Uni-axial compression; Other BC’s
• Of particular interest (for use with stiffeners):
y
free
ss ss b
ss x
a
choice of λ is important!
• Use:
mx ny
w Amn sin sin w
a b
• to substitute in governing equation:
4w 4w 4w 2w 2w
D11 4 2( D12 2 D66 ) 2 2 D22 4 Nx 2 Ny 2
x x y y x y
y
Buckling shape versus y
Uni-axial compression; 3 sides ss,
one unloaded side free
• λ is chosen so that the “free” condition is
represented (e.g. if λ=1=> simply-supported)
ny λ=1 λ=1/2
sin
b
1
λ=5/12
0.5 1 y/b
Uni-axial compression; 3 sides ss,
one unloaded side free
• Current choice of w does not satisfy all BC’s:
w( x 0) w( x a) 0 OK
w( y 0) 0
OK
w
2
w 2
M x D11 D 0 at x 0, a
x 2
12
y 2 OK
2w 2w
M y D12 2 D22 2 0 at y 0, b ???!!
x y
• Expression for the buckling load
2 D11 m 4 2( D12 2D66 )m 2 2 ( AR) 2 D22 ( AR) 4 4
No
a2m2
Uni-axial compression; 3 sides ss, one
unloaded side free
• Comparison with exact solution
1
Nxcrit (N/mm)
0.9 λ=1/2
0.8 D11 1.13
0.7 D12 0.56 λ=5/12
D22 1.13 Nm
0.6
0.5
D66 0.68 exact
0.4
0.3 infinitely long
0.2 plate
0.1
0
0 1 2 3 4 5 6 7 8 9 10
Length/Width
Infinitely long plate, 3 sides ss, one
unloaded side free
• exact solution:
free
12 D66
N xcrit ss ss b
b2 ss
a
• approximate solution:
4 2 2 2 2 exact: Nxcrit=0.032 N/mm
N xcrit 2 D66 2 D12
b b λ Nxcrit Δ(%)
≈ 6.9-9.9 1/2 0.047 46.9
D11 1.13
D12
D22
0.56
1.13 Nm 5/12 0.036 12.5
D66 0.68
b=50.8 cm
Buckling of rectangular plate under
shear
• simply supported plate*
Nxy x
Nxy
a D16=D26=Bij=0
y
*Whitney, J.M., Structural Analysis of Laminated Anisotropic Plates,
Technomic Publishing, 1987, section 5.7
Buckling under shear (cont’d)
• Galerkin solution
• Governing equation:
4w 4w 4w 2w
D11 4 2( D12 2 D66 ) 2 2 D22 4 2 N xy
x x y y xy
• assume solution w in the form:
mx ny
w Amn sin sin characteristic
a b functions
• multiply governing equation by the characteristic functions,
integrate over the domain of the plate and set the result
equal to zero (satisfied for m,n=1,2,…)
4w 4w 4w 2w mx ny
11 x 4
D 2( D12 2 D66 )
x 2 y 2
D 22
y 4
2 N xy
xy
sin
a
sin
b
dxdy 0
Buckling under shear (cont’d)
• Substituting for w and carrying out the integrations:
D m 2( D 2 D )m n ( AR) D n ( AR) A 32mn( AR) b N T A 0
4
11
4
12 66
2 2 2
22
4 4
mn
3 2
xy ij ij
ij
Tij for m i odd and n j odd
(m i )(n 2 j 2 )
2 2
(AR)=a/b, aspect ratio
Tij 0 otherwise
• Equation breaks down to two sets of homogeneous
equations, one for m+n odd and one for m+n even; each
is of the form: [E]{Amn}= Nxy [H] {Amn}
• => two generalized eigenvalue problems
• Lowest eigenvalue corresponds to buckling load; note
that for specially orthotropic plates the eigenvalues are in
pairs of positive and negative numbers of same
magnitude
Buckling under combined loads
• Combined compression and shear, simply-
supported all around
Nxy
x
ss
b ss ss
ss
Nx Nxy Nx
a
y
Buckling under compression and
shear
• approximate solution (2 terms):
x y 2x 2y Single term
w w1 sin sin w2 sin sin won’t work for
a b a b shear!
• satisfies displacement (and force) BC’s:
w(x 0) w(x a) 0
w(y 0) w(y b) 0
2w 2w
M x D11 2 D12 2 0 at x=0,a no need to satisfy force BC’s for
x y
energy minimization approach
2w 2w
M y D12 2 D22 2 0 at y=0,b (but need more terms)
x y
• w1, w2 are unknown
Energy minimization
• Minimize:
1 2 w 2 w 2 w
2 2 2
2w 2w 2w 2w 2w 2w
c D11 2 2 D12 2 D22 2 4 D66 4 D16 2 4 D26 2 dxdy
2 x x y 2 y x y x x y y x y
w w w
2
1
2 Nx dxdy N xy
x x y
dxdy
• integrations carried over the entire plate
• Nx, Nxy constant
N xy
• k
Nx
• D16=D26=0
Some intermediate results
2
2w 2
4
2x 2y 2 16
4
4x 4y
2 w1 4
1 cos 1 cos w2 4
1 cos 1 cos
x 4b a b 4b a b
4 4 1 x 3x y 3y
2w1 w2 4 cos cos cos cos
b 4 a a b b
… similarly for other integrands
• carrying out the integrations:
2
2w 2 2 4
a b 4 4 2
2w 2 2 4
0 0 x 2 dxdy w1 4a 4 ab w2 a 4 ab
a b 4 4
0 0 y 2 dxdy w1 4b 4 ab w2 b 4 ab
a b
2w 2w 4 2 4
4
0 0 x 2 y 2 dxdy w1 4a 2b 2 ab w2 a 2 b 2 ab
2
2 2
2
w
a b 2 2
2
0 0 x dxdy w1 4a 2 ab w2 a 2 ab
a b
2w 4 2 4
4
0 0 xy dxdy w1 4a 2b 2 ab w2 a 2b 2 ab
2
a b
w w w1 w2 2 2a 2a 2b 2b w1 w2 2 2a 2a 2b 2b
0 0 x y dxdy 2ab 3 3 2ab 3 3
Final energy expression
2 4 2 4
4
2 4 2 4
4
11 1
D w 3
b w2 3
b 2( D12 2 D66 1
) w w2
1 4 a a 4 ab ab
c
2 2 4
2 4
4
22 1 4b3
D w a w 2 3
a
b
No 2 2 2
2
32
w1 4a b w2 a b kN o w1w2 9
2
No=Nxcrit=-Nx
• to determine w1, w2,
c
0
w1
c
0
w2
System of equations
• which lead to
1 w1 4 b 4 w1 w1 4 a w1 2 b 32
D11 3
2( D12 2D66 ) D22 3
No kN o w2 0
2 2a 2ab 2b 4 a 9
1 8w2 4 b 8 4 w2 8w2 4 a w2 2 b 32
D11 2( D12 2 D66 ) D22 No kN o w1 0
2 a3 ab b3 a 9
• homogeneous system of two eqns in the
two unknowns w1, w2
• trivial solution w1=w2=0 corresponds to in-
plane deformations of the plate
System of equations
• Setting
1 4b 4 4a
K1 D11 3 2( D12 2 D66 ) D22 3
4 a ab b
• and using matrix notation eigen value
2b 32
K1 0 w1 k w
4a 9 1
0 No
16 K1 w2 32 k 2 b w2
9 a
A x α B x
generalized eigen-value problem of the form:
Ax B x
~ ~ ~ ~
Solution to the eigenvalue problem
• pre-multiply both sides of the equation by B-1
1 1
B Ax B B x
~ ~ ~ ~ ~ ~
C I
~ ~
• obtain standard eigen value problem:
Cx x
~ ~ ~
• the eigenvalues are obtained as solutions to:
det C I 0
~ ~
For our specific example
2b 32
1 1 k
B a 9
~
4 b 2 32 32 k
2
2b
k
4a 2 9 9 4a
and after some rearranging, the standardized eigenvalue problem has
the form,
2b 32
16 k w 4 2
b 32
2
1 w1
a 9
N o
1
k
2 2
16 2 1 w2
32
k b w 4 a 9 K
9 4a
eigenvalue α
Determination of buckling load(s)
• the eigenvalue is the solution to:
2b 4 2 b 512(32) 2
k 0
a a 81
• solving for α and substituting in terms of No:
a2 a4
D 2( D12 2 D66 ) 2 D22 4
2 11
b b 65536 a 2 2
No 2 5 9 k
a 8192 a 2
81 b
4 2
2 k 2
81 b 2 4
two solutions: use the lowest
Special cases: Pure compression
• Nxy=0 => k=0
2 a2 a4
N o 2 D11 2( D12 2 D66 ) 2 D22 4
a b b
comparing with exact solution found earlier, this
expression is identical to the exact solution for m=1
(but not so accurate for m>1)
Special case: Comparison to exact
solution
3.5
Buckling load
3
(N/mm) width
2.5
2 length
1.5 approx.
1
0.5 exact
0
0 0.5 1 1.5 2 2.5 3
Aspect ratio (length/width)
D11 0.66
D12 0.47
D22 0.66 Nm
D66 0.49
Special cases: Pure Shear
• set k very large
a2 a4
D 2( D12 2 D66 ) 2 D22 4
2 11
b b 65536 a 2 2
No 2 5 9 k
a 8192 a 2 2 81 4 2
b
2 k
81 b 2 4
Nxy ± a2 a4
D 2( D12 2 D66 ) 2 D22 4
2 11
Nok 2
b b
a 32 a
9 b 2
9 4 b a2 a4
N xycrit D11 2( D12 2 D66 ) 2 D22 4
32a 3 b b
Typically, 27% higher than exact solution!
Shear buckling: Comparison of various
methods(1)
81047
including transverse
shear effects (see
sandwich buckling)
(1) S. Simonian & C. de Winter
K relation: equation just derived
Interpolation: see sandwich buckling under shear
Shear buckling: Comparison of various
methods
81047
including transverse
shear effects (see
sandwich buckling)
(1) S. Simonian & C. de Winter
K relation: equation just derived
Interpolation: see sandwich buckling under shear
Shear buckling: Comparison of various
methods
81047
including transverse
shear effects (see
sandwich buckling)
(1) S. Simonian & C. de Winter
K relation: equation just derived
Interpolation: see sandwich buckling under shear
Shear buckling: Comparison of various
methods
81047
including transverse
shear effects (see
sandwich buckling)
(1) S. Simonian
K relation: equation just derived
Interpolation: see sandwich buckling under shear
Interaction curve: Buckling under
combined compression and shear
• Even though the present solution is
approximate, it is expected to be quite
accurate in providing the interaction curve
when both compression and shear are
applied
Buckling interaction curve:
Combined compression and shear
Nxy/Nxycrit
1
0.9
0.8
0.7 exact
0.6
0.5
For a given applied Nx, can
0.4 determine the Nxy that will
approx.
0.3 cause the plate to buckle when
0.2 both Nx and Nxy act
0.1
0
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
Nx/Nxcrit
Here, Nxcrit and Nxycrit refer N xy
2 “exact”
Nx
to the buckling when 1 interaction
N xcrit N
each load is applied xycrit curve
individually
Buckling under various loads and
Boundary Conditions(1)
ss 2 D11 m 4 2( D12 2D66 )m 2 ( AR) 2 D22 ( AR) 4
ss b ss No
a 2m2
ss
2
ss No D11 D22 K
aD
1/ 4
b2
c c 22 4 2( D12 2 D66 ) 3
b ss b D11 K 2 0<λ<1.662
2
D11 D22 4
m 4 8m 2 1 2( D12 2 D66 ) 2 λ>1.662
K
2 (m 2 1) D11 D22 m2 1
c aD
1/ 4
2 m2 2( D12 2D66 ) 16 2
ss 22 No D11 D22 K K
b ss b D11 b2 2 D11 D22 3 m2
c
2 4 8( D12 2 D66 )
1/ 4 No D11 D22 K K 42 0<λ<1.094
c aD b 2
3 D11 D22
2
c 22
b c b D11 m 4 8m 2 1 2( D12 2 D66 ) 2
c K 2 2 2 λ>1.094
(m 1) D11 D22 m 1
free 1/ 4
2
ss b ss aD No D11 D22 K K
12 D66
1
22 2
2
2
b D11 b D11 D22
ss
(1) NASA/DoD Adv Composites Design Guide, vol II, 1983
Buckling under various loads and
boundary conditions(1)
2
1/ 4 Mo D11 D22 K
a D22 b 2
M M
ss b D11
ss b ss m 2 2( D12 2 D66 ) 2 m 2 8( D12 2 D66 ) 2
K 0.047 2 b 2 2 2 2 16 2
ss D11 D22 m D11 D22 m
ss 2 D11 m 4 2( D12 2D66 )m 2 n 2 ( AR) 2 D22 n 4 ( AR) 4
ss ss No
b ss a 2 (m 2 kn 2 ( AR) 2 )
1/ 4
4
N xycrit D D
3 1/ 4
(K )
aD
11 22
ss b2
22
1/ 4
ss ( D 2 D66 ) 1 D
b ss b D11 K 8.2 5 12 A 11
ss D11 D22 ( B )
D22
10
( D 2 D66 )
A 0.27 0.185 12
D11 D22
a=b; 2
( D12 2 D66 ) ( D 2 D66 )
0<β≤1 B 0.82 0.46 0.2 12
D11 D22 D11 D22
(1) NASA/DoD Adv Composites Design Guide, vol II, 1983
Effect of BC’s on buckling load of a
square plate under compression
2.5 No No
Buckling load b=a
(normalized
2
to SS all
around) a
1.5
0.5
0
SS all around CCL,SSU SSL,CCU CC all around
Biggest difference is less than 2.5 to 1 (compare to beams with 4 to 1 ratio)