Anda di halaman 1dari 17

Boundary Value Problems Finite Element Method

Introduction to the Finite Element Method


Lecture 02

P.S. Koutsourelakis

pk285@cornell.edu
369 Hollister Hall

August 30 2010
Last Updated: August 30, 2010

pk285@conell.edu Cornell University


Lecture 02
Boundary Value Problems Finite Element Method

Boundary Value Problem


Example
Consider a bar of length L, cross-sectional area A which is held fixed on the
left end and pulled with a force F on the right end and stretched with a
distributed force b(x) along its length. Whats will be the deformation of the
bar u(x) at each point x?

b(x)
A F
u(x)
L
Boundary Value Problem (BVP)
d 2 u(x)
EA + b(x) = 0 ∀x ∈ (0, L)
dx 2
du
with boundary conditions: u(0) = 0, EA dx |x=L =F
pk285@conell.edu Cornell University
Lecture 02
Boundary Value Problems Finite Element Method

Boundary Value Problem

Boundary Value Problem (BVP)

d 2 u(x)
EA + b(x) = 0 ∀x ∈ (0, L)
dx 2

with boundary conditions: u(0) = 0, E A du dx |x=L = F


Although it is straightforward to derive a closed-form solution
(right?) things are not necessarily so if:
elastic modulus varies E(x)
cross-sectional area varies A(x)
if we are considering two or three dimensional versions with
arbitrary boundary shapes/conditions.
we need a general computational method that is able to produce
efficiently, accurate solutions of BVPs.

pk285@conell.edu Cornell University


Lecture 02
Boundary Value Problems Finite Element Method

Boundary Value Problem

Boundary Value Problem (BVP)

d 2 u(x)
EA + b(x) = 0 ∀x ∈ (0, L)
dx 2
Approximate derivatives with finite differences, i.e.:
du
dx = limh→0 u(x+h/2)−u(x−h/2)
h
≈ u(x+h/2)−u(x−h/2)
h for 0 < h << 1

pk285@conell.edu Cornell University


Lecture 02
Boundary Value Problems Finite Element Method

Boundary Value Problem


Boundary Value Problem (BVP)

d 2 u(x)
EA + b(x) = 0 ∀x ∈ (0, L)
dx 2

define N grid points xi = i h where h = NL and let ui = u(xi )


if h is small enough then I can approximate:

d 2 u(x) ui+1 − 2ui + ui−1


2
|x=xi ≈
dx h2

substitute in Equation (1) for x = xi , ∀i to obtain N algebraic


equations w.r.t N unknowns ui .
this is the the Finite Difference Method
pk285@conell.edu Cornell University
Lecture 02
Boundary Value Problems Finite Element Method

Finite Difference Method (FDM)

d 2 u(x)
E + b(x) = 0 ∀x ∈ (0, L)
dx 2
⇓ (discretization)
ui+1 − 2ui + ui−1
EA + b(xi ) = 0 ∀xi , i = 1, 2, . . . , N
h2

Observe that in FDM we approximate the PDE itself


FDM is still used in a wide range of problems and we will use it in
time-dependent problems to discretize time-derivatives.
In the Finite Element Method (FEM) we approximate the solution
of an equivalent form of the PDE.

pk285@conell.edu Cornell University


Lecture 02
Boundary Value Problems Finite Element Method

Finite Element Method

Figure: The FREU(E)D roadmap to Finite Elements

pk285@conell.edu Cornell University


Lecture 02
Boundary Value Problems Finite Element Method

Finite Element Method

Roadmap to FEM approximations:


1 Function Spaces: We are going to define where we are going to to be
looking for solutions, i.e. which function space
2 Reformulate: We are going to reformulate the original problem, i.e.
the PDE and Boundary Conditions - Weak form.
3 Equivalence: We are going to show that this new form is actually
equivalent to the original, i.e. any solution of the former is a solution of
the latter and vice versa.
4 Unique: We are going to show that the solution is unique.
5 Equivalence 2 : We are going to look at some equivalent forms which
can be considered as special cases.
6 Discretization: We are going to propose ways to discretize all these
equivalent forms.
pk285@conell.edu Cornell University
Lecture 02
Boundary Value Problems Finite Element Method

Function Spaces

Function Spaces
Since we are going to be approximating solutions of PDEs i.e. functions,
it makes sense to recap some of the basic function spaces and their
properties. If Ω is an open subset of R (or Rn ) in general, then:
C(Ω) contains all functions defined on Ω which are continuous.
C k (Ω) contains all functions defined on Ω which have continuous
derivatives up to order k .
Cbk (Ω) same as C k (Ω) plus the function is bounded
L2 (Ω) contains all functions defined on Ω which are square integrable
i.e.: Z
u 2 (x) dx < +∞

H 1 (Ω) contains all functions in L2 whose derivatives are also square


integrable i.e.: Z
|du/dx|2 (x) dx < +∞

pk285@conell.edu Cornell University


Lecture 02
Boundary Value Problems Finite Element Method

Function Spaces

Function Spaces

Boundary Value Problem (BVP)

d 2 u(x)
EA + b(x) = 0 ∀x ∈ (0, L)
dx 2
du
with boundary conditions: u(0) = u0 , EA |
dx x=L
=F
We are going to look for solutions in the trial or candidate solution space
S: Z
du(x) 2
S = {u(x) : (0, L) → R|u(0) = u0 , E A dx < +∞}
(0,L) dx
Observe that:
u ∈ S satisfy exactly only one of the two boundary conditions. This
BC is called essential.
u ∈ S have finite strain energy!
u ∈ S are continuous and bounded.

pk285@conell.edu Cornell University


Lecture 02
Boundary Value Problems Finite Element Method

Function Spaces

Function Spaces
Boundary Value Problem (BVP)
d 2 u(x)
EA + b(x) = 0 ∀x ∈ (0, L)
dx 2
with boundary conditions: u(0) = u0 , E A du |
dx x=L
=F
We are going to look for solutions in the space S:
Z
du(x) 2
S = {u(x) : (0, L) → R|u(0) = u0 , E A dx < +∞}
(0,L) dx
Observe that:
The space S is much larger than what the PDE and BC would
imply.
Even though a 2nd order derivative of appears in the BVP, we are
looking for solutions that are guaranteed to have a 1st order
derivative.
Even though a “force” BC must be satisfied, we are looking for
solutions that are not a priori guaranteed to satisfy it.
we have RELAXED already the original problem
pk285@conell.edu Cornell University
Lecture 02
Boundary Value Problems Finite Element Method

Weak Forms

Finite Element Method

Roadmap to FEM approximations:


1 We are going to define where we are going to to be looking for solutions,
i.e. which function space
2 We are going to reformulate the original problem, i.e. PDE and BC.
3 We are going to show that this new form is actually equivalent to the
original, i.e. any solution of the former is a solution of the latter and vice
versa.
4 We are going to show that the solution is unique.
5 If that wasn’t enough, we are going to look at some equivalent forms
which can be considered as special cases.
6 We are going to propose ways to discretize all these equivalent forms.
pk285@conell.edu Cornell University
Lecture 02
Boundary Value Problems Finite Element Method

Weak Forms

Weak Forms
Boundary Value Problem (BVP)
d 2 u(x)
EA + b(x) = 0 ∀x ∈ (0, L) (1)
dx 2
du
with boundary conditions: u(0) = u0 , EA |
dx x=L
=F
We are going to look for solutions in the space S:
Z
du
S = {u(x) : (0, L) → R|u(0) = u0 , E A| |2 (x) dx < +∞}
(0,L) dx
An arbitrary u ∈ S will not satisfy Equation (1) exactly (unless it is the
solution) and in general there will be a residual R(x)
d 2 u(x)
R(x) = EA + b(x)6= 0
dx 2

du
There will also be a residual R(L) = E A |
dx x=L
− F because u ∈ S do
not a priori satisfy this BC
pk285@conell.edu Cornell University
Lecture 02
Boundary Value Problems Finite Element Method

Weak Forms

Weak Forms
How can we make those residuals zero?
d 2 u(x) du
R(x) = EA + b(x) R(L) = F − EA |x=L
dx 2 dx

(Bubnov)-Galerkin approach

Figure: Boris Galerkin (1871-1945)

pk285@conell.edu Cornell University


Lecture 02
Boundary Value Problems Finite Element Method

Weak Forms

Weak Forms
How can we make those residuals zero?
d 2 u(x) du
R(x) = EA + b(x) R(L) = F − EA |x=L
dx 2 dx
(Bubnov)-Galerkin or Weighted Residual approach: define another set
of functions called weight functions w ∈ W:
Z
dw(x) 2
W = {w(x) : (0, L) → R|w(0) = 0, E A| | dx < +∞}
(0,L) dx

Find u ∈ S such that for all w ∈ W:


 RL
 0 w(x)R(x) dx = 0
 du

w(L)R(L) = w(L) F − EA |
dx x=L
=0

Note that the residual is not zero in the STRONG sense i.e.
R(x) = 0 ∀x but the condition is enforced WEAKLY as above.
pk285@conell.edu Cornell University
Lecture 02
Boundary Value Problems Finite Element Method

Weak Forms

Weak Forms
After some mainipulation (see Integration by parts)

STRONG form:  2
 EA d dxu(x)

2 + b(x) = 0
 u(0) = u 0
 F = EA du |
dx x=L

WEAK form:
 RL RL
EA du
0
dw
dx dx
dx = 0
b(x)w(x)dx + w(L)F ∀w ∈ W
u∈S
Does the weak form remind you of something?
Principle of Virtual Work
The necessary and sufficient condition for a system in equilibrium is that
the work done by internal forces should be equal to the work done by
externals loads for any kinematically acceptable virtual displacement
pk285@conell.edu Cornell University
Lecture 02
Boundary Value Problems Finite Element Method

Weak Forms

Weak Forms
WEAK form:
 RL RL
EA du
0
dw
dx dx
dx = 0
b(x)w(x)dx + w(L) F ∀w ∈ W
u∈S

Principle of Virtual Work:


a virtual kinematically acceptable displacement w(x) is one that
does not violate displacement boundary conditions u(0) = u0 , i.e.
w(0) = 0.
Work of internal forces:
Z Z
du dw
Wint (w) = A σu (x) ǫw (x) dx = A E dx
| {z } | {z } dx dx
stress from u strain from w

Work of external forces:


Z
Wext (w) = b(x) w(x) dx + F w(L)

pk285@conell.edu Cornell University


Lecture 02

Anda mungkin juga menyukai