FINITE ELEMENT ANALYSIS OF BEAMS

1
3.1 ELEMENTARY BEAM
THEORY

2
 INTRODUCTION
• We learned Direct Stiffness Method in Chapter 1
– Limited to simple elements such as 1D bars
• In Chapter 3, Galerkin Method and Principle of Minimum
Potential Energy can be applied to more complex elements
• we will learn Energy Method to build beam finite element
– Structure is in equilibrium when the potential energy is minimum
• Potential energy: Sum of strain energy and potential of
applied loads
Potential of
• Interpolation scheme:  UV applied loads
Strain energy
v(x)  N(x)  {q}

Beam Interpolation Nodal

deflection function DOF
3
 Review of Beam Theory
• Euler-Bernoulli Beam Theory
– can carry the transverse load
– slope can change along the span (x-axis)
– Cross-section is symmetric w.r.t. xy-plane
– The y-axis passes through the centroid
– Loads are applied in xy-plane (plane of loading)

y y
Neutral axis

x z
A
L
F F

4
 BEAM THEORY cont.
• Euler-Bernoulli Beam Theory cont.
– Plane sections normal to the beam axis remain plane and normal to
the axis after deformation (no shear stress)
– Transverse deflection (deflection curve) is function of x only: v(x)
– Displacement in x-dir is function of x and y: u(x, y)

dv ¶ u d u0 d2v dv
u( x, y ) = u0 ( x ) - y exx = = -y 2 q=
dx ¶x dx dx dx
y
y(dv/dx)
Neutral axis

x y = dv/dx
L
F v(x)

5
 BEAM THEORY cont.
• Euler-Bernoulli Beam Theory cont. exx =
¶ u d u0
=
d2v
-y 2
– Strain along the beam axis: e0 = du0 / dx ¶x dx dx
– Strain xx varies linearly w.r.t. y; Strain yy = 0
– Curvature: -d2v / dx 2
– Can assume plane stress in z-dir basically uniaxial status
d2v
s xx = E exx = E e0 - Ey 2
dx
• Axial force resultant and bending moment
d2v
P= ò s xx dA = E e0 ò dA - E 2
dx ò yd A P = EAe0
A A A
d2v
d2v M = EI 2
M = -ò y s xx dA = -E e0 ò yd A + E 2 ò y 2 dA dx
A A
dx A

EA: axial rigidity

Moment of inertia I(x)
EI: flexural rigidity
6
 BEAM THEORY cont.
• Beam constitutive relation
– We assume P = 0 (We will consider non-zero P in the frame element)
– Moment-curvature relation:

d2v Moment and curvature is linearly dependent

M = EI 2
dx
• Sign convention +Vy
+M y +M
+P x +P

+Vy

– Positive directions for applied loads

y
p(x)

x C1 C2 C3

F1 F2 F3 7
 GOVERNING EQUATIONS
• Beam equilibrium equations
æ dVy ö÷ dVy
å fy = 0  p( x )dx + ççèVy + dx dx ÷÷ø - Vy = 0
ç
dx
= - p( x )

æ dM ö÷ dx dM
-M + çç M + dx ÷÷ - ( pdx ) + Vy dx = 0 Vy = -
è dx ø 2 dx

d4v
– Combining three equations together: EI = p( x )
4
dx
– Fourth-order differential equation
p
d Vy
Vy  dx
dx

dM
M Vy M dx
dx

dx 8
 STRESS AND STRAIN
• Bending stress
d2v d2v
s xx = -Ey 2 M = EI 2
dx dx

M ( x )y
s xx ( x, y ) = -
I Bending stress

– This is only non-zero stress component for Euler-Bernoulli beam

• Transverse shear strain
¶u ¶v ¶v ¶v dv
g xy = + =- + =0 u( x, y ) = u0 ( x ) - y
¶y ¶x ¶x ¶x dx
– Euler beam predicts zero shear strain (approximation)
VQ
t
– Traditional beam theory says the transverse shear stress is xy =
Ib
– However, this shear stress is in general small compared to
the bending stress

9
 POTENTIAL ENERGY
• Potential energy P = U +V
• Strain energy
– Strain energy density
æ 2 ö2 æ 2 ö2
1 1 2 1 ç d v ÷ 1 2ç d v÷
U 0 = s xx exx = E ( exx ) = E ç -y 2 ÷÷ = Ey ç 2 ÷÷
2 2 2 çè dx ø 2 èç dx ø
– Strain energy per unit length
æ 2 ö2 æ 2 ö2
1 2ç d v ÷ 1 dv
UL ( x ) = ò U0 ( x, y , z )dA = ò Ey ç 2 ÷÷ dA = E çç 2 ÷÷÷ ò y 2dA
2 çè dx ø 2 çè dx ø
A A A

æ 2 ö2 Moment of
1 çd v ÷
UL ( x ) = EI ç 2 ÷÷ inertia
2 çè dx ø
– Strain energy
L 1 L æ d2v ö÷2
ò0 EI çç 2 ÷÷ dx
2 ò0
U= U L ( x )dx =
çè dx ø

10
 POTENTIAL ENERGY cont.
• Potential energy of applied loads
NF NC
L dv ( xi )
V = -ò p( x )v ( x )dx - å Fi v ( xi ) - å Ci
0
i =1 i =1 dx
• Potential energy

1 L æ d2v ö÷2 L
NF NC
dv ( xi )
EI çç 2 ÷÷ dx - ò p( x )v ( x )dx - å Fi v ( xi ) - å Ci
2 ò0
P = U +V =
çè dx ø 0
i =1 i =1 dx

– Potential energy is a function of v(x) and slope

– The beam is in equilibrium when  has its minimum value


¶P
=0
¶v

v* v 11
3.2 FINITE ELEMENT INTERPOLATION

21
 FINITE ELEMENT INTERPOLATION
• Rayleigh-Ritz method approximate solution in the entire beam
– Difficult to find approx solution that satisfies displacement BC
• Finite element approximates solution in an element
– Make it easy to satisfy displacement BC using interpolation technique
• Beam element
– Divide the beam using a set of elements
– Elements are connected to other elements at nodes
– Concentrated forces and couples can only be applied at nodes
– Consider two-node beam element
– Positive directions for forces and couples
– Constant or linearly F1 F2
distributed load C1 C2
x

p(x)
22
 FINITE ELEMENT INTERPOLATION cont.
• Nodal DOF of beam element
– Each node has deflection v and slope 
– Positive directions of DOFs
– Vector of nodal DOFs {q} = {v1 q1 v 2 q2 }T
• Scaling parameter s
– Length L of the beam is scaled to 1 using scaling parameter s
v1 v2
x - x1 1 1 2
s= , ds = dx, x
L L
ds 1 L
dx = Lds, =
dx L x1 x2
s=0 s=1

• Will write deflection curve v(s) in terms of s

23
 FINITE ELEMENT INTERPOLATION cont.
• Deflection interpolation
– Interpolate the deflection v(s) in terms of four nodal DOFs
– Use cubic function: v (s ) = a0 + a1s + a2s 2 + a3s 3
– Relation to the slope: dv dv ds 1
q= = = (a1 + 2a2s + 3a3s 2 )
dx ds dx L
– Apply four conditions:
dv (0) dv (1)
v (0) = v1 = q1 v (1) = v 2 = q2
dx dx
– Express four coefficients in terms of nodal DOFs

v1 = v (0) = a0
a0 = v1
dv 1
q1 = (0) = a1 a1 = Lq1
dx L
v 2 = v (1) = a0 + a1 + a2 + a3 a2 = -3v1 - 2Lq1 + 3v 2 - Lq2
dv 1 a3 = 2v1 + Lq1 - 2v 2 + Lq2
q2 = (1) = (a1 + 2a2 + 3a3 )
dx L
24
 FINITE ELEMENT INTERPOLATION cont.
• Deflection interpolation cont.
v (s ) = (1 - 3s 2 + 2s 3 )v1 + L(s - 2s 2 + s 3 )q1 + (3s 2 - 2s 3 )v 2 + L(-s 2 + s 3 )q2
ìï v1 üï
ïï ïï
ïï q1 ïï
v (s ) = [N1(s ) N2 (s ) N3 (s ) N 4 (s )] í ý v (s ) = êë N úû {q}
ïï v 2 ïï
ïï ïï
îï q2 þï
• Shape functions 1.0

2 3 N1 N3
N1(s ) = 1 - 3s + 2s 0.8

N2 (s ) = L(s - 2s 2 + s 3 ) 0.6

N3 (s ) = 3s 2 - 2s 3 0.4

N4 (s ) = L(-s 2 + s 3 ) 0.2 N2/L

– Hermite polynomials 0.0

0.0 0.2 0.4 0.6 N4/L 0.8 1.0

– Interpolation property -0.2

25
 FINITE ELEMENT INTERPOLATION cont.
• Properties of interpolation
– Deflection is a cubic polynomial (discuss accuracy and limitation)
– Interpolation is valid within an element, not outside of the element
– Adjacent elements have continuous deflection and slope
• Approximation of curvature
– Curvature is second derivative and related to strain and stress ìï v1 üï
ïï ïï
dv2
1 dv2
1 ïï q1 ïï
= 2 2 = 2 [-6 + 12s, L(-4 + 6s ), 6 - 12s, L(-2 + 6s )] í ý
dx 2
L ds L ïï v 2 ïï
ïï ïï
2 ï q2 þï
î
dv 1 T
2
= 2
{B} {q} B: strain-displacement vector
dx L 1´4 4´1
– B is linear function of s and, thus, the strain and stress
– Alternative expression: d2v 1 T
2
= 2
{q} {B}
dx L 1´4 4´1
– If the given problem is linearly varying curvature, the approximation is
accurate; if higher-order variation of curvature, then it is approximate 26
 FINITE ELEMENT INTERPOLATION cont.
• Approximation of bending moment and shear force
d2v EI
M (s ) = EI 2 = 2 {B} T {q} Linear
dx L
dM d3v EI
Vy = - = -EI 3 = 3 [-12 -6L 12 -6L]{q} Constant
dx dx L
– Stress is proportional to M(s); M(s) is linear; stress is linear, too
– Maximum stress always occurs at the node
– Bending moment and shear force are not continuous between adjacent
elements

27
 EXAMPLE – INTERPOLATION
v2
v1
• Cantilevered beam
• Given nodal DOFs
T 2
{q) = {0, 0, - 0.1, - 0.2} 1
L
• Deflection and slope at x = 0.5L
• Parameter s = 0.5 at x = 0.5L
1 L 1 L
• Shape functions: N1( 2 ) = , N2 ( 2 ) = , N3 ( 2 ) = , N4 ( 2 ) = -
1 1 1 1
2 8 2 8
• Deflection at s = 0.5:
v ( 21 ) = N1( 21 )v1 + N2 ( 21 )q1 + N3 ( 21 )v 2 + N4 ( 21 )q2
1 L 1 L v Lq
= ´ 0 + ´ 0 + ´ v 2 - ´ q2 = 2 - 2 = -0.025
2 8 2 8 2 8
• Slope at s = 0.5:
dv 1 dv 1 æ dN1 dN2 dN3 dN4 ö÷
= = vç + q1 + v2 + q2
dx L ds L çè 1 ds ds ds ds ÷÷ø
1 1
= v1 (-6s + 6s 2 ) + q1 ( 1 - 4s + 3s 2 ) + v 2 (6s - 6s 2 ) + q2 ( -2s + 3s 2 ) = -0.1
L L
28
 EXAMPLE
• A beam finite element with length L
L3 L2
v1 = 0, q1 = 0, v 2 = , q2 =
3EI 2EI L
F
• Calculate v(s)
v (s ) = N1(s )v1 + N2 (s )q1 + N3 (s )v 2 + N4 (s )q2

v (s ) = (3s 2 - 2s 3 )v 2 + L(-s 2 + s 3 )q2

• Bending moment
d2v EI d2v EI
M (s ) = EI 2 = 2 2 = 2 [ (6 - 12s )v 2 + L(-2 + 6s )q2 ]
dx L ds L
EI é L3 L2 ù
= 2 ê (6 - 12s ) + L(-2 + 6s ) ú
L êë 3EI 2EI úû
= L(1 - s ) = (L - x ) Bending moment cause by unit force at the tip
29
 Exercise
• Calculate the beam shape functions when the natural
coordinate is given as s = [-1, +1]
– Hint: Assume v (s ) = a0 + a1s + a2s 2 + a3s 3
determine 4 coefficients using the following conditions:
dv (-1) dv (1)
v (-1) = v1 = q1 v (1) = v 2 = q2
dx dx

30
3.3 FE EQUATION FOR BEAM
ELEMENT

31
 FINITE ELEMENT EQUATION FOR BEAM
• Finite element equation using PMPE
– A beam is divided by NEL elements with constant sections
• Strain energy
– Sum of each element’s strain energy
NEL x2  NEL
e
LT
U   UL (x)dx    e UL (x)dx   U
e
0 x1
e1 e1

– Strain energy of element (e)

2 2
x 2 
e
1  d2 v  EI 1 1  d2 v 
U   EI  e   2  dx  3 0  2  ds
e
x1 2  dx  L 2  ds 

y p(x)

C1 x C2 C3 C4 C5
2 3 5
1 4

F1 F2 F3 F4 F5
x1(1) x 2(1) = x1( 2 ) x 2( 2 ) = x1( 3 ) x 2( 3 ) = x1( 4 ) x 2( 4 ) 32
 FE EQUATION FOR BEAM cont.
• Strain energy cont.
– Approximate curvature in terms of nodal DOFs
2
 d2 v   d2 v   d2 v  e T e
 2   2  2   { q } {B } {B } T
{ q }
 ds   ds   ds  1 4 41 1 4 41

– Approximate element strain energy in terms of nodal DOFs

 e
1  e T  EI 1  1  e T  e  e
{q }  3  {B}{B}T ds {q  } 
e
U(e)  {q } [k ]{q }
2 L 0  2
• Stiffness matrix of a beam element
 6  12s 
 
EI 1 L( 4  6s)
e
[k ]  3 
L  6  12s 
0
 6  12s L( 4  6s) 6  12s L( 2  6s) ds
 
L(  2  6s) 

33
 FE EQUATION FOR BEAM cont.
• Stiffness matrix of a beam element

 12 6L 12 6L  Symmetric, positive semi-definite

 2 
EI  6L 4L 6L 2L 
2
e Proportional to EI
[k ]  3
L  12 6L 12 6L 
 2 
Inversely proportional to L
 6L 2L2
6L 4L 

• Strain energy cont.

NEL
1 NEL  e T  e  e
U  U (e)
  {q } [k ]{q }
e1 2 e1
– Assembly
1
U {Q s } T [K s ]{Q s }
2

34
y EXAMPLE – ASSEMBLY
• Two elements
x 2EI EI
3 • Global DOFs
1 2
2L L {Qs }T  {v1 1 v 2 2 v3 3 }
F2 F3
v1 1 v2 2 v2 2 v3 3
 3 3L 3 3L  v1  12 6L 12 6L  v 2
 3L 4L2 3L 2L2    2 
EI  6L 4L 6L 2L  2
2
EI  2
[k   ]  3   1
1
[k ]  3
L  3 3L 3 3L  v 2 L  12 6L 12 6L  v 3
 2   2 
 3L 2L2
 3L 4L  2  6L 2L2
6L 4L  3

 3 3L 3 3L 0 0 
 3L 4L2 3L 2L2 0 0 
 
EI  3 3L 15 3L 12 6L 
[K s ]  3  
L  3L 2L2 3L 8L2 6L 2L2 
0 0 12 6L 12 6L 
 2 
 0 0 6L 2L2
6L 4L  35
 FE EQUATION FOR BEAM cont.
• Potential energy of applied loads
 F1 
– Concentrated forces and couples C 
ND  1 
V   Fv i i  Ci i 
      T
V  v v
 1 1 2 ......  F
ND   2  {Q s } {Fs }
i1   
 
CND 
– Distributed load (Work-equivalent nodal forces)
1
NEL x2  NEL  e
e

V     e p(x)v(x)dx   V (e)

x2

x1
V(e)   e p(x)v(x)dx  L(e)  p(s)v(s)ds
x1
e1 e1 0

1
V (e)  L(e)  p(s)  v1N1  1N2  v 2N3  2N4  ds
0

 (e) 1   (e) 1   (e) 1   (e) 1 

 v1  L  p(s)N1ds   1  L  p(s)N2ds   v 2  L  p(s)N3 ds   2  L  p(s)N4 ds 
 0   0   0   0 
 v1F1(e)  1C1(e)  v 2F2(e)  2C(e)
2

36
EXAMPLE – WORK-EQUIVALENT NODAL FORCES
• Uniformly distributed load
1 1 pL
F1  pL  N1(s)ds  pL  (1  3s2  2s3 )ds 
0 0 2
1 1 pL2
C1  pL  N2 (s)ds  pL  (s  2s  s )ds 
2 2 3
0 0 12
1 1 pL
F2  pL  N3 (s)ds  pL  (3s  2s )ds 
2 3
0 0 2
1 1 pL2
C2  pL  N4 (s)ds  pL  ( s  s )ds  
2 2 3
0 0 12

 pL pL2 pL pL2 
{F} T    
2 12 2 12  p
Equivalent
pL/2 pL/2

pL2/12 pL2/12
37
 FE EQUATION FOR BEAM cont.
• Finite element equation for beam

 12 6L 12 6L   v1   pL / 2   F1 
 6L 4L2 6L 2L2      pL2 / 12   C 
EI    1      1
 
L  12 6L 12 6L  v 2   pL / 2   F2 
3

 2 
  
 6L 2L2
6L 4L 
 2  pL2
/ 12 C2 

– One beam element has four variables

– When there is no distributed load, p = 0
– Applying boundary conditions is identical to truss element
– At each DOF, either displacement (v or ) or force (F or C) must be
known, not both
– Use standard procedure for assembly, BC, and solution
38
 PRINCIPLE OF MINIMUM POTENTIAL ENERGY
• Potential energy (quadratic form)
1
  U V  {Qs }T [Ks ]{Qs }  {Qs }T {Fs }
2
• PMPE
– Potential energy has its minimum when

[K s ]{Q s }  {Fs } [Ks] is symmetric & PSD

• Applying BC
– The same procedure with truss elements (striking-the-rows and
striking-he-columns)

[K ]{Q }  {F } [K] is symmetric & PD

• Solve for unknown nodal DOFs {Q}

39
 Exercise
• Calculate the tip deflection of cantilevered beam shown in the
figure. Use L=1m, EI = 104Nm2, and q=100N/m.
y
q

x 2EI EI
2 3
1
L L

40
3.4 BENDING MOMENT AND
SHEAR FORCE

41
 BENDING MOMENT & SHEAR FORCE
• Bending moment
d2 v EI d2 v EI
M(s)  EI 2  2 2  2 {B} T {q}
dx L ds L
– Linearly varying along the beam span
• Shear force  v1 
 
dM d3 v EI d3 v EI  
Vy (s)    EI 3   3 3  3 [  12 6L 12 6L]  1 
dx dx L ds L v 2 
 2 
– Constant
– When true moment is not linear and true shear is not constant, many
elements should be used to approximate it
My
• Bending stress x   
I
• Shear stress for rectangular section
1.5Vy  4y 2 
 xy (y)  1  2 
bh  h 
42
 y
EXAMPLE – CLAMPED-CLAMPED BEAM
• Determine deflection & x
slope at x = 0.5, 1.0, 1.5 m 1 2 3
• Element stiffness matrices 1m 1m
F2 = 240 N
v1 1 v2 2
v2 2 v3 3
 12 6 12 6  v1
 6   12 6 12 6  v 2
4 6 2  6
[k (1) ]  1000   1 4 6 2  2
 12 6 12 6  v 2 [k ]  1000 
(2) 
   12 6 12 6  v 3
 6 2 6 4  2  
 6 2 6 4  3

 12 6 12 6 0 0   v1   F1 
 
 6 4 6 2 0 0   1   C1 
 
 12 6 24 0 12 6  v 2  240 
1000      
 6 2 0 8  6 2 
 2  0 
 0 0 12 6 12 6  v 3   F3 
    
 0 0 6 2 6 4   3   C3 
43
 EXAMPLE – CLAMPED-CLAMPED BEAM cont.
• Applying BC
 24 0   v 2  240  v 2  0.01
1000      2  0.0
 0 8 
 2  0 

• At x = 0.5 s = 0.5 and use element 1

v( 21 )  v1N1( 21 )  1N2 ( 21 )  v 2N3 ( 21 )  2N4 ( 21 )  0.01 N3 ( 21 )  0.005m
1 dN3
( 21 )  v 2  0.015rad
L(1) ds s  21

• At x = 1.0 either s = 1 (element 1) or s = 0 (element 2)

v(1)  v 2N3 (1)  0.01 N3 (1)  0.01m v(0)  v 2N1(0)  0.01 N1(0)  0.01m
1 dN3 1 dN1
(1)  (1) v 2  0.0rad (0)  v  0.0rad
L ds s 1 L(2) 2
ds s0

Will this solution be accurate or approximate?

44
 EXAMPLE – CANTILEVERED BEAM
p0 = 120 N/m
• One beam element
• No assembly required
EI = 1000 N-m2
• Element stiffness C = –50 N-m
L = 1m
 12 6 12 6  v1
 6 4 6 2  1
[K s ]  1000  
 12 6 12 6  v 2
 
 6 2 6 4  2
• Work-equivalent nodal forces
 F1e   1  3s2  2s3   1/ 2   60 
C  
1 (s  2s  s )L 
2 3     
 1e   L / 12   10 
   p0L 0   ds  p0L   
 F2e   3s  2s
2 3
  1/ 2   60 
C2e   ( s2  s3 )L  L / 12 10 

45
 EXAMPLE – CANTILEVERED BEAM cont.
• FE matrix equation
 12 6 12 6   v1   F1  60 
 6 4 6 2   1   C1  10 
1000      
 12 6 12 6  v 2   60 
   
 6 2 6 4   2  10  50 
• Applying BC
12 6   v 2   60  v 2  0.01m
1000     

 6 4   2   60  2  0.03 rad

• Deflection curve: v(s)  0.01N3 (s)  0.03N4 (s)  0.01s3

• Exact solution: v(x)  0.005(x 4  4x3  x 2 )

46
 EXAMPLE – CANTILEVERED BEAM cont.
• Support reaction (From assembled matrix equation)
1000  12v 2  62   F1  60 F1  120N
1000  6v 2  22   C1  10 C1  10N m
• Bending moment
EI
M(s)  2 
B  {q}
L
EI
 2 ( 6  12s)v1  L( 4  6s)1  (6  12s)v 2  L( 2  6s)2 
L
 1000[ 0.01(6  12s)  0.03( 2  6s)]
 60s N m
• Shear force
EI
Vy   3 12v1  6L1  12v 2  6L2 
L
 1000[ 12  ( 0.01)  6( 0.03)]
 60N
47
 EXAMPLE – CANTILEVERED BEAM cont.
• Comparisons
0.000 0.000
FEM FEM
-0.002 Exact -0.005 Exact

-0.010
-0.004
-0.015
v


-0.006
-0.020
-0.008
Deflection
-0.025
Slope
-0.010 -0.030
0 0.2 0.4 x 0.6 0.8 1 0 0.2 0.4 x 0.6 0.8 1
10
FEM
0
Exact
-10
-20
M

-30
-40
-50 Bending moment Shear force
-60
0 0.2 0.4 x 0.6 0.8 1

48
3.5 BUCKLING OF BEAMS

55
 Review of Buckling of a Beam
• Tip deflection of a cantilevered beam
– Without P, CL2
v (L ) =
2EI
– Axial tension makes it difficult to bend
– Axial compression makes the deflection larger
• Free-body diagram with P
– Bending moment at x: M ( x ) = C - P ( d - v ( x ) ) Normally ignored
in beam bending
d2v
M ( x ) = EI 2 = C - P ( d - v )
dx
d2v
EI 2 - Pv = C - P d (2nd-order D.E.)
dx

56
 Beam Deflection under Tensile Force
d2v
• Solution of EI 2 - Pv = C - P d
dx
æC ö P
v ( x ) = A sinh l x + B cosh l x - çç - d ÷÷÷ l2 =
èP ø EI
• BCs to determine A & B
– v(0) = 0 & dv(0)/dx = 0
æC ö÷
ç
v ( x ) = ç - d ÷÷( cosh l x - 1)
èP ø
• Tip deflection  = v(L)
Cæ 1 ö÷
ç
d = v (L ) = ç 1 - , P>0
Pè cosh lL ø÷÷
CL2
• As P→∞, →∞ and →0. When P→0, →0, and d
2EI

57
 Beam Deflection under Compressive Force
• When P < 0 (compressive force)
æC ö P
v ( x ) = çç + d ÷÷÷( 1 - cos l x ), l = , P <0
çè P ø EI
• Tip deflection
C æ 1 ö
d = v (L ) = çç - 1÷÷÷, P < 0
P è cos lL ø

– Tip deflection is unbounded when lL  p / 2

P p p 2EI EI
l= = or P = Pc = » 2.47
EI 2L 4L2 L2

• Pc : critical load for buckling of the beam

– Pc is independent of C P > 0: stress stiffening
– Depend on EI and L  structural property P < 0: stress softening

58
 Energy Method for Beam Buckling
• Shortening of beam due to coupling of P and the flexural
deformation
– Assume no stretching due to P
– Axial deformation  due to bending (positive for shortening)
2 q dv
dD = dx ( 1- cos q ) » 2dx sin sin q »
2 dx
2 2
1æ dv ÷ö L 1 L æ dv ö÷
dD » çç ÷÷ dx D=ò dD = ò çç ÷÷ dx
2 è dx ø 0 2 0 è dx ø
End shortening of the beam

59
 Rayleigh-Ritz Method for Buckling
• Approximate the deflection of the beam v ( x ) = Ax 2
– Satisfies essential BC
• Strain energy
æ d2v ö÷2
L
1
EI çç 2 ÷÷ dx = 2EILA 2
2ò
U=
çè dx ø
0

• Potential energy of applied loads

dv
V = -C - ( P )( -D ) −: Same dir. with positive P
dx x =L
L
1 2 2 3 2
V = -2CLA + P ò ( 2 Ax ) dx = -2CLA + PL A
2 3
0

• Total potential energy

2 2 3 2
P = U + V = 2EILA - 2CLA + PL A
3
60
 Rayleigh-Ritz Method for Buckling
• PMPE:
dP æ 4 3 ÷ö C
ç
= 0  ç 4 EILA - 2CL + PL A ÷÷ = 0  A =
dA è 3 ø 2EI + 2PL2 / 3

• Tip deflection (tensile force)

2 CL2
d = v (L ) = AL = 2
, P>0
2EI + 2PL / 3

• Tip deflection (compressive force)

CL2
d= 2
, P <0
2EI - 2 P L / 3
• Critical load (unbounded deflection)
EI
Pcr = 3 2
L
exact EI
– About 20% larger than exact value Pcr » 2.47 2
L
61
 Rayleigh-Ritz Method for Buckling
• Non-dimensionalization
d
d =
CL2 2EI
1
= 2
, lL > 0
1 + 0.333 ( lL )
1
= 2
, lL < 0
1 - 0.333 ( lL )

Unbounded when l L  3

• Non-dimensionalized exact solution

2æ 1 ö÷
dexact = ç 1- ÷÷, lL > 0
2ç
è cosh l ø Unbounded when lL  p 2
( lL ) L
2 æ 1 ö÷
= çç - 1÷ø, lL < 0
2 è cos lL
( lL )

62
 FE Method for Buckling
• Already have matrix form of [K]{Q} = {F}. So only new term…
• Potential (work done) by axial load
NE
P(e): element axial force
Vinc = -å P (e ) (e )
(-D )
e=1 Need to express (e) in terms of nodal DOFs

• Axial shortening
1 x j æ dv ö2 1 1æ dv ö2
D(e) = ò çç ÷÷÷ dx = ò çç ÷÷÷ ds
2 xi è dx ø 2L 0 è ds ø
dv
ds
= { }
dN
ds
{q}

(e ) 1 T
D = {q} [kinc ]{q}
2
é 36 3L -36 3L ù v i
1 ê 2 2
ú
1 1 ê 3L 4L -3L -L ú qi
[k inc ] = ò ¢ ¢ T
{N }{N } ds = ê ú
L 30L ê -36 -3L 36 -3L ú v j
0 ê ú
êë 3L -L -3L 4L úû q j
2 2

Incremental stiffness matrix (element) 63

 FE Modeling for Buckling
• Potential (work done) by axial load
NE NE
1 T (e ) P(e): element axial force
Vinc = -å P (e )
(-D ) = å P
(e ) (e )
{q} [k inc ]{q}
e=1 e=1 2 Pr: global axial force
NE
P (e ) 1 T (e ) 1
Vinc = å Pr {q} [kinc ]{q} = {Q}T [Pr Kinc ]{Q}
e=1 Pr 2 2

ìï NE P ( e ) ü
ï (e ) ï Global incremental
[Pr K inc ] = Assembly í å Pr [k inc ] ïý
ïï e=1 Pr ïï stiffness matrix
î þ
• Total potential energy
1 1
P= {Q } T [K ]{Q } - {Q } T {F } + {Q } T [Pr K inc ]{Q }
2 2
1
= {Q } T [K + Pr K inc ]{Q } - {Q } T {F }
2
– [Kinc] add a positive definite matrix to the stiffness matrix, thus
increasing the strain energy of the system
– Negative axial force makes the beam softer or more compliant
64
 Eigenvalue Problem for Buckling
• Apply PMPE
dP
=0 [K + Pr K inc ]{Q } = {F}
d{Q }
[KT ] : total stiffness matrix
– If the axial force Pr is such that |KT| = 0, then {Q} is unbounded
– Since [K] is positive, Pr < 0 (compressive) to make |KT| = 0
• Critical load (Pcr): The negative value of Pr to make |KT| = 0

K - Pcr K inc = 0

– Usually the lowest Pcr is of concern

• For calculating Pcr, external forces do not matter
[K - Pcr Kinc ]{Q} = {0} Eigenvalue problem for buckling
[K]{Q} = Pcr [Kinc ]{Q} {Q}: Mode shape of the buckled beam

65
 Cantilever Beam Example
 12 6L 12 6L  é 36 3L -36 3L ù v i
 6L 4L2 6L 2L2  ê ú
EI
[k   ]  3 
2 2
e
 1 ê 3L 4L -3L -L ú qi
L  12 6L 12 6L  [k inc ] = ê ú
 2 
30L ê -36 -3L 36 -3L ú v j
 6L 2L 6L 4L 
2
ê ú
êë 3L -L -3L 4L úû q j
2 2

y
3
C
2.5
x P
v(x) [m]

1.5
L
1

0.5

0
0 0.2 0.4 0.6 0.8 1

x [m]

v (P=0) v (P=+2000 N) v (P=-2000 N)

66
 Example: Buckling of a Cantilevered Beam
• Cantilevered beam with L = 1m, EI = 1,000 Nm2, C =
1,000Nm, calculate Pcr using one beam element
y
C
• Exact solution:
2
x P
p EI L
Pcrexact
1 = 2
= 2,467N
4L
9p 2EI Pcr 1 p Pcr 2 3p
Pcrexact
2 = = 22,207N l1 = = , l2 = =
4L 2 EI 2L EI 2L

• Finite element solution

– Single element, apply BC in element level
é 12 -6 ù v2 1 é 36 -3 ù ìï 0 üï ì
ïv2 ü ï
[K] = 1000 ê ú [Kinc ] = ê ú {F} = í ý {Q} = í ý
ëê -6 4 ûú q2 30 êë -3 4 úû îïï1,000þïï ïï
î q2 ïï
þ

67
 Example: Buckling of a Cantilevered Beam cont.
• Eigenvalue problem
é 12000 - 1.2Pcr -6000 + 0.1Pcr ùì ïv 2 üï ìï0üï
[K - Pcr K inc ]{Q } = 0  ê úí ý = í ý
êë -6000 + 0.1Pcr 4000 - 0.133Pcr úû ïî
ï q2 ïþ
ï ïî ïï
ï0þ
– Let  = Pcr/1000
2
| K - Pcr K inc |= 0  ( 12 - 1.2b )( 4 - 0.133b ) - ( -6 + 0.1b ) = 0

b1 = 2.468, b2 = 32.180  Pcr 1 = 2,486N,Pcr 2 = 32,180N

Pcrexact
1 = 2,467N,Pcrexact
2 = 22,207N

– Error in the 1st critical load = 1%, 2nd critical load = 45%
– More elements for accurate higher buckling loads
– FE critical loads are higher than true values (FE model is stiffer than
actual stiffness)

68
 Example: Buckling of a Cantilevered Beam cont.
• Mode shape for Pcr1 = 2,467N
é 9017 - 5751ù ïì v 2 ïü ìï 0 üï
[K - Pcr 1K inc ]{Q 1 } = ê úí ý = í ý
êë - 5751 3669 úû ï
îï q2 ï
þï ïîï 0 ïþï

– Matrix is singular  cannot solve the equation  infinitely many solns

– Can only get the relation between v2 and 2 q2
= 1.57
– Choose v2 = 1, 2 = 1.57 v2
v (s ) = v 2N3 (s ) + q2N4 (s ) = 1.43s 2 - 0.43s 3 (first mode)
• Mode shape for Pcr2 = 32,180N (v2 = 1, 2 = −9.57)
v (s ) = v 2N3 (s ) + q2N4 (s ) = 12.57s 2 - 11.57s 3 (second mode)
2.5

1.5
v(x)

0.5

0
0 0.2 0.4 0.6 0.8 1
x [m]

Mode 1 Mode 2
69
 Exercise: Buckling of a Clamped-Hinged Beam
• Calculate the buckling loads and corresponding mode shapes
of the beam using: (a) one element and (b) two elements of
equal length. Assume beam length 2L=2m and EI=1,000N-m2

70
3.6 FE MODELING
PRACTICES FOR BEAMS

76
 Stress and Deflection Analysis of a Beam
• Simply supported beam with distributed load at overhangs
• Determine the maximum bending stress σ and the deflection
δ at the middle portion
• Standard 30in wide-flange beam, A = 50.65in2, Izz = 7892in4,
w = 10,000lb/ft, E = 3.0x107psi, L = 20ft, a = 10ft, h = 30in.
w w

z
h

a L a

1 2 3 4
x
1 2 3 4 5

77
 Stress and Deflection Analysis of a Beam cont.
• bending moment in the middle portion (pure bending)
wa2
M= = -6 ´106 lb ⋅ in
2
• Stress on the top surface of the middle portion
M h2 -6 ´106 ´15
sxx =- =- = 11,404psi
Iz 7892
• Deflection: EIzz y ¢¢ = M = constant
y = a 0 + a1x + a2x 2 y(0) = y(240) = 0

y ( x ) = a2 x ( x - 240), y ¢¢ = 2a2
M
a2 = = -1.2671´ 10-5
2EIz

y (120) = -1.2671´ 10-5 ´ 120 ´ (120 - 240) = 0.1825in

78
 Stress and Deflection Analysis of a Beam cont.
• Single element for middle section (pure bending), but we will
use 2 elements to get deflection at the middle section + 2
elements for overhang

EI 6
M (s ) = ê B
2 ë û
ú {q} = - 6 ´ 10 lb ⋅ in
L

79
 Portal Frame under Symmetric Loading
• I-Beam sections with a uniformly distributed load ω = 500lb/in
across the span
• determine the maximum rotation and maximum bending
moment
• Icol = 20,300in4 and Ispan = 101,500in4.

qmax = 1 wa 3 / EI col
27

M max = 19 wa 2
54
 Buckling of a Bar with Hinged Ends
• Determine the critical buckling load of an axially loaded long
slender bar of length L with hinged ends, as shown in Figure
3.. The bar has a square cross-section with width and height
set to 0.5 inches. Determine the critical buckling load of an
axially loaded long slender bar of length with hinged ends.
The bar has a square cross-section with width and height set
to 0.5 inches
 Exercise
• Calculate the deflection curve v(s), bending moment M(s),
and shear force Vy(s) of the simply-supported beam :
p0 2 3
N1(s ) = 1 - 3s + 2s
N2 (s ) = L(s - 2s 2 + s 3 )
N3 (s ) = 3s 2 - 2s 3
E,I,L = 1
N4 (s ) = L(-s 2 + s 3 )

 12 6L 12 6L   v1   pL / 2   F1 
 6L 4L2 6L 2L2      pL2 / 12   C 
EI    1      1
 
L  12 6L 12 6L  v 2   pL / 2   F2 
3

 2 
  
 6L 2L2
6L 4L 
 2  pL2
/ 12 C2 