Journal of Constructional Steel Research 66 (2010) 772779

Contents lists available at ScienceDirect

Journal of Constructional Steel Research

journal homepage: www.elsevier.com/locate/jcsr

Displacement-based nonlinear finite element analysis of composite

beamcolumns with partial interaction
Joo Batista M. Sousa Jr. , Claudio E.M. Oliveira, Amilton R. da Silva
Dept. Civil Engnrg., Escola de Minas, Universidade Federal de Ouro Preto, 35400-000, Ouro Preto, MG, Brazil

article info abstract

Article history: Several structures employ members with a deformable connection between two components, also known
Received 18 September 2009 as partial interaction or interlayer slip. Classical examples include composite steelconcrete beams and
Accepted 27 December 2009 glued or nailed timber beams. Most of the numerical research in this topic, however, has been focused
on composite beams with material nonlinearity. In some situations these members may be subjected to
Keywords: compression and bending, for which 2nd order effects should be taken into account. The purpose of this
Composite beams
work is to develop and test a displacement-based finite element model for the nonlinear material and
Partial interaction
Interlayer slip
geometrical analysis of composite beamcolumns with interlayer slip. The finite element is based on a
Nonlinear analysis total Lagrangian description, in the context of large displacements, small strains and moderate rotations.
The robustness and accuracy of the proposed scheme is verified against examples from the literature.
2010 Elsevier Ltd. All rights reserved.

1. Introduction uncommon, as elements with a deformable connection may

It is common practice in civil engineering to employ members
which consist of an assemblage of two or more different elements
connected by a deformable interface. Examples are composite
steelconcrete beams, where the connection is normally done by
shear connectors welded to the steel beam and embedded in the
concrete, and composite wood beams, glued or connected by nails
or other devices.
Early works dealing with composite beams with interlayer
slip focused on the development of closed-form solutions. The
differential equations for the case of linear elastic composite beams
with linear connection stiffness were originally developed in the
early 50s [1]. Numerical methods have been employed to solve
more general problems of composite beams, mostly by the finite
element method. Displacement-based [2,3], force-based [4] and
mixed [5,6] finite elements have been largely used in problems
with linear and nonlinear material behaviour. A comparative study
of different schemes is given by Ranzi et al. [7].
Partly because of the assumptions commonly used in the
analysis of composite steelconcrete beams, few works have
dealt with geometrical nonlinear analysis on beamcolumns
with interlayer slip, or even with the influence of the normal
force without 2nd order effects. However, such situation is not
Corresponding author. Tel.: +55 31 35591564; fax: +55 31 35591548.
E-mail addresses: joao@em.ufop.br, joaobatistasousajr@hotmail.com
(J.B.M. Sousa Jr.), claudioernani@yahoo.com.br (C.E.M. Oliveira),
amilton@metalicas.em.ufop.br (A.R. da Silva).
0143-974X/$ see front matter 2010 Elsevier Ltd. All rights reserved.
doi:10.1016/j.jcsr.2009.12.015
eventually be part of a frame resisting vertical and lateral loads.
Nonlinear geometric effects were addressed by Girhammar
and Gopu [8], who obtained the differential equations and
boundary conditions for beamcolumns with interlayer slip
and linear material behaviour. Later, Girhammar and Pan [9]
extended the work on these closed-form solutions, discussing
the consideration of different boundary conditions and buckling
lengths of beamcolumns with interlayer slip. Xu and Wu [10]
developed the analytical formulation for geometric nonlinear
problems within the context of Timoshenko beam theory.
Numerical solutions for the geometric nonlinear problem of
beams with deformable connection were investigated by as
et al. [11]. These authors developed a nonlinear finite element
scheme based on strain interpolation, under Reissner beam
theory. Pi et al. [12] developed a total Lagrangian formulation
for nonlinear geometrical and material analysis of composite
beam columns and provided an extensive discussion on aspects
such as steel and concrete plasticity, importance of shear strains
and consistent linearisation of the virtual displacement principle.
The relative slip between the steel and concrete components
due to a flexible bond at the interface between the steel
and concrete components was considered as an independent
displacement. The implementation of a FE model and the
respective incremental-iterative solution was described in a
companion paper [13]. Krawczyk and coworkers [14,15] addressed
the problem analytically and developed a corotational formulation
for the nonlinear analysis of composite beams with interlayer
slips, based on Timoshenko beam theory. Very recently, Battini
et al. [16] developed the corotational transformations which can
be employed to derive formulations for large displacements and
 J.B.M. Sousa Jr. et al. / Journal of Constructional Steel Research 66 (2010) 772779 773

rotations of beams with interlayer slips. Within the corotational To develop a finite element formulation for the static analysis
framework different local elements may be employed. In their of composite beamcolumns under the above assumptions the
work, they exemplified the procedure applying it to a local element Virtual Work Principle will be employed. For an isolated element
based on the exact solution and presented several examples. of length `, the internal virtual work is given by the sum of the
There is still room for research, however, for cases of combined contributions of the two components plus the interface connection
material and geometrical nonlinearities. The purpose of the
`
Z !
present work is to develop a total Lagrangian displacement- X Z
Wint = x x dA + sS dx (6)
based finite element formulation for the nonlinear material and 0 =1,2 A
geometrical analysis of composite beamcolumns with interlayer
slip, which is able to handle nonlinear effects such as those and the external virtual work comes from the contribution of the
present in composite steelconcrete beamcolumns. It is based on surface and body loads
previously developed finite element formulations, such as the ones Z !
X
developed in [2], taking into account small strain and moderate Wext = u p x + v py dV
rotation effects. The slip is obtained from the beam displacements V =1,2
and rotation, so there is no need for a separate slip interpolation. Z !
At the end, some examples of nonlinear material and geometrical X
composite beams are presented to illustrate the accuracy and + u t x + v ty d (7)
=1,2
robustness of the finite element formulation. The results are
verified against analytical as well as numerical solutions. with px , py denoting forces per unit volume and tx , ty forces per unit
area.
2. Formulation Equality of internal and external virtual works for arbitrary
compatible displacement and strain fields is equivalent to the
The kinematical hypotheses for the composite beam are based weak form of equilibrium equations. Strong forms (i.e. differential
on the EulerBernoulli beam theory, namely, cross sections remain equations) may then be devised employing integration by parts
plane and normal to the reference axis after deformation. This of the virtual work equation, see [2] for the linear case. From the
gives rise to expressions involving trigonometric functions on the definition of the axial strain (4), the incremental strain is given by
beam cross section rotations which are essential if one wishes to differentiation as
take into account the general finite deformation case. However, x = u0,x (y y )v,0xx + vx0 vx0 . (8)
in civil engineering applications and especially with reinforced
concrete, the usual framework is that of small strains and moderate Introducing (8) in the internal virtual work expression, and defin-
rotations. DallAsta et al. [17] and Zona et al. [18] have discussed the ing the normal force and bending moment on each component as
mathematical aspects of these simplifications on the strain field for Z
nonlinear geometrical and material analysis of prestressed beams, N = x dA (9)
and equivalent assumptions are applied in the present work. A

Within the simplified kinematical scheme, with superscript 0

Z
denoting values measured at the reference axis, the in-plane beam M = x (y y )dA (10)
displacements u and v are mathematically expressed as
A

the internal virtual work may be rewritten in matrix form as

u (x, y) = u0 (x) (y y ) (x) = u0 (x) (y y )v,0x
Z `
= 1, 2 (1) Wint = t dx (11)
0
v (x, y) = v (x)0
(2)
with the generalised strains and stresses given by
for the displacements of the components and
T = 10 20 T = N1 S .
   
s and N2 M (12)
s(x) = u02 (x) u01 (x) + h v
0
,x (3)
Collecting the displacements in vector u
for the slip s between the components (Fig. 1), with h = y2 y1 .
uT = u01 u02 v0
 
These equations have been employed in several formulations for (13)
composite beam analysis in the geometric linear regime. In the the generalised strains are given by
case of geometric nonlinearity, large displacements and moderate
rotations also affect the kinematics of the interlayer slip, and  = u (14)
the horizontal and vertical components of s should be evaluated with matrix given by
separately. However, if it is assumed that the slips themselves
1

are moderate (as the rotations), Eq. (3) remains valid for practical x 0 v,0x x
engineering applications and as such was employed by Girhammar 2
1

and coworkers [8,9] in their analytical formulations. = 0

x v x .
0
(15)
,x
Considering a total Lagrangian formulation, the only strains in 2
the beams are the GreenLagrange axial strains, which in the small 0 0 xx

strain and moderate rotation context are given by 1 1 h x
1 The generalised strain variations are given in a similar fashion by
x = u0,x (y y )v,0xx + (v,0x )2 . (4)
2
 =
u (16)
2
The term involving v 0 ,x is responsible for the nonlinear geometric with matrix given by
effects. Introducing the membrane and bending parts the strain
x v,0x x

may also be expressed as 0
x = 0 + (y y ) (5)
0
= x v,0x x
. (17)
0 0 xx
with = v 0
,xx representing the curvature. 1 1 h x
774 J.B.M. Sousa Jr. et al. / Journal of Constructional Steel Research 66 (2010) 772779

Fig. 1. Displacement field of composite beam.

with
uT,x 0 vT v,x v,
T

x
0 uT,x vT v,x v,
T
x
=
B

(22)
T

0 0 v,xx

uT uT hv,
T
x

which after introduction into the virtual work expression allows

Fig. 2. Degrees of freedom of composite beam.
the identification of the internal force vector g
Z `
2.1. Displacement interpolation T dx.
g= B (23)
0
To carry out the development of the FE formulation the The external virtual work is a function of the applied loads (body
displacement interpolation must be defined. For the axial dis- forces, surface and member end loads). These terms will give rise
placements u0 and transverse displacements v 0 the interpolation to equivalent loads r to be applied to the discrete finite element
functions are introduced model and their evaluation follows standard procedures.
The tangent stiffness matrix is obtained from the derivative of
u0 = uT u and v 0 = vT v (18)
the internal force with respect to the displacement vector
with identical interpolations for virtual quantities u0 and v 0 . The ` `
g B T
Z Z
axial interpolation functions u may be polynomials starting from k= = dx + T
B dx. (24)
the linear case. The transverse interpolation functions must reflect d 0 d 0 d
the continuity of v 0 and its first derivative, therefore they should The second integral on the r.h.s. of Eq. (24) is obtained using
be at least cubic (hermite polynomial). 
The choice of the interpolation functions for the case of beam = = CB (25)
elements with interlayer slip has been the subject of intense d  d
debate over the past years, as non-matching polynomial terms where matrix C collects the cross section stiffness coefficients
may arise as the slip expression involves values of u and first

EA1 0 ES1 0
derivatives of v 0 . This causes numerical problems which have 0 EA2 ES2 0
C= (26)
been named slip locking or curvature locking [19]. To avoid this ES1 ES2 EI1 + EI2 0
abnormal behaviour the axial displacement interpolation should 0 0 0 K
be at least quadratic if the transverse interpolation is cubic. In the
which, if Et = dd is defined as the material tangent modulus, are
case of geometric nonlinear analysis there is additionally a poor
defined as
balance of terms due to the factor v,2x in the axial strain which Z
theoretically would require a very high polynomial description for dN
EA = = ET dA (27)
u0 . However, numerical experience has shown that the quadratic d0 A
axial interpolation already renders very reasonable results even Z
dN dM
with this inadequate balance. In view of these facts the degrees of ES = = = ET (y y )dA (28)
freedom of the element are those depicted in Fig. 2. d d 0 A
Collecting the element degrees of freedom in vector d dM
Z
EI = = ET (y y )2 dA (29)
dT = u01 u02 d
 
v (19) A

along with the connection tangent stiffness

the matrix expression for the displacement interpolation is u =
8d where dS
K = . (30)
ds
u
T
0 0
8 = 0 uT 0 . (20)
2.2. Material nonlinearity
0 0 vT
Employing the displacement interpolation one gets The integrations in Eqs. (9), (10) and (27)(29) are influenced by
material nonlinearity and may be performed in a number of ways.
 = 8 d = B d (21) Numerical integration is usually employed either with Gauss,
 J.B.M. Sousa Jr. et al. / Journal of Constructional Steel Research 66 (2010) 772779 775

1 kN/m 30
17.5 kN 17.5 kN
5
37.5 kN 37.5 kN
15

400

Fig. 3. Example 1: Beamcolumn and cross section.

GaussLobatto or midpoint rules in the so-called fibre method, 50

which in the plane case becomes a layer method.
This work employs the analytical evaluation of the resultant
40
forces and tangent stiffnesses of the cross sections, using an
algorithm described by Sousa Jr. and Muniz [20]. This algorithm
performs the exact evaluation of the area integrals, based on the 30

Load (kN)
application of Greens theorem, as long as the uniaxial stressstrain
relation is given by a polynomial expression. In its favour the exact
value of the integrals may be obtained in a reliable and fast manner. 20
On the other hand, the procedure shall be restricted to monotonic
loading as it is difficult to describe the evolution of the plastic
10
variables.

2.3. Geometric stiffness 0

0 5 10 15 20 25 30 35 40
The first term in the integral on the r.h.s. of Eq. (24) may be Displacement at midspan (cm)
written as
Fig. 4. Example 2: Membrane effect on beam.
(B T ) 

2 N2 + B 3 (M1 + M2 ) + B 4 S (31)


= B1 N1 + B This example has been analysed by Girhammar and cowork-
d =const
 d
where B i is the transpose of the ith line of B.
Performing the
derivatives only the two first terms in Eq. (31) are dependent on
d and one gets
` `
B T
Z Z
dx = BNL (N1 + N2 ) dx
0 d 0 Table 1 presents the results for displacement at midspan and
where
0 0 0 the present work, with meshes of 2, 4 and 8 elements. It may be
BNL = 0 0 0 . (33)
0 0 v,x v,x T
Eq. (32) may be identified as the geometric stiffness matrix. It
may be employed for the evaluation of linearised critical loads by
evaluation of the standard eigenvalue problem.
3. Applications
In this section a set of examples taken from the literature
is analysed with the proposed finite element scheme. The first
applications will focus on the assessment of the accuracy and
robustness of the element in the case of geometric nonlinearity
alone. Later the more general case of material and geometric
nonlinear effects is dealt with, emphasising the composite
steelconcrete beamcolumns.
3.1. Example 1: Beamcolumn with axial and transverse loads
A simply supported composite beam, with linear elastic
materials, loaded by a uniform transverse load and compressive
forces applied on the two components is depicted in Fig. 3. The
upper component is made of concrete with elastic modulus 12 GPa
and the lower component is made of wood with elastic modulus
8 GPa. The linear connection stiffness K is 50 MPa. The compressive
loads are applied at the centroids of the respective cross sections
and the total load is applied proportionally to each member area.
Nonlinear geometrical effects are present even for small values of
the applied load, as pointed out by Battini et al. [16].
ers [8,9] with their analytical solution. Their formulation takes into
account basically the same kinematical hypotheses as the present
numerical formulation.
Numerical results for the same structure were given by as
et al. [11], Krawkzyck and Rebora [15], and Battini et al. [16] with
(32) their respective finite element formulations.
for the maximum shear force at the interface Ksmax , for the closed-
form solution, for the numerical analyses by other authors, and for
seen that very good agreement for the displacement as well as for
the maximum shear force is already obtained for a coarse mesh
(2 elements only).
3.2. Example 2: Analysis of the membrane effect
An important aspect in the use of geometrical nonlinear
formulations for beams is the ability to capture the effects of
membrane action, which tends to stiffen a transversely loaded
beam when its supports are capable of reacting horizontally.
This effect is of particular importance in the analysis and design
for fire safety as it is interesting to take into account the large
displacements that the beam may undertake before failure.
In order to investigate the membrane effect, a beam with the
same geometry and materials as the one in the previous example,
but loaded only with a uniform transverse load, is considered here.
In each support the two beam components may be horizontally
restrained.
Initially the beam is analysed with just one restriction for
horizontal movement, to prevent a rigid body movement. In this
situation no membrane action is developed and the behaviour is
linear. Then the horizontal displacements are fixed only at the
lower component of the beam. In the third case all the horizontal
displacements are prevented. An increasing membrane action is
expected leading to stiffening of the beam.
Fig. 4 displays the loaddisplacement at midspan relationship
for the beam for three different support conditions. As expected,
when all the horizontal displacements are restricted the gain in
776 J.B.M. Sousa Jr. et al. / Journal of Constructional Steel Research 66 (2010) 772779

341
9
P
44 e
88

2400 38

Fig. 5. Example 3: Nonlinear buckling (dimensions in mm).

6
5
4
Displacement (mm)

3
2
1
0
1
2
3
4
5
0 0.5 1 1.5 2
Distance (m)

Fig. 6. Example 3: Beam deformed configuration.

Table 1
Example 1: Results for displacements and maximum connection force.
[8] [15] [16] 2 4 8

vmax (mm) 9.276 9.276 9.249 9.285 9.280 9.280

Fmax (kN/m) 13.878 13.865 13.798 14.143 13.920 13.890
Fig. 8. Example 4Cross section (dimensions in mm).

stiffness is pronounced. To obtain the solution, ten elements are Table 2

employed, with 4 integration points each. Example 3: Results for displacements at midspan in mm.
It is evident from the results that if there is horizontal restraint, e (mm) [21] [15] [16] Present
the membrane action may decrease significantly the deformability
11.50 5.2451 5.2387 5.2143 5.2451
of the beam. 13.25 0.3823 0.3823 0.3827 0.3824
15.00 4.4805 4.4738 4.4477 4.4803
3.3. Example 3: Nonlinear buckling
points for each element. Convergence was obtained within a single
In this example a two-layered beam with a deformable
load step with the standard NewtonRaphson scheme. Once again
connection is subjected to a compressive load P of 40 kN applied
an excellent agreement was observed between the formulations.
at different eccentricity levels of the lower member (Fig. 5). This
requires the member axis to be defined in a position different from
the centroid. The upper member has elastic modulus of 4.9 GPa 3.4. Example 4: Continuous composite steelconcrete beam (from [2])
and the lower member has an elastic modulus of 7.84 GPa. The
connection stiffness is 49 MPa. A symmetric two-span continuous steelconcrete composite
The lack of symmetry of the structure produces small lateral beam (Fig. 7) with nonlinear material behaviour and subjected
displacements even for small values of the compressive load. to uniform load was analysed by DallAsta and Zona [2] with a
Interesting deformation patterns occur when the eccentricity displacement approach and later with mixed elements [22].
e varies: The beam displacements are upward for e equal to These authors recognised this example as a difficult test
11.50 mm and downward for e equal to 15.0 mm, and an for composite beam elements, due to high slip gradient, strain
intermediate value of e equal to 13.25 mm produces a deformed localisation, concrete cracking over the support and softening in
configuration with three half waves. Fig. 6 displays the three the sagging region. The cross section geometry is shown in Fig. 8.
deflection patterns. Exact and numerical results are given in the Stressstrain relations for the materials and the connection
literature. Table 2 shows the values of the midspan displacements are shown in Fig. 9. The concrete uniaxial constitutive law is
for a second order analytical solution given by Kamiya [21], taken from the CEB-FIP Model Code [23]. The assumed concrete
along with the numerical results from [15,16], and the present peak stress is 33 MPa, and null strength is assumed for concrete
formulation with a mesh of ten elements and four integration under tension. Elastic-perfect plastic-hardening constitutive laws

30 m 30 m

Fig. 7. Example 4Continuous beam.

 J.B.M. Sousa Jr. et al. / Journal of Constructional Steel Research 66 (2010) 772779 777

0.0022 0.0039

Fig. 9. Example 4Constitutive and forceslip relations.

Fig. 10. Example 4Loaddisplacement curves.

are adopted for the beam steel and reinforcements bars. For the
beam steel fy is 275 MPa, fu 330 Mpa, sh is 4% and u 11%. For the
reinforcement fy is 430 MPa, fu 473 Mpa, sh is 2% and u 10%. The
connection force is given by the Ollgaard constitutive law

the behaviour of the beamcolumn, generating a much lower
S = Smax 1 e s
where Smax is 1240 kN/m, is 0.558, and is 1 mm . The 1 also reduces considerably. With equal to 1, the behaviour of the
connection ultimate displacement su is assumed as 6 mm.
In order to apply the analytical integration of cross section
forces and tangent moduli the expressions for the constitutive laws
for the concrete and beam steel were restated in terms of piecewise
polynomials. This process has been carried out by the authors in
a previous work and provided an excellent agreement with the
original results from [2,5].
The main purpose of this example is to investigate the influence
of the application of an axial load on the behaviour of the beam.
Without axial force, the beam presents a ductile behaviour with
large transverse displacements and reaches a maximum load q
close to 240 kN/m. A compressive load P with increasing values
qL, where L is the length of each span, was applied at the level
of the centroid of the concrete section. Eight finite elements were
employed in a single span, with three integration points each,
taking into account the symmetry of the structure.
Fig. 10 displays the results of the beam midspan displacement
with increasing load levels and five different values for . The loads
Fig. 12. Example 5Cross section (dimensions in mm).
are applied monotonically from zero to their maximum values, and
displacement control is employed to solve the nonlinear system of
equations. It may be seen that the presence of axial force changes
(34)
ductile loaddisplacement relationship. The load-carrying capacity
beam does not change a lot with respect to the beam without
axial load and they are able to sustain approximately the same
maximum transverse load. The main difference is that the beam
now is not able to reach large displacements. Increasing , the
beam gradually loses its load-bearing capacity and the maximum
displacements obtained are smaller.
For some fixed value of the transverse load, say 150 kN/m, it
may be seen that the displacement with equal to 4 is about
three times the displacement without axial load, which shows the
importance of second-order effects on this example.
3.5. Example 5: Continuous composite steelconcrete beam (from [4])
Salari and Spacone [4] presented the example of a two-span
steelconcrete composite beam with a concentrated load applied
at each mid-span (Fig. 11), with a cross section as depicted in
Fig. 12, and performed nonlinear material analyses employing

1675 mm 1675 mm

3350 mm 3350 mm

Fig. 11. Example 5Continuous steelconcrete beam with concentrated load.

778 J.B.M. Sousa Jr. et al. / Journal of Constructional Steel Research 66 (2010) 772779
(MPa)
Fig. 13. Example 5Uniaxial stressstrain and connection forceslip relations.
160
140
120
Load P (kN)
100
80
60 =0
=5
40 =10
=15
20
0 as a reliable and simple option for nonlinear numerical analysis of
0 0.01 0.02 0.03 0.04
Displacement (m)
Fig. 14. Example 5Loaddisplacement curves.
force and displacement-based finite elements. Later, Sousa Jr. and
Silva [24] employed interface and displacement-based elements in
the analysis of the same structure.
In the present example the beam will be subjected to different
levels of compressive load in order to assess the influence of the
compressive load in the overall resistance of the beam. Yam and
Chapman [25] reported an experimental ultimate load Pu equal to
148 kN, for the case without axial load.
The material uniaxial stressstrain relationships are depicted
in Fig. 13. In the present work polynomial approximations were
employed in order to evaluate the sectional properties analytically.
The stressstrain relation for concrete is parabolic-constant for
compression and a straight line for tension. The peak concrete
stress is reached at a strain equal to 0.0025, and the concrete strain
is conventionally limited to 0.0035.
Fig. 14 shows the results of the loaddisplacement behaviour
for a continuous beam without axial force and for values of axial
force equal to 5, 10 and 15 times the transverse load P. Eight finite
elements with four integration points each are employed for a
single span.
Once again the effect of the axial force is to reduce the load-
carrying capacity of the beam. Although in this example the
second-order effects are not very pronounced as the beam is not
slender, the presence of axial force has the effect of producing
much larger displacements for a fixed load level.
4. Summary and conclusions
This work presented the development of a displacement-
based finite element formulation for the analysis of compos-
ite beamcolumns with interlayer slip, or partial interaction.
Although this is a case for practical interest, very few works have
been published for the case of both geometrical and material non-
linearities.
(MPa) (N/mm)
(mm)
Departing from the EulerBernoulli beam assumptions in the
context of large displacements, moderate rotations and small
strains, the expressions for the element internal force and
tangent stiffness were developed within an adequate displacement
interpolation scheme.
As for the nonlinear material analysis, at the cross section level
resistant forces and tangent moduli were evaluated exactly using
an integration algorithm based on Greens theorem.
Nonlinear geometrical and material problems were solved by
means of the proposed finite element scheme. The results obtained
with this simple yet robust formulation were very accurate in
comparison with analytical and numerical results from different
finite element formulations. This qualifies the proposed element
0.05
composite beamcolumns with interlayer slip for civil engineering
applications.
Acknowledgements
The authors wish to thank brazilian agencies CNPq (Conselho
Nacional de Desenvolvimento Cientfico e Tecnolgico) and
FAPEMIG (Fundao de Amparo Pesquisa do Estado de Minas
Gerais) for the financial support.
References
[1] Newmark NM, Siess CP, Viest IM. Tests and analysis of composite beams with
incomplete interaction. Proc Soc Exp Stress Anal 1951;9(1):7592.
[2] DallAsta A, Zona A. Non-linear analysis of composite beams by a displacement
approach. Comput Struct 2002;80:221728.
[3] Gara F, Ranzi G, Leoni G. Displacement-based formulations for composite
beams with longitudinal slip and vertical uplift. Internat J Numer Methods
Engrg 2006;65(8):1197220.
[4] Salari MR, Spacone E. Finite element formulations of one-dimensional
elements with bond-slip. Eng Struct 2001;23:81526.
[5] DallAsta A, Zona A. Three-field mixed formulation for the non-linear analysis
of composite beams with deformable shear connection. Finite Elem Anal
Design 2004;40:42548.
[6] Ayoub A, Filippou FC. Mixed formulation of nonlinear steelconcrete
composite beam element. J Struct Eng 2002;126(3):37181.
[7] Ranzi G, Gara F, Leoni G, Bradford MA. Analysis of composite beams with
partial shear interaction using available modelling techniques: A comparative
study. Comput Struct 2006;84:93041.
[8] Girhammar UA, Gopu VKA. Composite beamcolumns with interlayer slip
Exact analysis. J Struct Eng, ASCE 1993;119(4):126582.
[9] Girhammar UA, Pan DH. Exact static analysis of partially composite beams and
beamcolumns. Int J Mech Sci 2007;49:13955.
[10] Xu R, Wu Y. Static, dynamic and buckling analysis of partial interaction
composite members using Timoshenkos beam theory. Int J Mech Sci 2007;
49:113955.
[11] as B, Saje M, Planinc I. Non-linear finite element analysis of composite planar
frames with interlayer slip. Comput Struct 2004;82:190112.
[12] Pi YL, Bradford MA, Uy B. Second order nonlinear inelastic analysis of
composite steelconcrete members. I: Theory. J Struct Eng, ASCE 2006;132(5):
75161.
[13] Pi YL, Bradford MA, Uy B. Second order nonlinear inelastic analysis of
composite steelconcrete members. II: Applications. J Struct Eng 2006;132(5):
76271.
 J.B.M. Sousa Jr. et al. / Journal of Constructional Steel Research 66 (2010) 772779
[14] Krawczyk P, Frey F, Zielisky AP. Large deflections of laminated beams with
interlayer slips Part 1: Model development. Eng Comput 2007;24(1):1732.
[15] Krawczyk P, Rebora B. Large deflections of laminated beams with interlayer
slips Part 2: Finite element development. Eng Comput 2007;24(1):3351.
[16] Battini JM, Nguyen QH, Hjiaj M. Non-linear finite element analysis of
composite beams with interlayer slip. Comput Struct 2009;87:90412.
[17] DallAsta A, Ragni L, Zona A. Analytical model for geometric and material
nonlinear analysis of externally prestressed beams. J Eng Mech 2007;133(1):
11721.
[18] Zona A, Ragni L, DallAsta A. Finite element formulation for geometric and
material nonlinear analysis of beams prestressed with external slipping
tendons. Finite Elem Anal Design 2008;44(15):9109.
[19] DallAsta A, Zona A. Slip locking in finite elements for composite beams with
deformable shear connection. Finite Elem Anal Design 2004;40:190730.
779
[20] Sousa Jr JBM, Muniz CFDG. Analytical integration of cross section properties
for numerical analysis of reinforced concrete, steel and composite frames. Eng
Struct 2007;29:61825.
[21] Kamiya F. Buckling strength of sheathed walls: Linear analysis. J Struct Eng,
ASCE 1987;113:200922.
[22] DallAsta A, Zona A. Comparison and validation of displacement and mixed
elements for the non-linear analysis of composite beams. Comput Struct 2004;
82:211730.
[23] CEB-FIP model code 1990. Final draft. CEB Bulletin dInformation vol. 203,
Comit Euro-International du Beton. Lausanne. 1991.
[24] Sousa Jr JBM, Silva AR. Nonlinear analysis of partially connected composite
beams using interface elements. Finite Elem Anal Design 2007;43:95464.
[25] Yam LCP, Chapman JC. The inelastic behavior of continuous composite beams
of steel and concrete. Proc Inst Civ Eng 1971;2:487501.