Proceedings of the

Annual Stability Conference

Structural Stability Research Council
St. Louis, Missouri, April 16-20, 2013

Compression Members with Hollow Sections and Concentric Slotted Gusset

Plates – Behavior and Recommended Design Model
H. Unterweger1, A. Taras2

Abstract
In practical applications, compression members in trusses or bracing systems are often composed
of hollow sections with slotted gusset plates on both ends. Either a bolted connection with splice
plates on both sides or a welded connection are used to achieve a concentric configuration. In
recent application cases, members of this type were designed with unusually long gusset plates at
their ends, leading to reduced load bearing capacities.
In the present paper, the load carrying behavior of such members is shown by means of realistic
numerical calculations. The resulting compression member capacities are compared with the
design models for flexural buckling as they are commonly employed in practice.
It will be shown that these models significantly overestimate the compression member capacity –
particularly in cases with low slenderness. Interestingly, imperfection forms similar to the second
eigenmode often lead to the most critical design situation. The influence of residual stresses due
to the welding of the gusset plate to the slotted hollow section is also studied in detail.
On the basis of these numerical results, an improved engineering design recommendation for the
practical verification of the gusset plate stability could be developed, which should be used in
addition to the conventional member buckling verification.

1. Introduction and motivation for a comprehensive study

Members in trusses as well as members of bracings are often designed with hollow sections
(rectangular – RHS or circular – CHS) and slotted gusset plates on both ends. Either a bolted
connection with splice plates on both sides or a welded connection are used to achieve a
concentric configuration. Two typical examples are shown in Fig. 1.

1
Full Professor, Graz University of Technology, <h.unterweger@tugraz.at>
2
Assistant Professor, Graz University of Technology, <taras@tugraz.at>
 Figuree 1: Examples for member joints – weldedd and bolted, rrespectively (ffrom [1])

A boltedd connectioon leads to a greater lenngth L1, witth reduced bending

b stifffness out off plane, as
is also visible
v in F
Fig. 1. Based on a studdy of comm mon detailingg solutions,, usual valuues for L1,
dependiing on the mmember deptth h, are givven in Fig. 22.

Figure 2: Studied RHS – member joiints; a.) boltedd, b.) welded

Fig. 2 aalso shows tthe two limiiting cases ffor memberr support ouut of plane. In Fig. 2a in axis I, a
pinned end
e must bee assumed, w whereas in Fig. 2b – a connection to a concreete wall –a ppractically
clampedd support is present. Onnly these two cases are studied in ddetail.

The autthors’ motivvation for a comprehensive studyy of the behhavior of coompression members
with slootted gusset plates was given by thhe fact that the literaturre – to the aauthors’ knoowledge –
lacks a ccomprehenssive analyticcal and num
merical studyy dealing w with bucklingg of such m members. It
must bee pointed oout that for these holloow sectionss out of plaane flexuraal buckling is always
predomiinant, whicch can alsoo be interprreted as ann interactionn of globall buckling and local
bucklingg of the gussset plate.

2
Within the literature, a lot of studies are available dealing with the load transfer from the gusset
plate to the hollow section (e.g. Zhao R. (2009) and references there), however the focus is
placed on tension applications, excluding buckling phenomena.
In past detailing practice, the buckling resistance was sufficiently warranted by performing a
buckling check for the hollow section itself, since the gusset plates were usually very short and
had a negligible effect on the buckling length and/or were not the subject to significant second-
order out of plane bending. However, more recent detailing practice led to free lengths L1 of the
gusset plates that are significantly larger than old design tradition.
As will be shown in the following, the limited cross-sectional bending capacity of the gusset
plate can also lead to a significant reduction of the overall compression strength of the member.
Only for very slender members a conventional buckling verification – even when taking account
of the potentially increased buckling length due to the reduced bending stiffness in the joint
region – is sufficient to guarantee structural safety.

If, in addition to a long gusset plate, an eccentric joint configuration is chosen (e.g. directly
bolted without splice plates) a further, significant reduction of the member compression capacity
is observed. In this case design recommendations are given in Unterweger (2010).

2. Simplifications and parameter range of the study

In order to study the load bearing capacity of hollow sections with slotted gusset plates on both
ends, some simplifications were introduced, shown in Fig. 3. In this figure, the studied geometry
with the chosen symbols and notations is presented as well. The two studied border cases for
boundary conditions of the gusset plate are: - pinned (BC1) and, - fully clamped (BC2) ends,
with rigid support out of plane (in both cases).

In the study described here, squared hollow sections were analyzed, ignoring the fillets at the
edges (b = h) and leading to the area A0, second moment of area Jz0 and radius of gyration rz0.
However, the final results can be shown to also be valid for rectangular and circular hollow
sections. The free length L1 of the gusset plate was varied between L1 = 1,0 · h ÷ 2,0 · h. The
height of the gusset plate h1 was fixed with h1 = 1,3 · h, which is typical for practical
applications.

The slotted length Ls, which was chosen to be Ls = 1,75 · h, is of great significance and
represents a minimum value: this length is at least necessary in order to utilize the full axial and
bending capacity of the member at the end of the gusset plate, due to the load introduction from
the gusset plate into the hollow section. This was also verified within the comprehensive FEM-
study, which will be presented in section 4.

3
 Figure 3: Sketch
S of studiied joint confiiguration; geom
metry and idealized boundaary conditions

2.1 Releevant stiffneess ratio I1/II0 and membber length raatio L1/L0
The bucckling capaccity of the m member strrongly depends on the stiffness ratio I1/I0 of tthe gusset
plate (I1 according tto Eq. 1) annd member ((I0 = Iz0)

∙
(1)

Based oon the assum

med geomettry of the gusset
g platess (h1 = 1,3 · h) and quuadratic RHS-sections
used in Europe forr the stiffneess ratio I1/II0, Eq. (2) is
i valid (h = depth andd t = thicknness of the
memberr respectivelly; A0 = areea of the mem mber, A1 = area of the gusset platee)

4,68 ∙ ∙ (2)

In practice, A1 ~ A0 is often vaalid leadingg to very small stiffness ratios of abbout I1/I0 ~ 00,01.

Anotherr important parameter iis the membber length ratio L1/L0. F

For practicaal cases, L1/L
/ 0 = 0,05
÷ 0,10 is often validd.

2.2 Loadding of the m

member
In this sstudy, only constant axxial compreession forcees are consiidered. Geoometric impperfections
were takken into acccount, basedd on executtion tolerancces. The efffect of residdual stressess was also

4
studied, in particular those ressidual stressses caused bby the weldded connecttion betweenn member
and gusset plate.

kling verificcation in prractice and danger of ooverestimation of mem

3. Buck mber capaccity

3.1 Buckkling verificcation in praactice

Due to the significantly reducced bending stiffness off the gussett plate at thhe end of the member
(length L1), for the buckling verificationn of the meember out of plane – based on a constant
bendingg stiffness Izz,0 over the whole
w lengthh L0 – an inncreased effeective lengthh Lcr =  · L0 must be
used. Thhis is illustrrated in Fig.. 4. For mem
mbers with slotted gussset plates, ssometimes tthe second
bucklingg mode is reelevant for tthe ultimatee buckling sttrength of thhe member, so that Lcr,22 based on
2 replaaces Lcr,1 callculated withh 1.

Figuure 4: a.) Firstt and b.) seconnd bifurcation buckling modde shapes and tthe associatedd equivalent coolumns

The relaationship beetween effecctive length and ideal buckling loadd Ncr for each mode i iss given by
the welll known Eq.. (3)

∙ ∙
, (3)
∙

In the liiterature, a number of analytic forrmulas for aan iterative solution foor the effective length
factors i and ideaal buckling loads Ncr,ii can be found (Pflügeer (1964)). Simplified solutions
(Dimitroov (1953), also foundd in Peterseen (1982)),, lead to vvery unsafe results. Apppropriate
analyticcal formulas for the studied
s casses (see Figg. 4) are ssummed upp by the aauthors in
Unterweeger, Taras (2011).

For pracctical appliccations, all the relevantt results forr the effective length faactors i aree summed
up in Fiig. 5. In these figures, tthe parametter I is equaal to the stiff
ffness ratio ((I = I1/I0). T
The results
are plottted for I = 00,01 , 0,02 aand 0,04, covvering all prractical casees.

5
 mination of buuckling lengthss (1st and 2nd bbuckling modee) using the eqquivalent
Figuree 5: Factors i for the determ
m
member for, a.)) pinned ends and b.) fixed ends
e

3.2 Oveerestimation of the buckkling strengtth of a member with slootted gusset plate
Before all the detaails and ressults of the comprehennsive study are presennted in the following
section, a simplified examplee is shown,, in order tto emphasiize the signnificant diffference in
bucklingg behavior when comppared to a siimply suppoorted consttant-sectionn member (ssection Iz,0
over whhole length L0).

Fig. 6 shhows an alteernative appproach for thhe calculation of the ulltimate comppression strrength of a
simply supported member.
m Baased on an eequivalent imperfection
i n eequ and aapplication of second
order thheory, the innternal forcees at the releevant sectioon m (N = P
P, M = P · w
w) are calcuulated. The
increaseed deformattion w depennds on the actual load P and the ideal bucklinng load Pcr, using the
well-knoown secondd-order elasttic amplificaation factor fII in Eq. (44).

⁄
(4)

Making use of the (elastic) m

member capaacity in sectiion m (maxx = Fy), the uultimate com mpression
strengthh of the mem
mber can be determineed analyticaally from seecond-order theory. Som me codes,
like the Eurocode (2012), givee formulas foor eequ that llead to the same
s comprression strenngth of the
memberr as the one given by thhe buckling curves codified in the ssame standaard.

6
 Figure 6: Compressive Strength of a column – alternative approach

Based on the second order theory approach in Fig. 6, it is now possible to show the utilization of
the relevant gusset-plate section capacity in axis 1. In Fig. 7, a typical example of a hollow
section member (RHS: h = b = 200, t = 10 mm) of length L0 = 8000 mm is shown.

The member compression strength of Pn = 840 kN given in Fig. 7a is based on the proposed
verification in Fig. 4, using the effective length factors 1 in Fig. 5 and the buckling curve a of
the Eurocode (2012). The shape of the imperfections is based on the 1st buckling eigenmode
(simplified in a sine-wave form) in Fig. 7a and on the 2nd mode (simplified, with eequ in section 1
of Fig. 7) in Fig. 7b.

When the load is thus set equal to the theoretical resistance Pn = 840 kN, only the hollow
member has sufficient section capacity. In the gusset plate section 1 – particularly for the
imperfections based on the 2nd mode – the utilization factors UF1 are significantly higher than 1,
and this even if the increased plastic section capacity is used (UF1,pl).

This means that the compression strength based on the buckling verification of the member with
increased, equivalent buckling length (see Fig. 4) significantly overestimates the buckling
capacity of the member with slotted gusset plates, because the gusset plate section 1 (and not the
mid-span cross-section of the HS member) determines the ultimate capacity, as well as the fact
that imperfections based on the 2nd mode must also be considered.

7
 h1=260
t1=28

b)

Figure 7: Practical example of a hollow section member with slotted gusset plates; Reduced buckling strength due to
reduced section capacity in section 1 of the gusset plate

4. Simulation of the realistic load bearing behavior of members with slotted gusset plates
The realistic load bearing behavior of members with slotted gusset plates, shown in Fig. 3, was
studied in detail, based on numerical FEM-calculations.

8
4.1 FEM M-model
The num merical calcculations weere done witth the Finitee Element SSoftware AB BAQUS. In Fig. 8 the
FEM-m model at the member ennd is shownn. A combinnation of voolume, shelll and beam m elements
was choosen. The ggusset plates were moddeled using volume eleements withh eight isopparametric
elementts with lineaar interpolattion functionns over the gusset platee thickness. The RHS-mmember in
the joinnt region waas modeled using shell elements w with 20 elemments for eaach section wall. The
welds bbetween the gusset platte and mem mber are noot modeled in detail, im mplying meeaning the
welds. The ssealing platees (see Fig. 1, 3) were omitted. Thhe rest of thhe member
executioon of butt w
outside the joint waas modeledd using beam m elements with a rigidd kinematic coupling too the shell
elementts.

The bouundary condditions in axxis I were bbased on thee studied caases, pinned or fully claamped out
of planee.

Within the calculattions membber lengths ofo L0 = 4,8 and 12 m were considdered with a member
section RHS 200/2000/10 mm w without filleets (leading tto A0 = 76000 mm², rz0 = 78 mm).

In all caalculations, a linear elaastic-ideal pplastic materrial behavioour without strain-hardeening was
chosen, with a yieldd stress Fy = 235 N/mm m². This leads to an ultiimate sectioon capacity of N0,pl =
A0 · Fy = 1.786 kN N. An elasticc modulus oof E = 210.0000 N/mm² and a Poissson’s ratio oof  = 0,3
were sellected.

Figuure 8: FEM nuumerical modeel for the calcuulation of the load bearing caapacity

4.2 Lineear bucklingg Analyses

For each studied m member, a linear buckkling analysses (LBA) w was perform
med prior too the non-
linear caalculations, due to the fact that thhe calculatedd buckling mmodes also formed thee basis for
the seleccted geomettric imperfeections (1st aand 2nd modde studied).

9
Fig. 9 shhows the firrst and secoond bucklingg modes forr a member with L0 = 8000 mm annd pinned
as well aas fully clam
mped ends oof the gusseet plate.

The com mparison of the ideal buckling loads Ncr,i bbased on thhe LBA anaalyses withh the ones
calculatted using bbeam-theoryy (Eq. (3) with effecctive lengthh factors of o Fig. 5) shows a
overestiimation by the latter.. This is uunderstandaable as a cconsequence of the llocal load
introducction from tthe gusset pplate to the member, making
m the first
f portion of the RHS
S member
bendingg stiffness paartially inefffective, leadding to an overestimatioon of Ncr,i.

Based oon a compreehensive stuudy by the aauthors, it iis suggestedd - for practtical applicaations - to
increasee the – com
mputational - free lengthh of the gussset plate L1 (used to ddetermine Ncr,i on the
basis off beam-theorretical methhods such ass the one reppresented byy Fig. 5 ) byy a distance L1 = b/5
T leads too an increaseed length L1* = L1 + h//5 if the effeective lengthh factors off Fig. 5 are
= h/5. This
used (neew parameteer L1*/L0).

Figure 9: Different bbuckling modees (FEM-Modeel) for a membber with L0 = 8 m (Profile R

RHS 200/200/110; gusset
plate 260/28;
2 L1 = 4400 mm)

4.3 Reallistic comprression loadd bearing beehaviour

For the calculationn of realisticc compressiion strengthh values of the membeer with slottted gusset
plates, a series of GMNIA A-analyses (geometricc and matterial nonliinear analyyses with
imperfections) was carried outt.

In a first step, LBBA-analyses were perfoormed. Thee calculatedd eigenmodees (1st and 2nd) were
selectedd for the impperfection sshape alongg the membeer. These shhapes were then scaledd, with the
maximuum value off the imperfe fection choseen with e0,m
max = L0/750 t code of execution
0, based on the
EN 10990-2 (2008) now used in Europe, which givees measurem ments that are relevannt for both
eigenmoode shapes.

10
 fect of an innclination anngle of maxx = 1/100 off the gusset plate was considered
In addittion, the effe c
as well, leading to tthe four diffferent shapees of imperffection, sum
mmarized in FFig. 10.

Figure 110: Realistic assumptions off geometric im

mperfections, based
b on execuutive codes; a..) EN 1090-2 ((2008) and
b.) E
EN ISO 13920 (1996)

4.4 Impaact of additional residuual stresses due to weldding

The imppact of addiitional residdual stressess was also sttudied in thhis paper, duue to relevaance of the
weldingg connectionn between ggusset platee and the m member. Forr the distribbution of thhe residual
stresses, a rather conservativee approach w was selecteed, leading tto residual stresses reaaching the
yield strress along thhe elementss near the weeld (simulattion of the rresidual stresses by a temperature
load caase T). Fiig. 11 show ws the resuults for a m member wiith L0 = 40000 mm. The T stress
Fig. 11c (ressidual stressses E and
distributtions in secttion A, B, C of the gussset plate aree plotted in F
ultimatee stresses 0 and u at top and botttom surfacee of the gussset plate). In I Fig. 11b,, the load-
deformaation behavvior – with and withoout residual stresses – is presenteed. Althouggh a very
conservvative approach for the residual strress pattern was made, only a modderate decreease of the
ultimatee capacity oof approxim mately 10 % (609 kN innstead of 6668 kN in Fiig. 11) was observed.
This eff
ffect will bee considereed in the deesign approoach in section 5 in fform of an increased
geometrric imperfecction (factorr fequ).

11
Figure 111: Load bearinng capacities aand consideredd residual stressses E for a m
member with L0 = 4000 mm
m and pinned
ends (geometric imperfectiions based on the 1st bucklinng mode shape)

5. Suggestion of a design mod del for pracctical appliccation

Based oon the results of the GMNA-calc
G culations foor the studieed members with slottted gusset
plates, the
t authors developed a simple design model for practicaal design. Thhe backgrouund of this
model, summed
s up in section 55.2, will be shown in thhe followingg.

5.1 Design model fo for compresssive strengthh considerinng the limiteed section capacity
c of thhe gusset
plate
As show wn for the example in F Fig. 7, the compressive strength off the membeer is often liimited due
to the lim
mited sectioon capacity of the gusseet plate in itts cross-secttion 1.

The maain idea of tthe design m model, baseed on 2nd orrder theory,, is summedd up in Figg. 12 for a
memberr with pinneed ends. The plot show ws the deform mation U3 ((out-of-planne) in the guusset plate
at the eend of the m member (seection 1) wwith increasiing axial looad N. The circles represent the
“realistiic” behaviouur, determinned by GMNNIA-calculaations.

Startingg from the geeometric im mperfection e1,0 at the member

m end in section 1, the deform mation (U3
nd
in Fig. 112) increasees followingg the elasticc 2 order aamplificationn factor givven by Eq. (4), i.e. the
factor fIII that accouunts for the correct
c ideaal buckling lload Pcr, neaarly up to thhe ultimate lload (U3 =
e1,0 · fII).

In addittion, the ssection cappacity of thhe gusset pllate is also representeed in Fig. 112 by the
descendding lines, lleading to reduced axxial load caarrying capaacites N iff the deform mation U3
II
increasees (due to addditional mooment M = N · U3). A As border linnes the (tooo high) fullyy (i.e. non-
linear) pplastic and tthe (too low
w) linear-elaastic cross-ssectional cappacity curvees for the reectangular

12
gusset plate section are shown. If a reduced plastic capacity curve – with a linear interaction
between axial force and bending moment - is used for the gusset plate, the intersection point with
the deformation curve (U3 = e1,0 · fII) gives the result of the design model (N/Npl,RHS = 0,28 in
Fig. 12).

This compressive strength is in good agreement with the real ultimate load of approximately
N/Npl,RHS = 0,29, albeit the latter features higher deformations U3 = 20 mm.

Figure 12: Deformation and cross-sectional resistance at the relevant gusset plate section for pinned ends

This simple design model works for all types of geometrical imperfections (as described in Fig.
10) and for both buckling modes. It is worth mentioning that - for the deformation U3 based on

13
the simple elastic 2nd order amplification factor, the relevant ideal buckling load Ncr,i must
always be considered.
Table 1 summarized some additional examples. The results of the simple engineering design
model (NEng.model) are compared with the FEM-calculations (NGMNIA).

It can be seen that - for the eigenmode-affine imperfections, with equal amplitudes, see Fig. 10a-
nearly in all cases the 2nd modes lead to the lower ultimate compressive strength.

In the case of a member with clamped ends at the gusset plate (BC2 in Fig. 3), the design model
is more complex due to the fact that the overall moment Mges caused by the deformation w1 at the
member end must be splitted in two individual parts for the two relevant sections I and 1, see
Fig. 13.

Table 1: Load bearing capacities NR [kN] for the variation of geometric imperfection shapes only (based on Fig. 10);
RHS 100/200/10, t1 = 28, L1 = 400 mm
eigenmode-affine; e0,max = L0/750 Gusset plate inclination
end length Ncr,1
e1,0 NGMNIA NEng.model e1,0 NGMNIA NEng.model
support L0 [m] Ncr,2
[mm] [kN] [kN] [mm] [kN] [kN]
BC 1 4,0 1.184 4,51 670 680 (+1,5 %) 4,0 697 707 (+1,4 %)
pinned 1527 5,33 707 708 (+0,1 %) 4,0 772 786 (+1,8 %)
8,0 954 4,45 611 615 (+0,7 %) 4,0 646 633 (-2,0 %)
1.313 10,6 514 497 (-3,3 %) 4,0 726 740 (+1,9 %)
12,0 595 2,78 497 497 (0 %) 4,0 521 467 (-10,4 %) 1
1.213 14,6 430 412 (-4,2 %) 4,0 700 716 (+2,3 %)
BC 2 4,0 4.642 3,27 1.386 1.361 (-1,8 %) 4,0 1.373 1.300 (-5,3 %)
fully 5.199 5,32 1.241 1.217 (-2,0 %) 4,0 1.340 1.313 (-2,0 %)
clamped 8,0 1.791 1,13 1.213 1.357 (+12 %) 4,0 1.312 1.064 (-19 %) 1
4.697 8,65 1.086 1.012 (-6,8 %) 4,0 1.351 1.302 (-3,6 %)
1. not relevant, because buckling of the member determined (section m)

The proposed design model for this cases focuses now on section I, because at this location a
design verification of the welded connection is needed as well. The moment in section I can
conservatively be estimated to amount to about 70 % of Mges, due to the fact that the simple
engineering model underestimates the deformation w1 by using fII. In partial compensation of
this error, it can be shown that at section I the full plastic section capacity can be utilized.

14
 Figure 13: E
Engineering m
model for the caalculation of thhe load bearinng capacity Nd for fixed endds

5.2 Summmary of thee suggested design moddel for the coompression strength off members w
with
slotted ggusset platees
Summinng up the results of section 4 and 5.1, a design moodel for thhe calculatioon of the
compresssion strenggth of membbers with slootted gusset plates is noow availablee.

This design model can be adaapted to diffferent natioonal codes, which mayy have diffeerent rules
(particuularly, buckling curvess and toleraance requireements) forr column bbuckling verrifications
against axial comprression forcces Nd. For a better undderstanding of the proposal, in the following
represenntation of thhe design moodel all safeety factors aare omitted.
The verrification oof the mem mber contaiins two inddividual chhecks - botth must be fulfilled.
Alternattively, for each veriffication a m maximum compressioon strength can be determined
d
(utilizattion factor eequal to 1,0)) – the miniimum of booth gives thee resulting compression
c n capacity
of the m
member.

 Veriification 1: bbuckling chheck of the member

m
 Connventional bbuckling check, based on an incrreased effecctive buckliing length ffor the 1st
buckkling mode (L = 1 · L0, see Fig. 5))
 For the calculaation of thee effective length facttor 1 usingg e.g. the ggraphs in FFig. 5, an
increeased lengthh L1* = L1 + h/5 shouldd be used
 Notee: Verificattion 1 checkks the mem mber in the region neaar midspan against seccond-order
internal fforces (comppression + bbending) exxceeding thee section cappacity

 Veriification 2: aadditional vverification oof the sectioon capacity of the gusseet plate
a) geom metric impeerfection at tthe gusset pplate (sectionn 1)
e1,0 = L1/100 ≥ 2 mm

15
 Note: Based on execution codes EN 1090-2 (2008) and EN ISO 13920 (1996). Different
imperfection amplitudes may be considered on the basis of other national or industry-specific
tolerance requirements.

b) geometric equivalent imperfections at the gusset plate (section 1)

e1,equ = fequ · e1,0 = 2.0 · e10 = 2.0 · L1/100 ≥ 4 mm
Note: The factor fequ is a very conservative assumption, based on the study of the effects of
residual stresses due to welding of the connection between gusset plate and member.

c) verification
 pinned ends (BC1): (only reduced plastic section capacity available)
∙ , ∙ , , ∙ 1 ⁄ , , (5)
⁄ ,

 fully clamped ends (BC2):

0,7 ∙ ∙ , ∙ , , ∙ 1 ⁄ , , (6)
⁄ ,
- End moment MI (basis for weld connection design)
0,7 ∙ ∙ , ∙ ⁄
(7)
,

Note 1: - the ideal buckling load Ncr,1 for the 1st mode must always be used
Note 2: - the plastic section capacity M1,pl,Rd and N1,pl,Rd of the gusset plate are defined (with
Fyd … design value of the yield stress), respectively:
∙
, , ∙ ; , , ∙ ∙

References
Zhao, R., Huang, R., Khoo, H.A., Cheng, J.J.R. (2009). “Parametric finite element study on slotted rectangular and
square HSS tension connections”. Journal of constructional steel research, 65 611-621.
Unterweger, H. (2010). “Load bearing capacity of bracing members with almost centric joints”. Proceedings of Int.
conference stability and ductility of steel structures – SDSS, Rio de Janeiro, 603-610.
Unterweger, H., Taras, A. (2011). “Hohlprofile mit beidseits zentrisch eingeschlitzten Knotenblechen –
Drucktragverhalten und Bemessungsvorschlag.” Stahlbau, Band 680, Heft 11, 839-851.
Eurocode EN 1993-1-1 (2012). “Eurocode 3; Design of Steel Structures – Part 1-1: General rules and rules for
buildings.”
Eurocode EN 1090-2 (2008). “Execution of steel structures and aluminium structures – Part 2: Technical
requirements for steel structures.”
EN ISO 13920 (1996). “Welding – General tolerances for welded constructions – Dimensions for lengths and
angles, shape and position.”
Petersen, Ch. (1982). “Statik und Stabilität der Baukonstruktionen”. Vieweg & Sohn, 2. Auflage.
Pflüger, A. (1964). “Stabilitätsprobleme der Elastostatik”, Springer Verlag, 2. Auflage.
Dimitrov, N. (1953). “Ermittlung konstanter Ersatz-Trägheitsmomente für Druckstäbe mit veränderlichen
Querschnitten”. Der Bauingenieur, 28, H6, 208-211.

16