an external
moment, which does not have to be taken into account when determining the slip resistance. For
bolts M22 in class 10.9 the design value of the slip resistance Fs,Rd follows from equation (3.16),
with a preload force Fp,C that is determined from equation (3.1):
Fp,C = 0,7 fub A s = 0,7·1000·303·10 –3 = 212 kN
Fs,Rd =
(
k snμ Fp,C – 0,8Ft,Ed ) = 1,0·1·0,5·(212 – 0,8·55, 0) = 67,2 kN
γM3 1,25
Fs,Ed 65,0
= = 0,97 1,0 (OK)
Fs,Rd 67, 2
So no slip at the ultimate limit state.

Shear and tension
When checking the bolts the extra tension force in the upper two bolts due to the eccentricity
moment must be taken into account. It may be assumed that the rotation point is located at
the level of the lower bolts (see also Connections 5 (Design of connections), section 5.2.2). The
maximum tension force Ft,Ed in the upper preloaded bolts is, for each bolt:
FEd,hor aFEd,vert 220 100·260

Ft,Ed = + = + = 142 kN
4 2s 4 2·150
The maximum shear force on each bolt is Fv,Ed = FEd,vert/4 = 260/4 = 65,0 kN. The resistance to
combined shear and tension follows from equation (3.5) and (3.13) respectively:
α v fub A s 0,5·1000·303·10 –3
Fv,Rd = = = 121 kN
γM2 1,25
k 2 fub A s 0,9·1000·303·10 –3
Ft,Rd = = = 218 kN
γM2 1,25
The check for combined shear and tension at the ultimate limit state is, according to equation (3.17):
Fv,Ed Ft,Ed 65,0 142

+ = + = 1,00 1,0 (OK)
Fv,Rd 1,4Ft,Rd 121 1,4·218
Bearing will not be critical, therefore this check is not considered here.
3.2 Connections with fit bolts and injection bolts

The stack of steel plates in a connection may be clamped together by preloading the bolts,
which prevents the plates moving (slipping) relative to each other. Movement can also be prevented
by using fit bolts, or injection bolts. Both bolt types are considered in this section.
3.2.1 Fit bolts

Fit bolts are not preloaded, so no special treatment of contact surfaces, and no controlled tightening
procedures, are required. The holes for fit bolts are initially drilled, in the fabrication shop, to
be approximately 2 mm smaller than the nominal bolt diameter. During assembly the holes are
enlarged to their nominal diameter.
Fit bolts come with a specified tolerance and fit with a tolerance of approximately 0,1 mm. As a
result, almost no movement can occur in the bolt holes. However, making the initial holes larger
is time consuming and costly, especially when this must be done on site. Alternatively injection
bolts can be used, see section 3.2.2.

In EN 1993-1-8, cl. 3.6.1, the following conditions are given for the use of fit bolts:
– paragraph (6): fit bolts must be designed as bolts in normal sized holes;
– paragraph (7): the shear plane must not pass through the thread of the bolt;
– paragraph (8): the length of the thread in bearing may not exceed 1/3 of the plate thickness
t
≤
1
t (fig. 3.15);
3
– paragraph (9): the hole tolerance must comply with EN 1090-2, cl. 6.6.2, in which the tolerance
class H111 is prescribed according to EN-ISO 286-2.
3.15 Maximum length of the thread for fit bolts. For the design of a connection with fit bolts the same rules as for ‘normal’ non-preloaded bolts
apply, see Connections 2 (Bolts in clearance holes), section 2.5 and equation (3.4) to (3.8), and
(3.13) and (3.14).
3.2.2 Injection bolts

Injection bolts are bolts of class 8.8 or higher with an opening in the head of the bolt (fig. 3.16).
A synthetic resin is injected through this opening into the space between the bolt shaft and the
sides of the hole. In this way, a connection is created in which relative movement of the plates is
prevented. Special washers under the head and the nut are necessary to be able to inject the resin.
When the bolt and washers are assembled, attention must be paid to the correct positioning
of the washers. The closing washer, which has a recessed space and a bevelled edge, must be
underneath the bolt head, and the washer with the venting groove must be underneath the nut.
Appendix K of EN 1090-2 provides rules for the execution of injection bolts.
detail of washer
1,5
d0 = d + 0,5 mm
Ø 5,5 mm
Ø 3,2 mm washer with recessed space
and beveled edge (at bolt
head side); see detail
synthetic resin
washer with venting

groove (at nut side)
3.16 Non-preloaded injection bolt.

One disadvantage of injection bolts is that injection is not possible in unfavourable weather
conditions, and at low temperatures. The bolts must be fully tightened before injection, and the
sides of the hole must be clean and dry. The preparation of the synthetic resin and the injection
process itself are rather simple, although due care is necessary.
Injection bolts can be preloaded, in which case a connection is created in which the load is not
only transferred by friction between the contact surfaces but also by bearing of the synthetic
resin and shear of the bolt shaft. Non-preloaded injection bolts are mainly used in maintenance
and repair work, and in slip-resistant connections in which the slip factor of the contact surfaces
is low. The latter is for example the case with hot-dipped galvanized or painted structural ele-
ments. Preloaded injection bolts are particularly suitable for dynamically loaded structures, or for
structures with varying loads, such as traffic and railway bridges.
Non-preloaded injection bolts

Load transfer for non-preloaded injection bolts is achieved through bearing and shear. Non-
preloaded bolts belong to category A. The design values of the tension resistance Ft,Rd and shear
resistance Fv,Rd are equal to those of non-preloaded bolts in normal holes, see Connections 2,
section 2.5.1 and 2.5.2.
The design value of the bearing resistance Fb,Rd,resin depends on the properties of the synthetic
resin. Under the influence of excessive constant pressure creep would occur in the resin, meaning
that in the long term the bolt would come to lie eccentrically in the bolt hole, with resulting slip
(fig. 3.17). To prevent creep, the design value for the bearing resistance of the synthetic resin
should be based on a long term bearing stress for which the slip does not exceed 0,3 mm.
To determine the design value for the bearing resistance, the bearing stress must be known for
each connected part. In the connection of figure 3.18, for example, the bearing stress σ1 in the
intermediate plate and the bearing stress σ2 in the outer plates must both be known. The bea-
ring stress σ1 will be critical when t1/t2 < 2. The bearing stress increases with a decreasing ratio
of t1/t2 and this effect is taken into account with a factor β. A similar situation occurs when four
or more plates are connected, but this situation is not treated in EN 1993-1-8. Conservatively,
one can assume that β = 1.
According to EN 1993-1-8, cl. 3.6.2.2(5), the design value for the bearing resistance of the synthetic
resin Fb,Rd,resin is determined as follows:
d0 d0
k t ksdt b,resin βfb,resin d d
Fb,Rd,resin = with γM4 = 1,0 (3.18)
γM4
3.17 Shift of an injection bolt in the bolt hole

shift of
due to creep of the synthetic resin. centre

Where:
kt factor depending on the limit state:
kt = 1,0 for the serviceability limit state, long duration;
kt = 1,2 for the ultimate limit state;
ks factor for the clearance:
ks = 1,0 for holes with normal clearances;
ks = 1,0 – 0,1m for oversized holes, in which m is the difference (in mm) between normal
and oversized hole dimensions;
d bolt diameter
tb,resin effective bearing thickness of the resin:
t1
for 2 t b,resin = 2t 2 1,5d
t2
t1
for <2 t b,resin = t1 1,5d
t2
β coefficient depending of the thickness ratio t1/t2 of the connected plates (see figure
3.18);
fb,resin bearing strength of the resin, determined by a test procedure similar to the procedure
for determining the slip factor of contact surfaces in Annex G of EN 1090-2. A commonly
used value for the bearing strength for long term loading is fb,resin,long term = 130 N/mm2.
Recent research [10] shows that for the two-component resin Araldite SW404 with hardener
HY2404 one may take fb,resin,long term = 200 N/mm2 en fb,resin,short term = 280 N/mm2.
3.18 Relationship between β and t1/t2.

β
t1
β = 1,66 – 0,33
t2
1,33
1,0
σ1 σ1 σ2
0 1,0 2,0
t1 / t2
t2 outer plate σ2
t1 σ1 intermediate
plate
t2 outer plate σ2

Preloaded injection bolts
When an injection bolt is preloaded, friction occurs between the contact surfaces of the
connecting parts. When determining the shear resistance Fv,Rd of a connection with preload
injection bolts, the slip resistance Fs,Rd and the bearing resistance Fb,Rd,resin may be added
according to EN 1993-1-8, cl. 3.6.2.2(4).
The slip resistance Fs,Rd of preloaded injection bolts is equal to that for preloaded bolts, see
equation (3.3). The same applies for combined shear and tension. The bearing resistance
Fb,Rd,resin of preloaded injection bolts is equal to that of non-preloaded injection bolts, see
equation (3.18).
Checks
EN 1993-1-8, cl. 3.6.2.2(4) lists the relevant checks for injection bolts. As a function of the
category of the connection, table 3.19 gives the criteria for a connection with injection bolts
loaded in shear. In this table:
Fv,Ed,ser design value of the shear force on each bolt at the serviceability limit state;
Fb,Rd,resin,ser design value of the bearing resistance of the resin at the serviceability limit
state;
Fs,Rd,ser design value of the slip resistance of each bolt at the serviceability limit state.
In addition, a check for tension resistance (failuree of the net cross-section or yielding of the
gross cross-section) is needed.
category criteria check
bearing of resin Fv,Ed ≤ Fb,Rd,resin
A (non-preloaded) bearing of steel Fv,Ed ≤ Fb,Rd
shear Fv,Ed ≤ Fv,Rd
bearing of resin Fv,Ed,ser ≤ Fb,Rd,resin,ser + Fs,Rd,ser
B (preloaded) bearing of steel Fv,Ed ≤ Fb,Rd
bearing of resin Fv,Ed ≤ Fb,Rd,resin + Fs,Rd
C (preloaded) bearing of steel Fv,Ed ≤ Fb,Rd
3.19 Criteria to check for a connection with injection bolts loaded in shear.

Example 3.3
• Given. Two steel plates with a 280x25 mm cross-section in grade S235 are connected end-to-end
using 12 non-preloaded M24 injection bolts (class 8.8 in normal holes) and two double-sided 20 mm
thick cover plates, also in grade S235 (fig. 3.20). The design strength of the synthetic resin
fb,resin = 130 N/mm2. The connection is loaded with a tension force TEd = 1340 kN.
• Question. Check the bolted connection (bolts in double shear), for which it may be assumed
that bearing of the steel is not critical (Fb,Rd,resin < Fb,Rd).
• Answer. Non-preloaded injection bolts belong to category A, for which the following resistances
must be checked; bearing of the synthetic resin, bearing of the steel (not critical), and shearing of the
bolt (see table 3.19). In addition, the tension resistance of the cross-section of the connected parts
must be checked for the connection as a whole.
Bearing of resin
The hole diameter d0 = d + 2 = 24 + 2 = 26 mm with ks = 1. For normal holes t1/t2 = 25/20 = 1,25,
so for β (see fig. 3.18):
t1
β = 1,66 – 0,33 = 1,66 – 0,33·1,25 = 1,25
t2
The bearing resistance Fb,Rd,resin follows from equation (3.18) with kt = 1,2 for the ultimate limit state:
t b,resin = t1 = 25 mm 1,5d = 1,5·24 = 36 mm
k t ks dt b,resin βfb,resin 1,2·1,0·24·25·1,25·130·10 –3

Fb,Rd,resin = γM4 = = 117 kN
1,0
For each of the twelve individual bolts:
3.20 Bolted connection in double shear with section I strip 280x25

non-preloaded injection bolts in normal holes.
TEd TEd
cover plate 270x20

Shear
The shear resistance Fv,Rd of an (injection) bolt in single shear with the shear plane running
through the thread according to equation (3.5) is:
α v fub A s 0,6·800·353·10 –3
Fv,Rd, single shear = γM2 = = 136 kN
1,25
In this connection the bolts are in double shear, and therefore:
Fv,Rd,double shear = 2Fv,Rd,single shear = 2·136 = 272 kN
Fv,Ed = 112 kN ≤ Fv,Rd,double shear = 272 kN (OK)
Tension resistance of the cross-section

Yielding of the gross cross section and fracture of the net cross section of the connected plate
must both be checked. The plate will be more critical than the cover plates, because Aplate =
25x280 = 7000 mm2 ≤ Acover plates = 2·20·270 = 10800 mm2.
• Fracture of the plate at section 1 is checked using equation (3.8):
( )
A net = b – 2d0 t1 = ( 280 – 2·26 ) ·25 = 5700 mm2
0,9A net fu 0,9·5700·360·10 –3

Nt,Rd = γM2 = = 1477 kN
1,25
The design value of the tension normal force Tt,Ed in section 1 is:
11
TEd,section 1 = TEd – contribution left bolt = T
12 Ed
11
= ·1340 = 1228 kN Nt,Rd = 1477 kN (OK)
12
• Yielding of the gross cross-section of the plate at section 1 is checked using equation (3.7):
Afy
7000·235·10 –3
Nt,Rd = Npl,Rd = = = 1645 kN
γM0 1,0
TEd,section 1 = 1228 kN Nt,Rd = 1645 kN (OK)

rivet head
grip
length
round
second
head
flat countersunk
before riveting after reveting
round countersunk
3.21 Example of a riveted steel bridge structure. 3.22 Types of rivets.
3.3 Rivets
Although rivets are practically no longer used in modern steel structures, some knowledge is
desirable because existing riveted structures still have to be maintained frequently (fig. 3.21).
The quality of a riveted connection strongly depends on how it is made, namely by hammering
or by pushing (pressing). When riveting, the rivet itself is first heated to a bright red, almost white
glowing state. The shank is then inserted through the hole, which is 1 mm larger than the shank
diameter of the rivet. The shank is then pushed in using a (pneumatic) hammer or press. The
hole is thereby filled first, after which a second head is formed (fig. 3.22). A small gap between
shank and hole wall is inevitable (not shown in fig. 3.22), because the rivet cools more than the
connecting parts. The large axial tension stress that occurs in the rivet – the two heads prevent
shrinkage of the shank as it cools down – also reduces the thickness of the shank.
Rivets are currently available in a limited numbers of lengths, which means that the total thick-
ness of the parts that can be connected (the grip length) is also limited. The grip length depends
on the method of riveting. EN 1993-1-8, cl. 3.6.1(16) limits the grip length as a function of the
diameter d of the rivet:
– for hammer riveting: ≤ 4,5d
– for press riveting: ≤ 6,5d
Design
Rivets are suitable for connections with rivets in single shear, double shear and multiple shear. It is
assumed that the load transfer takes place exclusively through shear and bearing of the rivet shank.
The considerable frictional force that is created between the connecting plates, which occurs due

to the large tension force in the rivet, is not taken into account in the design of a riveted connection.
Design values for the resistances of rivets are shown in table 3.4 of EN 1993-1-8. The shear resistance
Fv,Rd per shear plane is given by:
0,6 fur A 0
Fv,Rd = γM2 with γM2 = 1,25 (3.19)
Where:
A0 area of the rivet hole in mm2;
fur design value of the tension strength of the rivet material.
The bearing resistance Fb,Rd of a rivet is equal to that of a ‘normal’ bolt, with d the nominal
diameter of the rivet:
k1α b fudt
Fb,Rd = (= 3.6)
γM2
The tension resistance is:
0,6fur A 0
Ft,Rd = γM2 (3.20)
Equation (3.20) suggests – because of the reduction factor of 0,6 – that rivets are less suitable for
connections with an axial tension force on the nail head. For rivets in combined tension and shear
the same interaction formula is used as that for non-preloaded bolts in normal holes:
Fv,Ed Ft,Ed
+ 1,0 (= 3.17)
Fv,Rd 1,4Ft,Rd

In case of a widely varying load – such that the maximum shear load is always reached at the
serviceability limit state – rivets can slide in their holes. As a result, the tension force in the rivet
shank partly, or even completely, disappears and the rivets become loose. This can be prevented
by assuming a lower limit for the shear resistance. A safe, or reduced value for use in practice is:
0,26fur A 0
Fv,Rd,red = γM2 (3.21)
For rivets a check for punching resistance is not necessary, see table 3.4 of EN 1993-1-8.

3.4 Pins
Pins are mainly used in (most often exposed) hinged connections, or in structures that must be
easily disassembled (fig. 3.23). The connections simply comprise a round pin in a hole.
Connections using pins are discussed in EN 1993-1-8, cl. 3.13, where table 3.9 provides the
requirements for the plate geometry (fig. 3.24). Table 3.10 of EN 1993-1-8 defines the checks
that must be carried out for connections made with pins. For pins that must be replaceable an
additional design rule for the serviceability limit state applies for both bearing resistance Fb,Rd
and bending resistance MRd, see equation (3.24) respectively (3.26):

(3.22)
(3.23)

(3.24)
(3.25)
(3.26)
(3.27)
3.23 Example of connections made with pins

used in the diagonals of a truss.

Where:
A cross-sectional area of the pin;
fup ultimate tension strength of the pin;
t thickness of the connected part;
d diameter of the pin;
fy lower of the design strengths of the pin and the connected part;
fyp yield strength of the pin;
Wel elastic section modulus of the pin (Wel = πd3/32).
In addition to these checks, EN 1993-1-8, cl. 3.13.2(3) includes a requirement for the contact
bearing stress sh,Ed for a pin that must be replaceable:
σ h,Ed = 0, 591
(
EFEd,ser d0 – d ) 2,5 fy
with γM6,ser = 1,0 (3.28)
2
d t γM6,ser
Where:
E elastic (or Young’s) modulus;
FEd,ser design value of the force that is transferred by bearing, determined for the characteristic
load combination for the serviceability limit state.
3.24 Requirements for the dimensions of

at a given plate thickness t at a given geometry a connection made with pins.
d0 a 1,6d0
c 0,75d0
FEd d0 FEd d0 2,5d0

1,3d0
c 0,75d0
0,3d0
FEd t FEd t
FEd γM0 2d0 FEd γM0 d0 FEd γM0

a≥ + and c≥ + t ≥ 0,7 and d0 ≤ 2,5t
2 t fy 3 2 t fy 3 fy

3.5 Literature
2. EN 1993-1-8 (Eurocode 3. Design of steel structures. Part 1-8. Design of joints), 2005 (incl. AC,
2009).
3. EN 14399-1 (High-strength structural bolting assemblies for preloading. Part 1. General require-
ments), 2015.
4. EN 14399-2 (High-strength structural bolting assemblies for preloading. Part 2. Suitability for
preloading), 2015.
5. EN 14399-3 (High-strength structural bolting assemblies for preloading. Part 3. System HR.
6. EN 14399-4 (High-strength structural bolting assemblies for preloading. Part 4. System HV.
7. EN 14399-5 (High-strength structural bolting assemblies for preloading. Part 5. Plain washers),
2015.
8. EN 14399-10 (High-strength structural bolting assemblies for preloading. Part 10. System
HRC. Bolt and nut assemblies with calibrated preload), 2018.
9. EN ISO 286-2 (Geometrical product specifications (GPS). ISO code system for tolerances
on linear sizes. Part 2. Tables of standard tolerance classes and limit deviations for holes and
shafts), 2010 (incl. AC, 2013).

4
Connections
Welds

emeritus professor steel structures, Faculty of Civil Engineering and Geosciences, Delft University of Technology
welds | connections 4 | 1
Welds
Together with bolts, see Connections 2 (Bolts in clearance holes), welding is a primary joining
technique for steel structures in the built environment. This chapter gives an overview of important
welding procedures and of frequently applied weld types. Subsequently, an overview of the
procedures used to achieve a good weld, and the methods to identify and assess possible
welding imperfections, is given. Finally, the principles for designing fillet welds and butt welds
are considered.
4.1 Welding processes

Welding is a process whereby, under the supply of heat or pressure – or a combination of both – a
continuous connection is formed between two or more metallic elements (fig. 4.1). Current welding
processes are mainly distinguished by the way the heat is supplied, if – and how – extra weld
material is applied, and the way the metallic elements are pressed together or onto each other.
Figure 4.2 gives an overview of the classification of welding processes according to EN ISO 4063,
with corresponding numerical codes. Not all of these processes are used for the manufacturing
of steel structures. Those most relevant to steel construction are shown in red in figure 4.2. They
belong to two main groups, namely arc welding and resistance welding, with the former group
being the most common. Only arc welding and resistance welding are considered below.
4.1.1 Arc welding

When using arc welding, both parts – the parent material – are locally heated to a liquid state.
Usually, the (weld) filler material is applied in a liquid state, and its composition is as close as
possible to that of the parent material at any given time. After solidification, a robust connection
with a minimum strength corresponding to that of the parent material is created. An exception
to this general rule is that for welding high-strength steel grades often (weld) filler material with a
lower strength than that of the parent material is used, in order to improve deformation capacity.
4.1 Welder at work.
2 | connections 4 | welds
welding processes
arc welding resistance welding gas welding pressure beam welding

(1) (2) (3) welding (4) (5)
spot (21) oxy-acetylene ultrasonic (41) electron beam
seam (22) (311) friction (42) (51)
submerged projection (23) friction stir (43) laser (52)
miscellaneous arc (12) gas arc (13) TIG (14) plasma arc (15) flash (24)
stud arc (78) MIG solid TIG solid stud/dowel (26)
manual metal electrode (131) wire/rod (141) high frequent (27)
arc (111) MIG flux TIG without
self-shielded tubular- electrode (132) filler material(142)
cored arc (114) MAG solid TIG flux-cored
electrode (135) wire/rod (143)
MAG flux electrode
(136)
MAG metal powder
flux electrode (138)
4.2 Classification of welding processes according to EN ISO 4063 with

corresponding number code shown in brackets. The more common
When arc welding is used it is essential that both the liquid weld metal welding processes for modern steel structures are shown in red.
and the parent material are protected against atmospheric influences,
in particular the action of oxygen and nitrogen. Depending on the way
this action happens, four types of (electric) arc welding process can be
distinguished (fig. 4.2):
– arc welding with gas and slag protection (manual metal arc welding,
number 111);
– arc welding with slag protection (submerged arc welding, number 12);
– arc welding with gas protection (gas arc, number 13);
– stud welding (number 78).
electrode holder
Arc welding with gas and slag protection covered electrode
This welding process – referred to as SMAW (shielded metal arc welding)

clamp arc
– is mainly used in hand welding, whereby the welder performs all neces- elektrode cable
sary operations manually. A welding transformer or rectifier, that main- workpiece
tains a circuit with a (open) voltage of maximum 80 V (fig. 4.3), functions
arc welding
as the power source. The workpiece, the welding arc, and a covered power source
electrode – often called rod – are part of the circuit. The electrode is
clamped in an electrode holder that is connected to one pole of the
workpiece cable
power source via a cable. The other pole is connected to the workpiece
with a clamp. By ‘rubbing’ or ‘tapping’ the electrode on the workpiece,
the arc starts to burn. The welding arc remains intact as long as the elec- 4.3 Setup for hand welding (arc welding with gas and slag protection.
1 core rod
2 cover
3 end of core rod in liquid state
4 protection by (liquid) slag
5 electric arc
6 workpiece (parent metal)
7 weld pool
8 solidified slag
1
9 liquid slag
10 chalice, directs gas flows and metal drops 2
11 gas flow due to combustion of components of the
cover 3
12 drop of weld material (surrounded by slag) welding
13 solidified weld material 8 9 4 direction
12 10
5
13 7 11
4.4 Principle of arc welding with gas and slag trode is moved at the correct distance over the workpiece. The electricity in the arc is converted
protection.
to light and heat and as a result the parent material and the electrode are both heated to a liquid
state (fig. 4.4). The cover of the electrode melts and both gasses and slag are created to protect the
forming weld against the action of the outside air. In addition, a number of additives are released
from the coating to fulfill several functions, including:
– desoxidation of the liquid material;
– removal of impurities from the molten bath;
– stabilization of the arc;
– increase the recovery and quality of the molten material by the addition of alloying elements.
The slag that is generated from the coating is intended to float to the surface of the weld where
it forms a heat-insulating layer and protective layer against atmospheric influences. As a result,
the weld cools down less rapidly and this improves the structure of the steel. The slag is easily
removable with, for example, a wire brush or a chipping hammer.
Common diameters of electrodes (thicknesses of the core rod) typically range from 1,5 to 8 mm,
depending on the thickness of the weld to be realized. The length of an electrode is normally
around 400 mm. Electrodes are supplied with various coatings, for example basic, rutile (titanium
oxide), iron oxide, and cellulose. When selecting the electrode, one must pay attention to the
mechanical properties of the parent material, the desired appearance of the weld, and the welding
position. For structural applications a basic electrode is normally used. This results in a weld with
good mechanical properties, although its appearance will be less attractive than when using
a rutile electrode.
power source
reel
+ –
reel
power
source flux container electrode wire
weld slag
–
solidified unused supply
slag flux of flux
welding
machine welding
direction
solidified
workpiece weld material
welding
direction
backing strip
layer of flux parent material weld pool
The advantages of arc welding with gas and slag protection (covered electrode) are: 4.5 Scheme of submerged arc welding (SAW).
– possibility of application in all positions;
– low-heat input;
– good mechanical properties;
– low investment;
– suitability for any type of material;
– short time that is needed to set up the equipment;
– limited sensitivity to steel pollution.
The disadvantages of this process are:

– need for removal of the slag;
– weld spatter;
– risk of slag inclusions;
– low productivity.
Arc welding with slag protection

In this welding process – referred to as SAW (submerged arc welding) – an arc is drawn between
a bare wire or strip and a workpiece, and the arc is completely covered by a layer of flux (powder).
Submerged arc welding is an automatic process: the welding machine sits, for example, on a trolley
that runs next to the weld and has an adjustable speed. The flux is placed through a funnel to cover
the weld. The bare electrode is supplied from a large reel and is submerged in the flux, passing
4.6 Submerged arc welding of two large hollow
through a current inductor (fig. 4.5 and 4.6). The entire welding process is therefore effectively sections. Here, the welding equipment is fixed and
invisible. A part of the flux forms into slag, the remaining powder is wiped away and re-used. the hollow section moves (rotates).
Submerged arc welding uses a much higher current, compared with hand welding. This increases
the melted pool and penetrates the weld deeper into the parent material. This can often partially
(or entirely) avoid the need for pre-processing of butt welds. Since a lot of weld material can be
melted down in a short time, this process is very suitable for the installation of long horizontal
welds in, for example, bridge decks and plate girders.
The advantages of submerged arc welding are:

– low heat input;
– high welding speed;
– good penetration;
– little discoloration;
– no weld spatter;
– no radiation, because the arc is not visible;
– little smoke;
– long operating time (infrequent stopping), so time spent actually to welding is maximized.

– can only be used horizontally;
– significant investment;
4.7 Scheme of arc welding with gas protection
– long setup time;
(MIG/MAG welding). – not suitable for thin material.
power source Arc welding with gas protection

reel
Arc welding with gas protection is a fully or semi-automatic welding
+ – process, in English often known as GMAW (gas-shielded metal arc
welding). The gas protects the liquid weld and parent material
against atmospheric influences (fig. 4.7), meaning that no slag is
MIG welding inert gas formed unless a flux cored wire elecrode is used. The following
(e.g. argon)
MAG welding active gas electrode wire types can be distinguished (see fig. 4.2):
(e.g. CO2) – MIG welding (number 131, 132 and 133);
– MAG welding (number 135,136, and 138).
• MIG welding (metal inert gas). This welding process uses an inert
gas nozzle contact tip
gas and a melting electrode. An inert gas, also known as a noble gas,
does not react with the weld metal. The electrode is a continuously
gas shielding
solidified welding supplied welding wire surrounded by a nozzle, through which the
weld material direction
protective gas is fed. Originally, MIG welding was developed using
pure argon, or a mixture of argon and helium. It is mainly used for
parent material weld pool welding aluminium.
• MAG welding (metal active gas). This welding procedure is similar to MIG welding; however,
instead of noble gases carbon dioxide (CO2) or a mixed gas is used, because noble gases are
not suitable for welding steel. Carbon dioxide was mostly used in the past, with the welding
procedure sometimes known as CO2 welding. In MAG welding the protective gas is not therefore
inert. In general, both solid and flux-cored electrode wires are used. A flux-cored wire consists of
a soft steel shell containing a slag-forming powder. The advantage of a flux-cored wire is, inter
alia, a high efficiency, thus more molten material in a given time (per minute), and an improvement
in the appearance of the weld. Compared to MIG welding, the probability of weld imperfections –
such as binding errors and inadequate penetration – is considerably larger with MAG welding.
The advantages of arc welding with gas protection (MAG welding) are:
– no cover, and therefore no slag;
– low heat input;
– higher welding speed;
– long operating time (infrequent stopping), so time spent actually to welding is maximized;
– good protection against oxidation, also adjacent to the weld.

– sensitive to wind;
– a backing strip is necessary for the placement of a good root seam, because of the difficult
manageability of the welding procedure, see figure 4.5 and 4.10;
– difficult to control due to the high melting rate;
– annoying welding arc for the welder and the people around.
Stud welding
Stud welding (number 78) is used to weld studs, threaded rods, or anchors onto a steel element.
The product to be welded is inserted in a special holder that is then located by the welder.
Forming of the arc and pressing the product into position is fully automatic. A ceramic ring at the
base functions as a shield to create the right weld environment. The most important application
in steel construction is welding shear connectors (headed studs) for composite structures (fig. 4.8);
see also EN ISO 14555 and EN ISO 13918. When steel decking is used to form a composite
floor, the studs can be welded directly through the steel sheet onto the steel beam (thru-deck
welding), see also [3].
4.8 Stud welding of studs for a

composite beam. The picture shows
the welded connection itself.
water cooled
copper electrode
4.9 Examples of resistance welding. a. spot welding b. seam welding
4.1.2 Resistance welding

With resistance welding, the parts to be connected are brought in a semi-molten (dough-like)
state by a passing current and pressed together. This results in a connection without the addition
b. seam welding
of extra material. Resistance welding is used in the fabrication of building elements comprising
thin steel plates, adopting either point welding or roll welding (fig. 4.9). In point welding (number
21) two thin plates are pressed together with copper electrodes and at the same time an electric
current is passed through. The plates are locally heated due to the resistance encountered by
the electric current. In roll welding (number 22), two disc-shaped electrodes are used and as
the plates move between the rollers a continuous series of point welds is formed due to current
pulses.
4.2 Weld shapes

In arc welding, the weld itself is created by adding liquid material to a weld seam. The weld seam is
the enclosed space between prepared, or unprepared, plate edges before the start of the welding.
The cross-section of the weld seam determines the weld shape, for example an X, K or V weld.
EN 1993-1-8, cl. 4.3 distinguishes the following weld shapes:
– fillet welds;
– fillet welds all round;
– full penetration butt welds;
– partial penetration butt welds;
– plug welds;
– flare grove welds.
The most common types of connections and welds are defined in EN ISO 17659. Figure 4.10
shows some common terms used in welding, with a butt weld taken as the example. The relevant
information concerning a particular weld should be indicated on the construction drawings using
a standardized coding system according to EN ISO 2553. The basic notation consists of an arrow
with arrowhead and reference line. The type of weld and further information on its shape are
indicated with symbols, of which a few are given in figure 4.11. In addition, the throat thickness
(cross-section) of the weld, the longitudinal dimensions, and the welding process (through a code
number) are indicated on the drawings. Figure 4.12 shows an example of an interrupted double
fillet weld, where the MAG welding process with solid welding wire (number 135) is to be used.
included or
groove angle undercut
closed-V weld
with root faces
root face heat affected zone
root
reinforcement
V weld with
backing
back run intermediate

backing strip stringer beads
penetration
weld bead
4.10 Some commonly used terms in welding
crater (examples of butt welds).
basic symbols specific shape

1 single-V butt 1 single-V butt with convex section
2 double-V butt 2 double-V butt, polished flat on both sides
3 single bevel butt 3 single bevel butt, polished flat
4 double bevel butt 4 double bevel butt with convex section
5 fillet 5 concave fillet
6 double fillet 6 double concave fillet
1 2 3 4 5 6 1 2 3 4 5 6
dimensions of
possible cross-section longitudinal
special longitudinal weld dimension of
denotation weld symbol weld
number
welding
process
arrow line dashed reference line reference line
4.11 Coding system for indicating welds on steel

structure drawings according to EN ISO 2553. Only
the most important welding symbols are indicated.
a=4 2x20 (30)

135
2x20 (30)
20
30 mm
20
symbolic notation drawing 4.12 Example of weld notation.
4.2.1 Fillet welds
A fillet weld is a weld that is placed in the corner between two plates. The aim is not to melt the
plates over their full thickness (fig. 4.13). A distinction is made between a single and a double
fillet weld (fig. 4.14a/b). The nominal weld thickness a is equal to the throat thickness (figure
4.14a). The angle between the plates for a fillet weld should be between 60˚ and 120˚ (fig. 4.14c),
according to EN 1993-1-8, cl. 4.3.2.1(1). In general, it is not necessary to prepare (preprocess) the
plate edges. This is in contrast to butt welds, where the plate edges must normally be bevelled
(see fig. 4.10).
A fillet is the most commonly used weld type for connecting structural elements that are perpen-
dicular to each other. According to EN 1993-1-8, cl. 4.5.2(2), the minimum thickness of a fillet
weld is 3 mm. The most favourable position for obtaining a good fillet weld is obtained when the
4.13 Cross-section of a (single) fillet weld.
two plates are positioned at an angle of 45˚ during welding, and the surface of the weld seam
runs horizontal. This welding position is also called welding ‘in flat position’ (fig. 4.14d). Creating
a standing fillet weld, where the surface of the weld seam is 45˚, is difficult to achieve but pos-
sible due to the driving force of the arc (fig. 4.14e).
Fillet welds that are either ‘in flat position’ or ‘standing’ can be laid in a single pass provided they
are no greater than 6 mm weld thickness. For thicker fillet welds, multiple passes are needed
with the solidified slag removed before the next layer is placed (fig. 4.14f). Multiple passes are
necessary for ‘vertical’ and ‘overhead’ welds because in these cases the amount of molten filler
material in each layer must be limited.
a (nominal
throat thickness) ≤ 120˚
≥ 60˚
a. single fillet weld b. double fillet weld c. fillet welds with an angle
between the plates
flat position: in multiple

layers for a > 6 mm
always in multiple layers
d. flat position welding e. standing fillet weld f. filled welds in multiple layers
4.14 Applications of fillet welds. (upper weld)
Fillet welds are executed either continuously or non-continuously. Non-continuous fillet welds –
known as intermittent fillet welds – are only used in connections with a large weld length, or in
connections where the force to be transferred is limited. According to EN 1993-1-8, cl. 4.3.2.2(1),
intermittent fillet welds may not be used in corrosive environments because they would allow
moisture to penetrate the gap between the connected plates. Fillet welds have a low fatigue
strength and so are less suitable for structures loaded in fatigue. According to EN 1993-1-8, cl.
4.3.3(1), fillet welds in round and slotted holes – known as 'fillet welds all round' – may only be
used to transfer shear forces to prevent buckling or separation of lapped plates. According to
EN 1993-1-8, cl. 4.3.5(1) the same applies for plug welds (see section 4.2.4).
4.2.2 Full penetration butt welds

Butt welds are normally intended to achieve full penetration of the parent material (fig. 4.15),
unless there is a conscious decision to use a partial penetration butt weld, see section 4.2.3. Butt
welds are used for connecting plates which lay either in the same plane or meet with an angle.
For most butt welds, it is necessary to prepare (pre-process) the edges to be welded. The dimen-
sions of butt welds are standardized to ensure that weld symbols on structural drawings are suf-
ficient to fully define them in practice (see fig. 4.11). Non-standard weld shapes must always be
clearly indicated on drawings, for example at a 1:1 scale. The choice of the weld shape depends
on the plate thickness, the accessibility of the weld, and the available welding process. Butt weld
4.15 Cross-section of a butt weld.
shapes for automated welding processes are slightly different from the usual weld shapes for
manual welding. In general, less pre-processing can be sufficient depending on the automatic
welding procedure. The various weld types are briefly explained below (fig. 4.16).
• The simplest weld shape is the open square weld. Such welds are applied single-sided for
plate thicknesses up to 3-4 mm, and double-sided for plate thicknesses up to 6 mm (fig. 4.16a).
By using, for example, a special deep penetration electrode or an automatic welding procedure
– such as submerged arc welding – plate thicknesses of 15 mm may be welded.
40-60˚
4.16 Different weld shapes for butt welds. a. open square t h. double-V with
t = 3-6 mm root face
gap width t = 15-40 mm ±3 t
40-60˚
±3 mm
b. single-V
t = 6-25 mm t 40-60˚
2-3 mm i. asymmetric
2
double-V 3
t
50-70˚ t > 20 mm
1
c. single-V with root 3
t
gap
t = 6-25 mm t ±3 mm
10˚
±2 mm
j. single-U
45˚ t ≥ 40 mm t
d. single bevel
t = 5-25 mm t
±2 ±3 mm
±2 mm 10˚
45˚ k. double-U
t ≥ 60 mm
e. single bevel with ±3 mm t
prewelded bead
t
±2 mm 45˚
40-60˚
l. double bevel
f. single-V without
t = 15-40 mm t
root gap
t = 6-25 mm t
1,5-3 mm ±3 mm
40-60˚
40˚
m. double bevel with

g. double-V prewelded bead
t = 15-40 mm t
t
±3 mm
• The single-V weld is applied for plate thicknesses up to approximately 25 mm. A single-V weld
with pre-opening is suitable when gouging of the first layer is impossible (fig. 4.16b/c). Unaccept
able welding imperfections can be repaired by carving or cutting out, and re-welding. The single
bevel weld is mainly used in connections to existing structures (fig. 4.16d/e). The V weld without
root gap (fig. 4.16f) is a common weld shape. When complete penetration is required, gouging
of the first layer and counter welding is necessary. A disadvantage of the V weld is its asymmetric
shape, which can result in an unacceptable rotation of the connected elements due to transverse
shrinkage (fig. 4.17). Such rotations can be remedied by hot or cold forming.
• The double-V weld is more suitable than the single-V weld for plate thicknesses of 15-40 mm
(fig. 4.16g/h), partly because deformations due to transverse shrinkage are much less. In addition,
the weld content of a double-V weld is only half that of a corresponding single-V weld. For
double-V welds both edges of the work piece must, of course, be freely accessible. In some cases,
for example in tubular structures or at locations that are hard to access, it might be useful to choose
an asymmetric double-V weld (fig. 4.16i). Although a double-V weld has advantages – including
less weld filler material and lower labour and electricity costs – there are higher costs associated
with preparing the weld edges and turning the workpiece. Another disadvantage of a double-V
weld is the attention that must be paid to the accuracy of the plate preparation. It is important
that the points (noses) of adjacent plate edges lay precisely opposite each other during welding.
A better physical dimensioning is possible for both the open and closed single-V and double-V
weld by avoiding sharp edges. Gouging before the back layer is laid is also necessary for double-V
welds when complete penetration is required.
• For plate thicknesses greater than 40 mm, a wide single-U weld or single-U groove is sometimes
specified (fig. 4.16j). Plate preparation for this weld is costly, but the costs for actual welding and
the consumption of weld filler material are lower than those of a corresponding single-V weld. The
double-U weld is used only for plate thicknesses of 60 mm or larger (fig. 4.16k); for plate thickness
up to 60 mm an double-V weld is cheaper, because the extra preparation costs for a single-U weld
outweigh the savings in the costs of the welding processes and consumables.
• A double-bevel weld is suitable for connecting plates that either lay in line with each other
(fig. 4.16l) or are perpendicular to each other. For plates
perpendicular to each other the penetration in the con-
tinuous plate is not optimal, because the applied heat
flows away rapidly. Better weld penetration occurs if
some weld beads are first laid on the continuous plate,
before the continuous weld is laid (fig. 4.16m).
4.17 Rotation of a V weld due to transverse shrinkage.
4.2.3 Partial penetration butt welds
A partial penetration butt weld is a butt weld in which only limited penetration is achieved.
Full penetration butt welds are rather expensive to execute due to the checks required
on welding imperfections, and the associated probability of rejection and repair. This cost
may not be justified because it is not necessary for all applications that butt welds achieve
full penetration. The drawings must show when partial penetration is acceptable. The
cross-section of the weld is smaller than that of the plate for partial penetration butt welds
a. butt joint with b. T-joint with partial penetration,
partial penetration reinforced with superimposed (fig. 4.18a). The partial weld must clearly be strong enough to withstand the acting forces.
fillet welds
According to EN 1993-1-8, cl. 4.7.2, a partial penetration butt weld should be designed
4.18 Partial penetration butt welds. using the same method as for a fillet weld with deep penetration, given in EN 1993-1-8,
cl. 4.5.2(3).
When plates lie perpendicular to each other a partial penetration butt weld can be reinforced
with superimposed fillet welds (fig. 4.18b). When the requirements of EN 1993-1-8, cl. 4.7.3(1)
are met, a partial penetration butt weld reinforced with superimposed fillet welds may be
treated as a full penetration butt weld. Partial penetration butt welds are not suitable in cases
where a varying load may cause fatigue failure.
doubler plate plug weld fillet weld
column web
4.2.4 Plug welds and flare groove welds

A plug weld is a weld in a round or slotted hole, in which the hole is fully welded. Plug
welds are, amongst other things, used in beam-to-column joints to connect doubler plates
that strengthen the column web (fig. 4.19). According to EN 1993-1-8, cl. 4.3.5, plug
welds may only be used:
– to transfer shear forces;
– to avoid buckling or separating of two parts laying on top of each other;
– to connect the components of composite elements.
detail
Plug welds may not be used to transfer tension forces.
4.19 Example of the application of plug welds.
A flare groove weld (also called a flare V weld) is a weld placed on the curved edges
of sections (fig. 4.20), or between a circular bar and a plate. The name comes from the
fanning movements that must be made by the welder to execute this type of weld. The
effective throat section for a flare groove weld is given in figure 4.2 and figure 7.5 of
EN 1993-1-8.
4.20 Throat cross-section of a flare groove weld between two

rectangular hollow sections.
4.3 Welding procedures
Using modern welding processes, it is possible to produce welds that meet strict quality require-
ments with respect to the shape and the uniformity of the weld. The quality of a weld depends not
only on the welding process. Equally important is a suitable material choice, a correct weld shape,
and dimensional accuracy of pieces before welding. The weld shape and the welding sequence
largely determine the amount of weld shrinkage. When deformation due to weld shrinkage is
prevented significant stresses can occur in the structure and in the weld. Cracking could occur.
Therefore, good cooperation between the structural designer and welding specialist is important
when considering significant welds as expertise from both can affect the quality and strength of
the welded connection.
Every welded structure must be sufficiently reliable and safe, and at the same time the costs of
the welding must be limited. In practice, this means that a compromise must be found between
optimum reliability and safety, and minimizing costs. Checking of welds increases the assurance
of quality, but also results in extra costs.
EN 1090-2, cl. 7.4 prescribes when and how the variables that influence the quality must be
specified and documented. The level of requirements depends on the execution class, see also
Annex A3 and Annex B of EN 1090-2. Depending on the execution class, use is made of:
– welding procedure specification (cl. 7.4.1.1);
– welding procedure qualification (cl. 7.4.1.2);
– welder’s qualification.
Welding procedure specification

A welding procedure specification (WPS) describes the details of the welding procedure and
serves as the basis for the execution of the welds. It forms a part of the welding plan that is
mentioned in EN 1090-2, cl. 7.2.2a. The welding procedure specification contains information
including the number of passes, the weld filler material, the process, the welding position, the
weld shape, the welding current, and (if needed) pre-heating temperature. All parties involved
(workshop, client, and inspecting authority) sign the welding procedure specification for approval.
Procedures should be visible in the work place so it is easy to confirm that a welder is working in
accordance with the agreed procedure. The welding procedure specification is often based on prior
knowledge and experience. When a new welding procedure is to be used – for an uncommon detail,
such as when connecting very large material thicknesses – it may be necessary to determine
from trial welds if the proposed welding procedure is good. This is called a welding procedure
qualification.
Welding procedure qualification

A welding procedure qualification is the report on testing of a trial weld. Under the supervision
of an independent expert, a trial weld is executed using the procedure that has to be qualified.
Subsequently, test pieces are cut from the trial element to be used for both destructive and non-
destructive examinations. The results of this testing are set out in the welding procedure qualification,
and this document is attached to the welding procedure. A welding procedure qualification need
not be made for every new project, as it is transferrable (within the specified range of validity).
Welder qualification
A welder qualification is a report that indicates if a welder, or operator of a mechanized proce-
dure such as arc welding, is capable of welding according to a certain standard. An independent
body carries out the assessments for the welder qualification. Usually, a welder qualification is
valid for two years, and after that period it must be demonstrated that the welder has delivered
‘satisfactory work’ for the certificate to be extended for a new period of time. In the case of
significant welds, the welder qualification alone is not sufficient to assure quality; in some cases,
existing structures are also tested randomly using non-destructive methods.
4.4 Weld quality

The level of the quality requirements placed on the welds depends on the implementation class
of the structure or structural element. Annex B of EN 1090-2 specifies how the implementation
class can be determined. Completely fault-free welds are barely feasible economically, so it is
important to know which type of welding imperfections can be dangerous – see section 4.4.1 –
and when these imperfections should be repaired. EN 1993-1-8, cl. 4.1(3) and EN 1090-2, cl. 7.6
both declare that EN ISO 5817 is applicable to ensure a certain quality level. This ISO-standard
provides the acceptable limits of the imperfections for three quality levels: D (least strict), C, and
B (strict). EN 1090, cl. 7.6 indicates which quality levels apply to different implementation classes.
4.21 European standards for welding inspection
Welding imperfections can be detected with either destructive or non-destructives tests, see
and assessment of welds. section 4.4.2.
EN 1993 NEN 1090-2

design of steel structures execution
EN ISO 5817
quality levels
EN ISO 17635
quality level coherence
and acceptance criteria
radiografic pulsed current magnetic penetrant ultrasonic visual

non-destructive tests EN ISO 17636-1/-2 EN ISO 17643 EN ISO 17638 EN ISO 3452 EN ISO 17640 EN ISO 17637
characterization errors EN ISO 10675-1 – EN ISO 23278 EN ISO 23277 EN ISO 23279 EN ISO 5817
acceptance critaria EN ISO 10675-1 in specification EN ISO 23278 EN ISO 23277 EN ISO 11666 EN ISO 5817
The execution of non-destructive testing methods is regulated by European standards. The accep-
tance criteria are coupled to the quality levels in EN ISO 5817 through EN ISO 17635. Figure 4.21
gives an overview of the standards system.
4.4.1 Welding errors

Welding imperfections usually occur due to a combination of circumstances. The most important
factors are:
– welding process and welding procedure used;
– properties and quality of the weld filler material (for example, the type of electrode that is
used);
– weldability of the parent material;
– competence of the welder.
Some additional circumstances often play a role in the occurance of welding imperfections. A
number of these are briefly discussed below.
• Poor accessibility of the weld. Figure 4.22 shows examples of welds that are poorly accessible
(1), and therefore cannot be placed properly. When the angle between two sections is smaller
than 30˚ this can also cause major problems for welding (2). In these cases, serious welding errors
are almost inevitable.
• Excessive seam tolerances. Examples of these are very large or small openings and bevels.
• Poor welding position. Practice has shown that the likelihood of weld imperfections increases
for vertical or ‘overhead’ welding. Design a welded steel structure so that horizontal (or flat) welding
is possible.
• Unfavorable welding conditions. Examples include temperature of the environment, bothersome
wind, moisture, and narrow spaces without or with insufficient smoke extraction. Unfavourable
conditions in the workplace quickly reduce the attention of the welder, and thereby increase the
likelihood of weld imperfections.
• Inadequate supervision of the welder. Particularly for difficult welding work, proper support in
advance – for example, a metallurgist explaining and discussing weld pool behaviour – can be of
great help.
α α
2 2
4.22 Poor accessibility of the weld due to
incorrect design.
dp gp,si Figure 4.23 shows the most important imperfections found in a welded connection:
u o u
lack of fusion insufficient adhesion between the weld filler material
lof w
ip
and parent material: the parent material is not melted
c dn np
sufficiently;
c undersized the weld surface is below the surface of the parent
c crack np negative penetration material;
dn incompletely filled groove o overlap
dp excess weld metal si slag inclusion
oversized the weld sits higher than the parent material;
gp gas pore u undercut gas pore inclusion of gas;
ip incomplete penetration w wormhole
lof lack of fusion wormhole channel through which gas has escaped;
negative penetration fillet weld does not lie exactly in the corner;
4.23 Possible welding faults.
incomplete penetration material is not fused over the entire thickness;
overlap weld sits over the parent material;
undercut at the transition between the parent material and so-
lidified weld, the parent material is melted away, and
therefore this sits lower than the surrounding parent
material;
crack crack in the heat-affected zone or in the weld filler
material;
slag inclusion slag material which is enclosed by the weld.
a. b. c. Figure 4.24 shows some radiographs of weld imperfections. Research and expe-
rience show that some weld imperfections have little or no impact on the strength
of the weld, while other imperfections significantly reduce the strength. Not only
the size and shape of the imperfections plays a role, but also whether the faults
have sharp edges. The latter, for example, leads to conditions with respect to
joining faults that are considerably more stringent than those for gas and slag
inclusion. Cracks are never allowed.
The scope of EN ISO 5817 covers full penetration butt welds, fillet welds and
partial full penetration butt welds. The ISO-standard gives limits for each type of
imperfection for the three quality levels (B, C, and D) in table format. A distinction
is made between imperfections that occur at the surface and internal imperfec
tions. The quality levels refer to the quality in the production and not to the fitness
for purpose. The acceptability of certain welding imperfections depends, among
others, on the following factors:
– levels of stress or imposed strains;
– toughness of the weld filler material;
– load condition: a static load or a fatigue load;
4.24 Radiographs (photos of radiographic inspection)
– consequences of failure;
of welds with faults: (a) crack, (b) gas pores, and (c) slag
inclusion. – contractual agreements.
4.4.2 Welding inspection
Welding imperfections can be detected with either destructive tests (DT) or non-destructive
tests (NDT). Destructive tests provide direct information on the weld quality and the influence of
imperfections in the weld, but are not suitable for checking existing steel structures. In the case
of non-destructive tests, the weld is examined without causing any damage. Destructive tests
may be used in combination with non-destructive tests, for both welding procedure qualification
and welder qualification, and sometimes for mandatory production tests.
Types of non-destructive test are: visual, magnetic, penetrant, ultrasonic, and radiographic
inspection.
Visual inspection
With visual inspection welds are examined with the bare eye. A competent inspector can
tell from the appearance of the weld whether or not a competent welder did the work and 4.25 Visual inspection provides sufficient
information to reject the weld.
used the correct welding method (fig. 4.25). For statically loaded structures, visual inspection
is often enough. In addition to the examination of the welds, the following aspects can be
included.
• Prior to the assignment: check whether the workshop has a good reputation, and there-
fore may be considered to be suitable for the fabrication of a given type of welding work;
• Prior to the start of the work: check if the welders are competent for the required welding
work and possess the relevant welder qualification (certificates). In case of doubt over the
quality of the welder, or for the workshop in general, the client can ask for trial specimens or
samples which can be tested in advance;
• Prior to welding: check the pre-processing (plate preparation), setting, adherence, and
root gap, and check if the specified tolerances are respected;
• During welding: supervise the execution of the welding procedure to ensure it is carried
out correctly; 1 2
crack penetrate with liquid

• After welding: assessment of the appearance of the weld.
Penetrant inspection
Penetrant inspection is a widely used and relatively simple method to detect cracks. More
than a century ago, cracks in shafts were detected by immersing them for twenty-four hours
in oil, then drying them off and rubbing with chalk. The cracks were visible as damp spots in 3 4
the chalk. Penetrant investigation is based on this principle. First, the penetrant – a liquid – is drying the surface apply developer (chalk)
applied to the object to be examined. After the liquid has been able to penetrate for some
time, the material is properly dried off and then sprayed with a developer (mixture of chalk,
fig. 4.26). Only surface imperfections can be detected with this method.
4.26 Principle of penetrant inspection. crack becomes visible
Magnetic inspection
With magnetic inspection, the test piece is rubbed with a white, chalky liquid.
Subsequently, it is covered with a suspension of iron particles in oil. When an
electric magnet is placed on the piece the iron particles align in the direction of
the magnetic field (fig. 4.27).
When the weld includes a crack, the magnetic field lines are locally deflected.
The indication is most clear when the crack is perpendicular to the magnetic
crack
field. Using this method, very small cracks up to 0,001 mm can be detected by
4.27 Principle of magnetic inspection. the naked eye, although it is clearly only possible to see surface imperfections.
Ultrasonic inspection
transmitter/receiver With ultrasonic inspection, a transmitter is placed on the material to be examined.
weld defect The transmitter emits ultrasonic waves with a frequency of tens of kilohertz. These
waves are reflected from interfaces: not only the opposite surface, but also the
o inner boundary surfaces formed by weld defects. The location of the imperfection
ech
can be determined by considering the length of time that the sound is in transit
(fig. 4.28). By moving a transmitter/receiver around the imperfection, a reasonably
weld contour ultrasonic waves workpiece clear picture can be created of the location and nature of it.
4.28 Principle of ultrasound inspection.

Radiographic inspection (X-ray inspection)
In radiographic inspection, a source that emits X-rays or gamma rays is placed on
one side of the weld to be tested. A film strip or a digital sensor is placed on the
other side and is illuminated by the radiation that is sent through the weld (fig. 4.29).
Because steel reduces the intensity of the radiation, in places where less metal is
present – for example, due to imperfections in the weld – the intensity changes
and dark spots appear on the film. The level of contrast is a measure of the
decrease of material, and thus of the imperfection dimension in the transverse
direction.
An advantage of radiographic inspection is that a directly readable shadow
image of the weld is produced. Inclusions and other three-dimensional welding
X-ray variations in shape, length, and location in width and longitudinal direction are
clearly visible (see fig. 4.24). As long as the deviations are not projected across
each other, they can also be identified as separate imperfections. The dimensions
horizontal crack vertical crack of an imperfection in the transverse direction can only be estimated based on
(doubling) will not appears as a
be detected by sharp black stripe levels of contrast. A disadvantage of radiographic inspection is that sealed cracks,
radiographic inspection
and cracks perpendicular to the radiation direction, are not found.
film strip or
digital sensor 4.29 Principle of radiographic inspection (X-ray inspection).
4.5 Weld design
EN 1993-1-8, chapter 4, provides rules for designing welded joints for steel grades up to S460
and for material thicknesses of at least 4 mm. For smaller material thicknesses reference is made
to EN 1993-1-3. The scope of the rules for welded hollow section joints in EN 1993-1-8, chapter 7,
is limited to wall thicknesses of at least 2,5 mm. EN 1993-1-12 gives additional rules for designing
welds in the higher steel grades S500 to S700. The design of shear connectors (studs) is covered
by EN 1994-1-1. EN 1993-1-9 applies for welds subject to fatigue, which therefore implies implicitly
that the rules in EN 1993-1-8 are applicable for mainly statically loaded welded connections.
EN 1993-1-8, cl. 4.3.1. distinguishes between the following five types of welds, the key features
of which are discussed above in section 4.2:
– fillet welds;
– fillet welds all round;
– partial penetration butt welds;
– plug welds;
– flare groove welds (flare V welds).
This section covers the design of fillet welds and butt welds. For the design of welded hollow
section joints reference is made to [2] and [4].
4.5.1 Fillet welds

The design of a fillet weld is based on the following assumptions:
– the critical cross-section is the throat section a (fig. 4.30a);
– the stresses are evenly distributed over the throat cross-section, unless this is contrary to equi-
librium, for example when the load is eccentric to the throat cross-section (figure 4.31).
FEd FEd 4.30 Throat section of a double

weld with the assumed and
t1 actual stress distributions.
a a a a a a
t2
a. throat section b. assumed stress distribution c. actual stress distribution
FEd For the design of a double fillet weld the stresses due to a tension force FEd are assumed to be
evenly distributed over the throat section a (fig. 4.30b). In reality, this is not the case (fig. 4.30c);
an even stress distribution will form only after local yielding. A single fillet weld is not suitable
e
for resisting a tension force perpendicular to the weld direction, because an unfavorable stress
distribution occurs due to the eccentric position of the weld (fig. 4.31).
EN 1993-1-8, cl. 4.12(1), states that local eccentricities of the applied force should be avoided.
If this is not possible these eccentricities must be considered in accordance to EN 1993-1-8,
cl. 4.12(2), when tension appears in the root of the weld. For single fillet welds over the circum-
ference of a hollow section, local eccentricities do not have to be considered in accordance with
–
+
EN 1993-1-8, cl. 4.12(3).
4.31 Eccentric tension force acting on a fillet The effective length ℓeff of a fillet weld is the length over which the weld has its full thickness.
weld.
To allow for reduced thickness at the start and end (head and run off) of the weld, EN 1993-1-8,
cl. 4.5.1, reduces the total length by two times the throat thickness. According to EN 1993-1-8,
cl. 4.3.2.1(4) wherever possible a fillet weld should continue around a corner (fig. 4.32), and then
the effective length may be taken as equal to the total weld length.
According to EN 1993-1-8, cl. 4.5.1(2) fillet welds with an effective length ℓeff < 30 mm or ℓeff < 6a
may not be included when determining the resistance of a connection. When fillet welds are
loaded in shear, the stresses are not evenly distributed (fig. 4.32). EN 1993-1-8, cl. 4.11 specifies
how this should be taken into account for overlay connections with long welds (ℓeff > 150a), see
also Connections 5 (Design of connections), section 5.2.4.
When, for example, a plate is welded perpendicular to an unstiffened flange of a section, the
stress distribution in the plate depends on the stiffness ratio between the web and flange of
the section (fig. 4.33). Since the web loaded in-plane is much stiffer than the flange loaded in
bending, an uneven stress distribution occurs. A similar effect occurs when a plate is connected
τ|| actual stress distribution

schematization
1
b
2 eff
FEd FEd tw
beff bp bp
r
1
b
2 eff
weld
around ℓeff
the corner tw
tp tp
4.32 Unequal stress distribution in the longitudinal FEd tf FEd

tf
direction of a weld. The fillet weld should continue
around the corner in order to utilize the full weld
length. 4.33 The effective width of an unstiffened T-connection.
to the wall of a hollow section. According to EN 1993-1-8, cl. 4.10 this uneven stress distribution
must be taken into account by using an effective width beff, see also Connections 5, section
5.5.1.
Depending on the load direction, the following stresses act on the throat cross-section of a fillet
weld (fig. 4.34):
– normal stresses σ perpendicular to the throat section;
– shear stresses τ and τ|| in the plane of the throat section perpendicular or parallel to the direc-
tion of the weld respectively.
Perpendicular to the cross-section acts a normal stress σ||. This normal stress may always be
neglected, because the cross-section of the weld will be relatively small compared to the cross-
section of the connected elements. When the weld yields, the longitudinal force in the weld
(Aweld σ || ) is then resisted by the parent material near the weld. EN 1993-1-8, cl. 4.5.3, gives two
methods for designing fillet welds:
– directional method (cl. 4.5.3.2);
– simplified method (cl. 4.5.3.3). 4.34 Stress components in a fillet weld.
Sometimes, it is also possible to determine the dimensions of a fillet weld directly from the stress
state in the plate material. This method is discussed at the end of this section.
Directional method
The check according to the directional method is based on the yield criterion of Von Mises (see
also [1]: Structural basics 7 (Resistance of cross-sections), section 7.5.2), in which the yield limit
fy is replaced by the tension strength fu.This replacement is consistent with other checks such as
the resistance of the net cross-section and the tension resistance of bolts. The check is as follows:
( )
fu
σ 2 + 3 τ 2 + τ||2 (4.1)
β w γM2
and
0,9fu
σ γM2 (4.2)
steel grade βw
Where: S235 0,80
fu nominal tension strength of the weakest connected element; S275 0,85
βw correlation factor, depends on the tension strength of the parent material;
S355 0,90
γM2 partial safety factor (γM2 = 1,25).
S420 1,00
S460 1,00
The correlation factor βw takes into account the difference between the strength of the molten
weld material and that of the parent material (table 4.35). 4.35 Correlation factor βw for fillet welds.
Equation (4.2) is only critical when σ is clearly higher than τ . The requirement in equation
(4.2) is necessary to avoid that the side area (see fig. 4.34) becomes more critical than the
360
throat section. This requirement is not critical in normal cases for symmetric fillet welds. For
asymmetric fillet welds and partial penetration butt welds, that can be designed as fillet
welds, equation (4.2) may be critical. The factor 0,9 in equation (4.2) is based on fracture
mechanics studies on partial penetration butt welds.
259
Figure 4.36 shows the relationship between the stresses σ and τ for grade S235 steel and
for τ|| = 0. With γM2 = 1,25, βw = 0,8 and fu = 360 N/mm2 equation (4.1) is:
180
fu 360
σ 2 + 3τ 2 = = 360 N/mm2
β w γM2 0,8·1,25
Some typical values are:
α = 45˚ σ = 0 N/mm2 3τ 2 = 360 τ = 208 N/mm2
0 144 180 208

σ = τ 4τ 2 = 360 σ = τ = 180 N/mm2
4.36 Relationship between σ and τ , for τ|| = 0, for
0,9 f 0,9·360
grade S235 steel. σ = γ u = = 259 N/mm2 2592 + 3τ 2 = 360 τ = 144 N/mm2
M2 1,25
In a similar way, these characteristic values can also be determined for other steel grades.
Figure 4.37 gives practical formulae for calculating the stresses for some common load cases.
Simplified method
In the simplified method, according to EN 1993-1-8, cl. 4.5.3.3 a check is made to ensure
that the mean stress – independent of the orientation of the plane of the throat section
relative to the applied force – is smaller than the design value of the shear strength of the
weld fvw,d:
fu
Fw,Ed 3
fvw,d = (4.3)
a∑ℓeff β w γM2

Where Fw,Ed is the design value of the applied force per unit length.
load case 1 load case 2 load case 3
ℓeff ℓeff FEd
a a
FEd
ℓeff
a
FEd
FEd 2 FEd 2 FEd

; τ|| = 0 ; τ|| = 0 τ|| =
4aℓeff 4aℓeff 2aℓeff
load case 4 load case 5 load case 6
ℓeff ℓeff
MEd
a a
B
C
A
b b ℓeff
MEd
a
MEd
MEd 2,12MEd
; τ|| = 0 In A and B: ; t|| = 0 Method using two forces (for ℓeff ≤ 2b and ℓeff ≤ 150a):
2aℓeffb aℓ2eff
FEd MEd
In C: τ|| = with FEd =
aℓeff b
Method using two moments (for ℓeff > 2b and ℓeff ≤ 150a):
2,12MEd
; τ|| = 0
aℓ2eff
4.37 Six common load cases with the corresponding stresses in the weld.
Example 4.1
2 • Given. A 120x15 mm strip of grade S235 steel, loaded by a tension force FEd = 420 kN, attached
1 to a 200x10 mm gusset plate (fig. 4.38).
• Question. Calculate the size of the fillet welds using the directional method according to
FEd 3 (U shaped) FEd EN 1993-1-8, when the two plates are connected with:
– only side fillet welds, each with ℓeff = 120 mm (fig. 4.38a);
– side fillet welds and an end fillet weld with ℓll,eff = 120 mm and ℓ ,eff = 120 mm (fig. 4.38b).
Check both strip and gusset plate.
120 mm
120x15 • Answer. The check of the fillet welds shall be in both cases only according to equation (4.1).
For an equilateral fillet weld, equation (4.2) is not critical.
200x10
a. only side fillet welds

Only side fillet welds
First of all, according to EN 1993-1-8, cl. 4.5.1(2), check if the effective length of the fillet weld
gusset plate strip
is sufficient to carry load:
ℓeff = 120 mm 30 mm (OK)
200 120 ℓeff 6a a 20 mm (OK, see below)
The shear stress τ|| in the plane of the throat cross-section is critical. At the given tension force
120 mm
FEd, σ = 0 and τ = 0 applies, so that:
120x15
200x10 (
σ 2 + 3 τ 2 + τ||2 )= 3τ||2 = 360 τ|| = 208 N/mm2
b. side fillet welds and an end fillet weld FEd 420·103 1750
τ|| = 208 = = = a = 9 mm
4.38 Connection between two plates (a) 2aℓeff 2·a·120 a
lateral fillet welds and (b) lateral fillet welds
and perpendicular fillet weld.
Side fillet welds and an end fillet weld
Initially, assume for the side fillet welds that the throat thickness a|| = 6 mm. The side fillet welds
can transfer a force up to:
F||,Rd = 2a||ℓ||,eff τ||,max = 2·6·120·208·10–3 = 300 kN
The end fillet weld must then be able to transfer:
F ,Ed = FEd – F||,Rd = 420 – 300 = 120 kN
The stresses in the end fillet weld are:
1 1 120·103 707
σ =τ = 2 = 2· = N/mm2
2 a ℓ ,eff
2 a ·120 a
According to figure 4.36 the maximum stresses are σ = τ : so σ ,max =τ ,max = 180 N/mm2.
1
The minimum throat thickness is: 2
2 Fend,Ed
15 mm
707
180 = = 4 mm Fend,Ed
1
2 Fend,Ed 10 mm
However, it is recommended to adopt an end fillet weld that is not smaller than the side fillet welds. 2
Tests have shown that when the thickness of the end fillet weld is smaller than the thickness of
the side fillet welds – failure will occur before the load is equally distributed over all welds. In
this example it is better to apply a|| = 5 mm and a = 6 mm (result of a new calculation) or – for
4.39 Decomposition of the force on a
practical reasons – for all welds a = 6 mm. perpendicular fillet weld.
Check the strip and gusset plate

The tension resistance Nt,Rd of the strip and the plate must be checked considering sections 1, 2
and 3 (fig. 4.38a). The checks for sections 1 and 2 concern yielding of the cross-section:
section 1 Nt,Rd = Astripfy = 120·15·235·10–3 = 423 kN ≥ FEd = 420 kN
section 2 Nt,Rd = Aplatefy = 200·10·235·10–3 = 470 kN ≥ FEd = 420 kN
The check for section 3 concerns (block) tearing of the gusset plate. EN 1993-1-8 does not ex-
plicitly treat this failure mode for welded connections. It is assumed that the design value for the
yield resistance is equal to the summation of the shear resistance of the parts of section 3 (U shaped)
parallel to the load and the tension resistance of the part of section 3 transverse to the load:
fy
235
section 3 Nt,Rd = 2ℓeff t p + bt p fy = 2·120·10· ·10 –3 + 120·10·235·10 –3
3 3
= 608 kN FEd = 420 kN
The shape of the failure line has a lot of similarities with that of a bolt group in block tearing,
which can be checked using EN 1993-1-8, cl. 3.10.2, see Connections 2 (Bolts in clearance
holes), figure 2.35 and equation (2.11). As an alternative, this method can also be used for the
check of section 3:
fy A nv 235·2·120·10·10 –3
f A 3 360·120·10·10 –3
3
Nt,Rd = u nt + = +
γM2 γM0 1,25 1,0
= 671 kN FEd = 420 kN
Example 4.2
• Given. The welded connection of an 120x15 mm strip to a 200x10 mm gusset plate in example 4.1
(fig. 4.38).
• Question. Calculate the size of the fillet welds using the simplified method according to EN
1993-1-8, when the strip and the plate are connected with:
– only side fillet welds, each with ℓeff = 120 mm (see fig. 4.38a);
– side fillet welds and an end fillet weld with ℓ||,eff = 120 mm and ℓ ,eff = 120 mm(see fig. 4.38b).
• Answer. The stresses in the fillet welds should be determined in both cases using equation (4.3).
Only side fillet welds

The effective lengths of the fillet welds are sufficiently large to carry the load, see example 4.1.
The throat cross-section is:
fu 360
Fw,Ed 420·10 3
3 = 3
= = fvw,d =
a∑ℓeff a·2·120 β w γM2 0,8·1,25
1750
= 208 a = 9 mm
a
This result is the same as with the directional method, because the welds are only subject to
shear stress.
Side fillet welds and an end fillet weld

The effective length of all welds together is ∑ℓeff = 2ℓ||,eff + ℓ ,eff = 2·120 + 120 = 360 mm. The
size of the fillet weld follows from the weld cross-section A:
Fw,Ed 420 ·103

A = a∑ℓeff = = = 2019 mm2
fvw,d 208
A 2019
a= = = 6 mm
∑ℓeff 360
The calculated weld thickness is larger than with the directional method, because the simplified
method does not take into account the greater resistance to normal stresses; compare equation
(4.1) with equation (4.3).
Example 4.3
• Given. A fin plate, with a length of 200 mm and in grade S235 steel, is connected to a column
flange with double fillet welds (a = 5 mm), figure 4.40. The welds are loaded by a horizontal force
FH,Ed = 350 kN and a vertical force FV,Ed = 190 kN. 20
• Question. Check the size of the fillet welds using the directional and simplified methods, a = 5 mm
according to EN 1993-1-8.
• Answer. Each fillet weld is subjected to a horizontal force of 0,5FH,Ed and a vertical force of
0,5FV,Ed.
Directional method FH,Ed

200
The stresses σ and τ in the throat section of one fillet weld can be found by taking components
of the force 0,5FH,Ed (see fig. 4.39): FV,Ed
1 1
2 ·0,5FH,Ed · 2 ·0,5·350·103
σ =τ = 2 = 2 = 124 N/mm2
aℓeff 5·200
1
F
2 H,Ed
The stress τ|| on the throat section of one fillet weld is given by: 1
2
FV,Ed
0,5FV,Ed 0,5·190·103
τ|| = = = 95 N/mm2
aℓeff 5·200
The check is carried out using equation (4.1) and (4.2): 1

FV,Ed
2
1
F
2 H,Ed
(
σ 2 + 3 τ 2 + τ||2 = ) (
1242 + 3· 1242 + 952 )
fu 360
= 298 N/mm2 = = 360 N/mm2 (OK) 4.40 Fin plate with double fillet welds
β w γM2 0,8·1,25 connected to a column.
and
0,9 fu 0,9 ·360
σ = 124 N/mm2 = = 259 N/mm2 (OK)
γM2 1,25
Simplified method
In the simplified method, the forces FH,Ed en FV,Ed are combined into one force Fw,Ed per unit
length:
2 2
Fw,Ed = FH,Ed + FV,Ed = 350 2 + 190 2 = 398 kN
Check with equation (4.3):

fu
Fw,Ed 398·10 3
3
= = 199 N/mm2 fvw,d = = 208 N/mm2 (OK)
aΣℓeff 5· ( 200 + 200) β w γM2
Example 4.4
FEd sin 60˚ FEd • Given. A gusset plate, with a length of 400 mm and in grade S235 steel, is connected
to a beam with double fillet welds (a = 4 mm). A force FEd = 400 kN acts at an 60˚ angle
FV,Ede1
in the eye of the gusset plate (fig. 4.41).
FEd cos 60˚
C • Question. Check the size of the fillet welds for the load cases shown in figure 4.37.
• Answer. The following forces are acting at the centre of the weld (point C):
FEd
FH,Ed = FEd cos 60˚ = 400·0,500 = 200 kN
a=4 a = 60˚ FV,Ed = FEd sin 60˚ = 400·0,866 = 346 kN

B C A
e1 = 75 IPE 240 MEd = FV,Ede1 = 346·75·10 –3 = 26,0 kNm
200 200 mm
To check the welds, load cases 1, 3, and 5 apply:
load case 1
+
FV,Ed 2 346·103· 2
load case 1 σ (1) = τ (1) = = = 76 N/mm2
+ 4aℓeff 4·4·400
load case 5
–
FH,Ed 200·103
+
load case 3 τ||(3) = = = 63 N/mm2
2aℓeff 2·4·400
– load case 1 + 5
ℓeff = 400 mm 150a = 150·4 = 600 mm (OK)
4.41 Gusset plate with a double fillet weld 2,12MEd 2,12·26,0·106

connected to a beam. load case 5 σ (5) = τ (5) = = = 86 N/mm2
aℓ2eff 4·400 2
The maximum stresses at point A:
σ = σ (1) + σ (5) = 76 + 86 = 162 N/mm2
τ = τ (1) + τ (5) = 76 + 86 = 162 N/mm2
τ|| = τ||(3) = 63 N/mm2
Check with equation (4.1) and (4.2):
( )
σ 2 + 3 τ 2 + τ||2 = (
1622 + 3· 1622 + 632 )
fu 360
= 342 N/mm2 = = 360 N/mm2 (OK)
β w γM2 0,8·1,25
and
0,9 fu 0,9 ·360
σ = 162 N/mm2 = = 259 N/mm2 (OK)
γM2 1,25

Design of fillet welds direct from the stresses in the plate material ℓeff
In some cases, it is possible to determine the weld thickness (throat cross-section) of fillet welds
directly from the stresses in the plate material. For example, take a plate loaded in tension with
thickness t and width b = ℓeff (fig. 4.42a). The normal stress in the plate is σx. The double fillet σx
weld is loaded by a tension force FEd = tℓeff σx. This value is introduced into the equation for load t
case 1 (see fig. 4.37):

FEd
FEd 2 t ℓeff σ x 2 t
σ = τ = = = σx (4.4) b
4aℓeff 4aℓeff 2a 2
a. tension stresses
If there is also a shear force VEd = tℓeff τxz acting on the plate, then for load case 3:
VEd tℓeff τ xz t
τ|| = = = τ (4.5)
2aℓeff 2aℓeff 2a xz
τxz
For combined tension and shear forces the required weld size a follows from equation (4.1):
VEd
b. shear stresses

4.42 A plate loaded in tension and shear with

fillet welds.
(4.6)
For a plate loaded in tension (or compression) only (load case 1), then τxz = 0, and so:
tβw γM2 t βw γM2 2 σ x

a 2σ 2x = (4.7)
2fu 2 fu
When the steel grade is known – and so the correlation factor βw – equation (4.7) simplifies for
load case 1 as follows:
σx
for grade S235 steel a ≥ 0,70t (4.8)
fu
σx
fu
σx
fu
S235 S275 S355 When joints in statically indeterminate structures are designed according to elastic theory it is
essential that yielding of the connected elements occurs before weld failure, see EN 1993-1-8,
σx = 0,8f y a ≥ 0,37t a ≥ 0,38t a ≥ 0,46t
cl. 4.9(4) and 4.9(6). In other words, the welds must be stronger than the parent material.
σx = f y a ≥ 0,46t a ≥ 0,48t a ≥ 0,58t
This condition is satisfied when a normal stress is assumed in the plate material of σx ≥ 0,8fy.
Table 4.43 Minimum weld size a of perpendicular When the structure is designed according to plastic theory, and the connection is situated
fillet welds for three different steel grades, for two
at a location where a plastic hinge can occur, according to EN 1993-1-8, cl. 4.9(5), the
characteristic values of σx.
minimum weld size should be determined on the basis of σx = fy.
Table 4.43 gives, for perpendicular fillet welds, the required weld size for both values of
σx. Here, t is the thickness of the connected plate or plate element.

Example 4.5
• Given. A bracket with a thickness of 12 mm and a length of 300 mm in grade S235 steel
is welded to a column. The bracket is loaded by a vertical force FEd = 280 kN (fig. 4.44).
• Question. Determine the dimensions of the fillet welds at points 1, 2, and 3 from the
stresses in the plate material.
• Answer. A shear force VEd = FEd = 280 kN acts on the weld, and a bending moment
MEd = FEde = 280·200·10–3 = 56,0 kNm. For the maximum cross-section of the bracket at
point 3 (free centre of rotation) the second moment of area Iy and the first moment of area
Sy are determined from simple mechanics:
e = 200 mm
12
FEd z
σx τxz
1
50 +
2
e1 = 175
e2 = 125 150
3 x y
350
–
4.44 Bracket connected with double fillet welds to
a column.
The following normal and shear stresses are found in the bracket:
MEde1 56,0·10 6 ·175

point 1 σx = = = 228 N/mm2
Iy 6
42,9·10
2
τ xz = 0 N/mm
MEde 2 56,0·106 ·125

point 2 σx = = = 163 N/mm2
Iy 42,9·106
VEdS 280·103·90000
τ xz = = = 49,0 N/mm2
tIy 12·42,9·106
point 3 σ x = 0 N/mm2
VEdS 280·103 ·183750
τxz = = = 100 N/mm2
tIy 12·42,8 ·10 6
The sizes of the fillet welds at points 1, 2, and 3 are found using equation (4.8) and (4.7)
respectively:
σx 228
a1 0,70 t = 0, 70·12· = 6 mm
fu 360
t βw γM2 12·0,8·1,25
a2 2σ 2x + 3τ 2xz = 2·1632 + 3·49,02 = 5 mm
2fu 2·360
t βw γM2 12·0,8·1,25
a3 3τ 2xz = 3·1002 = 3 mm
2fu 2·360
The required weld thickness at point 1 determines the thickness of the welds. More examples
are given in Connections 5.
4.5.2 Butt welds
EN 1993-1-8, cl. 4.7 covers the design of the following three types of butt welds:
– full penetration butt welds (cl. 4.7.1);
– partial penetration butt welds (cl. 4.7.2);
– T-connections with butt welds (cl. 4.7.3).
Full penetration butt welds

With full penetration butt welds the weld cross-section is equal to the thickness of the plate:
the weld fully penetrates the plate. In general, assuming a proper execution of the weld, the
minimum strength of the molten weld material is higher than that of the plate material. This
means that the weld does not govern the strength of the connection. An exception is the case
in which higher grade steel elements are connected using electrodes with a lower strength than
the parent material (under matching). Reference is made to EN 1993-1-12, cl. 4.2 and 4.5.3.2.
a Partial penetration butt welds

For partial penetration butt welds the same design rules apply as for fillet welds with a deep
penetration. Therefore the throat thickness must be determined. The weld thickness for a fillet
weld with deep penetration must be determined as shown in figure 4.4 of EN 1993-1-8, cl. 4.7.2
(fig. 4.45). Here, a is the nominal weld thickness (prescribed in specifications). When determining
the design value of the resistance for a partial penetration butt weld with deep penetration,
the extra weld thickness due to penetration may be taken into account. Welding tests must be
4.45 Partial penetration butt weld. undertaken to show that the required penetration can be systematically achieved.
T-connections with butt welds

For butt welds in T-connections, EN 1993-1-8, cl. 4.7.3(1) and 4.7.3(2), make a distinction be
tween a weld with a small partial penetration and one with a large partial penetration (fig. 4.46).
For a weld with a small partial penetration, the nominal length cnom of the non-welded part must
satisfy cnom ≤ 3 mm and cnom ≤ t/5, where t is the thickness of the connected plate. When the
total weld thickness anom,1 + anom,2 is at least equal to t, the weld may be considered as a full
penetration butt weld. This means that ultimate limit state failure will be governed by one of
t
the connected elements, not the welds. Therefore, the welds are not critical. So, in that case, in
theory, there is no difference between a full penetration butt weld and one with a small partial
penetration. If the geometrical requirements are not met for a small partial penetration, then the
anom,1
weld must be designed as a fillet weld, with or without deep penetration.
anom,2
cnom
anom,1 + anom,1 ≥ t
t
cnom ≤ or cnom ≤ 3 mm
5 4.46 Effective penetration for a T-connection.
4.6 Literature
structures for buildings according to Eurocode 0, 1 and 3 (Steel Design 1), Bouwen met
Staal, Zoetermeer 2019.
2. J.W.B. Stark, Joints. Analysis and design of bolted and welded connections in steel frames
and in tubular structures according to Eurocode 3 (Steel Design 5), Bouwen met Staal,
Zoetermeer, to be published.
3. J.W.B. Stark and R.J. Stark, Composite structures. Analysis and design of composite steel
and concrete structures for buildings according to Eurocode 4 (Steel Design 4), Bouwen met
Staal, Zoetermeer, to be published.
4. J. Wardenier, J.A. Packer, X-L. Zhao and G.J. van der Vegte, Hollow sections in structural
applications, Bouwen met Staal, Zoetermeer 2010 (2nd edition).
6. EN 1993-1-3 (Eurocode 3. Design of steel structures. Part 1-3. General rules. Supplementary
rules for cold-formed members and sheeting), 2006 (incl. AC, 2009).
7. EN 1993-1-8 (Eurocode 3. Design of steel structures. Part 1-8. Design of joints), 2005 (incl.
AC, 2009).
8. EN 1993-1-9 (Eurocode 3. Design of steel structures. Part 1-9. Fatigue), 2005 (incl. AC, 2009).
9. EN 1993-1-12 (Eurocode 3. Design of steel structures. Part 1-12. Additional rules for the
extension of EN 1993 up to steel grades S 700), 2007 (incl. AC, 2009).
10. EN 1994-1-1 (Eurocode 4. Design of composite steel and concrete structures. Part 1-1. General
rules and rules for buildings), 2004 ( incl. AC, 2009).
11. EN ISO 2553 (Welding and allied processes. Symbolic representation on drawings. Welded
joints), 2019.
12. EN ISO 3452 (Non-destructive testing. Penetrant testing. Part 1. General principles), 2013
(corrected 2014).
13. EN ISO 4063 (Welding and allied processes. Nomenclature of processes and reference
numbers), 2010.
14. EN ISO 5817 (Fusion-welded joints in steel, nickel, titanium and their alloys (beam welding
excluded). Quality levels for imperfections), 2014.
15. EN ISO 10675-1 (Non-destructive testing of welds. Acceptance levels for radiographic testing.
Part 1. Steel, nickel, titanium and their alloys), 2016.
16. EN ISO 11666 (Non-destructive testing of welds. Ultrasonic testing. Acceptance levels),
2018.
17. EN ISO 13918 (Welding. Studs and ceramic ferrules for arc stud welding), 2018.
18. EN ISO 14555 (Welding. Arc stud welding of metallic materials), 2017.
19. EN ISO 17635 (Non-destructive testing of welds. General rules for metallic materials), 2016.
20. EN ISO 17636-1 (Non-destructive testing of welds. Radiographic testing. Part 1. X- and
gamma-ray techniques with film), 2013.
21. EN ISO 17636-2 (Non-destructive testing of welds. Radiographic testing. Part 1. X- and
gamma-ray techniques with digital detectors), 2013.
22. EN ISO 17637 (Non-destructive testing of welds. Visual testing of fusion-welded joints),
2016.
23. EN ISO 17638 (Non-destructive testing of welds. Magnetic particle testing), 2016.
24. EN ISO 17640 (Non-destructive testing of welds. Ultrasonic testing. Techniques, testing
levels, and assessment), 2018.
25. EN ISO 17643 (Non-destructive testing of welds. Eddy current examination of welds by
complex plane analysis), 2015.
26. EN ISO 17659 (Welding. Multilingual terms for welded joints with illustrations), 2004.
27. EN ISO 23277 (Non-destructive testing of welds. Penetrant testing of welds. Acceptance
levels), 2015.
28. EN ISO 23278 (Non-destructive testing of welds. Magnetic particle testing. Acceptance
levels), 2015.
29. EN ISO 23279 (Non-destructive testing of welds. Ultrasonic testing. Characterization of
discontinuities in welds), 2017.
5
Connections
Design and detailing
of connections

design and detailing of connections | connections 5 | 1

Design and detailing
of connections
This chapter deals with the design of complete connections that are
loads
typically used in steel building structures, as described in Connections
EN 1991 1 (Connections in steel structures). Design procedures use the rules
for various individual connectors that are described in Connections
2 (Bolts in clearance holes), Connections 3 (Slip-resistant joints, rivets
representative structure structural modelling and pins), and Connections 4 (Welds). A complete connection – often
NEN 1993-1-1,
chapter 5 also called a ‘joint’, see also Connections 1 – comprises all the individual
steel parts – such as end-plates, stiffeners, and angles – and the indi
vidual connectors, which together ensure the transfer of the forces and
moments
N shear forces moments between the elements that are connected.
axial forces It is necessary to determine how the forces in the individual parts of
V M
the connection can be derived from the external loads applied to the
connection. The design approach is discussed in section 5.1, and the
packing plate design
beam-to-column connection assumptions distribution of forces in a connection in section 5.2. The load path in
NEN 1993-1-8,
cl. 2.5 the complete connection can often be reduced to five basic components
of load transfer; these are dealt with in the section 5.3 to 5.7.
M N
F1 = –
h 2
h internal forces
V 5.1 Design approach
M N
F2 = +
h 2
Figure 5.1 shows schematically the approach to be taken for the
design rules for bolted design and checking of a connection in a steel structure. For a better
beam-to-column connection and welded connections
EN 1993-1-8, understanding, a connection has been chosen in which the bending
chapter 3 and 4
moment and the shear force are resisted by discrete parts of the
connection. This does not mean that such a detail would be an optimal
solution in practice. First of all, the structure is represented in such a
checking components
3 1 4 1 bolt loaded in way that it is appropriate for calculation. The individual loads and load
(fin plate example)
bearing and shear
2 net cross-section combinations are then determined for the representative structure.
fin plate
2 3 weld
4 beam web
Chapter 5 of EN 1993-1-1 provides the rules and procedures that are
used to determine the moments, shear forces, and axial forces that
fin plate connection act on a connection. Subsequently, an internal force distribution in
the connection is determined, which meets the requirements stated
5.1 The process for designing and checking connections. in EN 1993-1-8, cl. 2.5:
2 | connections 5 | design and detailing of connections

M A
F1 = + N flange
h A
M A
F2 = – N flange
h A
A
F3 = N web
A
axial bending shear

F1 force N moment M force V
F3 N h +
V
V M
–
F2
stress distribution in the beam 5.2 Force distribution in a beam-to-column
packing plate
connection with welded and bolted plates.
– the assumed internal forces must be in equilibrium with the external loads and moments acting
on the connection;
– the force in each part must not exceed the design value for the resistance of that part;
– the deformations that are necessary to achieve the assumed force distribution must not exceed
the deformation capacity of the bolts, welds, and other parts of the connection.
When a connection contains several bolts and/or welds it is statically indeterminate, and therefore
the force distribution cannot be derived exclusively from equilibrium. The actual distribution of
the forces to be transferred over the different bolts and/or welds depends on the stiffness and the
deformation capacity of the connection as a whole and of the different parts of the connection.
However, it is not necessary – and usually not possible – to determine the actual force distribution
very precisely. It is sufficient to assume a representative distribution of forces that is:
– in equilibrium; and
– reasonably consistent with the actual deformations.
The choice of a representative distribution of forces is often the most difficult part of the design
process. It requires an understanding of the way the connection will behave under the influence
of loadings. The following aspects are important.
• The distribution of forces in the parts to be connected. Take for example a bolted fin plate
connection between a beam and a column, as shown in figure 5.2. A bending moment, a shear
force, and axial force, all act on the connection. The stress distribution in the beam is shown in
figure 5.2. The shear force is resisted almost entirely by the web and the bending moment primarily
(say 85%) by the flanges. A reasonable representation of the force distribution in the connection
is then that the shear force is completely transferred by the fin plate, the bending moment

completely by the flange plates, and the tension force by a combination of the fin plate and the
flange plates. In practice, some of the shear force will be transferred by the fin plate and some by
the flange plates, but the relative stiffnesses of these elements (to vertical shear) mean that the
flanges only attract a small proportion of the total shear force. Likewise, in practice the web will
transfer part of the bending moment, but it will be a small part given relative lever arms. The
tension force will be transferred by both parts in accordance with the stress distribution in the beam,
because the stiffnesses of the web and flanges are almost equal with regard to the tension force.
• The relative stiffnesses in the connection. EN 1993-1-8, cl. 2.5(d), implies that when determining
the internal force distribution, the fact that stiffness due to an in-plane load (axial force deformation)
is much greater than the stiffness due to an out-of-plane load (bending deformation) must be taken
into account. To better understand the deformation behaviour of a connection, it is useful to consider
what happens after only a small deformation. Section 5.2.1 discusses this in more detail, partly based
on some examples.
• The consistent application of the assumed force distribution to all parts of the connection.
An incorrect force distribution can occur when separate calculations are performed for different
parts of the connection. An example of this is given in figure 5.3: the indicated force distribution
for the design of the bolts is not consistently applied to the design of the end-plate welds. The
welds in the upper part of the connection would be overloaded and could fail.
The forces in a connection can be simplified down to a number of basic components of load transfer.
This is demonstrated for a column-to-beam connection (T-connection) in figure 5.4. For this
connection, the following basic components are important:
– introduction of tension forces;
– introduction of compression forces;
– transfer of axial compression and tension forces;
– transfer of (vertical) shear forces;
– transfer of shear forces (web panel).
With these five basic components it is possible to model many different types of connections.
This makes it possible– despite the large number of different detailing options for connections –
to present a design method clearly and without repetitions. This is why in EN 1993-1-8, cl. 2.4(1),
is chosen for a common approach with different basic components. For each of the constituent
parts of a connection a number of possible failure modes can be identified. These failure modes
can relate to:
– the type of connection (welds or bolts);
– the connecting structural elements (like beams and columns);
– the additional parts (like stiffeners, end-plates, and angles).

WRONG
assumed wrong distribution

bolt force of asociated 5.3 Example of an inconsistent force distribution
distribution weld stresses for bolts and welds.
3
introduction of
tension force
1
transfer of axial tension
or compression force
5 2
web panel N transfer of
in shear shear force
V M
packing plate
transfer of axial tension

or compression force
1
4
introduction of
compression force 5.4 Six basic forms of load transfer in a
beam-to-column connection.

5.2 Force distribution in connections
The distribution of forces and moments in a connection depends on the relationship between
the stiffnesses of its parts. The location of the centre of rotation is also of importance for bolted
connections. In addition to a linear force distribution, plastic theory may also be assumed for bolted
connections whereby the bolt forces are distributed non-linearly. For long (bolted or welded)
connections, an unequal force distribution must be taken into account.
5.2.1 Influence of relative stiffnesses

The stiffness ratios within a connection must be taken into account when the forces and moments
applied to it are distributed. The stiffness loaded by in-plane forces in a plate is, for example,
much greater than loaded in bending. The relative stiffnesses of two elements of equal dimensions,
loaded in tension and loaded in bending, are calculated below for the example in figure 5.5a.
Consider a plate with a width b = 100 mm and a thickness t = 10 mm. The plate is connected to a
support along one edge to form a cantilever of L = 100 mm. At the plate tip a line load q = 10 kN/m
acts perpendicular to the plate. The maximum deflection w is:
FL3 bq·L3 100·10·1003

w= = = = 0,2 mm
3EI 1
3E· bt 3
1
3·2,1·105· ·100·10 3
12 12
An identical plate is now loaded with an in-plane line load of the same magnitude (figure 5.5b).
The load causes a displacement of:
FL bq·L 100·10·100
ΔL = = = = 0,0005 mm
EA E·bt 2,1·105 ·100·10
If the two plates are now connected to each other at their free ends, then they must move together.
So the displacements must be equal: w = ∆L. If a vertical line load of 10 kN/m is applied to the
structure comprising both plates (fig. 5.6), it will distribute between the plates relative to their
stiffnesses. The plate in tension is w/∆L = 0,2/0,0005 = 400 times stiffer than the plate loaded in
bending, which means that practically the plate in tension may be assumed to carry all the load.
The effect of differing stiffnesses is apparent from the behaviour of a welded connection between
a plate and a rectangular hollow section (RHS) (fig. 5.7). The plate is loaded in-plane, and the
RHS wall is loaded out-of-plane. Because the central part of the RHS wall is flexible compared to
its sides (near the corners) the stress distribution in the plate directly next to the weld is irregular.
A similar stress distribution occurs when a plate is welded to an unstiffened flange of an I section.
Another example of the influence of differing stiffnesses on the force distribution, is a bracing
connection in a floor structure (fig. 5.8). The tension force in the bracing can be distributed to
both bolt groups (1) and (2) of the gusset plate, in such a way that equilibrium is met.

10
100 mm 0
10
100 mm
∆L
0
10
10
10 kN/m
10 kN/m
5.5 Cantilever plate, loaded in bending and

a. cantilever plate, loaded in bending b. cantilever plate, loaded in tension tension.
9,975 kN/m
0,025 kN/m
10 kN/m 10 kN/m
5.6 Force distribution between two connected

plates acting in a combination of bending and
structure force distribution tension.
1
b
2 eff
b1 beff b1
1
b
2 eff
FEd FEd
5.7 Uneven stress distributions due to differences
in stiffness.

deformation
1 beam web 1
Fv
∆s1
1
deformation
Ft F end-plate ∆t1
2
∆1
A A ∆t2
2 Ft 2
Ft ∆s2
2 Fv 1
F ∆2
2
1
F F
force distribution deformations
5.9 Incorrect assumption of the force distribution.
1 B
1
1 Fv = 2F 1
2
∆s1
Ft < 0 ∆t1 ∆1
cross-section B-B
2
2 Ft < 0 Ft < 0 2 ∆t2
∆s2 ∆2
1
Fv = 2F
2
F F
cross-section A-A F force distribution deformations
5.8 Connection of bracing in a floor structure. 5.10 Correct assumption of the force distribution.
Assume that the force is distributed simply in accordance with its direction (so evenly
if the brace is at 45˚) (fig. 5.9). However, for compatibility the deformation at (1) due to
Fv must be equal to the deformation at (2) due to Ft, but given relative stiffnesses this
would require Fv to be much larger than Ft. Therefore, the assumed evenly distributed
forces cannot be correct. In figure 5.10 the correct force distribution is given: the force F
causes shear in the bolt groups at (1) and (2).
Differences in stiffness can also influence the modelling of a structure. An example of this
is a bolted connection between an IPE beam, which is nominally pinned to an HE column
by means of a welded fin plate (fig. 5.11a). In this case, the deformation of the bolted
connection will be larger than the deformation of the fin plate or the column web.
A bending moment M = F·(e + 0,5h) and a compression force F act on the column,
and a bending moment M = F·e and a shear force V = F act on the welded connection.
However if the column was a hollow section (fig. 5.11b), then the deformation of the
hollow section wall would be critical. The zero moment point is now assumed to be at

a. HE section as column b. hollow section as column
F
IPE
IPE
0,5h 0,5h e
zero moment zero moment

point point
M = F(0,5h + e) system line system line
M = F·0,5h
M = F·e
5.11 Connection between a beam and an HE column
(a) and a hollow section column (b).
the location of the hollow section wall. In this case, a moment M = F·0,5h and a compression
force F act on the column, and only a shear force V = F acts on the welded connection.
5.2.2 Centre of rotation

It has already been shown that in a bolted connection the distribution of forces between different
bolts depends on the relative stiffnesses of the parts of the connection. The distance of a given
bolt from the location of the centre of rotation is also of importance. A distinction can be made
between free and forced rotation, depending on the detailing of the connection.
Free centre of rotation

Consider a connection with a random bolt pattern loaded by an eccentric force PEd (fig. 5.12a). In
such a connection loaded in shear, it can be assumed that there are no stiffness differences given
equal bolt and hole diameters, and therefore the centre of rotation coincides with the centre of

a.
centre of rotation (cr) gravity of the bolt group. The load can be replaced by a force PEd acting at the centre of gravity
of the bolt group (centre of rotation, referred to as ‘cr’) and a moment MEd = PEd ·e. The bolt
forces must always be in equilibrium with the external loading.
When the material behaviour is elastic and the displacement of the bolts in the holes is not taken
into consideration, the magnitude of the bolt force REd due to the moment MEd is proportional
PEd to the distance r from the centre of rotation. The maximum force REd,max occurs in the bolt at the
distance rmax. For an arbitrary bolt at a distance ri (fig. 5.12b):
e
ri
Ri = R
z rmax Ed,max (5.1)
b.
This bolt force can be broken down into two components:
zi yi
y Ri,y = R and Ri,z = R (5.2)
rmax Ed,max rmax Ed,max
MEd = PEde
The resultant of all the bolt forces in the y- and z-directions due to the moment MEd must be
equal to zero in order to maintain equilibrium:
z
Ri
Ri,z REd,max
ΣRi,y = Σzi = 0 Σzi = 0 (5.3)
rmax
zi
Ri,y
ri
Also ∑yi = 0, so that the centre of rotation is indeed at the centre of gravity of the bolt pattern.
centre of
rotation (cr) yi
y The maximum bolt force REd,max follows from considering moment equilibrium of the bolt group:
5.12 Load on a bolt group.

r REd,max
( )
MEd = Σ riRi = Σ ri r i REd,max =
max
rmax
Σri2

MEdrmax MEdrmax
REd,max = = (5.4)
Σri2 (
Σ y i2 + zi2 )
For practical use of equation (5.4) it is easier to split REd,max into REd,y,max and REd,z,max:
MEd zmax MEd y max

REd,y,max = and REd,z,max = (5.5)
Σ ( y i2 + zi2 ) Σ ( y i2 + zi2 )
To determine the most unfavourable (maximum) bolt force FEd,max, the vertical force PEd /n must
be added to REd,z,max (due to the moment MEd), where n is the number of bolts in the bolt group:
2
2
PEd
FEd,max = REd,y,max + REd,z,max + (5.6)
n

z
For a complicated asymmetric bolt pattern care is needed to ensure that the direction of the REd,y,max
forces is correct and in accordance with the assumed axial system. In practice, asymmetric bolt
patterns do not often occur because symmetric patterns are less costly and the design and
detailing is easier. For relatively large narrow bolt patterns – with the z-axis in the direction of the
bolt row – the vertical component REd,z due to a moment MEd can be neglected with respect to
zmax z1 cr y
the horizontal component REd,y. As a result, a simplified approach is sufficient for determining
REd,y,max. For one row of bolts, with yi = 0 (fig. 5.13): MEd
MEd zmax
REd,max = REd,y,max = (5.7)
Σzi2
5.13 Eccentrically loaded bolt group.

Example 5.1
• Given. A vertical bolt group with two bolt rows at a spacing s = 35 mm is loaded by a force
PEd = 250 kN with an eccentricity e = 10 mm from the right-hand-side bolt row (fig. 5.14).
• Question. Determine the maximum force in bolts 1 and 6.
• Answer. The location of the centre of gravity (centre of rotation) of the bolt group in the
y-direction (for the seven bolts) follows from consideration of moment equilibrium of the vertical s = 35 e = 10
bolt forces, for example around bolt 7, in which the resultant of all seven bolts (= PEd) is applied
at the centre of gravity:
1
3s 3·35
3s = 7y1 y1 = = = 15 mm e*
7 7
z3 = 150 mm
2
Knowing this, the position of the centre of gravity in the y-direction is: y2 = s – y1 = 35 – 15 = 20 mm z
z2 = 100
and e* = e + y1 = 10 + 15 = 25 mm. In the z-direction there is symmetry, so that the y-axis runs 3
z1 = 50
through the centre of gravity of the bolt.
4 y
Acting on the bolt group are a vertical force PEd = 250 kN and a bending moment MEd = PEde* = cr
z1 PEd
250·25·10–3 = 6,25 kNm. To calculate the bolt forces the sum of the squared values of the bolt
5
z2
distances from the centre of gravity is important:

z3
6
Σ y i2 = 4y12 + 3y 22 = 4·152 + 3·202 = 2100 mm2
( ) ( )
7
Σ zi2 = 2 z12 + z 22 + z 23 = 2· 502 + 1002 + 1502 = 70000 mm2
( )
Σ y i2 + zi2 = 2100 + 70000 = 72100 mm2 y2 = 20 y1 = 15
5.14 Eccentrically loaded narrow bolt group.

From equation (5.5) and (5.6) the following results are obtained for bolt 1:
MEd z 3 6,25·103 ·150

REd,1,y = = = 13,0 kN ( )
(
Σ y i2 + zi2 ) 72100
MEd y1 6,25·10 3·15

REd,1,z = = = 1,30 kN
(
Σ y i2 + zi2 ) 72100
PEd 250
The proportion of the vertical force is: = kN
n 7
2
PEd 250
2
FEd,1,max = 2
REd,1,y + REd,1,z +
n
= (+13, 0 )2 + –1,30 –
7
= 39,2 kN
For bolt 6 the results are:
MEd z 2 6,25·103·100
REd,6,y = = = 8,67 kN ( )
(
Σ y i2 + zi2 ) 72100
MEd y 2 6,25·103·20
REd,6,z = = = 1,73 kN ( )
(
Σ y i2 + zi2 ) 72100
2
PEd 250
FEd,6,max = 2
REd,6,y + REd,6,z +
n
= ( –8, 67)2 + +1,73 –
7
= 35,1 kN
Using the simplified equation (5.7) for this relatively large narrow bolt pattern, the following
maximum force in bolt 1 is found:
MEd z 3 6,25·103 ·150

REd,1,y = = = 13,4 kN ( )
Σzi2 70000
2
PEd 250
2
FEd,1,max = 2
REd,1,y +
n
= ( +13, 4 )2 + –
7
= 38,2 kN
This value is only 3% smaller than FEd,1,max = 39,2 kN, determined with equation (5.6).

Forced centre of rotation e
A so-called forced centre of rotation is created with an eccentrically loaded

bolted connection, when the stiffness of the part of the connection loaded PEd
in tension differs from the stiffness of the part loaded in compression. Take
for example the bracket of figure 5.15. Due to the force PEd, the bottom F1
of the bracket is in compression and the top is in tension. The compression
F2
force is transferred directly, without any noticeable deformation, to the web
of the column. Deformations do occur due to the tension force at the top, s1
F3 s2
through bending of the end-plate and column flange and less significant
s3
through elongation of the bolts. The difference in tension and compression
F4 s4
stiffnesses results in the creation of a rotation point in the connection at the
A FA A
centre of compression A.
A similar force distribution is created when beams are connected with end-
plates, loaded in bending. Such a detail is frequently used in practice (forced) centre of rotation
or centre of compression
(fig. 5.16a). Due to the deformation of the end-plates, the distribution of
forces in the bolts does not correspond to the stress variations in the cross- 5.15 Bolted connection with a forced centre of rotation.
sections of the beam at some distance from the connection. This means that
when checking the beam parts directly adjacent to the connection the force
distribution drawn for the bolt group must be assumed. However, if the
beam to beam connection is instead detailed with web and flange plates
(fig. 5.15b), a free centre of rotation occurs in the heart of the bolt group
cr
and then the force distribution in the group corresponds to the stress dis-
–
tribution in the beam cross-section. MEd MEd
If, for the bolt group in figure 5.15, a linear force distribution is assumed,
then the following applies:
+
F1s 2 F1s 3 F1s 4 bolt forces stresses

F2 = ; F3 = and F4 = a. forced centre of rotation (cr) in the beam
s1 s1 s1
For the bolt group: –

MEd MEd
Mbolt group = F1s1 + F2 s 2 + F3 s 3 + F4 s 4
cr
s2 s2 s2
( )
F
= F1s1 + F1 2 + F1 3 + F1 4 = 1 ∑ s12 + s 22 + s 23 + s 24
s1 s1 s1 s1 +
In general, for n bolts: bolt forces stresses

b. free centre of rotation (cr) in the beam
F1 n 2
Mbolt group = ∑ s (5.8) 5.16 Bolt forces in a beam splice: (a) with end-plates and (b) with web
s1 i=1 i
and flange plates.

The size of the compression force FA at the centre of compression A follows from horizontal
equilibrium:
s1 s s s F
FA = F1 + F2 + F3 + F4 = F1
s1
+ F1 2 + F1 3 + F1 4 = 1 Σ s1 + s 2 + s 3 + s 4
s1 s1 s1 s1 ( )
In general, for n bolts:
F1 n
FA = s (5.9)
s1 i =1 i
The maximum bolt force F1 follows from moment equilibrium with respect to the centre of rota-
tion (= centre of compression A):
F1 n 2 s1
MEd = PEd e = Mbolt group = s F1 = MEd (5.10)
s1 i =1 i n
si2
i =1
As discussed above the distribution of bolt forces depends on the relative stiffnesses of the parts
of a connection. A linear distribution only occurs when the thickness of the connecting plate
parts is relatively large with respect to the diameter of the bolts. Structurally this is not a desirable
solution because then the bolts will dictate the strength of the connection (and they are not
ductile). With thin end-plates or thin column flanges large deformations can occur due to plate
bending. Moreover, the cantilevering tips of an end-plate are less stiff than the part between the
flanges of the beam. With an extended end-plate the force distribution in the top two bolt rows
is not linear (fig. 5.17a). With a thin end-plate the bolt force F2 may even be larger than the bolt
force F1 (fig. 5.17b). With practical dimensioning of the end-plate it is reasonable to assume that
the tension forces in row 1 and row 2 are equal. In addition, prying forces can occur in the case of
thin end-plates and thin flanges, and they should not be neglected (see section 5.5.2).
5.17 The influence of the thickness of the end-

plate on the force distribution.
1 F1 1 F1
2 F2 2 F2
centre of compression
cr
or centre of rotation (cr)
bolt forces bolt forces
a. thick end-plate b. thin end-plate

Example 5.2
• Given. A bracket is connected to a column with six bolts and is loaded by a vertical load
PEd = 200 kN at a distance of e = 150 mm from the centre of the connection (fig. 5.18).
• Question. Determine the maximum tension force Ft,Ed and the maximum shear force Fv,Ed on
the bolts, and the compression force FA at the centre of compression. Prying may be neglected
in this example.
• Answer. The column flanges at the location of the lower bolt row are not stiffened. The centre
of compression A (centre of rotation) is assumed to be at sA = 50 mm under the lower bolt row.
The bending moment on the bolt group is MEd = PEde = 200·150·10–3 = 30,0 kNm. For the bolts
s1 = 250 mm, s2 = 150 mm, and s3 = 50 mm. The maximum tension force in the upper bolt row
follows from equation (5.10):
s1 250·10 –3
F1 = MEd = 30, 0· = 85,7 kN
Σsi2 (250 2
)
+ 1502 + 502 ·10 –6
For the most heavily loaded upper bolt:
F1 85, 7 PEd 200

Ft,Ed = = = 42,9 kN and Fv,Ed = = = 33,3 kN
2 2 n 6
The force FA at the centre of compression A follows from equation (5.9):
F1 85, 7
FA = Σs = · ( 250 + 150 + 50 ) = 154 kN
s1 i 250
PEd
A stress check must be carried out to confirm whether or not
the choice of the location of point is realistic; the check of the
introduction of the compression force is considered in section 5.6.
F1 60
In order to be able to resist the force FA = 154 kN at the centre of
compression, in S235 an area of A = 154·103/235 = 655 mm2 is 100
F2
needed. With a symmetric spread from point A to the web of
100
the bracket, an effective web height of 2·20 = 40 mm is needed. F3
The web of the bracket must therefore have a thickness of at sA = 50 mm 70
A
FA
least t = 655/40 = 18 mm (approximately). In order to limit the
stresses at the centre of compression A, it is often assumed that
the centre of rotation is located in the bottom bolt row (see
example 5.3). e = 150
5.18 Bolted bracket to a column with a forced centre of rotation.

Free and forced centre of rotation
An axial (relative to the beam) tension force can act on an end-plate connection in combination
with a bending moment. In that case, either a forced or a free centre of rotation can occur depen-
ding on the relative size of the horizontal force (fig. 5.19).
When HEd is small, a forced centre of rotation occurs, and HEd is transferred at the centre of com-
pression A. As HEd increases the compression force due to the moment decreases. The moment
around point A is MA = MEd + HEdr, with r being the distance between the line of HEd and the
forced centre of rotation A. Using equation (5.8):
F1
s 2 = MEd + HEd r (5.11)
s1 Σ i
MA =
The resulting compression force D follows from horizontal equilibrium:
F1
s1 Σ i
D = FA – HEd = s – HEd (5.12)
The transition from a forced centre of rotation to a free centre of rotation occurs when D = 0.
Then from the combination of equation (5.11) and (5.12):
MEd Σ si2
r+ = (5.13)
HEd Σ si
A forced centre of rotation occurs when:
MEd Σ si2
r+ > (5.14)
HEd Σ si
1 F1
i Fi cr
si
HEd
MEd r
A FA D
cr
5.19 End-plate connection loaded with a forced centre free centre

bending moment and a tension force. of rotation (cr) of rotation (cr)

A free centre of rotation occurs, located at the centre of gravity of the bolt group, when:
MEd Σ si2
r+ < (5.15)
HEd Σ si
For r + MEd HEd = Σsi2 Σsi, both positions of the centre of rotation result in the same distribution
of bolt forces.
Example 5.3
• Given. The bracket from example 5.2, with a vertical load PEd = 200 kN and a horizontal force
HEd = 120 kN at a distance of 80 mm above the upper bolt row (fig. 5.20).
• Question. Determine the maximum tension force Ft,Ed in the bolts, assuming that the centre
of rotation is located at the lower bolt row. Any prying forces are ignored for simplicity. Prying
may be neglected in this example.
• Answer. The bending moment on the bolt group due to PEd equals MEd,P = PEde = 200·150·10–3
= 30,0 kNm. For the bracket r = 280 mm, so that:
MEd,P 30,0·103 Σ si2 2002 + 1002

r+ = 280 + = 530 mm > = = 167 mm
HEd 120 Σ si 200 + 100
This means that a forced centre of rotation occurs at location A, see equation (5.14). The moment
on the bolt group with respect to this centre of rotation is MEd = PEde + HEdr = (200·150 +
120·280)·10–3 = 63,6 kNm. The maximum tension force on the top row of bolts follows from 5.20 Bracket bolted to a column with a forced
centre of rotation.
equation (5.10):
e = 150
s1 200
F1 = MEd = 63,6·103 · = 254 k N PEd
Σ si2 2002 + 100 2
The maximum tension force per bolt is: HEd

F1 60 80
1 1 100
Ft,Ed = F = ·254 = 127 kN r = 280 mm
2 1 2 1
F
2 1
100
A
70

e = 150
1 F1,M F1,H 2Ft,Ed,1

100
rc HEd 2 s1 = 200 mm F2,M rc + F2,H = 2Ft,Ed,2
r = 100 s2 = 100 100
PEd 3 F3,M F3,H
2Ft,Ed,3
moment tension force moment + tension force
5.21 Gusset plate bolted to a column with a Example 5.4

free centre of rotation.
• Given. A gusset plate is bolted to a column, loaded by a vertical force PEd = 200 kN at a
distance e = 150 mm to the centre of the connection, and a horizontal force HEd = 600 kN applied
at the centre of the bolt group (fig. 5.21).
• Question. Determine the maximum tension force Ft,Ed in the bolts.
• Answer. The distance between the horizontal force and the lower bolt row is r = 100 mm. The
bending moment on the bolt group due to PEd is MEd = PEde = 200·150·10–3 = 30,0 kNm. From
the given loads and geometry:
MEd 30,0·103 Σsi2 2002 + 1002

r+ = 100 + = 150 mm < = = 167 mm
HEd 600 Σ si 200 + 100
This means that a free centre of rotation is created, see equation (5.15), located at the centre
of gravity of the bolt group. Due to the moment MEd and the horizontal force HEd the following
forces can occur in the bolt rows:
MEd 30,0·103
F1,M = = = 150 kN; F2,M = 0 kN; F3,M = – F1,M = –150 kN
s1 200
HEd 600
F1,H = F2,H = F3,H = = = 200 kN
n 3
The forces in bolts 1, 2, and 3 are:
F1,M + F1,H 150 + 200 PEd 200

Ft,Ed,1 = = = 175 kN and Fv,Ed,1 = = = 33,3 kN
2 2 6 6
F2,M + F2,H 0 + 200 PEd 200
2 2 6 6
F3,M + F3,H –150 + 200 PEd 200
2 2 6 6

Fy,max FRd FRd
Fy,max FRd
centre of
rotation
5.22 Elastic and plastic force distribution in

a bolt group. elastic elasto-plastic plastic
5.2.3 Non-linear distribution of bolt forces

F
So far in this chapter a linear (‘elastic’) force distribution has been assumed. However, when
F
designing bolted connections plastic behaviour may also be assumed according to EN 1993-1-8, cl.
3.12(1). As the applied load is increased the force in the most heavily loaded bolt at some point V
F VEd MEd
reaches its limiting value FRd (fig. 5.22). As the applied load is increased further, the force in this
F
bolt will remain constant, while the deformation of the connection increases. This deformation
results in increasing forces in the bolts that are nearer to the centre of rotation. Eventually, these
bolts will also reach their limiting load capacity. The resistance of the connection as a whole can
increase until all bolts reach their limiting load capacity, provided the deformation capacity of
the bolts is adequate.
In principle, with a plastic distribution of bolt forces, it may be assumed that each bolt force acts F
perpendicularly to a radius from the centre of rotation. A non-linear distribution of bolt forces
V
can also occur when the design is based on the equilibrium method. Any force distribution that
V
satisfies equilibrium – without the limiting load capacity of any bolt being exceeded and the
VEd MEd
V
deformation capacity is adequate – provides a lower bound for the ultimate load capacity of the
F
connection (fig. 5.23). This allows some simplifying assumptions to be made, as explained below.
When a bending moment and a shear force act on a bolt pattern it may be assumed that the
bolts close to the centre of rotation contribute little towards the moment transfer because of
their small lever arms. It is assumed that these bolts only transfer the (entire) shear force. The
more distant bolts are assumed to only transfer the (entire) moment. These assumptions are
only valid if the components of the connection have sufficient deformation capacity, as is the F1
case when the plate material governs the resistance of the connection, rather than the bolts or F2
V
welds. The deformation capacity of the plate material, either in bending or bearing, is much
V
greater than that of the bolts. That is why EN 1993-1-8, cl. 3.12(2), indicates that a linear force F2 VEd MEd
V
distribution must be used when the shear resistance Fv,Rd of the bolts is smaller than the bearing
F1
resistance Fb,Rd. When applying a plastic approach the following basic principles are important:
– choose a realistic internal force distribution that balances the external load, so that the three
equilibrium conditions are met: ∑V = 0, ∑H = 0 and ∑M = 0;
– make sure that the bolts are strong enough, and use relatively large diameter bolts so that their 5.23 Possible force distributions over the bolts in a
shear resistance Fv,Rd is larger than the bearing resistance Fb,Rd. fin plate connection loaded in bending and shear.

Lj
5.2.4 Force distribution in long connections

Ft,Ed Ft,Ed
For connections loaded in shear, the length over which an even force distribution
may be assumed is limited. This applies to both bolted connections and welded
connections.
assumed (plastic) elastic
force distribution force distribution
Bolted connections
In the elastic phase, the outer bolts of a bolted connection loaded in shear
attract considerably more load than the inner bolts (fig. 5.24). The difference
1
Ft,Ed
2 increases with the number of bolts, or in other words: with the length of the
Ft,Ed
1
2
Ft,Ed connection. However, since bolts can deform plastically, the outer bolts will
eventually yield as the applied load increases, and this will result in a redis-
5.24 Distribution of forces between the bolts in a long connection. tribution of the bolt forces. Depending on the deformation capacity, an even
distribution of bolt forces may occur before failure. The required deformation
capacity is greater for longer connections.
βLf According to EN 1993-1-8, cl. 3.8(1), an even distribution of load should only
be assumed when the distance Lj between the outer bolts is not greater than
1,00
15d. This roughly corresponds to five bolts spaced at p1 = 3,4d0, or six bolts
0,75 spaced at p1 = 2,7d0.
0,50
For longer connections, with Lj > 15d, it cannot be assumed that the defor-
mation capacity of the outer bolts is sufficient to allow an even distribution of
0,25
force. An uneven force distribution must then be taken into account, and this
Lj is done by reducing the design value of the shear resistance Fv,Rd (per shear
0 15d 65d
plane and bolt) using a factor βLf ≤ 1,0 (fig. 5.25).
5.25 Reduction factor βLf as function of the connection length Lj. For a connection with a length Lj ≥ 65d, only 75% of the bolt strength may be
taken into account. For a connection with a length 15d < Lj < 65d the reduction
is:
L j – 15d
βLf = 1 – (5.16)
200d

The reduction values shown in figure 5.25 are based on an assumption that the
bolts have sufficient deformation capacity. Even though EN 1993-1-8 does not
τmax
define this, it is recommended to ‘size’ the bolts in such a way that bearing is
VEd
decisive in long connections (shear resistance Fv,Rd > bearing resistance Fb,Rd).
According to EN 1993-1-8, cl. 3.8(2), the reduction factor βLf need not be taken
distribution of into account if the load is applied more-or-less evenly over the length of the
shear stress
connection. Such is the case for the bolted end-plate connection shown in
figure 5.26.
5.26 A connection in which the applied shear force is evenly
distributed.

Example 5.5
• Given. A double-sided bolted connection in grade S355 steel with M20 class 8.8 bolts (in
double shear), loaded by a force Ft,Ed = 1300 kN (fig. 5.27). The shear plane runs through the
threads of the bolts.
• Question. Determine the minimum number of bolts, and the maximum spacing, possible.
It may be assumed that neither bearing nor yielding of the net cross-section are decisive.
• Answer. The design shear resistance per shear plane follows from equation (2.14) of Connections 2
(Bolts in clearance holes), with αv = 0,6:
α v fub A s 0,6·800·245·10 –3
Fv,Rd = = = 94,1 kN
γM2 1,25
The minimum number of bolts n required, with two shear planes (m = 2), is:
Ft,Ed 1300
n= = = 6,9 n=7
mFv,Rd 2·94,1
For each bolt, the reduction factor βLf must be not less than 6,9/7 = 0,987 in order that 7 bolts
can carry the applied load. From equation (5.16):
L j – 15d
βLf = 1 – 0,987 Lj 17,6 d = 17,6 · 20 = 352 mm
200d
The maximum spacing p1,max is:
Lj 352 352
p1,max = = = = 59 mm
n–1 7–1 6
The minimum spacing follows from table 2.24 of Connections 2: p1,min ≥ 2,2d0 = 2,2·22 = 48
mm. So 7 bolts with a spacing between 48 mm and 59 mm is OK.
Lj
5.27 Long double-sided bolted connection.
Ft,Ed Ft,Ed
plate 125x25 mm plate 150x50 mm
p1
1
F
2 t,Ed
Ft,Ed
1
F
2 t,Ed

elastic stress Welded connections
distribution
An uneven stress distribution occurs over the length of a lap connection
τ||
with fillet welds that are loaded longitudinally (fig. 5.28). This situation
is comparable with a long bolted connection (see fig. 5.24). For long
thin welds there is a danger that the deformation capacity at the start
and end of the weld is not sufficient to justify an even stress distribution.
Ft,Ed Ft,Ed b
Cracks may occur at the ends before the middle part of the weld yields.
This phenomenon is known as ‘unzipping’.
For a lap connection with fillet welds, according to EN 1993-1-8, cl.
Lj,eff
4.11(3), the maximum length Lj for which an even stress distribution may
be assumed is 150a, where a is the effective throat length. When deter-
5.28 Uneven stress distribution in the longitudinal direction of the weld. mining the size of fillet welds with a length Lj > 150a, one must assume a
reduced effective weld length βLw,1Lj, with:
0, 2L j
βLw,1 = 1,2 – 1,0 (5.17)
150a
According to EN 1993-1-8, cl. 4.11(2) there is no need for this reduction

when the load is applied evenly. This is the case with a connection subject
to shear force (such as end-plates, see also fig. 5.26), or with a connection
between the flange and the web of a fabricated beam.
5.2.5 Combination of different types of connections

Each type of connection (bolted or welded) shows a characteristic relation
ship between load and deformation. Therefore, the failure load of each
type is associated with a different level of deformation (fig. 5.29).
Sometimes different connectors, with different stiffnesses, are used in
one connection for example when this is necessary for manufacturing or
load erection. According to EN 1993-1-8, cl. 2.4(3) the connectors with the
greatest stiffness must be designed to carry the total design load.
weld preloaded bolt
(category C) EN 1993-1-8, cl. 3.9.3, makes an exception for the combination of welds
preloaded bolt
(category B) and preloaded 8.8 and 10.9 bolts that are designed as slip-resistant at
the ultimate limit state ULS (category C), provided that final tightening of
the bolts is carried out after the welding is complete.
bolt in an oversize hole,

loaded in bearing and shear
5.29 Characteristic relationship between the load and deformation for

deformation different types of connection.

5.3 Transfer of axial compression or tension forces
Axial forces in an element can be transferred in different ways, namely through butt welds, via welded
plates, or with gusset plates. The most direct way of load transfer is through butt welds, although these
are normally avoided because of costs (see Connections 1 (Connections in steel structures), section 1.2).
Welded plates are required when fillet welds or bolts are to be used. Gusset plates are often used for
connecting multiple elements meeting at a connection at different angles.
Rules for the design of a splice connection (see fig. 5.31c) are given in EN 1993-1-8, cl. 6.2.7.1, par. (10)
to (16). In the case of a splice located in a beam at a point where a plastic hinge must be able to occur,
EN 1993-1-8, cl. 6.2.3(5) gives guidance on weld design.
EN 1993-1-8, cl. 6.2.7.1, par. (10) to (12) describes the design principles for a splice connection. A
distinction is made between splices in members in compression, in paragraph (13) to (15), and splices
in beams in bending, in paragraph (16). Often, a splice in compression is not prepared in the fabrication
shop for transfer of forces by direct contact. When there is no preparation paragraph (13) indicates that
the elements of the connection (connectors, plates) must be able to transfer the internal forces and
moments. Moments due to eccentricities, initial imperfections, and second order deformations must
be taken into account. The internal forces and moments must be assumed to be no smaller than:
– a shear force equal to 2,5% of the axial force resistance of the weakest connected profile, in the
direction of both axes;
– a bending moment equal to 25% of the bending moment resistance of the weakest connected
profile, about both axes.
Paragraph (14) indicates that for column-to-column connections that are prepared for transferring
forces by complete contact, splice elements must nevertheless be added. These elements must be
connection
able to transfer at least 25% of the maximum axial force in the column.
For elements loaded in bending (beams), paragraph (16) provides specific rules. In contrast to a column
splice, there is no specific requirement that for a beam splice at least 25% of the moment capacity of the
weakest connected section, around both axes, must be taken into account. However paragraph (16a)
states that compression flanges should be treated as compression members and this implies that the
MEd MEd
rules of paragraph (13) apply. Paragraph (16) explicitly stipulates that the moment due to an eccentricity
of bolt patterns must always be considered.
basic form
5.3.1 Butt welds
The most direct (although costly) way of transferring an axial force in an element – or in the flange of
Ft,fl,Ed Ft,fl,Ed
a beam in bending – is by means of a butt weld. A welded connection as shown in figure 5.30 does
not need to be designed, because butt welds are at least as strong as the connected parts, see also
Connections 4 (Welds), section 4.5.2. Welded butt joints are not suitable for erection connections.
Another disadvantage of butt welds is that dimensional deviations in the connecting elements are Ft,fl,Ed Ft,fl,Ed
difficult to accommodate. Fillet welds are generally preferred because they are easier to form, and
therefore less expensive, during fabrication than butt welds. 5.30 Transfer of axial force through a butt weld.

5.3.2 Plates
Figure 5.31 shows a number of connections in which the tension force in the flange, or in the
web of the beam, is transferred by plates, connected with bolts or fillet welds. The plates can be
applied single-sided or double-sided, in which case eccentricity of the load, and therefore the
associated deformation, is avoided. When plates are used, dimensional variations and deviations
can be accommodated by the use of packing plates (fig. 5.31b).
Welded plates
In a flange plate connection such as that shown figure 5.31a, the strength of the connection can
be governed by either the plate, the weld group, or the flange itself, see also example 4.1 in
Connections 4.
5.31 Transfer of axial forces by plates. connection basic form
ℓp
bp
MEd VEd VEd MEd
tp
tf
A
a. single-sided plate, welded
bp
VEd MEd
tp
tf
packing plate
b. single-sided plate, bolted
bp
MEd VEd VEd MEd tp,2 tp,1

tf
c. double-sided plate, bolted

• For the plate, yielding of the gross cross-section according to equation (2.4) of Connections 2
(Bolts in clearance holes) will always be critical. The limiting tension force Ft,E is:
Afy bp t p fy
Ft,Ed Nt,Rd = Npl,Rd = =
γM0 γM0 (5.18)
• For the weld group, the design resistance of the side fillet welds may be taken in combination
with the end fillet weld. It is recommended to make the size of the end fillet weld not less than that
of the side fillet welds, see example 4.1 of Connections 4. Experiments have shown that the use of
different weld sizes – where the end weld is smaller than the side welds – does not result in an even
force distribution over the welds just prior to failure of the welds. Under no circumstances should
the end fillet weld be smaller than half the size of the side fillet welds. Using a small fillet weld to
seal the joint (see point A in fig. 5.31a), for example to prevent corrosion, is not recommended for
the same reason. When an end fillet weld is required at the ‘front side’ for structural reasons, this
weld must be at least as big as the other welds.
Using the so-called simplified method according to EN 1993-1-8, cl. 4.5.3.3 equation (4.3) from
Connections 4 applies. For the tension force Ft,Ed:
Ft,Ed (
Fw,Ed = a Σℓeff fvw,d = a 2ℓp + bp fvw,d ) with fvw,d =
fu
3 β w γM2
(5.19)
The result for grade S235 steel, with βw = 0,8, fu = 360 N/mm2, and γM2 = 1,25, is that fvw.d =
208 N/mm2.
When the so-called combined stress method, according to EN 1993-1-8, cl. 4.5.3.2 – see
Connections 4, section 4.5.1 – is used the resistance of the side fillet welds remains the same,
but the resistance of the end fillet weld increases by a factor 3 2 (table 5.32).
For the weld group shown in figure 5.31a:
Ft,Ed 2ℓp afw,d,|| + bp afw,d, (5.20)
For grade S235 steel, equation (5.20) can be simplified to:
Ft,Ed 2ℓpa·208 + bpa·255 (5.21)

S235 S275 S355
side fillet welds: f w,d,|| 208 234 251

5.32 Design values for the resistance of side
fillet welds and end fillet welds (N/mm2) for the end fillet welds: 255 286 308
combined stress method.

• The flange – to which the welds are connected – must clearly be able to resist the forces in the
welds. This can be achieved by choosing welds that are not too big. The limiting weld thickness is a
function of the strength of the weld and the shear strength of the parent material:
fyd 0, 58fyd
afw,d,|| tf a tf (5.22)
3 fw,d,||
For grade S235 steel, equation (5.22) can be simplified to:
0, 58·235
a t f = 0,65t f (5.23)
208
Bolted plates
In case of a connection with a single-sided plate (fig. 5.31b), both the plate and the bolt group must
be checked, assuming that the resistance of the connected section is not critical.
• For the plate, either yielding of the gross cross-section according to equation (2.4), or failure of
the net cross-section according to equation (2.5) of Connections 2, will be critical:
Afy bp t p fy
yielding gross cross-section Ft,Ed Nt,Rd = Npl,Rd = = (5.24)
γM0 γM0
failure net cross-section Ft,Ed Nt,Rd = Nu,Rd =

0,9 A net fu
=
( )
0,9 bp – 2d0 t p fu
(5.25)
γM2 γM2
packing plate
(thickness tp) • For the bolt group, according to EN 1993-1-8, cl. 3.7, the design value of the resistance of the
bolt group may be taken as the sum of the bearing resistances Fb,Rd of all bolts in the group, provided
that the shear resistance of each bolt is larger than the bearing resistance. When the shear resistance
Fv,Rd of any bolt is smaller than the bearing resistance Fb,Rd, then the resistance of the bolt group shall
be taken as the number of bolts multiplied by the smallest resistance, see Connections 2, section 2.5.1.
The bearing resistance must be checked both for the plate with thickness tp and the flange with
thickness tf. When the length Lj of the connection is longer than 15d, the shear resistance shall
be reduced according to equation (5.16), see also section 5.2.4. In a connection with double-sided
5.33 Packing plates to accommodate size
plates (fig. 5.31c), it is generally assumed that each plate transfers half of the connecting force. In
differences. this case, the bolts have two shear planes and the shear resistance is twice as large as with a single-
sided plate.
When packing plates (fig. 5.33) with a thickness tp > d/3 (d is the nominal bolt diameter) are used,
according to EN 1993-1-8, cl. 3.6.1(12) the design value of the shear resistance shall be reduced by
a factor βp:
9d
βp = 1 (5.26)
8d + 3t p

number and position
of bolts M24 (8.8)?
Ft,Ed Ft,Ed
dimensions plate 260x25 mm 5.34 Bolted connection to be designed

cover plate?
using two cover plates.
Example 5.6
• Given. A 260x25 mm plate in grade S235 steel, loaded by a tension force Ft,Ed = 1500 kN
(fig. 5.34).
• Question. Design a bolted connection with double-sided cover plates.
• Answer. Assume M24 class 8.8 bolts with the shear planes going through the threaded part of
the bolt. Practical spacings are: e1 = 55 mm, e2 = 50 mm, and p1 = p2 = 80 mm, see Connections 2,
table 2.25. The design value of the shear resistance per shear plane per bolt follows from equation
(2.14) of Connections 2:
0,6 fub A s 0,6·800·353·10 –3

Fv,Rd = = = 136 kN
γM2 1,25
In order to determine the design value of the bearing resistance of the plate Fb,Rd, the factors
k1 and αb must first be determined. The factor k1 follows from equation (2.17) and (2.18) of
Connections 2:
2,8 e2 2,8·50
end bolts k1 = – 1,7 = – 1,7 = 3,7
d0 26
1,4p 2 1,4·80
k1 = – 1,7 = – 1,7 = 2,6
d0 26
k1 = 2,5 (critical)
1,4p 2 1,4·80
inner bolts k1 = – 1,7 = – 1,7 = 2,6
d0 26
k1 = 2,5 (critical)
The factor αb depends on the factor αd, which is given by equation (2.20) and (2.21) of Connections
2:
e1 55
end bolts αd = = = 0,705
3d0 3·26
p1 1 80 1
inner bolts αd = – = – = 0,775
3d0 4 3·26 4

The factor αb is the smallest of either αd, the ratio fub/fu, or 1,0. The ratio of material strengths
fub/fu = 800/360 = 2,2 > 1,0, and is therefore not critical. So for the end bolt αb = 0,705. The
smallest value of the bearing resistance is then:
k1α b fud t 2,5·0,705·360·24·25·10 –3

Fb,Rd = = = 305 kN
γM2 1,25
Comparing shear (for a bolt in double shear) and bearing resistance shows that shear is critical,
since 2Fv,Rd = 2·136 = 272 kN < Fb,Rd = 305 kN. The resistance of the bolt group must therefore
be determined using the smaller design value, and the required number of bolts n is:
TEd 1500
n= = = 5,5 n=6
2Fv,Rd 2·136
The chosen bolt pattern is shown in figure 5.35. Now that the number and locations of the bolts
is known, the connected plate must be checked for yielding of the gross cross-section according
to equation (5.24), and failure of the net cross-section according to equation (5.25).
Yielding of the gross cross-section is checked according to equation (5.24):
bp t p fy
260·25·235·10 –3
Npl,Rd = = = 1528 kN
γM0 1,0
Ft,Ed = 1500 kN Nt,Rd = Npl,Rd = 1528 kN (OK)
Failure of the net cross-section is checked using equation (5.25). For the three sections 1, 2, and 3:
section 1 2 3 50 plate 260x25
50
b = 260 4x40 Ft,Ed Ft,Ed
50
60 80 80 55
cover plate 250x16
bolts M24 (8.8)
5.35 Selected bolt pattern for the bolted using two cover plates.

section 1 ( )
A net = b – d0 t = ( 260 – 26 ) ·25 = 5850 mm
0,9A net fu 0,9·5850·360·10 –3

Nu,Rd,1 = = = 1516 kN
γM2 1,25
section 2 ( )
A net = b – 2d0 t = ( 260 – 2·26 ) ·25 = 5200 mm
0,9A net fu 0,9·5200·360·10 –3

Nu,Rd,2 = = = 1348 kN
γM2 1,25
section 3 ( )
A net = b – 3d0 t = ( 260 – 3·26 ) ·25 = 4550 mm
0,9A net fu 0,9·4550·360·10 –3

Nu,Rd,3 = = = 1179 kN
γM2 1,25
Checking the three sections with NEd = Ft,Ed:
NEd 1500
section 1 = = 0,99 1,0 (OK)
Nu,Rd,1 1516
NEd 5 1500
section 2 = · = 0,93 1,0 (OK)
Nu,Rd,2 6 1348
NEd 3 1500
section 3 = · = 0,64 1,0 (OK)
Nu,Rd,3 6 1179
All three sections are satisfactory. Calculation will show that none of the diagonal ‘sections’ (running
through the lines of bolts) are critical.
Finally, the minimum thickness of the cover plates must be determined. The maximum tension
force Ft,Ed = 1500 kN acts at section 3 (at which point complete load transfer has taken place).
Check yielding of the gross cross-section of both cover plates using equation (5.24):
2bp t pfy Ft,Ed γM0 1500·103 ·1,0

Ft,Ed Npl,Rd = tp = = 12,3 mm
γM0 2bfy 2·260·235
Check failure of the net cross-section of both cover plates with equation (5.25):
0,9A net fu Ft,Ed γM2 1500·103 ·1,25

Ft,Ed A net = = 5787 mm2
γM2 0,9 f u 0,9·360
( )
A net = 5787 mm2 = 2 b – 3d0 t = 2· ( 260 – 3·26 ) t t 15,9 mm (critical)
The minimum required width of the cover plate at the location of sections 1 and 2 can be deter-
mined in the same way.

undeformed Design recommendations
When designing a splice connection the following general guidelines should
deformed
be taken into account.
a. eccentric connection • Connect each part according to the force distribution that follows from the
analysis. Design the connection in such a way that practical values for bolt spacings
and edge distances are taken into account, see table 2.25 in Connections 2. For
example, for the spliced HE section shown in figure 5.31c, this means that the
b. symmetric connection
flanges and the webs are connected so that they transfer the forces in the flanges
5.36 Eccentrically and symmetrically connected plates. and the webs respectively.
• Avoid eccentricities in the connection as much as possible. In the examples of
figure 5.36, the eccentrically and symmetrically connected connections may have
approximately the same resistance, but the eccentric connection (fig. 5.36a) is
subject to large deformation before the ultimate strength is reached. Symmetry
can prevent deformation (fig. 5.37a). When angles are used in a connection they
have sufficient stiffness to prevent excessive deformations due to eccentricity. Cover
plates (at the outside) are satisfactory for connecting to the angle (fig. 5.37b).
a. symmetrically connected cover plates
5.3.3 Gusset plates

Stability bracing elements are often connected to the primary structure with
b. eccentrically connected cover plates
gusset plates, see also Connections 1, section 1.3.6. Gusset plates are also
used in trusses to connect braces to chords (fig. 5.38). Figure 5.39 shows some
5.37 Solutions with single-sided plates, in which deformation due solutions for welded and bolted connections of a tension element to a gusset
to eccentricity is prevented.
5.38 Examples of gusset plates in a bracing system (left) and in a truss (right).

plate. Welds are most suitable for connections made in the workshop, and
bolts for connections made on site, see also Connections 1, section 1.1.
When angles are used the in plane eccentricities e1 and e2 (fig. 5.40) of the line
of gravity with respect to the welds or the bolts must be taken into account.
In practice, the out-of-plane eccentricity e* (see fig. 5.40 and 5.41) is usually a. centrically welded
neglected when designing a connection. This practice is not in accordance with

EN 1993-1-8, cl. 2.7(2), stating that for angles and T-sections ‘every possible
eccentricity’ must be taken into account, including the out-of-plane eccentricity
e*. As an exception, EN 1993-1-8, cl. 3.10.3(2) states that for an angle con-
nected with one row of bolts loaded in tension the eccentricity e* need not to
be taken into account. b. centrically bolted
The following items are discussed in this section:

– welded connection to a gusset plate;
– bolted connection to a gusset plate;
– design of gusset plates.
c. eccentrically welded
Welded connections
For a welded connection between an element and a gusset plate (fig. 5.39a
and c), the axial force should not be larger than the resistance of the cross-
section of the element itself, the weld group, or the governing cross-section
of the gusset plate.
d. eccentrically bolted
5.39 Welded and bolted connections of a plate or angle in tension.
1 FEd,1
1
e1
FEd
1
b
e2
b FEd
1
2 FEd,2
ℓeff tp 1
ℓeff tp
FEd
e* FEd
e*
tp tp
5.40 Welded connection of an angle to a gusset plate. 5.41 Welded connection of a channel section to a gusset plate.

• The resistance of the fillet welds in the weld group can be determined in the same way as
described for plates in section 5.3.2. One can choose between the two methods as described in
Connections 4 (Welds), section 4.5.1, namely the combined stress method or the simplified method.
In a welded connection of an angle to a gusset plate without an end fillet weld (fig. 5.40), the
forces in both side fillet welds are in equilibrium with the axial force in the element. Due to the
eccentricity of the line of gravity of the angle with respect to the welds, the forces FEd,1 and FEd,2
are not equal:
e2 e1
FEd,1 = FEd and FEd,2 = FEd (5.27)
b b
Side fillet weld 1 must have a larger throat thickness than side fillet weld 2. Or, if both welds have
the same throat thickness, weld 2 can be shorter. When an end fillet weld is used to supplement
the side fillet welds, the throat thickness of the end weld – in order to assure adequate deformation
capacity – should be the same as that of the thicker side fillet weld 1. Usually, it is assumed that
the eccentric moment FEde* is carried by the angle, since the angle is stiff compared to the gusset
plate. The shear force on the welds then acts in the plane of the welds, with no additional stresses.
• For the check of the gusset plate, failure of the critical cross-section 1-1 must be considered
as a possible failure mode (fig. 5.40 and 5.41), see also Connections 4, section 4.5.1. The design
value of the tension force FEd (associated with yielding in tension over the length b and yielding
in shear over the length ℓeff), is:
2ℓeff t p fy
bt p fy +
3
FEd Nt,Rd = (5.28)
γM0
An alternative design approach – based on the method for determining the tearing out of a bolt
group (block tearing) – is to check the design block tearing resistance of a bolt group Veff,1,Rd ac-
cording to EN 1993-1-8, cl. 3.10.2; see also Connections 2, section 2.4.5:
A nv fy 2ℓeff t p fy
A nt fu bt p fu
3 3
FEd Nt,Rd = Veff,1,Rd = + = + (5.29)
γM2 γM0 γM2 γM0

Example 5.7
ℓeff,1
• Given. A tie, formed from double 150x100x10 mm angles, is connected angle

1 150x100x10
with fillet welds to a 20 mm thick gusset plate (fig. 5.42). The steel grade is FEd,1
S235 and the tension force to be transferred is FEd = 1100 kN. e1 = 48
• Question. Determine the minimum length of the fillet welds for: FEd
b = 150
– side fillet welds only, with a thickness a1 = 7 mm and a2 = 4 mm e2 = 102
(fig. 5.42a)
FEd,2
– side fillet welds and an end fillet weld with thicknesses a1 = a3 = 7 mm 2
and a2 = 4 mm (fig. 5.42b).
ℓeff,2
• Answer. In both cases, first the (minimum) weld lengths are deter- 20 mm
mined, followed by a check to ensure that the parent material of the a. with side fillet welds only
gusset plate can resist the weld forces.

ℓeff,1
Side fillet welds only

1
The forces FEd,1 and FEd,2 in the side welds follow from equation (5.27): FEd,1
e1 = 48
e2 102 e3 = 27
FEd,1 = FEd = ·1100 = 748 kN FEd,3 FEd
ℓeff,3 = 150
b 150 e2 = 102
3
e1 48
FEd,2 = FEd = ·1100 = 352 kN FEd,2
b 150 2
ℓeff,2
The resistance of a side fillet weld is fwd,|| = 208 N/mm2 (see table 5.32). b. with side fillet welds and an end fillet weld
The minimum effective length for each is:
5.42 Welded connection of double angle tie
1 1 to a gusset plate.
F
2 Ed,1 2
·748·103
ℓeff,1,min = = = 257 mm
a1fwd,|| 7·208
1 1
F
2 Ed,2 2
·352·103
ℓeff,2,min = = = 212 mm
a2 fwd,|| 4·208
For both weld lengths the requirements that ℓeff ≥ 6a and ℓeff ≥ 30 mm are satisfied, and so the
calculated values of effective length may be used to determine the design resistance. Also,
for both welds ℓeff ≤ 150a, so the weld lengths do not have to be reduced, see Connections 4,
section 4.5.1.
The connected parent material of the gusset plate must be able to resist the force transferred by
the welds. This is checked according to equation (5.22) and (5.23):
FEd,1 fyd
748·103 235
= = 146 N/mm2 > = = 136 N/mm2 (not OK)
tℓeff,1,min 20·257 3 3

The gusset plate at the location of weld 1 cannot resist the force. The choice of a slightly longer
weld, for example ℓeff,1 = 280 mm, would reduce the stress in the parent material. When the
choice of weld size a1 is based on equation (5.23), then immediately it should have been clear
that (for a double weld) the weld thickness was too large given the plate thickness, because: 2a1
= 2·7 = 14 mm > 0,65t = 0,65·20 = 13 mm.
The gusset plate at the location of weld 2 is adequate to resist the forces transferred by weld 2.
Side fillet welds in combination with an end fillet weld

The length of the end fillet weld ℓeff,3 is equal to the width b = 150 mm of the section. The minimum
lengths required for the side fillet welds 1 and 2 can be determined from consideration of the
capacity of the weld group. The resistance of a side fillet weld is fwd,|| = 208 N/mm2, and of an
end fillet weld fwd, = 255 N/mm2 (see table 5.32). The capacity of the three (each double) welds
becomes:
FRd,1 = 2a1ℓeff,1fwd,|| = 2·7·ℓeff,1 ·208 = 2912 ℓeff,1
FRd,2 = 2a2 ℓeff,2 fwd,|| = 2·4·ℓeff,2 ·208 = 1664 ℓeff,2
FRd,3 = 2a3 ℓeff,3 fwd, = 2·7·150·255·10 –3 = 536 kN
From the horizontal equilibrium and from the moment equilibrium (for a random point on the
line of action of force FEd ):
horizontal equilibrium FRd,1 + FRd,2 + FRd,3 = 2912· ℓeff,1 + 1664· ℓeff,2 + 536·103
= FEd = 1100·103 N
moment equilibrium FRd,1e1 = FRd,2e 2 + FRd,3e 3
2912· ℓ eff,1· 48 = 1664· ℓeff,2 ·102 + 536·103·27
From these two equations with two unknowns the solution ℓeff,1 = 165 mm and ℓeff,2 = 50 mm
can be derived.
A check is then made to ensure that the parent material of the gusset plate can resist the force
in weld 1. With FEd,1 = FRd,1 = 2912·ℓeff,1 = 2912·165·10–3 = 480 kN:
FEd,1 fyd
480·103 235
= = 145 N/mm2 > = = 136 N/mm2 (not OK)
t ℓeff,1,min 20·166 3 3
The gusset plate at the location of weld 1 cannot resist the force, so a longer weld length must
be chosen, for example ℓeff,1 = 180 mm.

Bolted connection schematic point
Figure 5.43 shows an example of a bolted connection between angles and a

gusset plate in a truss. With angles it is not possible to place the bolts along
the line of gravity of the section, because there is not sufficient space for the
head or the nut (and certainly for an optional washer), and the tightening tool. system line
(line of gravity
There would also be insufficient space for the drill or punch used to form the of angle)
holes in the in the fabrication shop.
According to EN 1993-1-8, cl. 2.7 the eccentricity e must be taken into when
angles and T-sections are connected with one or two rows of bolts. The eccentric equilibrium at
moment F(w – e) causes additional forces in the bolts (fig. 5.44). The maximum schematic point
bolt force in this case is:

5.43 Bolted connection between multiple angles in a truss.
2 2 F F ( w – e)
R= V +H with H= and V = (5.30)
3 2p1
In practice, eccentricity is usually taken into account by multiplying the horizontal

force H by a magnification factor ψ. The following values are used:
H
– with two bolts: ψ = 1,20;
– with three bolts: ψ = 1,10; V R
– with four or more bolts, the effect of eccentricity is often neglected. p1 p1
EN 1993-1-8, cl. 3.10.4 provides rules for a so-called lug angle (fig. 5.45); this
is a short length of angle connecting the free flange of the continuous angle
w F
to the gusset plate. e
In order to prevent deformations due to eccentricity between the lines of action

of the forces in the elements and gusset plate, it is recommended to use 5.44 Bolt forces in an eccentric connection of an angle
double angles and channel-sections (fig. 5.46). For secondary members, such
as those used in lightly loaded bracings, single angles could be used. This is
for example the case for roof bracings of single storey industrial buildings.
continuous angle
lug angle
5.46 Details using two angles to prevent deformation. 5.45 Lug angle connecting.

Example 5.8
• Given. A tie consisting of a double angle 90x90x9 mm is connected with three M20 class 8.8
bolts to a gusset plate with a thickness of 15 mm (fig. 5.47). The steel grade is S235 and the force
to be transferred is FEd = 350 kN.
• Question. Check the connection and determine the required width b of the gusset plate.
• Answer. In order to check the connection, the following parts must be considered:
– bolts: shear and bearing;
– angle: yielding of the gross cross-section and failure of the net cross-section;
– gusset plate: yielding of the gross cross-section and failure of the net cross-section.
Bolts
Assume the shear plane runs through the threaded part of the bolts. Practical spacings and end/
edge distances are: e1 = 45 mm, e2 = 45 mm and p1 = 70 mm; see also Connections 2 (Bolts in
clearance holes), table 2.21 and table 2.25.
The design value of the shear resistance per shear plane follows from equation (2.14) of
Connections 2 with αs = 0,6 for bolt class 8.8:
α s fub A s 0,6·800·245·10 –3
Fv,Rd = = = 94,1 kN
γM2 1,25
The design value of the bearing resistance follows from equation (2.16) of Connections 2:
k1α b fud t
Fb,Rd =
γM2
The factor k1 depends on the edge distance and spacing perpendicular to the direction of the
force. There is a single row, so only edge bolts are involved. The value of k1 is k1 = 2,5, because
e2 = 45 mm ≥ 1,5d0 = 1,5·22 = 33 mm; see Connections 2, section 2.5.1.
The factor αb depends on the end distance and spacing parallel to the force direction (e1 and p1)
and is the smallest value of αd; fub/fu or 1,0.
The ratio fub/fu equals 800/360 = 2,22 >1,0 and therefore is not critical. The factor αd is:
e1 45
end bolt αd = = = 0,682
3d0 3·22
p1 1 70 1
inner bolt αd = – = – = 0,810
3d0 4 3·22 4
The bearing resistance is:
k1α bfudt 2,5·0,682·360·20·15·10 –3

end bolt Fb,Rd,eb = = = 147 kN
γM2 1,25
k1α bfudt 2,5·0,810·360·20·15·10 –3
inner bolt Fb,Rd,ib = = = 175 kN
γM2 1,25

gusset plate
double angle section
bolts M20 (8.8) 90x90x9
ys = 25,4 mm
45 FEd
b
45
45 70 70 45 15 mm 5.47 Double angle sections bolted to a

gusset plate.
A comparison between the resistance of bolts in double shear and bearing shows that for every
bolt bearing is critical, since 2Fv,Rd = 2·94,1 = 188 kN > Fb,Rd,ib = 175 kN. In that case, according
to EN 1993-1-8, cl. 3.7, the overall resistance is the sum of the design values of the individual
bolts.
The eccentricity of the force FEd with respect to the bolt row is taken into account using a magni-
fication factor ψ = 1,10. Equation (2.22) of Connections 2 is used to check the bolts in bearing:
Fv,Ed = ψFEd = 1,10·350 = 385 kN ≤ Fb,Rd = Fb,Rd,eb + 2Fb,Rd,ib = 147 + 2·175 = 497 kN (OK)
Angles
The design value of the tension resistance Nt,Rd of an angle section with three bolts is the smaller
of the yielding resistance Npl,Rd of the gross cross-section and the failure resistance Nu,Rd of the
net cross-section.
• The yielding resistance of the gross cross-section follows from equation (2.4) of Connections 2:
Afy
2·1552·235·10 –3
Npl,Rd = = = 729 kN
γM0 1,0
• The failure resistance of the net cross-section follows from equation (2.10) of Connections 2.
The reduction factor β3 can be determined from figure 2.33 of Connections 2:
( )
A net = 2 A – d0 t = 2· (1550 – 22·9 ) = 2704 mm2
p1 70
β 3 = 0,3 + 0,08 = 0,3 + 0,08· = 0,555
d0 22
β 3 A net fu 0,555·2704·360·10 –3
Nu,Rd = = = 432 kN (critical)
γM2 1,25
Check of the angles in tension:
FEd 350
= = 0,81 1,0 (OK)
Nu,Rd 432

Gusset plate
To determine the required width b of the gusset plate, it is assumed that Nt,Rd = FEd = 350 kN.
The minimum required width is initially calculated considering net cross-section failure. A check
is then made to ensure that yielding of the gross cross-section is not critical. For the net cross-
section, from equation (2.5) of Connections 2:
Nt,Rd γM2 350 ·103 ·1,25

A net = = = 1350 mm2
0,9fu 0,9·360
A net 1350
bmin = + d0 = + 22 = 112 mm
t 15
Yielding of the gross cross-section is checked with equation (2.4) of Connections 2:
Afy
112·15·235·10 –3
Nt,Rd = = = 395 kN
γM0 1,0
FEd 350
= = 0,88 1,0 (OK)
Nt,Rd 395
Example 5.9
• Given: A tension bracing element, comprising a circular hollow section 219,5x6,3 mm, is connected
to a gusset plate (tp = 20 mm). The connection is achieved using a slotted plate (tp = 20 mm) that is
slotted into the hollow section and connected to the gusset plate by nine M24 class 8.8 bolts
(d0 = 26 mm), see figure 5.48. The plate is slotted into the hollow section over a length of 250 mm,
and welded on both sides with a double fillet weld (a = 5 mm). At the end of the tube on both sides
of the slotted plate end-plates are welded to the tube and the slotted plate. The steel grade is
S235. The bolts are not preloaded (category A), and the shear plane runs through the threaded
parts of the bolts. The connection is in a non-corrosive environment. The design value of the tension
force in the bracing element is TEd = 850 kN.
• Question: Check the connection between the slotted plate and the gusset plate, assuming
that the gusset plate is not critical.
• Answer: The bolts are loaded in shear and bearing and for the plate either the gross cross-section
or the net cross-section could determine the tension resistance. The resistance of the connecting
plate must also be sufficient to resist block tearing of the bolt group. To begin, the bolt positions
must be checked against the requirements in EN 1993-1-8, cl. 3.5(10), see Connections 2 (Bolts
in clearance holes), table 2.24:
end distance e1 = 40 mm ³ 1,2d0 = 1,2·26 = 31 mm (OK)

edge distance e2 = 45 mm ³ 1,2d0 = 31 mm (OK)

slotted plate
tp = 20 mm TEd = 850 kN
gusset plate
tp = 20 mm
slotted plate
fillet weld (a = 5 mm) tp = 20 mm
45
centre of gravity
half tube
80
circular
250 end-plate hollow section
219,1x6,3
80
45
40 80 80 60 250 mm e 5.48 Welded connection of a circular hollow

section to a gusset plate.
spacing p1 = 80 mm ³ 2,2d0 = 2,2·26 = 57 mm (OK)
p2 = 80 mm ³ 2,4d0 = 2,4·24 = 62 mm (OK)
The connection is located in a non-corrosive environment, so no additional requirements are set

for the maximum distances.
Shear
The shear plane of the bolt runs through the thread. The shear resistance Fv,Rd for each bolt per
shear plane follows from equation (2.14) in Connections 2 with αv = 0,6 for 8.8 bolts.
α s fub A s 0,6·800·353·10 –3
Fv,Rd = = = 136 kN
γM2 1,25
The design value of the total resistance Tv,Rd of the nine bolts in single shear equals: Tv,Rd =
9Fv,Rd = 9·136 = 1224 kN. From equation 2.15 in Connections 2:
Fv,Ed = TEd = 850 kN ≤ Fv,Rd = Tv,Rd = 1224 kN (OK)

Bearing
The bearing resistance of the bolts Fb,Rd follows from equation (2.16) in Connections 2, in which
the bearing factor k1 and the factor αb depend on the positions of the bolts. In the direction
of the force, there are three end bolts and six inner bolts. Perpendicular to the direction of the
force, there are two rows of edge bolts and one row of inner bolts, with three bolts in each row.
The bearing factor k1 depends on the distance perpendicular to the direction of the force according
to equation (2.17) and (2.18) of Connections 2:
2,8e 2 2,8 ·45

edge bolt k1 = – 1,7 = – 1,7 = 3,2 or
d0 26
1,4p 2 1,4 · 80
k1 = – 1,7 = – 1,7 = 2,6 or
d0 26
k1 = 2,5 (critical)
1,4p 2 1,4 · 80
inner bolt k1 = – 1,7 = – 1,7 = 2,6 óf k1 = 2,5 (critical)
d0 26
So for all bolts k1 = 2,5.

The factor αb depends on the distance parallel to the direction of the force according to equation
(2.19) and (2.21) of Connections 2:
e1 40
end bolt αd = = = 0,51
3d0 3·26
fub 800
α b = α d = 0,51 (critical) or αb= = = 2,22 or α b = 1,0
fu 360
p1 1 80 1
inner bolt αd = –= – = 0,78
3d0 4 3·26 4
f 800
α b = α d = 0,78 (critical) or α b = ub = = 2,22 or α b = 1,0
fu 360
The bearing resistances of the end bolts and inner bolts are therefore:
k1α b fudt p
2,5·0,51·360·24·20·10 –3
end bolt Fb,Rd = = = 176 kN
γM2 1,25
k1α b fudt p
2,5·0,78·360·24·20·10 –3
inner bolt Fb,Rd = = = 270 kN
γM2 1,25
For all bolts the shear resistance Fv,Rd is smaller than the bearing resistance Fb,Rd. Therefore, the
shear resistance of the bolt group is critical.
Tension resistance of the slotted plate

The tension resistance Nt,Rd of the slotted plate follows from equation (2.4) and (2.5) of
Connections 2, depending on yielding of the gross cross-section or failure of the net cross-
section:

A = bt p = 250·20 = 5000 mm2
A net = A – 3d0 t p = 5000 – 3·26·20 = 3440 mm2
Afy
5000·235·10 –3
yielding gross cross-section Nt,Rd = Npl,Rd = = = 1175 kN
γM0 1,0
0,9 A net fu
failure net cross-section Nt,Rd = Nu,Rd =
γM2
0,9·3440·360·10 –3
= = 892 kN (critical)
1,25
The verification is:
NEd TEd 850

= = = 0,95 1,0 (OK)
Nt,Rd Nt,Rd 892
Block tearing of the bolt group

The bolt group is symmetric and loaded by a centric load. There are two different possible tearing
patterns: A and B (fig. 5.49). For both patterns the net cross-section loaded in shear Anv is:
Anv = 2( 2p1 + e1 – 2,5d0 ) tp = 2·( 2·80 + 40 – 2,5·26) ·20 = 5400 mm2
The net cross-section loaded in tension Ant is different for each pattern:
pattern A Ant = 2( p2 – d0 ) tp = 2·( 80 – 26) ·20 = 2160 mm2
pattern B Ant = 2( e2 – 0,5d0 ) tp = 2·( 45 – 0,5·26) ·20 = 1280 mm2 (critical)
The design value of the resistance Veff,1,Rd follows from equation (2.11) of Connections 2:
fy A nv
235·5400·10 –3
fu A nt 3 360·1280·10 –3 3
Veff,1,Rd = + = + = 1101 kN
γM2 γM0 1,25 1,0
pattern A pattern B 5.49 Two possible tearing patterns of the

plate slotted into the hollow section.

NEd TEd 850

= = = 0,77 1,0 (OK)
Nt,Rd Veff,1,Rd 1101
Check welds between slotted plate and hollow section

There is an eccentricity e between the stress resultant in each half tube and the longitudinal fillet
weld on each side, see figure 5.48. However the welded end-plates prevent the displacement of
the half tube transverse to the slotted plate and compensate for the effect of the eccentricity. So the
longitudinal welds may be designed for the shear load only.
The welds are checked using the simplified method. The design value of the shear resistance of
a fillet weld fvw,d follows from equation (4.3) of Connections 4 (Welds). For S235 with βw = 0,8
(see table 4.35 of Connections 4):
fu 360
3 3
fvw,d = = = 208 N/mm2
βw γM2 0,8·1,25
The weld size (throat thickness) is a = 5 mm and the weld length ℓ = 250 mm. The effective length is
the total weld length reduced by twice the throat thickness, giving ℓeff = ℓ – 2a = 250 – 2·5 = 240 mm.
It is also necessary that ℓeff = 240 ≥ 6a = 6·5 = 30 mm, so that the requirement for the minimum
length is met. The check is made using equation (4.3) of Connections 4:
Fw,Ed 850·103
= = 177 N/mm2 fvw,d = 208 N/mm2 (OK)
a Σ ℓeff 5·4·240
Block tearing resistance of hollow section wall

It is assumed that the shear planes in the hollow section wall are located on both sides of the
double fillet weld. The calculation is identical to that of pull-out resistance along section 3 in
example 4.1 of Connections 4. The tension resistance at the end of each pulled-out plate part,
with a width of approximately 20 mm (thickness of the slotted plate), is neglected. The total shear
area is Av = 4ℓt p = 4·250·6,3 = 6300 mm2. The shear resistance Sv,Rd is:
fy
235
Av 6300· ·10 –3
3 3
Sv,Ed = = = 855 kN
γM0 1,0
TEd 850
= = 0,99 1,0 (OK)
Sv,Ed 855

Example 5.10
• Given. The ties of a cable-stayed roof are connected by pins to a gusset plate with a thickness
of tp = 30 mm, which is connected to the head of a pylon (fig. 5.50). The design value of the larger
tension force is FEd = 800 kN. The ties are circular hollow sections 193,7x6,3 mm with fork plates at
their ends. The forks comprise two plates each of thickness tf = 15 mm. A ring of neoprene with a
thickness of tn = 5 mm is attached on both sides between the fork and the gusset plate. The steel
grade of all parts is S355. The pin connection is permanent and does not have to be replaced.
• Question. Check the pin connection.
• Answer. The geometry of the given pin connection differs from the requirements for the dimen-
sions of a pin connection with a given geometry in figure 3.24 of Connections 3 (Slip-resistant
connections, rivets and pins). This means that the requirements for a pin connection with a given
thickness t are applicable. A check must be carried out to confirm that a and c are sufficient for pla-
te thickness t = 15 mm.
1
FEd γM0 2d0 2
·800·103 ·1,0 2·64
a = 97 mm + = + = 80 mm (OK)
2tfy 3 2·15·355 3
1
FEd γM0 d0 2
·800·103 ·1,0 64
c = 65 mm + = + = 59 mm (OK)
2tfy 3 2·15·355 3
5.50 Connection between a circular hollow
section and a gusset plate using a pin
The checks for a pin connection are according to equation (3.22) to (3.27) of Connections 3. connection.
detail 1
a = ·193,7 = 97 64 fork
2
1
tp = 30 mm c = ·(193,7 – 64) = 65
2
circular
d0 = 64 hollow section
193,7x6,3 mm
pin
60 mm
circular hollow section circular hollow section tg = 15

193,7x6,3 mm 193,7x6,3 mm
tp = 30
tg = 15
circular hollow section neoprene

406,4x8,8 mm tn = 5 mm

• Shear resistance pin. The cross-section of the pin is A = πd2/4 = π·602/4 = 2827 mm2. The
pins are in double shear, so Fv,Ed = FEd/2 = 800/2 = 400 kN. For a pin with a diameter of 60 mm ac-
cording to table 3.1 of EN 1993-1-1: fup = 470 N/mm2. The shear resistance Fv,Rd is:
0,6Afup
0,6·2827·470·10 –3
Fv,Rd = = = 638 kN Fv,Ed = 400 kN (OK)
γM2 1,25
• Bearing resistance plate and pin. The total thickness of both fork plates is equal to the thickness
of the gusset plate. So for the gusset plate and the fork:
1,5tdfy
1,5·30·60·355·10 –3
Fb,Rd = = = 959 kN Fb,Ed = FEd = 800 kN (OK)
γM0 1,0
• Moment resistance pin. According to EN 1993-1-8, cl. 3.13.2(2) the moment resistance of the pin
must be calculated assuming an evenly distributed load over the contact length between the pin and
the connecting elements. It must also be assumed that the connecting elements for the pin act as
hinged supports. The maximum moment in the pin, according to figure 3.11 of EN 1993-1-8, is:
MEd =
1
( 1
)
F t + 4t n + 2t g = ·800· ( 30 + 4·5 + 2·15 ) ·10 –3 = 8,00 kNm
8 Ed p 8
The moment resistance of the pin MRd is:
πd3 π·603
Wel = = = 21206 mm3
32 32
1,5Wel fyp
1,5·21206·335·10 –6
MRd = = = 10,7 kNm MEd = 8,00 kNm (OK)
γM0 1,0
• Combined shear and moment resistance pin. The check for combined shear and moment of
the pin is:
2 2
MEd Fv,Ed 2 2
8, 00 400
+ = + = 0,95 1 (OK)
MRd Fv,Rd 10, 7 638

Design of gusset plates
When designing a gusset plate the dispersion of forces in the plate should be taken into account.
For braces it is assumed that the force in the gusset plate disperses at an angle of 50-60˚ from
the first bolt (fig. 5.51). The force on cross-section A-A acts centrically and the stresses are evenly
distributed. The length A-A must be sufficient to resist the force in the brace. It is not possible to
achieve an evenly distributed stress distribution for all connected elements, particularly for the
connection of chords to a gusset plate. In this case, the full height (h) of the plate cross-section is
required to transfer the force in the chord. Section a-a (fig. 5.51) is eccentrically loaded and the
stress distribution is assumed to be linear.
Regarding the stress distributions in both the gusset plate and the connected elements, the fol-
lowing recommendations are important:
– let the lines of gravity of the element cross-sections coincide with the system lines of the
structure;
– make sure that the system lines of chords and braces meet at one point, to avoid bending
moments in the connected elements;
– avoid sharp re-entrant corners in gusset plates to avoid stress concentrations and notch effects.
It is only possible to approximate the actual stresses in a gusset plate. The highest stresses occur
at the edges of the plate, around the location of the first bolt hole for any given brace.
equilibrium
schematic point
F4 F1 F3
F2
F2
F3 a
α = 50˚-60˚
A
h
A A
A
F4 F1
assumed stress
schematic point a distribution in
cross-section a-a
5.51 Transfer of forces in a gusset plate.

D2 D2
D1 D1
z z
stresses in stresses in
a cross-section a-a a cross-section a-a
– –
1 1
h h
2 2
e2 e2
y y
e1 1
h e1 β 1
h
O1 α β O2 2 O1 O2 2
α +
+
σedge σedge
a 1 a
1 O
(O2 – O1) 3 1
3
+ stresses at the bottom D1cosα
– edge of the gusset plate
D1sinα
D2cosβ D1cosα
e2
D1 D1sinα
e1
D2sinβ b O2
equilibrium a 2
O
schematic point D2 3 1
O1
forces at section a-a
R R = O2 – O1 = D1cosa + D2cosb
5.52 Bolted gusset plate connection with a continuous chord. 5.53 Bolted gusset plate connection with a discontinuous chord.
As an approximation, the following maximum stress occurs at the bottom edge of the gusset
plate shown in figure 5.52:
1
ΣH M· 2 h
σ edge = + (5.31)
A net Inet
Where:
ΣH summation of all horizontal forces;
Anet net area of the cross-section being considered;
M bending moment in the cross-section being considered;
h height of the gusset plate;
Inet second moment of inertia of the net cross-section of the gusset plate.
Consider equilibrium of the gusset plate to the left of section a-a, for which moment equilibrium
has to be considered about the centre of gravity (intersection between the y-axis and section
a-a). The three bolts in the bottom chord together transfer the chord force O2 – O1 into the
gusset plate:
Σ H = D1 cosα – 1 O2 – O1
3
( ) and Σ V = D1 sinα
(5.32)
M=
1
( )(
O – O1 e1 + e2 + D1cos α·e 2
3 2
)

D1
D2 B1
R F1
B2
a. equilibrium with only F1
acting on the upper chord F1
B2 B1
F2 D1
D2 B1
R B2
b. equilibrium with only F2 F2
acting on the gusset plate D2 D1
cover plate
5.54 Reinforcement of the chords with cover plates. 5.55 Equilibrium in a truss with an external force F1 acting on the chord (a) or F2 acting
on the gusset plate (b).
Considering the forces at the intersection point, the resulting chord force R = O2 – O is equal to
the sum of the decomposed horizontal forces in the elements D1 and D2 (see fig. 5.52).
Stresses in the gusset plate are much more onerous when the chord is not continuous (fig. 5.53).
The entire chords forces O1 and O2 must be transferred by the gusset plate. In this case, for
cross-section a-a:
2
Σ H = D1cosα – O1 and Σ V = D1sinα
3
(5.33)
M=
2
( )
O e + e 2 + D1 cos α·e 2
3 1 1
In most cases the moment causes excessive stresses at the edge of the gusset plate, resulting
in the need to strengthen the joint. A structurally sensible solution is a cover plate fixed to the
upward flanges of the chord. This achieves that the force in these flanges is directly through the
cover plate. With that in fact the same situation is created as with a continuous chord. In some
cases it is even necessary to provide the flanges that are directly attached to the cover plate (the
adjacent flanges) with a gusset plate (fig. 5.54).
The resultant force R on the chord bolts follows from the polygon of forces. The magnitude and
the direction of R depends on the way in which the external load acts on the node. For example,
when an external force F1 acts on the chord, this force must be transferred through the bolts to
the gusset plate (fig. 5.55a). When instead of F1 a force F2 of the same magnitude is applied on
the gusset plate, then no additional load is created on the bolts (fig. 5.55b).

V1
angle angle
60x60x6 70x70x7
D1
gusset plate
tp = 12 mm
150 mm
z Z
45˚
105
150
double angle section
100x65x9 50 O1 O2
cover plate
tk = 10 mm
5.56 Connection in a truss with a discontinuous

chord.
Example 5.11
• Given. A gusset plate connection in a truss of which the bottom chord is discontinuous and
connected to a gusset plate (tp = 12 mm) with M20 bolts (d0 = 22 mm) (fig. 5.56). The flanges of the
chord are, where necessary, connected with M16 bolts and a cover plate (tc = 10 mm). The design
values of the forces in the element are: O1 = 360 kN, O2 = 552 kN, D1 = 272 kN and V1 = 192 kN.
• Question. Determine the required number of M16 and M20 class 8.8 bolts (assume the shear
plane runs through the threads) for connecting the chords O1 and O2, and check the thickness of
the gusset plate.
• Answer. Firstly, the most heavily loaded chord O2 is checked to see if the applied force can
be transferred to the gusset plate through its upward flanges. When it is found that this is not the
case, the outstand flanges of chord O2 are used to transfer a part of the force directly to element
O1 by means of a cover plate. The number of bolts can then be determined knowing the size of
force to be transferred through each group. The gusset plate is then checked for the maximum
stresses.
Check cross-section chord O2

For the flange that is adjacent to the gusset plate choose a bolt spacing p1 = 5d0. The reduction
factor for an angle with two or more bolts in one flange is β3 = 0,7; see figure 2.33 in Connections 2.
The net cross-section of the chord, consisting of two angles 100x65x9 mm, is: Anet = 2(A – d0t) =
2·(1415 – 22·9) = 2434 mm2. The design value of the tension resistance Nu,Rd of element O2 with
three bolts is derived from equation (2.10) of Connections 2:
β3 A net fu 0,7·2434·360·10 –3
Nu,Rd = = = 491 kN < Nu,Ed = O2 = 552 kN (not OK)
γM2 1,25
The tension resistance Nu,Rd is inadequate, so the outstand flanges of the angles must be used
to transfer a part of the force O2. A cover plate with a thickness of tc = 10 mm is connected to
these flanges.

The forces in the outstands depend on the relative cross-sectional areas of the flanges. The width
of the outstand flanges is bout = 65 mm and the width of the adjacent flanges is badj = 100 mm.
Assuming that the total force in the element is distributed between flanges based on the ratio of
their cross-sections, the forces in the outstands are approximately:
b out 65
O1,out = O1 = ·360 = 142 kN
b out + b adj 65 + 100
b out 65
O2,out = O2 = ·552 = 217 kN
b out + b adj 65 + 100
Number of bolts in the upward flanges

The maximum tension force O3 in the cover plate is assumed to be the mean value of the forces
in the outstands of O1 and O2:
O3 =
1
2( ) (
O1,out + O2,out = 1 · 142 + 217 = 180 kN
2
)
The forces that must be transferred through the upward flanges are therefore:
O1,adj = O1 – O3 = 360 – 180 = 180 kN
O2,adj = O2 – O3 = 552 – 180 = 372 kN
According to table 2.45 of Connections 2 the resistance of a M20 class 8.8 bolt in double shear,
with the shear plane through the threads, is Fv,Rd = 2·94,0 = 188 kN. The bearing resistance is
Fb,Rd = 9,6t = 9,6·12 = 115 kN. So bearing is critical: Fb,Rd < Fv,Rd. The required number of M20
bolts is therefore:
O1,adj
180
chord O1 n= = = 1,6 n=2
Fb,Rd 115
O2,adj
372
chord O2 n= = = 3,2 n=4
Fb,Rd 115
Number of bolts in the outstand flanges

According to table 2.45 of Connections 2 the shear resistance of an M16 class 8.8 bolt in single
shear, with the shear plane through the threads, is Fv,Rd = 60,2 kN. The bearing resistance is Fb,Rd
= 7,7t = 7,7·9 = 69,3 kN. So in this case shear resistance is critical: Fv,Rd < Fb,Rd. The required
number of M16 bolts is therefore:
O3 180
cover plate O3 n= = = 3,0 n = 4 (due to symmetry)
Fv,Rd 60, 2

Check thickness cover plate
Assume the width of the cover plate is b = 2·65 + 12 = 142 mm, so A = btc = 142·10 = 1420 mm2
and Anet = (b – 2d0)tc = (142 – 2·18)·10 = 1060 mm2. The tension resistance follows from equation
(2.4) and (2.5) of Connections 2:
Afy
1420·235·10 –3
yielding gross cross-section Nt,Rd = = = 334 kN
γM0 1,0
Nt,Rd = 334 kN Nt,Ed = O3 = 180 kN (OK)
0,9 A net fu 0,9·1060·360·10 –3

failure net cross-section Nt,Rd = = = 275 kN
γM2 1,25
Nt,Rd = 275 kN Nt,Ed = O3 = 180 kN (OK)
Check thickness gusset plate

The following forces act on the left hand part of the gusset plate, at the location of cross-section
a-a (fig. 5.57): O2,upw = 372 kN and V1 = 192 kN. With respect to the centre of gravity Z of the
gusset plate, the moment M = O2,upwe = 372·121,8·10–3 = 47,6 kNm. The maximum edge stress
at the lower side of the gusset plate is:
1 1
Iy = t h3 = ·12·300 3 = 2700·104 mm4 and A = t ph = 12·300 = 3600 mm2
12 p 12
1 1
O2,upw M· 2 h 372·103 47,6·106 · 2 ·300
σ edge = + = + = 368 N/mm2 fy = 235 N/mm2
A Iy 3600 4
2700·10
The stress at the edge of the gusset plate is too high. The stress level could be reduced by:
– choosing a gusset plate in grade S355 steel instead of grade S235;
– increasing the plate thickness tp = (368/235)·12 = 20 mm (rounded);
– using cover plates to reduce the force in the adjacent flanges, which would make the chord
5.57 Forces acting on a gusset plate. (almost) continuous.
a Finally, a check on the elastic shear stress must be

V1
carried out according to equation (7.28) of Structural
basics 7 (Resistance of cross-sections):
D1
150 mm
3V1 3·192·10 3
τEd = = = 80 N/mm2
h Z Z 2ht p 2·300·12
e = 121,8 150 τEd 80

= = 0,59 1,0 (OK)
O1,adj line of gravity O2,adj
fy 235
angle
3·1,0
section a-a a 3 γM0

5.4 Transfer of shear forces
Figure 5.58 shows a number of structural solutions for a beam-to-column connection between hot-
rolled sections, designed to transfer shear force only. EN 1993-1-8, cl. 5.1.1(2) uses a classification
system whereby connections (‘joints’ in Eurocode terminology) are divided into three (simplified)
models, depending on their moment resistance and rotational stiffness. The connections shown in
figure 5.58 are all nominally pinned, see also Joints (Steel Design 5).
When designing a pinned connection with partial depth end-plates (fig. 5.58a), it can be assumed
that the connectors (welds and bolts) are only loaded in shear. In addition to the resistance of the
connectors, the shear resistance of a part of the beam web and the shear resistance of the end-plate
must be checked. In the case of a pinned connection with a fin plate or with angles (fig. 5.58b and c),
the connection with the column or the beam must be designed taking into account an eccentricity
moment. The choice of the location of the point of zero moment determines which connectors are
subject to this eccentricity moment. This choice is influenced by the stiffness ratios of the different
parts of the connection; see section 5.2.1 and fig. 5.11. Figure 5.58 shows the location of the zero
moment point for the three examples.
For a beam-to-column connection using angles, usual practice is to locate the point of zero moment
at the face of the column, see [4]. The bolt group joining the web of the beam must then be checked
for both the shear force and the eccentricity moment. When normal clearance holes are used, the
stiffness of this bolt group is usually of the same order of magnitude as that of the connection to the
column flange. Therefore, the zero moment point may also be located in the bolt group joining the
web. In this case, the bolt group in the column flange must be designed for both the shear force and
the eccentricity moment. For a more detailed explanation on the design and detailing of nominally
pinned connections, see Joints.
a. with a short end-plate b. with a fin plate c. with angles
e e e
choice
e
5.58 Examples of a nominally pinned beam-to-column connection.

Figure 5.59 shows three examples of a nominally pinned connection between two beams (primary
and secondary beams). Taking into account the limited torsion stiffness of the I section primary
beam, it should be assumed that the line of action of the shear force coincides with the web of the
primary beam. The design and detailing of this type of connection is considered in detail in Joints.
Splices designed to transfer axial force are considered in section 5.3. Generally, a splice in a beam
loaded in bending must transfer a shear force. Often, the splice is positioned at a point of zero
moment so it is sufficient to simply connect the webs, where the shear forces are concentrated
(see also section 5.1). Figure 5.60 gives three solutions for a beam splice. Two of these solutions
adopt a cover plate (the one with two bolt rows is the more common) (fig. 5.60a). Bolt groups 1
and 2 have an equal stiffness, so the centre of the splice should be positioned at a point of zero
moment. Both bolt groups must be verified for a shear force V and an eccentricity moment V·e.
A number of possible options for the distribution of bolt forces are shown in figure 5.23.
In single storey industrial buildings, beam splices often take the form of an end-plate connection over
the full depth of the beam (fig. 5.60c). When lateral torsional buckling is to be considered – or, for
example in roofs of light weight steel structures, the stability of the lower beam flange – a splice with
only web cover plates is not a good solution.
5.59 Examples of a nominally pinned beam-to- a. with an end-plate b. with a fin plate c. with angles
beam connection.
1 2 1 2
5.60 Possible details for a beam splice at a point

of zero moment.
V V V
e e e e
moment line
M = V·e M = V·2e
a. web cover plate b. web cover plate c. full depth end-plate

with two bolt rows with one bolt rows

Example 5.12 qEd = 50 kN/m
splice
• Given: An IPE 500 beam in grade S235 steel, continuous over two equal spans each
12 m 12 m
of 12 m (fig. 5.61). The beam is subject to a uniformly distributed load qEd = 50 kN/m.
3m
A splice is applied at 3 m from the internal support. The splice consists of two web
cover plates 200x400x6 mm in steel grade S235, connected with eight M20 bolts in
bolt class 8.8. bolts web cover plate
IPE 500 M20 (8.8) 200x400x6
• Question: Check the splice.
• Answer: For a continuous beam with two equal spans, the zero moment points (inter- 50
100
nal hinges) are located at one quarter span from the middle support in case of an elastic
500 mm Anv 100 350
moment distribution, thus at 3 m. At the splice, the moment equals Ms,Ed = 0 kNm and
100
the shear force Qs,Ed = 3qEdL/8 = 3·50·12/8 = 225 kN.
Ant 50
The web cover plates must be checked for block tearing failure (block tearing of the
web itself is prevented by the flanges), and the bolts must be checked for bearing 200 50 100 50
and shear. In addition, the splice must be verified for an eccentric moment Ms,exc,Ed
= Qs,Ede = 225·50·10–3 = 11,3 kNm, assuming a structural model in accordance with 5.61 Splice in a two span continuous beam.
figure 5.60a.
Web plates
The bolt group is not symmetrically loaded due to the eccentricity of the shear force.
The block tearing resistance for each web plate Veff,2,Rd follows from equation (2.12)
of Connection 2 (Bolts in clearance holes):
1 1
A nt = 50 – d t = 50 – ·22 ·6 = 234 mm2
2 0 2
1 1
A nv = 350 – 3d0 – d t= 350 – 3·22 – ·22 ·6 = 1638 mm2
2 0 2
fy A nv
235·1638·10 –3
0,5 fu A nt 3 0, 5·360·234·10 –3
3
Veff,2,Rd = + = + = 256 kN
M2
γM0 1,25 1,0
The shear force per web plate equals Vs,Ed = Qs,Ed/2. The verification is therefore:
1
Vs,Ed 2
·225
= = 0, 44 1,0 (OK)
Veff,2,Rd 256

Bolts
The bolts must be checked for shear and bearing. The shear resistance Fv,Rd for bolts with the
shear plane through the thread follows from equation (2.14) of Connections 2:
α v fub A s 0,6· 800·245·10 –3

Fv,Rd = = = 94,1 kN (per section)
γM2 1,25
For the bearing resistance Fb,Rd, the thickness of the beam web (tw = 10,2 mm) is critical com-
pared to the thickness of the two cover plates (t = 2·6 = 12 mm). The equilibrium method (see
section 5.2.3) is assumed for the distribution of forces, in which the two inner bolts resist the
shear force and the two outer bolts the eccentricity moment. A limitation on the application of
the equilibrium method is that the shear resistance Fv,Rd of a bolt must be larger than its bearing
resistance Fb,Rd.
• Bearing due to shear force. The forces on the two inner bolts due to the shear force act vertically
(see fig. 5.23). The bearing factor k1 depends, according to equation (2.17) of Connections 2, on
the bolt spacing perpendicular to the direction of the force, where a single row of edge bolts is
involved and p2 is not relevant:
2, 8e2 2,8·50
k1 = – 1,7 = – 1,7 = 4,7 or k1 = 2,5 (critical)
d0 22
The factor αb depends, according to equation (2.19) and (2.21) of Connections 2, on the spacing
parallel to the direction of the force, where only inner bolts are involved:
p1 1 100 1
αd = – = – = 1,27
3d0 4 3·22 4
fub 800
α b = α d = 1,27 or α b = = = 2,22 or α b = 1,0 (critical)
fu 360
The bearing resistance Fb,Rd follows from equation (2.16) of Connections 2:
k1α b fudt 2,5·1,0·360·20·10, 2·10 –3

Fb,Rd = = = 147 kN
γM2 1,25
The bolts are loaded in double shear so for each bolt Fv,Rd = 2·94,1 = 188 kN > Fb,Rd = 147 kN.
Therefore, bearing is critical. The check with Vu,Rd = 2Fb,Rd is as follows:
Vs,Ed 225
= = 0,77 1,0 (OK)
Vu,Rd 2·147

• Eccentricity moment. The forces on the two outer bolts due to the eccentricity moment act
horizontally (see fig. 5.23). Only edge bolts are involved. The bearing factor k1 follows from
equation (2.17) of Connections 2, in which e2 is not relevant:
1,4 p2 1,4 ·100

k1 = – 1,7 = – 1,7 = 4,7 or k1 = 2,5 (critical)
d0 22
The factor αb depends, according to equation (2.19) and (2.21) of Connections 2, on the spacing
parallel to the direction of the force:
e1 50
αd = = = 0,76
3d0 3·22
fub 800
α b = α d = 0,76 (critical) or α b = = = 2,22 or α b = 1,0
fu 360
The bearing resistance Fb,Rd follows from equation (2.16) of Connections 2:
k1α bfudt 2,5· 0,76·360·20·10,2·10 –3

Fb,Rd = γM2 = = 112 kN
1,25
Bearing is also critical in this case.

The shear force on the lower and upper bolts due to the eccentricity moment equals:
Ms,exc,Ed 11, 3·103

Vs,Ed = = = 37, 7 kN
3p1 3·100
Vs,Ed 37,7
= = 0, 34 (OK)
Fb,Rd 112
For all bolts, bearing is therefore critical, so the structural model based on the equilibrium
method is valid.

5.62 Examples of a splice designed to transfer a When a splice is required to transfer a moment Msplice in addition to a shear force V, it is necessary
shear force and a bending moment.
to include cover plates on the flanges (figure 5.62). The elastic distribution of the moment Msplice
1 2 between the cover plates depends on the relative contribution of the web and flanges on the
second moment of area I of the beam:
Ffl
Iflange
flanges Mflange = Msplice
I
V (5.34)
Iweb
web Mweb = Msplice
Ffl I
e e
The forces in the flanges are Fflange = Mflange/(h – tf). For verification of the flange cover plates
Mflange see section 5.3.2. Bolt group 2 in figure 5.62a must be verified for a shear force V and a bending
Msplice
Mweb V·e moment Mweb + V·e.
When the moment Msplice is not too large, it may be assumed that it is transferred total by the
a. with a relatively large bending moment
flange cover plates (fig. 5.61b). The force in the flanges then equals Fflange = Msplice/(h – tf).
1 2 The web plates exclusively transfer the shear force V and the eccentricity moment V·e, and are
loaded as for the connection in figure 5.60a.
Fflange
V
5.5 Introduction of tension forces
Fflange Careful consideration should be given to connections in which a plate or beam flange, which is
loaded in tension, is connected perpendicularly to the flange of another section (be that flange
e e
stiffened or unstiffened).
Mflange
This section provides design methods for such cases, considering both welded and bolted
V·e connections.
b. with a relatively small bending moment

5.5.1 Welded connections
When a plate (or beam flange) is welded perpendicular to an unstiffened flange of a section, the
stress distribution in the attachment depends on the stiffness ratio of the web and the flange of
the section (see also section 5.2.1). The web is loaded in-plane by a tension force in the plate and
is much stiffer than the flange loaded in bending. Therefore, an uneven stress distribution occurs
in the attached plate (fig. 5.63). EN 1993-1-8, cl. 4.10, takes this effect into account by limiting
the effective width of the attached plate.
The design model is based on a schematization with yield lines (fig. 5.64). A part of the tension
force is transferred by normal stresses over a width tw + 2r, without the supporting flange being
subjected to bending. This part of the tension force equals:
Ft,1 = fy,ptp(tw + 2r) (5.35)

width depending on
the flange thickness tf
Ft,1 Ft,2
stress distribution:
realistic
schematized
1
b
2 eff
tf r
beff bp bp
tw
+
1
b
2 eff
tp
tw + 2r F = cmp
5.63 Uneven stress distribution due to differences in stiffness.
Where fy,p is the yield strength of the attached plate. The remaining part of the tension force 5.64 Model for determining the resistance of an
unreinforced column flange.
must be transferred by the flange to the web of the supporting section (column) through bending.
This can be modelled as a plate clamped at three sides with a line load perpendicular to the free
end of the plate. By applying elementary yield line theory, a failure load is found that is linear
proportional to the plate plastic moment resistance mp. From this follows:
1 1
Ft,2 = 2Cmp = Cf t 2 with mp = f t 2 (5.36)
2 y,f f 4 y,f f
The factor C is taken as C = 14 based on tests. The total force that can be transferred is:

( 1
) (
Ft = Ft,1 + Ft,2 = fy,p t p t w + 2r + ·14 fy,f t 2f = fy,p t p t w + 2r + 7fy,f t 2f
2
) (5.37)
From equation (5.37), the relationship for the effective width beff given in EN 1993-1-8, cl. 4.10(2)
can be derived:
Ft fy,f t
f
b eff = = t w + 2r + 7kt f with k= 1 (5.38)
fy,p t p fy,p t p
The part of the flange at each side of the web where the tension force is transferred directly,
without the flange being subjected to bending, is generally indicated as s.
For a hot-rolled section s = r and for a welded section s = a 2. So generally for an unstiffened
I or H section:
beff = tw + 2s + 7ktf (5.39)

According to EN 1994-1-8, cl. 4.10(4), the effective width of a plate connected per-
pendicularly to a hollow section, in which the plate has the same width as the hollow
section, equals:
beff = 2tw + 5tf (5.40)
In order to guarantee sufficient deformation capacity, EN 1994-1-8, cl. 4.10(3), states

that ‘yielding over the plate width’ must occur before ‘failure of the effective width’.
This defines a minimum effective width:
force in this part of the welded
plate is transferred to the
column web by the stiffener fy,p
b eff bp (5.41)
fu,p
force on
the stiffener
In this, fu,p is the tension strength of the welded plate. For grade S235 steel, for
example, equation (5.41) means that the force to be transferred must be equal to at
tw + 2r least 65% of the yield strength of the plate. Equation (5.39) and (5.40) for the effective
width beff only apply when the plate is welded over its entire width. Moreover, according
to EN 1993-1-8, cl. 4.10(5), the weld thickness shall be determined based on the yield
stress in the plate.
When the effective width beff does not meet the requirement of equation (5.41), the
connection must be reinforced by welding stiffeners between the flanges of the (column)
section in line with the welded plate. Partial (column) depth stiffeners are preferred for
fabrication, because these do not have to fit exactly between the flanges of the section.
5.65 Welded stiffeners to reinforce the column flange. In figure 5.65, the forces in the coloured parts of the welded plate are introduced into
the column web via the stiffeners.
In addition to the flange, the web of the section can also govern the maximum tension
force that may be transferred by the plate without stiffeners. EN 1993-1-8, cl. 4.10(1),
refers to EN 1993-1-8, cl. 6.2.6.3. The force F may be distributed into the column web
rc
at an angle of 1:2.5 over the effective width beff,t,wc, in which the start of the root radius
of the web is assumed to be the critical cross-section. For a rolled section (fig. 5.66):
Ft
( )
b eff,t,wc = t fb + 2 2ab + 5 t fc + rc (5.42)
rc tfc
For a more detailed consideration of the transfer of tension forces in beam-to-column
connections, refer to Joints.
ab
beff,t,wc Ft
tfb
slope 1:2,5 5.66 Effective width of the column web in case of

an unstiffened flange loaded in tension.

Example 5.13
• Given. A bracket for a crane runway beam, formed from an HEA 300, is con- HEB 300
nected with fillet welds to an HEB 300 column (fig. 5.67). The steel grade is S235.
The welded connection is subject to a bending moment MEd = 190 kNm. It may be σx,fl
assumed that the shear force is completely resisted by the web of the bracket. +
e = 138 mm
• Question. Determine the required size of the fillet welds at the upper flange
x
and check if stiffeners in the column are necessary at the level of the upper flange
VEd MEd
of the bracket. –
• Answer. The upper flange of the bracket, in tension, forms a T-connection com-
HEA 300
prising a plate connected to the unstiffened flange of an H-section (see fig. 5.63).
The effective width beff of the flange of the bracket follows from equation (5.38):
beff
fy,f t z
f 235 19
k= = · = 1,36 1 k=1
fy,p t p 235 14
5.67 HEA 300 bracket for a crane runway
b eff = t w + 2r + 7kt f = 11 + 2·27 + 7·1·19 = 198 mm beam welded to an HEB 300 column.
The effective width must be larger than the minimum value of equation (5.41):
fy,p
235
b eff = 198 mm bp = ·300 = 196 mm (OK)
fu,p 360
The mean stress sx,fl in the flange of the bracket is:
MEde bb 190·10 6 ·138 300

σ x,fl = = · = 218 N/mm2 fy = 235 N/mm2 (OK)
Ibeam b eff 4 198
18263·10
According to EN 1993-1-8, cl. 4.10(5), the dimensions of the end fillet welds for the flange must be
determined based on σx,fℓ = fy. The minimum weld size follows from table 4.43 of Connections 4 (Welds):
a 0,46t = 0,46·14 = 6,4 mm a = 7 mm
The effective width of the column web follows from equation (5.42):
( )
b eff,t,wc = t fb + 2 2ab + 5 t fc + rc = 14 + 2· 2·7 + 5· (19 + 27) = 264 mm
The maximum stress in the column web σwc follows from the tension force in the upper flange of
the beam Tfl :
Tfl = b eff t p σ x,fl = 198·14·218 = 604·103 N
Tfl 604·103
σ wc = = = 208 N/mm2 fy = 235 N/mm2 (OK)
t w b eff,t,wc 11·264
Stiffeners are not required.

5.5.2 Bolted connections and prying forces
5.68 Transfer of a tension force through bending Unlike a welded connection loaded in tension, in a bolted connection the entire tension force must
of the column flange in a bolted connection.
be transferred, by bending of the flange, to the web of the supporting (column) section (fig. 5.68).
As a result of bending of the supporting column flange, so-called prying forces Q occur, and these
cause additional forces in the bolts.
EN 1993-1-8, cl. 6.2.5.1, uses an equivalent T-stub to determine the design value of the resistance
FT,Ed FT,Ed of a flange loaded in tension. In figure 5.69, the forces in such a T-stub are shown for one bolt
row, assuming yielding of the flange governs failure. The applied forces in the bolts Fb,Ed are then
smaller than the tension resistance of the bolts Ft,Rd. EN 1993-1-8, cl. 6.2.6, provides a method to
model a bolted connection as equivalent T-stubs, including the values to use for ℓeff, m and n. This
method is discussed comprehensively in Joints.
In addition to yielding of the flange, two other failure modes can occur, depending on the resistance
of the flange relative to the resistance of the bolts. The following three failure mechanisms for a T-stub
are discussed below:
Fb,Ed
– failure mechanism 1: yielding of the flange;
– failure mechanism 2: bolt failure combined with yielding of the flange;
– failure mechanism 3: bolt failure.
Failure mechanism 1: yielding of the flange

With this failure mode, plastic hinges occur in the flange near the start of the root radius (either
side of the web), and at the locations of the bolts (fig. 5.70). The location of a plastic hinge near
FT,Ed the root radius is assumed to be at a distance of 0,8r to the side of the web (see fig. 5.69); for
fillet welds, this distance equals 0,8 a 2 . From the moment equilibrium around the plastic hinge
at the location of the bolts follows:
1
4Mpl,Rd 4· ℓeff t 2f fy ℓeff t 2f fy
1 4
F m = 2Mpl,Rd FT,1,Rd = = =
2 T,1,Rd m m m
FT,1,Rd m
tf = (5.43)
Q Q
ℓeff fy
(prying force) Fb,Ed Fb,Ed (prying force)
The prying force Q1 follows from:
Mpl,Rd
Q1n = Mpl,Rd Q1 = (5.44)
n

The tension force applied to a bolt Fb,Ed,1 follows from vertical equilibrium:
FT,1,Rd
Fb,Ed,1 = + Q1 (5.45)
2
For N bolt rows, equation (5.45) is modified:

ℓ eff
FT,Ed
tw
yield line plastic
r hinge
bolt
Q 1
Q t + 0,8r Q
2 w
(prying force) Fb,Ed Fb,Ed (prying force)
Q
m n 5.69 Model of an equivalent T-stub in which
(e) yielding of the flanges is critical.
FT,Ed = FT,1,Rd FT,Ed = FT,2,Rd

FT,Ed = FT,3,Rd
1 1 1 1
2
FT,1,Rd 2
FT,1,Rd 2
FT,2,Rd 2
FT,2,Rd
Ft,Rd Ft,Rd
Q1 Fb,Ed Fb,Ed Q1 Q2 Ft,Rd Ft,Rd Q2
n m m n n m m n n m m n
tw + 1,6r tw + 1,6r tw + 1,6r
Mpl,Rd Mpl,Rd MEd

moment
diagram
MEd
Mpl,Rd
Q1 Q2
shear force
diagram
1 1 1
F
2 T,1,Rd
+ Q1 2
FT,2,Rd + Q2 2
FT,3,Rd
5.70 Schematic force distribution for the three
failure mechanism 1: failure mechanism 2: failure mechanism 3: possible failure mechanisms of an equivalent
yielding of the flanges bolt rupture and yielding of the flanges bolt rupture T-stub.
FT,1,Rd Q1
Fb,Ed,1 = + (5.46)
2N N
When determining Mpl,Rd of the flange, the net cross section at the location of the bolts does
not have to be taken into account. This is based on tests, which have shown that the yield line
runs through and around the head (or nut) of the bolt. In addition, n = e applies, in which e is
the distance from the centre of the bolt to the edge of the flange. However, for e > 1,25m then
n = 1,25m. This limiting value of n is based on tests, showing that the prying force occurs at a
distance of maximum 1,25m of the bolts.

Failure mechanism 2: bolt failure with yielding of the flanges
This failure mechanism occurs when at the start of the root radius of the T-stub a plastic hinge
occurs, but the bolts are not strong enough to allow the creation of a plastic hinge at the location
of the bolts. For this failure mode (fig. 5.70):
1
moment equilibrium F m = Mpl,Rd + MEd = Mpl,Rd + Q 2n
2 T,2,Rd
(5.47)
1
vertical equilibrium Q2 + F = Ft,Rd
2 T,2,Rd
From equation (5.47):
2Mpl,Rd + 2nFt,Rd
FT,2,Rd = (5.48)
m+n
In a connection with two or more bolts or bolt rows, according to table 6.2 of EN 1993-1-8:
2Mpl,Rd + n ΣFt,Rd
FT,2,Rd = (5.49)
m+n
Where ΣFt,Rd is the sum of the resistances of all bolts in the T-stub. For two bolt rows, with for
example two bolts in each row, ΣFt,Rd = 4Ft,Rd.
Failure mechanism 3: bolt failure

This failure mode occurs when the bolts are not strong enough to allow a plastic hinge to form at
the beginning of the root radius of the T-stub. In this case, the bolts are therefore critical and no
prying forces occur. From vertical equilibrium it follows that for a T-stub with two bolts (fig. 5.70):
FT,3,Rd = 2Ft,Rd (5.50)
For N bolt rows equation (5.50) becomes:
FT,3,Rd = ΣFt,Rd = 2NFt,Rd (5.51)

FT
(bolt resistance)
ΣFt,Rd
failure The lowest value of FT,Rd of each of the three failure mechanism
mechanism
3
1 defines the critical failure mode. The relationship between FT,Rd
2
2n
and the critical failure mode can also be represented graphically
m + 2n as a function of β; this is the ratio between the plate resistance
1 (= 4Mpl,Rd/m) and the bolt resistance (= ΣFt,Rd), see figure 5.71.
(plate resistance)
2
4Mpl ℓeff tf fy 5.71 Relationship between the plate resistance
1 2 b= =
2n m ΣFt,Rd m ΣFt,Rd and the bolt resistance of an equivalent T-stub
m + 2n with more than one bolt row.

The deformation capacity of bolts loaded in tension is much less than that of plates loaded in
bending. This means that the deformation capacity associated with failure mechanism 3 (bolt rupture)
is limited, whereas the deformation capacity associated with failure mechanism 1 (yielding of the
flange) is large. Therefore, it is recommended to choose bolts that have sufficient strength to ensure
that failure mechanism 1 or 2 will occur.
The use of the equivalent T-stub model for the design of beam-to-column connections is treated
in Joints.
Example 5.14
• Given. A welded T-section in grade S235 steel, loaded by a tension force FEd = 150 kN, is bolted
to the lower flange of an IPE 300 beam with six class 8.8 M16 bolts. The T-section comprises two
plates with a thickness tp = 10 mm, joined by a double fillet weld (a = 5 mm) (fig. 5.72).
• Question. Check the bolted connection loaded in tension.
• Answer. The design value of the tension resistance of an 8.8 bolt M16 follows from table 2.46
of Connections 2 (Bolts in clearance holes): Ft,Rd = 90,4 kN. The plastic moment resistance of the
plate is:
1 1
l t2 f
4 eff p y 4
·230·102 ·235·10 –6
Mpl,Rd = = = 1,35 kNm
γM0 1,0
IPE 300 5.72 A T-section bolted to the lower flange of

a beam, and loaded in tension.
tp = 10
a = 5 mm tp = 10
FEd FEd
35 90 35 45 70 70 45
e p2 e

The distances m and n follow from the geometry of the connection:
p2 – tp
90 – 10
m= – 0,8a 2 = – 0,8·5· 2 = 34,3 mm
2 2
n = e = 35 mm < 1,25m = 1,25·34,3 = 42,9 mm
For the three possible failure mechanisms, the design value of the tension resistance of the
flange of the T-section FT,Rd follows from equation (5.43), (5.49), and (5.50):
4Mpl,Rd
4·1, 35
failure mechanism 1 FT,1,Rd = = = 157 kN (critical)
m 34,3·10 –3
2Mpl,Rd + n ΣFt,Rd
2·1,35 + 35·10 –3 ·6·90,4
failure mechanism 2 FT,2,Rd = = = 313 kN
m+n ( 34,3 + 35)·10 –3
failure mechanism 3 FT,3,Rd = Σ Ft,Rd = 6·90,4 = 542 kN
Failure mechanism 1 is therefore critical. So the check becomes:
FEd 150
= = 0,96 1,0 (OK)
FT,Rd 157
The bolt force Fb,Ed,1 due to the load FT,1,R follows from equation (5.44) and (5.45):
Mpl,Rd
1,35
Q1 = = = 38,6 kN
n 35·10 –3
1 1 1 1
Fb,Ed,1 = F + Q1 = ·157 + ·38,6 = 39,0 kN
6 T,1,Rd 3 6 3
Without prying forces, the tension force in each bolt is 150/6 = 25,0 kN. Due to the prying force,
at failure the bolt force is increased by a factor 39/25 ≈ 1,5.

5.6 Transfer of compression forces
When a compression force acts perpendicular to an unstiffened I section or H section, yielding and
local buckling can occur in the web of the section adjacent to the flange (fig. 5.73). EN 1993-1-8, cl.
6.2.6.2, presents a calculation model in which the web of the section buckles locally over a length
beff. Because an effective area is used, the stresses that are in equilibrium with the compression
force are taken as the yield stress. The equations for beff are based on a distribution of the compres-
sion force through the flange and the root radii at a slope of 1:2,5, and through the thickness of any
connecting elements, such as an end-plate or a contact plate, at a slope of 1:1.
The design value of the resistance of an unstiffened web is then:
beff t w fy
Fc,wc,Rd = γM0 (5.52)
When the web of the section is loaded in shear in addition to compression force, according to
EN 1993-1-8, cl. 6.2.6.2 possible interaction effects must be taken into account using a reduction
factor ω, depending on the size and the direction of the applied force. Also, when large axial
compression is present in the section the resistance to local buckling is decreased. When the com-
pression force in the longitudinal direction results in a stress σcom,Ed ≤ 0,7fy, then the influence of
axial compression may be neglected. If σcom,Ed > 0,7fy, then the influence of axial compression
forces must be taken into account using a reduction factor kwc:
σ com,Ed
k wc = 1,7 – (5.53)
fy
befftwfy
FEd = Fc,Rd =
γM0
in flange and web: in plate:

slope 1:2,5 slope 1:1
tf
r
tw fy
beff beff
5.73 Buckling of a web adjacent to a flange.

In addition to the region adjacent to the flange, yielding and local buckling of the
web may occur over the full width of the web (plate buckling), see failure mechanisms
b and c in figure 5.74. In order to take into account plate buckling, EN 1993-1-8, cl.
6.2.6.2(1) defines a reduction factor ρ:
a. yielding and buckling of the web adjacent to the flange

λp 0,72 ρ = 1,0

λ p – 0,2 beff,c,wc dwc fy,wc
λ p > 0,72 ρ= with λ p = 0,932 (5.54)
2
λp Et 2wc
Where:
beff,c,wc effective width of the web of the section in compression;
b. plate buckling dwc clear depth of the web;
twc thickness of the column web;
fy,wc yield strength of the column web;
λp plate slenderness.
For a hot-rolled section and a welded I-section the following apply:
c. buckled shape with lateral displacement (sway)

hot-rolled section (
dwc = hc – 2 t fc + rc )
(5.55)
welded section (
dwc = hc – 2 t fc + 2ac )
5.74 Three failure mechanisms of a web loaded in compression
perpendicular to the unstiffened flange. Where:
hc depth of the section;
tfc thickness of the flange of the section;
rc root radius of the section;
ac throat thickness of the welds between flange and web.
For a more detailed consideration of the transfer of compression forces in beam-

to-column connections, including worked examples, refer to Joints (Steel Design 5).
5.7 Transfer of shear forces

In the shear zone of a beam-to-column connection (see fig. 5.4) only one check is
necessary, namely shear of the column web panel. EN 1993-1-8, cl. 6.2.6.1(2) presents
an equation for the design value of the plastic shear resistance:
0,9 fy,wc A vc
Vwp,Rd =
(5.56)
3 γM0

Where Avc is the area of the shear plane of the column. For I-sections and H-sections in classes
1 or 2 loaded parallel to the web (in-plane), see equation (7.19) in Structural basics 7 (Resistance
of cross-sections):
( )
A vc = A – 2bt f + t w + 2r t f (5.57)
Equation (5.56) is applicable when the slenderness of the web meets the following requirement:
d 235
69ε with ε= (5.58)
tw fy
If the shear resistance of the connection is insufficient, it can be increased by adding diagonal
stiffeners or doubler plates (either single or double sided); see figure 1.8 of Connections 1
(Connections in steel structures). Reinforcement of the column web with a doubler plate increases
not only the shear resistance, but also the tension and compression resistances. According to
EN 1993-1-8, cl. 6.2.6.1(10) the length of a doubler plate must be such that the plate covers the
full effective width (depth) of the web loaded in tension and compression. The verification of a
shear panel, and possible reinforcement options, is treated in Joints.
5.8 Literature
structures for buildings according to Eurocode 0, 1 and 3 (Steel Design 1), Bouwen met
Staal, Zoetermeer 2019.
2. J.W.B. Stark, Joints. Analysis and design of bolted and welded joints in steel frames and in
tubular structures according to Eurocode 3 (Steel Design 5), Bouwen met Staal, Zoeter-
meer, to be published.
3. EN 1993-1-3 (Eurocode 3. Design of steel structures. Part 1-3. General rules. Supplementary
rules for cold-formed members and sheeting), 2006 (incl. AC, 2009).
4. EN 1993-1-8 (Eurocode 3. Design of steel structures. Part 1-8. Design of joints), 2005
(incl. AC, 2009).

WILEY END USER LICENSE AGREEMENT
Go to www.wiley.com/go/eula to access Wiley’s ebook EULA.

Steel Design 3 Connections

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Steel Design 3 Connections

Uploaded by

Copyright:

Available Formats

Behaviour of connections in steel structures and design of mechanical

fasteners and welds according to Eurocode 3

Connections J.W.B. Stark

Connections J.W.B. Stark

preliminary work | connections | 1

published by Bouwen met Staal

The publication of this textbook has been made possible by:

© Bouwen met Staal 2021

2 | connections | preliminary work

Aeroview/Dick Sellenraad 1.39 MCB verbindingstechniek 2.14

preliminary work | connections | 3

2 Bolts in clearance holes 2-2

4 | connections | preliminary work

3 Slip-resistant connections, rivets and pins 3-2

preliminary work | connections | 5

5 Design and detailing of connections 5-2

6 | connections | preliminary work

1 Structural safety H.H. Snijder and H.M.G.M. Steenbergen, Structural basics.

Steel Design series

preliminary work | connections | 7

World Steel Association

8 | connections | preliminary work

Connections in steel structures

prof.ir. J.W.B. Stark

connections in steel structures | connections 1 | 1

1.1 The function of connections

2 | connections 1 | connections in steel structures

– When the steel structure has to be connected to structural elements formed

Figure 1.2 shows an example of a steel frame containing four storeys, in

Connections are an important part of any steel structure. The structural

1.2 Locations in a framework where

connections in steel structures | connections 1 | 3

preprocessing: weld metal over

4 | connections 1 | connections in steel structures

clearance holes and hand-tight tightened, see Connections 2, bolts loaded

or in a combination of tension and shear (fig. 1.5), depending on

When this movement of the bolts, together with the displacement of

possible in statically loaded structures. Special treatment of the

angle partial end-plate flange plate

full packing plate

angles end-plate flange plates with fin plate

connections in steel structures | connections 1 | 5

horizontal stiffeners backing plate horizontal and doubler plate welded on

1.8 Possible types of stiffening of a beam-to-column connection.

6 | connections 1 | connections in steel structures

connections in steel structures | connections 1 | 7

weld 1.3.2 Column base connection

1.10 Column splice between two column parts

cover plate HEM 240

8 | connections 1 | connections in steel structures

shear tab column base plate stiffener

connections in steel structures | connections 1 | 9

with a partial depth end-plate with angles with a fin plate

10 | connections 1 | connections in steel structures

end-plate – most strong and stiff, – high manufacturing costs due to

angles – automatic production of – lower strength than end-plate connections

fin plate – dimensional variations can – relatively low strength

1.3.3 Beam-to-column connection

Nominally pinned connection

connections in steel structures | connections 1 | 11

1.19 Rigid moment connections between beams and columns.

1.20 Semi-rigid (flexible) moment

12 | connections 1 | connections in steel structures

column with a single continuous with welded fin plates