242

Finite Element Analysis

14
Analysis of Beams and Rigid Frames
14.1 INTRODUCTION
In Chapter 3 direct stiffness formulation of beams was discussed, which is actually stiffness matrix method for the analysis. In Chapter 5 shape functions were derived for two noded beam element by considering polynomial interpolation function and also using Hermitian functions both ultimately result into same shape functions. Assembling of strain displacement matrix [B-matrix] was presented in Chapter 4. In this chapter assembling of stiffness matrix for a two noded beam element by variational approach is presented. Taking simple problems for hand calculation, formulation of system equations and solution procedure is explained. The members of rigid frames are similar to beam elements, but their orientation in global system are different. Transforming the stiffness equations of beam element to global system is required before assembling global system. This aspect is presented and illustrated with numerical problem taking simple frame analysis. It was found difficult to extend the beam theory discussed above to plate bending. A new beam theory was developed in which lateral deflection w and rotation x were decoupled and treated as independent variables. However it needs inclusion of shear deformation and hence shear strain energy. Hence the beam theory becomes Co-continuity problem. This is known as Timoshenko Beam theory and the element developed on this theory is also presented in this chapter.

14.2 BEAM ANALYSIS USING TWO NODED ELEMENTS

The typical beam element is shown in Fig. 14.1. Note the orientation of axes is as per the right hand thumb rule.
y le x x = le s=1
=1

x=0 z s=0
= 1

le 2 1 s 2 =0

Fig. 14.1

A typical beam element

The nodal variable vector is

l q

= 1 2

3 4 = w1 1 w2 2

Analysis of Beams and Rigid Frames where wilateral displacement at node i.

243

i Rotation at node i.
In Chapter 5 the shape functions for such element have been determined as,

N1 = l

3x 2
2 le

2x3
3 le

N2 = x 3x 2
2 le

2x2 x3 + 2 le le 2x3
3 le

N3 =

and N 4 =

x2 x3 + 2 le le

(14.1)

If non-dimensiolising is done using s =

x , the shape functions are (Chapter 4, Art 4)., le

N1 = 1 3s2 + 2s3 N2 = le s(s 1)2 N3 = s(3 2s) N4 = le s2(s 1) (14.2)

If we use non-dimensiolising concept as used in isopatametric formulation i.e. varying from 1 to 1, then the shape functions are

N1 = N2 = N3 =
where

2 3 + 3 4 le 1 2 + 3 2 4 l 1 + 2 + 3 2 + 3 3 and N4 = e 2 4 4
(14.3)

2x 1. le

Moment Curvature Relation

From basic solid mechanics we know
M = EI

2y x 2

Since

y= N

lq

= N1

N2

N3

N4

lq

244

Finite Element Analysis

we get

M = EI = D

LM N MN x
2 2

2 N2 x
2

2 N3 x
2

2 N4 x
2

l q

OP l q PQ

e e

(14.4) (14.5)

where

[D] = EI

B =

LM N NM x
2 2

2 N2 x 2

2 N3 x 2

2 N4 x 2

OP QP

(14.6)

which is stress resultant curvature matrix. Strain Energy From basic solid mechanics we know strain energy dUe in an elemental length dx is given by

dU e =

1 M2 dV 2 EI 1 2
1 2

Ue =

z z
le 0 0

REI y U | | S x V | | T W
2 2

I dx EI
1 2

le

EI

2y x
le 2

dx =

1 = 2

lq

T e

z
0

z
0

le

EI B e

d l qi
le d 2

lq

dx

EI B

lq

since

2x 1 le

l U e = EI e 2

l q
1 2

T e

z
0

le

lq
2

(14.7)

Now,

B =

LM N MN x
2

2 N2 x
2

2 N3 x

2 N4 x
2

OP PQ

since

2x 1 le

Ni Ni = x

Ni = x

2 le

Analysis of Beams and Rigid Frames

245

2 Ni x
2

2 Ni = x x le 4
2 le

FG IJ H K
2 Ni
d 2
2

F2 Gl H
e

Ni

I J K

B =

LM N N N N OP l M P N Q 4 L 6 = MM 4 b1 43g l 64 b1 +43 g l OPP l N Q 1 6 b1 3 gl 6 b1 + 3 g l = l

4
2 2 2 1 2 3 4 2 e 2 2 2 2 2 e e e 2 e e e

(14.8)

LM 6 OP 1 Mb1 3 gl P 1 = PP l 6 b1 3gl 6 b1 + 3g l l M 6 MM l b1 + 3g P N Q LM36 b1 3 gl b6 g 36 OP 6 b1 + 3 gl l b1 3 g 6 b1 3 gl b1 + 3 gb1 3 gl P 1 M = MM PP 36 6 b1 + 3 gl SYM l MN PQ l b1 + 3 g

e 2 e 2 e e e e 2 2 e e 2 e 2 e 2 e 4 e 2 e 2 e 2 1

(14.9)

Now noting that

z
1

d = 2,

z
1

2 d = 0, and

z
1

3d =

2 3

We can write

1 U e = EI 2

lq

T le e 2

1 4 le

LM12 1 EI M = l q 2 l M MN
T 3 e

LM24 MM MN

12le 2 8le SYM

24 12le 24

12le 2 4le 12le 2 8le

12 6le 6le 2 2 6le 21le 4le SYM 12 6le 2 4le

OP PP lq PQ

OP PP l q PQ

1 2

lq

T e

l q

(14.10a)

246

Finite Element Analysis Where [k]e is element stiffness matrix and

LM12 EI M is equal to l M MN
3 e

12 6le 2 6le 21le SYM 12 6le 2 4le 6le 2 4le

OP PP PQ

(14.10b)

Potential Energy Potential energy of an element is equal to strain energy minus the work done by the external forces acting on the element. Thus

1 e = Ue 2

z
le 0

pydx Pm Ym

Mk

F dy I G dx J H K

(14.11)
k

Where pDistributed load per unit length PmConcentrated load at point m MkExternal moment applied at k. The strain energy term Ue has been already derived. The work done by external loads can be assembled as explained below: Due to Uniformly Distributed load P/unit length:

pl = e 2

z
1

LM 2 3 + NM 4

py dx =

p N1

N2

N3

N4

l q

le d 2

le 1 2 + 3 2 4

2 + 3 3 4

le 1 + 2 + 3 2 4

Noting that

z z z
1 1 1

C d = 2C,

z z
1 1

OP l q d QP
e

d =

3 d = 0 and

2 d =

2 , we get 3 ple l 2 22 e 2 8 2 3

py dx =

LM N

F G H

I J K

22

le 2 2 + 2 3

F G H

I OP l q JQ K

LM pl NM 2

2 ple 12

ple 2

2 ple 12

OP lq QP

(14.12)

This equivalent load on the element is shown in the Fig. 14.2. The point loads like Pm and Mk are readily taken care by introducing nodes at the points of application.

Analysis of Beams and Rigid Frames Thus work done by external load is assembled. Let it be represented by

247

F
i

= F1 1 + F2 2 + F3 3 + F4 4 =

l q lFq
T e

(14.13)

y
p/unit length

x le z ple 2 ple 2

le pl2 e
12

pl2 e 12

Fig. 14.2

Equivalent nodal loads of UDL

Minimization of Potential Energy It is to be noted that, in solid mechanics, minimization of potential energy of entire structure is to be assembled. In finite element analysis the total potential energy of the system is considered as the summation of total potential energy of the elements. Thus
= e = Ue

l q lFq
T e

1 F 2

l q k l q l q lFq
T T

where { } and {F} are nodal unknown vector and load vector respectively. From the principle of minimization of potential energy we get,
d =0 d

k F=0
or

lq

k =F

lq

(14.15)

In finite element analysis, element stiffness matrix [k]e is assembled and placed in global matrix at appropriate place. When this process is completed for all the elements, we get global stiffness matrix [k]. Similarly global load vector {F} is assembled. The necessary boundary conditions are imposed by (i) Elimination Method if the hand calculations are made or by (ii) Penalty Method if computers are used.

248

Finite Element Analysis

The solution of equation 14.14 gives the displacement vector { }. The required stress resultants are determined for each element. Moment at nodes:

M = EI B
= EI
2 le

lq

1 3 le

b1 + 3 gl l q
e

(14.15)

Shear forces at nodes

V=
=

dM dM d 2 dM = = dx d dx le d
6 3le 6 3le

2 EI
3 le

l q
k

EI
3 le

12 6le

12 6le

l q

(14.16)

The reactions at supports are nothing but end equilibrium forces. Hence

lRq =

lq

(14.17)
R4

R2

le R1 R3

Fig. 14.3 Positive reactions

For uniformly distributed load the above equation will be

RR U | | | | |R | | | = EI S V l |R | | | | | |R | | | T W
1 2 3 3 e 4

LM 12 MM MM 6l MM12 MM MN 6l
e e

6l e
2 4l e

12 6l e 12 6l e

6l e
2 2l e

OP R PP | | 2l P | PP | S 6l P | PP | | 4l P | T Q|
6l e
2 e e 2 e

U R pl | |2 | | pl | | | | 12 V S pl | | | |2 | | pl | | | T 12 W |
e 2 e e

2 e

U | | | | V | | | | | W

Note that the above reactions are to be interpreted as per sign conversion i.e. R1, R2 are positive in positive direction of z-axis (i.e. downward) and R2, R4 (the end moments) are positive when they are clockwise (refer Fig. 14.3). Example 14.1: Analyse the beam shown in Fig. 14.4 (a) by finite element method and determine the end reactions. Also determine the deflections at mid spans given E = 2 105 N/mm2 and I = 5 106 mm4

Analysis of Beams and Rigid Frames

24 kN/m 12 kN/m

249

5m (a)
1 3

5m

1
2

1
4

2
6

(b)

Fig. 14.4

Solution: Using kN and m units throughout,

E = 2 105 N/mm 2 = 2 105

106 103

= 2 108 kN/m 2

I = 5 106 mm4 = 5 106 m4 EI = 2 108 5 106 = 1000kN-m2 Let the two elements be numbered as shown in Fig. 14.4 (b). The nodal displacement vector is

l q

= 1 2

3 4 5 6
6l1 12 6l1 12 30 12 30 2 2 4l1 6l1 2l1 1000 30 100 30 50 = 3 6l1 12 6l1 12 30 12 30 5 2 2 2l1 6l1 4l1 30 50 30 100
3 4 Global Numbers

k1 =

E1 I1
3 l1

LM 12 MM 6l MN612 l
1 1 2

OP PP PQ

LM MM MN

OP PP PQ

Similarly,

LM 12 30 =8M MM12 N 30 LM 12 30 =8M MM12 N 30

3

30 12 30 100 30 50 30 12 30 50 30 100

OP 1 PP 2 3 PQ 4

30 100 30 50

12 30 12 30

k2

OP 50 P 30P P 100 Q
30

Global Numbers

3 4 5 6

250

Finite Element Analysis

LM 12 MM 30 12 = 8M MM 30 MM N
LM 12 MM 30 12 = 8M MM 30 MM 0 N0
1

30 12 100 30 12 30 +12 50
30 +30

30 50
30 +30 100 +100

12 30 12 30

12 30

30 50

OP P 30 P P 50 P P 30P 100 P Q
6

global numbers

1 2 3 4 5 6

30 12 30 0 0 100 30 50 0 0 30 24 0 12 30 50 0 200 30 50 0 12 30 12 30 0 30 50 30 100

OP 1 PP 2 PP 3 4 PP 5 PQ 6

global numbers

Consistent load vector is given by

R pl |2 | pl | | F = S 12 | pl |2 | pl | | 12 T
e 2 e e

2 e

U | | | | V | | | | | W
2

R b12g 5 | 2 | | b12g 5 | 12 | Fb g = S | b12g 5 | 2 | b12g 5 | | T 12

1

U | | R30U 1 | | | | |25| 2 |=S V V |30| 3 | | | | T 25 W 4 | | | W

Global Numbers

Analysis of Beams and Rigid Frames

251

R b24g 5 | 2 | | b24g 5 | 12 | Fb g = S | b24g 5 | 2 | b24g 5 | | 12 T

2 2

U | | R60U 3 | | | | |50| 4 |=S V V |60| 5 | | | | T 50 W 6 | | | W

Global Numbers

R30U |25| |90| | | lF q = S |25V |60| | 50 | | | W T |

The stiffness equation is

LM 12 MM 30 12 8M MM 30 MM 0 N0

30 12 30 0 0 100 30 50 0 0 0 12 30 30 24 50 0 200 30 50 0 12 30 12 30 0 30 50 30 100

OP R PP | | PP | S PP | | | QP | T

1 2 3 4 5 6

U R30U | |25| | | | | = |90| V S25V | | | | |60| | | 50 | | | | W T W

Boundary conditions: In the given problem the boundary conditions are

1 = 2 = 3 = 5 = 0
Imposing them by elimination method, we get

i.e.,

i.e.,

LM200 50 OP R U = R25U N 50 100Q S V S 50 V T W T W L4 1OP R U = R25U 400 M N1 2Q S V S 50 V T W T W R U = 1 LM4 1OP R25U S V 400 N1 2Q S 50 V T W T W
4 6

4 6

4 6

252

Finite Element Analysis

=
End reactions

2 1 1 1 400 8 1 1 4

LM N

OP R25U = 1 LM100OP Q S 50 V 2800 N 225 Q T W

Answer

RR U LM 12 |R | M 30 | |=8 SR V M12 | | M |R | N 30 T W
1 2 3 4

12 30 100 30 50 30 12 30 50 30 100 30

OP R 0 U R30U | | PP | 0 | |25| S 0 V |30| S V PQ | 100 | | 25 | | 2800 | | | T W T W

Answer

since

1 = 2 = 3 = 0 and 4

. R 21429 U | 10.714 | 100 | | = =S V 2800 |38.571| |53571| T . W

For element 2

RR U LM 12 |R | M 30 | |=8 SR V M12 | | M |R | N 30 T W
3 4 5 6

30 12 30 100 30 50 30 12 30 50 30 100

0 | | | | OP R 100 U R60U R70.714U | | S V | . | PP | 2800 | |50| = |53571| S 0 V |60| S49.286V PQ | 225 | | 50 | | 0 | | | T W | | T W | 2800 | T W

Answer

Deflection at mid span

y = N1 =

N2

N3
3

N4

LM 2 3 + MN 4

l q

le 1 2 + 3 2 4

2 + 3 3 4

le 1 + 2 3 2 4

OP l q PQ

For mid span = 0

y centre = 0.5 0125le .

For element 1,

0.5 0125le .

l q

y1 centre = 0.5 0125 5 0.5 .

R 0 U | 0 | | | 0125 5 S 0 V . | 100 | | 2800 | T W

= 0.02232 m = 22.32 mm

Analysis of Beams and Rigid Frames For element 2,

253

y2 centre = 0.5 0125 5 0.5 .

R 0 U | 100 | | | | | 0125 5 S 2800 V . 0 | | 225 | | | | 2800 W T

= 0.02790 m = 27.9 mm = 27.9 mm, downward Answer Example 14.2: A beam of length 10 m, fixed at one end and supported by a roller at the other end carries a 20 kN concentrated load at the centre of the span. By taking the modulus of elasticity of material as 200 GPa and moment of inertia as 24 106 m4, determine: 1. Deflection under load 2. Shear force and bending moment at mid span 3. Reactions at supports Solution: The beam is shown in Fig. 14.5 (a). Its finite element idealization is shown in Fig. 14.5 (b). In this problem, E = 200 GPa = 200 109 N/m2 = 200 106 kN/m2
20 kN

5m (a)
1 3

5m

(b) 6.25 37.25 0 31.25 13.75 (c)

13.75

6.25

Fig. 14.5

254

Finite Element Analysis Nodal displacement vector is

l q
ki = Ei I i
3 le

= 1 2

3 4 5 6

Stiffness matrix for an element is given by

LM 12 MM 6l MN612 l
e e

6le 2 4le 6le 2 2le

2

12 6le 12 6le
3

6le 2 2le 6le 2 4le

4

OP PP PQ
, Global Numbers

k1 =

4800 125

LM 12 MM 30 12 MN30
4

30 100 30 50

12 30 12 30

OP 50 P 30 P P 100 Q
30 3 4 5 6

1 2 3 4

Similarly,

k2

LM 12 4800 M 30 = 125 M12 MN 30 LM 12 MM 30 4800 12 = M 125 M 30 MM 0 NM 0

1

30 100 30 50

12 30 12 30

OP 50 P 30P P 100 Q
30
4

, Global Numbers

30 12 30 0 0 100 30 50 0 0 30 24 0 12 30 50 0 200 30 50 0 12 30 12 50 0 30 50 30 100

OP PP PP PP QP

, Global Number

1 2 3 4 5 6

The consistent load vector is directly available as

lFq
The stiffness equation is The boundary conditions are,

= 0 0 20 0 0 0 k = F

lq lq

1 = 2 = 5 = 0

Analysis of Beams and Rigid Frames Imposing these boundary conditions by elimination method, the stiffness equation reduces to

255

EI 125
By using direct inversion, we get

LM24 MM 0 N30

0 200 50

OP R | 50 P S 100P | QT
30

3 4 6

U R20U |=| 0 | V S V | |0| W T W

1500 6000 2400 900 1200 4800 1200

R | S | T

3 4 6

U 125 LM20000 2500 1 |= V EI 24b20000 2500g + 30b6000g M 1500 | MN 6000 W LM17500 1500 6000OP R20U 125 1 | | = MM 1500 1500 1200PP S 0 V EI 240000 T W N6000 1200 4800 Q | 0 | R182.292U 1 R182.292U R 0.03798 1 | = S 15.625 | = 4800 | 15.625 | = |3.2552 10 V S EI | | | 62.5 V S 0.01302 | | T 62.5 W T W T

OP R20U | | PP S 0 V T W Q |0|

U | V | W

Deflection under the load = 3 = 0.03798 and rotation under the load = 4 = 3.2552 106 radians Shear Force and bending Moment at midspan: Considering element (1), for mid span of beam = 1

LM 0 OP 0 1 M EI P 6 2l 6 4l M= EI M 182.292P l MN 15.625 PQ LM 0 OP 0 1 = 6 10 6 20 M MM182.292PPP = 31.250 KN m Answer 25 N 15.625 Q LM 0 OP 0 1 M EI P = 31.75 kN Answer 12 6 5 12 6 5 V = 125 EI M182.292P MN 15.625 PQ
2 e e e

Considering element 2, shear force at mid span may be found as 6.25 kN

256

Finite Element Analysis

End Reactions At support on left hand side, from element 1,

R U SR V T W
R1
2

0 0 EI 12 30 12 30 1 = 0 125 30 100 30 50 EI 182.292 15.625

LM N

LM OP M Q MM N

OP PP R U SV PQ T W

RR U = R13.75U SR V S37.50V T W T W
1 2

Considering element 2, the reactions at right hand supports can be obtained

R U S V T W

R5 EI 12 30 12 30 1 = R6 125 30 100 30 100 EI

LM N

LM182.292OP OP M 15.625 P R0U Q MM 0 PP S0V TW 62.5 Q N

R6.25U S0V T W

The reactions and shear force and bending moment at midspan are shown in Fig. 14.5 (c). Example 14.3: Derive the expression for consistant load, which varies linearly from p1 at node 1 to p2 at node 2 on a beam element of length le. Solution: The element with given load is shown in Fig. 14.6.

p2 p1
1 2

le

Fig. 14.6

The load intensity at a point x distance from node 1 is given by

p = p1 + p2 p1

g lx
e

= p1 + p2 p1
= p1

g b1 + g since lx = 1 + 2 l
e
e

1 1+ + p2 2 2

(14.18)

Analysis of Beams and Rigid Frames

257

lFq =
e

z z
1

N T pdx NT p le d = 2

N T p1

1 le d + 2 2

N T p2

1 + le d 2 2
3

z
1

R 2 3 + U | | 4 | | | 1 + l | | | 4 2 | 1 l | S 2 + 3 V 2 p 2 d + | | | 4 | | 1 + + l | | | | 4 2| W T
3 2 3 e e 2 1 2 3 e 1

z
1

R 2 3 + U | | 4 | | | 1 + l | | | 4 2 |1+ | S 2 + 3 V 2 p | | | 4 | | 1 + + l | | | | 4 2| W T
2 3 e 2 2 3 e 2

le d 2

Now

z
1

F 2 3 + GH 4

I b1 g d = e2 5 + 3 JK
1 1

Noting that

1 1

and

2 3 + 3 2 2 28 1 d = 2 2 0 + 3 + 0 = 4 3 5 5 1 The first term in p1 is, p1le pl 28 = 1 e 21 16 5 60 Similarly the other terms may be evaluated. Finally we get,

z
1

z z z
1 1

+ 3 + 4 d

c d = 2c,

1 3

z z
1 1

d =

3 d = 0

L O d = M P MN 3 PQ L O d = M P MN 5 PQ
5

=
1

2 3

=
1

2 , 5 we get

b g

R 21 U R 9 U | | | | p l | 3l | p l | 2l | Fq = l 60 S 9 V + 60 S 21 V | | | | |2l| |3l| T W T W
1 2 e

(14.19)

258

Finite Element Analysis

Example 14.4: Determine the consistent nodal vector due to loads acting on the beam shown in Fig. 14.7.
100 kN 20 kN/m

4m (a)

6m

F1

F3

F5

F2

F4
(b)

F6

Fig. 14.7

Solution: The beam is idealized with two elements as shown in Fig. 14.7 (b). Due to concentrated load, the nodal vector is directly obtained since there is a node directly under the load. It is given by

Due to udl, it is given by the expression

e

R0U |0| | | |100| F =S |0V |0| |0| | | W T |

R pl U |2 | | | | pl | | | lFq = | 12 | S pl V | | |2 | | pl | | | | 12 | T W R 20 4 | 2 | 16 | 20 | 12 lFq = S | 20 4 2 | |b20g 16 | 12 T

2 e e e 2 e 1

U | | R 40 U 1 | |26.667| 2 |=| V V S 40 | 3 | | | | | 26.667 | 4 W | T | W

Global Numbers

Analysis of Beams and Rigid Frames

259

R 20 6 U | 2 | | | 20 36 | R60U 4 | |60| 3 | | 12 | | lFq = S = 6 V S60V 5 | 20 2 | | | | |b20g 36 | | 60 | 6 | T W | | 12 W T

e

Global Numbers

Due to udl,

Hence due to udl and the concentrated load

R 40 U R 40 U | 26.667 | |26.667| | | b40 + 60g | = | b100g | lFq = S26.667 60| |33.333| V S V | | | | 60 | | 60 | | | 60 | | 60 | | T | | W | W T

R 40 U |26.667| | 200 | lFq = |33.333| V S | | 60 | | | 60 | | | W T

14.3 ANALYSIS OF RIGID PLANE FRAMES USING 2 NODED BEAM ELEMENTS

Answer

The members of a rigid frame differs from the beam in the following two respect: (i) They carry axial loads also and hence their deformation in axial direction also is to be considered. (ii) They are oriented in any direction in the plane. The typical frame element is shown in Fig. 14.8. The element selected is in x y plane. The right hand thumb rule is used for the selection of Cartesian coordinates. Let x y be the local coordinate system. The nodal displacement vector in local coordinate system is

l q

T e

= 1 2

5 3 4 6

(14.20)

and the nodal displacement vector in global system is,

lq

T e

= 1 2

3 4 5 6

(14.21)

260

Finite Element Analysis

'5 '6 '4 x'
5

x'
6 4

y y' '2

y '1 y'
2

'3

3 1

x
(a) (b)

Fig. 14.8

A typical frame element

It may be observed that,

1e = 1e cos + 2 e sin = 1e l + 2 e m
e = 1e sin + 2e cos = 1e m + 2e l 2 e = 3e 3
Similarly

e = 4 e l + 5e m 4 5e = 4 e m + 5e l

e = 6e 6
where l, m are direction cosines

l q = L l q
e

(14.22a)

where

LM l MM m 0 =M MM 0 MM 0 N0

m 0 0 0 0 l 0 0 0 0 0 1 0 0 0 0 0 l m 0 0 0 m l 0 0 0 0 0 1

OP PP PP PP PQ

(14.22b)

1 and are like the degrees of freedom of bar element while , 3 , 5 and are like the degrees 4 2 6 of freedom of beam element. Hence the stiffness matrix of frame element in the local coordinate system can be obtained by appropriately placing the two stiffnesses as

Analysis of Beams and Rigid Frames

261

LM EA MM l MM 0 MM 0 =M MM EA MM l MM 0 MM 0 N
e e

0 12EI
3 le 6 EI 2 le

0 6 EI
2 le 4 EI

EA le 0 0

0 12EI
3 le 6 EI 2 le

le 0 6 EI
2 le 2EI le

0 12EI
3 le 6 EI 2 le

EA le 0 0

0 12EI
3 le 6 EI 2 le

OP P 6 EI P l P P 2EI P l P P 0 P P 6 EI P P l P 4 EI P P l P Q
0
2 e 2 e 2 e e

(14.23)

From equation 14.10, we know strain energy of the element is given by

Ue = 1 2

But from equation 14.22

e

l q = L l q U = e L l q j = l q L
e e T e

b g
e

T e

l q

e T

k k

lq L l q = l q
L
e e

T e

where

k e= L

(14.24)

[k]e is element stiffness matrix in global coordinate system. Using stiffnesses matrices of all elements, system stiffness matrix [k] can be assembled.

Load Vector
ple 2
2 ple 12

x'

y y' ple 2 x

y pl2 e 12 x

Fig. 14.9

Element load vector due to distributed load

262

Finite Element Analysis Figure 14.9 shows the uniformly distributed load acting on the typical element. Noting that the load

acting is in y -direction, the nodal force system in local coordinate system is

By transforming it to global system, we get, {F}e = [L ]T {F }e . Using such expressions for all elements the load vector of the system due to the distributed loads can be assembled. The load vector due to concentrated loads and external moments can be directly added to global load vector, since at all such loads nodes are selected. The final system equation is [ k ] { } = {F} as usual. After introducing boundary conditional, the system equations are solved to get nodal displacement vector { }. Then the required stress resultants can be assembled. Example 14.5: Assemble element stiffness matrices for the rigid frame shown in the Fig. 14.10 (a). Explain how do you proceed further to solve the problem. Take, E = 200 GN/m2 Io = 40 106 m4 A = 4 103 m2
20 kN/m

lF q = LMM0 N
e

ple 2

2 ple 12

ple 2

2 ple 12

OP QP

(14.25)

30 kN 2 2I 0

3m

I0 1

I0

4 4m (a)

5 64

8 7

2 1 3 (b)

11 12 10

Fig. 14.10

Analysis of Beams and Rigid Frames

263

Solution: Figure 14.10 (b) shows the positive directions of the 12 nodal displacements. Thus in this problem

lq = l
T

2 ... 12

q
0 12 EI
3 le 6 EI 2 le

For any element, stiffness matrix is given by

LM EA MM l MM 0 MM 0 =M MM EA MM 0l MM MM 0 N
e e

0 12 EI
3 le

0 6EI
2 le

EA le 0 0

6EI
2 le

4 EI le 0 6EI
2 le

0 12 EI
3 le

EA le 0 0

0 12 EI
3 le 6 EI 2 le

6EI
2 le

2 EI le

OP P 6 EI P l P P 2 EI P l P P 0 P P 6 EI P P l P 4 EI P PP l Q
0
2 e e 2 e e

Let us take element and nodal connectivity as given below:

Element 1 2 3 Node 1 1 2 4 Node 2 2 3 3

For element No. 1 le = 3, EA = 200 109 4 103 = 800 106 N = 800 103 kN EI = EI0 = 200 109 40 106 = 800 103 N m2 = 800 kN m2

LM 333.333 MM 0 0 = 800 M 333.333 MM MM 0 N 0

0 0 0 0 333.333 0.444 0.667 0 0.444 0.667 0.667 1333 0 0.667 0.667 . 0 0 333333 0 0 . 0 0.444 0.667 0.444 0.667 0.667 0.667 0 0.667 1333 .

OP PP PP PP PQ

For this element = 90, l = cos = 0 m = sin = 1

The transformation matrix is

264

Finite Element Analysis

LM 0 MM1 0 L=M MM 0 MM 0 N0

1 0 0 0 0 1

0 0 0

0 0 0 0

0 0 0 1 0 0 1 0 0 0 0 0

OP 0P 0P P 0P 0P P 1P Q
. 333333 0 0 0.444 0 0.667 0.667 0 0.667 . 0 333333 0 0.667 0.444 0 . 0.667 0 1333

k L

LM 0 MM0.444 0.667 = 800 M MM 0 MM00.444 N .667 LM0 MM1 0 =M MM0 MM0 N0

1 0 0 0 0 0
T

333.333 0 0 0.667 . 0 1333 . 333333 0 0.667 0 0 0.667

OP PP PP PP PQ

0 0 1 0 0 0

0 0 0 0 0 0 0 1 1 0 0 0

0 0 0 0 0 1

OP PP PP PP QP
4 5 6 , Global Numbers

k L = L
1

d k L i
3

LM 0.444 MM 0 0.667 = 800 M MM0.444 MM0.0 N 667

0.667 0.444 0.667 0 0 . 333333 333.333 0 0 0 . 0 1333 0.667 0 0.667 0 0.667 0.444 0 0.667 333333 . 0 0 333.333 0 . 0 0.667 0.667 0 1333

OP PP PP PP PQ

1 2 3 4 5 6

For element No. 2, local and global coordinates are the same, and le = 4m, Ie = 2Io

Analysis of Beams and Rigid Frames

265

= k

LM 250 MM 0 0 = 800 M MM250 MM 0 N0

0 0 0 250 0.750 0 0.375 0.750 0.750 2.0 0 10 0.75 . 0 0 250 0 0 0 0.375 0.75 0.375 0.75 0.75 10 0 2.0 0.75 . 0 0.375

OP PP PP PP QP

, Global Numbers

4 5 6 7 8 9

For element No. 3, if we take it as member 43, we get the stiffness matrix identical to element 1. However nodal displacement vector for it is [ 10

11 12 7 8 9 ]. Hence
7 8 9

LM 0.444 MM 0 0.667 = 800 M MM0.444 MM0.0 N 667

10

11

12

0 0 0.667 0.444 0.667 10 11 333.333 0 0 0 333.333 0 1333 0.667 0 0.667 12 . 0 0.667 0.444 0 0.667 7 8 0 0 333333 0 333.333 . 0 0.667 0.667 0 1333 9 .

OP PP PP PP PQ

, Global Numbers

Brief Procedure for Further Analysis

(i) Stiffness matrix of the system [k] is of size 12 12. It can be assembled by placing the elements of the element stiffness matrices [k]1 [k]2, and [k]3 in the appropriate positions of 12 12 matrix. (ii) There are six boundary conditions, namely 1 = 2 = 3 = 10 = 11 = 12 = 0 . They are imposed either by elimination method or by penalty method. When hand calculations are made the elimination method is ideal. In this method, the elements corresponding to rows and columns corresponding to 1, 2, 3, 10, 11 and 12 get eliminated and the stiffness matrix reduces to 6 6 size corresponding to displacement vector { }T = [ 4 5 6 7 8 9 ]. If penalty method is used, which is suitable for computer applications, 12 12 size of stiffness matrix is maintained but diagonal elements corresponding to the rows 1, 2, 3, 10, 11 and 12 are increased by very large numbers. (iii) The equivalent nodal forces due to the applied loads are as shown in Fig. 14.11. Aprat from these loads, there is a 30 kN load in the direction 4. Hence the load vector is

{F}T = [0 0 0 30.0 40 26.667 0 40 26.667 0 0 0]. In case of elimination method of imposing boundary conditions, we get reduced load vector corresponding to the directions 4, 5, 6, 7, 8 and 9 as

lFq

= 30.0 40 26.667 0 40 26.667

266

Finite Element Analysis

b 20g 4 2
2

40

20

GH 4 JK 2

40

4 b 20g 12
2

3 26.667

4 b 20g 12
2

26.667

Fig. 14.11

(iv) The system equations may be solved to get the nodal displacement vector. (v) The member forces may be calculated as usual. The tensile force as positive is given by,

P=

EA 4 1 le

g R | b1 + 3 gl | S | | T
e 2 3 5 6

(14.26)

The bending moment and shear forces are obtained by:

M=

EI
2 le

1 3 le

U | | V | | W

(14.27)

V=

EI
3 le

12 6le

12 6le

R | | S | | T

2 3 5 6

U | | V | | W

(14.28)

(vi) The end reaction vector, +ve in the +ve direction of degrees of freedom is obtained by the expression

lRq = k l q lFq

(14.29)

14.4 A THREE DIMENSIONAL RIGID FRAME ELEMENT

The typical element is shown in Fig. 14.12. In this, there are six degrees of freedom at each node. Hence total degrees of freedom are 12. It may be noted that the rotations are about the axes but not in the directions of axes. Thus,

lq

= 1 2
= u1 v1

3 ... 12
w1 x1 y1 z1 u2 v2 w2 x 2 y 2 z 2

(14.30)

Analysis of Beams and Rigid Frames

ym ym

267

y1

y2

y x z

v1 u1 w1 i
x1

v2 u2 w2 j
x2

xm

z1

z2

zm

zm

Fig. 14.12

The 12 12 stiffness matrix in local coordinate system, may be written down as shown in equation 14.31 by carefully noting the contribution of each nodal displacement.
y ym xm ym y

zm
(b)

(a)

zm

Fig. 14.13

(a) Orientation of beam element (b) Orientation of principle axes about xm-axis

268

Finite Element Analysis

LM EA MM l MM 0 MM 0 MM MM 0 MM 0 MM MM 0 = M EA MM l MM MM 0 MM 0 MM MM 0 MM 0 MM MM 0 N
e e

0 12 EI z
3 le

0 0 12 EI y
3 le

0 0 0 GI x l 0 0 0 0 0 GI x le 0 0

0 0 6 EI y
2 le

0 6 EI z
2 le

EA le 0 0 0 0 0

0 12 EI z
3 le

0 0 12 EI y
3 le

0 0 0 GI x le 0 0 0 0 0 GI x le 0 0

0 0 6 EI y
2 le

0 6 EI z
2 le

0 0 0 6 EI z
2 le

0 0 0 4 EI z le 0 6 EI z
2 le

0 0 0 6EI z l2 0 12 EI y
3 le

0 0 0 2 EI z le 0 6 EI z
2 le

0 6 EI y le2 0 0 0 12 EI y
3 le

0 4 EI y le 0 0 0 6 EI y
2 le

0 6 EI y
2 le

0 2 EI y le 0 0 0 6 EI y
2 le

0 0 0 12 EI y
3 le

0 12 EI z
3 le

EA le 0 0 0 0 0

0 0 0 6 EI z
2 le

0 0 0 2 EI z le

0 0 0 6EI z
2 le

0 0 0 4 EI z le

0 6 EI y le2 0

0 2 EI y le 0

0 6EI y
2 le

0 4 EI y le 0

OP PP PP PP PP PP PP PP PP PP PP PP PP PP PP PP PP PP PQ

(14.31) If the member axes do not coincide with global axes, we need transformation matrix. Figure 14.13 (a) shows arbitrary orientation of the member axis in global system and Fig. 14.13 (b) shows the orientation of principal axes of the member. Then the transformation matrix T works out to be

LM L 0 =M MM 0 NM 0

0 L 0 0

0 0 L 0

0 0 0 L

OP PP PP Q

(14.32)

where L is a 12 12 matrix and each sub matrix on right hand side is a 3 3 matrix. [L ] matrix is given by

LM MM M C C =M MM MM C C MN
x x

Cx y cos C z cos
2 2 Cx + C y y

Cy
2 Cx + Cz2 cos

sin Cz cos
2 Cx

Cz2

2 Cx + Cz2 sin

OP PP C C C cos + C sin P PP C +C P C C sin C cos P PP C +C Q

z y z x 2 x 2 z y z x 2 x 2 z

(14.33)

Analysis of Beams and Rigid Frames In the above expression,

269

Cx =

x j xi le

, Cy =
2

y j yi le
2 j i

and Cz =

z j zi le
2 i

le =

dx

xi

i + d y y i + dz z i
j

For the derivation of above transformation matrix reorders may refer to Krishnamurthy[1]

14.5 TIMOSHENKO BEAM ELEMENT

In the theory used so far, the assumption was made that the plane section before bending remains plane even after bending. The plane section remain plane even after bending, means shear deformations are neglected. In most of the beams bending is associated with the shear. Elementary bending theory shows that shear stress is zero at extreme fibers and is maximum at the centroid of the cross section. These longitudinal stresses cause varying strain in the plane section. As a result of these shear stresses, plane section will not remain plane after bending. If the beam thickness is small, these stresses are small and hence the assumption that plane section is plane even after bending, gives good results; but in thick beams this assumption will not give good results. Timoshenko beam theory recognieses that the action of the shear force causes a shear strain. This causes warping of the beam element as shown in Fig. 14.14. The shear stress in general can be express as
b

Fig. 14.14

Warped beam element

xz = G xz
and the shear force as

(14.34)

Q = xz dz b
where xz is the shear strain at distance z from neutral axis at section x.

(14.35)

However the use of general equations 14.34 and 14.35 complicates the problem and a simplified approach is possible. In order to account for non-uniform stress distribution at a cross section while still retaining one dimensional approach, the equations 14.34 and 14.35 are modified using a shear correction factor as follows:

270

Finite Element Analysis

xz = G x
and
Q = xz A = AG x

(14.36) (14.37)

The shear correction factor ' ' is a function of the cross sectional shape and Poisson ratio . The term ' A' is the shear area of the section associated with shear and may be denoted as As. Thus
As = A where < 1

(14.38)

Values of for various cross sectional shapes are given in the solid mechanics books by DYM C.L. & Shames I.H. [2] and Ugural A.C. and Fenster S.K. [3]. The value of for a rectangular section is 5/6. Hence,

Q G As

(14.39)

Timoshenko beam theory averages the effect of shear strain over the cross section i.e. it takes,

w x

(14.40)

where is the angle through which the face of the cross section rotates after deformation as shown in Fig. 14.15.

w x

Before deformation

After deformation

Fig. 14.15

Now, in a beam element

M = EI

d dx

Strain energy due to flexure = =

M2 EI dx = 2 EI 2

z
le 0

F d I G dx J H K

dx and strain energy due to shear

1 shear stress shear strain dV 2

Analysis of Beams and Rigid Frames

271

z z

1 xy 2

dV
1 dV = G As 2

1 G 2 2

z
le 0

F w I G x J H K F w I G x J H K

dx

Total strain energy of the element is 1 U = EI 2

z
le 0

F d I G dx J H K

1 dx + GAs 2

z
le 0

dx

(14.41)

Finite Element Formation

Since according to Timoshenko theory =
dw , and w are decoupled i.e. they are independent of each dx other, at every node there are two independent displacement components and w. In a two noded beam element they vary linearly. Thus, w = N1 w1 + N2 w2

and In matrix form,

= N1 1 + N 2 2

RwU = LM N S V N 0 TW

0 N1

N2 0

Rw U 0 O | | S V PQ |w | N | | | | T W
1 1 2 2 2

(14.42)

where N1 and N2 are interpolation functions i.e. N1 = 1 Using isoparametric concept, we have
x=

x x and N2 = le le

N x
i

(14.43)

Now,

d = dx

i=1

Ni i x
w = x

(14.44)

Ni i

i =1

i =1

Ni wi x

(14.45)

Strain vector

lq = R U S V TW

272

Finite Element Analysis

where is curvature =

d dx dN1 dx N1 0 dN 2 dx dN2 dx x

L R U = MM 0 l q = S V T W MM dN N dx

OP PP = l1 LM0 1 PQ N
e

le x 1

OP xQ

The stress resultants M and Q are related to strain as

=
Now M = EI For rectangular section I =
D11 =

RM U = D lq SQ V T W

1 bh 3 12

E 3 bh 12

and

Q = As

xy = As

h = A

LE RM U = MM12 bh l q = S V T Q W MM 0 N
D = Ebh h 2 12 0

b g O 0 P P R U = Ebh LMMh0 S V 12 N bh P T W EP 2 Q

E bh E = 21+ 2

0 6

OP R U TW QP S V
(14.48)

Assuming = 0, D matrix for rectangular section is

= e

LM 0 1 OP 1 M1 l x P Ebh Lh = MNM 0 60 OPQP LMN0 MM 0 1 PP 12 l 1 N1 x Q LM 0 1 OP h Ebh M1 l x P L 0 = MM 0 1 PP MMN6 6 bl xg 12 l N1 x Q

2 e 0

z
le 0

LM NM

OP 6 Q P
0

D B dx

z
le

1 0 1 dx le x 1 x

OP Q

2 e 0

z
le

0 6

h2 dx 6x

OP PQ

Analysis of Beams and Rigid Frames

273

Ebh

2 12 le 0

z
le

LM 6 MM6 bl xg 6 MM 6 x N
e

OP 6 bl x g h + 6xbl x gP h + 6 bl x g PP dx 6 bl x g 6x 6 h + 6x h + 6xbl x g 6x PQ
6 le x
e 2

6
e

6x

Separating bending and shear terms, we can write

LM Ebh M = 12 l M MM N
2 e

z
le 0

LM0 MM0 MN0 0

se

0 h 0 h 2
2

1 le x 0 0 2 2 le x le x 0 h + 6 1 le x 0 0 2 x x le x 0 h

OP PP PQ

LM MM MM N

OPOP b g bl xg xbl xgPP dx x PP b g 1 PPPP x b g x QQ

1 x
e e 2

i.e.,

be

(14.49)

where [ ] be and [ ] se are the contributions of bending and shear to the total stiffness. The integrations can be performed to get [ ] be and [ ] se as shown below:

se

se

le very large), h the results obtained by this formulation are not correct. The shear term, which should tend to zero in such cases, do not tend to zero. This is called shear locking. The shear stiffness is increasingly constrained. This is called spurious constraint. There are two popular remedies for the elimination of this type of errors: (i) Reduced Integration Technique (ii) Using field consistency element
The above formulation gives good results for moderately thick beams. For thin beams ( (i) Reduced Integration Technique: If shear stiffness in equation 14.49 is integrated with one point Gaussian technique, we get

LM0 0 0 Ebh M0 1 0 = 12l M0 0 0 MN0 1 0 LM 1 l MM l 2 l Gh M 2 3 M = l M 1 l MM 2 l l MM N2 6

3 e e e 2 e e e e 2 e

0 1 0 1

OP PP PQ
le 2 le 2 2 le 6 l e 2 2 le 3

(14.50)

1 le 2

OP PP PP PP PP PQ

(14.51)

274

Finite Element Analysis

se

LM 1 MM l Gh M 2 M = l M 1 MM MM l N2
e e e

le 2 2 le 4 l e 2 2 le 4

1 le 2 1 le 2

le 2 2 le 4 l e 2 2 le 4

OP PP PP PP PP PQ

(14.52)

since and

x=

f d = 2 f = 0

bg

1 1+ , dx = d 2 2

In case of two point integration sampling points are at Hence

1 3

and weight function Wi = 1.

2 se

LM 1 MM l Gh M 2 M = l M 1 MM MM l N2
e e e

le 2 2 le 3 le 2 2 le 6

1 le 2

1 le 2

le 2 2 le 6 l e 2 2 le 3

OP PP PP PP PP PQ

(14.53)

Now consider the analysis of a cantilever beam with single element (refer Fig. 14.16). The stiffness matrix of the beam is

k = k

= k

be

+ k

se

P M b=1

Fig. 14.16

Analysis of Beams and Rigid Frames

275

Applying the boundary conditions that w 1 = 0, 1 = 0 and defining =

Eh 3 Eh 3 and = 12 le 12 l

Gh Gh = , we get the equilibrium equation from one point quadrature as, le l

LM MM MM 2l MM MM MM 2l N
i.e.,

l 2

l 2

l2 4

l 2 l2 4

l 2

OP R 0 U R 0 U P| | | | | | | | l P | | | | 0 + P 0 4 P | | | | l PS V=S V P |w | | P | 2 PP | | | | | | | | l + P | | | M | | | | | 4 QP T W T W
l 2
2 2 2 2

LM MM l MN 2
Solving the above equation, we get

OP PP R U = R M U Sw V S P V l + P T W T W 4 Q
l 2
2 2 2

w2 =

F l + 1I P + 1 M G 4 J 2 H K
2

(14.54)

and

l P+ M 2 = 2
In case of thin beam, . Hence equation 14.54 reduces to
w2 = 1 lP + M 2 2

(14.55)

and 2 remains same as equation 14.55. Thus the beam deformation is solely due to bending as shown by equation 14.55 and 14.56. Hence for thin beams one point integration gives correct results. If two point quadrature is used for integration, after applying the boundary conditions we get the equations as

LM N

OP Q

(14.56)

LM MM l MN 2

OP PP R U = R M U Sw V S P V l + P T W T W 3 Q
l 2
3 2 2

276

Finite Element Analysis Solving for w2 and 2 , we get

l2 lM 3 w2 = P+ 2 l l2 2 + + 12 12

F G H

I J K

F G H

I J K

and

l P 2 2 = l2 + 12 M +
In case of thin beams . Hence we get

w2 =

4P + 6

M l

(14.57)

and

2 =

6 l P + 2M l
2

(14.58)

The above two equations show that the free end deformation depends on the coefficient correspongind to shear deformation, which is not true in case of thin beams. Hence two point integration (equations 14.57 and 14.58) lead to erroneous results. Thus the reduced integration (1 point Gaussian integration instead of 2 point Gaussian integration) gives better results than the exact integration. The reduced integration technique is used to get good results for thin beams. (ii) Field Consistant Element Formulation: Let us first see how the element formulation is field inconsistent in the case of very thin beams. In the formulation we have taken,

w = N1
and

N2 N2

= N1
N1 =

Rw U = N w + N w Sw V T W R U = N + N S V T W
1 1 1 2 2 1 2 1 1 2

where

1 1+ and N 2 = 2 2

and rearranging the terms, we get

2x 1 le

w=

1 1+ w + w2 w w1 + 2 = a1 + a2 w1 + w2 = 1 2 2 2 2

Analysis of Beams and Rigid Frames

277

where similarly where Now bending strain

a1 =

w1 + w2 w2 w1 and a2 = 2 2

= b1 + b2
b1 =

1 + 2 1 and b2 = 2 2 2

=
and shear strain

d d d 2 = = b2 dx d dx l

dw d dw 2 = b1 + b2 = b1 + b2 a2 dx d dx l

Strain energy due to bending is given by,

Ub =
and strain energy due to shear is
Us =

EI 2

F d I G dx J H K
2

dx =

EI 2

F2b I Gl J H K
2 2

dx

(14.59)

GAs 2

bg

dx =

GAs 2

F b 2 a + b I G l J H K
1 2

dx

(14.60)

As the thickness approaches zero, the shear strain energy should vanish and bending strain energy should exist. i.e.,

b1

2 a2 + b2 0 l 2 a2 0 l
(14.61) (14.62)

i.e., and

b1

b2 0

The terms corresponding to condition 14.61 correspond to both the strain fields flexure and shear. Hence it is called field consistant term. The constraint corresponding to equation 14.62 contains the term corresponding only to flexure field. If b2 0 , the strain energy due to bending tends to zero, which should not happen. Hence this constraint requirement in the limiting case is spurious and it is this requirement which causes shear locking. To get rid of this situation in the limiting case, the function smoothening is required i.e. alter the terms to over come this situation. This is achieved by making b2 = 0 in the shear strain field i.e. by taking

= b1 =
Thus,

1 + 2 in the shear field 1

= b1 =

1 + 2 1 = 1 2

LM N

1 2

OP R U = Q S V T W
1 2

N1

N2

R U S V T W
1 2

278

Finite Element Analysis

i.e. in shear field interpolation function for will be taken as N1

LM 1 1 OP . This smoothening is N2 2 Q required not only in the shear strain field but even in the shear stress field also.
N2 =

The greatest advantage of Timoshenko beam element is C1 continuity problem is reduced to Co continuity problem. The extension of this concept by Mindlin to plate bending problems, (to be discussed in next chapter) is a great achievement in the plate analysis.

QUESTIONS
1. Derive the stiffness matrix for a beam element. 2. Analyse the beam shown in Fig. 14.17 using FEM technique. Determine the rotations at the supports. Given E = 200GPa and I = 4 106 mm4.
20 kN/m

A
6m

B
4m

Fig. 14.17

(Ans. A = 0.3, B = 0.375; C = 01625 ) . 3. Assemble the stiffness matrix for a plane beam element oriented at angle to the x-axis. Explain its use in FEA. 4. Assemble element stiffness matrix for the member of plane frame shown in Fig. 14.18, if it is oriented at angle 30o to the x-axis. Take E = 200GPa, I = 4 10-6 m4 and A = 4 103 m2.
y

5m

30

Fig. 14.18

LM150.025 MM M (Ans.) M MM NM

. . 86.558 012 150.015 086.558 012 50.070 0.208 86.558 50.070 0.208 . 0.208 0.8 012 0.4 . SYM 150.015 86.558 012 50.070 0.208 0.8

OP PP PP PP QP

Analysis of Beams and Rigid Frames

279

5. Explain Timoshenko Beam Theory. Discuss its advantages and disadvantages. 6. Explain the term Timoshenko Beam Theory and briefly explain the stiffness formulation for such element. 7. Explain the term shear locking as used in Timoshenko Beam Theory. How this problem is overcome? 8. Taking the example of a single element cantilever beam, show that single point Gaussian integration overcomes the problem of shear locking. 9. Explain what is meant by field consistant formulation to overcome shear locking in Timoshenko Beam Theory. Use the example of a single element cantilever beam.

References
1. Krishnamoorthy C.S., Finite Element AnalysisTheory and Programming, Tata McGraw-Hill Publishing Co. Ltd., New Delhi. 2. DYM C.L. and Shames I.H., Solid Mechanics, International Students Edition, McGraw-Hill Kogakusha Ltd., 1973. 3. Ugural A.C. and Fenster S.K., Advanced Strength and Applied Elasticity, American Elsevier Publishing Co., New York, 1975.

280

Finite Element Analysis

15
Bending of Thin Plates
15.1 INTRODUCTION
Plate is a flat surface having considerably large dimensions as compared to its thickness. Slabs in civil engineering structures, bearing plates under columns, many parts of mechanical components are the common examples of plates. In this chapter, we are considering bending of such plates under lateral loads. The bending properties of a plate depend greatly on its thickness. Hence in classical theory we have the following groups (i) thin plates with small deflections (ii) thin plates with large deflections, and (iii) thick plates In thin plates with small deflections theory, the following assumption are made (a) There is no deformation in the middle plane of the plate. This plane remains neutral during bending. (b) Points of the plate lying initially on a normal to the middle surface of the plate remain on the normal to the same surface even after bending. (c) The normal stresses in the direction transverse to the plate are negligible. This theory is satisfactory for plates with ratio of thickness to span exceeding deflection to thickness less then
1 and the ratio of maximum 10

1 . Many engineering problems lie in the above category and satisfactory 5 results are obtained by classical theories of thin plates. Stresses in the middle plane are negligible, if the deflections are small in comparison with thickness. If the deflections are large, the in plane stresses developed in the so called neutral surfaces are to be considered. This gives rise to theory of thin plates with large deflections, in which geometric non-linearity is incorporated.

1 th the assumption (a) and (b) listed under theory of 10 thin plates will not hold good. Such plates need three dimensional analysis. Theory developed for the analysis of such plates may be called as thick plate theory. In this chapter, analysis of thin plates with small deflections with finite element approach is explained. Since many stress-strain-displacement relation in the theory of thin plates with small deflection are used, the limitations of classical theory remains in this finite element analysis also. The advantage of finite element analysis is that it can handle the structures with different end conditions and shapes easily.
If the plate has thickness to span ratio less than