Exact Shape Functions For Timoshenko Beam Element: IOSR Journal of Computer Engineering May 2017
Exact Shape Functions For Timoshenko Beam Element: IOSR Journal of Computer Engineering May 2017
net/publication/317546589
CITATIONS READS
9 6,977
5 authors, including:
Shota Kiryu
Nihon University
7 PUBLICATIONS 36 CITATIONS
SEE PROFILE
All content following this page was uploaded by Han Ay Lie on 02 August 2017.
Abstract: This paper derives exact shape functions for both non-uniform (non-prismatic section) and
inhomogeneous (functionally graded material) Timoshenko beam element formulation explicitly. In this paper,
the shape functions formula embedded the explicit functions and its derivatives describing the non-uniformity
and inhomogeneity of a beam element. The shape functions are made interdependent by requiring them to
satisfy three homogeneous differential equations associated with the Timoshenko’s beam theory. With the
formulated axial, transverse and rotational displacement shape functions, the stiffness and mass matrices and
consistent force vector for a two-node Timoshenko beam element are developed based on Hamilton’s principle.
Comparison studies with reference work on the accuracy and computational efficiency for non-uniform and
inhomogeneous Timoshenko beam structures are highlighted. Static and vibrational analyses of the beams
element by using the exact shape functions can predict the displacement, and natural frequencies of non-
uniform and inhomogeneous Timoshenko beams by using only one/the least element accurately.
Keywords : Exact shape functions, Timoshenko beam, FEM, Functionally graded material
I. Introduction
In the state-of-the-art of computational beam element, the polynomial function assumptions are most
usually used. However, these functions were developed base on the uniform cross-section and homogenous
material. By using these shape functions, the analyses of beam elements require a significant number of element
divisions to study the behavior of beam with non-uniformity and inhomogeneity problems accurately [1-12].
The latest accomplishment for developing the exact solutions of the shape functions makes use of the power
series method [13-14], basic displacement function [15] and Taylor series expansion [16-17] approximations are
reported.
Given the fact that the finite element method has been a widely used technique for structural analysis, it
is necessary to develop reliable and accurate displacement functions [18-26] for the beam element. With the
proper shape functions in the finite element formulations, the mass, stiffness matrices and loading vectors can be
constructed consistently. In this paper, the formulation for exact shape functions for both non-uniform and
inhomogeneous Timoshenko beam element are derived based on Hamilton’s principle.
w
G ( x) A( x) 0 (1b)
x x
w
E ( x) I ( x) G ( x) A( x) 0 (1c)
x x x
where, u, w and are the axial displacement, transversal displacement and rotation. E( x), A( x), G( x) and are
the elastic modulus, section area and shear modulus as the function of x and shear correction factor, respectively.
By integration, (1a) becomes,
u
E ( x) A( x) k1 (2)
x
where k1 is the indefinite integration integral constant. Next, solving for u(x),
1
u ( x) k1 dx k2 k1 1 k2 (3)
E ( x) A( x)
defining,
1
1 dx . (4)
E ( x) A( x)
Imposing the boundary conditions of horizontal displacements at both end nodes of the beam results in,
u( x 0) u1 , u( x L) u2 .
Solving for k1, k2 from (2) and (3), then substituting back the results into (2), the exact displacement
functions and its first order derivative of the axial displacement of the beam can be obtained as
u
u ( x ) N u1 Nu 2 1
u2
1 1L
N u1
10 1L
1 10
Nu 2 (5)
10 1L
dN u1 1
dx 10 1L
dN u 2 1
dx 10 1L
where 10 1 , 1L 1 x L .
x 0
w
G ( x) A( x) c1
x
w( x) c1 4 ( x)dx c4 (7)
1 w
( x) c1
G ( x) A( x) x
Replacing (6) into (7), results in
a3 2
( x) c1 4 a1 a2 x x
2
c1 4 a2 a3 x (8)
x
2
c1 4 a3
x 2
w( x) c1 4 ( x)dx c4
a
c1 4 c1 4 a1 a2 x 3 x 2 dx c4 (9)
2
a a
a1 x 2 x 2 3 x 3 c4
2 6
Hence, (1c) can be written as
E ( x) I ( x) c1 0
x x
x 1
c1 c2 c1 3 c2 22 (10)
x E ( x) I ( x) E ( x) I ( x)
( x) c1 3 c2 2 c3
the second derivative can be given by,
2
c1 3 c2 2
x 2
where ci (i=1,2,3,4) and ai (i=1,2,3,4) are the polynomial constants. Defining,
1
2 dx,
E ( x) I ( x)
x
3 dx, .
E ( x) I ( x)
1
4 dx
G( x) A( x)
Collecting rotation and its derivatives from (8) and (10), as well as transverse displacement between (6)
and (9), we have relationship between ai and ci
Imposing the boundary conditions of vertical displacements and rotations at both end nodes of the
beam results in, w( x 0) w1 , w( x L) w2 and ( x 0) 1 , ( x L) 2 .
c1 w1
c
2 1 1
M wθc (12)
c3 w2
c4 2
By substituting ci, the expressions for transverse and rotational shape functions are as follow,
w1
w( x) N w1 N w2 N w3 N w4 1
w2
2
(13)
w1
( x ) N 1 N 2 N 3 N 4 1
w2
2
where the rotational shape functions and their derivatives are given by
1
N 1 6( 2 L ( 3 30 ) 20 ( 3 3L ) 2 ( 30 3L ))
L
1 ( 3 3 L ) L(3 2 L 2L L) 2 (6 4 L L(3 3L 3 4L
N 2
3L L 4L L)) 2 L (6 4 L L(3 3L 3 4L 3L L 4L L))
N 3 N 1 (14)
6 2 L ( 3 30 ) 6 2 30 6 2 3 L 6 2 4 L 3 2 L 3 L
2
1 3 2 L 30 L 3 2 3L L 3 2 4L L 2L 3 L 2L 30 L
2
N 4
2 3L L2 2 4L L2 20 (6 3 6 3 L 6 4 L 3 3L L
3 4L L 3L L2 4L L2 )
3 4x 6 3 6 4
1 x( 20 2 L ) 2
6 x( 20 3 L 2 L 30 )
N w1 3 3 x 3x 4x
2
L
( 30 3L )(6 2 x 3
2 x 2
2 x 3
)
3 2 L 3 L 3 2 L 3 L L
2
6 4 L 3 3L L 3 2 L 4 L 2L 3 L
(3 x x 2 ) 3 L L2 6 x L2 L2
1 2 L 3L
2 2 4L 3L 2L 4
N w2 L2 (6 L (3
6 4L
2 4L 3L
3 4L 3L L 4L L)
6 4 L L(3 3 L 3 4 L
3 3 L 3 4 L
6 2 L L L)
x(3 2 L 2 L )
3L 4L 3 L x 4L x
1 x(( 30 3 L )(6 2 3 2 x 2x ) ( 2 L 20 )(6 3 6 4 3 3 x
2
N w3
L 3 4x 3x 2 4x 2 ) 6( 20 3 L 2 L 30 )
6 30 6 3 L 6 4 L 6 3 L 6 30 L
2
(6 2 2x ) 3 3L L 3 4L L (3 2 L 2L ) 6 4 L 3 3L x
L2 L2 3 x x 2 x 2
x 3L
4L 4L 3L 4L
N w4
6 6 4 L 3 3L L 6 3 6 3 L 6 4
(3 2 x 6 20 ) 3 4L L 6( 2 L 20 ) 3 3 x 3 4x (15)
L2 L2 x 2 x 2
3L 4L 3 4
N w1 1
6( 2 ( 30 3L ) 20 ( 3 3L 4 ) 2 L ( 3 30 4 ))
x L
6 4 L L(3 3L
N w 2 1 ( 3 3 L 4 ) L(3 2 L 2L L) 2
3 4L 3L L 4L L)
x
2 L 6 4 L L(3 3L 3 4L 3L L 4L L)
(17)
N w3 N
w1
x x
(6 2 L 6 20 2L )( 3 30 4 )
N w 4 1
( 2 20 ) 6 4 L 3 3L L 3 4L L 3 3L L2 4 4L L2
x
6 2 ( 30 3 L ) 3 2 L 3 L 30 L 4 L
where,
( 30 3 L ) L(3 2 L 2L L)
2 L (6 4 L L(3 3L 3 4L 3L L 4L L))
20 (6 4 L L(3 3L 3 4L 3L L 4L L))
20 2 x 0 , 2 L 2
2 x 0 , 2 L 2 x L , 20 xL
2
, 20 x 0 , 2L 2 xL
30 3 x 0 , 3L 3
3 x L , 30
x 0 , 3L 3
x L , 30 3
x 0 , 3 L 3 xL
40 4 x 0 , 4L 4 xL
4
, 40 x 0 , 4 L 4 xL
4
, 40 x 0 , 4L 4 xL
H S E K E WE dt 0
t2
(18)
t1
where, H , SE , KE and WE are the variation of total energy, strain energy, kinetic energy and external
work of the beam, respectively.
The undamped equilibrium equation can be obtained by substituting (5) and (13) into the variational
equations of (18), which results in
Kd Md f 0 (19)
where, K, M , f , d and d are the stiffness matrix, mass matrix, loading vector, general displacement vector and
general acceleration vector, respectively. The stiffness matrix, mass matrix and loading vector are given by
Nu Nu
T
E ( x) A( x) 0 0
N N dx
K
L
v
N N 0 G ( x ) A( x ) 0 v (20)
N N
0
0 0 E ( x) I ( x)
( x) A( x)
T
Nu 0 0 Nu
N dx
M
L
Nv 0 ( x) A( x) 0 v (21)
N ( x) I ( x) N
0
0 0
Nu1 P1 Nu1 p( x)
N Q N q( x)
v1 1 v1
N 1 M1
L N 1 m( x)
f 0 dx (22)
Nu 2 P2 Nu 2 p( x)
N v 2 Q2 N v 2 q( x)
N 2 M 2
N 2 m( x)
where, Pi , Qi and M i are the concentrated axial load, transversal load and moment at node i, respectively. The
terms p( x), q( x) and m( x) are the distributed axial load, transversal load and moment, respectively.
In constructing the matrices and vector of (20-22), the Gaussian quadrature integration scheme is
recommended. Also, the shape functions 1 , 2 , 3 , 4 and their derivatives inside the equations can be
evaluated numerically by using the Gaussian quadrature integration scheme. These shape functions will carry on
the non-uniformity and inhomogeneity characteristics of the beam.
P=1
E=100
0.75 G=40 0.3
κ=5/6
0.2
0.2 L=1
Figure 1. Taper cantilever beam example
The beam is analyzed by using one element by using the exact displacement functions to construct the
stiffness matrices of the static equilibrium equation.
As shown in Table 1., the reaction forces and displacement resulting from the shear deformation
consideration are in good agreement with the theoretical results of [25].
P=1, w
E=100
G=40 0.3 h
κ=5/6
L L
0.2
Figure 2. Clamped taper beam example
Deflections which include shear deformation (ws) and neglect shear deformation (wns) at the mid-span
of the beams are listed in Tables 2, respectively. The present computed deflections are compared with the results
of [28].
Only using two element divisions, the present results gave an excellent agreement with the reference
(twelve elements) which can be observed from the table.
h0 h1
b
E=210 Gpa
b a) Clamped-Free (C-F)
E/(κG)=3.12
Κ=2/3
h0 h1
L=0.8660254 m
b
b b) Clamped-Clamped (C-C) ρ=7850 kg/m3
r2=I0/(A0L2)=0.08
h0 h1
b
b c) Clamped-Pinned (C-P)
L
Figure 3. Various boundary conditions of taper beam free vibration example
The cross-session is rectangular with constant width b and linear-varying height h defined as
h1 h0
(23)
h0
A0 L4
i i (24)
E I0
From Table 3, the results of the non-dimensional frequencies of the present formulation show lower
convergence results under various boundary conditions compared to the results in [15]. In the free vibration
problem, the beam element should be divided into sub-elements to accommodate the free degree of freedoms. In
this study, only four element divisions for the beam were employed to obtain the good results.
Table 3. Non-dimensional natural frequency of tapered beam with various boundary conditions
β Boundary µ1 µ2 µ3 µ4 µ5
Condition
0.0 C-F Present 3.22720 14.47780 31.50240 48.19680 62.87280
[15] 3.22713 14.46893 31.50254 47.90902 62.34702
C-P Present 11.08680 27.17590 45.11830 59.44520 63.59720
[15] 11.08250 27.11438 44.84353 59.20303 63.33950
C-C Present 13.83890 28.58680 45.95580 61.42410 68.86960
[15] 13.83476 28.51793 45.66595 61.86205 68.28361
-0.2 C-F Present 3.34530 14.32890 30.91900 48.13030 65.05240
[15] 3.33065 14.28921 30.71080 47.75021 64.99695
C-P Present 10.75020 26.31190 44.03880 62.05360 68.56940
[15] 10.68689 26.10717 43.59072 61.65596 68.42075
C-C Present 13.31860 27.95770 45.12020 61.90550 72.36990
[15] 13.22227 27.77822 44.69713 61.80658 72.55473
x
E ( x) ( Eleft Eright ) 1 Eright
L
(25)
x
G ( x) (Gleft Gright ) 1 Gright
L
The Simply-Supported (S-S) rectangular - cross-section beam with geometric data (width b=0.4 m;
height h=0.9 m and the total length L=20 m) is employed in the computation of fundamental frequencies.
The results (with only one element division) are compared with [12] which is computed with ten
element divisions. Very good agreement between the frequencies obtained in the present work with that of the
reference is given in Table 4 where Er=Eleft/Eright; μ1, μ2 are respectively the first and the second non-
dimensionalized fundamental frequencies of the beam. The non-dimensionalized fundamental frequency is
given by
AL4
i i (26)
Eleft I
A,I A,I
left right
L
Figure 4. Simply supported rectangular FGM beam example
V. Conclusions
Exact shape functions of a non-uniform and inhomogeneous Timoshenko beam element has been
formulated. These shape functions can be implemented in finite-element codes to create the mass and stiffness
matrices within the context of Timoshenko beam element where the shear deformation can be taken into account
in the analyses. The correctness of the formulated shape functions was verified through the numerical examples
of the static and free vibration analyses of non-uniform and inhomogeneous (FGM) beam elements. With the
formulated shape functions, highly accurate results can be obtained by using the least element number of
element division.
References
[1] D.K. Nguyen, B.S. Gan and T.H. Le, Dynamic Response of Non-Uniform Functionally Graded Beams Subjected to a Variable
Speed Moving Load, Journal of Computational Science and Technology, 7(1), 2013, 12-27.
[2] D.K. Nguyen and B.S. Gan, Large deflections of tapered functionally graded beams subjected to end forces, Applied Mathematical
Modelling, 38(11-12), 2013, 3054-3066.
[3] D.K. Nguyen, B.S. Gan and T.H. Trinh, Geometrically nonlinear analysis of planar beam and frame structures made of functionally
graded material, Structural Engineering and Mechanics, 49(6), 2014, 727-743.
[4] T.H. Le, B.S. Gan, T.H. Trinh and D.K. Nguyen, Finite element analysis of multi-span functionally graded beams under a moving
harmonic load, Mechanical Engineering Journal, 1(3), 2014, 1-13.
[5] B.S. Gan and D.K. Nguyen, Large Deflection Analysis of Functionally Graded Beams Resting on a Two-Parameter Elastic
Foundation, Journal of Asian Architecture and Building Engineering, 13(3), 2014, 649-656.
[6] D.K. Nguyen, B.S. Gan, T.H. Trinh, and S. Alexandrov, Post-buckling Behavior of Elastic-Plastic Functionally Graded Beams
Subjected to Eccentric Axial Load, Journal of Multidisciplinary Engineering Science and Technology, 2(5), 2015, 1129-1135.
[7] A.L. Han, B.S. Gan, S. As'ad and M.M.A. Pratama, Parametric Study of the Load Carrying Capacity of Functionally Graded
Concrete of Flexural Members, International Journal of Engineering and Technology Innovation, 5(4), 2015, 233-241.
[8] T.H. Trinh, D.K. Nguyen, B.S. Gan and S. Alexandrov, Post-buckling responses of elastoplastic FGM beams of nonlinear elastic
foundation, Structural Engineering and Mechanics, 58(3), 2016, 515-532.
[9] D.K. Nguyen, T.T. Tran, B.S. Gan and V.T. Bui, Influences of Dynamic Moving Forces on the Functionally Graded Porous-
Nonuniform Beams, International Journal of Engineering and Technology Innovation, 6(3), 2016, 173-189.
[10] T.H. Trinh, V.T. Bui, N.H. Nguyen, D.K. Nguyen and B.S. Gan, Dynamic Behavior of Functionally Graded Beams in Thermal
Environment due to a Moving Harmonic Load, International Journal of Mechanical Systems Engineering, 2(119), 2016, 1-8.
[11] B.S. Gan, D.K. Nguyen and T.H. Le, Effect of Intermediate Elastic Support on Vibration of Functionally Graded Euler-Bernoulli
Beams Excited by a Moving Point Load, Journal of Asian Architecture and Building Engineering, 16(2), 2017, 363-369.
[12] B.S. Gan, T.H. Trinh, T.H. Le and D.K. Nguyen, Dynamic response of non-uniform Timoshenko beams made of axially FGM
subjected to multiple moving point loads, Structural Engineering and Mechanics, 53(5), 2015, 981-995.
[13] A.Y.T. Leung, W.E. Zhou, C.W. Lim, R.K.K. Yuen and U. Lee, Dynamic stiffness for piecewise non-uniform Timoshenko column
by power series - part I: Conservative axial force, International Journal for Numerical Methods in Engineering, 51(5), 2001, 505-
529.
[14] A.Y.T. Leung, W.E. Zhou, C.W. Lim, R.K.K. Yuen and U. Lee, Dynamic stiffness for piecewise non-uniform Timoshenko column
by power series - part II: Follower force, International Journal for Numerical Methods in Engineering, 51(5), 2001, 531-552.
[15] R. Attarnejad, S.J. Semnani and A. Shahba, Basic displacement functions for free vibration analysis of non-prismatic Timoshenko
beams, Finite Elements in Analysis and Design, 46(10), 2010, 916-929.
[16] T.H. Trinh, T. Hibino and B.S. Gan, Consistent Shape Function for Free Vibration of Linearly Tapered Timoshenko Beam Element,
Bulletin of Nihon University College of Engineering, 57(1), 2015, 7-14.
[17] T.H. Trinh and B.S. Gan, Development of Consistent Shape Functions for linearly solid tapered Timoshenko Beam, Journal of
Structural and Construction Engineering, 80(713), 2015, 1103-1111.
[18] S.P. Timoshenko, On the correction for shear of the differential equation for transverse vibrations of prismatic bars, Philosophical
Magazine, 41, 1921, 744-746.
[19] S.P. Timoshenko, On the transverse vibrations of bars of uniform cross-section, Philosophical Magazine, 43, 1922, 125-131.
[20] J.S. Archer, Consistent mass matrix formulation for structural analysis using finite element techniques, American Institute of
Aeronautics and Astronautics Journal, 3, 1965, 1910-1918.
[21] R.T. Severn, Inclusion of shear deflection in the stiffness matrix for a beam element, The Journal of Strain Analysis for Engineering
Design, 5(4), 1970, 239-241.
[22] D.M. Egle, An approximate theory for transverse shear deformation and rotary inertia effects in vibrating beams, NASA CR-1317,
1969.
[23] Z. Friedman and J.B. Kosmatka, An Improved Two-Node Timoshenko Beam Finite Element, Computers & Structures, 47(3), 1993,
473-481.
[24] J.B. Kosmatka, An improved two-node finite element for stability and natural frequencies of axial-loaded Timoshenko beams,
Computers & Structures, 57(1), 1995, 141-149.
[25] M. Eisenberger, Explicit stiffness matrices for non-prismatic members, Computers & Structures, 20(4), 1985, 715-720.
[26] B.S. Gan, T.H. Trinh and D.K. Nguyen, Post-buckling behaviour of axially FGM planar beams and frames, Procedia Engineering,
171, 2017, 147-158.
[27] J.N. Reddy, Energy principles and variational methods in applied mechanics (Second edition, Hoboken, NJ: John Wiley & Sons,
2002).
[28] M. Eisenberger, Stiffness matrices for non-prismatic members including transverse shear, Computers & Structures, 40(4), 1991,
831-835.