Professional Documents
Culture Documents
*w *
_s
Commission of the European Communities
J.-B. Schleich
Ingnieur principal
Contract No 7210-SA/507
(1.7.1987-31.12.1990)
Final report
Directorate-General
Science, Research and Development
1992
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
TITLE OF RESEARCH
AGREEMENT N
EXECUTIVE COMMITTEE
Light Structures F6
BEGINNING OF RESEARCH
01.07.1987
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-
VII
IX
-XI
CONTENTS.
Summary.
1. Introduction.
2. Experimental investigations.
9
12
16
16
17
19
23
23
25
25
25
26
29
29
31
37
37
39
40
3. Theoretical developments.
41
41
42
47
53
57
58
4. Numerical developments.
69
69
73
79
XIII
61
61
61
62
63
63
65
65
67
81
81
81
82
83
84
85
85
85
85
87
88
90
91
94
94
. 95
96
97
99
101
103
105
106
107
107
108
109
109
110
1 io
Ill
Ill
7. Conclusions.
113
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
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 .
2.7.2.1. Description gnrale des rsultats.
2.7.2.2. Influence des paramtres.
2.7.3. Tests cruciformes tout acier.
2.7.3.1. Description gnrale des rsultats.
2.7.3.2. Influence des paramtres.
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.
41
42
47
53
57
4. Dveloppements numriques.
69
4.1.
4.2.
4.3.
69
73
79
XVII
58
61
61
61
62
63
63
65
65
67
85
87
88
90
91
94
94
95
96
97
99
101
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
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
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
. connection feformability
<t>=eb-ec
. shear deformability of
column web panel
Y=ec-ef
. joint deformability
e=eb-9f
/-=^
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
1.6
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.
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
'
h-
^
-n
=T
3. Steel joints with 2 cleats
per connection (type B).
CRUCIFORM JOINTS.
10
2.2
u
/
T
S
7. Steel joints with 2 cleats
per connection (type C).
EXTERIOR JOINTS.
2.3
11
12
2.4
number of tests/
type of connection
total number
of tests
tip
fo
.B
t{
25
\h
1 test per BD
1 test per BD
and per AT
31
56
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
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
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
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
2.7
15
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
16
2.8
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
/ / / / / / / / / / / / / / / / / / / / / / /
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=<|>
- 9
-joint
<>
t = e+y
or <>
t = F(l, 2and3)
ed = e -e s -eu
- column flange
6^(8-6)^
- cleat included bolt elongation
6a = (j- ecf
ea = ( 9 - 7 ) / d
2.13
21
Instrumentation.
Defomability curves.
^
L b =2.3m
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)
2xHb
22
2.14
2.15
23
fari.
canrcNENTS
JOINT
I
CONNECTION
VES SH EA
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
24
2.16
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.
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.
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
i'' i '7
|2 -C 0
|4.00
6.00
8.00
2.19
27
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
-J
1
I
1
1
1
1
J -
2-00
4.00
6.00
8-00
AD
24T3C2
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
|
,0
-2
36T3C2
|10.00
AO
Figure 2.15. - Column web shear deformability for joints with IPE 360 beams.
28
2.20
2.21
29
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
(*)
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
kNm
BARS OF
18 MM
150X90X10 CLEATS"
, -'
TiO
150X00X10 CLF* T S
BARS OF 10 MM
BARS OF M MM
BARS O c 18 MM
2.25
33
M - W CURVE
uuuLJuuJ
(c Nr.
10
3000.
2 CLEATS -
13 MM
BARS OF 10 tin
3 CLEA'S -
-1
10 MM
-f'
pio
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
34
2.26
M - W CURVE
LUJCLLUJ
k NU
IO
joon
3
CLEATS
13 MM
BARS OF
CLEATS
CLEATS
1 MM
13 MM
10 MM
'
f'Pio
2000
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
2.27
35
M - W CURVE
fujJiJujJ
kNm
3 CLEATS
10 MM
CLEMS
10 MM
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
36
2.28
Mmax/Mp
(%)
59
74
57
58
36
45
Test
number
24x2c4
24x2c8
30x2c8
30x2c9
36x2c4
36x2c8
Mmax/Mp
(%)
12
16
17
17
13
-
2.29
37
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
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
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
de
42
Mci + Mc2
Qbi + Qb2 de
(3-D
3.2
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
P
V
1)M
])M
3.3
43
,4
On
1^=^"
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*-
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
3.7
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
4.30
:
!
1
:
!1
r ] " " ^t
mm m
3. SS
;
!
^A^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
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.
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 .
lttaui*hmrdnAg|
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
Mbppl -
Mbe
<Ie
>st
<pu
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
3.13
53
I!?I_?fi?_:_J9INT
ti
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 v^i
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
54
3.14
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
MbPHI Curves:
Theoretical
0.12
.a
t.M
Experimental
"
56
3.16
Mh-PHI Curvas:
Theoretical
Experimental
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
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..
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.
3.19
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
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
60
3.20
(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)
m.u
74.43
.*"""
4.M
- **'
1/
a.M
Mb-PHI Run,.
1/
1
Bl-linear
il
11
I.M
(.M
.M
.M
Experimental
.M
62
3.22
Mb-PHI Curva :
Bilinear
Experimental
Experimental
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
^AO
r.
HI CURVE
JOINT
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
3.25
65
r-y
concrete
:
*", lower rebars
upper flange cleat
web cleat
lowerflangecleat
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
(4.2.)
(4.3.)
Formula (4.3) constitutes only a rough approximation of the actual shear loading
giving by (4.2.).
I
beam
structure /
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
&
!- k
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.
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
\ 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
B1
FC
8.863
450
28
BL.1
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
PI
..i'JTTTT}<
Willi//
PI
. P2
TTTTV
m/inii
111//1111
1i/ii 11 ir
76
4.8
,P1
,P1
, PI
,P1
PI
PI
PI
,PI
4.9
77
"-
-L
78
4.10
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
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
y.<t>(10" 2 rad)
10'
-v
MM
3.0
6.0
1-
9.0
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
I
I
O
/
a \
beam element
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
Mp
Me
f\
EI/L
82
5.2
5.3
83
Predesign
Iterative
procedure
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.
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
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
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
13.5 <MV
18.7 < My
6.9 <MV
li
7.3 < M V ;
2.0
88
5.8
XF
.F=100kN;
TTTTTT
E
v
II
J3
b=5m
,M = * *
Mpb
<|>c
<j)p
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
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.).
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
_ . GLOB. BI-LIN.
,s54
/*
ft
'.
MOOELLIN&S
SO.O
MM
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
12.0
1 HH
5.13
93
^^^
Mv
i r \Ksf
/
/
/
\ K
\
\
94
5.14
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.)
5.15
95
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)
98
5.18
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
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
"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
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
104
6.2
FL j 11 n 111111 u 11 n n i j j j -L
J.M 11I11JLU ,1111 m i i u i
F3
3.75m
3.75m
3.75m
6.3
105
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)
106
6.4
L 120x120x12 ; 5x5x1/2
Bolts M22 ; 7/8"
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
6.5
107
3.5m
X
3.5m
3.5m
r
3.5m
108
6.6
-v
Symmetric
\
AT
6.0 m
200 mm
L 120x120x12 ; 5x5x1/2
6 bolts M24 ; 1"
Figure 6.8. - Beam with pinned connections.
6.7
109
o)--
<- S
O) - r
L120x120x12
L 5x5x1/2
Bolts M24 ; 1"
6.0 m
Solution
Beams
Pinned
Rigid
PE30
IPE 300
Semi-rigid
IPE 270
Connection
Price
110
6.8
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
114
7.2
Connection^)
oo
Diffenciation
yes
no
oncntrton^
^
rx,
O
z
2
52
pi
4ont
Connection
Panel
ii*
B<|iction f iasviotir
frcdictioB of behaviour
ii ] *
IHI*ilHI1IIIIHalllllt1M1litiiiniiinmn
*f
5pe(3f. Characteristics
aottH^o 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
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
116
7.4
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
120
Ref
-2
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
P,
q
r
i-VJ t I t I i
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
_F% rr>n
>
>
9!w>
4;
i__i
\ 1 1 1\
\\W\\
Ijftjf*
>
VTTTf
Mi
9SW?
0.02 E
i
f y = 240Mpa
E = 210.10 3 Mpa
:y
10 Y
Annex 2.4
131
H/B<1.2
/K^-0.5
H/B> 1.2 - . 0 . 3
T.
i
H
B-
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 ).
/////
132
Annex 2.5
M b (kNm)
78.0
58.5
39.0
19.5
IJ
2.6
1.9
5.2
65
9. .0(10 'tad)
1
7.6
78.0
Connection
58.5
39.0
-i
1.3
2.6
1
33
1
5.2
1
6.5
e.r.<t>no"2rad)
*7.8
Annex 2.6
133
M b (kNm)
Ui.O
36.0-
9.J,MO Sad)
3.0
S.O
9.0
12.0
15.0
18.0
K4.0"
108.0
72.0
8. I > t 1 0 " 7 r d )
t
3.0
6.0
9J0
12.0
1S.0
18.0
134
Annex 2.7
jMjikmm
134.7511751-
ns.soisoi
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
. f . ( ' r i )
2.0
;JJ
1.0
IJO
Annex 2.8
135
J.15UC1-
32.70(721-
: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.
. f.H0*'UI
2.0
Ut
t.0
1.0
136
Annex 2.9
r
beam
jiiiimmj
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
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
braced
A
unbraced
braced
B
C
D
unbraced
braced
unbraced
braced
unbraced
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).
140
Annex 2.13
3.3
3.1
1.1
1.3
V-
-!
HOOELISATIONS
_ _ EXACTE
.-B.. GLOBALISEE HH
44.3
30.
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
Annex 2.14
141
MODELISATIONS
_ , - EXACTE h 3
. , - . GLOBALISEE
Hf
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
142
Annex 2.15
NCC6LISAT IONS
EXACTE H 3
GLOBALtSEE HM
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
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
HOOeUSATIONS
_ EXACTE
. . a - . GLOBALISEE
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
MODELISATIONS
EXACTE
. . . . . QLOB ALISEE
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
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
Annex 3.2
151
24x2c1
24x2c1g
24x2c2
24x2c2g
24x2c3
24x2c3g
24x2c5
24x2c5g
24x2c6
24x2c6g
24x2c7
24x2c7g
24x3c1
24x3c1g
24Y3R2
lt)
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
Column : HE200B
Beam :|IPEa6
~~'
: 12.45
Help )
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
^e
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
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
name
"fig"
general data
V
column
name of the column
height of the profile
width of the profile
thickness of the web
thickness of the web
radius of the fillets
caracteristic data
16
fyCw
E
Est
est
2
2
2
2
2
fucf
fyCf
E
Est
est
2
2
2
2
2
name
h
b
tw
tf
r
y
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
4
4
4
4
4
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
5
5
5
5
5
7
7
7
7
7
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
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
tethamin
tethamax
step
precision
5 0
15
14 j
14 I
15
15
15
data for
calculation
Table 1.
data
name
general data
V
column
name
h
b
tt
tf
r
Lx=lA
HEB200
200
200
9
15
18
51.62
fucw
fycw
E
416
276
210000
column web
168
Annex 3.19
Est
est
4200
0
fticf
419
274
210000
4200
0
IPE360
360
170
8
12.7
18
15
418
289.5
210000
4200
0
E
Est
est
416.5
279.5
210000
4200
0
Est
ni
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
-30
1200
120
12.45
150x90x13
800
37
4
37
16
3
34
220000
353
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
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
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.
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
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
Annex 3.23
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/y^Est,epsst
Annex 3.24
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
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 x^icorjcw
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
Ayb
A stb
A ub
Ayb=*L
eyb=|
Astb =(st - yb)*L + Ayb
Fub = nb*A*fu
Aub^Est + ^ ^ ^ 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,
diagram F - A of the spring :
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
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
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;
*-4s *wa
wcf=2*na
z-
tf3*wcf
"^ *
0*7*
kl+2*k4
Zrf*acf2+2*Za*aa+^j
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
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
*~
"
F v c f =min (F rfl ; F ^ ; F ^ )
n' a = min (na ; 1.25*01',,)
mpa = 0.25*s2*fy
^ma = 1
Frii = 2*Bt*
m pa *b ma +2*B;*n' a
(m a + n' a )
hrf2_
mpa*b|na*( 4 * n ' a - e )
Id3
tf
rd3*n' a +
pa* b n*
2^nV^
^Nbd
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
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
( cleat)
E
if ( y * 0 ) and ( F va < F
if ( F v l i
,up.)
~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
_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
(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
2*K*m c *n , c -e*(K-l)*(m c +n , c )
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
tw
fuwb
a
ef
b
sd
s
wd
rl
bd
1
Mt
fu
Ab
fy
y2
E
mb
Est
fub
est
gap
Fyc
Annex 3.34
183
without slip
Fugt
Fygt
, uge
gap
Fyge
gap
ugt
Fuge
V\MAr
F
i yC
=fv*S
iy
184
lJ
Annex 3.35
fy * 1
yc-~~
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
if t, 20 mm
then ia =
+3*^-2
^T + 3 *- 2
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_
ugc=(FugCK,Fygc)^*gap
tw-y2
M-
=> if d 1.5
then the web cleat hole istoo near from the side!
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
(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
Annex 3.38
187
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
A yc A sto
, uc
diagram 1
188
Annex 3.39
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
_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
StC=LC*(Est-eyc)+Ayc
Fuc = fu*S
Auc
Lc
lEst
Egt
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 %
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
(h-2*(r + t f ))
C *F
_ v "cu "cppl )
cu
V
^cst
"cuy
^p Mv **cw
cj
P F
__ r cj rcppl
'i=
cst -
Kcj
ce
+ A
"
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 %. = ^
then^ c =2.5
if tf20mm
then ^
tf30mm
*^j
- 2
=^
+ 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 ;
(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
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
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
: / .
it
>/
\i
'
Annex 4.2
233
234
Annex 4.3
CONNECTION
TYPE
TEST NO PAGE
103.003
1
AUTHOR(S)
TEST CENTRE
TEST DATE
October 1988
SOURCE OF
DATA
SPECIMEN DETAILS
TESTARRANGEMENT
:
TEST
:
MATERIAL STANDARD :
SPECIMEN REFERENCE:
COLUMN ORIENTATION:
BEAM:
IPE 300
Fe 360
COLUMN:
HE 200 B
Fe 360
CONNECTION:
cleat 150*90*10, St 37-2
cleat 150*90*10, St 37-2
IPE 300
HE 200 B
BEAMS FRAMING INTO THE COLUMN FLANGE
COMPOSITE BEAM: no
TEST NO PAGE
2
103.003
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]
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)
/H
units: [mm]
type 1:
1 mm
type 2:
1 mm
r.
-V
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
237
TEST NO PAGE
103.003
5
in
m
m
2239.0
units:
[mm]
INSTRUMENTATION
TYPE
LOCATION
DEFLECTION MEASUREMENT
ROTATION MEASUREMENT
LOAD MEASUREMENT
TEST NO PAGE
6
103.003
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
_^^i
^-^-^
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
105.0000 [kNm]
239
GENERAL OBSERVATIONS
COMMENTS
TEST NO PAGE
103.003
7
unloading stiffness:
240
TEST NO PAGE
103.025
1
CONNECTION
TYPE
AUTHOR(S)
TEST CENTRE
SOURCE OF
DATA
SPECIMEN DETAILS
TEST DATE
20.3.89
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
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
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
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:
1 mm
1 mm
rt
TEST NO PAGE
103.025
3
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
SLAB
fck
[N/mm a ]
[N/mm a ]
15.0
243
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
TEST NO PAGE
103.025
5
<
2417.0
o
o
o
OJ
2417.0
"
units: [mm]
INSTRUMENTATION
TYPE
LOCATION
DEFLECTION MEASUREMENT
ROTATION MEASUREMENT
LOAD MEASUREMENT
245
CONNECTION
M [kNm]
400
300--
200
100--
FAILURE MODE
buckling of the column web in the compression zone
246
358.0000 [kNm]
TEST NO PAGE
103.025
6
GENERAL OBSERVATIONS
COMMENTS
TEST NO PAGE
103.025
7
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 !
unloading stiffness:
247
it
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.
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.
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
Autres distributeurs /
Overige verkooppunten
Librairie europenne/
Europese boekhandel
Rue de la Loi 244/Wetstraat 244
B-1040 Bruxelles / B-1040 Brussel
Tl. (02) 231 04 35
Fax (02) 735 08 60
Jean De Lannoy
Avenue du Roi 202 /Koningslaan 202
B-1060 Bruxelles / B-1060 Brussel
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
B-1000 Bruxelles / B-1000 Brussel
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
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
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
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
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
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
UI
o
a
t
IO
H
ISBN l E f l S b a i E b D