CIVIL AND ENVIRONMENTAL ENGINEERING REPORTS

ISSN 2080-5187 CEER 2017; 25 (2): 107-123

DOI: 10.1515/ceer-2017-0023
Original Research Article

TRUSS JOINT WITH POSITIVE ECCENTRICITY

EXPERIMENTAL RESEARCH

Małgorzata GORDZIEJ-ZAGÓROWSKA, Elżbieta URBAŃSKA-

GALEWSKA1, Patryk DENIZIAK, Łukasz PYRZOWSKI
Gdansk University of Technology, Gdańsk, Poland

Abstract
Due to the technological reasons in modern lightweight steel trusses, fabricated from cold-
formed sections, positive eccentricities appear in the truss nodes what induce additional
forces in the truss chords. To account for the real load-carrying capacity of truss node area
the steel structure research in scale 1:1 were conducted. The experiments consisted of two
parts: preliminary and proper one, when conclusions from the first part were applied.
Carrying out preliminary studies helped to identify of the research station drawbacks and
eliminate most of them, what ensure the appropriate research results. The initial numerical
analysis were also conducted what was presented in the paper.
Keywords: steel structure, thin-walled structures, cold-formed sections, eccentricities,
truss, experimental research

1. INTRODUCTION

1.1. General
Interest in lightweight steel structures made from cold-formed sections, due to
their many advantages is growing year on year. A lot of experimental studies
which constitute the basis for numerical analysis are being conducted and new
standards are being elaborated all over the world [1, 5, 8, 10, 11]. Designers put

1Corresponding author: Gdansk University of Technology, Department of Metal Structures and

Construction Management, G. Narutowicza st 11/12, 80-233 Gdańsk, Poland, e-mail:
[email protected], tel.:+483471751

Unauthenticated
Download Date | 7/4/17 11:29 PM
108 Małgorzata GORDZIEJ-ZAGÓROWSKA, Elżbieta URBAŃSKA-GALEWSKA,
Patryk DENIZIAK, Łukasz PYRZOWSKI

more and more emphasis on the rational development of both the elements cross-
sections and designed with these elements thin-walled structures. The aim of this
paper is to present the methodology of proceeding with the preparation of
experimental studies mapping the behaviour of truss elements made of cold-
formed sections and research results presentation.

1.2. Research problem

Research issue concerns design solution used in structural systems of single storey
steel buildings made of cold-formed open cross-sections as a whole. In the
construction of a typical, mono-pitch, trapezium lattice girder, assuming a fixed
spacing between the nods, the angle of inclination of brace members change due
to technological reasons. The angle increase causes the appearance of the positive
eccentricity in the truss joint and therefore occurrence of the bending moment in
the presence of shear. In accordance with the current state of knowledge [7], when
dimensioning the lattice girder's chord, one should take into consideration the
eccentricity occurrence that leads to the necessity for local increase in the chord's
section in the joint area. Practically, this is done by strengthening the chords
through the use of channel-section covers (Fig. 1) or by increasing of the cross-
section wall thickness over the entire length of the element.
Both solutions are uneconomical what caused that research on real load-carrying
capacity of the compressed and bent chord of the lattice girder made of cold-
formed open cross-sections was undertaken.
The study aimed to clarify whether the bolted connection of the brace members
to the walls of the hat cross-section chord locally stiffen the joint area providing
increase of its capacity relative to this determined by traditional methods.

Fig. 1. Node with strengthening cover plates: a) view and cross-section, b) Photo. M.
Gordziej - Zagórowska [2]
Preliminary numerical model of the structure designed for research purposes was
created before experimental studies. The aim of the preliminary numerical

Unauthenticated
Download Date | 7/4/17 11:29 PM
 TRUSS JOINT WITH POSITIVE ECCENTRICITY EXPERIMENTAL RESEARCH 109

analysis was to verify both the geometry and the static research model, as well as
to confirm the research assumptions. In the future, the research results will
constitute the basis for the preliminary numerical model validation.

2. PRELIMINARY NUMERICAL ANALYSIS

2.1. Research model

The stability test of the compressed member, additionally loaded with a local
bending moment was planned in a 1:1 scale. To realise this goal the research
model of lattice girder section with dimensions 2000 x 1315 mm was designed
from thin-walled cold-formed open cross-sections (Fig. 2). Cold-formed elements
made of galvanized steel S350GD with zinc surface Z275MA were applied. The
chords were designed from the hat-section H 39/117/106/117/39x2, and the brace
members from channel-sections C 17/75/100/75/17x2. All the connections within
the truss joints were designed as bolted. Fig. 2a exemplifies location of the
connections which are made with six M12 kl. 8.8 bolts (three bolts on each hat-
section web).
The proper construction of the nodes, due to the lack of space for bolts, demanded
shifting the brace members ends outside, what caused the positive bending
moment appearance in the node. The research model was loaded with two
concentrated forces (Fig. 2b). Force H induced pressure on the lower chord, while
force P caused local bending of the lower chord in the joint area.

Fig. 2. Research model: a) view of the construction, b) static scheme [3]

Unauthenticated
Download Date | 7/4/17 11:29 PM
110 Małgorzata GORDZIEJ-ZAGÓROWSKA, Elżbieta URBAŃSKA-GALEWSKA,
Patryk DENIZIAK, Łukasz PYRZOWSKI

2.2. Results of the preliminary numerical analysis

On the basis of the research model presented above appropriate numerical model
with the use of Marc Mentat - MSC Software was built. For computational
modeling 4-node 'thin shell-139' elements were used. This is the thin shell element
with global displacements and rotations as degrees of freedom. Bilinear
interpolation is used for the coordinates, displacements and the rotations. The
membrane strains are obtained from the displacement field; the curvatures from
the rotation field [4].
The way of preliminary numerical model building, assumed models of material
features and assumptions concerning way of members connections in truss nodes
were presented in [3] in detail. The purpose of a preliminary numerical analysis
was to determine the member’s cross-sections shape (corresponding to the real
structure), the proper static scheme and sequence of structure loading. The model
parameters were selected so as to reconstruct the stress state in a real truss with
the eccentric joint. Due to the time-consuming nature of the process of numerical
modelling, welded joints were assumed in the preliminary analysis. Welds were
made with the use of rigid surface elements (3- and 4-node), to assure that they
will play role of non-deformable connections under loading. The complicated
issue of the joints flexibility in this type of connections will be developed during
further numerical analysis with the use of relations established by Zaharia &
Dubina [12], Słowiński at al. [9].
Preliminary numerical analyses were conducted for two types of truss models:
sensitive and insensitive to local instability (in accordance with the classification
of cross-section). Insensitive models were done from 5mm thick shell elements
(brace members) and 6mm thick (chords) thereby creating class 1 cross-section
(according to EN 1993-1-1, 2006 [6]). Models sensitive to local instability were
performed from 2 and 3mm thick shell elements in two variants (Table 1).
Table 1. Types of analyses
Types of models
Type of elements
insensitive sensitive
to local instability to local instability
brace members
member 5 mm 2 mm 2 mm
C 17/75/100/75/17
wall
thickness chords
6 mm 3 mm 2 mm
H 39/117/106/117/39
Analysis code: „5/6” „2/3” „2/2”

Unauthenticated
Download Date | 7/4/17 11:29 PM
 TRUSS JOINT WITH POSITIVE ECCENTRICITY EXPERIMENTAL RESEARCH 111

Next, all types of models were submitted for three types of analysis: Mechanically
Nonlinear Analysis (MNA), Geometrically Nonlinear Analysis (GNA) and
Geometrically and Mechanically Nonlinear Analysis (GMNA). Conducted
analysis confirmed, that 3 and 2 mm thick walls are sensitive to local instabilities.
Outcomes of these analyses were presented in detail in the paper [2], [3].
Furthermore, the results confirmed that research hypothesis conceived by authors
is valid as not only none of the obtained structure failure models occurred in the
compressed and bending area of the hat section but also the increase in material
effort was not observed. The model with 2 mm thick walls (both chords and brace
members) sensitive to local instability was chosen for further experimental
studies. Numerical analyses were also a tool to determine the approximate
maximum value of the compression force H and bending force P, necessary at the
experimental stage.

3. PRELIMINARY TEST

3.1. Research station

In order to conduct the experimental research the special stand, presented in Fig. 3
was designed. The following conditions were taken into consideration when the
shape of the laboratory stand was elaborated:
 due to the limitations of the technical possibilities within the laboratory the
research model had been turn by 90° during experimental phase (dead load of
the construction was assumed as an negligible);
 the supporting joints and the places where the concentrated forces were
applied, were properly strengthen;
 to ensure a free movement of the “C” joint under the bending force P (Fig. 4),
the introducing of the compression force H couldn’t been done through an
actuator, as such a solution would result in the change of the static scheme of
the research model that was anchored as an cantilever. In order to achieve the
free displacement of the truss in the “C” joint, a special compression head was
designed (Fig. 5).
With respect to above conditions the compression force H and bending force P
were applied respectively by means of two turnbuckles (Fig. 5) and an expansion
bolt (Fig. 3).
The value of the bending force P was registered by a dynamometer
N.B.C. Elettronica, with a nominal range of ±100 kN while the value of the
compression force H was measured with a dynamometer CZAH CT-25 with
nominal range of 250 kN. The last one was located between the compression head
and the compressed section of the tested model (Fig. 6).

Unauthenticated
Download Date | 7/4/17 11:29 PM
112 Małgorzata GORDZIEJ-ZAGÓROWSKA, Elżbieta URBAŃSKA-GALEWSKA,
Patryk DENIZIAK, Łukasz PYRZOWSKI

Fig. 3. Design of the first version of the research station [3]

Fig. 4. Research model scheme of work when using: a) stationary hydraulic cylinder,
b) compression head

Unauthenticated
Download Date | 7/4/17 11:29 PM
 TRUSS JOINT WITH POSITIVE ECCENTRICITY EXPERIMENTAL RESEARCH 113

Fig. 5. Design of the compression head - isometric view

Three inductive sensors with measuring range of ±100 mm located on the
structure in accordance with the scheme presented in Fig. 6 were used during the
experimental test. The inductive sensors X and Y measured the movements of the
research model in the plane of bending while the sensor Z measured the
movement from the bending axis. For measurement of the strains, strain gauges
LY11-6/120A were applied in four measuring cross-sections. Three cross-
sections were located on the compressed hat-section chord respectively: 1-1
above the objective joint, 2-2 in the axis of the joint and 3-3 below the joint. The
fourth cross-section was on the brace member as it is shown in Fig. 6.

Fig. 6. Placement of strain gauges and inductive sensors on the tested model [3]

Unauthenticated
Download Date | 7/4/17 11:29 PM
114 Małgorzata GORDZIEJ-ZAGÓROWSKA, Elżbieta URBAŃSKA-GALEWSKA,
Patryk DENIZIAK, Łukasz PYRZOWSKI

3.2. Results of the preparatory experimental tests

Prior to proper testing of the research model a static uniaxial tensile test for steel
material was carried out on a ZWICK / ROELL Z400 machine with 50 kN head
to achieve proof stress, Young's modulus and real stress-strain relationship. Next,
the loading of the research model started. At first the compression force H was
applied gradually increasing its value until reaching 115 kN. The value of the
compression force H was not changed since that moment. Next, the introduction
of the bending force P began. As can be seen in the loading diagram in time
presented in Fig. 7 the increase in the value of the bending force P caused
simultaneous decrease in the value of the compression force H.
It should be noted that maintaining the compression force H at the constant level
while introduce the bending load (force P) was not possible due to the
instrumentation used. A detailed description of the test was the theme of the
paper [3].

Fig. 7. The loading diagram in time (concentrated forces H and P) [3]

Strains measured in cross-sections 1-1, 2-2 and 3-3 were the basis of the normal
stresses quantification presented in Fig. 8. The results from the strain gauge „A”
showed that introduction of the compression force H caused torsion of the
compressed chord of the research model. This unintentional torsion was pointed
out throughout all three measurement sections. Introduction of the bending
force P caused further differentiation of strains resulting in its reduction or
increase in particular cross-sections.
Moreover, displacement control conducted during tensioning showed that the
researched model had moved from the axis of the compression and bending [3]
already at the moment of application of the compression force H. This was a
consequence of the way of the compression force H introduction what resulted in
the torsion of the tested element. Furthermore, all the measured displacements
were much higher than those calculated by preliminary numerical analysis.

Unauthenticated
Download Date | 7/4/17 11:29 PM
 TRUSS JOINT WITH POSITIVE ECCENTRICITY EXPERIMENTAL RESEARCH 115

Fig. 8. Graph of normal stresses in time obtained on the basis of measured strains in
cross-sections: a) 1-1, b) 2-2, c) 3-3. [3]

Unauthenticated
Download Date | 7/4/17 11:29 PM
116 Małgorzata GORDZIEJ-ZAGÓROWSKA, Elżbieta URBAŃSKA-GALEWSKA,
Patryk DENIZIAK, Łukasz PYRZOWSKI

3.3. Conclusions
The conducted research of the first model and obtained results allowed for
verification of geometry and static scheme of the truss research model as well as
the design of the research stand. It was recognized, that structure of the research
model was correct and did not require any changes. However, it was essential to
introduce changes to the construction of the first version of the research stand
prior to testing the remaining 5-five research models. Firstly, the structure of the
compression head should be changed in the way to protect the compressed truss
chord from random torsion. Next, the significant limitation of the model
displacements (from the bending plane) should be ensured.

4. EXPERIMENTAL RESEARCH

4.1. Research station modification

The design of the improved research station with the changes marked in the red
color is presented in Fig. 9 and 10.
The change of the way of introducing the force H to eliminate random torsion of
the compressed truss chord during experiments was the primarily authors concern.
In the first version of the laboratory stand the compression force H was introduced
by the help of the compression head structure presented in Fig. 5.

Fig. 9. Design of the final version of the compression head

Unauthenticated
Download Date | 7/4/17 11:29 PM
TRUSS JOINT WITH POSITIVE ECCENTRICITY EXPERIMENTAL RESEARCH 117

Fig. 10. Design of the modificated version of the research station

Unauthenticated
Download Date | 7/4/17 11:29 PM
118 Małgorzata GORDZIEJ-ZAGÓROWSKA, Elżbieta URBAŃSKA-GALEWSKA,
Patryk DENIZIAK, Łukasz PYRZOWSKI

Stretching elements were tensioned by means of two hand-tightened turnbuckles.

At the new version of the stand the force H was applied by means of low profile
hydraulic cylinder (BVA HF2005) with a capacity of 20 tons, which was located
between flat plate of the compression head and dynamometer (Fig. 11), exactly in
the hat section center of gravity axis. Thereby the possibility of unequal uploading
of the model was eliminated, as it was when two turnbuckles were used. The
proposed modification ensured the free movement of the joint “C” during
uploading as in the first version.
Anyway, the turnbuckles were left in the compression head structure to use them
for the leveling of the compression head and for adjusting and alignment the initial
tension of the stretching elements. Additionally, it was decided to exchange the
flat bars which were parts of stretching elements into rectangular hollow section
80x40x6,3 to ensure their higher torsional stiffness (Fig. 9).
The consecutive modification of the laboratory stand aims at limitating of the
stand excessive rotation and displacements. The strengthening of the stand
anchoring beams fixed to special anchoring rails in the laboratory floor plate was
necessary. Additional stiffeners of the pulled anchoring beam under joint “B”
were provided and anchoring bolts spacing was reduced (Fig. 10).

Fig. 11. Placement of measurement equipment on the research model (introduced

changes are marked in red colour)

Unauthenticated
Download Date | 7/4/17 11:29 PM
 TRUSS JOINT WITH POSITIVE ECCENTRICITY EXPERIMENTAL RESEARCH 119

The way of model loading by bending force P and recording its value were not
changed as the number and location of strain gauges.
Nevertheless authors decided to record measurements of possible horizontal and
vertical displacements of support joints “A” and “B” to determine real stiffness
of the base joint. To that end, additional three inductive sensors (W1, T1, T2) with
measuring range of ±100 mm and one dial indicator (W2) also with measuring
range of ±100 mm were applied. Location of the whole measuring equipment on
the research model after changes is presented in the Fig. 11.
The introduction of all the above changes during the modification of the
laboratory stand made it possible to eliminate most of the drawbacks of the
preliminary test.

4.2. Results of experimental research

The laboratory tests of 5 models in scale 1:1 were conducted on the modificated
research station. Research methodology was the same as at the first time. Strains
values of the four indicated cross-sections in the area of the analysed node
(Fig. 11), under the applied load were recorded continuously (in real time).
Graphs of normal stresses in time obtained on the basis of strains measured in the
all cross-sections and registered during the model No. 4 research are presented in
the Fig. 12.
Values of the stresses were stated based upon measured strains and with the use
the mean value Young's modulus determined on the basis of the uniaxial tensile
test for steel material of channel-sections and hat-sections equal respectively
199 GPa and 210 GPa.
As can be seen in Fig. 12 during a first loading step, that is during the introduction
of compression force H the coincidence of readings from strain gauges A, B, C
and D in the all measurement sections occurred. This is the evidence of the
uniform distribution of the stresses in the compressed sections. It should also be
noted that no shape deformations were observed during the first phase of research
models loading. At the beginning of the second phase of loading (before a loss of
stability), symmetrical increases (strain gauges A and C) or decreases (strain
gauges B and D) of cross-section deformations, with respect to the local axis y-y,
were observed. Therefore, the obtained results substantiated the purposefulness
of carrying out modification of the research station first version. However, within
the scope of the loss of the local stability it was pointed out that shape
deformations had lost their symmetrical forms (as can be seen in Fig. 12).

Unauthenticated
Download Date | 7/4/17 11:29 PM
120 Małgorzata GORDZIEJ-ZAGÓROWSKA, Elżbieta URBAŃSKA-GALEWSKA,
Patryk DENIZIAK, Łukasz PYRZOWSKI

Fig. 12. Graph of normal stresses in time obtained on the basis of strains measured in
cross-sections respectively: a) 1-1, b) 2-2, c) 3-3, d) 4-4, (from the model of No. 4)
The outstand elements of the cross sections (free flanges) had different directions
of the strains (measured by strain gauges A and C) what constitutes the evidence
of the typical halve-wave shape local buckling occurrence. Lack of deformations
symmetry with respect to a local axis z-z is caused by various inherent
imperfections of steel structures Buckling forms of the hat section and the
channel-section observed during the research are presented respectively in
Fig. 13a) and Fig. 13b). These deformations occurred at the final stage of the load
(P=40 kN, Hodp=55 kN).

Unauthenticated
Download Date | 7/4/17 11:29 PM
 TRUSS JOINT WITH POSITIVE ECCENTRICITY EXPERIMENTAL RESEARCH 121

Fig. 13. Buckling forms of: a) the hat section flange, b) distorsion of the channel-section
(for the final stage of the load)
Values of the bending force P, and the corresponding compression force H the
presence of which initiated the beginning of the stability losing in the area of the
analysed joint are shown in Table 2. Beginning of instability always started at the
cross-section 2-2 and the deformations were enlarging until failure while the
loading was growing.
Moreover it had been observed that described above research stand modifications
caused significant decrease of its displacements from the truss bending plane,
what was recorded by sensor Z. Measured displacement values were in the range
from -2,28 mm to 4,65 mm, what was recognized as negligible.
However, values of horizontal and vertical displacements of the support joints
“A” and “B” proved that modified research stand structure did not ensure full
rigidity of the base joints just as at the first version of the structure. Due to gained
data, the real stiffens of the base joints will be applied in the foreseen advanced
numerical analysis.
Table 2. Values of the forces P i Hcor initiated the beginning of the stability losing
Research model number Force P [kN] Force Hcor [kN]
Model 2 8,5 87
Model 3 7,7 89
Model 4 5,0 95
Model 5 6,7 87
Model 6 5,0 96

Unauthenticated
Download Date | 7/4/17 11:29 PM
122 Małgorzata GORDZIEJ-ZAGÓROWSKA, Elżbieta URBAŃSKA-GALEWSKA,
Patryk DENIZIAK, Łukasz PYRZOWSKI

5. CONCLUSIONS
Presented in the paper process of research station calibration exemplifies the
importance of both the preliminary numerical analysis and the preliminary
experiment. The former is helpful in designing the research station and research
model. The latter helps to verify correctness of the project assumptions. While
recognising the areas of errors researchers have a chance to remove them. The
presented approach is time consuming but eventually provides the higher quality
of achieved research results. In that case, the results gained after introduction of
changes to the research stand and carrying out of the final tests on 5 models will
constitute solid ground for validation of the numerical model of the analysed joint
and will enable full verification of the research hypothesis, which will be
presented in the next publication.

REFERENCES
1. Garstecki, A., Kąkol, W., Rzeszut, K.: Global and local stability analysis of
thin-walled beams, in: Metal Structures: Design, Fabrication, Economy, edit.
K. Jarmai, J.Farkas, Miscolc, Hungary, Millpress (2003), 61-66.
2. Gordziej-Zagórowska, M., Urbańska-Galewska, E., Jankowski R.: Modelling
of truss with cold-formed section and positive eccentricity in the node,
Structure and Environment, 4 (2014): 24-34.
3. Gordziej-Zagórowska, M., Urbańska-Galewska, E., Pyrzowski Ł., Deniziak
P., A. Łukowicz: Preliminary experimental research on stability of truss' joint
with positive eccentricity, in: Recent Progress in Steel and Composite
Structures: Proceedings of the XIII International Conference on Metal
Structures (ICMS 2016), edit. M. Giżejowski et.al., Zielona Góra, Taylor &
Francis Group, Balkema Book (2016), 425-432.
4. Marc 2013.1, Volume B: Element Library, MSC Software, program
instruction, (2013).
5. Mądry, D.: Niestateczność miejscowa trapezowego przekroju zginanego
z uwzględnieniem sprężystego utwierdzenia ścianek, Inżynieria
i Budownictwo, 1 (2006): 55-58.
6. EN 1993-1-1. 2005. Eurocode 3: Design of steel structures - Part 1-1: General
rules and rules for buildings.
7. EN 1993-1-8. 2005. Eurocode 3: Design of steel structures - Part 1-8: Design
of joints.
8. Rzeszut, K., Garstecki, A.: Zagadnienia niestateczności prętów z kształ-
towników giętych, Inżynieria i Budownictwo, 3 (2006) 166-169.

Unauthenticated
Download Date | 7/4/17 11:29 PM
 TRUSS JOINT WITH POSITIVE ECCENTRICITY EXPERIMENTAL RESEARCH 123

9. Słowiński, K., Swierczyna, S., Wuwer, W., Zamorowski, J.: Podatność

połączeń zakładkowych w konstrukcjach stalowych z kształtowników giętych,
Inżynieria i Budownictwo, 5-6 (2010): 327-331.
10. Szychowski, A.: The stability of eccentrically com-pressed thin plates with
o longitudinal free edge and with stress variation in the longitudinal
direction, Thin-Walled Structures, 46, 5 (2008): 494-505.
11. Szychowski, A.: Stability of cantilever walls of steel thin-walled bars with
open cross-section, Thin-Walled Structures, 94 (2015): 348-358.
12. Zaharia, R., Dubina, D.: Stiffness of joints in bolted connected cold-formed
steel trusses, Journal of Constructional Steel Research, 62, 3 (2006): 240-249.

BADANIA EKSPERYMENTALNE WEZŁA KRATOWNICY Z MIMOŚRODEM

DODATNIM

Streszczenie
W nowoczesnych konstrukcjach stalowych dźwigarów kratownicowych, wykonywanych
z kształtowników giętych na zimno, ze względów technologicznych w węzłach kratownic
występują mimośrody dodatnie, które powodują powstanie dodatkowego zginania ze
ścinaniem w rejonie węzła. W celu określenia rzeczywistej nośności takiego węzła
przeprowadzono badania na modelach konstrukcji kratownicy w skali 1:1. Badania
przeprowadzono w dwóch etapach: pierwszym wstępnym oraz drugim, w którym
wykorzystano wnioski z etapu pierwszego. Przeprowadzenie badań wstępnych było
pomocne w identyfikacji błędów oraz ich późniejszym usunięciu, co przyczyniło się do
uzyskania poprawnych i wiarygodnych wyników badań. W artykule przedstawiono
również wstępne analizy numeryczne badanego modelu kratownicy.
Słowa kluczowe: konstrukcje stalowe, konstrukcje cienkościenne, kształtowniki gięte
na zimno, mimośrody, kratownice, badania eksperymentalne
Editor received the manuscript: 04.07.2016

Unauthenticated
Download Date | 7/4/17 11:29 PM