Exam1 Sol PDF
Exam1 Sol PDF
Exam1 Sol PDF
a) Provide a precise weak statement for the problem and based on this give the final
expressions for the element stiffness matrix and load vector.
b) For each element, compute the element force vector Fe and element stiffness
e
matrix K and then provide the assembled F and K.
c) Apply the essential boundary conditions and compute the finite element nodal
displacements.
d) Provide expressions for the finite element stress in each element.
Possibly useful formulas: Using standard notation, the basis functions in x-coordinate
1 e
Ne = x2 − x x − x1e .
Le
for a two-node linear element are:
Based on this weak form, the element stiffness matrix and load vector are
=K e EA∫ BT Bdx
= , fe ∫ bN
T
dx + N T P |concentrated
load
location
1 1
N 1 = ( x − 2)( x − 3) −( x − 1)( x − 3) ( x − 1)( x − 2)
2 2
dN 1 5 3
B1 = = x− 4 − 2x x −
dx 2 2
Therefore, the element stiffness matrix is
2
5 5 5 3
x − x − ( 4 − 2 x ) x − x −
2 2 2 2
3 3
( 4 − 2x) ( 4 − 2 x ) x − dx
3
E1 A1 ∫ B1 B1dx =
K1 = 2 × 106 ∫ •
T 2
1 1
2
2
3
• • x−
2
2.3333 −2.6667 0.333
= 10 × −2.6667 5.333 −2.6667
6
1 x=B
10 ( x 1)( x 3) dx ( x 1)( x 3) x 3 13.333
1 1 3.333
( x − 1) ( x − 2) ( x − 1) ( x − 2)
2 2
For element 2:
1 −1 7 1
51 1 −1
0.05( x − 1)(210 ) ( −1=
1) dx 106 ∫
5 5
= ∫ B 2 B 2 E 2 A= ∫ ( x − 1)dx
T
K2
1
( x)dx
2 1 3 4 −1
3 3 2
1.5 −1.5
= 106 ×
−1.5 1.5
1 1 u= 6.4 × 10−5
EB 2 {d 2 } =
σ 2 ( x) = 2 × 107 −
4
=−640 Pa
2 2 u2 = 0
Solution:
Using the four basis functions,
1 1
N1 (ξ ,η ) = (1 − ξ )(1 − η ), N 2 (ξ ,η ) = (1 + ξ )(1 − η )
4 4
1 1
N 3 (ξ ,η ) = (1 + ξ )(1 + η ), N 4 (ξ ,η ) = (1 − ξ )(1 + η )
4 4
the mapping is
4
3
x =∑ x j N j =3 N 3 + 3 N 4 = (1 + η )
j =1 2
4
1
y = ∑ y j N j = N 2 + N3 = (1 + ξ )
j =1 2
Explain why the above condition is satisfied (and not only for this element!).
Solution:
According to the basis function defined in problem 2, we have
4
1 1 1 1
∑ N (ξ ,η )=
i =1
i
4
(1 − η − ξ + ξη ) + (1 − η + ξ − ξη ) + (1 + η + ξ + ξη ) + (1 + η − ξ − ξη )
4 4 4
1 1 1 1
= + + + =1
4 4 4 4
Generally, the above condition is always satisfied. Let’s consider a constant function
f (ξ ,η ) = a . By FEM interpolation, this function value at any point inside the element is
Nnodes Nnodes
=f i
=i 1 =i 1
∑ a N (ξ ,η ) a
= ∑ N i (ξ ,η )
Since f is a constant field, and we use at least linear interpolation, this interpolation
should be exact, i.e. f= f= a .
Therefore, we have
Nnodes Nnodes
a= a
=i 1 =i 1
∑ N i (ξ ,η ) ⇒ ∑ N i (ξ ,η ) = 1
The key idea here is that the FEM basis functions used form a complete set of functions –
can represent both rigid body motion and `constant strain’. It is the completeness of the
FE basis that makes this equation true.
Problem 4 (10 points) – For the 4-node element Ω shown in the figure below, compute
2
∂N
at the point .
(ξ ,η ) = (0.5,0.5) in Ω
∂y
a) Provide a precise weak statement for the problem and based on this give the
final expressions for the element stiffness matrix and load vectors.
b) Compute the element stiffness matrix. This will require a number of tasks
summarized below
1. Using the basis functions, give an expression in terms of (ξ ,η ) for
∂x e ∂y e
the Jacobian matrix
∂ξ ∂ξ
J =
e
∂x e ∂y e
∂η ∂η
2. Compute expressions for the determinant of the Jacobian and the
inverse of this Jacobian matrix.
3. Compute the matrix of the derivates of the shape functions wrt natural
coordinates, i.e.
∂N1e ∂ N e2 ∂ N 3e ∂ N e4
∂ξ ∂ξ ∂ξ ∂ξ
e
∂N1 ∂ N e2 ∂ N 3e ∂ N e4
∂η ∂η ∂η ∂η
∂T e
∂x
Give an expression for the B matrix relating and the
e
4.
∂T e
∂y
nodal temperatures.
5. Use Gauss quadrature to compute the stiffness matrix and load vector.
These will also be your global stiffness and load vector.
d) Apply essential BCs in the final system of equations and solve for the
unknown nodal temperatures.
1 1
e) Compute the heat flux components at the Gauss point ( ,− ).
3 3
Solution:
(a) The weak statement for this problem takes the form:
Find a function T ( x, y ) ∈ H 1 ( Ω ) , with =
T ( s ) T ( s ), s ∈ ΓT such that for any w ∈ H 1 ( Ω )
with w = 0 on ΓT the following equations holds:
∫ Ω
k ∇T ⋅ ∇=
wd Ω ∫Ω
fwd Ω − ∫ qwd Γ
Γq
∑ xi Ni =
x= 1 −η
2 N1 + 2 N 2 =
i =1
4
1
y = ∑ yi N i = N 2 + N 3 = (1 + ξ )
i =1 2
The Jacobian matrix can be obtained as
∂x ∂y
∂ξ ∂ξ 1
=J = 0 2
∂x ∂y
∂η ∂η −1 0
1
2. The determinant of the Jacobian is | J |= .
2
The inverse of this Jacobian matrix is
1
−1 0 − 0 −1
= 2=
2 0
J 2
1 0
∂N1 ∂N 2 ∂N 3 ∂N 4
∂ξ ∂ξ ∂ξ ∂ξ 1 η − 1 1 − η 1 + η −η − 1
=
3. ∂N
1 ∂N 2 ∂N 3 ∂N 4 4 ξ − 1 −1 − ξ 1 + ξ 1 − ξ
∂η ∂η ∂η ∂η
∂N1 ∂N 2 ∂N 3 ∂N 4
∂ξ 1 1 − ξ 1+ ξ −1 − ξ ξ −1
−1 ∂ξ ∂ξ ∂ξ
4. B J= 2 η − 1 2 1 − η 2 1 + η −2 η + 1
∂N1 ∂N 2 ∂N 3 ∂N 4 4 ( ) ( ) ( ) ( )
∂η ∂η ∂η ∂η
∑∑ k ( B B | J |)
1 1 2 2
= ∫ kB=
Bd Ω ∫ ∫ | d ξ dη
kB B | J=
T T T
K WiW j
Ω −1 −1 ξ ξi=
= ,η ηi
=i 1 =i 1
(c) 1.
T3 17.3529
T = 22.6471 C
4
1 1
(e) The heat flux components at the Gauss point ,− are
3 3
∂T T1 10
− k ∂x T
0.1057 0.3943 −0.3943 −0.1057 10
q= = − kB 1 2 = − 4 −0.7887 0.7887 0.2113 −0.2113 17.3529
− k ∂T T3
1
ξ = ,η = −
∂y
3 3
T4 22.6471
16.9434
= W /m
4.4752