For dynamic analysis of eccentrically stiffened laminated com-posite plates and shallow shells, Satish Kumar and Mukhopadhyay [9]studied the transient response analysis of laminated stif
Trang 1Nonlinear dynamic analysis of eccentrically stiffened imperfect functionally
graded doubly curved thin shallow shells
Dao Huy Bicha, Dao Van Dunga, Vu Hoai Namb,⇑
a
Vietnam National University, Ha Noi, Viet Nam
b
University of Transport Technology, Ha Noi, Viet Nam
a r t i c l e i n f o
Article history:
Available online 29 October 2012
Keywords:
Functionally graded material
Dynamic analysis
Critical dynamic buckling load
Vibration
Shallow shells
Stiffeners
a b s t r a c t
This paper presents a semi-analytical approach to investigate the nonlinear dynamic of imperfect eccen-trically stiffened functionally graded shallow shells taking into account the damping subjected to mechanical loads The functionally graded shallow shells are simply supported at edges and are rein-forced by transversal and longitudinal stiffeners on internal or external surface The formulation is based
on the classical thin shell theory with the geometrical nonlinearity in von Karman–Donnell sense and the smeared stiffeners technique By Galerkin method, the equations of motion of eccentrically stiffened imperfect functionally graded shallow shells are derived Dynamic responses are obtained by solving the equation of motion by the Runge–Kutta method The nonlinear critical dynamic buckling loads are found according to the Budiansky–Roth criterion Results of dynamic analysis show the effect of stiffen-ers, damping, pre-loaded compressions, material and geometric parameters on the dynamical behavior of these structures
Ó 2012 Elsevier Ltd All rights reserved
1 Introduction
Eccentrically stiffened shallow shell is a very important
struc-ture in engineering design of aircraft, missile and aerospace
indus-tries There are many researches on the static and dynamic
behavior of this structure with different materials
Studies on the dynamics first were carried out with
eccentri-cally stiffened shallow shells made of homogeneous material
Khalil et al.[1]presented a finite strip formulation for the
nonlin-ear analysis of stiffened plate structures subjected to transient
pressure loadings using an explicit central difference/diagonal
mass matrix time stepping method Shen and Dade[2]investigated
dynamic analysis of stiffened plates and shells using spline gauss
collocation method The free vibration of stiffened shallow shells
was studied by Nayak and Bandyopadhyay[3] By using the finite
element method, the stiffened shell element was obtained by the
appropriate combinations of the eight-/nine-node doubly curved
isoparametric thin shallow shell element with the three-node
curved isoparametric beam element Sheikh and Mukhopadhyay
[4] applied the spline finite strip method to investigate linear
and nonlinear transient vibration analysis of plates and stiffened
plates The von Karman’s large deflection plate theory has been
used and the formulation was done in total Lagrange coordinate
system Dynamic instability analysis of stiffened shell panels
sub-jected to uniform in-plane harmonic edge loading and partial edge
loading along the edges was studied by Patel et al.[5–8] In these studies, the new formulation for the beam element requires five degrees of freedom per node as that of shell element
For dynamic analysis of eccentrically stiffened laminated com-posite plates and shallow shells, Satish Kumar and Mukhopadhyay
[9]studied the transient response analysis of laminated stiffened plates using the first order shear deformation theory Parametric study on the dynamic instability behavior of laminated composite stiffened plate was studied by Patel et al.[10] The same authors
[11]investigated the dynamic instability of laminated composite stiffened shell panels subjected to in-plane harmonic edge loading
By using the commercial ANSYS finite element software, Less and Abramovich [12] studied the dynamic buckling of a laminated composite stringer stiffened cylindrical panel Bich et al.[13] pre-sented an analytical approach to investigate the nonlinear dynamic
of imperfect reinforced laminated composite plates and shallow shells using the classical thin shell theory with the geometrical nonlinearity in von Karman–Donnell sense and the smeared stiff-eners technique
For functionally graded materials (FGM), many researches fo-cused on the dynamical analysis of un-stiffened shallow shells Liew et al.[14]presented the nonlinear vibration analysis of the coating FGM substrate cylindrical panel subjected to a temperature gradient arising from steady heat conduction through the panel thickness Matsunaga et al.[15]investigated free vibrations and stability of FGM doubly curved shallow shells according to a 2-D higher order deformation theory Chorfi and Houmat[16] investi-gated nonlinear free vibrations of FGM doubly curved shallow 0263-8223/$ - see front matter Ó 2012 Elsevier Ltd All rights reserved.
⇑Corresponding author.
E-mail address: hoainam.vu@utt.edu.vn (V.H Nam).
Contents lists available atSciVerse ScienceDirect
Composite Structures
j o u r n a l h o m e p a g e : w w w e l s e v i e r c o m / l o c a t e / c o m p s t r u c t
Trang 2shells with an elliptical plan-form Nonlinear vibrations of
func-tionally graded doubly curved shallow shells under a concentrated
force were studied by Alijani et al.[17] Nonlinear dynamical
anal-ysis of imperfect functionally graded material shallow shells
sub-jected to axial compressive load and transverse load was studied
by Bich and Long[18], Dung and Nam[19] The motion, stability
and compatibility equations of these structures were derived using
the classical shell theory The nonlinear transient responses of
cylindrical and doubly-curved shallow shells subjected to excited
external forces were obtained and the dynamic critical buckling
loads were evaluated based on the displacement responses using
Budiansky–Roth dynamic buckling criterion
Recently, some authors have studied static and dynamical
behaviors of some kind of shells Najafizadeh et al.[20]have
stud-ied static buckling behaviors of FGM cylindrical shell Bich et al
[21]have studied the nonlinear static post-buckling of
eccentri-cally stiffened imperfect functionally graded plates and shallow
shells The nonlinear dynamical analysis of imperfect eccentrically
stiffened FGM cylindrical panels based on the classical theory with
the von Karman–Donnell geometrical nonlinearity are investigated
by Bich et al.[22]
Following the idea of work[22], in this paper, dynamic
govern-ing equations takgovern-ing into account the effect of dampgovern-ing, nonlinear
vibration and dynamic critical buckling loads of eccentrically
stiff-ened imperfect FGM doubly curved thin shallow shells are
estab-lished Effects of stiffeners, material, geometric parameters and
damping on the dynamic behavior of structure are considered
2 Eccentrically stiffened FGM shallow shells (ES-FGM shallow
shells)
Consider a doubly curved functionally graded shallow thin shell
(seeFig 1) of thickness h and in-plane edges a and b The shallow
shell is assumed to have a relative small rise as compared with its
span Let the (x1,x2) plane of the Cartesian coordinates overlaps the
rectangular plane area of the shell Note that the middle surface of
the shell generally is defined in terms of curvilinear coordinates,
but for the shallow shell, so the Cartesian coordinates can replace
the curvilinear coordinates on the middle surface
The volume fractions of constituents are assumed to vary
through the thickness according to the following power law
distribution
VcðzÞ ¼ 2z þ h
2h
k
; VmðzÞ ¼ 1 VcðzÞ; ð1Þ
where k is the volume fraction exponent (k P 0), z is the thickness coordinate and varies from h/2 to h/2; the subscripts m and c refer
to the metal and ceramic constituents respectively Effective prop-erties Preffof FGM shell are determined by linear rule of mixture as
Preff ¼ PrmðzÞVmðzÞ þ PrcðzÞVcðzÞ: ð2Þ
According to Eqs.(1) and (2), the modulus of elasticity E and the mass densityqcan be expressed in the form
EðzÞ ¼ EmVmþ EcVc¼ Emþ ðEc EmÞ 2z þ h
2h
k
;
qðzÞ ¼qmVmþqcVc¼qmþ ðqcqmÞ 2z þ h
2h
k
:
ð3Þ
Poisson’s ratiomand linear damping coefficienteare assumed to
be constants
Assume that the shell is reinforced by eccentrically longitudinal and transversal homogeneous stiffeners with the elastic modulus
E0and the mass densityq0of stiffeners In order to provide the continuity between the shell and stiffeners, the full metal stiffeners are put at the metal-rich side of the shell thus E0andq0take the value E0= Em,q0=qmand conversely the full ceramic ones at the ceramic-rich side, so that E0= Ec,q0=qc
3 Theoretical formulation The strains at the middle surface and curvatures relating to the displacement components u,v, w based on the classical shell the-ory and von Karman–Donnell geometrical nonlinearity assumption are of the form[24]
e0¼ @u
@x1
k1w þ1 2
@w
@x1
2
; v1¼@
2w
@x2;
e0¼@v
@x2
k2w þ1 2
@w
@x2
2
; v2¼@
2w
@x2;
c0
@x2þ@v
@x1þ@w
@x1
@w
@x2
; v12¼ @
2
w
@x1@x2
;
ð4Þ
in which k1¼ 1
1; k2¼ 1
2and R1, R2are curvatures and radii of cur-vatures of the shell respectively
Trang 3The strain components across the shell thickness at a distance z
from the middle surface are given by
e1¼e0 zv1; e2¼e0 zv2; c12¼c0
From Eq.(4), the strains must be relative in the deformation
compatibility equation
@2e0
@x2 þ@
2e0
@x2 @
2c0
12
@x1@x2
¼ @
2w
@x1@x2
!2
@
2w
@x2
@2w
@x2 k1
@2w
@x2 k2
@2w
@x2: ð6Þ
Hooke’s stress–strain relation is applied for the shell
rsh
1 m2ðe1þme2Þ;
rsh
1 m2ðe2þme1Þ;
ssh
2ð1 þmÞc12;
ð7Þ
and for stiffeners
rst
1¼ E0e1;
rst
Taking into account the contribution of stiffeners by the
smeared stiffeners technique and omitting the twist of stiffeners
because these torsion constants are smaller more than the
mo-ments of inertia[24]and integrating the stress–strain equations
and their moments through the thickness of the panel, lead to
the expressions for force and moment resultants of an ES-FGM
shallow shells as[22]
N1¼ A11þE0A1
s1
e0þ A12e0 ðB11þ C1Þv1 B12v2;
N2¼ A12e0þ A22þE0A2
s2
e0 B12v1 ðB22þ C2Þv2;
N12¼ A66c0
12 2B66v12;
ð9Þ
M1¼ ðB11þ C1Þe0þ B12e0 D11þE0I1
s1
v1 D12v2;
M2¼ B12e0þ ðB22þ C2Þe0 D12v1 D22þE0I2
s2
v2;
M12¼ B66c0
12 2D66v12;
ð10Þ
where Aij, Bij, Dij(i, j = 1, 2, 6) are extensional, coupling and bending
stiffness of the un-stiffened shell
A11¼ A22¼ E1
1 m2; A12¼ E1m
1 m2; A66¼ E1
2ð1 þmÞ;
B11¼ B22¼ E2
1 m2; B12¼ E2m
1 m2; B66¼ E2
2ð1 þmÞ;
D11¼ D22¼ E3
1 m2; D12¼ E3m
1 m2; D66¼ E3
2ð1 þmÞ;
ð11Þ
with
E1¼ EmþEc Em
k þ 1
h; E2¼ ðEc EmÞkh
2
2ðk þ 1Þðk þ 2Þ;
E3¼ Em
12þ ðEc EmÞ
1
k þ 3
1
k þ 2þ
1 4k þ 4
h3;
I1¼d1h
3
1
12 þ A1z
2; I2¼d2h
3 2
12 þ A2z
2:
ð12Þ
and
C1¼ E0A1z1
s1
; C2¼ E0A2z2
s2
;
z1¼h1þ h
2 ; z2¼
h2þ h
2 ;
where the coupling parameters C1, C2are negative for outside stiff-eners and positive for inside ones, s1and s2are the spacing of the longitudinal and transversal stiffeners, A1, A2are the cross-section areas of stiffeners, I1,I2 are the second moments of cross-section areas, z1, z2are the eccentricities of stiffeners with respect to the middle surface of shell, and the width and thickness of longitudinal and transversal stiffeners are denoted by d1, h1 and d2, h2, respectively
For later use, the strain-force resultant reverse relations are ob-tained from Eq.(9)
e0¼ A22N1 A12N2þ B
11v1þ B
12v2;
e0¼ A11N2 A12N1þ B21v1þ B22v2;
c0
12¼ A66þ 2B66v12;
ð13Þ
where
A11¼1
D A11þE0A1
s1
; A22¼1
D A22þE0A2
s2
;
A12¼A12
D ; A66¼ 1
A66
;
D¼ A11þE0A1
s1
A22þE0A2
s2
A212;
B
11¼ A22ðB11þ C1Þ A12B12;
B
22¼ A11ðB22þ C2Þ A12B12;
B
12¼ A22B12 A12ðB22þ C2Þ;
B
21¼ A11B12 A12ðB11þ C1Þ; B
66¼B66
A66
:
ð14Þ
Substituting Eq.(13)into Eq.(10)yields
M1¼ B11N1þ B21N2 D11v1 D12v2;
M2¼ B12N1þ B22N2 D21v1 D22v2;
M12¼ B66N12 2D66v12;
ð15Þ
where
D
11¼ D11þE0I1
s1 ðB11þ C1ÞB11 B12B
21;
D
22¼ D22þE0I2
s2 B12B
12 ðB22þ C2ÞB22;
D
12¼ D12 ðB11þ C1ÞB12 B12B
22;
D21¼ D12 B12B11 ðB22þ C2ÞB21;
D
66¼ D66 B66B
66:
ð16Þ
Based on the classical shell theory and the Volmir’s assumption
[23] u w and v w; q1@ 2 u
@t 2! 0 and q1@ 2v
@t 2! 0, the nonlinear motion equations of a shallow thin shell with damping force is written in the form
@N1
@x1þ@N12
@x2 ¼ 0;
@N12
@x1
þ@N2
@x2
¼ 0;
@2M1
@x2 þ 2@
2M12
@x1@x2
þ@
2M2
@x2 þ N1
@2w
@x2þ 2N12
@2w
@x1@x2
þ N2
@2w
@x2
þ k1N1þ k2N2þ q0¼q1@
2w
@t2 þ 2q1e@w
@t ;
ð17Þ
whereeis damping coefficient and
Trang 4Z h=2
h=2
qðzÞdz þq0 A1
s1þA2
s2
¼ qmþqcqm
k þ 1
h þq0 A1
s1
þA2
s2
:
The first two of Eq.(16)are satisfied automatically by
introduc-ing a stress functionuas
N1¼@
2u
@x2; N2¼@
2u
@x2; N12¼ @
2u
@x1@x2
Substituting Eq.(13) into the compatibility Eqs (6) and (15)
into the third of Eq.(17), taking into account expressions(4) and
(18), yields
A
11
@4u
@x4þ A66 2A12
u
@x2@x2þ A22
@4u
@x4þ B21
@4w
@x4
þ B
11þ B
22 2B
66
@x2@x2þ B
12
@4w
@x4þ k1
@2w
@x2þ k2
@2w
@x2
¼ @
2w
@x1@x2
!2
@
2w
@x2
@2w
q1@
2
w
@t2 þ 2q1e@w
@t þ D
11
@4w
@x4þ D12þ D21þ 4D66
w
@x2@x2
þ D22@
4w
@x4 B21@
4u
@x4 B 11þ B22 2B66 @4u
@x2@x2 B12@
4u
@x4
k1
@2u
@x2 k2
@2u
@x2@
2u
@x2
@2w
@x2 þ 2 @
2u
@x1@x2
@2w
@x1@x2
@
2u
@x2
@2w
@x2 ¼ q0: ð20Þ
For an initial imperfection shell: The initial imperfection of the
shell considered here can be seen as a small deviation of the shell
middle surface from the perfect shape, also seen as an initial
deflection which is very small compared with the shell dimensions,
but may be compared with the shell wall thickness Let
w0= w0(x1, x2) denote a known small imperfection, proceeding
from the motion Eqs.(19) and (20)of a perfect FGM shallow shell
and following to the Volmir’s approach[23] for an imperfection
shell, we can formulate the system of motion equations for an
imperfect eccentrically stiffened functionally graded shallow shell
(Imperfect ES-FGM shallow shell) as
A11@
4u
@x4þ A 66 2A12 @4u
@x2@x2þ A22@
4u
@x4þ B 21
@4ðw w0Þ
@x4
þ B11þ B22 2B66
ðw w0Þ
@x2@x2 þ B12
@4ðw w0Þ
@x4
þ k1
@2ðw w0Þ
@x2 þ k2
@2ðw w0Þ
@x2 @
2
w
@x1@x2
!2
@
2
w
@x2
@2w
@x2
2 4
3 5
2w0
@x1@x2
!2
@
2w0
@x2
@2w0
@x2
2
4
3
Uq1@
2w
@t2 þ 2q1e@w
@t þ D
11
@4ðw w0Þ
@x4 þ D
12þ D
21þ 4D 66
@4ðw w0Þ
@x2@x2 þ D22
@4ðw w0Þ
@x4 B21
@4u
@x4
B11þ B22 2B66
u
@x2@x2 B12
@4u
@x4 k1
@2u
@x2 k2
@2u
@x2
@
2u
@x2
@2w
@x2þ 2 @
2u
@x1@x2
@2w
@x1@x2
@
2u
@x2
@2w
@x2 q0¼ 0; ð22Þ
where w is a total deflection of shell
The couple of nonlinear Eqs.(19) and (20)or Eqs.(21) and (22)
in terms of two dependent unknowns w anduare used to investi-gate the nonlinear vibration and dynamic stability of ES-FGM shells
4 Nonlinear dynamic analysis
An imperfect ES-FGM shallow thin shell considered in this pa-per is assumed to be simply supported and subjected to uniformly distributed pressure of intensity q0and axial compression of inten-sities r0and p0respectively at its cross-section Thus the boundary conditions are
w ¼ 0; M1¼ 0; N1¼ r0h; N12¼ 0; at x1¼ 0; a;
w ¼ 0; M2¼ 0; N2¼ p0h; N12¼ 0; at x2¼ 0; b:
ð23Þ
The mentioned conditions(23)can be satisfied identically if the buckling mode shape is chosen by
w ¼ f ðtÞ sinmpx1
a sin
npx2
where f(t) is time dependent total amplitude and m, n are numbers
of haft waves in x1and x2directions respectively
The initial imperfection w0is assumed to have the same form of the shell deflection w, i.e
w0¼ f0sinmpx1
a sin
npx2
where f0is the known initial amplitude
Substituting Eqs (24) and (25)into Eq (21) and solving the resulting equation for unknownuyield
u¼u1cos2mpx1
a þu2cos2npx2
b u3sinmpx1
a sin
npx2
b
r0hx
2
2 p0hx
2
where
u1¼ n
2k2f2
32m2A 11
2f2
32n2k2A22;
u3¼
B
21m4þ B 11þ B22 2B66
m2n2k2þ B12n4k4a 2
p 2k1n2k2þ k2m2
A
11m4þ A 66 2A12
m2n2k2þ A22n4k4 :
ð27Þ Substituting the expressions(24)–(26)into Eq.(22)and apply-ing the Galerkin’s method to the obtained equation
Z b 0
Z a 0
Usinmpx1
a sin
npx2
b dx1dx2¼ 0
lead to
M€f þ 2Me_f þ D þB2
A
!
ðf f0Þ þ8mnk
2
3p2 d1d2
B
Aðf f0Þf
þ H f2 f2
þ K f2 f2
f a
2h
p2 r0m2þ p0n2
k2
f
þ4a
4h
mnp6d1d2ðk1r0þ k2p0Þ q0
4a4
mnp6d1d2¼ 0; ð28Þ
where the coefficients are given by
Trang 5M ¼a
4
p4q1;A ¼ A11m4þ A 66 2A12
m2n2
k2þ A22n4
k4;
B ¼ B
21m4þ B
11þ B
22 2B 66
m2n2
k2þ B
12n4
k4a
2
p2 k1n2
k2þ k2m2
;
D ¼ D
11m4þ D12þ D21þ 4D66
m2n2k2þ D22n4k4;
H ¼2mnk
2
3p2
B
21
A11þ
B 12
A22
d1d2 a
2
6p4mn
k2n2k2
A11 þ
k1m2
A22
!
d1d2; ð29Þ
K ¼ 1
16
m4
A
22
þn
4k4
A
11
!
;k¼a
b;d1¼ ð1Þ
m
1; d2¼ ð1Þn 1:
The governing Eq.(28)is a basic equation for dynamic analysis
of imperfect ES-FGM doubly-curved shallow thin shells in general
Based on this equation, the nonlinear vibration of perfect and
imperfect FGM shallow shells can be investigated and the dynamic
buckling analysis of shells under various loadings can be
performed
Particularly for a spherical panel k1= k2= 1/R, for a cylindrical
panel k1= 0, k2= 1/R, and for a plate k1= k2= 0 are taken in Eqs
(28) and (29)
4.1 Nonlinear vibration
Consider an imperfect ES-FGM shallow shell under uniformly
lateral pressure q0= Q sinXt and pre-loaded compressions r0, p0,
Eq.(28)becomes
M€f þ 2Me_f þ D þB2
A
!
ðf f0Þ þ8mnk
2
3p2
B
Ad1d2ðf f0Þf
þ H f2 f2
þ þK f2 f2
f a
2h
p2r0m2þ p0n2k2
f
þ4a
4h
mnp6d1d2ðk1r0þ k2p0Þ ¼ 4a
4
p6mnd1d2Q sinXt: ð30Þ
By using this equation, the fundamental frequencies of natural
vibration of ES-FGM shell and FGM shell without stiffeners, and
frequency-amplitude relation of nonlinear vibration and nonlinear
response of ES-FGM shell are taken into consideration The
nonlin-ear dynamical responses of ES-FGM shells can be obtained by
solv-ing Eq.(30)by the fourth order Runge–Kutta iteration method with
the time step Dt and with initial conditions to be assumed as
f ð0Þ ¼ 0; _fð0Þ ¼ 0
If the vibration is free and linear, and without damping, Eq.(30)
reduces to
M€f þ D þB
2
A
!
The fundamental frequencies of natural vibration of imperfect
ES-FGM shallow shells can be determined by
xmn¼
ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
1
M D þ
B2
A
! v
u
If the shell is perfect ES-FGM and the vibration is nonlinear
forced vibration without pre-loaded compressions r0= p0= 0, Eq
(30)can be rewritten a
€
f þ 2e_f þx2
mnf þ H1f2þ H2f3
¼ F sinðXtÞ; ð33Þ
in which
H1¼
8mnk 2
3p2 Bd1d2þ H
Mx2
mn
; H2¼ K
Mx2 mn
; F ¼4a
4d1d2Q
Mp6mn : ð34Þ
Seeking solution as f(t) =gsin (Xt) and applying procedure like Galerkin method to Eq.(33), the frequency–amplitude relation of nonlinear forced vibration is obtained
X24e
p X¼x
2
3p gþ
3H2
4 g2
F
wheregis the amplitude of nonlinear vibration
By introducing the non-dimension frequency parameter n ¼ X
x mn,
Eq.(35)becomes
n2 4e
pxmn
n¼ 1 þ8H1
3p gþ
3H2
4 g2
F
gx2 mn
In the case of the nonlinear forced vibration without damping, Eq
(36)leads to
n2¼ 1 þ8H1
3p gþ
3H2
4 g2 F
gx2 mn
If F = 0 and without damping, the frequency–amplitude relation of the nonlinear free vibration, from Eq.(37), is as
n2¼ 1 þ8H1
3p gþ
3H2
4.2 Nonlinear dynamic buckling Investigate the nonlinear dynamic buckling analysis of imper-fect ES-FGM doubly-curved panels under lateral pressure varying
as linear function of time q0= ct (c is a loading speed) and pre-loaded compressions r0= const, p0= const, Eq.(28)becomes
M€f þ 2Me_f þ D þB2
A
!
ðf f0Þ þ8mnk
2
3p2
B
Ad1d2ðf f0Þf
þ H f2 f2
þ K f 2 f2
f a
2h
p2 r0m2þ p0n2
k2
f
þ 4a
4h
mnp6d1d2ðk1r0þ k2p0Þ ¼ 4a
4
By solving Eq.(39), the dynamic critical time tcrcan be obtained according to Budiansky–Roth criterion[25] This criterion is based
on that, for large value of loading speed, the amplitude-time curve
of obtained displacement response increases sharply depending on time and this curve obtain a maximum by passing from the slope point and at the corresponding time t = tcrthe stability loss occurs Here t = tcris called critical time and the load corresponding to this critical time is called dynamic critical buckling load qdcr= ctcr For static buckling analysis of ES-FGM shallow shell, the explicit expressions of the upper and lower buckling load were obtained in
[21]
5 Numerical results and discussions 5.1 Validation of the present formulation
To validate the present study, the fundamental frequency parameter of unstiffened FGM shallow shells is compared with other studies
Table 1shows the present results in comparison with those pre-sented by Matsunaga [15] based on the two-dimensional (2D) higher-order theory, Chorfi and Houmat [16] accorded to the first-order shear deformation theory and Alijani et al.[17] used Donnell’s nonlinear shallow shell theory In this comparison, the fundamental frequency parameter ~x¼xmnh ffiffiffiffiffiffiffiffiffiffiffiffi
qc=Ec
p
of the perfect un-stiffened FGM shallow shell (a/b = 1, h/a = 0.1) with simply sup-ported edges The material properties are Aluminum and Alumina, i.e Em= 70 109N/m2;qm= 2702 kg/m3and Ec= 380 109N/m2;
q = 3800 kg/m3respectively The Poisson’s ratio is chosen to be
Trang 60.3 As can be seen, a very good agreement is obtained in the
com-parison with the result of Ref.[17], but there are a little differences
with those of Refs [15,16] because they use above mentioned
other theories
5.2 Nonlinear vibration results
To illustrate the proposed approach of eccentrically stiffened
FGM shallow shells, the stiffened and un-stiffened FGM shallow
h = 0.01 m; The shells are simply supported at all its edges The
combination of materials consists of Aluminum Em= 70 109N/
m2,qm= 2702 kg/m3and Alumina Ec= 380 109N/m2qc= 3800
-kg/m3 The Poisson’s ratio is chosen to be 0.3 for simplicity
Mate-rial of stiffeners has elastic modulus Es1= Es1= 380 109N/m2,
qs1=qs2= 3800 kg/m3 The height of stiffeners is equal to 50 mm,
its width 2.5 mm, the spacing of stiffeners s1= s2= 0.1 m
The obtained results inTable 2show that effects of stiffeners on
the fundamental frequencies of natural vibration are considerable
Obviously the natural fundamental frequencies of un-stiffened and
stiffened FGM spherical panels observed to be dependent on the
constituent volume fractions, they decrease when increasing the
power index k, furthermore with greater value k the effect of
stiff-eners is observed to be stronger This is completely reasonable due
to the lower value of the elasticity modulus of the metal
constitu-ent The natural frequencies of stiffened spherical panels are
great-er than one of un-stiffened panels
Table 3shows high frequencies of natural vibration of spherical
panels with R1= R2= 5 m, k = 1 Clearly, all modes of stiffened
spherical panel are greater than ones of un-stiffened panel
Espe-cially, the difference is larger with higher modes
Table 4shows the fundamental frequencies of doubly curved
FGM shallow shells In this case, with k = 1, It seem that the natural
frequencies of un-stiffened panels depend to the value jk1+ k2j, i.e
they have the same value for all panels of same jk1+ k2j, the natural
frequencies of un-stiffened panels increases when this value
in-creases, when k1+ k2= 0 (R1= 5, R2= 5 or R1= 5, R2= 5) the
nat-ural frequencies of un-stiffened panels are equal to the natnat-ural frequencies of un-stiffened plates (seeTables 2 and 4), but as can
be seen, this phenomenon is not observed for stiffened panels For stiffened panels of the same value jk1+ k2j have different natu-ral frequencies Except in special cases when k1+ k2= 0, the natural frequencies of stiffened panels have the same value with stiffened respective plate
The frequency–amplitude curves of nonlinear vibration of FGM spherical panels without damping are presented inFig 2 This fig-ure shows that the frequency–amplitude curves of forced vibration are asymptotic with the frequency–amplitude curve of free vibra-tion and the extreme point of stiffened panel is greater than one
of un-stiffened panel In forced vibration, a value of n corresponds
to the maximum five distinguished values ofg
Fig 3shows the effect of excitation force Q on the frequency– amplitude curves of nonlinear vibration of spherical panels As can be seen, when the excitation force decreases, the curves of forced vibration are closer to the curve of free vibration
Figs 4 and 5show the effect of volume-fraction index and ra-dius of panels on the frequency–amplitude curve of ES-FGM spher-ical panel without damping Clearly, the extreme points of frequency–amplitude curve decrease when the volume-fraction in-dex or the radius decreases
Four cases of Gauss curvature of stiffened and un-stiffened pan-els are taken into consideration: k1k2> 0 and k1+ k2> 0 when
R1= 3 m, R2= 10 m, k1k2> 0 and k1+ k2< 0 when R1= 3 m,
R2= 10 m, k1k2< 0 and k1+ k2> 0 when R1= 3 m, R2= 10 m and k1k2< 0 and k1+ k2< 0 when R1= 3 m, R2= 10 m (Figs 6–9)
It seems that the frequency–amplitude curve depends on the value
k1+ k2 The minimal point of frequency–amplitude curve increases when this value decreases Especially, when k1+ k2< 0, the fre-quency–amplitude curve of free vibration does not exist extreme points (seeFigs 7 and 9)
Table 1
Comparison of ~xwith results reported by Matsunaga [15] , Chorfi and Houmat [16]
and Alijani et al [17]
b/R 2 a/R 1 k Present Ref [15] Ref [16] Ref [17]
FGM plate
FGM spherical panel
FGM cylindrical panel
FGM hyperbolic paraboloidal panel
Table 2 Fundamental frequencies of natural vibration (rad/s) of spherical panels.
Table 3 Frequencies of natural vibration (rad/s) of spherical panels with R 1 = R 2 = 5 m, k = 1.
Un-stiffened Stiffened
x2 (m = 1, n = 2) and (m = 2, n = 1) 2437.29 6743.12
x4 (m = 1, n = 3) and (m = 3, n = 1) 3931.47 14576.20
x5 (m = 2, n = 3) and (m = 3, n = 2) 4920.50 16063.76
Trang 7Fig 10investigate the frequency–amplitude curve of nonlinear
vibration of stiffened plate and shallow shells with jR1j = jR2j, the
phenomenon is similar withFigs 6–9, however when k1+ k2= 0
with R1= 5, R2= 5 and R1= 5, R2= 5 the frequency–amplitude
curves coincide to the frequency–amplitude curve of plate
Consider the un-stiffened and stiffened perfect FGM spherical
panel without damping under the uniformly harmonic load q0
(t) = Q sin (Xt), nonlinear responses are obtained solving the Eq
(30)by fourth order Runge–Kutta method with the time stepDt
Nonlinear responses of un-stiffened FGM spherical panel with difference time steps are presented inFig 11 As can be observed, difference of nonlinear responses of time stepsDt = 2 104and
Dt = 104is very small Therefore, the next results are calculated with time stepDt = 104to ensure the accuracy of this method Nonlinear responses of stiffened and un-stiffened functionally graded spherical panel with k = 1, R = 5 are presented inFigs 12 and 13 Natural frequencies of un-stiffened and stiffened spherical panel are 1809.95 rad/s and 2560.43 rad/s, respectively (see
Table 2) The excitation frequencies are much smaller (Fig 12,
q0(t) = 105sin (100t)) and much greater (Fig 13, q0(t) = 105sin (104t)) than natural frequencies These results show that the stiffeners strongly decrease vibration amplitude of the shell when excitation frequencies are much smaller or much greater than natural frequencies
When the excitation frequencies are near to natural frequen-cies, the interesting phenomenon is observed like the harmonic beat phenomenon of a linear vibration (Figs 14 and 15) The exci-tation frequencies are 2510 rad/s and 2530 rad/s which are very near to natural frequencies 2348.57 rad/s of stiffened spherical pa-nel These results show that the amplitude of beats of stiffened panels increases rapidly when the excitation frequency approaches the natural frequencies The maximal amplitude of harmonic beat increases and the response time of beat decreases when the excita-tion force increases as shown inFig 15
The deflection–velocity relation has the closed curve form as in
Fig 16 Deflection f and velocity _f are equal to 0 at initial time and final time of beat and the contour of this relation corresponds to the period which has the greatest amplitude of beat
Table 4
Fundamental frequencies of natural vibration (rad/s) of doubly curved shallow shells.
jR 1 j jR 2 j k R 1 > 0, R 2 > 0 R 1 > 0, R 2 < 0 R 1 < 0, R 2 > 0 R 1 < 0, R 2 < 0
Unstiffened Stiffened Unstiffened Stiffened Unstiffened Stiffened Unstiffened Stiffened
Fig 2 The frequency–amplitude curve of nonlinear vibration of FGM spherical
panels (R = 5 m, k = 1, Q = 10 5
N/m 2
).
Fig 3 Effect of excitation force Q on the frequency–amplitude curve of stiffened
spherical panel (R = 5 m, k = 1).
Fig 4 Effect of index k on the frequency–amplitude curves of stiffened spherical panels (R = 5 m, Q = 10 5 N/m 2 ).
Trang 8Figs 17 and 18show the effect of pre-loaded compressions and
of known initial amplitude on the nonlinear responses of ES–FGM
spherical panel In these investigations, pre-loaded compressions
and known initial amplitude slightly influence on the amplitude
of nonlinear vibration of panels
Effect of damping on nonlinear responses is presented inFigs 19
and 20with linear damping coefficiente= 0.3 The damping
influ-ences very small to the nonlinear response in the first vibration
periods (Fig 19) however it strongly decreases amplitude at the
next far periods (Fig 20)
5.3 Nonlinear dynamic buckling results
To evaluate the effectiveness of stiffener in the nonlinear dynamic buckling problem, we consider an imperfect ES-FGM cylindrical panel and spherical panel under lateral pressure and pre-loaded compression Materials of shells and stiffeners used in this section are the same in the previous section
The effect of stiffeners to the critical buckling of perfect FGM shallow shells under only lateral pressure is investigated for two cases of cylindrical and spherical panel under uniformly lateral
Fig 5 Effect of radius R on the frequency–amplitude curves of stiffened spherical
panels (k = 1, Q = 10 5 N/m 2 ).
Fig 6 The frequency–amplitude curve of nonlinear vibration of un-stiffened
shallow shells (k = 1, Q = 10 5
N/m 2
).
Fig 7 The frequency–amplitude curve of nonlinear vibration of un-stiffened
shallow shells (k = 1, Q = 10 5
N/m 2
).
Fig 8 The frequency–amplitude curve of nonlinear vibration of stiffened shallow shells (k = 1, Q = 10 5 N/m 2 ).
Fig 9 The frequency–amplitude curve of nonlinear vibration of stiffened shallow shells (k = 1, Q = 10 5
N/m 2
).
Fig 10 The frequency–amplitude curve of nonlinear vibration of stiffened plate and shallow shells (k = 1, Q = 10 5
N/m 2
).
Trang 9pressure varying on time as q0= 105t (N/m2) (seeFigs 21 and 22).
The critical buckling load corresponds to the buckling mode shape
m = 1, n = 1 in all cases These figures also show that there is no
def-inite point of instability as in static analysis Rather, there is a
re-gion of instability where the slope of t vs f/h curve increases
rapidly According to the Budiansky–Roth criterion, the critical
time tcrcan be taken as an intermediate value of this region
There-fore one can choose the inflexion point of curve i.e.d 2 f
dt2
t¼t cr
¼ 0 as in Ref.[26] As can be seen, the dynamic buckling loads of stiffened
panels are greater than one of un-stiffened panels
Effect of volume-fraction index k and loading speed c on critical dynamic buckling of stiffened cylindrical panels and spherical pa-nel are showed inTable 5 Clearly, the critical dynamic buckling
of panel decrease when the volume-fraction index increases or the loading speed decreases As can be also observed inTable 5, the critical dynamic buckling load is greater than the static critical load
Table 6shows the effect of thickness h on the critical dynamic buckling load of cylindrical and spherical panels The critical dy-namic buckling loads of un-stiffened and stiffened panels increase when the thickness of panels increases In addition,Table 6also
Fig 11 Nonlinear responses of un-stiffened FGM spherical panel with difference
time steps (R = 5 m, k = 1, q(t) = 10 5
sin 100t).
Fig 12 Nonlinear responses of FGM spherical panel (R = 5 m, k = 1,
q(t) = 10 5 sin 100t).
Fig 13 Nonlinear responses of FGM spherical panel (R = 5 m, k = 1,
q(t) = 10 5
sin 10 4
t).
Fig 14 Harmonic beat phenomenon of stiffened spherical panel (R = 5 m, k = 1, q(t) = 10 5
sinXt).
Fig 15 Harmonic beat phenomenon of stiffened spherical panel (R = 5 m, k = 1, q(t) = Q sin 2530t).
Fig 16 Deflection–velocity relation of stiffened spherical panel (R = 5 m, k = 1, q(t) = 2 10 5
sinXt).
Trang 10shows that the effect of stiffeners decreases when the thickness
increases
Table 7gives the results on the effect of known initial amplitude
f0on the nonlinear buckling of stiffened cylindrical and stiffened
spherical panels Clearly, the known initial amplitude slightly
influences on the critical dynamic buckling loads of these
struc-tures when they only subjected to lateral pressure
Fig 23shows the dynamic response of stiffened spherical
pan-els under combination of lateral pressure varying on time q0= 105t
(N/m2) and pre-loaded compressions r0= const, p0= const As can
be observed, the pre-loaded compressions strongly influence on
Fig 17 Effect of pre-loaded compressions on nonlinear responses of stiffened
spherical panel (R = 5 m, k = 1, q(t) = 10 5
sin 10 5
t).
Fig 18 Effect of known initial amplitude on nonlinear responses of stiffened
spherical panel (R = 5 m, k = 1, q(t) = 10 5
sin 100t).
Fig 19 Effect of damping on nonlinear responses of stiffened spherical panel
(R = 5 m, k = 1, q(t) = 5 10 3 sin 2530t).
Fig 20 Effect of damping on nonlinear responses of stiffened spherical panel (R = 5 m, k = 1, q(t) = 5 10 3
sin 2530t).
Fig 21 Effect of stiffeners on dynamic buckling of cylindrical panel under lateral pressure (R = 6 m, f 0 = 0, d 1 = d 2 = 3 10 3 m, h 1 = h 2 = 3 10 2 m, s 1 = s 2 = 0.25 m).
Fig 22 Effect of stiffeners on dynamic buckling spherical panel under lateral pressure (R = 10 m, f 0 = 0, d 1 = d 2 = 3 10 3
m, h 1 = h 2 = 4 10 2
m,
s 1 = s 2 = 0.25 m).
Table 5 Effect of volume fraction index k, loading speed c on the critical dynamic buckling load of ES-FGM shallow panels (10 5 N/m 2 ).
Cylindrical panels
Dynamic q 0 = 10 5
Spherical panels
Dynamic q 0 = 10 5
Dynamic q 0 = 10 6 t 17.8372 12.2372 6.6619 5.4012
R = 5 m, d = d = 3 10 3 m, h = h = 35 10 3 m, s = s = 0.25 m.