BAHIR DAR UNIVERSITY

INSTITUE of TECHNOLOGY



MECHANICAL and INDUSTRIAL ENGINEERING
DEPARTEMENT
SCHOOL OF GRADUATE STUDIES
IN MANUFACTURING ENGINEERING PROGRAM

ADVANCED FINITE ELEMENT

Submitted to: Dr. Ing. Demiss Alemu
Prepared by: Yohannes Reggasa
ID /M.Sc/R/00032/04
Feb, 2012
BAHIRDAR, ETHIOPIA
2

ASSIGNMENT PART TWO
1. compute the element of stiffness matrix of bilinear isopermetric element (linear
rectangular element using two point gauss quadrature.

|
|
.
|

\
|

= dxdy
y
N
y
N
x
N
x
N
k K
2 1 2 1
2 , 1

Solution
From the equation
y x
y
Nj
y
Ni
x
j
x
Ni
x
x
y
y
j i k

|
|
.
|

\
|

=

2
1
2
1
,
(1)

|
|
.
|

\
|

= dxdy
y
N
y
N
x
N
x
N
k K
2 1 2 1
2 , 1

Where; X = ( )xi Ni
i
,
4
1

=
and Y= ( )yi Ni
i
,
4
1

=
(2)
For rectangular isopermetric element we have the following phyisical cordinate value at each
node
( -1,1) 4 3 (1,1)


(-1 ,-1) 1 2 (1,-1)
From hermitan shape function for rectangular isopermetric element we have
( )( ) = 1 1
4
1
1
N ( )( ) + = 1 1
4
1
2 N

( )( ) + + = 1 1
4
1
3 N
( )( ) + = 1 1
4
1
4 N



3

X
N
X
N
X
N

1 1 1
and
X
N
X
N
X
N

2 2 2

Y
N
Y
N
Y
N

1 1 1
and
Y
N
Y
N
Y
N

2 2 2

when we write in matric form,
(
(
(

Y
N
X
N
1
1
=
(
(
(

Y Y
X X

1
1
N
N
and
(
(
(

Y
N
X
N
2
2
=
(
(
(

Y Y
X X

2
2
N
N

( )
N X N
X
2 1 1
, + = ( ) ,
2
X + ( ) ( ) , ,
4
3
3 N N
X +
4
X
( )
N Y N
Y
2 1 1
, + = ( ) ,
2
Y + ( ) ( ) , ,
4
3
3 N N
Y +
4
Y
X = ( )( ) 1 1 [
4
1
(-1) + ( )( ) + 1 1 (1) +( )( ) + + 1 1 (-1) +( )( ) + 1 1 (-1)] =
4
4
=
Y = ( )( ) 1 1 [
4
1
(-1) +( )( ) + 1 1 (-1) +( )( ) + + 1 1 (1) +( )( ) + 1 1 (1)] =
4
4
=

X
=1 ,

X
= 0 ,

Y
= 0 ,

Y
=1 in matrix form
(
(
(

Y Y
X X


= J=
(

1 0
0 1


(

22 21
12 11
R R
R R

= R =J
-1
=

(

1 0
0 1




(

|
|
.
|

\
|

|
|
.
|

\
|

+
|
|
.
|

\
|

|
|
.
|

\
|

=
1
1
2
22
2
12
1
22
1
21
2
12
2
11
1
12
1
11
1
1
2 , 1
* *

N
R
N
R
N
R
N
R
N
R
N
R
N
R
N
R K
K

y x *

dxdy=det (J)= J
11
J
22
-J
12
J
2 1 =
1
1 * * *
1
1
2
22
1
22
2
11
1
11
1
1
2 , 1

(

|
|
.
|

\
|

|
|
.
|

\
|

+
|
|
.
|

\
|

|
|
.
|

\
|

=

N
R
N
R
N
R
N
R K
K


4

( ) ( ) , ,
2
1
1
1
1
1
2 , 1


=

= =
i
f wj wi f
K

The next job is to find the ( ) , f

The derviation each shape function with respect to sayi and eta will be,
( )( )

1 1
4
1
1
N
=
4
1
and ( )( )
4
1
1 1
4
1
2

= + =

N

( )( )
4
1
1 1
4
1
1

= =

N
( )( )
4
1
1 1
4
1
2

= + =

N

( ) =

=


2 1 2 1
,
N N N N
f
|
.
|

\
|
4
1

|
.
|

\
|
4
1
+
|
.
|

\
|
4
1
|
.
|

\
|
4
1

=
16
2
2 2
+

Now from the table given for Gauss point integration from -1 to 1
We have 57735 . 0
1
= and 57735 . 0
2
=
2691 . 0
1
= and 89626 . 0
2
= we have also w
1
=1=w
2
Thus we can easily compute the give question by using the equation of integration for Gauss
two point quadrature isopermetric rectangular element.
( ) ( ) ( ) ( )
2 2 2 2 1 2 1 2 2 1 2 1 1 1 1 1
, , , , f w w f w w f w w f W W I + + + =
( ) = , f =
16
2
2 2
+
by using this equation we can calculate the fellowing value
( ) =
1 1
, f =0.01733 , ( ) =
2 1
, f 0.14140 , ( ) =
1 2
, f 0.03363 and ( ) =
2 2
, f 0.14140
Therfore, ( ) ( ) ( ) ( )
2 2 2 2 1 2 1 2 2 1 2 1 1 1 1 1
, , , , f w w f w w f w w f W W I + + + = =0.33376






5

Preprocessing: Defining the Problem
1. Form Geometry
Boolean operations provide a means to create complicated solid models. These
procedures make it easy to combine simple geometric entities to create more complex
bodies. Subtraction will used to create this model; however, many other Boolean
operations can be used in ANSYS.
a. Create the main rectangular shape
Instead of creating the geometry using key points, we will create an area
(using GUI)
Preprocessor > Modeling > Create > Areas > Rectangle > By Dimension
Fill in the window of left corner with x1,x2 = 0,1.8 and y1,y2=0,0.5 then we will
get as follows.

2. Create the small rectangle that to be subtracted from big one to form angle plate
Preprocessor > Modeling > Create > Areas > rectangle > by dimension
Fill in the window that appear with x1,x2= 0.9,1.8 and y1,y2=0.3,0.5 then we will get..


6


a. Subtraction
Now we want to subtract the small rectangle from the big rectangle. To perform the Boolean
operation, from the Preprocessor menu select:
Modeling > Operate > Booleans > Subtract > Areas. Then select first the big area then apply
then select the smaller area which looks as follows after all it form angle plated form






The interest of our geometry looks like this


7

Meshining /mesh tool/
Preprocessor > Meshing > mesh tool > set the edge length of by 0.1 >ok>select
tri>area>mesh> pick all
Modify mesh>refine at>node then select the node where there is possibility for occurrence
of high stress the it looks like this






Save the work by utility menu>save as>check job name >ok

Solution phase applying boundary condition and force/moment
Solution> analysis type>new analysis>static>ok
Solution >define load>apply>displacement> on line the select the left end of object to constrain
in all degree of freedom >all DoF>apply >select bottom end of object >constrain UY>ok



8

Apply load
Solution >define load>apply>force/moment>on node
Then assign load to the right end of object with that of first node with FX=25e3 second and third
node with 50e3 and last end node with 25e3 ok


Save the work
Solving the System
Solution > Solve > Current LS


9

Post processing:
The graphical analytical solution for distributed stress and deflection
Deformation
General post processor>plot result>deformed shape >Deformed+undeformed>ok

UTILITY MENU >PLOT>NODE
General postprocessor>plot result>plot contour>nodal solution>DoF>Displacement vector
sum

10


Numerical value of deformation

Deformation due to Stresses
General Postproc > Plot Results > Nodal Solution>stress>von misses>ok

11

Numerical value of stress by von misses criteria solution
Utility menu>general post processor>list result>nodal solution>stress>von misses>ok


Plane stress
Utility menu>general post processor>list result nodal solution>stress>xy shear stress>ok






12

4. Find the stress disribution on the given plate using ansys,if the width of the plate is 0.35m,
diameter hole is 0.2m and length of plate is 0.5m for the same loading and the thichness of
palte is 2mm E= 2.1X10
11
Pa and v=0.3

Verification Example


Model by Pro/Eng.






Solution for problem suggested as follows /using Ansys/

Preprocessing: Defining the Problem
1. Give the Simplified Version a Title
Utility Menu > File > Change Title

13

2. Form Geometry
Boolean operations provide a means to create complicated solid models. These
procedures make it easy to combine simple geometric entities to create more complex
bodies. Subtraction will used to create this model; however, many other Boolean
operations can be used in ANSYS.
a. Create the main rectangular shape
Instead of creating the geometry using key points, we will create an area
(using GUI)
Preprocessor > Modeling > Create > Areas > Rectangle > By 2 Corners
Fill in the window as shown below left corner. This will create a rectangle where
the bottom left `corner has the coordinates 0, 0, 0 and the top right corner has the
coordinates 500, 350, 0.

We will get
Ok
14


a. Create the circle
Preprocessor > Modeling > Create > Areas > Circle > Solid Circle

Fill in the window as shown above. This will create a circle where the
center has the coordinates 250, 175, 0 (the center of the rectangle) and the
radius of the circle is 100 mm.
15

The zoomed one with the interest to show white circle



b. Subtraction
Now we want to subtract the circle from the rectangle. Prior to this operation, your image should
resemble the following: To perform the Boolean operation, from the Preprocessor menu select:
Modeling > Operate > Booleans > Subtract > Areas
16


At this point a 'Subtract Areas' window will pop up and the ANSYS Input window will
display the following message:
Therefore, select the base area (the rectangle) by clicking on it. Note: The selected area will
turn pink once it is selected.
The following window may appear because there are 2 areas at the location you clicked.
Ensure that the entire rectangular area is selected (otherwise click 'Next') and then click
OK.
Click 'OK' on the 'Subtract Areas' window.
Now you will be prompted to select the areas to be subtracted, select the circle by
clicking on it and then click 'OK'.
You should now have the following model:
17


3. Define the Type of Element
It is now necessary to define the type of element to use for our problem:
Preprocessor Menu > Element Type > Add/Edit/Delete
Add the following type of element: Solid (under the Structural heading) and the Quad 82
element, as shown in the above figure.
18


PLANE82 is a higher order version of the two-dimensional, four-node element PLANE82 is an
eight nodes quadrilateral element which is better suited to model curved boundaries.
For this example, we need a plane stress element with thickness, therefore
Click on the 'Options...' button, from bar of element type Click and hold the K3 button, and
select 'Plane strs w/thk', as shown below.

19

4. Define Geometric Properties
That is - assigning Real constant /providing thickness for solid plate/
Preprocessor menu > Real Constants > Add/Edit/Delete
Enter a thickness of 2 as shown in the figure below. This defines a plate thickness of
2mm)

5. Element Material Properties

Preprocessor>Material Props>Material models>Structural>Linear>Elastic>Isotropic


20

6. Meshining /mesh size/ To tell ANSYS how big the elements should be,
Preprocessor > Meshing > Size Cntrls > Manual Size > Areas > All Areas
And Select an element edge length of 50. We will return later to determine if this was adequate
for the problem.
Mesh:- Now the frame can be meshed. In the 'Preprocessor' menu select
Meshing > Mesh > Areas > Free and select the area /pick all /when prompted bar appears
You should now have the following ,




When the meshed plate magnified it looks like as follows,
21


I make more refine at this node due to there is high stress at this point
Saving Your Job
Utility Menu > File > Save as... select place where we need >ok close

Solution Phase: Assigning Loads and boundary constraints
You have now defined your model. It is now time to apply the load(s) and constraint(s) and solve
the resulting system of equations.
1. Define Analysis Type Ensure that a Static Analysis will be performed
22

Solution > Analysis Type > New Analysis





2. Apply Constraints
Solution > Define Loads > Apply > Structural > Displacement > On Lines
Select the left end of the plate and click on 'Apply' in the 'Apply U, ROT on Lines' window.
Fill in the window that all DOF's are constrained. Therefore, select 'All DOF' by clicking on it
and enter '0' in the Value field as shown above. You will see some blue triangles in the graphics
window indicating the displacement constraints.

23


1. Apply Loads
As shown in the diagram, there is a load of 450,000N/m distributed on the right hand side
of the plate. To apply this load:
Solution > Define Loads > Apply > Structural > Pressure > On Lines
When the window appears, select the line along the right hand edge of the plate and click 'OK'
The pressure is acting away from the surface of the plate, and is therefore defined as a negative
pressure.
24


1. Solving the System
Solution > Solve > Current LS



25

Post processing:
1. Viewing the result. At this point we need to find whether or not the final result has
converged. We will do this by looking at the deflection and stress at particular nodes while
changing the size of the meshing element.
Since we have an analytical solution for the maximum stress point, we will check the stress at
this point. First we need to find the node corresponding to the top of the hole in the plate. First
plot and number the nodes
Utility Menu > Plot> node
The graphical analytical solution for distributed stress and deflection
calculation
Deformation
General post processor>plot result>deformed shape >Deformed+undeformed>ok

26


Deformation @nodal level
General postprocessor>plot result>plot contour>nodal solution>DoF>Displacement vector
sum

Numerical value of Deflection
General Postprocessor>List Result >Nodal Solution>DOF>Displacement Sum >Ok
27



Deformation due to Stresses
General Postproc > Plot Results > Nodal Solution>stress>von misses>ok

Utility menu>general post processor>list result>nodal solution>stress>von misses>ok
28


Deformation due to Principal stress




29

Numerical value of plane stress
General post processor>List result>nodal solution>stress>principal stress >ok
Here I show the maximum and minimum nodal stress