FE Summary Examples
FE Summary Examples
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
●
„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
[ ][ ] [ ]
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
[ ]
●
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
[ ]
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
[ ] [ ]
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
[ ]
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
[ ] [ ]
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
[ ]
●
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
+ϱ
EA
L
1
)−−
EA
L (
2
L
EA
L
EA
L
L
+ϱ
[]
0
F
ϱ +
ϱ
= F FL
EA
Tension, BC, spring: example a
[ ]
●
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
[ ]
●
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
[ ]
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
[ ]
●
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
[ ]
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
[ ]
●
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
[ ]
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
[ ][ ]
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
●
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
[ ][ ]
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
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
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.
[]
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.
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.
[ ][ ][ ]
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
[ ]
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
[ ]
●
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
[ ]
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
[ ]
●
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
[ ]
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
[ ]
●
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
4
Frame: example
[ ]
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
[ ]
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
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
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
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
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
[ ][ ] [ ]
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
[ ][ ] [ ]
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
2
2 3 3
4
1
●
Node #2 and #4 can rotate!
not a proper discretization!
Frame: example
[ ][ ] [ ]
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
[ ][ ] [ ]
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.
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)
+ − + −
+ − + −
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