Convergence and accuracy of dynamic relaxation technique in determination of central deflection of composite rectangular laminates

www.ijsret.org
502
International Journal of Scientific Research Engineering & Technology (IJSRET), ISSN 2278 – 0882
Volume 5, Issue 10, October 2016
Convergence and Accuracy of Dynamic Relaxation Technique in
Determination of Central Deflection of Composite Rectangular Laminates
Osama Mohammed Elmardi Suleiman Khayal
Department of Mechanical Engineering, Faculty of Engineering and Technology, Nile Valley University, Atbara – Sudan
ABSTRACT
The convergence and accuracy of the Dynamic
Relaxation (DR) solutions for elastic large deflection
response of isotropic, orthotropic, and laminated plates
are established by comparison with various exact and
approximate solutions. The present Dynamic
Relaxation method (DR) coupled with finite differences
method shows a fairly good agreement with other
analytical and numerical methods used in the verification
scheme.
It was found that: The convergence and accuracy of the
DR solution is dependent on several factors including
boundary conditions, mesh size and type, fictitious
densities, damping coefficients, time increment and
applied load. Also, the DR large deflection program
using uniform finite differences meshes can be
employed in the analysis of different thicknesses for
isotropic, orthotropic or laminated plates under uniform
loads. All the comparison results for simply supported
(SS4) edge conditions showed that deflection is almost
dependent on the direction of the applied load or the
arrangement of the layers.
Keywords - dynamic relaxation; numerical technique;
converged solutions; deflections and stresses; first order
shear deformation theory.
I. INTRODUCTION
The term composite could mean almost anything if taken
at face value, since all materials are composites of
dissimilar subunits if examined at close enough details.
But in modern engineering materials, the term usually
refers to a matrix material that is reinforced with fibers.
For instance, the term “FRP” which refers to Fiber
Reinforced plastic, usually indicates a thermosetting
polyester matrix containing glass fibers, and this
particular composite has the lion’s share of today
commercial market.
In the present work, a numerical method known as
Dynamic Relaxation (DR) coupled with finite
differences is used. The DR method was first proposed
in 1960s and then passed through a series of studies to
verify its validity by Turvey and Osman Refs. [4], [8]
and [9] and Rushton [2], Cassel and Hobbs [10], and
Day [11]. In this method, the equations of equilibrium
are converted to dynamic equations by adding damping
and inertia terms. These are then expressed in finite
difference form and the solution is obtained through
iterations. The optimum damping coefficient and time
increment used to stabilize the solution depend on a
number of factors including the matrix properties of the
structure, the applied load, the boundary conditions and
the size of the mesh used.
Numerical techniques other than the DR include finite
element method, which widely used in the present
studies i.e. of Damodar R. Ambur et al [12], Ying Qing
Huang et al [13], Onsy L. Roufaeil et al [14]… etc. In a
comparison between the DR and the finite element
method, Aalami [15] found that the computer time
required for finite element method is eight times greater
than for DR analysis, whereas the storage capacity for
finite element analysis is ten times or more than that for
DR analysis. This fact is supported by Putcha and Reddy
[16] who noted that some of the finite element
formulations require large storage capacity and computer
time. Hence, due to less computations and computer
time involved in the present study. The DR method is
considered more efficient than the finite element
method. In another comparison Aalami [15] found that
the difference in accuracy between one version of finite
element and another may reach a value of 10% or more,
whereas a comparison between one version of finite
element method and DR showed a difference of more
than 15%. Therefore, the DR method can be considered
of acceptable accuracy. The only apparent limitation of
DR method is that it can only be applied to limited
geometries. However, this limitation is irrelevant to
rectangular plates which are widely used in engineering
applications.
The Dynamic Relaxation (DR) program used in this
paper is designed for the analysis of rectangular plates
irrespective of material, geometry, edge conditions. The
functions of the program are to read the file data;
compute the stiffness of the laminate, the fictitious
densities, the velocities and displacements and the mid –
plane deflections and stresses; check the stability of the
numerical computations, the convergence of the
solution, and the wrong convergence; compute through –
thickness stresses in direction of plate axes; and

www.ijsret.org
503
transform through – thickness stresses in the lamina
principal axes.
The convergence of the DR solution is checked at the
end of each iteration by comparing the velocities over
the plate domain with a predetermined value which
ranges between
9
10
for small deflections and
6
10
for
large deflections. When all velocities are smaller than a
predetermined value, the solution is deemed converged
and consequently the iterative procedure is terminated.
Sometimes DR solution converges to an invalid solution.
To check for that the profile of the variable is compared
with an expected profile over the domain. For example,
when the value of the function on the boundaries is zero,
and it is expected to increase from edge to center, then
the solution should follow a similar profile. When the
computed profile is different from the expected values,
the solution is considered incorrect and can hardly be
made to converge to the correct value by altering the
damping coefficients and time increment. Therefore, the
boundary conditions should be examined and corrected
if they are improper.
The errors inherent in the DR technique are cited in
Refs. {[17] – [24]} which include the discretization error
which is due to the replacement of a continuous function
with a discrete function, and there is an additional error
because the discrete equations are not solved exactly due
to the variations of the velocities from the edge of the
plate to the center. Finer meshes reduce the
discretization error, but increase the round – off error
due to the large number of calculations involved.
II. LARGE DEFLECTION THEORY
The equilibrium, strain, constitutive equations and
boundary conditions are introduced below without
derivation.
Equilibrium Equations
Strain Equations
The large deflection strains of the mid – plane of the
plate are as given below:
x
z
x
w
x
u
x
















2
2
1

y
z
y
w
y
v
y
















2
2
1

























xy
z
y
w
x
w
x
v
y
u
yx




 



y
w
zx

 



x
w
zy

The Constitutive Equations
The laminate constitutive equations can be represented
in the following form:
 














































zx
zy
x
y
j
j
jiji
jiji
i
i
AA
AA
Q
Q
DB
BA
M
N




5545
4544
3
Where: iN Stands for xN , yN and yxN . iM
Denotes xM , yM and yxM . jiA , jiB and jiD
 6,2,1, ji are respectively the membrane rigidities,
coupling rigidities and flexural rigidities of the plate. 
j
Denotes
yx 


 
, and
xy 



 
. 4544
, AA And 55
A
denote the stiffness Coefficients and are calculated as
follows:-
 



1
5,4,,
1
k
k
z
z
ji
n
k
jiji jizdckkA
Where ji
c the stiffness of a lamina is referred to the
plate principal axes, and i
k , j
k are the shear
correction factors.

www.ijsret.org
504
Boundary Conditions
Four sets of simply supported boundary conditions are
used in this paper, and are denoted as SS1, SS2, SS3,
and SS4 as has been shown in Fig. (1) Below:
Fig. (1) Simply supported boundary
conditions
III. DYNAMIC RELAXATION OF THE
PLATE EQUATIONS
An exact solution of the plate equations is obtained
using finite differences coupled with dynamic relaxation
method. The damping and inertia terms are added to
equations (1). Then the following approximations are
introduced for the velocity and acceleration terms:
 
t
ttt
ttt
ba
ba




























/
4
2
1
2
2


In which  ,,,, wvu . Hence equations (1) become:
The superscripts a and b in equations (4) and (5) refer
respectively to the values of velocities after and before

www.ijsret.org
505
the time increment t and 1*
2
1 
   tkk . The
displacements at the end of each time increment, t ,
are evaluated using the following integration procedure:
 6
t
t
b
ba





Thus equations (5), (6), (2) and (3) constitute the set of
equations for solution. The DR procedure operates as
follows:
(1) Set initial conditions.
(2) Compute velocities from equations (5).
(3) Compute displacement from equation (6).
(4) Apply displacement boundary conditions.
(5) Compute strains from equations (2).
(6) Compute stress resultants and stress couples from
equations (3).
(7) Apply stress resultants and stress couples boundary
conditions.
(8) Check if velocities are acceptably small (say 6
10
).
(9) Check if the convergence criterion is satisfied, if it
is not repeat the steps from 2 to 8.
It is obvious that this method requires five fictitious
densities and a similar number of damping coefficients
so as the solution will be converged correctly.
IV. VERIFICATION OF DYNAMIC
RELAXATION (DR) METHOD USING
LARGE DEFLECTION THEORY
Table (1) shows deflections, stress resultants and stress
couples in simply supported in – plane free (SS2)
isotropic plate. The present results have been computed
with 66 uniform meshes. These results are in a fairly
good agreement with those of Aalami et al [1] using
finite difference method (i.e. for deflections, the
difference ranges between 0.35% at 8.20q and 0 %
as the pressure is increased to 97). A similar comparison
between the two results is shown in Table (2) for simply
supported (SS3) condition. It is apparent that the center
deflections, stress couples and stress resultants agree
very well. The mid – side stress resultants do not show
similar agreement whilst the corner stress resultants
show considerable differences. This may be attributed to
the type of mesh used in each analysis. A set of thin
plate results comparisons presented here with Rushton
[2] who employed the DR method coupled with finite
differences. The present results for simply supported
(SS4) square plates were computed for two thickness
ratios using a 88 uniform mesh are listed in table (3).
In this instant, the present results differ slightly from
those found in Ref. [2]. Another comparison for simply
supported (SS4) square isotropic plates subjected to
uniformly distributed loads are shown in Tables (4) and
(5) respectively for deflection analysis of thin and
moderately thick plates. In this comparison, it is noted
that, the center deflection of the present DR analysis, and
those of Azizian and Dawe [3] who employed the finite
strip method are in fairly good agreement (i.e. with a
maximum error not exceeding 0.09%).
A large deflection comparison for orthotropic plates was
made with the DR program. The results are compared
with DR results of Turvey and Osman [4], Reddy’s [5],
and Zaghloul et al results [6]. For a thin uniformly
loaded square plate made of material I which its
properties are stated in Table (6) and with simply
supported in – plane free (SS2) edges. The center
deflections are presented in Table (7) where DR showed
a good agreement with the other three.
A large deflection comparison for laminated plates was
made by recomposing sun and chin’s results [7] for [

44 0/90 ] using the DR program and material II which
its properties are cited in Table (6). The results were
obtained for quarter of a plate using a 55 square
mesh, with shear correction factors 6/52
5
2
4  kk . The
analysis was made for different boundary conditions and
the results were shown in Tables (8), and (9) as follows:
The present DR deflections of two layer antisymmetric
cross – ply simply supported in – plane fixed (SS4) are
compared with DR results of Turvey and Osman [8] and
with sun and chin’s values for a range of loads as shown
in Table (4-8). The good agreement found confirms that
for simply supported (SS4) edge conditions, the
deflection depends on the direction of the applied load or
the arrangement of the layers. Table (9) shows a
comparison between the present DR, and DR Ref. [8]
results, which are approximately identical. The
difference between laminates  
90/0 and  
0/90 at
5/ ab is 0.3% whilst it is 0% when 1/ ab .
The comparison made between DR and alterative
techniques show a good agreement and hence the present
DR large deflection program using uniform finite
difference meshes can be employed with confidence in
the analysis of moderately thick and thin flat isotropic,
orthotropic or laminated plates under uniform loads. The
program can be used with the same confidence to
generate small deflection results.
Table (1) Comparison of present DR, Aalami and
Chapman’s [1] large deflection results for simply
supported (SS2) square isotropic plate subjected to
uniform pressure  3.0,02.0/  vah

www.ijsret.org
506
q S cw
 
 2
1
y
x
M
M  
 2
1
y
x
N
N
20.8
1
2
0.7360
0.7386
0.7357
0.7454
0.7852
0.8278
41.6
1
2
1.1477
1.1507
1.0742
1.0779
1.8436
1.9597
63.7
1
2
1.4467
1.4499
1.2845
1.2746
2.8461
3.0403
97.0
1
2
1.7800
1.7800
1.4915
1.4575
4.1688
4.4322
S (1): present DR results ( 66 uniform mesh over
quarter of the plate)
S (2): Ref. [1] results ( 66 graded mesh over quarter of
the plate)
  0,
2
1
1  zayx
Table (2) Comparison of present DR, Aalami and
Chapman’s [1] large deflection results for simply
supported (SS3) square isotropic plate subjected to
uniform pressure  3.0,02.0/  vah
S (2): Ref. [1] results ( 66 graded mesh over quarter of
the plate)
 
0)4(;0,
2
1
,0)3(
;0,
2
1
)2(;0,
2
1
1


zyxzayx
zyaxzayx
Table (3) Comparison of present DR, and Rushton’s
[2] large deflection results for simply supported (SS4)
square isotropic plate subjected to uniform pressure
 3.0v
q S cw  11
8.2
1
2
3
0.3172
0.3176
0.2910
2.3063
2.3136
2.0900
29.3
1
2
0.7252
0.7249
5.9556
5.9580
3 0.7310 6.2500
91.6
1
2
3
1.2147
1.2147
1.2200
11.3180
11.3249
11.4300
293.0
1
2
3
1.8754
1.8755
1.8700
20.749
20.752
20.820
S (1): present DR results ( 88;02.0/ ah uniform
mesh over quarter of the plate)
S (2): present DR results ( 88;01.0/ ah uniform
mesh over quarter of the plate)
S (3): Ref. [2] results (thin plate 88 uniform mesh
over quarter of the plate)
  hzayx
2
1
,
2
1
1 
Table (4) Comparison of the present DR, and Azizian
and Dawe’s [3] large deflection results for thin shear
deformable simply supported (SS4) square isotropic
plate subjected to uniform pressure
 3.0,01.0/  vah
q S cw
9.2
1
2
0.34693
0.34677
36.6
1
2
0.80838
0.81539
146.5
1
2
1.45232
1.46250
586.1
1
2
2.38616
2.38820
S (2): Azizian and Dawe [3] results.
Table (5) Comparison of the present DR, and Azizian
and Dawe’s [3] large deflection results for
moderately thick shear deformable simply supported
(SS4) square isotropic plates subjected to uniform
pressure  3.0,05.0/  vah
q S cw
0.92
1
2
0.04106
0.04105
4.6
1
2
0.19493
0.19503
6.9
1
2
0.27718
0.27760
9.2
1
2
0.34850
0.34938

www.ijsret.org
507
S (2): Azizian and Dawe [3] results.
Table (6) Material properties used in the orthotropic
and laminated plate comparison analysis.
Table (7) Comparison of present DR, DR results of
Ref. [4], finite element results Ref. [5] and
experimental results Ref. [6] for a uniformly loaded
simply supported (SS2) square orthotropic plate
made of material I  0115.0/ ah
S (1): present DR results ( 55 uniform non –
interlacing mesh over quarter of the plate).
S (2): DR results of Ref. [4].
S (3): Reddy’s finite element results [5].
S (4): Zaghloul’s and Kennedy’s Ref. [6] experimental
results as read from graph.
Table (8) Deflection of the center of a two – layer anti
symmetric cross ply simply supported in – plane
fixed (SS4) strip under uniform
pressure 01.0/,5/  ahab .
S (1): present DR results
S (2): DR results Ref. [8].
S (3): Values determined from sun and chin’s results
Ref. [7].
(1):    www /100 1 
(2):    www /100 2 
(3):   221 /100 www 
Table (9) Center deflection of two – layer anti –
symmetric cross – ply simply supported in – plane
free (SS1) plate under uniform pressure and with
different aspect ratios  18;01.0/  qah .
V. CONCLUSIONS
A Dynamic relaxation (DR) program based on finite
differences has been developed for large deflection
analysis of rectangular laminated plates using first order
shear deformation theory (FSDT). The displacements are
assumed linear through the thickness of the plate. A
series of new results for uniformly loaded thin,
moderately thick, and thick plates with simply supported
edges have been presented. Finally a series of numerical
comparisons have been undertaken to demonstrate the
accuracy of the DR program. These comparisons show
the following:-
1. The convergence of the DR solution depends on
several factors including boundary conditions, meshes
size, fictitious densities and applied load.
2. The DR large deflection program using uniform finite
differences meshes can be employed with confidence in
the analysis of moderately thick and flat isotropic,
orthotropic or laminated plates under uniform loads.
3. The DR program can be used with the same
confidence to generate small deflection results.
4. The time increment is a very important factor for
speeding convergence and controlling numerical
computations. When the increment is too small, the
convergence becomes tediously slow; and when it is too
large, the solution becomes unstable. The proper time
increment in the present study is taken as 0.8 for all
boundary conditions.
5. The optimum damping coefficient is that which
produces critical motion. When the damping coefficients
are large, the motion is over – damped and the
convergence becomes very slow. At the other hand when
the coefficients are small, the motion is under – damped

www.ijsret.org
508
and can cause numerical instability. Therefore, the
damping coefficients must be selected carefully to
eliminate under – damping and over – damping.
ACKNOWLEDGEMENT
The author would like to acknowledge with deep thanks
and profound gratitude Mr. Osama Mahmoud of Dania
Center for Computer and Printing Services, Atbara, who
spent many hours in editing, re – editing of the
manuscript in compliance with the standard format of
IJSRET Journal. Also, my appreciation is extended to
Professor Mahmoud Yassin Osman for revising and
correcting the manuscript several times.
REFERENCES
[1] Aalami, B. and Chapman J.C. 'large deflection
behavior of rectangular orthotropic plates under
transverse and in – plane loads, proceedings of the
institution of civil Engineers', 1969, 42, 347 – 382.
[2] Rushton K. R., ‘large deflection of variable –
thickness plates’, International journal of mech.
Sciences, vol. 10, 1968, 723 – 735.
[3] Azizian Z. G. and Dawe D.J., ‘Geometrically Non –
linear analysis of rectangular Mindlin plates using the
finite strip method’, computers and structures, vol.22,
No.3, 1985, 523 – 436.
[4] Turvey G.J. and Osman M.Y., ‘large deflection
analysis of orthotropic Mindlin plates, proceedings of
the 12th Energy resources technical conference and
exhibition’, Houston, Texas 1989, 163 – 172.
[5] Reddy J.N., ‘Energy and variation methods in
applied mechanics’, John Wiley and sons, New York,
1984, 379 – 387.
[6] Zaghloul S.A. and Kennedy J.B., ‘Nonlinear
behavior of symmetrically laminated plates’, Journal of
applied mechanics, vol.42, 1975, 234 – 236.
[7] Sun C.T. and Chin H., ‘on large deflection effects in
unsymmetrical cross – ply composite laminates’, Journal
of composite materials, vol.22, 1988, 1045 – 1059.
[8] Turvey G.J. and Oman M. Y., ‘large deflection
effects in anti-symmetric cross – ply laminated strips and
plates’, I.H. Marshall, composite structures, vol.6,
paisley college, Scotland, Elsevier science publishers,
1991, 397 – 413.
[9] Turvey G. J. and Osman M. Y., ‘Elastic large
deflection analysis of isotropic rectangular Mindlin
plates’, International journal of mechanical sciences, vol.
22, 1990, 1 – 14.
[10] Cassel A.C. and Hobbs R.E., ‘Numerical Stability
of dynamic relaxation analysis of nonlinear structures’,
International Journal for Numerical Methods in
Engineering, Vol.35, No.4, 1966, 1407 –1410.
[11] Day A.S., ’An introduction to dynamic relaxation’,
the engineer, vol. 219, No. 5668, 1965, 218 - 221.
[12] Damodar R. Ambur, Navin Jounky, Mark
Hilburger, Carlos G. Davila,’ Progressive failure
analysis of compression loaded composite Curved panels
with and without Cutouts,’ composite structures, vol.65,
2004, 143 – 155.
[13] Ying Qing Huang, Shenglin Di, Chang Chun Wu,
and Huiya Sun, ’Bending analysis of composite
laminated plates using a partially hybrid stress element
with interlaminar continuity’, computer and structures,
vol.80, 2002, 403 – 410.
[14] Onsy L. Roufaeil, Thanh Tran – Cong,’ finite strip
elements for laminated composite plates with transverse
shear strain discontinuities’, composite structure, vol. 56,
2002 , 249 – 258.
[15] Aalami B., ‘Large deflection of elastic plates under
patch loading’, Journal of structural division, ASCE, vol.
98, No. ST 11, 1972, 2567 – 2586.
[16] Putcha N.S. and Reddy J.N., ‘A refined mixed shear
flexible finite element for the non – linear analysis of
laminated plates,’ computers and structures, vol. 22, No.
4, 1986, 529 – 538.
[17]Osama Mohammed Elmardi Suleiman, ‘Verification
of dynamic relaxation method in isotropic, orthotropic
and laminated plates using small deflection theory’,
International Journal of Advanced Science and
Technology, volume 72; 2014, 37 – 48.
[18] Osama Mohammed Elmardi Suleiman, ‘Nonlinear
analysis of rectangular laminated plates using large
deflection theory’, International Journal of Emerging
Technology and Research, volume 2, issue 5; 2015, 26 –
48.
[19] Osama Mohammed Elmardi Suleiman, ‘Validation
of Dynamic Relaxation (DR) Method in Rectangular
Laminates using Large Deflection Theory’, International
Journal of Advanced Research in Computer Science and
Software Engineering , volume 5, issue 9; September
2015, 137 – 144.
[20] Osama Mohammed Elmardi Suleiman, ‘Nonlinear
analysis of rectangular laminated plates’, Lap Lambert
Academic Publishing, Germany, and ISBN: (978-3-659-
76787-6); 2015.
[21] Osama Mohammed Elmardi Suleiman, ‘Text Book
on Dynamic Relaxation Method’, Lap Lambert
Academic Publishing, Germany, and ISBN: (978-3-659-
94751-3) ; 2016.
[22]Osama Mohammed Elmardi Suleiman,
‘Introduction and Literature Review on Buckling of
Composite Laminated Plates’, Lap Lambert Academic
Publishing, Germany, and ISBN: (978-3-659-86387-5);
2016.

www.ijsret.org
509
[23] Osama Mohammed Elmardi Suleiman,
‘Bibliography and Literature Review on Buckling of
Laminated Plates’, International Journal of Science and
Engineering, volume 2, issue 8; August 2016, 104 – 112.
[24] Osama Mohammed Elmardi Suleiman, ‘Theories of
Composite Plates and numerical Methods Used on
Bending and Buckling of Laminated Plates’,
International Journal of Engineering Research And
Advanced Technology (IJERAT), volume 2, issue 10;
October 2016, 1 –12.
Author
Osama Mohammed Elmardi
Suleiman was born in Atbara,
Sudan in 1966. He received his
diploma degree in mechanical
engineering from Mechanical
Engineering College, Atbara,
Sudan in 1990. He also received a
bachelor degree in mechanical
engineering from Sudan
University of science and technology – Faculty of
engineering in 1998, and a master degree in solid
mechanics from Nile valley university (Atbara, Sudan)
in 2003. He contributed in teaching some subjects in
other universities such as Red Sea University (Port
Sudan, Sudan), Kordofan University (Obayed, Sudan),
Sudan University of Science and Technology
(Khartoum, Sudan) and Blue Nile university (Damazin,
Sudan). In addition he supervised more than hundred and
fifty under graduate studies in diploma and B.Sc. levels
and about fifteen master theses. He is currently an
assistant professor in department of mechanical
engineering, Faculty of Engineering and Technology,
Nile Valley University. His research interest and favorite
subjects include structural mechanics, applied
mechanics, control engineering and instrumentation,
computer aided design, design of mechanical elements,
fluid mechanics and dynamics, heat and mass transfer
and hydraulic machinery. He works also as a consultant
and a technical manager of Al – Kamali workshops
group in Atbara – Sudan old and new industrial areas.

Convergence and accuracy of dynamic relaxation technique in determination of central deflection of composite rectangular laminates

More Related Content

What's hot (20)

Similar to Convergence and accuracy of dynamic relaxation technique in determination of central deflection of composite rectangular laminates (20)

More from Osama Mohammed Elmardi Suleiman (20)

Recently uploaded (20)

Convergence and accuracy of dynamic relaxation technique in determination of central deflection of composite rectangular laminates