0 Up votes0 Down votes

259 views44 pagesDerivation of the shell element, Ahmed element, Midlin element in the finite element analysis, Hani Aziz Ameen

Aug 06, 2011

© Attribution Non-Commercial (BY-NC)

PDF, TXT or read online from Scribd

Derivation of the shell element, Ahmed element, Midlin element in the finite element analysis, Hani Aziz Ameen

Attribution Non-Commercial (BY-NC)

259 views

Derivation of the shell element, Ahmed element, Midlin element in the finite element analysis, Hani Aziz Ameen

Attribution Non-Commercial (BY-NC)

- Mae160_chap14
- A Handbook 0f Building Structure = Allan Hodgkinson
- The finite element method
- Finite Element Method
- 7207839 Ansys Structural Analysis Guide
- Automobile Engg-UNIT-2
- NISTNCSTAR1Draft
- Rapid Prototyping
- Mae160_chap13
- Nonlinear Micro Finite Element Analysis of Human Trabecular Bone 2003
- Chapter 9
- Chapter 2
- 17192061-Phd-Thesis-1997-D
- MMM351 - Machine Dynamics
- Chapter 3
- Chapter 4
- Chapter 5
- Chapter 6
- Chapter 8
- Nonlinear Solid Mechanics

You are on page 1of 44

Midlin Element in Finite Element Analysis

Ahmed Element , Midlin Element

in Finite Element Analysis

Technical College - Baghdad

Dies and Tools Eng. Dept.

E-mail:haniazizameen@yahoo.com

www.mediafire.com/haniazizameen

1- Definition of a Shell

Shell is defined as an object which, for the purpose of stress

analysis may be considered as the materialization of a curved surface [1].

This definition implies that the thickness of a shell must be small

compared with its other dimensions, but it does not require the smallness

be extreme. Most shells, of course, are made of a solid material, and

generally, it will be assumed that the material is isotropic and elastic.

1

Derivation of the shell element , Ahmed Element , Dr. Hani Aziz Ameen

Midlin Element in Finite Element Analysis

In most cases, a shell is bounded by two curved surfaces, the faces. The

thickness "ts" of the shell may be assumed the same everywhere or it may

vary from point to point. The middle surface of a shell is defined as the

surface which passes midway between the two faces. If the shape of the

middle surface and thickness are known, then the shell is geometrically

fully described.

The analysis of shells with an arbitrarily defined shape presents an

intractable analytical problem. If, in addition, the shell is a thick one in

which the shear deformation is significant, the applicability of a classical

approach becomes a question. In many engineering structures, these

difficulties might be overcome, if satisfactory (and hopefully optimized)

designs are ever to be achieved. Over the years, much has been written on

the various attempts to produce efficient, accurate, and reliable shell

elements in (FEM). Three distinct classes of shell elements have emerged

[2]:

1. Flat, plate-like elements which are sometimes called facet elements

because they approximate the curved shell by a faceted surface.

2. Curved shell elements founded on some shell theory.

3. Degenerated shell elements based on the three-dimensional continuum

theory.

In the first approach, the shell is replaced by an assemblage of flat plate

elements which are either triangular or quadrilateral in shape, as the

triangle shown in figures (2) (a) and (b) [3]. Each plate element is

connected in some fashion to those surrounding it and undergoes both in-

plane (membrane) and bending (flexural) deformations. The method has

2

Derivation of the shell element , Ahmed Element , Dr. Hani Aziz Ameen

Midlin Element in Finite Element Analysis

with each element. The coupling only appears indirectly through the

degrees of freedom at the nodal points linking the adjacent elements.

Consequently, a large number of elements must be used to achieve

satisfactory accuracy. Although of the certain shortcomings in the

approach, facet elements are very efficient for the approximate analysis

of many shell structures.

The second type includes curved shell elements based upon thin shell

theories of classical mechanics, consisting of the analysis of deep or

shallow shells. A commonly used theory of deep shells is based upon the

strain-displacement relationships of Novozhilov [4]. On the other hand,

specialized theories of shallow shells follow the simplified strain-

displacement relationships of Vlasov [5]. The later method is more

approximate than the former, but accurate results have been obtained,

even when shallow-shell concepts were applied to deep shells, Cowper

et.al. [6]. Figure (3) shows the geometry for an arbitrary shallow shell

element. Although the above type of shell elements are quite popular, but

they also suffer from various limitations associated with the lack of

consistency in many shell theories and also with the difficulty in finding

appropriate deformation idealization which allows truly strain-free rigid

body movement.

Finally, the curved elements (third class) for shell analysis can be

devised by specializing three dimensional solid elements to be thin in one

direction while introducing constraint conditions on nodal displacements.

As examples, the hexahedron and the pentahedron in figures (4) (a) and

(b) can be specialized to become quadrilateral and triangular shell

elements that are curved in three dimensional spaces. The characteristics

and analysis of this type is illustrated in the following section.

3

Derivation of the shell element , Ahmed Element , Dr. Hani Aziz Ameen

Midlin Element in Finite Element Analysis

z y

3

qi 2

i

1 u q i1

3

2

(a)

z y

3

qi 3

w qi 5

1 i

qi 4

3

2

(b)

x

Figure (2) Flat-facet element: (a) Membrane components (b) Flexural

3

components.

v

w

u

2

1

( , )

z

y

3

1

c

b

a

2

x

Figure (3) Shallow shell element geometry and coordinate system [6].

4

Derivation of the shell element , Ahmed Element , Dr. Hani Aziz Ameen

Midlin Element in Finite Element Analysis

Among all of the shell elements, the Ahmad [7] type

"degenerated" isoparametric shell element based on an independent

translational and rotational displacement interpolation, has become the

most popular in shell analyzation. In this element, the Mindlin [8] theory

is employed, where the "normal" to the middle surface of the element is

constrained to remain straight (but no longer normal) after deformation in

order to overcome the numerical difficulty associated with a large

stiffness ratio through the thickness direction. The strain energy

associated with the stress perpendicular to the middle surface is also

neglected. By adopting the isoperimetric geometric description, the

element can be used to represent thin and thick shell components with

arbitrary shapes.

Figure (5) (a), shows the original isoperimetric hexahedron element

which has a quadratic formula defining its geometry, where ui, vi, and wi

are translations in the global coordinates as demonstrated by Weaver and

Johnston [9]. In order to convert this hexahedron to a thin curved

quadrilateral element for the analysis of shell, one can first form a flat

rectangular solid by making the curvilinear coordinate's ξ, η, and ζ

orthogonal and the ζ dimension is small. The resulting element appears in

figure (5) (b) is the rectangular parent of shell element before constraints.

Note that groups of three nodes occur at the corners, while pairs of nodes

are at the mid-edge locations of the element. By invoking the former

constraints, each group and pair of nodes could be converted to a single

node on the middle surface, as shown in figure (5) (c), where αi and βi are

small rotations about two local tangential axes. The Relationships

between the nodal displacement at a corner and mid-edge with a node of

5

Derivation of the shell element , Ahmed Element , Dr. Hani Aziz Ameen

Midlin Element in Finite Element Analysis

shell elements can be seen more clearly in figures (6) (a), (b) and (c). So,

the nine nodal, translations in figure (6) (a), can

(b)

(a)

(a) Hexahedron (b) Pentahedron.

z

wi

Y 5

i x 17 13 22

vi 1

6 8

ui 9 12

,w 14 18

16

19

2 7

4

2a 2b

12 15 11

3

, v (a)

wi

ts i

,u (b) j

vi

i

ui

k i (c)

(a) Isoparametric hexahedron (b) Rectangular parent as shell element

before constraints(c) Constrained nodal displacements.

6

Derivation of the shell element , Ahmed Element , Dr. Hani Aziz Ameen

Midlin Element in Finite Element Analysis

following 9×5 constraint matrix:

1 0 0 0 0

0 1 0 0 0

0 0 1 0 0

ts

1 0 0 0

2

ts

Gai 0 1 0

2

0

………………………………………. (1)

0 0 1 0 0

ts

1 0 0 0

2

ts

0 1 0 0

2

0 0 1 0 0

Similarly, the six nodal translations in figure (6) (b) are related to

the five nodal displacements in figure (6) (c) by the 6×5 constraint

matrix.

ts

1 0 0 0

2

ts

0 1 0 0

2

Gbi 0 0 1 0 0 …………………………………………. (2)

ts

1 0 0 0

2

ts

0 1 0 0

2

0 0 1 0 0

number of nodal displacements could be reduced from 49 46 60 to

85 40 . In the following articles, the direct formulations of shell

element in the manner described by Cook [10] will be pursued.

7

Derivation of the shell element , Ahmed Element , Dr. Hani Aziz Ameen

Midlin Element in Finite Element Analysis

6

j

z 4 5

y ts / 2

3

x

i

2

1

ts / 2

k 8

7

9

(a)

3

j 2

k

5

4

6

(b)

1 5

i 3

2 4

k (c)

Mid-edge of rectangular element (c) Node of shell element, [9].

8

Derivation of the shell element , Ahmed Element , Dr. Hani Aziz Ameen

Midlin Element in Finite Element Analysis

Figure (7) shows the geometric layout of the shell element, in

which the global coordinates of any point take the form,

x 8 xi 8 3i

ts

y N i yi N i m3i ………………………………………. (3)

z i 1 z i 1 2

i n3i

illustrated and explained in appendix (A1).

In addition, the terms 3i , m3i , n3i are the direction cosines of vector V3i that

is normal to the middle surface and spans the thickness "t s" of the shell at

node i . Figure (7) (b) shows this vector which is obtained as:

x j x k 3i

V3i y j y k m3i t s ……..……………………………………..... (4)

z z n

j k 3i

Point j and k in the figure are at the upper and lower surfaces of the shell,

respectively. In a computer program, the direction cosines for V3i must be

given as data.

Generic displacements at any point in the shell elements are taken

to be in the directions of global axes. Thus, the generic displacements

vector is:

u

u v …………………………………………………………….. (5)

w

(in global directions) as well as two small rotations i and i about two

local tangential axes x and y , as indicated in figure (7).

Hence,

9

Derivation of the shell element , Ahmed Element , Dr. Hani Aziz Ameen

Midlin Element in Finite Element Analysis

4

8

7

5

6

z, w

2

ts

z ,

(a) wi y

y, v i

vi

x, u i

z , ui i

i x

i

v

j v 3i

y

u

i

t v 2i

i

2

v1i

i

ts

x

k

(b)

Figure (7) (a) Shell element (b) Nodal vectors, [9].

12

Derivation of the shell element , Ahmed Element , Dr. Hani Aziz Ameen

Midlin Element in Finite Element Analysis

ui

v

i

i wi …………………………………………………………. (6)

i

i

i 1,2,3,.......,8

terms of nodal displacements are:

u 8 ui 8

i

v N i vi N i i

ts

……………………………….... (7)

w i 1 w i 1 2 i

i

2i 1i

i m2i m1i ……………………………………………………... (8)

n2i n1i

of the second tangential vector V2i , and column 2 has the direction cosines

for the first tangential vector V1i (see figure (7)). These vectors are

orthogonal to the vector V3i and to each other. As infinity of vector

directions, normal to a given direction can be generated, a particular

scheme has been devised to ensure a unique definition. This is given in

appendix (A2) [11].

Figure (7) shows the local generic translations u and v (in the

directions of V1i and V2i ) due to the nodal rotations i and i , respectively.

Their values are:

ts ts

u i & v i …………………………….………. (9)

2 2

11

Derivation of the shell element , Ahmed Element , Dr. Hani Aziz Ameen

Midlin Element in Finite Element Analysis

given by the second summation in equation (7).

The displacement shape functions in equation (7) may be cast into

the matrix form:

ts ts

1 0 0 2 2i

2

1i

N i 0 1 0 t s m2i ts

m1i N i .................................... (10)

2 2

0 0 1 t s n n1i

ts

2

2i

2

i 1,2,......8

1 0 0 0 0

N Ai 0 1 0 0 0 Ni ………………….………………………… (11)

0 0 1 0 0

And,

0 0 0 2 i 1i

N Bi 0 0 0 m2i m1i s N i ……………………..…………...… (12)

t

2

0 0 0 n2i n1i

Then,

Ni N Ai N Bi ………………………………………………...… (13)

And, the shape function matrix becomes:

N N A NB ……………………………………………….…… (14)

The last of these formulas will later be used to drive the consistent mass

matrix.

The 3×3 Jacobian matrix required for this element is given by:

12

Derivation of the shell element , Ahmed Element , Dr. Hani Aziz Ameen

Midlin Element in Finite Element Analysis

x y z

J x y z

……………………………………………...…. (15)

x y z

the Jacobian matrix are found as follows:

x 8

N 8

N t

i x i i s 3i

i 1 i 1 2

x 8

N 8

N t

i x i i s 3i

i 1 i 1 2

x 8

t

N i s l 3i and so on.

i 1 2

x x x

J 1 J *

……………………………………...….. (16)

y y y

z z z

13

Derivation of the shell element , Ahmed Element , Dr. Hani Aziz Ameen

Midlin Element in Finite Element Analysis

5- Strain Calculations

Certain derivatives of the generic displacements (equation (7)) with

respect to local coordinates are needed. These derivatives are listed in a

column vector of nine terms as follows:

u

N i N i N i

u 0 0

2i

1i

u N i 0 0

N i

2i

N i

1i

0 0 0 N i 2i N i 1i u i

v

0 N i N i N i

0 m2 i m1i vi

v 8 w

t i ...........................17

0 N i

N i

N i

i 1

0

m2 i

m1i s

2

i

v N i m2 i

0 0 0 N i m1i t s

w N i N i N i

0 0 n2i i n1i 2

0 N i N i

n1i

Ni

w 0 n2i

0 N i n2i N i n1i

0 0

w

Transformation of these derivatives to global coordinates requires that the

inverse of the Jacobian matrix be applied. Therefore,

u u

x

u J 0 0 u

*

y 0 J * 0 ………………………………….…….. (18)

0 0 J * .....

w w

z

Multiplying the terms in this equation yields,

14

Derivation of the shell element , Ahmed Element , Dr. Hani Aziz Ameen

Midlin Element in Finite Element Analysis

u

x

u

y ai 0 0 d i l 2i d i l1i

u b

z i 0 0 ei l 2i ei l1i

v ci 0 0 g i l 2i g i l1i ui

x

8

0 ai 0 d i m2 i d i m1i vi

v

0 bi 0 ei m2i ei m1i wi ......................................................19

y i 1 0 g i m2 i

g i m1i i

v ci 0

z 0 0 ai d i n2 i d i n1i i

w

0 0 bi ei n2i ei n1i

x 0 g i n2 i g i n1i

w 0 ci

y

w

z

In which,

Ni * N i

ai J11

*

J12

N i * N i

bi J 21

*

J 22

N i * N i

ci J 31

*

J 32 …………………………………………. (20)

di

ts

2

ai J 13

*

Ni

ei

ts

2

bi J 23

*

Ni

gi

ts

2

ci J 33

*

Ni

The strain displacement vector may be written as:

15

Derivation of the shell element , Ahmed Element , Dr. Hani Aziz Ameen

Midlin Element in Finite Element Analysis

u

x

v

x

y

y w

z u z v ………………………………………………. (21)

xy

yz y x

v w

xz

z y

w u

x z

B …………………………………………………………...… (22)

Where, is the nodal displacement vector (equation (6)), and B is the

strain-displacement matrix. The ith part of matrix B may be written as:

ai 0 0 di 2i d i 1i

0 bi 0 ei m2i ei m1i

0 g i n2i

Bi

0 ci

ei 2i d i m2i

g i n1i

ei 1i d i m1i

………………………. (23)

bi ai 0

0 ci bi g i m2i ei n2i g i m1i ei n1i

ci 0 ai d i n2i g i 2i d i n1i g i 1i

could be isolated to get,

Sub-matrices BAi and BBi are composed from equations (20) and (23),

but the actual details are omitted. Altogether, one can have

which will be convenient when determining the stiffness matrix for the

shell element.

16

Derivation of the shell element , Ahmed Element , Dr. Hani Aziz Ameen

Midlin Element in Finite Element Analysis

6- Stress Calculations

material take the form:

Where D is the stress-strain matrix or the elasticity matrix in local axes

[10] or,

1 0 0 0 0

x 1 0 0 0 0 x

y 0 0 0 0 0 0 y

z 1

0 0 0 0 0 z …………………… (27)

xy 1 xy

2

2

1

yz 0 0 0 0 0 yz

2k

zx 1 zx

0 0 0 0 0

2k

The factor k included in the last two shear terms is taken as 1.2, and its

purpose is to improve the shear displacement approximation [12].

matrix D by using the 6×6 strain transformation matrix T .

Thus,

Where,

17

Derivation of the shell element , Ahmed Element , Dr. Hani Aziz Ameen

Midlin Element in Finite Element Analysis

2

2 m22 n22 2 m2 m2 n 2 n2 2

23 m32 n32 3 m3 m3 n3 n3 3

T ….(29)

2 1 2 2m1m2 2n1n2 1m2 2 m1 m1n2 m2 n1 n 1 2 n2 1

2 2m2 m3 2n2 n3 2 m3 3 m2 m2 n3 m3 n2 n 2 3 n3 2

2 3

2 3 1 2m3 m1 2n3 n1 3 m1 1m3 m3 n1 m1n3 n 3 1 n1 3

respect to the global axes. The concept of coordinate transformation is

illustrated in appendix (A4) [9]. To evaluate the matrix T at an

integration point, the directional cosines for vector V1,V2 ,V3 must be found

at that point. This may be done with the following sequence of

calculations:

In these expressions, the vector J1 norm. denotes the first row of the

Jacobian matrix normalized to a unit length, and so on.

Equation (28) would be more efficient if the third row and column of

matrix D (corresponding to z and z ) were deleted, along with the

third row of matrix T .

strain energy U. Applying the principle of variational approach [13], the

strain energy for the element may be written as:

U

1

2 V

eT e dV …………………………………………...…….. (30)

18

Derivation of the shell element , Ahmed Element , Dr. Hani Aziz Ameen

Midlin Element in Finite Element Analysis

1

U eT B DBdV e …………………………………………. (31)

T

2 V

Or

U eT K e e …………………………………………………….. (32)

1

2

V

equations (25) and (34) into equation (33) yields,

1 1 1

K e Ba Bb T DBa Bb J ddd …………………… (35)

1 1 1

the first power appears in certain terms. Zienkiewic [14] suggested that

these terms might be neglected in comparison with terms to which they

were added for the purpose of simplifying integration through the

thickness. Thus, the matrix [J] becomes independent of , and the explicit

integration through the thickness is possible. The products Ba T DBb

19

Derivation of the shell element , Ahmed Element , Dr. Hani Aziz Ameen

Midlin Element in Finite Element Analysis

through the thickness. While, the products Ba T DBa and 2 Bb DBb

may be integrated with respect to at once. Thus, equation (35) is

reduced to

1 1

1 1 3

Hence, the first part of matrix K e in equation (36) is due to the transverse

shearing deformations, whereas the second part is associated with flexural

deformations. To evaluate the integrals in equation (36) numerically,

Gauss-quadrature technique [15] is adopted using two integration points

in each of and coordinates. This method is explained in appendix

(A5).

8- Consistent Mass Matrix [M]e for Shell Element

energy KE as follows:

For any body of infinitesimal mass dm and velocity vector qe , the kinetic

energy is:

1 T

2

KE qe qe dm ………………………………………………....….. (37)

1

KE

2 V

qeT qe dV ……………………………………………..…… (39)

22

Derivation of the shell element , Ahmed Element , Dr. Hani Aziz Ameen

Midlin Element in Finite Element Analysis

d

Where, e e ………………………………...……………...….. (41)

dt

1

KE eT N N dV e ……………………………………...…. (42)

T

2 V

Or,

1

2

V

1 1 1

M e N A N B T N A N B J ddd ……………... (45)

1 1 1

1 1

1 1 3

and the second part gives a rotational (or rotary) inertia. The integrals in

equation (46) are evaluated in the same manner as in integrals equation

(36).

21

Derivation of the shell element , Ahmed Element , Dr. Hani Aziz Ameen

Midlin Element in Finite Element Analysis

The equation of motion (or dynamic equation) can be derived, using the

energy balance principle which involves that "the summation of the

structure energies is stationary", i.e., the summation of kinetic energy,

dissipation energy, strain energy and potential energy is stationary, or

then,

KE DE U PE 0 ………………………….………………. (48)

The first and third terms of equation (47) are obtained by equations (43)

and (32), respectively. Now, the second and the fourth terms will be

created. The dissipation energy DE depends upon the nature of damping,

and for the case of viscous damping, a damping matrix ce can be defined

such that:

1

2

Finally, the potential energy PE (with the absence of body forces) can be

written as:

PE W eT Fe t ………………………………………………… (50)

Substituting equations (32), (43), (49), and (50) in equation (48) gives,

22

Derivation of the shell element , Ahmed Element , Dr. Hani Aziz Ameen

Midlin Element in Finite Element Analysis

1 T

e M e e e C e e e K e e e Fe (t ) =0……………. (51)

1 T 1 T T 1 T

e 2 2 2 2

follows:

1 T d e 1 T

e M e e

e 2

d

e M e e

dt e e 2 dt

M e e M e e

The other terms can be easily derived to get the final form of the dynamic

equation of finite element.

23

Derivation of the shell element , Ahmed Element , Dr. Hani Aziz Ameen

Midlin Element in Finite Element Analysis

Appendices

Appendix (A1): The Shape Functions

Shell Element Shape Functions.

In the finite element analysis, the region of interest is subdivided

into a number of sub-regions known as elements, which are defined by

the locations of their nodal points. The main concept here is that the

geometry of the element is defined using the nodal coordinates and the

shape functions, which are used to interpolate the main unknowns (i.e.,

displacement) with an isoparametric formulations in terms of a non-

dimensional element coordinates , , which varies from -1 to +1 over

the element called natural coordinates. This coordinate system is

particularly useful when the adoption of numerical integrations is

considered to evaluate any integrals which are required during the

stiffness matrix calculations for example. Figure (A1.1) shows the

rectangular parent element (a) of the isoparametric quadrilateral element

(b) which is geometrically similar to the shell element used. Since 8-node

elements have been employed, and according to Pascal's triangle, the 8

terms polynomials are assumed for the displacement function as follows

u c1 c2 c3 c4 2 c5 c6 2 c7 2 c8 2 …………………… (A.1.1)

Several methods could be used in obtaining the displacement shape

function. Hence, a direct substituting method will be used by applying the

above equation to each node in the element. Thus,

u1 c1 c21 c31 c412 c511 c612 c7121 c8112

And so on, substituting the values of i ,i (where i the node number,

i 1,2,…….8) which are listed in table (A.1.1) into the above equations

24

Derivation of the shell element , Ahmed Element , Dr. Hani Aziz Ameen

Midlin Element in Finite Element Analysis

can be calculated. Substituting these constants into equation (A.1.1), the

displacement shape functions are obtained as follows:

1

7

4 7 3

1 1

8 6

y 7

x 1 5 2

1

(a)

3

4 7

6

8

y

2

x 1 5

(b)

25

Derivation of the shell element , Ahmed Element , Dr. Hani Aziz Ameen

Midlin Element in Finite Element Analysis

u u1 N1 u2 N 2 u3 N 3 u4 N 4 u5 N 5 u6 N 6 u7 N 7 u8 N8 …………… (A.1.2)

8

u N i ui , and similar for other displacements.

i 1

N1

1

1 1 1

4

N2

1

1 1 1

4

N3

1

1 1 1

4

N4

1

1 1 1

4

N5

1

2

1 2 1

N6

1

2

1 1 2

N7

1

2

1 2 1

N8

1

2

1 1 2

These shape functions must satisfy two conditions:

8

1- N ( , ) 1

i 1

i

1 if i j

2- N i ( i , j )

0 if i j

i 1 2 3 4 5 6 7 8

i -1 1 1 -1 0 1 0 -1

i -1 -1 1 1 -1 0 1 0

26

Derivation of the shell element , Ahmed Element , Dr. Hani Aziz Ameen

Midlin Element in Finite Element Analysis

displacement shape functions obtained. Physically, this means that the

natural coordinates , , are curvilinear, and all sides of the element

become quadratic curves.

Thus,

8 8 8

x N i xi , y N i yi , z N i zi

i 1 i 1 i 1

Figure (5) (a) in chapter three shows the isoparametric hexahedron

element used in the fluid finite element formulation. Both types of shape

functions for the fluid element could be obtained in the same manner as

for the shell element. Thus, the velocity shape functions N vi could be

written as :

N vi

1

1 0 1 0 1 0 0 0 0 2 i 1,2,......8

8

N vi

1

4

1 2 1 0 1 0 i 9,11,17,19

N vi

1

4

1 2 1 0 1 0 i 10,12,18,20

N vi

1

4

1 2 1 0 1 0 i 13,14,15,16

Where,

0 i , 0 i , 0 i

(A1.2).

27

Derivation of the shell element , Ahmed Element , Dr. Hani Aziz Ameen

Midlin Element in Finite Element Analysis

i i i i i i i i

1 -1 -1 -1 11 0 1 -1

2 1 -1 -1 12 -1 0 -1

3 1 1 -1 13 -1 -1 0

4 -1 1 -1 14 1 -1 0

5 -1 -1 1 15 1 1 0

6 1 -1 1 16 -1 1 0

7 1 1 1 17 0 -1 1

8 -1 1 1 18 1 0 1

9 0 -1 -1 19 0 1 1

10 1 0 -1 20 -1 0 1

N pi

1

1 0 1 0 1 0 i 1,2,......8

8

Where,

0 i , 0 i , 0 i

(A1.2).

28

Derivation of the shell element , Ahmed Element , Dr. Hani Aziz Ameen

Midlin Element in Finite Element Analysis

Reactor [56]

If a vector V3 is defined (by its three Cartesian components for

instance), it is possible to erect an infinity of mutually perpendicular

vectors orthogonal to it. Some scheme therefore has to be adopted to

eliminate this choice, and indeed quite arbitrary decisions can be made

here. A convenient scheme adopted in the present work related the choice

to the global x and y axis.

If i for instance is the unit vector along the x axis,

V1 i V3

V3 and the x axis. As V2 has to be orthogonal to both V1 and V3 , one can

have,

V2 V3 V1

divided by their scalar lengths, giving the unit vectors:

v1 , v2 , and v3 .

29

Derivation of the shell element , Ahmed Element , Dr. Hani Aziz Ameen

Midlin Element in Finite Element Analysis

In calculating the element strain, certain derivatives of the generic

displacement u, v, w with respect to the global coordinates x, y, z are

needed. But, since the shape functions are expressed in terms of the local

coordinates ,, , it is useful to use a convenient transformation as

follows:

The chain rule of partial differential calculus for differentiation of shape

functions N ,, with respect to , , and produces:

N N x N y N z

x y z

N N x N y N z

x y z

N N x N y N z

x y z

In matrix form:

N x y z N

x

N x y z N

y

N x y z N

z

For this arrangement, the terms in the coefficient matrix are easily

obtained by differentiating equation (3) . This array is called the Jacobian

matrix [J] which contains the derivatives of the global coordinates with

respect to the local coordinates. Thus,

N i N i

x

N i J N i

y

N N i

i

z

32

Derivation of the shell element , Ahmed Element , Dr. Hani Aziz Ameen

Midlin Element in Finite Element Analysis

Where,

x y z

x y z

Jacobian matrix [J] =

x y z

N i N i

x

N

i [ J ]1 N i

y

N i N

i

z

31

Derivation of the shell element , Ahmed Element , Dr. Hani Aziz Ameen

Midlin Element in Finite Element Analysis

The major purpose of the coordinate transformation is to permit the

material or element properties, known with reference to one coordinate

system, to be used in another coordinate system.

Let two Cartesian reference frames x y z and x y z be arbitrarily

oriented with respect to one another, figure (A4.1). Also, the directional

cosines of the angles between the various axes are tabulated in the same

figure. For example, 1 is the directional cosine of axis x with respect to

x axis and so on.

y

z

x x

x y z

x 1 m1 n1

y 2 m2 n2

z 3 m3 n3

cosines.

32

Derivation of the shell element , Ahmed Element , Dr. Hani Aziz Ameen

Midlin Element in Finite Element Analysis

system as and in x y z coordinates or as and in x y z

coordinates. Stresses and are arranged in the order:

x x

y y

z …………... (a) z ……………… (b)

xy xy

yz yz

zx zx

x x

y y

z …………..(c) z ……………….(d)

xy xy

yz yz

zx zx

[D] ……………………………………………………………. (A4.1)

Or,

[D] ………………………………………………………….. (A4.2)

Where, D and D are the stress-strain matrix (see sec.3.3.4) in the either

and vice versa can be implemented through the following approach [13].

For the convince in rotation of axes, the stress vector may be recast

into the form of a symmetric 3×3 matrix as follows:

x xy xz

yx y yz ………………………………………….……… (A4.3)

zx zy z

33

Derivation of the shell element , Ahmed Element , Dr. Hani Aziz Ameen

Midlin Element in Finite Element Analysis

Where, [R] is the rotation matrix and has the form

1 m1 n1

R 2 m2 n2 ………………………………………………….. (A4.5)

3 m3 n3

In this matrix, the terms 1 , m1 and so on, are the directional cosines.

Similarly, the strain vector may be recast as the symmetric 3×3 matrix:

x xy xz

yx

y yz ………………………………………….……… (A4.6)

zx zy z

R RT ………………………………………………………. (A4.7)

Now, rewrite the expanded result of equation (A4.7) as:

T …………………………………………………………..... (A4.8)

In this equation, the strains are in the forms of equation (A4.c) and (A4.d)

instead of equation (A4.6). The 6×6 strain transformation matrix T in

equation (A4.8) is as follows:

12 m12 n12 1m1 m1n1 n1 1

2

2 m22 n22 2 m2 m2 n 2 n2 2

23 m32 n32 3 m3 n3 3

T m3 n3

… (A4.9)

2 1 2 2m1m2 2n1n2 1m1 2 m1 m1n2 n2 m1 n1 2 n2 1

2 2m2 m3 2n2 n3 2 m3 3 m2 m2 n3 n2 m3 n2 3 n3 2

2 3

2 3 1 2m3 m1 2n3 n1 3 m1 1m3 m3 n1 n3 m1 n3 1 n1 3

The form of the stress transformation matrix T is derived from the

argument that during any virtual displacement, the resulting increment in

strain energy density U o must be the same regardless of the coordinate

system in which it is computed. Thus,

U o T T ……………………………………………. (A4.10)

34

Derivation of the shell element , Ahmed Element , Dr. Hani Aziz Ameen

Midlin Element in Finite Element Analysis

into equation (A4.10) to obtain:

T T T T ………………………………………………. (A4.11)

Hence, one can conclude that,

T ………………………………………………………… (A4.12)

Where,

T T T ………………………………………………………. (A4.13)

Thus, the stress transformation matrix T is proven to be the transposed

inverse of the strain transformation matrix T .

Now, to transform the stress-strain relationships from one set of

coordinates to another, substitute equation (A4.8) and equation (A4.12)

into equation (A4.1) to obtain:

T DT ………………………………………………….… (A4.14)

Then, premultiply equation (A4.13) by T 1 and use equation (A4.12) to

find:

T T DT …………………………………………………... (A4.15)

Or,

D …………………………………………………………... (A4.16)

Where,

D T T DT ………………………………………………….... (A4.17)

which represents the transformation of D to D .

The reverse transformation is:

D T DT T …………………………………………………... (A4.18)

35

Derivation of the shell element , Ahmed Element , Dr. Hani Aziz Ameen

Midlin Element in Finite Element Analysis

The process of computing the value of a definite integral (see

figure A5.1 (a)) from a set of numerical values of the integral is called

numerical integration.

x2

x1

formula and then integrating this formula between specified limits. When

applied to the integration of a function of a single variable, the method is

referred to as mechanical quadrature. The most accurate quadrature

formula in common usage is that of Gauss, which involves unequally

spaced points that are symmetrically placed. To apply Gauss's method,

the variable is changed from x to the dimensionless coordinate with its

origin at the center of the range of integration, as shown in figure (A5.1

(b)). The expression for x in term of is

x

1

1 x1 1 x2 ……………………………………………. (A5.2)

2

Substitution of equation (A5.2) into the function in equation (A5.1) gives,

f ( x) ( ) ………………………………………………………..... (A5.3)

Also,

1

dx ( x 2 x1 )d ………………………………………………….... (A5.4)

2

Then, substituting equations (A5.3) and (A5.4) into equation (1) and

changing the limits of integration yields,

1

1

I x ( x 2 x1 ) ( )d …………………………………………...… (A5.5)

2 1

of summing the weighted values of ( ) at n specified points as follows:

36

Derivation of the shell element , Ahmed Element , Dr. Hani Aziz Ameen

Midlin Element in Finite Element Analysis

f (x )

f (x )

0 x1 x2 x

(a)

( )

( )

(b)

1 0 1

Figure (A5.1) Gaussian quadrature.

x r

d d

y

dA

xb y

d

r

d

r

r d

y

d

x

r

xb

x

xb

d

y r

xb r d

j

xb

k x

xb i xb

xb

z

xb

37

Derivation of the shell element , Ahmed Element , Dr. Hani Aziz Ameen

Midlin Element in Finite Element Analysis

1 n

I ( )d R j ( j )

1 j 1

Or,

I R1 (1 ) R2 (2 ) ............... Rn (n ) …………………………..… (A5.6)

at which ( ) is to be calculated. The values of these parameters are listed

in table (A5.1).

For quadrilaterals in Cartesian coordinates, the type of integration to be

performed is:

y2 x2

y1 x1

the natural coordinates for a quadrilateral. One can accomplish this by

expressing the function f in terms of , and using the limits -1 to 1

for each of the integrals. In addition, the infinitesimal area dA dxdy must

be replaced by an appropriate expression in terms of d and d . For this

purpose, figure (A5.2) shows an infinitesimal area dA in the natural

coordinates. Vector r locates a generic point in the Cartesian coordinates

x and y , as follows:

r x y xi yi …………………………………………………..... (A5.8)

r x y

i j …………………………………………………...… (A5.9)

r x y

i j ………………………………………………….... (A5.10)

38

Derivation of the shell element , Ahmed Element , Dr. Hani Aziz Ameen

Midlin Element in Finite Element Analysis

n ± i Ri

1 0.0 2.0

2 0.5773502692 1.0

3 0.7745966692 0.5555555556

0.0 0.8888888889

4 0.8611363116 0.3478548451

0.3399810436 0.6521451549

5 0.9061798459 0.2369268851

0.53884693101 0.4786286705

0.0 0.5688888889

6 0.9324695142 0.1713244924

0.6612093865 0.3607615730

0.2386191861 0.4679139346

7 0.9491079123 0.1294849662

0.7415311856 0.2797053915

0.4058451514 0.3818300505

0.0 0.4179591837

8 0.9602898565 0.1012285363

0.7966664774 0.2223810345

0.5255324099 0.3137066459

0.1834346425 0.3626837834

39

Derivation of the shell element , Ahmed Element , Dr. Hani Aziz Ameen

Midlin Element in Finite Element Analysis

(A5.10) form two adjacent sides of the infinitesimal parallelogram of area

dA in the figure. This area may be determined from the following vector

triple product:

r r

dA d d k ……………………………………………. (A5.11)

produces,

x y x y

dA dd ………………………………………… (A5.12)

a 2×2 determinate. That is,

x y

dA dd J dd ……………………………………….. (A5.13)

x y

form of the integral in equation (A5.7) becomes,

1 1

I f ( , ) J dd ……………………………………………….. (A5.14)

1 1

n n

I R j Rk f ( j , k ) J ( j , k ) …………………………………..… (A5.15)

k 1 j 1

Where, R j and Rk are weighting factors for the function evaluated at the

quadrilateral are illustrated in figure (A5.3).

42

Derivation of the shell element , Ahmed Element , Dr. Hani Aziz Ameen

Midlin Element in Finite Element Analysis

(a) (b)

(c) (d)

(c) n 3 (d) n 4 (each way).

evaluated has the form:

I f ( x, y, z )dxdydz …………………………………………… (A5.16)

Before integrating, one can rewrite the functions in terms of the natural

coordinates , , and and using the limits -1 to 1 for each of the

integrals. In addition, the infinitesimal volume dV dxdydz must be

41

Derivation of the shell element , Ahmed Element , Dr. Hani Aziz Ameen

Midlin Element in Finite Element Analysis

the same procedure used for the quadrilaterals, dV can be written as:

x y z

x y z

dV ddd J ddd ………………………...… (A5.17)

x y z

revised form of the integral in equation (A5.16) becomes,

1 1 1

I f ( , , ) J ddd …………………………………….… (A5.18)

1 1 1

n n n

I R j Rk Rl f ( j , k , l ) J ( j , k , l ) ……………………...… (A5.19)

l 1 k 1 j 1

point ( j ,k , l ) . Integration points for n 1,2,3 and 4 (each way) are:

1,8,27, and 64, respectively.

42

Derivation of the shell element , Ahmed Element , Dr. Hani Aziz Ameen

Midlin Element in Finite Element Analysis

References

[1] Flugg W., "Stress in shell", Springer-Verlag, 4th ed., New York, 1967.

[2] Hou-Cheng H., "Static and Dynamic Analysis of Plate and Shells",

Springer-Verlag, Berlin, 1989.

[3] Zienkiewiz O.C. and Taylor R.L., "The Finite Element Method",

Butterworth-Heinemann, 5th ed., Oxford, 2000.

[4] Novozhlov V.V., "The Theory of Thin Shells", 2 nd ed., Noordhoff

Ltd., Groningen, Netherlands, 1964.

[5] Vlasov V.Z., "General Theory of shells and its Applications in

Engineering", NASA TTF-99, 1964.

[6] Cowper G.R., Lindberg G.M., and Olson M.D., "A shallow Shell

Finite Element of Triangular Shape", International Journal of

Solution Structures, Vol., 6, No.8, pp. 1133-1156, 1970.

[7] Ahmed S., Itons B.M. and Zienkiewic O.C., "Analysis of Thick and

Thin Shell Structure by Curved Finite Elements", International

Journal of Numerical and Mechanical Engineering, Vol.2, No.3, pp.

419-451, 1970.

[8] Mindlin R.O., "Influence of Rotary Inertia and Shear in Flexural

Motion of Isotropic and Elastic Plates", Journal of Applied

Mechanics, Vol.73, pp.31-38, 1951.

[9] Weaver W.Jr. and Johnston P.R., "Finite Element for Structural

Analysis", Prentice-Hall, Englewood Cliffs, N.J. 1984.

[10] Cook R.D., "Concepts and Applications of Finite Element Analysis",

2nd ed., Wiley, New York, 1981.

[11] Ahmad S., Irons B.M., and Zienkiewic O.C, "A Simple Matrix-

Vector Handling Scheme for Three-Dimensional and Shell

Analysis", IJNME, Vol.2, No.4, pp.509-522, 1970.

43

Derivation of the shell element , Ahmed Element , Dr. Hani Aziz Ameen

Midlin Element in Finite Element Analysis

Shells", 2nd ed., McGraw-Hill, New York, 1959.

[13] Washizu K., "Variational Methods in Elasticity and Plasticity", 2 nd

ed., Pergamon Press, New York, 1975.

[14] Zienkiewiz O.C., Taylor R.L., and Too J.M., "Reduced Integration

Technique in General Analysis of Plate and Shells", IJNME, Vol.3,

No.2, pp.275-290, 1971.

[15] Scarborough J.B., "Numerical Mathematical Analysis", 9th ed., Johns

Hopkins Press, Baltimore, 1986.

Elements", Prentic-Hall, Englewood Cliffs, N.J., 1987.

44

- Mae160_chap14Uploaded byaraveeti
- A Handbook 0f Building Structure = Allan HodgkinsonUploaded byNiques Dawson
- The finite element methodUploaded byem13un
- Finite Element MethodUploaded byKeri Mason
- 7207839 Ansys Structural Analysis GuideUploaded byPandharkar
- Automobile Engg-UNIT-2Uploaded byMuthuvel M
- NISTNCSTAR1DraftUploaded byjrod
- Rapid PrototypingUploaded byAnkit
- Mae160_chap13Uploaded byaraveeti
- Nonlinear Micro Finite Element Analysis of Human Trabecular Bone 2003Uploaded bySIMULIACorp
- Chapter 9Uploaded byradoslav121
- Chapter 2Uploaded byradoslav121
- 17192061-Phd-Thesis-1997-DUploaded byawalmanuhuruk
- MMM351 - Machine DynamicsUploaded bySam Wang Chern Peng
- Chapter 3Uploaded byradoslav121
- Chapter 4Uploaded byradoslav121
- Chapter 5Uploaded byradoslav121
- Chapter 6Uploaded byradoslav121
- Chapter 8Uploaded byradoslav121
- Nonlinear Solid MechanicsUploaded byAzam Rashdi
- 27324976-Mathcad-User-s-BookUploaded byshabin007
- Notes for Finite Element Method Thermal AnalysisUploaded byManikanth Valavoju
- Introduct Fatigue PJHUploaded byMartin Stone
- Nair ThesisUploaded bySuneel Kanhar
- FEM_TUDOUploaded byVitor Fc
- Materials Methodology Stress FactorsUploaded by538540
- B013-019Uploaded by538540
- Microsoft Word - Creep Behaviour of Reformer TubesUploaded byeibrahim1
- Strength and Material 1 Project 2 Photo Elasticity by Ryan ParkUploaded byRyan Horyang Park

- RESIDUAL ELASTO-PLASTIC STRESSES ANALYSIS OF POLYMERIC THICK –WALLED PRESSURIZED CYLINDER- Hani Aziz AmeenUploaded byHani Aziz Ameen
- Stess Elasticity Prof.dr.Hani Aziz AmeenUploaded byHani Aziz Ameen
- Formability of Laser WeldingUploaded byHani Aziz Ameen
- 8-High Rate Enfergy Forming-Prof.dr.HaniUploaded byHani Aziz Ameen
- Deep Drawing- Prof.dr.Hani Aziz AmeenUploaded byHani Aziz Ameen
- Professor Dr Hani Aziz AmeenUploaded byHani Aziz Ameen
- 7-Blanking and Piercing-Prof.dr.Hani AzizUploaded byHani Aziz Ameen
- 6-Forging Process- Prof.dr. Hani Aziz AmeenUploaded byHani Aziz Ameen
- Hani Aziz Ameen هاني عزيز امينUploaded byHani Aziz Ameen
- Hani Aziz AmeenUploaded byHani Aziz Ameen
- 4-Extrusion Process-Prof.dr.Hani Aziz AmeenUploaded byHani Aziz Ameen
- 2-Wire Drawing-Prof.Dr.Hani Aziz Ameen.pdfUploaded byHani Aziz Ameen
- Rolling Process-Prof.dr.Hani Aziz AmeenUploaded byHani Aziz Ameen
- The Effect of Coupled Field on the Vibration Characteristics and Stresses of Turbomachinery System- Hani Aziz AmeenUploaded byHani Aziz Ameen
- شهادة تقديرية - دكتور هاني عزيز أمين.pdfUploaded byHani Aziz Ameen
- Finite Element Analysis of Large Diameter Concrete Spherical Shell Domes - Hani Aziz AmeenUploaded byHani Aziz Ameen
- Experimental and Theoretical Investigation of Impact Dynamic Plasticity for CK45 - Hani Aziz AmeenUploaded byHani Aziz Ameen
- Effect of Post-Buckling on the Stiffness and Stress of Plate- Hani Aziz AmeenUploaded byHani Aziz Ameen
- Finite Element Analysis of the Dish Multi-Point FormingUploaded byHani Aziz Ameen
- Effect of Loading Path on Stress DistributionUploaded byHani Aziz Ameen
- Ansys analysis of concrete Planter round.pdfUploaded byHani Aziz Ameen
- ANSYS of Groundwater Flow Problem-Hani Aziz AmeenUploaded byHani Aziz Ameen
- Static and Dynamic Characteristics of Slotted Cylinder SpringUploaded byHani Aziz Ameen
- Manufacturing Cosine Die Profile for Tubes Hydroforming Test (Bulging)Uploaded byHani Aziz Ameen
- Slotted Cylinder SpringUploaded byHani Aziz Ameen
- Moving and Repeating Load in Ansys APDL - Hani Aziz AmeenUploaded byHani Aziz Ameen

- lec10_11 Z-transform.pptUploaded byIndomi double
- The Chinese Remainder TheoremUploaded byjamofrey
- transcript for e-portfolioUploaded byapi-253299910
- maths15.pdfUploaded byjhalakdutta
- c Star Algebra sUploaded bymandestature
- Complex NumbersUploaded bybendan
- 20. Mathieu Functions.pdfUploaded byAlexandra Nicoleta Teisi
- KC SINHAUploaded bysai kiran
- mathematics magazine 2003.pdfUploaded byΘωμαςΣτεφανιδης
- Rostock Delta KinematicsUploaded byjeanpremier
- Inequalities III - Ali Gurel - MOP (Blue) 2008Uploaded byTiago Souza
- IIT JEE MathematicsUploaded byUmang Mishra
- Solver Fortran FunctionsUploaded bymaliq12
- Plate and ShellUploaded bykumar9172
- The conformal group in various dimensionsUploaded byzwicky
- DirectionCosineMatrixUploaded byNeelam Shukla
- Polynomial Equations - Cardano v9Uploaded byNikitas Manolis
- RMOUploaded byPavan Teja
- NCERT SOLUTION 11TH CLASS MATHUploaded bywww_mohilaj
- LDA TutorialUploaded bySiddharth Mishra
- Carslaw - A Problem in the Linear Flow of Heat Discussed From the Point of View of Theory of Integral EquationsUploaded byRoberson Silva
- MTS 101 (C)Uploaded byUsman Samuel Babalola
- algebra_iUploaded byFahriGöktaş
- Problem Set 15 SolutionsUploaded byFABIANCHO2210
- Basic Math SymbolsUploaded byj1031salazar4545
- Paulsen_An Introduction to the Theory of Reproducing Kernel Hilbert SpacesUploaded bygrzerysz
- DE_Second_Order_Concepts.pdfUploaded bymalini72
- Oliveboard _ Number System.pdfUploaded byVreddy Reddy
- Dihedral AngleUploaded byMarto Fe
- MOP Homework 2013Uploaded byNishant

## Much more than documents.

Discover everything Scribd has to offer, including books and audiobooks from major publishers.

Cancel anytime.