Journal of Research in Engineering and Applied Sciences

KANTOROVICH VARIATIONAL METHOD FOR THE

ELASTIC BUCKLING ANALYSIS OF KIRCHHOFF
PLATES WITH TWO OPPOSITE SIMPLY SUPPORTED
EDGES
1Charles Chinwuba Ike and 2Onyedikachi Aloysius Oguaghamba
1
Associate Professor, Department of Civil
Engineering, Enugu State University of Science
and Technology, Enugu, 400001, Nigeria;
2
Senior Lecturer, Department of Civil
Engineering, University of Nigeria, Nsukka,
410001, Nigeria.
1
Email: { [email protected],
2
[email protected]}

Abstract
We present the Kantorovich variational method for the elastic buckling solutions of Kirchhoff plates with two opposite simply supported edges. Uniform
compressive loading is applied on the two simply supported edges, x = 0 and x = a. Three cases of edge support considered along the y = 0 and y = b edges are
(i) both are clamped (ii) both are simply supported, and (iii) edge y = 0 is simply supported while edge y = b is free. The Kantorovich method which is based
on seeking to minimize the total potential energy functional Π with respect to the unknown deflection w(x, y) assumes the deflection in variable separable form
as an infinite series of the product of an unknown function fm(y) and a known sinusoidal function that satisfies the Dirichlet boundary conditions along the
simply supported edges. Euler-Lagrange differential equation is used to obtain the fourth order ordinary differential equation which fm(y) must satisfy to
minimize Π. The general solution for fm(y) is obtained as a combination of hyperbolic and trigonometric functions with four unknown integration constants.
The conditions for nontrivial solutions are applied to find the characteristic stability equations in each considered case. Exact solutions are obtained for the
elastic buckling equations which are identical with previously obtained solutions in the literature. The elastic buckling equations are solved by iteration
methods and by closed form solution methods for the Kirchhoff plate to obtain the elastic buckling loads expressed in terms of the elastic buckling load
coefficients. It was found that the method yields exact elastic buckling load and exact elastic buckling load coefficient for each boundary condition considered
in this study.

Key Words - Kantorovich variational method, Kirchhoff plate, Total potential energy functional, Euler – Lagrange differential
equation, Characteristic elastic buckling equation

1. Introduction in-elastic buckling of plates of various shapes, sizes

subjected to different types of loads and for different
Elastic buckling problems of plates occur for plates boundary conditions have been reported.
subject to in-plane compressive loads which may be Some research work that reported significant
uniformly or non-uniformly distributed [1–7]. contributions to the elastic and inelastic buckling research
Navier presented a derivation of the governing partial include: Timoshenko [1], Timoshenko and Gere [2], Bulson
differential equation PDE for the problem of isotropic, [3], Chajes [4], Gambhir [5], Wang et al [6], Shi and Bezine
homogeneous, thin rectangular plates. His derivation which [7], Shi [8], Abodi [9], Batford and Houbolt [10], Wang et al
laid the groundwork for much of thin plate elastic buckling [11], Ullah et al [12, 13], Xiang et al [14], Yu Chen [15] and
research used the Kirchhoff’s assumptions and included the Abolghasemi et al [16].
twisting term which represented a significant contribution to Ibearugbulem [17] used the Ritz direct variational
the knowledge and theory of elastic stability at the time. method to determine buckling loads of isotropic,
Saint Venant formulated an improvement on the Navier’s homogeneous, thin flat plates under to in-plane uniform
governing domain PDE of elastic buckling of isotropic compressive loading for various boundary conditions.
homogeneous thin rectangular plates. Though one parameter shape functions were used he
Since the work of Navier and Saint Venant several other
ground breaking research work on the theme of elastic and
ISSN (Print): 2456-6411 | ISSN (Online): 2456-6403 596 JREAS, Vol. 08, Issue 03, July 23
obtained solutions that were not significantly different from Finite Fourier sine integral transform with respect to x
the previous results from literature. coordinate variable converted the governing domain PDE to
Nwadike [18] also used the Ritz method to find buckling an integral equation over the plate domain, which further
loads of isotropic, homogeneous thin flat plates under in- simplified to an ordinary differential equation (ODE). They
plane uniaxial uniform compressive forces for various solved the ODE subject to the clamped edge conditions
boundary conditions. Using one parameter shape functions, obtaining the characteristic elastic buckling equations using
he obtained reasonably good results that compared well with the requirements for non-trivial solutions on the resulting
solutions from previous literature. homogeneous set of equations. They solved the characteristic
Oguaghamba [19] investigated the buckling and post buckling equation which is a transcendental equation using
buckling loads of isotropic, homogeneous thin rectangular computer based iteration tools to obtain exact solutions for
flat plates. Oguaghamba et al [20] studied the buckling and the buckling loads that agreed with previous solutions in the
post buckling load characteristics of isotropic, homogenous literature.
thin rectangular flat plates with all edges clamped. Iyengar [25] solved the problem of buckling of thin
Abodi [9] presented a finite difference method for plates for various support conditions, and obtained exact
solving the elastic buckling problem of simply supported solutions. for the case of Dirichlet boundary conditions.
thin rectangular plates under in-plane patch loads, and Onyia et al [26] applied Galerkin-Kantorovich method
obtained the elastic buckling loads, by solving the resulting for solving stability problems of thin SCSC plates. Onyia et
finite difference equations. al [27] used the one-dimensional finite Fourier sine integral
Shi and Bezine [7] and Shi [8] solved the stability transform method to obtain exact solutions to the elastic
problems of orthotropic plates with the aid of boundary stability problems of rectangular thin SSSS and SSCF plates.
element method. Onyia et al [28] has used the Galerkin-Vlasov variational
Ibearugbulem et al [21] and Osadebe et al [22] used the method for solving stability problems of rectangular thin
Taylor – Maclaurin’s series to formulate displacement shape SSCF and SSSS plates, and obtained exact solutions.
functions for simply supported Euler – Bernoulli beams and Oguaghamba and Ike [29] used the Galerkin-Vlasov
applied such formulated shape functions in the Galerkin’s method for the analysis of the stability problems of thin
method to find the elastic buckling loads of simply supported plates with three simply supported edges and one free edge
isotropic, homogeneous thin rectangular plates under in- for the case of uniform uniaxial compression.
plane uniaxial, uniform compressive loads. In this work, the Kantorovich variational method is
Nwoji et al [23] applied two dimensional finite Fourier presented for solving stability problems of Kirchhoff plates
sine integral transform method to determine the elastic with two opposite simply supported edges subjected to
buckling loads of thin rectangular plates under uniaxial and uniform in-plane compressive load; while the other two
biaxial uniform compressive loads for the case of Dirichlet edges are: clamped (SCSC or CSCS plate) or simply
boundary conditions along all the four edges. For the case of supported (SSSS plate) or one edge is simply supported and
uniaxial compressive load, they considered the load applied the other edge is free (SSSF plate). Hence Kantorovich
along the two opposite edges x = 0 and x = a, (where a variational method is presented in this work for solving the
corner point is considered the origin) while for biaxial elastic buckling problems of SCSC (or CSCS). Kirchhoff
uniform compressive load, the loads were applied along x = plates, SSSS Kirchhoff plates and SSSF Kirchhoff plates.
0, x = a, y = 0 and y = b. They found the integral kernel
function of the double finite Fourier sine transformation 1.1. Justification for the study
ideally suited to the problem since the Dirichlet boundary
Review of literature shows that the variational Kantorovich
conditions were identically satisfied along all the boundaries.
method had not been previously applied to the solution of the
They found that the double integral transform converts the
elastic buckling problems of Kirchhoff plates with two
elastic buckling problem represented in equilibrium form as
opposite simply supported edges. The study is thus novel as
a PDE to an integral equation, and ultimately to an algebraic
it demonstrates the effectiveness of the variational
problem that gives closed form expressions for stability
Kantorovich methodology as a potent mathematical tool for
problem for all the buckling modes.
solving the elastic buckling problem of thin plates posed in
Onah et al [24] used one-dimensional finite Fourier sine
variational form.
integral transform method to obtain the elastic buckling loads
of thin rectangular plates with two opposite edges (x = 0 and
x = a) simply supported and the other edges (y = 0 and y = b) 2. Theoretical Framework
clamped with uniform compressive force applied at the The total potential energy functional Π for a rectangular
simply supported edges. They found the one-dimensional Kirchhoff plate subjected to in-plane edge compressive
finite Fourier sine integral transform used ideally fitted to the forces Nx, Ny and shear force Nxy is:
problem since the sinusoidal kernel function used satisfied
the Dirichlet boundary conditions. Π = U +V (1)

ISSN (Print): 2456-6411 | ISSN (Online): 2456-6403 597 JREAS, Vol. 08, Issue 03, July 23
in which U is the strain energy, and V is the potential energy
of the loading.
For isotropic homogeneous thin rectangular plates, U is
expressed as:
U=
1
2 ∫∫
A
{ 2
(
D ( ∇ 2 w ) − 2(1 − µ ) w xx w yy − ( w xy )2 )} dxdy (2)

wherein w ( x, y ) denotes deflection and A represents the

Figure 1: Rectangular thin plate with simply supported
plate domain, µ denotes the Poisson's ratio of the plate boundaries (x = 0, x = a) under uniform compressive load
material, D is the plate flexural rigidity given by while edges are subject to various boundary conditions
Eh3
D= (3) The problem considered shown in the Figure 1 is a thin
12(1 − µ 2 )
plate having the two opposite sides x = 0 and x = a simply
wherein E denotes the Young’s modulus of elasticity, h is
supported, and the other sides y = 0, and y = b subject to
the thickness of the plate.
three cases of boundary conditions namely: (i) both are
The potential function of the external forces is:
clamped, (ii) both are simply supported and (iii) edge y = 0
1  ∂w ∂w 
2 2
∂w   ∂w 
V = − ∫∫  N xx 
is simply supported and edge y = b is free. The compression
 + N yy   + 2 N xy dxdy
2   ∂x   ∂y  ∂x ∂y  force is applied only along the simply supported sides while
the other sides are not subject to any loading. Thus Nyy = 0,
…(4) and Nxy = 0.
Hence,
The energy functional then simplifies to:
D  ∂ 2 w ∂ 2 w    2 2 
2
∂ w ∂ 2 w ∂ 2 w  dxdy −
Π= ∫∫  2 + 2  + 2(1 − µ)    − 2
∂x ∂y 2  
 D
{
Π = ∫∫ (∇ 2 w )2 − 2(1 − µ )( w xx w yy − w xy
2
}) dxdy −
  ∂x∂y 
2  ∂x ∂y 
 2 A
1
2 ∫∫
1   ∂w 
2
 ∂w 
2
∂w ∂w  N xx ( w x )2 dxdy (7)
∫∫
2 
 N xx 
 ∂x 
 + N yy 
 ∂y 
 + 2 N xy  dxdy
∂x ∂y 
A
The objective of stability problems of thin plates is to
…(5) find minimum total potential energy functional with respect
Simplifying, to w(x, y).
D
2
{
Π = ∫∫ (∇ 2 w )2 + 2(1 − µ )( w xy 2
}
− w xx w yy dxdy − Three cases considered are (i) the sides y = 0 and y = b
are clamped (Figure 2) in which case the clamped boundary
1
2 ∫∫ { }
N xx ( w x )2 + N yy ( w y )2 + 2 N xy w x w y dxdy (6) conditions are:
w ( x, y = 0) = w( x, y = b ) = 0
where ∇2 is the Laplacian expressed by: (8)
w y ( x, y = 0) = w y ( x, y = b ) = 0
∂2 ∂2
∇2 = 2 + 2
∂x ∂y and (ii) the sides y = 0 and y = b are simply supported
∂ w
2 (Figure 3) with corresponding boundary conditions
w xy = w ( x, y = 0) = w( x, y = b ) = 0
∂x∂y (9)
w yy ( x, y = 0) = w yy ( x, y = b ) = 0
∂w
wx =
∂x
∂w
wy =
∂y
∂2w
w xx =
∂x 2
∂2 w
w yy = 2
∂y
The elastic buckling problem considered in this work is Figure 2: CSCS Kirchhoff plate under uniform uniaxial
shown in Figure 1. compression Nxx on the simply supported sides

ISSN (Print): 2456-6411 | ISSN (Online): 2456-6403 598 JREAS, Vol. 08, Issue 03, July 23
 ∞
mπx
w ( x, y ) = ∑ fm ( y)sin
a
(12)
m =1
The problem then reduces to finding the unknown
function fm(y) that extremizes the total potential energy
functional Π and also satisfies the boundary conditions
along the edges y = 0 and y = b.
Figure 3: SSSS Kirchhoff plate under uniform uniaxial
compression Nxx on the simply supported sides 3.2. Total potential energy functional

By differentiation of w(x, y), we obtain:

∞
(iii) Three sides x = 0, x = a, and y = 0 are on simple mπ mπx
supports while the side y = b is free as shown in Figure 4. wx = ∑ fm ( y )cos (13)
m =1 a a
∞ 2
 mπ  mπx
w xx = ∑ −  a  fm ( y )sin a (14)
m =1
∞
 mπ  mπx
w xy = ∑  a 
 fm′ ( y )cos
a
(15)
m =1

dfm ( y )
Figure 4: SSSF Kirchhoff plate simply supported where fm′ ( y ) =
dy
along three edges x = 0, x = a and y = 0,
∞ ∞
and free along y = b mπx df ( y ) mπx
wy = ∑ fm′ ( y )sin a
=∑ m
dy
sin
a
(16)
m =1 m =1
The SSSF Kirchhoff plate is under uniform uniaxial ∞ ∞
d 2 fm ( y ) mπx mπx
compression along the simply supported sides shown in
Figure 4.
w yy = ∑ dy 2
sin
a
= ∑ fm′′ ( y ) sin
a
(17)
m =1 m =1
The boundary conditions are: ∞  2

 mπ  mπx
w ( x = 0, y ) = w ( x = a, y ) = 0 ∇ 2 w = w xx + w yy = ∑  fm′′( y) −   fm ( y )  sin
a   a
m =1
∂2w ∂2 w
( x = 0, y ) = ( x = a, y ) = 0 (10) …(18)
∂x 2 ∂x 2 2
w ( x, y = 0) = 0  ∞  mπ 
2
 mπx 
(∇ w ) =  ∑  fm′′( y ) − 
2 2
 fm ( y )  sin  (19)
M yy ( x, y = 0) = − D( w yy + µw xx ) =0  m =1   a   a 
x, y = 0
Then,
M yy ( x, y = b ) = − D( w yy + µw xx ) =0
x, y = b D   mπ 
2

Vy ( x, y = b ) = − D( w yyy + (2 − µ )w xxy ) =0 Π = ∫∫ ∑∑ ( fm′′ ( y ))2 − 2   fm′′ ( y ) fm ( y ) +

x, y = b 2   a 
∂3w ∂3 w  mπ  2  2 mπx
4
where w xxy = , w yyy =   fm ( y )  sin − 2(1 − µ ) ×
∂x 2 ∂y ∂y 3  a   a
Myy is bending moment, Vy is the effective shear force.  2
 mπ  2 m πx
∑∑ −   fm ( y ) fm′′ ( y )sin −
3. Methodology   a  a
 mπ 
2
2 m πx 

3.1. Deflection function −∑∑   ( fm′ ( y )) cos
2
 dxdy −
 a  a 
The expression for buckled deflection is chosen in 2
N xx  mπ  2 m πx
variable separable form as the infinite series:
∫∫ ∑∑  a  ( fm ( y)) cos 2 dxdy
2
(20)
2
∞
w ( x, y ) = ∑ fm ( y) gm ( x ) (11) Simplifying,
m =1
 b D  mπ 
2
m = 1, 2, 3,..., ∞ Π = ∑∑  ∫ ( fm′′ ( y ))2 − 2   fm′′ ( y ) fm ( y ) +
where gm(x) is chosen to apriori satisfy Dirichlet conditions  0 2   a 
for x = 0, and x = a and fm((y) is the function sought such
 mπ  2 
4 2
 mπ 
that Π is minimized. Thus w(x, y) is considered as the  + − µ  {( fm′′ ( y )) I 2 +
2
  mf ( y ) I1 2(1 ) 
infinite series expression.  a    a 

ISSN (Print): 2456-6411 | ISSN (Online): 2456-6403 599 JREAS, Vol. 08, Issue 03, July 23
 2
N xx  mπ  2   mπ 
2
 mπ 
2

fm ( y ) fm′′ ( y ) I1} −   fm ( y ) I 2  dy (21) 2  m f ( y ) I1 + 2(1 − µ )   fm ( y ) I1  = 0
2  a    a   a  
a
mπx …(29)
I1 = ∫ sin 2 dx (22) Simplifying,
a 4 2
 mπ   mπ 
0
a
mπx   2 fm ( y ) I1 + 2(1 − µ )   fm′′ ( y ) I1 −
where I 2 = ∫ cos2 dx (23)  a   a 
a 2 2
0
 mπ  N xx  mπ 
 b 
2  fm′′ ( y ) I1 −   2 fm ( y ) I 2 −
2Π
2
mπ   a  D  a 
= Π * = ∑∑  ∫ ( fm′′ ( y ))2 − 2   fm′′ ( y ) fm ( y ) +
D  0   a   mπ 
2
 mπ 
2
4(1 − µ )   fm′′ ( y ) I 2 + 2 fm ( y ) I1 − 2   fm′′ ( y ) I1 +
iv
 a   a 
 mπ  2 
4 2
 mπ 
 + − µ  ( ( fm′ ( y )) I 2 +
2
  mf ( y ) I1 2(1 )   mπ 
2
 a    a  2(1 − µ )   fm′′( y ) I1 = 0 (30)
2
  a 
N  mπ  2
fm ( y ) fm′′ ( y ) I1 ) − xx   fm ( y ) I 2  dy Simplifying further,
D  a   4 2
 mπ   mπ 
 2 fm ( y ) I1 − 4   fm′′ ( y ) I1 + 2 fm ( y ) I1 −
iv
b

= ∑∑ ∫ F ( y, fm ( y ), fm′ ( y ), fm′′ ( y ))dy (24)  a   a 
2 2
0
N xx  mπ   mπ 
where F ( y, fm ( y ), fm′ ( y ), fm′′ ( y )) is the integrand in the   2 fm ( y ) I 2 + 4(1 − µ )   fm′′( y ) I1 −
D  a   a 
expression for Π*. 2
 mπ 
4(1 − µ )   fm′′ ( y ) I 2 = 0 (31)
3.3. Euler – Lagrange differential equation for the  a 
extremum of Π* and Π Further simplification gives:
4 2
 mπ   mπ 
 2 fm ( y ) I1 − 4   fm′′ ( y ) I1 + 2 fm ( y ) I1 −
iv
For extremizing Π* and hence Π we have: 
 a   a 
∂F d ∂F d 2 ∂F N xx  mπ 
2
 mπ 
2
− + 2 =0 (25)   2 fm ( y ) I 2 + 4(1 − µ )   fm′′ ( y )( I1 − I 2 ) = 0
∂fm ( y ) dy ∂fm′ ( y ) dy ∂fm′′ ( y ) D  a   a 
Differentiation of F gives: …(32)
∂F  mπ 
4
 mπ 
2 But, I1 = I 2 = I* (33)
=  2 fm ( y ) I1 + 2(1 − µ)   fm′′( y ) I1 − Then,
∂fm ( y )  a   a 
4 2
 mπ   mπ 
 fm′′ ( y ) I* + 2 fm ( y ) I* −
2 2
N xx  mπ   mπ    2 f m ( y ) I* − 4 
iv
  2 fm ( y ) I 2 − 2   fm′′ ( y ) I1 (26)  a   a 
D  a   a  2
2 N xx  mπ 
∂F  mπ    2 f m ( y ) I* = 0 (34)
= 2(1 − µ )   2 fm′′ ( y ) I 2 (27) D  a 
∂fm′ ( y )  a 
Division by 2I* yields the following differential equation:
∂F   mπ 
2
 4 2
=  2 fm′′( y ) − 2   fm ( y )  I1 +  mπ   mπ 
 fm ( y ) − 2   fm′′ ( y ) + fm ( y ) −
iv
∂fm′′ ( y )   a   
 a   a 
2
 mπ 
2
N xx  mπ 
2(1 − µ )   fm ( y ) I1 (28)   fm ( y ) = 0 (35)
 a  D  a 
Then, we obtain: Simplifying, we have the system of ordinary differential
4 2
 mπ   mπ  equations:
  2 fm ( y ) I1 + 2(1 − µ )   fm′′ ( y ) I1 − 2
  mπ 4
 a   a   mπ 
fmiv ( y ) − 2   m ′′ +  −
  a 
f ( y )
 mπ 
2
N xx  mπ 
2  a  
2  fm′′ ( y ) I1 −   2 fm ( y ) I 2 −
 a  D  a  N xx  mπ  
2

   fm ( y ) = 0 (36)
d   d2  D  a  
2
 mπ 
 2(1 − µ)   2 fm′′ ( y ) I 2  + 2  2 fm′′ ( y ) I1 −
dx   a   dx 
4. Results

ISSN (Print): 2456-6411 | ISSN (Online): 2456-6403 600 JREAS, Vol. 08, Issue 03, July 23
4.1. Solution of the ODE fm ( y ) = c1m cosh λ1m y + c2 m sinh λ1m y +

The Euler – Lagrange differential equation is obtained as c3m cos λ 2 m y + c4 m sin λ 2 m y (48)
a system of homogeneous ODEs in fm(y). The ODE is where c1m, c2m, c3m and c4m are the four sets of integration
solved using the methods for mathematical solutions of constants which are obtained from the boundary conditions
ODEs. By the method of trial functions, let us assume a along the unloaded boundaries.
solution for fm(y) in the exponential form:
Case 1: Solution for CSCS (or SCSC) plates
fm ( y ) = exp λy (37)
For CSCS or SCSC plates with two simply supported
wherein λ is a yet to be found parameter.
edges and two clamped sides, we have from Equation (8)
Then the ODE becomes: the boundary
 4 N xx  mπ    λy fm (0) = 0
 mπ  2   mπ 
2 4 2
λ − 2   λ +   −    e = 0 fm ( b ) = 0
  a   a  D  a   
…(38) fm′ (0) = 0
λy
For useful solutions, e ≠ 0. The auxiliary equation is fm′ ( b ) = 0
obtained as the fourth degree polynomial in λ, given by: (49)
2 4 2 wherein,
 m π  2  mπ  N xx  mπ 
λ4 − 2   λ +  −   =0 (39) fm′ ( y ) = c1m λ1m sinh λ1m y + c2 m λ1m cosh λ1m y −
 a   a  D  a  (50)
The solution is obtained from simplification as: c3 m λ 2 m sin λ 2 m y + c4 m λ 2 m cos λ 2 m y
2 Enforcement of the boundary conditions yield:
  mπ  
2
N xx  mπ 
2
 λ2 −    =   (40) fm ( y = 0) = c1m + c3m = 0 (51)
  a   D  a 
2 2 fm′ ( y = 0) = c2 m λ1m + c4 m λ 2 m = 0 (52)
 mπ  N xx  mπ 
λ2 −   =   (41) fm ( y = b ) = c1m cosh λ1m b + c2 m sinh λ1m b +
 a  D  a  (53)
c3 m cos λ 2 m b + c4 m sin λ 2 m b = 0
2 2
 mπ  N xx  mπ  fm′ ( y = b ) = c1m λ1m sinh λ1m b + c2 m λ1m cosh λ1m b −
λ2 =   ±   (42)
 a  D  a  (54)
c3m λ 2 m sin λ 2 m b + c4 m λ 2 m cos λ 2 m b = 0
2 2 2 2
 mπ  N xx  mπ  N xx  mπ   mπ  In matrix form, the system of four homogeneous
λ =
2
 +   =   +  equations becomes:
 a  D  a  D  a   a 
…(43)  1 0 1 0   c1m   0 
    
 λ λ
N xx  mπ  
2 2 2 2  0 0   c2 m  =  0 
 mπ  N xx  mπ    mπ  1m 2m
λ2 =   −   = − −
  a  +    cosh λ1m b sinh λ1m b cos λ 2 m b sin λ 2 m b   c3 m   0 
 a  D  a   D  a   
λ λ λ λ −λ λ λ
    
 1m sinh 1m b 1m cosh 1m b 2m sin 2m b 2 m cos λ 2 m b   c4 m   0
…(44) …(55)
 N 2 2 For nontrivial solutions,
xx  m π   mπ  
λ 2 = −    −  (45)  c1m 
 D  a   a    
Hence the four roots of the auxiliary polynomial are:  c2 m  ≠ 0
1/ 2  c3 m 
 N 2 2
 
xx  m π   mπ  
λ1m =    +   (46)  c4 m 
 D  a   a   The condition for nontrivial solution is the vanishing of
and the complex conjugate roots are ±iλ 2 , where the determinant of the coefficients matrix yielding the
1/ 2 elastic stability equation as:
 N 2 2
xx  m π   mπ  
λ2m =    −   (47) 1 0 1 0
 D  a   a   0 λ1m 0 λ2m
i is the imaginary number, i = −1 =0
cosh λ1m b sinh λ1m b cos λ 2 m b sin λ 2 m b
λ1m sinh λ1m b λ1m cosh λ1m b −λ 2 m sin λ 2 m b λ 2 m cos λ 2 m b

4.2. General solution for fm(y) …(56)

The elastic stability equation can be obtained using a
The general expression for fm(y) is obtained as: simpler procedure that reduces the elastic stability problem

ISSN (Print): 2456-6411 | ISSN (Online): 2456-6403 601 JREAS, Vol. 08, Issue 03, July 23
to two equations in two unknowns. This is accomplished by λ12m sin λ 2 m b sinh λ1m b + λ1m λ 2 m sin 2 λ 2 m b = 0 (67)
expressing c3m in terms of c1m and c4m in terms of c2m.
Simplifying again,
Thus, using λ1m λ 2 m (cosh 2 λ1m b − sinh 2 λ1m b ) + λ1m λ 2 m (cos2 λ 2 m b + sin 2 λ 2 m b ) −
c3m = −c1m (57)
2λ1m λ 2 m cosh λ1m b cos λ1m b + λ12m sin λ 2 m b sinh λ1m b −
c2 m λ1m
and c4 m = − (58) λ 22 m sinh λ1m b sin λ 2 m b = 0 (68)
λ2m
Using the trigonometric identities,
The system of homogeneous equations becomes:
cosh 2 λ1m b − sinh 2 λ1m b = 1 (69)
(cosh λ1m b − cos λ 2 m b )c1m +
cos λ 2 m b + sin λ 2 m b = 1
2 2
(70)
 λ  (59)
c2 m  sinh λ1m b − 1m sin λ 2 m b  = 0 Then,
 λ 2m  λ1m λ 2 m (1 + 1 − 2 cosh λ1m b cos λ1m b ) =
and λ 22 m sinh λ1m b sin λ 2 m b − λ12m sin λ 2 m b sinh λ1m b (71)
c1m (λ1m sinh λ1m b + λ 2 m sin λ 2 m b ) + Simplifying,
(60)
c2 m (λ1m cosh λ1m b − λ1m cos λ 2 m b ) = 0 λ1m λ 2 m (2 − 2 cosh λ1m b cos λ1m b ) =
Hence in matrix form: λ 22 m sinh λ1m b sin λ 2 m b − λ12m sin λ 2 m b sinh λ1m b (72)
  λ1m  
 (cosh λ1m b − cos λ 2 m b )  sinh λ1m b − λ sin λ 2 m b    c1m   0  Dividing by λ1mλ2m gives:
    = 
2(1 − cosh λ1m b cos λ1m b ) =
2m
   c2 m   0 
 (λ1m sinh λ1m b + λ 2 m sin λ 2 m b ) (λ1m cosh λ1m b − λ1m cos λ 2 m b ) 
…(61) λ 22 m λ12m
sinh λ1m b sin λ 2 m b − sin λ 2 m b sinh λ1m b
c  λ1m λ 2 m λ1m λ 2 m
For nontrivial solution,  1m  ≠ 0, and the elastic stability
 c2 m  …(73)
equation is obtained as: Hence,
2(1 − cosh λ1m b cos λ1m b ) =
 λ1m 
(cosh λ1m b − cos λ 2 m b )  sinh λ1m b − λ sin λ 2 m b 
  =0
λ2m λ
2m sinh λ1m b sin λ 2 m b − 1m sin λ 2 m b sinh λ1m b
(λ1m sinh λ1m b + λ 2 m sin λ 2 m b ) (λ1m cosh λ1m b − λ1m cos λ 2 m b ) λ1m λ2m
…(62) …(74)
Expansion of Equation (62) gives: Simplifying further yields the transcendental equation:
λ λ 
(cosh λ1m b − cos λ 2 m b )λ1m (cosh λ1m b − cos λ 2 m b ) − 2(1 − cosh λ1m b cos λ1m b ) =  2 m − 1m  sinh λ1m b sin λ 2 m b
λ
 1m λ 2m 
 λ1m 
 sinh λ1m b − λ sin λ 2 m b  (λ1m sinh λ1m b + λ 2 m sin λ 2 m b ) = 0 …(75)
 2m  where λ1m and λ2m are expressed by Equations (46) and (47).
…(63)
The characteristic elastic stability equation is a
λ1m (cosh λ1m b − cos λ 2 m b )2 − transcendental equation which is solved using computer
 λ 2 m sinh λ1m b − λ1m sin λ 2 m b  software based iteration methods; and or manual iteration
 λ 2m  ( λ1m sinh λ1m b + λ 2 m sin λ 2 m b ) = 0 methods such as the Newton iteration, Newton – Raphson’s
 
…(64) iteration.
The solution for the elastic buckling load coefficients
λ1m (cosh λ1m b + cos λ 2 m b − 2 cosh λ1m b cos λ 2 m b ) −
2 2
obtained for µ = 0.25, and for aspect ratios (a/b) for CSCS
 λ λ sinh 2 λ1m b + λ 22 m sinh λ1m b sin λ 2 m b −  (or SCSC) plates are shown in Table 1, which also shows the
 1m 2 m  previous results obtained by Timoshenko [1], Onah et al [24]
 λ1m sin λ 2 m b sinh λ1m b − λ1m λ 2 m sin λ 2 m b 
2 2
 =0 and Onyia et al [26]. The corresponding solutions for a
 λ2m  Poisson’s ratio µ = 0.30 for various plate aspect ratios are
…(65) presented in Table 2.
Multiplication by λ2m gives:
λ1m λ 2 m (cosh 2 λ1m b + cos2 λ 2 m b − 2 cosh λ1m b cos λ 2 m b ) −
Table 1
( λ1m λ 2 m sinh 2 λ1m b + λ 22 m sinh λ1m b sin λ 2 m b − Elastic buckling load coefficients of SCSC (or CSCS) Kirchhoff
λ12m sin λ 2 m b sinh λ1m b − λ1m λ 2 m sin 2 λ 2 m b ) = 0 (66) plate under uniform in-plane compressive load applied at the edges
x = 0, x = a for (µ = 0.25)
Simplifying, a/b K ( a/ b ) = N cr b 2 / π2 D K(a/b)
λ1m λ 2 m cosh 2 λ1m b + λ1m λ 2 m cos2 λ 2 m b − 2λ1m λ 2 m cosh λ1m b cos λ 2 m b − Onah et al [24], Onyia et al
(Present study)
λ1m λ 2 m sinh 2 λ1m b − λ 22 m sinh λ1m b sin λ 2 m b + [26], Timoshenko [1]

ISSN (Print): 2456-6411 | ISSN (Online): 2456-6403 602 JREAS, Vol. 08, Issue 03, July 23
 0.4 9.448 9.448 Substitution into Equation (80) gives:
0.6 7.055 7.055 c1m λ12m + c1m λ 22 m = c1m (λ12m + λ 22 m ) = 0 (83)
0.8 7.304 7.304
Hence, since λ12m + λ 22 m ≠ 0,
1.0 7.691 7.691
1.2 7.055 7.055 c1m = 0 (84)
1.4 7.001 7.001 c3 m = 0 (85)
1.6 7.304 7.304 Then the system of homogeneous equations reduces to
1.8 7.055 7.055 c2 m sinh λ1m b + c4 m sin λ 2 m b = 0 (86)
2 6.972 6.972
c2 m λ12m sinh λ1m b − c4 m λ 22 m sin λ 2 m b = 0 (87)
Or,
Table 2  sinh λ1m b sin λ 2 m b   c2 m   0 
Elastic buckling load coefficients of SCSC (or CSCS) Kirchhoff  2   =  (88)
 λ1m sinh λ1m b −λ 2 m sin λ 2 m b   c4 m   0 
2
plate under uniform uniaxial in-plane compressive loads applied at
x = 0, x = a for µ = 0.30 c 
a/b K ( a/ b ) K(a/b) Onah et al [24] For useful solutions,  2 m  ≠ 0 the elastic buckling
 c4 m 
0.2 27.86 27.86
equation becomes:
0.4 9.46 9.46
sinh λ1m b sin λ 2 m b
0.8 7.44 7.44 =0 (89)
1.0 7.69 7.69 λ1m sinh λ1m b −λ 22 m sin λ 2 m b
2

2 7.04 7.04 Expanding the determinant gives:

2 6.99 6.99 −λ 22 m sin λ 2 m b sinh λ1m b − λ12m sinh λ1m b sin λ 2 m b = 0 (90)
6 7.02 7.02 Simplifying,
3.2 6.98 6.98 − sin λ 2 m b(λ 22 m sinh λ1m b + λ12m sinh λ1m b ) = 0 (91)
12 6.99 6.99
or, sin λ 2 m b sinh λ1m b( −λ 22 m − λ12m ) = 0 (92)
4.5 6.98 6.98
Solving, for nontrivial solutions,
sin λ 2 m b = 0 (93)
Case 2: Solutions for buckling of SSSS Kirchhoff plates Hence,
The boundary conditions are given alternatively using λ 2 m b = sin −1 0 = nπ (94)
fm(y) as follows: n = 1, 2, 3, 4, …
fm ( y = 0) = 0 nπ
λ 2m = (95)
fm ( y = b ) = 0 b
From Equation (47) we have
d 2 fm ( y = 0)
fm′′ ( y = 0) = =0
1/ 2
(76)  N 2 
dy 2 xx  m π   mπ   nπ
λ 2 m =    −  = (96)
d fm ( y = b )
2  D  a   a  b
fm′′ ( y = b ) = =0 Squaring both sides, we obtain:
dy 2
2 2 2
By differentiation  nπ  N xx  mπ   mπ 
λ 22 m =   =   −  (97)
fm′′( y ) = c1m λ12m cosh λ1m y + c2 m λ12m sinh λ1m y −  b  D  a   a 
c3 m λ 22 m cos λ 2 m y − c4 m λ 22 m sin λ 2 m y (77) N xx  mπ 
2
 m π   nπ 
2 2

Enforcement of boundary conditions gives:

Hence,   =  +  (98)
D  a   a   b 
fm ( y = 0) = c1m + c3m = 0 (78) Squaring both sides yields:
fm ( y = b ) = c1m cosh λ1m b + c2 m sinh λ1m b + 2
N xx  mπ 
2
  m π  2  nπ  2 
c3m cos λ 2 m b + c4 m sin λ 2 m b = 0 (79)   =   +   (99)
D  a   a   b  
fm′′( y = 0) = c1m λ12m − c3m λ 22 m =0 (80) Simplifying,
fm′′( y = b ) = c1m λ12m cosh λ1m b + c2 m λ12m sinh λ1m b − 2
 a    m π   nπ  
2 2 2
N xx = D     +   (100)
c3m λ 22 m cos λ 2 m b − c4 m λ 22 m sin λ 2 m b = 0 (81)  mπ    a   b  
From Equation (78), Then,
c1m = −c3m (82)

ISSN (Print): 2456-6411 | ISSN (Online): 2456-6403 603 JREAS, Vol. 08, Issue 03, July 23
 2
2
a 1   m 2  n 2 
N xx = D   2 ⋅ π4    +    (101) Case 3: Solution for elastic buckling of SSSF Kirchhoff
m π  a   b   plate
Hence,
2
The elastic buckling problem of SSSF Kirchhoff plate in
1   m 2  n 2  Figure 4 is considered.
N xx = Dπ 2 a 2 2   +    (102)
m  a   b   The boundary conditions along the edges y = 0 and y = b
So, are given by Equation (10).
1  m4 m 2 n2 n 4  The boundary conditions are expressed in terms of fm(y) as
N xx = Dπ2 a 2  4 +2 2 2 + 4  (103)
m2 a a b b  follows:
a fm (0) = 0
Let =r (104) 2
b  mπ 
fm′′(0) − µ   fm (0) = 0
a = br (105)  a 
Then, 2
 mπ  (112)
1  m4 2m 2 n2 n 4  fm′′( b ) − µ   fm ( b ) = 0
N xx = Dπ2 b 2 r 2 2  4 4 + 2 2 2 + 4  (106)  a 
m b r b b r b  2
 mπ 
Thus, fm′′′( b ) − (2 − µ )   fm′ ( b ) = 0
1  m4 2m 2 n 2 n 4 r 2   a 
N xx = Dπ2 b 2 2  4 2 + + 4  (107) fm′′′( y ) = c1m λ13m sinh λ1m y + c2 m λ13m cosh λ1m y +
m b r b4 b 
Simplifying, c3m λ 32 m sin λ 2 m y − c4 m λ 32 m cos λ 2 m y (113)
Dπ 2  m 2 n4 r 2  Enforcement of the boundary conditions give:
N xx = 2  2 + 2n 2 + 2  (108)
b r m  fm ( y = 0) = c1m + c3m = 0 (114)
2
Hence,  mπ 
Dπ 2 c1m λ12m − c3 m λ 22 m − µ   (c1m + c3 m ) = 0 (115)
 a  Dπ
2
N xx = K (r ) 2 = K   2 (109)  a 
b b b c1m λ12m cosh λ1m b + c2 m λ12m sinh λ1m b − c3m λ 22 m cos λ 2 m b −
a m n4 r 2 
2
K (r ) = K   =  2 + 2 n 2 + 2 
2
(110)  mπ 
b  r m  c4 m λ 22 m sin λ 2 m b − µ   ( c1m cosh λ1m b +
 a 
For m = n = 1,
c2 m sinh λ1m b + c3m cos λ 2 m b + c4 m sin λ 2 m b ) = 0 (116)
a  1 
K   =  2 + 2 + r2  (111) c1m λ13m sinh λ1m b + c2 m λ13m cosh λ1m b + c3m λ 32 m sin λ 2 m b −
b r 
2
K(a/b) is tabulated for the case of simply supported edges  mπ 
c4 m λ 32 m cos λ 2 m b − (2 − µ )   (c1m λ1m sinh λ1m b +
for various values of a/b in Table 3  a 
c2 mλ1m cosh λ1m b − c3m λ 2 m sin λ 2 m b + c4 m λ 2 m cos λ 2 m b ) = 0
Table 3
Dimensionless critical elastic buckling load coefficients for the …(117)
simply supported Kirchhoff plate under uniaxial uniform Hence,
compression loading applied at x = 0, and x = a for various values c1m = −c3m (118)
of a/b Then,
a/b (Present K(a/b) K(a/b) K(a/b)
  mπ  
2
 2  mπ  
2
work) Iyengar Nwoji et Onyia et al c1m  λ12m − µ    + c1m  λ 2m + µ   =0 (119)
K(a/b) [25] al [23] [27, 28]   a     a  
0.1 102.01 102.01 102.01 102.01 c1m (λ12m + λ 22 m ) = 0 (120)
0.2 27.04 27.04 27.04 27.04
Solving, c1m = 0 (121)
0.3 13.201111 13.2011 13.2011 13.2011
0.4 8.41 8.41 8.41 8.41 c3 m = 0 (122)
0.5 6.25 6.25 6.25 6.25 Then the equations simplify to:
0.6 5.137778 5.1378 5.1378 5.1378   mπ 
2

0.7 4.530816 4.5308 4.5308 4.5308 c2 m  λ12m sinh λ1m b − µ   sinh λ1m b  −
  a  
0.8 4.2025 4.2025 4.2025 4.2025
0.9 4.044568 4.0446 4.0446 4.0446   mπ 
2

c4 m  λ 22 m sin λ 2 m b + µ   sin λ 2 m b  = 0 (123)
1.0 4 4 4 4   a  

ISSN (Print): 2456-6411 | ISSN (Online): 2456-6403 604 JREAS, Vol. 08, Issue 03, July 23
  2
 Elastic buckling load coefficients for SSSF Kirchhoff plate subject
 mπ 
c2 m  λ13m cosh λ1m b − (2 − µ )   λ1m cosh λ1m b  − to compressive load Nxx along the boundaries x = 0, and x = a, for
  a   varying aspect ratios (a/b) and for Poisson’s ratio µ = 0.25.
  mπ 
2
 a/b Present study Oguaghamba & Timoshenko % Difference
c4 m  λ 32 m cos λ 2 m b + (2 − µ )   λ 2 m cos λ 2 m b  = 0 Ike [29] [1] between present
  a   K(a/b) results for n = 1
…(124) K(a/b) K(a/b) K(a/b) K(a/b) for m = 1 and Timoshenko
for for for for [1]
Hence, m=1 m=2 m=1 m=2
  mπ  
2  2  mπ  
2
0.4 6.6367 25.2899 6.6367 25.2899
c2 m  λ12m − µ    sinh λ1m b − c4 m  λ 2 m + µ    sin λ 2 m b = 0
  a     a   0.6 3.1921 11.4675 3.1921 11.4675
0.8 1.9894 6.3667 1.9894 6.3667
…(125) 1.0 1.4342 4.4036 1.4342 4.4036 1.44 −0.403
  mπ  
2
1.5 0.8880 2.2022 0.8880 2.2022
c2 m λ1m  λ12m − (2 − µ )    cosh λ1m b − 2 0.6979 1.4342 0.6979 1.4342 0.698 −0.0205
  a  
2.5 0.6104 1.0798 0.6104 1.0798 0.610 0.0656
  mπ  
2
3 0.5630 0.8879 0.5630 0.8879 0.564 −0.1773
c4 m λ 2 m  λ 22 m + (2 − µ )    cos λ 2 m b = 0 (126) 3.5 0.5343 0.7726 0.5343 0.7726
  a  
4 0.5161 0.6979 0.5161 0.6979 0.517 −0.1741
It can be shown that 4.5 0.5034 0.6469 0.5034 0.6469
2 2
 mπ   mπ  5 0.4944 0.6104 0.4944 0.6104
λ12m − µ   = λ 2 m + (2 − µ ) 
2
 = β1m (127) 5.5 0.4877 0.5835 0.4877 0.5835
 a   a  6 0.4826 0.5630 0.4826 0.5630
2 2
 mπ   mπ 
λ 22 m + µ   = β2 m = λ1m − (2 − µ ) 
2
 (128)
 a   a  5. Discussion
Then,
c2 m β1 sinh λ1m b − c4 m β2 sin λ 2 m b = 0 (129) The Kantorovich variational method has been successfully
used in this work for the stability problems of rectangular
c2 m λ1m β2 m cosh λ1m b + c4 m λ 2 m ( −β1m )cos λ 2 m b = 0 (130)
Kirchhoff plates with two simply supported boundaries and
In matrix form, subjected to uniaxial uniform compressive load. Three
 β1m sinh λ1m b −β2 m sin λ 2 m b   c2 m   0  different cases of edge supports the unloaded sides
  = 
 λ1m β2 m cosh λ1m b −β1m λ 2 m cos λ 2 m b   c4 m   0  considered are:
…(131) (i) the unloaded sides are both clamped (SCSC or CSCS
 c2 m  plate),
For nontrivial solutions,   ≠ 0. (ii) the unloaded edges are both simply supported hence the
 c4 m  plate, is SSSS plate, and
The characteristic elastic buckling equation is obtained as: (iii) the side y = 0 is on simple support while the side y = b
β1m sinh λ1m b −β2 m sin λ 2 m b is free, and the plate is SSSF plate.
=0 (132)
λ1m β2 m cosh λ1m b −β1m λ 2 m cos λ 2 m b The general problem of elastic buckling of thin plates is
Expanding yields: presented in variational form as the problem of
minimization of the functional (Equation (6)) for the elastic
−β12m λ 2 m sinh λ1m b cos λ 2 m b + λ1mβ22 m cosh λ1m b sin λ 2 m b = 0
thin plate problem. For the specific considered problem, the
…(133) total potential energy to be minimized is presented from a
Dividing by cosh λ1m b cos λ 2 m b, gives: simplification of Equation (6) as Equation (7). By the
λ1m β22 m tan λ 2 m b − λ 2 m β12m tanh λ1m b = 0 (134) Kantorovich variational method adopted in this work, the
buckling deflection is assumed in variable – separable form
or, λ 2 m β12m tanh λ1m b = λ1m β22 m tan λ 2 m b (135) as an infinite series as given by Equation (11) where gm(x),
The solution to the transcendental equation are obtained the coordinate (basis) functions in the x coordinate direction
using computer software based iteration methods based on is chosen to apriori satisfy all the Dirichlet conditions at the
Newton’s methods or Newton – Raphson’s methods. The simply supported boundaries. Thus the buckled deflection
buckling loads corresponding to the first two buckling was considered as the infinite series in Equation (12) where
modes and for a Poisson’s ratio µ = 0.25 are found for fm(y) is an unknown function of the y coordinate variable.
various aspect ratios and presented in Table 4. Table 4 also The elastic thin plate buckling problem hence simplifies to
presents benchmark solutions found previously by finding the unknown function fm(y) that extremizes
Timoshenko [1]. (minimizes) the functional Π and simultaneously satisfies
all conditions along the boundaries y = 0, and y = b for each
Table 4 of the three cases considered in this study.

ISSN (Print): 2456-6411 | ISSN (Online): 2456-6403 605 JREAS, Vol. 08, Issue 03, July 23
 Since the buckled deflection function w(x, y) is using computer software and iteration methods for µ = 0.25
expressed in terms of fm(y) as given by Equation (12), the and µ = 0.30 and for plate aspect ratios between 0.4 and 2,
functional is expressed using fm(y) and its derivatives by and presented in Tables 1 and 2. Table 1 shows that the
substitution of w(x, y) and the corresponding partial presented results agree with previous results presented by
derivatives to obtain after simplifications, Equation (21). Timoshenko [1], Onah et al [24] and Onyia et al [26].
The modified total potential energy functional Π* is given as
Equation (24); where the integrand contains fm(y), fm′ ( y ) Table 1 shows that the dimensionless elastic buckling
coefficients for CSCS Kirchhoff plates with µ = 0.25 under
and fm′′( y ). It is easy to observe that the extremum or
in-plane uniform compressive loading obtained using the
minimum of Π* and Π would give identical results. Kantorovich variational method are identical with past
The Euler – Lagrange differential equation given by results obtained by Timoshenko [1], Onah et al [24], and
Equation (25) is used to obtain the conditions for functional Onyia et al [26].
to be minimized. The Euler – Lagrange differential equation
Table 2 reveals that K(a/b) for CSCS thin plates (with
for extremizing the total potential energy functional is found
µ = 0.30 ) under in-plane uniform compressive loading
as Equation (29). Further simplifications of Equation (29)
gives the system of ODEs – Equation (36) with fm(y) as the obtained using the present method are exactly the same as
unknown function desired to be found. the solution obtained by Onah et al [24].

A general solution to the Euler – Lagrange ODE is 5.2. Discussion on elastic buckling of SSSS Kirchhoff
obtained via trial functions by assuming a solution for fm(y) plates
of the exponential form as in Equation (37). This reduced
the ODE to the algebraic problem in Equation (38). The The boundary conditions for the case of simply
conditions for nontrivial solutions gave the auxiliary or supported edges y = 0 and y = b were expressed in terms of
homogeneous equation expressed by the polynomial – fm(y) as Equation (76). Enforcement of the boundary
Equation (39). The solution of the auxiliary polynomial conditions yielded a system of homogeneous Equations –
gave the four roots in Equations (46) and (47). The general Equations (78) – (81). The solution of Equations (82) and
solution for fm(y) was thus found as Equation (48). The (83) gave the two unknown constants of integration as
elastic buckling loads are determined for each of the Equations (84) and (85). The system of homogeneous
considered cases of edge supports by the enforcement of the equations simplified to Equations (86) and (87) presented in
appropriate boundary conditions. matrix form as Equation (88).

5.1. Discussion on elastic buckling of SCSC (or CSCS) The condition for nontrivial solutions gave the
characteristic elastic buckling equation for SSSS Kirchhoff
Kirchhoff plates
plates as Equation (89), which upon expansion and
For CSCS (or SCSC) plates where the edges y = 0 and y simplification gave Equation (92). The solution of the
= b are both clamped, Equation (8) is used to obtain the elastic buckling equation for nontrivial solutions was found
system of four equations presented as Equation (49). for SSSS thin plates as Equation (95) which yielded the
elastic buckling load expression given by Equation (108) for
Enforcement of boundary conditions along the unloaded any buckling modes m, n and any plate aspect ratio, r. The
edges gave the system of four homogeneous equations – elastic buckling load coefficient K(a/b) was found as
Equation (51), (52), (53) and (54), which were expressed as Equation (110) and for the buckling modes m = 1, n = 1,
Equation (55). K(a/b) was found as Equation (111). K(a/b) is tabulated for
The nontrivial solution rule of the system of equations is SSSS Kirchhoff plates for various aspect ratios (a/b) and for
used to express the elastic stability (buckling) problems as m = n = 1 and presented in Table 3 which also shows that
Equation (56). Expansion of Equation (56) would be the presented results agree with previously obtained
difficult, and hence a simplified procedure that reduces the solutions for K(a/b) by Iyengar [25] and Nwoji et al [23].
elastic stability problem to two unknowns is used. This is Table 3 demonstrates that the critical elastic buckling
accomplished by elimination of two constants of integration coefficients K(a/b) for SSSS Kirchhoff plate are exact and
for the system of equations using Equations (57) and (58); for a/b lying between 0.1 and 1.0, K(a/b) is identical with
yielding the system of homogeneous equations as Equations the previous results of Nwoji et al [23], Iyengar [25], and
(59) and (60) or Equation (61) in matrix form. The condition Onyia et al [27, 28].
for nontrivial solutions then gave the elastic stability
equation as Equation (62). 5.3. Discussion on elastic buckling of SSSF Kirchhoff
Upon expansion of the determinant in Equation (62) and plate
simplifications, the elastic buckling equation for CSCS (or The boundary conditions for the SSSF Kirchhoff plate
SCSC) Kirchhoff plates was obtained as Equation (75);
along the y = 0 and y = b edges are given in Equation (10);
which is a transcendental equation. Equation (75) is solved
and expressed in terms of fm(y) as Equation (112). Enforcing

ISSN (Print): 2456-6411 | ISSN (Online): 2456-6403 606 JREAS, Vol. 08, Issue 03, July 23
the boundary conditions yielded a set of four equations – (iii) The assumption of the deflection expression in variable
Equations (114) – (117). The system of four equations – separable form as the infinite series of products of
simplify because two constants of integration are found to fm(y) and gm(x) where gm(x) is known simplifies the total
vanish from Equations (121) and (122) obtained from potential energy functional to a functional dependent on
solving simultaneously Equations (114) and (115). The the y coordinate variable, the unknown function fm(y)
system of homogeneous equations then simplified to and derivatives of fm(y) ( fm′ ( y ) and fm′′( y )).
Equations (123) and (124) which was simplified to the
matrix form in Equation (131). The characteristic elastic (iv) Euler – Lagrange differential equations which are the
buckling equation obtained for nontrivial solutions is given conditions necessary for a minimization of the total
by Equation (132); which upon expansion and potential energy functional are applied to obtain the
simplification gave Equation (135). differential equation for fm(y) to yield a minimum value
of Π. The Euler – Lagrange differential equation
Equation (135) which is the characteristic elastic obtained for the presented elastic buckling problem is a
buckling equation for SSSF Kirchhoff plates is a system of ODEs of the fourth order.
transcendental equation which was solved by iteration
methods for µ = 0.25 and µ = 0.30 and for various aspect (v) Methods for solving ODEs and system of ODEs –
ratios, r. The elastic buckling load coefficients for SSSF differential (D) operator methods, trial function
Kirchhoff plate subject to uniform uniaxial compressive methods, etc – are used for the resulting Euler –
load Nxx along the simply supported edges for m = 1 and m = Lagrange equation obtained.
2 and for various aspect ratios and for Poisson’s ratio µ = (vi) The method of trial functions reduces the fourth order
0.25 are shown in Table 4 which also shows close ODE to a fourth degree polynomial as the auxiliary or
agreement with the previous results obtained by characteristic equation required to be satisfied for
Timoshenko [1]. Identical results were previously obtained nontrivial solutions for an assumption of fm(y) in
by Oguaghamba and Ike [29] using Galerkin-Vlasov exponential form.
variational method for the elastic stability of SSSF plates.
(vii) The four eigenvalues (roots) of the auxiliary
For the case of elastic buckling of SSSF Kirchhoff plate polynomial are used to obtain the general solution for
subjected to uniform compressive load, along the simply fm(y) in terms of hyperbolic functions and trigonometric
supported edges, the Kantorovich variational results functions with four unknown integration constants
presented in Table 4 (for µ = 0.25) show that for the first which agrees with the fourth order nature of the ODEs.
buckling mode, the present results agree remarkably well
(viii) Enforcement of boundary conditions are used to find
with previous results obtained by Timoshenko [1] with
the characteristic stability equations for each of the
relative difference of less than 0.41% in absolute terms.
considered three cases of edge support conditions along
Table 4 further shows that K(a/b) for m = 1 and m = 2 are
the unloaded boundaries.
identical with previous results obtained using the Galerkin-
Vlasov method by Oguaghamba and Ike [29]. (ix) The characteristic stability equations found for the
considered three cases of edge supports along y = 0 and
y = b were found to be exact solutions to the elastic
6. Conclusion
buckling problem, and yields exact solutions for the
In conclusion: corresponding boundary conditions considered.

(i) The Kantorovich variational method has been (x) The characteristic elastic buckling equations obtained
successfully used in this paper to solve the elastic for SCSC (or CSCS) and SSSF Kirchhoff plates are
buckling problem of Kirchhoff plate under uniform transcendental equations which are difficult to solve for
compressive loading applied at the two opposite simply in closed form, but are solved by iteration techniques or
supported edges x = 0 and x = a for the cases where the computer software tools.
edges y = 0 and y = a are clamped; the edges y = 0 and (xi) The characteristic elastic buckling equation obtained for
y = b are simply supported and the edge y = 0 is simply SSSS Kirchhoff plates is solved in closed form for any
supported and the edge y = b is free. given buckling mode m, n and the solutions are
(ii) The Kantorovich variational method adopted is based identical with solutions obtained previously.
on seeking to obtain a minimized energy functional Π (xii) The present solutions obtained for SCSC (or CSCS)
with respect to the unknown deflection function where Kirchhoff plates are identical with previously obtained
the deflection function is assumed as the infinite series solutions by Onah et al [24] who used the single Fourier
of a product of unknown function fm(y) and known sine integral transform method, Onyia et al [26] who
function gm(y) considered as the sinusoidal functions implemented the Galerkin – Kantorovich method for
that satisfy the Dirichlet boundary conditions along the the elastic stability analysis of thin rectangular SCSC
two opposite simply supported edges x = 0 and x = a. plates.

ISSN (Print): 2456-6411 | ISSN (Online): 2456-6403 607 JREAS, Vol. 08, Issue 03, July 23
Conflict of Interest [16] S. Abolghasemi, H. Eipakchi & M. Shariati, “An analytical solution
for buckling of plates with circular cutout subjected to non-uniform
The authors hereby declare that they have no conflict of in-plane loading”, Archive of Applied Mechanics, 89, 2019, 2519–
2543. https://doi.org/10.1007/S00419-019-01592-3.
interest in the publication of the paper.
[17] O.M. Ibearugbulem, “Application of a direct variational principle in
Acknowledgments elastic stability analysis of thin rectangular flat plates,” PhD Thesis
Structural Engineering Submitted to School of Post Graduate Studies
Federal University of Technology, Owerri, May 2012.
The authors sincerely acknowledge the dedicated work of
members of the Editorial Board and the team of reviewers [18] A.N. Nwadike, “Buckling analysis of isotropic rectangular plates
using Ritz method,” M.Eng Project report Dept of Civil Engineering,
for their contributions towards an expeditious and positive Federal University of Technology Owerri, July 2014.
review process that led to the successful publication of this
[19] O.A. Oguaghamba, “Analysis of buckling and post buckling loads of
paper. isotropic thin rectangular plates,” PhD Thesis submitted to Post
Graduate School, Federal University of Technology Owerri, Nigeria,
References 2015.

[1] S. Timoshenko, Theory of Elastic Stability (New York: McGraw [20] O.A. Oguaghamba, J.C. Ezeh, M.O. Ibearugbulem & L.O. Ettu,
Hill Publishing Co Ltd, 1936). DOI: “Buckling and post buckling loads characteristics of all edges
https://doi.org/10.1007/S0368393100104389. clamped thin rectangular plate”, The International Journal of
Science (IJES), 4(11), 2015, 55–61.
[2] S.P. Timoshenko & J.M. Gere, Theory of Elastic Stability (New
York: McGraw Hill, 1985). [21] O.M. Ibearugbulem, N.N. Osadebe, J.C. Ezeh & D.O. Onwuka,
“Buckling analysis of axially compressed SSSS thin rectangular
[3] P.S. Bulson, The Stability of Flat Plates (London, U.K: Chatto and plate using Taylor – Maclaurin shape function”, International
Windus, 1970). Journal of Civil and Structural Engineering, 2(2), 2011, 676–681.
[4] A. Chajes, Principles of Structural Stability Theory (Englewood [22] N.N. Osadebe, V.C. Nwokike & O.A. Oguaghamba, “Stability
Cliffs New Jersey: Prentice-Hall, 1993). analysis of SSSS thin rectangular plate using multi-degrees of
freedom Taylor – Maclaurins series in Galerkin variational method”
[5] M.L. Gambhir, “Elastic Buckling of Thin Plates” In: Stability
Nigerian Journal of Technology NIJOTECH, 35(3), 2016, 503–509.
Analysis and Design of Structures (Berlin Heidelberg: Springer,
http://dx.doi.org/10.4314/nijt/v35i3.5.
2004). https://doi.org/10.1007/978-3-662-09996-4_8.
[23] C.U. Nwoji, H.N. Onah, B.O. Mama, C.C. Ike & E.U. Ikwueze
[6] C.M. Wang, C.Y. Wang, J.N. Reddy, Exact solutions for buckling of
“Elastic buckling analysis of simply supported thin plates using the
structural members (Boca Raton Florida: CRC Press LLC, 2005).
double finite Fourier sine integral transform method” Explorematics
[7] G. Shi & G.P. Bézine, “Buckling analysis of orthotropic plates by Journal of Innovative Engineering and Technology (EJIET), 01(01),
boundary element method”, Mechanics Research Communications, September, 2017, 37–47.
17, 1990, 1–8.
[24] H.N. Onah, C.U. Nwoji, C.C. Ike & B.O. Mama, “Elastic buckling
[8] G. Shi, “Flexural vibration and buckling analysis of orthotropic analysis of uniaxially compressed CCSS Kirchhoff plate using single
plates by the boundary element method”, International Journal of finite Fourier sine integral transform method”, Modelling,
Solids and Structures, 26, 1990, 1351–1370. Measurement and Control B, 87(2), June 2018, 107–111.
DoI:10.18280/mmc_b.870208.
[9] J.T. Abodi, “Elastic buckling analysis of plates under in-plane patch
loading using finite difference method”, Journal of Kerbala [25] N.G.R. Iyengar, Structural Stability of Columns and Plates
University, 10(2), 2012, 142–150. (Chichester: Ellis Horwood Limited 316pp, 1988). ISBN:
0745805396.
[10] S.B. Batford & J.C. Houbolt, “Critical combination of shear and
transverse direct stress for an infinitely long flat plate with edges [26] M.E. Onyia, E.O. Rowland-Lato & C.C. Ike, “Galerkin –
elastically restrained against rotation”, NACA Report, 847, 1946, 8– Kantorovich method for the elastic buckling analysis of thin
12. rectangular SCSC plates”, International Journal of Engineering
Research and Technology, 13(4), 2020, 613–619.
[11] C.M. Wang, Y Xiang & J. Chakrabarty, “Elastic/plastic buckling of
thick plates”, International Journal of Solids and Structures, 38, [27] M.E. Onyia, E.O. Rowland-Lato & C.C. Ike, “Elastic buckling
2001, 8617–8640. analysis of SSCF and SSSS rectangular thin plates using the single
finite Fourier sine integral transform method” International Journal
[12] S. Ullah, J. Zhang & U. Zhong, “New analytical solutions of of Engineering Research and Technology, 13(6), 2020, 1147–1158.
buckling problem of rotationally restrained rectangular thin plates”,
International Journal of Applied Mechanics, 11(10), 2019, 195101. [28] M.E. Onyia, E.O. Rowland-Lato & C.C. Ike, “Galerkin-Vlasov
DOI:10.1142/S1758825119501011. variational method for the elastic buckling analysis of SSCF and
SSSS rectangular plates” International Journal of Engineering
[13] S. Ullah, H. Wang, X Zheng, J. Zhang, Y. Zhong & R. Li, “New Research and Technology, 13(6), 2020, 1137–1146.
analytic buckling solutions of moderately thick clamped rectangular
plates by a straight forward finite integral transform method”, [29] O.A. Oguaghamba & C.C. Ike, “Galerkin-Vlasov method for the
Archive of Applied Mechanics, 89(9), September 2019, 1885–1897. elastic buckling analysis of Kirchhoff plate with one free edge and
three simply supported edges under uniform uniaxial compression”
[14] Y. Xiang, C.M. Wang & C.Y. Wang, “Buckling of rectangular ARPN Journal of Engineering and Applied Sciences, 15(14), 2020,
plates with an internal hinge”, International Journal of Structural 1574 –1581.
Stability and Dynamics, 1(2), 2001, 169–179.
[15] Yu Chen “Buckling of rectangular plates under intermediate and end
loads,” MSc Thesis submitted to Department of Civil Engineering,
National University of Singapore, 2003.

ISSN (Print): 2456-6411 | ISSN (Online): 2456-6403 608 JREAS, Vol. 08, Issue 03, July 23

View publication stats