Cgna14427enc 001

eu* w$
*w *
Commission of the European Communities
technical steel researcl

Properties and service performance
Semi-rigid action in steel frame structures
_s
Commission of the European Communities
Properties and service performance
Semi-rigid action in steel frame structures

J.-C. Gerardy
Project manager
J.-B. Schleich
Ingnieur principal
Service recherches et promotion technique structures

Arbed-Recherches
66, rue de Luxembourg
L-4002 Esch/Alzette
Contract No 7210-SA/507
(1.7.1987-31.12.1990)
Final report
Directorate-General
Science, Research and Development
1992
PARI EUROP. Biblioth.
n.c.tvfiiwi?ci.
CUR 14427 I N
Published by the
COMMISSION OF THE EUROPEAN COMMUNITIES
Directorate-General
Information Technologies and Industries, and Telecommunications
L-2920 Luxembourg
LEGAL NOTICE
Neither the Commission of the European Communities nor any person
acting on behalf of the Commission is responsible for the use which might
be made of the following information
Cataloguing data can be found at the end of this publication
Luxembourg: Office for Official Publications of the European Communities, 1992

ISBN 92-826-2926-0
ECSC-EEC-EAEC, Brussels Luxembourg, 1992
Printed in Belgium
TITLE OF RESEARCH
AGREEMENT N
SEMI - RIGID ACTION IN STEEL

FRAME STRUCTURES
7210-SA/507
EXECUTIVE COMMITTEE
Light Structures F6
BEGINNING OF RESEARCH
01.07.1987
SCHEDULED COMPLETION DATE
31.12.1990
BENEFICIARY
ARBED-Luxembourg
TECHNICAL SUPPORT
University of Lige
M.S.M. - Institut de Gnie Civil
Mr. J-P Jaspart, Mr R. Maquoi
Acknowledgements.
This research consisting in the analysis of the semi-rigid action in steel frame structures
has been performed by ARBED SA. during the years 1987 to 1990 and sponsored by
C.E.C., the Commission of the European Community (C.E.C. Agreement N 7210SA/507).
We want to acknowledge first of all the important financial support from the
COMMISSION OF THE EUROPEAN COMMUNITY, as well as the moral support
given this research by all the members of the C.E.C. EXECUTIVE COMMITTEE F6
"LIGHT STRUCTURES".
Special thanks are due to the Dr. Ir. J.P. JASPART, Assistant in the Department MSM
of the Lige University (Belgium), for his technical support as well as to the Professor
Dr. Ir. R. MAQUOI, of the Department MSM of the Lige University (Belgium).
Thanks also to the technical staff of the laboratory of this Department which have
executed successfully the 56 tests.
Thanks are finally due to all, who by any means may have contributed to this research
program, as for instance C. KERN for the dactylography and Y. CONAN for the layout
and the programmation.
- V-
Semi-Rjgjd Action in Steel F r a m e s Structures.

Ceca agreement 7210-SA/507.
Summary.
This research has shown that the cost of a frame structure could be reduced by
taking into account the semi-rigid behaviour of the joints between the beams and the
columns.
In the first part of this research, 18 steel joints and 38 composite joints have
been tested at the Laboratory of the University of Lige in Belgium. These tests have
permitted to determine all the components of deformability of a semi-rigid joint with
cleated connection. All the tests are stored in a databank named SERICON (SEmiRlgid CONnection) developped at the University of Aachen (Germany) and distributed
with this final report.
On basis of these tests results, a mathematical model of the behaviour of these
joints has been developped. This model is able to calculate the moment - rotation curve
of a composite joint with cleated connection. The program of this model is distributed
with this final report.
In the numerical part of this research, a new finite element taking into account
the actual behaviour of the joint (connection deformability and shear deformability) has
been introduced in the non-linear finite elements program FINELG.
This new element has been used in the frame of a parametric study of braced
and unbraced structures with the aim to find in which conditions the joint deformability
may be concentrated in a single flexural spring acting at this end of the beams.
The philosophy of simplified methods to design steel structures with semi-rigid
joints has been presented for braced and unbraced frames. These developments bring a
starting point to a new practical document for designers.
Lastly, a cost comparison between semi-rigid and traditional design has shown
the potential benefits of the semi-rigid design of building frames.
VII
Action Semi-Rigide dans les Structures en Acier.

Ceca agreement 7210-SA/507.
Sommaire.
Cette recherche montre que le cot d'une structure mtallique peut tre rduit
par la prise en compte du comportement semi-rigide des joints entre les poutres et les
colonnes.
Dans la premire partie de cette recherche, 18 joints tout acier et 38 joints
mixtes furent tests dans les laboratoires de l'Universit de Lige en Belgique. Ces tests
ont permis de dterminer toutes les composantes de dformabilit d'un joint semi-rigide
ralis au moyen de cornires. Tous les tests sont rpertoris dans la banque de donnes
SERICON (SEmi Rigid CONnection) programme l'Universit d'Aix-La-Chapelle eft
Allemagne et distribue avec le rapport final.
Les rsultats des tests ont permis d'laborer un modle mathmatique du
comportement de ces joints. Ce modle calcule la courbe moment-rotation d'un joint
mixte avec cornires. Le programme conu partir du modle est distribu avec le
rapport final.
Dans la partie numrique de cette recherche, un nouvel lment fini prenant en
compte le comportement rel du joint (dformabilit d'assemblage et de cisaillement) a
t ajout au programme d'lments finis non-linaire FINELG.
Ce nouvel lment est utilis dans le cadre d'une tude paramtrique de
structures contreventes et non-contreventes dans le but de trouver dans quelles
conditions la dformabilit du joint peut tre substitue par un ressort flexionnel simple
agissant aux extrmits des poutres.
La philosophie de la mthode simplifie du calcul des structures en acier avec
joints semi-rigides est prsente pour des structures contreventes et non-contreventes.
Ces dveloppements servent de base un nouveau document pratique pour les bureaux
d'tudes.
Finalement, une comparaison de cot entre les calculs traditionnels et semirigides a mis en vidence les avantages potentiels du calcul semi-rigide des structures.
IX
Halhsteife Auschlsse im Stahlbau.

Ceca agreement 7210-S A/507
Zuzammenfassung.
Dieses Forschungsvorhaben zeigt, dass der Gestehungspreis einer
Stahlkonstruktion durch die Bercksichtigung eines halbstreifen Verhaltens der
Ansschlsse zwischen Sttzen und Riegel reduziert werden kann.
Im ersten Teil dieses Forschungsvorhabens wurde das Vorhalten von 18
Ansschlssen in reiner Stahlbauweise und 38 Ansschssen in Verbundbauweise an der
Versuchsanstalt der Universitt Lttich (B) durch Versuche getestet Diese Versuche
haben es ermglicht alle Verformungskomponenten eines halbsteifen Anschlusses zu
definieren. Alle Versuchsergebnisse wurden in einer Datenbank SERICON (SEmi Rigid
CONnection) aufgenommen. Diese Datenbank wurde an der RWTH-Aachen entwickelt,
und soll mit Endbericht verteilt werden.
Die Ergebnisse des Forschungsvorhabens haben es erlaubt, ein mathematisches
Modell auszuarbeiten, welches die Momenten-Verdrehungs-Beziehung eines Anschlusses
mit Winkeln in Verbundbauweise hergibt. Das Programm, das diesem Modell zugrunde
liegt, wird mit dem Endbericht verteilt werden.
Im numerischen Teil dieses Forschungsvorhabens, wurde ein neuentwickeltes
finite Element dem FEM-Programm FINELG hinzugefgt. Es bercksichtigt das wahre
Verhalten des Anschlusses (Verformung des Anschlusses und die Scher). Dieses neue
Element wurde fr Parameterstudien an ausgesteiften und unausgesteiften Strukturen
benutzt. Ziel war es herauszufinden, unter welchen Bedingungen die Verformbarkeit des
Anschlusses durch eine einfachen Biegefeder, welche an den Endpunkten der Riegel
angreift, erstetzt werden kann.
Die Philosophie der vereinfachten Berechnungsmethode fr Stahlbauten mit
halbsteifen Anschlssen wird aufgezeigt fr ausgesteifte und unausgesteifte Strukturen.
Diese Entwicklungen stellen die Grundlage fr ein neues Bemessungshandbuch, welches
fr die Ingenieur-Bros bestimmt ist, dar.
Zum Schluss hat ein Preisvergleich zwischen der traditionellen und der
halbsteifen Berechnungsmethode eindeutig die potentiellen Vorteile der halbsteifen
Ausfhrungen dargelegt
-XI
CONTENTS.
Summary.
1. Introduction.
2. Experimental investigations.
2.1. Types of tested beam-to-column joints.

2.2. Parameters influencing the detailed choice of the connections.
2.3. Testing arrangement.
2.3.1. Components of the joint deformability.
2.3.2. Test specimens and testing arrangements.
2.4. Instrumentation and measurements.
2.5. Data acquisition.
2.6. Experimental characteristic curves.
2.7. Interpretation of the tests results.
2.7.1. Steel "T" tests.
2.7.1.1. General description of the results.
2.7.1.2. Influence of the chosen test parameters
2.7.2. Composite cruciform tests.
2.7.2.2. Influence of the chosen test parameters.
2.7.3. Steel cruciform tests.
2.8. Data sheets.
9
12
16
16
17
19
23
23
25
25
25
26
29
29
31
37
37
39
40
3. Theoretical developments.
41
3.1. Behaviour of column web panels.

3.1.1. Column web panels in shear.
3.1.2. Column web panels subject to load-introduction.
3.2. Behaviour of steel cleated connections.
3.3. Improvement and extending of Eurocode 3 rules.
3.3.1. Column web panels subject to shear and to transverse compression and
tensile forces.
3.3.2. Flange cleated connections.
3.3.2.1. Plastic capacity.
3.3.2.2. Secant stiffness.
3.3.2.3. Comparisons between proposed bi-linear model and experimental
results.
3.3.2.4. Conclusion.
3.3.3. Flange and web cleated connections.
3.4. Behaviour of composite connections.
3.4.1. General description of the model.
3.4.2. Characterization of the non-linear springs.
41
42
47
53
57
58
4. Numerical developments.
69
4.1. Description of a new finite element.

4.2. Description of the parametric study.
4.3. Conclusions of the parametric study.
69
73
79
XIII
61
61
61
62
63
63
65
65
67
5. Simplified methods for the design of steel structures with

semi-rigid joints.
81
5.1. Simplifying assumptions.

5.1.1. Behaviour of steel.
5.1.2. Behaviour of beam and column cross-sections.
5.1.3. Behaviour of beam-to-column joints.
5.2. Design requirements.
5.3. Simplified design methods for braced frames.
5.3.1. Elastic design.
5.3.1.1. Design principles.
5.3.1.2. Buckling length of linearly end-restrained columns.
5.3.1.3. Second-order effects.
5.3.1.4. Examples of application.
5.3.2. Plastic design.
5.3.3. Validity of the bi-linear modeling of the joint response.
5.4. Simplified design method for unbraced frames.
5.4.1. Design under service loads.
5.4.2. Design under factored loads.
5.4.2.1. Assessment of the critical multiplier XCT.
5.4.2.2. Assessment of the plastic multiplier Xp.
5.4.2.3. Assessment of the collapse multiplier Xf.
5.4.3. Examples of application.
81
81
82
83
84
85
85
85
85
87
88
90
91
94
94
. 95
96
97
99
101
6. Benefit of semi-rigid design.
103
6.1. Unbraced frame.

6.1.1. Design in case ofrigidjoints.
6.1.2. Design in case of semi-rigid joints.
6.1.3. Cost comparison
6.2. Braced frame.
6.2.1. Design in case of pinned connections at each end of the beam.
6.2.2. Design in case ofrigidconnections
6.2.3. Design in case of semi-rigid connections
6.2.4. Cost comparison.
6.3. Conclusion of sections 6.1 and 6.2.
6.4. Need for further studies.
105
106
107
107
108
109
109
110
1 io
Ill
Ill
7. Conclusions.
113
7.1. Frames with steel connections

7.2. Frames with composite connection
7.3. Databank SERICON and program SPRINGS
7.4. Benefit of the semi-rigid design
113
114
117
117
REFERENCES.
119
XIV
Annex 1.
Ph.D. Thesis of JASPART, J.P., "Etude de la semi-rigidit des noeuds poutre-colonne
et de son influence sur la rsistance et la stabilit des ossatures en acier".
125
Annex 2.
Numerical simulations of frames by FINELG program.
129
Annex 3.
Mathematical model for the prediction of the moment-rotation curves of composite
joints with cleated connections.
Floppy disc with the program SPRINGS.
151
Annex 4.
Databank SERICON (SEmi-RIgid CONnection) and data sheets of the 56 tests.
Floppy disc with the program SERICON.
XV
233
TABLE DES MATIERES

Sommaire.
1. Introduction.
2. Etudes exprimentales.
2.1.
Type des joints poutre-colonne tests.
2.2.
Choix des diffrents paramtres des assemblages.
2.3.
Procdures d'essai.
2.3.1. Composantes de dformabilit du joint.
2.3.2. Descriptifs et procdures des essais.
2.4.
Appareillage et mesures.
2.5.
Acquisition des donnes.
2.6.
Courbes exprimentales caractristiques.
2.7.
Interprtation des rsultats.
2.7.1. Tests en "T" tout acier.
2.7.1.1. Description gnrale des rsultats.
2.7.1.2. Influence des paramtres.
2.7.2. Tests cruciformes mixtes .
2.7.3. Tests cruciformes tout acier.
2.8. Tableau des donnes.
9
12
16
16
17
19
23
23
25
25
25
26
29
29
31
37
37
39
40
3. Dveloppements thoriques .
41
3.1.
3.1.1.
3.1.2.
3.2.
3.3.
3.3.1.
Comportement du panneau d'me de la colonne.

Cisaillement dans le panneau d'me de la colonne.
Introduction des charges dans le panneau d'me de la colonne.
Comportement des assemblages avec cornires tout acier.
Amlioration et extension de 1' Eurocode 3 .
Cisaillement, compression transversale et efforts de traction dans
le panneau d'me de la colonne
3.3.2. Assemblages des semelles par cornires.
3.3.2.1. Capacit plastique.
3.3.2.2. Rigidit scante.
3.3.2.3. Comparaisons entre le modle bi-lineaire propos et les rsultats
exprimentaux.
3.3.3. Assemblages des semelles et des mes avec cornires.
3.4.
Comportement des assemblages mixtes.
3.4.1. Description gnrale du modle.
3.4.2. Caractrisation des ressorts non-linaires.
41
42
47
53
57
4. Dveloppements numriques.
69
4.1.
4.2.
4.3.
69
73
79
Description d'un nouvel lment fini.

Description de l'tude paramtrique.
Conclusions de l'tude paramtrique.
XVII
58
61
61
61
62
63
63
65
65
67
5. Mthodes simplifies pour la conception des structures en acier 81

avec des joints semi-rigides.
5.1.
Hypothses simplificatrices.
5.1.1. Comportement de l'acier.
5.1.2. Comportement des sections transversales des poutres et colonnes.
5.1.3. Comportement des joints poutre-colonne.
5.2.
Rgles de dimmensionement.
5.3.
Mthodes simplifies de calcul pour structures contreventes.
5.3.1. Calcul lastique.
5.3.1.1. Principes de calcul.
5.3.1.2. Longueurs de flambement des colonnes restraintes
linairement aux extrmits.
5.3.1.3. Effets du second ordre.
5.3.1.4. Exemples d'application.
5.3.2. Calcul plastique.
5.3.3. Validit du modle bi-linaire sur le comportement des j oints.
5.4.
Mthodes simplifies de calcul pour structures non-contreventes.
5.4.1. Calcul sous charges de service.
5.4.2. Calcul sous charges pondres.
5.4.2.1. Dtermination du coefficient critique Xcr.
5.4.2.2. Dtermination du coefficient plastique Xp.
5.4.2.3. Dtermination du coefficient de ruine Xf.
5.4.3. Exemples d'application.
85
87
88
90
91
94
94
95
96
97
99
101
6. Avantages du calcul semi-rigide.
103
6.1.
6.1.1.
6.1.2.
6.1.3.
6.2.
6.2.1.
6.2.2.
6.2.3.
6.2.4.
6.3.
6.4.
105
106
107
107
108
109
109
110
110
111
111
Structures non-contreventes.
Calcul des joints
rigides.
Calcul des joints semi-rigides.
Comparaison des cots.
Structures contreventes.
Calcul des joints rotules.
Calcul des joints rigides.
Calcul des joints semi-rigides.
Comparaison des cots.
Conclusion des paragraphes 6.1 et 6.2.
Etudes futures complmentaires.
81
81
82
83
84
85
85
85
7. Conclusions.
113
7.1.
7.2.
7.3.
7.4.
113
114
117
117
Structures compose d'assemblages en acier.

Structures compose d'assemblages mixtes.
Banque de donnes SERICON et programme SPRINGS
Avantages du calcul semi-rigide.
REFERENCES.
119
-XVIII
Annexe 1.
Thse de doctorat de JASPART, J.P., "Etude de la semi-rigidit des noeuds poutrecolonne et de son influence sur la rsistance et la stabilit des ossatures en acier".
125
Annexe2.
Simulations numriques de structures par le programme FINELG.
129
Annexe 3.
Modle mathmatique pour la prdiction de la courbe moment-rotation de joints mixtes
avec cornires.
Disquette contenant le programme SPRINGS.
151
Annexe 4.
Banque de donnes SERICON (SEmi-RIgid CONnection) et impressions des donnes
des 56 tests.
Disquette contenant le programmeSERICON.
-XIX
233
1. Introduction.
During the last decade semi-rigid joints has focused the attention of several European
researchers, regarding not only their actual behaviour but also on how and to which
extent they alter the stability of frameworks. That is testified by the number of papers
dealing with this subject presented each year in scientific journals or on the occasion of
colloquiums. At the same time, ad-hoc working groups of ECCS were launched in 1984
and in 1988, aimed at investigating respectively the structural response of semi-rigid
frames and the local properties of joints in frames [32]. Lastly Eurocode 3 is paving the
way to "semi-rigid" construction, in contrast to national standards and design codes
which have perpetuated the alternative of either "simple" or "continuous" construction.
As a matter of fact, the current methods of structural analyses allow only either for rigid
joints "continuous construction or for pin-ended connections "simple construction"; the
detailing of the joints is made in view to comply as much as possible with the
corresponding assumptions.
A rigid beam-to-column joint should not allow for any relative rotation between the axis
of the respective intersecting members; as a consequence, any external bending moment
applied on this joint distributes amongst the connected members according to the member
flexural stiffness. In contrast, a beam that is pin-ended to columns should freely rotate at
the ends without transferring any bending to the column; this rotation at the ends is
depending on the beam loading only.
Such extreme behaviours are respectively represented, in a bending - relative rotation
curve, by the axis of abscissoe for a pin-ended beam-to-column connection.
Of course no joint is never either fully rigid or actually pinned. More especially, a
presumed rigid joint always allows for a relative rotation, even when the column web
panel is transversely and diagonally stiffened and the connection of the beam made with
stiff end-plate and preloaded bolts; however the rotation of such a joint remains usually
so small that the assumption of a rigid joint is quite acceptable. Presumed pin-ended
connections, for their own, are always somewhat restrained because the actual connecting
devices are never proper hinges; the corresponding constructive detailing is identified to
hinges for sake of conservatism. Let us mention in addition that it is distinguished
between full strength and partial strength joints; in the first case the joint is able to
transmit the full resistance of the connected members, while in the second one, the
connection is weaker than these members.
Rigourously speaking, any joint is ever neither fully rigid, nor pinned; therefore it should
be termed as semi-rigid.
In some cases, semi-rigid joints should favour a better balance of end- and span
moments, with the possible result of lighter beams to columns should contribute to a
decrease in effective length and result in less conservative methods of column design.
Modern steel construction aims, however at not only material savings but more especially
at cost savings. In this respect, the detailing of the joints is simplified: the number of
stiffeners in the joint is reduced, as well as the length of fillet welds; that results in an
appreciable decrease in labour cost and consequently favours the global economy of the
project At the structural viewpoint, that leads to beam-to-column joints which are much
more simple to execute but exhibit a fully non-linear behaviour and therefore a M-<|>
curve, which is sometimes very far from the characteristic curve of arigidjoint (Figure
1.1
Extended end plate
Pinned
Figure 1.1. - M- $ curves for different types of joints.
Incorporation of actual jointflexibilityin the routine design practice implicates the use of
microcomputers and appropriate software packages. Indeed the joint behaviour is
non-linear, so is the response of frames too. Except for research work, it cannot be
expected to account for this non-linearity with a great accuracy; therefore the design shall
be based on idealized simplified constitutive laws for the joint behaviour.
As it is seen, the analysis and the design of multi-storey frames with semi-rigid and
partially resistant joints raise some problems; in particular:
- the characterization of the joint response;
- the frame analysis in view to determine the distribution of internal forces;
- the evaluation of the resistance and of the stability for isolated elements, for
sub-structures or for the whole frame.
It is necessary to point out not only the different aspects of the problem but also their
interaction and their complementarity. The followingflowchart is plotted with this aim in
view; it is briefly commented hereunder.
The characterization of semi-rigid beam-to-column joint is neither easy nor obvious. As a
matter of fact, the deformability of a joint results from several sources which may be
regrouped (Figure 1.2.), for sake of simplicity, in two main components: the column
web panel (subject to shear force Vn) and the connection(s) between the beam(s) and the
column (subject to bending moment M).
1.2
Components of defoimability
Web panel
Connection^)
yes
Concentration
!
C/5
Prediction of behaviour
Prediction of behaviour
trn
Joint
Connection
Panel
Specific characteristics
according to the analysis
planned
Non-linear curve(s)
1st order
2d order
analysis
analysis
(manual or
(manual or
numerical)
numerical
ELAST. PLAST.
ELAST. PLAST.
Non-linear
c/i
elasto-plastic
numerical
analysis
1
c
1.3
Displacements and
Ultimate loads
resistance of
forULS
sections for SLS
. connection feformability
<t>=eb-ec
. shear deformability of
column web panel
Y=ec-ef
. joint deformability
e=eb-9f
/-=^
Connection (M- <f>)

Sheared panel (Vn- *#
Figure 1.2. - Main components of joint deformability.
Both components may be considered separately, what corresponds to the reality, or, on
the contrary, be added in order to characterize the joint as a whole (concentration of the
joint deformability). Whatever the option may be, the corresponding curves have to be
predicted only from the knowledge of the mechanical and geometrical joint properties.
Then we've got either to define characteristic non-linear curves - two when the
deformability sources are differentiated; only one when it is concentrated - or to calculate
specific values - initial stiffness, secant stiffness, design moment resistance,.. - according
to the analysis which is planned.
The use of non-linear joint moment-rotation curves seems to be restricted to sophisticated
numerical analysis methods which are often able to take also account of the two types of
frame non-linearities (material and geometrical). The alternative procedure based on the
evaluation of specific joint characteristics is required by the hand (or pseudo-hand)
design methods; it may also be associated to numerical analysis methods less
sophisticated than the previous ones but however sufficiently accurate for practice.
The use of non-linear programs has the advantage to provide all the quantitative
informations which are necessary to form an opinion of the performance of the studied
frame.
However, reliable results quite satisfactory for daily practice may be based on a
simplified joint modelling associated to first or second order analysis methods depending on whether the frame is braced or not against lateral displacements - ; the
designer has to choose between the approaches - elastic or plastic - according to his
desiderata or to the class of cross-sections used for the connected members.
1.4
The less important degree of sophistication of these methods requires to make use of the
results with the view to determining the displacements of the frame or to checking the
elastic resistance of the cross-sections under service loads (SLS - serviceability limitstates), on one hand, and to evaluating the ultimate frame resistance at collapse (ULS ultimate limit-states), on the other hand. The judgement relative to the aptitude for the
frame to satisfy the limit-states is the direct corollary of these calculations.
This flow-chart should help the reader to clearly understand the aims of the present
research which is basically related to the study of two different types of building frames:
- steel frames with cleated connections;
- composite frames with steel columns and composite floors; the composite floors are
composed of steel beams surmounted by a continuous concrete slab with steel
reinforcement bars, the composite action is ensured by the presence of steel connectors
welded to the upper beam flanges.
As the in-plane behaviour of the frames is particularly of concern in this research, it has
to be assumed that the three dimensional frame effects are limited and that the frame
response may be reduced to that of the main constitutive in-plane portal frames.
This goal is achieved, for composite frames, through the concept of effective width for
the composite beams.
That being, we can concentrate on the aims of the present research:
1) Paramedical study, through experimentation in laboratory, of the behaviour of 56
beam-to-column bolted joints:
- 6 tests on steel exterior joints with cleated connection (3 cleats: lower and upper
flange cleats, single web cleat - see figure 1.3.a)
-12 tests on steel exterior and interior joints with cleated connections (3 cleats - see
figure 1.3.b. - or 2 cleats: lower flange cleat and single web cleat - see figure
I.3.C.)
- 38 tests on composite interior joints with steel cleated connections (3 cleats - see
figure 1.3.e. - or 2 cleats - see figure 1.3.d.).
-
-
r
a. "T" steel joint with
3 cleats.
1.5
b. Cruciform steel joint with

3 cleats per connection.
c. Cruciform steel joint with

2 cleats per connection.
d. Cruciform composite joints

with 2 cleats per connection.
e. Cruciform composite joints

with 3 cleats per connection.
Figure 1.3 - Types of joints tested in laboratory.

2) Development of calculation methods for the prediction of the non-linear flexural
response of the beam-to-column joints tested in laboratory and described in 1).
3) Elaboration of a finite element able to simulate the semi-rigid behaviour of beam-tocolumn joints and introduction of this new element in the non-linear program
FINELG in order to analyse the actual response of frames.
As we will see all along this report, the objectives of the research have been completely
fulfilled and some chapters go far beyond the initial planning:
- Chapter 2 on "Experimental investigation" presents the results as well as an analysis of
the 56 tests performed at the M.S.M. laboratory of the University of Lige. The annex
4 relative to this chapter contains a floppy disk on which the reader wl find all the
geometrical and mechanical data relative to the tests performed, the moment-rotation
curves obtained as well as a program allowing for the treatment of these data. This
program is a part of the database called SERICON developed in collaboration with
ARBED, at the University of Aachen (Germany). SERICON allows to produce easily
standardized data sheets for the steel and composite joints tested; an example of such a
data sheet is given also in annex 4.
- Chapter 3 presents briefly the calculation methods for the prediction of the semi-rigid
behaviour of
- column web panels subject to shear,
- column web panels subject to transverse loads;
- steel cleated connections.
These models have been developed in the frame of the present research; their
description may be found in annex 1, as well as the comparison with experimental data.
The attention focuses then on the non-linear response of composite joints tested in
laboratory. A detailed description of the prediction model and of the associated
calculation formula is presented. At last, he annex 3 reports all the comparisons
between the experimental curves obtained in laboratory and these resulting from the
application of the model.
A program aimed at easily calculating the M-6 curve relative to a composite joint,
whose geometrical dimensions and mechanical properties of constitutive material are
known has been written at ARBED-Research. Thefloppydisk containing this program
is available in annex 3.
1.6
- Chapter 4 on "Numerical developments" deals with the introduction of a new finite

element able to simulate accurately the non-linear behaviour of connection and sheared
column web panels in the non-linearfiniteelement program FINELG. The mathematic
background of these numerical developments in given in annex 1; the attention focuses
more in this final report on the advantages of this new finite element with respect to
other formulations proposed in the past.
This new efficient tool is then used in the frame of a large parametric study, the aim of
which is to define in which conditions the joint deformability may be concentrated
(connection deformability + shear deformability) in single flexural springs acting at the
end of the beams, the conclusion of this parametric study, which was not planned in the
official program of the present research, developed initially for the study of steel frames
may be used in conjunction with the "concrete" or "reinforced concrete" element in
order to simulate the structural response of composite frames; the shrinkage, the creep
and the non-linear interaction between the steel beam and the concrete slab may be taken
into account in such numerical simulations.
- The flow chart described here above points out the necessity to propose to the
designers simplified, but nevertheless accurate, design methods for frames with
semi-rigid joints. This topic is not covered by the present research. It appeared however
important in this document to specify that such procedures exist for braced and
unbraced frames and to briefly describe their mode and range of application. Chapter 5
on "Simplified methods for the design of steel structures with semi-rigid joints"
achieves this goal.
The design methods which are summarized in this chapter are extracted from a doctoral
thesis recently presented at the University of Lige (see annex 1).
This chapter is obviously not directed destined to practicers which would like to apply
the simplified methods proposed. This emphasizes the need to prepare in the future a
document presenting the submentioned simplified procedures in a way really orientated
towards daily practice.
- Chapter 6 on "Benefit of semi-rigid design" points out the potential benefits resulting
from a semi-rigid design of building frames; this economical interest is well highlighted
for two particular frames through cost comparisons betweenrigid(or pinned) and semirigid design approaches. The need for other cost comparisons and the further
developments in thisfieldplanned at the University of Lige are also expressed
- Lastly chapter 7 gives the conclusions of this research and explains the performed work
in relation with the flow chart presented in this chapter (page 1.3)
It remains now to the reader to discover in a more detailed way the respective contain of
the six chapters constituting the present final report.
1.7
2. Experimental investigations.
2.1. Types of tested beam-to-column joints.
Present experimental research is aimed to analysing the behaviour till collapse of in
plane strong axis beam-to-colum joints subject to static loading.
The wording "strong axis" means that the end cross-section of the beam(s) is connected,
whatever the names, to one of the column flanges so that any loading experiences
strong axis of the column cross-sectional area (figure 2.1.a.).
The "composite" joints considered in this study allow to connect afloorcomposed of
steel beams surmounted by a reinforces-concrete slab to a steel column (figure 2.1b.).
a. "Strong axis"
b. "Composite"
Figure 2.1. - Definition of a strong axis beam-to-column composite joint as considered
in this study.
Two joint configurations named respectively "cruciform" or "interior" (figure 2.2.a.)
and "exterior" or "T" (figure 2.2.b.) have been considered. The main reasons of this
choice will be explained in sub-section 2.3.
a. "Cruciform" or "interior" joint

b. "Exterior" joint
Figure 2.2. - Possible joint configurations.
At the beginning of the experimental study, it was decided:
2.1
- to focus the attention on steel connections where the beams are connected to the
column by a lower flange cleat, a single web cleat and possibly a third cleat
connecting the column to the upper beam flange;
- with a view to assess the role played effectively by the so-called composite action, to
test associated joints without concrete slab.
These choices led potentially to the testing of the eight different types of beam-tocolumn joints represented infigure2.3.
The number has however been reduced to 5 (types A, B, D, E and F) on account of the
following considerations:
- Except in special joint configuration which are not considered here, the beam
continuity due to the presence of the reinforced concrete slab is only warranted for
cruciform joints; the testing of type G and H composite exterior joints is
consequently equivalent to that of the associated type C and F steel joints.
iiHinmn
TR
'
1. Composite joints with 2 cleats per

connections (type A).
h-
2. Composite joints with 3 cleats

per connection (type D).
^
-n
=T
3. Steel joints with 2 cleats
per connection (type B).

per connection (type E).
CRUCIFORM JOINTS.
10
2.2

per connections (type G).

per connection (type H).
u
/
T
S
per connection (type C).
EXTERIOR JOINTS.

per connection (type F).
Figure 2.3. - Types of beam-to-column joints tested in laboratory.

The deformability of the very flexible type C joints concentrates mainly in the
connection elements (web cleat in tension) and does not differ consequently from that
registered in type B joints, it is not the same for type F joints where the relatively
high bending moment carried over from the beam to the column is likely to lead to
the deformability in shear of the column web panel (see sub-section 2.3.1.).
2.3
11
1.1. Parameters influencing the detailed choice of the connections.

The numerous parameters which were likely to be studied cannot obviously be all
examined. The governing ones have only been kept and arefixedas follows:
- The type of sections used as column and beam: it was decided to use a single HE
section (HEB 200) as columns and IPE sections as beams, in accordance with a
current practice for buildings.
- The column bending axis: only strong axis connections will be considered for obvious
reasons as far as in-plan behaviour is examined.
- The relative beam-to-column stiffness: because of the role played by the beam depth
in the force distribution between the steel and concrete components of a composite
joint (with a specified column section), three different beam depths are chosen:
IPE 240 - 300 - 360.
- The type of loading: only static loading is considered in the present research.
- The sizes of the connecting cleats (angles) and the kind of connection. Only bolted
connections are tested. For constructional and structural reasons, unequal leg angles
are used, the larger leg being in contact with the beam. For sake of simplicity, all the
angles used in a specified connection are identical. Two different values of thickness
are used with a view to assess the influence of the angle flexibility. Therefore
following sizes are used:
a) angles 150 x 90 x 10
b) angles 150 x 90 x 13
- All the bolts have a same diameter; they are tightened up to a specified torque
moment, associated to a not-controlled hand tightening.
- A clearance of 1 mm is adopted between bolts and their holes while the clearance
between the beams and the adjacent column flange is 15 mm. The holes are drilled on
base of the actual section dimensions.
- Three values of the steel reinforcement percentage (concrete slab) are selected in order
to cover the current range 0,67 % -1,3 % - 2,1 %.
- The slab thickness is kept constant: 12 cm. Its breadth is fixed to 120 cm because of
the considerations dealing with the effective width.
- Concrete strength is kept as constant as possible and corresponds to normal concrete.
- The concrete slab is connected to the steel beam by means of shear stud connectors.
(Stud connectors are designed to ensure a full interaction between the beams and the
concrete slab until the joint collapse). Two exploratory tests are conducted with a
view to assess the influence of the layout of connectors of the joint flexibility.
- Taking account of the slight role played by the axial load in the column on the
flexibility of web and flange cleat connections, the effect of the axial load is
disregarded.
It has also to be mentioned that the transversal slab reinforcement has been designed in
order to avoid the collapse of the slab during the test by excess of longitudinal shear in
the concrete.
A general view as well as a complete description of the experimental program are given
in the followingfiguresand tables.
12
2.4
Lowerflangecleat + web cleat

Type of
^connection
parameter
Beam depth (BD):3
angle thickness
(AT): 2
reinforcement
percentage (RP): 3
layout of shear
connectors
number of tests/
type of connection
total number
of tests
tip
1 test per BD 1 test per BD

and per AT
and per AT
1 test per RP
per AT
and per BD
1 exploratory
test
(other parameters fixed)
19
fo
Both flange cleats + web cleat
.B
t{

and per AT
and per AT
1 test per RP
per AT
and per BD
1 exploratory
test
(other parameters fixed)
19
25
\h
1 test per BD
1 test per BD
and per AT
31
56
Table 2.1. - General view of the experimental program.

Number
Test number
Connection type
Fig.2.4.
Column
Beam
1
2
3
4
5
6
24T3C1
24T3C2
30T3C1
30T3C2
36T3C1
36T3C2
2.4. a
2.4. a
2.4. a
2.4. a
2.4. a
2.4. a
HE200B
HE200B
HE200B
HE200B
HE200B
HE200B
IPE240
IPE240
IPE300
IPE300
IPE360
IPE360
Thickness of 150 x 190

cleats
Flange cleats Web cleats
(mm)
(mm)
10
10
13
13
10
10
13
13
10
10
13
13
a. Steel "T" test arrangement.
2.5
13
Number
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
Test
Connection
number
type
Fig.2.5.
30X3C1
30X3C2
30X3C3
30X3C4
30X3C6
30X3C7
30X3C8
30X2C1
30X2C2
30X2C3
30X2C4
30X2C5
30X2C6
30X2C7
36X3C1
36X3C2
36X3C3
36X3C5
36X3C6
36X3C7
36X2C1
36X2C2
36X2C3
36X2C5
36X2C6
36X2C7
24X3C1
24X3C2
24X3C3
24X3C5
24X3C6
24X3C7
24X2C1
24X2C2
24X2C3
24X2C5
24X2C6
24X2C7
2.5. a
2.5. a
2.5. a
2.5. a
2.5. a
2.5. a
2.5. a
2.5. b
2.5. b
2.5. b
2.5. b
2.5. b
2.5. b
2.5. b
2.5. a
2.5. a
2.5. a
2.5. a
2.5. a
2.5. a
2.5. b
2.5. b
2.5. b
2.5. b
2.5. b
2.5. b
2.5. a
2.5. a
2.5. a
2.5. a
2.5. a
2.5. a
2.5. b
2.5. b
2.5. b
2.5. b
2.5. b
2.5. b
Column
HE200B
HE200B
HE200B
HE200B
HE200B
HE200B
HE2O0B
HE200B
HE200B
HE200B
HE200B
HE200B
HE200B
HE200B
HE2O0B
HE200B
HE200B
HE2O0B
HE200B
HE200B
HE200B
HE200B
HE200B
HE200B
HE200B
HE200B
HE200B
HE200B
HE200B
HE200B
HE200B
HE200B
HE200B
HE200B
HE200B
HE200B
HE200B
HE200B
Beam
IPE300
IPE300
IPE300
IPE300
IPE300
IPE300
IPE300
IPE300
IPE300
IPE300
IPE300
IPE300
IPE300
IPE300
IPE360
IPE360
IPE360
IPE360
IPE360
IPE360
IPE360
n>E360
IPE360
IPE360
IPE360
IPE360
IPE240
IPE240
IPE240
IPE240
IPE240
IPE240
IPE240
DPE240
IPE240
IPE240
IPE240
IPE240
Thickness of
Concrete
slab
150x90 cleats longitudinal
flange web reinforcemen
t
(mm) (mm) Fig.2.6.
10
10
C
10
A
10
10
B
10
10
B
10
13
13
A
13
13
C
B
13
13
10
10
B
10
10
A
10
10
C
10
10
B
A
13
13
B
13
13
13
13
C
10
10
A
10
B
10
10
10
C
13
13
A
B
13
13
13
13
C
10
10
B
10
A
10
10
10
C
13
13
C
13
B
13
A
13
13
10
10
A
10
B
10
10
10
C
13
A
13
13
13
B
13
13
C
10
10
A
10
10
B
10
10
C
13
13
A
13
13
B
13
13
C
Shear
connectors
and transversal
reinforcement
Fig.2.7
B
B
B
C
B
B
B
B
B
B
C
B
B
B
D
D
D
D
D
D
D
D
D
D
D
D
A
A
A
A
A
A
A
A
A
A
A
A
b. Composite cruciform test arrangement.
14
2.6
Number
Test number
Connection type
Fig.2.4.
Column
Beam
45
46
47
48
49
50
51
52
53
54
55
56
24X3C4
24X3C8
30X3C5
30X3C9
36X3C4
36X3C8
24X2C4
24X2C8
30X2C8
30X2C9
36X2C4
36X2C8
2.4.a
2.4.a
2.4.a
2.4.a
2.4.a
2.4.a
2.4.b
2.4.b
2.4.b
2.4.b
2.4.b
2.4.b
HE200B
HE200B
HE200B
HE200B
HE200B
HE200B
HE200B
HE200B
HE200B
HE200B
HE200B
HE200B
IPE240
IPE240
IPE300
IPE300
IPE360
IPE360
IPE240
IPE240
IPE300
IPE300
IPE360
IPE360
Thickness of 150 x 190

cleats
Flange cleats Web cleats
(mm)
(mm)
10
10
13
13
10
10
13
13
10
10
13
13
10
10
13
13
10
10
13
13
10
10
13
13
c. Steel cruciform test arrangement.

Table 2.2. - Detailed description of the experimental program.
2.4.a.
Figure 2.4.
=J
2.5.a.
2.5.D.
Figure 2.5. - Different connections types.
index
A
B
C
Diameters of bars
(mm)
10
14
18
% of long,
reinforcement
0.67
1.3
2.1
Figure 2.6. - Concrete slab longitudinal reinforcement.
2.7
15
Shear connectors for one beam

index
Diameter
(mm)
d
(mm)
A
B
C
D
16
19
19
22
110
110
110
110
Number for
each of the
two files
19
19
10
19
Lg
Lg
(mm)
1980
1980
990
1980
365 mm
Transversal
reinforcement
d
Diameter
(mm)
(mm)
8
8
8
8
10
8
8
6
Concrete slab
d
IPE beam
-I -
2500 mm
beam end where the assemblage i s realized
Figure 2.7. - Shear connectors and transversal reinforcement arrangement

2.3. Testing arrangement.
2.3.1. Components of the joint deformabilitv.
A joint is the whole region concerned by the assemblage of the beam(s) with the
column. It is composed by the very end portion of the beam(s), the facing adjacent
portion of the column, as well as by all the connecting accessories (end plate, cleats,
bolts, welds,...) required by a specified type of connection. The deformability of a
strong axis beam-to-column joint consists mainly in two parts, which are respectively
fed by several contributions:
a. The deformability of the connection area associated to the following phenomena:
- Deformation of the connection elements: end plate or cleats, bolts, rivets, welds,...
- Slip and/or hole clearance;
- Deformation of the column web, across its depth, in the so-called tensile and
compression zones, i.e. in the regions where the forces carried over by the beam(s)
have to diffuse in to the column web (this effect, termed "trapezoidal effect", is the
result of the respective lengthening and shortening of the column web depth).
b.The shear deformation of the column web panel, which is subject to a complex shear
stress distribution in the region of the joint.
Let us reason on the strong axis steel beam-to-column joints represented in figure 2.8.
The contribution (J> of the deformability of the connection area to the joint relative
rotation 0 is defined by the difference between rotations Gb and 8C. That of the shear
defomation of the column web is represented by the difference between 8C and 8f,
where Of is the flexural rotation of the column.
16
2.8
Figure 2.8. - Deformability of a strong axis steel beam-to-column joint.

Regarding the shear force in the column web panel, it shall be stressed that it results
from: i) the combined action of equal but opposite forces Fb in the beam flanges, the
resultant of which is statically equivalent to the bending moment of the beam, and ii)
the shear force resulting from the moment distribution in the column.
Because both sources of deformability are dealing with the response of two different
zones of the beam-to-column joint, their contributions shall be assessed with reference
made to their respective appropriate loadings. As these latter are not identical, it must
therefore be accounted for both defomability sources separately and laboratory tests
have to be instrumented in such a way that measurements of the deformability
components be possible.
2.3.2. Test specimens and testing arrangements.
Because only the response of in-plane semi-rigid joints is of concern, the experiments
are of course conducted not on whole structures but on structural subassemblages
composed of portion(s) of beam connected to a portion of column. These
subassemblages must be designed so that to allow for realistic types of loading, while
having sizes which prevent external load introduction effects from altering the response
of the joint itself.
Shear deformation of the column web panel is expected to be more pronounced in teejoints, where a single beam is connected to the column (figure 2.9.a.). In contrast, the
"trapezoidal effect" exhibits as well in tee-joints as in cruciform joints, where two
beams are connected to the column at the same level and are both similarly loaded
(figure 2.9.b.). Therefore it shall be distinguished between tee-joints, which are
representative of outer joints in a real framework, and cruciform joints which
correspond to inner joints of this structure (figure 2.10.).
The height of the column is chosen so that it represents roughly the depth of one storey.
The beam is connected at mid-height of the column, so that the ends of the latter may
be considered as points of contraflexure in the columns of a sway frame subject to
horizontal loads. For sake of testing, it is thus sufficient to apply axial load at the ends
of the column. Bending in the column as result of the loading of the beam(s) will be
produced by the horizontal support reactions at the ends of the column.
What about the length of the beam(s) in tee or cruciform joint specimens ? It is
determined in view to allow for bending-to-shear ratios in the connection similar to
those encountered in practice. In this respect, it is while mentioning that some tests
reported in the literature do not care at all for such condition, which should however be
considered as a requisite. Bending in the beam is produced by point load(s) applied at
the end of the cantilever(s).
2.9
17
If
I i
a. Tee joint
b. cruciform joint
Figure 2.9. - Testing arrangement.
Outer joint
V:
Y
Inner
joint
_s
N.^_ ^.y
/ / / / / / / / / / / / / / / / / / / / / / /
Figure 2.10. - Subassemblages and frames

When semi-rigid composite joints, the steel column is connected with the composite
beam(s), composed of hot-rolled section associated to a reinforced concrete slab, by
means of shear connectors. Slab has a spatial structural behaviour. Because only
in-plane tests are carried out the composite test specimens must be carefully designed in
order to get an appropriate composite action. In this respect, the distribution of direct
bending stresses across the slab width must be quasi uniform in the joint cross-section.
Because the bending moment is produced by point load(s) at the end of the cantilever
beam(s), the width of the slab in the test specimens cannot be smaller than the regular
effective width, which is aimed to allow for the aforementioned required stress
distribution. This condition must also be adopted as a prerequisite, it is checked by
means of measurements, during the tests, of the direct stresses in the steel
reinforcements in several cross-sections, and especially in the joint cross-section.
18
2.10
The test specimen is inserted in a steel testing rig, that is fixed on the 1.5 m thick
concrete testing floor of the laboratory. This rig is aimed at provided the reactions
required by the loads to be applied to the test specimen.
The ends of the column are supported by devices which are able to support vertical and
horizontal reactions while providing no moment restraint; they can thus be considered
as ideally hinged.
Guide-plates are located all along both the beam(s) and the column in view to present
from any spatial behaviour and out-of-plane displacement of the subassemblage whose
in-plane behaviour is investigated only.
Load is applied by means of hydraulic jacks fitted with load cells at the end of the
cantilever. For the cruciform tests performed in the frame of this research, it has been
however decided to keep equal the loads acting at both cantilever ends (symmetrical
loading of the joint - joint defomability, reduced to that of both connections). To ensure
the equality of the moments in the left and right connections, the external load is
applied directly in the columns (see table 2.4.). The load is increased up to the joint
collapse or to the maximum deflection of the cantilever ends according to what is first
reached. Unloadings are carried out during the tests in view to compare the
instantaneous stiffness with the initial one.
2.4. Instrumentation and measurements.
The tests are instrumented so that the measurements allow for determining the
amplitude of all the components of the joint deformability at any level of the loading. It
is useful to perform measurements of horizontal and vertical displacements as well as of
rotations in appropriate sections; for this purposes, electronic transducers are used. The
applied load are measured by means of load cells.
As it is required to identify and measure separately all the components of the joint
deformability, it is of paramount importance that the measurements be scheduled
accordingly. In other words, it is necessary to perform "redundant" measurements
which shall allow for computing a specified rotation by at least two different manners.
Doing so warrants to get results even when one transducer is misfunctioning or when
something wrong is likely to occur during the test The "direct" measure is of course
preferable and the most thrusting but the searched information must also be deduced
from "indirect" measures, when necessary.
Last the testing frame, which is anchored on the testing floor of the laboratory and is
aimed at resisting the applied load and the support reactions, cannot be undeformable,
whatever its rigidity. Therefore, because rather small quantities are to be measured, one
must avoid to use this rig as support for the measurement devices. Accessories
independent of the testing rig are therefore clamped on the testing slab whose rigidity
can be considered as infinite. Following measurements are made for steel tee-joints
(Table 2.3.):
a) vertical and horizontal displacements at the end(s) of the beam and the column;
b) rotations of the beam and of the column;
c) rotation associated to the load introduction deformability of column web;
d) rotation due to the slip at the junction cleat-beam flange;
e) proper deformation of the upper flange cleat;
f) proper deformation of the column flange in the tension zone of the joint.
In view to compute the rotation of the beam and of the column, six measurements were
made (Table 2.3.). The transducers (5) on the column flange are located symmetrically
to the axis of this column. Their spacing is as large allowed by the beam depth.
Transducers (4') are pointing to their small stitches welded transversely into the web;
2.11
19
they are also located symmetrically to the beam axis. Transducers (4) are located at the
lower flange of the beam as near as possible to the connection cross-section.
The measurements of the rotation resulting from the slip at the interface between the
flange cleat and the beam flange are made by measuring the relative displacement
between two points located at a so small spacing as possible, respectively on the cleat
and on theflangeon both upper and lowerflangesrespectively (Transducers 7 and 7' in
Table 2.3.).
Direct measurements of the deformation of the column web in the compression and
tension zone are made in view to get the deformation curve associated to the
"trapezoidal effect" in the column web panel (Transducers 6 and 6' in Table 2.3.).
The computation of the rotation due to the deformation of the column flange results
from the measurement of the relation displacement (8) between two points located
respectively on one of the bolts connecting the upper cleat to the column and the
column web (Table 2.3.).
The rotation due to the deformation of the cleats (including the bolt elongation) is
deduced from measurements of the relative displacements between the two points
located at a so small spacing as possible respectively on one of the bolts connecting the
upper cleat to the column and on the upper beam flange. (Transducers 9 in Table 2.3.).
When cruciform steel joints, the deflections and the rotations are determined
accordingly for both beam-to-column connections.
Composite connections require in addition the measurement of the slip between the
concrete slab and the adjacent steel flange onto which the slab is connected by studs.
All the details for measurement are given in table 2.4.
Strain measurements are aimed at investigating the stress distribution in the beam crosssection. For bare steel beam-to-column joints, only some tests have been fully
instrumented in this respect; it has soon appeared that strain measurement in both
flanges of the beam in two cross-sections is quite sufficient For composite joints of the
steel beam but also in the reinforcements have been made in several cross-sections in
view to explore the stress distribution, more especially in the concrete slab because of
the shear lag effect. As soon as it has been demonstrated that a uniform direct stress
distribution in the slab in the cross-section corresponding to the connection was
reached, the following tests have been conducted with a less important instrumentation.
The measurement of the beam rotation has revealed to be rather dubious in most of the
tests performed. That is the consequence of: i) the out-of-plane deformation of the beam
web during the joint loading as a result of the dissymetry of some connection crosssection and the section where the rotation measurement is performed. However that
does not at all prevent from determining the rotation of the beam because of the
redundancy of the measurements performed. It has been anyway concluded that a direct
measure by means of potentiometrical rotating transducers is far preferable and such
devices have been purchased.
20
2.12
Instrumentation.
2m
Rotation curves.
Main components.
L = 2.5m
2m
jack(lOOkN)
P
M = PxL
B3
- connection
8 = 8 b - e c (fig 2.8.) 9 b by4 4' or 4"
6C by 5 or 5'
- shear of column web panel
Y= 6b - Of (fig 2.8.) 6f theoritically
evalueted (O.K. if column remains
elastic)
or
Y=<|>
"T testing arrangement and location of

displacement transducers.
- 9
-joint
<>
t = e+y
or <>
t = F(l, 2and3)
Components of connection deformabilitv.

- load introduction
en=(6-6')/d
-slip
es = ( 7 - T ) / d b
- cleat and column flange
ed = e -e s -eu
- column flange
6^(8-6)^
- cleat included bolt elongation
6a = (j- ecf
ea = ( 9 - 7 ) / d
Location of the displacement transducers and of the

rotation transducers (4") in the vicinity of the joint.
Table 2.3. - Measurements and reported curves for steel joints.
2.13
21
Instrumentation.
Defomability curves.
^
L b =2.3m
M-w curve (w = vertical

displacement of the column)
w=2
3 +4
^ ^ ^ ^ ^ . J K U i ^ k V S ^ ^ ^
M-<|> curve (<>
| = relative rotation of
the connection)
1^=2.42111
L, = 2.42m
04
6+ 7
<(> = =
A
A_
jack(500kN)
Crucifonn testing arrangement and
location of displacement transducers.
(first assessment)
5 - ( 9 + 2x 12)
2xHb
(second assessment)
where Hb = depth of beams.

M - <>| curve (<j> = relative
rotation without slip of the
connection)
4> ="2TH
Location of the displacement transducers and

of the rotation transducers (6 and 7)
in the vicinity of the connections.
M - <>| curve (<j> = component of

the connection rotation <J> dueto
the compression in the column
web)
2xHb
M - <>| curve (<j> = component of

the connection rotation (J> due to
the deformation at the upper beam
flange)
* =-2THbTable 2.4. - Measurements and reported curves for composite joints.
22
2.14
2.5. Data acquisition.

Measurements have been recorded during tests by means of a Hewlett-Packard
datalogger, which is able to measure up to 200 different parameters (transducers, strain
gauges,...) at a time.
This data logger is automatically commended by a software running on a PC computer.
This software - SCANPACK - was developed in the Lige laboratory; it is especially
useful for tests on structures and allows to follow, in real time on the screen, the
evolution of any reference measurement versus the applied load. All the measurements
are stored in view to further analysis of the results. This package stores in addition all
the information regarding the test conditions (connections of measuring devices,
calibration factors,...); that is especially useful for repetitive tests.
2.6. Experimental characteristic curves.
- Because of the appropriate instrumentation, the following characteristic curves
associated to the different components of the joint deformabity have been recorded
for steel joints:
a. the joint relative rotation curve;
b. the connection relative rotation curve;
c. the column web panel rotation curve.
For what concerns especially the deformabity of the cleated connection, it is
composed by the addition of the following curves:
d. the load introduction rotation curve;
e. the connection slip rotation curve;
f. the cleat deformabity rotation curve;
g. the column flange rotation curve.
- As explained in table 2.4., the foUowing deformation have been reported in the case
of composite joints:
a. the vertical displacement of the column;
b. the relative rotation of the connection;
c. the relative rotation without sp of the connection;
d. the component of the connection relation due to the compression in the web;
e. the component rotation due to the deformation at the upper beam flange.
It must be noted that the "connection-relative rotation M-<(>ws curve without sp"
obtained by combination of the measurements number 5 and 9 (defined in table 2.4.)
differs from the one that would result from another test on an equivalent connection but
actuaUy with no sp (cleats welded to the beam, for instance); this may be explained by
the dependence of the measurement number 5 on the sp value at the lower beam
flanges.
For each test, the strain have been measured in the beam flanges and in the rebars (20
strain gauges) in order to explore the stress distribution in the two beam sections
located just near the steel connections. This has enabled the foUowing values to be
reported:
(6) the mean values of the strains in the steel reinforcement of the concrete slab;
(7) the mean values of the stresses in the steel reinforcement of the concrete slab;
(8) the mean values of the stresses in the beam flanges.
Figure 2.11. represents, for instance, the different curves recorded for a specified steel
test specimen.
2.15
23
TEST 24T3C2 < C3nP0HE>(S OF^HEÔlNÔeFOJHAB IU

ri
fari.
canrcNENTS
JOINT
I
CONNECTION
VES SH EA
2.1 La. Components of the joint deformability.
ES_243ca_^_esfa!e.HS_aF_THC_c3NHEcrioM DEFORMAS i u TY
VN.
90.33
COHMNENTS
. CONNECTJOH
LOAD
INTSOOUC.
CLEATS
8
-00
AO
COLUn FLANGE
2.11.b. Components of the connection deformability.

Figure 2.11. - Characteristic deformability curves for 'T" steel joints 24 T3 C2.
24
2.16
2.7. Interpretation of the tests results.

2.7.1. Steel "T" tests.
The diagram of figure 2.12. shows clearly the beneficial influence of the beam depth
and of the thickness of the cleats on the stiffness of the nodes, on their plastification
level as well as on the maximum bending moment supported during the tests.
We think it advisable to mention that the tests have been stopped for fear of collapse of
the tension bolts, what could have heavily damaged the measuring device.
JOINT RELATIVE ROTATION CURVES

M
1 .80
IcNi
10
I
11.
I
I
I
I
-I
1 .66
1 .50
1 .20
0.90
T"
0.60
jj I
r
J
SIX_LABORATORY_TESTS
0.30
*r~II
|.00
I 4. 00
I
I
|6.00
24T3CI
24T3C2
30T3CI
,0"
8.64
|8.00
T
I
I
30T3C2
36T3C1
10.00
36T3C2
AD
Figure 2.12. - Joint deformability curves for the six steel "T" tests performed.
The maximum moment supported by the joints at the end of the test represents an
important percentage of the plastic moment of the beam (table 2.5.).
Test
24T3C1
24T3C2
30T3C1
30T3C2
36T3C1
36T3C2
Mmax / Mp (%)
60
76
59
80
51
58
Table 2.5.
2.17
25
The leading part of the joint deformability is due to the connection. This is linked up to
the progressive yielding of the upper flange cleat and of the web cleat as it clearly
appears in figure 2.13. The slip between the beam and the cleats as well as the out-ofplane deformation of the column flange in the tension zone of the joint constitute the
two other sources of connection deformability.
The "trapezoidal" deformability of the column webs due to the introduction of loads
carried over by the beam has proved to be quite negligible for all tests.
For what regards the column web shear deformability, it may be noted that a beginning
of yielding has only been registered for the tests 30T3C2, 36T3C1 and 36T3C2 the
column webs remained elastic during the test.
Figure 2.13. - Characteristic deformed shape of the joints.

Figure 2.14. shows, for instance, the cleat thickness on the deformability curves
obtained for specimens constituted with IPE 240 beams. It may be noted that:
- the deformability of the cleats of different thickness is hardly different at the
beginning of the loading (elastic behaviour);
- the spread of yielding in the cleats varies significantly according to their thickness;
- the cleats present an appreciable strength reserve after their plastification;
- the difference of the bending moments carried over at the end of the test (for a same
level of deformability) is important; it amounts for instance to 27 % for the joints
with IPE 240 beams (35 % for IPE 300, 12 % for IPE 360);
- the behaviour of the column flanges in the tension zone of the joints varies according
to the cleat thickness contrarily to what could be imagined; this may be explained by
the appreciable variation of the prying forces in the bolts according to the thickness
of the cleats.
26
2.18
The compression and tension forces carried over by the flange cleats do not vary, for a
given bending moment in the beam, according to the cleat thickness; it is therefore
logical to have an identical behaviour for the column sheared webs of joints which only
differ by this particular parameter; that is for instance the case for the column webs of
joints 36T3C1 and 36T3C2, the characteristic curves of which are presented in figure
2.15.
CLEATS DEFORMABILI TY ROTATION CURVES

1
kN*
90.30
75.00
_l
1
U'
1
)
1
^'1
1 /"
1
1
1
1
1
> -
i/
60.00
_,/
y \
S
I
45.00
f
l
30.00
/ y
_i_Zi /
\f
'/
15.00 ii
'J
'
1 -^
. i
i_
^-A
i_. _
\i 1 j
/ (
'
1
i
ii
|
t~\
_LL_
S\
/
_ 1
i !
T~pt"T
TWO LABORATORY TESTS

_ 24T3C1
24T3C2
i'' i '7
|2 -C 0
|4.00
6.00
8.00
a. Cleats deformability curves.
2.19
27
COLUMN -LANGE DEFORMABILITY ROTATION CURVES

M
IcNm
90.00
45-00
'A
1
1
1
1
1
1
1
1
i
1
_l_
60.30
75-00
30.00
15-00
L____H
I
1

24T3C1
-J
1
I
1
1
1
1
J -
2-00
4.00
6.00
8-00
AD
24T3C2
b. Column flange deformability curves.

Figure 2.14. - Deformability curves for joints with IPE 240 beams.
COLUMN WF3 PAWEL DEFORMABILITY ROTATION CURVES

IcNm
2
10
1 .(50
1 .50
'
'
/i
1
L i
l_
1
1
1
1
|_
1
1
_|
(ti j
1 .20
!
j
j
-
.
|
|
'
1
1
J.
L
1
1
1
0.90
O.GO
0-JO
'
| |
'
'
'
'
'
|2-00
| 4.00
|6-30
1
i
|8.00
|

^fiT^ri
,0
-2
36T3C2
|10.00
AO
Figure 2.15. - Column web shear deformability for joints with IPE 360 beams.
28
2.20
2.7.2. Composite cruciform tests.

- The nature of the collapse for each specimen tested, as well as the ultimate bending
moment supported by the connections are represented in table 5.6. It may be seen that
these maximum moments represent an appreciable percentage of the theoretical
plastic moment of the composite beam sections. For the joints with EPE 240 beams,
the value of the theoretical plastic moment of the composite beam sections for
negative bending has been attained for each test.
- The collapse of the partial strength connections with EPE 300 and IPE 360 beams
tested is linked up to the buckling of the transversally compressed column web
(figure 2.16.) or to the excessive yielding of the steel bars, according to the
percentage of reinforcement in the concrete slab.
A brittle failure of sheared bolts has only been encountered for test 30x20; it seems
to result more from a local problem of load distribution between the bolts connecting
the lower cleats to the flange of the left beam then from an excess of stresses in these
four bolts.
- The collapse of the specimens with IPE 240 beams tested is linked to the
plastification of the composite beam sections, and not to the failure of the
connections, which may consequently be considered as full strength connections. The
important rotation capacity associated to this type of collapse is however limited at
the end of the test by the buckling of the compressed flange of the beam (figure
2.17.).
Figure 2.16. - Buckling of the column web.
2.21
29
Figure 2.17. - Buckling of the yielded lower beam flanges in compression.

The test results highlight the importance of the flexibility of the connections due to
the slip between the lower cleat and the beam flange. It is not very difficult or
expensive to prevent such slip with the result of improved rotational characteristics
for the composite connections tested in the scope of this research.
Two other components of the connection deformability registered during the tests in
the laboratory are related to:
- the compression in the column web;
- the variation of the distance between the upper flanges of left and right beams (this
measurement gives an idea of the concrete-slab axial deformation).
The relative importance of these two sources of flexibility is strongly dependent on
the percentage of reinforcement in the concrete slab.
30
2.22
Test number
24x3c1
24x3c2
24x3c3
24x3c5
24x3c6
24x3c7
30x3c2
30x3c3
30x3cl
30x3c6
30x3c8
30x3c7
3x2c2
30x2cl
30x2c3
30x2c5
30x2c6
30x2c7
30x3c4
30x2c4
Type of
collapse*
f,e
e
e
e
e
e
a,c
a
a
c(a)
a
a
c
d
a
c
a
a
a
a
Mmax/Mp
%
113
103
112
115
113
105
99
98
94
114
95
95
81
89
94
79
99
94
98
89
Test number
24x2cl
24x2c2
24x2c3
24x2c5
24x2c6
24x2c7
36x3ci
36x3c2
36x3c3
36x3c5
36x3c6
36x3c7
36x2c2
36x2cl
36x2c3
36x2c7
36x2c5
36x2c6
Type of
collapse*
e
e
e
e
e
e
a
a
a
a
a
a
b
a
a
c
a
a
Mmax/Mp
%
100
105
106
117
112
102
11
80
79
79
85
84
61
82
78
67
83
77
(*)
a: buckling of the column web at the level of the lower cleats

b: collapse under tension of the reinforcement bars in the concrete
slab
c. limitation of the maximum permissible vertical displacement of the column
(20 cm), due to excessive yielding of the reinforcement bars in the concrete slab
d: collapse under shear of the bolts connecting the lower cleats and the beam
flanges
e: buckling of the yielded lower beam flanges in compression
f: column web buckling
Table 2.6. - Description of the collapse for the performed composite tests.
2.7.2. Influence of the chosen test parameters.
The parameters investigated in the experimental part of this research are the following:
- the thickness of the cleats;
- the number of cleats;
- the percentage of reinforcement in the slab.
The results of the three test series (connections with IPE 240, IPE 300 and IPE 360
beams) lead to the conclusion that the influence of these parameters is similar according
to the type of beam. Only the IPE 240 test serie and the IPE 360 one (IPE 300 test serie
is characterized by the same collapse modes than the IPE 360 one) will be consequently
discussed here below.
Figure 2.16. for instance, shows clearly the beneficial influence of the percentage of
reinforcement on the rigidity and on the ultimate strength of the connections. The
substantial rotation capacity of the composite connections with IPE 360 beams and with
bars of 10 mm in the concrete slab is linked to the tensile yielding of the rebars; that of
the connections with EPE 240 beams is associated to the yielding of the lower beam
flanges.
2.23
31
Figures 2.17. to 2.19. allow one to highlight the influence of the cleat thickness on the
semi-rigid behaviour of the composite connections. It may be seen that the rotational
rigidity and the ultimate capacity of the connections are not strongly affected by this
factor, more especially as the differences registered may be partly explained by:
- the relative importance of the slip between cleats and beam flanges;
- the buckling direction of the column web (for tests with IPE 300 and EPE 360 beams),
which influences the value of the collapse buckling load, and which is dependent on
the initial out-of-plane deformation of the web, on its direction and on the position of
the single cleats connecting the beam webs to the column (they are submitted to
compression during the test and tend consequently to produce small bending moments
in the column web).
The necessity to connect the upper flange of the beam to the column by means of a
third cleat may also be discussed on the basis of the diagrams of figures 2.17. to 2.19.
The upper cleat does not affect significantly the behaviour of the connection (figures
2.18 and 2.19.) as long as the plastic resistance of the rebars is not reached in the
section of the connection and the associated plastic deformation has not developed. In
the other cases (figure 2.17.) an additional bending moment related to the resistance of
the upper cleat submitted to tension forces is carried over to the column by the
composite beams. This difference is less sensitive for connections with IPE 240 beams,
whose loss of stiffness is associated to the yielding of the lower beam flanges.
32
2.24
M - W CURVE
LuucJluJ
M IS IKE CONNECTION BENDINO MOMENT

U IS THE VERTICAL DISPLACEMENT OF THE COLUMN
kNm
BARS OF
18 MM
150X90X10 CLEATS"
, -'
TiO
2.16.a. Connections with IPE 240 beams.

TESTS JGX2C COMPARISON OF M - W DEcORMA3IL l TY CURVES
kNm
A. 00
150X00X10 CLF* T S
BARS OF 10 MM
BARS OF M MM
BARS O c 18 MM
2.16.D. Connections with IPE 300 beams.

Figure 2.16. - Influence of the percentage of reinforcement in the concrete slab.
2.25
33
M - W CURVE
uuuLJuuJ
M IS THE CONNECTION BENOINC HOHEM!

(c Nr.
10
3000.
2 CLEATS -
13 MM
BARS OF 10 tin
3 CLEA'S -
-1
10 MM
-f'
pio
2.17.a. Connections with IPE 240 beams.
TESTS 3GX COMPARISON OF M-W DEFORMABILI TT CURVES

IcNm
3.00
10
1.00
1.00
BARS OF 10 MM
3 CLEATS - 10 MM
3 CLEATS - 13 Mn
10
2 CLEATS - 10 MM
2 CLEATS - 13 MM
0.50
I.5
2.00
MM
2.17.b. Connections with IPE 360 beams.

Figure 2.17. - Influence of cleat thickness and number of cleat (rebars of 10 mm).
34
2.26
M - W CURVE
LUJCLLUJ
n IS HIE CONNECT I UN BENDING H OH EN T

U IS THE VERTICAL OlSPLACEMEH I OF IME COLUI
k NU
IO
joon
3
CLEATS
13 MM
BARS OF
CLEATS
CLEATS
1 MM
13 MM
10 MM
'
f'Pio
2000
a. Connections with IPE 240 beams.

TESTS 3GX COMPARISON OF M-W DEFORMASILI TT CURVES
10
3.71
2.SO
1. 2
BARS OF M MM
3 CLEATS - 10 MM
3 CLEATS - 13 MM
2
)0
0.40
/ 0-80
2 CLEATS - 10 MM
2 CLEATS - 13 MM
1.20
b. Connections with IPE 360 beams.

Figure 2.18. - Influence of cleat thickness and number of cleats (rebars of 14 mm).
2.27
35
M - W CURVE
fujJiJujJ
ri IS THE CONHECI I OH BENDING H OtlCNT

kNm
3 CLEATS
10 MM
CLEMS
10 MM
a. Connections with IPE 240 beams.
TESTS 30X C0MPAR1S0M OF M-W DEFORMABILI TY CL'.IVES

kNn
+.00
10
3.00
2.00
BARS OF 18 MM
1.00
3 CLEATS - 10 MM
3 CLEATS - 13 MM
,0
2 CLEATS - 10 MM
2 CLEATS - 13 MM
0.40
t .0
MM
b. Connections with IPE 360 beams.

Figure 2.19. - Influence of cleat thickness and number of cleats (rebars ofl8 mm).
36
2.28
2.7.3. Steel cruciform tests.

All the tests have been stopped by attainment of the maximum deflection of the column
(20 cm).
The maximum moment supported by the connections with 3 cleats represents an
important percentage of the plastic moment of the beam, contrary to what is observed
for connections with 2 cleats (table 2.7).
Test
number
24x3c4
24x3c8
30x3c5
30x3c9
36x3c4
36x3c8
Mmax/Mp
(%)
59
74
57
58
36
45
Test
number
24x2c4
24x2c8
30x2c8
30x2c9
36x2c4
36x2c8
Mmax/Mp
(%)
12
16
17
17
13
-
Table 2.7 - Moment capacity of the cruciform steel joints tested.

The leading part of the joint deformability is due to the connection. This is linked up to
the progressive yielding of the upper flange cleat, if present, and of the web cleat
(figure 2.20).
The slip between the beam and the cleats as well as the out-of-plane deformation of the
column flange (for connection with 3 cleats) in the tension zone of the joint constitute
the two other sources of connection deformability.
The "trapezoidal" deformability of the column webs due to the introduction of loads
carried over by the beam is quite negligible for all tests.
20.a. Connections with two cleats.
2.29
37
20.b. Connections with three cleats.

Figure 2.20 - Characteristic deformed shaped of cleated connections tested.
The introduction of the characteristic deformability curves of the steel cruciform tests in
classification diagrams proposed in annex chapter 6 of Eurocode 3 (version of February
1989) shows clearly that (see, for instance, figure 2.21):
- the connection with 2 cleats may be considered as pinned
- the connection with 3 cleats may be considered as semi-rigid and partially resistant.
CII UM I I III ! I | c m
WITH I f . ^*m
PKAM
kN
Figure 2.21 - Classification of the steel cruciform connections with IPE 240 beams
using respectively 2 or 3 cleats.
38
2.30

Theoretically, comparisons between cruciform and "T" tests performed on similar steel
cleated connections should lead to the equivalence of the corresponding connection
deformability curves obtained (as confirmed, for instance, by tests 24T3C2 and
24X3C8 in figure 2.22), seeing that:
- the deformability of the connection elements (cleats, bolts, column flange the low
value of the normal stresses in the column flange are not likely to alter the column
flange deformability) are independent of the joint configuration;
- the load-introduction deformability is negligible (see section 2.7.1.1.) and is not
likely to be altered, in "T" joints, by the low shear stresses existing in the column web
panel (see section 2.7.1.1.).
Figure 2.22., which presents such comparisons for joints using IPE 240 beams and three
cleats with a thickness equal respectively to 10 and 13 mm, does not confirm
necessarily this theoretical principle (see tests 24T3C1 and 24X3C4). A a matter of fact,
it has to be remembered (see 2.7.1.1.) that:
- the slip and bearing deformability between the beam and the cleats constitutes one of
the main sources of connection deformability;
- the amplitude of this component of deformability is highly depending on the initial
relative position of the sheared bolts in their holes (holes in the cleats and in the beam
flange).
Consequently it may be concluded that the response of the connection with 3 cleats
tested in laboratory is independent of the joint configuration, but is strongly influenced
by the amplitude of the slip and bearing deformability at each load level
ANO
cmjctroMPt
C O M H C T I O N *
WITH
I P U 4 O
M A H
Figure 2.22. - Influence of the cleat thickness and of the joint configuration on the
connection deformability curves.
2.31
39
Conclusions relative to the influence of the slip deformability and of the cleat thickness
similar to those drawn for connection with 3 cleats may be expressed for connections
with 2 cleats.
2.8. Data sheets.
The development of a database aimed at gathering all the available results of
experimental tests on steel and composite beam-to-column joints is in progress since
one or two years at the University of Aachen (Germany) in the department of Prof. G.
SEDLACEK. The database format used is dBase; the databank may be mounted on
IBM PC and compatibles.
With the financial assistance of ARBED, all the test results presented in this chapter
have been introduced in this databank called SERICON. (SEmi - Rigid CONnections).
In annex 4 of this final report, the reader will find a floppy disk containing, for each
test, all the necessary information relative to the geometry of the joints, the mechanical
properties of steel and the joint response in the form of M-8 curves characterizing the
behaviour of the connection of the sheared column web panel and of the whole joint
From this information, the database is able to produce directly data sheets, the format of
which has been discussed and agreed - during the collaboration meetings held each six
months since the beginning of the present research - by the industrial and scientific
representants of the three jointed projects on "Semi-rigid action in steel frame
structures", introduced in 1987 to the European Community. (7210-SA/507, 413 and
819). An example of such a datasheet is also reported in annex 4).
40
2.32
3. Theoretical
developments.
As specified in the introduction, the theoretical part of the present research aims to
elaborate mathematical models for the prediction of the semi-rigid behaviour of the bolted
steel and composite joints which have been tested in laboratory (see chapter 2).
The achievement of this goal requires the separate study of four main sources of
deformability:
- the ahead column web panels;
- the column web panels subject to transverse loads carried over by the connection
(head-introduction deformabiuty);
- the steel cleated connections;
- the composite connections by means of steel cleats;
We will mainly concentrate in this chapter on the non-linear response of composite
connections (see 3.4.). As a matter a fact, the mathematical models related to the three
other sources of deformability are commented and compared to experimental results in
annex 1 [1]; only a brief description of these models will be consequently given in section
3.1. to 3.3. of this chapter. The reader interested by this topic is also begged to refer to
[34] and to the following English publications [35 to 38]
3.1. Behaviour of column web panels.
An important parametric study has been realized at the Polytechnic Federal School of
Lausanne and at the University of Lige. All the results and all the conclusions of this
study may be found in [34].
This study is based on numerical simulations with the non-linear FE-program FINELG
(see chapter 4) of the loading up to failure of welded beam-to-column joints. Material and
geometrical non-linear effects are taken into account, although the latter is far less
important than the former. The specimens of the chosen joints are analysed in three
dimensions by using "shell" finite elements to model the webs andflangesof the profiles
and "beam finite element to model stiffeners. The adopted finite element meshes are
shown on figure 3.1, respectively for a "T" joint (one column, one beam) and a "cross"
joint (one column, two beams).
"Goss" joints are assumed to be symmetrically loaded.
1 1
nun
inni
DUE
i
i
T
~T
_-1
^ l
I.
Ill
1
'
T1
li
i l l 1!
1 Ili
III L
ri-
i
1
i^>-^
\ ^ >
i II
/I
Figure 3.1. - Joint finite element meshes.
3.1
41
The numerical simulations allow to study the propagation of the plasticity in the profiles
and to observe the exact failure modes.
Steel is supposed to follow a piecewise linear law shown on Fig. 3.2. The 2D
elastoplastic state of stress is dealt with by using the incrementalflowtheory and the von
MISES yield criterion. Parabolic patterns of rolling residual normal stresses in flanges
and webs are taken according to the ECCS recommendations [39]. Welding
imperfections are not considered. Complete data may be found in [40].
The moment-rotation curves characterizing the shear deformability (Vn - y curves) and the
load-introduction deformability (M - ty curves) of the column web panel have been
reported for every simulation.
The following parameters have been taken into account in the parametric study of the
joints:
a) the type of the beam(s);
b) the type of the column;
c) the loading of the joint;
d) the initial out-of-flatness of the column web;
e) the presence or not of transverse stiffeners on the column web.
i'
Nel
Qcl
MWTQI^
Qc2
Q b l Mbl
-i
T db
Nc2
Figure 3.2. - Stress-strain curve (mild steel).
de
Figure 3.3. - Loading of an interior joint
3.1.1. Column web panels in shear.

Only the conclusions relative to the behaviour of the sheared column web panels are
presented here.
a. The shear stresses in the column web panels may be considered as uniformly
distributed; this is due to the action of the column flanges.
b. The actual value of the shear force Vn may be obtained form the equilibrium equations
of the web panel.
It is given by the following formula (figure 3.3.):
V =
42
Mci + Mc2
Qbi + Qb2 de
(3-D
3.2
Some other researches refer to another formula:

V =
Mci + Mc2
(3.2)
The validity of the proposed formula (3.1) has been clearly demonstrated
c. The Vn - y curve for a given joint depends on the actual loading of the joint
Let us assume that the two unstiffened welded nodes offigure3.4 are subject to different
types of loading (figure 3.5) and let us report, for each node, the characteristic Vn - Y
curve in a common diagram (figure 3.6).
Beam
A
B
Beam
IPE300
HE500B
Column
HE 160 B
HE 300 B
Figure 3.4. - Definition of two welded joints ("T" arrangement).
P
V
1)M
])M
Simple bending (FS)

Pure bending (FP) Pure bending in the beam (MP)
Figure 3.5. - Different types of loading.
The shear force Vn takes account, by means of formula (3.1.), of the loading of the
joints; in consequence one could believe that the Vn - Y curves are identical for a given
node. Actually only a similarity exists in the elastic range of the web panel behaviour and
this demonstrates the validity of the proposed shear force definition (formula 3.1.).
The difference between the Vn - Y curves in the non-elastic range of the web panel
behaviour are not negligible.
In reality an unstiffened column web panel experiences three types of stresses in its most
stressed zone (figure 3.7.):
3.3
43
- the shear stresses T;

- the normal stresses a n resulting from the compression force and the bending moment
in the column;
- the normal stresses Oi resulting from the introduction of beam loads into the column
web.
The shear behaviour of a web panel shall obviously be affected, except in the elastic
range, by the relative importance of each of these stresses according to the type of joint
loading.
,4
On
1^=^"
Figure 3.7 - Different types of stresses in a web panel.

The existing methods for the prediction of the shear deformability of web panels do not
take the influence of the actual joint loading into account.
Figure 3.6. shows that this is questionable and has led to the elaboration of a new
approach able to integrate this effect.
MP
--FS
-FP
,8.00
,10.00
An
3.6.a. Joint A.
44
3.4
0.81
1.00
t.80
10
"
0.6O
l
MO
jm
*f 1 1 i !
0.20
i
!
1
2.00
!
1
4.00
!
1
COO
!
| B . 19
8.00
i
10.00
Po
FS
FP
3.6.b. Joint B.
Figure 3.6. - Characteristic Vn - y curves.
This study has led to the proposal of the multi-linear model for unstiffened column web
panels shown in figure 3.8. It will be referred to [34, 35] or to annex 1 for stiffened
column. The mathematical formulation of its characteristic values is not given here
because of lack of space but is presented in [1, 34, 35] which can be offered to any
interesting people.
1^Vn
Vnu
Vny
Vne
\
/ 1
/ 1
yst
7
5*-
Figure 3.8 - Shape of the shear model.

The model, which has been compared in [34] with all the already existing ones and which
is applied in figures 3.9 and 3.10 to the studied nodes A and B (figures 3.4 and 3.5),
allows to predict in a quite satisfactory way the main characteristic of the non-linear shear
curves.
3.5
45
its
4.0
;!
"
3. M
'
j^T
f y :
3.34
' ^
! !
J^^"
3. I f
3. M
! i!
JZ._,1
1.M
;!
4.W
3. M
v1
1
1.
',-n
3.M
1
4.M
1
j.M
I
(.M
1 MO
>
ja.M
4.M
CM
B.H
M.M
1 Ml
3.9.a. FS
3.9.b.FP
Figure 3.9 - Comparison between numerical simulations and model (joint A).
3.10.a.FS
3:i0.b.FP
Figure 3.10 - Comparison between numerical simulations and model (joint B).
The reproach which may be addressed to the multi-linear model is the sudden and
unrealistic modification of the stiffness at the intersection of two zones characterized by a
constant stiffness, as well as the gap, sometimes important, between the actual curve and
the model in the range of moderate rotations. These flaws are inherent in the multi-linear
model. The introduction of the defined characteristics of the shear curves in an adequate
mathematical expression allows to obtain full non-linear curves with a more continuous
shape [1] (see annex 1).
The validity of the proposed model for steel bolted joints has been also checked by means
of comparisons with shear moment-rotation curves resulting from the experimental tests
on joints with extended end plate connections which have been performed in Lige five
years ago and which have been presented and discussed in [41].
46
3.6
The close agreement between the theoretical model and the experimental results has been
demonstrated. Two examples are shown on figure 3.11.
The bending resistance of the column flanges in unstiffened joints has been neglected
because of its small importance. It is however increased, for joints with end-plate
connections, by the bending resistance of the end-plate; this explains the actual but not
significant difference between the model and the experimental result for the test 010, the
end-plate thickness of which is relatively important (20 mm) in comparison with that of
the column flange (12,5 mm).
Finally the use of the proposed shear model for joints with bolted connections seems to
be quite justified.
M b lkNm)
MblkNm)
(0.01.124.58)
120-100
I0.07.79.68)
80
60
20
find.)
0.02
0.04
0.06
0.08
0.10
0.02
osa
O06
0.08
3.11.a. Test 01.

3.1 Lb. Test 010.
Figure 3.11 - Comparison between experimental results and model (joints with end-plate
connections).
3.1.2. Column web panels subject to load-introduction.
The conclusions of the parametric study relative to the load-introduction behaviour of
unstiffened column web panels are the following:
a) As for the column web panels in shear, the M-<|> curves characterizing the
load-introduction deformability depend on the actual loading of the joint (see figures
3.12 and 3.13 relative to the three joints defined infigures3.4 and 3.5).
b) It is not allowed to define the plastic capacity of a web subject to transverse loads as
this may be done for sheared column web panels: the propagation of the plasticity in
the column web transversally loaded does not end indeed in the apparition of an
horizontal yield plateau - when the strain-hardening is omitted in the numerical
simulation, but rather in the development, till the attainment of the ultimate buckling
load, of a zone characterized by a progressive increase of the resistance and the
deformability of the web (figure 3.14.a).
It will be consequently referred to the so-called pseudo-plastic moment, Mbppi, that is
associated to a limit state of the column web panel due to the effect of transverse loads
3.7
47
(figure 3.14.a). This characteristic load level may obviously be similarly defined for
the M-(|) curves relative to "T" joints (see figure 3.14.b).
c) The propagation of the plasticity in a web subject to transverse loads is not affected by
the presence of CJn stresses (see figure 3.7) in the web insofar as their maximum value
does not exceed a relatively high limit which should have to be explicitly determined.
il mm.
1.30
!
1
1
1
1
1
1
!
1
1
I
1
1.4
'
v^"*
1.X
Jt**M^*'fA
.so
i .
i>
!
i
!
i
!
i
!
i
!
i
!
I t'
r*
I.B
1.30
, _ . FS
i
t.K
1
1.M
1
1.90
1
!
!
!
!
!
!
! *
! .'" !
3.M
3.91
1
jl.M
1 MO
_ FP
Figure 3.12. - M-(j) characteristic curves (joint A).
4.30
:
!
1
:
!1
r ] " " ^t
mm m
3. SS
;
!
Â^L^
3.70
1.
(.30
t'
r
i
t
i
P
i
i
i
i
i
i
i
i
t
i
i
i
i
i
i
i
i
i
i
i
i
_ Fl
i
i
i
i
i
i
i
i
i
i
i
i
i
i
it.n
1
i.
1i.
i
i
i
i
i
i
i
. . FP
3.41
3.0
3.M
1 MO
Figure 3.13. - M-<|> characteristic curves (joint B).

This conclusion seems to confirm the result of an experimental study carried out in the
Netherlands [42] and which tends to show that the influence, on the "plastic capacity"
of the web, of o n stresses not exceeding 50 % of the column web yield stress is not
significant.
48
3.8
ZOETEMEIJER [42] proposes, for larger values of Gn (a n > 0.5 fy), to reduce the
"plastic capacity" by means of a factor e given by:
e =1.25 -0.5
(3.3)
KATO considers for his own [43] that the attainment of stresses a n greater than 0.5 fy
is not of practical interest because of the relative low values of the loads transmitted to
the column in order to prevent instability.
" l b initially Hal ( w i t h
stramhardvmnql
t>ppl
Krai af th i t i t l i w i a di ' th
Itrainhardtning f th who4
thard p i n c i .
'b initially I M I lUllwith
lttaui*hmrdnAg|
3.14.a. "cross" joint.

3.14.b. "T" joint.
Figure 3.14. - Definition of the pseudo-plastic capacity of a web subject to transverse
loads.
d) The amplitude of the column web out-of-flatness influences only the shape of the M-<|>
curves of joints whose collapse is linked up to the buckling of the web; this initial outof-flatness affects the value of the web ultimate buckling load in a significant way but
modifies very slightly the deformability of the web as far as the coUapse load is not
reached (figure 3.14.a).
e) The comparison (figure 3.15) of M-<|> curves relative to a "T" joint and to the
corresponding (same column and same type of beam) "cross" joint symmetrically
shows clearly the similarity of both web behaviours in the elastic range (it is not
possible to compare the curves in the non elastic range on account of the different
stresses interacting in the column webs).
This leads to the conclusion that the introduction of transverse loads in a column web
constitutes, as far as the stability of the web is not concerned with, a local phenomena
limited to the vicinity of the column flanges. The influence of the joint loading on the
shape of the M-<J> curves - as discussed in (a) - is seen to be very significant in this case.
3.9
49
19.a
f
to
!
,
!
,
.vT
13.
s. o
" '
*
!
i
'_J
i
im^
m**
i m
i
tj
1
!
i
!
,
!
,
!
,
;'i
ri
!
i
i
i
1
!
!
i
l
i
1
!
!
i
l
i
1
^ * ^ i
^ i
*
!
i
l
i
1
!
.
i
l
i
1
f.
CONFIGURATION
"T"
3.0
.
i
0.05
0.1
O.IS
.1
0.2S
CRUCIFORME
! *
0.1
IRO
Figure 3.15. - Comparisons of M-<|> deformability curves (joint A-FS).

Theoretical investigations have led to the development of a mathematical model for the
prediction of the load-introductionresponseof unstiffened column web panels (see figure
3.16). Its characteristic values are given in annex 1[1], but also in [34] and [35].
The model has been validated for welded joints [1, 34, 35] by comparisons with
numerical simulations (see for instancefigure3.17). Its extension to web panels in bolted
and composite joints (often very limited modification) is justified in the submentioned
references.
Mbppl -
Mbe
<Ie
>st
<pu
Figure 3.16. - Shape of the load-introduction model.
50
3.10
D M
if
1.30
1.34
1.3t
O.SO
o.st
M - P H I Curves
.a
'
Numerical
Model
! SM
!.
t.M
i.oo
i.
a.ot
a.so
t CAO
3.17.a. FS
1.20
10
113
/ ^ " ^
^__
i
'
0.9O
O.SO
M-PHT Curves
t.30
,
i
'
O.SO
1
1.00
1
! .
i
1,S1
1.90
1
Numerical
Model
10
3.M
1 RAO
3.17.b. FP
3.11
51
1.48
'
'
'
i.
1.3i
.ss
1
!_ J
t.m
M - PHI Curves
, Numerical
.M
!
l
1.31
'
1.M
22 1
!
1.9*
1
2.0
1
!
!
!
<
Model
3.9
1 MO
3.17.C. MP
Figure 3.17. - Comparison between numerical simulations and the model for loadintroduction deformability (joint A)
It has to be mentioned the formula for the evaluation of the ultimate resistance Mbu cover
the collapse of the column web by:
- excess of yielding;
- buckling (figure 3.18.a);
- crippling (figure 3.18.b).
As for the sheared column web panels some of the characteristic values reported in figure
3.16 may be integrated into a non-linear formulation in order to obtain a more continuous
deformability curve.
(a)
(b)
Figure 3.18. - Instability modes for the column web: buckling (a) and crippling (b).
52
3.12
3.2. Behaviour of steel cleated connections.

Two types of steel joints with cleated connections have been tested in laboratory:
- connections with a lower flange cleat and a web cleat;
- connections with at lower and an upper flange cleat, and also web cleat.
The results of the experimental tests on joints with two cleats have been reported in the
previous chapter. It has been shown that the maximum moment carried over by the
connections is very limited, and represents a small percentage of the plastic moment of
the beam. The introduction of their characteristic deformability curves in the classification
diagram proposed in chapter 6 of Eurocode 3 shows clearly (see 2.7.3.1. chapter 2), that
these connections may be considered as pinned. In these connections, it appears to be not
really necessary to develop mathematical models for the prediction of their main
characteristics of resistance and deformability. The account of their semi-rigid behaviour
in a frame design would complicated the calculations and would not be of interest from an
economical point of view.
In practice, we advise to consider these connections as perfectly pinned, because of their
limited resistance and of their almost infinite rotation capacity.
The adding of a third cleat connecting the beam upper flange to the column increases
considerably the resistance and deformability characteristics of the connections.
Figure 2.21 in chapter 2 highlights the semi-rigid and partial strength behaviour of the
connection with 3 cleats and justify the development of mathematical models for
prediction.
The mathematical model proposed by JASPART for the theoretical prediction of the semirigid behaviour of joints with flange cleated connections (with a lower and an upper
flange cleat) is presented in annex 1 [1]; it has been validated by means of numerous
comparisons with experimental tests performed in different research centres (Lige,
Trento, Sheffield, Hamburg). Its application to steel connections with three cleats of 10
mm which have been tested in laboratory leads to the results reported in Figures 3.19.a to
3.19.C.
3.13
53
I!?I_?fi?_:_J9INT
ti
RELATIVE ROTATION CURVE
IcNra
70.00
M-PHiriiivp. s
Lab. test
Model
3.19.a.-Test24T3C1.
TEST 30T3CI JOINT RELATIVE ROTATION CURVE
M
kNn
2
10
1 .20
1.00
0.80
i
i
!i _
i
i
i
i
i
i
\--^c^X-\
i / i vî
1
1
1
1
y^\
\i
1
1
1
1
11
II
/I
X S ^
iKfJr~
1
/ \S\
i
1
1
i
1
1
i
1
I
i
1 /l
|
/|
i / i
'
'
'
'
'
' / '
i//
1/
11/
17
I/A
1.00
2.00
3.00
\yr
11
Jr
y / \
0.20
!i
/T
0.60
0140_
_i
i
4.00
5.00
6.00
M-PHICurves
Lab. test
Model
'0
7.00
! f
3.19.b. - Test 30T3C1.
54
3.14
TEST 36T3C1 JOINT RELATIVE ROTATION CURVE

l
1.50
1 .25
k Nm
2
10
1
1
J_
J
1
1
1
\j
-4--1
_|
I/I
1 .00
T
li/
fif
0.75
0.50
li
In
7/1
i/ i
r i
J
// i
// i
0.25
// i
//
j
17
|2 00
\A 0 0
'
'
o 00
i
i
ir
8. 30
"
i_
M - PHI Curves
Lab. test
Model
1
'
10.00
IJAD
3.19.C.-Test36T3C.l
Figure 3.19. - Comparison between experimental results and the theoretical model. "T"
joints - steel connection with 3 cleats.
The non-linear M-<t curves obtained by means of the theoretical model forflangecleated
connections (eithout a web cleat) underestimate the actual resistance of the connections
with 3 cleats. This discrepancy is linked to the fact that the web cleat is not accounted for
in the theoretical model. However, it appears, from the comparisons of figure 3.19, that
the difference between the actual and predicted curves is particularly limited. In these
conditions, it may be asked if it is really necessary, for daily practice, to complicate the
theoretical approach by considering the non-linear behaviour of a supplementary
connecting element, the resistance of which is not significant
For what concerns the submentioned model for the prediction of the response of flange
cleated connections (fully described in annex 1), we will merely point out here that it
requires:
- the evaluation of the initial and strain-hardening stiffnesses K and Kst as well as that
of the resistances Mv and Mu (figure 3.20);
- the introduction of these characteristics in a non-linear formulations what allows to
obtain non-linear connection deformability curves.
3.15
55
Mu _
Mv
<D
Figure 3.20 - Main characteristics of connection deformability.

Figure 3.21 presents for instance the application of this procedure to two flange cleated
connections tested in Lige some years ago [41]. In these diagram, the theoretical model
has to be compared to the experimental curve "without slips". A a matter of fact, the
growth of visible slips between the beam and the cleats during the test is only depending
on the initial position of the bolts in their holes. To take account of these slips in the
theoretical curves requires only the translation of the curve position corresponding to
bending moments higher than Mgi (bending moment of which slips appear).
A similar prediction model is also available in [1] for extended end plate connections
which are not within the scope of the present research.
MbPHI Curves:
Theoretical
0.12
.a
t.M
Experimental
Experimental without slip
"
3.21.a. - Test 06.
56
3.16
Mh-PHI Curvas:
Theoretical

Experimental
*Experimental without slip
3.21.b. - Test 012.

Figure 3.21. - Non-linear modelings of tests onflangecleated connections performed in
Lige [41].
3.3. Improvement and extending of Eurocode 3 rules.
The analysis and the design of building frames by means of sophisticated non-linear
programs require an accurate prediction of all the deformability components of beam-tocolumn joints, whereas only some simplified characteristics of the actual joint behaviour
such as the second or the initial stiffness and the plastic capacity are needed in the daily
design practice.
Formulae for the assessment of (figure 3.22.)
- the design resistance Mv (Vny);
- the second stiffness Kbs (Kns)
for:
- sheared column web panel;
- unstiffened column web subject to transverse compression and tensile forces
Goad-introduction);
- end plate connection,
are proposed in this respect in the annex J of the Eurocode 3.
Work performed in Lige has led to the improvement of this formula as well as to thenextending to another type of steel connections (flange cleated connections and,
consequently,flangeconnections with a web cleat).
3.17
57
3.3.1. Column web panels subject to shear and to transverse compression and tensile
forces.
The EC3 formula for the assessment of the plastic capacity of column web panels in
joints with welded or bolted (end plate or cleats) connections lead in many cases to
unacceptable over-estimating or under-estimating of the plastic capacity.
New improved formulae have been presented and discussed in annex 1 [1] and in [35]
(V ny and Mbpgi = M v ) in figures 3.8 and 3.16. They have been validated by means of
comparisons with experimental and numerical test results.
The EC3 formulae lead similarly to over-or-under estimating of the secant stiffness of the
column web panels. The new ones that we suggest are based on the study of the available
experimental and numerical test results which show that the secant stiffness KbS (Kns)
may be defined in the following way (figure 3.22):
Kns - Kni / n v
Kbs = Kbi / n v
(3.4)
where Kbi (Kn) represents the initial stiffness of the column web panel subject to
transverse loads (to shear).
For column web panels subject to shear or to transverse loads, n v may be defined as:
n v = 3,0
(3.5)
Mv (Vny) ~^<=~~1~~\
I
iKbs(Khs)
*W
Oi fy) i>v = nv <I>i (yv = nv y)
Figure 3.22 - Definition of the design resistance M v (V ny ), the secant stiffness Ks (K ns ).
The initial stiffness Kbi(Kni) may be easily evaluated by means of formulae presented in
the [1,34,35] for welded joints and extended in [1] and [35] to bolted joints.
Figures 3.23 and 3.24 highlight the accuracy of the new proposed formulae for two shear
curves resulting from experimental tests on joints with end plate connections [41] and for
two load-introduction curves resulting from numerical simulations of tests on joints with
welded connections.
58
3.18
ESSAI Ol i ROTATION DANS L'AME DE LA COLONNE

M..
0.07.
79.68).
Lira. TUT
TUO Co n o
llLJJll
Hott
(VIsCiftf3
+
interval la X
Jntarvalla Y
0 . 0 2 RAO.
10 kNm.
ROT. REL..
3.23.a. - Test Ol [41].
ESSAI 010 COURBE H - ROTATION OANS L'AME DE LA COLONNE

M..
C 0.04. 124.56).
LX. TJT
FteioM ai-ivi/t.
Hooit
_ _ _ _
f*OlO&4
ROT. REL.
intarvoll X
interval la Y
0 . 0 2 RAO.
10 Min.
3.23.. - Test 010 [41].

Figure 3.23. - Comparisons of bi-linear models with experimental shear deformability
curves.
3.19
59
NOEUD D E RIVE NR
NM
10 !
1.30
i'34
(FS1
J!
1.20
l \ \ ' \
.0.50
o.ra
IILL LJ
m\.
1 !
.
- -yy ; J~~
COURBES Mb - PHI
a.30
\i
0.30
1
t.00
1
1.39
1
3.00
i
2.30
| RAO
_ _ _
NUMERIQUE
MODELE
MoPoSED
OiLltJtftft HootX
3.24.a. - Joint A - FS (see figures 3.4 and 3.5).
N0EU0 DE RIVE NR4 (FP)

**""
1.20
io
1.13
\'
l
1
i
1
0.30
/ I
o.n
o.x
.!__.
i\
L<
//I i I !
LI
s
30
1.00
I
COURBES Mb - PHI
a
"a
s
1.30
1
NUMERIQUE
MODELE
PRorosco
Bi-MAR. MOCCI.
3.00
1 RAO
3.24.b. - Joint B - FP (see figures 3.4 and 3.5).

Figure 3.24. - Comparisons of bi-linear models with load-introduction deformability
curves.
60
3.20
3.3.2. Flange cleated connections.

3.3.2.1. Plastic capacity.
The critical part of a flange cleated connection is usually the tensile cleat and the adjacent
zone (bolts in tensions and column flange). Mainly two methods for assessing the
pseudo-plastic capacity of the tensile zone are available [44,45]. Both have been used for
a comparison with fully instrumented test results performed in Trento [46] on isolated
cleats and in Lige [41], Sheffield [47] and Hamburg [45] on cleated connections. It
results that Kishi's method is largely unconservative and Hotz's one much too
conservative. The large discrepancy is due to specific problems - linked to the cleated
connection - which are not properly accounted for by both methods. It follows from
experimental observations that [1]:
- The collapse of a cleated connection by formation of a plastic mechanism involves a
three-yield lines mechanism (two in the tensile cleat, one in the compression cleat);
- The initial clearance between the beam end and the column flange is likely to change the
location of one yield line in the tensile cleat; the latter always develops at the toe of the
cleat fillet, at one time in the vertical leg, at one tine in the horizontal leg.
In addition the sole way to define the experimental plastic capacity of a cleated connection
is not likely to correspond to the lowest strength of the connection components [1]. Let
us just mention that the degree of bolt preloading and the onset of an appreciable
membrane action are the main reasons.
A fully original approach has been developed - similar to that developed in [1] for endplate connections - which accounts for the more accurate location of the yield lines, the
sizes of the bolts/or washers, the bolt preloading and the plastic mechanism of the
connection in its whole. As a result a set of design formulae have been suggested, which
are relative to all the possible collapse modes; bolts fracture, mixed plastic mechanism in
the whole connection, and yield lines mechanism either in the cleats and in the column
flange [1].
The theoretical results computed in accordance with this new approach have been
compared with test data. It may be conclude that the method, that is founded on
sufficiently simple formulae to allow for the daily practice, has a wide range of
application and a very good accuracy. It could constitute an improvement to EC3, the
annex J of which is restricted to end-plate connections only.
3.3.2.2. Secant stiffness.
The definition of the secant stiffness is similar to that adopted for the column web panels
(figure 3.22).
Ks = K i / n v
(3.6)
Simplified formulae for the assessment of K, are proposed in annex 1 [1]. The coefficient
n v is expresses as:
n v = 3,0 if (Xg > 1
= -llcxg+14if0,20ag<l
(3.7.a)
(3.7.b)
(Xg.the value of the ratio between the bending moment Mgi in the beam for which slips
occur between the beam and the cleats and the plastic capacity of the connection,Mv :
3.21
61
ctg = Mgi / Mv
(3.8)
On base of the test results, it is suggested to neglect the prestressing of bolts when the
latter is less than 20 % of the plastic capacity of the connection (Og < 0,20).
In these conditions, the secant stiffness is defined as the ratio between the initial stiffness
of the connection, the bolts of which are considered as not prestressed, by the coefficient
nv, the value of which is:
nv = 3,5
(3.9)
3.3.2.3. Comparisons between proposed bi-linear model and experimental results.

A lot of comparisons between the proposed bi-linear model and experimental results of
tests performed in lige, Trento, Sheffield and Hamburg have been presented in [1].
Two of these comparisons are reported here (figure 3.25). The agreement between theory
and experimentation appears to be quite satisfactory in all cases.
m.u
74.43
.*"""
4.M
- **'
1/
a.M
Mb-PHI Run,.
1/
1
Bl-linear
il
>--- Experimental without slip
11
I.M
(.M
.M
.M
Experimental
.M
3.25.a. - Test 03 [41].
62
3.22
Mb-PHI Curva :
Bilinear

Experimental
Experimental
D Experimental without slip
Experimental without slip
3.25.b. - Test JT08 [47].

Figure 3.25. - Comparison between experimental results and proposed bi-linear model
(flange cleated connections).
The set of design formula forflangecleated connections, presented here and detailed in
[7], are found likely to be used as the background for a possible future additional Annex
of EC3 chapter 6 devoted to this type of connection.
3.3.3. Flange and web cleated connections.
The theoretical models for the non-linear prediction of the behaviour of flange cleated
connection have been compared in section 3.2. to experimental results of tests on flange
and web cleated connections. From these comparisons, it seems to appear that the
influence of the web cleat on the overall behaviour of the connection could be neglected.
On base of this conclusion, we have decided to apply also the bi-linear model for flange
cleated connections (characterized by the second stiffness and the plastic capacity) to tests
on connections with flange and web cleats. Figure 3.26. presents this comparison for
tests 24T3CI, 30TC1 and 36T3C1 (see chapter 2). The apparent low level of plastic
capacity does not result from a lake of accuracy of the related proposed formulae: the
supplementary strength associated to the behaviour of the actual connections is due to
strain-hardening but also - and mainly - to membranar effect linked to the important
deformability of the cleats (see [1]), the plastic capacity of which is for lower than that of
the other parts of the connection (column flange, column web in tension and
comparison,...).
This influence of membranar effects does not appear so strongly infigure3.25 relative to
flange cleated connections because of the similarity of the plastic capacities associated in
both cases to the collapse of the cleats and of the column flange.
3.23
63
TES'
2T3Ci
>Ki
'
JO NT P.L A T 1 V
eo.co
V wO
1
1
1
1
1
1
1
1
1
1
1
1
\ ^ \ \
ZiZ
1
1
1
O.CO
30.00
2000
! 1
rrirv"
7
10.00
CJSV
!
70.00
ROTATION
" 7 "
. . . . : . . . . .
f !
7M
">
il
_ _ _ /
J2 00
1
00
1
1
Hi
/i
i 6.30
1
1
i
1
8.
00
CUP.V;
JOINT
2
'
11C co
ÂO
3.26.a. - Test 24T3C1.

IE.! J 0 , N T "ELATIVE ROTATION CURVE
M IN.
r.
HI CURVE
JOINT
3.26.b. - Test 30T3C1.
64
3.24
ST 2 5 I ? 5 i . I . J O I N T
I .50
E L A T I VE ROTATION CURVE
10
25
/
/
c:
HI CURV
0.25
2.30
i
I^.OO
I
JOINT
i
16- 30
I
I
>5.3C
0.33
RAO
3.26.C. - Test 36T3C1.

Figure 3.26. - Comparison between experimental results and bi-linear model (flange and
web cleated connections).
3.4. Behaviour of composite connections.
3.4.1. General description of the model.
The shear stud connectors, that are welded on the upperflangesof the steel beams in the
experimentally tested composite joints, have been designed in order to obtain a full
interaction between the steel members and the reinforced concrete slabs. The resulting
absence of slips allows to assume the in-plane indeformability of the end sections of the
composite girders (fig. 3.27.a). The simplified structural model (fig. 3.27.b) which has
been developed to predict theoretically the non-linear response of the connections is based
on this assumption. This model consists in an infinitely rigid beam (the end section of the
girders) lying on an elastic-plastic foundation materialized by axial springs simulating
respectively the deformation and the resistance:
- of the reinforcement bars;
- of the concrete;
- in the zone of the upper flange cleat, of the web cleat and of the lower flange cleat:
bolts in shear, in tension and in bearing, column flange, column web, cleat, slips
between cleat and steel beam, ovalization of bolt holes.
Each of these springs is characterized by a specified non-linear force-displacement curve
which has to be predicted as reliably as possible. Methods of modelling have been
proposed for each source of deformability [1]. The associated formula as well as the
needed data are given with details in the annex3.
3.25
65
The moment-rotation curve corresponding to a particular composite connection is built up

step by step by distributing, at each level of bending moment, the loads between the
springs according to their actual relative stiffness and by evaluating the associated rotation
of the infinitely rigid beam.
A program, available in the form of afloppydisk in annex 3 of this chapter allows the
user to built up easily the required M-(j) connection deformability curve.
r-y
concrete
:
*", lower rebars
upper flange cleat
web cleat
lowerflangecleat
a) actual composite connection

b) equivalent structural model
Figure 3.27. - Definition of the theoretical model for connection behaviour prediction.
This procedure has been applied in figure 3.28. to two composite connections only
differentiated by the percentage of reinforcement.
According to the initial relative position of the bolts in their holes, the slip between the
cleats and the beam may or not occur during the connection loading, what justifies the
necessity to report two different theoretical curves corresponding respectively to the
development of the maximum permissible slip (which depends to the hole clearances) and
to the absence of slip. The influence of this parameter on the connection deformability is
seen to be relatively significant
i
i
i
i
i
i
i
i
i
7
ta
7
10
i
i
i
?
1
'
'
1
1
i
i
j
19.00
N M I curva
ncnsTiCM.
IM .IFS
J--L
iri
J0_.0O
nour t u n
10.00
1
<
1
* - m*t curvo
ncatniCAL
I11M JFt
. . fina/r
tun
^ _ OOWllfDffiU.
CfftXIMHTAL
9.OB ,
1
1
1
:. 01
0 02
!<.03
iO.04
0.09
JM)
i0
01
O.03
0.03
0.04
,0.09
KA0
i
i
i
a) rebars of 10 mm
b) rebars of 18 mm
Figure 3.28. - Comparison between experimental results and the response of the
equivalent structural model (composite guders with IPE 300 beams - no upper cleat
150 x 90 x 13 cleats).
66
3.26
A close agreement between the experimental curve and the response of the theoretical
model is obtained in the first example where the collapse corresponds to the yielding of
the rebars, as well in the second one, except for what concerns the prediction of the
ultimate resistance associated, in this case, to the buckling of the column web.
This safe but important divergence may be explained by the very low out-of-plane initial
imperfection of the column web measured on the specimens tested in laboratory in
comparison with that, chosen on base of rolling tolerances [3.3.], which has been
considered for the assessment of the theoretical buckling load of the web.
The interest of such an equivalent structural model is fourfold:
- to validate the individual mathematical models developed for each source of flexibility
of the composite connections: cleats, introduction of loads in column webs,...;
- to represent a valuable tool in view of intensive parametric studies;
- to constitute a foundation for further developments of a more simplified and practical
approach for the evaluation of the deformability and resistance characteristics of the
composite connections.
- to supply the non-linear finite element program FINELG which will be presented in
next chapter with non linear connection curves.
3.4.2. Characterization of the non-linear springs
The annex 3 of this final report contains a program to calculate the moment rotation
curves of cruciform composite joints with cleated connections. The tests performed in the
frame of this research are compared to this model and presented in this annex.
3.27
67
4. Numerical developments.
4.1. Description of a new finite element.
FINELG is a materially and geometrically non linear finite element program which has
been developed jointly at the University of Lige, Belgium, and at Sie Polytechnic
Federal School of Lausanne, Switzerland. It is used to solve problems such as:
- step-by-step structural response up to and beyond collapse;
- linear and non-linear instability with calculation of critical loads and instability
modes;
- calculation of eigen frequencies and eigen modes, possibly taking into account the
current stress state.
Its library is composed of spatial truss bar, plane beam, spatial beam, membrane plate,
thin and thick shells, springs, linear constraints,...
FINELG program has been implemented to simulate accurately the non-linear
behaviour of connections and sheared column web panels (see annex 1)[1].
The flexural behaviour of the connection and of the adjacent beam as well as that of the
column web panel are gathered into a single "plane beam + connection + sheared panel"
finite element. Any type of non-linear response may be associated to the behaviour of
the beam, of the connection and of the web panel respectively.
The main features of the "beam + connection spring + sheared web panel" finite
element are the following:
Beam:
1) The beam cross-section is constant and not distorted along the axis of the element.
2) Hollow rectangle, full or hollow circular, H or I-shape bent about strong or weak
axis, T-shape cross-section as well as composed cross-sections, the geometry of
which is defined by the user are available.
3) The residual stresses from rolling, cold forming, welding,... as well as the initial
geometrical deformation of the element along its axis are considered in the
computation.
4) The theory of non-linear analysis is used to take account of the instability
phenomena.
Connection:
5) The connection is represented by a rational spring with a non-linear constitutive
law.
6) The spring has a length equal to zero.
7) The spring is located at the interface between the column and the beam, what
corresponds to its actual position in a frame (A in figure 4.1.).
8) The flexural behaviour of the spring is expressed by:
dM = K()(d<!>b-d<i>s)
(4.1)
when M is the bending moment carried over by the connection, K(> is stiffness (4.1.),
function of the relative rotation (<>D - <t>s), if <>
| s and <b represent respectively the total
rotations of the rest of the structure and of the beam.
<>b is a "internal DOF"; it is eliminated by static condensation.
4.1
69
Sheared web panel:

9) Thefiniteelement accounts for the actual dimension of the web panels.
10) The loading of the panel in shear is expressed by (figure 4.2.):
Vn = (Mbl + Mb2) / db - (V c l + Vc2)
and not, as in (ref. 2, ref. 3, ref. 4 - figure 4.3) by:
Vn = (Mbl + Mb2)/db
(4.2.)
(4.3.)
Formula (4.3) constitutes only a rough approximation of the actual shear loading
giving by (4.2.).
I
beam
structure /
Figure 4.1. - Modeling of the semi-rigid connections.
Mclf
Nel
Vbl Mbl
Mb2 Vb2
db
Vc2.
Nc2
H de H
Figure 4.2. - Loading of a column web panel.
70
4.2
Ml
Ml
Mr
STUTZKI[2]
Figure 4.3.a.
TSCHEMERNEGG [3J
Figure 4.3.b.
Mr
ATAMAZ [4]
Figure 4.3 .C.
Figure 4.3. - Previously proposed elements for the numerical modeling of sheared
column web panels.
4.3
71
A forementioned element can be used according to one of the three different manners
sketched in figure 4.4.
Whatever the mode of use may be, the finite element "beam + connection + sheared
panel" only requires the definition of two nodes (i and k in figure 4.4.) at his both
extremities.
! - *
a. - "Beam" element.
i -
I k
&
b. - "Beam + connection" element.
!- k
c. - "Beam + connection + shear web panel" element.

Figure 4.4. - Use of the finite element.
It exhibits several superiorities:
- in contrast to many other approaches [2, 3, 4] (figure 4.3.), it allows to fulfil the
equilibrium equations of the web panel;
- it makes possible an accurate and realistic picture of the actual macroscipic behaviour
of the column web panel;
- it enables a direct use of the deformability curves obtained experimentally or by
means of theoretical prediction models. TSCHEMERNEGG's and STUZZKTS
models (figures 4.a and 4.b) implies the "translation" of M-8 curves into N-u curves
characterizing the axial behaviour of trusses or springs.
- It allows to refer to the same finite element mesh of the frame than that which would
be used for a fully rigid frame (by simply referring to the beam and column axis as
well as to their interaction points).
- It leads consequently to a large decrease of the number of equations which have to be
solved at each analysis iteration (The account of the non-linear behaviour of the
connection and of the web panel in the beam element does not imply, as for the
previously suggested models, the definition of specific finite elements).
- In combination with the "concrete" or "reinforced concrete" element [5], the "beam +
spring + web panel" element allows to simulate the response of composite frames by
taking account of phenomena such is the shrinkage, the creep or the non-linear
72
4.4
interaction between steel and concrete components and also to the different stages
during the frame construction.
4.2. Description of the parametric study.
For daily practice, it cannot be expected to account separately for both the flexural
behaviour of the connection and the sheared behaviour of the column w e b panel.
Therefore the possibility of concentrating the deformability of both connection and
column web panel into a flexural spring located at the beam end must be contemplated
(figures 4.5. and 4.6.). As a matter of fact, how such springs affect the frame response
can be reflected through appropriate design method (see annex 1 [1] and section 5 of
this final report.)
Mb
Mb.i
Mb.i
L^
Mb,i
<t>
a - Connection
b - Sheared panel
c - Spring
Figure 4.5. - Flexural characteristics of the springs.
Mb,g
(Y+<t>)
J
(Y + <t>0
V -
V
Mb,g
Mb,d
(()r
(Y+<W
-<@
V
Mb,d
a - Exterior joint.
b - Interior j o i n t
Figure 4.6. - Concentration of the joint deformability into flexural springs.
4.5
73
Regarding the location of the spring, two possibilities arise (figure 4.7.):
- either at the beam-to-column physical interface (points A), or
- at the intersection of both beam and column axes (points B).
The optimum location of the spring as well as the allowance for summing up the joint
deformability components cannot be demonstrated theoretically. The parametric study,
which consists in the numerical simulation of the behaviour up to collapse of braced
and unbraced steel frames with semi-rigid joints, is consequently aimed at determining
to which extent the actual and relatively complex behaviour of a joint (shear panel + 1
or 2 connections) may be represented, with a sufficient accuracy, by isolated springs of
appropriate characteristics.
Figure 4.7. - Possible locations for the springs.

The main characteristics of the frames used for the parametric study are presented in
Table 4.1. Geometry and loading pattern are reported in figure 4.8. Two types of
connections commonly used in practice are considered: flange cleated and end-plate
connections. Their non-linear deformability characteristics are given in annex 2. AU the
connections of a specific frame are presumed identical The residual stresses, the elastoplastic behaviour of steel and the initial deformed shape of the frames are taken into
account in the numerical simulations.
74
4.6
Type o trame
%> - $gor4JB)
'
fii *
f..
Type
Ooluror
Height between
beantXe&i*}
Type
8eai?tf$jr
Spa between
column axes (m]
4-'
braced
unbraced
braced
HE160B
HE160B
HE160B
8.0
6.0
7.0
IPE200
IPE200
Nomen"
dature
Typeofconnec
oris.
FGllangcfeatt
EP*end plate
loading {figure 4.&

PZ
\ m
im
fm
Denomi*
nailon
AL.1
A1
FC
8.242
397
A2
FC
10.303
A3
FC
12.363
108
AL.3
A4
EP
12.363
108
AL.4
left beam: 10.0
B1
FC
8.863
450
28
BL.1
right beam: 8.0
B2
FC
2.037
32
BL.2
B3
FC
4.050
36
24
BL.3
B4
EP
4.050
36
24
BL.4
C1
FC
14.305
120
CL.1
C2
FC
18.392
CL.2
C3
FC
41.537
10
CL.3
C4
EP
41.537
10
CL.4
D1
FC
11.245
160
DL.1
D2
FC
11.245
DL.2
D3
FC
41.178
D4
EP
41.178
5.0
AL.2
5.0
IPE300
:
unbraced
<
braced
HE160B
HE140B
12.0
5.0
IPE300
IPE270
12.0
6.0
,
unbraced
HE200B
4.0
IPE300
5.0
^ ' ^
J*,
VC
braced
5.0
IPE220
5.0
b
unbraced
HE120B
HE160A
4.0
IPE300
DL.3
DL.4
5.0
v
Table 4.1. - Data of the numerical simulations
PI
|P1
PI
PI I
PI
PI
..EjJTTTTto
PI
11//1111/
u/1111//
//IIII//I
4.8.b - Type C frame
4.8.a - Type A frame
PI
..i'JTTTT}<
Willi//
PI
. P2
TTTTV
m/inii
111//1111
1i/ii 11 ir
4.8.C - Type B frame
76
4.8
,P1
,P1
, PI
,P1
PI
PI
PI
,PI
--;T ^ rrfTTt JTTTT1; JTTTRi^
4.8.d - Type D frame

Figure 4.8. - Different types of braced and unbraced frames
4.9
77
Three numerical modelings of joints are examined:

a. the deformabilities of both connection and sheared web panel are represented
separately - type M3 (because being the most sophisticated modeling, the case is
used as reference).
connection
"-
-L
sheared web panel

b. the deformabilities of both connection and sheared web panel are concentrated into a
single spring located at the beam-to-column physical interface - type M4
axes of the connected members
infinitely rigid beam

spring
c. the deformabilities of both connection and sheared web panel are concentrated into a
single spring located at the intersection of beam and column axes - type M5.
78
4.10
4.3. Conclusions of the parametric study.

Preliminary calculations (carried out with the aim to define the program for the
parametric study) have indicated that the deformability of the joints may be
concentrated at the beam-to-column interface (points A infigure4.7.), the model M4.
In view to confirm these results, the numerical simulations relative to M3 and M4
modelings have been performed in the first instance.
The resulting curves are presented as follows:
- curves "load factor X vs. mid-span vertical displacement A for the beam, the transverse
displacement of which is the most important at each level of the frame loadings" for
braced and unbraced frames;
- curves "load factor X vs. horizontal displacement V at the top" for unbraced frames.
These curves allow to control the influence of the simplified numerical modeling M4
on:
- the displacement of the beams and of the frame under service loads (X, =1);
- the collapse load factor.
Two of these comparisons are plotted infigure4.9. Others are reported in annex 1.
M
2.5
2.2
2.0
1.5
1.0
exact
0.5
modelling
simplified
modelling
54.0
15.0
30.0
5.0
a. - Braced frame C
Loading CL.1
60.0
MM
MM
b. - Unbraced frame C
End-plate connections
Figure 4.9. - Comparisons between M3 and M4 modelings.

Except for type B unbraced frame with end-plate connections, the agreement between
the curves relative respectively to "exact" modeling M3 and simplified one M4 is
almost perfect. The physical explanation of the underestimating of the collapse load
factor resulting from the use of M4 modeling for the unbraced frame B (figure 4.10.) is
detailed in annex 1 [1]. This kind of discrepancy - anyway on the conservative side - is
linked to the high stiffness and resistance of the connections in the zone Mb > 110 kNm
(seefigure4.11.) compared to those of the corresponding sheared column web panel.
4.11
79
,,Mb(kNm)
Connection
1U.0T
Shear of the
web panel
108.0"
72.0-
exact
modelling
simplified
36.0-
modelling
j
15.0
30.0
45.0
60.0
Figure 4.10. - Comparison relative to

unbraced trame B with end-plate
connections.
y.<t>(10" 2 rad)
10'
-v
MM
3.0
6.0
1-
9.0
Figure 4.11. - Characteristics of the

joint with end-plate connection
relative to the unbraced frame B.
It has been shown that the use of simplified numerical modeling M 4 leads to an
accurate prediction of the actual response of braced and unbraced frames, except when
the beam-to-column connections are almost fully rigid. In the latter case, the actual
frame collapse load would be somewhat underestimated. This last remark is not very
restrictive in practice: few connections are reallyrigid,on one hand, and the fabrication
of such connection is often very expensive, on the other hand.
Numerical simulations using M5 simplified modeling have not been performed. They
could not lead to better results than those obtained with M4 modeling and, furthermore,
it is while stressing that M4 simplified modeling is fully representative of the actual
joint behaviour when the column web panel is stiffened for shear (or when very flexible
connection are used). The joint deformability then consists in the sole connection
deformability which is concentrated at the beam-to-column physical interface, in
complete accordance with M4 modeling.
80
4.12
5. Simplified methods for the design of steel structures with semi-rieid

joints- Usually, the multi-storey frames are subdivided in two categories:
i. The braces frames in which the bracing system is constituted by rigid elements able
to support the vertical actions, including those resulting from imperfections and
second order effects.
Eurocode 3 stipulates that a bracing system is efficient if it is able to reduce from at
least 80 % the horizontal displacements of the frame. When this condition is not
fulfilled, the frame has now to be considered as laterally fixed.
ii. The unbraced frames which one their stability to the so-called frame effect; they
require consequently the use of resistant and relatively less flexible beam-to-column
joints.
Simplified design methods for braced and unbraced frames are presented in annex 1
[1] of this final report; they will be summarized hereunder in sections 5.3. and 5.4.
- For reasons of complexity, it cannot be envisaged, in these simplified methods, to take
account separately of the shear deformability of column web panels and of the
flexural deformability of connections. The idea, which consists in substituting a single
flexural spring to the set "shear column web panel + connection" is at the source of
the parametric study for the joint modeling which has been described in sections 4.2.
and 4.3. of this report.
On the occasion of this study, the accuracy of this simplification of the joint modeling
(as represented in figure 5.1.) has been highlighted. It will consequently be referred to
it in this section. The reader will remember that the deformability of the column web
and of the connection has to be concentrated in a single flexural spring acting at the
physical extremity of the beam and not at the intersection of the beam and column
axes.
I
infinitely rigid element
I
I
O
/
a \
beam element
deformability "sheared web panel + connection"
Figure 5.1. - Simplified modeling for beam-to-column joints adopted for the frame
design.
5.1. Simplifying assumptions.
5.1.1. Behaviour of steel.
The actual stress-strain diagram of steel is replaced by a bi-linear elastic perfectly
plastic diagram, which neglects the strain-hardening effect (figure 5.2.).
5.1
81
HOOKE
Figure 5.2. - Bi-linear modeling of the steel stress-strain diagram.

5.1.2. Behaviour of beam andcolumn cross-sections.
The classical bi-linear modeling of the M-moment x -curvature curve represented in
figure 5.3. is characterized by a first linear part, the slope of which corresponds to the
elastic linear flexural stiffness of the cross-section followed by a yield plateau of
infinite length.
This yield plateau is at the root of the plastic hinge concept which will allow to perform
a plastic design of the frames. However, it has to be noted that all the cross-sections are
not able to transfer this plastic moment and do not possess an infinite (or sufficient)
rotation capacity, because of the local instability phenomena, which are likely to appear
in the slender walls in compression.
The use of the bi-linear model proposed in figure 5.3. requires consequently a
preliminary check of its class in bending (see Eurocode 3).
The HE and IPE hot-rolled cross-section of usual steel grade belong generally to the
classes 1 (plastic cross-sections) and 2 (compact cross-sections), what justifies here the
use of the bi-linear model represented in figure 5.3. (recall: because of their limited
rotation capacity, class 2 profiles, which are able to transfer the plastic moment of the
cross-section, can however not been used for a plastic design).
. M
i,
Mp
Me
f\
EI/L
Figure 5.3. - Bi-linear modeling of the moment-curvature diagram for cross-sections in

bending.
82
5.2
5.1.3. Behaviour of beam-to-column joints.

The non-linear behaviour of the flexural spring which characterizes the joint response
cannot be taken into account in the design practice; the corresponding moment-rotation
curve has consequently to be idealized. One of the most simple idealizations to which it
may be referred is the elasc-perfectly plastic one (Figure 5.4.). This modeling has the
advantage to be quite similar to that used traditionally for beam and column sections
subject to bending (Figure 5.3.).
The moment corresponding to the yield plateau is the joint plastic capacitiy M v (called
design resistance in Eurocode 3). The constant stiffness which is usually recommended
(for instance in Eurocode 3) is the secant stiffness. Simple methods for the evaluation of
both values are suggested in annex 1 [1] for joints with end-plate and flange cleated
connections.
BDLAARD and ZOETEMEIJER assert in [7] that the use of this bi-linear idealization
leads to a safe estimation of the frame resistance and of the frame stability. Their
argumentation is however far from being satisfactory. The lack of theoretical
justification for this concept is the starting point for a study presented in annex 1 [1],
the aim of which is to check the accuracy and the safe character of bi-linear modeling
for the design of braced and unbraced frames. The conclusions of this study will be
briefly presented in sections 5.3 (for braced frames) and 5.4 (for unbraced frames).
M
Mv
Figure 5.4. - Bi-linear idealization usually adopted for joints.
5.3
83
5.2. Design requirements.

The design of a building frame is an iterating procedure, the successive stages of which
are reported here below.
Conditions to fulfill
Predesign
Check of the limit states?
Iterative
procedure
Modification of the frame

Successive steps for a design.
Sections 5.3. and 5.4. of this final report focus on the simplified methods, which have
been suggested at the University of Lige for the check of the limite-states for frames
with semi-rigid joints.
In reality, the check of the limit states is a two-step procedure which implies
successively:
- the check of the serviceability limit states under service loads to ensure that the frame
will comply with durability conditions over its whole presumed life duration;
- the check of the ultimate limit states under factored loads which, in accordance with
the philosophy of limit states, consists to check that the actual frame collapse load is
higher than the value of the design factored loads; the design factured loads are
obtained by multiplying the service loads by a coefficient (higher than 1.0), the value
of which depends on the nature of the loads and on their combination (see for instance
Eurocode 3).
It is while stressing that no restriction regarding the degree of plastification of structures
under service conditions is mentioned explicitely amongst the principles governing the
check of civil engineering structures. That seems to be somewhat contradictory with the
daily practice where, for obvious practical reasons, the frames are usually considered as
84
5.4
elastic under service loads. Some considerations reported in annex 1 [1] indicate that
this usual design principle may be economically applied to braced and unbraced frames
with semi-rigid joints.
As a consequence, the check of the frames under service loads will easily and
economically performed by referring to an elastic frame in which the bending moment
is less than the plastic moment in any cross-section (M < M p in Figure 5.3.) and in
which the bending moment is less than the design resistance m any beam-to-column
joint (M < Mv in Figure 5.4.).
5.3. Simplified design methods for braced frames.
- The check of the braced frames under service loads will be simply conducted by
referring to the displacements resulting from an elastic linear analysis taking into
account the behaviour of the beam-to-column joints through the use of their secant
stiffness. This analysis will be easily performed by means, for instance, of the slopedeflection method, as modified by JOHNSTON and MOUNT [8].
- The following sections describe the mode of application of the two usual design
philosophies - elastic and plastic - to braced frames with semi-rigid joints, with the
view to determine their collapse load.
5.3.1. Elastic design.
5.3.1.1. Design principles.
- The elastic design of a braces frame requires a first order elastic linear analysis in
order to determine the internal forces.
The extending to the analysis of frames with semi-rigid joints of classical elastic linear
methods such as the slope-deflection method and the moment-distribution method has
been introduced in 19842 by JOHNSTON and MOUNT [8].
The design in itself is achieved according to a "weak column-strong beam" criterion
[9, 10] which consists to design beams and joints in such a way that their collapse
never precedes that of the columns. The stability check of the whole frame is then
reduced to the individual check of columns by means of usual interaction formulae for
in plane or space loaded columns (see for instance Eurocode 3).
- The buckling length of an isolated column, useful to its stability check, may safely be
chosen equal to the column height, commonly termed "system length" [11]. As
columns form however part of the frame, a more accurate estimation of their carrying
capacity is obtained by considering a buckling length, termed effective length [11],
smaller than the system length. This reduction results from the presence of end
restraints due to the rest of the structure and particularly to the surrounding beams and
joints, whose elastic behaviour until frame collapse provides restraints with a constant
character.
5.3.1.2. Buckling length of linearly end-restrained columns.
- The formulae for the stability check of bent and compressed columns apply to
assumed isolated columns. Their application to actual columns in braced frames needs
the definition of an equivalent isolated and restrained column (Figure 5.5.). The effect
of restraints is revealed by the presence, at the column ends, of flexural springs, the
rigidity of which is defined in such a way that it equals that of the rest of the structure.
5.5
85
The determination of the effective buckling length of actual columns will result from
the study of corresponding isolated and restrained columns.
The main problem lies obviously in the evaluation of the flexural characteristic of
springs.
BJORHOVDE [12] limits the influence of the structure on the studied column to the
beams (and the corresponding joints) ending at the considered extremity (Figure 5.6.).
He proposes the following expression for the stiffness of the equivalent flexural
spring at each column extremity.
R=I^
where:
(5.1.)
2EIg
1+ KL
s g
E= YOUNG modulus;
Ig = stiffness of the beam ending at the considered extremity;
Lg = length of the beam ending at the considered extremity;
" s = secant stiffness of joints between beam and column.
fe
The summation extends to all the beams ending at the considered extremity.
This equation assumed that the beams of the substructure are bent in single curvature
with equal and opposite end rotations. It may be easily modified according to the
actual beam end conditions.
^
RB
Figure 5.6. - Isolated columns.
Figure 5.7. - BJORHOVDE's substructure.
The BJORHOVDE's approach is accurate for single-storey frames but generally too
safe - it may also be unsafe in some extreme cases - for multi-storey frames, as
explained in annex 1 [1], because of the influence of the upper and lower columns. In
[1], the concept of substructure has been extended accordingly (see figure 5.7.) and a
simple procedure has been proposed to take account of the lower and upper columns
and of their axial loading relatively to that of the studied column.
86
5.6
Figure 5.7. - JASPARTs substructure [1].

- The practical assessment of the effective buckling length of isolated and linearly endrestrained columns may be achieved by means of simplified formulae resulting from a
study of elastic linear stability or from the use of buckling curves for end-restrained
columns.
A survey of the main existing approaches, as well as an original buckling length
evaluation method for columns with different restraints at the ends are proposed in
[13].
5.3.1.3. Second-order effects.
SNIJDER, BULA ARD and STARK [11] have highlighted the possible importance of
second-order effects on the behaviour of braced frames. Indeed the compression axial
forces acting in the columns produce a decrease in flexural stiffness; this so-called "e
effect" has an influence on the bending moment diagram and may cause the premature
collapse of beams and/or joints, what results, for the columns, in a reduction of the
amount of restraint at theirs ends and in a modification of their loading.
According to BIJLAARD and SNIJDER, the influence of these second-order effects
could be neglected when:
- the beam span to column height ratio is larger than 1.0;
- the moment capacity of the beam is larger than that of the column.
However, studies performed in Lige [14, 15] have not allowed to confirm these
conclusions.
As long as more reliable criteria are not available, it is suggested:
5.7
87
- to design the frame according to the principles expressed here above by referring to
the first order elastic linear analysis of the whole frame;
- then to check that the second-order moments in the frame at collapse do not exceed
the plastic moment Mpb of the section in the beams and the design resistance Mv in
the joints. It must be noted that the second-order elastic linear analysis of a braced
frame may be achieved in a simple, accurate and non iterative way by means of the
modified slope-deflection method developed by VANDEPITTE [16].
5.3.1.4. Examples of application.
In order to point out the influence of second order effects, the described approach has
been applied (Table 5.1.) to the planar frame of figure 5.8. in two different cases
(HE140B and HE100B beam) and the computed collapse load multipliers have been
compared with those resulting from a numerical simulation of the frame behaviour by
means of the finite element program FINELG (see section 4.1. of this report), which
takes accounts of all the material and geometrical non linearities.
Type of analysis
Results
First order
elastic
linear
analysis
Second
elastic
linear
analysis
Non-linear
analysis FINELG
Collapse load multiplier

for beams and:or joints
Collapse load multiplier
for columns
Frame collapse load
multiplier
Moment in the beam at
mid-span (kNm) at collapse
Moment in the joint (kNm)
at collapse
Moment in the beam at
mid-span (kNm) at "first
order" collapse
Moment in the joints (kNm)
at "first order" collapse
Frame collapse load
multiplier
HE14B beam:
Mpb=57.8kNm
Joints:
Mv=46.2kNm
4.5
HE100B beam:
Mpb=24.4kNm
Joints:
Mv=19.5kNm
2.3
3.2
2.29
3.2
2.29
43.1 < Mpb
24.3 < Mpb
13.5 <MV
18.7 < My
49.7 < Mpb
35.7 > Mpb
6.9 <MV
li
7.3 < M V ;
2.0
Table 5.1. - Collapse load multipliers (elastic design).
88
5.8
. beams and columns bent about strong

axis;
XF
.F=100kN;
TTTTTT
. o/h = 1/1000 (sinusoidal initial

deformation);
E
v
II
J3
. X, = 1 corresponds to the service load;
){-/ ho . elastic-perfectly plastic stress strain

diagram of steel (fy = 235 MPa);
HEB 100 HE
B 100 !
. non-dimensional characteristic curves
or
of the beam-to-column joints (end plate
HEB140 HE
B 140
mrm r
r
r
rm
connections): see figure 5.9.
b=5m
Figure 5.8. - Frame for application
,M = * *
Mpb
with Mc = moment in the joint

Mpb = plastic moment of the
beam
Bi-linear model
Experimental curve
<|>c
<j)p
with . <|>c= joint relative rotation

(fp =EI^&- where Ip = beam inertia
P
hb = beam depth
5hb
Figure 5.9. - Joint non-dimensional M-f curves.

From the first order analysis of the frame, it may be stated in both cases that the
buckling of the column precedes the yielding of the beams and of the joints and
determines consequently the collapse of the frame.
The second-order elastic linear analysis allows however to point out the non negligible
influence of the axial loads in the columns on the maximum bending moment in the
beam and in the connections.
In the frame with a HE140B beam, the "e effect" is not sufficient to give rise to plastic
hinges in the beam at mid-span. The assessment of the collapse load multiplier based on
the first order analysis may be considered as a valuable appraisal of the actual frame
collapse multiplier (see table 5.1.) despite the modification of the column loading due
5.9
89
to the second-order effects. The situation is quite different for what regards the frame
with a HE100B beam, where the premature development of a plastic hinge in the beam
during the loading sequence reduces the ultimate strength of the frame.
The load multiplier associated to the formation of this hinge has been evaluated in the
[14] (k = 1.7); it constitutes the ultimate elastic resistance of the frame and has to be
normally considered as its design resistance.
A less safe estimation of the collapse load may however be calculated by determining
the buckling resistance of the column assumed to be hinged at its upper extremity and
subjected to the first order internal forces.
This approach leads to a value of the frame collapse load multiplier X equal to 2.16; this
slightly unsafe result (the actual collapse multiplier equals 2.06 - see table 5.1.) is
principally linked up to the degree of accuracy of the stability check formula. However,
it must be noted that the formation of a hinge in the beam at mid-span for"K= 1.7
deletes actually the restraint at the upper column extremity in the particular cases so
that this approach may be generally considered as a safe one.
Mc
Mc
Figure 5.10. - Reduction to zero of the beam stiffness after the formation of a hinge at
mid-span.
Supplementary calculation in Lige of different structures through the FINELG and the
described elastic design method has allowed to confirm its accuracy. Its application to
the frames studied numerically in annex 2 of this report (frames, the collapse of which
is linked to column instability) as well as a comparison with the FINELG results, is
presented in table 5.2.
Simplified
approach
Non-linear
analysis
FINELG
Difference
ALI
102
BLI
CLI
DLl
1.99
2.11
119
2.19
4.3
9.1
6.4
7.7
Type of
frame
Loading
A
B
C
D
2.5
2.4
2.60
(%)
Table 5.2. - Comparison between FINELG and the elastic design approach
5.3.2. Plastic design.
The plastic design is achieved according to a "strong column-weak beam" criterion
[9,10], in which the frame collapse is associated to the formation of beam plastic
mechanisms. The check of the column is performed, in a similar way to that described
here above, in the structure submitted to collapse loads, a part of which remains elastic.
The problem of the rotation capacity and of the required minimum stiffness of joints for
a plastic design is dealt with, among other things, in [17].
The plastic design approach has been applied to the frame defined in figure 5.8.
constituted in this case of HE200B columns and of a HE140B beam.
90
5.10
The collapse load multipliers are reported in table 5.2. The plastic collapse of the beam
preceedes that corresponding to the column instability and determines therefore the
frame collapse. The column collapse multiplier has been evaluated by assuming hinge
conditions at upper extremity of the columns which are each subjected there to an
increasing axial load and to a concentrated constant bending moment equal to the
plastic moment of the joint (Mv) or to that of the beam (Mpb), according to which is
lesser. This agreement between the hand computed collapse multiplier and that
resulting from a numerical simulation by means of the FINELG program is seen to be
excellent.
Type of analysis
Results
First order
Collapse load
multiplier
for the beam
Collapse load
multiplier
for the columns
Frame collapse load
multiplier
Frame collapse load
multiplier
plastic
analysis
Non-linear
analysis
FINELG
HE14B beam: Mpb = 57.8kNm

Connections: Mv = 46.2kNm
5.1
6.18
5.01
5.1
Table 5.3. - Collapse load multipliers (plastic design).

Table 5.4. presents also similar comparison for the frames considered in annex 2 of this
report (frames, the collapse of which is associated to the formation of a plastic
mechanism in one of the beams). It has to be recalled (see legend of table a.2.4. in
annex 2) that ultimate load factors reported in table 5.4. corresponds to the collapse of
the beams by excess of transversal displacement at mid-span (1/50 of the beam length)
after the formation of the plastic mechanism.
Type of
Loading
Simplified
frame
A
B
L2
BL2
approach
2.65
2.91
Non-linear
analysis
FINELG
2.98
3.23
Difference
(%)
ii.l
9.9
Table 5.4. - Comparison between FINELG and the plastic design approach.
The differences between the numerical and theoretical estimations of the ultimate load
factors in table 5.4. are mainly due to the fact that the strain-hardening effect is
neglected in the theoretical computation of the ultimate load factor.
5.3.3. Validity of the bi-linear modeling of the joint response.
As clearly seen in section 5.3.1. and 5.3.2., the non-linear characteristic of the joints has
to be idealized in view of a hand design. The bi-linear elastic-perfectly plastic
idealization has the advantage to be quite similar to that commonly used for beam and
column sections; in practice, it is used (as, for instance, in Eurocode 3) to refer to the
second stiffness of the joint for the elastic part and to the plastic moment (design
5.11
91
resistance) for the yield plateau (Figure 5.4.). The use of this idealization has however
never been justified by an appropriate study.
This explains the realization of the University of Lige of a parametric study similar to
that presented in annex 2 of this report and devoted to braced frames, the aim of which
is to compare the frame response of different types of frames according to the joint
modeling: separate account and non-linear modeling of the deformability
(figure 5.11.a.) or concentration and bi-linearization of the joint defomability (figure
5.11.b.).
5.11 .a. Separate account and non-linear modeling of the deformability.
5.1 Lb. Concentration and bi-linear idealization of the joint deformability.

Figure 5.11. - Different joint modelings considered in the parametric study.
This parametric study has allow:
- to confirm the safe and sufficiently accurate frame response under service loads and at
plastic collapse (what can be justified theoretically);
- to point out the safe estimation of the ultimate load for the frames, the failure of which
is due to column instability (what cannot be justified theoretically); resulting from the
use of the simplified joint modeling represented in figure 5.1 Lb.
Two examples are reported in figure 5.12.
92
5.12
2.S
3.0
2.3
2.8
1.3
""\ i r
1.a
// '
/ /
/ /
//
'
1
i
i
l
l
'
i
i
1
l
l
1S.0
1
30.0
1
.0
1
EXACT i NON LIN.
_ . GLOB. BI-LIN.
,s54
/*
ft
'.
MOOELLIN&S
SO.O
MM
S.12.a. Frame D with loading DLl (collapse by column instability).
i
i
i
i
i
i
i
i
i
i
'
i
i
i
i
i
i
i
i
i
r m
3.3
3.0
3.0
2.3
i
i
/ /
/ /
//
/ /
i
i //
2.3
1.3
r
i
fii
/'i
A I
A i
1.0
"
0.5
'
i
i
i
i
i
i
i
i
i
i
i
i
i
i
/ L !
7"'
i
1
i
i
i
i
1 L i
T i
'
'
'
2.0
l
4.0
l
S.O
l
t.O
l
10.0
l
i
I
10 8
j
T
"T
EXACT NON LIN.

GLOB. B ILIN.
12.0
1 HH
5.12.b. Frame A with loading AL2 (plastic collapse in the beam).

Figure 5.12. - Comparisons between "actual" and simplified numerical simulations of
the structural response of braced frames (see annex 2).
5.13
93
5.4. Simplified design method for unbraced frames.

5.4.1. Design under service loads.
The check of unbraced frames under service loads will be conducted by using a
geometrically non linear elastic analysis taking into account the second order effects as
well as the behaviour of the joints (characterized by a constant stiffness).
VANDEPlllE has shown [16] that the second order effects on a sway frame with rigid
joints may be taken into account in a very simple and direct manner by means of an
amendment of the slope-deflection method - thus avoiding the usual iterating analysis and even that the decrease in stiffness due to the thrust in the columns can be
incorporated into the computation. The extending of the slope-deflection method to the
analysis of frames with semi-rigid joints has been introduced in 1941 by JOHNSTON
and MOUNT [8]. In consequence, the analysis of unbraced frames under service loads
will be easily performed provided we define the value of the constant stiffness of the
joints.
As a matter of fact, it has been shown in [1], that the use of the second stiffness as
defined infigure5.4. leads often to a very large overstimating (sometimes 50 %) of the
actual sway deflection under service loads. The fulfillment of the serviceability limit
states requires, if it is referred to the secant stiffness, the strengthening of the beam and
column sections, as well as that of the beam-to-column joints. The use of the second
stiffness appears consequently to be extremely safe in most of the cases and should not,
in these conditions, be recommended for economical reasons.
Preliminary studies performed in Lige indicate that it would be preferable to refer to a
fictitious linear stiffness, called KSf infigure5.13., the value of which is intermediate
between the initial stiffness K and the secant stiffness Ks. This stiffness Kgf depends on
the type of joints, but also on the type of frames in which these joints are used (number
of storeys, of bays, loading, lateral rigidity of the frame). Studies are presently in
progress in Lige in order to propose simple procedures for the practical evaluation of
thisfictitiousstiffness.
M
^^^
Mv
i r \Ksf
/
/
/
\ K
\
\
Figure 5.13. - Definition of thefictitiousstiffness Kgf.
94
5.14
5.4.2. Design under factored loads.

The check of ultimate limit states consists in warranting that the actual ultimate
carrying capacity of the structure remains at least equal to the design factored loads.
In [18], a generalization of the well-known Merchant-Rankine formula, has been
suggested, when assessing, in a simple way, the ultimate load of an unbraced frame
fitted with semi-rigid joints. According to Wood [13], the aforementioned formula
writes:
1
0.9
(5 2 )
with:
X,f : collapse load factor (collapse multiplier);
XCT; linear plastic critical load factor (critical multiplier);
Xp : first order plastic collapse load factor (plastic multiplier).
The critical multiplier X.cr is derived from the linear elastic instability analysis
conducted on the unbraced frame by assuming that each of the joints is characterized by
a rigidity specified equal to the initial stiffness Ki (figure 5.13.).
With a view to determine the plastic multiplier Xp, a first order plastic analysis is
carried out. It is assumed that at the location of any specified joint, the plastic moment
is equal either to that of the connected beam or to that of the joint (M v , see figure
5.13.), according to which is the lesser.
The range of validity of the Merchant-Rankine formula, as slightly modified by WOOD
to account for the unavoidable influence of strain hardening through numerical factor
0.9, is usually specified as:
4<^<10
A-p
(5.3.)
Obviously the advantage to refer to the Wood-Merchant-Rankine formula is linked to

the availability of simple methods allowing for the assessment of both critical and
plastic multipliers. In this respect, simple, fast and accurate procedures have been
recently developed at the University of Lige for what regards unbraced frames with
semi-rigid connections (see annex 1). [1]. The validitity of these procedures has been
demonstrated on the ground of an extended comparison with the results of numerical
simulations conducted on a wealth of frames (figure 5.14.). The design tool, that has
been used for this purpose, is the FINELG program. From the aforementioned
comparison, conclusions are drawn for what regards the influence of different
parameters, such as the type of connections, the nature of the loading, the vertical to
horizontal loads ratio,
The range limits for each parameter was chosen in
accordance with conditions usually met in the daily practice.
5.15
95
Figure 5.14. - Semi-rigid frames used for the numerical simulations

5.4.2.1. Assessment of the critical multiplier "Krr.
Figure 5.15. is aimed at presenting the comparison between the results of simple design
approaches for the critical multiplier X& and those get numerically by means of the
FINELG programme. The ratios between hand and numerically computed values are
plotted in these figures.
Method 1 has been largely explained in [18]. It consists in replacing the actual semirigid structure by a sensibly equivalent Ginter single bay frame with rigid joints, the
critical multiplier of which is computed as the lowest of the critical multipliers
associated respectively to each storey considered separately.
Though based on a similar background as Method 1, Method 2 differs for what
regards:
- a more accurate definition of the characteristics of the equivalent Ginter frame,
according to what is largely justified and explained in [20];
- the account of the very actual continuity of the storeys when computing the critical
load of each of the latter.
Last Method 3 is aimed at extending the specifications of the British Code BS 5950
[21] dealing with structures with rigid joints to structures with semi-rigid joints.
Amongst the three methods in consideration, Method 1 is far the most fast one. Indeed
Methods 2 and 3 first require the determination of the linear elastic deflection of the
structure.
Based on the critical appraisal of the results, some conclusions can be drawn:
a) all the three methods yield results which are either safe, or unsafe, compared to the
ones get from numerical simulation;
b) the most accurate assessment - with a discrepancy, with respect to the actual value,
less than 10 % for the whole wealth of frames - is got from Method 2. Therefore the
use of the latter is recommended.
96
5.16
>.cr/\c,l
1.40
ligo..11.31
:-.1
a;
1.20
1.10
*
. ; . . mA.
**
1.00
?_*_* *
f,*
J.
! I "
0.90
"J
0.90
0.70
KTK00 1
rCTMOO 2
KTOOO 3
0.03
0.60
REFERENCE
14.00
8.00
; 12.00
10.00
SO00
Xcrf
Figure 5.15. - Comparison between hand (Xr) and numerically (Xcrf) computed values
of the critical multiplier.
Anyway one must stress the fact that, due to the range of validity of the
Merchant-Rankine formula (specified by (5.3.)), even an appreciable error on the value
of the critical multiplier Xr affects only slightly the value of the collapse multiplier Xf.
Therefore a larger freedom is offered when choosing the method aimed at assessing X^-.
5.4.2.2. Assessment of the plastic multiplier X.p.
In contrast to what regards the critical multiplier Xcr, any error made when assessing
the plastic multiplier Xp will result in a closely similar error on the collapse load, when
the latter is derived from the use of the Merchant-Rankine formula. Therefore, a rather
accurate value of the plastic multiplier Xp is required.
Applying the cinematic theorem in conjunction with the method of combination of
plastic mechanisms will obviously yield the proper exact value of Xp. Because this
procedure is rather long and fastidious, it is not at all suitable for a preliminary design,
as far as the designer has no access to adequate computer programmes. The use of the
static theorem is not likely to offer a better solution in this respect.
The assessment method of X.p that is suggested is detailed in [1] and [20]. It consists in a
two steps procedure. First, the structure is divided in substructures, each of which is
associated to one storey (figure 5.16.) and defined on base of the assumption that points
of contraflexure are located at mid-depth of each column when unbraced frames are
5.17
97
loaded [22]. The values of the plastic multipliers of all the substructures defined
accordingly are then computed; the smallest one is aimed at assessing the plastic
multiplier of the whole structure provided it remains lower than the structure provided
it remains lower than the plastic multiplier associated to the formation of any panel
mechanism in the whole structure. The latter is the aim of the second step; it is obtained
by combining independent panel mechanisms and node mechanisms in the whole
structure till a minimum value of the plastic multiplier is reached.
&T7TT
ni nil
(a)
ifiiii
ni i ih
llllll
Ill/Ill
Ull/Il
(b)
l ni 11
/T77777
(C)
Figure 5.16. - Definition of substructures for the evaluation of Xp:

a) upper storey; b) intermediate storey; c) lower storey
The comparison of the results obtained in accordance with this simplified approach on
the one hand and with exact computations on the other one results in the validity of the
above assumptions and in the accuracy of this new suggestion (figure 5.17.).
98
5.18
6. Benefit of semi-rigid design.

To take the actual resistance and deformability characteristics of the beam-to-column
joints into consideration may lead to a substantial decrease of the global cost for steel
building frames. It is important to clearly point out two aspects of the problem: the
benefit may be linked either to the decrease of the weight of the frame or to the
decrease of the manpower contribution to the global cost.
The decrease of the weight results generally from the hypothesis of perfect hinges to
which it is traditionally referred for simple connections in braved frames. Figures 6.1
and 6.2 illustrate these principles.
The increase of the carrying capacity for the isolated column represented in figure 6.1 is
obtained without any supplementary cost for the fabrication of the cleated connections
acting at his extremities.
The situation is quite similar for the isolated beam schematized in figure 6.2., whose
bending moment and transverse displacement at mid-span is reduced under service
loads as well as at collapse.
Restrained end
0.25 0.50 0.75 1.00 1.25 1.50 1.75 2.00 2.25

Figure 6.1. - Influence of the "semi-rigidity" on the carrying capacity of columns.
Restrained end
Pinned end
Diagram of Bending Moments
Figure 6.2. - Influence of the semi-rigidity" on the distribution of internal forces in
beams.
5.19
99
It has also to be mentioned that the validity Merchant-Rankine approach is restricted to

frames which fulfil the serviceability conditions relative to the maximum sway
displacement.
Other less important restriction to the use of the Merchant-Rankine approach should
also be pointed out. They have been discussed in different reports and papers [23 to 25]
for rigid frames.
1.00
A!
>\
: * :
\
0.90
:
i
0.80
'
M
&
'
oj X J
i
i \
AA
M
U c: :
0.70
>sj o
0.60
rCOMNTMNCXrC
cor 1^0 K O M .
! i
0.50
0.
;i.00
2.00
13.00
4.00
;3.oo
;a.ao
7.00
to"1
cm
rcawtiv*
M r d rCCIM.
8.00 %
Figure 5.18. - Comparison between hand and numerically computed values of the
collapse multiplier.
Ap
beam mechanism
Figure 5.19. - Influence of the second-order effects, according to the type of collapse
mechanism, on the value of the plastic multiplier.
100
5.20
5.4.3. Examples of application.

The hand computer values of X^, Xp and Xf are compared to the actual ones in table 5.4.
for the unbraced frames which have been studied numerically in the annex 2 of this
final report.
These comparisons confirm:
- the validity of the suggested approach, within its application range.
Type of
frame
"B"
Loading
AL4
"ELT
TLT
T3LT
Simplified approach"
535
-S77T
TT3T
"335"
Xcr
FINELG
7.13
TW
Difference (%)
271
inr
"5T
1JT
a - critical multiplier
Type of
frame
Loading
A
B
C
b
AL4
BL4
CL4
DL4
Simplified approach
2.5
3.48
i.76
i-65
^p
FINELG
2.5
3.48
1.86
1.65
Difference (%)
.
.
5.4
b = plastic multiplier
Type of
frame
Loading
A
B
AL4
DL4
BL4
CU
Merchant-Rankine
1.00
2.68
1.67
1.43
FINELG
2.17
3.86 (3.29)
1.7
1.57
Difference (%)
8.3
3.6 (18.5)
1.7
8.92
c = ultimate multiplier
Table 5.5. - Comparison between FINELG and the suggested simplified design
approaches.
the validity of the suggested approaches for the calculation of Xa and Xp;
the validity of the Merchant-Rankine formula within its application range (see frame
C);
the too safe character of the Merchant-Rankine formula for ratios X^ I A.p lower than
4.0 (see frame A,B and D); it has however to be mentioned that the difference of 31 %
reported for frame B is due to particular conditions:
- the plastic mechanism associated to Xp is a beam mechanism;
- the ratio between Xcr and Xp is really low : 2.5;
- the concentration of the joint deformability into a single flexural spring acting at
beam end is particularly conservative (see annex 2) for connections, for resistance
and the stiffness of which are greater than those of the sheared column web panels.
5.21
101
The ultimate multiplier in brackets (table 5.4.c) corresponds to the numerical

simulation, by means of FINELG, of the frame B, the joint deformability of which
is concentrated into singleflexura!springs.
102
5.22
4.00
3.00
2.00
1
/*>
/**
1.00
REFERENCE
;i.oo
2.00
3.00
4.00
APPROXIMATION
\
V
Figure 5.17. - Comparison between hand (Xp) and exact (Xpf) computed values of the
plastic multiplier.
5.4.2.3. Assessment of the collapse multiplier Xf.
The generalization of the Merchant-Rankine formula to frames with semi-rigid joints is
proved to be quite justified on base of the results plotted in figure 5.18. Each plot is
representative of a numerical simulation conducted by means of the FINELG program
with taken account of standard residual stresses, out-of-straightness and out-of-plumb,
of geometrical second order effects, of actual behaviour of the joints, of material
yielding and of material strain hardening ([1] and [7]).
The influence of the type of plastic mechanism associated to Xp on the accuracy of the
generalized Merchant-Rankine formula is clearly shown on the figure 5.18.. Some
conclusions can be drawn for what regards the procedure:
- it is very accurate as far as the collapse mechanism is a complete one, i.e. of the
combined type;
- it is slightly conservative when a partial beam mechanism is governing;
- it is generally largely unsafe when a panel mechanism is commanding.
Such conclusions can be physically justified by the variable influence of the second
order effects, according to the type of collapse mechanism, on the value of the plastic
multiplier when the structure sways progressively (angle *P - figure 5.19.).
It can thus be concluded that the Merchant-Rankine formula, used with the suitable
changes required by the semi-rigid character of the joints, is able to predict the collapse
load of sway frames with a very good accuracy as far as the first order plastic collapse
load is not associated to a panel mechanism. In reality, this last condition is not very
restrictive (see annex 1 [1]).
6.1
103
The second potential source of benefit is certainly more important One of the main
parameters, which govern the cost of a steel frame is the ratio between the cost of
manpower and that of materials; the continual increase of the ratio R between the mean
cost of one hour manpower and that of one kilogram of steel is schematized in figure
6.3 [26]. In West Europe for instance, this ratio is more than eight times greater now
than it was at the end of the second world war. It has also be kept in mind that the cost
of material never represents more than 20 to 40 % of the total cost of the structure; as a
consequence, the optimum design of a frame is more the result of a manpower
"minimisation" than the result of a decrease of the weight.
The fulfilment of this aim requires then the use of very simple, and consequently cheap,
beam-to-column joints. For what regards hot-rolled sections, the recourse to either fully
bolted or partially welded (in workshop) and partially bolted (on site) joints constitute a
really economical solution. Obviously, these joints are of ten semi-rigid and partially
resistant and it may be then concluded that the economical benefit associated to the use
of simple and cheap beam-to-column joint implies the semi-rigid design of the frames.
Design and analysis procedures have been described in chapter 5; they are characterized
by a real simplicity of use, in order not to substitute a supplementary cost for analysis
and design to the decrease of the manpower contribution to the total cost of the frame.
1945
1960
1970
1980
Figure 6.3 - Evolution of the R ratio (R=mean cost of one hour

manpower / mean cost of one kg of steel).
The actual impact of semi-rigid design may only be evaluated by comparing the cost of
similar frames designed successively and respectively as pinned, semi-rigid and rigid.
Such comparisons have been made very recently by COLSON and BJORHOVDE for
one braced frame and one unbraced frame. The conclusions of this study has been
presented at the Second International Workshop on Connections held in April 1991 in
Pittsburgh (USA) [27]. The contain of this paper will constitute also one of the chapters
of a forthcoming document on "Connection and Frame Design for Economy"
emanating from the Task Working Group 10.2 "Semi-Rigid Connection" of the TC10
Committee "Structural Connection" of the European Convention for Constructional
Steelwork [28]; the text of this chapter is reproduced hereunder (section 6.1 to 6.3). The
cost comparison of structures with rigid flexible and semi-rigid joints is established in
terms of relative cost in order to keep enough generality.
104
6.2
6.1. Unbraced frame.

Obviously semi-rigid connections in unbraced frame must be used in case of low-rise
building submitted to reasonable horizontal forces. The structure which is studied is
given in figure 6.4. The distance between each portal frame is 6 meters. The data
concerning the loads and the requirements are given in table 6.1.
Dead loads:
Live loads:
Wind loads:
9 kN/ml (all beams)

13 kN/ml (all beams)
Fl=3.93kN
F2= 7.87 kN
Horizontal drift allowed:
1/400 (28.1 mm)
Beam deflection:
1/360 (27.7 mm)
Steel grade for all elements:
Fe E 235
Table 6.1. - Design requirements
FL j 11 n 111111 u 11 n n i j j j -L
J.M 11I11JLU ,1111 m i i u i
F3
,ii 1111.11U14, 411111111 m
3.75m
3.75m
3.75m
Initial out of plumb : 1/200

Out of plane buckling: buckling length = system length
Figure 6.4. - Unbraced frame.
6.3
105
6.1.1. Design in case of rigid joints.

All the connections are supposed to be rigid. The beam to column connections are
bolted extended end plate connections with transversal and diagonal stiffeners on the
column web (figure 65.). The column footing connections are fully fixed with thick end
plate and four anchor bolts. The results are given in table 62.
V^
Symmetric
for the inner
columns
IPE400<>
W 16x45
H E B 200 ;W 8x40
Figure 65. Rigid connection.
All columns:
All beams:
Horizontal drift:
Beam deflection:
HEB 200
IPE400
9.5 mm (28.1 allowed)
11.2 mm (27.7 allowed)
Table 6.2. - Results of the "rigid" design.
106
6.4
6.1.2. Design in case of semi-rigid joints.

Top and seat angle with web angle are used in this case (figure 6.6). The column
footing connections are the same than in the case of rigid connection design (column
HEB 200 & beams IPE 400). In fact, depending on the interpretation of the results the
beams could be all IPE 360 but it could be matter of discussion. The results in terms of
economy are almost significative. The use of IPE 360 should bring an extra saving.
In case of IPE 400 the horizontal drift is 16.8 mm (28.1 allowed) and the beam
deflection is 183 mm (27.7 allowed).
L 120x120x12 ; 5x5x1/2
Bolts M22 ; 7/8"
80x80x8 ; 3 1/2x3 1/2x3/8

Bolts M20 ; 3/4"
HEB 200
W8x40
Figure 6.6.- Semi-rigid connection.
The cost comparison has been made (table 63) from an enquiry near two french
companies. The indicated prices include all material, columns, beams, plates, angles,
welds, workmanship in shop, workmanship on site, erection of the bare structure
completed.
Solution
Beams
Columns
Connection
Price
Rigid
connection
Semi-rigid
connection
IPE 400
HEB 200
Extended end
100.00
IPE 400
HEB 200
Top&seat
angles
82.00
Table 63. - Cost comparison for the unbraced frame.
6.5
107
6.2. Braced frame.

For braced frames the vertical loads are the only one to take into account for the
design. The structure is shown in figure 6.7. In order to make the comparison in the
framework of the more traditional method of design an elastic design is used for the
beams associated with the beam line method of analysis.
The factored load on the current inner beam is 42.45 kN/ml (135 D + 1 5 L).
The allowed beam deflection at serviceability limit state is L/300.
Inner current beam
3.5m
X
3.5m
3.5m
r
3.5m
Figure 6.7. - Braced frame.
108
6.6
6.2.1. Design in case of pinned connections at each end of the beam.

This is the most common solution because the most simple to design and to make in
shop and on site. In most cases predesigned connections are used. In that case the
beam is an IPE 3460 and the double web angle connection are made with 120x120x12
angles (figure 6.8).
-v
HEB 240 ; W 8x58
Symmetric
\
IPE 360 ; W 14x38
AT
6.0 m
200 mm
L 120x120x12 ; 5x5x1/2
6 bolts M24 ; 1"
Figure 6.8. - Beam with pinned connections.
6.2.2. Design in case of rigid connections

This solution is uncommon unless that specific requirements exist on the depth of the
floor. To satisfy the same requirements than in the former case the beam must be an
IPE 300 associated with extended end plate connections and stiff eners on the column
web (figure 6.9).
HEB 240 ; W 8x58
10 bolts HR 10.9 ; M24 ; 1"
IPE 300 ; W 12x30

6.0 m
Figure 6.9. - Beam withfixedconnections.
6.7
109
6.2.3. Design in case of semi-rigid connections

Top and seat connections are used in this case (figure 6.10). The beam must be an IPE
270.
L 150x150x151=150.0
L 6x6x5/8
Bolts M24 ; 1"
y vi
o)--
<- S
O) - r
X IPE 270 ; w 10x26
L120x120x12
L 5x5x1/2
Bolts M24 ; 1"
6.0 m
Figure 6.10. - Beam with semi-rigid connections.

The same enquiry than in the unbraced frame case has been done (table 6.4). The
indicated price includes all the material, beams, columns, plates, angles, bolts, welds,
workmanship f or manufacturing and erection.
Solution
Beams
Pinned
Rigid
PE30
IPE 300
Semi-rigid
IPE 270
Connection
Price
Double web angle TOmr

Extended
end
120.00
plate
Top&seat angles +
96.00
web angles
Table 6.4. - Cost comparison for the braced frame.
110
6.8
6.3. Conclusion of sections 6.1 and 6.2.

Comparing braced frames & unbraced frames, the most substantial economy by using
semi-rigid joints is got for unbraced frame, even if there is no change on the size of the
elements. The saving is got directly form the saving on the connection itself. In case of
braced frame the price of the bracing is the same in all cases of connections, so the
relative influence of the saving on the connections or on the members is less important
than in the case of unbraced frame.
6.4. Need for further studies.
The comparisons performed by COLSON and BJORHOVDE constitute one of the first
elaborate studies aimed at carefully evaluating the potential benefit resulting from the
semi-rigid design of building frames. The conclusions drawn by the authors of the study
are all the more interesting because:
- the braced and unbraced frames are considered separately;
- the cost comparisons do not only reflect the effect of semi-rigid design on the steel
weight but also on the workmanship associated to manufacturing and erection.
The implication of these conclusions has however to be limited to the particular frames
studied and to the particular companies in which the cost comparisons have been
performed.
This study has consequently to be considered as a first and important preliminary step
of a more extensive study which should take place in the future and should:
- investigate the response of several braced and unbraced frames characterized by
various configurations (number of storeys and bays, types of beam-to-column joints)
by different collapse modes and by different, but anyway realistic sets of design
requirements;
- take account not only of the beam response in braced frames, but also of that of the
columns, the effective length of which is highly dependent, on the value of the end
restraints, as demonstrated several times in the past by different authors [29 to 31];
- be based on cost evaluations by a lot of different companies, in different countries.
Only these conditions are likely to ensure the generality of the drawn conclusions and
also to definitively persuade the potential users of the real economical interest of the
semi-rigid design for building frames.
Conscious of this problem. COLSON has beside suggested to the members of the
previously mentioned ECCS TWG 102. "Semi-Rigid Connections" to contact national
compames in order to produce new cost evaluations for the frames studied in 6.1 and
6.2.
At the Belgian level, the University of Lige has decided to contribute to this
movement, not only through new cost evaluations relative to the COLSON's and
BJORHOVDE's frames, but also through the study of several other braced and
unbraced frames. Results should be available in June 1992. In future, the semi-rigid
design is largely subordinate to such efforts.
6.9
111
7. Conclusions.
The flow chart presented in the introduction (page 1.3.) of this final report highlights
the three main steps to follow when studying a semi-rigid frame:
- the characterization of the joint response;
- the frame analysis;
- the check of the serviceability and ultimate limit states.
The work performed during the present research project relates to several boxes of this
flow chart
By way of conclusion, we will merely recall these original contribution by
distinguishing those related to steel frames (figure 7.1.) the hachured boxes point out
the work which has been carried out in the frame of the present project.
7.1. Frames with steel connections (figure 6.1.)
1. Theoretical models for the prediction of the behaviour of joints with cleated
connections have been proposed and validated by means of comparisons with
experimental results.
The joint deformability components (shear and load-introduction deformabilities of
the column web panel as well as the deformability of the connection elements) may
be modelled:
- in the form non-linear curves
- by means of specific characteristics (secant and initials stiffness, design
resistance,...) according to the analysis planned.
The validity of the concentration of the joint deformability in isolated flexural
springs acting at beam ends has been demonstrated for braced as well as for
unbraced frames.
2. The materially and geometrically non-linear finite element program FINELG has
been implemented in order to simulate accurately and separately the non-linear
response of connections and of sheared column web panels. The validity of the
program {which has been used to perform two parametrical studies:
- the first for the justification of the concentration of the joint deformability (see
chapter 4);
- the other for the justification of the bi-linearization of the joint response
v
v(see
annex 1)[11}
is shown in annex 1.
3. The not hachured boxes in figures 6.1. are related to the frame analysis and to the
check of the limit-states by means of the present research.
Such methods being available at the University of Lige (see annex 1[1]), it has
however been decided to describe them briefly; this is done in chapter 5'of the
present report.
7.1
113
7.2. Frames with composite connection (figure 6.2.)

4. An original approach is proposed in chapter 3 to predict the non-linear response of
composite connections with steel cleats. This procedure is validated in annex 3 by
means of comparisons with all the composite tests performed.
A computer program making the use of the model easier has been written; it is
distributed with mis report (annex 3).
5. The new finite element mentioned in 2. may be used in conjunction with the
"concrete" or "reinforced concrete" finite element to perform numerical simulations
of the behaviour till collapse of composite frames with semi-rigid joints; in this
configuration, the FTNELG is able to take into account the creep, the shinkage and
the non-linear interaction between the steel members and the concrete slab.
Experimental tests on composite frames with semi-rigid connections are still needed
in order to confirm definitively the validity of this high sophisticated numerical tool.
In a more general way now, it may be concluded that:
1. A detailed answer has been brought to the problems raised at the beginning of the
research project and which had been included in the research planning. The work
reported in this final report goes in reality for beyond this planning (development of
hand design methods for steel frames, joint concentration, justification for the bilinearization of the joint response in planar frames,...).
2. Two fully consistent procedures for the study of steel frames with semi-rigid joints
have been proposed:
- the first, which may be considered as the most exact one, lies on the prediction
of the non linear joint response and on the use of a sophisticated finite element
program;
- the second is more orientated towards practice in so far as it does not require the
use of sophisticated programs; it may be implemented by hand or with the help
of simple and easily available program for elastic linear frame analysis.
3. The development of specific tools for the non linear prediction of the connection and
from response allows now to carry out an accurate and complete study of composite
frames with semi-rigid connections by taking into consideration all the material and
geometrical non-linear effects with affect the frame behaviour from early loading
stages to collapse.
We would not like however to end this report without mentioning some topics which
should be more investigated in the future and which would allow in some cases to fulfil
the gaps in knowledge that this research has clearly highlighted:
1. Writing of a document presenting the simplified design methods briefly described in
chapter 5 for braced and unbraced frames in a form directly usable by designers;
such a document would greatly contribute to the promotion of the semi-rigid design
and consequently to that of the steel building construction.
2. Development of simplified procedures for the evaluation of the initial and secant
stiffness as well as of the design resistance for composite joints.
3. Development of simplified design and analysis methods for compare frames with
semi-rigid joints.
114
7.2
4. Carrying out of experimental tests on composite frames with semi-rigid joints in

order to check the analysis finite element program and to bring some help in the
development of simplified design methods.
5. Better highlighting of the economical interest of the semi-rigid design of frames;
work is in progress at the University of Lige and also at the European level.
ConjponeaSQffJcfonniity
Wefc panel
Connection^)
oo
Diffenciation
yes
no
oncntrton^
^
rx,
O
z
2
52
pi
MM I'm I HMH I HMM 11 IHWIItllOHIIHMMI*
4ont
Connection
Panel
ii*
B<|iction f iasviotir
frcdictioB of behaviour
ii ] *
IHI*ilHI1IIIIHalllllt1M1litiiiniiinmn
*f
5pe(3f. Characteristics
aottHô ae gnalys
No-oear carves)
miiiiiyi.unir
mill T
1st order
2d order
analysis
analysis
(manual or
(manual or
numerical)
numerical
ELAST. PLAST.
ELAST. PLAST.
Noa-inear
2
lato-piasttc.
nsmertcaL
analysts
<
H
0*5
Displacements and
Ultimate loads
resistance of
forULS
sections for SLS

t t t m t i i i i m i i m i H i i i n t i i m i n i m i n , uhi
Checks
Operational area
Figure 6.1. - Frames study with steel connections.
7.3
115
Components of deforma&iliy
Web panel
Connections)
co
yes
v.ywyy
|tfwjitw^?ly<>wr^x
foexfictipoofteiiavipur
>*IIHIIIMMtlMMIHjtMIMINH,IHIM*MtMi
Connection
11I M 11 T IIfc****j
<
a
S p e c ^ characteristic?,
according tft tne analypj
Kon-lies- careers)
iaiUWHUUMMWUUi
miiniiiiiiiiii*>MiiMniiiMiii>nJuiibHim
co
1st order
2d order
analysis
analysis
(manual or
(manual or
numerical)
numerical
ELAST. PLAST.
ELAST. PLAST.
IH
co
Displacements and
Ultimate loads
resistance of
forULS
sections for SLS
Hilf! Operational area

s^^j
Not yet fully operational
Figure 6.2. - Frames study with composite connections
116
7.4
7.3. Databank SERICON and program SPRINGS :

The databank SERICON (SEmi Rigid CONnections)distributed during the F6 meeting
on 28th and 29th november in Luxembourg contains the 56 steel and composite tests
performed in the frame of this research.(see annex 4)
The program SPRINGS also distributed during the F6 meeting is a program written for
Macintosh with the well known software HyperCard V2.0 as preprocessor and also
written for compatible IBM which simulates, with the mathematical model using nonlinear springs, the behaviour of a composite joint with cleared connection.(see annex 3)
7.4. Benefit of the semi-rigid design.
A study performed by COLSON and BJORHOVDE in the frame of the TC10 Comitee
"Structural Connection" of the European Convention for Constructional Steelwork (see
chapter 6) constitues a first approach of the potential benefit of semi-rigid design for
braced and unbraced frames.
In these two analysed examples the benefit of semi-rigid design for the unbraced frame
(see 6.1.3.) is 18% and 4% for the braced frame.(see 6.2.4.)
7.5
117
REFERENCES.
1. JASPART, J.P., Etude de la semi-rigidit des noeuds poutre-colonne et de son
influence sur la rsistance et la stabilit des ossatures en acier, Ph.D. Thesis, M.S.M.
Department, University of Lige, January 1991.
2. STUTZKI, Ch., LOPETEGUI, J., SEDLACEK, G. "Semi-rigid connections in
frames, trusses and grids".
Proceedings of a State-of-the-Art Workshop on Connections and the Behaviour, Strength
and Design of Steel Structures, Cachan, France, 25.27 May, 1987. Elsevier Applied
Science Publishers, February 1988, pp. 166-174.
3. TSCHEMERNEGG, F., "On the non-linear behaviour of joints in steel structures".
Ibidem, pp. 158-165.
4. ATAMAZ SIBAI, W., FRAY, F., "Numerical simulation of the behaviour up to
collapse of two welded unstiffened one-sideflangeconnections.",
Ibidem, pp. 85-92.
5. BOERAEVE, PH., "Contribution T'analyse statique non linaire des structures
mixtes planes formes de poutres, avec prise en compte des effets diffrs et des phases
de construction".
Ph.D. Thesis, M.S.M. Department, University of Lige, 1991.
6. "Ultimate Limit State Calculation of Sway Frames with Rigid Joints".
ECSC - Technical Committee 8 - Structural Stabiility - Technical Working Group 8.2. System. First edition, 1984.
7. BIJLAARD, F.S.K., ZOETEMEIJER, P., "Influence of joint characteristics on the
structural response of frames".
Proceedings of the International Conference "Steel Structures: Recent Advances and their
Application to Desing", Budva, Yougoslavia, Sept. 29. Octobre 1,1986, pp. 109-133.
8. JOHNSTON, B. and MOUNT, E.H., "Analysis of building frames with semi-rigid
connections", American Society of Civil Engineers, Transactions, 1942, 107, pp. 9931019.
9. VINNAKOTA, S., "Design of columns in planar frames. A few comments",
presented at the National Conference on Tall Buildings, held at New-Delhi, 1973.
10. VINNAKOTA, S., "Design of columns as part of frames - Some remarks", paper
submitted to TG 3: SSRC, 1983.
11. SNIJDER, H.H., BIJLAARD, F.S.K. and STARK , J.W.B., "Use of the elastic
effective length for stability checks of columns and consequences for checks on beams in
braced frames", Proceedings of the Michael R. HORNE Conference, edited by L.J.
MORRIS, London, Granada, 1983.
12. BJORHOVDE, R., "Effect of end restraint on column strength - Practical
applications", Engineering Jl., A.I.S.C, to be published.
13. MAQUOI, R. and JASPART, J.P. "Contribution to the design of braced frames with
semi-rigid connections"; Proceedings of the Fourth International Colloquium on Stability
of Metal Structures, North American Session, New York City, 17-19 April, 1989.
Ref. 1
119
14. ELSEN, Ph., "Mthodes manuelles de prdimensionnement des ossatures

mtalliques contreventes assemblages semi-rigides", Diploma Work, MSM
Department, University of Lige, 1989.
15. AMOUR, A. "Contibution T'tude de la capacit portante des colonnes de structures
contreventes noeuds semi-rigides", Matrise en Sciences Appliques, Service MSM,
Universit de Lige, 1988.
16. VANDEPITTE, D., "Non-iterative analysis of frames including the P-A effect".
Journal of Constructional Steel Research, 2 (1982), pp. 3 -10.
17. STARK, J.W.B. and BIJLAARD, F.S.K., "Monograph on beam-to-column
connections. Ch. 17: Design rules for beam-to-column connections in Europe" Inst TNO
for Buildg.Struct. and Buildg. Materials, Delft, Holland, 1984.
18. JASPART, J.P.: "Extending of the Merchant-Rankine formula for the assessment of
the ultimate load of frames with semi-rigid joints", Journal of Constructional Steel
research, 11 (1988).
19. WOOD, R.H.: "Effective lengths of columns in mud-storey buildings - Part 3"; The
Structural Engineer, Vol. 52,1974, pp. 341 - 346.
20. JASPART, J.P.:"Etude paramtrique du comportement des structures non
contreventes noeuds semi-rigides et dveloppement d"une mthode simple de
dimensionnement Rapport interne, Universit de Lige, N 181, Nov. 1988.
21. British Standard BS 5950 Pt 1
"Structural use of steelwork in building". Code of practice for design in simple and
continuous construction: hot rolled sections. 1985,
22. GERLEIN, M. and BEAUFAIT, F.W.:"An optimimum preliminary design of
reinforced concrete frames". Journal Computation and Structures, Pergamon Press, Vol.
11, 1980, pp. 515-524.
23. ANDERSON, D. and LOK, T.S.: "A limitation on the use of the MERCHANTRANKINE formula". Research Report CE13, Department of Engineering, University of
Warwick, October 1982.
24. ANDERSON, D. and LOK, T.S.: "Design studies on unbraced multistorey steel
frames".
The Structural Engineer, Vol 61B., N 2, June 1983, pp. 29 - 34.
25. ANDERSON, D., LOK, T.S. and WOOD, R.H.:"Studies of the MERCHANTRANKINE formula".
Research Report CE 12, Department of Engineering, University of Warwick, October
1982.
26. VAN DOUWEN, A.A., "Design for economy in bolted and welded connections".
Joints in Structural Steelwork, Proceedings of the International Conference held in
Teeside Polytechnic, Middelborough, Cleveland, 6-9 APril 1981, pp. 5.18 - 5.35.
27. COLSON, A. and BJORHOVDE, R.: "Economy of design with semi-rigid
connections".
Proceedings of Second International Workshop on Connections in Steel Structures:
Behaviour, Strength and Design, Pittsburgh, USA, 10-12 April, 1991. Elsevier Applied
Science, to appear.
120
Ref
-2
28. Technical Working Group 10.2 on "Semi-Rigid Connections" of the TC10

Committee "Structural Connections of the European Convention for Constructional
Steelwork", "Connections and frame design for economy", to appear.
29. JASPART, J.P.: "Etude de la capacit portante des colonnes restreintes leurs
extrmits", Diploma work, University of Lige, 1985.
30. LUI, E.M. and CHEN, W.F.:"Strength of H-columns with small end restraints",
The Structural Engineer, Vol. 61B, N 1, Part B, March 1983, pp. 17 - 26.
31. NETHERCOT, D.A. and CHEN, W.F.: "Effects of connections on columns",
Journal of Constructinal Steel Research 10 (1988), pp. 201 - 239.
32. Task Group set up by Technical Work Group 8.2 of the European Convention for
Constructional Steelwork, "Analysis and Design of Steel Frames with Semi-Rigid
Joints", to appear.
33. ALPSTEN, G.A.: "Variations in mechanical and cross-sectional properties of steel",
ASCE-IABSE International Conference in Planning and Design of Tall Building, Vol. Ib9,1972.
34. ATAMAZ SIB AI, W. and JASPART.J.P.: "Etude du comportement jusqu" la ruine
des noeuds souds", Rapport interne n 194, Universit de Lige, Facult des Sciences
Appliques, Service M.S.M. / Rapport interne n" 89/7, Ecole Polytechnique Fdrale de
Lausanne, IREM, octobre 1989.
35. JASPARTJ.P.: "Study of the shear and the load-introduction deformability of
column web panels in strong axis beam-to-column joints - EC3 formulae (annex J):
discussion and proposals for improvement." Rapport interne N 202, Universit de
Lige, Facult des Sciences Appliques, Services M.S.M., Avril 1990.
36. JASPARTJ.P. and MAQUOI.R.: "Study of the shear deformability of column web
panels in strong axis beam-to-column joints.", Proceedings of the SSRC Annual
Technical Session, St. Louis, U.S.A., 9-11 avril 1990.
37. MAQUOI, R. and JASPART, J.P.: "Study of the load-introduction deformability of
column webs in strong axis beam-to-column joints.", Proceedings of the IUTAM
Symposium on Contact Loading and Local Effects in Thin-Walled Plated and Shell
Structures, Prague, Tchcoslovaquie, 4-7 septembre 1990.
38. JASPART, J.P. and MAQUOI, R.: "Plastic capacity of end-plate and flange cleated
connections - Prediction and design rules.", Proceedings of the Second International
Workshop on Connections in Steel Structures: Behaviour, Strength and Design,
Pittsburgh, U.S.A., 10-12 avril 1991.
39. Manual on Stability of Steel Structures, ECCS Publication, N 22, Bruxelles, 1976.
40. ATAMAZ SIBAI, W., "Simulation numrique du comportement de deux
assemblages de rives souds non raidis", Internai Report 87/5, IREM, Polytechnic
Federal School of Lausanne, June 1987.
41. JANSS, J., JASPART, J.P. and MAQUOI, R., "Experimental study of the nonlinear behaviour of beam-to-column joints", Proceedings of a State-of-the-Art Workshop
on Connections, Behaviour, Cachan, France, 25-27 May, 1987.
Elsevier Applied Science Publishers, February 1988, pp. 26 - 32.
Ref. 3
121
42. ZOETEMEIJER, P., Report 6-80-5, Stevin Laboratory, Delft, February 1980.
43. KATO, B. "Beam-to-column connection research in Japan", Journal of Structural
Division, ASCE, vol. 108, n ST2, February 1982, pp. 343-360.
44. Kishi, N., Chen, W.F., Matsuoka, K.G. and Momachi, S.G.: Moment-rotation
relationship of top- and seat-angle with double angle connections. Connections in Steel
Structures. Elsevier Appi. Sc. Pubi., 1988, pp. 121-134.
45. Hotz, R.: Traglasversuche fr Sttze-Riegel-Verbindungen mit verbesserter
Wirtschaftlichkeit. Der Stahlbau, 11,1983, pp. 329-334.
46. BURSI, O., "Tests and modelling of bolted double clip angles of steel bracing
connections". Internal Report, Department of Structural Mechanics and Design
Automation, University of Trento, October 1990.
47. Davison, J.B., Kirby P.A. and Nethercot, D.A.: Rotational stiffners characteristics
of steel beam-to-column connections. Jl. of Constr. Steel Research, Vol, 8, 1987, pp.
17-54.
122
Re- 4
Annex 1.
Ph.D. Thesis of JASPART, J.P., "Etude de la semi-rigidit des noeuds
poutre-colonne et de son influence sur la rsistance et la stabilit des
ossatures en acier", M.S.M. Department, University of Lige, January
1991.
%.
\jt'.J:
'Mk
This Ph.D. Thesis (about 600 pages in French) can be provided by ARBED Recherches
(tel.: 352-55512177) to interested persons.
125
Annex 2.
Numerical simulations of frames by FINELG program.
c- Rifili A.
For daily practice, it is quite impossible to account separately for the flexural behaviour
of connections and the sheared behaviour of column web panels. From a practical point
of view, it is necessary to concentrate the deformability of the connections and of the
column web panels in flexural springs located at the extremity of the corresponding
beams.
The validity of this concentration of the deformability components as well as the
optimum location of the springs can not be defined theoretically. The parametric study
which has been presented in the section 4.2. of the final technical report is consequently
aimed at determining how the actual and relatively complicated behaviour of a joint
(shear web panel + 1 or 2 connections) may be reduced, with a sufficient accuracy, to
that of isolated springs, the influence of which on the frame response may be easily
accounted for in practical design methods (see section 5 and annex 1 [1] of the final
technical report).
As explained in section 4.2. of the final technical report, this parametrical study is
based on the numerical simulation with the non-linear finite element program FINELG
(see section 4.1. of the final technical report) of the response till collapse of braced and
unbraced steel building frames.
This annex contains a detailed description of the frames considered in the study and of
the numerical simulations which have been performed as well as a presentation of the
results.
Description of the studied frames.
Four types of frames are considered in the study; they differ by the number of storeys
and the number of bays (figure a.2.1.). They will be successively considered as braced
and unbraced, what explains the stippled design of transversal rollers on the drawings in
figure a.2.1.).
The loading of the frames in service conditions is described in figure a.2.1. and in table
a.2.1.
The stress.strain characteristic of steel as well as the pattern of residual stresses in the
beam and column cross-sections are similar to those represented in figures a.2.2. and
a.2.3.
The initial geometrical imperfection of the unbraced frames is chosen in accordance
with the ECSC Recommendations [6] (see figure a.2.4.a.).
The initial out-of-straightness of the columns in each braced frames is obtained by
referring to the eigen mode which corresponds to the lower eigen value of the frame
(elastic linear instability analysis); this eigen mode is multiplied by an appropriate
coefficient to obtain an out-out-straightness of the initial column equal to the 1/1000 of
its height (see figure a.2.4.b.).
Two types of beam-to-column joints have been selected for each type of frames; they
correspond respectively to flange cleated connections and extended end plate
connections (figures a.2.5. to a.2.8.).
Annex 2.2
129
Types of frames
(figure a.2.1)
Braced
A
Unbraced
Braced
B
Unbraced
Braced
C
Unbraced
Braced
D
Unbraced
columns
beams
Nb
height
type
span
between
between
beam axes
column axes
(m)
(m)
HE160B
8.0
IPE200
Al
5.0
A2
5.0
A3
HE160B
6.0
IPE200
A4
left beam
Bl
HE160B
7.0
IPE300
10.0
right beam B2
8.0
HE160B
12.0
IPE300
12.0
B3
B4
HE140B
5.0
IPE270
6.0
Cl
C2
HE200B
4.0
EPE300
5.0
C3
C4
HE120B
5.0
IPE220
5.0
Dl
D2
IPE300
5.0
HE160B
4.0
D3
D4
type
Figure
connection
nodes
q
(kN/m)
PI
(kN)
a.2.5.a
a.2.5.a
a.2.5.a
a.2.5.b
a.2.6.a
8.242
10.303
12.363
12.363
8.863
397
0
108
108
450
28
6
6
-
ALI
AL2
AL3
AL4
BLI
8.2.6A
2.037
32
BU
a.2.6.a
a.2.6.b
a.2.7.a
a.2.7.a
a.2.7.a
a.2.7.b
a.2.8.a
a.2.8.a
a.2.8.a
a.2.8.b
4.050
4.050
14.305
18.392
41.537
41.537
11.245
11.245
41.178
41.178
36
36
24
24
-
3
3
10
10
5
5
BL3
BL4
Loading
P2
F
Name
(kN) (kN)
120
0
0
0
160
0
0
0
CLI
CL2
CL3
CL4
DLl
DL2
DL3
DL4
Table a.2.1. - Data relative to the studied frames.

l
' t'-
P,
q
r
i-VJ t I t I i
a.2.1.a. - Frame type A
130
till
P,
1
\ ^
*g;
a
a.2.1.b. - Frame type C
Annex 2.3
L_L
r r u
17777S
'77M,
97777!
Ml
a.2.1.c. - Frame type B
_F% rr>n
>
>
9!w>
4;
i__i
\ 1 1 1\
\\W\\
Ijftjf*
>
VTTTf
Mi
9SW?
a.2.1.d. - Frame type D

Figure a.2.1. - Numerically studied braced and unbraced frames.
0.02 E
i
f y = 240Mpa
E = 210.10 3 Mpa
:y
10 Y
Figure a.2.2.: a - e curve for steel.
Annex 2.4
131
H/B<1.2
/K^-0.5
H/B> 1.2 - . 0 . 3
T.
i
H
B-
Figure a.2.3.: Pattern of residual stresses.
r
L
;
0
300 '
r1 =V
'
V1
ifL<5m
L > 5m
( L en m
number of columns in
. a storey
a.2.4.a. - Unbraced frames - initial transversal displacement of the whole frame 0P o ).
/////
a.2.4.b. - Braced frames - initial out-of-straightness of the columns.

Figure a.2.4. - Initial imperfections of the studied frames.
132
Annex 2.5
M b (kNm)
78.0
Shear of column web panel
58.5
Beam - to - column joint
39.0
19.5
IJ
2.6
1.9
5.2
65
9. .0(10 'tad)
1
7.6
a.2.5.a. - Joint with flange cleated connection.

M b lkNm)
78.0
Connection
58.5
39.0

19.5
-i
1.3
2.6
1
33
1
5.2
1
6.5
e.r.<t>no"2rad)
*7.8
a.2.5.b. - Joint with end plate connection.

Figure a.2.5. - Joints relative to braced and unbraced frame (type A).
Annex 2.6
133
M b (kNm)
Ui.O

108.0

72.0
36.0-
9.J,MO Sad)
3.0
S.O
9.0
12.0
15.0
18.0
a.2.6.a. - Joint withflangecleated connection.

M b (kNm)
K4.0"
108.0
72.0

36.0-
8. I > t 1 0 " 7 r d )
t
3.0
6.0
9J0
12.0
1S.0
18.0
a.2.6.b. - Joint with end plate connection.

Figure a.2.6. - Joints relative to braced and unbraces frames (type B).
134
Annex 2.7
jMjikmm
134.7511751-
ns.soisoi

M.2St12SI
Connection
n.oonooi
S7.7S a s i
l i .Ml 501
19.25(251-
a.2.7.a. - Joint with flange cleated connection for braced and unbraced frames,
i MikNmi
140
Connection
120
100
Beam to column joint
. f . ( ' r i )
2.0
;JJ
1.0
IJO
a.2.7.b. - Joint with end plate connection for unbraced frames.

Figure a.2.7. - Joints relative to type C frames.
Annex 2.8
135
J.15UC1-
32.70(721-
L _ Shear of column web panel
:7.:I6OI-
Connection
21.S0US)-!-
1SJSI36I-
10.JIH-
J.iSI'.SIt.r.*(10'r>)
2.3
.0
(.0
1.0
.2.8.a. - Joint with flange cleated connection for braced and unbraced frames.
* *,ikNmi Shear of column web panel

Connection
. f.H0*'UI
2.0
Ut
t.0
1.0
a.2.8.b. - Joint with end plate connection for unbraced frames.

Figure a.2.8. - Joints relative to type D frames.
136
Annex 2.9
Performed numerical simulations.

Tables a.2.2. and a.2.3. present the numerical simulations which were planned,
respectively for braced and unbraced frames.
Five different modelings were considered for each frame; they were subdivided in two
categories:
a. representation of the connection deformabilitv only.
(column web panel assumed to be fully rigid, a very flexible connection)
a.l. deformability of the connection concentrated in a single spring located at the
interface between the beam and the column - type Ml.
(corresponds to the reality > reference case)
external face of the

column flange
column axis
r
beam
a.2. deformability of the connection concentrated in a single spring located at the

level of the column axis - type M2.
jiiiimmj
b. representation of the connection and of the sheared column web panel

deformabilities.
b.l. deformabilities of the connection and of the sheared web panels are represented
separately - type M3
(corresponds to the reality - reference case)
sheared
column web panel
Annex 2.10
beam
137
b.2. deformabilities of the connection and of the sheared web panels concentrated, at
each extremity, in a single spring located at the level of the column axis - type
M4.
.nummi j
b.3. deformabilities of the connection and of the sheared web panels concentrated, at
each extremity, in a single spring located at the interface between the beam and
the column - type M5.
jmmmii
Preliminary calculations (performed at the time of the definition of the program for the
parametric study) had indicated that the deformability of the joints may be concentrated
at the beam-to-column interface. The type of the corresponding numerical model is M4.
In view to confirm these results, the numerical simulations relative to the hachured
boxes in tables (a.2.2.) and (a.2.3.) have been performed in the first instance (see results
hereunder). These simulations have allowed to confirm this first impressions.
Seeing that:
- numerical simulations using M5 simplified modeling could not lead to better results
than those obtained with M4 modeling;
- M4 simplified modeling is fully representative of the actual joint behaviour when the
column web panel is stiffened for shear or when very flexible connections are used
(simplified modeling Ml);
it has been decided not to go on with the other numerical simulations which are time
consuming and which could not bring more useful information.
138
Annex 2.11
Table a.2.2. - Numerical simulations of braced frames.
Stiffened column webs

Type of
Frames
Unstiffened column webs
UHHIHIl
Jnnmnle
mmm
IMUttlMLe
Ml
M2
M3
M4
romnQ
loading
M5
AL3
AL4
BL3
BL4
CL3
CL4
DL3
DL4
Table a.2.3. - Numerical simulations of unbraced frames.
Annex 2.12
139
Results of the numerical simulations.

The curves resulting from the numerical simulations using M3 and M4 joint modelings
are presented in figures a.2.9. to a.2.11. in form of:
- curves "load factors A, - mid-span displacement A for the beam, the transverse
displacement of which is the most important at each level of the frame loading",
- curves "load factors X - transverse displacement v at the upper column extremity for
unbraced frames".
These curves allow to control the influence of the simplified numerical modeling M4
on (see table a.2.4.):
- the transverse displacement of the beams and of the frame under service loads (load
factor X = 1);
- the collapse load factor.
Type of frame
braced
A
unbraced
braced
B
C
D
unbraced
braced
unbraced
braced
unbraced
Loading A for braced frames / X for unbraced

frames under service loads (mm)
M4
M3
Difference
simplified
modeling
(reference) modeling
(%>
9.84
ALI
1.5
9.99
AL2
10.54
10.72
1.7
AL4
0.5
1.98
1.99
BLI
14.60
16.06
10.0
BL2
5.3
20.98
22.09
BL4
0.4
3.39
3.40
CLI
13.92
3.0
13.51
CIA
9.3
4.80
5.25
DLl
9.96
5.7
10.53
-5.4
DL4
2.08
1.97
X ultimate
M3
modeling
(reference)
2.11
2.98*
2.17
2.19
3.23*
3.86
2.185
1.70
2.605
1.57
M4
simplified
modeling
2.105
2.98*
2.14
2.19
3.20*
3.29
2.185
1.69
2.605
1.57
Difference
(%)
0.2
0.0
1.4
0.0
0.9
14.7
0.0
0.8
0.0
0.0
- In this table, A represents the transversal displacement at mid-span with respect to the beam extremities;
in the figures a.2.9. to a.2.11., A corresponds to the total displacement of the beam at mid-span (including
the shortening of the columns in compression).
- Differences in %: (+) if on the safe side
(-) if on the unsafe side.
- *: the ultimate load factor corresponds to the collapse of one of the beams by excess of transversal
displacement at mid-span (1/50 of the beam span).
Table a.2.4. - Comparison of M3 and M4 modelings
140
Annex 2.13
3.3
3.1
1.1
1.3
V-
-!
HOOELISATIONS
_ _ EXACTE
.-B.. GLOBALISEE HH
44.3
30.
a.2.9.a. - Type A frame - Loading ALI.
3.3
3.3
i,
3.0
TT^:*
,
3T
S>
>*
1.3
/ '
l/ /
//
/r
1.1
/'
/'
/ '
.3
i
i
i
:
a
1
t
1
1
1
1
i
i
i
i
i
i
*
*
i
71 1 ! !""
/
:sa.
:t.*
i
:\s*.i
.
i
MODELISATIONS
_ , _ EXACTE H 3
. _ GLOBALISEE H H
:M.
im
a.2.9.b. - Type B frame - Loading BLI.
Annex 2.14
141
MODELISATIONS
_ , - EXACTE h 3
. , - . GLOBALISEE
Hf
a.2.9.c. - Type C frame - Loading CLL
3.0
" " " " I " " " " ! " " "
3.S
2.3
\r'\
2.0
>+
1.3
i
!
/*
i
x/
i
/*
//
ii j/ '
i
i
.
1.
/'
// i
/*
i
/*
/ *
s*
i
i
i
i
i
ii
i
i
i
i
i
i
i
i
i
!i
.*
St
i
i
i
7 [ I i
(.3
i *f
HOOELISAIONS
, _ EXACTE H 3
. _ , _ . GLOB ALISEE H H
i m
a.2.9.d. - Type D frame - Loading DLL
142
Annex 2.15
NCC6LISAT IONS
EXACTE H 3
GLOBALtSEE HM
a.2.9.e. - Type A frame - Loading AL2.
X
4.0
4.0
>
_J"""
3.0
2.0
1'
f,fc""T
It
'
/
It
1.0
_J^*
sr
'
T
'
1
'
1
1
1
1
1
1
HOOEUSATIOHS
/ i n r;
7.3
ia.!
aa.3
so.
i m
_ , _ EXACTE
H3
. . a GLOB ALISEE
H4
a.2.9.f. - Type B frame - Loading BL2.

Figure a.2.9. - Influence of M4 numerical modeling on the response of braced framescurves "load factor X. - beam mid-span displacement A".
Annex 2.16
143
3.3
2.0
I ! I
V\
A*
1.S
1.0
l [ ! i ]
HOCeLISATlONS
.s
_ _
/ ! 1 1 ! !-
3.1
l
4.a
l
s.
1
t.a
1
EXACTE
..... GLOBALISEE
M.i
1 MI
a.2.10.a. - Type A frame - Joints with end plate connections. AIA
HOOeUSATIONS
_ EXACTE
. . a - . GLOBALISEE
a.2.10.b. - Type B frame - Joints with end plate connections. BL4.
144
Annex 2.17
\
3.1
1.7
. . .
\ 'f^'l
I\
i.
MODELISATIONS
.3
'
, _ EXACTE
'
/ !1
ao.a
40.
. . . . . GLOB ALISEE
ss.a
H.f
a.2.10.c. - Type C frame - Joints with end plate connections. CL4.
MODELISATIONS
EXACTE
. . . . . QLOB ALISEE
a.2.10.d. - Type D frame - Joints with end plate connections. DIA

Figure a.2.10. - Influence of M4 numerical modeling on the response of unbraced joints
- Curves "load factor X, - beam mid-span displacement A".
Annex 2.18
145
3.a
H
Il
!'
^ ^ f l ! ! * * 1
!.__
3.3
/' i
1.3
HcceusAiTtOHS
i.
1.3
EXACTE
. . . . . GLOBALISEE
/
4.
a.o
ia.
ci
aa.t
i mM
a.2.1 La. - Type A frame - Joints with end plate connections. AL4.
\
4.0
i/
.'
'
'
"r
3.1
!"
'
i
'
1
i
1
1
|
7>'
^'
/ M
/'
Jt
2.0
1.0
//
:
i
1
1
|
71 1 1
HCCELISATIONS
/ 1 ! ! !..
13.0
3S.0
I
43.0
l
a EXACTE
..... GLOBALISEE
SO.O
1 IM
a.2.1 l.b. - Type B frame - Joints with end plate connections . BLA
146
Annex 2.19
X
3.1
1.3
>
7
fi
A
Jit..
/.
h
jt
/.
/.
/'
1
|i
1
i
I
1
i
If
H
'
II
w
M
r
l
1
i
i
II
i
nOOEL[SATIONS
J;*
_ . _ EXACTE
..... GLOBALISEE
'
!*
.i
i
B
3 .I
B
3 .I
1
4*.
1 1
a.2Lll.c. - Type C frame - Joints with, end plate connections. CL4.
WOEL ISAT IONS

,
EXACTE
..... GLOBALISEE
a.2.1 l.d. - Type D frame - Joints with end plate connections. DIA
Figure a.2.11. - Influence of M4 numerical modeling on the response of unbraced
frames - Curves "load factor X - transverse displacement of the frame V".
Itanex 2.20
147
Annex 3.
Mathematical model for the prediction of the moment-rotation curves of
composite joints with cleated connections.
Floppy disc with the program SPRINGS.
(The floppy disc is available at ARBED Recherches
tl. : (352)- 5551-2177)
3.1. Introduction.
The aim of this annex 3 is to explain the way to calculate the moment rotation curve of a cruciform
composite joints with cleated connections.
If you use a Macintosh computer with the program HyperCard V 2.0, you have the possibility to
introduce the data of the joint in a HyperCard pile and calculate directly the moment rotation curve
of the joint, (see section 3.2.)
If you use a IBM or compatible computer, you must create a ASCII file containing all the data of
the joint before running the main program calculating the moment rotation of the joint (see section
3.3.)
If you want understand how the program calculates the different springs, you can find information
in the section 3.4.
You can find the conclucions of the mathematical model in the section 3.5.
The hypotheses of this program are the following:
- Slip between concrete and the beams cannot appear. We suppose that we have enough shear
connectors to prevent the sps.
- For the moment, the model doesn't take into account the possibility to have slip in the upper cleat
due to a gap of the bolts in their holes. If you have a concrete slab, this phenomenous is prevented.
- The used values of yield points, ultimate strengths, dimensions and so one are exact measured
values. The program doesn't use "fa factors.
- The applied moment must be in the following direction:
V///////////////////////////////.
M
- The left moment equals the right moment.

- The program doesn't take into account of a normal loading in the column.
3.2. Macintosh program with HyperCard V 2.0.
Make a copy of the floppy disk before using it because with the program HyperCard, all the
changes you will do will be automatically stored on the floppy disk.
If you have the HyperCard V 2.0 program on you hard disk, you can put the floppy disk called
"SPRINGS MAC" in your driver and open the file SPRINGS.
Annex 3.2
151
You see on your screen the following card:
24x2c1
24x2c1g
24x2c2
24x2c2g
24x2c3
24x2c3g
24x2c5
24x2c5g
24x2c6
24x2c6g
24x2c7
24x2c7g
24x3c1
24x3c1g
24Y3R2
lt)
( Print ) (delete)( Quit ) ( Neu; )
This card represents an index card with all the simulated joints.
By clicking on a test name, you go to the main menu containing the different conponents of
the selected joint.
By clicking on the Hause button, you return to the pile HOME of the program hypercard.
By clicking on the PRINT button, you print this card on your printer.
By clicking on the DELETE button, you delete this card.
By clicking on QUIT button, you quit this pile and return to the system.
By clicking on NEW , you create a new file where you must define all the components of the
joint
You can use the lift cursor to see the other tests.
If you click for exemple on the name 36X2C5g, you arrive in the main menu and you see the
following card:
152
Annex 3.3
SEMI RIGID CONNECTIONS

Name of the test :]36x2c5g
Upper cleat :| 150x90x13
Column : HE200B
Beam :|IPEa6
~~'
Web cleat : 150x90x13g
Lower cleat : 150x90x13g
Reinforcement bars : D10

Concrete
: 12.45
Help )
( Print ) (delete ) ( Quit ) ( Neu
1A
'ndeH )
Calculation) ( O X ^ K
)
VISBED
This card represents all the components of the joint 36X2C5g: you have
- the column
-the beam
- the reinforcement bars
- the concrete
- the upper cleat
- the lower cleat
- the web cleat
By clicking on the HELP button you can see the architecture of this program. The pile
SPRINGS contains the different names of the tests and each card of this pile SPRINGS is
linked to 7 other piles containing the informations about respectively, the column , the
beam, the reinforcement bars, the concrete, the lower cleat, the upper cleat and the web
cleat.
By clicking on the RETURN button, you come back to the card 36X2C5g.
By clicking now on the name COLUMN and clicking YES in the message box which will
appear, you can go to the card containing the informations of the chosen column.
You are now in the pile of the columns. If you click on the INDEX button you can see that we
have for the moment only one column in this pile.
By clicking on the HE200B value, you return to the card of this column.
By clicking on the HELP button, you have information on the different values putting in the
different fields. You can see the other HELP cards by ckicking on the NEXT button or
PREVIOUS button. To quit the HELP, click on the RETURN button.
If you want add a new column, click on the NEW button of the file COLUMN.
If you want return to the SPRINGS pile, click on the MENU button.
When you are in the MENU card, you can directly type a new name of a column in the field
behind the word COLUMN and after that, click on the word COLUMN. If this name of
column doesn't exist, the program creates automatically a new one and add this name in
the index of the pile of the columns.
For the other components of the joints, you can use the same procedure.
If you don't want, for exemple, an upper cleat you must written "no" in the field behind the
word upper cleat. If you leave a blank field, the program will not work! ! !
Annex 3.4
153
If you create a new test with the NEW button of the pile MENU, you see a "?" in thefieldsof
the column, beam and lower cleat: that means that you must define a column, a beam and a
lower cleat to have a joint. The other fields, the reinforcement bars, the concrete,... are
optional. With a new test, the optionalfieldsare "no".
When you are ready with all the components of the joint, you can click on the button
CALCULATION of the card SPRINGS. If you have not made a mistake in the introduction of the
data, after a short moment, you can see the calculated values of the moment versus the rotation. At
the end, you can see the moment-rotation curve. The data of the moment-ration are stored on the
floppy disk in an ASCII file called <name of the test>+".MT". (With the test 36X2C5g you have
the file 36X2C5g.MT). Thisfilecan be open with an EDIT program.
If the program doesn't run, you can check in the file <name of the test>+."DAT" (36X2C5g.DAT
in our exemple) all the data that you have introduce. You can compare it with an existing file.
Hereafter, you have thefile36X2C5g.DAT and 36X2C5g.MT
3.3. IBM program or compatible.
If you use an IBM computer or an IBM compatible, you need to make a ASCII file containing all
the informations for Hie calculation of the moment rotation curve of your joint
In the section 3.3.1. you have a description of all the needed data for calculation.
In the section 3.3.2. you have an exemple of a needed data for the test 36X2C5g.
In the section 3.3.3. you have the ASCII file used for the calculation with comments.
3.3.1. Data for calculation.
Hereafter, you have a list of the data used in the program. You can find a help with the different
figures. The general data, the column data, the beam data, the lower cleat data and the data for
calculation are obligatory, the other data are optional.
154
Annex 3.5
COLUMN
;
figi
stc
uc
fig 2
Annex 3.6
155
BEAM
fig 3
^TTisto
Eb
Stb
Ub
ê
fig 4
156
Annex 3.7
fu
fy
Example of
reference
position 0
Layers
Position X 1 f
fig 6
Annex 3.8
157
CONCRETE
number of springs
(4 8 ori 6)
fu
fig 7
158
Annex 3.9
UPPER CLEAT
r
+
S
7^
7^
fig 8
Position x
na
ma
fig 9
Annex 3.10
159
CLEA
fig io
t washer
IK'VT"";
O washer
screw
ST
figli
160
Annex 3.11
WEB CLEAT
Position X 2 A
A Position X 1
"72
fig 12
HimiMK4rtn>'>fHlrtll>'wMl.
^ <s--\-
^1
fig 13
Annex 3.12
161
LOWER CLEAT
PI bolts per file
fig 14
position X
Friction
coefficient
fig 15
162
Annex 3.13
ections
IPEI
Lx. = IA (mm) 1
H5 {BB
"8W
mr
HE12
HE14A
HE160A
HEI8A
HE2A
HE22A
HE24A
HE2A
HE 28o A
H3A
HE32A
HE34A
HE3A
HE4A
H45A
HE5A
HE55A
HEOA
HE5A
HE7A
HEsooA
94712
"TOS"
HE900A
HE 1000 A
135"
IPE120
IPE160
"TOS"
IOT
gE180
3U5
IPE200
IOT
E>E14o
DE22
I>E24o
TFITo"
IPE300
IPE330
ff3
183T
4T8T
^2T
ISTTT
"3W
"58T"
IPE40
E>E45
IPE5O
IPE550
P6O
n>E 750*137
n>E 750*147
IPE 750*161
IPE 750*173
IPE 750*185
IPE 750*196
IPE 75*21
IPE 750*222
Sections
SS
"58776"
"73TTT
TOS"
"8338"
8T3T
"7535
STTT
Lx=lA(mm)
3nr
"32778"
T735"
T3T
45.49
TOT
"5434"
60.18
^23D"
mr
"7T54743D"
"TOT
~8T5T
"8T9T
9T2T
"SOT
10237"
"T530"
109.55
TI83T
"T2T47T
130.77
103.35
ToTT
fig 16 a
Annex 3.14
163
Sections
E1B
HE 120 6
HE 140 B
HE 160 B
HE 180 B
HE 200 B
HE 220 B
HE 240 B
HE26B
HE2S0B
HE300B
HE 320 B
HE34B
HE360B
HE4B
HE 450 B
HE50B
HE 550 B
HE 600 B
HE 650 B
HE 700 B
HE80B
HE MOB
HElB
L\= 1/X(mm)
sra
173T
"4T4T
"sTT
32T
"5030"
"55l8~
T836"TOT
"TETS"
"8W
"8T7T
"EST
101.08
"T0334109.63
mST
T2T6T
132.88
"T503"
HE120M
HE14M
HE1M
HE 180 M
HE200M
HE 220 M
HE24M
HE260M
HE2SM
HE 300 M
HE32M
HE34M
HE36M
HE40M
HE45M
HE5M
HE 550 M
HE 600 M
HE 650 M
HE70M
HE 800 M
HE90M
HEIOM
"8T2T
86.74
"50T
103.54
10315"
"TEBT
112.82
116.35
TT977T
140.24
"TO"
fig 16 b
164
Annex 3.15
All the data must he in mm and in N.

data
name
"fig"
distance column axis-plane section

Poisson coefficient of the steel
general data
V
column
name of the column
height of the profile
width of the profile
thickness of the web
radius of the fillets
caracteristic data
16
fyCw
E
Est
est
ultimate strenght of the web

yield point of the web
Young modulus of the web
Young modulus strain hardening
strain hardening deformation
2
2
2
2
2
fucf
fyCf
E
Est
est
ultimate strenght of the flange

yield point of the flange
Young modulus of the flange
2
2
2
2
2
name
h
b
tw
tf
r
y
name of the beam

distance
3
3
3
3
3
9
filwb

4
4
4
4
4
name
h
b
tw
tf
r
Lx=lA
column web
fticw
column flange
beam
beam web
fywb
E
Est
est
beam flange
Annex 3.16
165
E
Est
est
ultimate straight of the flange

4
4
4
4
4
number of layers (max 5)
relative position of the spring

number of bars in the effective width
mean area of one reinforcement bar
yield point of a re-bar
ultimate straight of a re-bar
Young modulus of a re-bar
fufb
fyfb
reinforcement
bars
layerni(l...n)
nb
A
fy
fu
E
Est
Est
concrete
nl
X
w
t
fu
upper cleat
name
a
b
s
ri
1
na
ma
yi
X
ne
Tli
W
bolts
166
fu
fy
E
Est
est
number of springs (4,8 or 16)

relative position of the top of the
slab
effective width of the slab
thickness of the slab
ultimate straight in compression
on cylinder
5
5
5
5
5
7
7
7
7
7
name of the upper cleat

length of the upper cleat
width of the upper cleat
thickness of the upper cleat
radius of the fillet
height of the upper cleat
distance
distance
distance
relative position of the upper cleat
distance
distance
distance
ultimate straight of the upper cleat
yield point of the upper cleat
Young modulus of the upper cleat
8
8
8
8
8
9
9
9
9
10
10
10
5
5
5
5
5
|
Annex 3.17
ftib
wd
wt
sd
In
It
Is
Mt
Ab
web cleat
name
a
b
s
ri
1
fii
bolts
fy
E
Est
est
xl
x2
H
sd
wd
bd
Mt
Ab
y2
mb
fiib
gap
Lower cleat
name
a
b
s
ri
fu
fy
E
Est
Annex 3.18
ultimate straight of the bolts

diameter of the washer
thickness of the washer
diameter of the screw
distance
distance
distance
tightening of the bolts
mean area of the bolts
name of the web cleat
length of the web cleat
width of the web cleat
thickness of the web cleat
height of the web cleat
ultimate straight of the web cleat
yield point of the web cleat
Young modulus of the web cleat
relative position of the upper bolt
relative position of the lower bolt
coefficient of friction
diameter of the bolts
distance
distance
gap of the bolts
name of the lower cleat
length of the lower cleat
width of the lower cleat
thickness of the lower cleat
height of the lower cleat
relative position of the lower cleat
ultimate straight of the lower cleat
yield point of the lower cleat
Young modulus of the lower cleat
11 1
11 1
11 I
13
13
13
13
12
5
5
5
S
5
12 I
12
11
11 I
13
12
13
I
H
14
14
15
5
S5 1
s1
167
est
V
fa
nf
nb
bd
Mt
Ab
y3
y4
fub
gap

number of friction areas
number of files (1 or 2)
number of bolts per files (1 or 2)
distance
distance
gap of the bolts
tethamin
tethamax
step
precision
first angle (rad) (different of 0)

last angle (rad)
incrementation of the angle
convergence when the horizontal
equilibrium is lower than precision
inN
5 0
15
14 j
14 I
15
15
15
data for
calculation
Table 1.
3.3.2. Exemple of data for the test 36X2C5g.
data
name
distance column axis-plane section

Poisson coefficient
general data
V
column
name
h
b
tt
tf
r
Lx=lA
name of the column

caracteristic data
HEB200
200
200
9
15
18
51.62
fucw
fycw
E
ultimate straight of the web

416
276
210000
column web
168
Annex 3.19
Est
est

4200
0
fticf

419
274
210000
4200
0
name of the beam

distance
IPE360
360
170
8
12.7
18
15

418
289.5
210000
4200
0
E
Est
est

416.5
279.5
210000
4200
0
number of layers (max 5)
Est
relative position of the spring

number of bars in the effective width
mean area of one reinforcement bar
yield point of B re-bar
ultimate strenght of B re-bar
Young modulus of B re-bar
ni
number of springs (4,8 or 16)
column flange
fyCf
E
Est
est
beam
name
h
b
tw
tf
r
y
beam web
fiiwb
fywb
E
Est
est
beam flange
fiifb
fyfb
reinforcement
bars
layerni(l...n)
nb
B
fy
fu
E
Est
concrete
Annex 3.20
0
6
82.5
578.25
489
210000
4200
0
60
6
82.5
578.25
489
210000
4200
0
169
w
t
fu
upper cleat
name
a
b
s
ri
1
na
ma
yi
X
bolts
ne
mc
w
fu
fy
E
Est
est
fub
wd
wt
sd
In
lt
s
Mt
Ab
web cleat
name
a
b
s
ri
1
fu
fy
E
Est
170
relative position of the top of the

slab
effective width of the slab
thickness of the slab
ultimate strenght in compression
on cylinder
-30
1200
120
12.45
name of the upper cleat

length of the upper cleat
width of the upper cleat
thickness of the upper cleat
height of the upper cleat
distance
distance
distance
relative position of the upper cleat
distance
distance
distance
ultimate strenght of the upper cleat
yield point of the upper cleat
Young modulus of the upper cleat
150x90x13
ultimate strenght of the bolts

thickness of the washer
distance
distance
distance
800
37
4
37
16
3
34
220000
353
name of the web cleat

length of the web cleat
width of the web cleat
thickness of the web cleat
height of the web cleat
ultimate strenght of the web cleat
yield point of the web cleat
Young modulus of the web cleat
150x90x13
150
90
13
12.5
170
40
50
55
-10
50
50
35
392.5
240.5
210000
4200
0
150
90
13
12.5
170
408
272.5
210000
4200
Annex 3.21
bolts
est
xl
x2
relative position of the upper bolt

relative position of the lower bolt
distance
distance
gap of the bolts
sd
wd
bd
Mt
Ab
y2
mb
fub
gap
Lower cleat
name
a
b
s
ri
1
X
fu
fy
E
Est
est
V-
fa
nf
nb
bd
Mt
Ab
y3
y4
fub
gap
data for
calculation
Annex 3.22
tethamn
tethamax
step
precision
name of the lower cleat

length of the lower cleat
width of the lower cleat
thickness of the lower cleat
height of the lower cleat
relative position of the lower cleat
ultimate strenght of the lower cleat
yield point of the lower cleat
Young modulus of the lower cleat
number of friction areas
number of files (1 or 2)
number of bolts per files (1 or 2)
distance
distance
gap of the bolts
first angle (rad) (different of 0)

last angle (rad)
incrementation of the angle
convergence when the horizontal
equilibrium is lower than precision
inN
171
For the values of Est and est, if you put "0", the program takes by default the following
values:
Est = E/50
est = 10 * ey
3.3.3. Exemple of the ASCII file with the 36X2C5g data.
Name of the General Data File:

36x2C5gX>AT
comments
1 help word at the first line
ldata
General Data
Length unity (mm),Load unity (N)
Distance Column axis-plane section
182.5
Poisson Coefficient
0.3
insert more than one blank line

keyword : it cannot change
2 help words separated by B comma
1 help word
ldata
1 help word
ldata
Name of the column

HEB200
h,b,tw,tfj J-x= 1/lamda
200,200,9,15,18,51.62
Web data: fucw,fyCw33st,epsst
416,276,210000,4200,0
Range data: fucf,fycf,E,Est,epsst
419,274,210000,4200,0

ldata
6 help words separated by comma
6 data separated by comma
Name of the beam

IPE360
h,b,tw,tfT,y
360,170,8,12.7,18,15
Web data: fuwb,fywb.EEstepsst
418,289.5,210000,4200,0
Flange data: fiiflfyfb,E,Est,epsst
416.5,279.5,210000,4200,0

ldata
Reinforcement bars
yes
Number of layers (max5)
2
position xl .number of bars^nean area
0,6,82.5
fu,fy,E3st,epsst
578.25,489,210000,4200,0
172

data: yes or no : if no, not take into account
ldata
insert 1 blank line
Annex 3.23
position x2,number of bars,mean area

60,6,82.5
fu,fy,E,Est,epsst
578.25,489,210000,4200,0
Concrete
yes
number of layers (4 8 or 1 opposition of the
top, width, thickness, fu
4,0,1200,120,12.45
Upper cleat
yes
Name of the cleat
150x90x13
B,b,s,rl,l
150,90,13,12.5,170
na^na,yl,position x^^nc.w
40,50,55,-10,50,50,35
fu,fy,E,Est,epsst
392.5,240.5,210000,4200,0
fub, washer diameter.washer thickness
800,37,4
screw diameterJn,lt,ls,tightening,mean area
37,16,3,34,220000,353
Web cleat
yes
Name of the cleat
150x90x13
B,b,s,rl,l
150,90,13,12.5,170
fu4yEst,epsst
408,272.5,210000,4200,0
position xl,position x2,friction coefficient,screw
diameter.washer diameter
220,320,0.5,37,37
bolt diameter,tightening,mean area,y2,mb,fub,gap
20,150000,245,80,50,640,1
Lower cleat
yes
Name of the cleat
150x90x13
B,b,s,rl,l,position x
150,90,13,12.5,170,456.5
fu/yÊst,epsst
Annex 3.24
insert 1 blank line between the layers

1 help data
ldata
1 help word
ldata
1 help word
ldata
173
408,272.5,210000,4200,0
friction coefficient^riction areali file (1 or 2), n
bolt (lor 2)
0.5,1,2,2
bolt diameter,tightening,mean area,y3,y4,fub,gap
24,220000,353,55,55,640,1
Data for calculation

first angle(rad),last angle(rad),step,precision
0.001,0.05,0.001,10

keyword : It cannot change
If you want remove the upper cleat of this test for exemple, you can put directly "no" instead of
"yes" without remove the characteristics of the upper cleat. Thus you have :
Upper cleat
no
Name of the cleat
150x90x13
B,b,sjl,l
150,90,13,12.5,170
na,ma,yl,position xîcorjcw
40,50,55,-10,50,50,35
fu4yfiEst,epsst
392.5,240.5,210000,4200,0
normal stress.washer diameter,washer thickness
800,37,4
screw diameter,ln,lt J8,tightening,mean area
37,16,3,34,220000,353
Web cleat
It is equivalent to :
Upper cleat
no
Web cleat
For other exemples, you can use an editor programs to see on the floppy disc all the programs with
the extension ".DAT"
To run the program, when you have made the <ascii file name> + .DAT, you must type
A:
SPRINGS and the key RETURN.
(It is also possible to make a copy of the program on a hard disk.)
The program ask you the name of your file (without .DAT).
The program will show you the values of the Moment versus the Rotation. These values will be
stored on the used drive with the name oscii file name> + ".MT"
At least, the program shows you the Moment Rotation curve of your joint.
174
3.25
3.4. Calculation of the springs.

3.4.1. Reinforcement bars.
dlisee tablei)
general data :
L
v
reinforcement bars : n
nb
A
fy
fu
E
E^
%
For each layer of reinforcement bars we have a spring with the following behaviour
F (load) - A (displacement)
diagram F - A of the spring :
Ayb
A stb
A ub
calculation of the values :

F y b = nb*A*fy
Ayb=*L
eyb=|
Astb =(st - yb)*L + Ayb
Fub = nb*A*fu
AubÊst + ^ ^ ^ L
Annex 3.26
175
3.4.2. Concrete,
data: (see table 1)
general data
concrete
nl
fu
The concrete slab is divided in layers (4 8 or 16). Each layer is represented by B spring,
calculation of the values:

_ fu*w*t
uc
~ nl
i
E = 9500*(fu + 8)3
F
c uc
uc=ec*L
2
ft = 0.3*fu3
Fot-
ft*w*t
ni
F ctg = 0.6*Fct
Um = 0.0015*L
U
U c r =*L
E, L
176
Annex 3.27
3.4.3. Upper cleat.

data: (see table 1)
general data
column
tf
beam
flange
Ecf
Est,cf fycf
web
E cw
Est,cw rUcw
upper cleat
mc
w
bolts
rub
wd
diagram F - A of the spring
AF
s
ru
wt
fucf
ri
fy
sd
1
E
In
na
Est
It
ma
est
Is
yl
Mt
Ab
nc
k4 = 0.1*l + 0.4*wt
k"3 =
(S + tf)
tf of the column
k2 = 18 + 1.43*1, + 0.91*ln + 0.8*wt

kl = 18+1.43*1,+ 0.71 *ln
k, = s + T*rl
La=yi-ka
m' c = mc - ( ^ + T*r J t^j of the column
ify = 0 then m'a = m a - k a
s
i f y * 0 then m'8 = m a - x
Annex 3.28
177
ify*0
then La = La
.
La*(L^ + 4kJ)
ify = 0 then 1* =
r-2(L a +kJ 3
l a = 2*(m.'+0.75*nJ
lcf=2*(mcf,+0.75*w)
n
a a = 0.75*-p
*a
a cf =0.75*p-
cf
2 (
'a^
-8*a;
aal = 1.5*aa*[l + y-2*ai
a cfl = 1.5*a cf -2*aj f

wa = 1 ( 1 of the upper cleat )
*-4s *wa
wcf=2*na
z-
tf3*wcf
if the bolts are prestressed (Mt = 0):

Zrf*acfl+2*Za*ga
q _ q s
"^ *
0*7*
kl+2*k4
Zrf*acf2+2*Za*aa+^j
if the bolts are not prestressed (Mt o 0):
178
Annex 3.29
Zcf*acfl+2*Za*(xa
q = qt=.
k2*k 3
Z rf *a cf2+ 2*Z a *a a+ 2*Ab*(k2+k 3)
K_f
cf
rZrf
^ * ( 0 . 2 5 + 0.5*q*acn)
Ka =
Ki_
Za (
Z*l*\\
J_ + _l_
Ka
Kcf
Si
( 1 - e ^ '* cos <&))
KH =
( in radian)
2*L>, ^ * E W \
( h - 2*(tf+r))
K,i =" 1
KT
1
K
K = 1.2
Nbd = 2*k*Mt
e =7 max (sd;wd)
b ^ = min (4*m'c + 1.25*nc ; 2*7i*m'c)
n' c = min (nc ; \.25*tt
R=1.00
mpc = 0.25*tf2*fycf
B ;=o. 9 *A b
Annex 3.30
179
Frdl = 2*B t
m pc l W *R(8n , c -2e)
2*m c *n' c -e*(m c +n' c )
Frd3*n' c + 2* mpc*bmc*R
if
2*Hv^
Nbd
*~
m ^ b ^ R ^ S * K*n' c -2*(K-l)*e) + 2*n'c*K*Mt

1x13
2*K*m c *n , c -e*(K-l)*(m c+ n' c )
"
F v c f =min (F rfl ; F ^ ; F ^ )
n' a = min (na ; 1.25*01',,)
mpa = 0.25*s2*fy
^ma = 1
( 1 of the upper cleat )
Frii = 2*Bt*
m pa *b ma +2*B;*n' a
(m a + n' a )
hrf2_
mpa*b|na*( 4 * n ' a - e )
Id3
"2*m a *n' a -e*(m a + n , a )

F
tf
rd3*n' a +
pa* b n*
2^nV^
^Nbd
mpa*bto.(4 K*n' a - (K-l)*e) + 2*n'a*K*Mt

1x13
2*K*m a *n' a -e*(K-l)*(m a +n , a )
F v a =min (FnnFrfaFnu)
**vli = "mc tycw *w
F v = min(F v c f ; F v a ; F v l i )
180
Annex 3.31
'M
K2
- ^ a
=
Kcf
K3 = KH
if F v = F va then
F
= 19F
' 7
v,up
Ki=Ki*l (cleat)
if ( F vcf < FViUJ then K2 = K2 * - g ^
if (v lFv,h- v< rFv , u p ^ uthen
icii
11
K, - JS-3
K~ *
IV3
E st
'wc
wc
if(y = 0 ) then Fty = F v

i f ( y * 0 ) then E
Ki = 2*Ki * -gs t (cleat)

F v s = min(F v c f ; F vli )
if ( FV)S < FV)Up) then
1
Ks
ty
My
~_L J_
K 3 K2
_ FV*K. + FV|S*K!
K s + K,
else
=
Mr
if F v = F vcf then
Fv,up= 1.65 F v
K2 = K 2 * % ^ f
b
cf
if ( y = 0 ) and ( F va < Fv>up) then Kj = Kj * -f
( cleat)
E
then K, = 2*Kj * - (cleat)

E,
< Fv>up) then K3 = K3 * -st^,wc
-
if ( y * 0 ) and ( F va < F
if ( F v l i
,up.)
if(y = 0 ) then Fty = F v

i f ( y * 0 ) then
FV8 = min(Fva;Fvli)
if ( F v s <FViUp) then
1
K
g
~J_ J_
K
Fv*K8 + FViS*K2
F,
= ty
K+K 2
else
My
Annex 3.32
Ey
181
if Fv = F vli then
F
= 1 65 F
K3 = K 3 * % ^
Es,
if(y = 0)and(F v a <F v u p ) then K ^ K j * - ^ (cleat)
Est
if ( y * 0 ) and ( Fva < F vnp ) then Kj = 2*Kj * -^ (cleat)

if ( F vcf < FV)Up) then K2 = K2 *
if(y = 0 ) then Fty = Fv

i f ( y * 0 ) then
F V8 = min(F vcf ;F v a )
if ( Fv>g < F vup ) then
1
E
_L _L
ty
K,8t
My
_ FV*K. + FV,.*K3
K 8 +K 3
else
=
*v
~J_ J_ _L
Kj K2 K3
Ftu = min(F u c f ;F u a ;F u l i )
B ^ A b + fub
muc=0.25*t?*fucf
2*(muc*bmc*R + B > ' c )

tM
(nie+n'e)
FKj3 =
muc*bmc*R*(8*n'c-2*e)
2*m c *n' c - e^nic+n'c)
Frf3*n,c + 2*muc*bmc*R
tf
2^7^
Nbd
ta
_ mnc*bmc*R*(8* K*n' c -2*(K-l)*e) + 2*n'c*K*M,

1x13
2*K*m c *n , c -e*(K-l)*(m c +n , c )
F ucf =min (Frf, ; FnQ ; FnB )
182
Annex 3.33
uli- b
* t * fii
nipU = 0.25*s2*fii
F
1
= 2*B"*(l.OOu
e
"n' ; ) -
ny* 1
n'.
Fty
Fl =
fl.OO-i]
K
tiJ
Ni
.
ut-
(FtO-Fty)
\C
^tst
3.4.4. Web cleat.

data: (see table 1)
general data
beam
web
webe ;leat
tw
fuwb
a
ef
b
sd
s
wd
rl
bd
1
Mt
fu
Ab
fy
y2
E
mb
Est
fub
est
gap
diagram F - A of the spring

n F
Fyc
traction - compression in the web cleat
Annex 3.34
183
without slip
Fugt
Fygt
, uge
gap
Fyge
gap
ugt
Fuge
slips, ovalisation, shear of the bolts

These diagrams represents the modlisation of one bolt of the web cleat. They must be added to
have the full diagram F - A (the addition is made only on the displacements. It is like two springs
in series). Our web cleat have two bolts : thus , we have two springs.
2 bolts
2 springs
'VWW-
V\MAr
calculation of the values :

1/2 web cleat in compression
Le=V2-(s+5*rlJ
F
i yC
=fv*S
iy
184
lJ
Annex 3.35
fy * 1
yc-~~
Ovalisaon , shear in the bolts, compression in the cleat

d = max (sd ; wd )
10
K=
Hf)*(
S=
Loo+a78
*)
Mt
0.18*bd
Ybi = 1.00
Yb2=1.25
if s 2 0 mm
then ^ c =TT
if 20 mm s 30 mm then ^ c = - 0 . 5 * i - ^ J
if s 30 mm
then Tfe = 2.5
if t, 20 mm
then ia =
+3*^-2
^T + 3 *- 2
if 20mmt w 30 mm then Tia = -- 5 *(

if ty, 30 mm
then ^ a = 2.5
K c = 12.5*Ybl*Yb2*Ytc*bd*fu
Kg = 12.5*Ybi*Yb2*Yu*bd*fuwb
K
L_
_L _L
Kc
Ks
Kb = 0.93*Ab*fub
Annex 3.36
185
L_
B"= 0.5*A b *fub

a = min[l.OO;-^J
=> Rjc = 2.5 *a*fu*bd * s
a = min i 1 . 0 0 ; ^ J => Rj s = 2.5 * a *fuwb * bd * t

F ygc = ^ * S
FIW = miii(B;;l c ;Rj 1 ) + F ygc
A
ugc=(FugCK,Fygc)^*gap
Ovalisaon, shear in the bolts, traction in the cleat :

Y bl. Y ta and Y te don't change
,
tw-y2
M-
=> if d 1.5
then the web cleat hole istoo near from the side!
= > i f l . 5 < d < 3 then Yb2= 0 - 2 5 *d +0.5

=>ifd3
then 7^ = 1.25
K c = 12.5*Ybi*Yb2*Ytc*bd*fu
K8=12.5*Ybi*Yb2*Yta*bd*fuwb
K,=
J_ J_
Kc
Kg
Kb = 0.93*Ab*fub
186
Annex 3.37
'-_l_
J_
Kt
Kb
B"= 0.5*A b *fub

K
(bd + gap)
t w -y 2
^
a = min i ; fu ' 3*(bd + gap) )
wb
* bd
KH *
*s
l=2.5*a*fiiwb*bd*tw
FVBt = n * S
F ugt = m i n ( B ^ ; R ; c ; R ; s ) + F ygt
_ ( F ugt
A a t1
"S
Kr
ygt)
+2* gap
1/2 web cleat in traction :

Fugt
U= K
Annex 3.38
187
3.4.5. Lower cleat.

data: (see table 1)
general data
column
web
tw
Ecw
fycw
flange
tf
fyfb
fUfb
a
fa
b
nf
s
nb
beam
lower cleat
ri
bd
Mt
ry
y3
E
y4
Est
fub
est
gap

Compression of the cleat :
F
A yc A sto
, uc
diagram 1
188
Annex 3.39
Compression of the column

if F cu > F ej
A est
F 1i
F ce
if F eu < F cj
y*
\ K cj |
/ 1
'
K ci
^-
CJ
diagram 2
Ovalisation, shear in the bolts :
Annex 3.40
189
without slip
diagram 3
These three diagrams must be added to represent the full diagram F - A of the lower cleat, (the
addition is only made on the displacements like 3 springs in series ).
calculation of the values:
lower cleat:
S = s*l
(1 of the web cleat}
_fy
^c - E
F yc = fy*S
Lc=y3 + y - ( s + 5 * r l J
Aye = Lc yC
if Egt = 0 then
A
EA = 10 * yC
(eA of the lower cleat )
StC=LC*(Est-eyc)+Ayc
Fuc = fu*S
Auc
Lc
lEst
Egt
Compression of the column web:

Lp = 2*s + (2 - VT)*rl + 5*(r + tf )
*cppl = 1-p ^ Vf
fe-
2*s + (2-VT)rl
2*LX
190
Annex 3.41
Si
m = (I_e(-i)*c0s(.))
( j inradian)
F cc = 2*tw*Mi*fy*Lx
k=1.00
Fccr=(h-2*tf)X*k*
*(l-v2) v h - 2 %
Fcb = î*V F ce* F ccr
F cu = max (Fcppi;Fcb)
Kci - 2*Lx*jii*Ecw*( h _ 2* (t,+ r ))
.
"ce
cc=T~
*-ci
if Egt = 0 then
Acst - Est*
Kcst
e a = IO * I -p I ( e^ of the column web )
(h-2*(r + t f ))
C *F
_ v "cu "cppl )
cu
V
^cst
"cuy
^p Mv **cw
if F cu > F cuy then Fcu = F cuy

Fj = Fcppi + Kcgt Aggi
if F c n <Fej then
cj
P F
__ r cj rcppl
'i=
cst -
Kcj
ce
+ A
"
Ovalisation, shear in the bolts, compression in the lower cleat
Annex 3.42
191
0.18*bd
Fgy = nfa*nf*nb*n * S
Calculation of K r :
ifnf=l then 7^ = 1.00
Yb2=1.25
ifnf=2 then 7bl=- 9
if Y4 ^ 20 mm then the 2 files of the bolts are too near !
d=
bd
if d < 3.5 then yb2 = 1.25 + 0.25 * d + 0.375
if d 3.5 then f = 1.25 + 1.25
if s 20 mm
s
then %. = ^
if 20mms 30mm then ^ c =0.5 * ( ^ )

if s 30mm
then^ c =2.5
if tf20mm
then ^
tf30mm
*^j
- 2
=^
if 20mmt f 30mm then ^ = - 0 . 5 * [ J Q J

if
+ 3
+3*JQ-2
t h e n ^ = 2.5
Kc == 12.5*nf*Ybl*Yb2*Yta*bd*fu
K8 == 12.5*nf*Ybi*Yb2*Yta*b<i*filfb
K,=" 1
Kc
1
+
1
K8
Kb = 0.93*nf*nb*Ab*fub
192
Annex 3.43
L_
Calculation of Fgu :
B"= 0.5*A b *fub
a = m i n
1 ;
3*(bd y +gap)- ) =* R5c = 2.5*a*fu*bd*s
Fgu = nfWmin ( B^ ; Rc ; Rj, ) + Fgy

u
(F - F )
jf
6aP
3.5. Conclusions.
The results of the simulations of the composite cleated connections are shown on the following
pages. We can make the following remarks:
- In all the cases, the mathematical model gives a safe value of the ultimate moment. This is due to
the fact that we have used the nominal tolerances of the web profiles to calculate the column web
buckling load. In all the tests, the actual initial deformation of the column web is lower than those
given by the nominal tolerances.(see tests 07(30X30), 08(30X3C2),.). When the collapse is not
due to a local buckling of the web column, the mathematical model gives a good assessment of the
ultimate moment(see tests 18(30X2C5), 28(36X2C2),....).
- The mathematical model gives a very good assessment of the rigidities. There is sometimes some
differences between the test and the model because of a misfunctioning of the rotative transducers
during the test.(see test 27(36X2C1)).
- With this mathematical model, we dispose of a tool to predict the moment rotation curves of
composite cruciform joints with cleated connections.
Annex 3.44
193
07(30x3c1)
350000000 -r
(O
300000000
250000000
200000000
test
E
E 150000000 ~
without slip
with slip
100000000 -f- /
/
50000000
o
K
UI
0,005
0,01
0,015
0,02
0,025
-50000000 -^
rad
0,03
0,035
0,04
0,045
0,05
08(30x3c2)
o
X
OJ
JCk
O)
300000000 T
250000000 --
200000000
test
150000000
E
E
z
without slip
with slip
100000000
50000000 --,
0,005
(O
0,01
0,015
0,02
0,025
50000000
rad
0,03
0,035
0,04
0,045
0,05
09(30x3c3)
(O
0>
250000000 -r
200000000 --
150000000 --
test
E
E
z
without slip
witht slip
100000000
50000000
X
O)
0,01
0,01
0,02
rad
0,03
0,04
0,05
10(30x3c4)
>
3
3
(D
X
k
00
300000000 -r
250000000
200000000 -test
E
E
z
150000000 -f-
without slip
//
'/
with slip
100000000
50000000
(O
0,01
0,01
0,02
rad
0,03
0,04
0,05
11(30x3c6)
(O
00
300000000 -r
250000000 --
200000000 --
150000000
test
E
E
z
without slip
100000000 --
with slip
50000000
3
<D
X
0>
k
0,005
0,01
0,015
0,02
0,025
-50000000 -1rad
0,03
0,035
0,04
0,045
0,05
12(30x3c7)
350000000 -r
X
CO
In
o
300000000
250000000
200000000
test
E
E
150000000
without slip
with slip
(O
(O
0,05
-0,01
-50000000
-1rad
13(30x3c8)
io
300000000 T-
250000000
200000000 --
150000000 --
test
E
E
z
without slip
with slip
100000000 --
0,05
0,01
w
in
-50000000 - 1 rad
14(30x2c1)
ui
ro
300000000 -r
250000000
200000000
test
E 150000000
without slip
with slip
100000000
50000000
ro
o
0,005
0,01
0,015
0,02
0,025
rad
0,03
0,035
0,04
0,045
0,05
15(30x2c2)
250000000 -r
ro
200000000
y^"
150000000 --
test
E 100000000
with slip
without slip
N
50000000 --;
0,005 0,01
w
in
w
0,015 0,02 0,025 0,03
50000000 -1-
rad
0,035 0,04
0,045 0,05
16(30x2c3)
350000000 -r
u
In
300000000
250000000 --
E
E
200000000
test
without
150000000 -f
with
i
I
I
I
I
100000000 -|- /
/
/
/
50000000 +J'
ro
o
w
-0,01
0,01
0,02
rad
0,03
0,04
0,05
slip
slip
17(30x2c4)
300000000
io
250000000
200000000 --
150000000
test
E
E
without slip
100000000
o
in
with slip
0,05
-0,01
ui
-50000000 -1rad
18(30x2c5)
250000000 nr
co
In
O)
200000000
150000000
test
without
E 100000000
slip
with slip
50000000
0,01
0,02
0,03
ro
o
o
-50000000 -1rad
0,04
0,05
19(30x2c6)
ro
8
300000000 -T
250000000
E
E
200000000 -test
without slip
150000000 --
with slip
u
in
-I
0,05
20(30x2c7)
w
Ol
OS
350000000 T
300000000
250000000
200000000
test
E
E
z
without slip
150000000
with slip
100000000
50000000
0,05
0,01
ro
o
-50000000 -rad
21(36x3c1)
300000000 -r
ro
o
oo
250000000 --
200000000 -/
/
150000000 -|-
test
E
E
z
without slip
with slip
100000000
50000000
0,005
0,01
0,015
0,02
0,025
bi
(O
-50000000 -*rad
0,03
0,035
0,04
0,045
0,05
22(36x3c2)
350000000 -r
E
E
300000000
250000000
200000000
test
without slip
with slip
150000000
100000000
ro
o
co
-0,01
0,01
0,02
rad
0,03
0,04
0,05
24(36x3c5)
350000000 -r
to
o
300000000 --
250000000 --
200000000
test
E
E
z
without slip
150000000 --
with slip
100000000 --
c*
b)
0,05
-0,01
-50000000 -1rad
25(36x3c6)
400000000 -r
o
b)
ro
350000000
300000000 --
250000000
test
E
E
without slip
200000000
with slip
150000000
100000000 -
ro
0,01
0,01
0,02
rad
0,03
0,04
0,05
26(36x3c7)
400000000 T
ro
ro
350000000
300000000 --
250000000
test
200000000
E
E
without slip
with slip
150000000 --
100000000 --
50000000 --
0,05
50000000 -1rad
27(36x2c1)
350000000 -r
300000000 --
250000000
_.
200000000
^ 150000000 -f
100000000 -\-J
test
/
/
/
I
I
without slip
with slip
50000000
(i
CO
0,005
0,01
0,015
0,02
0,025
-50000000 ^
rad
0,03
0,035
0,04
0,045
0,05
28(36x2c2)
300000000 -r
250000000 --
200000000 --
test
150000000
E
E
z
without slip
100000000 --
with slip
50000000
0,01
o
0,07
-50000000 -1-
b
rad
29(36x2c3)
400000000 -
O)
S*
350000000 /
/
300000000 -
/
/
'
'
--------
T^-
.-'
250000000 E
E
z
/
/
200000000 -
/
/ /
/ /
150000000 - 1 1
f/
t
9
9
1
test
without slip
1
1
1
1
f
f
with slip
1
t
1
100000000 -
L-'
50000000 - C
ro
I
I
-0,01
n
o
0,01
0,02
0,03
0,04
rad
1
0,05
30(36x2c5)
400000000 -T
ro
05
350000000 --
300000000 --
250000000
E
E
z
test
200000000 --
without slip
150000000
100000000
with slip
50000000
u
in
(i
0,005
0,01
0,015
0,02
0,025
50000000 -J-
rad
0,03
0,035
0,04
0,045
0,05
31(36x2c6)
bl
b>
oo
350000000 -r
300000000
250000000
E
E
200000000
test
150000000 -f
without slip
I
I
with slip
100000000 __/
50000000
0,01
ro
0,01
0,02
-50000000 -rad
0,03
0,04
0,05
32(36x2c7)
ro
oo
300000000
250000000
200000000
test
150000000
E
E
without slip
with slip
100000000
50000000
0,07
-0,01
OJ
50000000 -1-
b)
(O
rad
33(24x3c1)
200000000 -r
o
180000000
160000000 --
140000000
120000000
test
E
E
z
100000000 +
without slip
with slip
80000000
60000000
40000000
ro
co
0,01
0,01
0,02
rad
0,03
0,04
0,05
34(24x3c2)
ro
ro
o
250000000 -T-
200000000
150000000
test
E
without slip
with slip
100000000
0,01
0,01
0,02
rad
0,03
0,04
0,05
35(24x3c3)
300000000
o
to
250000000
200000000
E
E
z
witout
150000000 --
slip
with slip
l>0
0,01
0,05
36(24x3c5)
ro
io
ro
200000000 -180000000 -160000000 -140000000 -120000000 -E
test
100000000
without slip
with slip
80000000 +
60000000
40000000
20000000
0
0,005
0,01
0,015
0,02
0,025
-20000000 -1rad
0,03
0,035
0,04
0,045
0,05
37(24x3c6)
250000000 -r
200000000
150000000
test
E
E
without slip
with slip
0,05
-0,01
l\3
(O
50000000
-1-
rad
38(24x3c7)
IO
C
300000000 x
250000000 --
200000000 -test
E
E
with slip
150000000 --
without slip
100000000 --
0,01
0,01
0,02
Ol
rad
0,03
0,04
0,05
39(24x2c1)
O)
180000000
160000000 -140000000
120000000
100000000
E
E
z
Test
without slip
80000000
with slip
60000000
/
40000000
20000000
0M
0,01
20000000 -1-
0,01
0,02
Ol
rad
0,03
0,04
0,05
40(24x2c2)
250000000 -r
to
ro
O)
200000000 --
150000000
Test
E
E
z
without slip
with slip
100000000
50000000 --
co
0,005
0,01
0,015
0,02
0,025
rad
0,03
0,035
0,04
0,045
0,05
41(24x2c3)
300000000
o
09
E
E
250000000
200000000
150000000
test
without slip
100000000
with slip
0,05
-0,01
ro
ro
-50000000 -rad
42(24x2c5)
io
ro
ao
200000000 -r
180000000
160000000 --
140000000
120000000
test
E 100000000
wihout slip
with slip
80000000 --
60000000
40000000 -\-
20000000 -l
w
<o
0,005
0,01
0,015
0,02
0,025
rad
0,03
0,035
0,04
0,045
0,05
43(24x2c6)
250000000 -r
w
bo
o
200000000
150000000 -test
E
E
without slip
with slip
100000000 --
50000000
fO
ro
co
0,01
0,01
0,02
rad
0,03
0,04
0,05
44(24x2c7)
CO
O
300000000 nr
250000000
200000000
test
150000000 -E
E
without slip
100000000 --
with slip
50000000 --,V'
0,01
u
0,01
0,02
-50000000 -1-
00
rad
0,03
0,04
0,05
Annex 4.
Databank SERICON (SEmi-RIgid CONnection) and data sheets of the 56
tests.
Floppy disc with the program SERICON.
(The floppy disc is available at ARBED Recherches
tl. : (352)- 5551-2177)
m&m'
Ut, ':
zu-
.'fSa
:**V ;? v;
=- J -
'.Mt -
'.:.': m
: / .
TOI S lc? *"
it
>/
\i
'
4.1. Installation procedure of SERICON version 1.0.

1. Enter the floppy disk on drive A.
2. Type
A:
INSTALL
the databank system will Deinstalled on your hard disk.
3. Before you start SERICON, you must add in your system file C:\AUTOEXEC.BAT
the line
SET CLIPPER=F50
and set the value for FILES in your system file C:\CONFIG.SYS
FILES = 50
4. Reboot your system (press ALT+CTRL+DEL). Go to directory C:\SERICON by
typing:
CDC:
CD\SERICON
Then start the program by typing
SC
4.2. To start SERICON for the first time.
1. To start SERICON type
SC
2. Choose a type of connection (e.g. CLEAT) by selecting CLEAT with cursor keys
and confirm with <RETURN>
3. Press function key F9 to come to the MENU.
4. Choose configuration by selection with cursor key and press <RETURN>.
5. Enter the default directory where you have installed SERICON, e.g.:
CNSERICON
and press <RETURN> to confirm.
6. Enter the directory for the curves, in this case:
C:\SERICON\CURVES
and press <RETURN> to confirm (the directory name CURVES cannot be changed)
7. Press function key FIO to come back to main screen.
4.3. Problems to install?
If you have problem to start SERICON, please contact:
Dipl.-Ing. Klaus WEYNAND
Institute of Steel Construction
RWTH Aachen
Mies-van-der-Rohe-Str. 1
D-5100 Aachen
Germany
Telephon: +49 241805181
Telefax:
449 241805178
Annex 4.2
233
4.4. Exemples of data sheets.

The following pages show two exemples of data sheets printed with the program
SERICON. The printer must be a HP laser m.
234
Annex 4.3
CONNECTION
TYPE
SINGLE WEB AND UPPER AND LOWER FLANGE

CLEATED CONNECTION
TEST NO PAGE
103.003
1
AUTHOR(S)
TEST CENTRE
Schleich, J.B.; Grardy, J.C.; Jaspart, J.P

University of Lige
TEST DATE
October 1988
SOURCE OF
DATA
Semi-rigid action in steel frame Structures (CECA - No

7210 - SA/507)
SPECIMEN DETAILS
TESTARRANGEMENT
:
TEST
:
MATERIAL STANDARD :
SPECIMEN REFERENCE:
COLUMN ORIENTATION:
MEMBERS SIZES AND REFERENCES

canteiiver
in plane
Eurocode 3
3 0T3C1
strong axis
COLUMN WEB STIFFENING

: no
COLUMN FLANGE STIFFENING : no
BEAM:
IPE 300
Fe 360
COLUMN:
HE 200 B
Fe 360
NUMBER AND SIZE OF BOLTS : 12 M20, 4 M24

GRADE OF BOLTS
: 8.8 (DIN 931)
WASHERS
: yes
BOLT TIGHTENING PROCEDURE:
hand controlled tightening
CONNECTION:
cleat 150*90*10, St 37-2
cleat 150*90*10, St 37-2
DRAWING OF THE CONNECTION AND THE CONNECTED MEMBER
IPE 300
HE 200 B
BEAMS FRAMING INTO THE COLUMN FLANGE
COMPOSITE BEAM: no
DATA SHEET printed with SERICON Vers. 1.0, 21.11.91

235
TEST NO PAGE
2
103.003
ACTUAL CHARACTERISTICS OF THE JOINT ELEMENT
COLUMN AND BEAM(S)
column
beam_l
beam_2
beam_3
beam_4
202.8
( 200.0) ( 300.0)
201.9
( 200.0) ( 150.0)
tf
14.2
15.0) (
9.1
9.0) (
7.1)
18.0) (
15.0)
tw
r
10.7)
hi
\1
4c
t4c
1.
units: [mm]
measured average values of h, b, t and r

(-) nominal values
ENDPLATE, CLEATS, T-STUBS, WELDS, BOLTS, HOLES,
cleat 1
where web
a
cleat 2
flanges
149.4
149.4
( 150.0) ( 150.0)
cleat 3
90.0
90.0) (
90.0
90.0)
10.1
10.0) (
10.1
10.0)
ri
12.5) (
12.5)
r2
6*5) (
6.5)
measured average values of a, b, t and r

(-) nominal values
BOLTS
/H
units: [mm]
type 1:
12 M 20, 8.8 (DIN 931)

clearance between bolts and holes:
bolts with washers
hand controlled tightening (15 kgm)
1 mm
type 2:
4 M 24, 8.8 (DIN 931)

bolts with washers
hand controlled tightening (22 kgm
1 mm
bolts M 24 only used in flange cleats to connect

the cleat to the column
236
r.
-V
MECHANICAL CHARACTERISTICS OF BEAMS, COLUMN

AND CONNECTING STEEL ELEMENTS
grade
E
[N/mm 2 ]
COLUMN
flange
Fe 360
web
BEAM 1
flange
Fe 3 60
web
St 37-2
CLEAT 1
long leg
short leg
St 37-2
CLEAT 2
long leg
short leg
fy
[N/mm a ]
est
TEST NO PAGE
4
103.003
E st
[N/mm 3 ]
fu
[N/mm 3 ]
274.0
(210000) ( 235.0)
276.0
(210000) ( 235.0)
419.0
( 360.0)
415.0
( 360.0)
279.0
(210000) ( 235.0)
314.5
(210000) ( 235.0)
419.5
( 360.0)
428.5
( 360.0)
(210000)
(210000)
(210000)
(210000)
287.0
473.0
298.0
478.5
287.0
473.0
298.0
478.5
(-) nominal values
237
TEST NO PAGE
103.003
5
EXPERIMENTAL TESTING ARRANGEMENT
in
m
m
2239.0
units:
[mm]
INSTRUMENTATION
TYPE
LOCATION
DEFLECTION MEASUREMENT
ROTATION MEASUREMENT
LOAD MEASUREMENT
Remark: for more datails see SOURCE OF DATA (see page 1)

238
TEST NO PAGE
6
103.003
LOAD BEHAVIOUR CHARACTERISTICS
EXPERIMENTAL M - *
CURVES
1
2
3
CONNECTION
SHEAR
TOTAL ROTATION
M [kNm]
1?5
-,
T
1
1
1
1
1
1
2
100,
75
50 25 -
0 i
i
1
-.02
.00
^r
'
! 3
1
1
g ^
-|
1
1
1
1
1
1
1
1
.02
_^î
^-^-^
T
1
'
1
1
\\
)\
r
1
1
//
'
'
II
II
' n
1
1
' II
' II
1
1
'
'
'
.04
.06
.08
phi
[rad]
FAILURE MODE
excessive yielding of all the connection cleats
maximum moment obtained during the test:
105.0000 [kNm]
239
GENERAL OBSERVATIONS
COMMENTS
TEST NO PAGE
103.003
7
DEFINITION OF CURVE COORDINATES

definition of M
M is evaluated at the beam-to-column interface
definition of phi
connection:
phi corresponds to the deformability of the

connection elements (e.g. endplate, cleats, bolts)
and to the load-intoduction deformability in the
column web
shear:
sheared column web panel
total rotation: phi corresponds to the deformability of the
whole joint (connection + shear)
initial tangent stiffness:
unloading stiffness:
secant tangent stiffness:
corresponding value of phi:
240
TEST NO PAGE
103.025
1
CONNECTION
TYPE
SINGLE WEB AND UPPER AND LOWER FLANGE

CLEATED CONNECTION
AUTHOR(S)
TEST CENTRE
Schleich, J.B.; Grardy, J.C.; Jaspart, J.P

University of Lige
SOURCE OF
DATA
Semi-rigid action in steel frame Structures (CECA - No

7210 - SA/507)
SPECIMEN DETAILS
TEST DATE
20.3.89
MEMBERS SIZES AND REFERENCES
TESTARRANGEMENT
:
TEST
:
MATERIAL STANDARD :
SPECIMEN REFERENCE:
COLUMN ORIENTATION:
BEAM:
cruciform
in plane
Eurocode 3
3 6X3C6
strong axis
I PE 360
Fe 360
COLUMN:
HE 200 B
Fe 360
COLUMN WEB STIFFENING

: no
COLUMN FLANGE STIFFENING : no
CONNECTION:
NUMBER AND SIZE OF BOLTS : 4 M20, 12 M24
GRADE OF BOLTS
: 8.8 (DI N 931)
WASHERS
: yes
BOLT TIGHTENING PROCEDURE:
hand controlled tightening
cleat 150*90*13, St 37-2

cleat 150*90*13, St 37-2
DRAWING OF THE CONNECTION AND THE CONNECTED MEMBER
f. I .
1 "*
'1
59
40
'
1 I ""1
1_J
I I
70
39
OO
>-
19
i i -
11
99 4 O
38-
IF E 3
HE 200 B
BEAMS FRAMING INTO THE COLUMN FLANGE
COMPOSITE BEAM: yes
- plain concrete
- normalweight concrete
DATA SHEET printed with SERICON Vers. 1.0, 21.11.91
241
TEST NO PAGE
103.025
2
ACTUAL CHARACTERISTICS OF THE JOINT ELEMENT
COLUMN AND
column
beam_2
beam_l
202.8
359.6
( 200.0) ( 360.0)
170.0
201.9
( 200.0) ( 170.0)
14.2
15.0) (
12.0
12.7)
8.6
9.0) (
7.8
8.0)
18.0) (
18.0)
tf
tw
BEAM(S)
beam_3
tflwI"
*4
beam_4
3H>
units: [mm]
measured average values of h, b, t and r
(-) nominal values
ENDPLATE, CLEATS, T-STUBS, WELDS, BOLTS, HOLES,
cleat 1
cleat 2
flanges
where web
151.0
151.0
( 150.0) ( 150.0)
cleat 3
91.2
90.0) (
91.2
90.0)
12.8
13.0) (
12.8
13.0)
ri
12.5) (
12.5)
r2
6.5) (
6.5)
b
t
/H
units: [mm]
measured average values of a, b, t and r
(-) nominal values
BOLTS
type 1:
type 2:
4 M 20, 8.8 (DI N 931)

bolts with washers
hand controlled tightening (15 kgm)
1 mm
12 M 24, 8.8 (DI N 931)

bolts with washers
hand controlled tightening (22 kgm
1 mm
bolts M 20 used in the web cleat

bolts M 24 used in the flange cleat
242
rt
TEST NO PAGE
103.025
3
CONCRETE SLAB, STEEL REINFORCEMENT, CONNECTORS

1200.0
200
120.0
90.0
200
30.0
50
100
<- REBAR 1
<- REBAR 1
110
<- REBAR 2
<- REBAR 2
.1
97.0
23.0
.1.
110
units: [mm]
MECHANICAL PROPERTIES OF SLAB COMPONENTS
mechanical characteristics of reinforcement bars and connectors
Material
fy
[mm]
[mm2 ]
BE 40
14.0 152.84
REBAR 1
CONNECTOR KCO type KKB 22x100 mm
(22)
fu
[N/mm*]
[N/mm*]
[N/mm*]
210000
526.0
611.5
(210000) ( 350.0) ( 450.0)
mechanical characteristics of concrete slab

Material
SLAB
fck
[N/mm a ]
[N/mm a ]
15.0
Test procedure for fck: cube 157*157*157 mm3

(-) nominal values
243
MECHANICAL CHARACTERISTICS OF BEAMS, COLUMN

AND CONNECTING STEEL ELEMENTS
grade
E
[N/mm*]
COLUMN
flange
Fe 360
web
BEAM 1
Fe 3 60
flange
web
CLEAT 1
St 37-2
long leg
short leg
CLEAT 2
St 37-2
long leg
short leg
(-) nominal values
244
fy
[N/mm 2 ]
est
TEST NO PAGE
103.025
4
E st
[N/mm 2 ]
fu
[N/mm 2 ]
274.0
(210000) ( 235.0)
276.0
(210000) ( 235.0)
419.0
( 360.0)
415.0
( 360.0)
279.5
(210000) ( 235.0)
289.5
(210000) ( 235.0)
418.0
( 360.0)
418.0
( 360.0)
(210000)
(210000)
(210000)
(210000)
272.5
408.0
240.5
392.0
272.5
408.0
240.5
392.0
EXPERIMENTAL TESTING ARRANGEMENT
TEST NO PAGE
103.025
5
<
X '///////////A y//, y//////////A
2417.0
o
o
o
OJ
2417.0
"
units: [mm]
INSTRUMENTATION
TYPE
LOCATION
DEFLECTION MEASUREMENT
ROTATION MEASUREMENT
LOAD MEASUREMENT
Remark: for more datails see SOURCE OF DATA (see page 1)
245
LOAD BEHAVIOUR CHARACTERISTICS
EXPERIMENTAL M-* CURVES
CONNECTION
M [kNm]
400
300--
200
100--
FAILURE MODE
buckling of the column web in the compression zone
maximum moment obtained during the test:
246
358.0000 [kNm]
TEST NO PAGE
103.025
6
GENERAL OBSERVATIONS
COMMENTS
TEST NO PAGE
103.025
7
DEFINITION OF CURVE COORDINATES

definition of M
M is evaluated at the beam-to-column interface
definition of phi
connection:

connection elements (e.g. endplate, cleats, bolts)
and to the load-intoduction deformability in the
column web
CONCRETE RESISTANCE
- nominal 'fck' characterizes the strength after 28 days;
- measured 'fck' characterizes the strength on the day at
which the test is performed ( 15 days).
- bad concrete !
initial tangent stiffness:
unloading stiffness:
secant tangent stiffness:
corresponding value of phi
247
STRUCTURAL STEEL RESEARCH REPORTS

established by
RPS DEPARTEMENT / ARBED RECHERCHES
[101] Grardy J.C. .Schleich J.B.; Elasto Plastic Behaviour of Steel Frames with Semi-Rigid
Connections / NORDIC STEEL COLLOQUIUM on Research and Development within The
Field of steel Construction; Odense, Denmark, 9-11 September 1991, RPS Report No 101/91.
[102] Grardy J.C., Schleich J.B.;Semi-Rigid Action in Steel Frames Structures / CEC agreement
No 7210-SA / 507 ; Draft of Final Report, November 1991, RPS Report No 102/91.
[103] Ppin R.,Schleich J.B.; Seismic Resistance of Composite Structures, SRCS / CEC agreement
No 7210-SA / 506 ; Draft of Final Report, November 1991, RPS Report No 103/91.
[104] Chantrain Ph.,Schleich J.B.; Interaction Diagrams between Axial Load N and Bending
Moment M for Columns submitted to Buckling / CEC agreement No 7210-SA / 510 ; Draft of
Final Report, November 1991, RPS Report No 104/91.
[105] Schaumann P., Steffen A.; Verbundbrcken auf Basis von Walztrgern, Versuch Nr. 1
Einstegiger Verbundtrger / HRA, Bochum, Juli 1990, HRA Bericht A 89199, RPS Report No
105/90.
[106] Schaumann P., Steffen A.; Verbundbrcken auf Basis von Walztrgern, Versuch Nr. 2
Realistischer Verbundbrckentrger / HRA, Bochum, November 1991, HRA Bericht
A 89199-2, RPS Report No 106/91.
[107] Bruis A., Wang J.P. ; Composite Bridges with Hot Rolled Beams in High Strength Steel
Fe E 460 , and Spans up to 50 m / Service Ponts et Charpentes, Universit de Lige; Lige,
November 1991, RPS Report No 107/91.
[108] Schleich J.B., Witry A.; Acier HLE pour Ponts Mixtes Portes Moyennes de 20 50 m /
Journe Sidrurgique ATS 1991; Paris , 4 et 5 dcembre 1991, RPS Report No 108/91.
[109] Schaumann P, Steffen A.; Verbundbrcken auf Basis von Walztrgern, Versuch Nr. 5
Haupttrgerstoss mit Stahlbetonauflagerquertrger / HRA, Bochum, Januar 1992, HRA
Bericht A 90232-A, RPS Report No 109/92.
[110] Schaumann P, Schleich J.B., Kulka H., Tilmanns H.; Verbundbrcken unter Verwendung
von Walztrgern / Zusammenstellung der Vortrge anlsslich des Seminars
"Verbundbrckentag" am 12.09.90 an der Ruhruniversitt Bochum, RPS Report No 110/92.
[Ill] Schaumann P., Steffen A.; Verbundbrcken auf Basis von Walztrgern, Versuche Nr. 3 u. 4
Haupttrgerstoss mit geschraubten Steglaschen / HRA, Bochum 1992, HRA Bericht 90232-B,
RPS Report No 111/92.
[112] Schleich J.B., Witry A.; Neues Konzept fr einfache Verbundbrcken mit Spannweiten von
20 bis 50 m / DC. Leipziger Metallbau-Kolloquium; Leipzig, 27. Mrz 1992, RPS Report
No 112/92.
[113] Bergmann R., Kindmann R.; Auswertung der Versuche zum Tragverhalten von
Verbundprofilen mit ausbetonierten Kammern; Verbundsttzen / Ruhruniversitt
Bochum, Bericht No 9201, Februar 1992, RPS Report No 113/92.
[114] Bergmann R., Kindmann R.; Auswertung der Versuche zum Tragverhalten von
Verbund profilen mit ausbetonierten Kammern; Verbundtrger / Ruhruniversitt
Bochum, Bericht No 9202, Mrz 1992, RPS Report No 114/92.
[115] Schleich J.B., Wippel H., Witry A.; Untersuchungen an stegparallel versteiften
Rahmenknoten, ausgefhrt aus dickflanschigen hochfesten Walzprofilen . Entwurf
hochbelasteter Vierendeeltrger im Rahmen des Neubaus des Zentrums fr Kunst und
Medientechnologie ( ZKM ), Karlsruhe / RPS Report No 115/92.
[116] Chantrain Ph., Becker A., Schleich J.B.; Behaviour of HISTAR hot-rolled profiles in the
steel construction - Tests / RPS Report No 116/91.
249
it
For up-to-date information on

European Community research
consult
CORDIS
The Community Research
and Development
Information Service
C
CORDIS is an on-line service set up under the VALUE programme to give quick and easy
access to information on European Community research programmes.
The CORDIS service is at present offered free-of-charge by the European Commission Host
Organisation (ECHO). A menu-based interface makes CORDIS simple to use even if you are
not familiar with on-line information services. For experienced users, the standard Common
Command Language (CCL) method of extracting data is also available.
CORDIS comprises eight databases:
RTD-News: short announcements of Calls for Proposals, publications and events in the
R&D field
RTD-Programmes: details of all EC programmes in R&D and related areas
RTD-Projects: containing 14,000 entries on individual activities within the programmes
RTD-Publications: bibliographic details and summaries of more than 50,000 scientific
and technical publications arising from EC activities
RTD-Results: provides valuable leads and hot tips on prototypes ready for industrial
exploitation and areas of research ripe for collaboration
RTD-Comdocuments: details of Commission communications to the Council of Ministers
and the European Parliament on research topics
RTD-Acronyms: explains the thousands of acronyms and abbreviations current in the
Community research area
RTD-Partners: helps bring organisations and research centres togetherfor collaboration
on project proposals, exploitation of results, or marketing agreements.
For more information and CORDIS registration forms, contact
ECHO Customer Service
CORDIS Operations
BP 2373
L-1023 Luxembourg
Tel.: (+352) 34 98 11 Fax: (+352) 34 98 12 34
If you are already an ECHO user, please indicate your customer number.
European Communities - Commission

EUR 14427 - Properties and service performance - Semi-rigid action in
steel frame structures
J.-C. Gerardy, J.-B. Schleich
Luxembourg: Office for Official Publications of the European Communities
1992 - XIX, 249 pp., num. fig., tab. - 21.0x29.7 cm
Technical steel research series
ISBN 92-826-2926-0
Price (excluding VAT) in Luxembourg: ECU 27
This research has shown that the cost of a frame structure could be reduced
by taking into account the semi-rigid behaviour of the joints between the
beams and the columns.
In the first part of this research, 18 steel joints and 38 composite joints
have been tested at the laboratory of the University of Lige in Belgium.
These tests have permitted to determine all the components of deformability
of a semi-rigid joint with cleated connection. All the tests are stored in a
databank named SERICON (SEmi-Rlgid CONnection) developed at the University of Aachen (Germany) and distributed with this final report.
On the basis of these test results, a mathematical model of the behaviour
of these joints has been developed. This model is able to calculate the
moment-rotation curve of a composite joint with cleated connection. The
program of this model is distributed with this final report.
In the numerical part of this research, a new finite element taking into
account the actual behaviour of the joint (connection deformability and
shear deformability) has been introduced in the non-linear finite elements
program FINELG.
This new element has been used in the framework of a parametric study
of braced and unbraced structures with the aim of finding in which conditions
the joint deformability may be concentrated in a single flexural spring acting
at this end of the beams.
The philosophy of simplified methods to design steel structures with semirigid joints has been presented for braced and unbraced frames. These
developments present a starting point to a new practical document for
designers.
Lastly, a cost comparison between semi-rigid and traditional design shows
the potential benefits of the semi-rigid design of building frames.
Venta y suscripciones Salg og abonnement * Verkau f u nd Abonnement

Sales and subscriptions Vente et abonnements Vendita e abbonamenti
Verkoop en abonnementen Venda e assinaturas
BELGIQUE / BELGIE
FRANCE
SUOMI
TRKIYE
Moniteur belge /
Belgisch Staatsblad
Rue de Louvain 42 / Leuvenseweg 42
B-1000 Bruxelles / B-1000 Brussel
Tl. (02) 512 00 26
Fax (02) 511 01 84
Journal officiel
Service des publications
des Communauts europennes
26, rue D esaix
F-75727 Paris Cedex 15
Tl. ( 1 ) 4 0 58 75 00
Fax (1) 40 58 77 00
Akateeminen Kirjakauppa
Keskuskatu 1
PO Box 128
SF-00101 Helsinki
Tel. (0) 121 41
Fax (0) 121 44 41
Pres Gazete Kitap Dergi

Pazarlama Dagitim Ticaret ve sanavi
A
Narlibahe Sokak N. 15
Istanbui-Cagaloglu
Tel. (1) 520 92 96 - 528 55 66
Fax 520 64 57
Telex 23822 D SVO-TR
Autres distributeurs /
Overige verkooppunten
Librairie europenne/
Europese boekhandel
Rue de la Loi 244/Wetstraat 244
Tl. (02) 231 04 35
Fax (02) 735 08 60
Jean De Lannoy
Avenue du Roi 202 /Koningslaan 202
Tl. (02) 538 51 69
Tlex 63220 UNBOOK B
Fax (02) 538 08 41
Document delivery:
Credoc
Rue de la Montagne 34 / Bergstraat 34
Bte 11 / Bus 11
Tl. (02) 511 69 41
Fax (02) 513 31 95
DAN MARK
J . H. Schultz Information A / S
Herstedvang 10-12
DK-2620 Albertslund
Tlf. (45) 43 63 23 00
Fax (Sales) (45) 43 63 19 69
Fax (Management) (45) 43 63 19 49
NORGE
IRELAND
Government
4-5 Harcourt
Dublin 2
Tel. (1)61 31
Fax(1) 78 06
Supplies Agency
Road
11
45
ITALIA
Licosa SpA
Via D uca di Calabria, 1/1
Casella postale 552
1-50125 Firenze
Tel. (055)64 54 15
Fax 64 12 57
Telex 570466 LICOSA I
GRAND-DUCH D E LUXEMBOURG
Messageries Paul Kraus
1 1 , rue Christophe Plantin
L-2339 Luxembourg
Tl. 499 88 88
Tlex 2515
Fax 499 88 84 44
DEUTSCHLAND
Bundesanzeiger Verlag
Breite Strae
Postfach 10 80 06
D-W-5000 Kln 1
Tel. (02 21)20 29-0

Telex ANZEIGER BONN 8 882 595
Fax 2 02 92 78
Narvesen information center

Bertrand Narvesens vei 2
PO Box 6125 Etterstad
N-0602 Oslo 6
Tel. (2) 57 33 00
Telex 79668 NIC N
Fax (2) 68 19 01
BTJ
Tryck Traktorwgen 13
S-222 60 Lund
Tel. (046) 18 00 00
Fax (046) 18 01 25
SCHWEIZ / SUISSE / SVIZZERA
OSEC
Stampfenbachstrae 85
CH-8035 Zrich
Tel. (01) 365 54 49
Fax ( 0 1 ) 3 6 5 54 11
CESKOSLOVENSKO
NIS
Havelkova 22
13000 Praha 3
Tel. (02) 235 84 46
Fax 42-2-264775
NEDERLAND
MAGYARORSZAG
S D U Overheidsinformatie
Exteme Fondsen
Postbus 20014
2500 EA 's-Gravenhage
Tel. (070) 37 89 911
Fax (070) 34 75 778
EuroInfoService
Pf. 1271
H-1464 Budapest
Tel./Fax(1) 111 60 61/111 62 16
POLSKA
Business Foundation
GREECE/
PORTUGAL
G.C. Eleftheroudakis SA
International Bookstore
Nikis Street 4
GR-10563 Athens
Tel. (01) 322 63 23
Telex 219410 ELEF
Fax 323 98 21
Imprensa Nacional
Casa da Moeda, EP
Rua D . Francisco Manuel d e Melo. 5
P-1092 Lisboa Codex
Tel. (01) 69 34 14
ESPANA
Boletn Oficial del Estado
Trafalgar, 29
E-28071 Madrid
Tel. (91)538 22 95
Fax (91) 538 23 49
MundiPrensa Libros, SA
Castell. 37
E-28001 Madrid
Tel. (91) 431 33 99 (Libros)
431 32 22 (Suscripciones)
435 36 37 (D ireccin)
Tlex 49370-MPLI-E
Fax (91)575 39 98
Sucursal:
Librera Internacional A E D O S
Consejo de Ciento, 391
E-08009 Barcelona
Tel. (93) 488 34 92
Fax (93) 487 76 59
Llibrera de la Generalitt
de Catalunya
Rambla deis Estudis, 118 (Palau Moja)
E-08002 Barcelona
Tel. (93) 302 68 35
302 64 62
Fax (93) 302 12 99
Distribuidora de Livros
Bertrand, Ld."
Grupo B ertrand, SA
Rua das Terras dos Vales, 4-A
Apartado 37
P-2700 Amadora Codex
Tel. (01) 49 59 050
Telex 15798 BERD IS
Fax 49 60 255
ul. Krucza 38/42

00-512 Warszawa
Tel. (22) 21 99 93, 628-28-82
International Fax&Phone
(0-39) 12-00-77
STERREICH
Manz'sche Verlags
und Universittsbuchhandlung
Kohlmarkt 16
A-1014 Wien
Tel. (0222) 531 61-0
Telex 112 500 B O X A
Fax (0222) 531 61-39
410066"426709
ROY International
PO Box 13056
41 Mishmar Hayarden Street
Tel Aviv 61130
Tel. 3 496 108
Fax 3 544 60 39
CANADA
Renouf Publishing C o . Ltd
Mail orders Head Office:
1294 Algoma Road
Ottawa, Ontario K1B 3W8
Tel. (613) 741 43 33
Fax (613) 741 54 39
Telex 0534783
Ottawa Store:
61 Sparks Street
Tel. (613)238 89 85
Toronto Store:
211 Yonge Street
Tel. (416)363 31 71
UNITED STATES OF AMERICA

UNIPUB
4611-F Assembly D rive
Lanham, MD 20706-4391
Tel. Toll Free (800) 274 4888
Fax (301)459 0056
AUSTRALIA
Hunter Publications
58A Gipps Street
Collingwood
Victoria 3066
Tel. (3)417 5361
Fax (3)419 71S4
ROUMANIE
JAPAN
Euromedia
65, Strada D ionisie Lupu
70184 Bucuresti
Tel./Fax 0 12 96 46
Kinokuniya Company Ltd

17-7 Shmjuku 3-Chome
Shinjuku-ku
Tokyo 160-91
Tel. (03) 3439-0121
BULGARIE
Journal Department
PO Box 55 Chitse
D.J.B.
59. bd Vitocha
1000 Sofia
Tel./Fax 2 810158
UNITED KINGD OM
H M S O B ooks (Agency section)
H M S O Publications Centre
51 Nine Elms Lane
London SW8 5D R
Tel. (071) 873 9090
Fax 873 8463
Telex 29 71 138
ISRAEL
Tokyo 156
Tel. (03)3439-0124
RUSSIA
SINGAPORE
C C E C (Centre for Cooperation with

the European Communities)
9. Prospekt 60-let Oktyabria
117312 M o s c o w
Tel. 095 135 52 87
Fax 095 420 21 44
Legal Library Services Ltd

STK Agency
Robinson Road
PO Box 1817
Singapore 9036
CYPRUS
AUTRES PAYS
OTHER COUNTRIES
ANDERE LAND ER
Cyprus C h a m b e r of C o m m e r c e and
Industry
Chamber Building
38 Grivas D higenis Ave
3 D eligiorgis Street
PO Box 1455
Nicosia
Tel. (2)449500/462312
Fax (2) 458630
Office des publications officielles

des Communauts europennes
2, rue Mercier
L-2985 Luxembourg
Tl. 499 28 1
Tlex PUBOF LU 1324 b
Fax 48 85 73/48 68 17
UI
NOTICE TO THE READER

All scientific and technical reports published by the Commission of the European Communities
are announced in the monthly periodical 'euro abstracts'. For subscription (1 year: ECU 110)
please write to the address below.
o
a
t
IO
H
Price (excluding VAT) in Luxembourg: ECU 27

t* * *t
* ON *
OFFICE FOR OFFICIAL PUBLICATIONS

OF THE EUROPEAN COMMUNITIES
L-2985 Luxembourg
ISBN l E f l S b a i E b D

Cgna14427enc 001

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

Cgna14427enc 001

Uploaded by

Copyright:

Available Formats

eu* w$

Commission of the European Communities

technical steel researcl

Semi-rigid action in steel frame structures

Properties and service performance

Semi-rigid action in steel frame structures

Service recherches et promotion technique structures

PARI EUROP. Biblioth.

Cataloguing data can be found at the end of this publication

Luxembourg: Office for Official Publications of the European Communities, 1992

SEMI - RIGID ACTION IN STEEL

SCHEDULED COMPLETION DATE

Semi-Rjgjd Action in Steel F r a m e s Structures.

Action Semi-Rigide dans les Structures en Acier.

Halhsteife Auschlsse im Stahlbau.

2.1. Types of tested beam-to-column joints.

3.1. Behaviour of column web panels.

4.1. Description of a new finite element.

5. Simplified methods for the design of steel structures with

5.1. Simplifying assumptions.

6. Benefit of semi-rigid design.

6.1. Unbraced frame.

7.1. Frames with steel connections

TABLE DES MATIERES

Comportement du panneau d'me de la colonne.

Description d'un nouvel lment fini.

5. Mthodes simplifies pour la conception des structures en acier 81

6. Avantages du calcul semi-rigide.

Structures compose d'assemblages en acier.

Extended end plate

sections for SLS

Connection (M- <f>)

b. Cruciform steel joint with

c. Cruciform steel joint with

d. Cruciform composite joints

e. Cruciform composite joints

Figure 1.3 - Types of joints tested in laboratory.

- Chapter 4 on "Numerical developments" deals with the introduction of a new finite

a. "Cruciform" or "interior" joint

1. Composite joints with 2 cleats per

2. Composite joints with 3 cleats

4. Steel joints with 3 cleats

5. Composite joints with 2 cleats

6. Composite joints with 3 cleats

8. Steel joints with 3 cleats

Figure 2.3. - Types of beam-to-column joints tested in laboratory.

1.1. Parameters influencing the detailed choice of the connections.

Lowerflangecleat + web cleat

1 test per BD 1 test per BD

Both flange cleats + web cleat

1 test per BD 1 test per BD

Table 2.1. - General view of the experimental program.

Thickness of 150 x 190

a. Steel "T" test arrangement.

b. Composite cruciform test arrangement.

Thickness of 150 x 190

c. Steel cruciform test arrangement.

Figure 2.6. - Concrete slab longitudinal reinforcement.

Shear connectors for one beam

Figure 2.7. - Shear connectors and transversal reinforcement arrangement

Figure 2.8. - Deformability of a strong axis steel beam-to-column joint.

Figure 2.10. - Subassemblages and frames

"T testing arrangement and location of