Design Guide For Rectangular Hollow Section Joints

Construction with Hollow Steel Sections - Design guide for rectangular hollow section (RHS) joints under predominantly
static loading
Discuss me ...
@BODE
BE@O@N
FOR RECTANGULAR HOLLOW SECTION [RHS) JOINTS
UNDER PREDOMINANTLY STATIC LOADING
This material is copyright - all rights reserved. Use of this document is subject to the terms and conditions of the Steelbiz Licence Agreement
Created on 01 March 2012
Construction with Hollow Steel Sections - Design guide for rectangular hollow section (RHS) joints under predominantly static loading
Discuss me ...
CONSTRUCTION
WITH HOLLOW STEEL
SECTIONS
Edited by: Comite Internationalpour le Developpementet I’Etude

de la Construction Tubulaire
Authors: Jeffrey A. Packer, University of Toronto
Jaap Wardenier, Delft Universityof Technology

Yoshiaki Kurobane, Kumamoto University
Dipak Dutta, Technical Commissionof Cidect
Noel Yeomans, Chairman Cidect Joint and Fatigue Working Group
Discuss me ...
FOR RECTANGULAR
HOLLOW SECTION (RHS)
JOINTS UNDER
PREDOMINANTLY STATIC
LOADING
Jeffrey A. Packer, Jaap Wardenier, Yoshiaki Kurobane,
Dipak Dutta, Noel Yeomans
Verlag TUV Rheinland

Discuss me ...
Die Deutsche Bibliothek - CIP Einheitsaufnahme
Design guide for rectangularhollow section

(RHS) joints under predominantly static loading I
(ed. by Comite International pour le Developpement
et I’Etude de la Construction Tubulaire]. Jeffrey A.
Packer . . . - Koln: Verl. TUV Rheinland, 1992
(Construction with hollow steel sections)
ISBN 3-8249-0089-0, Reprinted edition
NE: Packer, Jeffrey A.; Comite International pour le
Developpement et I’Etude de la Construction
I Tubulaire
ISBN 3-8249-0089-0
0 by Verlag TUV Rheinland GmbH, Cologne 1996, Reprinted edition

Entirely made by: Verlag TUV Rheinland GmbH, Cologne
Printed in Germany
First edition 1992
Reprinted with corrections 1996
Discuss me ...
Preface
The square and rectangular hollow section is the youngest member in the family of steel
sections. Itsindustrial production started as late asin 1959 in England applying the processof
forming a round section first and then squaring the round in hot orcold operation conditions.
They possess practically all the favourable structural qualities of circular hollow sections in
various degrees, such as high radii of gyration aboutboth axes (higher compressive load than
that of conventional open sections), effective cross section for resisting torsional moments
(material is uniformly distributed about the polar axis), low surface area (less painting and
external fire protection leading to economy in maintenance), lack of ledges and edges (clean
structures), closed aerodynamicshape(reducedwindorwaveloads),andaesthetic
appearance (pleasing architecture). The inner space in both CHS and RHS gives possibilities
of increasing the strength of the sections byfilling them withconcrete, which, as the research
works in recent years show, also raises the fire resistance time. Other facilities
due to thehole
in the hollow sections are heating and ventilation of buildings by water or air circulation.
However, the distinct advantage of applying RHS in lieu of CHS emerges in the field of
fabrication, where the facility of end preparation with flat cuts for RHS is significantly easier
and more economical than that with profile cuts for CHS. Thetrend in fabrication techniques
with hollow sections, mainly by welding and bolting, is increasingly towards simple types of
joints i.e. whenever possible, joints without gussets, stiffeners or other means of reinfor-
cement, since the ratio of labour to material costs has increased rapidly in all industrialized
countries in the last thirty years.
Due to the very shapeof hollow sections andthe mere fact that they are closed sections, the
welding is by farthe most appropriate and commonfabrication technique for hollow Section
joints. Bolted connections remain nonetheles desirablein many cases and especially forsite
joints between prefabricated subassemblies. It is however of significant importance that the
behaviour of these joints depending on various geometrical parameters and loading types
should be understood before they can be correctly designed and calculated. This knowledge
is precisely necessary to establish, first of all, simple design formulae and then to select
particular structural arrangements conducive to a rational and economical application of
hollow sections.
In order to obtain this knowledge, which was nearly totally missing in the early sixties as
regards RHS joints, CIDECTfromitsvery inception in 1962concentrateditsresearch
activities predominantly on the various aspects of RHS jointing and some fifty individual
projects have been undertaken up to now in the research institutesof various partsof the world
with CIDECT involvement. Theresults of the CIDECT research works on the static behaviour
of RHS joints together with those obtained from other sources form the basis of this book
presenting the guide lines to design RHS joints in structures and the simple formulae to
calculate them. The aim is to disseminate to the architects, engineers and constructors the
most recent design methods RHS for joints along with worked-out examples, which will enable
them to design and construct secure and economical RHS structures.
It is worth mentioning that the CIDECT research results have been used in many national and
international design recommendations, e. g. DIN (Deutsche lndustrie Normung - German
-
Standard), NF (Norme Francaise French standard), BS (British Standard), CSA (Canadian
Standard), AIJ (Architectural Institute of Japan), AWS (America1 Welding Society) and IIW
(International Institute of Welding). The design formulae and guide lines given in this book
have been fullyincorporated into Eurocode 3 (ENV 1993-1 -1) “Design of steel structures, Part
1: Generalrules and rules for buildings, 1992”.
This design guide is the third of a series of eight in total, which CIDECT is in the process of
publishing recently:
5
Discuss me ...
- Design guide for circular hollow section (CHS) joints under predominantly static loading
(alreadey published)
- Structural stability of hollow sections (reprinted)
- Design guide forrectangular hollow section(RHS) joints under predominantly static loading
(reprinted)
- Design guide for hollow section columns exposedto fire (already published)
- Design guide for concrete filled hollow section columns under static and seismic loading
(already published)
- Design guide forstructural hollow sections in mechanical applications (already published)
- Designguidefor fabrication, assemblyanderection of hollowsectionstructures (in
preparation)
- Design guide for circular and rectangular hollow section joints under fatigue loading (in
preparation)
Weexpressour sincere thanks to thethreeverywellknownresearchers in thefield of
application technology of hollow sections - Professor Jeff Packer of University of Toronto,
Canada, Professor Jaap Wardenier of Delft University of Technology, The Netherlands and
Professor Yoshiaki Kurobane of KumamotoUniversity, Japan - who wrote this book in close
cooperation. Our special thanks go to Mr. D. Grotmann of Technical University of Aix-la-
Chapelle,Germany,whocheckedthecontents of thebookandmademanyimportant
comments.
Further we acknowledge the support of the CIDECT memberfirms with sincerethankfulness.
Dipak Dutta
Technical Commlssion
CIDECT
6
Discuss me ...
Contents
Page
l General:Advantages of hollowstructuralsections.andRHS relativeto
CHS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
2 Design of tubular
structures ..................................... 14
2.1 Truss configurations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
2.2 Trussanalysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
2.3 Effectivelengths forcompressionmembers .......................... 16
2.3.1 rules . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Simplified 16
2.3.2 Empirical methodfor bracing members . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
2.3.3 Long.laterallyunsupportedcompression chords . . . . . . . . . . . . . . . . . . . . . . 17
2.4 Trussdeflections . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
2.5 Generalconnectionconsiderations ................................. 17
2.6 Trussdesignprocedure .......................................... 18

2.7 Designofwelds . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
3 Weldeduniplanartrussconnections between RHS chordsandRHSor

CHSbracingmembers .......................................... 22
3.1 KandNconnections . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
3.2 T. YandXconnections . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29
3.3 KTconnections . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30
3.4 Graphicaldesign charts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32
3.5 Reinforcedconnections .......................................... 41
3.5.1 With stiffeningplates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41
3.5.1.1 KandNconnections . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41
3.5.1.2 T, YandXconnections . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42
3.5.2 With concrete filling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44
3.5.3 Design example ................................................ 46
3.6 Cranked-chord connections . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49
3.6.1 Design example ................................................ 49
4 Trussdesignexamples . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51
4.1 Uniplanar
truss ................................................. 51
4.2 Arch-formed
truss ............................................... 56
5 Boltedconnections . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57
5.1 Flange-plate
connections ......................................... 57
5.1.l Boltedon two sides of the RHS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57
5.1.2 Boltedonfoursides of the RHS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59
5.1.3 Designexamplefor bolted flange-plateconnection . . . . . . . . . . . . . . . . . . . . . 59
5.2 RHS to gusset-plateconnections ................................... 60
5.2.1 Netarea, effective netarea,andreduced effective netarea . . . . . . . . . . . . . . 61
6 RHS to
RHSmoment
connections ................................. 63
6.1 Vierendeel
connections .......................................... 63
6.1.1 introduction to Vierendeeltrusses . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63
6.1.2 Connection behaviourand strength . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64
6.1.2.1 In-planebendingmomentsfor T and X connections . . . . . . . . . . . . . . . . . . . . 70
7
Discuss me ...
Page
6.1.2.2 Out-of-plane bending moments for T and X connections . . . . . . . . . . . . . . . . 70
6.1.3 Connection flexibility . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71
6.1.4 Design
example ............................................... 72
6.2 Kneeconnections . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 74
7 Multiplanar welded connections . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 80
8 Otheruniplanarconnections . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 83
8.1 Trusses with RHS bracing members framing into the corners of the chord . 83
8.2 Trusses with flattened and cropped-end CHS bracing members . . . . . . . . . 84
8.3 Double chord trussconnections . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 88
8.4 Plate to RHS connections . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 89
9 List of symbols . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91
10 References .................................................... 94
CIDECT . International Committee for the Development and Study of Tubular

Structures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 100
i
Roof support structure for a swimming centre
8
Discuss me ...
1 General:Advantages of hollow structuralSections,

and RHS relative to CHS
The structural advantages of hollow sections have become apparent to most designers,
particularly for structural members loaded in compression or torsion. Circular hollow sections
(CHS) have a particularly pleasing shape and offer a very efficient distribution of steel about
the centroidal axes, as well as the minimum possible resistance to fluid, but specialized
profiling is needed when joining circular shapes together. As a consequence, rectangular
hollow sections (RHS) have evolved as a practical alternative, allowing easy connectionsto
the flat face, and they are very popular for columns and trusses.
Fabrication costs of all structural steelwork are primarily a function of the labour hours
required to produce the structural components. These need not be more with hollow section
design (RHS or CHS) than with open sections, and can even be less depending on conne
configurations. In this regardit is essential that the designer realize that the selection of hollow
structuralsectiontrusscomponents,forexample,determinesthecomplexity of the
connections at the panel points. It is not to be expected that members selected for minimum
mass can be joined for minimum labour time. That will seldom be the case because the
efficiency of hollow section connections is a subtle functionof a numberof parameters which
are defined by relative dimensions of the connecting members.
Handling and erecting costs can be less for hollow section trusses than for alternative trus
Their greater stiffness and lateral strength mean they are easier to pick up and more stable to
erect. Furthermore, trusses comprised of hollow sections are likely to be lighter than their
counterparts fabricated from non-tubular sections, as truss members are primarily axially
loaded and hollow structural sections represent the most efficient use of a steel cross-section
in compression.
Protection costs are appreciably lower for hollow section trusses than for other trusses. A
square hollow section has about213 the surface areaof the same size l-section shape, and
hollow section trusses may have smaller members as a result of their higher structural
efficiency. The absenceof reentrant corners makes the application of paint orfire protection
easier and the durability is longer. Rectangular (which includes square) hollow sections, if
closed at the ends, also have only four surfaces to be painted, whereas an l-section has eight
flat surfaces for painting. These combined features result in less material and less labour for
hollow section structures.
Regardlessof the typeof shape used to design a truss, it is generally false economy to attempt
to minimize massby selecting a multitude of sizes for bracing members. The increased cost
to source and to separately handle the various shapes more than offsets the apparent sav
in material. It is therefore better to use the same section size for aof group bracing members.
Circular hollow section connections are more expensive to fabricate than rectangular hollow
section connections. Joints of CHSrequire that the tube ends profile
be cut when the tubes are
to befitted directly together, unless the bracing tubes are much smaller than the chords. Mor
than that, the bevel of the end cut must generally be varied for welding access asone
progresses around the tube.If automated equipment forthis purpose is not available, semi
automatic or manual profile cutting has to be used, which is much more expensive than
straight bevel cuts on RHS.
In structures where deck or panelling islaid directly on the top chordof trusses, RHS offer
superior surfaces to CHS for attaching and supporting the deck. Other aspects to consider
when choosing between circular and rectangular hollow sections are the relative of fittingease
weld backing bars to RHS, of andhandling and stacking RHS. The latter is important because
material handling is said to be the highest cost in the shop.
Joint configurations are increasingly cheaper progressing from partial overlap, complete
overlap, to gap (see Fig. 1). Gap joints have the advantage of a single bevel cut, if the chord is
9
Discuss me ...
81 ' 82
e>d
(a) Gap connectlon nodlng (b) Gap connection with
e=O Dositive eccentricity
(cl Partial overlap connection ( d ) Total overlap connection

wlth negative eccentrlcity with negatwe eccentricity
e<O e< 0
-0.55<L
h0
< 0.25
Fig. 1 - Noding eccentricity, with permitted limits between which the resulting moment on the connection
can be ignored,for connection design
I
Aircraft museum; RHS columns and roof girders
10
Discuss me ...
Roof consjruction for a customs station
an RHS, and complete ease of fitting. Partial overlap joints have double cuts with minimum
flexibility infitting.
This detracts from the initial appeal of overlap connections, which the designerwill come to
recognize as usually having superior static and fatigue connection strength compared to gap
connections. Also, when the concealed portion of an overlappedbracing member needsto be
welded, it must be done before the overlapping bracing member is fitted. This preventsfitting
and tack welding of all members prior to final structural welding, an economical sequence
preferred by many fabricators.
Welding costs are sensitive to joint geometry, weld typeand weld size.Fillet welds usually do
not require the preparation of bevel surfaces that is inherent in almost all partial or full
penetration butt welds. A 12 mm fillet weld hastwice the resistance of a 6 mm fillet; however,
it has four times the volume. Therefore cost per unit resistance is clearly better with smaller
member thicknesses and thus with smaller size welds. An RHS bracing member whose width
is the sameas anRHS chord member presentsa condition where the side wallsof the bracing
line up with the round corners of the chord. Depending on the corner radius and the wall
thicknesses, at best there is a flare bevel joint (more awkward than a fillet), or more likely,
11
Discuss me ...
especially for cold-formed sections a flare bevelwith a gaprequiring a customfitted backing

bar. Thus, less-than-full width RHS bracing to chordconnections are more economical.
Connection piecessuch as gusset plates obviously add materialand labour costs. Welding is
essentially doubled because loads are transferred twice instead of once,first from a member
to the connectingpiece, then from that piece to another member. Hence, direct connection of
one hollow sectionto another is tobe preferred.
Stiffeners and other reinforcement, which similarly increase costs of material and labour,
should always be keptto a minimum and used only when truly needed (e. g. for repair).
The design recommendations presented in the Guide are applicable for hollow sections
having a nominal yieldstress of up to355 Nlmm‘ and anominal yield to ultimate stress ratio
(fYi/fui)S 0.8. The scope ofthis Guide covers connections subjected to “predominantly static
loading” and alternativedesign philosophiesmaybe required under different loading
conditions. (For example, under earthquake loading, when connection resistances greater
than member resistances may be required.)
Automobile transporter
12
Discuss me ...
I
Container lifting vehicle

13
Discuss me ...
2 Design of tubular structures
This chapter deals with

the design philosophy applicable triangulated
to (e. g. Warren or Pratt)
planar RHS trusses with bracing members directly welded to single-section chord members.
2.1 Trussconfigurations
Some of the common truss types are shown in Fig. 2. Warren trusses will generally provide the
most economicalsolution since their long compression bracing members can take advantage
of the fact that RHS are very efficient in compression. They have about half the number of
bracing members and halfthe number of connections compared to Pratt trusses, resulting in
considerable labour and cost savings. The panel points of a Warren truss can be located at the
load application points on thechord, if necessary with an irregulartruss geometry; however,
even if the chord is loaded in bending, that disadvantage is usually less significant with RHS
chords than withalternatives. If support is required at all load points to a chord (for example.
to reduce the unbraced length), a modified Warren truss could be used rather than a Pratt
truss by adding vertical members as shownin Fig. 2(a).
Warren trusses provide greater opportunities to use gapjoints, the preferred arrangement at
panel points. Also, when possible, a regular Warren truss achieves a more "open" truss
suitable for practical placement of mechanical, electrical, and other services.
Truss depth is determined in relation to the span,loads,maximum deflection, etc., with
increased truss depth reducingthe loads in the chord members andincreasing the lengths of
the bracing members. The ideal span to depth ratio is usually foundto be between 10 and 15
[l]. If the total costs of the building are considered, a ratio nearer 15 will represent optimum
value.
(b) (d)
Fig. 2 - Common RHS trusses

(a) Warren trusses (modified Warrenwith verticals)
(b) Pratt truss (shownwith a sloped roof, but may have parallel chords)
(c) Finktruss
(d) U-framed truss
2.2 Trussanalysis
Elastic analysis ofRHS trusses is frequently performed by assuming that all members arepin
connected. Nodal eccentricities between the centrelines of intersecting members at panel

points should preferably be kept within the limits shown in Fig. 1 . These eccentricities produce
primary bending momentswhich, for a pinned joint analysis, need only be taken into account
in member design when proportioning the compression chord, by treating it as a beam-
14
Discuss me ...
column.This is doneby distributingthe panel point moment(sumofthehorizontal

components ofthe bracingmember forcesmultipliedby the nodal eccentricity) to the chordon
the basis of relative chord stiffnesson either sideof the connection (i.e. in proportion to the
values of moment inertia
of divided by chord length the to next panel point, on eitherof side the
connection). Theeccentricitymoments can be ignored for design the of the tension chord and
bracing members.
Eccentricity moments can be ignored for the design of the connections provided that the
eccentricitiesare within thelimits shown in Fig. 1. If these eccentricity limits are violated, the
eccentricity momentmayhave adetrimental effectonconnectionstrengthandthe
eccentricity moment mustbe distributedbetween the members at a connection. If moments
are distributedto the bracing members, the connection capacity must then be checked for the
interaction between axial load andbending moment, for eachbracing member.
A rigid joint frame analysis is not recommended for most planar, triangulated, single-chord,
directly-weldedtrusses, as it generally tends to exaggerate bracing member moments, and
the axial force distributionwill still be similarto that for a pin-jointedanalysis.
Transverseloadsappliedtoeither chord awayfrom the panel points produceprimary
moments which must always be taken into account whendesigning thechords.
Computer plane frame programs are regularly used for truss analysis.In this case the truss
can be modelled byconsidering acontinuous chord with bracingmembers pin connected to
it at distances of + e or - e from it (e being the distance from the chord centreline to the
intersection of the bracing member centrelines). The links to the pins are treated as being
extremely stiff asindicated in Fig. 3.The advantage of this model is that a sensible distribution
of bending moments is automatically generated throughout the truss, for cases in which
bending moments needto be takeninto account in thedesign of the chords.
Noding condition
for
most
overlap
Extremely
stiff ,Pin
ding condition
members for many gap
connections
Fig. 3 - Plane frame joint modellingassumptions to obtain realisticforces for memberdesign
Table 1 - When moments need to be consideredfor RHS truss design

type of moment primary secondary
moments nodal eccentricity transverse member secondary effects

due to loading such as local
deformations
compression chord Yes no

design
design of other no no
members
design of connections no, provided yes, influences f (n) no, provided

eccentricity limitsare parametric limits of
not exceeded validity are met
15
Discuss me ...
Secondary moments,resulting from endfixity of the bracingmembers toa flexible chord wall,
cangenerally be ignored for both membersandconnections, provided thatthere is
deformation and rotation capacity adequateto redistributestresses after somelocal yielding
at the connections [2, 31. This is the case when the prescribed geometriclimits of validity for
design formulae given in the next chapter are followed. Welds in particular need to have
potentialfor adequate stress redistributionwithout prematurefailure, and this will be achieved
with therecommendations given in Section 2.7.
Table 1 summarizes when moments need to be considered for designing an RHS truss.
Plastic design could be used to proportion the chords of a truss by considering them as
continuous beams withpin supportsfromthe bracing members.Insuch a designthe
plastically designed members must be plastic design sections and the welds must be sized to
develop the capacityof the connected bracing members.
2.3 Effective lengths for compression members

To determine the effective length KL fora compression memberin atruss, the effective length
factor K can always be conservatively taken as 1.0. However, considerable end restraint is
generally present for compression members in an RHStruss, and it has been shown that K is
generally appreciablyless than 1.O [4, 5,6,71.This restraint offeredby members framing into
a connection could disappear, or be greatly reduced, if all members were designed optimally
for minimum mass, thereby achieving ultimatecapacity simultaneously under static loading
[8]. In practice, design for optimal or minimum mass will rarely coincide with minimum cost;
the bracing members are usually standardized to a few selected dimensions (perhaps even
two), to minimize the number of section sizes for the truss. In the unlikely situation that all
compression bracing members are proportioned on the basis of a single load combination,
and all reach their compressive resistances at approximately the same truss loading, an
effective lengthfactor of 1.O is recommended.
CIDECT has sponsored and co-ordinatedextensive research work to specifically address the
determinationof effective lengths in hollow section trusses,resulting inreports from CIDECT
Programs 3E-3G and Monograph No. 4 [4,9]. More recently,a re-evaluation ofall test results
has beenundertakento produce recommendations for Eurocode 3 [5,6,10].This has resulted
in the following effective lengthrecommendations, which are implemented in the truss design
example given in this guide (Chapter 4).
2.3.1Simplifiedrules
For RHS chord members:

In the plane
of the truss
KL = 0.9L where L is distancebetween chord panel points.
In the plane perpendicularto the truss
KL = 0.9 L where L is the distance between points of lateral support for the chord. (2.2)
For RHS or CHS bracing members:
In either plane
KL = 0.75 L where L is the panelpoint topanel pointlength of the member.(2.3)
These values of K are only valid for RHS members which are connected around the full
perimeter of the member, withoutcropping or flattening of the members. Compliance with the
connection design requirements restrictive control on
of Chapter 3 will likely place even more
the member dimensions.
16
Discuss me ...
2.3.2 Empirical method for bracing members
For trusses withidentical width RHS top and bottom chord members and with a:
circular bracingmember weldedto rectangular chords
K = 2.35(dyLb,)0.25 but 0.55 K I0.75
rectangular bracing member weldedto rectangular chords
K = 2.3 (bFLb,)0.25 but 0.55 K I0.75
where L is again the panel point
to panel point length of the bracing member and
“rectangular”
includes “square”.
CIDECT MonographNo. 4 [4] presented a method fordetermining a bracing member effective
length intrusses which had different width, or section shape, members for
the topand bottom
chords. This has not been addressed in the latest provisions for Eurocode 3 [lo],so it is
recommended that theeffective lengthfactor K be calculated for the connectioncondition at
each end of the bracing member, and the higher value be used. Another conservative rider
from CIDECT Monograph No. 4, which impacts upon Eqns. 2.4and 2.5above, should also be
added to the above recommendations:
For RHS chord members,bois replaced by h, when h, < bo.

For RHS bracing members, bi is replaced by hi when hi > bi.
2.3.3 Long, laterally unsupported compression chords
Long, laterally unsupported compression chords can exist in pedestrian bridges such as
U-framedtrusses, and in roof trusses subjected to large wind uplift. The effective lengthof
such laterally unsupported truss chords can be considerably less than the unsupported
length. For example, theactual effective length of a bottom chord, loadedin compression by
uplift, depends onthe loadingin the chord, the stiffness of the bracing members, the torsional
rigidity of the tension chord, thepurlin to truss connections andthe bending stiffness of the
purlins. The bracing members act as local elastic supports at each panel point. When the
stiffness of these elastic supportsis known, theeffective length of the compression chord can
be calculated. A detailed method for effective length factor calculation has been given by
CIDECT Monograph No. 4 [4].
2.4 Trussdeflections
For the purpose of checking the serviceability condition of overall truss deflection under
specified (unfactored) loads, an analysis with all members being pin-jointed will provide a
conservative (over) estimate of truss deflections when all the connections are overlapped
conditions is to assume continuouschord members
[l 1 , 121.A better assumption for overlap
and pin-jointed bracing members. However, for gap-connected trusses, a pin-jointedanalysis
still generallyunderestimatesoveralltrussdeflections,because of the flexibility ofthe
connections [l 1 , 12, 13, At 141.
the service load level, gap-connected RHS trussdeflections
have been underestimated by around 12 to 15% [l 1, 131.Thus, a conservative approach for
gap-connected RHS trussesis to estimate the maximum trussdeflection by 1.15times that
calculated from a pin-jointedanalysis.
2.5 General connection considerations
It is essential that the designer has an appreciation of factors which make it for RHS
possible
members to be connected together at truss panel points without extensive (and expensive)
17
Discuss me ...
reinforcement.Apparenteconomiesfromminimum-massmemberselection will quickly

vanish at the connections if a designerdoesnothave a knowledge of the critical
considerationswhich influence connection efficiency.
1 . Except for100% overlap connections, chords should generally have thick walls rather than
thin walls. Thestiffer walls resist loads from the bracing members more effectively, and the
connection resistance thereby increases as width to thickness ratios decrease. For the
compression chord, however, a large thin section is more efficient in providing buckling
resistance, so for this member the final RHS wall slenderness will be a compromise
between connection strength and buckling strength, and relatively stocky sections will
usually be chosen.
2. Bracing membersshouldhave thin wallsratherthanthickwalls(exceptoverlap
connections), as connection efficiency increases as the ratio of chord wall thickness to
bracing wallthicknessincreases.Inaddition, thin bracing memberwalls will require
smaller fillet welds for a prequalifiedjoint.
3. Ideally, R H S bracing members should not bethe same width asRHS chord members, as
this can present an awkward the cornerof the chord

flare bevel weld situation for the atjoint
section. A preferred arrangement is bracing members just sufficiently narrower than the
chord to permit thebracing member, and someof the fillet weld, to sit on the“flat” of the
RHS chord member.
4. Gap connections (for
K and N situations) arepreferredto overlap connections because the
members are easier to prepare, fit and weld.
5. When overlapconnectionsare used, at least
a quarter of the width (in the plane ofthe truss)
of the overlappingmemberneedstobeengaged in theoverlap,however 50% is
preferable.
6. An angle of less than 30’ between a bracingmember anda chord creates serious welding
difficulties and is not covered by the scopeof these recommendations (see section2.7).
2.6 Trussdesignprocedure
In summary,the designof an RHS truss should be approached

in the followingway to obtain
an efficient and economical structure.
I. Determine the truss
layout, span, depth, panel lengths, truss spacing
and lateral bracing
by the usual methods, but keep the numberof connections toa minimum.
(I. Determine loads at connections and on members; simplify these to equivalent loads
at
the panel points.
111. Determine axial forces in all members by assuming thatjoints are pinned andthat all
member centrelines are noding.
IV. Determine chord member sizes by considering axialloading, corrosion protection and
tube wall slenderness.(Usual width to thickness ratios are15 to 25.) Aneffective length
factor of K = 0.9 can be used for the designof the compression chord.
V. Determine bracing member sizes based on axial loading, preferably with thicknesses
smaller than the chord thickness.
The effectivelength factor for the compression bracing
members caninitially be assumedto be 0.75(see section 2.3.1).
VI. Standardize the bracingmembers toa few selected dimensions (perhaps even two), to
minimize the numberofsectionsizesforthe structure. Consider availability of all
sections when making member selections. For aesthetic reasons, a constant outside
18
Discuss me ...
member width may bepreferredfor all bracing members, with wall thicknesses varying;
but this will require special quality control proceduresin the fabrication
shop.
VII. Layout the connections, trying gap joints first. Check that the connection geometry and
member dimensions satisfy the validity ranges the for dimensional parameters givenin
Chapter 3, with particular attention to the eccentricity limits. Consider the fabrication
procedure whendeciding on a connection layout.
VIII. Check the connection efficiencies withthe charts givenin Chapter 3. In some instances
(for example when using rectangular rather than square RHS members, or when a
slightly moreaccuratesolution is needed),directuse of theconnectionfactored
resistance equations givenin chapter 3 may be required.
IX. If theconnectionresistances(efficiencies)arenotadequate,modifytheconnection
layout (for example, overlap rather
than gap), or modify the bracing or chord members
as
appropriate, and recheck the connection capacities. Generally, only
a few connections
will need checking.
X. Check the effect of primary moments onthe design of the chords. For example, use the
proper load positions (rather than equivalent point
panelloading); determine the bending
moments in thechords by assumingeither: (a) pinned jointseverywhereor (b)

continuous chords with pin-ended bracing members. For the compression chord, also
determine the bending moments produced by any noding eccentricities, by using either
of the above analysis assumptions. Then check that the factored resistanceof all chord
members is still adequate, under theinfluence of both axialloads and primary bending
moments.
XI. Check truss deflections (see section 2.4) at the specified (unfactored) load level, using
the proper load positions.
XII. Designwelds.
2.7 Design of welds
Except forcertain K and N connections withpartially overlapped bracing members (as noted
below), a welded joint should be established aroundthe entire perimeterof a bracing member
by means of a butt weld, a fillet weld, or a combination of the two. Fillet welds which are
automatically prequalified forany bracing memberloadsshouldbedesignedtogive a
resistance that is not less than the bracing member capacity. This results in the following
minimum throat thickness (a) for fillet welds around bracing members, assuming matched
electrodes and IS0 steel grades[2]:
a 2 0.95 t, for S 235 (fyl = 235 N/mm2)
a 2 1.OO t, for S 275 (fyl = 275 N/mm2)
a 2 1.07 t, for S 355 (fyl = 355 N/mm2)
If welds are proportioned on the basis of particular bracingmember loads, the designer must
recognize thatthe entire length of the weld may not be effective, and the model for the weld
resistance must be justified in terms of strength and deformation capacity.
An effective length
of RHS bracing member welds in planar, gap, K and N connections subjected to predomi-
nantly static axial load, is given by[15]:
Effective length = (2 h,/sin 0,) + b,, for 0, 2 60°

Effective length = (2 h,/sin 0,) + 2 b,, for 0, S 50°
For 50° < 0, 60°, a linear interpolationhas been suggested [16].
19
Discuss me ...
Retractable arched roof structureof the Toronto SkyDorne
For overlappedK and N connections, limited experimental research on connections with 50°/o
overlap has shown that the entire overlapping bracing member contact perimeter can be
considered aseffective [15].
These recommendations foreffective weld lengths in RHS K and N connections satisfy the
required safetylevelsforuse inconjunctionwithboth Europeanand North American
steelwork specifications[l 51. Based onthe weld effectivelengths forK and N connections, an
extrapolation has been postulated RHS for T, Y and X connections under predominantly static
load [ 16, 171:
length Effective = (2 h,/sin e,) (2.8)
45
Detail A2 DetaiLC2
Fig. 4 - Recommended weld details (corner

radii shown are for hot-formedRHS)
20
Discuss me ...
With overlapped K and N connections, welding of the toe of the overlapped memberto the
chord is particularly important for 100% overlap situations. Forpartial overlaps, the toe of the
overlapped member neednot be welded, providing the components, normal to the chord, of
the bracing memberforcesdonot differ bymore than about 20%. Whentheseforce
components donot balance, the more heavily loaded bracing member should be the “through
member” and its full circumference should be welded to the chord. Generally, the weaker
member (defined by wall thickness times yield strength) should be attached to the stronger
member, regardless ofthe load type.
It is generally more economical to use fillet welds than butt (groove) welds,providingthe fillet
weld sits on the“flat” of the RHS member, however the upper limit on throat or
leg size forfillet
welds will depend onthe fabricator.Most welding specificationsonly allowfillet welding atthe
toe of a bracingmember if Oi 2 60’. Because of the difficulty of welding at the heel of a bracing
member at lowBi values, a lower limit for the applicabilityof the connection design rules given
herein has been set at 0, = 30’. Some recommended weld details[2] are illustrated in Fig. 4.
21
Discuss me ...
3 Welded uniplanartrussconnections between RHS

chords and RHS or CHS bracing members
All connection design formulae presented in this guide are given in limit states strength (or
LRFD*) terms. This means that the effect of the factored loads (the specified or unfactored
loads or characteristic loads multipliedthe byappropriate load factors), should not exceed the
factored resistance of the connection, N*. The connection factored resistance expressions
given in this guide, in general, already include a material and connection partial safety factor
(Y,,,,)or connection resistance factor (6).If allowable stress design (ASD), or working stress
design, is used, the connection resistance expressions should in addition be divided by an
appropriate load factor to obtain allowable connection resistances. In this case a load factor of
1.5 is recommended.
This guide uses terminology adopted by CIDECT and IIW (InternationalofInstitute Welding) to
defineconnectionparameters,whereverpossible.Fig. 5 showssome of thecommon
connection parameters for gap and overlap uniplanar K connections. Definitions of all terms
are given in the List of Symbols.
Ad2 Connection shown d

f i / ~ b 2 Isolated
below b,
Fig. 5 -Typical RHS chord truss K-connection
Eccentricity, e, is positive when measured towards the outside of a chord, and negative
towards the inside. The gap or overlap, g orq respectively, as well as the eccentricity,e, may
becalculatedbyEqns.3.1and3.2below.SeeFigs. 1, 5and 16 for illustration of the
connection parameters.
g=(.++) elsin
sin (0, + 0,)e2
sin - 2 h,sin
0, 2 h2
sin
0,
Note that a negative value

of gap (g) corresponds to an overlap(q).
h2 sin 0, sin 8, -
h,
-
sin (0, + 0,) 2 (3.2)
Note that g abovewill be negative for an overlap
These equations also apply for panel points which have a stiffening plate on the
of the
surface
h0 .
chord, in whichcasethe ierm 2 ISreplacedby , where t, is thestiffeningplate
thickness.
LRFD = Load and resistance factor design
22
Discuss me ...
3.1 K and N connections
The majority of RHS truss connections have one compression bracing member and one
tension bracing member welded to the chord as shown in Fig. 5. The Warren arrangementis
commonly referred to as a K connection and the Pratt as an N connection. The latter is
basicallyaparticularcase of theformer;bothcanbeeithergaptypeoroverlaptype
connections.
Experimental research (for example byWardenierandStark [18])on RHS welded truss
connections has shown that different failure modes can exist depending on the type of
connection, loading conditions, and various geometric parameters. Failure modes have been
described [l91for RHS as illustrated in Fig. 6.
Mode B: Punchlng shear fallure

of the chord face
Mode A: Flexural fallure of

the chord face
-_ _ ----
Mode D. Local buckllng of

the braclng member
____ Mode F: Local

buckltng of
5--- ~ the chord walls
Mode E Overall shear failure

of the chord
- +.-
Secrlo"
__ ____
"lew
Mode G Local buckllng of

the chord face
Fig. 6 - Failure modes forK and N-typeRHS truss connections
Mode A: Plastic failure of the chord face (one bracing member pushing the face in, and the
other pulling it out).
Mode B: Punching shearof the chord face around a bracing member (either compression or
tension).
Mode C: Rupture of the tension member or its welding.
Mode D: Local buckling of the compression bracing member.

Mode E: Shear failureof the chord memberin the gap.
Mode F: Chord wall bearingor local buckling under the compression bracing member.
Mode G: Local bucklingof the chord face behind the heel of the tension bracing member.
23
Discuss me ...
Table 2 - Factored resistance of axially loaded welded connections between square or circular
bracing members and asquare chord section
type of connection factored connection resistance (i = 1.2)
T-. Y - and X-connections S 4 0.85 yielding

face
chord
basis:
N; = + 4 (I- p ) O
1 . f(n)
K- and N-gap connections Y I1.0 plastification

facechord
basis:
L J
f ho
(i = 1,2)
K- and N-overlap connections' 25% I0, < 50% basis:

effective width
50% 5 0, < 80% basis:

effective
width
N,* = f, ' 1, [2h, - 4t, + be + be(,,J

I /
3, 2 80% width
effective
basis:
N,' = f, t, [2h, - 4t, + b, + be(ov)]
circular bracings multiply formulae by n / 4 and replace b1,2and hl,z by d,,2
functions
0.4
f (n) = 1.3 + - . nforn 0 (compression)
f(n) = 1 .O forn h 0 (tension)
P
but 5 1.0
Ib, I Ib,
Note l : Effective width computations need only be done for the overlapping bracing member. How-
ever the efficiency (the factored connection resistancedivided by thefull yield capacityof the
bracing member) of the overlapped bracing member is not to be taken higher than that of the
overlapping bracing member.
24
This material is copyright - all rights reserved. Use of this document is subject to the terms and conditions of the Steelbiz Licence Agreemen
Discuss me ...
Table 2 a - Range of validity of Table 2

type of connection parameters (i = 1 or 2, j = overlapped bracing)
connection
T blltl
compression tension
eccentricity
1.25 Ip I0.85a 5 1.25 dt I 35

b0
e 0.1 + 0.01 - I 35
10
p z 0.35 l I gzt,+t, e
0.55 5 - I0.25
fi
-
h0
K, N overlap e 0.25 5 1.1
~~ ____
for circular d, d2 limitations as above ford, = b,

bracings 0.4 I- I 0.8 - I 50
b0 t,
(web members)
Note: a) Outside this range of validity other failure criteria may be governing; e. g., punching shear, effective width, side wall failure, chord shear or loca
buckling. If these particularlimits of validity are violated the connection may
still be checked as onehaving a rectangular chordusing Table 3, providec
the limits of validity in Table 3 a are still met.
f,,, f, 5 355 Nlmm2, ,f (or fy,)lfu,I 0.8
b,
c)
9 > 1.5 (1
If - - p), treat as a T or Y connection
bn
Discuss me ...
Table 3 - Factored resistance of axially loaded welded connectionsbetween rectangular, square

or circular bracing membersand a rectangular chord section
tvDe of connection factored connection resistance (i = 1.2)

T-, Y- and X-connections BI0.85 basis: chord face yielding
N; = + 4(1 - P ) 0 5
1 .f(n)
use linear interpolation of

chord face yielding and chord
side wall criteria
p > 0.85 basis: effective width

N; fyl . t, [2 h, - 4t1 + 2 be]
0 . 8 5 ~ p 1 1- 1 / ~ basis: punching shear
K- and N-gap connections basis: chord face yielding
basis<hord shear
basis: effective width

N:= f, . ti [2 h, - 41, + b, + be]
p 5 1 - 1lY basis: punching shear
K- and N-overlap connections similar to connections of square hollow sections (Table 2)

circular bracinas multiply formulae by d 4 and replace b, and h, by dl
functions
tension: f, = ,,f compression: f, = ,f (T-, Y-connections) f, = 0.8 sin el .,f (X-connections)
,,f = buckling stress according to the relevant steelwork specification, using acolumn slenderness
ratio (KLlr) of 3.46 (ho/to- 2) (1/sin 1 9 , ) ~
f(n) = 1.O for n 2 0 (tension)

0.4
f(n) = 1 . 3 + - - . n f o r n < O
P
but I1.0 (compression) For square and rectangular bracings, A, = (2 h, + CY . bo) . to
For circular bracings, A, = 2 h,. to
Ib, I Ibi
Note': For X-connections with angles0 < 90°, the chord sidewalls must be checkedfor shear
26
Discuss me ...
Table 3 a - Range of validity of Table 3

1
type of connection parameters (i = 1 or2, j = overlapped bracing)

connection
gap/overlap eccentricity
b./b.
t./t.
B 0.25
K, N gap B 0.1
P 2 0.35
bo
+ 0.01 -
t0
535 1 535
-
5 35 0.5 (1 - p) 5
b0
I
g z t , +t,
1.5 (1 - p)b)
- 0.55 I
e
- 5 0.25
h0
K, N overlap P 0.25 I40
for circular d, limitations as aboveford, = b,

bracings 0.4 I- I0.8
b0
(web members)
Note: a) f,,, f,, I355 N/mm2, ,f (or fy,)/fu,I0.8
b) If 2 > I .5 (I - p), treat as a T or Y connection

b0
Discuss me ...
Failure in test specimens has also been observed to be a combinationof more than one failure
mode.
It should be noted here that Modes C and D are generally combined together under the term
"effective width failures", and are treated identically since the connection resistance in both
cases is determined bythe effective cross sectionof the critical bracing member, with some
bracing member walls possibly being only partially effective.
Plastic failureof the chord face (Mode A) is the most common failure mode for gap connections
with small to medium ratiosof the bracing member widths to the chord width p. For medium
width ratios ( p = 0.6 to 0.8),this mode generally occurs together with tearing in the chord
(Mode B) or the tensionbracingmember(Mode C) althoughthe latter onlyoccurs in
connections with relativelythin walled bracing members. ModeD, involving local bucklingof
the compression bracing member, is the most common failure mode for overlap connections.
Shear failure of the entire chord section (ModeE) is observed in gap connections where the
width (or diameter) of the bracing members is close to thatof the chord (p = 1.O), or where
h, < bo.Local buckling failure (ModesF and G) occurs occasionallyin RHS connections with
high chord width (or depth) to thickness ratios (bolt, or h,/t,). It has been found1191 that, in
some cases, oneor two governing modes can be used to predict connection resistance. For
example, ModeG failure, is excluded from the subsequent design expressions by restrictions
on the range permitted for geometric parameters.
Various formulae exist for the connection failure modes described above. Some have been
derived theoretically, while others are primarily empirical. The general criterion for design is
ultimate resistance, but the recommendations presented herein, and their limits of validity,
have been set so that a limit state for deformation is not exceeded by specified (service)
loads.
Limit states design recommendations are summarized in Tables 2 and 2 a (for square chords),
andTables3and 3 a (for rectangularchords). A number of observationsfor K and N
connections can be made from an examination of these tables:
A common design criterion for all K and N gap connections is Mode A, plastic failure of the
chordface.Theconstants in theresistanceequationsarederivedfromextensive
experimental data, and the other termsreflect ultimate strength parameters suchas plastic
moment capacity of the chord face per unit length (fyo t34), bracing to chord width ratio, p,
chord wall slendernessY,and the termf (n) which accounts for the influence of compression
chord longitudinal stresses.
Tables 2 and 3 show that the resistance of a gap K or N connection with an RHS chord is
largely independentof the gap size (no gap size parameter).
Table 2 which is restricted to square RHS chords was derived from the more general Table 3
and uses more confined geometric parameters. The resultis that gap K and N connections
with square RHS need only be examined for failure Mode A, whereas those with rectangular
Fig. 7 - Shear area (A,) of the chord in thegap region of an RHS K or N-connection
28
Discuss me ...
RHS must be considered failurefor Modes B,C or D, and E as well. This approach has allowed
the creation of helpful graphical design charts which are presented later for connections
between squareRHS.
In Table 3, the Mode E check for chord shear in the gap of K and N connections involves
dividing the chord cross sectioninto two portions. Thefirst is a shear areaA, comprising the
side walls plus part of the top flange, shown in Fig. 7, which can carry bothshear and axial
loads interactively.The secondis the remainingarea, A, - A, which is effective in carrying
axial forces but not shear.
Tables 2 and 3 present a range of resistances based on the concept of effective width for
square and rectangular RHS overlap connections, starting with 25% overlap, which is the
minimum to ensure overlap behaviour. The resistance increases linearly withoverlap from
25% to 50%, is constant from50% to 8O%, then is constant above80% at a higher level. Fig.
8 illustrates the physical interpretation of the expressions for effective width given in the
tables.
Fig. 8 - Physical interpretation of effective width terms
3.2 T, Y and X connections
In the sameway that an N connection can be considereda particularcase of the general K

connection, the T connection is a particularcase of the Y connection. The basic difference
betweenthetwotypes is thatthecomponent of load normal to the chord in T and Y
connections is resisted by shear and bending in the chord, whereas withK or N connections
thenormalcomponentfromone bracing member is balanced primarily bythesimilar
component in theother.
The limit states design recommendations for T, Y and X connections are summarized,in
Tables 2 and 2 a (for square chords), and Tables 3 and 3 a (for rectangular chords).As with K
and N connections, various observations can be made from the Tables:
Resistance equationsin Tables 2 and 3, forp S 0.85, are based ona yieldline mechanism in
the RHS chord face. By limiting connection design capacity under factored loads to the
connection yield load, one ensures that deformations will be acceptableat specified (service)
load levels.
For full width (p = 1 .O)RHS T, Y and X connections, flexibility is no longer a problem, and
resistance is based on either the tension capacity or the compression instability of the chord
side walls, for tension and compression bracing members respectively.
Compressionloaded, full-width X connectionsfor RHS are differentiated from T or Y

connections as their side walls exhibit greater deformation than Tconnections. Accordingly,
the valueoff, in the resistance equationis reduced to0.8 sin 0, of the value which is used for
Tor Y situations.In both instances, a linear progressionis followed for resistances from values
29
Discuss me ...
Office building
for p = 0.85 (where flexure of the chord face governs) to values forp = 1.O (where chordside
wall failure governs).
All RHS T, Y and X connections with highbracing width to chord width ratios (p 2 0.85) are
also checked for the "effective width" failure modes and for punching shear of the chord face.
For this range of width ratios, thebracing member loads are largely carried by their side walls
parallel to the chord while the walls transverse to the chords carry relatively little. The upper
limit of p = 1 - 1I r for checking punching shear is determined bythe physical possibility of
such a failure, when one considers that the shear hasto be between the outer limits of the
bracing width and the inner faceof the chordwall.
3.3 KT connections
As shown in Fig. 9, KT connectionsoccur in sometrusses, and the strength ofgap

connections can be related to K and N connections byreplacing (b, + ba12 boin Table 2 with
(b, + b, + b,)/3 bo, and by replacing (b, + b, + h, + hJ14 bo in Table 3 with (b, + b, + b,
+ h, + h, + h,)/6 bo. In gap KT connections the gap should be taken as the largest gap
between twobracing members havingsignificant forces acting in the opposite sense.
In the caseof KT connections with gap, the force components, normal to the chord,of the two
members acting in the same sense are added together to represent the load. The connection
resistance component, normal to the chord, of the remaining diagonal is then required to
exceed that load. For the examples shown in Fig. 9,
N;sin 8 , N,~ sin 8, + N, sin 8, (Fig. 9a)

N;sin 8, 2 N, sin O2 + N, sin 8, (Fig. 9 b)
where N; is calculated from Tables2 or 3 in which N; sin 8, = N;sin 8,.
30
Discuss me ...
When there is across-chord load (for instance from

a purlin or hanger), which acts
in the same
direction as the load componentsthat were combined above, the connection resistanceof the
remaining diagonal needsto be examineddirectly. For the examples shownin Fig. 9,
N; sin e, 1 N, sin O2 (Fig. 9c) (3.5)
N; sin 0, 1N, sin 0, (Fig. 9d) (3.6)
again, where N; is calculated from Tables2 or 3 in which N; sin 8, = N; sin 0,.
Fig. 9 - Four KT connections

Technology centre
31
Discuss me ...
One should also remember that for load cases such asinshown Figs. 9 c and 9d, the purlin or
hanger connection tothe chord may require an additional check.
I f the vertical bracing member in the gapKT connection shownin Fig. 9 had no force in it, the
gap should be taken as the distance between the of members
toes 1 and 2, and the connection
treated as a K connection withp = (b, + b, + h, + h,)/4 boor (dl + d,)/2 bo.
For overlap KT connections, which are actually more likely to occur, the resistance of an RHS
connection can be determined by checking each overlapping bracing member and ensuring
that N,"(from Table 2) 2 N,. (Overlapped bracing members would also have a restriction on
their connection efficiency as noted at the bottom of Table 2.) For the bracing member
effectivewidthterms,careshouldbetakentoensurethatthemembersequenceof
overlapping is properly accounted for.
Tests have shown that the resistance of K and N connections in the presenceof cross-chord
loading (i.e. X connection loading) is similar to that given by the resistance formulae K andfor
N connections, andthe same is true for KT connections. If all the bracing forces on one side
of a connection act in the same sense, or if only one bracing member is carrying load, the
connection should be checked as anX connection using an equivalent bracing member size.
3.4 Graphical design charts
Fig. 10 allows a quick evaluationof whether a K or N gap connection can be configured from
proposedsquarebracingsectionsonarectangularorsquarechordsection.Thechart
indicates the maximum value of p (average width of bracing members relative to the chord
width) which can be accommodated while remaining within the allowable bounds of nodal
0 and chord aspectratio ho/bo.
eccentricity. Input parameters are chord to bracing angles
ReusinkandWardenier [20]haveproducedaset of consistentdesignchartsforthe
preliminary design of K, N, T,Y, and X connections which are based onIIW [2] and therefore
Eurocode 3 [21] recommendations. Figs. 11 to 18 are for square HSS and they relate to
formulae in Table 2.
1.0
m
,E 0.5
I
D 0.4
0.35
2.0 1.6 1.8 1.4 1.2 1.0 0.8 0.6
-
Warren
hO/ bo (chord aspect ratio)
Restrictions:
--- (bl = h1 and b2 = h 2 )
Pratt
(8,= 8 2 if Warren)
(bl = b2 if Pratt)
U
b0
Fig. 10 -Allowablerange of bracingmemberwidthratiosfor RHS gap connections,based on the
allowable eccentricity limits
32
Discuss me ...
Efficiency coefficient CT, tension
I I I I I I
10
15
20
25
30
35
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0
P
I T, Y and X connections with bracings in tension: calculation example l

connection and symbols Assume a 4 5 O X connection with these members (IS0 sizes)
chord: 200 x 200 x 8.0 (A, = 6050 mm2)
bracings: 100 x 100 x 5.0 (A, = 1890 mm2)
fyo = fyl = 355 Nlmm2
8, = 4 5 O and sin el = 0.707
n = -0.48
p = 1001200 = 0.5
f(n) = 0.92 from Fig. 19
l
blltl = 20
ranges of validity
bolto = 25
0.25 S p S 1.0
boltoS 35
bllt, S 35 (for tension bracing)
= 0.33
f, S 355 Nlmm2
300 S e, S 900 .: N; = 0.33(1890) (0.355)= 224 kN
Fig. 11 - Bracing efficiency for square hollow section T, Y and X connections with a tension loaded
bracing
33
Discuss me ...
Efficiency coefficient CT, compr,

I I 1
bdtQ
10
15
20
25
30
35
- . .
0 0.1 0:2 0..3 0.4 0.'5 0.6 0..7 0.8 0.9 1.0
3l
T and Y connections with bracings in compression: calculation example 1

I connection and symbols Assume a Tconnection with these members(IS0 sizes):
chord: 100 x 100 x 8.0(A, = 2910 mm2)
bracings: 100 x 100 x 5.0(A, = 1890 mm2)
,,f = fyl = 355 N/mmz
B1 = 90°and sin B, = 1.0
n = -0.60
p = 1001100 = 1.0
f(n) = 1.0from Fig. 19
ranges of validity bllt, = 20 < 30.4

bolto= 12.5
0.25 Ip I1.0
boltoI35
N;
... - -
A1 $1
- 0.68 (S) (1.0)(1.0)
b,/tl I1.25 5 35 = 1.09:. take as 1 .O
(for cornpresslon bracing)
f,, cr 355 N/mm2 N; = 1.0(1890)(0.355)= 671 kN
30° IB, I90°
Note: This resistance is subject to a further bracing
"effective width" check, shown in Fig. 14.
Fig. 12 - Bracing efficiency for square hollow section T and Y connections with a compression loaded
bracing
34
Discuss me ...
Efficiency coefficient Cx, compr.

1.0 I I l l
X connections with bracings incompression: calculation example 1

connection and symbols Assume a 30° X connection with these members (IS0 sizes):
b. chord: 150 x 150 x lO.O(A,= 5450 mm2)
fl bracings: 140 x 140 x 8.0 (A, = 4130 mm2)
fyo = fy, = 355 N h m 2
B, = 30° and sin B, = 0.5
n = +0.38
p = 140l150 = 0.93
p z 1 - 1 I r , so punching shear check notnecessary
f (n) = 1 .O,as chord is intension
ranges of validity b,It, = 17.5< 30.4
0.25 r p S 1.0 bolto= 15
boltoS 35
b,lt, c 1.25 5 35
(for compresslon bracing) :. take as 1.0
fy, c. 355 NImm2
30° c B, I90°
:. interpolating linearly, for p = (

0.93, -
AINiy,) = O.’’
Note: This efficiency is stillsubject to further checks for:

(l) bracing “effective width” (Fig. 14)
(11) chord shear (Table 3)
Fig. 13 - Bracing efficiency for square hollow section X connections with a compression loaded bracing
35
Discuss me ...
1.0 1.3
111
1.'2 1.4 1.5 1.6 1.7 1.8 1.9 2.0
Effective widthcheck for T,Y and X connections: calculation example
connection and symbols The T connection from Fig. 12 will be checked for bracing
b. "effective width"
1, ,chord:100 x 100 x 8.0 (A, = 2910 mm2)
bracings: 100 x 100 x 5.0 (A, = 1890 mm2)
,,f = fyl = 355 N/mm*
0, = 90° and sin B , = 1.O
p = 100/100 = 1.0
bl/tl = 20 < 30.4
bolto= 12.5
ranges of validity
to/tl = 1.6
0.25 5 p I1.0
N;
... - - - 1.00
bolt, S 35 A, $1
b,/t, S 35 (for tension bracing) This is also the bracingefficiency obtained in Fig. 12 for the
b,/t, I1.25 m I35
for compression bracing)
chord side wall failure mode
,f S 355 N/mmz
Fig. 14 - Total efficiency for effectivewidth for T. Y and X connections with p > 0.85
36
Discuss me ...
Efficiency coefficientCK,
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
10 15 20 25 30 35
bo1to
K and N gap connections: calculation example 1

connection and symbols Assume a 45O K connection with these members
(IS0 sizes) and f, = 355 Nlmm,:
chord: 200 x 200 x 10.0 (A, = 7450 mm2)
compression bracing: 150 x 150 x 8.0 (A, = 4450 mm2)
tension bracing: 140 x 140 x 8.0 (A2 = 4130 mm2)
assume n = -0.8
From Fig. 10, a gap connection is feasible providing

p I1.0
eccentricity (e) = 0.20 h, = 40
ranges of validity
= 75 t 1, + t2 = 16
0.35 Ip I1.0 p = (140 + 150) (2 X 200) = 0.725
b,/b, t 0.1 + 0.01 (2); p t 0.35

0.5(1 - p) bo = 27.5 5 g 5 82.5 = 1.5 (1 - P ) bo
b, = 150 t 0.77 (140 + 150)/2 = 112
bo/to = 20,bzlt, = 17.5
b,/t, I1 . 2 5 a s 3 5 b,/t, = 18.8 I30.4,b,/b, = 0.75 > 0.3and t 0.35
n = - 0.8, so f(n) = 0.86 from Fig. 19.
b,/t, I35
15 Ibolt, I35 ... -A,N;

-$1 - 0.35 (S)(A) (0.86) = 0.53
b, t 0.77 (b, + b,)/2
N; N;sin 0,
0.5(1-P)Ig/b,11.5(1-P),gtt,+t, and -
A, ,f
- A,$, sin 0,
= 0.57
-0.55 Ielh, I0.25
Note: If 1, and 1, had been lower, the efficiency could
f,, I355 N/rnm2
have been higher.
300 Ie, I900

Fig. 15 - Bracing member efficiency for square hollow section K and N gap connections
37
Discuss me ...
The concept of "connection efficiency", defined as the connection factored resistance divided
by the full section yield load of the particular bracing member, is employed for thesecharts.
That is, connectionefficiency = N,Y(A,fyl). However,the efficiencies given by the charts for all
but the overlapped K connections - and the T, Y, X connection effective width checks - are
termed C, C, or C, depending onthe type of connection. These latterefficiencies need to be
multiplied by three factors to obtain the final connection efficiency in each case.
The firstfactor, correcting for differing strengths between the chord and thebracing member,
is (fy,,to)/(fy, t,). This generally reduces tot,/t, because the same gradeof steel would normally
be used throughout.
The second factor, adjusting for the angle between the bracing member and the chord, is
l / s i n 0, for square RHS T, X and gap K connections. One should note that such an angle
function does not exist for squareRHS overlap connections.
The third factor, correcting for the influence of compression chordlongitudinal stresses onthe
connection efficiency, is f (n) for RHS.f (n) isdefined in Table 2, Table 3, the List of Symbols
and isplotted in Fig.19. The function f (n) is 1.Ofor square RHS overlap connections. It should
be noted thatthe chord stressratio, n, is calculated using the maximum compression stress in'
the chord member, taking into account axial force and (where applicable) bending. Although
this reduction factor was originally determined from K connection tests, it has been shownto
be applicable to T connections [22], and also where the chord is subjected to high bending
moment and no axial force [23].
In the design charts a lowerbound of the variousfailure modes was generally drawn to result
in CK,T,xfunctions which depend only on the type of connection. Simplifying, conservative
assumptions,andnarrowervalidityparameterrangesweresometimesnecessary in the
process. Still, use of the charts produces results close to those given by the formulae in
Table 2.
Four charts are presented for T, Y and X connections (Figs. 11 to 14). The first applies to all
three types of connection when they are loaded in tension; the second applies to T and Y
connections, whenloaded in compression;the third to X connections, whenloaded in
compression; andthe fourth is an effective width failure mode check, necessary only when p
exceeds 0.85. The firstthree charts are identical for p values up to 0.85. When 8 exceeds 0.85
the behaviour of the chord side walls is different for the three situations, and three charts
become necessary.') They show linear interpolations between known resistance values atp
= 0.85 and p = 1.O. For gap K and N connections the efficiency is given by Fig. 15.
The range of overlap (defined graphically in Fig. 16) for square K and N connections is from
50% to 100% rather than from 25% asin Table 2.This avoids the more complex lower range
where resistance varies constantly with amount of overlap. In the 80% to 100% overlap chart,
Fig. 16 - Definition of overlap (0,) in Figs. 17 and 18

'1 European buckling curve "a" is implicit in these charts [87].

However, the net difference is small if the
curve from another code is used.
38
Discuss me ...
Ni"
Partial efficiency ( ___ )
A, ' f, eff.
1.o
0.9
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
35 30 25 1020 15
bo/to or b,/t,
K and N overlap connections with 50% 5 0,< 80%: calculation example

~ ~~ ~~ ___~ ~~

(IS0 sizes) and ,f = 355 Nlmm2:
chord: 200 x 200 x 10.0 (A, = 7450 mm2)
compression bracing: 150 x 150 x 8.0 (A, = 4450 mm2)
(overlapped member = j)
tension bracing: 140 x 140 x 8.0 (A2 = 4130 mm2)
(overlapping member = i)
eccentricity (e) = - 0.30 h, = - 60 mm
W
bo
= - 125 :. q = 125
ranges of validity
b,lb, 2 0.25
b,lt, I1.1 JE/f,, b,lb, = 1401150 = 0.93
b21t2 I35 t,lt, = 8.018.0 = 1.0; t,lt, = 1018 = 1.25
bolt, I40 bolto = 20,b21t2 = 17.5, b,/t, = 18.8 C. 26.8
t,lt, I1 .o b,lbo = 0.75, b,lb, = 0.70
b,lb, S 0.75
. -NI'
- - 0.38 + 0.36 = 0.74 :. N,*= 1085 kN
50% 5 0,< 80% ' ' A, ,f
-0.55 5 elh, I0.25
N'
hence I0.74
(see Table 2) :. N,'= 1169 kN
,f 5 355 Nlmm2 4 f,
300 S e, C. 900
Fig. 17 - Bracing member efficiency for square hollow section K and N overlap connections having
50% 5 0,< 80%
39
Discuss me ...
Totalefficiency ( -
Ni"
A , . f, ' eff.
10 15 35 2030 25
b,/t,
r K and N overlap connections with 80% I0,I100%: calculation example 1

(IS0 sizes) and f,, = 355 Nlmm2:
chord: 200 x 200 x 10.0 (A,, = 7450 mm2)
compression bracing: 150 x 150 x 8.0 (A, = 4450 mm')
(overlapped member = j)
tension bracing: 140 x 140 x 8.0 (A2 = 4130 mm')
(overlapping member = i)
eccentricity (e) = - 0.40 h, = - 80 mm
ranges of validity = - 165 :. q = 165

q x 100
165 x 100
b,lb, 2 0.25 0,=
- 1401sin
b,lt, I1.1 m P
b,lb, = 1401150 = 0.93
450 = 830'0
b,lt2 c 35 t,lt, = 8.018.0 = 1.0

bolto c 40 bolt, = 20,b21t2 = 17.5, b,lt, = 18.8 c 26.8
t,lt, I1.0 b,lbo = 0.75, b21bo= 0.70
b,lb, 2 0.75
. -N If
= 0.87 .: N,'=1276 kN
80% I0, c 100% ' ' Aif,,
-0.55 Ielh, I0.25
N'
hence -L c 0.87 (see Table 2) :. N;= 1374 kN
,f I355 Nlmm2 AI f,
300 e,
I I900
Fig. 18 - Bracing member efficiency for square hollow section K and N overlap connections having
80% I0,I1OOO/o
40
Discuss me ...
Fig. 18, the subscript j applies to the overlapped member while subscript i refers to the
overlappingmember. Unlikecircular hollowsectionoverlappedconnections,onlythe
overlapping member need be checked for square RHS connections; however, there is a check
on the connection efficiency for the overlapped bracing members, as notedin Table 2.Fig. 17
for partially overlapped connectionsrequires entering twice, once for the overlapthe on chord
(when 0 and i terms are used) anda second time for the overlap on the other bracing member
(when j and i terms are used). The two part-efficienciesare then added together for the total
efficiency. Validity ranges are the same as for2 Table except that overlap starts
at 50%. It can
be seen from the chartsthat efficienciesof K and N overlap connections always exceed 0.8 for
full overlap(Fig. 18) and 0.6 for partial overlap(Fig.17).Therefore,fullyoverlapped
connections are usually stronger than partially overlapped ones.
Functlon f (nl
-1.0 -0.8 -0.6 -0.4 -02 0

n
Fig. 19 - Function f (n) which describes the influence of chord stress on the total efficiency of square
hollow sectionT, Y, X, K and N gap connections.
3.5 Reinforcedconnections
Instances may occur when a truss connection has an inadequate resistance, and a designer
needs to resort to some form of connection reinforcement. aSuch if RHS
situation might arise
material were ordered on the basis of member selection only, without connection capacity
checks being performed. Alternatively, only one or a few connections of a truss may be
inadequate due to the selection of a particular chord member, and so just these critical
connections could be reinforced. The labour costs associated with connection reinforcement
are significant, and theresulting structuremay lose its aesthetic appeal, but
in many casesit
may be an acceptable solution.
3.5.1 Withstiffening plates
The most common method of strengtheningRHS connectionsis to weld a stiffeningplate (or

plates) to the RHS chord member. It is particularly applicable to gap K connections with
rectangularchordmembers,althoughan unstiffened overlapconnection is generally
preferable from the viewpoints of economy and fatigue. However, a gap connection with a
stiffening plate eliminates the necessity for double cuts on the bracing members, and in
41
Discuss me ...
certain cases may prove more acceptable to the fabricator. The addition
of a flat plate welded
totheconnectingface of the chordmembergreatlyreduceslocaldeformations of the
connection and consequently the overall truss deformations are reduced. It also permits a
more uniform stress distributionin thebracing members.
3.5.1 . l K and N connections
The type of reinforcement required depends upon the governing failure mode which causes
the inadequate connection capacity. Two types of plate reinforcement - in one case to the
-
chord connecting face andin theother to the chord side wallsare shownin Fig. 20. Bothof
these would be applicable to connections with RHS chord members and either CHS or RHS
bracing members. An alternative to stiffening a connection with plates is to insert a lengthof
chord material of the required thickness at the connection, the length of which would be the
same as L, given below. (This is equivalent to the use of a “joint can” in offshore steel
structures.)
The capacityof gap K connections is typically controlled by either the chord face plastification
criterion or the chord shear criterion, as summarized in Tables 2 and 3. When chord face
plastificationcontrols, the connectioncapacitycanbeincreasedbyusingflangeplate
reinforcement as shownin Fig. 20(a). This will usually occur whenp 5 1.O and the members
aresquare.Whenchordshearcontrols,theconnectioncapacitycanbeincreased by
reinforcing with a pair of side plates as shown in Fig. 20(b). This failure mode will usually
govern when p = 1 .Oand h, bo.
tP tP
( a ) Flange plate
relnforcement (b) Slde plate
relnforcement
Fig. 20 - Pratt truss connection with plate stiffening

(a) Flange plate reinforcement
(b) Side plate reinforcement
The first design guidance available forK connections stiffened with a flangeplate, as shown
in Fig. 20(a), was given by Shinouda [24]. However, this method was based on an elastic
deformation requirement of the connection plate under specified (service) loads. A more
logical limit states approach which is recommended for calculating the necessary stiffening
plate thickness for gap K connections is to use the connection resistance expressionsin Table
2 (for square or circular bracing members to square chord members), and Table 3 (for
rectangular members) by considering t, as the chord face thickness and neglecting 1, [19].
Also, the plate yield stress should be used.It is suggested that proportioning
of the stiffening
plate be based on the principle of developing the capacity of the bracing members (A,fy,).
Dutta and Wurker [25] consider that this 2 2t2. Careful attention
will be achieved providing tp
should be paid tothe stiffening plate-to-chord welds which should have a weld throat size at
least equal to the wall thickness of the adjacent bracing member [25]. The stiffening plate
should have a minimum lengthL, (see Fig. 20(a)), suchthat:
sin 0, (3.7)
42
Discuss me ...
A minimumgap betweenthe bracingmembers, justsufficient to permit welding of the bracing

members independentlyto the plate issuggested [26]. All-round weldingis generally required
to connect thestiffening plateto the chord member, and in order to prevent corrosion onthe
two inner surfaces.
In order to avoidpartial overlapping of one bracing member onto anotherin a K connection,
fabricators may elect to weld each bracing member to a vertical stiffener as shown in Fig.
21 (a). Another variation onthis concept is to use the reinforcement shownin Fig. 21 (b). For
both of these connections,t, 2 2t1and 2t, is recommended [25]. Designers should note that
the K connection shownin Fig. 21 (c)is not acceptable, as it does not develop the strength of
an overlappedK connection where one bracing member is welded only to thechord face.Also,
it is difficult to create and ensure effective
an saddle weld between the two bracing members.
If the capacity of a gap K connection is inadequate and the chord shear criterion is the
governing failure mode, then asmentionedbeforeoneshould stiffen withside plate
reinforcement,as shownin Fig. 20 (b).A recommended procedurein this case forcalculating
the necessary stiffening platethickness is to use the chord shear resistance expression in
Table 3, by calculatingA, as 2 h, (to + t,). The stiffening platesshould again have minimum
a
length, L, (see Fig. 20(b)), given by Eqn. 3.7 and have the same depth as the chord member.
-p- I----
- -~ ___
-___-
(C)
N o t acceptable
Fig. 21 - Some acceptable and unacceptable, non-standard truss K-connections
3.5.1.2 T, Y and X connections
Under tension or compression bracing loading, the capacity of a T, Y or X connection is

typically controlled by either chord face yielding
or chord side wallfailure, as summarized in
Tables 2 and 3. When chord face yielding controls, the connection capacity can be increased
by using flange platereinforcement similar to the connection shown in Fig. 20(a). This will
usually occur whenp S 0.85. When chord side wallfailure controls, the connection capacity
can be increased byreinforcing witha pair of sideplates similar to the connection shownin
Fig. 20(b). Thisfailure mode will usually govern whenp = 1 .O.
For T, Y or X connectionsstiffened withside plate reinforcement, arecommended
procedure forcalculating thenecessary stiffening plate thickness is to use the chord side wall
resistance expressionin Table 3, byreplacingtowith (to+ t,) for the side walls. Thestiffening
plates should have alength L, (see Fig. 20(b)), such that for T and Y connections:
L, 2 1.5 (hl/sin e,) (3.8)

For T,Y and X connections stiffenedwith a flangeplate there is a difference in behaviour of
of the loadin the bracingmember. With a tension
the stiffening platedepending on the sense
43
Discuss me ...
load in the bracing, the plate tends tolift off the chord member and behaveas a plate clamped
(welded) along its four edges. The strength of the connection thereby depends only on the
plate geometry and properties, and not on the chord connecting face. Thus, for tension
bracing loading, if one applies yield line theory to the plate-reinforcedT, Y or X connection
with rectangular members, the connection factored resistance can be reasonably estimated
by:
where
t, = thickness of the stiffening plate,
,f = yield strengthof the stiffening plate,
pp = width ratio of bracing member to plate = b,/Bp,
7, = bracing member depth to plate width ratio = h,/B,, and
B, = width of plate.
The similarity betweenEqn.3.9andthefactoredresistance of unstiffened T, Y and X
connections, based on chordface yielding as given by Table is 3,clearly evident. In order to
develop the yieldline pattern in the stiffeningplate implicit in Eqn. 3.9, the length of the plate
L, should be at least:
L, 2 (h,/sin 6,) +d m (3.10)
Also, the plate width (B,) should be such that a good transfer of loading to the side walls is
achieved; for example B,=bo (see Fig. 20).
For T, Y and X connections stiffened with a flange plate, and under compression bracing
loading, the plate and connecting chord face can be expected to act integrally with each
other. This typeof connection has been studied by Korol et al[27], also using yieldline theory.
Hence for P, I0.85 (a reasonable upper limit for application of yield line analysis also
employed for unreinforced connections), the following plate design recommendations are
made to obtain afull strength connection:
I. B, = flat width of chord face
II. L, L 2b0 (Increaseproportionately if 6, k 90° toallowforgreater bracing member
“footprint”).
Ill. t, 2 4t, - to.
The application of the above guidelines, for compression-loaded T, Y and X connections,
should ensure thatthe connection capacity exceedsthe bracingmember capacity, provided
chord side wall failure by webcrippling is avoided 1271.
3.5.2 With concrete filling
A less visible alternative

to adding stiffening plates to theexterior of an RHSis to fill the hollow
section chord with concrete or grout. Filling the chordmembers of an RHS truss, either along
the full length of the chord or just in the vicinity of critical connections,hastwomain
disadvantages: the concretewill increase the dead weight ofthe structure, and it involvesa
secondary tradewith its associatedcosts.On the otherhand, the strength of certain
connections may increase, andif the members are completely filled, there are further benefits
ofenhancedmembercapacity and improved fire endurance.Concrete-fillingofchord
members can bedone in the fabrication shop bytilting the trussand using a concrete or grout
with a highwater-to-cement ratio. If flange plates existin the chord member, thefilling can be
restricted to just the distancefrom the truss endto the flangeplate.

The connectionswhich benefit most fromconcrete-fillingare X connections with thebracing
members loaded incompression; that is, connections at which a compression forceis being
transferred through theRHS. Examplesof such connections are truss reactionpoints, truss
44
Discuss me ...
Fig. 22 - Applications for which concrete-filling ofRHS may improve connection resistance
connections at which there is a significant external concentrated load, and beam-to-RHS

column moment connections,as illustrated in Fig. 22. An experimental study
of concrete-filled
RHS members subjected to transverse compression loads has been made by Packer and
Fear [28], whoconcluded that:
I. Concrete-filling of RHS greatly enhances their
performance
under
transverse
compression. The RHS provides some confinement for the concrete which allows it to
reach bearing capacities greater than its crushing strength,as determined by cylinder
compression tests.
II. For limit states design, the factored resistance of a concrete-filled RHS, compression-
loaded, X connection could be taken as:
N; = (6cf,'A,/sin e,) dA2/A1) (3.11)
where
6, = resistance factor for concretein bearing (0.6 in Canada, for example)
f,' = crushing strength of concrete by cylinder tests,
A, = bearing area over which the transverse loadis applied, and
A, = dispersed bearing area.
A, should be determined by dispersion
of the bearing loadat a slope of2 : 1 longitudinally
along the chord member, as shown in Fig. 23. The valueof A, is limited by the length of
concrete, andd m )cannot be taken to be greater than 3.3.
111. The followingare also recommended for general
design applicationof Eqn. 3.11:
h,/b, c 1.4
L, 2 (h,/sin e,) + 2 h,
where L, = length of concrete in RHS chord member.
IV. It has also been shown that shrinkage
of the concrete (or grout) away from the RHS inside
walls does not have a negative impact the
on strength of a concrete-filled connection.
Fig. 23 - Recommendedmethodfordeterminingbearingcapacity of concrete-filled RHS loaded in

transverse compression
45
Discuss me ...
3.5.3 DesignExample
The 45OX connection given inFig. 24 is subjected to the factored loads shown. The resistance
of the connectionwill be checked to see if it is adequate. The members are hot-formed hollow
sections with dimensions conforming to ISOlDlS 657-14 [29]. The steel grade is Fe510,
conforming withIS0 630 [30], with a minimum specified yield strength of 355 N/mm2.
Refer to Design Chart, Fig. 13 withp = 1.O, for a “Chord Side Wall Failure” check.
N; = cx (tolt,) (f(n)) (A,) (fy1)
With p = 150/150 = 1.0and bolto = 150/10 = 15, CX = 0.49
n = ( - 1200)/5450(0.355) - 0.620
... f(n) = 1.3 + (0.4/1.00)(-0.620), but 5 1.0
= 1.0
:. N; = 0.49(1 .O) (1.O) (5450)(0.355)
= 948 kN 5 1200 kN :. Inadequate.
Also refer todesign chart, Fig. 14 for an“Effective Width”check on thebracing member.
t,) = 1.O and bolto = 15, N; = 0.815 A, fyl.

With (fyOtO/fyl
:. N; = 0.815 (5450) (0.355) = 1580 kN > 1200 kN :. O.K.
Hence,the connection is still inadequate due tochordsidewallcapacityandmust be
reinforced, either by using platereinforcement or concrete-filling.The chord should alsobe
checked for shear (through the chord near the bracing member toes),as noted at the bottom
of Table 3, but reinforcement is required anyway.
(a)Side plate reinforcement

A pair of side plates will be added to the chord side walls, withihe side plates also havinga
yield strengthof 355 N/mm*.
From Table 3, for“Chord Side Wall Failure” with p = 1.O,
N; = 0.8,,f (to + t,) [2 h,/sin 0, + lot,]
Assume that a plate thickness of 10 mm (same as the chord), is chosen. Assuming that the
chord sidewall andplate actindependently,they will haveapproximatelythesame
compression resistance. Hence, it can be seen that the connection resistance will double
when reinforced inthis manner.
Hence N; = 1900 kN > 1200 kN :. O.K.
For the length of the side plates, L,, the intent of Eqn. 3.8 forT and Y connections is that the
plates extend 50°/o further than the bracing member “footprint”. Applying this same guidance
to the X connection of Fig. 24, with two offset bracing member “footprints”,
L, 2 1.5 (150/tan45O + 150/sin 45O) = 543 mm
:. Make thestiffener plates 600 mmlong x 150 mm high x 10 mmthick, and weldall around
the plate perimeter.
NI = 1200 kN
150x 150x 10 RHS
0kN kN
No =
1200
___
L 4 5 0 ;;g&,
Total foot prmt /’
___
;450
150x 150x l0 RHS

No =
A0 = 5450 mm2
N1 = l 2 0 0k N‘ 1 5 0 x 1 5 0 ~ 1 0 RHS
A , = 5450 mm2
Fig. 24 - RHS X connection design example
46
Discuss me ...
(b) Concrete filling

Fill the chord member of the connection shown in Fig. 24 with concrete having a crushing
strength,
;f = 40 N/mm2.
N; = (4,f;A,/sin e,) (3.1 1)
= 0.6 (0.040) (31,820) (1.767)lsin 450
= 1908 kN > 1200 kN :. O.K.
In this calculation d w ) was limited to 1.767 because A2 was interpreted to be (total
footprint length shown on Fig. 24 + 300 mm) (150 mm) = 99 300 mm2. Hence A2/A,= 3.12.
For this chord member h,/b, = 1.O which is within the limit of experimental verification. An
appropriate minimum length of concrete would be the “total footprint”(see Fig. 24) plus 2 h,
= 0.7 metres.
Covering structureof a building with RHS
47
Discuss me ...
Warehouse rack structure
48
Discuss me ...
3.6 Cranked-chordconnections
“Cranked-chord” connections arisein certain Pratt or Warren trusses such as the one shown
in Fig. 25 and are characterized by a crank or bendin the chord member at the connection
noding point. The crank is achieved by butt (groove) welding two common sections togethe
the appropriate angle, and the intersectionof the three member centre-lines is usually made
coincident. The uniquenessof this cranked-chord connectionlies both in its lack of a straight
chord member and the role of the chord member as an “equal width bracing member”. An
experimental research program with square and rectangular members has revealed that
unstiffened,welded,cranked-chord RHS connectionsbehavegenerally in amanner
dissimilar toRHS T or Y connections, despite their similar appearance (they all have a single
bracingmemberweldedtoauniform-sizechordmember).Instead,cranked-chord RHS
connections have been shown to behave as overlapped K or N connections, and their capacity
can be predicted using the current recommendations given in Table 2, subject to the
limits of
application in Table 2a [31]. Cranked-chord connections can be interpreted as overlapped K
connections as shown in Fig. 26, wherein one chord member can be given an imaginary
extensionandthecranked-chordmemberisconsideredtobetheoverlappedbracing
member.
lmagmary extension
of chord mb member
v
Same s u e member
Fig. 26 - Cranked-chordconnection
represented as an overlapped
Fig. 25 - Cranked-chord connection in a Pratt truss K or N-connection
3.6.1 Designexample
The 4 5 O cranked-chord connection givenin Fig. 27is subjected to the factored loads shown.
The resistanceof the connectionwill be checked to seeif it is adequate. TheRHS members
have dimensions conforming toISOlDlS 657-14 1291. The steel grade isFe510, conforming
with IS0630 1301, with a minimum specified yield strength of 355 N/mm2.
NI = 1200 kN
1 N2 = 1700 kN
1 5 0 x 1 5 0 x 1 0 R H S ~ ~ ~
(Braclng I )
A. = 5450 mm2 180x 18Ox 10 RHS
(Brdclng I )
A , = 6650 mm2
22.50
NO
1200
‘Weld
18Ox 18Ox 10 RHS
A. = 6650 m m 2
Fig. 27 - RHS cranked-chord connection design example

Imagine the horizontal chord member extending as shownin Fig. 26 and both of the other
members joining on to the topof the extended chord member.
49
Discuss me ...
Overlap(seeFig. 16) = q/p x 100% = 112.3 x 100/150 = 75%

Check range of validity for an overlappedK connection in Table 2 a:
b,/bo = 0.83and 1.00 z 0.25 ... O.K. b,/t, = 15 I1.1 V’(E/f,,) :. O.K.
b,/t, = 18 I35 :. O.K. bolto = 18 I40 :. O.K.
t,/t, = 1.00 I1.oo :. O.K. b,/b, = 0.83 2 0.75 :_ O.K.
0, = 75% 2 25%
I100% :. O.K. e = 0 :. O.K.
... From Table 2, for 50% 5 0, < 8O%,
N; = fy, t, 12 h, - 4 tl + be + be(ov)l
where be = (10118) (1.OO)150 = 83.3 mm = be(ov)
also.
:. N,* = (0.355)(10) [300 - 40 + 2 (83.3)]
= 1510 kN 2 1200 kN .: O.K.
Alternatively, one could use Fig. 17 to calculate NI:

For bo/to= 18and fyoto/fy,t, = 1.0, partial efficiency = 0.37
Forb,/t, = 18and fy,t,/fy,t, = 1.O, partial efficiency = 0.37

Total efficiency = 0.37 + 0.37 = 0.74.
or N,’= 0.74 (0.355) (5450) = 1430 kN z 1200
(somewhat more conservative than direct use of equations).
Check (see Noteto Table 2) on bracing efficiencies:
The efficiency of the overlapped bracing j should not be greater than that of the overlapping
bracing, hence:
= 1840 kN 2 1700 kN
:. still O.K.
50
Discuss me ...
4 Truss design examples
4.1 Uniplanartruss
Truss layout andmember loads

An example has been selected illustrate
to the useof the connection design methods given in
Chapter 3, as well as the truss design principles described in Chapter 2. A Warren truss
consistingof squareRHS is presented since that configurationis often the preferredsolution.
A Warren configuration with low bracing member angles, such as used here, keeps the
number of connections to a minimum. All members chosen are hot-formed hollow sections
with dimensions conforming to ISOlDlS 657-14 [29]. The steel grade throughout is Fe510
conforming with IS0 630 [30], with a minimum specified yieldstrength of 355 N/mm2.
108 k N
5
140
1 k8
0Nk8 N
kN I 1 0185
048
kNkN
Fig. 28 - Example Warren truss showing applied loads and resulting member forces (in kN)
Fig. 28 shows the trussand factored loads alongwith member axial forces, determined by a
pinjointed analysis. The trusses are spaced at 12 metre intervals andthe top (compression)
chord is considered to be laterally supported at each
purlin position. The span-to-depthratio
is 15, whichis around the optimal upperlimit consideringservice load deflections and overall
costs (see Section 2.1).
Design of members
For member selection one could use either member resistance tables for the compression
members,withtheappropriateeffectivelength,or the applicable strut buckling curve/
equation. In practice one would also pay attentionto the availability of member sizes selected.
For this truss design example, compression member resistance has been determined in
accordancewithEurocode 3 using buckling curve “a”. Theresistancehasalsobeen
calculated assuming yu = 1.O.(i.e. no partial safety factor or resistance factor, since this
factor will be different for various countries (1.O and higher)). Since the connections at the
truss ends are generally critical, the chord walls selected should not be too thin, as a single
size member will be used for the top chord and another single size member selected for the
bottom chord.
Top chord
Use a continuous section with an
effectivelength, forboth in-plane and out-of-plane
buckling,
of 0.9 L = 0.9(6000) = 5400 mm, as noted in Section 2.3.1, Eqns. 2.1 and 2.2.
Maximum force = 1148 kN (compression).

Possible section sizes are shownin Table 4 below, along with their compressive resistances.
As noted in Section2.6,use bolto ratioswhicharebetween 15 and 25, so select the
180 x 180 x 8.0 RHS at this stage.
51
Discuss me ...
Table 4 - Possible section sizesfor top (compression) chord

-
II I
fY0
Nlmm2 I
No
kN
II KL
m I
possible sections
mmxmmxmm
A, bolto x
I l I 355 1148
5.4 200
l x 200 x 8.0
180 x 180 x 8.0
200 x 200 x 6.3
1230
1230
Bottom chord
Table 5 - Possible section sizes for bottom (tension) chord
NO possible sections A0 bolt0 fyo A0
fY0
Nlmm2 kN mm x mm x mm mm2 kN
355 1215 150 x 150 x 6.3 3560 1260 23.8

160 x 160 x 5.6 28.6 34801240
180 x 180 x 5.0 3460 1230 36.0
It is best, for connection capacity, to keep the tension chord as compact and stocky as
possible; so select the 150 x 150 x 6.3 RHS at this stage.
Diagonals
By aiming for gap connections, reference to the chart in Fig. 15 shows that the highest
connection efficiencywill be achieved whenthe ratio (f,,t,/f,,t,) is maximized. So try to select
bracing members such that (fyOtO/fy,t,)2 2.0, which in this case implies t, 5 3.15 mm, or near
this thickness if possible. For the compression bracing members use an effective length of
0.75 L (Eqn. 2.3, Section 2.3.1) = 0.75 d(2.4' + 3.02) = 2.88 m.
Compression diagonals
Table 6 - Possible sectionsizes for compression diagonals
Nlmm2
2.881
possible sections
mm x mm x mm
120 x 120 x 4.0

A,
mm2
1850
bllt' Ix I x $1 A1
kN
51 6
100 x 100 x 5.0 1890 448
2.881 100 x 100 x 3.6 1380 27.8 0.98 332
90 x 90 x 3.6 1240 25.0 1.10 263

2.881 70 x 70 x 3.2 850 21.9 1.42 123
80 x 80 x 3.2 978 25.0 1.23 178
52
Discuss me ...
Tension diagonals
Table 7 - Possible section sizesfor tension diagonals
fY2 N2 possible sections A2 bJt2 $2 A2
N/mm2 kN mm x mm x mm mm2 kN
355 432 x 90 x 3.6 1240 90 25.0 440
70 x 70 x 3.2 259 355 21.9 850 302
355 86 30 x 30 x 2.5 272 12.0 97
Member selection
The number of sectional dimensions depends on the total tonnage to be ordered. In this
example, only two different sections
will be selected for the bracing members. A comparison
o! the members suitable for compression diagonals and tension diagonals shows that the
following are most convenient:
Bracings: 120 x 120 x 4.0 RHS
80 x 80 x 3.2 RHS Note: 80 x 80 selected rather than 70 x 7 0 to
conform to b, 2 0.77 (b, + b,)/2 limit.
Top chord: 180 x 180 x 8.0 RHS
Bottom chord: 150 x 150 x 6.3 RHS
For square chords and O1 = 0, = 38.7', Fig. 10 shows that gap connections are possible
providing p I1.O, so the members selected allow gap connections.
The location of the sections selected, along with connection numbers, is shown
in Fig. 29.
1 3 I I 1,4
L
6
Bolted site connectlon
120~120~4.0 150~150~6.3
Fig. 29 - Member dimensions and connection numbers
Commentary
At connections 1 and 4 the top chord memberis welded to a flange plate for connecting to a
column and a neighbouring chord member, respectively. At connection 1, a gap of 4 to is
chosen between the toe of the tension bracing member and the plate. This connection is
checked as a K (or specifically N) connection, rather thanY , because the flange plate provides
similar restraint to the chord face as a neighbouring compression bracing of the
member
same
size asthe tension bracing. Connection4 is also checked asa K connection sincethe plates
(see Fig.29) again stiffen the connection, despite to an X connection.
the loading being similar
Hence all connections are checked as K (or N) connections and the chart in Fig. 15 can be
used.
Table 8 gives the connection resistance calculations and it can be seen thatall connections
are adequate. This was possible due to an astute selection of member sizes,in which theratio
(fyoto/fy,t,) was kept as high as possible. Furthermore, realizing that a large bracing member
53
ConstructionwithHollowSteelSections-Designguideforrectangularhollowsection(RHS)jointsunderpredominantlystaticloading
Discussme...
Thismaterialiscopyright-allrightsreserved.UseofthisdocumentissubjecttothetermsandconditionsoftheSteelbizLicenceAgreement
Table 8 - Check for connection resistances

connection
parameters
connection efficiency remarks
(check
validity
range, Fig. 15) efficiency (see Fig. 15)
connection chord bracings b, b,

- e
N,*z N,
mm x mm x mm mm x mm x mm P T t, mm
1 180 x 180 x 8.0 plate 0.67 - 0 0.18 - - - -

bolto= 22.5 x 120 x 4.0 30 120 0.66
0.34
1.60 > 1.0 Yes
2 x 180 x 8.0 x 120 x 4.0 180 30 0.56 120 0 1.54

0.29
0.66
0.36 0 89
bolto= 22.5 80 x 80 x 3.2 25 1.54
0.44
0.75 > 1.0
3 180 x 180 x 8.0 120 x 120 x 4.0 0.56 30 1.390.29
0 0.390.36 0.81
bolto = 22.5 80 x 80 x 3.2 25 1.39
0.44
0.25 > 1.0
4 x 180 x 8.0 8 O x 80 x 3.2 0.44

180 25 1.20 0.25
0 0.54 > 1.0
>l 0
I I I
bolto= 22.5 80 x 80 x 3.2 25 0.25
l 5 6
150 x 150 x 6.3 120
bolto= 23.8 120
150 x 150 x 6.3

x 120 x 4.0
x 120 x 4.0
120 x 120 x 4.0

0.80
0.86 ~ 301.60 3o 0.66
30 0.67
18
18 t
0.27 ~ 1 20
bolto= 23.8 80 x 80 x 3.2 25 0.42 0.75
7 x 150 x 6.3 80 x 80 x 3.2 150 25 0.53 0.33

18 0.250.69
bolto= 23.8 80 x 8Ox 3.2 25 0.33 0.25
Createdon01March2012
Discuss me ...
would be adjacentto a much smallerbracing member at connections 2,3 and 6, the 80 x 80

RHS was selected instead of the 70 x 70 RHS to satisfy bi h 0.77 (b, + b,)/2. Along the
compression chord, all connections have zero noding eccentricity, which is usually the first
choice ofdesigners providing a sufficient gap results. On thetensionchord,a noding
eccentricity has beenintroducedat all the connections; butthis does not influence the
design
of the tension chord or the connections (see Table1, Section 2.2).
Purlin connections
Depending on the type of purlins, various purlin connections are possible. If light gauge
purlins are used, such ascold-formed channelshapes for example, a popular form of purlin
cleat is asection of angle weldedto thetop faceof the chordmember, extending across the full
width of the RHS. The purlin would then bebolted to the outstanding legof the angle.
In this example, long span purlins are used; so they are likely to be l-sections,in which case
angle cleats could be welded to each side of the RHS chord member and the purlin bolted
through its flange tothe outstanding leg of the angle as shownin Fig. 30.
__- __--
+.-%
J
g
T
7
rq
q --l-
Fig. 30 - Possible purlin cleat connection at truss connection 2no.

Arched roof trusses for a multi-use hall
55
Discuss me ...
4.2 Arch-formed truss
The connections of arched trusses can be designedin asimilar way to thoseof straight chord
trusses. If the arched chords are made by bending at the connection locations only,as shown
in Fig. 31 (a), the chord members can also be treated
in a similar way to thoseof straight chord
trusses providedthat the bending radiusremains within the limits to avoid distortion of the
cross section [25,321. If the arched chords are made by continuous bending, the chord
members have a curvedshape betweenthe connection locations, as shown in Fig. 31 (b). In
this case the curvature should be taken into account in the member design by treating the
chord as a beam-column. (Moment = axial force x eccentricity.)
Fig. 31 - Arched truss

56
Discuss me ...
5 Boltedconnections
5.1 Flange-plateconnections
RHS flange-plateconnections have generally beenbolted along all four edges of the plates;
however the optionof bolting along only two edges has been
investigated during the1980s
and shown to be effective.
5.1 .l Bolted on two sides of the RHS
Preliminary tests on flange-plate connections bolted along two sides of the RHS as in Fig. 32
were performed by Mang [33] and Kat0 and Mukai [34], followed by a more extensive study by
Packer et al. [35].Thelattershowedthatonecould,by selecting specificconnection
parameters, fully develop thetensile resistance of the member bybolting along only two sides
of the tube. This form
of connection lends itselfto analysis as a 2-dimensionalprying problem,
but the application of traditional prying models developed for T-stubs was found to not
correlatewell with thetest results. One theof mainreasons was attributedto the locationof the
hogging plastic hingelines, which tended to form within the width of the tube as shown on
Fig. 32.
Small spacers
improve Derformance
Bolthole
dlameter =d
Typlcal location
of hogging
plastic hinges
Section A - A
Fig. 32 - Flange-plate connectionwith bolts along two sides of RHS
A modified T-stub design procedure was consequently proposed [36] and verified against a
set of possible failure mechanisms which were based on the observed failure modes. The
design procedure involved redefinition
a of various parametersin theT-stub design method of
Struik andde Back [37] which is now adopted by manystructural steelwork codes. To reflect
theobservedlocationof the inner(hogging) plastic hinge line,andalsorepresentthe
connection behaviour illustrated by the more complexanalytical models, the distance b was
adjusted to b' as shown onFig. 32, where:
b' = b - (d/2) + t, (5.1)

The term(Y has been usedin Struik and de Back's T-stub prying model to represent the ratio
of the bending moment per unit platewidth at the bolt line, to the bending moment per unit
plate width at the inner (hogging) plastic hinge. Thus, for the limiting case of a rigid plate
57
Discuss me ...
(Y = 0, and for the limiting case of a flexible plate in double curvature with plastic hinges
occurring bothat the bolt line and the edge of the T-stub web 01 = 1.O. Hence, the term01 in
Struik and de Back's model was restricted to the range0 5 01 5 1.O.For bolted RHS flange-
plate connections, this range of validity for a! was changed to simply 01 2 0. This implies that
the sagging moment perunit width at the boltline is allowed to exceed the hogging moment
per unit width, which is proposed because the RHS member tends to yield adjacent to the
hogging plastic hinge and participatethe in general failure mechanism.
Thus, a suitable design method forthis connection would be to initially estimate the number,
grade and sizeof bolts required, knowing the applied tensile force N, and allowing for some
amount of prying. Ingeneral, theapplied external load per bolt should be only 60% to 80% of
the bolt tensile resistance in anticipation of bolt load amplification due to prying. Hence,
determine a suitable connection arrangement. The bolt pitch pshould generally be about 3 to
5 bolt diameters (although closer pitches are physically possible if wanted), and the edge
distance aabout 1.25b. Prying decreases asa is increased up to 1.25 b, beyond which there
is no advantage. Hence, maximumeffective afor use in Table 9, is 1.25 b. Then proceed as
shown in Table 9.
- Design
Table 9 RHS flange-plate connections with bolts along two sides

steps for
commentary design equations
1. From a layout of the connection,

calculate 6, the ratioof the net area
at bolt lineto grossarea atthe tube
face
2. Determine a trial flange-plate (KP,/(l + 6))O 5 ,t 5 (KP,)O (5.3)

thickness 1,
where P, = N,/n = external factored tensile loadlbolt
and K = 4 b'l(0.9 f,p) (5.4)
3. Calculate the ratio 01 necessary for U = [(K T,/t2,) - l ] [(a + d/2)/6 (a + b + t,)] (5.5)
equilibrium, assuming bolts are
where a S 1.25 b; T, = bolt tensile resistance
loaded to their tensile resistance
4. Calculate the connection factored N,'= tg (1 + 601) nlK

resistance, N,: by using 01 from Eqn.
where n = number of bolts
5.5, except set CY = 0 if 01 0.
N,' must be 2 N,
4a. If required, the total bolt tension, T, = P, [ l + (b'la') (6011(1+ 6a))l (5.7)
including prying(T,) can also be
where a' = a (but 5 1.25 b) + d/2
calculated. (Y from Eqn. 5.8 is not
necessarily the same as that from and CY = [(KP,/ti) - l ] (1/6) (5.8)
Eqn 5.5 which assumed bolts were
loaded to theirfull tensile resistance.
Eqn. 5.8 becomes Eqn. 5.3 for
CY = O o r l . 0
This design method should be restrictedto the range of flange-plate thicknesses over which it
has been validated experimentally and analytically[35, 361, namely 12 to 26 mm. It should be
borne in mind that when a connection with bolts in tension is subject to repeated loads, the
flange must be made thick enough and stiff enough so that deformation of the flange is
virtually eliminated ((Y5 0). Most structural steelwork specifications require that bolts with
tensile loads be pretensioned,a requirement that is essential for fatigue situations. Spacers
placed between the plates in line with the RHS sidesparallel to the bolt lines, as shownin Fig.
32, can avoid prying actionand improve fatigue performance considerably [38].
58
Discuss me ...
5.1.2 Bolted on four sides of the RHS
CIDECT research programs on flange-plate connections bolted along all four sides, asin Fig.
33, have been undertaken by Mang[33], and Kat0 and Mukai[39], but a reliableconnection
design procedure has not yet evolved. Kat0 and Mukai proposed a complex model based on
yield line theory with an estimateof the prying force. Depending on the relative strengths of the
flange plate to the bolts, the ultimate strength of the connectionwas determined by one of six
failure modes. Failure modes 1 to 3 involved failure of the flange plates while modes4 to 6
involved bolt failure. Two recent connection testsof this type (401, however, have indicated
that the model overestimated the strength of the connections by about 25%, so further
investigation of this connection is still warranted.
Fig. 33 - Four and eight bolt configurations for bolting on all sides of anRHS
Kat0 and Mukai summarized, for connections on 150 and200 mm square RHS, that the flange
plate thickness should not be less than the bolt diameter for connections with four bolts, nor
less than the bolt diameter
plus 3 mm for connections having eight bolts. They 16,20
used and
24 mm diameter bolts in their tests and decided thatthe “yield load”of the connection was
approximately 0.8 times the sum of the original tensions in the bolts. Kat0 and Mukai also
recommended that the external load per bolt be limited to no nore than 75% of the bolt’s
factored resistanceT,.
Kat0 andMukai’s method for proportioning flange-plate thickness does not consider the plate
yield strength. This fact and the evidence that some connection strengths are overestimated
suggest that a more conservative approach be taken pending additional experimental work
and comprehensive recommendations.
5.1.3 Design example for bolted flange-plate connection
For the uniplanar truss design example

in Section 4.1, the bottom chord splice to the
right of
connection 7 (see Fig. 29) can be made readily with flange plates bolted along two sides as
shown in Fig. 34.
No = 1215kW b =35mm
d - 27 mm p = 75mm 45 357575 35 45
d’ ~ 30 mm to = 6.3mm
a = 45mm
a (effectwe) : 43.8 mm ( a 1.25 b )
Fig. 34 - Tension chord splice in uniplanar truss design example
59
Discuss me ...
Refer to Section5.1.1 and makea trial arrangement. Load is 1215 kN.

Try 27 mm diameter high gradebolts, having a factored tensileresistance (T,)of 300 kN
If 6 bolts areused, P, = 202.5 kN and P,ITr = 0.68
FromTable 9, 6 = 1 - d‘lp (Eqn. 5.2)
= 1 - 30/75 = 0.60
b’ = b - d/2 + to (Eqn. 5.1)
= 35 - 13.5 + 6.3 = 27.8 mm
K = 4 b’/(0.9 f,p) (Eqn. 5.4)
= 4 (27.8)/[0.9(0.355)751 = 4.64
Plate with a yield strength of 355 N/mm2 has been assumed
,t = (KP,/(l + (Eqn. 5.3)
= (4.64 (202.5)/(1+ 0.60)]0.5= 24.2 mm
,,,t = (KP,)0.5 (Eqn. 5.3)
= 30.7 mm
= (4.64 (202.5))0-5
Try a 28 mm plate
Determine a,the ratio of the “sagging” plate moment at the edge of the bolts to the
“hogging”
plate moment within the tube.
a = [(KT$;) - l ] [(a + d/2)/6 (a + b + to)] (Eqn. 5.5)
= [(4.64(300)/28‘) - l ] ((43.8+ 13.5)I0.60 (43.8 + 35 + 6.3)]
= 0.87
Calculate the splice tensile
resistance.
N,’ = t; (1 + 6a)nlK (Eqn. 5.6)
= 28’ (1 + 0.6 (0.87))6l4.64
= 1543 2 1215 ... connection is adequate.
For general interest, calculate the actual total bolt tension, including prying
force.
T, = P,[l + (b‘/a‘) (6a/(l + 6 4 ) ] (Eqn. 5.7)
a’ = a (effective) + d/2 = 43.8 + 13.5 = 57.3 mm
= [(KPllt;) - l ] [1/6] (Eqn. 5.8)
= [(4.64(202.5)/28‘) - l ] [ 10.61
l = 0.33
TI = 202.5 [l + (27.8l57.3)(0.60)(0.33)/(1+ (0.60)(0.33))]
= 219 kN > 202.5 kN (P,)
< 300 kN (T,)
5.2 RHS to gusset-plate connections
RHS bracing members can befield bolted to gusset plates which have been shop welded to
RHS chord members, thus producing bolted shearconnectionsasshown in Fig. 35. If
Weld along RHS at
4 corners w i t h
weld length = L
facilltate
( a ) Slmple shear spllce (b) Modlfled shear spllce

Fig. 35 - Bolted RHS gusset-plate connections
60
Discuss me ...
dynamic loading is a design consideration, this type of connection has an advantage over
bolted flange-plate connections in that flange plates must be proportioned to eliminate all
pryingwhenfatigueloadsarepresent. In general static loadapplications,however,the
gusset-plate connection is less aesthetically pleasing and often more expensive than its
flange-plate counterpart. An important limitation to the use of RHS gusset-plate connections
is the need to have closely matching member widths. Equal width members are connected
directly as in Fig. 35 (a), but the gussets often need to be spread slightly by jacking after
welding is complete in order to allow field assembly (welding contraction tends to pull the
gussets inwards). Small width differences can be adjusted by ofthe filler
use
plates welded on
the sides of the bracing member. Larger differences allow the further option of extra plates,
Fig. 35(b), which can be more convenientin the field.
5.2.1 Net area, effective net area, and reduced effective net area
The conceptof gross area, net area, effective net area, and reduced effective net area can be
used to describe various failure modesa tension
for member with holes or openings
[41], and
these conceptswill be utilized herein. The three basic checks are:

oss T, = dJA,f,
on (yielding (5.9)
a) netT, effective
= 0.85 dJA,,fon
, (rupture
T, = 0.85 +Kef, (rupture on
effective
net
area
reduced for
shear
lag)
(5.1 1)
where dJ is a reduction factor which can be taken as 0.9 (Note: @I = 1I y M ) .
The0.85factorrepresents a suitableminimummarginbetweenfactoredloadsandthe
factored ultimate resistance for failure modes established by of fracture
the tension members.
The 0.85@I term = 0.77 is comparable to that used in other limit state codes(e. g. Eurocode 3
[21] requires a factor of 117, = 0.80).
The effective net area A,, is the sum of individual net areas A, along a potential critical
section of the member. Such a critical section may comprise net area segments loaded in
tension, segments loaded in shear, and segments witha combination of the two loads. The
method given provides a means of checking against “block shear”failures, whereby a chunk
of material tears away from the piece a combination
by of shear and tension ruptures [21,41].
An illustrative example which includes all three typesof net area segments is the gusset plate
Y connection in Fig. 36, where the “block shear” area is calculated from the proposed failure
line AM.
The tension segment, perpendicular to the load (AB), has A, = w,t = (g, - d’12) t
Shear segments parallel to the load(GH, JK, LM), have in total A, = 0.6 L,t
= 0.6 (L - 2.5 d’) t
so that sectionsfailing in shear and
The 0.6 factor relates shear strength to tensile strength,
tension can be “added together”.
diameter = d’
Fig. 36 - Calculation of
effective net areaA,,,
for a gusset plate ___- ___ -
61
Discuss me ...
Each inclined segment (CD or EF) has

,, = (W, + S2/4g2)t = (g2 - d’) t + (S2/4g2)t
A
The effective net areaA, of the gusset plate for the potentialcritical section being examined
is therefore thesum of all the net area segments above,A, + A ,, + 2A,.
For bolted connections the effectivenet area, reduced for shear lag, A
,; is theeffective net
area A
,, multiplied by a shear lag factor. This factor comes into operation when a member is
connected by some but not all of its cross-sectionalelements, if the critical net section includes
elements which are not connected. The critical net section in this instance may include net
area segmentsA, which are perpendicular the to loador inclined to it, but not those which are
parallel to it. (This is not a check against block shear tear out.) The shear lag factor to be
applied to A
(,; = shear lag factor x A ), can be taken as:
0.90 when shapes like l-sections (or teescut from them)are connected only by their flanges
with at least three transverse rows of fasteners(flange width at least two
thirds the depth),
0.85 for all other structural shapes (e.g. RHS) connected with three
or more transverse rows
of fasteners,
0.75 for all members (e.g. RHS) connected withtwo transverse rowsof fasteners.
For example,if the bracingmember in Fig. 36 was an RHS and it was bolted togusset plates
on two sides as shownin the figure with each side having eightbolts, in three rows, then the
reduced effective net area(A;,) would be 0.85 A., In this instance the effective net areaA ,,
would bethe lesser of the failure pathsAB-CD-EF-GN and AB-CF-GN (see Fig. 36).
The effective net areareduced for shear lag,A,; also applies towelded connections, when
a member is not welded all around its cross-section. An example is Fig. 35(b) where bolting
plates were weldedto thesides of the bracingmember. For weldsparallel to thedirection of
load (as welds would bein Fig. 35(b), along the corners of the RHS), the shear lag factoris a
function of the weld lengths and the distance between them. The distance between such
welds would be bi. The shear lag factor to be appliedto A(A
, ;, = shear lag factor x A), is:
1.OO when the weld lengths(L) along the RHS corners are 2 2 b,,
0.87 when the weld lengths (L) along the RHS corners are 1.5 bi 5 L < 2 bi.
0.75 when the weld lengths (L) along the RHS corners are b,S L < l .5 b,.
The minimum lengthof welds (L) is the distancebetween them.
Another failure mode ofthe gusset plate which must be checked is yielding across an effective
dispersionwidth of the plate, which can be calculated using Whitmore
the [42] effectivewidth
concept illustrated in Fig. 37. For this failure mode (for two gusset plates),
NI*= 4fy,2t,(g + l.l5,Yp),
where4 = 0.9 (5.12)
The useof Nrindicates thatthis check applies toboth tension and compression load cases. If
the member is in compression, buckling of the gusset plate must also be prevented. The term
Cp represents the sumof thebolt pitchesin a boltedconnection or thelength of the weldin a
welded connection.
Fig. 37 - Whitmore criterion for gusset-plate yielding
62
Discuss me ...
6 RHS to RHS moment connections
6.1 Vierendeelconnections
6.1 .l Introductionto Vierendeel trusses
Arthur Vierendeel first proposed Vierendeel trusses in 1896. They are comprised of chord
members connectedto bracing members which are nearly always at 90° to the chords. The
typical design premisewith Vierendeel trusses has been to assume full connection rigidity, but
this is very rarely the case with RHS to RHS Vierendeel connections. Unlike triangulated
Warren or Pratt trusses, in which the connections approach a pinned condition at their
ultimate limit state and cause the bracing members to be loaded by predominantly axial
forces, Vierendeel connections have bracing members subjected to substantial bending
moments as well as axial and shear forces.Until very recently, most of the testing performed
on Vierendeel connections has been on isolated connection specimens with a lateral load
applied to the vertical bracing member while the connection is in an inverted T position as
shown in Fig. 38. Thus, the connection strength and moment-rotation behaviour have been
assessed mainly by researchers under momentplus shear loading.
Fig. 38 - RHS Vierendeel connection types [51]

(a) Unreinforced
(b) With bracing plate stiffeners
(c) With chord plate stiffener
(d) With haunch stiffeners
(e) With truncated pyramid stiffeners
63
Discuss me ...
Square and rectangular RHS single chord connections loaded by in-plane bending moments
have been studied by Duff(431, Redwood [44], Cute et al. [45], Mehrotra and Redwood[46],
Lazar and Fang (471, Mehrotra and Govil(481, Staples and Harrison (491, Brockenbough (501,
Korol et al. [51], Korol and Mansour[52], Giddings [53], Kanatani et al. [54], Korol et al. [27],
Korol and Mirza [55], Mang etal. [22], Davies and Panjehshahi [56], Szlendak and Brodka [57,
58, 591, Szlendak [SO, 611, and Kanataniet al. 1621. Researchers concur that both the strength
and flexural rigidity of an unstiffened connection decrease as the chord slenderness ratio
boltoincreases, and as the bracing to chord widthratio b,/bo (or p ) decreases. Connections
with p = 1.O and a lowboltovalue approachfull rigidity, but all other unstiffened connections
can be classed as semi-rigid. For such semi-rigid connections, Fig.to38 (b)
(e) gives a variety
of means of stiffening which have been used to achieve full rigidity. From these alternatives,
Figs. 38 (c) and (d) are recommended since the resistance of Fig. 38 (b) is limited by the
effective width criterion and Fig. 38 (e) is rather expensive to fabricate.
6.1.2 Connection behaviour and strength

The connection ultimate moment capacityin tests is typically recorded, and Korol et al. [51]
even developed an empirical formula for estimating the maximum connection moment, but
this moment typically occurs at excessively large connection deformations. Thus, for all
practical design purposes, the moment capacity of a connection can be determined in a
manner similar to that used for axially-loaded RHS T connections, whereby the strength is
characterized by an ultimate bearing capacityor by a deformation or rotationlimit [19]. This
design approach is more apparent if one considers the possible failure modes for such
connections, which are shownin Fig. 39. The failure modes representedin Fig. 39 presume
that neither the welds nor the members themselves are critical (e.g. local buckling of the
bracing is precluded). Cracking in the chord (chord punching shear) has not actually been
observed in any test, andchordshearfailureis strictly amember failure, so analytical
solutions for failure modes(b) and (e) are not considered herein.
(a1 (b) (dl

Fig. 39 - Possible failure modes forRHS connections loaded by in-plane bending moments
1191
(a) Chord face yielding
(b) Cracking in chord
(c) Cracking in bracing member
(d) Crippling of the chordside walls
(e) Chord shear failure
For Mode (a), the moment capacity of connections with low to moderate p values can be
determined by the yield line model in Fig. 40. Neglecting the influence of membrane effects
and strain hardening, the moment capacity is given by Wardenier[19]:
for p 5 0.85.
64
Discuss me ...
Fig. 40 - Yield line mechanism for chord face yielding under in-plane bending (failure
mode (a))
The term f(n) is a function to allow for the reduction in connection moment capacity in the
presence of large compression chord forces, accordingto de Koning and Wardenier1631:
f(n) = 1.3 + );O( -

but n, cr 1.0; (6.1 a)
where n = axial compression loadin the chord (negative) expressedas a fraction of the chord
yield load = No/A,,fy0 +
MO
m.
For tension chords,f(n) = 1.O. Eqn. 6.1 ais shown graphically inFig. 41
Functlon f (ni
-1.0 -0.8 -0.6 -0.4 -0.2 0

n
Fig. 41 - Connection strength reduction factor f(n), as a function of the compressive load in the chord
(
expressed as a fraction of the chord yield load n = N,/A,fy,
For chord tension loading f (n) = 1 .O
+-
)
so . $0
Nearly all Vierendeel connections have thebracing to chord angle 0 = 90°, which simplifies
Eqn. 6.1 to:
for p I0.85.
65
Discuss me ...
For Mode(c), an effective width approachis used torelate thereduced capacity of

the bracing
member (consideredto be the same onthe tensionand compression flangesof the bracing
member) tothe applied bracingmoment as follows [ 191:
%) b, t, (h, -
(l- b l
The term be inEqn. 6.3 is the effectivewidth of the bracingmember flange, andis given by:
For Mode (d),a chordside wall bearing or buckling capacity can conservatively be given by
Eqn. 6.5 [l91 which is illustrated in Fig. 42.
M,', = 0.5 fkto(h, + 5t0)' (6.5)
M,,
____
-- -- - - -- -
Fig. 42 - Chord side wall bearingor buckling failure model under in-plane bending (failure mode (d))
This moment is derived fromstress blocks oftwice (two walls) fkto (h, /2 + 2.5 to)acting as a
couple at centres of h1/2 + 2.5t0. Since the compression is very localized, tests [63] have
shown that buckling is less critical for moment-loaded T connections thanaxially loaded T
connections.Hence, within theparameter rangeof validity given,the chord yield stress can be
used insteadof the buckling stress for T connections. For X connections this is reduced by
20% to be consistentwith Table 3.
Hence, for design purposes an estimateof the connectionmoment capacitycan be obtained
from thelower ofthe M; values obtained from Eqns.6.2,6.3, and6.5. It can be seen thatthe
moment capacity predicted by Eqn tends 6.2 towardsinfinity as p tends towards unity, and so
this failure mode, which corresponds to a state of complete connection plastification and
hence high connection deformations, is not critical in the highp range. This accounts forthe
p c: 0.85 limit attached to Eqn. 6.2 and, for high p values, the web crippling failure criterion
expressed by Eqn. 6.5 will likely govern. A summary of the design equations for in-plane
moment loading is given inTable 10.
From the above expressions for M; it can be seen that full width (p = 1.O) unstiffened RHS
Vierendeel connections are capable of developing the full moment capacity of the bracing
member, providing b,/t, is sufficiently low. For h, = bo = h, = b, and h,/t, S 16, the chord
side wall webcrippling capacityis given by Wardenier [19]:
M,', = 12fyOt~hl (6.6)
a square RHS bracing member is given approximately by
Since the plasticmoment capacity of
M, = 1.5 b:t, fyl (6.7)
66
Discuss me ...
Table 10 - Factored resistance of welded T and X connections of rectangular hollow sections

under in-plane and out-of-plane bending moments
type of connection factored connection resistance
T and X connections under p S 0.85 yielding

face
chord
basis:
in-plane bendingmoments
0.85 < p S 1.O basis:

effective
width
T
.-’ M,, 0.85 < p 5 1.O basis: chord side
wall
faiture
M,; = 0.5fkt0(h, + 5t0)‘

T and X connections under p 5 0.85 basis: chord face yielding
out-of-plane bending moments
e - goo
0.85 p I1.Oa) basis:

effective
width
M&, = fyl [Z, ~ 0.5 t, (b, - (6.13)
0.85 p S 1.Oa) basis:

chord
failure
wall
side
Mp; = fkto(h, + 5t0) (bo - to)”’ (6.14)
functions
‘(n) = 1.Ofor n 2 0 (tension)
0.4
‘(n) = 1.3 +- . nforn < 0
P f, = ,,f for T connections under in-plane bending
but I1.0 f, = 0.8 fyo for X connections under in-plane bending
NO MO For out-of-plane bending optainf, from Table3.
l=-+-
A0 fyo so ‘ $0
range of validity
y, I 355 N/mm2 b,/t, 5 1.1 6

to and ho/to5 35 8, = 90”
Notes: a) For design on thisbasis,distortion of thechordsectionmustbeprevented.Rhomboidal

distortion of the chord can be preventedby the useof diaphragm stiffening.
67
Discuss me ...
I
B
Cycle path footbridge over railway and carpark
a truss p = 1, dimensional ratiosof bolto= 16 and to/t,

So, for the same steel grades used throughout
= 2 will produce a connection with a moment capacity approximately equal to the plastic
moment capacity of the bracing [l91when hot-formed material is used. In this case, the
bracing member cross-section is fully effective (be = b, in Eqns 6.3 and 6.4). The above is
similar to the recommendation by Korol et al. [51] for hot-formed sections that boltobe less
than 16 with p = 1 for full moment transfer to be assumed at theconnection.
Any resistance factor (4) or partial safety factor (Y,,.,) is alreadyincluded, where necessary,in
the above resistance expressions of M; for their use in a limit states design format. Even
though a rigorous evaluation of the proposed equations against all available experimental
data has not yet been performed, experience suggests that these proposed equations will
prove to be practicallower bounds on the connection moment capacity. The expressions for
M; also havea limited range of validity which correspondsto the limits of the test data against
which the equations have been checked. Thisvalidity range is: boltoI35, ho/toI35, 8, =
90°, fyi nominal I355 N/mm2, and the compression bracing member is restricted to plastic
design sections (see Table10).
68
Discuss me ...
The weldsin RHS moment connectionsare loaded in a highly non-uniformmanner and should
also be capable of sustaining significant connection 'rotations. To enable adequate load
redistribution to take place, the fillet weld sizes should be at least as large as those now
specified for axially-loaded RHS truss connections to develop the capacity of the bracing
member (see Section2.2).
The previous expressions for moment capacity areonbased moment loading only, whereasin
Vierendeel trussessignificant axial loads may also of
exist in the bracing members. The effect
the axial load onthe connection momentcapacity depends on the critical failure mode, andso
a complex setof interactions is developed. Consequently,it isconservatively proposed that a
linear interaction relationship be used to reduce the in-plane moment capacity of a Vierendeel
connection as follows:
Mi, and N, are the applied bending moment and axial load respectively in the bracing
member, Mi., is thelower ofthe values obtained from Eqns. 6.2,6.3 and 6.5 (Table lo), and N;
is the connection resistance with only an
axial load applied to the bracing member (Table 3).
Offices and production unit;RHS lattice columns, external
69
Discuss me ...
The resistance of an RHS T connection under bracing member axial load is given and
of P = 1. There
discussed in Section 3.2, but is reproduced below for the most relevant case
are two failure modes to be checked; web crippling
of the chord member side walls is again the
likely governing failure mode, and can be estimated by [2]:
N; = fkto(2h, + loto) (6.10)
where f, is determined in Table 3. The value for f, in Eqn. 6.10 assumes that the bracing
member is in compression; if the bracing is in axial tension f, = ,,f which corresponds to
chord wall tensile yielding. The other failure mode to check,RHS P=
for Tanconnection with
1, is premature failure ofthe bracing member or connecting weldment. This is also termed an
“effective width” check on the bracing member, and is expressed by [2):
N; = fylt, (2h1 - 4t, + 2be) (6.11)
where be is given by Eqn. 6.4. Thus the connection resistance as an axially-loaded RHS T
connection, for P = 1, will be given by the lowerof the NTvalues from Eqns. 6.10 and 6.1
1.
6.1.2.1 In-plane bending moments for T and X connections
The designcriteria for RHS T connections with the bracing member subjected to an in-plane
bending moment (Mip) are described in Section 6.1.2. These are summarized in Table 10. For
RHS X connections,subjecttoequalandopposite(self-equilibrating)in-planebending
moments (Mip) applied to the bracing members, connection resistance formulae are also
given in Table 10. These arethe same as for RHS T connections except a reduced bearing
strength is used forthe chord side wall failure mode.
In thecase of the stiffened connection shownin Fig. 38(c), the effect of the stiffening can be
treated in a similar way to that
of axially-loaded, plate-reinforced T connections(i. e. modify the
formulae in Table 10 in a similar way to Section 3.5.1.2). For haunched connections with
p > 0.85 as shown in Fig. 38(d), recommended minimum haunch dimensions are shown on
the figure and connection resistance should be checked using Eqn. 6.5 in Table 10 with a
modified value of h,. For haunched connections with p I0.85 use Eqn. 6.2 with a modified
value of h,.
6.1.2.2 Out-of-plane bending moments for T and X connections
For RHS T connections with the bracing member subjected to an out-of-plane bending
moment (Mop),such as shownin Fig. 43, there is very
little test evidence available to support
any design models. However, one can postulate analogous failure modes to those described
above for in-plane moment loading, which has been doneAWS for [16].
(a) For B S 0.85, design would likely be governed by chord face yielding as shown
in Fig. 43.
For this yieldline mechanism,
(6.12)
where f(n) is given by Eqn.6.1 a. It should be noted that for this failure
all deformation takes
place in the chordface and thechord will therefore not distort as a rhombus.
(b) For 0.85 @ S 1.O, design would likely be governed by the more critical failure mode
between: Reduced bracing member capacity (or an “effectivewidth” failure mode), and chord
side wall bearing orbuckling capacity (see Fig. 44).
For
“effective width” failure: Mip = f, [Z, - 0.51,
(b, - be)2] (6.13)
70
Discuss me ...
- Hoggingyleld lines
_ _ _ Sagging yield lines
MO,
- MOP
Fig. 43 - T connection subject to an out-of-plane bending moment, showing the chord face
yielding failure mode forp S 0.85
Z, in Eqn. 6.13 is the plastic section modulus about the correct axisof bending, and plastic
design sections should be selected for thebracing member. The termbe is defined by Eqn.
6.4.
chord
For
side
wal!
failure: M& = fkto (h, + 5t0) (bo -(6.14)
to)
The term f, is the bearing strength of the chord side walls for T connections, and can be
obtained from Table 3. These design provisions RHSfor T connections subject to out-of-plane
bending are summarizedin Table 10.
(a) (b)
Fig. 44 - T connection subject to an out-of-plane bending moment, showing the basis

of design models
for: (a)Effective width failuremode
(b) Chord side wall failure mode
One could see that the designcriteria for RHS X connections, subject to equal and opposite
(self-equilibrating) out-of-plane bending moments (Mop) applied tobracing
the members, are
again the same as those given above for T connections with one exception. The difference
would be that when determining the resistance the for chord side wallfailure mode, fk should
be reduced by 20% as indicatedin Table 3. The design formulae are also covered by Table 10
6.1.3 Connectionflexibility
In the foregoing it was shown thatunstiffened RHS connections withp = 1 and certain bolto
and to/t, values could achieve thefull moment capacityof the bracing member, butit should
be remembered that any connection moment resistance calculated (Mi;) must be reduced to
71
Discuss me ...
take account of the influence of axial load in the bracing member (see Eqn. 6.9). Such
connections, which still develop a moment resistance exceeding the moment capacity of the
bracing member, can be consideredfully as rigid for the purposeof analysis of the Vierendeel
truss. All other connections (which covers most possible connection combinations) should be
considered as semi-rigid.
To analyse a frame, which is connected by semi-rigid connections, one needs the load-
deformation characteristics of the connections being used, and these can be obtained by
either reliable finite element analysisor from laboratory tests.
6.1.4 Design example
P of 17 kN.
The Vierendeel truss shownin Fig.45 is to be designed for a factored panel load
All the connectionpointsare laterally braced,perpendiculartothetruss,bysecondary
(P (P (P (PI2 (P (PI2
B C E G I K
A D F H J L
Fig. 45 - Example Vierendeel truss
members. The top and bottom chord members will be the same, and one section size
will be
used for all vertical (bracing) members. A statically admissible set of moments and shears
follows in Fig. 46. Memberswill be designed using plastic analysis.All members chosen are
hot-formed hollow sections with dimensions conforming to ISOlDlS 657-14 [29]. The steel
grade throughout is Fe510 conforming with IS0 630 [30], with a minimum specified yield
strength of 355 N/mm2. Reductions in plastic moment capacity due to axial force or shear
force can be shown to be negligible [64].
P P P
-1.5P- (E-3.9P. - lp-5.1~- - II
-
r-
1.5P
JI
1.25P
2.4P
-L
7- O.%P LAzp d!2> -lo 0.45P
-1.75P -0.5P -0.5P -0.5P
Member axial forces and shear forces
1.875P
\
1.875P
1.125P
1.125P
r
0.375P
0.375P
0.375P
\ r ) r
e1 7
1.875P
\I c-l
3P
. 1.5P
0
1.875P 3P 1.5P
1.875P
1.875P
1.125P
1.125P
0.375P
0.375P
0.375P
Bending moments
Fig. 46 - Forces and moments within Vierendeel

truss (forces and momentsshown applied to nodes)
72
Discuss me ...
Select 150 x 150 x 10 RHS for the chord.

Confirm thatthis section is Class
1 (suitable forplastic design) at the worst axial load
condition.
Maximum moment = 1.875 P = 31.9 kNm
Plastic moment of resistance = Z, .,,f = (282) (0.355) = 100.1 kNm > 31.9 :. O.K.
Note: The resistance above has been calculated assuming yU = 1.O (i. e. no partial safety
factor or resistancefactor),tobeconsistentwithChapter 4.1. Designersshould
introduce theappropriate partial safety factor/resistance factor for member design.
Therefore, 150 x 150 x 10 RHS is suitable for the chords.
Select 150 x 150 x 6.3 RHS for the verticalmembers.

Again, confirm that this section is Class 1 (suitable for plastic design) at the worstaxial load
condition.
Maximum moment = 3 P = 51.0 kNm
Plastic moment of resistance = (191) (0.355) = 67.8 kNm 2 51.O :. O.K.

(This again ignores anypartial safety factorlresistance factor to be consistent with member
design elsewhere.)
Therefore, 150 x 150 x 6.3 RHS is suitable for the verticalmembers.
Plastic collapse mechanism

Fig. 47 illustrates the collapse mechanism. Let X' be the additional multiplication factor by
which the already factored loads have to be increased to cause plastic collapse. By the
principle of Virtual Work,
17X' (30 + 60 + 60 + 60 + 30) = 100.1 (40) + 67.8 (80)
:. X' = 2.31
Therefore, adequate reserve capacity exists for
ultimate strength as X' 2 1.O.
Fig. 47 - Plastic collapse mechanism for Vierendeel truss
Connection capacity check

As p is 1.O, the moment resistanceof the connectioncould be limited by Mode (c),cracking in
the bracingmember, or Mode (d),chord side wall buckling (see Table 10).
73
Discuss me ...
Mode (d)
MI; = 0.5 fkto(h, + St,)’
f, = ,,f for T connections (Table 10)
:. MI; = 0.5 (0.355)(10)(150 + 50)’
= 71 kNm 2 51.0 ... O.K.
:. The limiting moment resistance is67.8 kNm
Now check that the moment andaxial force interaction is satisfied according to:
(6.9)
N; = fkto(2hl + lot,) (6.10)

= 0.330 (10)(300 + 100) = 1320 kN.
or N; = ,f t, (2 h, - 4 t, + 2 be) (6.11)
be = [lO/(b,/to)] (to/tl) b, = b,as before
SO N; = 0.355 (6.3)(300 - 25.2 + 300)
= 1286 kN
:. governing value of N; = 1286 kN

Hence, oneshould check the connectionsto the outside posts (maximum axial compression
force of 1.75 P = 29.8 kN), and theconnections tothe most critical interior vertical (having a
maximum moment of 3 P = 51 kNm).
For outside posts:
(29.8/1286)+ (31.9/67.8)
= 0.49 5 1 .O :. O.K.
For interior verticals:
(8.5/1286)+ (51.0167.8)
= 0.76 5 1.0 :. O.K.
Therefore, connection resistance is adequate and trussis satisfactory.

Thememberswouldalsobe suitable byelastic design procedures,andevenwiththe
introduction of a partial safety factor (resistance factor) applied to member resistance. By
either design method, thechord thickness is still enhanced to provide adequate connection
strength. The endconnections(at A, B, M and N) can be made by welding verticalthe posts to
the chord to form T connections, and then adding cappingplates to the ends of the chord
sections.
6.2 Kneeconnections
Research onmitred RHS knee connections (such as those in Fig. 48) has been performed by
Mang et al. [65, 66) at the University of Karlsruhe. Their recommendations have also been
reported by Wardenier [19],CIDECT [l], Dutta and Wurker [25]and Eurocode 3 [21].They
cover both stiffened and unstiffened knee connections, and are intended for use in corner
connections of rigid frames.
The original test results and moment vs. rotation diagrams are not widely available, but
CIDECT [ l ] makes its design recommendationsapplicablefor “flexurally-rigid frame
corners”. However, it could be expected thatthe rotation capacity ofsome unstiffened

connectionsmightbe low,and instructuresin whichreasonable rotational capacity is
required, a stiffened knee connection should be used [19]. In the Karlsruhe tests, simple
unstiffenedknee connections tended fail to by excessive deformation of the lateral RHS cross-
74
Discuss me ...
Section A - A Section B - B
fino
a=t 4
t>Zmrn
Detail E
Detail C n A
W M
Detail D
3 Detail F
(a) (b)
Fig. 48 - Details of RHS knee connections[25]
(a)Unstiffened
(b) With a transverse stiffening plate
wall in compression. On the other hand, for connections with a stiffening plate, excessive
deformationsappearedonlyforvery thin walledmembers.Forthickerhollowsections,
complete plastification was reached in the course of the tests [l]. In view of the uncertain
moment vs. rotation propertiesit is suggested to use, forunstiffenedconnections, only RHS
members which satisfyplastic designrequirements forrigid frames.
Analysis of the test results showed that, for design purposes,it was possible to estimate the
total flexural and axial loadcapacity of the connection by applying a reductionfactor to the
material yield stress. Thus, adequate connectionstrength will be obtained forboth stiffened
and unstiffened 90° mitre connections providing Eqns. 6.15and 6.16are satisfied [21].
N, + -
M,,
S a , for i = 1 and 2
Fig.
(see 48) (6.15)
Nr, - Mr,
N,, is used here to refer to theaxial resistance of memberi, either in compression or tension
as applicable, M,, refers to the moment resistance of member i. The term CY is a stress
reduction factor which can be taken as 1 .O for mitre connections with stiffening plates. For
mitre joints without stiffening plates, (Y is a functionof the cross-sectional dimensions andis
given in Figs. 49 and 50.For connections withoutstiffening plates, NI should also not exceed
0.2Nr,[21].
The shear forceacting at the connection V should also meet the requirement
[21]:
V
- 5 0.5 (6.16)
v,
where V, is the shear yield load in the member under consideration. This can be takenthe as
yield stress in pure shear (fy,/fi) multiplied by the cross-sectional area of the RHS webs
(2h, t,).If Eqn. 6.16 is notstatisfied, the connection strengthcould still be deemed adequate,
providing the combined stress does not produce failure according to the Von Mises failure
criterion; in doing this check, the normal stresses(axial and bending) should be increased by
a factor of 1 la.
75
Discuss me ...
m
m
m
m
m
m
m
I
I
H
H
I
I
r
Triangular multiplanar trussfor a pipeline bridge
For stiffened knee connections,the plate size

should complywith [21]:
t, 2 1.5ti (i = 1 or 2),and t, 1 10 mm. (6.17)
The fabrication detailsshown in Fig. 48 are recommended. The weld size can be considered
to be adequate when the throat thickness (a) is equal to the
connected wall thickness, plus the
factor (Y (for unstiffened knee connections) is S 0.71. This rule pertains to RHS having
fyi = 355 Nlmm’. If RHS with fyi = 235 N/mm2are used forunstiffenedknee connections,this
can be adjustedto (Y S 0.84.
76
Discuss me ...
If mitred knee connections are usedwith an obtuse angle betweenthe RHS members (i.e.
e > 90° in Fig. 48), the same
design checkscan be undertaken as for right-angleconnections,
since obtuse angle knee connections behave more favourably than right-angle [l]. onesFor
unstiffenedknee connections with 90° < 0 c 180°, this strength enhancement can be used to
advantage in Eqn. 6.15 by increasing thevalue of a as follows:
a = 1- $)(1 - ayg
(ecos = (6.18)
aH= is the value obtained from Fig. 49 or Fig. 50.

An alternative form of connection reinforcement (other than a transverse stiffening plate) is a
haunch onthe inside ofthe knee. This haunch piece needs to be ofthe same width as the two
main members, andcan easily be provided by taking a cutting from one of the RHS sections.
Provided the haunch length is sufficientto ensure that the bending moment does not exceed
the section yield moment (Siryi) in either main member, the connection resistance will be
adequate and does not require checking[l].
for a pipeline bridge

Quadrangular multiplanar truss
77
Discuss me ...
8L
Discuss me ...
hlb Q=
5 10 252015 30 35 40 45
bl t
Fig. 49 - Stress reduction factors01, for RHS subjected to bending about the major axis
in 90' unstiffened
mitred knee connections[65]
5 10 15 20 25 30 35 40 45
blt
Fig. 50 - Stress reduction factors for RHS subjected to bending aboutthe minor axisin 90' unstiffened
mitred knee connections[SS]
79
Discuss me ...
7 Multiplanar welded connections
Multiplanar connections are frequently used in tubular structures such as towers, space
frames, off-shore jacket structures, triangular trusses, quadrangular trusses and many other
applications. Design rules for suchconnections, however, are given by specifications or codes
only for CHS, andonly by AWS [l61 and CIDECT [67] at present.
RHS KK connections
Even less attention has been devotedto multiplanar connections between RHS compared to
connections between CHS. Initial tests by Coutie et al [68] on RHS multiplanar connections
(KK) found a small decrease in the strength of the in-plane K connection due to out-of-plane
loaded bracing members, as found for CHS connections. Bauer and Redwood [69], for KK
connections to the single RHS chord of a triangular truss as shown in Fig. 51, deduced that
Fig. 51 - RHS triangular truss with double compression chords.
there was little interactive effect produced by identical loading (same sense) on an adjacent
wall of the chord. Bauer and Redwood's study concentrated on connections with low to
medium width ratios (p) between the bracing members and chord, for which the yield line
method represented an ideal form of analysis. On this basis, it was suggested thatin cases
where the angle between bracing member planes (p in Fig. 52) was less than 90"leading to
an increase in the effective value of p at the chord face, and when the bracing members were
attached to the chord face off-centre (as drawn in Fig. 52), then the strength of a triangular
truss tension chord face will begreater than that of planartruss
a chord face with the same size
members. As further failure modes may exist over a wider rangeof connection parameters
than thosestudied by Bauer and Redwood1691, and asuniplanar, gap, K connection strength
is traditionally assessed nowadayson the basis ofultimate strength rather thanpredicted yield
strength, for simplicity it is suggested that a reduction factor of 0.9 be applied to the uniplanar
K connectiondesignformulae in Tables 2 and 3. This is the same reduction factor as
recommended by CIDECT for CHSKK connections [67]. This recommendation for RHSKK
connections is made for 60° 5 p I90°, where p is shown on Fig. 52 121). In addition, it is
suggested that one always performs a chord shear check forKK connections (see Table
gap
1l ) , even for square RHS members [21].
Fig. 52 - Elevation viewof KK connection to triangular truss tension chord
80
Discuss me ...
Multiplanar RHS connection
Triangular trusses, such asillustrated in Fig. 51, have several advantages over single plane
trusses, such as the increased lateral stability offered by twin, separated, but connected
compression chords. They are frequently used as exposed structures and considered
equivalent in appearance, but less expensive, than space frames.In general,purlins are
also not necessarywith triangular trusses asthe usual practice is to space the top chordsof
the trusses at a distance suitable for the roof deck, and then fasten directly to the
the roof deck
flat surfacesof the RHS top chords.
RHS TT and XX connections

It has been found thatCHS for TT connections thereis no change in the T connection strength
due to out-of-plane loaded braces, while for CHS XX connections there is a significant
increase in strength forpairs of bracings loadedin the same sense (or significantdecrease for
pairs of bracings loaded in the opposite sense)[67].For RHS TT and XX connections however,
Davies andMorita [70]have shown theoretically that very little difference exists between the
81
Discuss me ...
Erection of an RHS pipeline bridge
design strengths of planar and multiplanar conmections,for 90° TT and XX connection

shapes. Although lacking in experimental evidenceat this stage, it isrecommendedthat a
correction factor of 0.9 be applied to the uniplanar T and X connection resistances
(Tables 2 and 3)to account for out-of-plane loaded bracings.
A summary of the correction factors for multiplanar RHS connections is given in Table 11.
-
Table 11 Correction factors forRHS multiplanar connectlon resistances
type of connection correction factor to uniplanar connection resistance from Table 2 or Table 3
0.9
KK gap foralso, KK connections,

that:
check
6oO~cp~9OO
TT, xx 0.9
82
Discuss me ...
8 Otheruniplanarconnections
8.1 Trusses with RHS bracing members framing into the cornersof the chord
With some multiplanar (or even planar) RHS trusses it is possible to have the truss bracing
members framing into the cornersofan RHS chord member, as shown in Fig. 53. This
necessitates verycareful profiling of the bracing member ends,particularly where cornerradii
are large, into so called “bird-mouth’’ or “bill-shaped’’ joints. Such a member arrangement
has been used occasionally in North America, for example in the Minneapolis Convention
Center Roof andin the Minneapolis/St. Paul Twin Cities Airport Skyway. It has also been used
in Japan, wherein this case a robot was developed profile
to the ends of the bracing members.
By framing into the corners of theRHS chord member high a connection strength and stiffness
is achieved, regardless of the bracing to chord member width ratio. Ono et al. [71] have
undertaken an experimental study of such square RHS T andK connections and have found
that these connectionsare much stronger than their conventional RHS counterparts with the
chord and bracings rotated through 45O about the member axis. All of the T connections
tested (25) hadthe bracing loaded in compression, andthe K connections (16) had all bracing
members inclined at 45O to the chord. The orientation of the bracing and chord member is
p J 0
(a) !b)
-
Fig. 53 RHS “bird mouth” T and K connections
shown in Fig. 53, and one should also note that the bracing member is axially rotated. Ono
et
al. concluded that the connection ultimate strengths could be given by:
For T connections:
Nl” = Gf,
1
0.211 - 0.147(b,/bO)
+ 1.794 - 0.942(b,/bo)
) f (n‘)
For K connections:
where the effective area coefficient CY is given for 45O K connections in Fig. 54. f(n’) is a
function usedforCHSconnectionstoallowfor the influence ofnormalstresses in
compression chords, and is given by [67]:
f(n’) = 1 + 0.3n’ - 0.3n”, for n’ < 0 (compression)

(8.3)
and f (n‘) = 1.O, for n’ 2 0 (tension) (8.3a)
where n‘
(8.3 = fop/fyo. b)
83
Discuss me ...
a
1.0
0.9
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0 8 0.9 1.0
-
bt
b0
Fig. 54 - Effective area coefficient a,for "bird mouth" 45O K connections
As these equations are based on a regression analysis of the test data, one should be careful
to ensure that they are only applied within the approximate bounds of parameter ranges
examined in thetests (i.e. 16 S boltoI42 and 0.3S b,lbo I1.0 forTconnections;
16 Ibolt, S 44, 0.2 Ib,lbo I0.7 and 0 = 45O for K connections). Resistance factors or
partial safety factors are necessary for application of Eqns. 8.1and 8.2to limit states design,
giving thefactored resistance expressions (with roundingof the constants), below:
For T connections:
N; = 0.9tgfy0
1
0.21 - 0.15 (b,lbo) + 1.79- 0.94(b,/bo)
) f(n')
For K connections:
with (Y given by Fig. 54.
8.2 Trusses with flattened and cropped-end CHS bracing members
For statically loaded hollow section trussesof small to moderate span, cropping - a procedure
in which a CHS bracing member is simultaneously flattened and sheared -
can simplify
fabrication and reduce cost. The procedure is faster than sawing profiling,
or the conventional
methods of preparingCHSbracingmembersforweldingtoRHS and CHS chords,
respectively, and it simplifies the welding process. Typical cropped-bracing Warren truss
connections to an RHS chord member are shown in Fig. 55. Note thatthe flattened endsof the
bracing member can be aligned in the directionof the truss or transverse it. toFor all trusses
with flattenedor cropped-bracing members aneffective lengthfactor of 1.0 should be

used for design of the bracing members.
Flattening the CHS bracings in the plane of the truss does not provideas good a structural
performance, nor the economies of fabrication, compared to out-of-plane flattening [72].
84
Discuss me ...
~ ~
( b ) Square chord transverse cropping
Fig. 55 - Cropped-bracingconnections to an RHS chord
Although this has been argued for CHS chord members, the out-of-plane flattening of CHS
--
bracing members and welding to RHS chord members is the basis of the “Strarch” roof sysem
[73]. At this stage no design guidance is available for such connections.
Asshown in Fig. 56, various types of flattening can be performed on the CHS bracing
members. In the case of full or partial flattening, the maximum taper fromthe tube to the flat
should remain within 25% (or 1 :4) as shownin Fig. 56 [67]. Fordoltoratios exceeding 25 the
flattening will reduce the bracing member compressive strength[l]. For welded connections
the length of the flat part should be minimized for compressionbracing members toavoid local
buckling in the flattened region.
4
c-* l Y
e-
Cropping ( A ) Full flatterllng (B. C) Partlal flattenmg ( D )
Fig. 56 -Various types of flattening for CHS bracing members
Considerable research has been performedon in-plane cropped-end CHS bracings to RHS
chord members. Tests of 32 Pratt isolated connections [74], eight similar test connections
incorporated into twotrusses[75]and 15 isolatedWarrenconnections [76] havebeen
reported. The latter tests had the geometry shown
in Fig. 57, in which the toes ofthe flattened
Sectlon Elevation
Fig. 57 - Cropped-bracing,zero gap Warren connections
85
Discuss me ...
.
Separated double chord truss

construction
Discuss me ...
Construction with Hollow Steel Sections - Design guide for rectangular hollow section (RHS) joints under predominantly static loadin
(9'8)
L8
0.02
Discuss me ...
8.3 Double chord truss connections
Limitations onthe largest available RHS member size have restricted the applicationofrange
RHS structures. For very long span roof trusses, such as sports centres and auditoria, the
of double RHS chord members will enable longer clear spans than those available from single
chord trusses. Immediate advantages of double chord RHS trusses include not only their
greater span capacity, but also more efficient stiffer
andconnections compared to some single
chord trusses. Enhancedlateral stiffness can reducelateral bracing requirements as well as
facilitate handling and erection of the structural components.
Research has been undertaken in Canada [78,79,80,81,82,83] on isolated connections and
trusses of the types shownin Fig.58. The two separated chord truss types requireall that
the
bracing members have the same width;in such cases the bracing member sizes can be varied
by changing the bracing member wall thickness (t,) or depth (h,). For the separated chord
bolted connection (Fig. 58(b)), it is recommended that tie bars be used between the RHS
chord members onthe outside of the truss as they significantly increase the truss stiffness
by
maintaining the alignment of the sections. However, connections without thesetie bars are
almost as strong, just more flexible.
Fig. 58 -Types of RHS double chordconnections

(a) Separated chord welded connection
(b) Separated chord bolted connection
(c) Back-to-back chord connection
For RHS double chord trusses it is recommended that a pin-jointed analysis be used with
effectivelengthfactors (K) asgiven in Section2.3.1,whendesigningthecompression
members. A microcomputer program DCTRUSS* forthe analysis and design ofRHS double
chord trusses is available[82, 831. Connectionresistanceexpressionshaveonlybeen
proposed for the separated chord welded connection, and these can be simplifiedto:
This software program has been made based on Canadian specification 1411
88
Discuss me ...
where AV = 2.6h0t0
h,/bo for 2 1 (8.8a)
and AV = 2 hoto for ho/bo< 1 (8.8b)
Eqns. 8.8a and 8.8 b take into account the reduced effectiveness of the chord outer side walls
in resistingshear forces, at different chord aspect ratios.
The interactionof axial force and shear force in the gapregion of the doublechord connection
should also be checked. The connection eccentricityhas been found to have little effect onthe
connection strength and pin-jointed analysis is recommended for the truss analysis, so that
moments acting on the connection can be ignored.axial The forcelshearforce interactioncan
be checkedin a manner similarto that used in Table 3, such that:
In Eqn. 8.9,A, is the area of one chord member, A, is given by Eqns.8.8a and 8.8 b,V is the
vertical shear force appliedto the connection (N, sin Oi assuming no “purlin load”), and V, is
given by:
fyo AV
v, =- (8.10)
fi
A recent economic comparison of single chordand doublechord RHS trusses [83]has shown
that for short spans single chord trussesthewere
lightestand most economical, being around
20% less expensive than back-to-back double chord trusses. (Back-to-back double chord
trusses are generally the heaviest and most expensiveoption for welded trusses.) Thus, for
long spans separateddouble chordwelded connections are preferable andwill again prove
more economical than back-to-back connections.
8.4 Plate to RHS connections
Plates are sometimes weldedto the faces of RHS members for bracing connections, purlin
cleat connections, hanger connections, and a pair of plates can even be used to represent the
flanges ofan l-section beamto RHS column momentconnection. For the latter, the
connection moment resistance can be obtained by multiplying the plate axial force resistance
by the beam depth (h, - t,).
Plates can be welded longitudinalor transverse to the RHS member axis.A longitudinal plate
will always have a very low p value and hence be anextremely flexible connection, with the
axial resistance governed by the formation of a yield line mechanism which represents a
control on the connection deformation. This is reflected in the design resistance given in Table
12, but this plate orientation is not recommended.
A transverse plate with a low to mediump value will also developa yieldline mechanism in the
RHS connection face. For large p values, but wherep is still less than 1 - 1/r,punchingshear
failure of the chordface is the most probable failure mode, and for somewhat lesser p values
a combination of flexural failureand punching shear will likely occur. This combinedfailure
mode has beenstudied byDavies and Packer184) but is too complicated for routine design.
However the bracing (plate) load capacity can also be reduced by a non-uniform stress
distribution in the plate, which is termed a bracing “effective width” failure criterion [85,861.
Moreover, it has been shown[63]that bracing effective width always governs over chord face
yielding as a critical failure mode for p 5 0.85, so the chord face yielding failure mode is
omitted in Table 1 2 and a bracing effective width checkis included for all p values. At p = 1 .O,
89
Discuss me ...
the plate will bear directly on the RHS side walls andso side wall failureis the pertinent failure
mode for whichthe connection must be designed.If the bracing is loaded in compression the
RHS chord side wall buckling stresscan still be takenastheyieldstress(fy0)asthe
compression is verylocalized. The foregoing design criteria are all summarized in Table 12.
Table 12 - Factored resistanceof plate to RHS connections

type of connection factored connection resistance
longitudinal plate p 5 0.85 basis: chord face yielding
wherep = t,/b,,f(n) = 1.3(1 + n)andq = hl/bo

For n < 0 (cornpr.): f(n) = 1.3(1 + n)
For n 2 0 (tens.): f(n) = 0
/3 5 0.2
transverse plate p = 1.0 basis:failure
chord
wallside
N; = fyoto [2t, + IOt,]

1 - llr
0.855p~ basis: punching shear
all p basis: effective width
N; = fy,t, be
functions
fyoto .
b = - .l o b, but 5 b, be = -.10 - b, but 5 b,
bolto bolt0 $1 tl
I
range of validity
b o / t o s30
90
Discuss me ...
9 List of symbols
Throat thicknessof a fillet weld; edge distanceof plate fromcentre of bolt hole (see
Fig. 32)
a (effective) + d12; a (effective) = a, but S 1.25 b (see Fig. 32)
Bearing area over which a transverse load is applied = h, b,
Dispersed bearing area = (h, + 2wJ b,
Cross-sectional area of memberi (i = 0, 1, 2, 3)
Gross cross-sectional area of member= A,
Effective net cross-sectional areaof tension member
Effective net cross-sectional area reduced for shear lag
Effective shear area of the chord (see Table 3 and Fig. 7 )
Distance from boltline to the hollow section face (see Fig. 32)
b - d12 + ti (see Fig. 32)
Effectivewidth of a bracingmember (see Tables2 and 3, and Fig.8)
Effective width for overlapping bracing member connected to overlapped bracing

member (see Tables2 and 3, and Fig. 8)
Effective punchingshear width (see Table 3 and Fig. 8)
Externalwidth of square or rectangular hollow section (RHS) member i (90° to plane
of truss) (i = 0, 1, 2, 3, j)
Width of plate (see Fig. 20)
Uncorrected connection efficiency, for K, T, and X-type connections respectively,
expressed asa proportionof the yield load (Aifyl) for a particularbracing member.
Nominal bolt diameter
Bolt hole diameter
External diameter of circular hollow section (CHS) for memberi (i = 0, 1, 2, 3)
Noding eccentricity -
for a connection positive being towards the outside of the truss
(see Fig. 1)
Modulus of elasticity
Crushing strengthof concrete
Axial stress in member i (i = 0, 1, 2, 3)
Buckling stress according to steelwork specification, using acolumn slenderness
ratio of KL/r
Yield stress of memberi (i = 0, 1, 2, 3) or j
Yield stress of plate
Ultimate tensilestress of member i (i = 0, 1, 2, 3)
Functions in the connection resistance formulae which incorporate the influenceof
normal stresses in compression chords
Function in “bird-mouth” connection resistance formulae which incorporates the
influence of prestressin compression chords
Maximum applied axial stress in chord (or maximum stress due to axial force and
bending moment where momentis taken into account)
Gap between the bracing members (ignoringwelds) of a K, N, or KTconnection,at the
face ofthe chord (see Fig. 5); bolt gauge,or distance between bolt lines (see Fig. 36)
Gap divided by chord wall thickness, g’ = glt,
External depth of square or rectangular hollow section (RHS) member i (in plane of
truss) (i = 0, 1, 2,3)
Subsript to denote member of connection; i = 0 designates chord; i = 1 refers in
general to the bracing for T, Y and X connections, or it refers to the compression

bracing member for K, N, and KT connections; i = 2 refers to the tension bracing
member for K,N and KT connections; i = 3 refers to the vertical for KT connections;
i = i refers to the overlappingbracing member forK and N type overlap connections
91
Discuss me ...
Moment of inertia of member

Subscript to denote the overlapped bracing memberfor K andN-typeoverlap
connections
Effective length factor
Length of member
Length of concrete in RHS chord member
Net length
Length of plate
In-plane bending moment applied to bracing member
Connection resistance for in-plane bending, expressed as a bending moment in
bracing member
Out-of-plane bending moment applied to bracing member
Connection resistance for out-of-planebending, expressed as a bending moment in
bracing member
Plastic moment capacity of member i
Moment resistance of member i
Bending moment in chord member

fo No MO ; number
n ~
fyo
- __
-
Aofyo
+- bolts
of
Sofyo
n
Axial force applied to member i (i = 0 , 1, 2, 3)

Connection resistance, expressed as an axial force in member i
gap) Reduced axialload resistance, due to shear,in the cross section of the chord at the
gap
No, Chord "preload" (additional axial force in the chord member ata connection which is
not necessary to resist the horizontal components ofthe bracing member forces)
NI' Axial resistance of member i
0" Overlap, 0, = q l p x 1000/0 (see Fig. 16)
P Length of projected contact area between overlapping bracing member and chord
without presenceof the overlapped member (see Fig. 16); length of flange-plateattri-
buted to each bolt, or bolt pitch (see Fig. 32 and Fig. 37); subscript to denote aplate
External tensile load applied to a bolt
Length of overlap betweenbracing members of a K or N connection at the chordface
(see Fig. 16)
r Radius of gyration
S Bolt spacing (see Fig. 36)
S, Elastic section modulus of member i
t,, t, Thickness of hollow section member i (i = 0 , 1, 2,3) or j
Thickness of plate
Tensile force applied to member or component
Tensile resistance of member or component or bolt
Ultimate tensile capacity of a bolt
Applied shear force
Shear yield capacity of a section (see Table3)
Net width
Dispersion width (see Fig. 23)
Plastic section modulus of memberi
lt Non-dimensional factor for the effectiveness of the chord flange in shear; ratioof the
equilibrating moment perunit plate width at the bolt line to the flange moment atthe
inner
plastic hinge; axial forcelbending interaction factor for mitred knee
connections; effective area coefficient for "bird mouth" connections
92
Discuss me ...
Width or diameter ratio between bracing member(s) and chord

bldl d, + d,
b, + b, + h, + h,
fi =bo-
bo
, - (T, Y, X), fi =
2b0 ' 4 bo (K! N)
d, + d, + d, b, + b, + b, + h, + h,+h,
8= (KT)
3 bo 6 b0
BP Widthordiameterratiobetween bracing memberand plate, 0, = b,/Bp
0 Connectionresistancefactor(inverseof partial safetyfactor y,,,,)
v Anglebetweenplanes of multiplanar joints (see Fig. 52)
b0
? Halfwidthtothicknessratioof the chord y = ~
2 t0
'I Bracingmemberdepth to chordwidthratio 7 = h,/bo
'IP Bracingmemberdepth to plate widthratio vP = h,/Bp
8, Included anglebetween bracing member i(i = 1 , 2, 3) and the chord
x Slenderness of
a
member under compression 187)
Reduction factor
for buckling curves [B71
li
Note: When mechanical or geometric propertiesof members in the Listof Symbols are used
in limit states design equations, or in conjunction with design charts, the nominal or
specified values are to be used.
93
Discuss me ...
10 References
CIDECT: Construction with hollow steel sections. British Steel plc, Corby, Northants,
England, 1984, ISBN 0-9510062-0-7.
International Instituteof Welding, Subcommission XV-E.: Design recommendations for
hollow section joints - Predominantly statically loaded. 2nd. ed., IIW Doc. XV-701-89,
International Institute of Welding Annual Assembly, Helsinki, Finland, September 1989.
131 Wardenier, J., and Packer,J. A.:Connections between hollow sections. Chapter 3.5 in
“Constructional Steel Design - An International Guide”, Elsevier, London, England,
1992.
[41 Mouty, J. (ed.): Effective lengthsof lattice girder members. CIDECT Monograph No.4,
Boulogne, France, April 1981.
151 Rondal, J.: Effective lengthsof tubular lattice girder members- statistical tests. CIDECT
Report 3K-88/9, August 1988.
161 Rondal, J.: Addendum to report 3K-88/9. CIDECT Report 3K-89/9, August 1989.
171 Dumonteil, P,: In-plane bucklingof trusses. Canadian Journalof Civil Engineering, Vol.
16, 1989, pp. 504-518.
PI Galambos, T.V. (ed.): Guide to stability design criteria for metal structures, 4th ed.,
Structural Stability Research Council, John Wiley& Sons, New York, 1988.
191 CIDECT:Bucklinglengths of HSSwebmembersweldedtoHSSchords.CIDECT
Programs3E-3G,SupplementaryReport - RevisedVersion,CIDECTDoc.80/3-E,
January 1980.
Sedlacek, G., Wardenier, J., Dutta, D., and Grotmann,D.: Evaluation of test results on
hollow section lattice girder connections in order to obtain strength functions and
suitable model factors. Background report to Eurocode 3: Common unified rules for
steel structures. Document 5.07, Eurocode3 Editorial Group, 1989.
Philiastides, A.: Fully overlapped rolled hollow section welded jointsin trusses. Ph. D.
Thesis, University of Nottingham, England, 1988.
N.: Testing of full-scale lattice
Coutie, M.G., Davies, G., Philiastides, A., and Yeomans,
girders fabricated with RHS members. Conference on Structural Assessment based on
Full and Large-scaleTesting, Building Research Station, Watford, England, April 1987.
11 31 Frater, G. S.: Performance of welded rectangular hollow structural section trusses.
Ph. D. Thesis, Universityof Toronto, Canada, 1991.
11 41 Czeckowski, A., Gasparski, T., Zycinski, J., and Brodka,J.: Investigation into thestatic
behaviour and strength
of lattice girders madeof RHS. InternationalInstitute of Welding
Doc. XV-562-84, Poland, 1984.
Frater, G.S., and Packer, J. A.: Design of fillet weldments for hollow structural section
trusses. CIDECT Report No. 5AN/2-90/7. Universityof Toronto, Canada, October 1990.
- Steel. ANSVAWS D l .l-92, 13th.
American Welding Society: Structural Welding Code
ed., Miami, Florida, U S A . , 1992.
Packer, J.A., andWardenier, J.: Designrulesforwelds in RHS K,T, Y and X
connections.
IIW
International
Conference
on
Engineering
Design in Welded
Constructions, Madrid, Spain, September 1992, Proceedings pp. 113-120.

Wardenier, J., and Stark, J. W. B.: The static strength of welded lattice girder joints in
structural hollow sections: Parts 1-10. CIDECT Final Report 50-78/4, Delft Universityof
Technology, The Netherlands, 1978.
94
Discuss me ...
1191 Wardenier, J.: Hollowsection joints. DelftUniversityPress, Delft, TheNetherlands,

1982.
1201 Reusink, J. H., and Wardenier, J.: Simplified design charts for axially loaded joints of
square
hollow
sections.3rd.
International
Symposium on
Tubular
Structures,
Lappeenranta, Finland, September 1989, Proceedings pp. 54-61.
[21] Eurocode 3: Design of steel structures: Part1. l - General rules and rules for buildings,
ENV 1993-1-1: 1992.
122) Mang, F., Bucak,O., and Wolfmuller, F.: The developmentof recommendations for the
design of welded joints between steel structural hollow sections
(T- and X-typejoints).
University of Karlsruhe,FederalRepublic of Germany,FinalReportonECSC
Agreement 7210SA1109 and CIDECT Program 5AD, December 1983.
1231 Zhao,X-L.,andHancock, G.J.: Tubular T-joints subject to combined actions. 10th.
InternationalSpecialtyConferenceonCold-FormedSteelStructures, St. Louis,
Missouri, U.S.A., October 1990, Proceedings pp. 545-573.
[24]Shinouda, M.R.: Stiffenedtubular joints. Ph.D.Thesis,University of Sheffield,
England, 1967.
[25] Dutta, D., and Wurker, K.: Handbuch Hohlprofile in Stahlkonstruktionen. Verlag TUV
Rheinland GmbH, Koln, Federal Republicof Germany, 1988, ISBN 3-88585-528-3.
[26] Stelco: Hollow structural sections- design manual for connections, 2nd ed., Stelco
Inc.,
Hamilton, Canada, 1981.
127) Korol, R. M,, Mitri, H., and Mirza,F. A.: Plate reinforced square hollow section T-joints
of
unequal width. Canadian Journal of Civil Engineering, Vol.9, No. 2, 1982, pp. 143-148.
128) Packer,J.A.,andFear, C.E.: Concrete-filledrectangularhollowsection X andT
connections.4th.
International
Symposium on
Tubular
Structures, Delft, The
Netherlands, June 1991, Proceedings pp. 382-391.
1291 ISO: Hot-rolledsteelsections - Part14:Hot-formedstructuralhollowsections -
Dimensions and sectional properties. Draft International Standard ISOlDlS 657-14,
International Organization for Standardization, Switzerland, 1991.
(301 ISO: StructuralSteels. I S 0 630-1980(E), 1st. ed.,InternationalOrganizationfor
Standardization, Switzerland, 1980.
1311 Packer,J.A.:Cranked-chordHSSconnections.Journal of StructuralEngineering,
American Society of Civil Engineers, Vol. 117, No. 8, August 1991,pp. 2224-2240.
1321 Kennedy, J. B.: Minimumbending radii forsquareandrectangularhollowsections
(3-roller cold bending). CIDECT Report l 1C-88/14-E, University of Windsor, Canada,
1988.
[33]Mang,F.:Investigation of standardboltedflangeconnectionsforcircularand
rectangularhollowsections.CIDECTReport8A-81/7-E,University of Karlsruhe,
Federal Republic of Germany, 1980.
1341 Kato, B., andMukai,A.:Boltedtensionflangessquarehollowsectionmembers
Supplement: Bolted at two sidesof flange. CIDECT Program 8B Report, University
of
Tokyo, Japan, 1985.
1351 Packer, J. A., Bruno, L., and Birkemoe, P. C.: Limit analysis
of bolted RHS flange plate
joints. Journal of Structural Engineering, American Society
of Civil Engineers, Vol. 115,
No. 9, September 1989, pp. 2226-2242.
[36] Birkemoe, P. C., and Packer, J. A.: Ultimate strength design of bolted tubular tension
connections. Conference on Steel Structures- Recent Research Advances and their
Applications to Design, Budva, Yugoslavia, 1986, Proceedings pp. 153-168.
95
Discuss me ...
[37] Struik, J. H. A., and de Back, J.: Tests on bolted T-stubs with respect to a bolted beam-
to-column connection. Stevin LaboratoryReport6-69-1
3,
Delft
University of
Technology, The Netherlands, 1969.
Bouwman, L. P,: Fatigue of bolted connections and bolts loaded in tension. Stevin
Laboratory Report 6-79-9, Delft Universityof Technology, The Netherlands, 1979.
Kato, B., and Mukai,A.: Bolted tension flanges joining square hollow section members.
CIDECT Report 8B-82/3-E, University of Tokyo, Japan, 1982.
Caravaggio,A.:Tests on steel roof joints forTorontoSkyDome. M.A.Sc. Thesis,
University of Toronto, Canada, 1988.
CSA: Limit states
designof
steel
structures,
CAN/CSA-S16.1-M89.
Canadian
Standards Association, Toronto, Canada, 1989.
Whitmore, R. E.: Experimental investigation of stresses In gusset plates. Bulletin 16,
University of Tennessee Engineering Experiment Station, U.S.A., 1952.
Duff, G.: Joint behaviour of a welded beam-column connection in rectangular hollow
sections. Ph. D. Thesis, The College of Aeronautics, Cranfield, England, 1963.

Redwood, R.G.: Thebehaviour of joints betweenrectangularhollow structural
members. Civil Engineering and Public Works Review, October 1965, pp. 1463-1469.
Cute, D., Camo, S., and Rumpf, J. L.: Welded connections for square and rectangular
structural steel tubing. Research Report No. 292-10, Drexel Institute of Technology,
Philadelphia, U.S.A., 1968.
Mehrotra, B. L., and Redwood, R.G.: Load transfer through connections between box
sections. Canadian Engineering Institute, C-70-BR and Str. 10, Aug./Sept. 1970.
Lazar, B. E.,and Fang, P. J.: T-type moment connections between rectangular tubular
sections. Research Report, Sir George Williams University, Montreal, Canada, 1971.
Mehrotra, B. L.,and Govil, A. K.: Shear lag analysis of rectangular full-width tube
connections. Journal of the Structural Division, American Society of Civil Engineers,
Vol. 98, No. ST1, January 1972, pp. 287-305.
Staples, C. J. L., and Harrison, C.
C.: Test results of 24
right-angled branches fabricated
from rectangular hollow sections. University of Manchester Institute of Science and
Technology (UMIST)Report, England.
Brockenbough, R. L.: Strength of square tube connections under combined loads.
Journal of the Structural Division, American Society of Civil Engineers, Vol. 98, No.
ST12, December 1972, pp. 2753-2768.
Korol, R.M., El-Zanaty, M,, and Brady, F. J.: Unequal width connections of square
hollow sections in vierendeel trusses. Canadian Journal of Civil Engineering, Vol. 4,
1977, pp. 190-201.
Korol, R. M,, and Mansour, M.H.: Theoretical analysis of haunch-reinforced T-joints in
square hollow sections. Canadian Journal of Civil Engineering, Vol. 6, No. 4, 1979, pp.
601 -609.
Yeomans, N. F. and Giddings, T. W.: The design offull width joints in RHS Vierendeel
girders. Report No. 800/0/72, British Steel plc., Corby, England, November 1988.
Kanatani, H., Fujiwara, K., Tabuchi, M., and Kamba, T.: Bending tests on T-joints of
RHS chord and RHS or H-shape branch, CIDECT Report 5AF-80/15, 1980.
Korol, R. M., and Mirza, F. A.: Finite element analysis of RHS T-joints. Journal of the
Structural Division, American Society of CivilEngineers, Vol. 108, No. ST9, September
1982, pp. 2081-2098.
96
Discuss me ...
[56] Davies, G., and Panjehshahi, E.: Tee joints in rectangular hollow sections (RHS) under
combined axial loading and bending. 7th International Symposium on Steel Structures,
Gdansk, Poland, 1984.
157) Szlendak, J., and Brodka, J.: Strengtheningof T momentof RHS joints. Proceedings of
the Institution of Civil Engineers, Part2, Vol. 79, December 1985, pp. 717-727.
1581 Szlendak,J.,andBrodka, J.: Design of strengthenframe RHS joints. International
Meeting on Safety Criteria in Design of Tubular Structures, Tokyo, Japan, July 1986,
Proceedings pp. 159-168.
1591 Szlendak, J., andBrodka, J.: Replytodiscussionby J.A. Packeronthepaper,
“Strengthening of T-moment of RHS joints”. Proceedings of the Institution of Civil
Engineers, Part 2, Vol. 81, December 1986, pp. 721-725.
[60] Szlendak, J.: Interaction curves for M-N loaded T RHS joints. International Meeting on
Safety Criteria in Design of Tubular Structures, Tokyo, Japan, July 1986, Proceedings
pp. 169-1 74.
1611 Szlendak, J.: Beam-Column weldedRHS

connections.
Journal of Thin-Walled
Structures, Vol. 12, 1991, pp. 63-80.
[62] Kanatani, H., Kamba, T., and Tabuchi, M.: Effect
of the local deformationof the jointson
RHS vierendeel trusses. International Meeting on Safety Criteria
in Design of Tubular
Structures, Tokyo, Japan, July 1986, Proceedings pp. 127-137.
[63] De Koning, C. H. M., and Wardenier, J.: The static strength of welded joints between
structural hollow sections or between structural hollow sections and H-sections. Part 2:
Joints between rectangular hollow sections. Stevin Report 6-84-19. Delft Universityof
Technology, Delft, The Netherlands, March 1985.
1641 Horne, M. R., andMorris, L. J.: Plasticdesign of low-riseframes.Collins,London,
Chapter 3, 1985.
1651 Mang,F.,Steidl,G.,andBucak, 0.: Design of welded lattice jointsandmoment
resisting knee joints made of hollow sections. International lnsitute of Welding Doc.
XV-463-80, University of Karlsruhe, Federal Republicof Germany, 1980.
1661 Deutsches lnstitut furNormung:Stahlbauten:TragwerkeausHohlprofilenunter
vorwiegend ruhender Beanspruchung. (Steel structures: made from hollow sections
subjected to predominantly static loading.)DIN 18808, Berlin, Federal Republic of
Germany, October 1984.
167) Wardenier, J., Kurobane, Y., Packer, J.A., Dutta, D., and Yeomans, N.: Design guide
for circular hollow section (CHS) joints under predominantly static loading. CIDECT an
Verlag TUV Rheinland, Koln, Federal Republic of Germany, 1991, ISBN 3-88585-975-0.
[68] Coutie, M. G., Davies G., Bettison, M,, and Platt, J.: Development of recommendations
for the design of welded joints between steel structural hollow sections or between steel
structural hollow sectionsand H sections. Final Report,3:part Threedimensional joints.
Report on ECSC Contract 7210.SN814, Universityof Nottingham, England 1983.
1691 Bauer, D., and Redwood, R. G.: Triangular truss joints using rectangular tubes. Journal
ofStructuralEngineering,AmericanSociety of Civil Engineers, Vol. 114,No.2,
February 1988, pp. 408-424.
[70] Davies, G., and Morita, K.: Three dimensional cross joints under combined axial branch
loading, 4th. International Symposium on Tubular Structures,Delft, The Netherlands,
June 1991, Proceedings pp. 324-333.

[71] Ono, T., Iwata, M,, and Ishida, K.: An experimental study on jointsof new truss system
using rectangular hollow sections. 4th International Symposium on Tubular Structures,
Delft, The Netherlands, June 1991, Proceedings pp. 344-353.
97
Discuss me ...
[721 Grundy, P,, and Foo EeKiu, J.: Performance of flattenedtubeconnections.4th

International Symposium on Tubular Structures, Delft, The Netherlands, June 1991,
Proceedings pp. 251-258.
Papanikolas, P. C., Nowak,J. A., and Kulak, G.L.: Flexible flattened HSS connections.
Report by Northern SteelInc., Edmonton, Canada, November1990.
Thiensiripipat, N., Morris, G. A., and Pinkney, R. B.: Statical behavior of cropped-web
tubular truss joints. Canadian Journal of Civil Engineering, Vol. 7,No. 3, 1980,pp.
51 3-522.
Ghosh,A.,andMorris, G.: Behavior of tubularsteeltrusseswithcroppedwebs.
Canadian Journal of Civil Engineering, Vol. 8,No. 1 , 1981,pp. 51-58.
Lau, B. L., Morris, G.A., andPinkney, R. B.: Testing of Warren-typecropped-web
tubulartruss joints. CanadianSocietyfor Civil EngineeringAnnualConference,
Saskatoon, Canada, May1985.
Morris, G.A., and Packer, J.A.: Yield line analysis of cropped-web Warren truss joints.
Journal of StructuralEngineering,AmericanSociety of Civil Engineers,Vol. 114,
No. 10,October 1988,pp. 2210-2224.

Korol, R. M. andChidiac,M. A.: K-joints of double-chordsquarehollowsections.
Canadian Journal of Civil Engineering, Vol. 7,No. 3, 1980,pp. 523-539.
Korol, R.M., Mirza, F.A., and Chiu, E.T.C.: An experimental investigation of double-
chord HSS trusses. Canadian Journalof Civil Engineering, Vol.10, pp. 248-260.
1983,
Korol, R. M,:The behaviourof HSS double chord Warren trusses and aspects
of design.
Canadian Symposium onHSS, CIDECT, April 1983.
Korol, R. M.,and Mitri, H.S.: Strengthanalysis of RHS doublechordjointswith
separated chords. Canadian Journal of Civil Engineering, Vol. 12,No. 2, 1985,
pp.
370-381.
Luft, R.T.: DCTRUSS - A computer program to aid in the analysis and design of
separated HSS double chord Warren trusses.
Civil Engineering Department, McMaster
University, Hamilton, Canada, 1990.
1831 Luft, R. T., Korol,R. M., and Huitema,H.: An economic comparisonof single chord and
double chordRHS Warren trusses. 4th International Symposium on Tubular Structures,
Delft, The Netherlands, June 1991,Proceedings pp. 11-20.
1841 Davies, G., and Packer, of branch plate- RHS connections
J. A.: Predicting the strength
for punching shear. Canadian Journal of Civil Engineering, Vol. 9,No. 3, 1982,pp.
458-467.
Wardenier, J., Davies G., and Stolle, P.: The effective width of branch plate to RHS
chord connectionsin cross joints. Stevin Laboratory Report No. 6-81-6,Delft University
of Technology, Delft, The Netherlands, 1981.
Giddings, T. W., and Wardenier, J. (eds.): The strength and behaviour of statically
loaded welded connections in structural hollow sections. CIDECT Monograph No. 6,
British Steel plc, Corby, Northants, England, 1986.
Rondal, J., Wurker, K. G., Wardenier, J., Dutta, D., Yeomans, N.: Structural stability of
hollow sections, CtDECT and Verlag TUV Rheinland, Cologne, Federal Republic of
Germany, 1991,ISBN 3-8249-0075-0.
98
Discuss me ...
Tree top connections
Acknowledgements for photographs:

Theauthorsexpresstheirappreciationtothefollowingfirmsformakingavailablethe
photographs usedin this Design Guide:
British Steel
Mannesmannrohr GmbH
Nippon Steel Metal Products
Rautaruukki
Mannstaedt Werke
SIDERCAD
Tubeurop
Voest Alpine Krems
Discuss me ...
p,
Cornite International pour le Developpernent et I’Etudelade
Construction Tubulaire
International Committee
for the Development and Study
of Tubular Structures
CIDECT, founded in1962 as an international association, joins together the research
resources of major hollow steel section manufacturers to create a major force in th
research and applicationof hollow steel sections worldwide.
The objectives of CIDECT are:

0 to increase knowledgeof hollow steel sections and their potential application by
initiating and participating in appropriate researches and studies
0 to establish and maintain contacts and exchanges between the producers

of the
hollow steel sections and the ever increasing ofnumber
architects and engineers
using hollow steel sections throughout the world.
0 to promote hollow steel section usage wherever this makes for good engineering
practiceandsuitablearchitecture,ingeneralbydisseminatinginformation,
organizing congresses etc.
0 to co-operate with organizations concerned with practical design recommen-
dations, regulations or standards at the national and international level.
Technical activities
The technical activities
of CIDECT have centred on the following research aspects
of
hollow steel section design:
0 Buckling behaviour of empty and concrete-filled columns
G Effective buckling lengths of members in trusses
0 Fire resistance of concrete-filled columns
0 Static strength of welded and bolted joints
0 Fatigue resistance of joints
0 Aerodynamic properties
0 Bending strength
0 Corrosion resistance
0 Workshop fabrication
The results of CIDECT research form the basis of many national and international
design requirements for hollow steel sections.
100
Discuss me ...
CIDECT, the future

Current work is chiefly aimed at filling up the gaps in knowledge regarding the
structuralbehaviour of hollowsteelsectionsandtheinterpretationandimple-
mentationofthecompletedfundamentalresearch.Asthisproceeds,anew
complementary phase is opening that will be directly concerned with practical,
economical and labour saving design.
CIDECT Publications
The current situation relating to CIDECT publications reflects the ever increasing
emphasis on the disseminationof research results.
Apart from the final reportsof the CIDECT sponsored research programmes, which
are available at the Technical Secretariat on demand
at nominal price, CIDECT has
published a number of monographs concerning various aspects of design with
hollowsteelsections.TheseareavailableinEnglish,FrenchandGermanas
indicated.
-
Monograph No. 3 Windloads for Lattice Structures (G)
-
Monograph No. 4 Effective Lengths of Lattice Girder Members (E, F, G)
Monograph No.5 - Concrete-filled Hollow Section Columns (E, F)
Monograph No. 6 - TheStrengthandBehaviour of StaticallyLoadedWelded
Connections in Structural Hollow Sections (E)
-
Monograph No. 7 Fatigue Behaviour of Hollow Section Joints (E, G)
A book “Construction with Hollow Steel Sections”, prepared under the direction of
CIDECT in English, French, German and Spanish, was published with the sponsor-
ship of the European Community presenting the actual state of the knowledge
acquired throughout the world with regard to hollow steel sections and the design
methods and application technologies related to them.
Inaddition,copies of thesepublicationscanbeobtainedfromtheindividual
members given belowto whom technical questions relating to CIDECT workor the
design using hollow steel sections should be addressed.
The organization of CIDECT comprises:

0 President: J. Chabanier (France)
Vice-president: C. L. Bijl (The Netherlands)
7 A General Assembly of all members meeting once a year and appointing an
Executive Committee responsible

for administration and execution
of estabished
policy
0 Technical Commission and Working Groups meeting at least once a year and
directly responsiblefor the research and technical promotion work
101
Discuss me ...
0 Secretariat in Paris responsible for the day to day running

of the organization.
Present membersof CIDECT are:

(1 996)
0 British Steel PLC, United Kingdom

G CS1 Transformados S.A., Spain
0 EXMA,France
0 ILVA Form, Italy
0 IPSCO Inc., Canada
0 Laminoirs de Longtain, Belgium
0 Mannesmannrohr GmbH, Federal Republicof Germany
0 Mannstadt Werke GmbH, Federal Republicof Germany
0 Nippon Steel Metal Products Co. Ltd., Japan

0 Rautaruukki Oy, Finland
0 Sonnichsen AIS, Norway
0 Tubemakers of Australia, Australia
0 Tubeurop, France
0 VOEST Alpine Krems, Austria
Cidect Research Reports can be obtained through:

Mr. E. Bollinger
Office of the Chairman of the CIDECT Technical Commission
c/o Tubeurop France
lmmeuble Pacific
TSA 20002
92070 La Defense Cedex
Telephone: (33)1 /41258181
Telefax: (33) 2/41 258800
Mr. D. Dutta
Marggrafstrasse 13
40878 Ratingen
Germany
Telephone: (49) 21 02/842578
Telefax: (49) 21 02/842578
Carehasbeentakentoensurethat all dataandinformationhereinisfactualandthat

numerical values are accurate. To the best of our knowledge, all information in this book is
accurate at the timeof publication.
CIDECT, its members andthe authors assume no responsibility for errors or misinterpretation
of the information containedin this book or in its use.
102

Design Guide For Rectangular Hollow Section Joints

Uploaded by

Document Information

Original Title

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

Design Guide For Rectangular Hollow Section Joints

Uploaded by

Copyright:

Available Formats

Construction with Hollow Steel Sections - Design guide for rectangular hollow section (RHS) joints under predominantly

Edited by: Comite Internationalpour le Developpementet I’Etude

Jaap Wardenier, Delft Universityof Technology

Verlag TUV Rheinland

Die Deutsche Bibliothek - CIP Einheitsaufnahme

Design guide for rectangularhollow section

0 by Verlag TUV Rheinland GmbH, Cologne 1996, Reprinted edition

2.6 Trussdesignprocedure .......................................... 18

3 Weldeduniplanartrussconnections between RHS chordsandRHSor

7 Multiplanar welded connections . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 80

CIDECT . International Committee for the Development and Study of Tubular

Roof support structure for a swimming centre

1 General:Advantages of hollow structuralSections,

(cl Partial overlap connection ( d ) Total overlap connection

Aircraft museum; RHS columns and roof girders

Roof consjruction for a customs station

especially for cold-formed sections a flare bevelwith a gaprequiring a customfitted backing

Container lifting vehicle

2 Design of tubular structures

This chapter deals with

Fig. 2 - Common RHS trusses

connected. Nodal eccentricities between the centrelines of intersecting members at panel

column.This is doneby distributingthe panel point moment(sumofthehorizontal

Fig. 3 - Plane frame joint modellingassumptions to obtain realisticforces for memberdesign

Table 1 - When moments need to be consideredfor RHS truss design

moments nodal eccentricity transverse member secondary effects

compression chord Yes no

design of connections no, provided yes, influences f (n) no, provided

2.3 Effective lengths for compression members

For RHS chord members:

2.3.2 Empirical method for bracing members

For RHS chord members,bois replaced by h, when h, < bo.

2.3.3 Long, laterally unsupported compression chords

2.5 General connection considerations

reinforcement.Apparenteconomiesfromminimum-massmemberselection will quickly

this can present an awkward the cornerof the chord

In summary,the designof an RHS truss should be approached

moments in thechords by assumingeither: (a) pinned jointseverywhereor (b)

2.7 Design of welds

Effective length = (2 h,/sin 0,) + b,, for 0, 2 60°

Retractable arched roof structureof the Toronto SkyDorne

Fig. 4 - Recommended weld details (corner

3 Welded uniplanartrussconnections between RHS

are given in the List of Symbols.

Ad2 Connection shown d

Fig. 5 -Typical RHS chord truss K-connection

Note that a negative value

LRFD = Load and resistance factor design

3.1 K and N connections

Mode B: Punchlng shear fallure

of the chord face

Mode A: Flexural fallure of

Mode D. Local buckllng of

____ Mode F: Local

Mode E Overall shear failure

Mode G Local buckllng of

Fig. 6 - Failure modes forK and N-typeRHS truss connections

Mode D: Local buckling of the compression bracing member.

type of connection factored connection resistance (i = 1.2)

T-. Y - and X-connections S 4 0.85 yielding

K- and N-gap connections Y I1.0 plastification

K- and N-overlap connections' 25% I0, < 50% basis:

50% 5 0, < 80% basis: