Вы находитесь на странице: 1из 7

1

CHAPTER 4

NUMERICAL SOLUTION OF PARTIAL DIFFERENTIAL EQUATIONS

Partial differential equations (PDE) are those which involve two or more independent variables
that determine the behavior of the dependent variable as described by a differential equation,
usually of second order or higher order. They find application in the fields of heat flow, fluid
flow, electrostatics, diffusion of matter, and so on.

Many physical phenomena are a function of more than one independent variable and must be
represented by a PDE, by which term we mean an equation involving partial derivatives. Most
scientific problems have mathematical models that are second-order equations, with the
highest order of derivative being the second.

The distinction as to elliptic, parabolic, or hyperbolic for second-order PDE depends on the
coefficients of the second derivative terms.

Second Order PDE:

𝜕 2𝑢 𝜕 2𝑢 𝜕 2𝑢 𝜕𝑢 𝜕𝑢
𝐴 2
+ 𝐵 + 𝐶 2
+ 𝐷 (𝑥, 𝑦, 𝑢, , ) = 0
𝜕𝑥 𝜕𝑥𝜕𝑦 𝜕𝑦 𝜕𝑥 𝜕𝑦

Where u is a function of the two independent variables x and y.

Classification: Elliptic, if B2 – 4AC < 0;

Parabolic, if B2 – 4AC = 0;

Hyperbolic, if B2 – 4AC > 0.

Elliptic Equations: In a steady-state problem the dependent variable does not vary with time.

Laplace’s Equation: Equation for two dimensional steady-state heat flow problem (A = C = 1
and B = 0).

𝜕2 𝑢 𝜕2 𝑢
+ 𝜕𝑦 2 = 0 𝑜𝑟 ∇2 𝑢 = 0.
𝜕𝑥 2

𝜕2 𝜕2
Where ∇2 = 𝜕𝑥 2 + 𝜕𝑦 2 is called the Laplacian operator. For three-dimensional heat flow
problems, we have

𝜕 2𝑢 𝜕 2𝑢 𝜕 2𝑢
+ + = 0 𝑜𝑟 ∇2 𝑢 = 0.
𝜕𝑥 2 𝜕𝑦 2 𝜕𝑧 2
For classroom use only. Please don’t share.
2

Poisson’s Equation: ∇2 𝑢 = 𝑓(𝑥, 𝑦)

Representation as a Difference Equation

One scheme for solving PDE’s is to replace the derivatives by difference quotients,
converting the equation to a difference equation. We then write the difference equation
corresponding to each point at the intersections (nodes) of a gridwork that subdivides the
region of interest at which the function values are unknown. Solving these equations
simultaneously gives values for the function at each node that approximate the true values. We
begin with two dimensional case.

Difference Equation:

OR: 4 𝑢𝑖,𝑗 − 𝑢𝑖+1,𝑗 − 𝑢𝑖−1,𝑗 − 𝑢𝑖,𝑗+1 − 𝑢𝑖,𝑗−1 = 0

Pictorial Operator that represents the linear combination of the u-values:

−1
1
OR: ∇2 𝑢𝑖𝑗 = ℎ2 {−1 4 − 1} 𝑢𝑖𝑗 = 0
−1

For classroom use only. Please don’t share.


3

The relation makes the equations easy to write and the constant (1/h2) can be divided out.

Laplace’s Equations on a Rectangular Region

PROBLEM: A typical steady-state problem.

A thin steel plate is 10 x 20 cm rectangle. If one of the 10-cm edges is held at 1000C and the
other edges are held at 00C, what are the steady-state temperatures at interior points? For
steel, k = 0.16 cal/sec-cm2.0C/cm.

Solution: We are asked to find u(x,y) such that

𝜕 2𝑢 𝜕 2𝑢
+ =0 ,
𝜕𝑥 2 𝜕𝑦 2

With the ff. boundary conditions;

u(x,0) = 0; u(x,10) = 0; u(0,y) = 0; u(20,y) = 100.

Sketch of the boundary conditions:

Difference Equation:

We find the solution to the set of difference equations. We choose, h = 5cm.

We use the following notations.

For classroom use only. Please don’t share.


4

The system of equations is

1
(0 + 0 + 𝑢2 + 0 − 4𝑢1 ) = 0,
52
1
(𝑢 + 0 + 𝑢3 + 0 − 4𝑢2 ) = 0,
52 1
1
(𝑢 + 0 + 100 + 0 − 4𝑢3 ) = 0,
52 2
The pictorial operator is

1
1
∇2 𝑢𝑖𝑗 = ℎ2 {1 − 4 1} 𝑢𝑖𝑗 = 0.
1
ANS: 𝑢1 = 1.786, 𝑢2 = 7.143, 𝑢3 = 26.786

Redoing the problem with smaller h should improve the precision.

When h = 2.5, we obtain 21 interior points where u1=v9, u2 = v11, u3 = v13.

For classroom use only. Please don’t share.


5

Solution in Matrix Form:

Solution vector:

For classroom use only. Please don’t share.


6

Where u1=v9, u2 = v11, u3 = v13.

ITERATIVE METHODS FOR LAPLACE’S EQUATION

Liebmann’s Method – the Gauss-Seidel iterative technique applied to the Laplace’s equation.

The pictorial relation of the Laplace’s equation results to a system of linear equations that is
diagonally dominant if proper ordering of the equations is made.

The first step in applying Liebmann’s method is to rearrange these equations into a new form
by solving each equation for the variable on the diagonal.

For classroom use only. Please don’t share.


7

Example: We illustrate this procedure with the previous example.

Difference equations when h=5cm:


𝑢2
−4𝑢1 + 𝑢2 = 0 → 𝑢1 = ,
4
𝑢1 + 𝑢3
𝑢1 − 4𝑢2 + 𝑢3 = 0, → 𝑢2 =
4
𝑢2 + 100
𝑢2 −4𝑢3 = −100 → 𝑢3 =
4
𝑢𝑖+1,𝑗 + 𝑢𝑖−1,𝑗 + 𝑢𝑖,𝑗+1 + 𝑢𝑖,𝑗−1
Form: 𝑢𝑖𝑗 = 4

For classroom use only. Please don’t share.

Вам также может понравиться