Você está na página 1de 7

Southern Cross University

ePublications@SCU
23rd Australasian Conference on the Mechanics of Structures and Materials

2014

1D frame element formulation for analysis of


layered composite beams
N Khorsandnia
University of New South Wales

H Valipour
University of New South Wales

S Foster
University of New South Wales

K Crews
University of Technology, Sydney

Publication details
Khorsandnia, N, Valipour, H, Foster, S, Crews, K 2014, '1D frame element formulation for analysis of layered composite beams', in ST
Smith (ed.), 23rd Australasian Conference on the Mechanics of Structures and Materials (ACMSM23), vol. II, Byron Bay, NSW, 9-12
December, Southern Cross University, Lismore, NSW, pp. 795-800. ISBN: 9780994152008.

ePublications@SCU is an electronic repository administered by Southern Cross University Library. Its goal is to capture and preserve the intellectual
output of Southern Cross University authors and researchers, and to increase visibility and impact through open access to researchers around the
world. For further information please contact epubs@scu.edu.au.

23rd Australasian Conference on the Mechanics of Structures and Materials (ACMSM23)


Byron Bay, Australia, 9-12 December 2014, S.T. Smith (Ed.)

1D FRAME ELEMENT FORMULATION FOR ANALYSIS


OF LAYERED COMPOSITE BEAMS
N. Khorsandnia*
School of Civil and Environmental Engineering, The University of New South Wales (UNSW), NSW,
2052, Australia. N.Khorsandnia@unsw.edu.au (Corresponding Author)
H. Valipour, S. Foster
School of Civil and Environmental Engineering, The University of New South Wales (UNSW), NSW,
2052, Australia. H.Valipour@unsw.edu.au, S.Foster@unsw.edu.au
K. Crews
School of Civil and Environmental Engineering, University of Technology, Sydney (UTS), NSW,
2007, Australia. Keith.Crews@uts.edu.au

ABSTRACT
In this paper a numerical model for non-linear analysis of layered composite beams is developed. It
takes advantage of an efficient 1D frame element with forced-based formulation. The developed
element is equipped with lumped translational springs at the nodal points to model the partial shear
interaction between the layers. Also, the proposed FE model can capture the material non-linearities as
well as non-linear shear-slip behaviour of connections between the layers. The sections of frame
element are discretised into fibres and the stiffness matrix of each section can be determined by
integrating the stiffness of fibres over the section depth based on Navier-Bernoulli hypothesis. The
superior performance of the developed FE model for capturing the behaviour of composite beams was
demonstrated by comparing the numerically predicted results with experimental data. The proposed
model can efficiently capture the global response (i.e. load vs deflection, slip, strain and stress) of the
composite beams up to failure with adequate accuracy.
KEYWORDS
1D frame FE model, composite elements, deformable connections, lumped non-linear springs.
INTRODUCTION

The timber-concrete composite (TCC) represents a construction technique which consists of


connecting an existing or new timber beam to a concrete slab by means of a connection system.
The full-scale experimental set-up for TCC and timber-timber composite (TTC) beams is typically
time consuming and costly. Accordingly, developing numerical methods such as finite element (FE)
analysis can be a good alternative for predicting the short- and long-term behaviour of TCC members.
To date, a wide range of FE models with different levels of accuracy and complexity have been
developed. The existing FE models for analysis of TCC beams can be generally categorised as
continuum-based (Bou Sa
d et al. 2004; Dias et al. 2007; Gutkowski et al. 2010) and discrete frame
elements (Fragiacomo 2005; Fragiacomo & Ceccotti 2006; Lukaszewska et al. 2010), which have
different domains of applicability. The continuum-based models can be employed for detailed analysis
considering local effects; however, they are time-demanding for capturing global behaviour of
structures. Frame elements offer sufficient accuracy and efficiency for overall response of composite

This work is licensed under the Creative Commons Attribution 4.0 International License. To view a copy of this license, visit
http://creativecommons.org/licenses/by/4.0/

795

beams and many authors have formulated various linear and non-linear models over the last few
decades (Ayoub & Filippou 2000; Heidarpour & Bradford 2009; Ranzi & Bradford 2006).
The main difference of the developed composite frame elements is the shear interaction between the
layers which can be provided via two main mechanisms; i.e. mechanical fasteners and friction. To the
best of the authors knowledge, no connection with full composite action has not been reported in the
literature for TCC and TTC members and therefore, partial interaction between the layers should be
considered (Ceccotti et al. 2006; Deam et al. 2008; Yeoh et al. 2011). Furthermore, the shear
interaction can be treated as continuous (e.g. glued or mesh connections) or discrete (e.g. mechanical
fasteners). Many frame FE models with the assumption of continuous shear transfer have been
proposed (Fragiacomo 2005; Fragiacomo & Ceccotti 2006; Schnabl & Planinc 2011); however,
developing multi-layered composite frame model with discrete connections for analysis of TCC/TTC
is still missing in the literature. The assumption of discrete connection can better represent the global
behaviour of TCC beams with mechanical connections such as nails, screws and notches.
The frame FE models can be formulated based on displacement, force and a combination of them (i.e.
mixed). Although displacement-based formulations are easier to derive, they may not be as efficient as
force and mixed formulation, since more number of elements is needed when a same level of accuracy
is required. The reason is that the equilibrium equations in force or mixed formulations are always
satisfied in exact sense and they are typically locking free (Ayoub & Filippou 2000; Valipour &
Bradford 2009).
In this paper, a 1D frame FE model for analysis of layered composite beams is developed. For
capturing the shear-slip behaviour between the layers, lumped non-linear springs in both ends of the
element are employed. The element is cast within the scope of force-based formulation using fibre
elements. The accuracy of the proposed model for analysis of TCC beams is verified against
experimental data from literature. The proposed model is efficient and simple and can be used for nonlinear analysis of layered composite beams, especially when time consuming analyses such as longterm and/or parametric studies are of concern.
ELEMENT FORMULATION
A composite element AB (see Figure 1a) with four nodes, eight degrees of freedom and two non-linear
springs with different stiffness at both ends (KA and KB) is considered. Based on the equilibrium
equations for free body length of Ax (see Figure 1b) in conjunction with cantilever configuration
clamped at end B, the section generalised forces, D(x), can be expressed as
D( x) b( x) Q A K CA q A DA ( x) DcpA

(1)

where QA and qA are generalised nodal forces and displacements, respectively; b(x) is the force
interpolation matrix; KCA, D*A (x) and DcpA are the matrix containing the stiffness of connector, section
forces due to element loads and section forces due to plastic slip in connector at end A, respectively.

(a)
(b)
(c)
Figure 1. (a) frame element AB with length l, (b) free body of length Ax., (c) discretisation of section
into fibres and strain distribution over the depth
The compatibility equation with the assumption of deformable shear connection according to NavierBernoulli hypothesis (see Figure 1c) can lead to
xi ri yi
(i 1, 2)
(2a)

ACMSM23 2014

796

b r 2 r1 H

(2b)
where xi, ri, and b represent the total strain, strain at the mid-plane of the layer, curvature of section
and slip strain, respectively.
The constitutive law based on the total secant approach can be written as
(3)
xi Eei exi Eei ( xi pxi )
(i 1, 2)
where xi, Eei, exi and pxi are the total stress, elastic secant modulus, elastic and plastic strains at the
monitoring point, respectively. The subscripts 1 and 2 denote the corresponding layer.
Considering equilibrium over the section and using Equations (2a) and (3) gives
D( x) k s ( x) d( x) D p ( x)

(4)

where ks(x), Dp(x), and d(x) = [r1 r2] are the secant stiffness matrix, residual plastic force and
section generalised strains, respectively. Furthermore, the compatibility condition with the assumption
of small strains leads to
T

q A bT ( x) d( x) dx

(5)

Rearranging Equation (4) with consideration of Equations (1) and (5) gives

I B A q A FAA Q A q*A q pA qcpA

(6)
q*A,

where BA is the stiffness matrix due to stiffness of the connector at end A;


qpA and qcpA represent
the nodal generalised deformation due to element loads, plastic strains and plastic slip in the connector
at end A, respectively; and I is identity matrix.
Pre-multiplying both sides of Equation (6) by KAA = [FAA]-1, yields
K eAA q A Q A Q*A Q pA QcpA
e

(7)
Q*A,

in which K AA is the stiffness sub-matrix of end A for the element AB shown in Figure 1a;
QpA and
QcpA are the nodal generalised force due to element loads, plastic strains and plastic slip in the
connector at end A, respectively.
Employing global equilibrium of the element as well as consideration of and T as transformation
matrices, can lead to
K eBA q A Q B Q*A Q pA

(8)

in which K BA = K AA + T KCA and QB is the nodal generalised force at end B.


e

Similar equations for cantilever frame element clamped at end A can be obtained and the governing
equation of the element can generated as
K eAA K eAB q A Q A Q*A Q pA QcpA
(9)
e

*
e
K BA K BB q B Q B Q B Q pA QcpB
A direct iteration solution scheme can be used for solving the Equation (9) (Valipour & Bradford 2009;
Valipour & Foster 2009). It is noteworthy that for elements with more than two layers, corresponding
longitudinal (i.e. along the element axis) degrees of freedom should be considered and the formulation
can be readily derived with a similar approach as above. Further details of the formulation can be
found in Khorsandnia et al. (2014).
ADOPTED MATERIAL MODEL
A uniaxial constitutive law for different components of TCC beam (timber, concrete and connection)
is adopted. Among the available constitutive laws for timber under compression loads, Gloss model
(1981) is adopted in this paper (see Figure 2a). The strain corresponding to ultimate compressive
strength and ultimate strain are assumed as c0 0.004 and cu 0.012 , respectively. In tension, a
linear elastic response up to tensile strength followed by a linear softening branch is adopted. It should

ACMSM23 2014

797

be noted that for different mesh sizes, the slope of softening branch is adjusted by using constant
fracture energy.
For concrete under compression, a uniaxial stress-strain relationship based on the CEB-FIP model
code 1990 (CEB-FIP 1993) is employed for ascending branch which is followed by a linear softening
part after compressive strength. Furthermore, a linear quasi-brittle failure model with exponential
softening curve is adopted for tension (see Figure 2b). In this study, the strain corresponding to
compressive strength and the ultimate compressive strain are taken as c0 0.002 and cu 0.01 ,
respectively. It is noteworthy that for unloading/reloading regime, a damage model with no plastic
strains under compression and tension is assumed for both timber and concrete (see Figure 2a and b).
The shear-slip behaviour of springs at both ends can be different and can be obtained from push-out
tests. Different types of connections have been tested in the literature and various models have been
developed (Yeoh et al. 2011). Amongst them, the one with non-linear pre-peak response followed by a
linear post-peak with elasto-plastic unloading/reloading regime is adopted in this study (see Figure 2c).

(a)
(b)
(c)
Figure 2. Schematic outline of adopted (a) stress-strain relationship for timber, (b) stress-strain
relationship for concrete, (c) shear-slip relationship for connection
ERROR ESTIMATION
To evaluate the accuracy of the proposed model, the convergence rate for an arbitrary cantilever beam
(see Figure 3a) is obtained. Based on the obtained results, the maximum error occurs when the
stiffness of the spring is infinity, whereas there is no error in the case of free slip between the layers
(i.e. stiffness of spring is zero). Accordingly, the error of the vertical displacement at end A (q2) is
demonstrated in Figure 3b with respect to the number of elements used in the model for different
values of Ks/Kb (i.e. spring to beam stiffness). Furthermore, the value of Ks/Kb for TCC/TTC beams is
normally less than 200 for maximum length of 2 m. Therefore, the developed model has sufficient
accuracy for the beams with low level of composite action; however, more number of elements (6 to
10 elements) should be employed for the beams with near full-composite behaviour.

(a)
(b)
Figure 3. (a) Outline of the cantilever configuration adopted for error estimation, (b) error in vertical
displacement (q2) versus number of elements for different values of Ks/Kb

ACMSM23 2014

798

VERIFICATION OF DEVELOPED MODEL


For verification of the developed FE model, a short-term test results reported in the literature is
employed. The experimental program includes a 6 m simply-supported TCC beam with glued re-bar
connections and was conducted at University of Florance (Italy) by Ceccotti et al. (2006). During the
test, different characteristics of the beam such as vertical deflection at mid span, maximum slips over
the supports and strains at mid span were recorded. The geometrical details of the beam as well as the
material properties taken from the test data are given in Table 1.
The numerical predictions from the model in this study and FE model developed by Ceccotti et al.
(2006) are compared with experimentally measured data and shown in Figure 4a to d, which
respectively, include load versus mid span deflection, maximum slip over the support, upper and lower
timber stress and upper concrete stress at mid span. Based on the obtained results, the frame FE model
developed in this study can adequately capture the short-term behaviour of the TCC beam and in the
meantime, it has similar accuracy with the model developed by Ceccotti et al. (2006).
Table 1. Geometrical details and adopted material properties of the TCC beam (Ceccotti et al. 2006)
Timber
Load
Span
Dist.
E
fcu
ftu
bh
(m)
(m) (mmmm) (GPa) (MPa) (MPa)
5.7 1.8 (2)125500 10
40
29

Concrete
E
fcu
ftu
bh
(mmmm) (GPa) (MPa) (MPa)
3.3
1500 30.4 30.4

(a)

Connection
Load-slip
No.
Relationship
-0.7s 0.9
192 P= 40(1-e )

(b)

(c)
(d)
Figure 4. Verification of developed FE model against test and FE results from Ceccotti et al. (2006); (a)
load vs deflection, (b) load vs slip, (c) load vs stress in top and bottom of timber, and (d) load vs stress
in top of concrete
CONCLUSIONS
This paper deals with developing a 1D frame FE model for non-linear analysis of TCC/TTC beams.
The FE model is formulated in the framework of the force interpolation and takes advantage of nonlinear discrete connections at the nodal points for modelling the partial interaction between the layers.
Each section of the element is divided into fibres (i.e. integration points) and the shear deformations

ACMSM23 2014

799

are assumed to be negligible (validity of Navier-Bernoulli hypothesis). The maximum and minimum
error of the model with respect to spring stiffness is obtained and good convergence rate is observed. It
seems when the TCC beam has high composite action (stiff connections), using at least 6 elements
over the beam length can provide adequate results. Furthermore, the numerical predictions are
validated against experimental results reported in the literature and it is concluded that the proposed
model can sufficiently capture the short-term response of TCC/TTC beams up to failure. The
significance of the model is its simplicity and efficiency which can be a great advantage for timedemanding procedures such as design oriented parametric studies and/or long-term analysis.
REFERENCES
Ayoub, A. & Filippou, F.C. (2000) "Mixed formulation of nonlinear steel-concrete composite beam
element", Journal of Structural Engineering, Vol. 126, No. 3, pp. 371-81.
Bou Sad, E., Jullien, J.F. & Ceccotti, A. (2004) Long term modelling of timber-concrete composite
structures in variable climates, 8th World Conference on Timber Engineering (WCTE), Lahti,
Finland, 14-17 June.
CEB-FIP (1993) CEB-FIP model code 1990: Design code, Thomas Telford, London.
Ceccotti, A., Fragiacomo, M. & Giordano, S. (2006) "Long-term and collapse tests on a timberconcrete composite beam with glued-in connection", Materials and Structures, Vol. 40, No. 1, pp.
15-25.
Deam, B.L., Fragiacomo, M. & Buchanan, A.H. (2008) "Connections for composite concrete slab and
LVL flooring systems", Materials and Structures, Vol. 41, No. 3, pp. 495-507.
Dias, A.M.P.G., Van de Kuilen, J.W., Lopes, S. & Cruz, H. (2007) "A non-linear 3D FEM model to
simulate timber-concrete joints", Advances in Engineering Software, Vol. 38, No. 8-9, pp. 522-30.
Fragiacomo, M. (2005) "A finite element model for long-term analysis of timber-concrete composite
beams", Structural Engineering and Mechanics, Vol. 20, No. 2, pp. 173-90.
Fragiacomo, M. & Ceccotti, A. (2006) "Long-term behavior of timber-concrete composite beams. I:
finite element modeling and validation", Journal of Structural Engineering, Vol. 132, No. 1, pp.
13-22.
Glos, P. (1981) Zur Modellierung des Festigkeitsverhaltens von Bauholz bei Druck-, Zug- und
Biegebeanspruchung, Berichte zur Zuverlssigkeitstheorie der Bauwerke, SFB 96, Munich,
Germany.
Gutkowski, R.M., Balogh, J. & To, L.G. (2010) "Finite-Element Modeling of Short-Term Field
Response of Composite Wood-Concrete Floors/Decks", Journal of Structural Engineering, Vol.
136, No. 6, pp. 707-14.
Heidarpour, A. & Bradford, M.A. (2009) "Generic non-linear modelling of a bi-material composite
beam with partial shear interaction", International Journal of Non-Linear Mechanics, Vol. 44, No.
3, pp. 290-7.
Khorsandnia, N., Valipour, H., Foster, S. & Crews, K. (2014) "A force-based frame finite element
formulation for analysis of two- and three-layered composite beams with material non-linearity",
International Journal of Non-Linear Mechanics, Vol. 62, pp. 12-22.
Lukaszewska, E., Fragiacomo, M. & Johnsson, H. (2010) "Laboratory Tests and Numerical Analyses
of Prefabricated Timber-Concrete Composite Floors", Journal of Structural Engineering, Vol.
136, No. 1, pp. 46-55.
Ranzi, G. & Bradford, M. (2006) "Analytical solutions for the time-dependent behaviour of composite
beams with partial interaction", International Journal of Solids and Structures, Vol. 43, No. 13,
pp. 3770-93.
Schnabl, S. & Planinc, I. (2011) "The effect of transverse shear deformation on the buckling of twolayer composite columns with interlayer slip", International Journal of Non-Linear Mechanics,
Vol. 46, No. 3, pp. 543-53.
Valipour, H.R. & Bradford, M.A. (2009) "A steel-concrete composite beam element with material
nonlinearities and partial shear interaction", Finite Elements in Analysis and Design, Vol. 45, No.
12, pp. 966-72.
Valipour, H.R. & Foster, S.J. (2009) "Nonlocal damage formulation for a flexibility-based frame
element", Journal of Structural Engineering, Vol. 135, No. 10, pp. 1213-21.
Yeoh, D., Fragiacomo, M., De Franceschi, M. & Buchanan, A.H. (2011) "Experimental tests of
notched and plate connectors for LVL-concrete composite beams", Journal of Structural
Engineering, Vol. 137, No. 2, pp. 261-9.
ACMSM23 2014

800

Você também pode gostar