0% found this document useful (0 votes)
179 views16 pages

Lecture 3

1) The document summarizes the derivation of the truss element equations using Logan's finite element method. It involves setting the element type, selecting displacement functions, defining stress-strain relationships, and deriving the stiffness matrix. 2) A key step is transforming the element equations, which use a local coordinate system, into the global coordinate system used to solve load-deformation problems. This involves using a transformation matrix to relate displacements and forces between the two systems. 3) The global stiffness matrix is assembled by transforming the element stiffness matrices using the transformation matrices, resulting in a system of equations that relates nodal forces to displacements in the global coordinate system.
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
0% found this document useful (0 votes)
179 views16 pages

Lecture 3

1) The document summarizes the derivation of the truss element equations using Logan's finite element method. It involves setting the element type, selecting displacement functions, defining stress-strain relationships, and deriving the stiffness matrix. 2) A key step is transforming the element equations, which use a local coordinate system, into the global coordinate system used to solve load-deformation problems. This involves using a transformation matrix to relate displacements and forces between the two systems. 3) The global stiffness matrix is assembled by transforming the element stiffness matrices using the transformation matrices, resulting in a system of equations that relates nodal forces to displacements in the global coordinate system.
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/ 16

Lecture 3: Development of the Truss Equations.

3.1 Derivation of the Stiffness Matrix


for a Bar in Local Coordinates.
y
In 3.1 we will perform Steps 1-4 of
Logans FEM.
Derive the truss element equations.
1. Set the element type.

2. Select a displacement
y
functions.
3. Define the stress/strain

relationships.
x
4. Derive the stiffness matrix.
x
Steps 1-4 completed using an x y
elemental/local coordinate system.
Need to put element equations in

terms of global coordinate system


to solve truss load-deformation
problems.

MECH 420: Finite Element Applications


Lecture 3: Development of the Truss Equations.
Step 1: Set the element type. y
Constant cross section (A).
Constant material props on
length L (E).
element affords stretch in only the
axial direction, x .
External loads are applied at the
nodes.
State of the truss/bar element is
given by displacements at nodes:
d , d
1x 2x

Note: at this point we are not


concerned with any particular
truss geometry.

MECH 420: Finite Element Applications


Lecture 3: Development of the Truss Equations.
Step 2: Select the displacement y
function (see section 3.2).
u(x): polynomial interpolating
nodal displacements.
Conforming element:
displacement continuous within
the element.
continuity in the model DOFs
between elements.
element should be capable of
rigid body motion.
Completeness:
The necessary level of
differentiation is possible through
u(x) choice (we can have a
constant strain in the bar).
see step 3 of this procedure and
the MWR to come later in the
course.

MECH 420: Finite Element Applications


Lecture 3: Development of the Truss Equations.

Step 3: Define the stress-strain relationships.


As we are focused on the load-displacement problem, the
stress strain relationships are of primary interest.
Our conforming du
element can be x = and x = E x Set by element type.
applied within this dx
DE.
d2 x d1x
u ( x ) = d1x +
x
L

d2 x d1x d2 x d1x Set by displacement


x = and x = E function.
L L

MECH 420: Finite Element Applications


Lecture 3: Development of the Truss Equations.

Step 4: Derive the element stiffness matrix, k .


Relate nodal forces to the nodal displacements through
definition of internal tension, T.
f1x = T and f2 x = +T Same concept as with
our spring element.
The sense of the
nodal loads is
d2 x d1x governed by the
T = A x =AE
L reference frames x
direction.

AE 1 1 d1x f1x
= Element equations.
L 1 1 d2 x f2 x

MECH 420: Finite Element Applications


Lecture 3: Development of the Truss Equations.
3.3 Transformation of Vectors in
Two Dimensions.
The displacement, d, at a node
point can be communicated using
any reference frame.
When we assemble truss elements, +
we will have different elemental x
directions for each element.
Need to represent the nodal d and
f vectors in a common frame.
A direction, i, can be formed from
a combination of basis vectors of
another frame.

MECH 420: Finite Element Applications


Lecture 3: Development of the Truss Equations.
3.3 (contd)
d1x C S d1x
=
d1 y S C d1 y

R
C S
r1 = , r2 =
S C

ri i rj = ( C 2 + S 2 ) = 1 for i = j
ri i rj = ( CS SC ) = 0 for i j

R 1 = RT The planar rotation operator is an


orthonormal (orthogonal) matrix.

MECH 420: Finite Element Applications


Lecture 3: Development of the Truss Equations.
3.4 (Forming the) Global Stiffness
Matrix. y
Up until now we have ignored the
diy and d jy nodal displacements.
Clearly a truss node must displace
normal to an elements axis. y
We augment/pad the element
equations to account for the
transverse degrees of freedom. x
The truss assembly will have to x
ensure that all transverse motions are
restrained.
Whether or not this occurs can be
detected by the determinant of the
stiffness matrix (singular?).

MECH 420: Finite Element Applications


Lecture 3: Development of the Truss Equations.

0 d1x f1x

We already 1 0 1
defined R as a 0 d1 y f1 y
AE 0 0 0
transformation =
L 1 0 1
0 d2 x f2 x
from the
elemental frame 0 0 0 0 d f
2y 2y
to the global
frame. d1x f1x
d1x f1x
d1 y RT2 x 2 02 x 2 d1 y f1 y RT2 x 2
02 x 2 f1 y
= T and =
2x 2x2
d 0 R 2x2
d 2 x f 2 x 02 x 2 RT2 x 2 f 2 x
d 2 y f 2 y
d
2 y f2 y

C S 0 0
RT2 x 2 0
02 x 2 S C 0
T = = Eq. (3.4.15) of Logan.
02 x 2 RT2 x 2 0 0 C S

0 0 S C

MECH 420: Finite Element Applications


Lecture 3: Development of the Truss Equations.

Since T is block diagonal and is formed by a submatrix that is


orthogonal:
1
T =T T

Each set of element equations can be written in terms of the


global reference frame by using the T for each element:

{ }
T f = k {T d }

f = T k Td
T

f =k d
Element stiffness matrix
in terms of global
coordinates (no carat).
MECH 420: Finite Element Applications
Lecture 3: Development of the Truss Equations.

The change in coordinate system does not change the generic


sense of the element stiffness matrix.

C2 CS C 2 CS
Assembly
AE CS S2 CS S 2 N
= K = k (e)
(e)
k
L C 2 CS C2 CS e =1

CS S 2
CS S2

Notes:
C and S are particular to each element.
The rotation takes the global frame into alignment with the element
frame (+ clockwise about z axis protruding from the page).

MECH 420: Finite Element Applications


Lecture 3: Development of the Truss Equations.

Steps 5 and 6: Assemble Global Equations, apply boundary


conditions, solve.
Boundary conditions are likely in terms of global displacement values at
each node, dix and diy .
3.5 Computation of Stress for a Bar in the xy Plane.
Steps 7 and 8: Recovery and Interpretation.
Any piece of information (stress and strain) can be recovered from the
nodal displacements.
[ 1 1] 1x
E d
(1) =
L1 d 2 x
d2 x
{d d Ny }
E
= [ 1 1]
T
F = KD d1 y d2 x (2)
d3 x
1x L2

E d( N 1) x
(N )
= [ 1 1]
LN d Nx
MECH 420: Finite Element Applications
Lecture 3: Development of the Truss Equations.
Element strains and stresses are in terms of local displacements
(displacement along the elements axial direction).
The nodal displacements in the specific element frame can always be
recaptured from the orthogonal transformation, T.
d1x 0 d1x
C S 0

d1 y S C 0 0 d1 y
=
d2 x 0 0 C S d2 x
0 0

S C d 2 y
d 2 y
d1x
d1x C
S 0 0 d1 y See equations (3.5.6) thru
=
d 2 x 0
0 C S d 2 x (3.5.8) pg.# 79.
d 2 y
C S 0 0
T* = A shorthand matrix
0 0 C S operator. Not a true linear
transformation.
MECH 420: Finite Element Applications
Lecture 3: Development of the Truss Equations.
3.6 Solution of a Plane Truss.
Problem 3.20 E = 1 106 psi
A = 5 in 2
Boundary conditions aligned L = 100 in
with a global reference frame. y x
Applied loads also easily
expressed in global coordinate
system x
y

C (1)2 C (1) S (1) C (1)2 C (1) S (1) C (2)2 C (2) S (2) C (2)2 C (2) S (2)

AE C (1) S (1) S (1)2 C (1) S (1) S (1)2 AE C (2) S (2) S (2)2 C (2) S (2) S (2)2
k (1) = k (2) =
L1 C (1)2 C (1) S C (1)2 C (1) S (1) L2 C (2)2 C (2) S C (2)2 C (2) S (2)
(1) (1) (2) (2)
C S S (1)2 C (1) S (1) S (1)2 C S S (2)2 C (2) S (2) S (2)2

( 5 in ) 1 10
2 6lbf

AE
= in 2
= 35.36
kip
2L 2 100 in in

MECH 420: Finite Element Applications


Lecture 3: Development of the Truss Equations.

Orientation of each element calculated independently:


C (1) = cos (1) = cos(45 ) = 0.707
S (1) = cos (1) = sin(45 ) = 0.707

C (2) = cos (2) = cos(135 ) = 0.707


S (2) = cos (2) = sin(135 ) = 0.707

Step 5: Assembly:
f1x (1) F1x 0.5 0.5 0.5 0.5 0 0 d1x
F d1 y Note:

f1 y (1)
0.5 0.5 0.5 0 0

1y

f 2 x + f 2 x F2 x
(1) (2)
kip 1.0 0.0 0.5 0.5 d 2 x K is symmetric
(1) = = 35.36 d 2 y and singular.
f + f (2)
F in 1.0 0.5 0.5

2 y 2 y 2 y

f3 x (2) F3 x 0.5 0.5 d3 x


(2) SYM 0.5 d3 y
f 3 y F3 y

MECH 420: Finite Element Applications


Lecture 3: Development of the Truss Equations.

Step 6: Apply boundary conditions and solve.


d1x = d1 y = d3 x = d3 y = 0.0 F2 y = 10.0 kip F2x = 0.0 kip
Reducing the assembled system equations
F1x 0.5 0.5 0.5 0.5 0 0 d1x
F d1 y
0.5 0.5 0.5 0 0

1y

F2 x kip 1.0 0.0 0.5 0.5 d 2 x


F = 35.36 d 2 y
in 1.0 0.5 0.5

2 y

F3 x 0.5 0.5 d3 x
SYM 0.5 d3 y
F3 y

We solve directly for the remaining state variables:

F2 x kip 1.0 0.0 d 2 x d 2 x 0.0


F = 35.36 d =
2y in 0.0 1.0 d 2 y 2 y 0.283

MECH 420: Finite Element Applications

You might also like