Hamk Metnet Lulea 2013 Ekirja

Virdi & Tenhunen (Editors)
METNET seminars deal with technical aspects of metal construction

as well as issues of concern to industry on management, planning
and sustainability of projects.
Metnet cooperation promotes regional innovation environments and
strengthens knowledge structures, especially in the European regions
represented by its members. At the annual seminars, the participating
partner organization represents its regional innovation environment
but simultaneously is a partner of a larger and stronger international
innovation world created by active cooperation between the institutes
and enterprises.
Proceedings of the METNET Seminar 2013 in Lule
This book covers the papers presented in the annual Metnet Seminar
in October 2013 held at Lule University of Technology. The seminar
continued the METNET tradition of presenting scientific and
development papers of high calibre.
Metnet Annual Seminar in Lule, Sweden, on 2223 October 2013
electronic
ISBN
978-951-784-642-4 (PDF)
ISSN 1795-424X

HAMKin e-julkaisuja 1/2014
Kuldeep Virdi and Lauri Tenhunen (Editors)

HAMK
printed
ISBN 978-951-784-641-7
ISSN 1795-4231

HAMKin julkaisuja 1/2014

HAMK University of Applied Sciences
Editors:
Kuldeep Virdi, Aarhus University
Lauri Tenhunen, HAMK University of Applied Sciences
printed
ISBN 978-951-784-641-7
ISSN 1795-4231

electronic
ISBN ISBN 978-951-784-642-4
ISSN 1795-424X

HAMK UAS and writers

PUBLISHER
PO Box 230
FI-13101 Hmeenlinna, FINLAND
tel. +358 3 6461
julkaisut@hamk.fi
www.hamk.fi/julkaisut
This publication has been produced in cooperation with ESF
partly funded EU project BOAT.
Hmeenlinna, January 2014
Index
PREFACE.................................................................................................................................. 6
CONNECTIONS AND MATERIALS............................................................................................ 9
Michal Jandera and Lukas Ledecky
Czech Technical University in Prague
DESIGN OF LINER TRAY WITH DISTANCE SCREW CONNECTION

IN THE NARROW FLANGE.................................................................................................... 10
Nikolay I. Vatin, Jarmo Havula, Lassi Martikainen, Alexey S. Sinelnikov,
Anna V. Orlova and Stepan V. Salamakhin
Saint-Petersburg State Polytechnical University
THIN-WALLED CROSS-SECTIONS AND THEIR JOINTS:

TESTS AND FEM-MODELLING.............................................................................................. 13
Jukka Joutsenvaara
Kemi-Tornio University of Applied Sciences
FORMING LIMIT CURVE OF HIGH STRENGTH STEEL WITH THE HELP OF

DIGITAL IMAGE CORRELATION............................................................................................ 17
EXPERIMENTAL TECHNIQUES..............................................................................................33
Pavel Ryjek and Miroslav Vok
MONITORING OF STEEL RAILWAY BRIDGE AND CONTINUOUS WELDED RAIL.................34

P.A., Kuznetsov, O.V. Vasilyeva, I.V. Kudryavtseva and A.V. Tereschenko
FSUE CRISM Prometey, Saint-Petersburg
LASER SYNTHESIS TECHNOLOGY FOR THE CREATION,

RESTORATION AND REPAIR OF COMPLEX MACHINE PARTS.............................................. 37
COMPUTATIONAL METHODS................................................................................................43
Kristo Mela and Markku Heinisuo
Tampere University of Technology
WEIGHT AND COST OPTIMIZATION OF WELDED HIGH STRENGTH STEEL BEAMS...........44
Teemu Tiainen, Kristo Mela, Timo Jokinen and Markku Heinisuo

HIGH STRENGTH STEEL IN TUBULAR TRUSSES..................................................................56

Markku Heinisuo and Timo Jokinen
TUBULAR COMPOSITE COLUMNS IN A NON-SYMMETRICAL FIRE .....................................60

Michal Malendowski and Adam Glema
Poznan University of Technology
INFLUENCE OF NATURAL FIRE SCENARIO BY CFD AND FEM ANALYSIS

COUPLING ONTO STEEL CONSTRUCTION STRUCTURAL RESPONSE.................................64
Alexander Tusnin and Olga Tusnina
Moscow State University of Civil Engineering
MEMBRANE STRUCTURES SUPPORTED BY A FRAME

OF THIN-WALLED CLOSED SECTIONS.................................................................................. 70
Olga Tusnina
Moscow State University of Civil Engineering
NUMERICAL STUDIES OF Z-PURLINS SUPPORTED BY SANDWICH PANELS...................... 78

OPEN TOPICS........................................................................................................................ 87
S.I. Matreninskiy, E.M. Chernyshov and V.Y. Mischenko
Voronezh State University of Architecture and Civil Engineering, Russia
PROBLEMS OF FUNCTIONING AND ELABORATION OF THE MASS HOUSING

DEVELOPMENT AREAS OF THE CITIES AND SETTLEMENTS...............................................88
Lauri Tenhunen and Arto Ranta-Eskola
Rautaruukki Oyj
THE BLUE BOX AND THE TECHNOLOGICAL TRAJECTORIES IN CONTEXT

OF UNIVERSITY-BUSINESS COOPERATION (UBC)............................................................. 100
Tarja Merist and Jukka Laitinen
Laurea University of Applied Sciences
SUSTAINABILITY AS A BUSINESS OPPORTUNITY TODAY AND TOMORROW:

TRIPLE HELIX PERSPECTIVE...............................................................................................111
Anatoly Perelmuter and Vitalina Yurchenko

Kyiv National University of Civil Engineering and Architecture
OPTIMIZATION OF STEEL TOWERS FOR LARGE WIND TURBINES................................... 120

Olli Ilveskoski
STRUCTURAL BRACING FORCES AND STIFFNESS......................................................... 125

Hans-Albert Staedler
Alcoa Fastening Systems Industrial Products
INNOVATION IN MECHANICAL FASTENING TECHNOLOGY

FOR MAINTENANCE-FREE JOINTS..................................................................................... 143
Christine Heistermann, Anh Tuan Tran, Milan Veljkovic and Carlos Rebelo
Lule University of Technology
University of Coimbra
FLANGELESS CONNECTIONS IN STEEL TUBULAR WIND TOWERS................................... 157
PREFACE
Some 40 universities, research institutes and enterprises have been cooperating
from the year 2006 within Metnet network. They actively utilize their mutual
connections, competitive strengths and the various cooperation possibilities.
Metnet cooperation opens new options for universities and research organizations
to customer- and needs-based product development, in collaboration with
enterprises in construction and metal product industries in Europe. It makes
it possible to foster innovative processes in partner organizations. Participating
organizations learn from each other and exchange relevant pieces of information
with target industries in Europe and beyond.
One of the outcomes of Metnet cooperation is the movement of researchers
internationally from one organization to another. This clearly arises from the
needs of increasing international presence of the metal industry. It also makes
it possible for the mother organizations to broaden the scope of marketing and
product development on larger geographical areas. Metnet partners help the
transfer of qualified personnel to matching jobs where they can best utilize their
abilities and to make the best use to their new working environment.
strengthens their knowledge structures, especially in the involved European
regions. Domestically, a participating partner organization represents its
regional innovation environment but simultaneously is a partner of a larger and
stronger international innovation world created by active cooperation between
the institutes and enterprises.
The annual Metnet Seminar concentrates on presenting new research results
in relevant scientific areas. In the Metnet Lule Seminar, in October 2013, a
new standard was brought about for the scientific and development papers. The
presented papers covered a number of themes covering both experimental and
numerical contributions.
The theme of computational mechanics included a paper on weight and cost
optimisation of high strength steel beams. Another presentation was made
on the optimisation of steel towers for large wind turbines. Experimental and
computational results were the focus of a paper describing innovative flangeless
connections for wind turbine towers. Two papers dealt with the fire limit state.
In one paper, numerical results were presented on the increasingly important
aspect of columns not exposed to fire on all four sides. As a step towards more
realistic analysis of structures exposed to fire, a method was presented for
the coupling of thermal response and the consequent mechanical response.
Problems related to the use of new grades of steel such as S960 were discussed
in another paper dealing with tubular trusses.
Thin walled sections were the topic of several papers. Finite Element analysis
and test results were presented on thin walled cross-sections and their joints,
on z-purlins, and on membrane structures supported by a frame of thin wallled
boundary elements. One paper dealt with the influence of the spacing of screw
connections on the design of liner trays.
Tests were described on the monitoring of a steel railway bridge, studying the
influence of continuous welded rails, on the use of digital image correlation
in formulating the stress-strain relationship for high strength steel, and
on a technique for creating and repairing complex machine parts using
3-dimensional laser synthesis technology. Another presentation described
recent developments on maintenance-free bolts.
METNET seminars deal with not only technical aspects of metal construction,
but also issues of concern to industry on management, planning and
sustainability of projects. Thus, one paper covered the topic of university and
business co-operation and another on sustainability as a business opportunity.
Finally, a comprehensive technical note was presented on forces and stiffness
issues relating to structural bracing.
Kuldeep Virdi
Lauri Tenhunen
CONNECTIONS AND MATERIALS
10
DESIGN OF LINER TRAY

Design of Liner Tray
Distance
Screw
Connection in the Narrow
WITHwith
DISTANCE
SCREW
CONNECTION
Flange FLANGE
IN THE NARROW
Michal Jandera
Michal
Jandera
Czech Technical
University
in Prague
Lukas Ledecky
CzechLukas
Technical University
in Prague
Ledecky
The paper describes an experimental and numerical study on structural

performance of cold-formed liner trays cladding system. In the described
system, the outer sheeting is connected by distance screws to the narrow
The paper describes
and
studythermal
on structural
flanges of an
the experimental
liner tray profile.
It numerical
allows additional
insulationperformance
of cold-formed
linerthetrays
cladding
the (see
described
the outer
between
liner tray
and outer system.
trapezoidal In
sheeting
example onsystem,
Figure
1).
This
is
beneficial
for
the
thermal
insulation
and
acoustics
performance
of
sheeting is connected by distance screws to the narrow flanges of the liner tray
the system.
As widely known
and insulation
used in the current
Eurocode
1993-1-3
profile. It allows
additional
thermal
between
the(EN
liner
tray and outer
2006), when the narrow flange (restrained by the sheeting) is in compression,
trapezoidal sheeting
(see example on Figure 1). This is beneficial for the thermal
the resistance of the liner tray is influenced by the longitudinal spacing of
insulation andfasteners
acoustics
performance
of the
system.
As widely
known
supplying
lateral restraint
to the
narrow flanges.
In the
design and used in
the current Eurocode
(EN 1993-1-3
the narrow
(restrained
by the
procedure described
here, the2006),
length ofwhen
the distance
screw is flange
also taken
into
sheeting) is account.
in compression, the resistance of the liner tray is influenced by the
longitudinal spacing of fasteners supplying lateral restraint to the narrow flanges. In

the design procedure described here, the length of the distance screw is also taken
into account.
1 liner tray
2
support
1 liner
tray / column
3
foundation
2 support/column
4 facade profile
3 foundation
5 sealing tape
4 facade
profiletape
6
sealing
7
polyurethane
seal
5 sealing
tape
8
mineral
wool
6 sealing tape
7 polyurethane seal
8 mineral wool
Figure 1. Liner tray wall system using distance screw.
Figure 1. Liner tray wall system using distance screw.
Six bending tests of simply supported liner trays were carried out. For all of them, the
narrow flange was in compression and distributed load was simulated by loads at
four points of equal spacing. The width of the supports was designed so as not to
11
Six bending tests of simply supported liner trays were carried out. For all
of them, the narrow flange was in compression and distributed load was
simulated by loads at four points of equal spacing. The width of the supports
was designed so as not to affect the bending resistance. The liner tray profiles
were 150 mm height and 600 mm width of the thickness 0.75, 0.88 and 1.0
mm. The screw provided 40 mm distance between the outer sheeting and the
trays. Total depth of the system was then 190 mm. Spacing of the screws was
330 mm. Outer sheeting was made of 0.63 mm thick trapezoidal sheet. Failure
occurred between the points of application of loads.
An FE modelAninFE
software
Abaqus
waswas
made,
GMNIA
analysis.
The model was
model in software
Abaqus
made,using
using GMNIA
analysis.
The model
was the
validated
using the results.
experimentDespite
results. Despite
simplifications
validated using
experiment
some some
simplifications
ininthe numerical
the numericalwas
model,
the agreement
was found
to beThe
very good.
The resistance
model, the agreement
found
to be very
good.
resistance
calculated by the
calculated by the model was, on average, about 2% lower with 4.8% standard
model was, deviation.
on average,
about
2
%
lower
with
4.8
%
standard
deviation.
The FE
The FE model failure mode was also found to correspond well with
model failurethemode
found
test, aswas
shownalso
in Figure
2. to correspond well with the test, as shown in
Figure 2.
Figure 2. Failure mode of the test and FE model.
Figure 2. Failure mode of the test and FE model.
on the parametric study, a general design formula was developed.

Later, basedLater,
on based
the parametric
study, a general design formula was developed.
Material strength up to fy = 420 MPa was considered. Liner tray profiles of
Material strength
up to fymm
= and
420thickness
MPa was
considered.
Liner
tray of
profiles
of height
height 100150
0.751.0
mm were used.
The length
the
100 150 mm
and
thickness
0.75
1.0
mm
were
used.
The
length
of
the
distance
distance screw (the gap between the liner tray and the outer sheeting itself)
was between
considered as
0, 20,
40 tray
and 60
mm.the
These
parameters
therefore
setwas
the considered
screw (the gap
the
liner
and
outer
sheeting
itself)
proposed
formula.
as 0, 20, 40 limits
and of
60themm.
These
parameters therefore set the limits of the proposed
formula.
Variation in the spacing between the screws along the liner tray was also considered.
The spacing was assumed between 165 mm (screw in each rib of the outer sheeting)
and 660 mm. The influence of the spacing was retained as given by Eurocode
(EN 1993-1-3 2006) by the correlation factor b. The design formula was finally
obtained as follows:
(1)
12
Later, based on the parametric study, a general design formula was developed.
Material strength up to fy = 420 MPa was considered. Liner tray profiles of height
100 150 mm and thickness 0.75 1.0 mm were used. The length of the distance
screw (the gap between the liner tray and the outer sheeting itself) was considered
as 0, 20, 40 and 60 mm. These parameters therefore set the limits of the proposed
formula.
Variation in the spacing between the screws along the liner tray was also
considered.
The spacing
between
165 mm
(screw
inconsidered.
each rib of
Variation
in the spacing
betweenwas
the assumed
screws along
the liner
tray was
also
The spacing
wassheeting)
assumed and
between
165 mm
in each
rib spacing
of the outer
the outer
660 mm.
The(screw
influence
of the
was sheeting)
retained
and 660
influence(EN
of the
spacing
wasbyretained
as givenfactor
by Eurocode
as mm.
given The
by Eurocode
1993-1-3
2006)
the correlation
b. The
(EN 1993-1-3
2006) by
the
correlation
factor
b. The design formula was finally
design formula
was
finally
obtained
as follows:
obtained as follows:
(1)
where b is correlation factor:
for s1 300 mm or
for 300 mm < s1 1000 mm, s1 is the screw spacing along
the liner tray profile.
The formula was designed to be safe in all cases. For the liner tray of 0.75 mm
thickness Equation 1 under-predicts the resistance by approx. 5 % on average. For
The formula was designed to be safe in all cases. For the liner tray of 0.75 mm
liner trays of greater thickness, lower height, or lower yield strength than 420 MPa,
thickness
Equation
1 under-predicts the resistance by approx. 5% on average.
the formula
is even
more conservative.
For liner trays of greater thickness, lower height, or lower yield strength than
420 MPa, the formula is even more conservative.
ACKNOWLEDGMENTS
The technical and financial support of RUUKKI CZ, namely Dan Marha and
Jan Samec is gratefully acknowledged.
REFERENCES
EN 1993-1-3, 2006. Design of steel structures Part 13: General rules
Supplementary rules for cold-formed members and sheeting. Brussels: CEN.
THIN-WALLED CROSS-SECTIONS AND THEIR JOINTS:

TESTS AND FEM-MODELLING
Nikolay I. Vatin
Saint-Petersburg State Polytechnical University, vatin@mail.ru
Jarmo Havula
HAMK University of Applied Sciences, jarmo.havula@hamk.fi
Lassi Martikainen
HAMK University of Applied Sciences, lassi.martikainen@hamk.fi
Alexey S. Sinelnikov
Saint-Petersburg State Polytechnical University, alexey_sinelnikov@mail.ru
Anna V. Orlova
Saint-Petersburg State Polytechnical University, a_v_orlova@mail.ru
Stepan V. Salamakhin
Saint-Petersburg State Polytechnical University, ssalamakhin@gmail.com
This summary report describes the experimental and numerical research on a thin-walled
cross-sections compression resistance and the shear strength of their joints. This work
is continuation of an earlier paper (Vatin and Sinelnikov 2013). Current situation in the
Russian market concerning the usage of cold-formed thin-walled cross-sections is aimed
at starting up stipulation of such elements in the building industry.
In this area a number of Doctoral theses have been defended during recent
years in Russia (Tusnin 2009, Belyy 1983, Astahov 2006). Theoretical research
and laboratory tests were done only for specific types of thin-walled crosssections. Jyrki Kesti contributed significantly to the development of local and
distortional buckling of perforated steel wall studs. Today thin-walled coldformed steel structures have a good presence in the Finnish building industry.
Steel galvanized C- and U-profiles and thermo-profiles are types of thinwalled cross-sections which are normally used in small house construction.
Thermo-profiles have slots in webs that decrease the thermal flow through the
web, but have a negative effect on the strength of the profiles.
Reticular-stretched thermo-profile, a new type of thin-walled cross-section,
is the subject of research on reticular-stretched thermo-profiles and their
joints. The following profiles are discussed:
1.
Specimen S1 (stud) C 200-45-2,0;
2.
Specimen S2 (rack) 200-50-2,0.
13
14
Steel used for specimen production has the following parameters:

1.
Steel grade S350GD (yield strength not less than 350 H/mm2);
2.
Coating mass, 350 g/m2;
3.
Coating thickness, 25 microns.
The research goal was to form the theoretical rationale for usage of the
reticular-stretched thermo-profile including buckling and shear strength
analysis based on the laboratory tests.
Tests were carried out in the Sheet Metal Centre at HAMK, using contemporary
laboratory stand (Instron 3250).
First stage included preparation of specimen. Support blocks were placed
inside the profile at the ends. The specimen was placed in vertical position in
the test rig using level board (Stayer Standard 3460-150_z01).
Different boundary conditions were used for the two ends of the specimen.
Three specimens had the following boundary conditions. Upper end of the
specimen had a hinged support: the load of a hydraulic cylinder was applied
through hinged block and then a thick steel plate was attached to the upper
end of the specimen. Lower end of the specimen stood on the floor and was
semi-rigid in nature.
Specimen S3 had hinged boundary conditions for both end of the specimen.
Upper end of the specimen was like for other specimen described above.
Lower end of the specimen had hinged support.
The specimen was compressed under various loads and deformations were
recorded. To obtain the buckling force a load-deformation diagram was
plotted and analyzed. From the load-deformation diagram it was determined
that buckling failure was achieved at approximately 54 kN. The buckling mode
was one half wave of the sinusoid. Some short downfall of the applied load was
observed, demonstrating that first local buckling of the web takes place and
then sudden overall buckling of the profile takes place.
The load-deformation pattern for self-drilling screws joint shows shear
behavior through the following stages: First, all the clearances are closed;
Second stage demonstrates elasto-plastic strain; Third stage is noted by the
yield, strengthening, bearing failure of the sheet and achievement of the
ultimate strength; and the Fourth stage crushing of the joint. At the higher
load level the end of the stud contacts the rack profile placed on a rigid base.
The ultimate strength achieved was a force equal to 93 kN.
Numerical modeling of thin-walled cross-sections and their stud-to-rack
joint was done with contemporary analysis software (SCAD Office and Lira)
using finite element method (FEM). FEM-models parameters were the same
as for the tests described above. During the modeling process the thin-walled
profile based on shell- and bar-elements and joint based on solid-elements
were created and buckling/shear analysis tasks showed good results.
Point load (80kN) was applied to the flexural center (FC) and points nearby
it to justify different type of profile deformation. It was shown that when the
load point is situated before FC (6.0, 9.0 mm from the outside surface of the
web) thin-walled profile bends inside itself. The differences between test and
FEM-modelling results are equal to 0.33 and 3.75% accordingly.
When the load point is situated after FC (11.5 mm from the outside surface of
the web). In this case the the thin-walled profile bends outwards, which was
different from the test result. The difference between test and FEM-modelling
results is greater at 6.03%.
As for screw connection it was decided to model it using two-node FE with
unilateral elastic constraint between nodes. Point load was applied to the
flexural center (FC) step by step (5.0, 10.0, 15.0, 20.0, 25.0, 30.0 kN). Real
bearing failure of the screw connection takes place due to crushing of the
element material under a load of 25 kN. The FEM-model showed that the main
stress of steel nearby connection place was approaching 335 MPa (more than
the yield strength of the steel, accordingly to Table 3.1b Eurocode 3 Part 13).
This is followed by large plastic deformations and round-form screw hole
changes to oviform.
Results of experimental investigation into the behavior of thin-walled crosssections by compression (buckling analysis) and shear strength of their
joints have been reported. For both tests numerical analysis was carried out
including bar/shell finite elements for compression and solid finite elements
for shear strength analysis.
Compression bar buckling has resulted in the axial failure of profile specimens
S3 at a buckling force 53.82 kN. Results of numerical analysis (shell finite
elements) differ from compression tests by about 4%. When bar finite element
are used, the results differed from the tests by about 22%. The analysis clearly
demonstrated that existing Russian design guidelines for thin-walled crosssection modeling by bar finite elements is not exact and could be used only by
taking into account an extra safety factor of 1.2.
Actual bearing failure the screw connection takes place due to crushing of the
element material. Results of numerical analysis (solid finite elements) differ
from shear strength tests by only about 2%.
The experimental work was commented upon by Arto Ranta-Eskola, Director
of Research, Rautaruukki Oyj (Finland).
15
16
REFERENCES
Astakhov, I. V (2006). Prostranstvennaia ustoichivost elementov konstruktsii
iz kholodnognutykh profilei, Dissertation, St.-Petersburg, 123 p.
Belyi, G. I (1983). Raschet uprugoplasticheskikh tonkostennykh sterzhnei
poprostranstvenno-deformiruemoi
scheme,
Stroitelnaia
mekhanika
sooruzhenii: Mezhvuz. temat. sb. tr, LISI. 42, pp. 4048.
Kesti, J (2000). Local and distortional buckling of perforated steel wall studs,
Dissertation for the degree of Doctor of Science in Technology, Espoo, 101p.
+ app.19p
Tusnin, A. R (2009). Chislennyi raschet konstruktsii iz tonkostennykh
sterzhnei otkrytogo profilia, Moscow: Izd-vo ASV, 143 p.
Vatin, N.I. and Sinelnikov, A.S (2013). Strength and Durability of Thin-Walled
Cross-Sections, Design, Fabrication and Economy of Metal Structures.
International Conference Proceedings, 2013, Miskolc, Hungary, April 2426,
Miskolc, pp. 165170
FORMING LIMIT CURVE OF HIGH STRENGTH STEEL

WITH THE HELP OF DIGITAL IMAGE CORRELATION
Jukka Joutsenvaara
Kemi-Tornio University of Applied Sciences, Technology RDI Department,
Kemi 94600, Finland
ABSTRACT
Requirements for modern forming operations require in-depth knowledge of material
behaviour in various stress-strain stages. Conventional method of evaluating material
properties is the well documented and standardized uni-axial tensile test. The tensile test
provides an easy way of comparing different materials. However, the test normally does
not include data after the uniform elongation. The strain state after the uniform elongation
transforms into diffuse necking and after that to a localized necking. In order to add more
usability of acquired data it is often necessary to test material in various strain states even
up to onset of localized necking. This is due the notion that most of the forming processes
induce thinning to the part. The accuracy of estimations of thinning plays an important
role in forming process design phase, e.g. in feasibility study. One way of acquiring such
data is to define forming limit diagram (FLD) for the material. The forming limit diagram
is constructed based on forming limit curve (FLC). The tests described here were made
according to the current standard SFS-EN ISO 12004-2 with Erichsen Universal Sheet Metal
Testing Machine, model 145-60. In conjunction, the GOM ARAMIS digital image correlation
(DIC) system was used to measure the strain field over the whole visible area of the
specimen during the test procedure. The tested steel sheet material for the following study
was thermo-mechanically hot rolled, cold formable, structural steel with the minimum yield
strength of 650 N/mm2 and 3 mm in thickness. In addition to standardized test results, the
strain rate in the A-geometry is observed in the presented paper for evaluation purposes.
The usability of A-geometry strain data for a bulge stress analysis is also evaluated. The
stress strain curves obtained from these various tests are then compared. Secondly, the
strain rate of a standard tensile test specimen is compared with the strain rate of the FLC
E-geometry specimen in major strain direction. This data is gathered for future use in
research of the effect on results based on strain rates of different tests.
INTRODUCTION
Higher strengths of materials usually means lower formability. With modern
steels it is possible to increase the formability by alloying or by novel production
methods. The use of high strength steels instead of regular mild steels in
construction and end products is one possibility to decrease the carbon foot
print. The design and usage guidelines from the manufactures offer valuable
information on how the materials potential is best utilized in a safe manner.
17
18
Complex forming problems, as typically found in automotive industry as

well as in various other end products, require more detailed information of
material behaviour during the forming process. Advanced forming methods,
such as deep drawing or hydroforming, are processes that require more
information than a standard tensile test can provide. A series of tests is
made with variable geometry to better represent different strain stages that
a material could exhibit during a forming process. The FLC specimen goes
through the same elongation phases as in normal tensile test specimen, but
with slightly different strains depending upon the biaxiality of the test and
on the specimen geometry. In an FLC test the specimen in pushed out of
plane by a spherical punch and the edges of the specimen are restrained by
sheet holder. The specimen experiences elongation as well as displacement
from its original position during the test due the test set-up. After the uniform
elongation is reached, diffuse necking can be clearly seen especially on
narrower specimens as with the tensile test specimens. The diffuse necking
in full round specimens are seen as a band of increased elongation around the
dome at certain distance from assumed centre point.
As the forming process progresses the correct evaluation of thinning plays
an important role. The thinning of the material determines the usability of
the end product. For test purposes, the thinning in narrowest FLC specimen
is similar to that in regular tensile test specimen. The difference in strain
behaviour and evolution grows when testing wider specimens and eventually
fully round specimen, namely A-geometry of the FLC specimens.
The forming limit is primarily material specific but also thickness of the
specimen affects the position of the forming limit curve. Also, the conventional
tensile test is widely used for comparison purposes, although the more indepth analysis requires additional equipment and software, for instance,
Digital image correlation (DIC) with automated strain analysis. Numerous
studies have been made in recent years about the different formability tests in
order to improve material models for the computational evaluation of forming
processes. (Ghadbeigia et al. 2010, Kim et al. 2005)
In this study the formability of Class 650 high strength steel was investigated.
The forming limit curve tests were carried out with Erichsen Universal Sheet
Metal Testing Machine, model 145-60 according to SFS-EN ISO 12004-2
which is based on the Nakajima test set-up. The maximum sheet holder and
drawing force of the test machine are 600 kN. Digital image correlation
equipment was used for measurement and evaluation of surface strains on
the specimens during the test. A continuous image acquisition during the test
was needed for complete analysis, from start to failure of the specimen. From
the selected specimens of the FLC test series additional analyses were made
and the results were recorded. The strain evolution in different specimen
geometries was also investigated in the major axis direction.
19
EXPERIMENTS
Material and Specimen Specifications
The tested material is single-phase carbon steel with minimum yield strength
of 650 N/mm2. Adequate forming properties of the material are achieved by
controlling the grain size during the manufacturing process. The controlled
thermo-mechanical hot rolling process is also a key factor when determining
the strength properties of the final product. Typically the rolling process is
done below the re-crystallization temperature of that alloy in question. The
use of this steel class, instead of regular constructional steel, could provide
cost savings in materials for various constructions (Vierel 2012).
The specimen geometries for determining the forming limit curve are shown in Figure 1. The
The specimen geometries for determining the forming limit curve are shown
dimensions are selected
appropriate
forarethe
Erichsen
test machine.
specimens
were cut from
in Figureas
1. The
dimensions
selected
as appropriate
for theThe
Erichsen
test
the sheets with laser.
TheThe
rolling
direction
specimens
also laser.
determined
by the used standard
machine.
specimens
were in
cutthe
from
the sheetsiswith
The rolling
direction
specimens
is also determined
used standard
and itthorough cleaning
and it is visible in
Figure in1.the
The
test preparation
phase ofby
thethespecimens
includes
is visible
in Figure
The
testis
preparation
phasewith
of the
of the surface with
ethanol.
The 1.
test
area
then sprayed
anspecimens
even coatincludes
of matt white finishing
thorough cleaning of the surface with ethanol. The test area is then sprayed
and on top of that a stochastic spray pattern is made with black paint. The analyser software
with an even coat of matt white finishing and on top of that a stochastic spray
requires the high-contrast
stochastic
pattern
onThe
the analyser
investigated
surface.
pattern is made
with black
paint.
software
requires the highcontrast stochastic pattern on the investigated surface.
RD
1. Thegeometries
FLC-specimen
geometries
FigureFigure
1. The FLC-specimen
with dimensioning.
with dimensioning
In the other part of the test series, standard-sized tensile test specimens were used. The test series
are partly based on the previous work of the author, so the tested material was high strength, singlephase, Class 650 steel with nominal thickness of 3 mm. The actual dimensions for the specimens
Figure 1. The FLC-specimen geometries with dim

20
In the other part of the test series, standard-sized tensile test specim
In the other part of the test series, standard-sized tensile test specimens were
areThepartly
based
on based
the previous
work
theauthor,
author,
used.
test series
are partly
on the previous
workof
of the
so theso the tested ma
tested
material
was
high
strength,
single-phase,
Class
650
steel
with
nominal
phase, Class 650 steel with nominal thickness of 3 mm. The actua
thickness of 3 mm. The actual dimensions for the specimens are shown in
are2.shown in Figure 2.
Figure
Specimen dimen
a0Thickness of sp
b0Width of gauge
LtTotal length of
LoGauge length
LcLength of strai
RRadius of Fille
WgSpecimen wid
Specimen dimensioning
Figure 2. Dimensions of the tensile test spe

a0 Thickness of specimen 3.04 mm
b0 Width of gauge length 12.5 mm
Lt Total
lengththe
of specimen
285 mm tensile tests, the specimens had an extra thinni
As
with
previous
Lo Gauge length 80 mm
the
gauge length on both sides. The extra thinning was used in orde
Lc Length of straight part 120 mm
middle
the
RRadius of of
Fillet 37
mm specimen during the tensile test. The predetermined
helps
thewidth
set-up
Wg Specimen
30 mm phase of the optical measurement equipment for the
for the tensile test is the same as with FLC-specimens.
Figure 2. Dimensions of the tensile test specimen.
21
As with the previous tensile tests, the specimens had an extra thinning of
0.05 mm in the middle of the gauge length on both sides. The extra thinning
was used in order to steer the deformation to the middle of the specimen
during the tensile test. The predetermined location of maximum thinning
helps the set-up phase of the optical measurement equipment for the test. The
specimen preparation for the tensile test is the same as with FLC-specimens.
Use of digital image correlation as a tool for more in-depth research has been
validated in numerous publications. Especially in formability evaluations,
such as described here, Nakajima-test and others, computerization provides
a substantial improvement on accuracy and on work load compared with,
for instance, a manual measurement of strain field on a specimen. With
the ARAMIS system, a stochastic pattern is used but in other commercial
solutions prefabricated grids, namely square or circle, are used for strain
analysis. (Banabic et al. 2010, Cordero et al. 2005, Yang et al. 2010)
Use of digital image correlation as a tool for more in-depth Proceedings

researchofhas
been validated in numerous
the METNET Seminar 2013 in Lule
21
publications. Especially in formability evaluations, such as described here, Nakajima-test and
others, computerization provides a substantial improvement on accuracy and on work load
compared with, for instance, a manual measurement of strain field on a specimen. With the
ARAMIS system, a stochastic pattern is used but in other commercial solutions prefabricated grids,
limit
namely squareForming
or circle,
arecurve
usedtest
for strain analysis. (Banabic et al. 2010, Cordero et al. 2005, Yang
et al. 2010)
During the test, the specimen is punched out-of-plane with a spherical punch
while
thetest
edges are locked with sheet holder. Drawbeads in the tooling can also
Forming limit
curve
be used to enhance the locking of the edges of the specimen. The tests were
done at room temperature. Based on the standard, a single punch speed was
During the test, the specimen is punched out-of-plane with a spherical punch while the edges are
used, namely 90 mm/min, or 1.5 mm/s, during the test. In order to accurately
locked with sheet
holder. Drawbeads in the tooling can also be used to enhance the locking of the
capture the ultimate or fracture strain on the specimen, a fast frequency of
edges of the specimen.
The
tests wereThe
done
at room
temperature.
Based
onHz
theimage
standard, a single
image taking
is required.
standard
suggests
a minimum
of 10
punch speed capture
was used,
namely
90
mm/min,
or
1.5
mm/s,
during
the
test.
In
order
at the time of failure. Due to the memory limits of the DIC computer, to accurately
capture the ultimate
or fracture
strain
on the
specimen,
fast frequency
image1.5taking
the beginning
of the
test was
recorded
with alower
frequency, of
namely
Hz. is required.
The standard By
suggests
a minimum
of of
10images,
Hz image
capture attime
theistime
of failure. Due to the
minimizing
the number
the processing
also minimized.
memory limits
of the DICview
computer,
the beginning
the test
A schematic
of the set-up
is shown inofFigure
3. was recorded with lower frequency,
namely 1.5 Hz. By minimizing the number of images, the processing time is also minimized. A
schematic view of the set-up is shown in Figure 3.
Light source
Camera
Camera
Optional drawbead
R
Lubrication
layer
Figure 3. Schematic view of the FLC test set-up
Figure 3. Schematic view of the FLC test set-up.
The ARAMIS camera system with 65mm focal length lenses was used to capture the surface
strains. The selected lenses were used due to the dimensions of the test crevice. The lighting for
the test set-up was provided by a projector placed between the two cameras. The post-processing,
i.e. strain analysis, of the images was done with the 3-D Analysis software ARAMIS from
Gom gmbh (Germany). An actual test set-up is shown in Figure 4.
22
22
The ARAMIS camera system with 65 mm focal length lenses was used
to capture the surface strains. The selected lenses were used due to the
dimensions of the test crevice. The lighting for the test set-up was provided by
a projector placed between the two cameras. The post-processing, i.e. strain
analysis, of the images was done with the 3-D Analysis software ARAMIS
from Gom gmbh (Germany). An actual test set-up is shown in Figure 4.
Figure 4. The test equipment.
Figure 4. The test equipment
Tensile
Tests
and Strain Measurements
Tensile Tests and
Strain
Measurements
A different test set-up of DIC equipment is required for the tensile test due to
A different test set-up

of DIC
forlength
the tensile
testpolarized
due to filters
the proximity of the
the proximity
of equipment
the specimen.isArequired
23 mm focal
lens with
specimen. A 23mm
focalsources
lengthwere
lensused
withforpolarized
filters
and light
sources
were used for the test.
and light
the test. The
ARAMIS
system
was connected
The ARAMIS system
was
connected
the Zwick
Roellthrough
250kN I/O
tensile
machine
through I/O
with the
Zwick
Roell 250with
kN tensile
test machine
portstest
in order
to
get
synchronised
data
from
the
load
cell
and
extensometer
during
the
test.
The
ports in order to get synchronised data from the load cell and extensometer during the test. The test
test speeds
matched
thethe
strain
rate from
the FLC-test E-geometry.
speeds were matched
with were
the strain
ratewith
from
FLC-test
E-geometry.
The equipment for tensile
The equipment for tensile test is shown in Figure 5.
test is shown in Figure 5.
specimen. A 23mm focal length lens with polarized filters and light sources were used for the test.
The ARAMIS system was connected with the Zwick Roell 250kN tensile test machine through I/O
ports in order to get synchronised data from the load cell and extensometer during the test. The test
speeds were matched with the strain rate from the FLC-test E-geometry. The equipment for23tensile
test is shown in Figure 5.
Figure 5. Set-up for the tensile test.
Figure 5. Set-up for the tensile test
Images were recorded at two frequencies, namely 1.5 Hz and 10 Hz as with

FLC-specimens. The resolution of images for the analysis was 2448
2048 pixels. The post-processing of the captured images was done with a
3-D Analysis software ARAMIS from
23 GOM gmbh (Germany). After postprocessing, the displacements and strain information of the whole visible
surface of the specimen is acquired. The accuracy or level of detail, of the
measurements depends also on the selected measuring volume which in turn
is defined by the proper selection of lenses and their aperture values. (GOM
gmbh 2008)
In the ARAMIS software, the same facet size was used for the both of the
measurements. The facet size, i.e. the smallest software generated local
extensometer, is also one of the factors that affect the scale of the results.
The smaller facets provide more detailed information but also require more
computational time in the main computer.
The DIC measurements have been widely used in various studies and research.
Advances in photogrammetry equipment have also improved the quality of
the data acquired from these measurements. The measurement procedure
with non-contacting optical method provides faster testing with continuous,
high resolution data. For instance, the scope of usage can be seen in various
publications and conference proceedings. (Proulx 2013, Banabic 2010)
RESULTS
Curves from FLC-tests
The Forming Limit Diagram (FLD) is constructed based on the forming limit
curve and corresponding strains in the two principal axes. The FLD describes
various publications and conference proceedings. (Proulx 2013, Banabic 2010)
24
RESULTS
Curves from FLC-tests
the material
and behaviour
in strain
inducing
The Forming
Limitproperties
Diagram (FLD)
is constructed
based
on the forming
forming processes.
limit curve and
corresponding
strainsFLD
in the
principal
axes.6. The FLD describes the material properties and
The acquired
is two
shown
in Figure
behaviour in strain inducing forming processes. The acquired FLD is shown in Figure 6.
Figure
The Forming
Limit Diagram of tested material.
Figure 6. The Forming
Limit 6.
Diagram
of tested material.
24FLC-tests
Surface Strains Surface
and Strain
Evolution
FLC-tests
Strains
and StraininEvolution
in
With full-field
optical
analysis, it
possiblesurface
to acquire
surface
strains
of thesurface. From
With full-field optical
analysis,
it is possible
toisacquire
strains
of the
visible
visible
surface.
From
thatbe
image
sequencedepending
various dataoncould
be constructed
that image sequence
various
data
could
constructed
the research
topic. The surface
depending
on
the
research
topic.
The
surface
strains
tell
how
the
material
strains tell how the material behaves in particular forming process. For successive forming process,
in particular forming process. For successive forming process, it
it is important tobehaves
know the
limits for safe and unsafe strain combinations in principal directions.
is important to know the limits for safe and unsafe strain combinations in
When the strainsprincipal
supersede
the forming limit curve, various failures are expected, for instance,
directions. When the strains supersede the forming limit curve,
wrinkling, excessive
thinning,
In for
Figures
7andwrinkling,
8 can beexcessive
clearly thinning,
seen the evolution of
various failures shearing.
are expected,
instance,
surface strains. shearing. In Figures 7 and 8 can be clearly seen the evolution of surface strains.
Figure
The major
strainstrain
evolutionevolution
of selected points
specimen surface.
Figure
7.7.The
major
ofonselected
points on specimen surface.
Figure 7. The major strain evolution of selected points on

Surface Strains and Strain Evolution in FLC-tests
25
With full-field optical analysis, it is possible to acquire surface strains of the visible surface. From
that image sequence various data could be constructed depending on the research topic. The surface
strains tell how the material behaves in particular forming process. For successive forming process,
it is important to know the limits for safe and unsafe strain combinations in principal directions.
When the strains supersede the forming limit curve, various failures are expected, for instance,
wrinkling, excessive thinning, shearing. In Figures 7and 8 can be clearly seen the evolution of
surface strains.
evolution of selected points on specimen surface.

1
1
Figure 7. The major strain evolution of selected points on specimen surface.
2
2
Figure 8. The surface strains of A-geometry specime

25
3
Figure
The strains
surface
strains specimen
of A-geometry
Figure
8. The8.surface
of A-geometry
during the test.specimen during the test.
25
26
The strain rate information is also easily acquired due to the combination
of image acquisition and analysis equipment and software. In Figure 9 are
The strain rate information
is also easily acquired due to the combination of image acquisition and
shown the strain rates of selected points on the specimen surface during the
analysis equipment
and
In that
Figure
9 aretheshown
thelimits,
straintherates
of the
selected
points on the
test.
As software.
with any tests
evaluate
forming
end of
test
specimen surface shows
duringsharp
the changes
test. Asinwith
any
tests
that
evaluate
the
forming
limits,
the
end of the
strain rate. This is closely related to the onset of the
diffusion in
and
local rate.
necking
of the
material.related
In the Figure
are shown
test shows sharp changes
strain
This
is closely
to the 10
onset
of thefew
diffusion and local
images of
testimages
overlaidof
with
strain rate
necking of the material.
Inthe
theanalysed
Figure surface
10 are during
shownthe
few
thethe
analysed
surface during the
information.
test overlaid with the strain rate information.
Figure 9. The major strain rate evolution during the test.
Figure 9. The major strain rate evolution during t

27
The strain rate information is also easily acquired due to the combination of image acquisition and
analysis equipment and software. In Figure 9 are shown the strain rates of selected points on the
specimen surface during the test. As with any tests that evaluate the forming limits, the end of the
test shows sharp changes in strain rate. This is closely related to the onset of the diffusion and local
necking of the material. In the Figure 10 are shown few images of the analysed surface during the
test overlaid with the strain rate information.
jor strain rate evolution during the test.

1
1
Figure 10. The variation of major strain rate of A-geometry spe
26
3
Figure 10.Figure
The10.
variation
ofofmajor
strain
of A-geometry
specimen
during the test (1-3).
The variation
major strain
rate rate
of A-geometry
specimen during
the test (13).
or strain rate of A-geometry specimen26during the test (1-3).
28
Figure 10. The variation of major strain rate of A-geometry specimen during the test (1-3).
Surface Strains and Strain Evolution in Tensile Tests
Surface Strains and Strain Evolution in Tensile Tests
One of the most commonly used forms of data collected from a tensile test is
stress-strain
information
is used
different
One of the most acommonly
usedcurve.
forms Itofholds
data collected
fromthat
a tensile
testtoiscompare
a stress-strain
curve.
materials
or
to
determine
their
stress
bearing
capacity.
The
same
surface
It holds information that is used to compare different materials or to determine their stress
bearing
parameters
along withalong
strainwith
andstrain
other and
derivative
data can be
from
capacity. The same
surface parameters
other derivative
dataprocessed
can be processed
the optically constructed surface analysis. For instance, a strain in different
from the optically constructed surface analysis. For instance, a strain in different parts of the tensile
parts of the tensile test specimen can be easily viewed with the help of fulltest specimen can be easily viewed with the help of full-field analysis. Example views from
field analysis. Example views from standard tensile test are seen in Figure 11.
standard tensile test are seen in Figure 11.
3
2
1
Figure 11. The major strain (1) and major strain rate (2) evolution during the tensile test.
Figure 11. The major strain (1) and major strain rate (2) evolution during the tensile test.
27
Surface Strains and Strain Evolution in E-geometry of FLC-test

The E-geometry specimen of FLC-test series is the closest to a tensile test
specimen, so it is used to evaluate the strain rate. From the analysis and by
visual estimation, it is seen that the test speed resulting from FLC-testing
would require tenfold increase in speed of the crosshead of the tensile test
machine, in order to match the test time length. The major strain and major
strain rate of two selected points on the E-geometry specimens surface are
shown in Figure 12.
Surface Strains and Strain Evolution in E-geometry of FLC-test

The E-geometry specimen of FLC-test series is the closest to a tensile test specimen, so it is used to
evaluate the strain rate. From the analysis and by visual estimation, it is seen that the test speed29
resulting from FLC-testing would require tenfold increase in speed of the crosshead of the tensile
test machine, in order to match the test time length. The major strain and major strain rate of two
selected points on the E-geometry specimens surface are shown in Figure 12.
Figure 12.
majormajor
strain (1)
and major
(2) evolution
on E-geometry
specimen.on E-geometry specimen.
Figure
12.TheThe
strain
(1) strain
and rate
major
strain rate
(2) evolution
27
In addition to the surface strains shown in previous figures, side view of the
recognized surface show clearly how the diffuse necking has evolved during the
surface test.
strains
inand
previous
figures,
view of
the recognized
The shown
curvature
evolution
of strainside
localization
is shown
in Figure 13.surface
on to the
show
how the diffuse necking has evolved during the test. The curvature and evolution of strain
ion is shown in Figure 13.
13. Curvature of the

specimen before specimen
failure.
Figure Figure
13. Curvature
ofE-geometry
the E-geometry
before failure.
ress Curve Based on FLC-test A-geometry
30
Flow Stress Curve Based on FLC-test A-geometry

Normally, the flow stress is determined by hydraulically bulging of the specimen
out of plane in a test rig. At the same time surface strains are obtained with
non-contacting optical measuring system. The forming pressure is recorded
by a pressure gauge. The momentary curvature of the surface is required to
complete the equation for the flow stress. The curvature is acquired by using
an analytical best-fit circle on the recognized surface. With the GOM ARAMIS
system, the flow stress is obtained by executing analysis macro with some user
interaction. A liquid forming medium is normally used for the test in order to
eliminate the effect of friction that would normally occur when using solid,
contacting punches and dies. (GOM gmbh 2008)
The FLC-tests described here were done with complex lubrication system which
offers nearly frictionless conditions during the test. Therefore, the usability of
full circle, A-geometry specimen with surface strains, curvature and punch
force data is evaluated for flow stress estimation. With aforementioned analysis
macro and the relation of bulge pressure and punch force from previous tests,
a flow stress curve was obtained. The flow stress is shown in Figure 14 along
with other stress strain curves obtained by the author in previous work.
(Joutsenvaara 2012)
DISCUSSION AND CONCLUSIONS

In this paper, the forming limit diagram was constructed for the tested structural
steel with the nominal yield strength of 650 N/mm2 and 3 mm in thickness. The
forming limit curve was made according to the current standard SFS-EN ISO
12004-2 with Erichsen Universal Sheet Metal Testing Machine, model 145-60.
The test parameters were given in order to ensure similar conditions for future
tests.
Figure 14. Comparison of various stress-strain curves obtained by the author.
Figure 14. Comparison of various stress-strain curves obtained by the author.
The strain values on the surface of the A-geometry were recorded as well as the
strain rate during the whole test phase. To further clarify the strain evolution
during the test phase, selected graphs and images were presented. In order to
add comparison to a standardized tensile test, the strains on the tensile test
and E-geometry specimen surface were presented. The strain rate is roughly
three times higher in FLC E-geometry specimen compared to the tensile test
specimen. In the case of the A-geometry, the major strain rate is roughly twice as
high as in the tensile test specimen. For estimation purposes, the higher strain
rates should indicate higher stresses and lower ultimate elongation. Based on
the results presented here, the comparison of different material tests based on
their strain rates is one of the future research topics. This requires also the
evaluation for different kind of real forming processes with real products.
ACKNOWLEDGEMENTS
The author would like to acknowledge the financial support of Tekes the
Finnish Funding Agency for Technology and Innovation for their support
for the METNET Network and Osaamista ajoneuvoteollisuuden kanssa
ConceptCar and MineSteel -projects. Research equipment mentioned in this
paper at Kemi-Tornio University of Applied Sciences is partly funded by the
Regional Council of Lapland (European Regional Development Fund).
REFERENCES
Banabic D., Barlat F., Cazacu O., Kuwabara T., (2010) Advances in anisotropy
and formability, International Journal of Material Forming , vol. 3, no. 3, pp.
165189.
Ghadbeigia H., Pinnaa C., Celottob S., Yatesa J.R., (2010). Local plastic
strain evolution in a high strength dual-phase steel, Materials Science and
Engineering, Vol. 527 pp. 50265032.
Gom Hardware manual, (2008). Gom gmbh.
Gom Software manual, (2008). Gom gmbh.
Joutsenvaara, Jukka, (2012). Investigation of strain localization of high strength
steel with the help of digital image correlation, Proceedings of the METNET
Seminar 2012 in Izmir, pp. 1930.
Kim Hyoung Seop, Kim Sung Ho and Ryu Woo-Seog, (2005). Finite Element
Analysis of the Onset of Necking and the Post-Necking Behaviour during
Uniaxial Tensile Testing, Materials Transactions, Vol. 46 pp. 21592163.
Proulx, Tom (Ed.), (2013). Application of Imaging Techniques to Mechanics of
Materials and Structures, Volume 4.
Vierel Raimo, (2012). Edge formability research for hot-rolled steel Optim 650
MC.
31
32
Experimental Techniques
33
34
MONITORING OF STEEL RAILWAY BRIDGE AND

CONTINUOUS WELDED RAIL
Monitoring of steel railwayPavel
bridge
Ryjekand continuous welded rail
Thakurova 7, Prague 6, Czech Republic
Pavel Ryjek

Thakurova 7, Prague 6, Czech Republic
Miroslav Vok
MiroslavVok
Czech Technical University
in Prague, Klokner Institute
Czech Technical
in Czech
Prague,
Klokner Institute
olnovaUniversity
7, Prague 6,
Republic
olnova 7, Prague 6, Czech Republic
Continuous welded rail (CWR) is used as standard on almost all modern

railway bridges. However, this results in the interaction between CWR and the
bridge
structure,
generating
not only
longitudinal
onall
themodern
substructure,
Continuous
welded
rail (CWR)
is used
as standard
on forces
almost
railway bridges.
significant
additional stresses
in CWR
the rail.
Unfortunately,
the input generating
However, but
this also
results
in the interaction
between
and
the bridge structure,
parameters for
performing
analysis
easily accessible
and stresses in
not only longitudinal
forces
on the numerical
substructure,
but are
alsonot
significant
additional
they vary significantly
in different
sources.
of that,
extensiveanalysis
longthe rail. Unfortunately,
the input
parameters
forBecause
performing
numerical
are not
time monitoring
of the
rail-bridge
interaction
was performed
on aBecause
bridge
easily accessible
and they
vary
significantly
in different
sources.
of that,
the unusual
combination
ballast track,
direct fastening
system andon a bridge
extensive with
long-time
monitoring
of theofrail-bridge
interaction
was performed
barrier
in
the
ballast.
The
paper
presents
the
results
of
the
monitoring
and the
with the unusual combination of ballast track, direct fastening system and
barrier in the
analysis and evaluation of formulas, defining the fastening nonlinear stiffness
ballast. The paper presents the results of the monitoring and the analysis and evaluation of
of track to bridge connection.
formulas, defining the fastening nonlinear stiffness of track to bridge connection.

The old railway bridge in Koln with a clearance of 4.37 m was, for a long time,
The old railway bridge in Koln with a clearance of 4.37 m was, for a long time, a critical
a critical point on the Elbe water way. In order to increase the clearance to
point on the
Elbe water way. In order to increase the clearance to 5.25 m, and in the
5.25 m, and in the future up to 7.0 m, the bridge was completely refurbished
future up to
7.0
m, the bridge was completely refurbished in 2010.
in 2010.
Figure 1. The bridge elevation.
Fig. 1 The bridge elevation
The new bridge consists of the fixed part, made as truss steel structure with the ballasted
The potential
new bridge
of the
fixedfastening.
part, madeThe
as truss
steelfor
structure
with fastening
bed, and the
liftconsists
span with
direct
reason
using direct
the
ballasted
bed,
and
the
potential
lift
span
with
direct
fastening.
The
reason
is the very limited clearance. The superstructure consists of four simply supported spans
for using direct fastening is the very limited clearance. The superstructure
of 32+49+28+20
m. Building height of the fixed part is 1500 mm with a lift span of
consists of four simply supported spans of 32+49+28+20 m. Building height of
1000 mm (see Fig. 1).
the fixed part is 1500 mm with a lift span of 1000 mm (see Fig. 1).
One of the most difficult problems during the bridge design was caused by the continuous
welded rail. It was influenced by a number of complications:
the small rail radius of the curvature, which resulted in the use of sleeper anchors;
the combination of the ballasted bed and direct fastening;
35
One of the most difficult problems during the bridge design was caused by the
continuous welded rail. It was influenced by a number of complications:
the small rail radius of the curvature, which resulted in the use of
sleeper anchors;
the combination of the ballasted bed and direct fastening;
solving the transition from ballasted bed to direct fastening, including
the barrier in the ballasted bed (see Fig. 2);
long dilatation length, which had exceeded the standard limits.
The unusual solution of the CWR resulted in the necessity of establishing the
long term monitoring of the CWR behaviour, including its interaction with
the bridge. The monitoring started in summer 2010 and finished in October
2011. Thus, one full year was measured, including summer hot and winter
The unusual
of the CWR resulted in the necessity of establishing the long term
cold solution
temperatures.
monitoring of the CWR behaviour, including its interaction with the bridge. The monitoring
started in The
summer
2010
and
in October
2011.
Thus, one
full year
was measured,
rail stress
level
wasfinished
established
for different
temperatures.
Extreme
values
achieved
overwinter
Pier 1, where
the dilatation length is extreme. Interesting
including were
summer
hot and
cold temperatures.
result is that the stress in the ballasted bed is higher than stresses reached
The ratio of the change in the axial force in the

track dNP1, S2 to the bearing translation ub, P1 [kN]
on the direct
It can be for
sometimes
observed,
that the stiffness
The rail stress
level fastening.
was established
different
temperatures.
Extremeofvalues were
ballast
the dilatation
temperature.
That is
dependence
be significant,
achieved the
over
Pier depends
1, whereonthe
length
extreme. can
Interesting
result is that the
whenballasted
the waterbed
in the
and increase
its stiffness.
Todirect
check fastening.
that
stress in the
is gravel
higherfreezes
than stresses
reached
on the
It can
phenomenon
on
Koln
bridge,
a
total
of
200
days
were
chosen
in
order
to
be sometimes observed, that the stiffness of the ballast depends on the temperature.
That
evaluate
of temperature
the longitudinal
resistance,
showing
dependence
can the
beinfluence
significant,
when theonwater
in the gravel
freezes
and increase its
of the bridge temperature
had occurred.
stiffness. significant
To check change
that phenomenon
on Koln bridge,
a total of 200 days were chosen in
order to evaluate the influence of temperature on the longitudinal resistance, showing
significant change of the bridge temperature had occurred.
Fig.Figure
2. The
ratio between the change in the axial force increment N
to the
2. The ratio between the change in the axial force increment NP1,S2 to the displacement ub,P1, P1,S2
displacement
utemperature
b,P1, compared
compared
to the mean day
Tm,meanP1. to the mean day temperature Tm,meanP1
It is clear from Fig.2, that the stiffness of the ballast is almost independent on the
temperature. It means that the freezing of the ballast did not occur on that bridge, although
a lot of snow was noticed during the monitoring.
36
It is clear from Fig. 2, that the stiffness of the ballast is almost independent on
the temperature. It means that the freezing of the ballast did not occur on that
bridge, although a lot of snow was noticed during the monitoring.
The 3D FEM computer model in software Scia Engineer 2008 was created, in
order to verify the experimental data and evaluate the longitudinal stiffness
and resistance of the track. The model included the bridge structure and the
adjacent 100m of the rail and substructure. The connection between the rail
and the bridge was modelled with fictive stiff members with nonlinear stiffness
in all directions, representing the vertical, transversal and longitudinal rail
stiffness.
The relation between longitudinal resistance and the track/bridge displacement
was established for the ballasted bed and also the direct fastening. The results
were then compared with the experimental data. The longitudinal resistance
function is shown in Fig. 3.
Fig.3
between
longitudinal
track on
resistance
on relative
displacement
FigureThe
3. Therelation
relation between
longitudinal
track resistance
relative displacement
for ballasted
track and for
ballasted
track
and
direct
fastening
direct fastening.
The following
conclusions
emerged
from the
monitoring
and the analysis
results: of
The following
conclusions
emerged
from
the monitoring
and theofanalysis
the coefficient of the equivalent thermal expansion were higher, than can be found in
results:
the literature,
the functions of the longitudinal track resistance were evaluated, both for ballasted
thedirect
coefficient
of the equivalent thermal expansion were higher, than
track
and
fastening,
can no
be signs
foundofinthe
thetemperature
literature, effect on the ballast stiffness,
there were
no significant impact of the ballast barrier on rail stresses was observed,
the extreme
stressesofinthe
thelongitudinal
rail were found
on the direct
fastening, both
but in the
the functions
tracknot
resistance
were evaluated,
ballast. for ballasted track and direct fastening,
Research reported
in this
paperof was
supported by
Competence
Centres
program of
there were
no signs
the temperature
effect
on the ballast
stiffness,
Technology Agency of the Czech Republic (TA CR), project Centre for Effective and
Sustainable Transport Infrastructure (no. TE01020168) and by TA CR 02031453.
no significant impact of the ballast barrier on rail stresses was observed,
the extreme stresses in the rail were found not on the direct fastening,
but in the ballast.
Research reported in this paper was supported by Competence Centres
program of Technology Agency of the Czech Republic (TA CR), project Centre
for Effective and Sustainable Transport Infrastructure (no. TE01020168) and
by TA CR 02031453.
LASER SYNTHESIS TECHNOLOGY FOR THE CREATION,

RESTORATION AND REPAIR
OF COMPLEX MACHINE PARTS
P.A., Kuznetsov, O.V. Vasilyeva, I.V. Kudryavtseva and A.V. Tereschenko
Department of Nano-structured Materials
FSUE CRISM Prometey, 191017 Saint-Petersburg, Russia
ABSTRACT
Creation of new products and restoring worn items by traditional process is quite hard
to perform and is expensive both in material and energy consumption, as also on time
resources. The main objective of this study is to establish a technique and conduct repair
of intricate parts with high surface hardness (up to 5060 HRC). As a result, after the
material selection, different brands of powders based on Fe, Ni, Co-Cr were tested. To
find the optimum sintering and welding conditions in the process the following parameters
were varied: laser power, the depth of sintered layers and the time of exposure to laser
radiation. After selecting optimal operating system, restoration of a number of details was
carried out. Results of hardness measurement of restored parts showed the correct level
(about 5355 HRC). The experimental results demonstrate that the technology of selective
laser sintering and laser cladding allows obtaining products of a complex dimensional
form with micro hardness parameters that approach the value characteristic of the products
obtained by traditional methods.
METHODS AND MATERIALS

At the present time global market demands new approaches and technologies
which allow creation of high quality goods and to reduce energy and material
consumption. One of the most important tasks at present is the problem of
hardening and recovery of heavy-duty parts and products in various fields of
engineering, including arctic and agricultural machinery. For these purposes
use is made of different technologies of cladding and surfacing. There is a
number of cladding and surfacing technologies which may be used in different
ways of energy supply to the base metal and using different powder materials.
Most of the technologies have relatively low deposition locality energy input,
resulting in a heat affected zone of great thickness. This, in turn, causes a
significant reduction in the properties of the hardened layer. In addition, the
hardened layer is subjected by internal stresses, resulting in deformities of the
work pieces. The above mentioned causes significantly restrict the use of such
surfacing technologies as plasma, arc and flame.
Today, technologies of producing complex parts and functional coatings from
metal powder materials by laser radiation are developing rapidly. In order to
37
38
obtain functional coatings, creation and restoration of parts and products of

complex geometry, including those damaged during operation, the technology
of laser cladding and sintering of bulk objects is unique, allowing achievement
of required operational characteristics, defined geometric dimensions of parts
and coatings, as well as the opportunity to receive new composite materials,
even from immiscible components. Relating to parts recovery technology
and coatings processes technology from metal powders, laser bulk cladding
is considered as a promising method. One such method is the LENS (Laser
Engineered Net Shaping) technology.
In the
schematic
laser cladding
process,
a high-power
In the schematic bulk
laser
claddingbulk
process,
a high-power
laser
beam (Fig.laser
1) isbeam
focused on a metal
(Fig. 1) is focused on a metal surface that leads to formation of a local
surface that leads to formation of a local microscopic liquid melt bath. Inert carrier gas argon is
microscopic liquid melt bath. Inert carrier gas argon is injected into the melt
injected into the melt
bath portion
of the
metal
powder.
After
laserthebeam
the molten
bath portion
of the metal
powder.
After
laser beam
exposure,
moltenexposure,
metal
metal immediatelyimmediately
solidifies, and
by
the
addition
of
the
powder
into
the
molten
bath
the
part
surface
solidifies, and by the addition of the powder into the molten
bath
the
part
surface
thickness
is
increased.
Thus,
as
a
result
of
a
systematic
thickness is increased. Thus, as a result of a systematic scanning of the surface of the part by a laser
scanninginjection
of the surface
the part by
beamfirst
with and
simultaneous
injection
beam with simultaneous
of ofpowder
area laser
formed
subsequent
layers of the part
of powder are formed first and subsequent layers of the part which is being
which is being created. This method allows the use of almost any metal powder and alloy powder
created. This method allows the use of almost any metal powder and alloy
that can be melted powder
by a laser
thatbeam
can bewithout
melted byevaporation.
a laser beam without evaporation.
Figure 1. Scheme of LENS technology.
Due to ultrafast crystallization of the melting bath the building part has ultra-
Due to ultrafast crystallization

of the melting
the building
part
has ultra-fine-grained or
fine-grained or amorphous
structure.bath
According
to the results
of comparative
amorphous structure.
to theproperties,
results ofthese
comparative
testscases
on its
mechanical
properties,
tests According
on its mechanical
parts in some
become
far
superior
the partsfar
obtained
fromthan
similar
alloys
by traditional
these parts in some
casesthan
become
superior
the
parts
obtained production
from similar alloys by
methods
(casting,
stamping).
traditional production
methods
(casting,
stamping).
Table 1. Chemical composition of powders

Due to ultrafast crystallization of the melting bath the building
part
ultra-fine-grained
or
Proceedings
of thehas
METNET
Seminar 2013 in Lule
amorphous structure. According to the results of comparative tests on its mechanical properties,
these parts in some cases become far superior than the parts obtained from similar alloys by
traditional production methods (casting, stamping).
Table 1. Chemical composition
of powders.
Table
1. Chemical composition of powders
Element
R65
1560
PR-H30SRNDYU
Ni
Fe
Cr
B
Al
Si
V
W
Mo
Cu
Mn
80
3,5
1,7
7
6
-
73,3
3,7
14,8
3,1
4,3
0,8
-
1,46
55,1
32
1,9
0,43
3,6
4,9
0,48
1,2
For experiments on protective coatings deposition by bulk laser cladding, the

following powders were selected: R6M5, 1560 Hganas, PR-H30SRNDYU
JSC Polema. The chemical composition of the powders was investigated by
the X-ray fluorescence analysis and results are presented in Table 1. Plates
For experiments
protective
coatings
deposition
bulk
laser cladding, the following powders
were madeonfrom
steel 20
and were
used asbythe
substrates.
were selected: R6M5, 1560 Hganas, PR-H30SRNDYU JSC Polema. The chemical composition
For
experiments
on protective
coatings
by bulk laser
cladding,
the following
powdersin
of the
powders was
investigated
by thedeposition
X-ray fluorescence
analysis
and results
are presented
were
1560from
Hganas,
PR-H30SRNDYU
Tableselected:
1. PlatesR6M5,
were made
steel 20
and were used asJSC
the Polema.
substrates. The chemical composition
of the powders was investigated by the X-ray fluorescence analysis and results are presented in
Table 1. Plates were made from steel 20 and were used as the substrates.
Table 2. Powders cladding conditions.
Table 2. Powders cladding conditions
Powder
R65
Powder
1560
R65
PR-H30SRNDYU
1560
PR-H30SRNDYU
Table
2. Powders
claddingscanning
conditions Feed rate, g/s
Laser
power,
W Substrate
speed, m/s
Laser power,
scanning Feed 0,23
rate, g/s
400 W Substrate
0,013
speed,
m/s
320
0,015
0,18
400
0,013
0,23
350
0,015
0,20
320
0,015
0,18
350
0,015
0,20
Hatch,
mm
Hatch,
0,23
mm
0,23
0,23
0,23
0,23
0,23
Table 3. Results of Rockwell hardness measurements.

Tablehardness
3. Results
of Rockwell hardness measurements.
Table 3. Results of Rockwell
measurements.
Powder
R65
1560
PR-H30SRNDYU
measurement Powder
R65
1560
PR-H30SRNDYU
1
61,6
65,1
67,4
measurement
2
61,8
64,6
68,6
13
61,6
65,1
67,4
61,4
65,9
67,9
2
61,8
64,6
68,6
Average
61,6
65,2
68,0
3
61,4
65,9
67,9
Average
61,6
65,2
68,0
RESULTS AND DISCUSSION
RESULTS
DISCUSSION
Selection ofAND
optimum
modes is to choose the laser power, substrate scanning speed, powder feeding
rate, and the distance between the laser beam passes. Based on the experience of working with
Selection
of optimum
modes ispowders,
to chooseLENS
the laser
power,
substrate
speed,
feedingin
similar chemical
composition
main
parameters
forscanning
each powder
arepowder
summarized
rate,
and
the
distance
between
the
laser
beam
passes.
Based
on
the
experience
of
working
with
Table 2. The results of Rockwell hardness measurements are summarized in Table 3. SEM graphs
similar
chemicalprepared
composition
powders,direction
LENS main
parameters
each 2-4.
powder are summarized in
of the coatings
in transverse
are presented
onfor
Figures
Table 2. The results of Rockwell hardness measurements are summarized in Table 3. SEM graphs
of the coatings prepared in transverse direction are presented on Figures 2-4.
39
2
3
40
Average
61,8
61,4
61,6
64,6
65,9
65,2
68,6
67,9
68,0
LTS AND DISCUSSION

RESULTS AND DISCUSSION
Selection of optimum modes is to choose the laser power, substrate scanning
on of optimum modes
is to choose
the and
laser
speed, powder feedin
speed, powder
feeding rate,
thepower,
distance substrate
between the scanning
laser beam passes.
Based on the
of working
with
similaron
chemical
composition of working wit
nd the distance between
theexperience
laser beam
passes.
Based
the experience
powders,
LENS
main
parameters
for
each
powder
are
summarized
in Table 2.
chemical composition powders, LENS main parameters for each powder
are summarized i
The results of Rockwell hardness measurements are summarized in Table 3.
2. The results of SEM
Rockwell
are direction
summarized
in Table
graphs ofhardness
the coatingsmeasurements
prepared in transverse
are presented
on 3. SEM graph
Figures
24.
coatings prepared in transverse direction are presented on Figures 2-4.
gure 3. SEM graphs of the coating with 1560 powder.

Figure 2. SEM graphs of the coating with JSC Polema powder.
Figure 2. SEM graphs of the coating with JSC Polema powder.
Figure 3. SEM graphs of the coating with 1560 powder.
Figure 4. SEM graphs of the coating with R6M5 powder.
gure 4. SEM graphs of the coating with

R6M5
powder.
Figure
3. SEM
graphs of the coating with 1560 powder.
From the obtained results it is clearly seen that the coatings obtained with
powders 1560 and PR-H30SRNDYU have the highest hardness but are fragile
as evidenced by the cracks (Fig. 2 and 3). Their formation is explained as
follows: cladding with the direct feed of powder in the area of laser radiation
leads to the formation of large amount of carbide and boride compounds in
the structure of coating. It affects the physical and mechanical properties
of coatings. The formation of these compounds due to the presence in the
chemical composition of powders carbide and boride generating elements
confirms the tendency to cracking. Sample obtained from the powder brand
R6M5 has stable hardness at all points of measurement and the absence in the
chemical composition of carbide and boride generating elements reduced the
risk of cracking to a minimum (Fig. 4).
SUMMARY
Analysis of the results leads to the conclusion that it is most appropriate to use
laser cladding powder with low-carbide and boride generating elements (e.g.,
R6M5), giving a high surface hardness and a reduced tendency to cracking.
Laser cladding technology LENS by high coating properties gives the prospect
of increasing the resistance of the most heavily loaded local areas of different
parts of the Arctic and agricultural equipment. In addition, the laser cladding
can be successfully used for repair and rehabilitation of worn out sections of
parts due to the high mechanical properties of the coating and the absence of
deformation.
ACKNOWLEDGMENT
The work was done under a partnership agreement with Lappeenranta
University of Technology Development of materials and technologies for the
Arctic dated 02.01.2012.
REFERENCES
[1]
Shishkovsky I.V. Laser synthesis of functionally gradient mesostructures

and bulk products Moscow: Fizmatlit, 2009 p. 424.
[2]
Tereshchenko A.V., Bobyr V.V., Savin V.I., Kudryavtseva I.V., Vasilyeva

O.V. Restoring a given geometry of machine parts and tools for
technology selective laser sintering // Proceedings 14th International
Scientific-Practical Conference Technology of hardening, coating and
repair: theory and practice 1720 April 2012 Abstracts. St. Petersburg.
Part 1. p. 178181.
[3]
Tereshchenko A.V., Bobyr V.V., Savin V.I., Kuznetsov P.A., Kudryavtseva

I.V. Creation of complex geometry parts by selective laser sintering of
powder materials. // Proceedings of the 9th International Scientific and
Technical Conference Modern metal materials and technologies on
June 2224, 2011. St. Petersburg, 2011.
41
42
Computational Methods
43
44
WEIGHT AND COST OPTIMIZATION

OF WELDED HIGH STRENGTH STEEL BEAMS
Kristo Mela
(kristo.mela@tut.fi)
Markku Heinisuo
(markku.heinisuo@tut.fi)
Department of Civil Engineering
P.O.Box 600, FI-33101 Tampere, Finland
ABSTRACT
Weight and cost optimization of welded beams in steel grades S355, S500 and S700 is
considered in the paper. The aim is to study the utilization of high strength steels in welded
geometrically double-symmetric I-beams by optimizing simply supported beams under
uniform lateral load. The homogenous S355 prole works as the reference case. For higher
steel grades, S500 and S700 are available. Shear and bending resistance are included,
but lateral torsional buckling is neglected. Hybrid beams are are also studied. In a hybrid
cross-section, the steel grades of the anges are higher than for the web. Cross-section
class may vary from 14 following Eurocodes. The resistance rules of Eurocodes at room
temperature are used as constraints in the problem when they are available. For hybrid
cross-sections rules available in the literature are used to check the resistances. Weight
and cost are used separately as optimisation criteria, such that each considered beam is
optimized both for weight and cost. The geometries of the members are extremely simple
and do not include any ttings or other features for joints. The costs include fabrication,
transportation and erecting on site. The fabrication costs include materials, ame cutting,
welding, sawing and painting with all costs needed in the workshop, such as real estate,
energy, labor etc. The optimization is performed by the Particle Swarm Optimization
method. The design variables are the cross-section dimensions of the members in the
range which is typically used in practice, meaning discrete values. The results show that
while signicant weight reductions can be achieved by HSS, the cost savings remain
moderate.
INTRODUCTION
High strength steels (HSS) are currently used in other applications than
buildings. Nevertheless, HSS may be used cost effectively in buildings, as well.
Proper information on the economy of HSS in building applications including
not only material costs but also other costs, such as fabrication, transport and
erecting on site is missing. The purpose of this paper is to provide insight on
how to utilize the higher strength of available steels. This question is answered
by nding the weight and cost optimal solutions for geometrically doublesymmetric welded I-beam that is used as an example structure. By employing
optimization, the full potential of HSS can be discovered, as the solutions can
be totally different from conventional designs.
For cost calculations, the method proposed in (Haapio, 2012) is used along
with typical Finnish price levels. As publicly available experimental and
measured data of fabrication cost is scarce, relative material cost data for
different steel grades based on (Johansson, 2005) and cost factors for welding
and sawing with relation to fabrication of S355 steel members are employed.
The beams made of S355 are used as a reference to which the solutions
obtained by HSS are compared. S500 and S700 are chosen as the HSS grades.
Hybrid beams are also considered, i.e. the anges can be of different grade
than the web. Furthermore, the top ange can be of different grade than the
bottom ange.
The paper is organized as follows. First, the minimum weight and minimum
cost optimization problems are formulated. This part consists of writing the
expressions for the objective functions and introducing the constraints as well
as the design variables. Then, the results of optimization are presented with
discussion on the characteristics of the optimum designs. The paper ends with
conclusions drawn from the study.
PROBLEM FORMULATION
The geometry of a simply supported beam under uniform later load considered
for weight and cost minimization is shown in Figure 1. The beam is optimized
for several loads and spans. The cross-section is geometrically doublesymmetric I-prole.
Formulating an optimization problem requires identifying the design
variables, the objective function and the constraints. In the present study, the
design variables are the dimensions of the cross-section. The design variables
are collected into a vector, denoted by (see Figure 1)
={ }
(1)
T = {5, 6, 8, 10, 12, 14, 15, 16, 18, 20, 22, 25, 30, 35, 40, 50, 60, 80, 100} [mm]
(2)
Variables and can take values from the discrete set T, which contains the
following 19 elements:
The ange width and web height are to be chosen from the discrete set of
integer values ranging from 100 mm to 1000 mm in 10 mm intervals. That is
, ={100, 110, 120, , 990, 1000}.
45
The ange width and web height are to be chosen from the discrete set of integer values rangin
46
from 100 mm to 1000 mm in 10 mm intervals. That is "#, &*={100, 110, 120, , 990, 1000}
.
Figure 1.Simply supported

beam to be optimized and the geometrically double-symmetric cross
Figure 1. Simply supported beam to be optimized and the geometrically double-symmetric cross-section.
Note yield
that the yield
strengths of the
not be equal,
i.e.
allowed. i.e. #3$ #3" is allowed.
section. Note that the
strengths
ofanges
the need
anges
need
not be isequal,
Objective functions
In this study, both the weight and fabrication cost are chosen as objective
functions. The weight is written as
= 2+=2+
47
(3)
where = 7850 kg/m3 is the density of steel and is the length of the beam.
The manufacturing cost function is based on the work of (Haapio, 2012). The
manufacturing process is divided into cost centers, each containing a set of cost
components such as labour, material, equipment, energy, and real estate. The
main task is then to devise appropriate expressions for the cost components of
every cost center. These expressions depend on the specic technologies used,
and they contain many parameters that must be obtained by measurement.
Unless otherwise stated, the default values found in (Haapio, 2012) are used
for the parameters appearing in the cost function. Here, the parameter values
are combined to single numbers, when possible.
The cost function can be written as
=+++++++()
(4)
=(,+,+,)
(5)
The material cost is
where =0.7 /kg, and , and are the cost factor of the steel grade and
the cross-sectional area of plate . For S355, , = 1.0, for S500, , = 1.15
and for S700, ,= 1.30.
The blasting cost depends on the number of plates and the length of the beam.
It is
=33.6447104
(6)
47
The cutting cost can be expressed as

=1.3227++(+ )
(7)
=()8.92122486.87+8115.8
(8)
where the non-productive time =3.0 min. Depending on the plate

thickness, either flame cutting or plasma cutting is used. For plate thickness
up to 30 mm, plasma cutting is used. The productive time is
=<9F!
=79F
)4.1939
where is the
thickness
of(!
the
plate. The$+658.67
cost of cutting consumables is
=0.38 /min and the cost of energy is =0.12 /min.
?9F! =0.22+4.18(1105$2+0.001$+0.0224)
(9)
(10)
For flame cutting, productive time and cost of consumables are
whereas the torch's energy consumption is neglected, i.e. ?>G9F=0.

=()4.1939+658.67
(9)
The length of the

cut is 79F!=2("#+7 ) for the flanges
and 79F!=2(&+7
) for the web.
=0.22+4.18(11052+0.001+0.0224)

(10)
whereas
the torchsthe
energy
consumption
is neglected,
i.e. =0.
The cost
of welding
plates
to form the
actual beam
can be expressed as
The length of the cut is =2(+) for the flanges and =2(+) for
the =1.3627
web.
95!
=E*5+=<*5! +=<*5! (?9*5+ ?>G*5)
(11)
The cost of welding the plates to form the actual beam can be expressed as
where ?9*5=0.08/min and ?>G*5=1.36/min. Furthermore, =E*5=6.25 min. The

=1.3627++(+
)
(11)
productive time is written
as
where =0.08 /min and =1.36 /min. Furthermore, =6.25
=<*5!
7&O&
214=7.85106
7&9&2$&214
min. The=7.85106
productive time
is written
as
=7.85106214=7.851062214
(12)
(12)
where 7&=7 is the length of the weld, and 9& is the weld size factor that depends on the grade of
is the length
of the weld,
andwelding
is the weld
size flange.
factor that
depends
the web (Table 1).where
The =
expression
of 9&(!)
is for
of one
For
the whole beam, both
on
the
grade
of
the
web
(Table
1).
The
expression
of
()
is
for
welding
of
flanges need their own expressions. In order to take into account the expense one
of welding high
flange. For the whole beam, both flanges need their own expressions. In order
strength
steels, cost factors are introduced. These factors are given in Table 1.
to take into account the expense of welding high strength steels, cost factors
are introduced. These factors are given in Table 1.
Table
1. Cost
parameters
different steel grades.
Table 1. Cost
parameters
for different steelfor
grades.
Grade
S355
S500
S700
9& Welding factor Sawing factor

0.55
1.0
1.0
0.80
1.25
1.15
0.82
1.5
1.3
sawingas
cost is written as
The sawing cost isThe
written
=1.2013++(()+ )
9;! =1.2013=E;+=<;! +=<;! (?9;(! )+ ?>G;)
(13)
(13)
The non-productive time, =E;=4.5+7 /20000 min, and the cost of energy is ?>G;=0.02 /min.
48
The non-productive time, =4.5+/20000 min, and the cost of energy

is =0.02 /min. The productive time depends on the position of the
cross-section, when it is sawn. It is assumed that the geometrically doublesymmetric prole is laid on its side such that the anges are considered to be
sawn vertically and the web horizontally. For taking into account the higher
steel grades, cost factors are introduced to the productive time, see Table 1.
The cost of consumables includes the wear of the saw blade. It depends on the
plate thickness and on the productive time. For details of productive time and
sawing consumables, see (Haapio, 2012, pp. 3942).
The painting cost includes also the cost of drying and it can be written as
=4.17106()+0.3611031103
(14)
where =3+4+22 is the painted area per unit length, and

= is the smallest width dimension of the beam (it is assumed that
).
The transportation cost is expressed as
=0.0106+1.2729, if ()/()2644105+4.8103,
otherwise
(15)
where =2+ is the volume occupied by the beam and () is the

weight of the beam. The conditional definition states that if the weight to volume
ratio is below the given limit, then the transportation cost is determined by
the volume of the beam. Otherwise, the cost will be determined by the weight.
Finally, the erecting cost of the beam can be written as
=+
(16)
=30000+27+20.50.42+36
(17)
where =3.1 /min is the cost of labor, =1.3460 /min, and =0.36.
The time needed for erecting the beam is
where =15 m is the distance from the lifting area to the final position, and
=6 is the number of bolts per joint. The erecting time consists of moving the
main lift from one end to the other, lifting the beam, joining the beam to the
main structure, and returning the hook to the lifting area. It is assumed that
five workers are included in the erecting process, and that the beam is lifted
by a crane with lifting capacity of 25 tons. It is assumed that the beam is lifted
to 10 m. This is reflected in the equipment cost. Note that the erecting cost
depends only on the length of the beam, as the number of bolts per joint and
the capacity of the crane are fixed.
Constraints
The constraints of the optimization problems are derived from the Eurocodes
(EN 199311, 2005; EN 199315, 2006). They ensure that the solution has
sufcient moment and shear force resistance, and that the top ange does not
buckle against the web. Finally, the displacement of the beam is limited.
The moment resistance constraint is
()
(18)
If the cross-section belongs to Class 1 or 2, plastic design is applied and

is the plastic moment of the cross-section. If the web is in Class 3 and the
top flange in Class 1 or 2, then the plastic moment of the cross-section with
effective web is used. For Class 3, elastic bending resistance is used and for
Class 4, effective cross-section is determined.
For elastic design of hybrid cross-sections, the partial plastication of the web
is taken into account. In general, the minimum of the yield strengths of the
anges determines the maximum stress that can appear in the cross-section.
For a linear stress distribution, it is common that the yield strength of the
web is attained, and it will plastify in part. By allowing the web to plastify,
the higher steel grades can be utilized better. For determining the plastied
region of the web, it is assumed that the neutral axis is located at the centroid
of the effective cross-section. This assumption leads to a small error in the
force equilibrium of the cross-section, but it is considered negligible. Then,
the location where the yield stress in the web is reached is determined from
the linear stress distribution. The part of the web above this point yields. See
(Ongelin and Valkonen, 2010, pp. 130140) for more details.
The shear resistance constraint is
,()
(19)
Plastic shear resistance is employed and shear buckling is taken into account.
To prevent the top ange from buckling in the plane of the web, the following
constraint is needed (EN 199315, Clause 8(1)):
,
(20)
where =210000 MPa, = is the area of the web, , is the effective

area of the top flange, and =0.4 if plastic resistance if used and =0.55 if
elastic resistance is used.
49
50
The maximum displacement occurs at the midpoint of the span. The

displacement constraint is written as
()
(21)
=53844()
(22)
where
is the displacement in the middle of the span and =/200 is the maximum
allowed displacement. The load in service limit state, =0.75.
Problem statement
The developments shown above can be combined to the following optimization
problem
minsuch
that101,0,0()
10, ,
(23)
where is either the weight, or the cost . Note that the constraints have
been normalized and all terms are transferred to the right.
Both the minimum weight and minimum cost problems are nonlinear
discrete optimization problems. With only four design variables and four
constraints, the problems are small-scale, but the mathematical properties
of the constraints makes finding the global optimum rather difficult. For
example, when a cross-section is altered such that it goes from belonging to
Class 2 to Class 3, the expression for moment resistance changes considerably.
From a mathematical point of view, the moment resistance constraint is in
general discontinuous in the regions where the cross-section class changes.
This makes it difficult to apply classical methods of nonlinear discrete
optimization that require continuity of the constraint and objective functions.
On the other hand, heuristic, population-based algorithms can handle such
problems. In this study, such an algorithm is used to solve the optimization
problems at hand.
RESULTS OF OPTIMIZATION
Both the weight minimization and minimum cost problems were solved by the
Particle Swarm Optimisation (PSO) algorithm (Kennedy and Eberhart, 1995;
Poli et al., 2007). In the literature, many possibilities regarding the details
of the algorithm are available. In this study, an implementation of the PSO
method was produced, with chosen constraint handling mechanism, velocity
update rules with inertia term, craziness effect and elite particle. Furthermore,
whenever the best feasible solution was updated, a neighborhood search
was performed, i.e. better solutions were searched in a neighborhood of the
Optimisation (PSO) algorithm (Kennedy and Eberhart, 1995; Poli et al., 2007). In the literature,
many possibilities regarding the details of the algorithm are available. In this study, an
the METNET Seminar
2013 in Lulemechanism,
51
implementation of the PSO method was produced, with Proceedings
chosen ofconstraint
handling
velocity update rules with inertia term, craziness effect and elite particle. Furthermore, whenever
the best feasible solution was updated, a neighborhood search was performed, i.e. better solutions
were searched in a neighborhood of the solution. Detailed description of the algorithm is beyond the
solution. Detailed description of the algorithm is beyond the scope of this
scope of this paper.
paper.
Three spans of 6m,

8m,
andof10m
considered
with three
different
loads,loads,
X=20 kN/m, X=60
Three
spans
6 m, 8were
m, and
10 m were considered
with
three different
=20
kN/m,
=60
and =100steel
kN/m.
For the
threeS500,
available
kN/m, and X=100
kN/m.
For
the kN/m,
three available
grades,
S355,
andsteel
S700, altogether 14
grades,
S355,
S500,
and
S700,
altogether
14
combinations
could
be
formed
for the grade of the
combinations could be formed for each span and load. The only restriction was that
each span and load. The only restriction was that the grade of the web cannot
web cannot be greater
thanthan
the the
grade
ofofeither
anges.
be greater
grade
eitherof
of the
the anges.
Weight minimization
Weight minimization
As a rst step, it is interesting to compare the homogenous S500 and S700 cross-sections with the
As a rst step, it is interesting to compare the homogenous S500 and S700
homogenous S355cross-sections
prole, which
as the reference
case.
In works
Tableas2,thethe
ratios of the minimum
withworks
the homogenous
S355 prole,
which
reference
weights of S500 to
S355
and
S700
to
S355
are
shown
for
different
loads
and
spans. It can be seen
case. In Table 2, the ratios of the minimum weights of S500 to S355 and S700
to
S355
are
shown
for
different
loads
and
spans.
It
can
be
seen
that
for
that for the lowest load, X =20 kN/m, only marginal weight savings can bethe
achieved by HSS.
lowest load,
=20
kN/m, only
marginal
weight savings
cannot
be favourable
achieved by to HSS. On the
This is partly explained
by the
allowable
plate
dimensions,
that are
HSS. This is partly explained by the allowable plate dimensions, that are not
other hand, the HSS
solutions
areOn
mostly
restricted
byHSS
thesolutions
displacement
constraint
favourable
to HSS.
the other
hand, the
are mostly
restrictedwhere the higher
strength does notby play
any
role.
For
the
S355
solution,
the
moment
constraint is
the displacement constraint where the higher strength does not resistance
play any
role. For
S355 solution,
resistance constraint
dominant, even though
thetheutilization
ratiothe
of moment
the displacement
constraintisisdominant,
also more than 90%.
even though the utilization ratio of the displacement constraint is also more
than 90%.
For higher loads, weight savings provided by higher strength gets more signicant. For the
intermediate loadFor
X=60
S500 gives
saving of
about strength
15%, whereas
S700 reduces the
higherkN/m,
loads, weight
savingsa provided
by higher
gets more
weight by about 25%.
For the
load, and
largest
span,S500
the weight
reductions
signicant.
For largest
the intermediate
load
=60 kN/m,
gives a saving
of aboutare 23% for S500
15%, whereas S700 reduces the weight by about 25%. For the largest load, and
and 35% for S700.
largest span, the weight reductions are 23% for S500 and 35% for S700.
Table 2. Results of weight minimization. The numbers are the ratios 5500/5355, and
2. Results
of weight is
minimization.
The numbersweight
are the ratiosofthe
500/homogenous
355, and 700/cross-section
355,
5700/5355, Table
where
5#3
the minimum
with yield
where is the minimum weight of the homogenous cross-section with yield strength .
strength #3.
S500
Span (m)
Load
6
8
10
20
0.98 0.97 0.90
60
0.85 0.84 0.84
100 0.84 0.84 0.77
S700
Span (m)
Load
6
8
10
20
0.98 0.97 0.90
60
0.76 0.74 0.73
100 0.71 0.67 0.65
Cost Minimization
52
52
Cost Minimization
For the results of cost optimization, it is again interesting to compare the
For the results of

cost optimization, it is again interesting to compare the homogenous S500 and
homogenous S500 and S700 solutions with the S355 case. Also, it is interesting
S700
with
the
case.
Also,
isperform
interesting
see,
how the
hybrid
cross-sections
perform
Forsolutions
the results
of
cost
optimization,
it is it
again
interesting
to
compare
homogenous
S500 and
to see,
howS355
hybrid
cross-sections
in costto
optimization.
In Table
3 the
S700
solutions
with
the
S355
case.
Also,
it
is
interesting
to
see,
how
hybrid
cross-sections
perform
in cost optimization.
optima
of compared
S500, S700
and
relative In
costTable
optima3of the
S500,relative
S700 andcost
the best
solution
to S355
are the best solution
in cost to
optimization.
In
Table
theHSS
relative
optima
of S500,
S700
and
the best solution
given.
It can
be
seen
solutions
only
for the
largest
compared
S355
are
given.
It 3that
can
be
seen cost
thatbecome
HSS benecial
solutions
become
benecial
only for the
most
cases,
the
costsof
of
HSS are
too
compared
to load.
S355
given.
can material
be and
seenand
thatfabrication
HSS solutions
become
benecial
for the with
largest
load. In
mostInare
cases,
the Itmaterial
fabrication
costs
HSS
are
toohigh
highonly
compared
compared
with the
S355.
largest
load.
In
most
cases,
material
and
fabrication
costs
of
HSS
are
too
high
compared
with
S355.
S355.
Table 3. Results of cost minimization.
Table 3. Results of cost minimization.
Table 3.
S500
Span
S500
Span(m)
(m)
Load
6
8
10
Load
6
8
10
2020 1.15
1.15 1.14
1.14 1.10
1.10
6060 1.06
1.06 1.03
1.03 1.01
1.01
100
1.04
1.04
0.94
100 1.04 1.04 0.94
Results of cost minimization.
S700
S700
Load
Load
20
20
60
60
100
100
Span(m)
(m)
Best SpanSpan
Span
Best
(m) (m)
Load 6 6 8 8 10 10
66
8 8 1010 Load
1.26 1.25
1.25 1.22
1.22 20 20 1 1 1 10.99 0.99
1.26
1.10 1.06
1.06 1.03
1.03 60 60 0.950.95
1.10
0.930.93
0.90 0.90
1.04
0.96
0.94
100
0.93
0.91
0.90 0.90
1.04 0.96 0.94 100 0.93 0.91
lowest
load,the
theS355
S355solution
solution is
except
for for
the largest
span.span.
Then,Then,
the the
ForFor
thethe
lowest
load,
is the
themost
mosteconomical,
economical,
except
the largest
For
the
lowestof
load,
theand
S355
solution
isS355
the most
economical,
exceptWith
for the
solution
with
both
anges
S500
the
web
of
is
more
economical.
increased
load,
solution with both
anges of
S500 solution
and thewith
webboth
of anges
S355 is
more economical.
With increased load,
largest span.start
Then,tothe
of minimum
S500 and the
web
of S355
the hybrid cross-sections
have a positive
effect
on the
cost.
The
cost savings vary
the hybrid cross-sections
start to have
a positiveload,
effect on the minimum
cost.
The
cost savings vary
is more
With
start
toexcept
from 5% to 10%.
In economical.
all cases, the
topincreased
ange is madetheofhybrid
S700.cross-sections
The web is S355,
for the
from
5%
to
10%.
In
all
cases,
the
top
ange
is
made
of
S700.
The
web
is
S355,
exceptforfor the
have
a
positive
effect
on
the
minimum
cost.
The
cost
savings
vary
from
5%
largest load and the largest span, in which case it is S500. The bottom ange is of S700, except
10%.
In all span,
cases, in
thewhich
top ange
is made of S700. The web is S355, except
largest
loadX=60
and tothe
largest
kN/m,
7=6 m, where
it is ofcase
S355.it is S500. The bottom ange is of S700, except for
the case
for the largest load and the largest span, in which case it is S500. The bottom
the case X=60 kN/m, 7=6 m, where it is of S355.
ange is of S700, except for the case =60 kN/m, =6 m, where it is of S355.
The cost distribution of the minimum cost solution is shown in Figure 2. It can be seen that the
cost is
50%
ofdistribution
the
cost.
Erecting
cost
plays
important
with
Beam
Thematerial
cost distribution
the total
minimum
cost solution
is an
shown
in Figure
2.2.aItIt16%
can share.
be seen
that the
The
costof
of the
minimum
cost
solution
is shown
inrole
Figure
can
welding
and
painting
are
also
relatively
important,
with
11%
share
each.
Sawing,
blasting
and
be seen
material
cost is 50%cost
of the
totalan
cost.
Erecting cost
material cost is 50%
ofthat
the the
total
cost. Erecting
plays
important
roleplays
withan
a 16% share. Beam
important
with
athe
16%
share.
Beam welding
and painting
are also
relatively
transport
very
little
torelatively
total
cost.
welding
andcontribute
painting
arerole
also
important,
with
11% share
each.
Sawing, blasting and
important, with 11% share each. Sawing, blasting and transport contribute
transport contribute
very little to the total cost.
very little to the total cost.
Figure 2.Cost distribution of the minimum cost beam. X=60 Z E/Q, 7 =8 Q
Discussion
Figure 2.Cost distribution of the minimum cost beam. X=60 Z E/Q, 7 =8 Q
Figure 2. Cost distribution of the minimum cost beam. =60 /, .
In the above, the criterion values of the solutions were investigated. It is also interesting to study the
actual designs that lead to the optimum solutions, i.e. to focus on the design space. A common
Discussion
feature of optimum designs is that the width of the ange is kept relatively low. The maximum
Discussion
In the above, the criterion values of the solutions were investigated. It is also
interesting to study the actual designs that lead to the optimum solutions, i.e.
to focus on the design space. A common feature of optimum designs is that the
width of the ange is kept relatively low. The maximum ange width among
the minimum weight solutions is 210 mm. Out of 126 cases, 83 solutions have
ange width of 100 mm, i.e. the minimum value. Similarly, the thickness
of the web is kept close to the minimum value of 5 mm. The maximum web
thickness appearing in the minimum weight solutions is 8 mm and, in 99 of
the 126 problems, the optimum web thickness is 5 mm. Thus, the strength
and stiffness of the cross-section were increased mainly by increasing the web
height and ange thickness. Consequently, the top ange was most often in
Class 1 or 2, whereas the web is in Class 3 or 4. Similar behaviour is observed
for the minimum cost designs.
In 68 of 126 cases, the minimum weight and minimum cost solutions coincide.
There are also cases, where the minimum weight solution is more economical
than the minimum cost solution and vice versa. This indicates, that the PSO
was not able nd the correct minimum cost or minimum weight solution. In
all cases, the differences between the minimum cost and minimum weight
solutions are very small.
On the other hand, for given load and span, the minimum weight and
minimum cost solutions do not coincide in most of the cases, when comparing
the different steel grade combinations. In Figure 3, the minimum cost and
minimum weight solutions for =60 kN/m, and =8 m are depicted in the
criterion space. It can be seen that the minimum cost solution is obtained
in Case 4 (both flanges S700, web S355), and the minimum weight solution
is obtained in Case 6 (flanges and web S700). The cost of the minimum
weight solution is 556.55 , whereas the minimum cost is 488.25 . Thus,
the minimum weight solution is 14% more expensive to manufacture. The
minimum weight is 302.70 kg. The weight of the minimum cost solution is
317.14 kg, which corresponds to 4.7% increase. The solution of Case 5 (flanges
S700, web S500) represents a compromise solution between minimum weight
and minimum cost. The weight is 304.58 kg, which corresponds to only 0.6%
increase from the minimum weight, whereas the cost 529.93 corresponds
to 4.8% decrease in manufacturing cost compared with the minimum weight
design. Such trade-off information is useful, if there is uncertainty in the
parameters of the cost function.
53
4.7% increase. The solution of Case 5 (flanges S700, web S500) represents a compromise solution
between minimum weight and minimum cost. The weight is 304.58 kg, which corresponds to only
0.6% increase
fromofthe
minimum
weight,
Proceedings
the METNET
Seminar 2013
in Lule whereas the cost 529.93 corresponds to 4.8% decrease in
54
manufacturing cost compared with the minimum weight design. Such trade-off information is
useful, if there is uncertainty in the parameters of the cost function.
Figure 3. Minimum cost and minimum weight solutions of the WI-beam. X=60 Z E/Q, 7 =8
Figure 3. Minimum cost and minimum weight solutions of the WI-beam. =60 /, =8 .
Q.
CONCLUSION
CONCLUSION
The examples considered

in the
paper indicate
thatindicate
for geometrically
double double
symmetric WI-beams
The examples
considered
in the paper
that for geometrically
symmetric
under
uniforminlateral
load,
savingsbyin employing high
under uniform lateral
load,WI-beams
substantial
savings
weight
cansubstantial
be achieved
weight can be achieved by employing high strength steel. On the other hand,
for smaller loads, the weight reduction can be relatively small, even marginal.
With current material and fabrication
54 costs, HSS tends to produce more
expensive solutions than the conventional S355 cross-section for small loads.
However, for larger loads, the cost difference is marginal (few percent), and
for high loads and large spans, HSS solutions are more economical. Hybrid
cross-sections can reduce the manufacturing cost up to 10% according to the
case studies of this paper. In order to get real information on the cost savings
provided by HSS it is essential to get measured up-to-date data of all actions
needed in fabrication. The cost data used in this study is taken from (Haapio,
2012) and it is, in a way, mean data available for S355 steel. For HSS, this data
is extrapolated using the given factors for material, welding and sawing costs.
The most economical combination of steel grades depends strongly on the load
and span of the beam. It is normally very difcult to nd the most economical
solution by trial and error. For the designer, optimization provides an efcient
way of obtaining good solutions quickly.
ACKNOWLEDGEMENTS
This study was carried out as part of a European research project RUOSTE,
funded by the Research Fund for Coal and Steel.
REFERENCES
Bernt Johansson. Buckling resistance of structures of high strength steel. In
Hans-Peter Gnther, editor, Use and Application of High-Performance Steel
for Steel Structures, pages 120128. IABSE, 2005.
EN 199311. Eurocode 3: Design of Steel Structures. Part 11: General rules
and rules for buildings. CEN, 2005.
EN 199315. Eurocode 3: Design of Steel Structures. Part 15: Plated
structural elements. CEN, 2006.
Jaakko Haapio. Feature-Based Costing Method for Skeletal Steel Structures
Based on the Process Approach. PhD thesis, Tampere University of
Technology, 2012.
James Kennedy and Russell Eberhart. Particle swarm optimization. In IEEE
International Conference on Neural Networks, pages 19421948, 1995.
Petri Ongelin and Ilkka Valkonen. Welded proles EN 1993-handbook.
Rautaruukki Oyj, 2010. In Finnish.
Riccardo Poli, James Kennedy, and Tim Blackwell. Particle swarm
optimization an overview. Swarm Intelligence, 1:3357, 2007.
55
56
HIGH STRENGTH STEEL IN TUBULAR TRUSSES

Teemu Tiainen
Kristo Mela
Timo Jokinen
Markku Heinisuo
Tampere University of Technology, Tampere, Finland
ABSTRACT
This paper deals with weight and cost optimization of Warren-type welded tubular roof
trusses with and without verticals. Specially, the effect of steel grade is studied. The trusses
are optimized being either S355, S500, S700 or hybrid solutions. Costs are calculated
based on features of the trusses. The starting point is the exact geometry of the truss
from which finite element analysis model is derived. This approach allows the resistance
and other requirements of Eurocodes for both for the members and the joints to be used
as constraints. Design variables are the height of the truss, the locations of the joints,
gaps at the joints and the member sections (cold-formed square tubes). The resulting
mixed-integer optimization problem is solved using the particle swarm optimization (PSO)
algorithm. The results imply a significant saving in weight when using high strength steel.
Cost reduction is smaller but existent. The results motivate the use and further research of
high strength steels in building products.
INTRODUCTION
Tubular steel trusses are widely used in buildings to carry roof loads due
to nice appearance, good economic and load bearing performance. Active
research in the area has produced efficient and reliable analysis methods. The
natural step after analysis is optimization. The three basic approaches found in
structural optimization literature are sizing, shape and topology optimization
[1] and this classication applies to trusses as well. In sizing optimization of
a tubular truss, the design variables are the sections of each member. The
problem can be formulated either continuous or discrete. For continuous
design variables many optimization approaches have been proposed [1, 2].
However, in practical design problems the prole choice has to typically be
made from steel suppliers catalogue resulting in discrete problem.
Truss optimization and cost analysis of steel structures have both been
under active research work (see for example [3, 4, 5]). What the authors nd
missing so far in the literature is a consideration of using high strength steels
(HSS) in tubular trusses. Therefore, the aim of this study is to nd out by
using optimization if the use of HSS in tubular steel trusses results in more
economical solutions than with regular (S355) steel. When comparing trusses
made of different steel grades it is essential to consider the costs, not only
their weights. Steel grade has effect both on the material costs and on the
fabrication costs. Weight affects transport and erection costs thus making it
important as well.
In contrast with typical approaches, the starting point of the analysis is the
parametric geometrical presentation of the truss. Sizing variables as well
as geometrical dimension variables are present leading into mixed-integer
problem. The example structures used are Warren type trusses with two
possible layouts.
leading into mixed-integer problem. The example structures used are Warren type trusses with two
possible layouts.
OPTIMIZATION FORMULATION
2. OPTIMIZATION FORMULATION
In this optimization formulation two types of Warren type tubular trusses

(Fig. 1) are considered with Eurocode based constraints together with weight
In this optimization formulation two types of Warren type tubular trusses (Fig. 1) are considered
and cost as objective functions. Starting point of the optimization is the vector
with Eurocode based constraints together with weight and cost as objective functions. Starting
of design variables (19 or 23 elements) and other parameters that dene the
point of the optimization is the vector of design variables (19 or 23 elements) and other parameters
exact geometry of the truss. From this geometry, a nite element model using
that dene the
exact geometry of the truss. From this geometry, a nite element model using EulerEuler-Bernoulli beam elements is derived. The braces of both models are
Bernoulli beam
elements is derived. The braces of both models are aligned but in the nite element
aligned but in the nite element model brace elements are extended to point
model bracewhere
elements
extended
pointother
where
two intersect.
braces next
each
other
would
intersect. In
two are
braces
next totoeach
would
Into
this
case,
a very
rigid
this case, a very
rigid
eccentricity
element
connects
the
braces
to
chord.
eccentricity element connects the braces to chord.
Figure 1: Two types of trusses considered
Figure 1. Two types of trusses considered.
From
the nite
element
stress
acquired
and are
From the nite
element
model,
stressmodel,
resultants
!,resultants
"# and $%,are
and acquired
they are utilized to
and
they
are
utilized
to
perform
the
resistance
check
of
members
and [6] and
perform the resistance check of members and joints as required in standards EN 1993-1-1
joints
as
required
in
standards
EN
1993-1-1
[6]
and
EN
1993-1-8
[7]
dealing
EN 1993-1-8 [7] dealing with members and joints, respectively. For high strength steel and hybrid
withofmembers
and
Forsome
highcases
strength
steel
hybrid
solutions some
the design
hadjoints,
to be respectively.
extrapolated in
giving
veryand
conservative
results.
solutions
some
of
the
design
had
to
be
extrapolated
in
some
cases
giving
very
In cost calculation, a very general method proposed by Haapio [8] is adopted.
conservative results. In cost calculation, a very general method proposed by
Haapio [8] is adopted.
The optimization problem of structural weight can be written as
The optimization problem of structural weight can be written as

min = =1,
s. t.
min &# = )*+=1-./+0+,2

s. t. #
where, x is where,
the vector
design
variables
including
bothincluding
discrete sizing
variablessizing
and continuous
x isofthe
vector
of design
variables
both discrete
geometry variables
and
is
the
feasible
set;
a
set
of
solutions
that
comply
all
variables and continuous geometry variables and is the feasible set; athe
setconstraints
2+(# )0 derived
from that
Eurocodes.
The
optimization
problem
theEurocodes.
same constraints as
of solutions
comply all
thecost
constraints
()0
derivedhas
from
cost
optimization
formulationThe
1 and
can
be written asproblem has the same constraints as formulation 1 and
can be written as
min &8= 9:$++=1-.(9<+9:+9=)+9>
min = +=1(++)+
s. t. #
s. t.
3. RESULTS
57
58
RESULTS
The optimization problem was solved with particle swarm optimization
method. It is a heuristic method without rigorous mathematical basis and
it usually yields good solutions but sometimes fails badly. The runs were
repeatedproblem
10 to 20
times
butwith
stillparticle
the results
be considered
but
The optimization
was
solved
swarmcannot
optimization
method. optimal
It is a heuristic
best found.
Themathematical
main ndings
canand
be it
seen
in Figure
and 3
which present
the
method without
rigorous
basis
usually
yields 2good
solutions
but sometimes
best
found
weight
and cost
those
valuescannot
of regular
steel (S355)
fails badly.
The
runs were
repeated
10 tocompared
20 times buttostill
the results
be considered
optimal
but best found.
main2,ndings
can be It
seen
Figure
and weight
3 whichreduction
present thebybest
found
for typesThe
1 and
respectively.
caninbe
seen2that
using
weight and
coststrength
comparedsteel
to those
values of regular
steel
(S355)
forcosts
typesare
1 and
2, respectively.
high
is signicant
(20% to
30%).
The
lower
especially It
can be seen
that
weight
reduction
by using high strength steel is signicant (20 % to 30 %). The
with
the
higher
loading.
costs are lower especially with the higher loading.
1,4
1,2
1
Cost (23.5 kN)
0,8
Cost (47 kN)

Mass (23.5 kN)
0,6
Mass (47.0 kN)
0,4
S700/
S500
S700/
S355
S500/
S355
S700
S500
S355
0,2
Figure 2. Type 1 truss results compared to S355/S355 solution.

Figure 2: Type 1 truss results compared to S355/S355 solution
4.
CONCLUSIONS
CONCLUSIONS
The starting
of thepoint
optimization
should be the geometrical
presentation
of thepresentation
truss in order to
Thepoint
starting
of the optimization
should be the
geometrical
check allofrequirements,
as geometrical
of codes andsuch
correct
structural analysis
the truss insuch
order
to check allrules
requirements,
as geometrical
rulesmodel
of
with eccentricities.
In sizingstructural
and shape analysis
optimization
thiswith
will result
in non-linear
mixed-integer
codes and correct
model
eccentricities.
In sizing
and
optimization for which heuristic methods can used. As the problems are very demanding and the
shape optimization this will result in non-linear mixed-integer optimization
method does not include checks for optimality, the optimized results may need extra manual
for which
heuristic
methods
cantoused.
the problems
are very
enhancement.
Weight
reductions
compared
S355 As
trusses
were of around
20 %demanding
for S500 and
and
the
method
does
not
include
checks
for
optimality,
the
optimized
30 % for S700. The cost reduction on the truss type Warren type truss without verticals results
performed
need
manual enhancement.
Weight
reductions
compared
to S355
better, atmay
around
15extra
%. Reductions
for hybrid trusses
were
between these
numbers,
but for
trusses
of around
20% for
and were
30% for
S700.
The cost
reduction
S700/S500
(chordswere
S700,
braces S500)
the S500
reductions
about
the same
as for
S700/S700
trusses. on the truss type Warren type truss without verticals performed better, at
around 15%. Reductions for hybrid trusses were between these numbers,
but for S700/S500 (chords S700, braces S500) the reductions were about the
same as for S700/S700 trusses.
59

1,2
Cost (23.5 kN)
0,8
Cost (47 kN)

0,6
Mass (23.5 kN)
Mass (47.0 kN)
0,4
0,2
S700/S500
S700/S355
S500/S355
S700
S500
S355
Figure 3. Type 2 truss results

compared to S355/S355 solution.
The results demonstrate that the use of HSS is economical in tubular trusses
Figure 3: Type 2 truss results compared to S355/S355 solution
applying the design rules of the present Eurocodes even though some of the
rules penalize high strength steels. These issues should be solved in the
future and
fabrication
costs
of HSS in
tubular
can be reduced
and
The results demonstrate
that the
the use
of HSS is
economical
tubulartrusses
trusses applying
the design
more
savingseven
achieved.
rules of the present
Eurocodes
though some of the rules penalize high strength steels. These
issues should be solved in the future and the fabrication costs of HSS tubular trusses can be reduced
and more savings achieved.
REFERENCES
REFERENCES
[1]
[1]
[2]
[3]
[4]
[5]
[6]
[7]
[8]
Uri Kirsch. Structural optimization. Springer-Verlag, 1993.
Uri Kirsch. Structural optimization. Springer-Verlag, 1993.

[2] Raphael T. Haftka and Zafer Grdal. Elements of structural optimization.
Raphael T. Haftka and Zafer Grdal. Elements of structural optimization. Kluwer academic
Kluwer academic publishers, 1992.
publishers, 1992.
Jozsef Farkas
Kroly Farkas
Jrmai. and
Analysis
and Jrmai.
OptimumAnalysis
Design ofand
Metal
Structures.
A. A. of
[3] and Jozsef
Kroly
Optimum
Design
Balkema, Rotterdam,
1997.
Metal Structures. A. A. Balkema, Rotterdam, 1997.
D. A. Nethercot. Towards a standardization of the design and detailing of connections.
Journal of constructional
Steel Research,
46:34,
1998.
[4] D. A. Nethercot.
Towards
a standardization
of the design and detailing
L. Pavlovi, A. Krajnc,
and
D.
Beg.
Cost
function
analysis inSteel
the structural
optimization
of
of connections. Journal of constructional
Research,
46:34, 1998.
steel frames. Structural and Multidisciplinary Optimization, 28:286295, 2004.
CEN. EN-1993-1-1.
3: A.
Design
of steel
1-1:
General analysis
rules and rules
[5] L. Eurocode
Pavlovi,
Krajnc,
andstructures.
D. Beg. Part
Cost
function
in the
for buildings., 2006.
structural optimization of steel frames. Structural and Multidisciplinary
CEN. EN-1993-1-8.
Eurocode 3: Design
of steel2004.
structures. Part 1-8: Design of joints., 2006.
Optimization,
28:286295,
Jaakko Haapio. Feature-Based Costing Method for Skeletal Steel Structures based on the
Process Approach.
PhD thesis,
Tampere Eurocode
University of
2012.structures. Part 11:
[6] CEN.
EN-1993-1-1.
3: Technology,
Design of steel
General rules and rules for buildings., 2006.
[7]
CEN. EN-1993-1-8. Eurocode 3: Design of steel structures. Part 18:

Design of joints., 2006.60
[8]
Jaakko Haapio. Feature-Based Costing Method for Skeletal Steel

Structures based on the Process Approach. PhD thesis, Tampere
University of Technology, 2012.
59
60
TUBULAR COMPOSITE COLUMNS

IN A NON-SYMMETRICAL FIRE
Markku Heinisuo
(markku.heinisuo@tut.fi)
Timo Jokinen
Tampere University of Technology, Tampere, Finland
ABSTRACT
A considerable number of studies have been conducted worldwide on columns with fires
that act on all four sides (symmetrical fire). In real buildings, columns are often embedded.
Results of numerical analyses for reinforced concrete filled square steel tubes columns in
non-symmetrical fires are presented for a total of 150 cases. An ISO 834 fire acts constant
along the column on one, two adjacent or three sides. Three embedding systems are
considered for the remaining sides: adiabatic, concrete wall and sandwich panel. When
fire acted on one, two adjacent or three sides, the fire resistance times were on average
about 3.4, 2 and 1.3 times longer than in a symmetrical fire. The final failure mode and
corresponding resistance time depend on the direction of the initial bow imperfection.
Experimental tests are needed to verify the results.
INTRODUCTION
Reinforced concrete filled tubular (CFT) columns combine the advantages of
both steel and concrete materials, such as: attractive appearance, structural
efficiency, fast construction technology and high fire resistance [1], [2]. A
considerable number of numerical simulations and tests have been conducted
worldwide on fires that act on all four sides of a column. These cases are used
for the validation of the analysis models developed in this study. In [3] and
related papers of the authors, composite W-shaped columns with different
embedding systems are examined both experimentally and analytically.
Likewise, in [4] concrete filled rectangular steel tubes in non-symmetrical
fire are examined both experimentally and analytically. The paper excludes
square columns and all columns are without reinforcement. Reference [4]
reports that some columns finally collapsed towards the fire.
This paper presents results of numerical analyses for tubular composite
columns in non-symmetrical fires. Reinforced concrete-filled square steel
tubes of three sizes and two buckling lengths are considered at given central
axial loads. The fire is assumed to be constant along the column length. The
sides on which fire does not act are assumed to be embedded. Three different
embedding systems are considered: adiabatic (see above), concrete wall, and
steel-mineral wool-steel sandwich panel. Axial loads are defined so that they
represent the maximum ultimate concentric loads for symmetrical cases with
respect to specified fire resistance times: 30, 60, 90 and 120 minutes. A total
of 150 cases were analysed.
Thermal and mechanical analyses were conducted using the finite element
software ABAQUS/Standard. Geometrically identical 3D finite element
method (FEM) models were used both for thermal and mechanical analyses.
The material models for steel tubes, reinforcement and concrete are presented
in Eurocode 4 [5]. The effect of an initial bow imperfection is considered. The
shape of the imperfection corresponded to the buckling mode of the column
for the lowest buckling load in ambient conditions. All cases were analysed
using an amplitude of span/666 for the imperfection towards the fire. Some
cases were also studied with the initial imperfection away from the fire. All
columns were hinge supported at both ends. The end nodes of the columns
were fixed to the plane using the coupling command of ABAQUS in the
mechanical analysis.
The reinforcement was modelled using one-dimensional stringer elements.
3D continuum finite elements were used both for the tube and the concrete.
Thermal analysis was conducted first and the temperatures were stored. Then,
mechanical analysis was done with a constant central axial load by increasing
the temperatures inside the column according to the stored temperatures.
Thermal analysis was done using a 3D continuum FE model for the entire
column. To verify the numerical model, the temperatures and fire resistance
times in the symmetrical fire were compared to results available in Reference
[6] and [7]. Temperatures were also calculated using the Safir software [47]. In
summary, the thermal analysis model seemed to work rather well in all cases
considered.
Material models for concrete, tubes and reinforcement were taken from [5].
The Elastic and Plastic options of ABAQUS were used for steel. The Elastic
and Concrete Sheared Cracking options of ABAQUS were used for concrete.
Bi-linear and tri-linear stress-strain relationships, including those of [8],
did not work properly, which led to convergence problems. Similar trials
have been reported in [8]. Finally, a simple elastic stress-strain relationship
was applied to concrete on the tension side. The resistance times were based
on points in time when convergence was no longer reached using ABAQUS
implicit and RIKS methods. The used model gives results which are close to
[7] but for the smallest columns. Table 3 shows a comparison of fire resistance
times between symmetric and non-symmetric fires.
61
n [8]. Finally, a simple elastic stress-strain relationship was applied to concrete on the tension side.
The resistance times were based on points in time when convergence was no longer reached using
ABAQUS
andof the
RIKS
Proceedings
METNETmethods.
Seminar 2013 inThe
Lule used model gives results which are close to [7] but for
62 implicit
he smallest columns. Table 3 shows a comparison of fire resistance times between symmetric and
non-symmetric fires.
Table 3. Ratios
fire
resistance
for non-symmetric
Table 1. of
Ratios
of fire
resistance timestimes
for non-symmetric
versus symmetric fires.versus symmetric fires.

#
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20

CFT column
150x150, 2m
150x150, 2m
150x150, 3m
150x150, 3m
250x250, 3m
250x250, 3m
250x250, 3m
250x250, 3m
250x250, 5m
250x250, 5m
250x250, 5m
250x250, 5m
400x400, 3m
400x400, 3m
400x400, 3m
400x400, 3m
400x400, 6m
400x400, 6m
400x400, 6m
400x400, 6m
Target
R30
R60
R30
R60
R30
R60
R90
R120
R30
R60
R90
R120
R30
R60
R90
R120
R30
R60
R90
R120
Load
[kN]
476
174
303
110
1975
1094
728
443
1171
634
406
233
8031
5942
4868
4053
5624
3564
2625
2045
4S
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
3S
1.43
1.37
1.23
1.00
1.53
1.44
1.29
1.19
1.50
1.17
1.33
1.15
1.27
1.18
1.05
0.93
1.46
1.50
1.60
1.53
2S
1.85
1.73
1.43
1.49
2.49
2.51
2.08
1.95
1.94
2.54
2.51
1.89
2.05
2.23
2.24
2.37
2.72
3.17
3.18
2.35
Ratio to 4S case [-]

1S C3S C2S C1S
3.56 1.52 2.37 4.12
2.68 1.50 2.36 3.24
3.73 1.20 1.42 3.56
2.92 1.07 1.91 3.30
4.63 1.51 2.26 5.97
4.22 1.54 3.15 4.22
2.81 1.33 2.70 2.81
2.01 1.23 2.01 2.01
2.89 1.47 1.72 3.13
3.65 1.15 2.29 4.26
2.68 1.19 1.89 2.68
1.89 1.53 1.89 1.89
4.60 1.30 2.11 4.99
4.31 1.23 2.34 4.31
3.11 1.03 2.40 3.11
2.62 0.93 2.57 2.62
3.69 1.37 3.26 4.41
4.36 1.76 3.99 4.36
3.18 1.74 3.18 3.18
2.35 1.55 2.35 2.35
S3S
1.35
1.30
1.22
0.99
1.51
1.41
1.28
1.17
1.47
1.17
1.15
1.14
1.26
1.15
1.18
1.08
1.44
1.46
1.57
1.50
S2S
1.73
1.68
1.39
1.43
2.35
2.42
2.07
1.90
2.14
2.27
2.40
1.87
2.01
2.09
2.14
2.24
2.59
3.06
3.17
2.35
The main conclusions from this study are that composite tubular columns
resisted the same axial load longer in non-symmetric fires than in symmetric
fires. If a column is embedded in concrete on three sides (fire acts on one side),
fire resistance time increases 2.7-fold, in the case of two adjacent embedded
sides it increases 1.4-fold, and with one embedded side 1.1-fold compared to
a symmetric fire. In the case of embedding with sandwich panels, the fire
resistance times did not increase as much as with concrete embedding. The
increase is about the same as in the adiabatic case. Slender columns tended
to collapse towards the fire, stocky columns finally collapsed away from the
fire. It is recommended that initial bow imperfection should be considered in
directions both away from and towards the fire. Tests are needed to verify the
calculations.
ACKNOWLEDGMENTS
CSC Oy deserves special thanks for providing ABAQUS license, computing
time and general assistance. The financial support of Seinjoen Seudun
Elinkeinokeskus (SEEK) is also gratefully acknowledged.
S1S
3.45
2.60
2.89
2.53
5.14
4.18
2.81
2.01
2.89
3.65
2.68
1.89
-
-
-
-
-
-
-
-
REFERENCES
[1]
Espinos A., Romero M., Hospitaler A. Advanced model for predicting

the fire response of concrete filled tubular columns, Journal of
Constructional Steel research, 66(89), pp. 10301046, 2010.
[2]
Ding J., Wang Y.C. Experimental study of structural behavior of steel

beam to concrete filled tubular column assemblies with different types
of joints, Engineering Structures, 29(12), pp. 34853502, 2007.
[3]
Pires T., Correia A., Rodrigues J. Silva J., CHS and partially encased
columns subjected to fire, Proceedings of EUROSTEEL 2011 6th
European Conference on Steel and Composite Structures, ECCS,
Brussels, pp. 15691574, 2011.
[4]
Yang H., Liu F., Zhang S. Fire performance of CFST columns in

non-uniform fire, Proceedings of EUROSTEEL 2011 6th European
Conference on Steel and Composite Structures, ECCS, Brussels, pp.
15631568, 2011.
[5]
EN 1994-1-2 Eurocode 4: Design of composite steel and concrete

structures Part 12: Structural fire design, CEN, Brussels, p. 110,
2006.
[6]
CIDECT, Improvement and extension of the simple calculation method

for fire resistance of unprotected concrete filled hollow columns, SaintRemy-les-Chevreuse, France, CTICM, CIDECT research project 15Q12/03, 101 p, 2004.
[7]
TRY, Betonitytteisen tersliittopilarin suunnitteluohje, Julkaisumonistamo Etelranta Oy, Helsinki, 2004. (in Finnish)
[8]
Hong, S., Varma, A., Analytical modeling of the standard fire behavior
of loaded CFT columns, Journal of Constructional Steel Research,
65(1), pp. 5469, 2009.
63
64
INFLUENCE OF NATURAL FIRE SCENARIO

BY CFD AND FEM ANALYSIS COUPLING ONTO STEEL
CONSTRUCTION STRUCTURAL RESPONSE
Michal Malendowski
(michal.malendowski@put.poznan.pl)
Adam Glema
(adam.glema@put.poznan.pl)
Institute of Structural Engineering
Poznan University of Technology, Poznan, Poland
ABSTRACT
The problem of numerical analysis of steel structure under natural fire is considered. The
emphasis is put on the importance of coupling between computational fluid dynamics (CFD)
and finite element analysis (FEA) codes with respect to proper calculation of temperature
field inside the structural members. Models preparation, simulation analysis and results
verification precede research on comparing of computations of structural analysis within
steel construction. Computations consist of two sequential numerical simulations: CFD
analysis of fire evolution with the use of FDS software, introducing temperature and heat
flux to thermo-mechanical part of structure respond with the use of ABAQUS finite element
system. Behaviour of the steel construction under fire is comprehensively discussed to
get a safe design with the checks of ultimate limit state. The main goal of the paper is to
analyse different fire scenarios within the building by taking into account, for example,
fire load size and placement, and its influence on the fire development, temperature and
heat distribution, as well as history and state of deformation and stresses in steel structure.
A set of numerical examples is selected and computed for a range of representative fire
scenarios. Conclusions are formulated to present the ability to control the structure under
the fire accident and give directions for designing of safe construction.
INTRODUCTION
Nowadays the usage of natural fire approach for the design of new structures
is gradually rising. In this paper the most general type of this approach
is considered, namely designing using coupled CFD-FEA procedures.
Difficulties are caused especially in the translation of heat condition resulted
from CFD fire analysis into solid mechanics boundary conditions, when the
model differences between those two codes are significant. Therefore, to study
the influence of fire scenario onto the structural response of structures under
natural fire, the coupling procedure and boundary conditions translation
method is developed.
In this work firstly the standard approaches used in current fire engineering
practice are briefly introduced. Then the theoretical background of the
heat transfer problem is invoked as the base for the development of
further translation method. Finally the description of proposed method,
implementation, its verification and a test case of steel structure under
different natural fire scenarios is discussed.
THEORETICAL BACKGROUND OF THE HEAT TRANSFER PROBLEM

Calculation of heat transfer at the interface between solid and gas phase
needs to take into consideration two main and inseparable, in that sense, heat
transfer sources: convective and radiative heat flux. The net heat flux into the
unit surface is the algebraic sum of incoming heat fluxes by convection and
radiation, and can be expressed by the well know expression:
=+
(1)
As far as the two incoming heat quantities are inherent for the general
definition of the boundary conditions in heat transfer problem consideration,
they two different physical phenomena and from the mathematical point of
view can be derived separately.
CONVECTIVE HEAT FLUX

The convective heat flux is basically proportional to the difference between
the gas and surface temperatures at the interface between solid and gas phase,
and is usually expressed as:
=
Wm2
(2)
Where , g, and s are respectively the coefficient of heat transfer by

convection [W/m2 K], gas and solid surface temperatures [K].
RADIATIVE HEAT FLUX

From the structural point of view, the radiative heat flux into the solid phase
corresponds to the absorbed radiation resulted from the incident radiation
coming into the solid surface. From the point of view of CFD field modelling,
the constitutive radiative heat transfer equation can be obtained from the
equilibrium of intensity change in a particular direction across elemental
volume of fluid medium.
65
66
Finally, the radiant incoming heat flux vector , can be calculated as an

integral from the incident intensity over the spherical domain of interest. The
net radiant heat flux absorbed by the solid body can be then calculated as
a difference between the incident (,) and outgoing heat, using the well
known equation for the surface radiation:
=,4
(2)
where is the emissivity of solid surface and is the Stephan-Boltzmann

constant.
T FLUX AND SECTION TEMPERATURES CALCULATION METHOD

HEAT FLUX AND SECTION TEMPERATURES CALCULATION METHOD
proposed approach, the section temperature calculation takes into account both co
In the proposed approach, the section temperature calculation takes into
diative heat fluxes
facing the particular section surfaces. As long as convective hea
account both convective and radiative heat fluxes facing the particular section
dent only on thesurfaces.
constant
of flux
theis heat
transfer
convection and fluid
As longcoefficient
as convective heat
dependent
only on by
the constant
coefficient of the heat transfer by convection and fluid-surface temperature
rature difference,
calculation of temperature does not cause problems. This me
difference, calculation of temperature does not cause problems. This means
l attention mustthat
bespecial
paidattention
for the
recognition
of radiations
mustproper
be paid for
the proper recognition
of radiationsdirections. To
directions.
To
do
that,
authors
introduce
the
definition
of
face
view
angles
as
s introduce the definition of face view angles as lower and upper
limit
of radiation
lower and upper limit of radiation direction that can reach particular point on
an reach particular
on the(Fig.
section
surface (Fig.1.).
the point
section surface
1.).
D
,
2
E
,
1A
,
1
,
2
B
C
Figure 1. Geometrical visualization of view angles 1 and 2 for points A and B and corresponding angle of
radiation vector .
1. Geometrical visualization of view angles ,1and ,2 for points A and B and correspo
angle of radiation vector .
T TRANSFER FROM AMBIENT BOUNDARY INTO THE SECTION
HEAT TRANSFER FROM AMBIENT BOUNDARY INTO THE SECTION

The above approach is useful for the derivation of boundary conditions at
the sections surface. Assuming that the net stored heat is equal to the net
received heat, the relevant partial differential equation can be reduced to a
zero-dimensional case and becomes an ordinary differential equation with
only time dependence:
=
(3)
where is the net heat flux coming into the element and , c, and are the
thickness of sections wall, specific heat and density, respectively.
IMPLEMENTATION
Implementation of the proposed approach is made by coupling the CFD code
called FDS McGrattan et al. [1] and FEM program Abaqus [2]. The formulation
is appropriate for sequential incremental coupling. However, single boundary
transfer is considered in the following investigations.
COUPLING METHOD
When the structural elements are not present in the CFD model, the coupling
must be arranged in such a way, so as to provide sufficient information for
external calculation of section temperatures. This means that even if elements
themselves are not a part of CFD analysis, their position must be somehow
marked in that model. The idea is then to represent the existence of structural
member into the CFD model by introduction of the output request in the point
corresponding to the real position of the member.
BEAM FINITE ELEMENTS

Since the main goal of most structural engineering studies within the field
of fire engineering is the proper prediction of structural response in fire, the
approach taken is to use beam finite element throughout applying the above
procedure for the beams integration points.
COMPUTATIONAL EXAMPLE
The structure under consideration is a single-compartment frame structure
with dimensions 8m by 20 m, height equal to 3,20 m at the ridge and 2,50 m
at the eaves. The geometry of that structure comes from the natural fire full
scale test conducted by Pyl et al. [3].
The structure is a standard steel assembly for small halls. All columns, girders
and purlins are made from I-sections: IPE140, IPE160, IPE120, respectively.
The distance between frames is equal to 5,0 m.
67
68
LOADS
Two contrasting fire scenarios are considered: a localized fire in the middle
of the compartment and a uniformly distributed fire load spread throughout
the floor. The fire is described by given heat release rate per unit area. The
mechanical load is primarily reduced to dead load only (taking into account
0,3 kN/m2 of cladding system), which corresponds to the experimental
conditions (no snow and wind) and favours the goals of this contribution.
Finally also three additional load cases are introduced, to check the structural
behaviour of construction in more severe conditions.
RESULTS
The results are obtained both for thermal and mechanical data comparison
between two selected fire scenarios. A special emphasis is also made to verify
the coupling method used for translation of heat condition between FDS and
Abaqus.
According to simulations, structure lasts more than 60 minutes, when no
other dead load acts on the frame. When mechanical load increases, fire
resistance time decreases, which is to be expected. In accordance with the
limited number of test cases, no pattern of dependence between mechanical
load and fire resistance time was observed for both fire scenarios. A summary
of results is presented in Table 1.
Table 1. Comparison between fire resistance times for different fire scenarios
Table 1. Comparison between fire resistance times for different fire scenarios.
Pool
distributed
pool and
distributed
1,0
dead load
> 60
> 60
-
fire resistance time (minutes)

1,5
2,0
dead load
dead load
> 60
54
60
29
-
1.86
2,5
dead load
22
15
1.47
CONCLUSIONS
CONCLUSIONS
Development
of contemporary
structural engineering
design standards,
Eurocodes, gives
Development
of contemporary
structural engineering
designlike
standards,
engineers more
and more possibilities
for evaluation
structural
safety infor
fire.
Nevertheless, it is
like Eurocodes,
gives engineers
more andofmore
possibilities
evaluation
still a major
to examine,
certain level
accuracy,
or assembly
of issue
structural
safety inassuming
fire. Nevertheless,
it isofstill
a major whole
issue tobuilding
examine,
in natural fire
simulation
using
3D
approaches.
The
proposed
method
is
a
flexible
tool
assuming certain level of accuracy, whole building or assembly in naturalresponding
to the need fire
for incorporating
natural
fire simulations
for the evaluation
safety.
simulation using
3D 3D
approaches.
The proposed
method isof
a structural
flexible tool
responding to the need for incorporating natural 3D fire simulations for the
The heat transfer

analysis
on the interface
evaluation
of structural
safety. between solid and fluid phase are formulated in the paper
and simplified, but take into account both radiative and convective heat fluxes and temperatures as
well, at specific points on the section surface. The results obtained show significant difference in
fire resistance time between pool fire scenario and uniformly distributed fire. The simulations show
that distributed fire scenario is more severe for the structure than a local fire, even if the local fire
per unit area power is approximately four times greater.
The heat transfer analysis on the interface between solid and fluid phase are
formulated in the paper and simplified, but take into account both radiative
and convective heat fluxes and temperatures as well, at specific points on
the section surface. The results obtained show significant difference in fire
resistance time between pool fire scenario and uniformly distributed fire. The
simulations show that distributed fire scenario is more severe for the structure
than a local fire, even if the local fire per unit area power is approximately four
times greater.
The next step in implementation of the method is to extend scripts and run
more exact calculations of heat transfer coefficients for convection on the fluidsolid interface and to take into account the radiation in sections cavities with
the view factors application. Special laboratory tests of different structural
elements under natural fire wait to be performed for experimental validation
of the method under investigation.
REFERENCES
[1]
McGrattan, K., Hostikka, S. and Floyd, J., Fire Dynamics Simulator

(Version 5) Technical Reference Guide, NIST Special Publication
10185, 2010.
[2]
McGrattan, K., Hostikka, S. and Floyd, J., Fire Dynamics Simulator

(Version 5) Users Guide, NIST Special Publication 10195, 2010.
[3]
Pyl, L., Schueremans, L., Dierckx, W., Georgieva, I., Fire safety
analysis of a 3D frame structure based on a full-scale fire test, ThinWalled Structures, 61, pp. 204212, 2012.
69
70
MEMBRANE STRUCTURES SUPPORTED BY A FRAME

OF THIN-WALLED CLOSED SECTIONS
Alexander Tusnin
Professor, Dr. Sc.
Olga Tusnina
PhD Student
Moscow State University of Civil Engineering (MGSU), Moscow, Russia
ABSTRACT
In the paper the work of membrane structures on the rectangular plan with plane support
contour when membrane eccentrically fastened is considered. Features of strain-stress
state of the membrane structure with support contour made of closed section profiles are
studied. An analysis of the differential equation of the twist angles of the contour has
allowed to establish behaviour of the system when membrane is fastened eccentrically
to support contour. This is defined by the relations that include the pure and restraint
torsional stiffness of the contour, span, thickness and elastic modulus of the membrane,
eccentricity of its fastening to contour and a sectorial coordinate of the membrane edge.
Test numerical analyses of the membrane models validated the derived laws.
INTRODUCTION
Membrane structures are widely used as roofing elements in many kinds of
buildings [4, 8, 9]. Typically, a membrane structure is the thin metal shell
(membrane) fastened to the support boundary. Supports can be made of steel,
reinforced concrete or concrete-filled steel tubular members. In the membrane
structures with spans 618 m supports are often made of thin-walled welded
closed profiles. For simplification the membrane structure can be fastened to
the edges of the rectangular boundary frame (Fig. 1). In this case, because of
the eccentrically applied membrane forces, the frame begins to twist.
Many researchers are involved in studies of membrane structures [5, 7]. A
study of the behaviour of membrane structures with thin-walled support
frames of closed profile experiencing restrained torsion is of considerable
practical interest.
THEORETICAL CONIDERATIONS
Torsion of closed thin-walled rectangular profiles is accompanied by their
cross-sectional warping. With the restraining of free warping (in the places of
fastening and connections with other members), additional normal stresses
occur in thin-walled profile and its torsional stiffness increases.
71
Figure
View ofof
membrane
covering.covering.
Schemes of theSchemes
membrane sheet
fastening
to support boundary
of
Figure
1. 1.View
membrane
of the
membrane
sheet frame
fastening
to support
closed profile.
boundary frame of closed profile
For the analysis of thin-walled open profile rods the well-known Vlasov theory
For the analysis

of thin-walled
profileto the
rodstheory
the well-known
Vlasovbeams
theory is successfully
is
successfully
applied
[10].open
Analogous
used for thin-walled
Figure
1. View of
membrane
covering.
Schemes of
the membrane
sheet fastening
to support
applied [10].
Analogous
to
the
theory
used
for
thin-walled
beams
with
open
with open profile for beams
with aframe
closeof
profile,
warping stiffness ofprofile
the for beams with
boundary
closed the
profile
a close profile,
the
warping
stiffness
of
the
beam
is
introduced
in
addition
to the pure torsion
beam is introduced in addition to the pure torsion stiffness. The deformations
stiffness.of The
deformations
of need
the supporting
need to be
taken
in consideration when
the supporting
frame
to be taken frame
in consideration
when
analysing
For thin-walled
the
analysis closed
of closed
thin-walled
open beams,
profile
rods
the
well-known
Vlasov
is successfully
profile
beams,
especially
when
the
membrane
istheory
connected
analysing
thin-walled
profile
especially
when
the
membrane
is connected to the
applied
[10].
Analogous
to
the
theory
used
for
thin-walled
beams
with
open
profile
for
beamsframe
with with the
to
the
support
boundary
with
eccentricity.
The
interaction
of
twisted
thinsupport boundary with eccentricity. The interaction of twisted thin-walled support
a
close
profile,
the
warping
stiffness
of
the
beam
is
introduced
in
addition
to
the
pure
torsion
walled
support
frame
with
the
membrane
has
some
special
features
and
membrane has some special features and requires a further study.
stiffness.
Thea further
deformations
requires
study.of the supporting frame need to be taken in consideration when
analysing thin-walled closed profile beams, especially when the membrane is connected to the
support boundary with eccentricity. The interaction of twisted thin-walled support frame with the
METHODS
membrane has some special features and requires a further study.
METHODS
The simple
model is studied with the aim of determination the main parameters influenced on the
The simple model is studied with the aim of determination the main parameters
work METHODS
of influenced
the system.
Thework
system
of system
contourconsists
beam ofwith
the beam
span with
2a with fixed edges
on the
of theconsists
system. The
contour
(constraints
are
set
on
the
linear
displacements
in
the
plane
perpendicular
to
the
contour
span
2a with
fixed with
edges
are set on
linear
displacements
The the
simple
model
is studied
the(constraints
aim of determination
thethe
main
parameters
influenced
on theand on the
angelswork
ofinrotation)
and
flat
withcontour
the
sizes
and
thickness
t (fig.
of
system.
Themembrane
system to
consists
of contour
beam
with
the span
2a
with2).
fixed edges
thetheplane
perpendicular
the
and2aa
on the
angels
of rotation)
and
(constraints
are set onwith
the linear
displacements
the planetperpendicular
to the contour and on the
flat membrane
the sizes
2aa and in
thickness
(fig. 2).
angels of rotation) and flat membrane with the sizes 2aa and thickness t (fig. 2).
a)a)
b)
Figure 2. The studied system. a) general view; b) cross-section of the contour

Figure
2. The
studied
system.
a) general
view;
b) cross-section of the contour
Figure
2. The studied
system,
a) general
view; b) cross-section
of the
contour.
96
96
b)
72
The sheet is fastened with eccentricity e at the height of the boundary beam.
The uniformly distributed load p acts on the free edge of the sheet with the
coordinate y=a. Numerical studies have been carried out using the program
MSC NASTRAN. The analysis is carried out for half of the structure taking into
account symmetry of the system by the introduction appropriate constraint at
the axis of symmetry.
The
sheet is the
fastened
with eccentricity
e at
the height
of of
the BEAM
boundaryfinite
beam.elements
The uniformly
Initially,
problem
was solved
with
the use
for
distributed load p acts on the free edge of the sheet with the coordinate y=a. Numerical studies
the
modelling
of
support
boundary
excluding
torsional
restraint.
The
steel
have been carried out using the program MSC NASTRAN. The analysis is carried out for half of
membrane
wasintomodelled
as PLATE
finite byelements
(mesh
2020)constraint
with
the
structure taking
account symmetry
of the system
the introduction
appropriate
atthickness
the axis of symmetry.
of 1 mm. The connection between the boundary beam and the
membrane was modelled by using bars of high stiffness, with the eccentricity
Initially, the problem was solved with the use of BEAM finite elements for the modelling of support
of membrane fastening. The size a=6 m, the load is p=120 kN/m, cross-section
boundary excluding torsional restraint. The steel membrane was modelled as PLATE finite
of boundary
beamwith
is 2001006
Theconnection
design scheme
considered
elements
(mesh 2020)
thickness of 1 mm.
mm. The
between of
the the
boundary
beam and
membrane
model
is shown
in Fig.
use of
this
the following
the
membrane was
modelled
by using
bars3.ofWith
high the
stiffness,
with
themodel,
eccentricity
of membrane
fastening.
The obtained.
size a=6 m,
thedisplacement
load is p=120ofkN/m,
cross-sectionedge
of boundary
beam is
results were
The
the membrane
in the middle
2001006
mm.beam
The design
scheme
of the of
considered
membrane
model
shown
in Fig.
With
of contour
in the
direction
the Y-axis
is equal
tois17.5
mm,
and3.the
the use of this model, the following results were obtained. The displacement of the membrane edge
of rotation
ofbeam
the Boundary
beam
about
itsislongitudinal
axisand
(X-axis)
inangle
the middle
of contour
in the direction
of the
Y-axis
equal to 17.5 mm,
the angleisof
0.526.
rotation
of the Boundary beam about its longitudinal axis (X-axis) is 0.526.
Figure 3.Figure
The numerical
model whenmodel
contourwhen
is modelled
by BEAM
finite elements.
3. The numerical
contour
is modelled
by BEAM finite elements
The
same
membrane
system
withsizes
theand
same
and
butbeam
with
the
The
same
membrane
system with
the same
loads sizes
but with
the loads
boundary
modelled
using
PLATE finite
were using
analysed.
The design
scheme
for thiswere
case ofanalysed.
contour modeling
boundary
beamelements
modelled
PLATE
finite
elements
The
isdesign
represented
on thefor
Fig.this
4. case of contour modeling is represented on the Fig. 4.
scheme
For the midspan of a boundary beam, the following results were obtained in this case. The
For the midspan
of a boundary
beam,
following
results
obtained
in
displacement
of the membrane
edge in the
Y-axisthe
direction
is 24 mm.
The were
angle of
rotation about
the
longitudinal
axis
is 0.0498.
this
case. The
displacement
of the membrane edge in the Y-axis direction is 24
mm. The angle of rotation about the longitudinal axis is 0.0498.
A comparison of the results for these two models allowed us to make following conclusions. Twist
angle of the beam for the first model (0.526) exceeds twist angel in the second model (0.0498)
A comparison of the results for these two models allowed us to make following
conclusions. Twist angle of the beam for the first model (0.526) exceeds twist
97 than 10 times. Displacements for the
angel in the second model (0.0498) more

first model (17.5 mm) are smaller than for the second model (24 mm) by 27%.
73
more than 10 times. Displacements for the first model (17.5 mm) are smaller than for the second
model
(24
mm)
by
27%.
more
than
10 times.
Displacements
for the first
(17.5
mm) (17.5
are smaller
the second
more
than
10
times.
Displacements
for model
the first
model
mm)than
are for
smaller
than for the second
model
(24
mm)
by
27%.
model (24 mm) by 27%.
more than 10 times. Displacements for the first model (17.5 mm) are smaller than for the second
model (24 mm) by 27%.
Figure
4. The
numerical
modelmodel
when the
boundary
is modelledisusing
PLATE using
finite elements.
Figure
4. The
numerical
when
the boundary
modelled
PLATE finite elements
Thus, Figure
required4.accuracy
of the analysis
not achieved
when isuse
simple BEAM
elements finite
for elements
The numerical
model iswhen
the boundary
modelled
using PLATE
modelling
the supportaccuracy
beams. Because
of inabilityisofnot
theachieved
beam finitewhen
elements
for
the boundary,
its
Thus,
required
of
the
analysis
use
simple
BEAM
Figure
4. Themembrane
numericalneed
model
when
the and
boundary
is modelled
using
PLATE
interaction
with
to be
studied
the basic
parameters
influencing
on finite
the elements
elements
forthe
modelling
support
beams.
Because
of
inability
ofsimple
the beam
Thus,
required
accuracy the
of the
analysis
is not
achieved
when use
BEAM
elements for
behavious of structure need to be derived.
finite elements
for the
boundary,
its interaction
with
membrane
needBEAM
to the elements
modelling
the support
beams.
Because
of is
inability
of
the the
beam
finite
for
boundary,for
its
Thus,
required
accuracy
of the
analysis
not achieved
when
use elements
simple
interaction
with
theThe
membrane
need
tooftwist
be
studied
theisbasic
parameters
influencing
the
beusstudied
and
the
basic
parameters
influencing
the
behavious
of
structure
Letmodelling
consider
thesupport
differential
equation
of
the
angles
ofand
thin-walled
closed
profile
the
beams.
Because
inability
of
the
beam
finite
elements
for
the boundary,
its
Figure
4.
numerical
model
when
the on
boundary
modelled
usingsupport
PLATE
finite on
elements
behavious
of
structure
need
to
be
derived.
need
to
be
derived.
beam
about
its
longitudinal
axis
which
is
generally
represented
in
the
same
view
as
for
thin-walled
interaction with the membrane need to be studied and the basic parameters influencing on the
open
profile section
[9]:
behavious
of structure
need to be derived.
Thus, required accuracy of the analysis is not achieved when use simple BEAM elements for
Let
consider
equation
of the
twisttwist
angles
of beam
thin-walled
closed profile
Let us
usmodelling
considerthe
thedifferential
differential
equation
of
of thin-walled
the
support
beams.
Because
of the
inability
ofangles
the
finite elements
for thesupport
boundary, its
beam
its
longitudinal
axis
which
is
generally
represented
in
the
same
view
as
for
thin-walled
closed
profile
support
beam
about
its
longitudinal
axis
which
is
generally
Let
usabout
consider
the
differential
equation
of
the
twist
angles
of
thin-walled
closed
profile
supporton the
,
(1)
interaction with the membrane need to be studied and the basic parameters influencing
open
profile
section
[9]: view
represented
inlongitudinal
the
same
as
forbethin-walled
profile in
section
[9]: view as for thin-walled
beam
about
its
axis
which
is
generallyopen
represented
the same
behavious
of structure
need
to
derived.
open
profile
section
[9]:
- the stiffness of the support contour on the pure torsion;
- contour sectorial
where,
Let us consider the differential equation of the twist angles of thin-walled
(1) ofclosed profile
(1) support
stiffness; - twist angel of the contour; m - ,intensity of the distribution of torque; b intensity
beamofabout
its longitudinal axis which is generally represented in the same view as for thin-walled
the distribution
bi-moment.
,
(1)
open profile section [9]:
Equation
ist valid
in the
absenceof
of the
of the
beam. on
For
this
where,
the
stiffness
thebending
support
contour
pure
- contour sectorial
where,(1)GI
- the
stiffness
the
support
contour
on the
theproblem:
puretorsion;
torsion; EI
contour
sectorial
stiffness;
twist
angel
of
the
contour;
m
intensity
of
the
the stiffness
of the
support
torsion; of torque;
- contour
sectorial
where,
stiffness; , - -twist
angel of the
contour;
m ,-contour
intensityonofthe
thepure
distribution
b intensity
of(1)
,
distribution
of of
torque;
intensity
of the
of bi-moment.
the
distribution
bi-moment.
stiffness;
- twist
angelbof the
contour;
m - distribution
intensity of the
distribution of torque; b intensity of
(2)
the distribution of bi-moment.
Equation
(1)
is
valid
in
the
absence
of
the
bending
of
the
beam.
For
this
Equation
(1)
is
valid
in
the
absence
of
the
bending
of
the
beam.
For
this
problem:
the
stiffness
of
the
support
contour
on
the
pure
torsion;
- contour sectorial
where,
,
,
problem:
Equation
(1) is valid
in the absence of the bending
of the beam. For this problem:
stiffness;
, - twist angel of the contour;
, m - intensity of the distribution of torque; b intensity of
the distribution of bi-moment.
modulus and the thickness of, the membrane;
- Poisson ratio; u and v where, m , tm Elastic
,
(2)
the displacements of membrane through its fastening to a contour parallel to the axes and
Equation (1) is valid in the absence of the bending of the beam. For this problem:
(2)
(2)
98 ,
,
,
,

,
,
where, m , tm Elastic modulus and the thickness of the membrane;
- Poisson ratio; u and (2)
vthe displacements
of membrane
its fastening
to
a
contour
parallel
to
the
axes
and
where,
m , tm , Elastic
modulus through
and the thickness
of
the
membrane;
Poisson
ratio;
u
and
v
,
where,
,
t
Elastic
modulus
and
the
thickness
of
the
membrane;
m
m
98
Poisson ratio; u and v the displacements of membrane
through its fastening

98
to a contour
parallel
to
the
axes
and
respectively;
the
eccentricity
where, m , tm Elastic modulus and the thickness of the
membrane;of the- Poisson ratio; u and v the sectorial coordinate
membrane
fastening
to
contour
on
the
vertical;
98

74
- the
respectively;
the
eccentricity
the membrane
fastening
to contourNyon the
of cross-section
of the
contour of
where
the membrane
is fastening,
thevertical;
sectorial
coordinate
ofeccentricity
cross-section
of the
contour
where
isthe
fastening,
respectively;
per
the
the
membrane
fastening
to membrane
contour
vertical; Ny - the
normal force
unit
length inofmembrane
acting
along the
Y-axis
in theonplace
respectively;
per
the
of the
membrane
fastening
tomembrane
contour
on
vertical;Ny with
- the
the
normal
force
unit
lengthinterface;
in membrane
acting
along
Y-axis
in the
place
sectorial
coordinate
ofeccentricity
cross-section
ofNthe
where
theper
isthe
fastening,
of contour
with
membrane
the shear
force
unit
length
in of contour
xy contour
the
sectorial
coordinate
of
cross-section
of
the
contour
where
the
membrane
is
fastening,
N
y
the
shear
force
per
unit
length
in
membrane
in
the
place
of
contourmembrane
interface;
N
normal
forcein per
unit xylength
in membrane acting
along Y-axis in the place of contour with
membrane
the place
of contour-membrane
interfacing.
normal force
per unit
in membrane
along inY-axis
in thein place
of contour
with
interfacing.
the shear
force peracting
unit length
membrane
the place
of contourmembrane
interface;
Nxylength
membrane
interface;
Nxy the shear force per unit length in membrane in the place of contourmembrane
interfacing.
eicity
fastening
contour on
the vertical;
- the
Given
thetoto
well-known
dimensionless
parameters
ofofthe
membrane
fastening
contour
on
the
- -the
ricity
theto
membrane
fastening
contour
on
thevertical;
vertical;parameters
the [6]: [6]:
membrane
interfacing.
Given
the well-known
ur
where ofthe
membrane
is
fastening,
Ny thedimensionless
-section
contour
the
isisfastening,
the
s-section
ofthe
the
contourwhere
where
themembrane
membrane
fastening,NNy ythe
the
well-known
dimensionless
parameters
[6]:
along
Y-axis in
the Given
place
contour
with
hg in
acting
along
Y-axis
inin the
with
gth
in membrane
membrane
acting
alongof
Y-axis
the, place
place ofof contour
contour
with, [6]:
(3)
, well-known
, parameters
(3)
Given
the
dimensionless
tthelength
in
membrane
in
the
place
of
contourhe
shear
force
per
unit
length
in
membrane
in
the
place
of
contourshear force per unit length in, membrane ,in the place, of contour-,
(3)
,
,
,
,
(3)
Using relations (2), equation (1) can be represented as follows:
]:
sionless parameters
Usingrelations
relations(2),
(2),equation
equation
can
represented
as follows:
nsionless
parameters[6]:
[6]:Using
(1)(1)can
bebe
represented
as follows:
Using relations (2), equation (1) can be represented as follows:
(3)
,,
,,
(3)
(3)
,
(4)
(4)
,
(4)
,
(4)
as
(1)
can
(1)follows:
canbe
berepresented
representedasasfollows:
follows:
where,
- the relative stiffness of the contour on pure torsion;
- the
where,
- the relative stiffness of the contour on pure torsion;
- the
,
- the
contour
on on
pure
torsion;
- the
where,
where,
therelative
relativestiffness
stiffness
the
contour
pure
torsion;
,(4)
(4)
,
(4) ofofthe
relative sectorial torsional stiffness of the thin-walled open profile contour;
- the relative
relative sectorial torsional stiffness of the thin-walled open profile contour;
- the relative
relative sectorial
torsional
stiffness of
the thin-walled open profile contour;
- the relative
sectorial
coordinate
of the membrane
edge.
contour
on
pure
torsion;
the
the
relative
sectorial
torsional
stiffness
of
the
thin-walled
elative
stiffness
of
the
contour
on
pure
torsion;
the
sectorial
coordinate
of
the
membrane
edge.
relative stiffness of the contour
on pure torsion;
- the

sectorial
coordinate
of the membrane
edge. that
In
the derivation
of equation
(4) it is assumed
(for the twist angles
the error
(for
the
twist
angles
the error
In
the
derivation
of
equation
(4)
it
is
assumed
that
led
open
contour;does
- the
relative
open
profile
contour;
relative
coordinate
the
ffness
ofofprofile
the
profile
contour;
relative
tiffness
thethin-walled
thin-walled
open
profile
contour;
-the
the
relative
not
exceed
5%),
and relative
angle
is equal
to
. the twistofangles
Inopen
the
derivation
of equation
(4)-ittwist
isthe
assumed
that sectorial
(for
the error
membrane
edge.
does
not exceed
5%), and relative twist angle is equal to
.
mbrane edge.
does not exceed 5%), and relative twist angle is equal to
.
embrane
edge.
The
relations
include
the contour,
In the
derivation
of torsional
equationstiffness
(4) it isofassumed
that span,
cos thickness
= 1 (for and
the Elastic
twist modulus of the
(for
the
twist
angles
the
error
membrane,
eccentricity
of
its
fastening
to
contour
determine
the
behaviour
of
the system.
Relative
The
relations
include
torsional
stiffness
of
the
contour,
span,
thickness
and
Elastic
modulus
of the
the
twist
angles
the
error
(for
18
the
error
does not exceed
5%), and relative twist angle is equal
n(4)
(4)ititisisassumed
assumedthat
that angles
(for
the
twist
angles
theof
error
The relations
include
torsional
stiffness
thechanging
contour,
span,
thickness
andofElastic
modulus
of
the
membrane,
eccentricity
ofdoitsnot
fastening
towith
contour
determine
the
behaviour
thethe
system.
Relative
values
,
and
change
of
absolutely
values
of
contour
span,
al to twist
.angleisisequal
membrane,
eccentricity
of
its
fastening
to
contour
determine
the
behaviour
of
the
system.
Relative
.
ative
to
to
.
.
lative
twistangle
equalthickness,
to
,Elasticand
doofnot
with and
changing
of eccentricity
absolutely values
the contour
span,
values
modulus
thechange
membrane
contour,
of the of
membrane
fastening
, the above
and a simplified
do not change with
changing of
of membrane
absolutely structures
values of can
the be
contour
span,
values
etc.
GivenElastic
of the
developed.
thickness,
modulus of the method
membrane
andanalysis
contour, eccentricity
of the membrane
fastening
thickness,
Elastic
modulus
of
the
membrane
and
contour,
eccentricity
of
the
membrane
fastening
our,
span, thickness
andIn
Elastic
modulus
of
the
the
frame
this
method,
the
values
of analysis
relative of
parameters
determine
and
etc.
Given
the of
above
a Elastic
simplified
method
the
membranewill
structures
can bestresses
developed.
al
stiffness
ofofthe
span,
thickness
and
ofofof
the
nal
stiffness
thecontour,
contour,
span,
thickness
and
Elasticmodulus
modulus
the
etc.
Given
theof
above
simplified
method
ofofThe
the
analysis
of span,
membrane
structures
be
developed.
determine
behaviour
of the
the
system.
Relative
displacements
occurring
inthethe
structure.
productivity
of such
an determine
approach
proved
by
In
frame
this aofmethod,
the
values
of
relative
parameters
will
stresses
and
totocontour
the
system.
Relative
The
relations
include
torsional
stiffness
the
contour,
thickness
andcan is
sfastening
fasteningthe
contourdetermine
determine
thebehaviour
behaviour
ofthe
system.
Relative
In ofthethemodulus
frame
of [8,
thisthe
method,
the
values
of isrelative
parameters
will
determine
stresses
and
ging
of
absolutely
values
contour
span,
practice
of
design
9].
The
aim
of
this
paper
to
confirm
the
validity
of
the
equations
derived
displacements
occurring
in
the
structure.
The
productivity
of
such
an
approach
is
proved
by
Elastic
of
membrane,
eccentricity
of
its
fastening
to
contour
not
change
with
changing
of
absolutely
values
of
the
contour
span,
not change with changing
of absolutely
values of
the
contour
span,
displacements
occurring
in
the
structure.
Theisproductivity
of validity
such anofapproach
is proved
by
practice
of
design
[8,
9].
The
aim
of
this
paper
to
confirm
the
the
equations
derived
,
and
).
for
the
relative
parameters
(
ontour,
eccentricity
of
the
membrane
fastening
determine
the
behaviour
of
the
system.
Relative
values
m,
m
and
do
not
f the
eccentricity
of
membrane
fastening
of
themembrane
membraneand
andcontour,
contour,
eccentricity
ofthe
the
membrane
fastening
practice
of
design
[8,
9].
The
aim
of
this
paper
is
to
confirm
the
validity
of
the
equations
derived
alysis
of
membrane
structures
be developed.
for
the
relative
parameters
, be
and values
). of the contour span, thickness,
change
with
changing
of( can
absolutely
ified
ofofthe
membrane
structures
lifiedmethod
method
theanalysis
analysis
ofcan
membrane
structures
can
developed.
for of
the
relative
parameters
( contour
, bedeveloped.
andthecontour,
).
ative
parameters
will
determine
stresses
and
Elastic
modulus
of
the
membrane
and
eccentricityprofile
of the[8]
membrane
For
the
thin-walled
support
of
the following expressions
od,
the
values
of
relative
parameters
will
determine
stresses
and
hod, the values of relative parameters will determine stresses closed
and rectangular
roductivity
of
such
an
approach
is
proved
by
fastening
etc.
Given
the
above
aissimplified
method
of the analysis
membrane
are
valid:
For
the thin-walled
support
contour
of the closed
profileof[8]
the following expressions
the
of
an
approach
by
the structure.
structure. The
The productivity
productivity
of such
such
an
approach
is proved
proved
by rectangular
For
the
thin-walled
contour
ofderived
the
closedofrectangular
profile
the following
expressions
confirm
thepaper
validity
of
the
equations
derived
structures
be support
developed.
In the
frame
this method,
the[8]
values
of
are
valid:
etoaim
isistoto
confirm
the
validity
ofofthe
he
aimofofthis
this
paper
confirm
thecan
validity
theequations
equations
derived
are valid:
The
inertia
moment
on
the
pure
torsion
,
relative
parameters
will
determine
stresses
and
displacements
occurring
in
,,
and
and ).).
The
moment
the pure torsion
- an approach
, is proved by practice of
the inertia
structure.
The on
productivity
of such
The inertia moment on the pure torsion ,
design
[8,
9].
The
aim
of
this
paper
is
to
confirm
the
validity of the equations
ectangular
profile
[8]
the
following
expressions
ontour ofofthe
profile
[8]
contour
theclosed
closedrectangular
rectangular
profile
[8]the
the-following
followingexpressions
expressions,
section inertia
moment
derived for the relative parameters (m, m and ).
section inertia moment ,
section inertia moment ,
,
section
coordinate
of
the
membrane
edge
.
re torsion
,,
ure
torsion- section coordinate of the membrane edge .
section coordinate of the membrane edge .
,,
. edge- mbrane
mbraneedge
..
99

99
99

values
,
and
do not change with changing of absolutely values of the contour span,
,Elasticand
doofnot
with and
changing
of eccentricity
absolutely values
of
the contour
span,
values
thickness,
modulus
thechange
membrane
contour,
of the
the membrane
membrane
fastening
thickness, Elastic modulus of
the
membrane
and contour,
eccentricity of
fastening
thickness,
Elastic
modulus
of
the
membrane
and
contour,
eccentricity
of
the
membrane
fastening
etc. Given
Given the
the above
above aa simplified
simplified method
method of
of the
the analysis
analysis of
of membrane
membrane structures
structures can
can be
be developed.
developed.
etc.
etc.the
Given
theof
above
simplifiedthe
method
of of
the
analysis parameters
of
membrane
structures
can be
developed.
In
the
frame
of
thisa method,
method,
the
values
of Proceedings
relative
parameters
will
determine
stresses
and
of
the
METNET Seminar
2013
in Lule
75 and
In
frame
this
values
relative
will
determine
stresses
In
the
frame
of
this
method,
the
values
of
relative
parameters
will
determine
stresses
displacements
occurring
in
the
structure.
The
productivity
of
such
an
approach
is
proved
by
displacements occurring in the structure. The productivity of such an approach is proved and
by
displacements
occurring
in
the
structure.
The
productivity
of
such
an
approach
is
proved
by
practice
of
design
[8,
9].
The
aim
of
this
paper
is
to
confirm
the
validity
of
the
equations
derived
practice of design [8, 9]. The aim of this paper is to confirm the validity of the equations derived
practice
of design
[8, 9]. (The, aim of
this paper
is to confirm the validity of the equations derived
for
the
relative
parameters
and
).
for the relative parameters ( ,
and ).
for the relative parameters ( ,
and ).
ForFor
thethe
thin-walled
support
contour
of
closed
rectangular
profile[8]
[8]the
thefollowing expressions
thin-walled
support
contour
ofthe
the closed
closed rectangular
rectangular profile
profile
For the thin-walled support contour of
the
[8] the
following expressions
following
expressions
are
valid:
Forvalid:
the thin-walled support contour of the closed rectangular profile [8] the following expressions
are
valid:
are
are valid:
TheThe
inertia
moment
on
The
inertia
moment
onthe
thepure
puretorsion
torsion -,,
inertia
moment
on
the
pure
torsion
The inertia moment on the pure torsion ,
section
inertia
moment-section
inertia
moment
section
inertia
moment
section inertia moment -
,,
,
..
.
section coordinate
coordinate of
of the
the membrane
membrane edge
edge -section
section
coordinate
of of
thethe
membrane
section
coordinate
membraneedge
edge-
In the expressions for the inertia moments and section coordinate, the
In the texpressions
for the
moments
andthe
section
the following
of
contour for
wall,
binertia
and coordinate,
hmoments and
following symbols are used:
k the thickness
In inertia
the
section co
99 the
expressions
99
the width and the height- the
of the
profileofonthethe
middle
surface
of -the
crossthickness
contour
wall,
b and
the
widthwall,
and the
height
of width
the pro
thickness
-99
of hthe
contour
b and
h - the
the
section.

surface of the cross-section.
tia moments and section coordinate, the following symbols are used: surface of the cross-section.
In the expressions for the inertia moments and section coordinate, the following symbols are used:
ur wall, b and h - the width
andthis
the height
of theTaking
profile
on
middle
Taking
into account
the relative
parameters
of the
membrane
structure
thisthe
into
account
the
relative
parameters
of
membrane
structure
formt
- the thickness of the contour
wall,
b and
h - the
width
andinto
the
height
thethe
profile
on
the middle
Taking
this
accountofthe
relative
parameters
of the
surface
of the cross-section.
equal:
,
profile contour
contour with
with eccentricity
eccentricity
for thin-walled
closed profile
are eccentricity
equal:
are equal:
profile contour with
elative parameters of the membrane structure for thin-walled closed
Taking this into account the relative parameters of the membrane structure for thin-walled closed
. ,
.
city
are equal:
,
profile contour with eccentricity
are equal:
,
,
The correctness of these

for the
generalized
parameters
is confirmedp
Therelations
correctness
of these
relations
for the generalized
The correctness .of thesenumerical
relations studies
for the of
generalized
parameters
is confirmed
thenumerical
modelling
of
boundary
by
the
PLATE
finite
elements.
studies of the modelling of boundary
by the PLT
by the results of numerical
studies
of the
modelling
of boundary
the PLATE
approach
was
confirmed
by the
analysis
ofby
structures
the eccentric
fastening
approach
was confirmed
by thewith
analysis
of structures
with
finite
elements.isThe
validity
ofboundary
the results
approach
was confirmed
byInthe
analysis
of
tions for the generalized
parameters
confirmed
by
the
of pure
to the
with
torsion
[3].
addition
to
the
above
derived
relation
to
the
boundary
with
pure
torsion
[3].
In
addition
to
the
The
correctness
of these
relations
for
the
generalized
is confirmed
by
the results
lling of boundary by the
PLATE
finite
Thefastening
validity
of
structures
with elements.
the
eccentric
ofthe
the
membrane
toparameters
the stiffness
boundary
with
generalized
parameters
for the parameters
bending
thebending
contour
in the of
plane
generalized
forofthe
stiffness
of theof
numerical
studies
of
the
modelling
of
boundary
by
the
PLATE
finite
elements.
The
validity
of
the
e analysis of structurespure
with torsion
the eccentric
fastening
of to
thethe
membrane
[3]. In
addition
above derived relations the well-known
approach
wasderived
confirmed
by thethe
analysis
. of structures with the. eccentric fastening of the membrane
sion [3]. In addition to
the above
relations
well-known
generalized
parameters
for the bending
stiffness of the contour in the plane
the
boundary
with
pure torsion [3]. In addition to the above derived relations the well-known
e bending stiffness of to
themembrane
contour
in[8,
the
of
9]:plane of membrane [8,9]:
generalized parameters for the bending stiffness of the contour in the plane of membrane [8,9]:
.
RESULTS
RESULTS
The numerical studies were

carried outstudies
with constant
valuesout
of with
generalized
The numerical
were carried
constantparam
valu
geometric sizes and characteristics
of
the
structure
and
the
values
of
the load.
T
geometric
sizes
and
characteristics
of
the
structure
and
th
RESULTS
RESULTS
modelled using PLATE modelled
finite elements
(Fig.
4).
using PLATE finite elements (Fig. 4).
arried out with constant values of generalized parameters but various
The
numerical
studies
were
carried
out
with
constant
values
of
generalized
parameters
butcan
various
The
numerical
studies
were
carried
out
with
constant
values
of of
generalized
istics of the structure and the values of the load.TheThe
boundary
was The forces
forces
and displacements
for the
middle
the boundary
span
and
displacements
forThe
the boundary
middle
ofbe
theexpresse
bounda
geometric
sizes
and
characteristics
of
the
structure
and
the
values
of
the
load.
was
parameters
but
various
geometric
sizes
and
characteristics
of
the
structure
elements (Fig. 4).
modelled
using
PLATE
finite
elements
(Fig.
4).
;
and the values of the load. The boundary
was modelled ;using
; PLATE finite ;
elements
(Fig.
for the middle of the boundary
span
can4).
be expressed as follows:
The forces and displacements for the middle of the boundary span can be expressed as follows:
;
;
; ;
, ;
,
(5)
(5)
,
,
,
depends
only
generalizeo
where, correction factors
,
,
, on, the depends
;
,
equation (5),
- bending
moments
contourmoments
in the plane
membrane
equation
(5), in the
- bending
in theofcontour
in t
flexural-torsional
in the beam;bi-moment
and - in
horizontal
displacement
and
the beam;
and - horiz
,
,
,
depends only on the generalized
parameters.bi-moment
In flexural-torsional
,
, the results
depends
only
on
the
generalized
parameters.
In
beam.,In Table
1,
of
numerical
analyses
are
represented.
Geometrical
beam. In Table 1, the results of numerical analyses are rep
moments in the contour in the plane of membrane
of the span;
-thickness
cross-section,
loads,
ofofmembrane
were
varied
and the
equation (5),
- bending
moments in
the contour
inand
the span
plane
membraneand
of span
the
span;
- gener
cross-section,
loads,
thickness
of membrane
we
n the beam; and - horizontal displacement and
angle The
of the
weretwist
constant.
results
represented
for
two
cases
of
membrane
structure
ofinme
f
were
constant. The
results represented
for two
cases
flexural-torsional
bi-moment
in
the
beam;
and
horizontal
displacement
and
twist
angle
of
the
of numerical analyses are represented. Geometrical
sizesofof
the beam
section
beam
is 2001006 mm, and in the second - 40020012 mm.
to the
to boundary
the boundary
withwith
purepure
torsion
torsion
[3]. [3].
In addition
In addition
to the
to above
the above
derivd
generalized
generalized
parameters
parameters
for the
for the
bending
bending
stiffness
stiffness
of the
of the
contour
contour
in thin
76
RESULTS
The numerical studies were carried out with constant values of generalized parameters but various
geometric sizes and characteristicsRESULTS
ofRESULTS
the structure and the values of the load. The boundary was
modelled using PLATE finite elements (Fig. 4).
The forces and displacements for the middle of the boundary span can be
The The
numerical
numerical
studies
studies
werewere
carried
carried
out with
out with
constant
constant
values
values
of genera
of gen
expressed as follows:
The forces and displacements for the
middle
of the
boundary
span can of
bethe
expressed
as and
follows:
geometric
geometric
sizes
sizes
and
and
characteristics
characteristics
of structure
the structure
and
the values
the values
of t
modelled
modelled
using
using
PLATE
PLATE
finite
finite
elements
elements
(Fig.(Fig.
4). 4).
;
;
The The
forces
forces
and and
displacements
displacements
for the
for middle
the middle
of the
of boundary
the boundary
spanspan
can cb
(5) (5)
; ;
; ;
,

,
,
,
depends
only on the
generalized
parameters. In
; ; only
,
where, correction factors 1, 2, 1 , 2 depends
on the , generalized
equation
(5),
bending
moments
in
the
contour
in
the
plane
of
membrane
of
parameters. In equation (5), Msp bending moments in the contour in the the span;
plane of membrane
of the in
span;
Bsp flexural-torsional
in and
the twist angle of the
flexural-torsional
bi-moment
the beam;
and - horizontalbi-moment
displacement
, beam.
,Geometrical
, In
,, Table
, depends
depends
onlybeam
only
on the
on
where,
where,
correction
correction
factors
factors
beam;In
v and
1,horizontal
and
twist
angle
of the
beam.
Table
the results displacement
of numerical
analyses
are
represented.
sizes
of the
bending
bending
moments
moments
in
the
in
contour
the
contour
in
the
in
plane
the
plane
of m
equation
equation
(5),
(5),
cross-section,
loads,
thickness
and
span
of
membrane
were
varied
and
the
generalized
parameters
1, the results of numerical analyses are represented. Geometrical sizes of the
were
The results
represented
for
two
cases
of
membrane
structure
in
first
case
the
crossbeamconstant.
cross-section,
loads,
thickness
and
span
of
membrane
were
varied
flexural-torsional
flexural-torsional
bi-moment
bi-moment
in the
in beam;
the beam;and and- horizontal
- horizontal
displac
dis
section
beam is 2001006
mm,
and
in
the
-1,40020012
and theofgeneralized
parameters
were
The
results
for
beam.
beam.
Inconstant.
Table
Insecond
Table
1, the
results
the
results
of represented
numerical
ofmm.
numerical
analyses
analyses
are represented.
are represented
Ge
two cases of membrane structure
cross-section,
in first case
the
cross-section
of span
beam
cross-section,
loads,
loads,
thickness
thickness
and and
span
of membrane
ofismembrane
werewere
varied
varied
and
Analysis
of the
data
ininTable
1 obtained
with
varying
torsional
stiffnessfor
oftwo
thetwo
contour,
span
and strucs
2001006
mm,
and
the second
constant.
40020012
mm.
werewere
constant.
The The
results
results
represented
represented
for
cases
cases
of membrane
of membrane
thickness of the membrane shows section
thatsection
at constant
values
of generalized
, - 40020012
and mm
of beam
of beam
is
2001006
is 2001006
mm,mm,
andparameters
and
in the
in second
the second
-, 40020012
Analysis
of the
in Table
1 obtained
with varying
torsional
stiffness
of exceed
the
the values
ofdata
correction
factors
in Equation
(5), (maximum
error
doesnt
5.3%) do not
of the
of data
the
data
in that
Table
incharacteristics
Table
obtained
1 obtained
with
with
varying
varying
torsional
st
contour,onspan
of Analysis
theofAnalysis
membrane
shows
at1constant
values
depend
the and
spanthickness
and thickness
membrane,
stiffness
of the
beam
and torsional
load.stiffne
of generalizedthe
parameters
m,parameters
mthickness
and
values
of correction
factors
in the
of,the
the
of membrane
the
shows
shows
that
that
at constant
at constant
values
values
of generalized
of of
general
, nthickness
Consequently,
generalized
, membrane
and
really
characterize
behaviour
Equation (5), (maximum error doesnt
exceed
5.3%)
do
not depend
onEquation
the
span (5), (5),
the
values
the
values
of
correction
of
correction
factors
factors
in
in
Equation
(maximum
(maximum
errorerrd
the system
when the of
membrane
eccentrically
fastened
to the thin-walled
closed
profile
support
and thickness
membrane,
stiffness
characteristics
ofand
thethickness
beam
and
load.
depend
depend
on
the
on
the
span
span
and
thickness
of
membrane,
of
membrane,
stiffness
stiffness
characteris
charac
beam.
Consequently, the generalized parameters m, m, n and really characterize
realc
Consequently,
Consequently,
the generalized
the generalized
parameters
parameters, , , ,and and really
100
the behaviour of the system when the membrane
eccentrically fastened to the
thin-walled closed profile support beam.
100 100
Table Table
1 - Parameters
thestress-strain
stress-strain
the membrane
1. Parametersof
of the
statestate
of the of
membrane
model. model
Parameter
, m
tm, m
tk, mm
p, kN/m
GIt, kNm2
EI , kNm2
1 case
3.24910
0,1
1.54310-6
0.00278
Beam bending moment

Flexural-torsional
bi-moment in the beam
sp, kNm
sp,
, mm
Deformation of the beam
CONCLUSIONS
-5
, rad
6
0.001
0.006
120
1250
2.29
0,741
1.02910-3
0,057
1539.15
24.0
5.35610-3
8.69210-4
2.51210-3
2 case
12
0.002
0.012
240
20000
73.2
6.009
1.04210-3
0,456
1538.81
48,4
5.38510-3
9.1810-4
2.65310-3
Error,%
1.3
0.022
0.54
5.3
CONCLUSIONS
On the basis of the numerical analyses presented, it is shown that the
redistribution of stresses in the structure depended on its geometrical
characteristics. Twisting of contour beam leads to the irregular displacements
of the membrane edge along the length of beam when membrane is fastened
eccentrically to the beam. This causes redistribution of the normal forces
in membrane through its interface with the boundary beam. In the middle
of the span forces perpendicular to the beam decrease and on the edges of
the boundary beam, they increase. Tangential forces occurred along the
membrane edge. Redistribution of forces in the membrane is reflected in the
stress-strain state of the beam: the bending moments decreases and additional
compression occurs.
REFERENCES
1.
Brudka Ya., Lubinski M. 1974 Legkie stalnye konstrukcii (Light steel

structures). Moscow: Stroiizdat, 125 p.
2.
Bychkov D. V. 1948 Raschet balochnyh I ramnyh system iz tonkostennyh

elementov. Moscow: Stroiizdat, 208 p.
3.
Eremeyev P.G., Tusnin A.R. 1990 Vliyaniye excentrichnogo kreplenia

membrani k opornomu konturu na perepaspredeleniye usilii v sisteme
// Stoitelnay mechanika I raschet sooruzhenii No.1, 813 pp.
4.
Farfel M.I. 2009 Razrabotka I issledovanie konstrukcii duskatnogo

bloka iz membrannich panelei Moscow, 176 p.
5.
Hideki Magara, Kiyochi Okamura, Mamory Kawaguchi 1984 An

analysis of membranes structures engineering. London, 112 pp.
6.
Kornishin M. S. 1968 Nelineinye zadachi teorii plastin I pologih

obolochek I metody ih resheniya Moscow: Nauka, 191 p.
7.
Taylor R.L., Oriate E., Ubach P. 2005 Finite element analysis of

membrane structures // Computational methods in applied sciences
Vol. 3, 47-68 pp.
8.
Trofimov V.I., Eremeyev P.G. 1990 Membrannie konstrukcii zdanii I

sooruzhenii. Chast 1. Moscow: Stroizdat, 248 p.
9.
Trofimov V.I., Eremeyev P.G. 1991 Membrannie konstrukcii zdanii I

sooruzhenii. Chast 2. Moscow: Stroizdat, 198 p.
10.
Vlasov V.Z. 1961 Thin-Walled Elastic Beams. Israel Program for

Scientific Translations, Jerusalem, 493 p.
77
78
NUMERICAL STUDIES OF Z-PURLINS

SUPPORTED BY SANDWICH PANELS
Olga Tusnina
PhD Student
Moscow State University of Civil Engineering, Moscow, Russia
ABSTRACT
In this paper numerical analysis of thin-walled Z-purlins in the cladding of buildings made
of sandwich panels are considered. The software used is MSC NASTRAN. The influence of
the mesh of elements on the results of calculations is studied. The results of numerical
analyses allowed developing the method for determining the stiffness of the connection
of the purlin with the sandwich panels.
INTRODUCTION
Nowadays Z-, C- and -purlins are widely used as supporting elements in roof
coverings of modular buildings based on the light steel skeletons. Sandwich
panels are used to cover the building envelope. A purlin-sheeting interaction
plays a significant role in the performance of these types of structures.
When loading the structure by distributed loads, the thin-walled coldformed purlin undergoes a restrained torsion because of the eccentric load
application. The sheeting does not allow free rotation of the purlin and its
supports. The bearing capacity of the purlin is significantly different for cases
when purlin is considered as a free beam and as a beam which is attached to
sheeting because of complex interaction of the purlin with sheeting. This is the
reason the purlin cannot be considered separately from the sheeting fastened
to it.
The type of fasteners also plays a role in the behaviour of the system. Today,
screws are used as fasteners of the sheeting to purlin as a rule. Alternatively,
rivets can be used as fasteners. Screws are installed through the thickness of
the sandwich panel (Figure 1). Rivets, however, form the connection between
the bottom flat sheet of the sandwich panel and the flange of the purlin
(Figure 2).
The investigation of the complex behaviour of such systems may be performed
experimentally or theoretically. One of the way of the implementation of
theoretical studies is to use numerical analysis. In this paper the results of the
numerical analysis of the system are presented.
79
Figure 1. Scheme of screw fastening

Figure 1. Scheme of screw fastening.
Figure 2. Scheme of rivet fastening.
Figure 2. Scheme of rivet fastening
.
BACKGROUND
Many researchers
have studied similar systems. Ciurej and Piekarczyk [1] carried out numerical
BACKGROUND
calculations using the MSC.Marc program. For the modelling of the purlin, the authors used plate
elements.Many
Also,researchers
laboratory tests
carried
out. systems.
The test model
section of the
have were
studied
similar
Ciurej was
and aPiekarczyk
[1] load-bearing
construction
of aout
roof
coveringcalculations
being in useusing
for instance
in lightweight
single-storey
buildings.
carried
numerical
the MSC.Marc
program.
For the
Three purlins
Z250ofwith
length
m were used
applied.
steel sheet
was fixed to each
modelling
the the
purlin,
the6 authors
plateTrapezoidal
elements. Also,
laboratory
tests were carried
out. The test
model wasresults
a section
the load-bearing
purlin. Comparison
of experimental
and numerical
was of
done.
The analysed beam-sheet
of amodelled
roof covering
being in
usethe
for use
instance
in lightweight
singlestructureconstruction
was effectively
as a whole
with
of shell
finite elements
for steel profiles
buildings.
Three purlins
Z250 with
the length
6 m wereFEM
applied.
and sheetstorey
roof and
brick elements
for insulation
blankets.
The elaborated
model of a roof
steel
was fixed
to each purlin.
Comparison
of this
experimental
coveringTrapezoidal
is appropriate
forsheet
simulating
the behaviour
of real
structures of
type.
and numerical results was done. The analysed beam-sheet structure was
effectively
a whole
the usepurlins
of shell
finite elements
for for different
In the articles
[4, 5]modelled
stability as
analysis
of with
thin-walled
restrained
by sheeting
steel profiles
andofsheet
and brickwas
elements
forout.
insulation
blankets.
Theis focused on
beam lengths
and type
the roof
cross-section
carried
Special
attention
elaborated
FEM model
a roof
covering
is appropriate
for simulating
the
modelling
the restrained
purlinsof in
general
purpose
finite element
program ABAQUS.
The
behaviour
of
real
structures
of
this
type.
comparison of results obtained with the use of analytical and numerical models is carried out. The
examples illustrate the importance of proper modelling the laterally restrained purlins.
In the articles [4, 5] stability analysis of thin-walled purlins restrained by
sheeting for different beam lengths and type of the cross-section was carried
out. Special attention is focused on modelling the restrained purlins in general
METHODS
purpose finite element program ABAQUS. The comparison of results obtained
with theanalyses
use of analytical
and specimens
numerical models
is carried
out. The
The numerical
are for the
for which
test results
[3]examples
were available. The
illustrate
of proper modelling
the laterally
purlins.
investigated
modelthe
is importance
the one recommended
by European
standard restrained
[2] for design
characteristics of
testing.
Numerical analyses are done with the use of program NASTRAN. The model (Figure 3) consists of
RUUKKI purlin Z200x2, 1 m long, sandwich panel SPC120/80PU and connectors screws GT6
175-5.5/6.3 mm. Purlins are modelled using PLATE finite elements (E=210000 MPa, =0.33).
104
80
METHODS
The numerical analyses are for the specimens for which test results [3] were
available. The investigated model is the one recommended by European
standard [2] for design characteristics of testing.
Numerical analyses are done with the use of program NASTRAN. The model
(Figure 3) consists of RUUKKI purlin Z200x2, 1 m long, sandwich panel
SPC120/80PU and connectors screws GT6 175-5.5/6.3 mm. Purlins are
modelled using PLATE finite elements (E=210000 MPa, =0.33). Sandwich
panel consists of 3 layers flat and trapezoidal metal sheets (PLATE finite
element, thickness 0.5 mm (E=210000 MPa, =0.33) and polyurethane core
(SOLID
finite of
element,
E=3.5
MPa,
=0.15).
Connection
purlin and
Sandwich panel
consists
3 layers
flat
and
trapezoidal
metalbetween
sheets (PLATE
finite element,
panel is modelled
with theand
use polyurethane
of RIGID finite
element,
which
thickness 0.5sandwich
mm (E=210000
MPa, =0.33)
core
(SOLID
finite element,
connects one
row of nodes
on the purlin
purlin with
on the
panel
E=3.5 MPa, =0.15).
Connection
between
and nodes
sandwich
panel
is(independent
modelled with the use of
is located
on connects
the purlin,one
connecting
degreesonofthe
freedom
all on the panel
RIGID finitenode
element,
which
row of nodes
purlin(DOF)
with are
nodes
linear
displacements).
Screws
are
modelled
as
BEAM
elements.
On
the
(independent node is located on the purlin, connecting degrees of freedom (DOF) are all linear
external side of sandwich panel screw-nut is modelled as 4 intersecting
displacements). Screws are modelled as BEAM elements. On the external side of sandwich panel
beam elements to distribute stresses and to realize more real behaviour of
screw-nut is modelled
as 4 intersecting beam elements to distribute stresses and to realize more real
construction.
behaviour of construction.
Figure 3. Investigated system.
Figure 3. Investigated system
Only linear analysis has been carried out. The values of force applied to the free flange were
obtained from
reported
tests. The
causesout.
the The
displacement
of the
free to
flange
along Y-axis at
Only
linear analysis
has force
been carried
values of force
applied
the free
obtained
from Analyses
reported tests.
the displacement
h/10 (h theflange
heightwere
of the
purlin).
have The
beenforce
donecauses
for different
directions of the load
the free
flange
h/10 (hpositions
the height
of the
purlin).
(uplift T1S,ofT2S,
gravity
along
T3P, Y-axis
T4P), at
different
of the
purlin
andAnalyses
different positions of
been done panel
for different
of the load (uplift T1S, T2S, gravity
the screws onhave
the sandwich
(Figuredirections
4).
T3P, T4P), different positions of the purlin and different positions of the
screws on the sandwich panel (Figure 4).
T1S
T2S
T3P
T4P
Figure 3. Investigated system

Only linear analysis has been carried out. The values of force applied to the free flange were
Proceedings ofof
thethe
METNET
in Lule
obtained from reported tests. The force causes the displacement
freeSeminar
flange 2013
along
Y-axis at81
h/10 (h the height of the purlin). Analyses have been done for different directions of the load
(uplift T1S, T2S, gravity T3P, T4P), different positions of the purlin and different positions of
the screws on the sandwich panel (Figure 4).
T1S
T2S
T3P
T4P
C
Figure 4. Schemes used in analysis.
Figure 4. Schemes used in analysis.

105

To determine the optimal size and number of elements, test numerical

were done for different parameters of the mesh for cases T1SA and
To determineanalyses
the optimal
size and number of elements, test numerical analyses were
T4PA. The mesh size of the element along the height of the purlin was equal to
different parameters
of the
1 cm, 2 cm,
4 cmmesh
and 8 for
cm. cases T1SA and T4PA. The mesh size of the element
height of the purlin was equal to 1 cm, 2 cm, 4 cm and 8 cm.
done fo
along th
RESULTS
RESULTS
The values of displacements in the direction of load acting in the top point
of the purlin web, the axial forces in the screws and the angel of the purlin
The values ofrotation
displacements
in thearedirection
in the
top point
of the purlin web, th
about it edge
comparedofinload
the acting
results for
different
meshes
(Figure
5). This comparison
is represented
in the Table
1. The
resultsitare
alsoare compared in th
axial forces in
the screws
and the angel
of the purlin
rotation
about
edge
represented
in graphs
(Figures
7, 8).comparison is represented in the Table 1. The result
results for different
meshes
(Figure
5). 6,
This
are also represented in graphs (Figures 6, 7, 8).
Figure 5. Scheme of purlin showing parameters of deformation.
Table 1 - Comparison of the results for different meshes

Size of the Displacement, mm Axial force in the
element,
screws, N
Angle of
rotation, rad
82
5. ofScheme
ofdifferent
purlinmeshes.
showing
Table Figure
1. Comparison
the results for
parameters of deformation.
Table 1 - Comparison of the results for different meshes
TableSize
1 - Comparison
of the results
different
of the Displacement,
mm forAxial
forcemeshes
in the
screws,
Sizeelement,
of the Displacement, mm Axial
forceNin the
m
T1SA
T4PA
T1SA
element,
screws,T4PA
N
18.5
23.4
18.5
23.4
m 1
T1SA
T4PA
T1SA
T4PA
17.7
22.4
17.7
22.4
1 2
18.5
23.4
18.5
23.4
4
15.9
20.4
15.9
20.4
2
17.7
22.4
17.7
22.4
14.0
18.6
14.0
18.6
4 8
15.9
20.4
15.9
20.4
8
14.0
18.6
14.0
18.6
Angle of
rotation,
Anglerad
of
T1SA
T4PA
rotation,
rad
0.098
T1SA 0.121
T4PA
0.094
0.098 0.117
0.121
0.086
0.094 0.108
0.117
0.071
0.086 0.100
0.108
0.071
0.100
Figure 6. Graph showing the dependence of displacement Y on the size of the element
106Y on the size of the element.
Figure 6. Graph showing the dependence of displacement
Figure 6. Graph showing the dependence of displacement Y on the size of the element
106

Figure 7. Graph showing the dependence of angle of rotation on the size of the element.
Figure 7. Graph showing the dependence of angle of rotation on the size of the element
83
Figure 7. Graph showing the dependence of angle of rotation on the size of the element
Figure 8. Graph showing the dependence of the axial force in screw N on the size of the element.
Figure 8. Graph showing the dependence of the axial force in screw N on the size of the element
As can be seen from the graphs, with the decrease in the element size and corresponding increase in
As can
be the
seenvariation
from theof
graphs,
with the decrease
in the
element
size
the density of the
mesh,
the considered
parameters
reduces
and
theand
results converge
increase
of the mesh,
of theIn such a mesh
to some limit. corresponding
We can conclude
thatina the
sizedensity
of the element
of 2 the
cm variation
is sufficient.
considered parameters reduces and the results converge to some limit. We
quite accurate results
are obtained and the dimension of the problem is still not as large as with the
can conclude that a size of the element of 2 cm is sufficient. In such a mesh
use of 1 m elements.
Soresults
we have
sufficiently
accurate
results
with
reasonable
quite accurate
are obtained
and the
dimension
of the
problem
is still time spent to
analyse the FEnot
model.
Based
on
this
conclusion,
the
analyses
of
all
other
schemes
as large as with the use of 1 m elements. So we have sufficiently accurate (T2SA, T2SB,
T2SC, T3PA, T3PB,
have beentime
carried
these
of theon
mesh.
results T3PC)
with reasonable
spentout
to with
analyse
the parameters
FE model. Based
this
conclusion, the analyses of all other schemes (T2SA, T2SB, T2SC, T3PA, T3PB,
T3PC)
have been of
carried
out with these[3]
parameters
of the mesh.
In the Table 2 the
comparison
the experimental
and numerical
results is represented.
Table 2 the comparison
of theresults
experimental
[3] and
results
Comparison ofIn
thetheexperimental
and numerical
shows that
the numerical
error between
NASTRAN and
is represented.
test results does not exceed 25%. Thus, the results, obtained with assumed mesh can be used for
further studies Comparison
and determination
of the supporting spring stiffness.
of the experimental and numerical results shows that the
error between NASTRAN and test results does not exceed 25%. Thus, the
results, obtained with assumed mesh can be used for further studies and
determination of the supporting spring stiffness.
107

84
Table 2. Displacement
of free flange of
andfree
error %flange
of numerical
compared
with laboratoryanalysis
tests results.
Table
2. Displacement
andanalysis
erroras %
of numerical
as compared with laboratory tests results
Displacement, mm
Force,
Scheme
Error, %
N
NASTRAN
Test
T1S
323.96
17.7
20
11.5
T2S
348.10
20.5
20
2.5
T3P
308.09
18.6
20
7
T4P
393.42
22.4
20
12
T1S
319.39
17.7
20
11.5
T2S
307.64
18.5
20
7.5
T3P
346.38
20.2
20
1
T4P
344.91
17.4
20
13
T1S
233.02
15.9
20
20.5
T2S
211.2
15.3
20
23.5
T3P
244.16
18
20
10
T4P
278.52
19.8
20
1
CONCLUSIONS
The compiled CONCLUSIONS
numerical model in NASTRAN is valid for studies of the work for this kind of
structures. With the use of this numerical model the engineering method of determination of the
The compiled numerical model in NASTRAN is valid for studies of the work for
supporting spring
stiffness was developed in Reference [6].
this kind of structures. With the use of this numerical model the engineering
method of determination of the supporting spring stiffness was developed in
Reference [6].
REFERENCES
1.
2.
3.
4.
5.
6.
Ciurej H.,REFERENCES
Piekarczyk M., Pieciorak E. 2010 Experimental research and integrated FEM
modelling for sheet roof covering with thermal insulation// Mechanics and Mechanical
1.
Ciurej H., Piekarczyk M., Pieciorak E. 2010 Experimental research
Engineering, Vol. 14, No. 2, pp 183-192.
and integrated FEM modelling for sheet roof covering with thermal
EN 1993-1-3:2004
Eurocode
3: Design
of steel Engineering,
structures. Vol.
Part14,1-3:
rules insulation
// Mechanics
and Mechanical
No. General
2,
Supplementary pp
rules
for
cold-formed
members
and
sheeting.
183192.
Kujawa M., Werochowski W., Urbaska-Galewska E. 2008 Restraining of the cold-formed Z2. sandwich
EN 1993-1-3:2004
Eurocode
Design ofPoland,
steel structures.
purlins with
panels Final
report,3: Gdansk,
126 p. Part 13:
General rules A.
Supplementary
rules braced
for cold-formed
members
and
Rzeszut K., Czajkowski
2011 Laterally
thin-walled
purlins
in stability
th
sheeting.
problems//Proceedings of the 19 International Conference on Computer Methods in
Mechanics, Warsaw, Poland, pp 202-203.
3.
Kujawa M., Werochowski W., Urbaska-Galewska E. 2008 Restraining
Rzeszut K., Studzinsky
R. 2010 Stability analyses of thin-walled laterally braced purlins//The
of
the cold-formed Z-purlins with sandwich panels Final report,
10th International
Conference
Modern
Building Materials, Structures and Techniques,
Gdansk, Poland, 126
p.
Vilnius, Lithuania, pp 758-762.
Tusnina O.
Design
of thin-walled
Z-purlins
with sandwich
4. 2013
Rzeszut
K., Czajkowski
A. 2011
Laterallyconnections
braced thin-walled
purlins in panels in roof
th
stability
problems
//
Proceedings
of
the
19
International
Conference
made by rivets// Design, Fabrication and Economy of Metal Structures, International
on Computer
MethodsHungary,
in Mechanics,
Warsaw, Poland, pp 202203.
Conference Proceedings,
Miskolc,
pp 157-163.
5.
Rzeszut K., Studzinsky R. 2010 Stability analyses of thin-walled

laterally braced purlins // The 10th International Conference Modern
Building Materials, Structures and Techniques, Vilnius, Lithuania, pp
758762.
6.
Tusnina O. 2013 Design of thin-walled Z-purlins connections with

sandwich panels in roof made by rivets // Design, Fabrication and
Economy of Metal Structures, International Conference Proceedings,
Miskolc, Hungary, pp 157163.
85
86
Open Topics
87
88
PROBLEMS OF FUNCTIONING AND ELABORATION

OF THE MASS HOUSING DEVELOPMENT AREAS
OF THE CITIES AND SETTLEMENTS
S.I. Matreninskiy
Candidate of Technical Sciences, Associate Professor
E.M. Chernyshov
Academician of the Russian Academy of Architectural and Construction Sciences,
Doctor of Technical Sciences, Professor
V.Y. Mischenko
Doctor of Technical Sciences, Professor
Voronezh State University of Architecture and Civil Engineering, Russia
ABSTRACT
In a number of cities and settlements the existing condition of mass housing development
areas, including numerous buildings and constructions, engineering and network
infrastructure, road and transport communications etc. demands reconstruction and
conformation with modern norms. The purpose of this project is elaboration of the system
model of mass housing development areas, and also bases of methodological approach
to reorganization of these areas with providing comfortable conditions for residence and
life activity of the population.
The suggested approach to functioning and elaboration of the mass housing development
area includes 2 levels of specification of description of the given problem. The first level
philosophical or cognitive-theoretic the verbal description of a plan or a concept, with
formation of structure and composition of the hierarchical system with definition of the
aim of its development and reorganization which is to provide favorable conditions for
residence and life activity of the society. The second level introduction of the system and
methods of its reorganization in the scientific-research language in the form of models
such as generation analysis choice for determination of the action variants on
reconstruction of territories, as well as variants of technical and technological conditions
providing their implementation.
INTRODUCTION
In a number of cities and settlements, the established condition of
mass housing development areas, including numerous buildings and
constructions, branched engineering and network infrastructure, road and
transport communications etc., demands reorganization and conformation
in compliance with modern norms and life activity of the population of these
areas. Thus, for making effective administrative as well as organizational
and technological decisions on maintenance and reorganization of the mass
housing development areas including a great number of various objects, it is

necessary to elaborate a system model of these mass housing development
areas as well as the bases of methodological approach to reorganization of
these areas, providing comfortable conditions for accommodation and life
activity of the population of these areas.
Before choosing a certain alternative of modeling of mass housing development
areas of the cities and settlements, and also development of methodological
bases of decision-making on reorganization of these areas it is necessary
to make a thorough analysis of specific features of the concrete city or its
part. Such analysis of the reality must be made on the basis of the modern
theory and contain not only statistical data, determinative, comparative to
normative, dimensional, constructive, ecological, social and other factors,
but also include the methodology of making effective administrative and
production decisions about reconstruction of mass housing development
areas.
Mass housing development area is considered as a component part of
the city intended for habitation and life activity of people, according
to the modern definition of a city. A city is one of the types of social and
dimensional organizations of population, arising and developing on the
basis of concentration of industrial, scientific, cultural, administrative
and other functions (Golyshev et al. 2006). That is why for mass housing
development areas as parts of the city or other center of population it is
already not enough to be seen only in a spatial refraction as areas of land, set
of buildings, constructions and streets. Mass housing development area as a
component part of the city is a complex multilevel system with its own laws
and peculiarities of existence.
METHODS
Realizing the system approach to existence and development of the city and
its component parts, it is reasonable to see the mass housing development
area as system complex city planning formation hereinafter referred to as
CPF. In the system analysis for representation of complex systems, one of
which is definitely CPF, stratified representation of such systems is normally
used (Mesarovich et al. 1973). According to this approach, the system is set
by a family of models, each of which describes its behaviour from the point of
view of a certain abstracting level.
Originally, in the most general, fundamental understanding, at the
semantic, notional level CPF is represented in philosophic or theoreticalcognitive language as a verbal description of a project or conception. Thus,
system complex city planning formation (CPF) can be defined as the set of
interconnected, controllable, conditioned by existing economic and material
and technical potential of the given territory, spatial, architectural and
building, engineering decisions of the environment of population groups
(society), providing favourable conditions for habitation and life activity
89
90
of people. CPF component is implied as its component part, characterizing

functional content and purpose of system complex city planning formation.
Introduction of the definition of CPF with formation of its structure is necessary
for dynamic evaluation and analysis of different parts and elements of a
modern human environment, observing interconnection of CPF components
and their interaction with environment, applying system theory and system
analysis for solution of the problems of functioning and elaboration of mass
housing development areas.
Thus, for further analysis of CPF and determination of possible rational lines
of its development considering unavoidable contradictions between separate
components it is necessary to represent the system in the language of selected
scientific theory in the form of various kinds of models which help deeper
understanding and discovering of the main idea of the system Figure 1.
For this purpose, we make a description of the system in scientific research
language of the theory of complex systems synthesis, when CPF is represented
as a complex multilevel hierarchical system interacting with the environment
under influence of which it is being formed. With that, CPF is defined as the
set of interconnected components architectural and constructional (AC),
engineering and network (EN), engineering and transport (ET), territorial
and spatial (TS), functioning in environmental conditions for the purpose of
providing the required favourable conditions for habitation and life activity of
the society on the given mass housing development area in accordance with
Figure 1. In this case, the environment implies existing natural and ecological
conditions, generated material and technical potential of the certain territory,
existing managing administrative and financial system and the society of the
given territory.
defined as the set of interconnected components architectural and constructional (AC),

engineering and network (EN), engineering and transport (ET), territorial and spatial (TS),
functioning in environmental conditions for the purpose of providing the required favourable
conditions for habitation and life activity of the society on the
given mass
development
Proceedings
of thehousing
METNET Seminar
2013 in Lule
area in accordance with Figure 1. In this case, the environment implies existing natural and
ecological conditions, generated material and technical potential of the certain territory, existing
managing administrative and financial system and the society of the given territory.
Figure
complex
city planning
formationformation
CPF in the form
system
interacting
Figure1.1System
System
complex
city planning
CPF ofinmultilevel
the formhierarchical
of multilevel
hierarchical
system interacting with the environment
with the environment.
RESULTS
Realizing the methods of system approach, the CPF components are considered as subsystems of
the First level with their subsequent division into subsystems of the Second level (common objects)
RESULTS
which
in turn consist of elements - private objects, making in total the subject structure of CPF
according to Figure 2.
Realizing the methods of system approach, the CPF components are

considered as subsystems of the First level with their subsequent division
111
into subsystems of the Second level (common
objects) which in turn consist

of elements private objects, making in total the subject structure of CPF
according to Figure 2.
Engineering and network component (EN) pipeline, wire and wireless
communications included into the arrangement of residential community
involving networks of water supply, water drain, heat supply, gas supply,
electric supply, communication and alarm, and also constructions designed
for service of engineering networks and cooperation with them. These include,
engineering and transport component (ET) buildings and constructions for
service of public and individual transport (depots, garages, parking places,
car-washing facilities, auto repair shops), transport communications (roads,
bridges, trestles etc.), included into the arrangement of residential community.
Other components are, territorial and spatial component (TS) the set of
undeveloped and modified areas of land having particular functionality or
being a reserve for beautification and development of the given residential
91
heat supply, gas supply, electric supply, communication and alarm, and also constructions designed
for service of engineering networks and cooperation with them. These include, engineering and
transport component (ET) buildings and constructions for service of public and individual
92
transport (depots, garages, parking places, car-washing facilities, auto repair shops), transport
communications (roads, bridges, trestles etc.), included into the arrangement of residential
community. Other components are, territorial and spatial component (TS) the set of undeveloped
and modified areas of land having particular functionality or being a reserve for beautification and
and constructional component (AC) the set
developmentcommunity.
of the givenArchitectural
residential community.
Architectural and constructional component (AC)
of residential, public and administrative buildings and constructions
the set of
residential,
public
and
administrative
and constructions
(monuments, stadiums, fountains etc.) locatedbuildings
in the territory
of the given (monuments,
stadiums, fountains
etc.)
located
in
the
territory
of
the
given
residential
community.
residential community.
Figure 2. The objective structure of the city planning formation as a system including its components and objects.
112

As it was already mentioned, one of the most important system design rules
is that when designing the system it is required to consider its interaction
with external environment, but the environment, in its turn, should also
be regarded as a system. In our case, the External environment system
includes the following subsystems: Natural and ecological conditions of CPF
the set of natural components (geographical coordinates, climate, waters,
soils etc.) forming natural basis of the given residential community. Material
and technical potential of CPF the set of building materials used in the given
region, industrial enterprises producing these materials and goods, building
organizations, construction and mounting organizations, branches serving
the construction, and also normative basis regulating construction of the CPF
objects, their use and maintenance. Administrative and financial system the
set of federal, regional, municipal authorities, regulating the relations among
the CPF components as well as the relations between the whole CPF and the
environment, and providing financing at the expense of the federal funds,
tax and voluntary subscriptions from different enterprises and organizations
(industrial, commercial, banking and financial, entertainment, service etc.),
including the monetary funds of the population of the given territory for
development and reorganization of CPF. CPF society the total community of
people (population) living in the territory of the given residential community
or connected with it by a certain kind of activity or way of life, which is
expressed by a system of stable social mutual relations.
As it was noted before, the aim of CPF functioning when interacting with
environment is creation of favorable conditions for life activity of the society.
At the same time it is necessary to consider the permissible level of resource
expenditure. This aim as it is formulated is quite difficult to define on the
number scale which makes it necessary to introduce the goal achievement
efficiency index or criteria the factors for evaluation of correspondence
between functioning of the system and the given result. With regard to CPF it
is reasonable to take as efficiency index such global factors as comfort and
resource consuming which in their turn depend on the particular factors.
Interconnection of CPF efficiency indices is shown on Figure 3.
Figure 3. Composition and structure of CPF efficiency indices.
93
94
Comfort is characterized by convenience of the habitation and life activity

environment, moral and physical depreciation of this environment. Resource
consuming is characterized by material and technical consumption, energy
consumption and labour expenditures, measured in money equivalent, for
providing current functioning or rearrangement of the habitation and life
activity environment in order to achieve the required level of its comfort.
The convenience of habitation and life activity of the society in the given
territory is evaluated by expert method, in accordance with corresponding
variants of realization. Moral depreciation is a gradual (in time) lag
of consumer properties of the earlier created urban funds (buildings,
engineering equipment, infrastructure etc.) comparing to achievements of
the modern operating, architectural and technical level in the given area.
Moral depreciation is measured in percentage terms. Determination of moral
depreciation consists in calculation of costs for elimination of depreciation,
concerning replacement cost of the examined object, in this instance the CPF
with all its subsystems. At that, as a variant of CPF with maximal comfort,
without moral depreciation it is possible to consider the sketch of architectural
and constructional project of mass housing development area developed
in accordance with modern achievements of science, technology, ecology,
architecture, that is in some way elite, close to ideal project. Physical
depreciation is a gradual loss of the initial qualities of urban funds (buildings,
engineering equipment, infrastructure etc.) in the process of their natural
aging and deterioration, with decrease of consumer properties, evaluated by
correlation of the cost of objectively necessary repair activities, eliminating
damages of the system, construction, element and their replacement cost.
Possible variants of actions concerning further functioning or reconstruction
of CPF, including all its material components and subsystems after rating
them by corresponding efficiency indices, are as follows:
1.
Maintenance of service properties current repair.
2.
Repair of the separate objects, areas and elements.
3.
Capital repair.
4.
Reconstruction and modernization.
5.
Construction of additional objects, areas and elements.
6.
Demolition and dismantling of old objects, areas and elements with

construction of new ones.
as follows:
1. Maintenance of service properties current repair.
2. Repair of the separate objects, areas and elements.
3. Capital repair.
95
4. Reconstruction and modernization.
5. Construction of additional objects, areas and elements.
6. Demolition and dismantling of old objects, areas and elements with construction of new
ones.
Let us look at the structure and components of the concept action with
respect
given variants
in accordance
with
Figureaction
4.
Let us look
at to
thethe
structure
and components
of the
concept
with respect to the given
variants in accordance with Figure 4.
Action
Subject of action
Object of
action
Aim of action
Order of action
implementation
Figure
4 Structure of the concept action.
Figure 4. Structure of the concept
action.
Subject of action are organizations and enterprises directly implementing examination, architectural
and building
projecting
reconstruction
Object
of action
are CPF components
Subject
of actionand
areCPF
organizations
and works.
enterprises
directly
implementing
which are
liable
to
reorganization
architectural
and
constructional
component
(AC), engineering
examination, architectural and building projecting and CPF reconstruction
and transport
component
(ET),
territorial
and
spatial
component
(TS),
engineering
works. Object of action are CPF components which are liable to reorganizationand network
component
(EN) also including their constituent subsystems and elements. Aim of action is the
architectural and constructional component (AC), engineering and transport
reorganization
of CPF, including its components, with achievement of the established level of
component (ET), territorial and spatial component (TS), engineering and
network component (EN) also including
114 their constituent subsystems and
elements. Aim of action is the reorganization
of CPF, including its components,
with achievement of the established level of comfort at the minimal resource
expenses. Order of action implementation is the complex of technical and
technological solutions, providing the achievement of the set aim.
Thus, the system of decision-making on reorganization of CPF has 3 levels

of specification of description or strata according to the terminology
accepted in the system analysis as shown in the picture 5 (Mesarovich et al.
1973). Concepts such as subject, object, aim of action are included into the first
level of specification of description of decisions functional stratum, where
the decisions (R1) are made on the establishment of the concrete content of
these concepts on all the components and subsystems of CPF. On the basis of
the first functional stratum, a different variant of actions on reorganization or
functioning can be chosen for the different CPF components depending on the
established efficiency indices, such as comfort and resource intensity and, in
the general case, the quantity of diverse combinations of variants can turn out
to be very considerable. For example, for the territorial and spatial component
(TS) the variant of actions such as construction of additional objects, sites
and elements can be accepted, for engineering and network component (EN)
it can be general overhaul, for engineering and transport component (ET)
and architectural and constructional component (AC) reconstruction and
modernization.
The next concept order of action implementation has two levels of
specification of description (two strata) technical and technological. The
second level of specification of description of decisions is technical level,
where the general technical decisions (R2) on achievement of the chosen
aims are formed. For example, the reconstruction and modernization
decision on architectural and constructional component, accepted in the
96
functional stratum, in its turn, in the technical stratum can be realized by

the development of design decisions on superstructure of floors, extension of
additional sections, change of inner layout, etc.
The third level of specification of description of decisions is technological that is definition of thorough and detailed design and technological decisions
(R3), providing the implementation of technical decisions.
Thus, generally, the values of the global efficiency indices such as comfort
(1) and resource intensity (2) depend on the chosen decisions R1, R2,
R3. In this case, the problem of CPF synthesis consists in definition of such
decisions on the functional, technical and technological strata which provide
minimal resource expenses for achievement of the established comfort.
Formally the problem is put down as follows. To find
(R1, R2, R3) optimal = arq min 2 (R1, R2, R3)

<R1, R2, R3>
At the restrictions

1 (R1, R2, R3) K required
R1 RD1 set of admissible actions.
R2 RD2 set of admissible technical decisions.
R3 RD3 set of admissible technological decisions.
At an estimation of CPF condition as a system, its global and private

efficiency indices cannot be expressed in the form of analytical dependences
on the considered variables (R1, R2, R3), that would allow the forecast
of the developments and to introduce recommendations on functioning
and reorganization of mass housing development areas on the basis of
corresponding mathematical models. To solve the problems of synthesis of
complex systems of this kind, nowadays multistage technological schemes of
their solution are developed (Mesarovich et al. 1973).
The decision-making process at any stratum is divided into 3 subprocesses:
generation of variants, estimation of variants analysis and choice. These
subprocesses are considered as decision-making layers.
Figure 5 - Technological
scheme of decision-making
on reorganization
of massdevelopment
housing development
Figure 5. Technological
scheme of decision-making
on reorganization
of mass housing
areas. areas
116

On the generation layer operations on formation of admissible variants of

actions on reorganization of each component and corresponding subsystems
of CPF, technical decisions on realization of each chosen action, technological
decisions on realization of each chosen technical decision, are realized.
On the analysis layer reception of quantitative and qualitative values of the
indices characterizing CPF efficiency on a set of chosen admissible variants
of actions, and chosen admissible variants of technical and technological
decisions, is carried out.
The choice layer includes operations of the search of preferences (in the
meaning of set criteria) of action variants, technical and technological
decisions, which determine rational composition and structure of CPF, and
also values of component, subsystem and element characteristics.
The technology of CPF synthesis as determination of variants of actions and
decisions on functional, technical and technological strata consists in time
analysis of the set of formal and heuristic operations on compromise decisionmaking on substantiation of CPF structure for the purpose of receiving the
necessary set of working documentation, regulating its reorganization.
At the first stage that is the draft design of reorganization of a mass housing
development area the functions (aims) of necessary actions concerning each
component and object of CPF are defined.
97
98
At the second stage architectural and building project technical decisions

which are necessary for achieving the generated aims are determined.
At the third stage concrete working drawings and documentation for
direct implementation of reorganization the detailed constructive and
technological decisions, realizing the chosen technical decisions, are defined.
As a result, the technological scheme of decision-making on reorganization
of mass housing development areas will look as it appears in the Figure 5.
Thus, the offered methodology allows to form a great number of admissible
variants of actions on reorganization of CPF, as well as to estimate the
efficiency of each variant by the series of basic indices and to make a
reasonable choice of an expedient variant by the accepted criteria.
In Figure 6 there is an example of determining the variants of actions on
reorganization of the CPF object Residential building with specification
on 3 strata, in assumption that by the set criterion minimal resource
expenses at the given comfort the action such as Reconstruction and
modernization, and technical decisions such as Superstructure of floors
and Built-in framework are chosen.
CONCLUSION
The introduced approach to the functioning and elaboration of mass housing
development areas as system complex city planning formations includes two
levels of detailed description specification of the given problem.
The first level philosophical or cognitive-theoretic the verbal description
of a plan or a concept, with formation of structure and composition of the
mass housing development areas as system complex city planning formation
CPF, with definition of the aim of its development and reorganization
which is to provide favourable conditions for residence and life activity of
the society.
The second level introduction of CPF system and methods of its
reorganization in the scientific-research language in the form of models
such as generation analysis choice for determination of the action
variants on reorganization of CPF, as well as variants of technical and
technological decisions responsible for their implementation.
The second level introduction of CPF system and methods of its reorganization in the scientificresearch language in the form of models such as "generation
analysis
- choice"
for determination
Proceedings of the- METNET
Seminar
2013 in Lule
99
of the action variants on reorganization of CPF, as well as variants of technical and technological
decisions responsible for their implementation.
Figure 6. Example of determination of the action variants with specification on 3 strata on reorganization of the
Figure
- Example
of determination of the action variants with specification on 3 strata on
CPF
object 6Residential
building.
reorganization of the CPF object - "Residential building
REFERENCES
REFERENCES
Golyshev, A.V., Kolchunov, Vl. I. and Kolchunov, V.I. (2006), Architectural and Construction
Golyshev, A.V.,
Kolchunov,
Vl. I. and
Kolchunov, V.I. (2006), Architectural and
Encyclopedia.
Reference
Dictionary.
oscow.
Construction Encyclopedia. Reference Dictionary. oscow.
Mesarovich, ., Mako, D. and Takahara, I. (1973), Theory of hierarchical multilevel systems.

Mesarovich,
oscow:
Mir. ., Mako, D. and Takahara, I. (1973), Theory of hierarchical
multilevel systems. oscow: Mir.
118

100
THE BLUE BOX AND THE TECHNOLOGICAL

TRAJECTORIES IN CONTEXT
OF UNIVERSITY-BUSINESS COOPERATION (UBC)
Lauri Tenhunen
HAMK University of Applied Science, Hmeenlinna, Finland
Arto Ranta-Eskola
Rautaruukki Oyj, Finland
ABSTRACT
Matthews (1992) presents a concept of Blue Box for integrating technology into business
strategy. His analysis serves to find and manage the development routes in the inevitable
replacement of old technologies by new ones. Matthews also suggests a practical
approach, a conceptual Management of Technology (MoT) analysis framework, to enable
continuous and intensive discussion and decision-making between business managers and
technologists. This represents one of our approaches to university-business cooperation
(UBC).
The second approach in this paper is to analyse the UBC possibilities from the point of view
of the technological trajectories. A technological trajectory is defined, quite flexible, as the
pattern of conventional problem solving activity within a given technological paradigm
(Dosi, 1982).
When ideas, markets and professions are institutionalized, a technology development can
get stuck within one technological trajectory, and organisations as well as individuals
are unable to adapt ideas and innovation from outside.
In this article we develop a method for analysing alternative options for UBC.
INTRODUCTION THE BLUE BOX CONCEPT

Matthews (1992) argues in his analysis that a companys business strategy and
technology strategy cannot successfully be developed in isolation. Cooperation
with universities and research institutes is needed and an overall business
strategy should be formulated by taking technological considerations into
account. He illustrates a conceptual framework within a diagram detailing
the central issues together with questions to be asked during the development
process. He also outlines further tools and techniques to probe more specific
issues which can then be related to the overall framework.
101
Hakkarainen (2006) describes the original Mathews ideas as follows (Figure

1). The Blue Box covers the interphase development stage where original ideas
are gradually developed into commercial solutions (strategic options stage
in Picture 1). This stage in technological development would fruitfully be
accomplished within UBC.
Figure 1. Types of research and development projects (Matthews 1990 and Hakkarainen 2006).
Figure 1. Types of research and development projects (Matthews 1990 and

Hakkarainen 2006)
ON TECHNOLOGICAL TRAJECTORIES
ON TECHNOLOGICAL TRAJECTORIES
Kristinsson-Rao classification
Kristinsson-Rao classification
The Kristinsson-Rao classification of technological trajectories is based on
industry
life cycles. Theof
literature
argues that
industries is
gobased
through
The Kristinsson-Rao
classification
technological
trajectories
ondifferent
industry life cycles. The
stages. Like an organic life form, industries are viewed as going through what
literature arguesisthat
industries go through different stages. Like an organic life form, industries are
usually described as the four phases of introduction, growth, maturity and
viewed as going
through
what has
is then
usually
described
as theand
four
of introduction, growth,
decline. Each phase
its own
market structure,
firmphases
characteristics
maturity and decline.
Each phase
own
structure,
and firm
characteristics such as
such as innovation
ratehas
andthen
type its
vary
withmarket
these stages
(Kristinsson
and Rao
innovation rate 2013).
and type vary with these stages (Kristinsson and Rao 2013).
In life-cycle analyses, the volume of an industry is usually described by either
In life-cycle analyses,
the volume of an industry is usually described by either the number of firms
the number of firms (Klepper and Simons, 1997) or the sales volume of an
(Klepper and Simons,
or &the
sales volume
industry
& is
Rennhak,
2006). If the
industry1997)
(Menhart
Rennhak,
2006). Ifof
thean
sales
volume (Menhart
of an industry
the
sales volume of an industry is the unit of analysis the industry life cycle is the sum of product life
cycles within the industry, as shown in Figure 2.
literature argues that industries go through different stages. Like an organic life form, industries are
viewed as going through what is usually described as the four phases of introduction, growth,
maturity and decline. Each phase has then its own market structure, and firm characteristics such as
102
innovation rate and type vary with these stages (Kristinsson and Rao 2013).
In life-cycle analyses, the volume of an industry is usually described by either the number of firms
(Klepper and Simons, 1997) or the sales volume of an industry (Menhart & Rennhak, 2006). If the
unit of analysis the industry life cycle is the sum of product life cycles within
sales volume of an
isshown
the unit
of analysis
the industry life cycle is the sum of product life
the industry
industry, as
in Figure
2.
cycles within the industry, as shown in Figure 2.
Market
volume;
sales
industry life cycle
product life
cycle
life cycle of product

generations
Time
FigureFigure
2. 2.Industry
lifeandcycle
product
(Kristinsson
& Rao
Industry life cycle
productand
life cycles
(Kristinssonlife
& Raocycles
2013, based
on Menhart & Rennhak
2006). 2013,
based on Menhart & Rennhak 2006)
Following the definition of Dosi (1982), a technological paradigm is a model
120
and a pattern of solution of selected technological problems, based on
selected principles derived from natural sciences and on selected material
technologies.
Correspondingly a technological trajectory is defined by Dosi (1982) as the
normal problem solving activity determined by a paradigm, represented
by the movement of multi-dimensional trade-offs among the technological
variables which the paradigm defines as relevant.
In other words, even within the same paradigm there is some room for
choices, where evolution of the technology depends on the directions taken by
the development of the technology, i.e. the technological trajectory.
In many young industries there are two or more alternative technological
trajectories competing for dominance. Firms have an interest in seeing that
the technological trajectory they are currently most knowledgeable about,
would achieve dominance within the industry in longer run. As the industry
matures, some technology trajectories are chosen through a socioeconomic
process, possibly leading to one trajectory becoming dominant. Several
alternate possibilities can emerge from the selection process.
Kristinsson and Rao (2013) have mapped situations where several rival
trajectories might achieve roughly equal and stable market share and two
similar trajectories might compete over a long period for the dominance
while neither achieves it. In Table 1, these and other scenarios are divided
into six phases each representing a different technological environment for

the firms in the industry. There is no clear order in which these phases will
present themselves; one could easily imagine an industry moving from a
dominant technology trajectory to an environment of two rival trajectories
vying for dominance. However, in general a trend from ambiguity towards
clear dominance might be expected.
Table 1. Emergence of a dominant technology trajectory (Kristinsson and Rao 2013).
Technology
Relatedness
Several Competing
Trajectories
Dominance
Competition
Dominance
Related
A Range of Trajectories
Similar Trajectories
Clear Dominance
Unrelated
Ambiguity
Alternate Trajectories
Dominance
Kristinsson and Rao (2013) have introduced six phases of the development of
a dominant trajectory within an industry as follows:
Phase I Ambiguity
Few pioneer firms. Basic research is done in universities. The possible
future technological trajectory is highly uncertain. The range of possible
technological trajectories cannot be identified. Other factors such as possible
demand, institutional factors and competition are also highly uncertain.
Phase II A Range of Trajectories
Pioneers in applied research are followed by other organizations. Research
programs take place in the area. Gradually the area is populated by several
actors each working on their own technological design that often represent
totally different technological trajectories.
Phase III Alternate Trajectories
As the industry matures, stabilizes and the major players become known.
Typically two main trajectories turn up. These alternative trajectories are
often incompatible and therefore there appears a competition for dominance.
Phase IV Similar Trajectories
The technological development can be described as having two main outcomes
which are related and therefore do not pose so much uncertainty to the firms
as in a regime of alternative trajectories.
Phase V Dominance
One technological trajectory has achieved dominance above all others. More
than half of the innovations in the industry belong to this trajectory. The
competing trajectories are relegated to minor roles.
Phase VI Clear Dominance
The majority of the remaining trajectories are related to the dominant one,
thus giving a very focused and homogenous R&D space.
103
One technological trajectory has achieved dominance above all others. More than half of the
innovations in the industry belong to this trajectory. The competing trajectories are relegated to
minor roles.
104
Phase IV Clear Dominance

The majority of the remaining trajectories are related to the dominant one, thus giving a very
focused and homogenous R&D space.
Pavitt-Hedge classification
Pavitt-Hedge classification
The university-industry innovation cooperation in the Pavitt-Hedge
The university-industry
innovation
cooperationtrajectories
in the Pavitt-Hedge
classification
technological
classification
of technological
take place
mainly in ofcompanies
trajectories take
place
mainly
in
companies
which
are
specialized
(e.g.
fabricated
metals,
which are specialized (e.g. fabricated metals, machinery, instruments,
machinery, instruments,
electrical,
electronics)
or
science
based
(e.g.
pharmaceuticals,
drugs,
electrical, electronics) or science based (e.g. pharmaceuticals, drugs,
chemicals, microelectronics)
Table
2.
chemicals, microelectronics) Table 2.
Table 2. Pavitt`s
(1984)
Hedge`s
(2004)
taxonomy
and expectations
Table 2.
Pavittsand
(1984)
and Hedges
(2004)sectorial
sectorial taxonomy
and expectations.
Core sectors
Firm size
Supplier
dominated
Agriculture,
food, wood
based,
textiles,
rubber &
plastics
Small
Scale
intensive
Automotive,
transport
equipment
Large &
Medium
Specialized
suppliers
Fabricated
metals,
machinery,
instruments,
electrical,
electronics
Science based
Information
intensive
Pharmaceuticals,
drugs,
chemicals,
microelectronics
All services
Small
Medium & Large
Small
Product &
Process
Primary
services
education,
transportation,
finance,
tourism,
health
Small &
Medium
Organisational
Type of
innovation
Process
Process
Product
Strategy
Cost
affectivity
Either cost
affectivity
(price) or
Differentiati
on (quality)
Product &
Process
Differentiation
(quality,
performance,
customization)
Differentiation,
Focus strategy
(innovation,
quality)
Differentiation
(quality, quick
delivery,
customization)
Mixed
External
sources of
innovation :
cooperation
Suppliers
and users
Suppliers
and users
Universities
and users
Universities and
users
Users
Financers
and users
Comparing the
Pavitt-Hedge classification to the Kristinsson-Rao classification, the following
Comparing the Pavitt-Hedge classification to the Kristinsson-Rao classification,
hypothesis could be made: The most important role of universities in creating innovations and
the following hypothesis could be made: The most important role of universities
developing new technology are in the early stages of the industry life cycle, especially in the area of
in creating innovations and developing new technology are in the early stages
specialized suppliers and science based industries.
of the industry life cycle, especially in the area of specialized suppliers and
science based industries.
122
Castellacci classification
To analyze technological paradigms, Castellacci (2008) has arranged industries
in four sectorial categories and eight sub-groups (Table 3):
1.
Advanced knowledge providers

Knowledge-intensive business services (e.g. Software; R&D,
Engineering, Consultancy)
Specialized suppliers manufacturing (e.g. Machinery, Instruments)
2.
Mass production goods

Science-based manufacturing (e.g. Electronics)
Scale-intensive manufacturing (e.g. motor vehicles)
3.
105
Supporting Infrastructure services

Network infrastructure services (e.g. Telecommunications, finance)
Physical infrastructure services (e.g. Transport, wholesale trade)
4.
Personal good and services

Supplier dominated goods (e.g. Textiles and wearing)
Supplier dominated services (e.g. Hotels and restaurants).
Table 3. Main characteristics of various sectorial groups in the Castellacci taxonomy (Castellacci, 2008).
Sectoral
category
Advanced
knowledge
providers
Mass
production
goods
Supporting
Infrastructure
Services
Personal
goods
and services
Sub-groups
within
each category
Typical
core sectors
KnowledgeSoftware: R&D;
intensive
Engineering;
business services Consultancy
Major function and

relationship to
technological paradigms
Technological
regimes
Technological
trajectories
The supporting
knowledge base
of the ICT paradigm
Opportunity levels: very high

External sources: users and Universities
Appropriability: Know-how; copyright
Dominant firm size: SMEs
Type of innovation: new services;

organisational innovation
Innovation expenditures and strategy:
R&D; training; cooperations
Specialised
suppliers
manufacturing
Machinery;
Instruments
The supporting
knowledge base
of the Fordist paradigm
Opportunity levels: high

External sources: users
Appropriability: patents, design know-how
Type of innovation: new products

R&D; acquisition of machinery;
sotftware purchase
Science-based
manufacturing
Electronics
The carrier industries

of the ICT paradigm
Opportunity levels: very high

External sources: Universities and users
Appropriability: patents, design; copyright
Dominant firm size: large
Type of innovation: new products;

organizational innovation
R&D; cooperations
Scale-intensive
manufacturing
Motor vehicles
The carrier industries

Opportunity levels: medium

External sources: suppliers and users
Appropriability: design; processy secrecy
Type of innovation: mixed products and

process innovation
R&D; acquisition of machinery
Network
infrastructure
services
Telecommunications;
Finance
The supporting
infrastructure
of the ICT paradigm

External sources: suppliers and users
Appropriability: standards; norms; design
Type of innovation: mixed process,

service and organizational innovation
R&D; acquisition of software; training
Physical
infrastructure
services
Transport;
Wholesale trade
The supporting
infrastructure
Opportunity levels: low

External sources: suppliers
Appropriability: standards; norms; design
Type of innovation: process

acquisition of machinery and software
Supplierdominated
goods
Textiles and
wearing
Supplierdominated
services
Hotels and
restaurants
They enhance
the quality
of final products and
services by acquiring
and embodying
technologies related
to different paradigms

External sources: suppliers and end users Innovation expenditures and strategy:
Appropriability: trademarks; design know-how
acquisition of machinery
Opportunity levels: low
External sources: suppliers
Appropriability: non-technical means

acquisition of machinery; training
In Table 3, university cooperation turns up only in:

Knowledge-intensive business services (e.g. Software; R&D,
Engineering, Consultancy), where innovations are expected to be
organizational or concern new services
Science-based manufacturing (e.g. Electronics), where innovations are
expected to be organizational or concern new products.
106
Formulating a Model
The types of University-Industry cooperation depend on the degree of the
maturity of the industry. See Figure 3.
Technology
Concentration
(number of
competing
trajectories)
The maturity
of an industry
Basic
Research;
Radical
Innovations
Mutual
Research
Programs
Applied
Research;
Minor
Innovations
Typical University-Industry Cooperation

Figure
3. Supplied
University Services
by Industry Life-Cycle.
Figure
3. Supplied
University
Services
by Industry Life-Cycle
On the other hand, some companies are in a continuous stage of quick
On the other hand,

some companies
are The
in anature
continuous
stage of quick
development
(e.g.
development
(e.g. electronics).
of each competing
trajectory
is
electronics). Thesupposed
nature oftoeach
competing
trajectory
is
supposed
to
be
deep
and
the
stage
of
basic
be deep and the stage of basic research and radical innovation
possibilities
covers
a long period.
companies,
services arewhose services
research and radical
innovation
possibilities
coversAlso
a long
period. whose
Also companies,
knowledge-based,
look
for
strong
partnerships
fostering
their
activities.
are knowledge-based, look for strong partnerships fostering their activities.
These qualitative considerations above give us a possibility to formulate
These qualitative aconsiderations

above give us a possibility to formulate a model to connect types
model to connect types of supplied university R&D services to typical
of supplied university
R&D
services
toThis
typical
cooperating
This is stated in Picture 4.
cooperating companies.
is stated
in Figure companies.
4.
Order of
Innovativiness
Services
Supplied
Remarkable
Scientific
Results
Basic
Research;
Radical
Innovations
Creating
Business
Knowledge
Mutual
Research
Programs
Producing
Pieces of
Information
Applied
Research;
Minor
Innovations
Making
Recruiting
Riskless
Daily Work
by Students
or Researchers
in Exchange
107
Scetch of
Typical
Cooperating
Companies
Science-based
Manufacturing
Knowledge
-intensive
Business
Services
All Companies
in the Industry
Figure 4. Supplied University R&D Services by Type of Industry.
Consideration of alternative trajectories in company R&D&I

Consideration of alternative trajectories in company R&D&I
Rautaruukki Oyj (Ruukki) solution, product and production process

development (solution and product management or SPM) is a five stage process
Rautaruukki Oyj (Ruukki) solution, product and production process development (solution and
with go/stop decisions made between the stages, Figure 5. Each product group
product management or SPM) is a five stage process with go/stop decisions made between the
and key production process has a cross-functional steering group composed
stages, Figure 5. Each product group and key production process has a cross-functional steering
of main stakeholders. These groups are responsible for the overall supervision
group composed of main stakeholders. These groups are responsible for the overall supervision of
of the development portfolio and for generating ideas for new projects in their
the development
portfolio and for generating ideas for new projects in their area.
area.
Figure
Ruukki SPM-process
for solution,
and production
Figure
5.5.Ruukki
SPM-process
forproduct
solution,
productprocess
anddevelopment.
production process development.
The SPM process is very efficient in managing production process and new product development.
108
Figure 5. Ruukki SPM-process for solution, product and production process development.
SPM
process in
is managing
very efficientproduction
in managing
production
and
The SPM process isThe
very
efficient
process
and process
new product
development.
new product development. As the steering groups are composed of persons
As the steering groups are composed of persons responsible of e.g. development, sales, marketing,
responsible of e.g. development, sales, marketing, production, quality and
production, quality production
and production
fast on
decisions
onfreezing
launching, freezing or
control,control,
they canthey
makecan
fastmake
decisions
launching,
terminating projectsorand
allocating
resources
to projects
with highest
economic
potential. However,
terminating
projects
and allocating
resources
to projects
with highest
potential.
However,
this also
represents aexclusively
problem as there
is a existing
risk
this also represents economic
a problem
as there
is a risk
of adhering
to the
technology
of
adhering
exclusively
to
the
existing
technology
trajectories.
Therefore
a
trajectories. Therefore a concurrent development process for radical innovations is included,
concurrent development process for radical innovations is included, Figure
Figure 6. In this process
new
disruptive
innovations
are explored.
mostly related to new
6. In this
process
new disruptive
innovations
are explored.They
They are
are mostly
collaboration networks,
or businessnetworks,
models services
than products
andmodels
solutions.
relatedservices
to new collaboration
or business
than The ideas are
productsto
and
solutions.
Theapplication
ideas are developed
thisthen
process
to concept and
developed in this process
concept
and
level in
and
transferred
to the normal SPM
application
level
and
then
transferred
to
the
normal
SPM
process.
Sometimes
process. Sometimes a new iteration round is needed if the concept is not tangible enough to be
a new iteration round is needed if the concept is not tangible enough to be
handled by the steering
group.
handled by the steering group.
6. Ruukki
Figure 6. RuukkiFigure
innovation environment.
innovation environment.
The concept and application development process in Figure 6 is aimed

at finding new technological trajectories. It is related mainly to basic
125
research and mutual research programmes depicted in Figure 3. University
cooperation and collaboration with other companies play an important role in
these activities as a company cannot have enough resources to cover the very
large research area needed. When the technology or solution proceeds to the
applied research phase, less interaction is possible as the product or solution
is approaching launching and competition normally prevents cooperation
unless commercialization can be made jointly.
SUMMARY
There exists a risk that companies adhere to their strong historical trajectories.
In this article we examine different options for avoiding this pitfall by utilizing
several theories of technological trajectories and integration of technology
development into business strategy.
In mature stage of industry where the technological trajectories are strong and
balanced, it is difficult to find new technical or commercial breakthroughs.
Therefore it is important that companies are continuously searching for
alternative trajectories and transform them into new businesses. These are
most likely to be found in the area of new collaboration networks, services or
business models.
Any technological trajectory has a limited lifespan. Thus companies that are
not able to renew their technology portfolio are in danger that their business
will shrivel. Active search for new trajectories is therefore needed. As the
resources of any company are limited and different cooperation networks and
e.g. UBC (university-business cooperation) are necessary for renewal of the
industry.
REFERENCES
Castellacci, Fulvio (2008). Technological paradigms, regimes and trajectories:
Manufacturing and service industries in a new taxonomy of sectoral patterns
of innovation. Research Policy, 37 (2008), pp. 978994.
Dosi G. (1982). Technological paradigms and technological trajectories: a
suggested interpretation of the determinants and directions of technical
change, Research Policy, vol. 11, no. 3, June, pp. 14762.
Hakkarainen, Kari (2006). Strategic Management of Technology. From
Creative Destruction to Superior Resilience. Acta Wasaensia, Number 162.
Industrial Management 11. Universitas Wasaensis 2006.
Hannula, Mika, Ulla Korsman, Eila Pajarre & Marko Seppnen (2003). A Guide
to Academic Writing [online]. Tampere, Finland. Department of Industrial
Engineering and Management, Technical University of Tampere [cited
11.4.2013]. Available: http://butler.cc.tut.fi/~pesone24/academicwriting.pdf
Hedge, Deepak (2004). Innovation and technology trajectories in a developing
country context: Evidence from a Survey of Malaysian Firms. A Thesis.
Georgia Institute of Technlogy, April 2004.
Klepper, S. and K. Simons (1997). Technological extinctions of industrial firms:
An inquire into their nature and causes, Industrial and Corporate Change, 6,
379460.
109
110
Kristinsson, Kari & Rao, Rekha (2013). The Emergence of Dominant

Technology Trajectories in the US Medical Device Industry. First Draft.
Aalborg University, DRUID and IKE. Laboratory of Economics and
Management, Scuola Superiore S. Anna (Available: www.druid.dk). An unfinished paper.
Matthews, William H. (1990). Kissing Technological Frogs: Managing
Technology as a Strategic Resource. Perspectives for Managers 5. Lausanne,
Switzerland: International Institute for Management Development (IMD).
Matthews, William H. (1992). Conceptual Framework for Integrating
Technology into Business Strategy. Int. J. of Vehicle Design 13:5/6, 524532.
Michael Menhart & Carsten Rennhak (2006). Drivers of the Lifecycle the
Example of the German Insurance Industry. Reutlingen Working Papers
on Marketing & Management 200603. School of International Business,
Reutlingen University, 2006.
Pavitt, K. (1984). Sectorial patterns of technical change: towards a taxonomy
and a theory. Research Policy 13 (1984), pp. 343373.
SUSTAINABILITY AS A BUSINESS OPPORTUNITY

TODAY AND TOMORROW: TRIPLE HELIX PERSPECTIVE
Tarja Merist
(tarja.meristo@laurea.fi)
Jukka Laitinen
(jukka.laitinen@laurea.fi)
FuturesLabs CoFi
Laurea University of Applied Sciences, Finland
ABSTRACT
In this paper we focus on SMEs from different industries in Baltic Region. The main goal
is to recognize the business potential of sustainable development in this region, to wake
up the companies to be aware of this potential and finally to design visionary business
concepts for number of cases in this field. In the first phase we have designed a tool
based on the reporting guidelines of Global Reporting Initiative (GRI). Our tool is meant
for companies to recognize their business opportunities and business potential generally
throughout the whole value chain as well as new business opportunities across the borders
of other industries.
INTRODUCTION
Sustainable development has been used as a word since Bruntlands report
1987 (United Nations 1987). Originally, it meant not only ecological, but also
social and economic sustainability. In a broad definition, all the activities
throughout the society, including business, have to meet three challenges, not
harming the future opportunities of the generation first coming to the planet.
At a regional level, in order to achieve sustainable development in practice,
all the Triple Helix dimensions have to be covered, which means actors from
public, private and educational sectors (Etzkowitz & Laydesdorff 1995). The
role of sustainability in the triple helix context have been inspected earlier
(e.g. Yang et al. 2012) but the focus has been in ecological sustainability. In
our paper we are focusing on all dimensions of sustainability, i.e. economic,
social and ecological sustainability.
In business literature sustainable development means different things, e.g.
corporate social responsibility, energy efficiency, lean resource management,
human rights policy or eco-efficiency are the expressions to companies
sustainable behaviour in the context of society and environment. From the
business perspective, the attitudes of consumers are an important part of
sustainable development. The focus from environmental and ecological
issues has moved towards more broad definition of sustainability including
not only environmental but also healthy and safety aspects of sustainability.
111
112
This is called the EHS criteria. Also, World Business Council for Sustainable
Development (WBCSD), which is a global association of some 200 international
companies, deals with the issues concerning business and sustainable
development.
This project is focused on SMEs from different industries in Baltic Region. The
main goal is to recognize the business potential of sustainable development
in this region, to make the companies aware of this potential and finally to
design visionary business concepts for a number of cases in this field. In the
first phase, we have designed a tool based on the reporting guidelines of Global
Reporting Initiative (GRI). The tool is meant for companies to recognize their
business opportunities and business potential generally throughout the whole
value chain as well as new business opportunities across the borders of other
industries.
SUSTAINABILITY
Sustainability means development with economic, social and ecological balance
not only today but for the future generations (United Nations 1987). In business
context this means also responsibilities in different time frames, i.e. short,
medium and long-term. Economic sustainability includes e.g. economic activity
(profitability, turnover), market position (market share) and indirect impacts
(e.g. via taxation); ecological sustainability in the companys eyes consists of
material use (how much, which kind of), energy use (efficiency, renewables),
water footprint, CO2 footprint, biodiversity and waste management among
the others. Social sustainability will cover personnel issues like occupational
health and safety and security, but also human rights, corruption, legislation
and product responsibilities.
TRIPLE HELIX MODEL

The concept of the Triple Helix of university-industry-government relationships
interprets the shift from a dominating industry-government dyad in the
Industrial Society to a growing triadic relationship between universityindustry-government in the Knowledge Society. The creative renewal arises
within each of the three institutional spheres of university, industry and
government, as well as at their intersections (Etzkowitz & Leydesdorff 1995).
In other words, the Triple Helix is a metaphor for university, industry, and
government interacting closely, while each maintains its independent identity.
All the Triple Helix actors benefit from some of the capabilities of the other,
even as each maintains its primary role and distinct identity (Etzkowitz 2007).
SUSTAINABILITY FROM TRIPLE HELIX PERSPECTIVES

We have combined the Triple Helix context with the sustainability concept
which means that all the three dimensions of sustainability have to run through
all the actors as a part of triple helix (Figure 1).
SUSTAINABILITY FROM TRIPLE HELIX PERSPECTIVES

Proceedings of theconcept
METNET Seminar
in Lule
We have combined the Triple Helix context with the sustainability
which2013
means
that all113
the
three dimensions of sustainability have to run through all the actors as a part of triple helix
(Figure 1).
Figure
The
Triple
Helix
combined
Figure 1. The
Triple 1.
Helix
context
combined
withcontext
the sustainability
concept.with the sustainability concept.

129
In other words, university, government and industry have to adopt

sustainability
as a living part
their everyday
On the university
In other words,
university, government
andofindustry
have toactivities.
adopt sustainability
as a living part of
side
this
means
e.g.
new
study
programs
focusing
on
sustainability
but focusing on
their everyday activities. On the university side this means e.g. new study issues,
programs
as a part of their work, sustainability could and should be a guiding principle.
sustainability issues, but as a part of their work, sustainability could and should be a guiding
On governmental side the decision-making process will focus on creating
principle. On
governmental side the decision-making process will focus on creating new
new opportunities for sustainable solutions. Government has to be seen as
opportunities an
for enabler
sustainable
solutions.
Government
has to be
as an enabler
for new,
for new,
sustainable
business practices.
On seen
the industry
side,
sustainable business
practices.
On
the
industry
side,
sustainability
will
also
give
win-win-situations
sustainability will also give win-win-situations with less is more solutions,
with less is more
where
cansociety
win, but
thewill
society
and nature will be
where solutions,
the company
can the
win,company
but also the
andalso
nature
be winners
winners even in
the
long-run.
even in the long-run.
Developed
further, this could
construction
a bit more
i.e. in corner the
Developed further,
this construction
be a bitcould
morebe
advanced,
i.e.advanced,
in governments
governments
corner
the
social
circle
(e.g.
equality)
has
more
focus
than
social circle (e.g. equality) has more focus than in the others; in industrial corner
the economic
in
the
others;
in
industrial
corner
the
economic
circle
is
in
the
middle
circle is in the middle and business profit has a bigger role. University perspective will bring
and business &
profit
has a bigger
role. to
University
perspective
Research & Development
Innovation
(R&D&I)
the table,
and focuswill
on bring
multidisciplinary
Research
&
Development
&
Innovation
(R&D&I)
to
the
table,
and
focus
on
studies combining e.g. natural and technical sciences to human and commercial perspectives
will
multidisciplinary
studies
combining
e.g.
natural
and
technical
sciences
to
put attention to ecological issues at many levels (Figure 2).
human and commercial perspectives will put attention to ecological issues at
many levels (Figure 2).
Figure
2. The
ofsustainability
the sustainability
can
vary
in triple
helix
Figure
2. The focus
focus of the
dimensions candimensions
vary in triple helix
actors.
Through
networking
a actors.
Throughis networking
a balanced sustainability is reached.
balanced sustainability
reached.
SUSTAINABILITY IN CONSTRUCTION INDUSTRY
114
SUSTAINABILITY IN CONSTRUCTION INDUSTRY

As an example, in this paper we will focus on construction industry and
its problems with sustainability issues. In Finland, construction industry
contributes about 10% of annual GNP (Rakennuslehti 2013). Because of
northern climate in Finland, housing is an essential part of everyones budget,
i.e. all the sustainable solutions in this side will have a significant impact to
the whole society.
Construction is traditionally a local business. It is closely related to cultural
tradition, design and materials used. Education for construction industry
will also rely on national norms and legislation, including not only individual
houses but all the urban planning and community structure with its
infrastructure. Therefore, triple helix approach combined with sustainability
approach is especially suitable within construction industry and its different
phases and activities. On the governmental side, local authorities have impact
on infrastructural decisions, but they also decide on social and economic
conditions concerning equality and welfare e.g. social housing production.
On the industrial side, the material and energy use, waste and footprints as
well as labour force with all the economic implications and responsibilities
are the key components of sustainability in companies. In the triple helix
context, new university programs will focus on sustainability issues, but they
also will educate new skills and competences for industry in different sectors
including construction industry among the others. The output is, of course,
new study programs and students with new skills and competences, but as a
part of an education task, new attitudes and values will be spread throughout
the society. Also R&D&I project in close co-operation with business people
will create new solutions, which are better by nature in terms of sustainability.
Indicators for assessing the sustainability performance of new or existing
buildings, related to their design, construction, operation, maintenance,
refurbishment and end of life have been defined in ISO 21929-1 standard. The
set of indicators provides measures to express the contribution of a building(s)
to sustainability and sustainable development. ISO 21929-1:2011 adapts
general sustainability principles for buildings and it includes a framework for
developing sustainability indicators for use in the assessment of economic,
environmental and social impacts of buildings (ISO 21929-1).
Sustainability reporting guidelines to the companies are also offered by
the Global Reporting Iniative (GRI) which a non-profit organization that
promotes economic, environmental and social sustainability. GRI provides all
companies and organizations with a comprehensive sustainability reporting
framework that is widely used around the world. Sustainability reporting
helps organizations to set goals, measure performance, and manage change in
order to make their operations more sustainable (Global Reporting Initiative
2013).
We have developed, as part of a sustainable business opportunities project,
a simple tool for companies to recognize their potential in sustainability
as well as to wake up their own awareness in this field. The tool consists of
115
the following parts: company-specific background questions, importance

of components of sustainability (economic, social, ecological), the status of
sustainable business today and in the future, the channels to get information
concerning sustainability as well the future focus area of the company in the
future. Testing of the tool and preliminary results from state of the art in
construction industry has been applied among top decision makers in Finnish
construction industry. Economic sustainability is, in the first place social
sustainability, in the second place an ecological sustainability and in the third
place among these repliers (Figure 3). Probably construction industry as an
energy- and material intensive industry has already worked a lot for ecological
sustainability and therefore the other dimensions will rise above that!
None of the repliers said that sustainability does not concern construction
business. All the answers were at least on the level of awareness and some
of them even at the level of active work today or visioning better future for
tomorrow (Figure 4).
Sustainability has been seen for construction industry a part of basic business
but also giving new business opportunities and competitive advantage
today (Figure 5). In the future, competitive advantage was the most often
mentioned reason to develop sustainability as a part of the business. No one
said sustainability drive is only empty words without real opportunities.
The dimensions
the sustainable
development
in which
the industry should
Figure 3.of
The dimensions
of the sustainable development
in which the industry should
invest.
3. The dimensions of the sustainable development in which the industry should i
The dimensions
theSeminar
sustainable
development in which the industry shou
Proceedings of of
the METNET
2013 in Lule
116
which the industry should invest.
Figure 4. The current role of sustainability in the construction industry.

onstruction industry.
Figure 5. How the sustainability in perceived in the construction industry now and in the future.
132
117
Figure 5. How the sustainability in perceived in the construction industry now and in the future.
Media seems to be the main channel to meet the challenges of sustainable development, but als
Media seems to be the main channel to meet the challenges of sustainable
education and authorities
arebutthe
giving
impulses
foractors
more
active
work (Figure 6). Stil
development,
alsoactors
education
and authorities
are the
giving
impulses
money talks and only
energy
issues
and product
service
will arise
for more
active
work (Figure
6). Still,and
money
talks concepts
and only energy
issuesclose to that level o
and
product
and
service
concepts
will
arise
close
to
that
level
of
economy.
economy. Economic activities such as profitability and turnover were perceived the most importan
Economic activities such as profitability and turnover were perceived the most
issues related to the
sustainability in the construction industry (Figure 7). Also energy, product
important issues related to the sustainability in the construction industry
and services and society
were
as products
essentialand
factors.
theseen
respondents
did not not
(Figure 7).
Alsoseen
energy,
servicesSurprisingly,
and society were
as
essential
factors.
Surprisingly,
the
respondents
did
not
note
ecological
issues
ecological issues such as water, sewage, debris and biodiversity at all. None among human right
such
water, sewage,
debriswere
and biodiversity
at all. None among human
personnel practices
or as
working
conditions
seen important.
rights, personnel practices or working conditions were seen important.
Figure 6. The channel of the sustainable development in the construction industry.

118
133in the construction industry.

Figure 7. The most important sustainability issues
Construction industry is in this paper has been covered only with opinions
of a few experts. In the near future, we will continue the work with in-depth
interviews in construction industry companies in Uusimaa County. Also
other industrial sectors in the region will be covered and data collected
for comparison. Especially from economic viewpoint of sustainability,
international interviews in Sweden and in Germany will be carried out with
focus on EUs Small Business Act and its opportunities for SMEs. All the
interviews will take place in autumn 2013.
CONCLUSION
This paper presents the first results from our work dealing with sustainability
and triple helix together. Framework developed in this work seems to be
promising. It shows the elements of sustainable development in business
context at local and regional level. It also gives alternative perspectives to a
balanced development in terms of sustainability. Actor-based model will be
replaced by network based model with new features. This gives an opportunity
to develop a dynamic model to examine new business opportunities for
sustainable growth in the way all the actors can focus on the strengths of their
own. This is only a beginning. The research work will continue in autumn
with field work (e.g. interviews), but also theoretical framework development
will be in progress. Sustainability is one of the most important issues in global
context. Every triple helix hub and every single actor have to be prepared
for the future keeping in mind these three dimensions of sustainability, i.e.
economic, social and ecological sustainability. Without dynamic balance in
this sense there are no business opportunities in the future at all.
REFERENCES:
Etzkowitz, H. (2007) University-Industry-Government: The Triple Helix
Model of Innovation. Proceedings of the 51st EOQ Congress, 2223 May,
2007, Prague, Czech Republic.
Etzkowitz, H., Leydesdorff, L., (1995) The Triple Helix: University Industry
Government Relations A Laboratory for Knowledge Based Economic
Development. EASST Review 14 (1), 1995.
Global Reporting Iniative (GRI) (2013) G4 Sustainability Reporting
Guidelines. https://www.globalreporting.org/resourcelibrary/GRIG4-Part1Reporting-Principles-and-Standard-Disclosures.pdf (viewed 30.8.2013).
ISO 21929-1 (2011) Sustainability in building construction Sustainability
indicators.
Rakennuslehti (2013) Rakennusmarkkinat Suomessa. (Construction
market in Finland) (in Finnish). http://www.rakennuslehti.fi/tietoa/
rakennusmarkkinat_suomessa/ (viewed 30.8.2013)
United Nations (1987) Our Common Future Report of the World Commission
on Environment and Development.
Yang, Y., Holgaard, J.E., & Emmen, A. (2012), What can triple helix
frameworks offer to the analysis of eco-innovation dynamics? Theoretical and
methodological considerations. Science and Public Policy, Volume 39, Issue
3, pp. 373385.
WBCSD World Business Council for Sustainable Development http://www.
wbcsd.org/home.aspx (viewed 30.8.2013).
119
120
OPTIMIZATION OF STEEL TOWERS

FOR LARGE WIND TURBINES
Anatoly Perelmuter
SCAD Soft Ltd, Kyiv, Ukraine
Vitalina Yurchenko
Kyiv National University of Civil Engineering and Architecture, Kyiv, Ukraine
ABSTRACT
Tall towers of wind-powered generators are relatively new types of steel structures widely
applied. At the present time several thousand steel towers for wind-powered generators
are manufactured every year in the world. The paper describes a technique for the
optimisation of steel towers.
INTRODUCTION
During last decade the typical capacity of the wind-powered generators
increased nearly ten times (Perelmuter 2013). Steel towers for wind turbine
go up to a height of 100 m and more and the tower cost approaches 1520%
of the construction budget for the whole wind-powered generator. Takeing
into account the increasing ratio between the cost of the steel tower and
the construction budget for the whole wind-powered generator, formulation
of an optimization problem for this kind of steel structure is considered as
beneficial.
PROBLEM FORMULATION
Optimization problem is formulated as searching of the optimum design of
the tower for wind-powered generator, which is designed as a steel conic shell
with diameter varied linearly along shell length.
The tower structure was divided into parts along the tower height while using
the finite element method for determination of the stress-and-strain state of
the towers conic shell. For each part of the tower a constant thickness and
a constant diameter of the middle surface at lower level were considered as
characteristic cross-sectional dimensions (Fig. 1).
The main parameters of the steel conic shell tower were considered as design
variables, namely: H the towers height; Dmin minimum diameter of the
middle surface of the towers conic shell at a height of H (the height of wind
wheel above earths surface level); Dmax maximum diameter of the middle
surface of the tower conic shell at the base level; t i the thickness of ith part of
towers conic shell, i = l...ns , where ns number of the tower shell parts; Dw
the diameter of the wind wheel.
121
Figure 1. Design variables of the steel conic shell tower.
Figure1.1.Design
Designvariables
variablesofofthe
thesteel
steelconic
conicshell
shelltower
tower
Figure
Thecapacity
load-bearing
of the steel
shellshould
structure
should
be for two wind load
Theload-bearing
load-bearing
thecapacity
steeltowers
towers
shelltowers
structure
beverified
verified
The
capacity ofofthe
steel
shell
structure
should be
for two wind load
verified for two wind load cases. The first wind load case is determined by
cases.
The
first
wind
load
case
is
determined
by
the
maximum
possible
wind
velocity
atataa
cases. The first wind
load case
is determined
byvthe
maximum
possible
wind
velocity
the maximum
possible
wind velocity
at a height
of H above
earths
H,max,1
surface
level (at
the height
of (at
installation
of the
wind
wheel) byof
wind wheel) by which
above
earths
surface
level
theheight
height
installation
thewind
wind
heightofof above
height
earths
surface
level
(at
the
ofofinstallation
ofwhich
the
wheel) by which
wheels
blades
of
the
wind-powered
generator
are
under
operating
conditions.
wind wheels
wheels blades
blades ofof the
the wind-powered
wind-powered generator
generator are
are under
under operating
operating conditions.
conditions. The
The
wind
The corresponding wind velocity v0,max,1 at a height 10 m above the earths
corresponding
wind
velocity
at
a
height
10
m
above
the
earths
surface
level
is
calculated
corresponding wind velocity
at a height 10 m above the earths surface level is calculated asas
m).
( ( ininm).
surface level is calculated as (H in m).
Thesecond
secondwind
windload
loadcase
caseisisdetermined
determinedby
bythe
themaximum
maximumpossible
possiblewind
windvelocity
velocity
The
atataa
The
second
wind
load
case
is
determined
by
the
maximum
possible
wind
height10
10mmabove
abovethe
the earthssurface
surfacelevel.
level. This
Thiswind
windvelocity
velocitycharacterizes
characterizesthe
the considered terrain
height
velocityearths
v0,max,2 at a height
10 m above
the earths surface
level. This windconsidered terrain
where
wind-powered
generator
will
be
installed.
where wind-powered
generator
will the
be installed.
velocity
characterizes
considered terrain where wind-powered generator
will be installed.
Theload-bearing
load-bearingcapacity
capacityofofthe
thesteel
steeltowers
towersshell
shellstructure
structureshould
shouldbe
beverified
verifiedfor
forboth
boththe
thewind
wind
The
The
load-bearing
capacity
of
the
steel
towers
shell
structure
should
be
verified
loadcases
casesmentioned
mentionedabove
abovesosolong
longasasititisisnot
notknown
knownbeforehand
beforehandwhich
whichofofthem
themisismore
morecritical.
critical.
load
for both the wind load cases mentioned above so long as it is not known
Under the
the lower
lower wind
wind velocity
velocity
distributed wind
wind load
load
Under
((
) ) distributed
togetherwith
withconcentrated
concentratedpulling
pullingload
loadofofthe
thewind
windturbine
turbine
applyon
onthe
thesteel
steeltower
tower
together
apply
structure. Under the higher wind velocity
the distributed wind load
on the
122
beforehand which of them is more critical. Under the lower wind velocity
v0,max,1 (v0,max,1 < v0,max,2) distributed wind load qi,1 (H, zi, Di ) together with
concentrated pulling load of the wind turbine Fx (v0,max,1) apply on the steel
tower structure. Under the higher wind velocity v0,max,2 the distributed wind
load qi,2 (H, zi, Di) on the tower structure is greater, but the pulling force of
Functionthe wind turbine is not applied to the towers top due to rotation of the blades
Function
Functionin a feathered position.
Function
Objective
Objective
Objective
Objective
The optimization problem of tower for wind turbine with target capacity
can be solved by
The
optimization
problem
of
tower
for
wind
turbine
with
target
capacity
can
be
by
The
optimization
problem
of
tower
for
wind
turbine
with
target
capacity
can wind
be solved
solved
by
of turbine
the wind
wheel
and
corresponded
velocity
choosing
the corresponding
diameter
The
optimization
problem
of
tower
for
wind
with
target
capacity
can
be
solved
by
Objective
Function
choosing
the
corresponding
diameter
of
the
wind
wheel
and
corresponded
wind
velocity
of the wind wheel and corresponded wind velocity
choosing the corresponding diameter
of the. wind
wheelfrom
and the
corresponded
velocity
choosing, which
the corresponding
diameter
depends on the
tower height
It follows
expression*wind
for produced
The optimization
problem
of tower for
turbine from
with target
capacity W for
,, which
on
height
.. wind
It
the
which depends
depends
on the
the tower
tower
height
It follows
follows
from
the expression
expression
for produced
produced
capacity, which
of the
wind-powered
generator,
which
estimated
based
on
model
depends
on thebytower
height
. Itisfollows
from
thetheexpression
for produced
can
be solved
choosing
the corresponding
diameter
Dw
of
windideal
wheelturbine
capacity
of
the
wind-powered
generator,
which
is
estimated
based
on
ideal
turbine
model
capacity
of
the
wind-powered
generator,
which
is
estimated
based
on
ideal
turbine
model
and corresponded
velocitytaken
vH (H),into
which
dependsaverage
on the tower
height
H. velocity
proposed by N.
E.
Zhukovskyj
and wind
A.
Betzs
account
annual
wind
capacity
of
the
wind-powered
generator,
which
is
estimated
based
on
ideal
turbine
model
proposed
by
N.
E.
Zhukovskyj
and
A.
Betzs
taken
into
account
average
annual
wind
velocity
It follows fromand
the A.
expression
produced
capacity W
of the wind-powered
proposed by N. E. Zhukovskyj
Betzs for
taken
into account
average
annual wind velocity
as:
proposed by N. E. Zhukovskyj
andis A.
Betzsbased
takenoninto
average
annual
wind
generator, which
estimated
idealaccount
turbine model
proposed
by N.
E. velocity
as:
as:
Zhukovskyj
and
A.
Betzs
taken
into
account
average
annual
wind
velocity
v
0
as:
as:
Naturally, at the same time it is aspired to the minimise the weight of the steel
Naturally,
at
time
is
minimise
Naturally,
at the
the same
same
time it
itconic
is aspired
aspired
tobethe
the
minimiseas:the
the weight
weight of
of the
the steel
steel
theoretical
the tower
shell canto
calculated
Naturally, weight
at the of
same
timeat itthe
issame
aspired
to
the
minimise
the weight
ofof the
steel
theoretical
weight
of
the
tower
conic
shell
can
be
calculated
as:
Naturally,
time
it
is
aspired
to
the
minimise
the
weight
the steel
theoretical weight of the tower conic shell can be calculated as:
theoretical weight oftower.
the tower
conic shell
canofbe
The theoretical
weight
thecalculated
tower conicas:
shell can be calculated as:
tower.
tower.
tower.
tower.
The
The
The
The
where
t/m33 the steel density.
where
t/m
the
density.
where
t/m33where
the steel
steel
= 7,8density.
t/m3 the steel density.
where
t/m thessteel
density.
In the problem formulation, the produced capacity
of the wind-powered generator being equal to
In the problem
formulation,
the produced
capacity
W of the wind-powered
In
the
problem
formulation,
the
produced
capacity
of
wind-powered
generator being
equal
In
the
problem
formulation,
the
produced
capacity
of the
the
wind-powered
being
equal to
to
the
target
capacity
was considered
asthe
additional
constraint
expressed
inasgenerator
the
form of
inequalities
* wind-powered
of
the
generator
being
equal to
In
the
problem
formulation,
the
produced
capacity
generator
being
equal
to
target
capacity
W
was
considered
additional
was
considered
as
additional
constraint
expressed
in
the
form
of
inequalities
the
target
capacity
the
capacity
was considered
constraint
in the
form of inequalities
( target
small
positiveconstraint
number).
expressed inas
theadditional
(Wexpressed
small positive
number).
the
capacity
was considered
as
additional
constraint
expressed
in the
(( target
small
positive
number).
small positive number).
(
small positive number).
System of constraints
System
System of
of constraints
constraints
Constraints of the ultimate limit state were formulated as verifications that the actual stresses arising
Constraints
of
the
limit
state
were formulated
as
verifications
the
arising
th
Constraints
the ultimate
limit state were
formulated asthat
verifications
thatstresses
Constraints
of cross-section
the ultimate
ultimate
limit
state
formulated
as shell,
verifications
the actual
actual
stresses
in
the design
of ofth
partwere
of towers
conic
here that
, caused
by arising
load
th
th
Constraints
of cross-section
the ultimate
limit
state
were
formulated
ascross-section
verifications
theofactual
stresses
arising
the actual
stresses
arising
in the design
of that
ith part
towers
th
th load
in
the
design
of
part
of
towers
conic
shell,
here
,
caused
by
in
the
design
cross-section
of
part
of
towers
conic
shell,
here
,
caused
by
load
th i =the
case
should
notofexceed
stresses
of
compression
(by the
conic shell,
here
, caused
by
kth load
casehere
combination,
should
notstability):
stowers
in
thecombination,
design cross-section
partl,...,n
ofcritical
conic
shell,
, local
caused
by th load
case
exceed
the
stresses
of
(by
case combination,
combination, should
should
not
exceedstresses
the critical
critical
stresses (by
of compression
compression
(by the
the local
local stability):
stability):
exceed not
the critical
of compression
the local stability):
case combination, should not exceed the critical stresses of compression (by the local stability):
where
the critical stress for th
the design axial force in
th part of the towers conic shell;
where
the
part of the towers conic shell;
the
where
the critical
critical stress
stress for
for th
the design
design axial
axial force
force in
in
th
th part of the towers conic shell;
part of the
tower
conicstress
shell;for
theofdesign
bending
moment
for th
partdesign
of theaxial
tower
conic
where
the
critical
part
the
towers
conic
shell;
the
force
in
th
th
th part of the tower conic shell;
th
the
the design
design bending
bending moment
moment for
for th part
part of
of the
the tower
tower conic
conic
th part of the towerthconic shell;
shell
subjected
to thconic
load case
axialfor
load applied
a height
of
part
of the tower
shell;taking into
theaccount
design concentrated
bending moment
part ofatthe
tower conic
shell
load
case
taking
into
account
concentrated
axial
load
applied
at
a
height
of
shell subjected
subjected to
to th
load
case
taking
into
account
concentrated
axial
load
applied
at
a
height
of
th
design
cross-section
of ofpartth of
towers
conicconic
shell,shell,
here here
, caused
by byloadth load
the design
cross-section
part
of towers
, caused
3
3
t/m the
steel
where
density.
thestresses
steelstresses
density.
mbination,
should
not
exceed
thet/m
critical
of compression
(by the
stability):
ase
combination,
should
not exceed
the critical
of compression
(bylocal
the local
stability):
he steel density.
123
ofthe
theproduced
wind-powered
generator
of the
being
wind-powered
equal to
generator be
em formulation, In
thethe
produced
problemcapacity
formulation,
capacity
of
the
wind-powered
generator
being
equal
to
on,
the
produced
capacity
apacity
was the
considered
target capacity
as additionalwas
constraint
considered
expressed
as additional
in the form
constraint
of inequalities
expressed in the form of
was
considered
as( additional
constraint
expressed in the form of inequalities
th
th number).
positive
number).
small
positive
stressstress
for for
part of
theoftowers
conicconic
shell;shell; the design
axial axial
force force
in in
here the critical
the critical
part
the towers
the design
where cr,i the critical stress for ith part of the towers conic shell; Ni the
mber).
th
h
th
of
theoftower
conicdesign
shell;axial
the
bending
moment
for
part
of
the
conicconic
force
in ithdesign
part
the
tower
conic
shell;
Mi,k for
the
design
bending
part
the tower
conic
shell;
theofdesign
bending
moment
part
oftower
the tower
th
th
moment for i part of the tower conic shell subjected to k load case taking
th
th
ubjected
to toloadinto
case
taking
into account
concentrated
axial axial
load applied
at a height
hell subjected
load
case taking
into account
concentrated
load applied
at a heightof of
account concentrated axial load applied at a height H of tower caused
,
safety
factors
for
self
weight
loads
caused
caused
by self
of
the
rotor;
,
safety
factors
for
self
weight
caused
wer caused
byweight
selfbyweight
of
the
rotor;
self weight of the rotor; f m r , f m sw safety factors for self weight loads loads
onstraints
System
of
constraints
caused by the rotor and steel tower conic shell respectively; Ai and Wi area
and steel
conic
shell respectively;
and and
areaand
moment
of inertia
yrotor
the rotor
and tower
steeland
tower
conic
shell respectively;
areasecond
and second
moment
of inertia
second moment of inertia of the cross-section of ith part of the tower conic
th
th
ross-section
of limit
part
ofpart
theof
conicconic
shell
respectively.
shell
respectively.
thethe
cross-section
of
oftower
the
tower
shell
of
ultimate
Constraints
state
were
the
formulated
ultimate
as
limit
verifications
staterespectively.
werethat
formulated
the actual
as stresses
verifications
arising
that the actual stre
th
th the actual stresses arising
th
ate
limit
state
were
formulated
as
verifications
that
n cross-section in
of the part
design
of cross-section
towers conicofshell, part
hereof towers conic
, caused
shell,
by here load
, caused
th
th
part
of
towers
conic
shell,
here
,
caused
by
load
ion
of
nation, should notcase
exceed
combination,
the
should
stresses
notofexceed
compression
the critical
(by stresses
the localofstability):
compression (by the local sta
RESULTS
OF critical
OPTIMIZATION
LTS
OPTIMIZATION
OFthe
OPTIMIZATION
dESULTS
notOF
exceed
critical stresses of compression (by the local stability):
A number of optimization problems have been formulated with different
*
values
ofproblems
target
capacity
of theformulated
wind-powered
generator.
These
parametric
ber
of optimization
problems
have
beenWbeen
formulated
with different
values
of target
capacity
number
of optimization
have
with
different
values
of target
capacity
optimization
problems
have been solved
by improved
methodhave
of projection
the
wind-powered
generator.
These
parametric
optimization
problems
been
solved
by
of the wind-powered
generator.objective
These parametric
optimization
have
solved
by
improved
method of projection
function thgradient
on theproblems
surface of
thebeen
active
constraints
th function gradient on the surface
objective
of of
the
active
constraints
using
using
the critical
stress
where
for
part
the
of
the
critical
towers
stress
conic
for
shell;
part
the
the
towers
design
conic
axial
shell;
force
in
the design
software
OptCAD
(Peleshko
et
al.
2009).
Initial
data
are
as
follows:
average
annual
wind a
th
software
OptCADconic
(Peleshko
et al. 2009).
Initial
data
are force
as follows:
average
l stress
for
part
of
the
towers
shell;
the
design
axial
in
th
th
th
m/s;
possible
velocities
fordesign
wind
wheel
and
for considered
hevelocity
tower conic shell;
partmaximum
ofwind
the
the
tower
design
shell;
moment
possible
the
for
partbending
of theblades
tower
moment
conic
for
part of the
annual
velocity
vconic
7 wind
m/s;
maximum
wind
velocities
for wind
0 =bending
th
nicterrain
shell;th
thewheel
design
bending
of
the
conic
blades
and forthmoment
consideredfor
terrainpart
vH,max,1
= 20tower
m/s and
v0,max,2 = 30
m/s
and
m/s
correspondingly;
the
design
resistance
load case
taking
into account
load
concentrated
case
takingaxial
intoload
account
applied
concentrated
at asteel
height
axial of
load applied at a h
ted to
shell
subjected
to
m/s correspondingly; the design steel resistance Ry = 3,823 t/sm2; safety
ad
case
into
account
concentrated
axial
load
applied
at
a
height
of
2 taking
d t/sm
by
self
weight
tower
of
the
caused
rotor;
by
self
,
weight
of
safety
the
rotor;
factors
for
,
self
weight
safety
loads
factors
caused
forself
selfweight
weight l
factors
wind
load f m w = 1,2,,for
weight
selftower
f m r 1,2; for tower
; safety factors
for for
wind
load
forrotor
rotorselfself
weight
; for
ight of the rotor; weight
, f m sw =1,05.
safety
factors
for
self
weight
loads
caused
Table 1 presents the results of optimization calculations.
and steel tower
conic
byThe
the1shell
rotor
respectively;
andbetween
steel
tower
conic
shell
respectively;
and
area
and
second moment
and
of
area
inertia
andbetween
second mome
. Table
presents
the results
of
optimization
Theiscorrelation
the
correlation
the tower
height and thecalculations.
generation capacity
shown
th respectively;
tharea and second moment of inertia
wer
conic
shell
and
part
of
the
tower
conic
shell
part
respectively.
of
the
tower
conic
shell
respectively.
-section
of
of
the
cross-section
of
in
Figure
2.
h tower height and the generation capacity is shown in Figure 2.
part of the tower conic shell respectively.
Tower's height, m
OF OPTIMIZATION
RESULTS
OF
OPTIMIZATION
Table 1. Results
of optimization.
Table
1. Results
of optimization
IZATION
,m
,m
,m
, with
t
, m problems
, Wt
of optimization problems
A number
have
of optimization
been formulated
have
different
beenvalues
formulated
of target
withcapacity
different values of tar
on problems have
been
formulated
with different
of parametric
target
capacity
1,0
75,65
39,76 values
1,881
4,769
65,65
wind-powered
generator.
of the
These
wind-powered
parametric
generator.
optimization
These
problems
have
optimization
been
solved problems
by
have bee
2,0
98,13
90,74
d generator. These parametric
optimization
problems 2,520
have been6,335
solved by88,13
3,0
114,57 147,71 2,975
7,482
104,57
4,0
128,02 209,00 3,347
8,422
118,02
5,0
139,60 273,80 3,667
9,233
129,60
6,0
149,89 341,55 3,951
9,955
139,89
7,0
159,22 411,90 4,209
10,609 149,22
180
160
140
120
100
80
60
40
20
0
0,5
1,5
2,5
3,5
Capacity, MWt
4,5
5,5
6,5
114,57 147,71
147,71 2,975
2,975
3,03,0
114,57
4,0
128,02
209,00
3,347
4,0
128,02 209,00 3,347
5,0
139,60
273,80
3,667
139,60
Proceedings5,0
of the METNET Seminar
2013 in Lule273,80 3,667
149,89 341,55
341,55 3,951
3,951
6,06,0
149,89
7,0
159,22
411,90
4,209
7,0
159,22 411,90 4,209
Tower's height, m
Tower's height, m
124
7,482 104,57
104,57
7,482
8,422
118,02
8,422
118,02
9,233
129,60
9,233
129,60
9,955 139,89
139,89
9,955
10,609
149,22
10,609 149,22
180 180
160 160
140 140
120 120
100 100
80 80
60 60
40 40
20 20
0 0
0,5 0,5 1
1 1,5 1,5 2
2 2,5 2,5 3
3 3,5 3,5 4
4 4,5 4,5 5
5 5,5 5,5 6
6 6,5 6,5 7
Capacity,
Capacity,
MWtMWt
2.2.Optimum
tower
height
capacity
of the
generator
Figure
Optimum tower
height
depending
ondepending
produced capacity
ofproduced
the generator.
Fig.Fig.
2.
Optimum
tower
height
depending
on on
produced
capacity
of the
generator
CONCLUSIONS
CONCLUSIONS
CONCLUSIONS
This
paper
presents
parametric
optimization
of steel
conic
shell
towers
of wind-powered
generators.
This
paper
presents
parametric
optimization
of steel
conic
shell
towers
of wind-powered
generators.
The
minimum
weight
of
the
steel
tower
has
been
considered
as
the
objective
function
when
The minimum weight of the steel tower has been considered as the objective function when
This
paper
presents
parametric
optimization
of
steel
conic
shell
towers
of
produced
capacity
wind-powered
generators
been
fixed
at the
target
level.TheThe
tower
produced
capacity
of of
thethe
wind-powered
generators
hashas
been
fixed
at the
target
level.
tower
wind-powered generators. The minimum weight of the steel tower has been
height,
diameters
of
the
middle
surface
of
the
towers
conic
shell
at
the
base
level
and
at
a
height
height, diameters of
the middle
surface
of thefunction
towerswhen
conicproduced
shell at the
base level
at a height of of
considered
as the
objective
capacity
of theand
windthe
wind
turbine
installation,
thicknesses
of
the
tower
conic
shell,
and
also
the
diameter
of the wind
the wind turbine installation,
thicknesses
thefixed
toweratconic
shell, level.
and also
diameter
powered generators
has of
been
the target
Thethe
tower
height,of the wind
wheel
were
considered
as
the
design
variables.
The
wind
loads
applied
on
steel
tower
shell
have
wheel were considered
as the
design
variables.
wind loads
applied
onbase
steel
tower
have
diameters
of the
middle
surface of The
the towers
conic shell
at the
level
and shell
been
considered
as
state
variables.
The
optimization
problem,
formulated
as
a
non-linear
of the wind The
turbine
installation, problem,
thicknessesformulated
of the toweras
conic
been considered atasa height
state variables.
optimization
a non-linear
programming
task,
been
solved
improved
gradient
method.
Optimum
designs
steel
shell,
and
also
the
diameter
of the wind
wheelmethod.
were
considered
as the
design
programming
task,
hashas
been
solved
by by
improved
gradient
Optimum
designs
of of
thethe
steel
variables.
The
wind
loads
applied
on
steel
tower
shell
have
been
considered
conic
shell
towers
for
the
wind-powered
generators
with
produced
capacity
of
wide
range
have
been
conic shell towers for the wind-powered generators with produced capacity of wide range have been
as state variables. The optimization problem, formulated as a non-linear
defined.
defined.
programming task, has been solved by improved gradient method. Optimum
designs of the steel conic shell towers for the wind-powered generators with
produced capacity of wide range have been defined.
REFERENCES
, . . 2013.
[The towers of million-kilowatt wind turbines],
[Industrial
building and engineering structires], 1.
Peleshko, I.; Yurchenko, V. 2004. An optimum structural computer-aided
design using update gradient method, Proc. of the 8th Int. Conf. Modern
Building Materials, Structures and Techniques (Lithuania, Vilnius, May
1921, 2004): 860865.
STRUCTURAL BRACING FORCES AND STIFFNESS

Olli Ilveskoski
HAMK University of Applied Sciences, Hmeenlinna, Finland
INTRODUCTION
The paper deals with the design of structural bracing used in beams,
columns, and frame structures. Bracing used in structural systems generally
serve two primary functions. They resist secondary loads on structures e.g.,
wind bracing and increase the strength of individual members by resisting
deformation in the weakest direction. For the latter case, structural bracing
forces higher modes of deformation by providing resistance to lateral and
rotational displacement. This is achieved through axial, shear, and flexural
deformations of the bracing member. Diaphragms, for instance, provide
restraint through their shear stiffness while diagonal cross-bracing relies on
axial stiffness.
European standards give brief instructions on bracing connection forces and
do not inform much about connection stiffness or the use of sandwich panels
as stabilizing elements. This short survey is made to check other guidelines
and recommendations. Especially the use of self-supporting sandwich panels
as stabilizing elements for single steel members such as beams or columns has
been recently investigated and Recommendations on the Stabilization of Steel
Structures by Sandwich Panels have been published. The possibility to use
sandwich panels as stabilizing elements reduces the weight and costs of the
frame and creates even the method to design frameless buildings.
Manufactures have big challenges to meet present requirements in standards.
The stabilizing elements shall fulfil the requirements shown by the CE-mark
of the product. Depending on the case and practice, the manufacturer shall
provide the additional information about the stiffness and the resistance of
the fastenings or about the shear stiffness and the torsional restraint of the
stabilizing system.
BRACING SYSTEMS
Bracing systems used to control instability fall into four general classifications:
relative, nodal, continuous, or lean-on. Relative bracing systems, such as
diagonal bracing or shear walls, prevent the relative lateral movement of
adjacent stories or of adjacent points along the length of a member. Nodal
systems control the movement only where they attach to the braced member
and do not directly interact with adjacent brace points. Cross-frames or
diaphragms between two adjacent beams are considered nodal braces.
Continuous systems provide uninterrupted support along the entire length of
a member, leaving no unbraced length. Shear walls and roof or floor deck are
125
126
examples of continuous bracing systems. Lean-on systems rely on adjacent

structural members to provide support. Lean-on bracing links together
adjacent structural members such that buckling of one member requires all
members in the system to buckle with the same lateral displacement [1].
Figure 1. Bracing systems [1].
Figure 1. Bracing systems [1]
Structural bracing
used tobracing
increase
of strength
members
must possess
both sufficient strength
Structural
usedthe
to strength
increase the
of members
must possess
both sufficient
and stiffness.
bracing
designtorules
such
and stiffness. Simple
bracingstrength
design rules
such asSimple
designing
a brace
resist
2% of the member
as
designing
a
brace
to
resist
2%
of
the
member
compressive
force
address
compressive force address only the strength criterion. The stiffness of the brace along with the outonly the strength criterion. The stiffness of the brace along with the outof-straightness ofof-straightness
the memberof has
a direct effect on the magnitude of the brace force. Design
the member has a direct effect on the magnitude of the
recommendationsbrace
based
onDesign
perfectly
straight members
should
not straight
be usedmembers
directly in design since
force.
recommendations
based on
perfectly
extremely large brace
shouldforces
not be and
used displacements
directly in designmay
sinceresult.
extremely large brace forces and
displacements may result.
Winter developed the concept of a dual strength and stiffness criterion for the design of bracing
Winter developed the concept of a dual strength and stiffness criterion for
used to control instability.
The required brace strength cannot be uniquely determined, but depends
the design of bracing used to control instability. The required brace strength
on both the magnitude
of the determined,
brace stiffness
and on
member
out-of-straightness.
The
cannot be uniquely
but depends
both theinitial
magnitude
of the
relationship between
parameters
is illustrated
for the relativeThe
column
brace in Figure 2 and
brace these
stiffness
and member
initial out-of-straightness.
relationship
is illustrated
can be extended between
to otherthese
typesparameters
of bracing
systems. for the relative column brace in Figure
2 and can be extended to other types of bracing systems.
used to control instability. The required brace strength cannot be uniquely determined, but depends
on both the magnitude of the brace stiffness and member initial out-of-straightness. The
of the
METNET Seminar
2013in
in Lule
relationship between these parameters is illustrated for theProceedings
relative
column
brace
Figure 127
2 and
can be extended to other types of bracing systems.
Figure 2. Required
brace strength,
strength, stiffness
and initial member
out-ofstraightness
[1]. out-ofstraightness [1]
Figure 2. Required
brace
stiffness
and initial
member
148
order to reach the Euler buckling load, Pe, the brace must possess a
brace must possess a minimum stiffness known as
order to reach theIn
Euler
buckling load, Pe, the
minimum stiffness known as
the ideal stiffness, i. Figure 2 shows the
ideal stiffness, irelationship
. Figure 2between
shows the
thebrace
relationship
between
the the
brace
stiffness and
force. When
idealstiffness
stiffness and force. When
is
used
(=
),
as
the
column
load
approaches
P
the
sway
deflections
become
Pe the sway
deflections become
ideal stiffness is used (=
i i), as the column load approaches
e
very
large.
Unfortunately,
this
results
in
very
large
brace
forces
since
P
br=. At twice the ideal
y large. Unfortunately, this results in very large brace forces since Pbr=.
At twice the ideal stiffness (=2i), the brace force equals 0.4% of the column
fness (=2i), theload
brace
force
0.4% of
the column
load when
P=Pe. For practical designs,
when
P=Peequals
. For practical
designs,
the deflections
and corresponding
deflections and corresponding
brace
forces
arebrace
keptstiffness
small by
using
brace
stiffness greater than
brace forces are kept
small
by using
greater
than
the ideal
ideal stiffness. stiffness.
The plots developed in the figure above were based on an assumed initial out-
e plots developedof-straightness
in the figureequal
above
were based
an assumed initial
equal
to 0.002L.
Largeron
out-of-straightness
values out-of-straightness
linearly
increase
the
magnitude
of
the
brace
forces
[1].
The
connections
details
used
to
0.002L. Larger out-of-straightness values linearly increase the magnitude of the brace forces [1].
attach
structural
bracing
members can
be of great
importance
e connections details
used
to attach
structural
bracing
members
canwhen
be designing
of great importance when
or evaluating the overall performance of a bracing system. If the connections
igning or evaluating
the overall
performance
a bracing
system.
If the connections are
are flexible,
the stiffness
of the overallof
bracing
system can
be significantly
xible, the stiffnessless
of than
the the
overall
bracing
system
can be
significantly
than the stiffness of the
stiffness
of the bracing
member
alone.
The stiffnessless
of a bracing
system
be evaluated
springs insystem
series using
cing member alone.
Thecan
stiffness
of aasbracing
can be evaluated as springs in series using
The system stiffness, sys, will always be less than the smaller of the brace
e system stiffness,member
sys, will
always be less than the smaller of the brace member stiffness, br,
stiffness, br, and the connection stiffness values conn [1].
d the connection stiffness values conn [1].
MEMBER BRACING ACCORDING TO EUROCODE [6]
rocode 3 Design of Steel structures Part 1-1 contains rules to prevent column-beams lateral
The system stiffness, sys, will always be less than the smaller of the brace member st
Thethe
system
stiffness,
sys, will
always
less than the smaller of the brace member s
and
connection
stiffness
values
connbe[1].
Proceedings
of
the
METNET
Seminar
2013
in
Lule
128
[1]. of the brace member stiffness, ,
and the
stiffness
conn
he system
stiffness,
sysconnection
, will always
be lessvalues
than the
smaller
br
nd the connection stiffness values conn [1].
3. MEMBER BRACING ACCORDING TO EUROCODE [6]
3. MEMBER BRACING ACCORDING TO EUROCODE [6]
MEMBER BRACING ACCORDING TO EUROCODE [6]
. MEMBER BRACING
EUROCODE
[6]
Eurocode 3 ACCORDING
Design of Steel TO
structures
Part 1-1
contains rules to prevent column-b
Eurocode
3
Design
of
Steel
structures
Part
11
contains
rules torules
prevent
Eurocodebuckling.
3 Design ofIt Steel
structures
Part 1-1assessment
contains
to prevent
column-b
torsional
includes,
e.g. simplified
methods
for beams
wi
column-beams lateral torsional buckling. It includes, e.g. simplified
torsional
buckling.
It
includes,
e.g.
simplified
assessment
methods
for
beams
w
buildings
(6.3.2.4)
urocode 3 Design
of Steel
structures
Part with
1-1 restraints
containsinrules
to prevent
assessment
methods forbeams
buildings
(6.3.2.4).column-beams lateral
buildings
(6.3.2.4) e.g. simplified assessment methods for beams with restraints in
orsional buckling.
It includes,
uildings (6.3.2.4)
(6.59)
rules can be found about lateral torsional buckling of members with

MoreMore
rules
can be found about lateral torsional buckling of members with plastic
plastic hinges (6.3.5) where significant axial compression in the stable length
More
rules
can be
found
about lateral
torsional
buckling
with plastic
wheremay
significant
axial
compression
in the
stable length
mayofbemembers
taken from
be taken from
significant
axial compression
in the stable
length may
takenhinges
from ( 6.3.5)
More rules can where
be found
about lateral
torsional buckling
of members
withbe
plastic
where significant axial compression in the stable length may be taken from
(6.68)
and effective restraint should be provided so that at each plastic hinge location, the
andcompression
effective
restraint
provided
that
each
plastic
location, th
the
flangeshould
to the be
resisting
element
that
point
andhinge
any intermediate
and
effective
restraint
should
be
provided
soso
that
atatat
each
plastic
hinge
location,
the connection
the
flange
to the
resisting
element
the compression
flange
todesigned
thecompression
resisting
element
at
that location,
point
diagonal
brace)
should
beof
to each
resist
a local
force
equal
toand
2,5%
ofintermediat
N f,Ed tran
nd effective restraint
should
be
provided
so
that
at
plastic
hinge
theany
connection
of
at
that
point
and
any
intermediate
element
(e.g.
diagonal
brace)
should
be
diagonal
brace)
should
be
designed
to
resist
a
local
force
equal
to
2,5%
of
N
tra
f,Ed
flange
in
its
plane
and
perpendicular
the
web
plane,
without
any
combination
with
he compression flange
to to
the
resisting
element
thatof Npoint
and any intermediate element (e.g.
designed
resist
a local force
equal toat2,5%
f,Ed transmitted by the flange
inbe
itsdesigned
plane
and
to
theequal
web
plane,
without
any
combination
in its
plane
and perpendicular
webforce
plane,
without
combination
withtransmitted
iagonal brace) flange
should
to perpendicular
resist to
a the
local
toany
2,5%
of N f,Ed
bywith
the
other
loads.
For
the
design
of
bracing
systems
it
should
be
verified
that
the
bracing
system
is
ange in its plane and perpendicular to the web plane, without any combination with other loads. ab
For theofdesign
of bracing
systemsatiteach
should
be verified
thatatthe
systemlocat
is a
effects
local forces
Qm applied
stabilised
member
thebracing
plastic hinge
For the design of bracing systems it should be verified that the bracing system
effects
of local forces
Qm applied
at eachthat
stabilised
member
at theisplastic
or the design of
bracing
it should
beforces
verified
bracing
system
able tohinge
resistloca
the
is able tosystems
resist the effects
of local
Qm applied the
at each
stabilised
member
at the
hingeatlocations,
where; member at the plastic hinge locations, where;
ffects of local forces
Qmplastic
applied
each stabilised
This verification ensures minimum required stiffness of the bracing system.
This verification ensures minimum required stiffness of the bracing system.
his verification ensures

minimum
required
stiffness
of the
bracing
This verification
ensures
minimum
required
stiffness
of thesystem.
bracing system.
149
Annex BB.3 gives information about stable lengths of149
segments containing
plastic hinges for out-of-plane buckling concerning uniform members made

149I-sections (BB.3.1).
of rolled sections or equivalent welded
Annex BB.3 gives information about stable lengths of segments containing plastic
Annex BB.3 gives information about stable lengths of segments containing plastic hinge
of-plane
buckling
concerning
uniform
members
made
of rolled
sections
or equiv
Annex BB.3
gives
information
about
stable
lengths
of Proceedings
segments
plastic
hinges
of thecontaining
METNET
Seminar
2013
in Lule
129for outof-plane
buckling
concerning
uniform
members
made
of
rolled
sections
or equivalent
sections
(BB.3.1).
of-plane buckling
concerning
uniform members made of rolled sections or equivalent welded Isections (BB.3.1).
sections (BB.3.1).
cks in member without a haunch ( BB.3) /6/
be regarded as being laterally restrained in the plane of the

onnected to a beam and the following condition is met according
Figure 3. Checks in member without a haunch ( BB.3) /6/
Figure 3. Checks in member without a haunch ( BB.3) /6/
Figure 3. Checks
in member
a haunch
BB.3) /6/
Figure
3. without
Checks
in (member
without a haunch ( BB.3) /6/
The beam at the connection may be regarded as being laterally restrained in t

The The
beam
at the
connection
may
be regarded
aslaterally
being restrained
laterally in
restrained in the pla
beam at if
thetrapezoidal
connection may
be regarded
as being to
sheeting
is as
connected
a beam
and the following
condition
The
beamsheeting
at the
connection
may
be sheeting
regarded
being
laterally
restrained
in the
plane
the
ciently
restraint
from
torsional
deformations
[6]
ifconnected
if
trapezoidal
sheeting
is
to
a
beam
and
the
following
condition
isofmet
the plane of the sheeting if trapezoidal sheeting is connected to a beam and
to
EN
1993-1-3
sheeting iftotrapezoidal
sheeting
is connected
to a to
beam
and the following condition is met according
EN
the1993-1-3
following
condition
is met according
EN 1993-1-3
to EN 1993-1-3
(BB.2)
es, designers have to master the theory and the statics in the
es. The information about e.g. product properties i.e. diaphragm
A properties
beam
may
be considered
sufficiently
restraint
from
torsional deformations [6]
A beam
may
considered
asassufficiently
restraint
from
torsional
deformations
stiffness and fastening
would
help
the as
design.
A beam
may
bebeconsidered
sufficiently
restraint
from torsional
deformations [6] if
[6]considered
if
A beam may be
as sufficiently restraint from torsional deformations [6] if
o get familiar with other available material e.g. American Institue
Guidelines Load and Resistance Factor Design Manual 2002,
ckling with restraint from sandwich wall elements and European
Despite
the Eurocodes
many rules, designers have to master the theory and th
ion of Steel Structures
by
Sandwich
Panels
Despite
thethe
Eurocodes
many2013.
rules, designers
have
to the
master
the
theory and the stat
Despite
Eurocodes
many
rules,
have
to
master
theory
and
background
to make
the
rightdesigners
choices.
The
information
about
e.g.the
product
propertie
Despite the
Eurocodes
many
rules,
designers
have
to
master
the
theory
and
statics
ini.e.
thed
background
to
make
the
right
choices.
The
information
about
e.g.
product
properties
the statics in the background to make the right choices. The information about
sandwich
walls,
connection
forces,
stiffness
and
fastening
properties
would
help
the
background
toe.g.
make
the right
choices.
The information
e.g.
product
properties
i.e. diaphragm
product
properties
i.e.forces,
diaphragm
sandwich
walls,
connection
forces,
sandwich
walls,
connection
stiffness
andabout
fastening
properties
would help
the design
sandwich walls,
connection
forces,
stiffness
and
fastening
properties
would
help
the
design.
stiffness and fastening properties would help the design.
To convince oneself it would pay to get familiar with other available material e.g. A
To convince oneself it would pay to get familiar with other available material e.g. America
To
convince
oneself
it would
to
get familiar
with other
available
of
Steel
Constructions
(AISC)
Guidelines
Load
andmaterial
Resistance
Factor Institue
Desig
To convince
oneself
it would
pay
to getpay
familiar
with
other
available
material
e.g.
American
of Steel
Constructions
Guidelines
Load
and Resistance
Factor
Design Man
e.g.Hedman-Ptursson,
American
Institue of(AISC)
Steel
Constructions
(AISC)
Guidelines
Load
and
E.:
Column
buckling
with
restraint
from
sandwich
wall
elemen
of Steel Constructions
(AISC)Column
Guidelines
Loadwith
andrestraint
Resistance
Factor
Design Manual 2002
Hedman-Ptursson,
buckling
fromE.:
sandwich
Resistance Factor E.:
Design Manual
2002, Hedman-Ptursson,
Column wall elements and
150
Recommendations
on the with
Stabilization
of
Steel
Structures
byelements
Sandwichand
Panels
201
Hedman-Ptursson,
E.:
Column
buckling
restraint
from
sandwich
wall
European
Recommendations on the Stabilization of Steel Structures by Sandwich Panels
2013.
Recommendations on the Stabilization of Steel Structures by Sandwich Panels 2013.
130
buckling with restraint from sandwich wall elements and European

Recommendations on the Stabilization of Steel Structures by Sandwich
Panels 2013.
4. MEMBER
BRACING
ACCORDING
AISC
4. MEMBER
BRACING
ACCORDING
TOTO
AISC
[1][1]

MEMBER BRACING
ACCORDING
AISC [1]
4. MEMBER BRACING
ACCORDING
TOTOAISC
[1]
The
LRFD/AISC
design recommendations
for
relative
and
nodal
column
The
LRFD/AISC
design
recommendations
relative
and
nodal
column
bracing
based
The
LRFD/AISC
design
recommendations
forfor
relative
and
nodal
column
bracing
areare
based
onon
an an
bracing
are
based
on
an
initial
out-of-straightness
=0.002L,
where
L is theare based on an
The
LRFD/AISC
design
recommendations
for
relative
and
nodal
column
bracing
initial
out-of-straightness
=0.002L,
where
L
is
the
column
length
and
a
brace
stiffness
equal
initial out-of-straightness
=0.002L, where L is the column length and a brace stiffness equal to to
column length and a brace stiffness equal to twice the ideal stiffness. AISC/
initial
out-of-straightness
=0.002L,
where
Lrequirements
is the column
length
and
a brace
stiffness
twice
ideal
stiffness.
AISC
/LRFD
brace
for
relative
column
bracing
are equal to
twice
thethe
ideal
stiffness.
/LRFD
brace
relative
LRFDAISC
brace requirements
forrequirements
relative columnfor
bracing
are column bracing are
twice the ideal stiffness. AISC /LRFD brace requirements for relative column bracing are
is the
required
compressive
strength
column,
is the
required
brace
where
= 0.75,
Pisu the
b the
where
= 0.75,
Pu where
required
compressive
strength
of of
thethe
column,
andand
Lb Lis
required
brace
= 0.75, Pu is the required compressive strength of the column,
and Lb
where
=
0.75,
P
is
the
required
compressive
strength
of
the
column,
and
L
is
the
required
brace
u
b
spacing.
spacing.
is the required brace spacing.
spacing.
Figure
4.Relative
Relative
column
bracing
Figure 4. Relative
column4.
bracing
[1].
Figure
column
bracing
[1][1]
Figure 4. Relative column bracing [1]
AISC/LRFD
brace
for nodal
column
bracing are:
AISC/
LRFD
brace
requirements
for
nodal
column
bracing
AISC/
LRFD
brace
requirements
forrequirements
nodal
column
bracing
are:are:
AISC/ LRFD brace requirements for nodal column bracing are:
where
= 0.75,
required
compressive
strength
column,
required
where
= 0.75,
PuPu
is is
thethe
required
compressive
strength
of of
thethe
column,
andand
LbLb
is is
thethe
required
where
=
0.75,
Pu
is
the
required
compressive
strength
of
the
column,
and
where
=
0.75,
Pu
is
the
required
compressive
strength
of
the
column,
and
Lb
is
the
required
brace
spacing.
The
formulas
safe
side
specified
after
case
case.
brace
spacing.
The
formulas
areare
onon
safe
side
andand
cancan
be be
specified
after
case
byby
case.
is the required brace spacing. The formulas are on safe side and can be
brace spacing. TheLbformulas
are on safe side and can be specified after case by case.
specified after case by case.
131
where = 0.75, Pu is the required compressive strength of the column, and Lb is the required
race spacing. The formulas are on safe side and can be specified after case by case.
Figure 5. Nodal column bracing [1].
Figure 5. Nodal column bracing [1]

151
Figure 6. Bracing examples [1]
AISC/ LRFD presents as well e.g. continuous metal- deck bracing and le
principles of beam bracing. The LRFD/AISC design recommendations
required bracing forces and the required bracing stiffness of beams and c
include an assumed initial out-of-straightness equal to L/ 500. As expec
are near the simple bracing design rules such as designing a brace to re
compressive force. The stiffness of Plan
the view
brace along with the out of- straigh
a direct
effect[1].on the magnitude of the brace force. If the connections are
Figure
6. Bracing examples
Figure 6.the
Bracing
examples
overall
bracing[1]
system can be significantly less than the stiffness of the b
AISC/LRFD presents as well e.g. continuous metal-deck bracing and leans as well e.g. continuous
metal- deck bracing and lean- on systems and the
on systems and the principles of beam bracing. The LRFD/AISC design
racing. The LRFD/AISC
design recommendations
offer
formulas
the and the
recommendations
offer formulas
for
the
required
bracingfor
forces
5.
STABILIZATION
OF
STEEL
STRUCTURES
BY SANDWICH PAN
s and the required required
bracingbracing
stiffness
of beams
andand
columns.
Theformulas
formulas
stiffness
of beams
columns. The
include an
tial out-of-straightness equal to L/ 500. As expected, the strength criteria
acing design rules Sandwich
such as designing
brace
to resist used
2% offor
theenclosures
member of buildings. They are
panels aare
commonly
e stiffness of the brace
along
with
the
out
ofstraightness
of
the
member
has and wind, to the substr
and they transfer transverse loads, for example snow
magnitude of the brace force. If the connections are flexible, the stiffness of
132
assumed initial out-of-straightness equal to L/500. As expected, the strength

criteria are near the simple bracing design rules such as designing a brace
to resist 2% of the member compressive force. The stiffness of the brace
along with the out of-straightness of the member has a direct effect on the
magnitude of the brace force. If the connections are flexible, the stiffness of
the overall bracing system can be significantly less than the stiffness of the
bracing member alone.
STABILIZATION OF STEEL STRUCTURES BY SANDWICH PANELS [4]

Sandwich panels are commonly used for enclosures of buildings. They are
fixed to a substructure and they transfer transverse loads, for example snow
and wind, to the substructure. When loaded by in-plane shear forces, sandwich
panels have high shear stiffness. Unlike for related building products such
as trapezoidal sheeting or cassette profiles, the shear stiffness of sandwich
panels is usually not taken into account for the design of the building. The
high shear stiffness can be used for different stabilizing effects.
Sandwich panels can restrain the lateral displacement of single components
(e.g. beams, columns). Therefore flexural and lateral torsional buckling is
prevented. By acting as diaphragm sandwich panels can also be used for
global stabilisation of complete building structures and for transferring
horizontal loads, e.g. wind loads. To act as a diaphragm sandwich panels have
to be connected to the substructure by direct fixings.
The European standard EN 14509 covers the manufacture and design of
industrially made self-supporting structural sandwich panels. The use of
sandwich panels as stabilizing elements extends the application area outside
the scope of EN 14509. Therefore, the extended application area shall be
regulated nationally. Sandwich panels used as stabilizing elements have to
fulfil the requirements shown by the CE mark of the product.
The European Recommendations on the Stabilization of Steel Structures by
Sandwich Panels 2013 gives information about the use of self-supporting
sandwich panels as stabilizing elements for single steel members such as
beams or columns.
Torsional restraint [4]

The torsional restraint is governed by the stiffness of the connection of the
sandwich panel to the supporting structure. Recent research carried out
showed that this stiffness significantly depends on the load transferred by the
sandwich panel. A design concept for the quantification and calculation of the
stabilizing effects on beams under predominantly static loading by sandwich
panels was developed within the framework of the EASIE project. Formulae
for calculating the parameters of this moment-rotation-relation are given for
sandwich panels with three different core materials and connections through
the upper or lower crimp.
the
stabilizing
effects
beams
under
predominantly
static
loading
by sandwich
panels was
the
transferred
by theonon
sandwich
design concept
for the
quantification
and calculation
of load
the stabilizing
effects
beams panel.
under A
predominantly
static
loading
by sandwich
panels was
veloped
within
theeffects
ofofthe
EASIE
project.
Formulae
forcalculating
calculating
parameters
of
the stabilizing
on beams
under
predominantly
static loading
by sandwich
panels
was
developed
within
theframework
framework
the
EASIE
project. Formulae
for
thethe
parameters
of of
Proceedings
of
the
METNET
Seminar
2013
in
Lule
developed
within
the
framework
of
the
EASIE
project.
Formulae
for
calculating
the
parameters
of
sthis
moment-rotation-relation
panelswith
withthree
threedifferent
different
core
materials
133 andand
moment-rotation-relationare
aregiven
given for
for sandwich
sandwich panels
core
materials
this
moment-rotation-relation
are
given
for
sandwich
panels
with
three
different
core
materials
and
nnections
through
the
upper
or
lower
crimp.
connections through the upper or lower crimp.
connections through the upper or lower crimp.
Figure 7. Purlin, sandwich panel and details of fastening [4]
Figure
7. Purlin,
panelsandwich
and details of fastening
[4].
Figure
Purlin,
sandwich
panel
and
fastening
[4][4]
Figure
7.7.sandwich
Purlin,
panel
anddetails
detailsofof
fastening
Figure
8. Stabilization:
Figure 8. Stabilization:
torsional
restraint [4].
torsional restraint [4]

Figure 8. Stabilization: torsional restraint [4]
Figure 8. Stabilization: torsional restraint [4]
Limitation of stabilization
Limitation ofmoment
stabilization moment
Limitation of stabilization moment
The torsional
restraint
given bypanels
sandwich
can be calculated
mitation
of stabilization
moment
The torsional
restraint
given
by sandwich
canpanels
be calculated
usingusing
the the
mechanical model
mechanical
model
based on apanels
torsion can
the spring
stiffness
The
torsional
restraint
given
by
bewith
calculated
using
the Cmechanical
model
,k.
based
on a torsion
spring
with
thesandwich
spring
stiffness
Cspring
,k. This spring stiffness is a combination
of the
This
spring
stiffness
is
a
combination
of
the
bending
stiffness
of
the
attached
based
on stiffness
a torsion
spring
withby
thepanel
springCC,k
stiffness
C,k
. This
spring
stiffness
isA,k
athe
combination
of the
ebending
torsional
restraint
given
sandwich
panels
can
be
calculated
using
mechanical
model
,
the
stiffness
of
the
connection
C
and
the
distortional
of
the
attached
panel CC,k, the stiffness of the connection CA,k and the distortional stiffness
bending
stiffness
of
the
attached
panel
C
C,k, the stiffness of the connection CA,k and the distortional
stiffness
C
B,k
of
the
beam
to
be
stabilized.
sed on a torsion spring
the to
spring
stiffness C,k. This spring stiffness is a combination of the
CB,k ofwith
the beam
be stabilized.
stiffness
CB,k ofof
thethe
beam
to be stabilized.
nding
stiffness
attached
panel CC,k, the stiffness of the connection CA,k and the distortional
Withindesign
the simplified
model introduced
[4] ausing
a secant
value
CA
WithinCB,k
the of
simplified
modeldesign
introduced
[4] using
secant
value
C
, ,the stabilization
ffness
the the
beam
to be stabilized.
stabilization moment m shall be limited to the contact moment m . The
Within
[4] using
secant
value CKmoment
, the stabilization
A
momentthem simplified
shall be design
limited model
to the introduced
contact
moment
m . aThe
stabilization
should be
stabilization
moment
should
be calculated
using
[4].
moment
m
shall
be
limited
to
the
contact
moment
m
.
The
stabilization
moment
should be
calculated
using
[4].
thin the simplified
calculated
using [4]. design model introduced [4] using a secant value C, the stabilization
A
oment m shall be limited to the contact moment m . The stabilization moment should be
culated using [4].
A
153
153
134
mitation of the rotation of the stabilized beam

Limitation of the rotation of the stabilized beam
Limitation
of the
rotation of thethe
stabilized
beam
cording
to the
investigations,
rotation
has to be limited to 0.08 . The rotation of th
According
to
the
investigations,
the using
rotation[4]:
has to be limited to 0.08. The
bilized beam can approximately be calculated
rotation of the stabilized
beamhas
can to
approximately
[4]: rotation of the
According to the investigations,
the rotation
be limited be
to calculated
0.08using
. The
stabilized beam can approximately be calculated using [4]:
Lateral restraint in-plane shear resistance [4]
2 Lateral restraint inplane shear resistance [4]
Sandwich
panels have
a highrstiffness
and [strength
when loaded in the plane of
5.2 Lateral restraint
inplane
shear
esistance
4]
the panel. This can be used to stabilize the supporting structure of the panels
ndwich panels have(beams,

a highpurlins,
stiffness
and The
strength
whenofloaded
inpanels
the plane
of the panel. This ca
columns).
deformation
sandwich
themselves
Sandwich panels have
a high
stiffness
andload
strength
when loaded
in theThe
plane
of theofpanel.
This can
caused
by
in-plane
shear
may
normally
be
neglected.
flexibility
the columns).
used
to
stabilize
the
supporting
structure
of
the
panels
(beams,
purlins,
Th
be used to stabilize
theusually
supporting
structure
of flexibility.
the panels
purlins,
columns). The
fixings
dominates
the shear
The (beams,
fixings must
be designed
formation
of ofsandwich
panels
themselves
caused
in-plane
loadthenormally
may normally
b
deformation
sandwich
panelsshear
themselves
caused
by by
shearshear
load
may
be
for the in-plane
load. In typical
cases,
itin-plane
is not necessary
to design
glected.
TheThe
flexibility
ofofpanels
the
usuallydominates
dominates
the
shear
flexibility.
The fixings
sandwich
for thisusually
additional
load, but itthe
is sufficient
to design
the
panels
neglected.
flexibility
the fixings
fixings
shear
flexibility.
The
fixings
must be must b
for theirshear
primary
loading
ofcases,
the distributed
snow
and wind
load
signed
for for
thethe
in-plane
load.
Inconsisting
typicalcases,
is not
necessary
to and
design
the sandwic
designed
in-plane
shear
load.
In
typical
it isit not
necessary
to design
the sandwich
against
the
forces
resulting
from
the
difference
of
the
temperature
between
panels
additionalload,
load, but itit isis sufficient
design
the the
panels
for their
loading loadin
nels
for for
thisthis
additional
sufficientto to
design
panels
for primary
their primary
the faces. However, this rule resulted from current experiments, which shall
consisting
of
the
distributed
snow
and
wind
load
and
against
the
forces
resulting
from
the
difference
nsisting of the distributed
snow andThe
wind
load and
thesandwich
forces resulting
from the differenc
not be generalized.
resistance
of theagainst
individual
panels to inof
the
temperature
between
the
faces.
However,
this
rule
resulted
from
current
experiments,
which whic
plane
shear
load
shall
be
studied
in
each
case.
The
shear
resistance
of
the
the temperature between the faces. However, this rule resulted from current experiments,
shall not be generalized.
Thepanels
resistance
of thebyindividual
sandwich
panels to bonding,
in-plane shear load
individual
is influenced
such
as incomplete
allshall
not be
be studied
generalized.
The
resistance
ofresistance
theimperfections
individual
sandwich
panels to in-plane shear loa
inin each
case.
The shear
of the individual panels is influenced by
addition
to material
properties
and thicknesses.
allimperfections
be studiedsuch
in each
case. The
shear
resistance
of theproperties
individual
panels is influenced b
as incomplete
bonding,
in addition
to material
and thicknesses.
perfections such asSandwich
incomplete
in addition
to the
material
properties
thicknesses.
panelsbonding,
are normally
connected to
supporting
structure and
at the
transverse
edges
only. Theyto
usually
do not have connections
the longitudinal
Sandwich panels are
normally
connected
the supporting
structure atatthe
transverse edges only.
edges. This is common practice, especially for wall panels. Each panel acts as
They usually
have connections
attothe
longitudinal
edges.
This
isthecommon
practice,
ndwich
panels do
arenot
normally
connected
the
supporting
structure
at
transverse
edges only
an individual element.
especially
for
wall
panels.
Each
panel
acts
as
an
individual
element.
ey usually do not have connections at the longitudinal edges. This is common practice
pecially for wall panels. Each panel acts as an individual element.
Figure 9. Displacement
of shear loaded
uni-directionally spanning
sandwich panels
and fastenings.
Figure 9. Displacement
of shear loaded
uni-directionally
spanning
sandwich
panels and fastenings
For stabilization
of eachofbeam
theloaded
shear stiffness
available is spanning sandwich panels and fastenings
gure
9. Displacement
shear
uni-directionally
of the METNET Seminar 2013 in Lule

135
Figure 9. Displacement of shear loaded uni-directionallyProceedings
spanning
sandwich panels and fastenings
For stabilization of each beam the shear stiffness available is
For stabilization of each beam the shear stiffness available is
check
B is the width of a sandwich panel.
B is the width of a sandwich panel

154
Figure
10.
Position
of fastenings
at the supports
of the sandwich panels [4]
Figure 10.
10. Position
of fastenings
at the supports of the
panels [4]. of
Figure
Position
of fastenings
at sandwich
the supports
the sandwich panels [4]
For the stabilization of a beam-column with axial force and bending moment
For the stabilization
of a beam-column with axial force and bending moment
/4/
/4/
For panels at the ends of the beam- column the moment M is approximately [4]:
For panels at the ends of the beam- column the moment M is approximately [4]:
S
the stabilization of a beam-column with axial force and bending moment

136
/4/
/4/
is
approximately
[4]: [4]:
panels
at theatends
of panels
the
beamcolumn
the
M
For
atbeamthe ends
of the beam-column
the
MS is approximately
For panels
the ends
of the
column
themoment
moment
Mmoment
is approximately
[4]:
S
Figure 11. Forces resulting in the moment MS [4]
Figure 11. Forces resulting in the moment MS [4].
Figure 11. Forces resulting in the moment MS [4]
results
in the
shear
in the fastenings.
The momentThe
MSmoment
results M
inS the
shear
forces
VSforces
the
TheseThese
forces act in longitudi
S Mfastenings.
M in V
155
forces
act
in
longitudinal
direction
of
the
panel.
The
highest
forces
arise
in
direction of the panel. The highest forces arise in the outer fastenings of athe
panel. The force in
outer fastenings of a panel. The force in the highest stressed fastenings is [4]:
highest stressed
fastenings is [4]:
155
In addition to the design of the fastenings, the displacements resulting from the stabilization sho
be limited. In addition to the design of the fastenings, the displacements resulting from
the stabilization should be limited.

137
addition to the design of the fastenings, the displacements
resulting from the stabilization
shou
limited.
In addition to the design of the fastenings, the displacements resulting from the stabilization shoul
be limited.
Figure 12. Individual components of a fastening /4/
FigureFigure
12. Individual
components
of of
a fastening
12. Individual
components
a fastening/4/
/4/
Stiffness of fastenings
of fastenings
ffnessStiffness
of fastenings
The translational stiffness of a connection with a self-drilling or self-tapping
The translational
stiffness
of a connection
withself-drilling
a self-drilling
self-tappingscrew
screw fastener
e translational
stiffness
of a can
connection
with
or orself-tapping
fastenercan
canb
screw fastener
be calculated
witha[4]:
calculated with [4]:
culated with [4]:
The fastenings
at the longitudinal joints have a wide influence on the stiffness and also on the loa
The fastenings at the longitudinal joints have a wide influence on the stiffness
bearing
capacity
of the
diaphragm.
also
on the
load
bearing
capacity
of the influence
diaphragm. on the stiffness and also on the lo
e fastenings at theandlongitudinal
joints
have
a wide
ring capacity of the diaphragm.
156
156
138
Figure 13. In the diagram, the shear stiffness S resulting from a stiffness of the fastenings between 0 and
10 kN/mm
is shown
Figure
13. In
the[10].
diagram, the shear stiffness S resulting from a stiffness
of the fastenings between 0 und 10 kN/mm is shown [10]
Figure 14. Concerning the forces of the fastenings there is a wide difference between diaphragms with and
without connections
at theof
longitudinal
joints. [10]. there is a wide difference between
Figure 14. diaphragms
Concerning
the forces
the fastenings
diaphragms with and diaphragms without connections at the longitudinal joints. [10]
139
Stabilisation of frameless sandwich structures

A new application is to use sandwich panels with flat or lightly profiled faces
framelessfor
buildings.
In smaller
buildings
such
as cooling
chambers,
arefordecisive
load-bearing
behaviour
and
capacity.
So the
fastenings have to be desi
climatic
chambers
and
clean
rooms
the
panels
are
applied
without
any
load
forces resulting from transfer of horizontal loads [10].
transferring substructure. Because the in-plane shear stiffness of the panels
are decisive for load-bearing

behaviour
thefastenings,
fasteningsonly
have
be designed for the
is very much higher
thanand
the capacity.
stiffness ofSothe
thetofastenings
are decisive
behaviour
forces resulting from
transferfor
ofload-bearing
horizontal loads
[10]. and capacity. So the fastenings have
to be designed for the forces resulting from transfer of horizontal loads [10].
Figure 15. Transfer of horizontal loads in frameless buildings
Figure
15.ofTransfer
of inhorizontal
loads
Figure
15. Transfer
horizontal loads
frameless buildings.
in frameless buildings
Design
forsandwich
sandwich
panels
subjected
to bending
and transvers
Design procedures
procedures for
panels
subjected
to bending
moments moments
and
Design procedures
for
panels
subjected
to general
bending
moments
and
transverse
forces
are
given
in sandwich
ENforces
14509,
there
are14509,
no
design
methods
panels
subjected
to axi
transverse
arebut
given
in EN
but there
are no
generalfor
design
given in EN 14509,
but
there
are
no
general
design
methods
for
panels
subjected
to
axial
loads
or
a
methods for of
panels
to axial loads
a combination
of axialmethod
and
combination
axialsubjected
and transverse
loadsoravailable.
A design
for axially load
combination ofpanels
axial
and
loadsAavailable.
A design
method
axially panels
loaded sandwich
transverse
loads available.
design
for axially
loadedfor
sandwich
hastransverse
been
developed
[10]. method
panels has been developed
[10]. [10].
has been developed
Figure 16. Wall panels with

typical loads [10].
Figure 13. Wall panels with typical loads [10]
140
loaded by axial load N and bending moment M, can be designed with

Panels loaded by Panels
axial load
N and bending moment M, can be designed with the formulae [10]::
the formulae [10]:
158
COLUMN BUCKLING WITH RESTRAINTS FROM SANDWICH WALL ELEMENT
OLUMN BUCKLING
WITH RESTRAINTS
FROM
SANDWICH
WALL ELEMENT
Eva Hedman-Petursson
from Lule University
of Technology
has researched
in the years 19982001 the restraining effect of a sandwich element wall

on steel beam-columns. The wall elements give restraint to beam-columns
Hedman-Petursson
from
Lulein University
has ofresearched
against
buckling
the wall plane of
andTechnology
that the resistance
the member in the years 19
1 the restraining effect
sandwich
wall ontests,
steeltheoretical
beam-columns.
thereby of
canabe
increased. Aelement
series of full-scale
modelling The wall elem
and
FE-calculations
have
been
performed
at
the
Division
of
Steel
Structures,
restraint to beam-columns against buckling in the wall plane and that the resistance o
Lule University of Technology.
mber thereby can be increased. A series of full-scale tests, theoretical modelling and
Her doctoral
thesis showed
that there
is a substantial
increase
in resistance toLule Universit
ulations have been
performed
at the
Division
of Steel
Structures,
gain that is worth taking into consideration with the element wall assembled
hnology.
in a conventional way, that is, without connection between the elements.
If the wall structure is recognized as a restraint, a brace, and designed
accordingly
a situation
support for beam-columns
be offered.to gain that is w
doctoral thesis showed
that
there with
is a full
substantial
increase in can
resistance
The constructive step towards this is to strengthen the horizontal element
ng into consideration
the element
wall cases
assembled
a conventional
joints.with
Consequently,
two different
have beeninstudied;
how to benefitway, that is, wit
fromelements.
a partial restraint
in thewall
designstructure
of beam-columns
with the wall elements
nection between the
If the
is recognized
as a restraint, a brace,
assembled
in
the
conventional
way;
and
how
to
design
the
element
wall to
gned accordingly a situation with full support for beam-columns
can be offered.
offer a full lateral support for a beam-column. Design recommendations were
tructive step towards
this
istwo
tocases
strengthen
the horizontal element joints. Consequently,
given for
those
[11].
rent cases have been studied; how to benefit from a partial restraint in the design of b
mns with the wall elements assembled in the conventional way; and how to design the ele
to offer a full lateral support for a beam-column. Design recommendations were given
e two cases [11]
Table 1. Results from full scale testing and FE-simulation.
Test nr.;
cross section
Moment Axial force

lever arm lab. test
Test 1;
IIEA 120
Test 2;
IIEA 120
0.05 m
Test 3;
IPE 200
Test 4;
IPE 200
0.25 m
Test 5;
IPE 200
Failure mode
FE-simulation EC3
413.0 kN
405.3 kN
411.5 kN
398.6 kN
223.6 kN
212.5 kN
220.1 kN
211.8 kN
407.1 kN
2)
413.4 kN
414.7 kN
2)
408.2 kN
484.1 kN1)
flexural buckling
217.0 kN1)
flexural buckling
415.6 kN1)
lateral torsional buckling
191.6 kN1)
flexural buckling
204.4 kN
184.0 kN
199.2 kN
182.7 kN
268.3 kN
254.2 kN
125.5 kN 3)
partially restrained buckling
272.5 kN
261.7 kN
412.3 kN
in weak direction
4)
1)
Full support for the compression flange in calculation of resistance according to EC3, equation 1
Testrig failure just before beam-column collapse
3)
Unrestrained column capacity, buckling in weak direction
4)
Fully restrained column, lateral torsional buckling
2)
SUMMARY
In this study bracing systems used to control the instability are presented. The
basic theory of a dual strength and stiffness criterion for the design of bracing
used to control the instability is studied. Eurocode is reviewed from the
bracing point of view. Despite the Eurocodes many rules, designers have to
master the theory and the statics in the background to make the right choices.
The information about e.g. product properties of diaphragm sandwich walls,
connection forces, stiffness and the fastening properties have to be solved
independently.
American Institute of Steel Construction (AISC) Guidelines on Load and
Resistance Factor Design Manual 2002, the work of Hedman-Ptursson, E.,
Column buckling with restraint from sandwich wall elements and European
2013 were examined to confirm the present bracing practices.
The European standard EN 14509 covers the manufacture and design of
industrially made self-supporting structural sandwich panels. The use of
sandwich panels as stabilizing elements extends the application area outside
the scope of EN 14509. Therefore, the extended application area shall be
regulated nationally. The sandwich panels used as stabilizing elements have
to fulfil the requirements shown by the CE mark of the product. The European
141
142
2013 gives information about the use of self-supporting sandwich panels as

stabilizing elements for single steel members such as beams or columns. A
new application is to use sandwich panels with profiled faces for frameless
buildings. A design method for axially loaded sandwich panels has been
presented.
REFERENCES
1
Structural Bracing. Brian Chen Joseph Jura.
Load and Resistance Factor Design Manual. American Institute of

Steel Construction (AISC)
Column buckling with restraint from sandwich wall elements.

Hedman-Ptursson, E.
European Recommendations on the Stabilization of Steel Structures

by Sandwich Panels 2013 CIB
EN 1990: 2002 + A1:2005 + A1:2005/AC:2010: Eurocode: Basis of

structural design.
EN 1993-1-1:2005 + AC:2009: Eurocode 3: Design of steel structures

Part 11:General rules and rules for buildings.
EN 1993-1-3:2006 + AC:2009: Eurocode 3: Design of steel structures

Part 13: General rules Supplementary rules for cold-formed
members and sheeting.
EN 14509:2006: Self-supporting double skin metal faced insulating

panels Factory made products Specifications.
FprEN 14509:2013: Self-supporting double skin metal faced insulating

panels Factory made products Specifications.
10
EASIE Ensuring Advancement in Sandwich Constructions Through

Innovation and Exploitation http://www.easie.eu/
11
Column buckling with restraint from sandwich wall elements HedmanPtursson, Eva Doctoral Thesis / 2001-08-30.
INNOVATION IN MECHANICAL FASTENING

TECHNOLOGY FOR MAINTENANCE-FREE JOINTS
Hans-Albert Staedler
Alcoa Fastening Systems Industrial Products, Telford, UK
ABSTRACT
The paper describes innovative mechanical bolts which are maintenance free.
INTRODUCTION
Alcoa Fastening Systems have experience of many years in mechanical
fastening technologies. Louis Huck (the founder of Huck fasteners) developed
in the middle of 1940s the very first lock bolt system. The functionality of
lock bolt systems is based on cold forming / extrusion technology. A socalled collar gets cold formed during installation by an anvil (part of a special
installation tool) into unique locking grooves of a pin. After the installation
the fastener is mechanically locked due to this volume-controlled swaging
process. The locking grooves are very often filled with a collar material. This
system results in a natural barrier against vibration and loss of tightening.
Lock bolt systems create vibration resistant and high fatigue-life joints. Since
the very first lock bolts, these systems have been continuously optimized.
A major step was, when Alcoa Fastening Systems came to the market with
pintail-less lock bolts. These systems not longer have a pintail end with blank
and rust effected surface. The installation is faster, smoother and with reduced
noise compared with traditional lock bolts and much faster than conventional
bolts. Germanys highest construction authority DIBt tested the Bobtail in
accordance with the EN 1993 regulation and approved it as appropriate for
generating totally maintenance-free joints when correctly installed and with
correlated parts. The Bobtail delivers also the performance of ISO 12944 C5
M corrosion protection.
Table 1 shows the evolution of lock bolt systems compared with high strength
conventional bolts. The table includes the latest innovation product, the Huck
360 (H360). The H360 bolt thread is similar to a lock bolt system, has
mild contours and has no stress concentration areas. In comparison with the
lock bolt system, the Huck 360 no longer needs a special installation tool.
It is a removable and reusable nut and bolt system (Figure 1) with vibration
resistance, high fatigue strength similar and installation flexibility to the
equivalent lock bolt. Due to the tightening installation method the tightening
range is larger than for the lock bolt system, but smaller than conventional
bolt systems. The ductile nut thread and its deformation during the tightening
process generate a mechanical lock with a hardened thread of the pin. This
makes the Huck360 the highest strength bolting system for connections
requiring field service with standard tools.
143
144

Table 1: Lock bolt evolution [Source Alcoa Fastening Systems Test Lab. Waco}
High Strength
Bolt & Nut
Traditional
Lock Bolt
Advanced
Lock Bolt
Table
Lock
evolution
[Source
Alcoa Fastening
Systems
Test Lab. Waco}
Table 1:
1. Lock
boltbolt
evolution
[Source
Alcoa Fastening
Systems Test
Lab. Waco].
High Strength
Bolt & Nut
Features
Clamp Load
Features
Clamp
Clamp Load
Variation
Clamp
Removability
Variation
16mmTool
Removability
Weight
16mmTool
Vibration
Weight
Resistance
Vibration
Resistance
Traditional
Lock Bolt
High
30%
High
6%
Yes
< 10Yes
lbs.
Semi-permanent
20 lbs.
Semi-permanent
Low
High
High
30%
< 10 lbs.
Low
High
6%
Advanced
Lock Bolt
High
6%
High
6%
Semi-permanent
20 lbs.
< 10 lbs.
Semi-permanent
< 10 lbs.
High
High
High
Huck 360
System
Huck 360
System
High
20%
High
20%
Yes
< 10 lbs
Yes
< 10 lbs
High
High
Figure
1.Figure
Parts
system;
//Source:
Alcoa
fastening
Systems
test lab Waco//
1:H360
Parts
H360
system;
//Source:
Alcoa
fastening
Systems
test
lablab
Waco//
Figure
1: ofParts
ofofH360
system;
//Source:
Alcoa
fastening
Systems
test
Waco//
The free running nut thread quickly spins down and tightens while minimizing damage to coatings.
TheThe
free preloaded
running nut
threadthread
quickly
down andbolt/
tightens
while
minimizing
damage
to coatings.
H360
andspins
conventional
nutand
thread
detail while
(Figure
2) shows
the
The
free running
nut
thread
quickly
spinsbolt/
down
tightens
minimizing
Thedifference
preloaded
H360
thread
and
conventional
nut
thread
detail
(Figure
2)
and reason for this high resistance against vibration. The nut thread is designed in shows
a style, the
damage
to enough
coatings.
Theduring
preloaded
H360
thread
conventional
bolt/
difference
and reason
for this
high
resistance
against
vibration.
Theanand
nut
thread
is designed
in a style,
which deforms
material
the tightening
process for
afficient
locking
mechanism.
nut
(Figure
2) and
shows
the
reason for
high
which
deforms
enough
material
during
theallows
tightening
process forand
anreusability.
afficient
locking
mechanism.
The
nutthread
thread
isdetail
in ductile
hardness
fordifference
removability
and
Thisthis
means,
that
Thethe
nut
threadisissuitable
in ductile
hardness
and
allows
removability
reusability.
Thisstrategies
means, that
system
to solve
work piece
surface
imperfections
and and
/ or screw
resistance
against
vibration.
The
nutfor
thread
is designed
ininstallation
a style,
which
the system
is suitable
solvewind
work
piecetowers.
surface
imperfections
and /for
or screw
installation
strategies
for
instance
in
largeto
flange
turbine
deforms
enough
material
during
the tightening
process
an afficient
locking
for instance
in largeThe
flange
wind
turbine
towers.
mechanism.
nut
thread
is in
ductile hardness and allows for removability
and reusability. This means, that the system is suitable to solve work piece
surface imperfections and/or screw installation strategies for instance in
large flange wind turbine towers.
145
Figure 2: Functionality of H360 system; //Source: Alcoa fastening Systems test lab Waco//
Figure 2. Functionality of H360 system; //Source: Alcoa fastening Systems test lab Waco//
Figure 2: Functionality of H360 system; //Source: Alcoa fastening Systems test lab Waco//
The Alcoa internal test results are shown in Figure 3. The conclusions drawn from the tests are
TheinAlcoa
test results are shown in Figure 3. The conclusions drawn
given
Tableinternal
2.
The Alcoa internal
results
are in
shown
from thetest
tests
are given
Tablein
2. Figure 3. The conclusions drawn from the tests are
given in Table 2.
Table 2: Summary of Huck 360 features and benefits

[Source
Alcoa
Fastening
Test Lab.
Table 2. Summary
of Huck
360
features andSystems
benefits [Source
Alcoa Waco}
Fastening Systems Test Lab. Waco].
Table 2: Summary of Huck 360 features and benefits

FeaturesAlcoa Fastening Systems Test Lab.Benefits
[Source
Waco}
Features
Vibration Proof
running thread
Vibration Free
Proof
Free running thread
Shallow low notch factor bolt thread
Reduction
Fast fit-up
in frequency of maintenance
No coating damage
Longer
equipment
uptime
Easy removal
Field serviceable
Fast fit-up
coating
damage
No High
fatigue
strength
Easy
Absorbs
high spike loads
removal
Field serviceable
Grade8 and Class10.9 tensile and shear

strength
Shallow low notch factor bolt thread

Installed
wrenches
and
Reduction in frequency of maintenance
Longer equipment uptime

Benefits
removed
Easy upgrade from conventional

nuts and bolts
High fatigue strength
Absorbs
high spike loads
Field serviceability
conventional
Grade8 and Class10.9 tensile and shear

strength
Easy upgrade from conventional

nuts and bolts
Installed
wrenches
Field serviceability
and
removed
conventional
146
Figure 3: Performance of H360 system; //Source: Alcoa fastening Systems Test Lab Waco//
Figure 4 shows all the metric and imperial size diameters which are actually available. In
accordance with construction industry needs the diameters are M24, M27, M30 and M42.
An example for usage of H360 system is in the mining industry. Morbark Inc. were having poor
insert (tooth) life onFigure
a Model
1300 Tub
Grinder.
stoverized
nuts Test
andLabGrade
3. Performance
of H360
system;Using
//Source:
Alcoa
fastening Systems
Waco// 8 Bolts. They
were losing upwards ofFigure
6 to 8 teeth
a day according to
operator. system;
Using the Huck
360 they
3: Performance
oftheH360
//Source:
Alcoa
were able to save approximately 120 hours over a period of five weeks. During that period of time
only 8-10 inserts were lost.
fasteni
Figure 4 shows
all the
metric andSystems
imperial size
diameters
which are actually
360 system; //Source:
Alcoa
fastening
Test
Lab Waco//
available. In accordance with construction industry needs the diameters are
M24, M27, M30 and M42.
example for usage of H360 system is in the mining industry. Morbark

ic and imperialAn
size
diameters which are actually available. In
Inc. were having poor insert (tooth) life on a Model 1300 Tub Grinder. Using
ndustry needs thediameters
areand
M24,
M27,
M30
stoverized nuts
Grade
8 Bolts.
Theyand
were M42.
losing upwards of 6 to 8
teeth a day according to the operator. Using the Huck 360 they were able to
save approximately 120 hours over a period of five weeks. During that period
the
mining
industry.
Morbark
Inc. were having poor
of time
only 810
inserts were
lost.
0 system is in
300 Tub Grinder. Using stoverized nuts and Grade 8 Bolts. They
8 teeth a day according to the operator. Using the Huck 360 they
y 120 hours over a period of five weeks. During that period of time
Figure
Available H360
H360 system
//Source:
AlcoaAlcoa
fastening
Systems//
Figure
4: 4.Available
systembolts
bolts
//Source:
fastening
Systems//
147
148
Figure 4: Available H360 system bolts //Source: Alcoa fastening Systems//
(a)
(b)
Figure 5: Morbark Inc. Grinder with H360; //Source: Alcoa fastening Systems //
Figure 5. Morbark Inc. Grinder with H360; //Source: Alcoa fastening Systems //
Other opportunities
for opportunities
using H360 arefor
in using H360 are in
Other
Steel construction (Figure 6)
Steel construction (Figure 6)
Truck and trailer
Truck
andmachinery
trailer
Agriculture,
harvest
Agriculture, harvest machinery
Rail and railcar
Car industry
Rail and railcar
Mining
Car industry
Construction machines and excavators
repair
Mining
After and
market where normal screws failed
Construction machines and excavators
After and repair market where normal screws failed
Figure 6: H360 in Steel construction

Figure 6. H360 in Steel construction.
Figure 6: H360 in Steel construction
or all applications in construction market an approval is necessary. In Germany, the approach is

hrough DIBt approval. DIBt is the German authority for the fulfilment of technical tasks in the
eld of public law. This authority sets National Technical approvals. The DIBt is member of
ForOrganization
all applications
in construction
market an approval is necessary. In Germany, the approa
OTA (European
for Technical
Approvals).
For all applications in construction market an approval is necessary. In

Germany, the approach is through DIBt approval. DIBt is the German
authority for the fulfilment of technical tasks in the field of public law. This
authority sets National Technical approvals. The DIBt is member of EOTA
(European Organization for Technical Approvals).
The actual status is of this approach is:
1.
The approach has been generated and the DIBt decided the accredited
test laboratory and the assessor.
2.
The test program has been developed and approved. The test program
includes the tests in accordance with categories A, B, C, D and E in
accordance with DIN EN 1993-1-8. The clear advantage is supposed for
high strength and friction controlled connections (category B, C and E
after DIN EN 1993-1-8.
3.
Preliminary and first tests with M27 diameter have been done and
results look good.
Preliminary test for bolt diameter M 27

The aim of these preliminary tests was to get a first impression from an
independent third party about the service which can be provided with
H360 bolt system. These preliminary results are the basis for the official
test program, which will start at the end of 2013. All test specifications have
been developed by Fraunhofer AGP Rostock and in accordance with DIBt
requirements. Specimen for tests material was S355. The Preload (clamp
force) sensor is from company Kistler.
1.
Characteristic preload and preload loss F p,C measured with Piezo

system (Figure 7)
2.
Slip resistance at serviceability and ultimate F v Rk.
3.
Cyclic load tests for identifying of self-loosening effects in friction

controlled joints.
4.
Preload after multiple retightening processes.
149
150
Table 3: Categories of bolted connections [Source: EN1993-1-8]

Table 3. Categories of bolted connections [Source: EN1993-1-8].
results of characteristic
loss Fp,C
Test results ofTest
characteristic
preload andpreload
preloadand
losspreload
Fp,C
The range
of installed
arises from
the tightening
The range of installed
preload
is arises preload
from theistightening
installation
method. installation
The loss of preload
Thesystem
loss ofand
preload
is the
veryminimal
close tosetting
a lock effects
bolt system
showssystem.
the
is very close tomethod.
a lock bolt
shows
in theand
installed
The
setting
effects
installed
The reason of
forpreload
this isloss
the Fp,C has
reason for this minimal
is the impact
of bolt
and in
nutthe
thread
style. system.
The determination
of one
bolt bolt
and nut
Thehours.
determination
loss been
Fp,C has
been identifiedimpact
for each
afterthread
24, 48style.
and 96
Therefore,ofapreload
sensor has
placed into
identified
for 24,
each48,
oneorbolt
48nut
andhas
96 hours.
Therefore,
a sensor
the preloaded been
connection.
After
96after
hours24,the
been removed
with
an automatic
has disconnection
been placed into
connection.with
After
48, or 96ofhours
the
screwdriver. The
of the
the preloaded
joint is synchronous
the24,
discharge
the sensor.
This
nut
has
been
removed
with
an
automatic
screwdriver.
The
disconnection
supplies a negative signal voltage, which is equivalent to the preload at the point of time before
joint is synchronous
with
the discharge
of the sensor.
This
disconnecting. ofAthe
continuous
determination
of preload
is impossible
with this
testsupplies
method.a For the
negative signal voltage, which is equivalent to the preload at the point of time
long time preload loss different tests methods are necessary.
before disconnecting. A continuous determination of preload is impossible
with this test method. For the long time preload loss different tests methods
are necessary.
Test results about slip resistance Fv Rk
These tests have been done on double shear specimen S355J2+N material with a hole tolerance
Test results about slip resistance F
d = 2,0 mm. The holes in diameter 29mm havev Rk
been drilled. Style and size of specimen deliver
bolt failure as criterion. The bolts have been equipped with hardened washers. The test speed was
These tests have been done on double shear specimen S355J2+N material
1mm/min untilwith
boltafailure.
To get specimen surface friction coefficient 0,5, Zinc-Silicate in
hole tolerance d = 2,0 mm. The holes in diameter 29mm have been
accordance with
DIN
18800-7
has been used. The bolt failure occurred as expected. The A linear
drilled. Style and size of specimen deliver bolt failure as criterion. The bolts
inductive displacement
transducer
the relative
and absolute
outer and
have been equipped picks
with up
hardened
washers.
The test expansion
speed was between
1mm/min
middle work piece.
Displacement
has
been
measured
up
to
200
m.
The
servohydraulic
until bolt failure. To get specimen surface friction coefficient 0,5, Zinc-machine
software delivered
machine
force and transfer
displacement.
Silicate
in accordance
with DIN
18800-7 has been used. The bolt failure
occurred as expected. The A linear inductive displacement transducer picks
up the relative and absolute expansion between outer and middle work piece.
151
Displacement has been measured up to 200 m. The servo-hydraulic machine

software delivered machine force and transfer displacement.
Figure 7: Draft of test results Preload and pre

Figure 7. Draft of test results Preload and preload lossTest
after 24 hours;
//Source: Fraunhofer AGP
Test
Laboratory
Rostock
, Prfberic
Laboratory Rostock, Prfbericht
Nr. P-FhAGP 7:
V1109-03,
Rostock 2012
// test results Preload and pre
Figure
Draft
of
Figure 7: Draft of test results Preload and preload loss after 24 hours; //Source: Fraunhofer AG
Test Laboratory Rostock , Prfbericht
P-FhAGP V1109-03,
Rostock
2012 //
Test Nr.
Laboratory
Rostock
, Prfberi
152
Figure 8: Draft test results for shear and slip resistance; //Source: Fraunhofer AGP Test
Laboratory Rostock , Prfbericht Nr. P-FhAGP V1109-03, Rostock 2012 /
Cyclic load tests to determine self- loosening effects in friction controlled connections
Friction controlled joints are preferred in reverse and /or cyclic loaded connections. Keeping the
preload is essential for long service life. Conventional bolt and nut connections tend to unscrew
with time. This effect reduces the fatigue life of the joints. To determine these effects for H360
system the following preliminary test has been done.
Specimen used in double shear S355J2-N plates, dry, no rust and uncoated.
Hole diameter 29mm with tolerance d = 2,0 mm, drilled.
Hardened washers have been used.
Figure 8: Draft test results for shear and slip resistance; /
153
Figure 8: Draft test results for shear and slip resistance; //Source: Fraunhofer AGP Test
Figure 8. Draft
test results for
shear and ,slip
resistance; //Source:
AGPTest
Laboratory Rostock 2012 /
Laboratory
Rostock
Prfbericht
Nr.Fraunhofer
P-FhAGP
V1109-03,
Rostock, Prfbericht Nr. P-FhAGP V1109-03, Rostock 2012 /
Cyclic load tests to determine self- loosening effects in friction controlled connections
Cyclic
load tests
to determine
self-loosening
Friction
controlled
joints
are preferred
in reverse effects
and /or cyclic loaded connections. Keeping the
in is
friction
controlled
connections
preload
essential
for long
service life. Conventional bolt and nut connections tend to unscrew
with time. This effect reduces the fatigue life of the joints. To determine these effects for H360
controlled
joints are
in reverse and /or cyclic loaded
systemFriction
the following
preliminary
test preferred
has been done.
connections.
Keeping
the preload
is essential
for long
Conventional
Specimen
used
in double
shear S355J2-N
plates,
dry,service
no rustlife.
and
uncoated.
bolt
and
nut
connections
tend
to
unscrew
with
time.
This
effect
reduces the
fatigue life of the joints. To determine these effects for H360 system the
Hardened
washers have been used.
following preliminary test has been done.
Specimen used in double shear S355J2-N plates, dry, no rust and uncoated.
Hardened washers have been used.
Bolts have been equipped with strain gauges which have been calibrated in
force range 50 up to 400 kN. Therefore, a force can be assigned for each signal
voltage. For statistical reasons 3 specimens were pre-tested. The testing came
out with individual sliding match 113 kN, 153 kN and 182 kN, the average was
149 kN. Further details are described in test report, revision 2 of Fraunhofer
AGP Rostock.
Bolts
have
been
equipped
with
strain
gauges
which
have
been
calibrated
force
range
Bolts
have
been
equipped
with
strain
gauges
which
have
been
calibrated
in in
force
range
5050
upup
to to
400
kN.
Therefore,
a
force
can
be
assigned
for
each
signal
voltage.
For
statistical
reasons
400 kN. Therefore, a force can be assigned for each signal voltage. For statistical reasons
3 specimens
were
pre-tested.The
The
testing came out with individual sliding match 113 kN, 153 kN
3 specimens
were
pre-tested.
testing
Proceedings
of
the METNET Seminar
2013
in Lule came out with individual sliding match 113 kN, 153 kN
154
and
182
kN,
the
average
was
149
kN.Further
Furtherdetails
detailsarearedescribed
describedin intest
testreport,
report,revision
revision2 2of of
and 182 kN, the average was 149 kN.
Fraunhofer
AGP
Rostock.
Fraunhofer AGP Rostock.
Bolts have been equipped with strain gauges which have been calibrated in force range 50 up
400 kN. Therefore, a force can be assigned for each signal voltage. For statistical reaso
3 specimens were pre-tested. The testing came out with individual sliding match 113 kN, 153 k
and 182 kN, the average was 149 kN. Further details are described in test report, revision 2
Figure
Draftof oftest
testresults
resultsself-loosening
self-looseningeffects
effectsunder
undercyclic
cyclicload;
load;/Source:
/Source:Fraunhofer
Fraunhofer
Figure
9: 9: Draft
AGP
test
laboratory
Rostock
, Prfbericht
P-FhAGP
V1109-03,
Rostock
2012
AGP
test
laboratory
Rostock
, Prfbericht
Nr.Nr.
P-FhAGP
V1109-03,
Rostock
2012
Bolts have been equipped with strain gauges which have been calibrated in force range 50 up to
400 kN. Therefore, a force can be assigned for each signal voltage. For statistical reasons
Proceedings of the
METNET Seminar
2013113
in LulekN, 153
155 kN
3 specimens were pre-tested. The testing came out with individual
sliding
match
and 182 kN, the average was 149 kN. Further details are described in test report, revision 2 of
Figure 9:
Draft of test results self-loosening effects unde

AGP test laboratory Rostock , Prfbericht Nr. P-
Figure 9. Draft of test results selfloosening effects under cyclic load;

/Source: Fraunhofer AGP test
laboratory Rostock, Prfbericht Nr.
P-FhAGP V1109-03, Rostock 2012.
Figure 9:
Draft of test results self-loosening effects under cyclic load; /Source: Fraunhofer
156
The results below in picture 10 show the accuracy of preload a

process. The process was by torque control and similar for b
Proceedings ofinstallation
the METNET Seminar 2013
in Luleand test conditions.
tool
360 system
Ability of retightening of H360 system

This first step of the test about the performance of H360 system for remove
and reuse does demonstrate the ability for instance to solve work piece
imperfections in large flange application.
The results below in picture 10 show the accuracy of preload after 5 time
complete retightening process. The process was by torque control and similar
for both systems in terms of usage of installation tool and test conditions.
about the performance of H360 system for remove and reuse does
stance to solve work piece imperfections in large flange application.
10 show the accuracyH360
of preload
after
5 angle
time complete retightening
preload in kN over
rotation
y torque control and similar for both systems in terms of usage of
ditions.
Conventional HV screw system M27 preload in kN over rotation angle
Figure 10:
Draft of test results retightening of H360; /Sourc

Rostock ,2012
CONCLUSIONS
Overall the H360 system is already in many applications in ra

United States and other countries and now on the approach to get
ability to generate maintenance free joints.
Figure 10. Draft of test results retightening of H360; /Source: Fraunhofer AGP test laboratory Rostock, 2012.
CONCLUSIONS
ults retightening
of H360; /Source: Fraunhofer AGP test laboratory
Overall the H360 system is already in many applications in rail, mining and
construction in the United States and other countries and now on the approach
to get the DIBt approval. The aim is the ability to generate maintenance free
joints.
FLANGELESS CONNECTIONS
IN STEEL TUBULAR WIND TOWERS
Christine Heistermann1
Anh Tuan Tran1
Milan Veljkovic1
Carlos Rebelo2
Lule University of Technology
Department of Civil, Environmental and Natural Resources Engineering
Division of Structural and Construction Engineering
Research Group of Steel Structures
Sweden
1
University of Coimbra
Department of Civil Engineering
Institute for Sustainability and Innovation in Structural Engineering
Steel and Mixed Construction Technologies
Portugal
2
ABSTRACT
Extensive research is currently being conducted on the improvement on the use of
renewable energy. One field is the use of wind energy, where the tower construction is one
of the main issues. This paper deals with new ideas and ongoing research in this area. To
raise the height of steel tubular towers, fatigue as the design limit and constraints due to
transportation issues have to be overcome. Changes in the cross-section are considered as
one of the possible solutions. This work presents an extensive finite element study dealing
with different ways to improve shell stability, which become the limiting criteria if a friction
connection substitutes the common flange connection between two tower segments. The
use of circular and polygonal cross-sections is briefly described and will be investigated
in an experimental programme.
INTRODUCTION
Wind is one of the most important sources of renewable energy. One of its
main advantages is very low CO2-emission while it generates power (Burton
et al. 2001) within a very short payback time. Wind power stations convert
kinetic energy from wind into electrical energy. During the last years, wind
energy had an increase of about 25%. The installed capacity for wind power
all over the world increased from 6 GW in 1996 up to 121 GW in 2008, of
which about 60% are produced in OECD-Europe. In September 2012 the
installed wind power capacity in the European Union hit the milestone of 100
GW, which equals the output of 39 nuclear power plants (EWEA 2012a). This
continuing trend requires growing production.
157
158
The European Wind Initiative (EWI) aims for making wind energy produced
onshore the most competitive energy source by 2020. Until 2030 offshore
wind energy shall follow (EWEA 2012b, 2013a, 2013b, Fichaux et al. 2011).
To improve the wind energy industrys competitiveness, technologies for
turbines and components, grid integration and management methods will be
focused on. In addition, the design and layout of wind farms as well as the
manufacturing process and structural concept is receiving attention as the
trend of building higher towers is continuing. Therefore, the project Highstrength tower in steel for wind turbines (HISTWIN) and the dissemination
of knowledge gained in it fills an important gap in promoting the design of
wind towers, which has been seen as the exclusive work of highly specialized
offices. The competitiveness of steel tubular towers as support for wind
turbines improves by using an innovative solution for the assembling joint:
a friction connection with long open slotted holes (Veljkovic et al. 2012).
Thereby, one of the most common design criteria of the traditional towers, the
fatigue resistance of the flange connection, is overcome and the design of the
shell stability becomes an interesting issue. This opens the possibility to use
higher strength steel grades if the tower shell is properly designed.
Other advantages of this type of connection are
Simpler design,
Reduction of construction time accompanied by a decrease of costs for
construction,
Lower production and maintenance costs,
Improvement of work conditions, as the bolts can be installed and
maintained always from inside the tower.
Hence, steel tubular towers with friction connections hold the potential to
outclass traditional flange connection towers, which has been proven clearly
within the scope of the HISTWIN project (Veljkovic et al. 2012). To connect
the tubular segments of steel towers in a flangeless way is one of the main
objectives of this project. The substitution of the expensive and fatiguesusceptible flange connections with friction connections with long open
slotted holes is illustrated in Figure 1. The specimen shown represents one of
the 1:2 scale prototypes developed within the framework of the projects.
tive solution for the assembling joint: a friction connection with

fatigue-susceptible
connections
with friction connections with
jkovic et al. 2012).
Thereby, one of flange
the most
common design
is illustrated
Figure
1. The specimen
shown
one of t
Proceedings
of the METNET
Seminarrepresents
2013 in Lule
159
ers, the fatigue resistance
of theinflange
connection,
is overcome
the framework
of the projects.
ability becomes andeveloped
interestingwithin
issue. This
opens the possibility
to
es if the tower shell is properly designed.
of connection are
tion time accompanied by a decrease of costs for construction,

maintenance costs,
conditions, as the bolts can be installed and maintained always
with friction connections hold the potential to outclass traditional

hich has been proven clearly within the scope of the HISTWIN
). To connect the tubular segments of steel towers in a flangeless
ectives of this project. The substitution of the expensive and
nnections with friction connections with long open slotted holes
he specimen shown represents one of the 1:2 scale prototypes
ork of the projects.
iction connection
a) Main concept of the friction connection
a) Main concept of the friction connection
b) Fe
Figure 1: The HISTWIN connection, friction connection for the flang
b) Feasibility test
b) Feasibility test
Figure 1. The HISTWIN connection, friction connection for the flangeless wind tower.
nection, friction connection for the flangeless wind tower
160
In this type of connection, the two tower segments easily slide over each oth
slots are cut into the lower segment, whereas the upper one contains normal cl
in which the bolts can be preinstalled and will be used for angular alig
positioning
of the upper
tower
Dueslide
to long
open
slots the lower towe
In this type of connection,
the two
towersegment.
segments easily
over each
other.
Long
open
slots
are
cut
into
the
lower
segment,
whereas
the
upper
one
contains
flexible and therefore less sensitive to assembling tolerances. During the tighten
normal clearance holes, in which the bolts can be preinstalled and will be
the
bolts, these flexible fingers will move towards the stiffer outer tower segm
used for angular alignment during positioning of the upper tower segment.
Due to long open slots the lower tower part is quite flexible and therefore less
sensitive toofassembling
tolerances. During
tightening
process
of the bolts,
Instead
single washers,
cover the
plates
per bolt
column
are used. They hol
these flexible
fingers
will moveand
towards
the stiffer
outer
tower
segment.
position
during
assembly
distribute
the
bolt
force
uniformly afterwards. Th
part
of the connection also improves the shell bending stiffness.
Instead of single washers, cover plates per bolt column are used. They hold
the bolts in position during assembly and distribute the bolt force uniformly
afterwards.
The overlapping
part ofconnection
the connection
also improves
the shell
The
resistance
of a friction
mainly
depends
on the slip behaviou
bending
stiffness.
which is directly related to the clamping force of the engaged bolts as well
conditions
clamped
plates.
Thedepends
performance
ofbehaviour
various bolt types has b
The resistanceof
of the
a friction
connection
mainly
on the slip
predict
the
behaviour
of
the
connection,
longand
short-term
tests were per
of the joint, which is directly related to the clamping force of the engaged bolts
as
well
as
the
surface
conditions
of
the
clamped
plates.
The
performance
of
these tests, a function for the loss of pretension in the bolts could be estab
various bolt types has been tested. To predict the behaviour of the connection,
correction
factor ks, which considers the hole geometry, and the correction facto
long- and short-term tests were performed. From these tests, a function for the
account
the
slipincould
becould
obtained
and improved.
loss of pretension
the bolts
be established
and the correction factor
ks, which considers the hole geometry, and the correction factor taking into
account the slip
couldcompared
be obtainedthe
and behaviour
improved. of the common flange
Additional
tests
connection to
new
type
of
friction
connection
and
show
that
the
bolts
keep
their
forces more
Additional tests compared the behaviour of the common flange connection
the
friction
connection.
The friction
connection
behaves
completely
elastically u
to the
one of the
new type of friction
connection
and show that
the bolts
keep
their
forces
more
constantly
in
the
friction
connection.
The
friction
connection
its ultimate resistance, which is higher than that of the common flange connectio
completely
abehaves
ductile
way. elastically until it reaches its ultimate resistance, which is
higher than that of the common flange connection, and fails in a ductile way.
From
thefatigue
fatigue
tests,
it was
found
the ultimate
resistance for frictio
From the
tests,
it was
found
that that
the ultimate
fatiguefatigue
resistance
for friction to
connections
according
to EN
1993-1-9 is very
conservative
according
EN 1993-1-9
is very
conservative
(Eurocode3,
1-9). In reality it w
(Eurocode3, 19). In reality it was found to be more than twice as high as the
more
than twice as high as the fatigue resistance for the common flange connect
fatigue resistance for the common flange connection.
DESIGN OF FRICTION
CONNECTIONS
DESIGN
OF FRICTION
CONNECTIONS
High strength friction grip connections, or simply friction connections, work
High
strength
friction
gripofconnections,
or help
simply
friction
by clamping
together
the plates
the joint with the
of high
strengthconnections, work
friction grip
bolts.
Theseof
bolts
usually
of grade
10.9 or
higher.
Load friction grip bolt
together
the
plates
theare
joint
with
the help
ofeven
high
strength
applied
on the plates
is lead 10.9
across or
the connection
by friction.
design ofon
a the plates is le
are
usually
of grade
even higher.
LoadForapplied
friction connection the general rules of, for example, Eurocode 3 part 18 can
connection
by friction.
be followed (Eurocode3,
18):For design of a friction connection the general rules of
Eurocode 3 part 1-8 can be followed (Eurocode3, 1-8):
(1)
Where
Fs,Rd
ks
n
m
is the static design resistance of the joint,

is the coefficient depending on the type of hole,
is the number of friction surfaces,
is the slip factor,
Where Fs,Rd
ks

n

m

M3

Fp,C
is the static design resistance of the joint,

is the coefficient depending on the type of hole,
is the number of friction surfaces,
is the slip factor,
is the partial coefficient, usually 1.25,
is the pretension force for the sum of all bolts in the joint.
The slip resistance of the connection can be regarded as ultimate limit state.
Of course, fatigue of the connection has to be checked as well. The friction
connection in steel tubular towers can be regarded as detail category 90
(Veljkovic et al. 2012; Eurocode3, 19).
The design for a friction connection can be carried out segment wise, so
that it is not necessary to take the complete cross-section of the tubular
tower into account but only a segment, which corresponds to the distance
between two rows of bolts. It is important that although the lap joint becomes
unsymmetrical by this simplification, it can still be regarded as constraint
in three dimensions by the complete tower cross-section. Obviously, the
serviceability limit state needs also to be checked, but it is not regarded as
critical for the design of a tubular tower.
ONGOING RESEARCH: HISTWIN 2

The substitution of the common flange connection by the use of friction
connections with long open slotted holes changes the design driving criteria
of the tower. The fatigue resistance of the flanges is no longer the main
constraint. Instead, the stability of the steel tube limits the resistance and
thereby the height of the structure. Also, transportation to the construction
site raises questions, as it defines the maximum diameter of the tower.
Since on-site welding is not an option, the tubes have to be transported to
the construction place as one piece of full cross section. To overcome this
problem, a modularized tower is suggested, in which the parts of the tubular
cross section are bolted to each other. Herein, friction connections will join
the tower segments even in vertical direction, see Figure 2. Ongoing research
is aimed at studying the behavior of circular and polygonal cross sections.
161
section. To overcome this problem, a modularized tower is suggested, in which the parts of
the tubular cross section are bolted to each other. Herein, friction connections will join the
Proceedingseven
of the METNET
Seminar 2013
in Lule
162 segments
tower
in vertical
direction,
see Figure 2. Ongoing research is aimed at studying
the behavior of circular and polygonal cross sections.
Figure
Circular
vs.tubular
polygonal
section
of the
steel modular tower,
Figure 2:
2. Circular
vs. polygonal
cross sectiontubular
of the steel cross
modular tower,
two segments
of modularized
tower are shown.
two segments of modularized tower are shown
As the door opening forms a hole in the tower structure, this represents one of the weakest
As the door opening
formsthe
a hole
in the
tower the
structure,
this represents
one
parts of the construction.
Therefore,
shell
around
hole has
to be stiffened
to transfer
of
the
weakest
parts
of
the
construction.
Therefore,
the
shell
around
the
hole
the applied load. There are two alternatives, which have been considered in a finite element
has to be stiffened to transfer the applied load. There are two alternatives,
analysis at LTU:
a) to increase the thickness of the shell structure surrounding the door
which have been considered in a finite element analysis at LTU: a) to increase
opening or b) totheimplement
a stiffener.
thickness of
the shell structure surrounding the door opening or b) to
implement a stiffener.
FINITE ELEMENT STUDY

An extensive finite element study was carried out in order to investigate the
influence of a door opening in the lowest tower segment (Tran et al. 2013).
A height of the lowest tower segment of 6.7 m, a diameter of 4.15 m at the
bottom and 3.919 m at the top and a wall thickness of 37 mm were considered.
A parametric study showed that for a door opening of 750 mm a stiffener with
a 37 mm thickness would be sufficient to obtain the same resistance as in a
tower segment without opening. Raising the steel grade from the common
S355 to S500 or even S650 reduces the thickness of the stiffener by 22% and
38%, respectively. Figure 3 shows that the different steel grades reach the same
moment resistance. But the reduced thickness has to be taken into account.
The study also showed that the position of the door opening within the crosssection does not affect the ultimate strength of the tower significantly.
S500 or even S650 reduces the thickness of the stiffener by 22 % and 38 %, respectively.
Figure 3 shows that the different steel grades reach the same moment resistance. But the
reduced thickness has to be taken into account. The study
alsoofshowed
that the
of the
Proceedings
the METNET Seminar
2013 inposition
Lule
163
door opening within the cross-section does not affect the ultimate strength of the tower
FINITE ELEMENT STUDY
significantly.
An extensive finite element study was carried out in order to investigate the influence of a
door opening in the lowest tower segment (Tran et al. 2013). A height of the lowest tower
segment of 6.7 m, a diameter of 4.15 m at the bottom and 3.919 m at the top and a wall
thickness of 37 mm were considered. A parametric study showed that for a door opening of
750 mm a stiffener with a 37 mm thickness would be sufficient to obtain the same resistance
as in a tower segment without opening. Raising the steel grade from the common S355 to
S500 or even S650 reduces the thickness of the stiffener by 22 % and 38 %, respectively.
Figure 3 shows that the different steel grades reach the same moment resistance. But the
reduced thickness has to be taken into account. The study also showed that the position of the
door opening within the cross-section does not affect the ultimate strength of the tower
significantly.
Figure 3: Comparison of various steel grades to stiffen the door opening.

Figure 3. Comparison of various steel grades to stiffen the door opening.
EXPERIMENTS
Figure 3:toComparison
of various
to stiffen
door opening.
EXPERIMENTS
Experiments are
planned
be performed
at steel
LTUgrades
in order
tothe
investigate
the effects of using
high strength Experiments
steel, cold are
forming
bend
in
manufacture
and
opening.
Coupon
tests will be
planned to be performed at LTU in order to investigate
carried out EXPERIMENTS
andtheconsidered
as input
data of steel,
numerical
models.
Coupon
specimens will be cut
effects of using
high strength
cold forming
bend
in manufacture
off from original
plate
specimens.
The
positions
of
coupon
specimens
will
reflect effects of
and opening. Coupon tests will be carried out and considered as input
data
Experiments are planned to be performed at LTU in order to investigate the effects of using
of
numerical
models.
Coupon
specimens
will
be
cut
off
from
original
plate
cold formedhigh
corner
on steel,
the strength
of material.
Figure 4and
shows
positions
coupon
strength
cold forming
bend in manufacture
opening.
Coupon of
tests
will be specimens
specimens. The positions of coupon specimens will reflect effects of cold
carried
out
and
considered
as
input
data
of
numerical
models.
Coupon
specimens
will
be cut
in circular and polygonal plate specimens.
formed
corner
on the
strengthThe
of material.
4 shows
positions
coupon
off from
original
plate
specimens.
positions Figure
of coupon
specimens
will of
reflect
effects of
specimens
in
circular
and
polygonal
plate
specimens.
cold formed corner on the strength of material. Figure 4 shows positions of coupon specimens
in circular and polygonal plate specimens.
4: Positions
specimens
Figure 4. Positions of Figure
coupon specimens.
Figure
4: Positionsof
of coupon
coupon specimens
164
gure 5 shows a coupon

performed
LTU. atALTU.
100AkN
machine is used
Figure 5 test
showsbeing
a coupon
test being at
performed
100 Dartec
kN Dartec
machine
is
used
for
these
tests.
Displacement
of
the
coupon
specimens
is
ese tests. Displacement of the coupon specimens is measured by an Epsilon Extensome
measured by an Epsilon Extensometer with a 25 mm of gage length and travel
ith a 25 mm of ingage
length
and travel
in mm
tension
compression of 12.5 mm a
tension
and in compression
of 12.5
and 6.25and
mm, in
respectively.
.25 mm, respectively.
Figure 5 shows a coupon test being performed at LTU. A 100 kN Dartec machine is used for
these tests. Displacement of the coupon specimens is measured by an Epsilon Extensometer
with a 25 mm of gage length and travel in tension and in compression of 12.5 mm and
-6.25 mm, respectively.
Figure 5. Set up of coupon tests.
Figure 5: Set up of coupon tests
36 L-specimens with different

angles
and
different
thicknesses will be tested
Figure 5:
Set
updifferent
of
coupon
tests
6 L-specimens with
different angles
and
thicknesses
will be tested in order
in order to investigate the effects of different angles in corners of the polygonal
vestigate the36effects
ofFigure
different
angles
in corners
of the
tower.
6 shows
tower.
shows
L-specimens
with angles:
90,polygonal
100,
120,
140,
160 Figure
L-specimens
with 6different
angles and
different
thicknesses
will be
tested
in order
to
and the
170.
4 mm
and
6 mm
plate
are
considered.
The
length
of L L-and 6 mm pl
investigate
effects
of different
angles
in thicknesses
corners
polygonal
Figure
ecimens with
angles:
90,
100,
120,
140,of the
160
andtower.
170.
46 shows
mm
specimens
is 600 90,
mm.100, 120, 140, 160 and 170. 4 mm and 6 mm plate
specimens
with angles:
icknesses arethicknesses
considered.
The length
ofofL-specimens
600 mm.
are considered.
The length
L-specimens is 600is
mm.
Figure 6: L- specimens with different angles

Figure there
6. L-specimens
different angles.
In addition,
are 32with
specimens
with opening and without opening made of different steel
grades, see Table 1.
Two types of cross sections are polygonal and circular. Figure 7 shows plate specimens with
opening and without opening.
In addition, there are 32 specimens with opening and without opening made
of different steel grades, see Table 1.
Two types of cross sections are polygonal and circular. Figure 7 shows plate
specimens with opening and without opening.
Figure 8 shows a compression test with a circular downscale specimen
without opening. The experiments are performed using an Instron system
with a maximum load of 4500 kN.
Table 1: List of plate specimens

Table 1. List of plate specimens.
Name of
Name of
specimens
specimens
Test-1
Test-1
Test-2Test-2
Test-3Test-3
Test-4Test-4
Test-5Test-5
Test-6
Test-6
Test-7
Test-7Test-8
Test-8
Table 1: List of plate specimens
Material
Material
650MC
650MC
650MC
650MC
650MC
650MC
650MC
650MC
650MC
650MC
650MC
650MC
650MC
650MC
650MC
650MC
Length Thickness
Number of
Length Thickness
Number
of
specimens
(mm)
(mm)
specimens
(mm)
(mm)
1000
6
4
1000
6
4
1000
1000
66
4 4
1000
1000
66
4 4
1000
66
4 4
1000
1000
4
4 4
1000
4
1000
4
4
1000
4
4
1000
4
4
1000
1000
44
4 4
1000
4
4
Figure 7: Plate specimens with opening and without opening
Figure 7: PlateFigure
specimens
with
opening
andopening.
without opening
7. Plate specimens
with opening
and without
165
166
Figure 7: Plate specimens with opening and without opening
Figure 8. Compression
test of plate specimen.
Figure
8: Compression
test of plate specimen
The failure modes obtained in the experiment and by FEA are rather similar
The failureasmodes
obtained
in the
and by FEA
arematerial
rather used
similar
it is shown
in Figure
9. experiment
It should be emphasized
that
in as it is shown in
numerical
simulation
is
S650.
The
properties
of
this
material
are
defined
in
Figure 9. It should be emphasized that material used in numerical simulation is S650. The
et al.
(2013). are defined in Tran et al. (2013).
properties Tran
of this
material
Figure 9. Deformation of opening circular plate specimen with 6 mm thickness.
Figure 9: Deformation of opening circular plate specimen with 6 mm thickness

Preliminarily results for the ultimate loads are presented in Table 2. Differences between
Figure 9: Deformation of opening circular plate specimen with 6 mm thickness

Preliminarily results for the ultimate loads are presented in Table 2.
Differences
between
results
range
from
to 7.6%.in Table 2. Differences between
Preliminarily
results
for the
ultimate
loads
are1.1%
presented
results range from 1.1% to 7.6%.
Table
2:ofPreliminary
Table 2. Preliminary initial
results
ultimate load. initial results of ultimate load
Test-Cir-1
Test-Cir-D-1
Test-Pol-1
Test-Pol-D-1
According to EC3
3292
2672
3177
2567
Ultimate load (kN)

According to experiment
3255
2484
3258
2444
Difference %
1.1
7.6
2.5
5.0
CONCLUSIONS
CONCLUSIONS
From the findings described above, the following can be concluded when using friction
connections in tubular steel wind towers:
From
the endurance
findings described
above,
thefriction
following
can be increases.
concluded when
1.
Fatigue
of the towers
with the
connections
using
friction
connections
in
tubular
steel
wind
towers:
2. Shell stability may often be the main design criterion.
3. The use of higher steel grades than the common S355 may be justified from the
1. structural
Fatigue
endurance
point
of view. of the towers with the friction connections increases.
4. The use of higher steel grades for the segment with the door opening leads to less
2. material.
Shell stability may often be the main design criterion.
3.
The use of higher steel grades than the common S355 may be justified
from the structural point of view.
4.
The use of higher steel grades for the segment with the door opening
leads to less material.
ACKNOWLEDGEMENTS
The authors gratefully acknowledge the financial support by the European
Research Fund of Coal and Steel, Grant-Agreement No. RFSR-CT-2010-00031,
and the Centrum for High-performance Steel (CHS) at Lule University of
Technology, Sweden.
REFERENCES
(Burton et al. 2001) Burton T, Sharpe D, Jenkins N and Bossanyi E: Wind
energy handbook, Chichester: Wiley, 2001.
(Eurocode3, 1-6) EN-1993-1-6 Design of Steel Structures, Part 16: Strength
and stability of shell structures, Brussels, Belgium, European Committee for
Standardization, 2007.
167
168
(Eurocode3, 18) EN-1993-1-8 Design of Steel Structures, Part 18: Design of

joints, Brussels, Belgium, European Committee for Standardization, 2005.
(Eurocode3, 19) EN-1993-1-9 Design of Steel Structures, Part 19: Fatigue
strength, Brussels, Belgium, European Committee for Standardization, 2004.
(EWEA 2012a) EWEA: United in tough times The European Wind Energy
Association, Annual report 2012.
(EWEA 2012b) EWEA: Green Growth The impact of wind energy on jobs
and the economy, April 2012.
(EWEA 2013a) EWEA: The European Wind Initiative Wind Power Research
and Development to 2020, 2013.
(EWEA 2013b) EWEA: Wind in power 2012 European statistics, February
2013.
(Fichaux 2011) Fichaux N, Beurskens J, Hjuler P, Wilkes J, Frandsen,
S, Dalsgaard Srensen J, Eecen P, Malamatenios C, Arteaga Gomez J,
Hemmelmann J, van Kuik G, Bulder B, Rasmussen F, Janssen B, Fischer T,
Vossanyi E, Courney M, Giebhardt J, Barthelmie R, Holmstrm O: UpWind
Design limits and solutions for very large wind turbines, EWEA, March 2011.
(Germ.Lloyd 2004) Germanischer Lloyd Wind Energy: Guideline for the
certification of wind turbines, Edition 2003 with supplement 2004.
(RFCS 2013) Success Stories of the Research Fund for Coal & Steel, http://
cordis.europa.eu/coal-steel-rtd/stories_en.html, last updated 2013-01-22.
(Tran et al. 2013) Tran A T, Veljkovic M, Rebelo C, da Silva L: Resistance of door
openings in towers for wind turbines, 3rd South-East European Conference on
Computational Mechanics, Kos Island, Greece, 1214 June 2013.
(Veljkovic et al. 2012) Veljkovic M, Heistermann C, Husson W, Limam M,
Feldmann M, Naumes J, Pak D, Faber T, Klose M, Fruhner K.-U, Krutschinna
L, Baniotopoulos C, Lavasas I, Pontes A, Ribeiro E, Hadden M, Sousa R, da
Silva L, Rebelo C, Simoes R, Henriques J, Matos R, Nuutinen J, Kinnunen
H: High-strength tower in steel for wind turbines (HISTWIN), final report,
RFCS-CT-2006-00031, 2012.
Virdi & Tenhunen (Editors)
METNET seminars deal with technical aspects of metal construction

as well as issues of concern to industry on management, planning
and sustainability of projects.
strengthens knowledge structures, especially in the European regions
represented by its members. At the annual seminars, the participating
partner organization represents its regional innovation environment
but simultaneously is a partner of a larger and stronger international
innovation world created by active cooperation between the institutes
and enterprises.
This book covers the papers presented in the annual Metnet Seminar
in October 2013 held at Lule University of Technology. The seminar
continued the METNET tradition of presenting scientific and
development papers of high calibre.
Metnet Annual Seminar in Lule, Sweden, on 2223 October 2013
electronic
ISBN
978-951-784-642-4 (PDF)
ISSN 1795-424X


HAMK
printed
ISBN 978-951-784-641-7
ISSN 1795-4231


Hamk Metnet Lulea 2013 Ekirja

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Hamk Metnet Lulea 2013 Ekirja

Uploaded by

Copyright:

Available Formats

Virdi & Tenhunen (Editors)

METNET seminars deal with technical aspects of metal construction

Proceedings of the METNET Seminar 2013 in Lule

Proceedings of the METNET Seminar 2013 in Lule

Metnet Annual Seminar in Lule, Sweden, on 2223 October 2013

Kuldeep Virdi and Lauri Tenhunen (Editors)

Proceedings of the METNET Seminar 2013 in Lule

HAMK University of Applied Sciences

HAMK UAS and writers

Hmeenlinna, January 2014

DESIGN OF LINER TRAY WITH DISTANCE SCREW CONNECTION

THIN-WALLED CROSS-SECTIONS AND THEIR JOINTS:

FORMING LIMIT CURVE OF HIGH STRENGTH STEEL WITH THE HELP OF

MONITORING OF STEEL RAILWAY BRIDGE AND CONTINUOUS WELDED RAIL.................34

LASER SYNTHESIS TECHNOLOGY FOR THE CREATION,

WEIGHT AND COST OPTIMIZATION OF WELDED HIGH STRENGTH STEEL BEAMS...........44

Teemu Tiainen, Kristo Mela, Timo Jokinen and Markku Heinisuo

HIGH STRENGTH STEEL IN TUBULAR TRUSSES..................................................................56

TUBULAR COMPOSITE COLUMNS IN A NON-SYMMETRICAL FIRE .....................................60

INFLUENCE OF NATURAL FIRE SCENARIO BY CFD AND FEM ANALYSIS

MEMBRANE STRUCTURES SUPPORTED BY A FRAME

NUMERICAL STUDIES OF Z-PURLINS SUPPORTED BY SANDWICH PANELS...................... 78

PROBLEMS OF FUNCTIONING AND ELABORATION OF THE MASS HOUSING

THE BLUE BOX AND THE TECHNOLOGICAL TRAJECTORIES IN CONTEXT

SUSTAINABILITY AS A BUSINESS OPPORTUNITY TODAY AND TOMORROW:

Proceedings of the METNET Seminar 2013 in Lule

Anatoly Perelmuter and Vitalina Yurchenko

OPTIMIZATION OF STEEL TOWERS FOR LARGE WIND TURBINES................................... 120

STRUCTURAL BRACING FORCES AND STIFFNESS......................................................... 125

INNOVATION IN MECHANICAL FASTENING TECHNOLOGY

FLANGELESS CONNECTIONS IN STEEL TUBULAR WIND TOWERS................................... 157

Proceedings of the METNET Seminar 2013 in Lule

Proceedings of the METNET Seminar 2013 in Lule

Proceedings of the METNET Seminar 2013 in Lule

Proceedings of the METNET Seminar 2013 in Lule

CONNECTIONS AND MATERIALS

Proceedings of the METNET Seminar 2013 in Lule

DESIGN OF LINER TRAY

Czech Technical University in Prague

The paper describes an experimental and numerical study on structural

longitudinal spacing of fasteners supplying lateral restraint to the narrow flanges. In

Figure 1. Liner tray wall system using distance screw.

Figure 1. Liner tray wall system using distance screw.

Proceedings of the METNET Seminar 2013 in Lule

Figure 2. Failure mode of the test and FE model.

Figure 2. Failure mode of the test and FE model.

on the parametric study, a general design formula was developed.

Proceedings of the METNET Seminar 2013 in Lule

THIN-WALLED CROSS-SECTIONS AND THEIR JOINTS:

Specimen S1 (stud) C 200-45-2,0;

Specimen S2 (rack) 200-50-2,0.

Proceedings of the METNET Seminar 2013 in Lule

Steel used for specimen production has the following parameters:

Coating mass, 350 g/m2;

Coating thickness, 25 microns.

Proceedings of the METNET Seminar 2013 in Lule

Proceedings of the METNET Seminar 2013 in Lule

Proceedings of the METNET Seminar 2013 in Lule

FORMING LIMIT CURVE OF HIGH STRENGTH STEEL

Proceedings of the METNET Seminar 2013 in Lule

Complex forming problems, as typically found in automotive industry as

Proceedings of the METNET Seminar 2013 in Lule

Figure 1. The FLC-specimen geometries with dim

Proceedings of the METNET Seminar 2013 in Lule