Scribd
Upload
89 views4 pages

HW2 Fem 2012

This homework assignment involves using finite element methods to analyze two beam problems. For the first problem, students are asked to analyze a cantilever beam with a varying moment of inertia using 1 and multiple elements, and compare the results to an exact solution. The number of elements needed to achieve less than 0.5% relative error is determined. For the second problem, students must substructure a plane sheet metal into elements, eliminate internal degrees of freedom, calculate nodal forces and stresses, and compare results to exact solutions.

Uploaded by

greatkurosh
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
89 views4 pages

HW2 Fem 2012

This homework assignment involves using finite element methods to analyze two beam problems. For the first problem, students are asked to analyze a cantilever beam with a varying moment of inertia using 1 and multiple elements, and compare the results to an exact solution. The number of elements needed to achieve less than 0.5% relative error is determined. For the second problem, students must substructure a plane sheet metal into elements, eliminate internal degrees of freedom, calculate nodal forces and stresses, and compare results to exact solutions.

Uploaded by

greatkurosh
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 4

KTH HLLFASTHETSLRA

SE1025 FEM for engineering applications

SE1025 FEM: Homework assignment 2, Per. 1, 2012/13


Information
The homework assignments are together with the computer workshops a compulsory part of
the course. The assignments should be solved in groups of two or three students. The deadline
of HW1 is Friday, Oct. 7, 6pm. A total of 10 points can be obtained on HW2.
Note! If programs like Maple and Matlab are used to solve the homework assignments,
printout from these program must only be added as an appendix to the solution. Solutions only
consisting of printout from Maple/Matlab will be graded with zero points.
Problem 1. (4 points) A cantilever beam of linear elastic material is loaded by a point force P0.
The beam has a length 2L and a linearly varying moment of inertia given by
4
I x = I 0 1 x L , where I0 is a constant (dimension m ) and a non-dimensional
parameter given in Table 1, where mm is the month in the "personal number" employed on the
front endpage. Calculate the deflection of the beam at its right end by use of the finite element
method. Use a two node beam element (two degrees of freedom at each node) with a constant
value of the bending stiffness in the element. An exact solution can be obtained with only one
element if = 0 . This is not the case for 0 , but a solution with a sufficient accuracy can
be obtained if enough elements are used. This strategy should be used here. Carry out and
present the analysis in the following steps:
(a) Solve the problem by use of only one element (account for all the computational steps).
(b) Repeat the analysis with a gradually increasing number of elements by use of the FEMbased Matlab program "beam1D.m", which can be downloaded from the home page (or
you can also use ANSYS).
(c) Compare with the exact solution given in Figure 1 and plot the relative error
FEM

Error Rel = 0

exact

exact

as a function of the number of element used in a log-log diagram. How many elements
are needed in order for the relative error to be less than 0.5%?
(d) ("bonus problem" + 1p) If the varying moment of inertia of the beam is considered,
stiffness matrix of the beam element can be written on the form k e = k e const + k e mod ,
where ke,konst represent the case with a constant value of I (here evaluated at the
element mid point), and ke,mod a stiffness correction term due the linear variation across
the beam. Show that, if the cantilever beam in this example is modelled by only one
element, the contributions to the stiffness matrix can be written as
EI 0 1 d 2 N T d 2 N
EI 1 2 N T d 2 N
- ------------ ---------- d , k e mod = -------0- d--------------------- d .
k e konst = ------3
2
2
3
2
2
L 1 d d
L 1 d d
z,w

x = L

mm

01 04

0.1

P0

05 08

0.2

Exact
P 0 L 4 2 1 2 1 +
exact
------------ --------------- + ------------------- ln -----------=

solution: 0
2
3
1
EI 0

09 12

0.4

x = L

Figure 1.

x
3

SE1025 FEM: Homework assignment 2, Per. 1, 2012/13

Table 1.

KTH HLLFASTHETSLRA

SE1025 FEM for engineering applications

Problem 2. (6 points) A way to facilitate FEM-analysis of very large complex systems (air
planes, vehicles, etc.) is to divide the structure into several sub-structures. The analysis is then
carried out following a two step procedure. In step (i) the sub-structures are analysed one at the
time with the purpose to eliminate all the internal d.o.f. of the sub-structure, such that the
compliance of each sub-structure can be described by the d.o.f. of its boundary. Thus, a
relatively accurate description can be obtained by only considering the limited number of d.o.f.
defined at the boundary. In step (ii) the sub-structures are assembled and the whole system can
now be analysed, because the sub-structuring has drastically reduced the number of equations
that needs to be solved simultaneously as compared to the original system. Figure 2 illustrates
in a schematic way how a system can be sub-structured. Assume that a certain sub-structure
Collect all displacement d.o.f.
in vectors:
DR (belonging to the boundary
of the sub-structure
m D.O.F.)

Complex
structure

DI (belonging to the internal


of the sub-structure,
n D.O.F.)

sub-structure
Figure 2.

has m d.o.f. on its boundary and n d.o.f. in its domain (inside the boundary), and thus a of total
m + n d.o.f. Usually, m n in practise. The system of equations for the sub-structure can now
be formulated as
K RR K RI D R
K IR K II

DI

FR
FI

(1)

where the matrices have the dimensions according to: KRR m m ; KII n n ; KIR n m ; KRI
m n . We shall now eliminate the d.o.f. inside the sub-structure, i.e. DI. From the set of
equations of the second row in (1) we obtain
K IR D R + K II D I = F I

D I = K II F I K II K IR D R ,

(2)

which inserted into the set of equations of the first row in (1) gives
K RR D R + K RI D I = F R

K RR K IR K II K IR D R = F R K IR K II F I ,

(3)

where K RI = K IR has been used (the stiffness matrix of the sub-structure is symmetric). The
mechanical response of the sub-structure can thus be modeled by the reduced system of
equations
T

K red D R = F red with K red = K RR K IR K II K IR and F red = F R K IR K II F I .

(4)

Here, only one sub-structure shall be analysed. The sub-structure consists of a plane quadratic
sheet metal with side l and thickness h. The material is linear elastic with elastic modulus E
and Poissons ratio = 1 3 . The sub-structure is modeled by four CST-elements under the
assumption of plane stress, see Figure 3. The metal sheet is subjected to a linearly varying
traction vector acting in the x-direction on the side x = l, according to
T
t = 1 1 y l + 2 y l 0 . Carry out the analysis in the following steps:
SE1025 FEM: Homework assignment 2, Per. 1, 2012/13

KTH HLLFASTHETSLRA

SE1025 FEM for engineering applications

(a) Eliminate the d.o.f. coupled to the internal nodes of the sub-structure, i.e. only node 5 in
the present case, by use of Eqs. (1)-(3) above (m = 8, n = 2). The full, non-reduced,
stiffness matrix of the sheet metal structure is given in Eq. (4) below.
(b) Calculate the consistent nodal forces acting on element 2. The shape functions of the
element are given in Figure 3 (local node numbering is employed).
(c) Implement the displacement boundary conditions and solve the reduced system of
equations. The solution can be checked by setting 1 = 2 . The lateral contraction of
the sheet metal should then agree with what would be expected from the value of
Poissons ratio.
(d) Calculate the stress in element 2. This requires that all displacements in the element is
solved for and known, and for this purpose Eq. (2) must first be used. The stresses for
elements 1, 3 and 4 are given in Figure 3. Compare with the exact solution, i.e.
xx = 1 1 y l + 2 y l and yy = xy = 0 .
y/l
4

3 (1,1)

(0,1)

e4

e3
3

e2

Element 1:

Element 3:
1

2 (1,0) 1 x/l

Element 2: N 1 = x l y l

xx

N 2 = x l + y l 1

7 1 + 2 2 9

yy = 1 + 2 18
xy

xx

2 1 + 7 2 9

yy =
xy

e1
1

STRESSES

1 2 18
0

xx

Element 4:
N3 = 2 1 x l

1 + 2 2
=
yy
0

xy
1
2 18

Figure 3.

4 2 1 0 0 0
4 0 1 0 0
4 2 1 0
4 0 1
3Eh
4 2
K = ---------16
4

SYMMETRIC

1
0
0
0
1
0
4

0
1
0
0
0
1
2
4

4
2
4
2
4
2
4
2
16

2
4
2
4
2 ,
4
2
4
0
16

SE1025 FEM: Homework assignment 2, Per. 1, 2012/13

D =

D1

u1

D2

v1

D3

u2

D4

v2

D5
D6

u3
v3

D7

u4

D8

v4

D9

u5

D 10

v5

(4)

KTH HLLFASTHETSLRA

SE1025 FEM for engineering applications

Front endpaper - HW 2: FEM for engineering applications


Name and personal number (1):
Name and personal number (2):
Name and personal number (3):
yy

Use one of the personal numbers


to determine problem parameters:

mm

dd

Summarize your answers below:


Answer problem 1: (give complete solutions for (a)-(c) and (d) if solved, on separate pages)
FEM

One element solution: 0

Number of elements needed to bring down the relative error less than 0.5% is =

Answer problem 2: (give complete solutions for (a)-(d) on separate pages)


D1 =

D2 =

D3 =

D4 =

D5 =

D6 =

D7 =

D8 =

xx =

yy =

stresses in element 2:

Front endpaper - HW 2: FEM for engineering applications

xy =

You might also like