Available online at www.sciencedirect.

com

ScienceDirect
Procedia Engineering 111 (2015) 789 – 796

XXIV R-S-P seminar, Theoretical Foundation of Civil Engineering (24RSP) (TFoCE 2015)

Selection of parameters for I-beam experimental model subjected to

bending and torsion
Alexander R. Tusnina *, Milan Prokica
a
Moscow State University of Civil Engineering, 26, Yaroslavskoe shosse, Moscow, 129337, Russia

Abstract

Thin-walled rods in steel structures can work in conditions of bending with torsion. Dealing with bending, conjugate with
warping torsion can be a challenging task. To make justified choice of construction parameters, a more detailed study of its
behavior and carrying capacity in this case is needed. Currently, the Russian norms have no guidance on accounting for plastic
deformations when calculating beams experiencing simultaneously bending and torsion. Research of bearing capacity at such
work, taking into account development of material plastic deformations, is a topical issue. To develop a methodology for the
design of beams experiencing both bending and twisting a careful theoretical and numerical analysis of these structures is needed.
Justification of numerical and theoretical results requires experimental evaluation. Currently in MGSU experimental studies of I-
beams bearing capacity considering their warping torsion are under preparation. The article considers the numerical calculations
of the experiment model of I-beam under torsion with bending. Behavior of the model was analyzed, considering mesh
partitioning into finite elements, diagram of the material, support conditions. On the basis of studies, the parameters of test
models, applied loads value were determined and planning of experiment was carried out.
© 2015 The Authors. Published by Elsevier B.V.
© 2015 Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license
Peer-review under responsibility of organizing committee of the XXIV R-S-P seminar, Theoretical Foundation of Civil
(http://creativecommons.org/licenses/by-nc-nd/4.0/).
Engineering
Peer-review (24RSP)
under responsibility of organizing committee of the XXIV R-S-P seminar, Theoretical Foundation of Civil Engineering (24RSP)

Keywords: Warping torsion; Bending; Numerical studies; I-beam; Test model

1. Introduction

Behavior of symmetric I-beams under the action of bending moment is well studied, up to the formation of plastic
hinge and loss of bearing capacity. In construction norms recommendations for plastic analysis of I-beams in

* Corresponding author. Tel.: +7-916-115-1421; fax: +7-495-287-4914.

E-mail address: valeksol@mail.ru

1877-7058 © 2015 Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license
(http://creativecommons.org/licenses/by-nc-nd/4.0/).
Peer-review under responsibility of organizing committee of the XXIV R-S-P seminar, Theoretical Foundation of Civil Engineering (24RSP)
doi:10.1016/j.proeng.2015.07.146
790 Alexander R. Tusnin and Milan Prokic / Procedia Engineering 111 (2015) 789 – 796

bending are given. Accounting of plastic deformations allows increasing the bearing capacity of profile. In thin-
walled open-section members warping torsion is causing additional sectorial stresses, which are equivalent to
internal force – bimoment. Under combined bending moment and bimoment actions, the transition to plastic
condition of member occurs earlier than in bending. Studding of combined bending and warping torsion actions with
development of plastic deformations will lead to practical methods for calculating thin-walled open-section members
in complex loading.

Nomenclature

l beam span
h section height
bf flange width
tf flange thickness
tw web thickness
IȦ warping torsion section constant
It uniform torsion section constant
k flexural-torsional characteristic of the rod
E modulus of elasticity
G shear modulus of elasticity
P load
M bending moment
T torque
B bimoment
Bpl plastic bimoment
Mpl bending moment at plastic collapse
ș rotation angle

Airumyan E.L., Rybakov V.A., Vatin N.I. [1-3] have shown the inapplicability of plane-sections hypothesis when
analyzing stress-strain state of thin-walled open section profile. The unclear accounting of bimoment as an additional
force factor, present in the Russian construction norms, was also noted. Theoretical and experimental studies
confirmed the importance of accounting bimoment as force factor, which contributes greatly to the normal stresses
of structural elements [4-10].
Studies of steel I-beams in warping torsion showed that loss of member bearing capacity is due to plastic collapse
of flanges and at large rotation angles [11]. Experimental results [12] and numerical studies [13] have showed that
strain-hardening and second order longitudinal stresses cause significant strengthening at a large rotations.
By the authors of this paper was suggested to perform the calculation of I-beam profiles in a form consistent with
Russian construction norms, and formulated an expression for bearing capacity check of I-beam in plastic stage
under the action of bimoment [14]. The calculation is performed using the plastic deformations coefficient for
warping torsion.
Experimental and theoretical studies of I-beam bearing capacity under bending and warping torsion should be
performed. From previous studies, it is known that the bearing capacity of I-section under the action of only
bimoment increases by about 1.5 times. Further numerical calculations have shown that if bending moment is added
to the action of bimoment, reserves of bearing capacity due to plastic deformations in warping torsion are reduced
[15]. Coefficient, accounting for the plastic deformations of steel, decreases depending on the proportion of bending
moment in normal stresses. Therefore, to confirm the numerical calculations, it is necessary to carry out
experimental studies. To define the parameters of experimental model, a preliminary calculation of beams is needed,
including the physical and geometric nonlinearity[16-19]. When selecting profiles for an experiment, cross section
height and spans, the characteristics of equipment used for the tests and particularities of investigated stress-strain
state should be considered.
 Alexander R. Tusnin and Milan Prokic / Procedia Engineering 111 (2015) 789 – 796 791

This article gives an analysis of the numerical calculations results, with physical and geometric nonlinearity. The
behavior of models at different material diagrams and mesh partitioning into finite elements is investigated. A
comparison of normal stresses value from bending and warping torsion is given, obtained by numerical calculations
and using known theoretical relationships.
Based on the results of studies, parameters of test specimens and constructive solution of beam support sections
were determined.

2. Methods

Theoretical studies are performed using numerical methods. In order to substantiate the accuracy of calculations
performed with physical and geometric nonlinearity, the convergence of results depending on mesh partitioning into
finite elements was evaluated. The rod was modeled by shell finite elements, in program complex Nastran.
The steel material was modeled by two diagrams, bilinear and unified shown in Fig. 1.
a) b)

Fig. 1. Stress-strain relationships: (a) Bilinear diagram; (b) Unified diagram.

A cantilever beam was calculated, from steel S245, section 50B3, 4m long, at its free end loaded with torque
7 N1ÂPDKLJKHUWKDQYDOXHRIWKHHODVWLFWRUTXHDQGFDXVLQJQRUPDOVWUHVVRQWKHDQFKRUHGHQGRIEHDP
the beyond yield point of steel.
Based on the calculation results, plastic bimoment on anchored end of cantilever and rotation angle of free end
were determined. Bimoment was calculated as equivalent value corresponding to the distribution of normal stresses
in considered section.
Plastic bimoment is theoretically defined as the product of plastic moments in flanges, on the distance between
them (Fig. 2):

B pl M pl ˜ h (1)

Fig. 2. Scheme for determining plastic bimoment

792 Alexander R. Tusnin and Milan Prokic / Procedia Engineering 111 (2015) 789 – 796

Figure 3 shows a vector distribution of normal stresses in the cross section of I-beam, at four, eight, sixteen and
thirty-two elements across the width of the flange, respectively.

a) b)

c) d)

Fig. 3. Normal stresses in the cross section flanges with: (a) 4 elements; (b) 8 elements; (c) 16 elements; (d) 32 elements.

Results of calculations, depending on the mesh partitioning into finite elements, are shown in Table 1.

Table 1. Results of numerical calculations.

Mesh partitioning* 2x6x40 4x10x80 8x20x160 16x40x320 32x80x600

Rotation angle ș
0.2962 0.2960 0.3001 0.3037 0.3049
(rad)
Bilinear diagram
Bimoment %Ȧ
13.019 14.889 16.558 16.793 16.811
(kNÂm2)
Rotation angle ș
0.2965 0.3025 0.3061 0.3096 0.3124
(rad)
Unified diagram
Bimoment %Ȧ
12.951 14.645 16.268 16.487 16.476
(kN·m2)
Rotation angle 94.8 94.8 96.1 97.2 97.6
The percentage of Bilinear diagram
results in Bimoment 79.0 90.4 100.5 101.9 102.0
comparison with
highlighted table Rotation angle 94.9 96.8 98.0 99.1 100.0
values Unified diagram
Bimoment 78.6 88.9 98.7 100.1 100.0
*Mesh partitioning indicates the number of finite element across the flange width, wall height and length of the beam, respectively.

Comparison of the results was performed relative to the values obtained with the densest mesh partitioning and
using unified diagram of steel. For clarity, the results of calculations are shown on the graph (Fig.4) where,
depending on mesh partitioning, the bimoment value varies with two diagrams of material.
 Alexander R. Tusnin and Milan Prokic / Procedia Engineering 111 (2015) 789 – 796 793

Fig. 4. Dependence of bimoment from mesh partitioning and material diagram.

Analysis of the results showed that already at eight elements across the width of the flange plastic bimoment in
the cross section tends to the limiting value, and with the further mesh refinement its value practically do not
increase. It should be noted, that previous studies of torsional calculations convergence in the elastic stage shown,
that even at the two elements across the width of flange, results close to theoretical are obtained. However, if
numerical calculations over steel yield strength are carried out, than mesh partitioning should be taken in which are
at least eight finite elements across the I-beam flange width.
The results obtained using the two diagrams differ by no more than 2%. Therefore, for calculations with physical
nonlinearity of steel, it is expedient to apply the bilinear diagram, since the calculating time is significantly reduced.
These results were used when choosing parameters for the numerical calculation of experimental models.
For the convenience of experiment calculation scheme shown in Fig. 5 was selected. Simply hinged beam, which
has on supports the rotation of cross sections excluded, and warping displacements of cross sections allowed. In the
mid-span, beam is loaded by a force acting in the vertical plane with eccentricity relative to the longitudinal axis,
which creates simultaneously torque and bending moment in the beam.
a) b)

Fig. 5. Beam experimental model: (a) Calculation scheme; (b) Deformed model in Nastran.

Preliminary calculations were performed using the known theoretical dependencies [20]:

T sh (kl / 2)
B ˜ (2)
2k ch (kl / 2)

P ˜l
M (3)
4
794 Alexander R. Tusnin and Milan Prokic / Procedia Engineering 111 (2015) 789 – 796

T § kl sh (kl / 2) ·
T 3
˜¨  ¸ (4)
2k EIZ © 2 ch (kl / 2) ¹

According to formulas (1) and (2), normal stresses from warping torsion and bending were calculated, and have
been used in determining the total normal stresses and displacements.
The behavior of thin-walled open-section rod in torsion is characterized by dimensionless parameter
(dimensionless characteristic number [21]), equal to the product of flexural-torsional characteristic of the rod on its
span:

GI t
kl l (5)
EIZ

Based on laboratory equipment (Instron 3382 with force up to 100 kN), a beam profile and spans were chosen in
such a way to cover the range of parameter, close to the practical. In view of this, beam has been selected, a
standard profile 12B1, with spans in the range of 60cm to 120cm. Beam material was accepted steel S345. For
selected spans calculations were performed, that allowed to determine the eccentricities of load application. The ratio
of warping torsion and bending normal stresses is within the range of 1.9 to 1.5.
Flat specimens were tested in tension to determine the yield stress, modulus of elasticity and diagram of the test
specimens material.
Based on average test data required parameters are defined, and the diagram of material is obtained. The material
test results are given in Table 2.

Table 2. Results of material test.

Speci Specimen Cross- Length, Length Relative Tensile Yield Force at Ultimate Force at
men thickness section (mm) after displacem modulus of strength yield tensile stress ultimate
width (mm) area, rupture, ent, (%) elasticity (MPa) point (MPa) tensile stress
(mm) (mm2) (mm) (MPa) (kN) (kN)

9.40 3.80 35.72 60.00 80.00 33.33 253978.20 329.46 11.77 482.93 17.25

Analysis of beams in program complex Nastran with geometric nonlinearity was carried out, using the previously
defined parameters of steel. Table 3 presents the main parameters of calculated beams and magnitude of the ultimate
loads under which a loss of bearing capacity occurs.
Calculations have set the value of ultimate load, at which profile losses its bearing capacity, it is within the range
of 25kN to 55kN, and corresponds to the power of press, available in laboratory. In this case, the eccentricity of load
is taken constant and equal to 2cm, since the ratio of the bending and sectorial stresses, will vary with the change the
of tested beam span. Also was chosen the thickness of stiffener installed in the beam middle-span, in a place of load
application. Its purpose is to ensure the rigidity of cross-section contour, according to the provisions of Vlasov V.Z.
theory [21].

Table 3. Parameters of test specimens.

Specimen no. 1 2 3 4 5 6
Span (cm) 60 70 80 90 100 120
h (cm) 11.76 11.76 11.76 11.76 11.76 11.76
tw (cm) 0.38 0.38 0.38 0.38 0.38 0.38
bf (cm) 6.40 6.40 6.40 6.40 6.40 6.40
tf (cm) 0.51 0.51 0.51 0.51 0.51 0.51
kl 1.224 1.428 1.632 1.836 2.040 2.447
Eccentricity (cm) 2.00 2.00 2.00 2.00 2.00 2.00
Bending moment (kNÂcm) 794.10 784.70 778.20 766.35 757.81 731.25
Torque (kNÂcm) 105.88 89.68 77.82 68.12 60.62 48.76
Bimoment (kNÂcm2) 1415.8 1347.9 1283.5 1210.5 1144.1 1004.8
Ultimate load [Nastran] (kN) 52.94 44.84 38.91 34.06 30.31 24.53
 Alexander R. Tusnin and Milan Prokic / Procedia Engineering 111 (2015) 789 – 796 795

After determining the main parameters of calculation scheme, load values and eccentricities, supports were
developed, load application method and the points are chosen for installation of strain gauges. Fig. 6 shows
assembled experiment installation and deformed beam after a loss of bearing capacity and unloading.
a) b)

Fig. 6. Tests of 12B1 beam in bending with warping torsion: (a) Experiment installation; (b) Deformed specimen.

Before the tests, strain gauges were set in places where normal stresses in the section were expected close to
maximal. Installation locations of strain gauges are shown in Figure 7. Also, during loading process, the rotation
angle and deflection of beam mid-span section were measured.
Graphs of strains, stress and rotation angle variation with load, obtained experimentally (dashed line) and by
numerical calculations in Nastran (solid line) are shown in Fig.7.

a) b)

c)

Fig. 7. Graphs of specimen no. 3 with 80cm span: (a) Load - strain; (b) Load - stress; (c) Load – rotation angle.
796 Alexander R. Tusnin and Milan Prokic / Procedia Engineering 111 (2015) 789 – 796

3. Conclusions

Numerical and experimental studies of I-beams under combined bending and torsion actions were performed, up
to the loss of element bearing capacity. Behavior analysis of experimental samples confirmed that a beam in bending
and torsion has significant bearing capacity reserves when compared with the elastic stage of work.
By results of preliminary tests should be noted that deformation pictures, obtained by calculations and by tests,
are qualitatively similar. The ultimate load value, at which test specimen losses bearing capacity, is 12% higher than
its calculated value. The values of rotation angle differ by no more than 20%.

References

[1] E.L. Airumyan, Osobennosti rascheta konstruktsiy iz tonkostennykh gnutykh profiley, Montazhnye i spetsial'nye raboty v stroitel'stve, (3),
2008, pp. 2-7
[2] V.A. Rybakov, O.S. Gamayunova, The stress-strain state of frame constructions’ elements from thin-walled cores, Internet Journal
"Construction of Unique Buildings and Structures", No.7(12), 2013, pp. 79-123
[3] N.I. Vatin, A.S. Sinelnikov, Strength and Durability of Thin-Walled Cross-Sections. Design, Fabrication and Economy of Metal Structures.
International Conference Proceedings, 2013, pp. 165-170.
[4] J. Becque, K.J.R. Rasmussen, Experimental investigation of local-overall interaction buckling of stainless steel lipped channel columns,
Journal of Constructional Steel Research, Vol. 65, Issues 8-9, 2009, pp. 1677-1684.
[5] N.S. Boulton, Plastic twisting and bending of an I-section in which warp is restricted, Int. J. Mech. Sci., Vol. 4, 1962, pp. 491-502
[6] K.S. Dinno, S.S. Gill, The plastic torsion of I-sections with warping restraint, Int. J. Mech. Sci., Vol. 6, 1964, pp. 27-43
[7] K.S. Dinno, and W. Merchant, A procedure for calculating the plastic collapse of I-sections under bending and torsion, The Struct. Engr.,
43(7), 1965, pp. 219–221
[8] C.R. Farwell, T.V. Galambos, Nonuniform torsion of steel beams in elastic range, J. Struct. Div., 95(12), 1969, pp. 2813–2829
[9] M. Heinisuo, J. Kukkonen, Resistance of cold-formed steel members by new eurostandard, Structural Mechanics, 39(2), 2006, pp. 3-21
[10] V.V. Lalin, V.A. Rybakov, S.A. Morozov, The Finite Elements Research for Calculation of Thin-Walled Bar Systems, Magazine of Civil
Engineering, Vol. 1, 2012, pp. 53-73
[11] Y.L. Pi, N.S. Trahair, Inelastic torsion of steel I-beams, Journal of Structural Engineering, 121(4), 1995, pp. 609–620
[12] A. Aalberg, An experimental study of beam-columns subjected to combined torsion, bending, and axial actions, Dr.ing. thesis, Trondheim:
Department of Civil Engineering, Norwegian institute of Technolohy, 1995
[13] N.S. Trahair, Non-linear elastic non-uniform torsion, Journal of Structural Engineering, 131(7), 2005, pp. 1135–1142
[14] A.R. Tusnin, M. Prokic, Resistance of I-beams in warping torsion with account for the development of plastic deformations, Vestnik MGSU
[Proceedings of Moscow State University of Civil Engineering], Vol. 1, 2014, pp. 75-82 (rus)
[15] A.R. Tusnin, M. Prokic, Bearing capacity of steel I-sections under combined bending and torsion actions taking into account plastic
deformations, Journal of Applied Engineering Science, Vol. 11, No 3, 2014, pp. 179-186
[16] N.I. Vatin, J. Havula, L. Martikainen, A. Sinelnikov, A. Orlova, S. Salamakhin, Thin-walled cross-sections and their joints: tests and FEM-
modelling, Advanced Materials Research, Vols. 945-949 , 2014, pp. 1211-1215
[17] B.W. Schafer, T. Peköz, Computational modeling of cold-formed steel: characterizing geometric imperfections and residual stresses,
Journal of Constructional Steel Research, 47(3), 1998, pp.193-210
[18] Z. Li, S. Ádány, B.W. Schafer, Modal identification for shell finite element models of thin-walled members in nonlinear collapse analysis,
Thin-Walled Structures, Volume 67, 2013, pp. 15-24
[19] D.S. Shatov, Finite modeling of thin-walled perforated studs, Magazine of Civil Engineering, Vol. 3, 2011, pp. 32-35.(rus)
[20] D.V. Bychkov, Stroitel'naya mekhanika sterzhnevykh tonkostennykh konstruktsiy [Structural mechanics of bar thin-walled systems],
Moscow: Gosstroyizdat Publ., 1962, 475 p.
[21] V.Z. Vlasov, Tonkostennye uprugie sterzhni [Thin-walled Elastic Beams], Moscow: Fizmatgiz Publ, 1959, 568 p.