Beruflich Dokumente
Kultur Dokumente
Gbor Szab
Abstract
Members with high slenderness are commonly used in steel structures. Those members load carrying
capacities are limited due to global instability. One way to reduce this handicap is through lateral
stabilization. Sheeting, which is commonly part of the structure and is primarily designed for exterior
load only, could be included in the member global stability check. This may lead to a more efficient
design of compression members.
In this thesis the positive effect of cassette wall cladding on the behaviour of structural members is
investigated. This is commonly known as stressed skin design. Experimental investigation and
numerical modelling were used to determine the favourable effect of the cladding made by cassette
profiles. A parametric study was carried out to extend the experimentally obtained results.
The experimental work consisted of two parts. In the first part, the cassette wall torsional stiffness is
determined. Altogether 17 specimens were tested with different sheeting thicknesses, fastener
arrangements and connected flange properties. As a result, an equation was worked out for the
calculation of the rotational stiffness of the cassette profile. This part of the thesis was used as the
input parameter for following work.
Before the second part of the experimental investigation, the numerical model was prepared and the
best arrangement of the model was worked out by ANSYS software package. The second type of
experimental investigation included beam-columns full-scale tests and considered the load-carrying
capacity and failure mode. The beam-columns were stabilized by cassette wall. Overall six specimens
were prepared and tested. The beam-column cross section used was IPE300 in steel grade S355. The
member's system lengths were 5330 mm and 5680 mm depending on the boundary condition. The
loading was applied by end cantilevers which ensured combined bending-compression loading. The
ratio between the bending moment and axial force was 1 m. Specimens were selected to include
different sheeting stiffness, sheeting position and boundary condition. The results demonstrated that
connected sheeting could fully stabilize the member against lateral-torsional buckling in cases when
the sheeting is connected at compressed flange. In other case (connected at tensioned flange) torsional
buckling failure occurred about imposed axis.
The above mentioned numerical model, which was used to predict the most suitable specimens, was
updated, improved and calibrated by results from experiments. A parametric study was carried out to
extend the experimentally obtained results with the help of numerical models. The parametric study
was executed for hot rolled IPE and HEA profiles for different lengths and more cassette walls and
bending moment distribution.
The main results of the thesis are wall stiffnesses which performed full or partial stabilization for
specific member. These stiffnesses depend on wall properties and were calculated according to ECCS
recommendation. Besides, the impact of rotational stiffness of sheeting was followed.
-1-
Gbor Szab
Pro ocelov konstrukce se bn pouvaj pruty s vysokou thlost. nosnost takovch prut je asto
omezena globln ztrtou stability. Jednm ze zpsob omezen tto nevhody je pn podepen.
Pl, kter je soust bn konstrukce a je navren primrn na penos vnjho zaten, me bt
zahrnut do posouzen globln stability prutu. Takov postup umouje zvit efektivitu nvrhu
tlaench a ohbanch thlch prut.
Dizertan prce vyetuje kladn inek kazetov stny, pipojen k ocelovmu prutu. Vyuv se
plov psoben stny. Pro analzu pznivho inku se pouilo experimentln vyetovn a
numerick modelovn. Byla t provedena parametrick studie pro rozen zskanch vsledk.
Experimentln st prce se skldala ze dvou st. V prvn se urovala rotan tuhost ppoje
kazetov stny na sloup. Bylo provedeno 17 zkouek s rznou tloutkou plechu kazety, uspodnm
roub a kou psnice pipojenho sloupu. Na zklad zkouek a numerickho modelovn byly
odvozeny rovnice pro uren rotan tuhosti ppoje kazetovho profilu k psnici sloupu. Zskan
vsledky slouily jako vstupn hodnoty pro dal vyetovn.
Druhou st experimentln prce pedchzelo numerick modelovn programem ANSYS pro zjitn
nejvhodnjho uspodn zkouky. Vyetoval se sloup skuten velikosti namhan ohybem a
tlakem, stabilizovan pipojenou kazetovou stnou. Zjiovala se jeho nosnost a zpsob poruen.
Bylo provedeno est zkouek. Byl pouit profil IPE300 z oceli jakosti S355. Dlka sloup byla 5330
mm a 5680 mm v zvislosti na okrajovch podmnkch. Zaten bylo aplikovno pomoc konzol pro
dosaen kombinovanho namhn tlakem a ohybem. Pomr osov sly k ohybovmu momentu byl 1
m. Vsledky potvrdily, e v ppad, kde je kazetov stna pipojen k tlaen psnici, me pln
stabilizovat prut proti ztrt stability z roviny ohybu. V ppad pipojen k taen psnici dochz
k ztrt stability s vnucenou osou oten.
Numerick model pouit k nvrhu experimentu byl vylepen, zkalibrovn podle vsledk
experiment a nsledn pouit pro parametrickou studii. Parametrick studie byla provedena metodou
GMNIA pro vlcovan profily ady IPE a HEA pro rzn dlky, jakosti oceli, tuhosti kazetov stny a
prbh ohybovho momentu po dlce prutu.
Z parametrick studie byly odvozeny potebn tuhosti kazetov stny pro plnou nebo stenou
stabilizaci prutu v zvislosti na ve uvedench vstupnch parametrech. Dle byly vypracovny
pomcky pro vpoet smykov i rotan tuhosti stny podle doporuen ECCS a vlastnch vsledk.
Byl t uren vliv rotan tuhosti ppoje stny k prutu na jeho nosnost.
-2-
Gbor Szab
Acknowledgements
This project was worked out in Steel and Timber Department at the Faculty of Civil Engineering in the
Czech Technical University in Prague during 2004 2008. I am grateful for all expert help of mine
supervisor Prof. Tomas Vrany.
I would like to acknowledge for the project financial support provided by the Czech Technical
University (internal grant IG VUT CTU0502311 - 2005), the Ministry of Education, Youth and
Sport of the Czech Republic (grant FRV 1820 - 2006) and the support of the research project of the
Czech Ministry of Education (grant no. 6840770003).
The experimental part of the project was conducted in the Experimental Centre of CTU. I am thankful
to all the technicians who contributed to the work.
I would like to thank Ing. Vitezslav Hapl with whom I carried out the experimental part of the project,
and for his advices during the work. Further thanks should also be given to all members of Steel and
Timber Department for they reconciliation, extremely to colleagues from room D1064.
I am also grateful to Dr. Martin Heywood of The Steel Construction Institute (Ascot, UK) who has
helped me with the language correction.
Finally, I will always be grateful to my wife Rose and my family for their support and encouragement
throughout my study.
-3-
Gbor Szab
-4-
Gbor Szab
Contents
Abstract ....................................................................................................................................................1
Acknowledgements ..................................................................................................................................3
Notation....................................................................................................................................................7
List of figures .........................................................................................................................................10
List of tables...........................................................................................................................................13
Chapter 1
Introduction ......................................................................................................................14
1.1.
Background...........................................................................................................................14
1.2.
Chapter 2
2.1.
2.1.1.
2.1.2.
2.1.3.
2.2.
2.3.
2.3.1.
Interaction.........................................................................................................................23
2.3.2.
2.3.3.
2.3.4.
Chapter 3
Chapter 4
Experimental study...........................................................................................................34
4.1.
Introduction ..........................................................................................................................34
4.2.
4.2.1.
Experiments overview......................................................................................................35
4.2.2.
4.3.
4.3.1.
Experiments overview......................................................................................................42
4.3.2.
Imperfection .....................................................................................................................47
4.3.3.
4.3.4.
Gbor Szab
4.4.
Additional tests..................................................................................................................... 55
4.4.1.
4.4.2.
Chapter 5
5.1.
Introduction .......................................................................................................................... 58
5.2.
5.2.1.
5.2.2.
5.3.
5.3.1.
5.4.
5.4.1.
Chapter 6
.................................................................................................... 68
6.1.
6.2.
6.2.1.
6.2.2.
Chapter 7
Conclusions...................................................................................................................... 91
References ............................................................................................................................................. 93
Appendix A Tables of measured data................................................................................................. 96
Appendix B Calculation example..................................................................................................... 103
-6-
Gbor Szab
Notation
b1, b2
distance between screw and cassette profile edge / column flange edge
Bo
Bu
ci,j
cA
cM
cP
Cexp
Ccalc
Cmy, Cmz
Factors
eL
eNy ,eNz
shift of the centroid of the effective area Aeff relative to the centre of gravity of the
gross cross section
es
modulus of elasticity
FL
Fp
Fs
fy
shear modulus
Gbor Szab
iy, iz, ip
gyration radius
I1
second moment of area of the wide flange of a cassette about the longitudinal axis
of the cross section per metre [mm4/mm] , see Fig.7.
Iy
Iz
Iz,G
second moment of area of the wide flange of a cassette about own centroid [mm4]
IT
warping constant
Kq
factor for considering the moment distribution and the type of restraint
bending moment
Mcr
Mpl,k
My,Ed, Mz,Ed
MPl,y,, MPl,z
Ncr
elastic critical force for the relevant buckling mode based on the gross cross
sectional properties, critical flexural buckling load
NEd
nf
np
ns
nsc
nsh
Pmax
member resistance
Rwithout_rot
Rwith_rot
S1
stiffness of the fasteners between the sheeting and the connected member
S2
S3
Sact
sp
Gbor Szab
ss
shear flexibility of the fastener in the cassette wall between adjacent cassettes
[mm/kN]
ssc
shear flexibility of the fastener between cassette and edge member [mm/kN]
Td
Tv
shear force in the diaphragm caused by the load in the serviceability limit state
TV
TV,L
shear flow in the ultimate limit state of fasteners between cassette and edge
member of wall
TV,Q
shear flow in the ultimate limit state of fasteners between cassette and framework.
TV,S
shear flow in the ultimate limit state of fasteners between adjacent cassettes
Vbuc
Weff,y,min, Weff,z,min minimum effective section modulus about y-y axis, z-z axis
*
fastener number factor between cassette and framework per sheet width
y, z
Factors
Poisson's ratio
LT
y, z
My, Mz
-9-
Gbor Szab
List of figures
Fig.1.
Cassette profile..................................................................................................................... 15
Fig.2.
Possible cassette wall constructions, a) direct connection, b) non-direct connection of the
covering panel........................................................................................................................................ 16
Fig.3.
Fig.4.
Fig.5.
Fig.6.
Idealization of interaction..................................................................................................... 25
Fig.7.
Fig.8.
Fig.9.
Assumed deflection of the element to stabilize with the resulting concentrated force at the
supports ..30
Fig.10.
Fig.11.
Fig.12.
Fig.13.
Fig.14.
Fig.15.
Fig.16.
Fig.17.
Fig.18.
Measured values................................................................................................................... 39
Fig.19.
Deformed specimen.............................................................................................................. 39
Fig.20.
Load deflection curves from experimental investigation, marked with respect of Tab. 3. .. 40
Fig.21.
Fig.22.
Specimen V2 ........................................................................................................................ 42
Fig.23.
Fig.24.
Fig.25.
Fig.26.
Fig.27.
Fig.28.
Gbor Szab
Fig.29.
Fig.30.
Fig.31.
Fig.32.
Fig.33.
Fig.34.
Fig.35.
Fig.36.
Fig.37.
Fig.38.
Fig.39.
Fig.40.
Fig.41.
Fig.42.
Fig.43.
Fig.44.
Fig.45.
Fig.46.
Fig.47.
Fig.48.
Force displacement curves for specimens KK, one screw (both sheet thickness 0,75mm).
..57
Fig.49. Force displacement curves for specimens KS, one screw (sheet thickness 0,75mm, steel
plate thickness 10mm). ..........................................................................................................................57
Fig.50.
Fig.51.
Fig.52.
Fig.53.
Fig.54.
Fig.55.
Fig.56.
Fig.57.
Fig.58.
Comparison of load-deflection diagrams from model and experiment for specimens V1, K1.
...64
Fig.59.
Comparison of load-deflection diagrams from model and experiment for specimens K2, V2.
...64
Fig.60.
Comparison of load-deflection diagrams from model and experiment for specimens K3, V3.
...65
-11-
Gbor Szab
Fig.61. Relative deflection of compressed and tensioned flange, comparison between model and
experiment, specimens K1, V1.............................................................................................................. 65
Fig.62. Relative deflection of compressed and tensioned flange, comparison between model and
experiment, specimens K2, V2.............................................................................................................. 66
Fig.63. Relative deflection of compressed and tensioned flange, comparison between model and
experiment, specimens K3, V3.............................................................................................................. 66
Fig.64.
Moment distribution............................................................................................................. 67
Fig.65.
Fig.66.
Fig.67.
Fig.68.
in MPa.
Numerical model of the connection loaded by positive bending moment, von Mises stress
..70
Fig.69. Ratio between load carrying capacity for rotationally supported and non-supported
members IPE boundary condition K, supported at tensioned or compressed flange ...................... 71
Fig.70. Ratio between load carrying capacity for rotationally supported and non-supported
members IPE boundary condition V, supported at tensioned or compressed flange ...................... 72
Fig.71. Ratio between load carrying capacity for rotationally supported and non-supported
members HEA boundary condition K, supported at tensioned or compressed flange.................... 72
Fig.72. Ratio between load carrying capacity for rotationally supported and non-supported
members HEA boundary condition V, supported at tensioned or compressed flange.................... 72
Fig.73. Ratio between load carrying capacity for rotationally supported and non-supported
members with length 6000 mm, impact of section torsional stiffness................................................... 73
Fig.74. Compressed flange lateral deflection for different wall properties; A three cases: without
support, cassette wall support, full support, B two cases: cassette wall support, full support (same
graph with larger scale) ......................................................................................................................... 74
Fig.75. Compressed flange lateral deflection for different wall stiffness, section IPE300
(tthickness, ns-number of screws, cl-cassette profile length)................................................................... 75
Fig.76. Compressed flange lateral deflection for different steel grade, sections: IPE140, L=4 m;
IPE300, L=8.4 m ................................................................................................................................... 76
Fig.77.
-12-
Gbor Szab
List of tables
Tab. 1
Factor K for considering the moment distribution and the type of restraint
Tab. 2
Tab. 3
41
Tab. 4
System lengths
43
Tab. 5
List of specimens
43
Tab. 6
50
Tab. 7
Steel properties
56
Tab. 8
Comparison table
65
Tab. 9
78
Tab. 10
79
Tab. 11
Tab. 12
81
Tab. 13
82
Tab. 14
Tab. 15
84
Tab. 16
85
Tab. 17
Tab. 18
87
Tab. 19
88
Tab. 20
-13-
28
Gbor Szab
Chapter 1
Introduction
1.1.
Background
Cassette walls are commonly used as cladding systems, which belong to the light weight facings.
Nowadays these types are widely used for framed structures. Among the main advantages of this
system is the simple assembly that arises from the low weight of the light gauge wall members. A
wide range and high quality of those members increase the demand for the cassette walls.
Load-bearing frames provide support for cassette walls. Wall members are connected to the outer
flange of the column. Cassette walls, which can provide load into own plane, contribute to the global
stability of the building. The shear stiffness of the whole wall consisting of cassette profiles and
trapezoidal panels depends on many variables. Support of column provided by cassette wall is
generally assumed as a continuous lateral elastic support.
1.2.
One of the progressive cladding systems which are nowadays used for the steel framed structures is
cassette wall. The main advantages of this system are [1]:
simple details,
The cassette profile contains one wide, two narrow flanges and two web elements, see Fig.1. Particular
elements of the cassette are commonly stiffened by longitudinal stiffeners because they are made from
light gauge steel. The members have high slenderness ratio and it causes susceptibility to local
buckling of the wall or flange, buckling due to shear forces or distortional buckling.
-14-
Gbor Szab
1.
cassette,
1.
cassette,
2.
trapezoidal panel,
2.
corrugated panel,
3.
thermal insulation,
3.
thermal insulation,
4.
4.
5.
5.
6.
6.
distance profile,
7.
self-tapping screws,
7.
distance profile
8.
8.
9.
framework,
9.
self-tapping screws,
10.
10.
-15-
Gbor Szab
11.
12.
framework,
-16-
Gbor Szab
Chapter 2
Literature review
2.1.
Gbor Szab
The stressed skin design of a cassette wall may be solved in a similar way to a diaphragm composed of
trapezoidal profiles. This idea has been proofed by Baehre and was based on the experimental
investigation. During the experimental investigation the behaviour of empty and with insulation filled
cassette members was established [4].
Similar way was used by Davies in his work [1] which was focused on the characterization of
diaphragm created by cassette profiles. The main differences between the behaviour of cassettes and
trapezoidal profiles used in diaphragm are as follows:
the distortion of the cross section caused by shear forces can be neglected in cassettes in
consideration of the stiffness of the section. Therefore the deformation of the cassette wall is
based on the deformation of the joints,
the load-carrying capacity is principally limited by local buckling of the wide flange against
other modes of failure, common for the trapezoidal sheeting,
the edge members of the wall are often not present (horizontal members at the bottom and the
top of the wall). Therefore the upper and lower cassette profile must be checked for the
additional compression and tension forces.
The most important parts of the cassette wall for the stressed skin action are joints which have critical
effect for the stiffness and for the load-carrying capacity of the wall. The basic possible modes of joint
failure are as follows:
failure between cassette and frameworks (at least three fasteners should be arranged, [11]),
failure between cassette and edge members due to shear forces (if edge members are present).
Gbor Szab
the distance between fasteners (es) in the adjacent webs is no more than 300 mm,
the distance between fasteners (a1) in narrow flange is no more than 1000 mm,
the fasteners in the webs must be as close as possible to wide flange (eu30 mm),
the other geometrical limits are defined as following: 30BO60 mm; 60H200 mm;
300Bu600 mm; 0.75t1.5 mm, Ia/bu10 mm4/mm, see Fig.4.
(2.1)
8.43EL 4 9
I1t
Bu2
(2.2)
Vbuc
The cassette wall shear flexibility is equal to the sum of individual flexibilities of the diaphragm (2.3).
The formula contains these sub flexibilities: shear flexibility of the diaphragm by shear deformation of
the sheet c1,2 (2.4), shear flexibility of the diaphragm by the deformation of the fasteners between
cassette and framework c2,1 (2.5), shear flexibility of the diaphragm by the deformation of the fasteners
between adjacent cassettes c2,2 (2.6) and shear flexibility of the diaphragm by the deformation of the
fasteners between cassettes and edge members c2,3 (2.7).
c c1,2 c2,1 c2,2 c2,3
(2.3)
(2.4)
c2,1 2 Bs p p / L2
(2.5)
(2.6)
(2.7)
-19-
Gbor Szab
Individual shear flexibility for screws should be determined experimentally. The safe values for
screws can be also found in tables [11].
The difference of Baehres approach [4] in comparison with ECCS [11] is only in the formula for the
shear flexibility of the fasteners between adjacent cassettes c2,2. The following published equation by
Baehre could be used to obtain c2,2:
c2,2 ss nsh 1 / ns
(2.8)
The stiffness of the diaphragm according to ECCS [11] may be determined from the following
simplified equation:
(2.9)
L
S act S act
In formula (2.9) Sact is the shear stiffness of the diaphragm which could be determined as a
multiplication of the shear stiffness per unit length (2.10) and the total length of the diaphragm.
Sact
aLBu
es B Bu
(2.10)
The serviceability limit state requirement for diaphragm composed from cassettes is following:
S act
T
V
375 L
(2.11)
i nf
(2.12)
1
i
(2.13)
the sheeting should first be designed for its primary purpose in bending according to EN 19931-3 [13] (either by calculation or by testing). It should then be checked that the maximum
shear stress due to diaphragm action does not exceed 25% of the design yield (normal) stress.
it may be assumed in design that transverse load on a panel of sheeting will not affect its
strength or flexibility as a shear diaphragm.
diaphragm forces in the roof or floor planes should be transmitted to the foundations by means
of braced frames, stressed skin diaphragms, or other methods of sway resistance.
structural connections of adequate strength and stiffness should be used to transmit diaphragm
forces to the main steel framework.
-20-
Gbor Szab
diaphragms should be provided with edge members. These members, and their connections,
should be sufficient to carry the flange forces arising from diaphragm action.
diaphragms should not be used to resist permanent external loads (except for lateral support to
beams and the dead load from light weight construction) but should be predominantly
restricted to resisting
o
loads applied through the cladding, such as wind loads and snow loads, and
seismic forces and other similar (in magnitude and frequency) transient loads,
crane forces,
Note: the force induced in any fastener or group of fasteners by horizontal surge or
braking effects from overhead cranes should not exceed 30% of the capacity of the
fastenings.
stressed skin diaphragms should be treated as structural components and should not be
removed without consideration of the effect on the stability of the building. Such
consideration should not invalidate planned removal of areas of sheeting for maintenance
purposes, provided the remaining areas are adequate as a diaphragm or temporary bracing is
provided during maintenance.
the calculations, drawings and contract documents should draw attention to the fact that the
building incorporates stressed skin diaphragms, subject to National rules.
openings totalling more than 3% of the area in each shear panel should not be permitted.
Openings of less than this amount may be permitted without special calculation provided the
total number of fasteners in each seam with openings is not less than that in a seam without
openings.
stressed skin diaphragms should be designed predominantly for short-term imposed loads,
unless long term phenomena such as creep are taken into account.
stressed skin buildings in which the frames have not been designed to carry the full unfactored
load without collapse should be braced during erection. Buildings which utilize the roof or
floors as stressed skin diaphragms should be erected so that the roof and floors are sheeted
before the walls are cladded.
the structural effects of building modifications on stressed skin buildings should be checked.
Changes in use or occupancy which might affect the original design assumptions should be
noted in the contract documents and notified to the appropriate authority.
The shear stiffness of the diaphragm could be determined in accordance with [19] according to
equation (2.10) if the following condition is valid for the second moment of area of wide flange about
its major axis:
B T
I z ,G 7.75 10 4 u v
t E
(2.14)
(2.15)
TV , L FL / eL
(2.16)
-21-
Gbor Szab
(2.17)
The value TV represents shear flow in the ultimate limit state of fasteners. Equations (2.15), (2.16) and
(2.17) represent the individual possible failures of fasteners.
Davies, who is one of the dominant personalities in field of cladding structures, pointed out some
weakness of the design procedures in [3]:
one of the basic assumption for the cassettes is that the wide flange need not be stiffened. In
the context of cassette wall construction with wide flanges exposed externally this is
architecturally undesirable. The design procedure for local shear buckling of the thin wide
flange is equally applicable to both stiffened and unstiffened flanges.
the most worrying aspect is that Eurocode does not require any formal design check in the
capacity of the fasteners.
2.2.
Members under axial compression and bending develop specific load-carrying behaviour in the
different ranges of slenderness. At very low slenderness the cross-sectional resistance dominates,
described by the well-known interaction formulae for elastic or plastic limit states:
M y , Ed N Ed eNy M z , Ed N Ed eNz
N Ed
1
Aeff f y / M Weff , y ,min f y / M Weff , z ,min f y / M
(2.18)
With increasing slenderness a second order effect appears, which is significantly influenced by both
geometrical imperfection and residual stresses. In the high slenderness range, member buckling is
dominated by elastic behaviour.
Depending on the emphasis given to the different ranges, different concepts of interaction formulae
were proposed in the past:
The present approach of Eurocode [12] was based on the linear-additive form. In this method the
effect of the axial force and the bending moments are linearly summed and the non-linear effects are
accounted for by specific interaction factors. The preference for this concept results from its userfriendliness, since it allows the evaluation of the individual effects of the axial force and bending
moments.
The code [12] includes two different formats of the interaction formulae called Method 1 [23] and
Method 2 [24]. The main difference between them is the kind of presentation of the different structural
effects, either by specific coefficients in Method 1 or by one compact interaction factor in Method 2.
This makes Method 1 more adaptable to identifying and accounting for the structural effects, while
Method 2 is mainly focused on the direct design of standard cases [25].
-22-
Gbor Szab
N Ed
y N pl , Rd
k
y LT
LT
N Ed
z N pl , Rd
k
z * LT
LT
Cmy M y , Ed
N
1 Ed
N
cr , y
C yy ,mod M pl , y , Rd
Cmz M z , Ed
N Ed
1 N C yz ,mod M pl , z , Rd
cr , z
Cmy M y , Ed
N
1 Ed
N cr , y
Czy ,mod M pl , y , Rd
Cmz M z , Ed
N Ed
1 N Czz ,mod M pl , z , Rd
cr
,
z
(2.19)
(2.20)
(2.21)
M y , Ed
C M
N Ed
k LT
k z mz z , Ed 1
y N pl , Rd
LT M pl , y , Rd
M pl , z , Rd
(2.22)
In [12] the unified formulation of above mentioned methods is present (2.23), (2.24). The difference
between both methods is handled by coefficients kyy, kzz, kyz, kzy. After substitution the according to
chosen method the corresponding formula is achieved.
M y , Ed M y , Ed
M
M z , Ed
N Ed
k yy
k yz z , Ed
1
y N Rk
LT M y , Rk
M z , Rk
M1
M1
M1
M y , Ed M y , Ed
M
M z , Ed
N Ed
k zy
k zz z , Ed
1
z N Rk
LT M y , Rk
M z , Rk
M1
2.3.
(2.23)
(2.24)
M1
M1
2.3.1. Interaction
Members of the secondary structure which are connected to the framework may have a significant
beneficial effect on the behaviour of the primary frame members. In particular, in the right conditions,
secondary members may either partially or fully stabilize the connected primary member against
lateral deflection.
The secondary structure may be of several types:
-23-
Gbor Szab
bracing,
covering systems,
Gbor Szab
The interaction between various covering systems and beams was the theme of several dissertation
thesis. The thesis of Petrusson was focused on the interaction between sandwich panels and steel
column [31], [32]. The thesis was worked out in Lule University of Technology. During the
investigation three types of experiments were carried out:
determination of the rotational stiffness provided with sandwich panels, where the panels
could act against the stabilized member rotational deflection (on Fig.6 labelled as C). The
obtained rotational stiffness for one pair of panels with two screws was 1kNm/rad. The
stiffness was linearly increasing according to the number of screws.
sandwich panel shear stiffness determination in own plane (on Fig.6 labelled as S).
full scale experiments with connected columns and sandwich panels. The experiments were
focused on the determination of sandwich panel stabilization effect.
The used members were HEA120 and IPE200. Two modes of loading were used during the tests: i)
axial compression, ii) compression and bending. The results of the experiments pointed out that
HEA120 was fully stabilized by the panels under both modes of loadings and lateral torsional buckling
was not appeared. The failure was caused by buckling. When IPE200 was tested the failure mode
depended on the type of loading. For axial compression load the member failed by buckling. Under
axial force and bending moment the lateral torsional buckling occurred. The conclusion of the work
was that in common cases for the investigated members the sandwich panels provide sufficient elastic
support and provide full lateral stabilization.
if shear stiffness S is sufficient for full stabilization of supported member then torsional
deformation does not occur and the rotational stiffness does not activate.
if the shear stiffness is not enough for full stabilization and act only as a partial support then
the torsional deformation appears and the rotational stiffness is activated and contributes on
stabilization.
-25-
Gbor Szab
In this arrangement of elastic supports the stabilization may be full. In the other case when the
supporting members are connected at tensioned flange the lateral-torsional buckling with imposed axis
of rotation in tensioned flange occurs. In this case the stabilization effect is softer.
2.3.4.1.
The design procedures which were published by Lindner are based on the experimental investigation.
The work was arisen from Fischers work [34] and from Lindners investigation. Sufficient stiffness S
(Fig.6) for full stabilization of beam was published in [26]. Special evaluation was made out for IPE
sections with a depth more than 200 mm.
The limiting shear stiffness for the full stabilization is following:
2
2 70
S EI 2 GIT EI z 2 h 2 .
L
4L
(2.25)
The same equation (2.25) may be found in codes [12] and [33].
Stiffness S for cassette profiles can be found out from experimental investigation, from the earlier
project of Rybn [17] or from ECCS recommendation [11].
Even if the shear stiffness S does not reach the limiting value of (2.25), a remarkable positive effect on
the ultimate lateral torsional buckling load is given in [25]. In such cases this effect can be taken into
account by calculating the elastic lateral torsional buckling moment Mcr under the assumption of the
given value for C (explained below) and S.
The next value which was investigated by Lindner is the rotational stiffness C (Fig.6) provided by
connected sheeting. The rotational stiffness acts against the rotational movement of stabilized beam.
The effect of stabilization is included at torsional stiffness of the member IT. The equation designed by
Lindner is following:
IT * IT
CL2G .
(2.26)
Stiffness C presents the rotational stiffness whereby the sheeting acts on the member. This stiffness is
defined as:
1
1
1
1 .
C c M c A c P
(2.27)
The equation for obtaining the rotational stiffness is based on the idea that each component included is
independent and could be defined separately. The individual stiffnesses are showed on Fig.7 and have
following meanings:
-26-
Gbor Szab
EI ,
a
(2.28)
where: k is the coefficient including the span of the sheeting. (4 for multi span, 2 for single or
double span),
a is the span of the sheeting.
The value of cA is commonly determined by experimental investigation. Equation (2.29) was
presented in [26] by Lindner for the fasteners between trapezoidal sheeting and hot rolled sections and
may be also found in [13].
b
c A c A
100
(2.29)
For cold formed sections value c A may be obtained according to the equation determined by Vran
[35].
The equation (2.30) may be used for sections I and U [33]. Constant c1 depends on the shape of cross
section (0.5 for I sections for random load; 0.5 for U sections for gravity load; 2 for U sections for
uplift).
c P
E
1
2
h
4 1 c b
1 3
tw3
tf
(2.30)
Alternative equation included in code [12], to check the sufficient torsional stiffness provided by
sheeting is following:
-27-
Gbor Szab
M pl2 , k
EI z
K K
(2.31)
where K is factor for considering the moment distribution, see Tab. 1, and K is equal to 1.00 for
plastic analysis and equal to 0.35 for elastic analysis.
Tab. 1. Factor K for considering the moment distribution and the type of restraint of
compression flange
Tab. 2. Factor K for considering the moment distribution and type of restraint according to
[25]
In [25] another table could be found (Tab. 2), which contains detailed value for K defined for each
buckling curve and also for other moment distribution.
-28-
Gbor Szab
The hot rolled profiles with depth lesser than 200 mm are always sufficient supported against lateral
torsional buckling. For sections with higher depth Lindner presents the limiting ratio for the
trapezoidal panel shear stiffness S and rotational stiffness C whereby should be achieved for full
stabilization of supported members, see Fig.8. The graph is operative for single span beams with end
uniformly distributed loads and end moments.
2.3.4.2.
The elastic support calculation according to Heil is based on the following idea. The point of
application of the uniformly distributed load is in the same plane as the point of the connected
secondary structure [28]. The stabilized member is sufficiently stabilized if equation (2.32) is valid.
2
2 3 M pl , y
3 2 EI z
S
3 2 2 3 LT2 h 2 3 L2
2
2 2
2
1 1 3 c ,
2
3 h
(2.32)
2 EI GI t L2 .
(2.33)
EI z
In this equation the stiffness C of the sheeting against the rotational movement of connected member
is not considered. The rotational stiffness factor c2 with respect of the sheeting rotational stiffness C
could be obtained from equation (2.34).
c2
2 EI GI t L2 L4C / 2
EI z
-29-
(2.34)
Gbor Szab
We can consider that the member is fully stabilized against lateral torsional buckling if LT 0.4 .
When this limit is substituted into (2.32) the obtained equation is following:
S 10.18
2.3.4.3.
M pl , y
h
4.31
EI z
c2
1
1
1.86
L2
h 2
(2.35)
In ECCS [11] the possibility is given to use shear panels as stability bracing to prevent flexural,
flexural torsional and lateral torsional buckling or combination of these behaviour. In this document
the procedure is given for checking the stabilized state of columns and beams made from doubly
symmetric I sections. The determination of the minimum strength and stiffness of the shear panel
according to the recommendation ECCS was mentioned above, see section (2.1.3).
The formulae in [11] are based on a sinusoidal form of lateral deflection, which leads to a concentrated
reaction force at the support (see Fig.9). The concentrated force can be regarded as an internal force in
the diaphragm provided the fasteners between the sheeting and the element to stabilize can sustain this
concentrated force on length of 1/8 of the span of the element. If these fasteners are unable to sustain
this concentrated force, then that force should be introduced into the edge members via the fasteners
between the diaphragm and the edge members.
Fig.9. Assumed deflection of the element to stabilize with the resulting concentrated force at
the supports
If the unstiffened flange of the beam is in tension the beam is fully stiffened when the shear stiffness
of the diaphragm S fulfils the following requirement:
S act S y
-30-
fy A
2
(2.36)
Gbor Szab
Sact is the shear stiffness of the diaphragm and Sy is the required shear stiffness of a diaphragm for full
stabilization. If the requirement for full stabilization is not fulfilled (Sact < Sy), the critical forces Mcr
and Ncr are calculated from following formulae:
M cr M
(2.37)
N cr N
(2.38)
where is the eigenvalue being the lowest value of 1 and 2. If both eigenvalues are negative, no
stability problem exists for the given stress resultants M and N.
1,2
k1
k1
h
1
WzWw S
2k2
2
2
k
k
ip
2
2
(2.39)
where S is the minimum of Sact or Sy. Coefficients included in (2.39) should be obtained from
following formulae:
Sh
k1 N Wz Ww M 2
i p
k2 N 2 M 2
1
i p2
EI 2
Wz z2 S
L
Ww
1
i p2
2
ECw 2
h
GI
S
T
2
2
L
i p2 i y2 iz2
Cw
1
I z h2
4
(2.40)
(2.41)
(2.42)
(2.43)
(2.44)
(2.45)
Cw is the warping constant for doubly symmetric I section. M is constant bending moment about y
axis, N is constant centric normal force to be introduced with the actual sign, for definition see Fig.10.
-31-
Gbor Szab
S o
1 L
(2.46)
where eo is the initial lateral bow imperfection of the stiffened flange according to the Eurocode [12].
The maximum shear force Td shall be smaller than the design resistance of the fasteners Pmax as was
mentioned above (2.1). Furthermore one of the following conditions shall be fulfilled:
the shear force Td shall be smaller than FpL/8p where Fp is the design strength per fastener
between the sheeting and the stabilized element, p is the pitch of the fasteners between the
sheeting and the stabilized element.
the fastenings of the sheeting to the edge elements shall be accounted for Td.
-32-
Gbor Szab
Chapter 3
Thesis objectives
The main objective of the PhD. thesis is to find out the effect of the sheeting from cassette profiles to
the load carrying capacity of the connected slender beam-column.
The thesis was divided into experimental, numerical and theoretical parts.
The objects of the experimental part were focused on:
to determine the stabilization effect of the cassette wall to the load carrying capacity of hotrolled beam-column.
The theoretical analysis is based on results of experimental and numerical work and the objectives are
also the main ones of the thesis:
to derive formulae for determining the rotational stiffness of the cassette profiles,
to work out the design tables for minimal cassette wall stiffness for full and partial
stabilization of connected beam-column.
-33-
Gbor Szab
Chapter 4
Experimental study
4.1.
Introduction
Experimental investigation belongs to the main part of the thesis. Experiments give the primary results
data, which is the base of the investigation.
During the thesis altogether three types of experiments were carried out:
4.2.
Rotational stiffness
The aim of these experiments were to determine the rotational stiffness of the connection between
cassette profile and hot-rolled beam-column from I profile about its x-x axis, see Fig.11.
-34-
Gbor Szab
The experiments were carried out because useful data were not found for cassette profiles. A similar
rotational stiffness value was worked out by Ptrusson for sandwich panels but these results couldnt
be used due to dissimilar behaviour between those two sheeting types [31].
The obtained values served as input parameter for following work. The experiments were focused on
the rotational stiffness influenced by local behaviour of cassette in region of connection. From the
statical points of view the investigated stiffness couldnt be determined by calculation method
available in standards or other literature. The experimental investigation was necessary due to the
complex behaviour of profile in the area of local deformations.
The experiments were not made to derive a general equation for cassette profiles but as input
parameters for following work. This was a reason why only specific specimen with similar properties
was used.
Trapezoidal sheeting was neglected during the experiments because it did not influence the
investigated stiffness.
2x2,
-35-
Gbor Szab
2x3.
300 mm,
250 mm,
150 mm.
Fig.14. Used steel plates replacing column flange, dimensions, position of fasteners
Screw arrangement with respect of the flange centriod, see Fig.14:
symmetrically to the flange axis [a], (for plate width 150 mm),
symmetrically to the flange axis near axis [b], (for plate width 250 mm and 300 mm),
symmetrically to the flange axis near edge [c], (for plate width 250 mm and 300 mm),
un-symmetrically to the flange axis [d], (only for plate width 250 mm and 300 mm, used non
symmetrical fasteners arrangement about flange axis).
The last variable, which had significant effect and was followed in experiments, was the cassette
profile. Two types of cassette profile were used with two different thicknesses; see Fig.15, [37]:
-36-
Gbor Szab
Cassette profiles:
The nominal yield strength of the light gauge steel used for cassettes was 320 MPa.
-37-
Gbor Szab
Altogether 17 specimens were made from combinations of above listed parameters. The list of the
specimens is presented in Tab. 3. (Arrangement of the screws which is in the table labelled by letters is
noted in Screw arrangement capitulation above in squared bracket). The specimens were assembled
by hand. The distance between the cassette profiles was 10 mm. Screws were tightened by torque
wrench. The assembly was carried out horizontally and lifted up onto the laboratory column. The
loading was applied by weights. A cantilever, which was welded to flange, served for the application
of load. A loading platform was hung from the end of the cantilever where the weights were put. The
gauge was clamped on the cassette. The deflection of the cantilever was measured by manual
deflectometer. The measuring equipment precision was 0.05 mm.
-38-
Gbor Szab
load.
The bending moment was calculated from load and force arm according to Fig.18.
-39-
Gbor Szab
(4.1)
and the rotation was determined from the cantilever movement and gauge position
p .
arctan
(4.2)
Following figure shows the values as rotational deflection and applied moment.
Fig.20. Load deflection curves from experimental investigation, marked with respect of Tab. 3.
The rotational stiffness is defined as:
C
M Nm
/ cassette width
rad
-40-
(4.3)
Gbor Szab
4.3.
The main goal of the project is to determine the stabilization effect of cassette wall to the load carrying
capacity of the beam column. The effect of the sheeting would have been determined directly from
numerical model but experimental study was necessary to prove the correctness of the model. These
facts lead to provide full scale experiments of beam-column stabilized by cassette wall.
This test arrangement was chosen so that lateral torsional buckling would be governing for the load
capacity. According to the theoretical assumption the connected sheeting should obstructed this failure
mode.
The stabilization effect depends on several parameters. Lets see which these are and how could
influence the results:
Position of the connected sheeting with respect of the member flanges has significant
influence to the stabilization effect. Cassette profile connected to the tensioned flange allows
free lateral movement of compressed flange but partially prevents the cross-section against
torsional deformation. The column resistance is less increased in this arrangement. In the other
case when the wall is connected at the compressed flange the effect of the wall is stronger.
The wall provides lateral restraint of the compressed flange. During the experiments both
cases were tested.
Shear stiffness of the cassette wall in its own plane depends on cassette profile properties and
amount of screws between adjacent cassettes. The stiffness of the wall commonly is not a
weak point of the construction. Defined maximum pitch of 300 mm according to [11] presents
sufficient intensity. The length of used cassette profiles was smaller as in normal case.
Therefore higher density of screws between adjacent cassettes was chosen. The screw density
was constant during the investigation.
Number of connection between the flange and the cassette profile influenced both lateral and
rotational stiffness of the connection. Connection with two fasteners was chosen for all
specimens.
Static scheme and the boundary conditions play an important part in the beam-column
behaviour. These conditions were chosen so that the lateral torsional buckling occurred.
-41-
Gbor Szab
Cantilever was used to achieve combined (M+N) internal forces in the member. Two types of
the beam-column boundary condition were chosen with different out of plane behaviour.
The specimen comprises more members than only beam-column, cantilevers, cassette profiles
and corrugated sheeting. An additional frame was made which supported cassette profiles in
the desired position. These necessary additional members affected the behaviour of whole
specimen and could increase the load carrying capacity of the beam-column. This effect was
taken into account during the numerical modelling and analysing of results.
Fig.22. Specimen V2
The parts of the specimen were as follows:
beam-column,
two cantilevers,
Gbor Szab
Altogether six specimens were produced, each of them with different properties. The variables were:
boundary conditions. Two types of boundary conditions were used. One was the typical frame
corner joint with partial warping and out-of-plane bending restraint which provided a
significant prevention against beam-column distortional deflection (V), see Fig.23. The
second one was a fork connection allowed free warping of beam-column end sections and free
out-of-plane rotation (K), see Fig.24. The connection consisted from tension contact
represented by pin and from compression contact, which was made by cubes where one of
them was modified to sphere. Rotation about x-x axis was restrained for both types. In plane
static scheme for both types of specimens was the same. For both types of specimen different
cantilevers and beam-column were produced. The system length of the beam-column and
cantilever for each type is noted in Tab. 4.
amount and position of cassette profiles. The cassette profiles were put on the specimen in
three different arrangements:
o
The following Tab. 5 contains all specimens with boundary conditions and cassette arrangements.
Tab. 4. System lengths
The specimens were produced in DT vhybkarna a mostarna, Prostjov - and built up in CTU
Experimental centre. During the experiments the same cantilevers were used for all specimens, only
the tested beam column was replaced. For the beam-column section IPE300 and for the cantilevers
HEA320 were used. Steel grade 355 MPa was used. The joints are shown in Fig.23 and Fig.24. The
specimens were symmetrical about vertical plane.
-43-
Gbor Szab
-44-
Gbor Szab
Gbor Szab
cassette profiles (delivered by Kovove profily s.r.o., type B K145/600, 0.75mm, S320 GD),
trapezoidal panels (delivered by Kovove profily s.r.o., type TR 35/207, 0.75mm, S320 GD),
-46-
Gbor Szab
screws:
o
Loading was applied by hydraulic jack. The acting point was on one of the cantilevers in 1 m distance
from centroid of beam-column. The loading was controlled by displacement. As first step 30% load of
theoretical load carrying capacity was carried out due to screws haul before main loading was applied.
The tests were conducted until collapse.
Fig.29. The loading point with hydraulic jack and the support
During the experiments three types of sensors were used:
potentiometer sensor,
inductive sensor,
strain gauge.
The position of the sensors and measured data are included in following chapters.
4.3.2. Imperfection
Global imperfection of beam-column has considerable influence to the buckling behaviour of the
member. This was the reason why the imperfection of the beam-column was examined.
The Czech standard CSN 73 2611 [38] (adequate to EN 1090-2) specifies the maximum allowable
production imperfection. To follow the specimen behaviour the imperfection in minor axis orientation
was important. According to the standard the members are sorted to three classes A, B and C. Each
class has got different limit:
A) 0,0015L, max. 15 mm,
B) 0,002L, max. 20mm,
C) 0,003L, max. 30mm.
-47-
Gbor Szab
The initial imperfection was measured. Measuring was carried out by level instruments for each beamcolumn, see Fig.30 to Fig.35.
-48-
Gbor Szab
-49-
Gbor Szab
-50-
Gbor Szab
U4, U5, U6 lateral deformation of first three cassette profiles counted from cantilever,
U9 longitudinal deformation of cantilevers in position of hydraulic jack; this value was used
for loading handles,
R1, R2, R3 lateral relative movement between cassettes and beam-column top flange.
-51-
Gbor Szab
T5, T6 top flange, both edges, 370 mm from end plate of beam-column from bearing side,
T7, T8 bottom flange, both edges, 370 mm from end plate of beam-column from bearing
side.
Gbor Szab
T7, T8, T9, T10 web of the beam-column, midspan, vertically arranged in one line.
V1, K1 lateral torsional buckling about imposed axis, see Fig.40, Fig.41. Cassette wall
connected at tensioned flange could not prevent beam-column torsion. Cassette profiles failed
locally in place of cassette-flange screws.
V2, K2 plastic moment resistance of beam-column reached. Cassette wall fully stabilized
the beam column with in plane stiffness, no relevant lateral movements at collapse, see Fig.42.
-53-
Gbor Szab
V3, K3 lateral torsional buckling. One sided cassette wall with decreased plane stiffness
buckled in connection with beam-column lateral movement, see Fig.43, Fig.44. Cassette
profiles failed by two ways, locally and globally.
Gbor Szab
Tables including all the measured values are presented in Appendix A. Fig.45 shows the loaddeflection (measured by gauge U3 see Fig.37 and Fig.39) curves for both types of specimens for each
combination of wall arrangements.
4.4.
Additional tests
In this section two types of tests are noted. The first type was focused on determination of the used
material properties of beam-column. The second one was about the stiffness of the screw connections.
-55-
Gbor Szab
-56-
Gbor Szab
Fig.48. Force displacement curves for specimens KK, one screw (both sheet thickness
0,75mm).
Fig.49. Force displacement curves for specimens KS, one screw (sheet thickness 0,75mm,
steel plate thickness 10mm).
-57-
Gbor Szab
Chapter 5
Numerical modelling
5.1.
Introduction
Numerical modelling comprises two parts. At first during the experiments preparation, the models of
beam-column with cassette wall were developed. The models served to test and tune the specimens.
Secondarily the real specimens were modelled. They confirmed the experimentally obtained results.
The numerical model was used for the parametric study additionally. The numerical modelling was
carried out by ANSYS 11.0 software package.
5.2.
Pre-modelling of specimens
The numerical modelling was first used for predicting the results of experiments and basically
facilitated the experiments design. The goal was to find out the profile of sufficient slenderness to
illustrate the effect of connected sheeting stiffness.
The model contains from following main parts:
beam-column,
cantilevers,
sheeting.
The cantilevers served for the loading and for generation of combined loading on beam-column. The
sheeting was connected at either compressed or tensioned flange of beam-column. The connections
between the cantilevers and the beam-column were made according to mentioned details marked as V
and K.
Gbor Szab
applications. The element is defined by four nodes (I, J, K, L) in rectangular position. The length to
width ratio may be up to 3. The geometry definition of the element is presented on Fig.50.
two parts of cassette profiles due to proper wall model, S3, S4; see Fig.52.
-59-
Gbor Szab
-60-
Gbor Szab
The section geometry and the length of the beam-column governed the slenderness of the
member which determined the type of failure. During the numerical modelling such kind of
section was searched which slenderness allows for lateral-torsional failure. The aim was to
prove the sheeting impact to the member failure. Section IPE300 was chosen against HEA
due to higher slenderness about weak axis.
The following ratios between bending moment M and axial force N were checked: 0.5 m, 1 m
and 2 m. For the experimental investigation the ratio 1 m was chosen when the influence of
bending moment was significant.
The boundary condition of the primary examined beam-column was rigid connection V
with cantilevers. This option prevents the distortion of beam-column end cross section. It was
decided to prepare the specimens with fork boundary conditions as well. The specimen type
K was designed based on this idea.
The results suggested that a cassette wall could be sufficient for full stabilization of the beam
column. Therefore, specimens with both side walls and one side wall were designed for
testing.
5.3.
As mentioned in the previous section, the experimental investigation was preceded by numerical
modelling. These numerical models were used as the basis for a modified model that was developed
after the experimental investigation. The main modifications are listed below:
-61-
Gbor Szab
Boundary conditions of beam column. Finally two types of beam column were tested, see
Chapter 4. For both of them the adequate models were created, see Fig.54.
The yield strength and the ultimate strength were calculated as a mean value from standard
tests, see 4.4.1, Tab. 7, and used for the modeling. For material properties of other parts
(cantilevers, pin), the nominal values delivered by manufacturer were used. Different material
models were used for beam-columns and for the other structures. For beam-columns
multilinear with kinematic hardening, for cantilevers and other parts bilinear material (elastoplastic) model were used, see Fig.55, Fig.56.
-62-
Gbor Szab
The cassette wall was modelled with a system of spring elements, Fig.52. Each segment of the
wall was used separately as a spring, and connected with adjacent elements according to the
mentioned system. The force deflection relations of screws were taken from experiments, see
chapter 4.4.2. The wall stiffness was calculated according to ECCS [11]. The whole system
gives sufficient results as proved by Hapl [36].
An additional structure was used to support the cassette wall on the specimens. A frame
providing the wall support was modelled using beams. Their presence in the model was
necessary because during the experimental investigation their effect on the load carrying
capacity was detected. The numerical model allowed their elimination from final results. The
frame was made from BEAM188 3-D linear finite strain beam, see Fig.57.
-63-
Gbor Szab
The boundary conditions of whole specimens were changed. The additional frame worked as
lateral support in plane of cassette wall. Therefore the lateral supports were removed.
Beam-column imperfection was used e0,y= L/300, e0,z=L/250, according to standard [12].
5.3.1. Calibration
The updated model was calibrated according to the experimentally obtained results. Three main
parameters were followed during the calibration:
load-carrying capacity defined as maximally achieved loading in both cases; experiments and
numerical model,
stiffness,
failure mode.
For comparison of stiffness the force-deflection curves at midspan were used, see Fig.58, Fig.59,
Fig.60.
Fig.58. Comparison of load-deflection diagrams from model and experiment for specimens V1,
K1.
Fig.59. Comparison of load-deflection diagrams from model and experiment for specimens K2,
V2.
-64-
Gbor Szab
Fig.60. Comparison of load-deflection diagrams from model and experiment for specimens K3,
V3.
Following Tab. 8 table contains comparison of load-carrying capacity of models and specimens.
Tab. 8. Comparison table of load carrying capacity obtained from numerical model and
experiments.
The mean value of numerically and experimentally obtained load carrying capacity is 0.956 and the
standard deviation is 0.027.
The lateral movement of the beam-column flanges was followed when the failure mode was studied,
namely if lateral-torsional bending occurred or not. Following graphs show the difference of
deflections between compressed and tensioned flange lateral movement (U2-U1, see Fig.37, Fig.39)
obtained from experiments and from numerical models, see Fig.61, Fig.62, Fig.63.
Fig.61. Relative deflection of compressed and tensioned flange, comparison between model
and experiment, specimens K1, V1.
-65-
Gbor Szab
Fig.62. Relative deflection of compressed and tensioned flange, comparison between model
and experiment, specimens K2, V2.
Fig.63. Relative deflection of compressed and tensioned flange, comparison between model
and experiment, specimens K3, V3.
The graphs show that tested specimens V2, K2 were fully stabilized by cassette profiles. In other cases
the lateral movement of the flanges was significant.
5.4.
Parametric study
During the experimental investigation six specimens were tested. From these results general
conclusions could not be made. Therefore, a parametrical study was executed to expand the result set.
The parametric study was carried out on the calibrated models. The data expansion covered the length
of the specimens, the profile types, numbers of screws between adjacent cassettes, thickness of
cassette profiles and the bending moment distribution. The altered parameters were as follows:
sections: series of IPE and HEA sections (IPE80 - IPE600, HEA140 - HEA700),
-66-
Gbor Szab
Shear stiffness of the wall was not influenced by global buckling of wide flange of cassette
profile.
-67-
Gbor Szab
Chapter 6
Analysis of results
The project leads to several results that could be sorted into two main parts; the rotational stiffness
provided by cassette profiles and the effect of the sheeting to the load carrying capacity of connected
hot-rolled beam-column member. The first part was based on the experimental investigation followed
by the theoretical analysis (see Chapter 6.1 bellow). The investigation on the sheeting restraint
presented the main part of the work and contained experiments, numerical modeling and parametric
study.
6.1.
This part of the project was focused on the determination of the rotational stiffness provided by
cassette profiles. These results were necessary for the modelling of the main specimens and were
therefore used as input parameters during the work.
Experimentally obtained results which lead to the rotational stiffness C of the connection are
summarised in Tab. 3, see Chapter 4.2.2.
The results were divided into two groups (connection with 2x2 and 2x3 screws, respectively). The
parameters on which the stiffness C depends were the thickness t and the distances b1 and b2. As
illustrated on deformed connection in Fig.65, the stiffness C depends on distances of screws between
columncassette profile compression contacts. Distances b1 and b2 were relevant for given sign of
torsion moment while distances a1 and a2 did not influence the connection behaviour.
-68-
Gbor Szab
Gbor Szab
The coefficient k was determined by FE model, see Fig.68. The different behaviour of positive and
negative bending moment of the cassette profile and its effect on the whole stiffness was followed by
the model. The constant n was selected to minimize the scatter of results.
Fig.68. Numerical model of the connection loaded by positive bending moment, von Mises
stress in MPa.
Two equations were derived, one for connection with 2x2 screws (6.2) and the other one for 2x3
screws (6.3).
C / t n kb1 b2
(6.1)
(6.2)
(6.3)
The mean value of ratio of experimentally and calculated (according to eq. (6.2) and (6.3)) acquired
torsional stiffness Cexp/ Ccalc is 0.992 for 2x2 screws, the standard deviation is 0.158. For screw
arrangement 2x3 the mean value is 1.004 and the standard deviation is 0.083.
6.2.
The second part of the project was focused on determination of the effect of the sheeting on the load
carrying capacity of the column and the understanding of the interaction between them. An
experimental investigation and numerical model were used to achieve the results.
-70-
Gbor Szab
lateral support stabilization effect on the load carrying capacity of the supported members.
6.2.2.1.
The cassette wall provides two independent supports to the connected member: lateral restraint and
rotational restraint. The effect of the support was investigated. Two cases were compared: wall
connected to the compression flange and wall connected to the tension flange. The effect of the
rotational stiffness is significant only when the wall is connected to the tension flange. In the other
case, when the compression flange is laterally restrained, rotational movement about the longitudinal
axis did not occur and the rotational stiffness was not activated. This example is valid for most cases
when the cassette wall connected at the compression flange is able to restrain the member against
rotational movement.
The following figures show the effect of the rotational stiffness to the stabilization effect. Load
carrying capacity was determined with and without rotational restraint (lateral restraint was always
applied). The study was conducted for IPE and HEA sections with S235 and S355 materials. Two
types of wall were used, one by sheeting thickness 0.75mm and the other for 1.5mm. Constant
moment distribution with length from 2010 mm up to 15000 mm was used. The boundary conditions
are K or V.
Fig.69. Ratio between load carrying capacity for rotationally supported and non-supported
members IPE boundary condition K, supported at tensioned or compressed flange
-71-
Gbor Szab
Fig.70. Ratio between load carrying capacity for rotationally supported and non-supported
members IPE boundary condition V, supported at tensioned or compressed flange
Fig.71. Ratio between load carrying capacity for rotationally supported and non-supported
members HEA boundary condition K, supported at tensioned or compressed flange
Fig.72. Ratio between load carrying capacity for rotationally supported and non-supported
members HEA boundary condition V, supported at tensioned or compressed flange
Figures Fig.69 - Fig.72 present value between load carrying capacity of members with and without
rotational support according to following equation:
Rwith _ rot
Rwithout _ rot
(6.4)
The rotational stiffness influence for supported compressed flange is neglectible, especially for
boundary conditions V. When the cassette wall is connected at tensioned flange the rotational
-72-
Gbor Szab
stiffness may grant by 133% - K and 77% - V members load carrying capacity for section IPE and
71% - K and 38% - Vfor section type HEA.
The impact of the rotational stiffness depends on the section properties as it is shown on Fig.73.
Sections with greater torsional stiffness are able to act against rotation easier. This behaviour is
reflected on the rotational stiffness impact.
Fig.73. Ratio between load carrying capacity for rotationally supported and non-supported
members with length 6000 mm, impact of section torsional stiffness
6.2.2.2.
Distribution of deformation and internal forces along the member length is influenced by connected
cassette wall. The cassette wall presented the diaphragm with complex behaviour. The wall includes
relatively weak and stiff elements as the fasteners between adjacent cassette profiles and the cassette
profiles respectively. This feature of cassette wall leads to additional local bending of the connected
members about their weak axis and is studied in this section.
The following graph shows the different behaviour of the compressed flange for three types of lateral
support, see Fig.74:
Wall: elastic lateral restraint is presented by cassette wall (thickness 0.75 mm, 6
screws/adjacent cassettes, 2 screws between cassette and flange, cassette length 3000 mm),
The investigated member properties IPE300, boundary conditions V, constant bending moment
distribution, L=9000 mm, cassette wall connected at compressed flange. Fig.74 A shows compressed
flange behaviour for all three cases Full, Wall, Without. Flange lateral deflection without any
restraint is smooth with maximum amplitude 138 mm. Fully supported flange graph shows that points
in place of support do not move laterally. In case three, cassette Wall provides (1.42 mm) nearly
same support as Full (from amplitude) which leads to the fact that the wall support must be
significant, see Fig.74 B. The lateral deflection comparison of all cases shows that Wall stabilize up
to 99% against lateral movement if boundary values Full is equal to 100% and Without is
assigned to 0%.
-73-
Gbor Szab
Fig.74. Compressed flange lateral deflection for different wall properties; A three cases:
without support, cassette wall support, full support, B two cases: cassette wall support,
full support (same graph with larger scale)
The cassette wall stiffness has a significant impact on the global deformation of the loaded member.
The local lateral deformation amplitude (the difference of minimum and maximum deformation within
the width of one cassette) is not influenced by cassette stiffness as shown in Fig.75. This result leads to
the conclusion that the additional local bending from the sheeting connection does not depend on cross
section properties and cassette wall stiffness.
-74-
Gbor Szab
Fig.75. Compressed flange lateral deflection for different wall stiffness, section IPE300
(t-thickness, ns-number of screws, cl-cassette profile length)
Fig.76 compares two identical members with same sheeting but different steel grade. The graphs prove
that lateral movement of compressed flange is mainly influenced by cross-section properties and the
steel grade not by the sheeting stiffness.
-75-
Gbor Szab
Fig.76. Compressed flange lateral deflection for different steel grade, sections: IPE140, L=4 m;
IPE300, L=8.4 m
6.2.2.3.
The last presented results are wall stiffnesses required for full (98% of load carrying capacity of
laterally restrained member) or partial (80%) member stabilization by connected cassette wall. The
results were worked out from parametric study.
Definition of full stabilization: supported column flange is restrained against lateral movement. The
expression stabilized does not mean the same for the two investigated cases; sheeting connected at
compressed flange and sheeting connected at tensioned flange. In the case when the compressed
flange is restrained, stabilized really means that the member is globally stabilized against out-ofplane buckling. In the other case, stabilization means lateral-torsional buckling about imposed axis.
Above mentioned cases could be presented with following formulae:
Fully stabilized beam-column resistance is equal to member loading capacity under combined
load (M+N) with ratio 1m and lower:
M y , Ed
N Ed
k yy
0.98 1.0
y fy A
LTWpl , y f y
M y , Ed
N Ed
k zy
0.98 1.0
z f y A
LT Wpl , y f y
(6.5)
-76-
(6.6)
Gbor Szab
when LT, z = 1; for cassette wall connected at compressed flange. For cassette wall connected at
tensioned flange the lateral-torsional buckling about imposed axis should be considered.
Partially stabilized beam-column resistance is equal to 0.8x member loading capacity under
combined load (M+N) with ratio 1m and lower. The definition of reduction factors are same
as for full stabilization:
M y , Ed
N Ed
k yy
0.8
y fy A
LTWpl , y f y
M y , Ed
N Ed
k zy
0.8
z f y A
LTWpl , y f y
(6.7)
(6.8)
type according to boundary conditions and beamresults for stabilized compressed and stabilized
are presented for full and for partial member
paragraph. Steel grade S235 and S355 are also
Values of wall stiffnesses for each wall arrangement were worked out by ECCS recommendation [25],
see eq. (2.3). The considered minimum wall stiffness was defined as follows:
screws number between adjacent cassette: IPE, HEA140-180: 3 pcs.; 200-300: 5 pcs.; 330600: 6 pcs.
The cassette profile length was also different, but as this does not influence the result, it was not
presented. The screws at the flange were placed at the quarter of flange width from both sides. The
maximum considered web stiffness was calculated with a cassette sheeting thickness of 1.5 mm. In the
following result tables one stiffness value is given for several types of profiles because the minimum
considered wall stiffness was sufficient for several members.
The numerical models were calculated according to GMNIA (Geometrically and Materially Nonlinear
Analysis of the Imperfect Structure) analysis. The rotational stiffness of the connection between
cassette profiles and column flange was applied during calculation. For members without values given
in the following tables, the wall stiffness from used range did not satisfy the stabilization condition.
For the stabilization higher wall stiffness is required, which could be achieved for example with higher
number of screws between adjacent cassettes.
-77-
Gbor Szab
-78-
Gbor Szab
Tab. 10. Required stiffness for V, section IPE supported tensioned flange
-79-
Gbor Szab
Tab. 11. Required stiffness for V, section IPE supported compressed and tensioned flange
-80-
Gbor Szab
Tab. 12. Required stiffness for K, section IPE supported compressed flange
-81-
Gbor Szab
Tab. 13. Required stiffness for K, section IPE supported tensioned flange
-82-
Gbor Szab
Tab. 14. Required stiffness for K, section IPE supported compressed and tensioned flange
-83-
Gbor Szab
Tab. 15. Required stiffness for V, section HEA supported compressed flange
-84-
Gbor Szab
Tab. 16. Required stiffness for V, section HEA supported tensioned flange
-85-
Gbor Szab
Tab. 17. Required stiffness for V, section HEA supported compressed and tensioned flange
-86-
Gbor Szab
Tab. 18. Required stiffness for K, section HEA supported compressed flange
-87-
Gbor Szab
Tab. 19. Required stiffness for K, section HEA supported tensioned flange
-88-
Gbor Szab
Tab. 20. Required stiffness for K, section HEA supported compressed and tensioned flange
Member with lower steel grade (S235 in this case) needs equal or lower wall stiffness for
stabilization. This behaviour could be explain:
LT
Wy f y
(6.9)
M cr
(6.10)
S 235
S 355
S 235
S 355
LT
LT
LT
LT
(6.11)
Rwall
LT
R full
(6.12)
The tables could be used for design. The sheeting is primarily designed for bending due to the exterior
loads. When the stabilization effect is also considered the sheeting is subjected to combined actions.
The project did not focus on this interaction but it is known that such combination is not usually
governing for the steel cold-formed sheeting.
-89-
Gbor Szab
-90-
Gbor Szab
Chapter 7
Conclusions
The rotational stiffness of the connection between cassette profile and hot-rolled beam-column
was determined according to experimental study (Chapter 4.2). Altogether 17 specimens were
tested (Tab. 3). Different cassette profile thickness, connected column flange width and screw
arrangement was investigated. Obtained results were used as input data for main experiments
(Chapter 4.3). Formulae were derived for rotational stiffness calculation of connection for
different screw arrangements, (Eq. (6.1).
A numerical model was developed and was used to prepare full scale experiments of beamcolumn connected with cassette wall (Chapter 5.2). The software package ANSYS was used.
The modelling was focused to find the most suitable hot-rolled section for experimental
investigation. The connected cassette walls were modelled by spring elements, beam-column
by shells. The stiffness of the cassette wall was calculated according to ECCS [11]. During the
numerical modelling the effect of the cassette wall position and their stiffness were followed.
The full scale experiments were prepared with respect of obtained results.
Full scale tests were carried out to determine the effect of the cassette wall to beam-column
load-carrying capacity (Chapter 4.3). Profile IPE300 with steel grade S355 was used. Two
types of specimens were used with different boundary conditions; rigid frame support, fork
support. Altogether 6 specimens were made; 3 for both boundary condition arrangements. The
system lengths of the specimens were 5330 mm and 5680 mm according to boundary
conditions. Three types of cassette wall arrangements were analysed: connected at compressed
flange, connected at tensioned flange and half sided wall at compressed flange. To achieve
combined load in member cantilevers were used for the load application. The ratio between
bending moment and axial force was 1m. The load was applied until specimen failure. Tests
pointed out that cassette wall could fully stabilize the connected beam-column (Chapter 4.3.4).
The numerical model was calibrated (Chapter 5.3.1) according to experimentally obtained
results. Necessary changes were made to adjust the model according to specimens. A
parametric study was carried out (Chapter 5.4) for sections IPE and HEA with following
altered parameters: section type, beam-column length, number of screws between adjacent
profiles, cassette thickness, bending moment distribution. Load carrying capacity and failure
mode were observed. Altogether 8400 models for section type IPE and HEA were calculated.
The effect of rotational connection between cassette wall and flange was examined.
The rotational stiffness influence for supported compressed flange is negligible,
-91-
Gbor Szab
The supported flange behaviour observation pointed out that the lateral deflection
amplitude between full supported and supported by cassette wall was 1%. Cassette
wall stiffness has a significant impact on the global deformation of the loaded
member. The local lateral deformation amplitude (the difference of minimum and
maximum deformation within the width of one cassette) is not influenced by cassette
stiffness (Chapter 6.2.2.2).
Design tables with required wall stiffness to ensure the full stabilization of beamcolumn were made for different cassette wall stiffnesses, compressed and tensioned
flange connection, bending moment distribution, sections and lengths (Chapter
6.2.2.3). Two types of sections were studied: IPE and HEA. Calculation tool was
made for cassette wall stiffness calculation, Fig.77. With help of the tables the
positive effect of cassette wall could be used for design.
-92-
Gbor Szab
References
[1]
Davies J.M.: Light gauge steel cassette wall construction, Nordic steel construction conference
98, p. 427-440, 1998.
[2]
Davies J.M., Fragos A.S.: The local shear buckling of thin-walled cassette infilled by rigid
insulation 1.Tests, 2. Finite element analysis, Eurosteel Coimbra 2002, p. 669-688,
Akadmiai kiad, Budapest, 2002.
[3]
Davies J.M.: Light gauge steel cassette wall construction theory and practice, Journal of
Constructional Steel Research 62, p. 1077 1086, 2006.
[4]
Baehre R.: Zur Schubfeldwirk und -bemessung von Kassettenkonstruktionen, Stahlbau 56, Heft
7, p. 197-202, 1987.
[5]
Baehre R., Ladwein T.: Diaphragm action of sandwich panels, Journal of Constructional Steel
Research 31, p. 305-316, Elsevier Science Limited, 1994.
[6]
Baehre R., Holz R., Vo R.P.: Befestung von Trapezprofiltalfeln auf StahlkassettenprofilStahlbau 57, Heft 10, s. 309-311, 1988.
[7]
Kisin S.: The realistic effect of stressed skin design of typical industrial hall structure,
Proceedings of the Conference Eurosteel 99, VUT v Praze, p. 243-246, 1999.
[8]
Davies J.M., Omadibi I.A.: An ultimate limit state theory for stressed skin design, Proceedings
of the Third International Conference on Structural Engineering, Mechanics and Computation,
s. 87-92, Cape Town, South Africa, 2007
[9]
Ladwein T.: Zur Schubfeldwirkung von Sandwichelementen, Stahlbau 62, Heft 11, p. 342-346,
Heft 12, s. 361-363, 1993.
[10] Baehre R.: Zur Schubfeldwirkung von Aluminiumtrapezprofilen, Stahlbau 62, Heft 3, p. 81-87,
1993.
[11] European convention for construction steelwork: European recommendations for the application
of metal sheeting acting as a diaphragm. ECCS Technical Committee 7, Technical working
group 7.5, Publication 88, 1995.
[12] Eurocode 3: Design of steel structures Part 1-1: General rules and rules for buildings,
European Committee for Standardization, Brussels, 2005.
[13] Eurocode 3: Design of steel structures Part 1-3: General rules -Supplementary rules for coldformed members and sheeting, European Committee for Standardization, Brussels, 2005.
[14] Strnad M.: Spolupsoben pl u lehkch ocelovch hal, Stavebn informan stedisko,
Praha, 1975.
[15] epika D.: Smykov spolupsoben pl z tenkostnnch profil, dizertan prce, FSv,
VUT v Praze, 2003.
[16] epika D., Machek J., May I.M.: Shear diaphragms in purlin roofs, Eurosteel Coimbra 2002,
p. 235-242, Akadmiai kiad, Budapest, 2002.
-93-
Gbor Szab
[17] Rybn J.: Plov psoben tenkostnnch kazet, dizertan prce, FSv, VUT v Praze, 2001.
[18] Nyberg G.: Diaphragm action of assembled C-shaped panles, Swedish council for buildings
research, Document D9, 1976.
[19] DASt-Richtlinie 016: Bemessung und konstruktive Gestaltung von Tragwerken aus
dnnwandigen kaltgeformten Bauteilen, 2. berarbeitet Auflage, Stahlbau-Verlagssgesellschaft,
1992.
[20] The Steel Construction Institute: Steel designers' manual, Sixth Edition, Blackwell Science,
2003.
[21] The Steel Construction Institute: Design of Single-Span Steel Portal Frames to BS 59501:2000, The Steel Construction Institute, 2004.
[22] Lindner J., Gregull T.: Torsional restraint coefficients of profiled sheeting, IABSE colloquium,
p. 161-168, Stockholm, 1986.
[23] Boissonnade N., Jaspart J.P., Muzzeau J.P., Villette M.: New interaction formulae for beamcolumns in Eurocode 3: The French Belgian approach, Journal of Constructional Steel Research
60, p. 421-431, 2004.
[24] Greiner R., Lindner J.: Interaction formulae for members subjected to bending and axial
compression in EUROCODE 3 the Method 2 approach, Journal of Constructional Steel
Research 62, p. 757-770, 2006.
[25] European convention for construction steelwork: Rules for Member Stability in EN 1993-1-1,
Background documentation and design guidelines, ECCS Technical Committee 8 Stability,
Publication 119, 2006.
[26] Lindner J.: Stabilisierung von Trgern durch Trapezbleche, Stahlbau 56, Heft 1, s. 9-15, 1987.
[27] Lindner J.: Stabilisierung von Biegetrgern durch Drehbettung eine Klarstellung, Stahlbau 56,
Heft 12, s. 365-373, 1987.
[29] Sochor R.: nosnost vaznc a padk stabilizovanch proti klopen inkem pl, Pozemn
stavby 3, s. 124-131, Praha, 1976.
[30] Trahair N.S.: Flexural torsional buckling of structures, E&FN Spon, London, 1993.
[31] Ptursson E.H.: Column buckling with restraint from wall elements, Nordic steel construction
conference 98, Volume 1, p. 157-166, Bergen,1998.
[32] Ptrusson E.H: Column buckling with restraint from sandwich wall elements, 9th Nordic steel
construction conference, p. 291-298, Helsinki, 2001
[35] Vran T.: Rotan podepen tenkostnn ocelov vaznice krytinou, Inaugural dissertation,
CTU Prague, 2002.
[36] Hapl V., Vran T.: Vliv spolupsobc konstrukce a konstruknho detailu na nosnost
ohbanho prvku, Teoretick a kontrukn problmy oceovch a drevench kontrukci ahk oceov kontrukcie, s. 111-116, Mojmrovce, 2005.
[37] www.kovprof.cz
-94-
Gbor Szab
-95-
Gbor Szab
-96-
Gbor Szab
-97-
Gbor Szab
-98-
Gbor Szab
-99-
Gbor Szab
-100-
Gbor Szab
-101-
Gbor Szab
-102-
Gbor Szab
Bu =600 mm,
t = 0.75 mm,
L = 3000 mm,
sp = 0.35 mm/kN,
(fasteners flexibility for screws between cassette profile and column flange)
ss = 0.3 mm/kN,
p = 500 mm,
ns = 10,
c2,1 2 Bs p p / L2
c2,2 ss nsh 1 / ns
After substitution:
Gbor Szab
S 1/ 0.82 1.22 kN / mm
This value can also be obtained with help of the calculation tool shown in Fig.77.
According to design tables the required wall stiffness for boundary conditions K, wall connected at
compressed flange, triangular bending moment distribution, member length 9000 mm, S355, for full
stabilization: 0.52 kN/mm, see Tab. 12.
Member check under combined axial force and bending moment, full stabilized, cassette wall
connected at compressed flange:
Check the member resistance under combined axial force and bending moment, supported compressed
flange, full stabilization (IPE450, S355, L=9m):
9000 1
0.64
184.8 76.06
N Ed
450.103
1.07
1.00
1
0.64
0.2
k yy Cmy 1 y 0.2
/
0.874
3508.10
/1
N
y
Rk
M
1
N Ed
450.103
k yy Cmy 1 0.8
1.00 1 0.8
1.12
3
0.874 3508.10 /1
y N Rk / M 1
k yy 1.07
M y , Ed
N Ed
k yy
1
y fy A
LT Wpl , y f y
-104-
Gbor Szab
450.103
450.106
0.94
1.07
0.874 3508.103 /1
1.0 1702.103 355 /1
0.94 1
Satisfactory
Member check under combined axial force and bending moment, full stabilized, cassette wall
connected at tensioned flange:
Check the member under combined axial force and bending moment, supported tensioned flange, full
stabilization (IPE450, S355, L=9m).
9000 1
0.64
184.8 76.06
LT
Wy f y
M cr
1702000 355
1.68
214.106
Gbor Szab
f y A 355.9882 3508 kN
Cmy 1.00 for 1.00
N Ed
450.103
1.00
1
0.64
0.2
k yy Cmy 1 y 0.2
1.07
3
/
0.874
3508.10
/1
y Rk
M1
N Ed
450.103
k yy Cmy 1 0.8
1.00 1 0.8
1.12
3
y N Rk / M 1
0.874 3508.10 /1
k yy 1.07
M y , Ed
N Ed
k yy
1
y fy A
LT Wpl , y f y
200.103
200.106
1.07
1.31
0.874 3508.103 /1
0.284 1702.103 355 /1
Not satisfactory.
N cr ,T
EI
1
GI t 2 w
2
io
lT
io 294.1 mm
N cr ,T
1
210.106 791.109
6
81.10
668700
700.7 kN
294.12
90002
Af y
N cr ,T
9882 355
2.24
700700
200.103
200.106
0.64
1.07
0.176 3508.103 /1
0.284 1702.103 355 /1
Not satisfactory.
-106-