Dr. Jawad Kadhim Uleiwi /Al-khwarizmi Engineering Journal ,Vol.2, No.

2 PP 42-55 (2006)
Al-khwarizmi
Engineering
Journal
Al-Khwarizmi Engineering Journal, Vol.2,No.2,pp32-45 (2006)

Buckling Analysis of Unidirectional Polymer

Matrix Composite Plates

Dr. Jawad Kadhim Uleiwi

Material Engineering. Dept. / University of Technology

(Received 14 November 2005; accepted 4 April 2006)

Abstract:-
This study deals with the estimation of critical load of unidirectional polymer matrix
composite plates by using experimental and finite element techniques at different fiber angles
and fiber volume fraction of the composite plate.
Buckling analysis illustrated that the critical load decreases in nonlinear relationship
with the increase of the fiber angle and that it increases with the increase of the fiber volume
fraction.
The results show that the maximum value of the critical load is (629.54 N/m) at ( =
0) and (Vf = 40 %) for the finite element method, while the minimum value of the critical
load is (49 N/m) at ( = 90) and (Vf = 10 %) for the experimental results.
The results also indicated that the maximum difference between the finite element
analysis and experimental work is about (11 % ) at (  = 0) and (Vf = 40 %).

1.Introduction unstable the load is called a critical load

( or buckling load ) [3].
Many aeroengine components
The nature of the buckle pattern
consist of thin-walled plates under loads
in plate depends not only upon the type
which can potentially cause failure by
of the applied loading but also upon the
buckling. These loads comprise
manner in which the edges are
combinations of compressive, shear and
supported, shape (dimension) of the
bending force [1].
problem and material properties.
The word "buckling" evokes an
The composite plate used in this
image of failure of structures which has
study was a unidirectional polymer
been compressed in some way. Pictures
matrix composite plate made from glass
and perhaps sound come to mind of
fiber – epoxy matrix composite with
sudden, catastrophic collapse involving
constant length to width ratio (L/w = 2)
very large deformations [2].
at different fiber angles and fiber volume
When the magnitude of the load
fraction.
on a structure is such that the
The aim of this work is to
equilibrium is changed from stable to
investigate the buckling analysis of the

42
 Dr. Jawad Kadhim Uleiwi /Al-khwarizmi Engineering Journal ,Vol.2, No. 2 PP 42-55 (2006)

composite plate by using the finite 2.Theoretical Analysis

element analysis and the experimental Using lamination theory, the
work. strains in structural laminate resulting
An enormous amount of effort from a given applied stress from
has been exerted on the subject of information on the basic lamina which
buckling since the beginning of the 20th are: glass and epoxy moduli, poisson's
century mainly the work performed by ratios, densities, and volume fractions of
NACA (National Advisory Committee fibers, in addition to geometric data for
for Aeronautics). the composite plate.
Rais-Rohani, M and Marcellier, The constitute relations may be
P. [4,5] studied the effects of elastic edge established for a lamina by assuming
conditions, aspect ratio, and face sheet that the lamina is a homogeneous
ply pattern on the buckling loads and orthotropic material in a state of plane
natural frequencies of anisotropic stress. The geometry of a single
rectangular sandwich plates by using filamentary lamina is shown in figure (1)
Rayleigh – Ritz method. [11].
Timothy L.C. Chen and Charles Hook's law for a composite
W. Bert [6] used the theory for buckling material with orthotropic properties in a
of laminated, anisotropic, thin plates to plane stress state is:
study buckling of the rectangular plates,  1  Q11 Q12 0   1 
laminated of composite material and    Q   
subject to uniaxial compressive loading  2   21 Q22 0     2  (1)
for optimal design plates laminated of  12   0 0 Q33   12 
glass/epoxy, boron/epoxy, and The components of the reduced stiffness
carbon/epoxy composite materials. matrix, [Q]. are:
Hyonny Kim and Keith T. Q11  E11 / 1  1221  (2)
Kedward [7] studied the global buckling
and local sublaminate buckling of the Q12  Q21  21  E11 /1  12 21 (3)
laminated plate containing Zone of
delamination by using Rayleigh – Ritz Q33  G12 (4)
method and finite element models. And the longitudinal and lateral
Timoshenko, Stephen P. and elastic modulus is predicted by the rule
James M. Gere [8] explained the of mixtures formula.
approximate method to obtain the results
of buckling stresses which are always
E1  E f  V f  Em  Vm (5)
higher than the exact solution. E f  Em
Yasui [9] presented what appears E2  (6)
to be the first in-depth parametric study E f  Vm  Em  V f
of buckling behavior of laminated And the major poisson's ratio is given
composite plates with a central circular by:-
cutout that has been obtained by using
the finite element method. 12   f  V f  m  Vm (7)
Rouse [10] presented an And the minor poisson's ratio 21 can be
experimental investigation of the estimated from
buckling and postbuckling behavior of
E2
square graphite-epoxy and graphite-
thermoplastic plates loaded in shear.
21  12  (8)
E1

43
 Dr. Jawad Kadhim Uleiwi /Al-khwarizmi Engineering Journal ,Vol.2, No. 2 PP 42-55 (2006)

For a given fiber angle ( ) the  M x  t / 2  x 

stress/strain relationship    
 x  Q11 Q12 Q13    x   M y     y 
       M xy  t / 2  xy 
   k
 y   Q21 Q22 Q23     y  (9) (12)
 xy  Q31 Q32 Q33   xy   
      n zk x
 
 z  dz     y   z  dz
Q11  Q11  cos 4   2  Q12  2Q33  
k 1 z k 1
 xy 
 
sin 2   cos 2   Q22  sin 4  After integration and rearranging
Q22  Q11  sin 4   2  Q12  2Q33   equations (11 and 12 ), the constitutive
equations can be obtained as follows:
sin 2   cos 2   Q22  cos 4 
Q12  Q21  Q11  Q22  4  Q33    N x   A11 A12 A13 B11 B12 B13    x 
 

sin 2   cos 2   Q12  sin 4   cos 4   N  
 y   A21 A22 A23 B21 B22 B23    y 
Q33  Q11  Q22  2  Q12  2  Q33    N xy   A31 A32 A33 B31 B32 B33   xy 
     
sin 2   cos 2   Q33  (sin 4   cos 4  )    
 M x   B11 B12 D13   k x 
Q13  Q31  Q11  Q12  2  Q33   sin  
B13 D11 D12
     
 M y   B21 B22 B23 D21 D22 D23   k y 
cos 3   Q12  Q22  2  Q33   sin 3   cos  M   B
 xy   31 B32 B33 D31 D32 D33   k xy 
Q23  Q32  Q11  Q12  2  Q33   sin 3 
(13)
 cos   Q12  Q22  2  Q33   sin   cos 3 
Where:


Aij   Qij  z k  z k 1  
n
Also the stress/strain equation in
the ply can be written in reference to the k 1
geometric midplane as follows:

 x  Q11 Q12 Q13    x 

1 n
  
Bij    Qij  zk2  zk21
2 k 1

     
 y   Q21 Q22 Q23     y  
Dij    Qij  zk3  zk31 
1 n
 xy  Q31 Q32
   Q33   xy 
(10)
3 k 1
Q11 Q12 Q13   k x 
   
Z  Q21 Q22 Q23    k y 
3.Experimental Work
Q31 Q32
 Q33  k xy  The experimental part was
carried out in the field to determine
To find the force resultants acting experimentally the critical load on many
on the total laminate, Equation (10) is specimens of the composite material of
integrated over the total thickness in the unidirectional glass fiber – epoxy matrix
z direction. composite plates with different fiber
 N x  t / 2  x  angles and volume fraction of fiber as
    following:
 N y     y  Vf = 10 %, 20%, 30% and 40%
 N xy  t / 2  xy   = 0, 15, 30, 45, 60, 75, 90
   k
(11) The dimension of the test
zk
 x  specimen has a length of (200 mm) and
n
 
 dz     y   dz width of (100 mm) and a thickness of
(1mm).
k 1 z k 1
 xy 
  Figure (2) represents the test
machine with a test specimen of the

44
 Dr. Jawad Kadhim Uleiwi /Al-khwarizmi Engineering Journal ,Vol.2, No. 2 PP 42-55 (2006)

composite plate with some accessories critical load of the unidirectional

like dial-gauge and load indicator to composite plate at (Vf = 10 % )
measure the lateral deflection of the It is clear from these figures that
composite plate. the buckling happens at the midline of
the plane and the values of (Pcr = 195.83
4.Finite element Analysis N/m, 139.82 N/m, 73.111 N/m and
The finite element analysis 55.142 N/m ) at fiber angle are (  = 0,
carried out as part of this work was 30, 60 and 90) respectively.
performed using the ANSYS package in Figure (6) shows the relationship
the buckling analysis of composite plate between the critical load and the fiber
to determine the critical load at which angle of the composite plate at different
the structure becomes unstable. fiber volume fraction and at ( L/w =
2)for the finite element analysis.
Element Selected It is clear from this figure that the
This element is used to modal the critical load decreases in nonlinear
composite plate. The element has six relationship with the increase of the fiber
degrees of freedom at each node angle in different manners depending on
translation in the x,y and z directions and the volume fraction. The maximum
rotations about the nodal x,y and z-axes. value of ( Pcr = 629.54 N/m ) at ( = 0
The geometry, node locations, and the and Vf = 40 %) while the minimum
coordinate system for this element are value of ( Pcr = 55.142 N/m ) at ( = 90
shown in figure (3). The element is and Vf = 10 %).
defined by eight nodes four thicknesses Figure (7) shows the relationship
and orthotropic material properties [12]. between the critical load and fiber
volume fraction of the composite plate at
5.Boundary Conditions And Mesh different fiber angles for the finite
Generation element method.
There are two main types of It can be seen that the critical
boundary condition (displacement and load increases with the increase of the
load boundary condition). The composite fiber volume fraction in different
plate is a simply – supported plate from manners depending on the fiber angle,
the two opposite edge and free from where the behavior of curve at ( = 0 )
other edges and the load is applied is the greatest one. At this angle the
axially on the two opposite edges. maximum value of (Pcr = 629.54 N/m) at
As for the mesh generation of the (Vf = 40 %) while the minimum value of
composite plate, it is necessary to (Pcr = 195.83 N/m) at (Vf = 10 %).
discretize it into a sufficient number of Figure (8) shows the relationship
elements in order to obtain a reasonable between the critical load and the fiber
accuracy. angle of the composite plate at different
Figure (4) represents the mesh fiber volume fractions and at ( L/w =
generation and boundary conditions of 2)for the experimental work.
the composite plate. It is clear from this figure that the
critical load decreases in nonlinear
6.Results and Discussions relationship with the increase of the fiber
The results obtained from the angle in different manners depending on
experimental work and the finite element the volume fraction. The maximum
analyses of the buckling analysis of the value of ( Pcr = 560 N/m ) at ( = 0 and
composite plate are discussed here. Vf = 40 %) while the minimum value of
Figure (5) shows the lateral
deflection contours and the value of the

45
 Dr. Jawad Kadhim Uleiwi /Al-khwarizmi Engineering Journal ,Vol.2, No. 2 PP 42-55 (2006)

( Pcr = 49 N/m ) at ( = 90 and Vf = 10 7.Conclusion

%). The main conclusions of the
Figure (9) shows the relationship buckling analysis of the composite plates
between the critical load and the fiber for finite element analysis and the
volume fraction of the composite plate at experimental work are:-
different fiber angle for experimental (1-) Critical load decreases in nonlinear
work. relationship with the increase of
It can be seen that the critical the fiber angle in different manners
load increases with the increase of the depending on the fiber volume
fiber volume fraction in different fraction and the maximum value of
manners depending on the fiber angle, (Pcr = 629.54 N/m) at ( = 0) and
where the behavior of the curve at ( = (Vf = 40 %) for finite element
0 ) is the greatest one. At this angle the analysis.
maximum value of (Pcr = 560 N/m) at (2-) Critical load increases with the
(Vf = 40 %) while the minimum value of increase of the fiber volume
(Pcr = 173 N/m) at (Vf = 10 %). fraction and the minimum value (
Figure (10) shows the Pcr = 55.142 N/m ) at ( = 90) and
relationship between the critical load and (Vf = 10 %) for finite element
the fiber angle at ( Vf = 40 %) and at analysis.
(L/w = 2) for the experimental and finite (3-) The behavior of the critical load
element analysis. with fiber angle and with fiber
It is clear from this figure that the volume fraction for experimental
critical load decreases in nonlinear work is similar to the behavior of
relationship with the increase of the fiber finite element analysis but here the
angle. It has also been found that the maximum value of (Pcr = 560 N/m)
maximum difference between the finite at ( = 0) and (Vf = 40 %) the
element analysis results and the minimum value of ( Pcr = 49 N/m )
experimental results ( 11 %) at ( = 0). at ( = 90) and (Vf = 10 %).
The reason for this difference is (4-) Maximum difference between the
attributed to the condition of preparing results of finite element analysis
the specimens. and experimental work results ( =
Figure (11) shows the 11 %) at ( = 0) and ( Vf = 40 %).
relationship between the critical load and
the fiber volume fraction at ( = 0) and
at (L/w = 2) for the experimental and 8.References
finite element analysis.
It is clear from this figure that the [1-] C.A. Featherston and C Ruiz, "
critical load increases in linear Buckling of Flat Plates under
relationship with the increase of the fiber Bending and Shear ", Proc Instn
volume fraction and it has also been Mech Engrs, Vol.212, Part C,
found that the maximum difference (1998).
between finite element analysis results [2-] David Bushnell, " Buckling of
and experimental results ( = 11 %) at ( Shells - Pitfall for Designers",
Vf = 40 %). AIAA80-0665R, Vol.19, No.9,
September, (1981).
[3-] N.G. Lyenger and S.Gupta, "
Programming Methods in Structure
Design", (1980).

46
 Dr. Jawad Kadhim Uleiwi /Al-khwarizmi Engineering Journal ,Vol.2, No. 2 PP 42-55 (2006)

[4-] Rais-Rohani, M and Marcellier, P.," [8-] Timoshenko, Stephen P. and James
Buckling and Vibration Analysis of M. Gere, " Theory of Elastic
Composite Sandwich Plates with Stability", 2nd Edition, McGraw-
Elastic Rotational Edge Hill Book Company, New York,
Restraints". AIAA Journal, Vol.37, (1961).
No. 5, (1999). [9-] Yasui, Yoshiaki," The Buckling of
[5-] Marcellier, P. and Rais-Rohani, M., Rectangular Composite Plates with
" Free Vibration and Buckling Cutout under Uniaxial and Biaxial
Analysis of Anisotropic Sandwich Compression", Proceedings the 8th
Plates with Edges Elastically International Conference on
Restrained Against Rotation", Composite Materials, July, (1991).
Proceedings of the 39th AIAA/ [10-] Rouse, Marshall," Effect of
ASME/ ASCE/ AHS/ ASC Cutouts or Low-Speed Impact
Structures, Structural Dynamics, Damage on the Postbuckling
and Materials Conference, Long Behavior of Composite Plates
Beach, CA, April 20-23, 1998. Loaded in Shear", Proceedings of
AIAA-98-2084-CP. the 31st AIAA, ASME, ASCE,
[6-] Timothy L.C. Chen and Charles W. AHS, ASC, Structures, Structural
Bert " Design of Composite Dynamics and Materials
Material Plates for Maximum Conference, April, (1990).
Uniaxial Compressive Buckling [11-] R.M., Jones, " Mechanics of
Loads", Mechanical and Nuclear Composite Material", McGraw-
Engineering, The University of Hill, New York, (1975).
Oklahoma, Norman, Oklahoma, [12-] Kohnke P.," ANSYS Theory
(1976). Reference Release 5.4", ANSYS,
[7-] Hyonny Kim and Keith T. Inc., (1975).
Kedward," A Method for Modeling
the Local and Global Buckling of
Delaminated Composite Plate ",
Composite Structures, Vol.44,
No.1, (1999).

47
 Dr. Jawad Kadhim Uleiwi /Al-khwarizmi Engineering Journal ,Vol.2, No. 2 PP 42-55 (2006)

Notation
Aij Extensional stiffness terms force intensity in the xy-
(N/m). axis/unit width of the
Bij Coupling stiffness terms laminate cross-section (N).
(N). Qij Reduced stiffness terms
Dij Bending stiffness terms (N/m2).
(N.m). Qij Transformed reduced
2
E11 Young's modulus in the stiffness terms (N/m ).
fiber direction (N/m2). Q Reduced stiffness matrix.
E22 Young's modulus in the Q  Transformed reduced
matrix direction (N/m2). stiffness matrix.
Ef Young's modulus of the Vf Volume fraction of fiber
fiber (N/m2). (%).
Em Young's modulus of the Vm Volume fraction of matrix
matrix (N/m2). (%).
G12 Shear Modulus along the z Coordinate measured from
principal axis (N/m2). the midplane (m).
kx, ky, Curvatures of the midplane  x ,  y ,  xy Midplane strains.
& kxy (m).
Mx Resultant moment intensity  Strain vector in ply.
about the y-axis per unit
 Stress vector in ply.
width of the laminate cross-
section (N.m). i Stress components (N/m2).
My Resultant moment intensity 12 Shear strength along the
about the x-axis per unit principal axis (N/m2)
width of the laminate cross- 12 Shear strain along the
section (N.m). principal axis.
Mxy Resultant twisting moment  Fiber angle ().
intensity about the xy- 12,21 Major and Minor poisson's
axis/unit width of the ratio.
laminate cross-section
(N.m).
Nx Resultant in-plane force
intensity in the x-axis/unit
width of the laminate cross-
section (N).
Ny Resultant in-plane force
intensity in the y-axis/unit
width of the laminate cross-
section (N).
Nxy Resultant in-plane shearing

48
Dr. Jawad Kadhim Uleiwi /Al-khwarizmi Engineering Journal ,Vol.2, No. 2 PP 42-55 (2006)

Figure (1): Location of Material (Principal) Coordinate

System on the Lamina [11].

Load Indicator

Unidirectional
Dial Gauge
Composite Plate

Figure ( 2 ): Test Apparatus and Test Specimen

49
Dr. Jawad Kadhim Uleiwi /Al-khwarizmi Engineering Journal ,Vol.2, No. 2 PP 42-55 (2006)

Figure ( 3 ): SHELL93 8-Node Structural Shell [12].

Figure ( 4 ): Mesh Generation and Boundary Conditions

of the Composite Plate.

50
 Dr. Jawad Kadhim Uleiwi /Al-khwarizmi Engineering Journal ,Vol.2, No. 2 PP 42-55 (2006)

 = 0  = 30

 = 60  = 90

Figure ( 5 ): Deflection Contours of the Composite Plates at (V f=10 %)

and at Different Value of Fiber Angle.

51
 Dr. Jawad Kadhim Uleiwi /Al-khwarizmi Engineering Journal ,Vol.2, No. 2 PP 42-55 (2006)

700

600 F.E.M.

Vf = 40 %

500 Vf = 30 %

Critical Load (N/m)

Vf = 20 %
Vf = 10 %
400

300

200

100

0
0 15 30 45 60 75 90
Fiber Angle (Deg.)

Figure ( 6 ): Relationship Between Critical Load and Fiber Angle for

Finite Element Analysis at Different Volume Fraction.

650
600
F.E.M.
550
 = 0
500  = 45
 = 90
Critical Load (N/m)

450
400
350
300
250
200
150
100
50
0
10 20 30 40
Fiber Volume Fraction (%)

Figure ( 7 ): Relationship Between Critical Load and Fiber Volume

Fraction for Finite Element Analysis at Different Fiber Angle.

52
 Dr. Jawad Kadhim Uleiwi /Al-khwarizmi Engineering Journal ,Vol.2, No. 2 PP 42-55 (2006)

600

500 Experimental

Vf = 40 %
Vf = 30 %
Critical Load (N/m) 400 Vf = 20 %
Vf = 10 %

300

200

100

0
0 15 30 45 60 75 90
Fiber Angle (Deg.)

Figure ( 8 ): Relationship Between Critical Load and Fiber Angle for

Experimental at Different Volume Fraction.

600
550
Experimental
500  = 0
450  = 45
 = 90
Critical Load (N/m)

400
350
300
250
200
150
100
50
0
10 20 30 40
Fiber Volume Fraction (%)

Figure ( 9 ): Relationship Between Critical Load and Fiber Volume

Fraction for Experimental at Different Fiber Angle.

53
 Dr. Jawad Kadhim Uleiwi /Al-khwarizmi Engineering Journal ,Vol.2, No. 2 PP 42-55 (2006)

700

600 F.E.M
Experimental

500

Critical Load (N/m) 400

300

200

100

0
0 15 30 45 60 75 90
Fiber Angle (Deg.)

Figure ( 10 ): Relationship Between Critical Load and Fiber Angle for

Finite Element Analysis Experimental at Fiber Volume
Fraction( = 40 %).

650
600
F.E.M.
550
Experimental
500
Critical Load (N/m)

450
400
350
300
250
200
150
100
50
0
10 20 30 40
Fiber Volume Fraction (%)

Figure (11): Relationship Between Critical Load and Fiber Volume

Fraction for Finite Element Analysis and Experimental
at Fiber Angle (= 0).

54
 ‫)‪Dr. Jawad Kadhim Uleiwi /Al-khwarizmi Engineering Journal ,Vol.2, No. 2 PP 42-55 (2006‬‬

‫تحلٌل األنبعاج للصفائح المركبة األحادٌة األتجاه ذات األساس البولٌمري‬

‫د‪ .‬جواد كاظم عليوي‬

‫قسم هندسة المواد ‪ /‬الجامعة التكنولوجية‬

‫الخالصة‪:‬‬
‫تهتم هذه الدراسة بتقٌم الحمل الحرج للصفائح المركبة األحادٌة األتجاه ذات األساس البولٌمري بأستخدام‬
‫الطرٌقة العملٌة وطرٌقة تقنٌات العناصر المحددة عند زاوٌا لٌف مختلفة وكسر حجمً مختلف لتلك‬
‫الصفائح‪.‬‬
‫اذ ٌبٌن تحلٌل األنبعاج بان الحمل الحرج ٌقل بعالقة غٌر خطٌة مع زٌادة زاوٌة اللٌف وٌزداد مع زٌادة‬
‫الكسر الحجمً للٌف‪.‬‬
‫بٌنت النتائج بان اقصى قٌمة للحمل الحرج لكل وحدة طول ( ‪ )Pcr = 629.5 N/m‬عند ( ‪) = 0‬‬
‫و (‪ )Vf = 40 %‬لطرٌقة العناصر المحددة‪ .‬بٌنما اقل قٌمة للحمل الحرج الحرج ( ‪ )Pcr = 49 N/m‬عند‬
‫(‪ ) = 90‬و (‪ )Vf = 10 %‬للطرٌقة العملٌة‪.‬‬
‫وكذلك بٌنت النتائج بان اقصى فرق بٌن طرٌقة العناصر المحددة والطرٌقة العملٌة (‪ ) = 11 %‬عند ( = ‪‬‬
‫‪ )0‬و (‪.)Vf = 40 %‬‬

‫‪55‬‬