0% found this document useful (0 votes)
32 views

FE Summary Examples

The document provides an overview of using the finite element method (FEM) to solve structural analysis problems. It discusses key steps in the FEM process including discretization, assembly of elementary stiffness matrices and load vectors, application of boundary conditions through various techniques like condensation, and calculation of secondary variables. Examples are presented to illustrate tension problems in bars with different element configurations and application of boundary conditions through condensation. The summary focuses on the high-level FEM workflow and techniques discussed in the document.
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
0% found this document useful (0 votes)
32 views

FE Summary Examples

The document provides an overview of using the finite element method (FEM) to solve structural analysis problems. It discusses key steps in the FEM process including discretization, assembly of elementary stiffness matrices and load vectors, application of boundary conditions through various techniques like condensation, and calculation of secondary variables. Examples are presented to illustrate tension problems in bars with different element configurations and application of boundary conditions through condensation. The summary focuses on the high-level FEM workflow and techniques discussed in the document.
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 96

Finite Element Method

for
Civil Engineers
Overview of Lecture #9

Using FEM to solve structures

Boundary conditions

Calculation of secondary variables

Examples
Solving a structure

Discretization

Elementary stiffness matrices

Elementary load vectors

Compilation

Boundary conditions

Solving the matrix equation

Calculating secondary variables
Meaning of stiffness matrix

State (equilibrium) equation: K v=q

K v =K e j ej
K ej ?

physical meaning: vector of nodal forces so that: v= e j
 special nodal forces that generates special displs

mathematical meaning: ‚j’-th column of K

‚j’-th column of K: set of nodal forces


to reach special displacements:
- unit displacement at the j-th DOF,
- zero displacement elsewhere
Meaning of load vector

real load (e.g. distributed) nodal forces


„equivalent”: the nodal forces (must) have the same
effect as the original loads
Boundary conditions

In general: BC is a prescription for the displacement

Mostly: „displ=0”, but can be different


2 + 1 strategies
– unknown displ replaced by unknown reaction force
 condensation
– springs
– (+1: no nodes at supports)
BC: no nodes at supports

Used mostly for bars (i.e., frames)

Illustration:


Advantage (?): slightly fewer DOF

Disadvantage: complicates the calculations (e.g.,
various versions of FE, compilation, etc.), not for
plates, shells, etc.
Boundary conditions

In general: BC is a prescription for the displacement

Mostly: „displ=0”, but can be different


2 + 1 strategies
– unknown displ replaced by unknown reaction force
 condensation
– springs
– (+1: no nodes at supports)
BC: displ-reaction replacement

„inner” and „outer” nodes (DOFs)

inner:
unknown (unprescribed) displ,
known (prescribed) force

outer:
(partly) unknown (unprescribed) force, (i.e., reaction)
known (prescribed) displ

v=
[]
vi
vo
q=
[ ]
qi
q o+r o
BC: displ-reaction replacement

the matrix equation, partitioned:

[ ][ ] [ ]
K ii K io v i
=
qi
K oi K oo v o q o +r o
unknown


still ‚n’ unknowns: partly displ, partly reaction


2 matrix equations:
K ii v i + K io v o= qi
K oi v i + K oo v o= qo + r o
BC: displ-reaction replacement

from first equation: nodal forces
equivalent to
K ii v i + K io v o= qi
prescribed
displacements
K ii v ii =q i− K io v o

~
q i=qi −K io v o

thus, the equation:
K v =~
ii q
i i v i = K ii ~
−1
qi

the reactions, from 2nd equation:

K oi v i+ K oo v o= qo + r o r o= K oi v i + K oo v o− qo
BC: problem condensation

Generalization of replacing unknown displs and
reactions  condensation

Assume, 1 single displ DOF ‚o’ is prescribed:

[ ][ ] [ ]
K aa k ao K ab v a qa
k oa k oo k ob v o = q o + r o
K ba k bo K bb v b qb

After rearranging:

[ ][ ] [ ] [ ]
K aa K ab v a
K ba K bb v b
=
qa
qb

k ao
k bo
vo
displs []
unknown v a
v b

smaller by 1 DOF  condensed


BC: problem condensation

Generalization of replacing unknown displs and
reactions  condensation

Then the reaction from the middle equation:

[ ][ ] [ ]
K aa k ao K ab v a qa
k oa k oo k ob v o = q o + r o
K ba k bo K bb v b qb
r o =k oa v a+ k oo v o + k ob v b− q o

More than single DOF prescribed  repeat the
condensation procedure
Boundary conditions

In general: BC is a prescription for the displacement

Mostly: „displ=0”, but can be different


2 + 1 strategies
– unknown displ replaced by unknown reaction force
 condensation
– springs
– (+1: no nodes at supports)
BC: elastic supports

If a DOF is restrained  no (or small) displacement

A possible model for the support: an elastic spring
attached to the restrained node

If the node is fixed  apply large spring stiffness

When the springed node displaced  a spring force
arises (acting on the node, negative)
BC: elastic supports

The equation system:

[ ][ ] [ ]
K aa k ar K ab v a qa
kra k rr k rb v r = qr −ϱ v r
K ba k br K bb v b qb

Can be rearranged:

[ ][ ] [ ]
K aa k ar K ab v a qa
kra k rr + ϱ krb v r = q r
K ba k br K bb v b qb

 the spring stiffness must simply be added to the


stiffness matrix!
Secondary variables

„primary variable”: nodal displacements

„secondary variable”: all others

Reaction force, either: r o= K oi v i + K oo v o – q o
or: from springs

Strain:
ε ( x )= [ ε ( x ) ] = B ( x ) v e

Normal force: element by
σ ( x )= [ N ( x ) ] = D B ( x ) v e element
 small-size

Stress: problems
N ( x)
σ ( x )=
A
Tension, BC, condensation: example a

Tensioned bar with a single 2-node element



stiffness matrix: we have calculated earlier

load vector: just 1 nodal force

[ ]

equation to solve: EA EA


L
EA
L
L v1
EA v 2
L
[][]
=
0
F


BC: left node is restrained
v 1= 0
Tension, BC, condensation: example a

Tensioned bar with a single 2-node element



condensation:
- delete first row
- delete first column at left-hand-side

[ ]
EA EA


L
EA
L
L v1
EA v 2
L
[][]
=
0
F [ ]
EA
L
[ v 2 ]= [ F ]


The unknown displ (at right end)
FL
v 2=
EA
Tension, BC, condensation: example b

Tensioned bar with two 2-node elements



compilation from elementary stiffness matrices

[ ] [ ]
EA EA 0 0 0
− 0
L/2 L/2 EA EA
0 −
K 1= EA EA K 2= L/2 L /2
− 0
L/2 L/2 EA EA
0 −
0 0 0 L/2 L /2


vector: just 1 nodal force
Tension, BC, condensation: example b

Tensioned bar with two 2-node elements



equation to solve:

[ ]
2 EA 2 EA
− 0

[][]
L L
v1 0
2 EA 4 EA 2 EA
− − v2 = 0
L L L
v F
2 EA 2 EA 3
0 −
L L


BC: left node is restrained v 1= 0
Tension, BC, condensation: example b

Tensioned bar with two 2-node elements



condensation:
- delete first row
- delete first column at left-hand-side

[ ] [ ]
4 EA 2 EA FL


L
2 EA
L
L v2
2 EA v 3
L
=
[][]
0
F []
v2
v3
=
2 EA
FL
EA

Unknown displs: at middle and right end
Tension, BC, spring: example a

Tensioned bar with a single 2-node element



similar to earlier example a

[ ]

equation to solve:
EA EA
+ϱ −


L
EA
L
L v1
EA v 2
L
=
[][]
0
F

[) ] [ ]

from which:
EA EA F

[]( v1
v2
=
EA
L

EA
L
1

)−−
EA
L (
2
L
EA
L
EA
L
L

[]
0
F
ϱ +
ϱ
= F FL
EA
Tension, BC, spring: example a

Tensioned bar with a single 2-node element

[ ]

nodal displacements: F

[]
v1
v2
ϱ
= F FL
ϱ + EA


if spring stiffness is increasing to infinity:
FL
v 1= 0 v 2=
EA

if spring stiffness is decreasing to zero:
rigid-body displ (real motion!, no equilibrium!)
Frame: example

2D frame

p=13 kN/m

3m EA=18000 kN
2
EI =24000 kNm

4m 4m
Frame: example

Discretization

4 nodes, three 2-node (Euler) elements

2
2 3 3

4
1

Nodal displacements:
T
v = [u 1 w 1 φ 1 y u 2
all
w2 φ 2 y u3 w 3 φ 3 y u 4 w4 φ4 y ]
Frame: example

Elementary stiffness matrix

[ ]

Clamped-clamped, local coordinate system:

EA EA
0 0 − 0 0
L L
12 EI 6 EI 12 EI 6 EI
0 − 0 − −
L3 L2 L3 L2
6 EI 4 EI 6 EI 2 EI
0 − 2 0
L L L2 L
K e= EA EA
− 0 0 0 0
L L
12 EI 6 EI 12 EI 6 EI
0 − 3 2
0 3 2
L L L L
6 EI 2 EI 6 EI 4 EI
0 − 2 0
L L L2 L
Frame: example

Elementary stiffness matrix



1-2 bar:

[ ]
6000 0 0 −6000 0 0
0 10667 −16000 0 −10667 −16000
12 0 −16000 32000 0 16000 16000
K =
e −6000 0 0 6000 0 0
0 −10667 16000 0 10667 16000
0 −16000 16000 0 16000 32000

e.g.:
EA 18000
EA = 18000 = =6000
L 3
L=3
12 EI 12⋅24000
E I = 24000 3
= 3
=10667
L 3
Frame: example

Elementary stiffness matrix


Pinned-clamped, local coordinate system:

[ ]

EA EA
0 0 − 0 0
L L
3 EI 3 EI 3 EI
0 0 0 − 3 − 2
L3 L L
0 0 0 0 0 0
K e= −
EA
0 0
EA
0 0
L L
3 EI 3 EI 3 EI
0 − 3 0 0
L L3 L2
3 EI 3 EI 3 EI
0 − 2 0 0
L L2 L
Frame: example

Elementary stiffness matrix



2-3 bar:

[ ]
4500 0 0 −4500 0 0
0 1125 0 0 −1125 −4500
23 0 0 0 0 0 0
K =
e −4500 0 0 4500 0 0
0 −1125 0 0 1125 4500
0 −4500 0 0 4500 18000


e.g.:
EA = 18000 EA 18000
= =4500
L=4 L 4
3 EI 3⋅24000
E I = 24000 3
= 3
=1125
L 4
Frame: example

Elementary stiffness matrix

[ ]

Clamped-clamped, local coordinate system:

EA EA
0 0 − 0 0
L L
12 EI 6 EI 12 EI 6 EI
0 − 0 − −
L3 L2 L3 L2
6 EI 4 EI 6 EI 2 EI
0 − 2 0
L L L2 L
K e= EA EA
− 0 0 0 0
L L
12 EI 6 EI 12 EI 6 EI
0 − 3 2
0 3 2
L L L L
6 EI 2 EI 6 EI 4 EI
0 − 2 0
L L L2 L
Frame: example

Elementary stiffness matrix



3-4 bar:

[ ]
3600 0 0 −3600 0 0
0 2304 −5760 0 −2304 −5760
34 0 −5760 19200 0 5760 9600
K =e −3600 0 0 3600 0 0
0 −2304 5760 0 2304 5760
0 −5760 9600 0 5760 19200


e.g.:
EA = 18000 EA 18000
= =3600
L= √ 4 2 + 3 2 = 5 m
L 5
12 EI 12⋅24000
E I = 24000 3
= 3
=2304
L 5
Nodal forces

for distributed loads:
L
q e =∫ N ( x) p ( x ) dx
T

0
Nodal forces, 2D frame

Meaning of elements of a nodal force vector:

[]
... nodal longitudinal force at node ‚i’
... nodal transverse force at node ‚i’
... nodal moment at node ‚i’
qe= nodal longitudinal force at node ‚j’
...
... nodal transverse force at node ‚j’
... nodal moment at node ‚j’


„force” can be „moment”, too!
Nodal forces, 2D frame

for UDL, clamped-clamped:

[ ]
p x L /2 nodal longitudinal force at node ‚i’
p z L/ 2 nodal transverse force at node ‚i’
2
− p z L /12 nodal moment at node ‚i’
qe=
p x L /2 nodal longitudinal force at node ‚j’
p z L/ 2 nodal transverse force at node ‚j’
2
p z L / 12 nodal moment at node ‚j’
Nodal forces, 2D frame

for UDL, pinned-clamped:

[]
p x L/2 nodal longitudinal force at node ‚i’
3 p z L/8 nodal transverse force at node ‚i’
0 nodal moment at node ‚i’
q e=
p x L/2 nodal longitudinal force at node ‚j’
5 p z L/8 nodal transverse force at node ‚j’
2
p z L /8 nodal moment at node ‚j’
Frame: example

Elementary vector of nodal forces



only the middle element is loaded, pinned-clamped:
T
q =[ 0 0 0 0 0 0 ]
12

[ ][ ]
e
T
q =[ 0 0 0 0 0 0 ]
34
e p x L/2 0
3 p z L/ 8 19.5
23 0 0
qe = =
p x L/2 0

with:
5 p z L/ 8 32.5
p x =0 2 26
p z L /8
p z =13
L=4
Frame: example

Stiffness matrix (of the whole structure)



So far: each elementary stiffness matrix in local
coordinate system

Before compilation: transform to global
Vector of nodal forces (for the whole structure)

So far: each elementary load vector in local coordinate
system

Before compilation: transform to global


transformation = rotation
Rotation of a vector

a 3D vector
g

in global XYZ system: v

in local xyz system: vl

local-to-global rotation:
g l
v =T v

with: T =[ ex ey ez]


where e x ey ez
- unit vectors parallel with the local xyz
- expressed by the global XYZ coordinates
Rotation of a vector

the rotation matrix (3×3) orthonormal:

T −1
T =T


therefore:
g l
v =T v
T g T l l
T v =T T v = v
Rotation of a vector

In 2D (x-z plane):

T =[ ex ey ez] T =[ e x e z ]

[ ]
X
α e z= −sin α
ez ex cos α

Z
z
x
e x=[ ]cos α
sin α

T= [
cosα −sin α
sin α cosα ]
Frame: example

Vector of nodal forces



only the middle element is loaded, pinned-clamped:
T
q =[ 0 0 0 0 0 0 ]
12

[ ][ ]
e
T
q =[ 0 0 0 0 0 0 ]
34
e p x L/2 0
3 p z L/ 8 19.5
force vector on 23 0 0
qe = =
node ‚i’ of element p x L/2 0
5 p z L/ 8 32.5
force vector on 2
p z L /8 26
node ‚j’ of element
Frame: example

Vector of nodal forces

to be transformed e.g.: [ ][
cos α −sin α ⋅ p x L/ 2
sin α cos α 3 p z L/ 8 ]

[ ][ ]
p x L/2 0
3 p z L/ 8 19.5
force vector on 23 0 0
qe = =
node ‚i’ of element p x L/2 0
5 p z L/ 8 32.5
force vector on 2
p z L /8 26
node ‚j’ of element
Frame: example

Vector of nodal forces

to be transformed e.g.: [ ][
cos α −sin α ⋅ p x L/ 2
sin α cos α 3 p z L/ 8 ]

but the angle is zero  practically: no transformation

[ ][ ]
p x L/2 0
3 p z L/ 8 19.5
23 0 0
qe = =
p x L/2 0
5 p z L/ 8 32.5
p z L 2 /8 26
Transformation of vector of nodal displ.

Elementary vector of nodal displacements

[]
ui
displ vector on wi
node ‚i’ of element φiy
ve=
uj
displ vector on wj
node ‚j’ of element φ jy
Transformation of vector of nodal displ.

Elementary vector of nodal displacements

to be transformed e.g.:
[ ⋅
sin α cos α w i][ ]
cos α −sin α u i

[]
ui
displ vector on wi
node ‚i’ of element φiy
ve=
uj
displ vector on wj
node ‚j’ of element φ jy
Transformation of vector of nodal displ.

Elementary vector of nodal displacements



transformation of an elementary nodal displ vector:

[ ][ ][ ]
Ui cos α −sin α 0 0 0 0 ui
Wi sin α cos α 0 0 0 0 wi
Φ iy 0 0 1 0 0 0 φiy
=
Uj 0 0 0 cos α −sin α 0 uj
Wj 0 0 0 sin α cos α 0 w
j
Φ jy 0 0 0 0 0 1 φ jy

g ̂ l
v = T ij v e
e
Transformation of stiffness matrix

Elementary stiffness matrix – makes connection
between „end forces” and „end displacements”

Both „forces” and „displacements”: in local system

Rotation of the displ vector: l ^ −1 g ^ T g
v = T ij v e =T ij v e .
e


The end forces can be written:
K e T̂ ij v e = q e
l l l l T g l
K e v e= q e

Pre-multiply:
T̂ ij K e T̂ ij v e = T̂ ij q e ̂ l ̂T
l T g l g g
T ij K e T ij v e = q e

Stiffness matrix transform: ^ g l ^T
K =T ij K e T ij
e
Transformation of stiffness matrix

Elementary matrix transform: ^
g l ^T
K =T ij K e T ij
e

Can be written:

[ ][ ][ ][ ]
T T T
T 0 ⋅ K e , ii K e , ij T 0 T K e ,ii T T K e , ij T
⋅ T = T T
0 T K e , ji K e , jj 0 T T K e , ji T T K e , jj T

 it is practical to perform the matrix multiplication on


3×3 blocks
Frame: example

Stiffness matrix transform



1-2 bar:
α =+ 90 deg

[ ]
0 1 0 0 0 0
−1 0 0 0 0 0
T̂ 1−2= 0 0 1 0 0 0
0 0 0 0 1 0
0 0 0 −1 0 0
0 0 0 0 0 1
Frame: example

Stiffness matrix transform

[ ]

1-2 bar:
6000 0 0 −6000 0 0
0 10667 −16000 0 −10667 −16000
K 12 = 0 −16000 32000 0 16000 16000
e
−6000 0 0 6000 0 0
0 −10667 16000 0 10667 16000
0 −16000 16000 0 16000 32000

[ ]
10667 0 −16000 −10667 0 −16000
0 6000 0 0 −6000 0
K 12 g
= −16000 0 32000 16000 0 16000
e
−10667 0 16000 10667 0 16000
0 −6000 0 0 6000 0
−16000 0 16000 16000 0 32000
Frame: example

Stiffness matrix transform



2-3 bar:
α=0

[ ]
1 0 0 0 0 0
0 1 0 0 0 0
T̂ 2−3= 0 0 1 0 0 0
0 0 0 1 0 0
0 0 0 0 1 0
0 0 0 0 0 1
 practically: no transformation
Frame: example

Stiffness matrix transform


2-3 bar:

[ ]

4500 0 0 −4500 0 0
0 1125 0 0 −1125 −4500
K 23 = 0 0 0 0 0 0
e
−4500 0 0 4500 0 0
0 −1125 0 0 1125 4500
0 −4500 0 0 4500 18000

[ ]
4500 0 0 −4500 0 0
0 1125 0 0 −1125 −4500
K 23g = 0 0 0 0 0 0
e
−4500 0 0 4500 0 0
0 −1125 0 0 1125 4500
0 −4500 0 0 4500 18000
Frame: example

Stiffness matrix transform



3-4 bar:
ta n α = 3 / 4

[ ]
0.8 −0.6 0 0 0 0
0.6 0.8 0 0 0 0
̂
T 3−4 = 0 0 1 0 0 0
0 0 0 0.8 −0.6 0
0 0 0 0.6 0.8 0
0 0 0 0 0 1
Frame: example

Stiffness matrix transform

[ ]

3-4 bar:
3600 0 0 −3600 0 0
0 2304 −5760 0 −2304 −5760
K 34 = 0 −5760 19200 0 5760 9600
e
−3600 0 0 3600 0 0
0 −2304 5760 0 2304 5760
0 −5760 9600 0 5760 19200

[ ]
3133 622 3456 −3133 −622 3456
622 2771 −4608 −622 −2771 −4608
34 g 3456 −4608 19200 −3456 4608 9600
Ke =
−3133 −622 −3456 3133 622 −3456
−622 −2771 4608 622 2771 4608
3456 −4608 9600 −3456 4608 19200
Frame: example

Compilation of the stiffness matrix


23 g
Ke 2 3
1 2 3 4
2
1
12 g
Ke 1 2 3
2
1
34 g
3 Ke 3 4
2
4 3

4
Frame: example

Compilation of the stiffness matrix

[ ]
all
K =
10667 0 −16000 −10667 0 −16000 0 0 0 0 0 0
0 6000 0 0 −6000 0 0 0 0 0 0 0
−16000 0 32000 16000 0 16000 0 0 0 0 0 0
−10667 0 16000 10667+4500 0+0 16000+0 −4500 0 0 0 0 0
0 −6000 0 0+0 6000+1125 0+0 0 −1125 −4500 0 0 0
−16000 0 16000 16000+0 0+0 32000+0 0 0 0 0 0 0
0 0 0 − 4500 0 0 4500+3133 0+622 0+3456 −3133 −622 3456
0 0 0 0 −1125 0 0+622 1125+2771 4500−4608 −622 −2771 −4608
0 0 0 0 − 4500 0 0+3456 4500−4608 18000+19200 −3456 4608 9600
0 0 0 0 0 0 −3133 −622 −3456 3133 622 −3456
0 0 0 0 0 0 −622 −2771 4608 622 2771 4608
0 0 0 0 0 0 3456 −4608 9600 −3456 4608 19200
Frame: example

Compilation of the stiffness matrix

[ ]
10667 0 −16000 −10667 0 −16000 0 0 0 0 0 0
0 6000 0 0 −6000 0 0 0 0 0 0 0
−16000 0 32000 16000 0 16000 0 0 0 0 0 0
−10667 0 16000 15167 0 16000 −4500 0 0 0 0 0
0 −6000 0 0 7125 0 0 −1125 −4500 0 0 0
all −16000 0 16000 16000 0 32000 0 0 0 0 0 0
K = 0 0 0 −4500 0 0 7633 622 3456 −3133 −622 3456
0 0 0 0 −1125 0 622 3896 −108 −622 −2771 −4608
0 0 0 0 −4500 0 3456 −108 37200 −3456 4608 9600
0 0 0 0 0 0 −3133 −622 −3456 3133 622 −3456
0 0 0 0 0 0 −622 −2771 4608 622 2771 4608
0 0 0 0 0 0 3456 −4608 9600 −3456 4608 19200
Frame: example

Boundary conditions

p=13 kN/m

3m EA=18000 kN
2
EI =24000 kNm

4m 4m


supports: u 1=w 1=φ 1y = u 4= w 4 =0

reactions: r 1x , r 1z , m r1y , r 4x , r 4z
Frame: example

Boundary conditions

reactions are added to nodal forces at supports:

[ ][ ] [ ]
10667 0 −16000 −10667 0 −16000 0 0 0 0 0 0 u1 0+r 1 x
0 6000 0 0 −6000 0 0 0 0 0 0 0 w1 0+ r1 z
−16000 0 32000 16000 0 16000 0 0 0 0 0 0 φ1 y 0+mr 1 y
−10667 0 16000 15167 0 16000 −4500 0 0 0 0 0 u2 0
0 −6000 0 0 7125 0 0 −1125 −4500 0 0 0 w2 19.5
−16000 0 16000 16000 0 32000 0 0 0 0 0 0 φ2 y 0
=
0 0 0 −4500 0 0 7633 622 3456 −3133 −622 3456 u3 0
0 0 0 0 −1125 0 622 3896 −108 −622 −2771 −4608 w 3 32.5
0 0 0 0 −4500 0 3456 −108 37200 −3456 4608 9600 φ 3 y 26
0 0 0 0 0 0 −3133 −622 −3456 3133 622 −3456 u 4 0+r 4 x
0 0 0 0 0 0 −622 −2771 4608 622 2771 4608 w 4 0+r 4 z
0 0 0 0 0 0 3456 −4608 9600 −3456 4608 19200 φ 4 y 0
Frame: example

Solution

considering zero displacements at supports,
then rearranging:

[ ] [ ][ ]
−10667 0 −16000 0 0 0 0 0+r 1x 10667 0 −16000 0 0
0 −6000 0 0 0 0 0 0+r 1z 0 6000 0 0 0

[]
16000 0 16000 0 0 0 0 0+m r 1 y −16000 0 32000 0 0

[]
u2
15167 0 16000 −4500 0 0 0 0 −10667 0 16000 0 0 u1
0 7125 0 0 −1125 − 4500 0 w2 19.5 0 −6000 0 0 0
φ2 y w1
16000 0 32000 0 0 0 0 0 −16000 0 16000 0 0 φ1 y
−4500 0 0 7633 622 3456 3456 u3 = −
0 0 0 −3133 −622
0 u4
0 −1125 0 622 3896 −108 −4608 w 3 32.5 0 0 0 −622 −2771
0 −4500 0 3456 −108 37200 9600 φ 3 y 26 0 0 0 −3456 4608 w 4
φ4 y
0 0 0 −3133 −622 −3456 −3456 0+r 4x 0 0 0 3133 622
0 0 0 −622 −2771 4608 4608 0+r 4 z 0 0 0 622 2771
0 0 0 3456 −4608 9600 19200 0 0 0 0 −3456 4608
Frame: example

Solution

considering zero displacements at supports,
then rearranging:

[ ] [ ][ ]
−10667 0 −16000 0 0 0 0 0+r 1x 10667 0 −16000 0 0
0 −6000 0 0 0 0 0 0+r 1z 0 6000 0 0 0

[]
16000 0 16000 0 0 0 0 0+m r 1 y −16000 0 32000 0 0

[]
u2
15167 0 16000 −4500 0 0 0 0 −10667 0 16000 0 0 u1
0 7125 0 0 −1125 − 4500 0 w2 19.5 0 −6000 0 0 0
φ2 y w1
16000 0 32000 0 0 0 0 0 −16000 0 16000 0 0 φ1 y
−4500 0 0 7633 622 3456 3456 u3 = −
0 0 0 −3133 −622
0 u4
0 −1125 0 622 3896 −108 −4608 w 3 32.5 0 0 0 −622 −2771
0 −4500 0 3456 −108 37200 9600 φ 3 y 26 0 0 0 −3456 4608 w 4
φ4 y
0 0 0 −3133 −622 −3456 −3456 0+r 4x 0 0 0 3133 622
0 0 0 −622 −2771 4608 4608 0+r 4 z 0 0 0 622 2771
0 0 0 3456 −4608 9600 19200 0 0 0 0 −3456 4608
Frame: example

Solution
after deleting the zero term:

][]

[
15167 0 16000 −4500 0 0 0 0
0 7125 0 0 −1125 −4500 0 19.5

[]
16000 0 32000 0 0 0 0 0
−4500 0 0 7633 622 3456 3456 u2 0
0 −1125 0 622 3896 −108 −4608 w 2 32.5
φ2 y 26
0 −4500 0 3456 −108 37200 9600 =
0 0 0 3456 −4608 9600 19200 u 3 0
−10667 0 −16000 0 0 0 0 w3 0+r 1 x
0 −6000 0 0 0 0 0 φ3 y 0+r 1 z
φ4 y
16000 0 16000 0 0 0 0 0+mr 1 y
0 0 0 −3133 −622 −3456 −3456 0+r 4 x
0 0 0 −622 −2771 4608 4608 0+r 4 z
Frame: example

Solution
partitioning:

][]

[
15167 0 16000 −4500 0 0 0 0
0 7125 0 0 −1125 −4500 0 19.5

[]
16000 0 32000 0 0 0 0 0
−4500 0 0 7633 622 3456 3456 u2 0
0 −1125 0 622 3896 −108 −4608 w 2 32.5
φ2 y 26
0 −4500 0 3456 −108 37200 9600 =
0 0 0 3456 −4608 9600 19200 u 3 0
−10667 0 −16000 0 0 0 0 w3 0+r 1 x
0 −6000 0 0 0 0 0 φ3 y 0+r 1 z
φ4 y
16000 0 16000 0 0 0 0 0+mr 1 y
0 0 0 −3133 −622 −3456 −3456 0+r 4 x
0 0 0 −622 −2771 4608 4608 0+r 4 z
Frame: example

Solution
from first 7 rows (i.e., equations): nodal displs

][]

[
15167 0 16000 −4500 0 0 0 0
0 7125 0 0 −1125 −4500 0 19.5

[]
16000 0 32000 0 0 0 0 0
−4500 0 0 7633 622 3456 3456 u2 0
0 −1125 0 622 3896 −108 −4608 w 2 32.5
φ2 y 26
0 −4500 0 3456 −108 37200 9600 =
0 0 0 3456 −4608 9600 19200 u 3 0
−10667 0 −16000 0 0 0 0 w3 0+r 1 x
0 −6000 0 0 0 0 0 φ3 y 0+r 1 z
φ4 y
16000 0 16000 0 0 0 0 0+mr 1 y
0 0 0 −3133 −622 −3456 −3456 0+r 4 x
0 0 0 −622 −2771 4608 4608 0+r 4 z

T
v=10 ⋅[ −3.8115 5.8647 1.9058 −6.0702 16.528 0.82054 4.6491 ]
−3
Frame: example

Solution

Displaced shape:

T
v=10 ⋅[ −3.8115 5.8647 1.9058 −6.0702 16.528 0.82054 4.6491 ]
−3
Frame: example

Solution
from last 5 rows (i.e., equations): reactions

][]

[
15167 0 16000 −4500 0 0 0 0
0 7125 0 0 −1125 −4500 0 19.5

[]
16000 0 32000 0 0 0 0 0
−4500 0 0 7633 622 3456 3456 u2 0
0 −1125 0 622 3896 −108 −4608 w 2 32.5
φ2 y 26
0 −4500 0 3456 −108 37200 9600 =
0 0 0 3456 −4608 9600 19200 u 3 0
−10667 0 −16000 0 0 0 0 w3 0+r 1 x
0 −6000 0 0 0 0 0 φ3 y 0+r 1 z
φ4 y
16000 0 16000 0 0 0 0 0+mr 1 y
0 0 0 −3133 −622 −3456 −3456 0+r 4 x
0 0 0 −622 −2771 4608 4608 0+r 4 z
Frame: example

Solution

from last 5 rows (i.e., equations): reactions

[ ][ ]
r1 x 10.164 kN (→)
r1 z −35.188 kN (↑)
mr 1 y = −30.492 kNm(↷)
p=13 kN/m r4 x −10.164 kN (←)
3m EA=18000 kN
r4 z −16.812 kN (↑)
2
EI =24000 kNm

4m 4m
Frame: example

Handling internal hinges



Discretization #1

2
2 3 3

4
1

Nodal displacements:
T
v = [u 1 w 1 φ 1 y u 2
all
w2 φ 2 y u3 w 3 φ 3 y u 4 w4 φ4 y ]
Frame: example

Handling internal hinges



Discretization #2

2
2 3 3

4
1

Nodal displacements:
T
v = [u 1 w 1 φ 1 y u 2
all
w2 φ 2 y u3 w 3 φ 3 y u 4 w4 φ4 y ]
Frame: example

Solution #1 vs. #2
Solution #1 Solution #2

1-2 element: fixed-fixed fixed-pinned

2-3 element: pinned-fixed fixed-fixed
● φ 2y at: 1-2 2-3

 the calculation details are modified accordingly


 calculation steps are the same
 solution is essentially the same
Frame: example

Solution #1 vs. #2

unsupported nodal displacements

v=[ u2 w 2 φ 2y u 3 w 3 φ 3y φ 4y ]
T


solution #1
T
v=10−3⋅[ −3.8115 5.8647 1.9058 −6.0702 16.528 0.82054 4.6491 ]


solution #2
v = 10 ⋅[ − 3.8115 4.6491 ]
T −3
5.8647 − 5.1312 − 6.0702 16.528 0.82054
Frame: example

Handling internal and external hinges



Discretization #3

2
2 3 3

4
1

Nodal displacements:
T
v = [u 1 w 1 φ 1 y u 2
all
w2 φ 2 y u3 w 3 φ 3 y u 4 w4 φ4 y ]
Frame: example

State equation (after condensation)



Discretization #3

[ ][ ] [ ]
7167 0 0 −4500 0 0 0 u2 0
0 7125 0 0 −1125 −4500 0 w 19.5
2
0 0 0 0 0 0 0 φ2 y 0
−4500 0 0 7011 1452 1728 0 u3 = 0
0 −1125 0 1452 2790 2196 0 w 32.5
3
0 −4500 0 1728 2196 32400 0 φ 3 y 26
0 0 0 0 0 0 0 φ4 y 0
Frame: example

State equation (after condensation)



Discretization #3

[ ][ ] [ ]
7167 0 0 −4500 0 0 0 u2 0
0 7125 0 0 −1125 −4500 0 w 19.5
2
0 0 0 0 0 0 0 φ2 y 0
−4500 0 0 7011 1452 1728 0 u3 = 0
0 −1125 0 1452 2790 2196 0 w 32.5
3
0 −4500 0 1728 2196 32400 0 φ 3 y 26
0 0 0 0 0 0 0 φ4 y 0


The matrix is singular
 φ 2y and φ 4y cannot be calculated, can be arbitrary
Frame: example

Handling internal and external hinges



Discretization #3

2
2 3 3

4
1

Node #2 and #4 can rotate!
 not a proper discretization!
Frame: example

State equation (after condensation)



Discretization #3

[ ][ ] [ ]
7167 0 0 −4500 0 0 0 u2 0
0 7125 0 0 −1125 −4500 0 w 19.5
2
0 0 0 0 0 0 0 φ2 y 0
−4500 0 0 7011 1452 1728 0 u3 = 0
0 −1125 0 1452 2790 2196 0 w 32.5
3
0 −4500 0 1728 2196 32400 0 φ 3 y 26
0 0 0 0 0 0 0 φ4 y 0


The problem can be solved by eliminating the zero
rows/columns
Frame: example

State equation (after further condensation)



Discretization #3

[ ][ ] [ ]
7167 0 −4500 0 0 u2 0
0 7125 0 −1125 −4500 w 2 19.5
−4500 0 7011 1452 1728 u3 = 0
0 −1125 1452 2790 2196 w 32.5
3
0 −4500 1728 2196 32400 φ3 y 26


The solution:
v =10 [ −3.812 5.865 −6.070 16.528 0.821 ]
T −3

 same solution
Frame: example

Secondary variables

nodal displacements calculated (in global system)

nodal displacements for each element

transformation back to local coordinate system

from displs  strains

from strains  stresses (stress resultants)

element by element !
Frame: example

Secondary variables

e.g., bar 3-4, local displacements
−3 T
= 10 ⋅[ − 6.0702 4.6491 ]
34 g
v 16.528 0.82054 0 0

[ ]
0.8 −0.6 0
34
T = 0.6 0.8 0

[ ]
34 T
34 T 0 34 g 0 0 1
v = 34 T v
0 T
T
v = [ 5.061 4.649 ]
34
16.865 0.8205 0 0
Frame: example

Secondary variables

e.g., bar 3-4, local displacements  displ functions
T
v = [ 5.061 4.649 ]
34
16.865 0.8205 0 0

[ ]
x x
1− 0 0 0 0
5 5
3 2 1 3 3 1 3 2 1 3
N ( x)= 0 1− x+ x −x + x 2− x 3 0 x− x 0
50 250 10 50 5 250
3 3 2 3 3 2 6 3 2
0 − x− x 1− x + x 0 − x+ x 0
25 250 5 50 5 250

[ ] [ ]
u( x) 5.061−1.012 x
34 −3
u( x)= w (x ) = N ( x) v =10 ⋅ 16.865−0.8205 x−0.7658 x 2+ 0.05105 x 3
φ y (x ) 0.8205+ 1.5315 x−0.15315 x 2
Frame: example

Secondary variables

e.g., bar 3-4, local displacements  strains
T
v = [ 5.061 4.649 ]
34
16.865 0.8205 0 0

[ ]
1 1
− 0 0 0 0
5 5
B( x)=L N ( x)=
3 3 3 3 6 3
0 − x − + x 0 − + x 0
25 125 5 25 5 125

ε ( x)=
[ ]
ε x ( x)
κy(x )
34 −3
=B(x) v =10 ⋅
[
−1.012
1.5315−0.3063 x ]
Frame: example

Secondary variables

e.g., bar 2-3, local displacements  displ functions
T
v 23 g = 10 −3⋅[− 3.8115 5.8647 − 5.1312 − 6.0702 16.528 0.82054 ]


no transformation

final displ = displ from nodes + displ from load
Frame: example

Secondary variables

e.g., bar 2-3, local displacements  displ functions
T
v 23 g = 10 −3⋅[− 3.8115 5.8647 − 5.1312 − 6.0702 16.528 0.82054 ]


axial displ function: only from nodes

−3
u ( x )= 10 ⋅(− 3.8115− 0.5647 x )
Frame: example

Secondary variables

e.g., bar 2-3, local displacements  displ functions
T
v 23 g = 10 −3⋅[− 3.8115 5.8647 − 5.1312 − 6.0702 16.528 0.82054 ]


transverse from nodes
w n ( x )= 10 −3⋅( 5.8647+ 5.1312 x− 0.36110 x 2 − 0.063812 x 3 )

transverse from load (requires some calculation...)
w p ( x )= 10 − 3⋅( 0.36111 x 2− 0.18056 x 3 + 0.022569 x 4 )

finally:
−3
( 2 3
w ( x )= 10 ⋅ 5.8647 + 5.1312 x + 0⋅x − 0.24437 x + 0.022569
Frame: example

Secondary variables

e.g., bar 2-3, strain, stress, etc.

 the effect from nodal displacements and (transverse)


load must be summarized
Performance analysis:
beams and membranes
Clamped-clamped beam
y
p = 1.3 kN/m
2.5
x
5m

Maximum defection (mm)


2
p
y 1.5
0.24 m

z
1
membrane
0.5 Timoshenko
0.1 m Timoshenko reduced
Euler-Bernoulli
0
2 4 8 16 32 64 128 256 512
E = 80000 kN/m2
Horizontal resolution
ν=0
Performance analysis:
membranes
y y
p = 1.3 kN/m
p
z x
0.24 m

0.1 m 0.96 m
Performance analysis:
membranes
Normal stresses (σx)

[N/m2]
Performance analysis:
membranes
Normal stresses (σy)

[10-1 N/m2]
Performance analysis:
membranes
Shear stresses (τxy)

[10-1 N/m2]
Performance analysis:
membranes

a b
2(− , )
y
a b
1( , )
N2= ( −)( )
1 x 1 y
+
2 a 2 b
y N1=( )( )
1 x 1 y
+ +
2 a 2 b
2 2 2 2

x x

a b a b
3(− ,− ) 4 ( ,− )
2 2
(
2 2 N 3 = 1− x
2 a )( 1 y

2 b ) N 4= ( )(
1 x 1 y
+ −
2 a 2 b )

[ ]
b b b b
+y 0 − −y 0 − +y 0 −y 0
2 2 2 2
1 a a a a
B= L N= 0 +x 0 −x 0 − +x 0 − −x
ab 2 2 2 2
a b a b a b a b
+x +y −x − − y − + x − + y − −x −y
2 2 2 2 2 2 2 2
Performance analysis:
membranes
y y
u 2−u3 v 1−v2 u 1−u4 v 1−v 2
a b a b + +
2(− , ) 1( , ) b a b a
2 2 2 2

x x
[γ xy ]

a b a b u 2−u3 v 4 −v 3 u 1−u4 v 4 −v 3
3(− ,− ) 4 ( ,− ) + +
2 2 2 2 b a b a

[ ] [ ]
1 1 u 1−u2
0 − 0 0 0 0 0
a a a
B ( )
a b
, =
2 2
0
1
b
0 0 0 0 0 −
1
b ( ) ( )
a b a b
ε , =B , v e =
2 2 2 2
v 1 −v 4
b
1 1 1 1 u1−u4 v 1−v 2
0 − 0 0 − 0 +
b a a b b a
Performance analysis:
membranes
Shear strain distribution y
u 2−u3 v 1−v2 u 1−u4 v 1−v 2
+ +
b a b a

+ ++ + x
[γ xy ]

u 2−u3 v 4 −v 3 u 1−u4 v 4 −v 3
+ ++ + + +
b a b a
(pure shear: uniform strain)

+ − + −

+ − + −

(pure bending: alternating strain)


Performance analysis:
membranes
y

x
σx
σy

x
τxy

V
Summary

General formulation and terminology for FEM


Frames:
– coordinate transformation by elements
– boundary conditions: possibly different discretization

Bars vs. membrane:
– more or less like a Timoshenko beam
– secondary variables: still discontinuous strain and stress field

You might also like