Partial differential equations

Hoang Hai Ha

HCMUT-OISP
Email: [email protected]

Hoang Hai Ha (HCMUT) Partial differential equations 1 / 15

Outline

1 Introduction

2 Numerical methods for solving PDEs

Hoang Hai Ha (HCMUT) Partial differential equations 2 / 15

Definition 1.1
Partial differential equations (PDE) are mathematical equations
that describe relationships between unknown functions with
multiple variables and its partial derivatives.

Hoang Hai Ha (HCMUT) Partial differential equations 3 / 15

Definition 1.1
Partial differential equations (PDE) are mathematical equations
that describe relationships between unknown functions with
multiple variables and its partial derivatives.
The order of a PDE is the highest order of partial derivatives
occurring in the equation.

Partial differential equations arise in the study of some physical

phenomena such as: vibrating strings, mechanics of membrane, rods,
heat flow...

Hoang Hai Ha (HCMUT) Partial differential equations 3 / 15

Some typical examples of PDE

∂ 2u 2
2∂ u
1) = c : one dimensional wave equation.
∂t 2 ∂x 2

We find a function u(x, t) that satisfies this equation, herein, x : the

spatial variable and t : the time .

∂ 2u ∂ 2u ∂ 2u
 
2) 2 = c 2 + : two dimensional wave equation.
∂t ∂x 2 ∂y 2

∂u ∂ 2u
3) = c 2 2 : one dimensional heat equation.
∂t ∂x

Hoang Hai Ha (HCMUT) Partial differential equations 4 / 15

Some typical examples of PDE

∂ 2u ∂ 2u
 
∂u
3) = c2 + : two dimensional heat equation.
∂t ∂x 2 ∂y 2

A solution of this equation is the function u(t, x, y ), where

(x, y ) ∈ Ω ⊂ R2 and t ∈ [0, ∞).

∂ 2u ∂ 2u
4) + = 0: the Laplacian equation.
∂x 2 ∂y 2

∂ψ ℏ2 ∂ 2 ψ
5) iℏ =− + V (x)ψ: one dimensional Schrödinger
∂t 2m ∂x 2
equation.

Hoang Hai Ha (HCMUT) Partial differential equations 5 / 15

Numerical methods for solving PDEs

In this part, we aim to find numerical solutions of a one dimensional

heat equation via the finite difference method. We want to know the
temperature distribution of a uniform bar of length L with an
insulated lateral surface and no internal sources of heat, subject to
certain boundary and initial conditions.

Hoang Hai Ha (HCMUT) Partial differential equations 6 / 15

Let u(x, t) represent the temperature of the point x on the bar and
at the moment t. We consider the problem
∂ 2u 2
2∂ u
= c , 0 < x < L, t ≥ 0, (1)
∂t 2 ∂x 2
satisfying the Dirichlet boundary conditions:

u(0, t) = 0, u(L, t) = 0, for all t ≥ 0 (2)

and the initial condition:

u(x, 0) = f (x), 0 ≤ x ≤ L. (3)

Hoang Hai Ha (HCMUT) Partial differential equations 7 / 15

The finite difference method: an explicit scheme

Creating grids for x and t: we divide [0, L] into N subintervals

L
with the step size h := , the grid points of x are
N
x 0 = 0, x i = ih, x N = L, i = 1, . . . N − 1; the grid points of t
are tj = jk, j = 0, 1 . . ., k is the time step size.

Hoang Hai Ha (HCMUT) Partial differential equations 8 / 15

The finite difference method: an explicit scheme

Creating grids for x and t: we divide [0, L] into N subintervals

L
with the step size h := , the grid points of x are
N
x 0 = 0, x i = ih, x N = L, i = 1, . . . N − 1; the grid points of t
are tj = jk, j = 0, 1 . . ., k is the time step size.
Let u(xi , tj ) ≈ ui,j . We aim to find ui,j .

Hoang Hai Ha (HCMUT) Partial differential equations 8 / 15

The finite difference method: an explicit scheme

Creating grids for x and t: we divide [0, L] into N subintervals

L
with the step size h := , the grid points of x are
N
x 0 = 0, x i = ih, x N = L, i = 1, . . . N − 1; the grid points of t
are tj = jk, j = 0, 1 . . ., k is the time step size.
Let u(xi , tj ) ≈ ui,j . We aim to find ui,j .
∂ 2u
Applying the centered difference formula, is approximated
∂x 2
by:
∂ 2u u (xi + h, tj ) − 2u (xi , tj ) + u (xi − h, tj )
2
(xi , tj ) ≈ , i = 1, . . . N −
∂x h2

Hoang Hai Ha (HCMUT) Partial differential equations 8 / 15

The finite difference method: an explicit scheme
∂u
Applying the forward difference formula, is evaluated by
∂t
∂u u (xi , tj + k) − u (xi , tj )
(xi , tj ) = .
∂t k

Substituting two difference formulas into the equation (1), we

obtain
ui,j+1 − ui,j ui+1,j − 2ui,j + ui−1,j
− c2 = 0. (4)
k h2

Hoang Hai Ha (HCMUT) Partial differential equations 9 / 15

The finite difference method: an explicit scheme
∂u
Applying the forward difference formula, is evaluated by
∂t
∂u u (xi , tj + k) − u (xi , tj )
(xi , tj ) = .
∂t k

Substituting two difference formulas into the equation (1), we

obtain
ui,j+1 − ui,j ui+1,j − 2ui,j + ui−1,j
− c2 = 0. (4)
k h2

2c 2 k
 
k
(4) ⇒ ui,j+1 = 1− 2 ui,j + c 2 (ui+1,j + ui−1,j ).
h h2

Hoang Hai Ha (HCMUT) Partial differential equations 9 / 15

The finite difference method: an explicit scheme
Using the initial condition u(x, 0) = f (x) (see (3)), we have
u0,0 = f (0), u1,0 = f (x1 ), . . . , uN,0 = f (xN ).

Hoang Hai Ha (HCMUT) Partial differential equations 10 / 15

The finite difference method: an explicit scheme
Using the initial condition u(x, 0) = f (x) (see (3)), we have
u0,0 = f (0), u1,0 = f (x1 ), . . . , uN,0 = f (xN ).

Using the boundary conditions u(0, t) = u(L, t) = 0 (see (2)), we

have
⋆ u0,1 = u(0, t1 ) = 0.

Hoang Hai Ha (HCMUT) Partial differential equations 10 / 15

The finite difference method: an explicit scheme
Using the initial condition u(x, 0) = f (x) (see (3)), we have
u0,0 = f (0), u1,0 = f (x1 ), . . . , uN,0 = f (xN ).

Using the boundary conditions u(0, t) = u(L, t) = 0 (see (2)), we

have
⋆ u0,1 = u(0, t1 ) = 0.
2c 2 k
 
k
⋆ u1,1 = 1 − 2 u0,0 + c 2 2 (u2,0 + u0,0 ) .
h h
2
 
2c k k
⋆ u2,1 = 1 − 2 u2,0 + c 2 2 (u3,0 + u1,0 ).
h h
⋆ ...
2c 2 k
 
k
⋆ uN−1,1 = 1 − 2 uN−1,0 + c 2 2 (uN,0 + uN−2,0 ).
h h
Hoang Hai Ha (HCMUT) Partial differential equations 10 / 15
The finite difference method: an explicit scheme
Since the values ui,0 ; (i = 1 . . . N − 1) are known, we can
evaluate ui,1 , (i = 1 . . . N − 1) and so ui,2 , (i = 1 . . . N − 1).
Now, we are going to describe this method by using the matrix form.
c 2k
We denote λ := 2 ;
h
U (0) := (f (x 1) , f (x 2) , . . . , f (x N −1))T ; and
U (j ) := (u 1,j , u 2,j , . . . , u N −1,j )T , j = 1, 2, . . .

Hoang Hai Ha (HCMUT) Partial differential equations 11 / 15

The finite difference method: an explicit scheme
Since the values ui,0 ; (i = 1 . . . N − 1) are known, we can
evaluate ui,1 , (i = 1 . . . N − 1) and so ui,2 , (i = 1 . . . N − 1).
Now, we are going to describe this method by using the matrix form.
c 2k
We denote λ := 2 ;
h
U (0) := (f (x 1) , f (x 2) , . . . , f (x N −1))T ; and
U (j ) := (u 1,j , u 2,j , . . . , u N −1,j )T , j = 1, 2, . . .
Thus we have
U (j ) = AU (j −1),
where A is a tri-diagonal matrix with the form
 
(1 − 2λ) λ 0 ... 0
 λ (1 − 2λ) λ 0 
A= . . .
 
 0 λ .. .. .. .

 .
.. . ..

 λ 
0 0 . . . λ (1 − 2λ)
Hoang Hai Ha (HCMUT) Partial differential equations 11 / 15
The finite difference method: an explicit scheme

Hoang Hai Ha (HCMUT) Partial differential equations 12 / 15

Figure: The value at blue point is evaluated by neighboring values at black points
on the previous time

Hoang Hai Ha (HCMUT) Partial differential equations 13 / 15

The finite difference method: an explicit scheme

Example 2.1
Approximate solutions of
∂u ∂ 2u
(x, t) − 2 (x, t) = 0, 0 < x < 1, 0 ≤ t
∂t ∂x
satisfying
the boundary conditions: u(0, t) = u(1, t) = 0, 0 < t
the initial conditions: u(x, 0) = sin(πx), 0 ≤ x ≤ 1.
Let the step size for x is h = 0.1 and for t is 0.0005. Compare the
2
numerical results with the exact solution: u(x, t) = e −π t sin(πx).

Hoang Hai Ha (HCMUT) Partial differential equations 14 / 15

The finite difference method: an explicit scheme

Example 2.1
Approximate solutions of
∂u ∂ 2u
(x, t) − 2 (x, t) = 0, 0 < x < 1, 0 ≤ t
∂t ∂x
satisfying
the boundary conditions: u(0, t) = u(1, t) = 0, 0 < t
the initial conditions: u(x, 0) = sin(πx), 0 ≤ x ≤ 1.
Let the step size for x is h = 0.1 and for t is 0.0005. Compare the
2
numerical results with the exact solution: u(x, t) = e −π t sin(πx).
k
We have c = 1. Thus λ = c 2 = 0.05 và N = 10.
h2
We approximate the values of solution at t = 0.5.

Hoang Hai Ha (HCMUT) Partial differential equations 14 / 15

The convergence condition

The sequence of values {ui,j } obtained by the explicit scheme

converges to the exact solutions if

λ ≤ 0.5.

Hoang Hai Ha (HCMUT) Partial differential equations 15 / 15

The convergence condition

The sequence of values {ui,j } obtained by the explicit scheme

converges to the exact solutions if

λ ≤ 0.5.

Example 2.2
Consider the above example with k = 0.01.

Hoang Hai Ha (HCMUT) Partial differential equations 15 / 15