0% found this document useful (0 votes)
44 views7 pages

Mixed Crack Propagate

This document describes a boundary value problem and numerical method for simulating mixed-mode crack propagation over time. It presents the governing equations, including boundary and initial conditions, for modeling the displacement field u and phase field z representing the crack shape. A semi-discrete scheme is used to approximate the time derivatives. The weak form of the displacement equation is derived using divergence theorem. Finite elements are then used to discretize the spatial domain and solve the coupled equations for u and z simultaneously at each time step.
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)
44 views7 pages

Mixed Crack Propagate

This document describes a boundary value problem and numerical method for simulating mixed-mode crack propagation over time. It presents the governing equations, including boundary and initial conditions, for modeling the displacement field u and phase field z representing the crack shape. A semi-discrete scheme is used to approximate the time derivatives. The weak form of the displacement equation is derived using divergence theorem. Finite elements are then used to discretize the spatial domain and solve the coupled equations for u and z simultaneously at each time step.
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/ 7

1 Boundary Problem

This is Boundary problem for mixed mode (I+II) crack propagation time dependent:
 � �


 0 = − div (1 − z)2 σ[u] + f (x, t) x∈Ω





 � �

 ∂z γ(x)
 α2 = � div(γ(x)∇z) − z + σ[u] : e[u](1 − z) x∈Ω
∂t � + (1.1)





 + B.C.






+ I.C.

Where we have Boundary Condition (B.C.):





 u = g(x, t) on ΓD





σ[u]n = 0 on ΓN (1.2)





 ∂z

 =0 on Γ
∂n
and Iniatial Condition (I.C.)

u(x, 0) = u0 (x) x∈Ω
(1.3)
z(x, 0) = z0 (x) ∈ [0, 1] x ∈ Ω

Suppose that:

1. Ω is a bounded two dimensional domain and using a Cartesian coordinate, where x =


(x1 , x2 ) ∈ R2 is parallel to the plate and x3 the coordinate perpedicular to the plate.

2. u(x, t) represent the small anti-plane displacement at the position x ∈ Ω and time t ≥ 0

3. The function z(x, t) is called the phase filed for crack shape, where variable z(x, t)
satisfies 0 ≤ z(x, t) ≤ 1 in Ω and represents the crack shape.

4. f (x, t) is external force

5. g(x) is a given anti-plane displacement on the boundary ΓD

6. q(x) is a given boundary load in the x3 -direction on the boundary ΓN

Remark 1
In this case, we want to simulation mode III crack growth with condition; f (x, t) = 0,
α2 = 10−3 , and � = 10−3 .

2
2 Semi Discretization in Time and Semi-discrete Scheme
To solve boundary problem in equation (1.1), we have to do a semi discretization in time.
Δt > 0 time increment,where tk := kΔt
0 = t 0 < t1 < t2 < · · ·
So, approximate with time disretization:


 uk (x) ≈ u(x, kΔt)



g (x)
k
≈ g(x, kΔt) (2.1)


z̃ k − z k−1
 ∂z

 ≈ = zt (x, kΔt)
Δt ∂t
Using semi dicretization in time (equation 2.1), we can define equation (1.1) and than can
find find uk (x) and z̃ k (x).
� �
0 = − div (1 − z k−1 )2 σ[uk ]
� � � �
z̃ k − z k−1 γ(x) k
α2 = � div γ(x)∇z̃ k − z̃ + σ[uk−1 ] : e[uk−1 ](1 − z̃ k )
Δt � +

z := max(z̃ , z )
k k k−1
(2.2)
uk = g k (x, t) on ΓD
σ[uk ]n = 0 on ΓN
∂ z̃ k
=0 on Γ
∂u
Because equation (2.2) is a couple equation, so to find uk (x) and z̃ k (x) we have to solve
equation (2.2) simultaneously.

3 Weak Form and Finite Element Method


3.1 Variable uk (x)
Weak Form
V u := {w ∈ H 1 (Ω) | w = 0 on ΓD }
We choose the test function w ∈ V u
for all w ∈ V u
� �
0 = − div (1 − z k−1 )2 σ[uk ]
� � � ��
k−1 2
0 = − div (1 − z ) σ[u ] w dx
k
Ω
� � � ��
k−1 2
0 = − ∂j (1 − z ) σij [u ] wi dx
k
(3.1)
Ω

3
To solve equation (3.1), recall Divergence Theorem:
Divergence Theorem

� � �
(∂j fj )g dx = fj nj g ds − fj (∂j g) dx
Ω ∂Ω Ω

so we get:
�� � � � � � �
0 = − (1 − z k−1 )2 σij [uk ] nj wi ds − (1 − z k−1 )2 σij [uk ] (∂j wi ) dx
Ω
� � ∂Ω � � � �
0 = (1 − z k−1 )2 σij [uk ] (∂j wi ) dx − (1 − z k−1 )2 σij [uk ] nj wi ds (3.2)
Ω ∂Ω

or equation (3.2):
� �
(1 − z k−1 )2 σij [uk ](∂j wi ) dx = (1 − z k−1 )2 σij [uk ]nj wi ds (3.3)
Ω ∂Ω

If we use Boundary Condition (equation 1.2) and z k−1 = 0 on ΓN , so equation (3.3) can be:

(1 − z k−1 )2 σij [uk ](∂j wi ) dx = 0 (3.4)
Ω

if σ ∈ R3×3
sym

σij [uk ](∂j wi ) = σ[uk ] : (∇T w)


1� �
σij [uk ](∂j wi ) = σ[uk ] : ∇T w + ∇w T (3.5)
2
= σ[uk ] : e[w]

Using equation (3.3), equation (3.6) is writen:



(1 − z k−1 )2 σ[uk ] : e[w] dx = 0 (3.6)
Ω

Remark 2
Because the material is homogeneous and isotropic. According Hooke’s Law, σ[u] can
be defined:

σ[u] = λ(div u)I + 2µe[u]


σ[u] = λ(tr(e[u]))I + 2µe[u]

where:

1. div u = tr(e[u]) = I : e[u]

2. λ and µ are Lame’s constanta

4
By Remark 2, equation (3.6) be:
� � �
(1 − z k−1 )2 (λ(div uk )I + 2µe[uk ]) : e[w] dx = 0
� Ω
� �
(1 − z k−1 )2 λ(div uk )I : e[w] + 2µe[uk ] : e[w] dx = 0
�Ω � �
(1 − z k−1 )2 λ(div uk )(div w) + 2µe[uk ] : e[w] dx = 0
Ω

So can be define:
� � �
aku (u, w) = (1 − z k−1 )2 λ(div uk )(div w) + 2µe[uk ] : e[w] dx
Ω

luk (w) = 0 ds

Find uk ∈ V u (g k ) (k = 1, 2, · · · ) such that:

au (uk , wh ) = luk (w) (∀ w ∈ V u ) (3.7)

Finite Element Method


We choose Vh ⊂ V = H � (Ω), dim(Vh ) < ∞
We define Vhu := Vh ∩ V u .
Let ghk ∈ Vh be an approximation of g k and we define that Vhu (ghk ) := Vhu + ghk
Find ukh ∈ Vhu (ghk ) (k = 1, 2, · · · ) such that

aku (ukh , wh ) = luk (wh ) (∀ wh ∈ Vhu )

where:

� � �
aku (ukh , wh ) = (1 − z k−1 )2 λ(div ukh )(div wh ) + 2µe[ukh ] : e[wh ] dx
Ω
luk (wh ) = 0

3.2 Variable z̃ k (x)


Weak Form
V z := {v ∈ H 1 (Ω) | v = 0 on Γ1N }
We choose the test function v ∈ V

5
for all v ∈ V
� �
∂z γ(x)
α2 = z + σ[u] : e[u](1 − z)
� div(γ(x)∇z) −
∂t �
� k � � �
z̃ − z k−1 γ(x) k
α2 = � div(γ(x)∇z̃ ) −
k
z̃ + σ[u ] : e[u ](1 − z̃ )
k−1 k−1 k
Δt �
Here, we set that v = min(z̃ k , 0)
� � k �
� � �
− z k−1
z̃ γ(x) k
α2 v dx = � div(γ(x)∇z̃ k ) − z̃ + σ[uk−1 ] : e[uk−1 ](1 − z̃ k ) v dx
Ω Δt Ω �
� � k � � � � � �
z̃ − z k−1 γ(x) k �
α2 v dx = � div(γ(x)∇z̃ k ) v dx − z̃ v dx
Ω Δt � Ω
�� � � ٠� �� �
I II
� � �
+ σ[u k−1
] : e[uk−1 ](1 − z̃ k ) v dx (3.8)
�٠�� �
III

from equation (3.8), we can solve:


1. Part I
� � � � � �
� div(γ(x)∇z̃ ) v dx = �
k
div(γ(x)∇z̃ k ) v dx
Ω Ω
�� � � � �
= � (γ(x)∇z̃ k ) · n v ds − (γ(x)∇z̃ k ) · ∇v dx
∂Ω Ω
�� � ∂ z̃ k � � �
= � γ(x)v ds − (γ(x)∇z̃ ) · ∇v dx
k
(3.9)
∂Ω ∂n Ω

∂ z̃ k
According Boundary Condition that = 0, so equation (3.9) be:
∂n
� � � �
� div(γ(x)∇z̃ ) v dx = −�
k
(γ(x)∇z̃ k ) · ∇v dx (3.10)
Ω Ω

2. Part II
� �
γ(x) �1�
z̃ v dx =
k
γ(x)z̃ k v dx (3.11)
Ω � � Ω
3. Part III
� � �
σ[uk−1 ] : e[uk−1 ](1 − z̃ k ) v dx
Ω

if σ[u k−1
] = λ(div u k−1
)I + 2µe[uk−1 ], so, the above equation be:
� �� �� ��
(λ(div uk−1 )I + 2µe[uk−1 ]) : e[uk−1 ] 1 − z̃ k v dx
� �� Ω
�� ��
λ(div uk−1 )I : e[uk−1 ] + 2µe[uk−1 ] : e[uk−1 ] 1 − z̃ k v dx
٠� �� � �2 �� ��
λ(div uk−1 )2 + 2µ e[uk−1 ] 1 − z̃ k v dx (3.12)
Ω

6
Using equation (3.10), (3.11), and (3.12), equation (3.9) can be:
� � k � �
z̃ − z k−1 1�
α2 v dx = −� (γ(x)∇z̃ k ) · ∇v dx − γ(x)z̃ k v dx
Ω Δt Ω
� � Ω
�� � �2 �� ��
+ λ(div uk−1 )2 + 2µ e[uk−1 ] 1 − z̃ k v dx
Ω
� � �
�Δt Δt
(z̃ k − z k−1 )v dx = − (γ(x)∇z̃ ) · ∇v dx −
k
γ(x)z̃ k v dx
Ω Ω α2 Ω �α2

Δt �� � �2 �� ��
+ λ(div uk−1 )2 + 2µ e[uk−1 ] 1 − z̃ k v dx
Ω α2

Separate the above equation:


� � � �
Δt � γ(x) � k−1 2
� � ��
k−1 2 �Δt
1+ + λ(div u ) + 2µ e[u ] z̃ v dx +
k
(γ(x)∇z̃ k ) · ∇v dx
Ω α2 � Ω α2
� �
Δt � � �2 �
= λ(div uk−1 )2 + 2µ e[uk−1 ] v dx + z k−1 v dx
Ω α2 Ω

where:
� � � �
Δt � γ(x) � � �2 �� �Δt
akz (z, v) = 1+ + λ(div uk−1 )2 + 2µ e[uk−1 ] z̃ k v dx + (γ(x)∇z̃ k ) · ∇v dx
Ω α2 � Ω α2
� �
Δt � k−1 2

k−1 2
� �
lz (v) =
k
λ(div u ) + 2µ e[u ] v dx + z k−1 v dx
Ω α2 Ω

Find z k ∈ V z (k = 1, 2, 3, · · · ) such that:

az (z k , vh ) = lzk (v) (∀ v ∈ V z ) (3.13)

Finite Element Method


We choose Vh ⊂ V = H � (Ω), dim(Vh ) < ∞
We define Vhz := Vh ∩ V z .
Let ghk ∈ Vh be an approximation of g k and we define that Vhz (ghk ) := Vhz + ghk
Find zhk ∈ Vhz (ghk ) (k = 1, 2, · · · ) such that

akz (zhk , vh ) = lzk (vh ) (∀ vh ∈ Vhz )

and we define that:


� � � �
Δt � γ(x) k−1 2 k−1 2
� �Δt
akz (zhk , vh ) = 1+ + (λ div u ) + 2µ(e[u ]) z̃h vh dx +
k
(γ(x)∇z̃hk ) · ∇vh dx
Ω α2 � Ω α2
� �
Δt � � �2 �
lzk (vh ) = λ(div uk−1 )2 + 2µ e[uk−1 ] vh dx + zhk−1 vh dx
Ω α2 Ω

7
4 Program of FreeFEM++
In this problem we set the boundary condition as Ω = (−1, 1)×(−1, 1) and ΓD = {(x1 , x2 )|x1 ∈
(−1, 1), x2 = ±1}(see fig. 1). The boundary condition for u is given by g = (1, 1, 0) for x ∈ ΓD
and t ≥ 0.

Γ D1 Γ D1
(1, 1) (1, 1)
(−1, 1) (−1, 1)
G2 G2

Σ Σ
ΓN G3 G1
ΓN ΓN G3 G1
ΓN
Ω Ω
G4 G4

(−1, −1) (1, −1) (−1, −1) (1, −1)

Γ D2 Γ D2
(a) Case I (b) Case II

Figure 1: Computational domain of mode I + II crack growth with ΓD1 (g = (u1 , u2 )) and
ΓD2 (g = (−u1 , −u2 )) (a) and ΓD1 (g = (−u1 , u2 )) and ΓD2 (g = (u1 , −u2 )) (b)

where:

� � � � � �
x 1 1
G1 : = t + (1 − t) (t : 0 → 1)
y 1 −1
� � � � � �
x −1 1
G2 : = t + (1 − t) (t : 0 → 1)
y 1 1
� � � � � �
x −1 −1
G3 : = t + (1 − t) (t : 0 → 1)
y −1 1
� � � � � �
x 1 −1
G4 : = t + (1 − t) (t : 0 → 1)
y −1 −1

You might also like