• Sonuç bulunamadı

The temperature u in a nonuniform rod solves the partial di¤erential equa- tion

N/A
N/A
Protected

Academic year: 2021

Share "The temperature u in a nonuniform rod solves the partial di¤erential equa- tion"

Copied!
3
0
0

Yükleniyor.... (view fulltext now)

Tam metin

(1)

6. Sturm-Liouville Eigenvalue Problems 6.1. Some Examples

Heat Flow in a Nonuniform Rod

The temperature u in a nonuniform rod solves the partial di¤erential equa- tion

c @u

@t = @

@x K 0 @u

@x + Q (1)

where Q denotes any possible sources of heat energy. The thermal coe¢ cient c; and K 0 depend on x: The method of separation of variables is applied if (1) is linear and homogeneous. Usually, we consider the case Q = 0: But, we will be slightly more general. Assume that the heat source Q is proportunal to u;

Q = au where a can depend on x (but not on t). So

c @u

@t = @

@x K 0

@u

@x + au: (2)

We apply the method of separation of variables to solve this equation. We assume that there is a homogeneous boundary condition (unspeci…ed) at end points x = 0 and x = L:

Consider

u (x; t) = X (x) T (t) : (3)

If we substitute this function in (2), we obtain c X (x) dT

dt = T (t) d

dx K 0 dX

dx + aX (x) T (t) : If we divide by c X (x) T (t) ; we have

1 T

dT dt = 1

c X d dx K 0

dX dx + a

c = (4)

where - is separation constant.

dT

dt = T

) T (t) = Ce t

It is seen that if > 0; it has exponentially decaying solutions; if < 0; solution grows and if = 0 solution is constant.

The spatial di¤erential equation is

1

(2)

d dx K 0

dX

dx + aX + c X = 0: (5)

If two homogeneous boundary conditions are given, it forms a boundary value problem. Here, the thermal coe¢ cients a; c; ; K 0 are not constant and the equation (5) is a di¤erential equation with nonconstant-coe¢ cient. In general, nonconstant-coe¢ cient di¤erential equations occur appear in situations where physical properties are nonuniform. Generally, we can not solve (5) in the variable-coe¢ cient case, but we can …nd a numerical approximate solution on the computer. Later, we will return to reinvestigate heat ‡ow in a nonuniform rod.

Circularly Symmetric Heat Flow

The di¤erential equations with nonconstant-coe¢ cient also arise if the phys- ical parameters are constant. In Section 1.5 we showed that if the temperature u in some plane two dimensional region is circularly symmetric, that is, u de- pends only on time t and on the radial distance r from the origin), then u is the solution of the linear and homogeneous partial di¤erential equation

@u

@t = k 1 r

@

@r r @u

@r (6)

where we assume that all the thermal coe¢ cients are constant. By the method of separation of variables, we seek for a solution in the form

u (r; t) = R (r) T (t) : After necessary calculations, we obtain

1 kT (t)

dT (t)

dt = 1

rR (r) d

dr r dR (r)

dr = ;

which gives two di¤erential equations dT (t)

dt = kT (t) (7)

and d

dr r dR (r)

dr + rR (r) = 0: (8)

Here, the separation constant is denoted - , because we expect solutions to exponentially decay in time when > 0: The solution of (7) is T (t) = Ce kt and also the equation (8) will be solved in terms of Bessel functions later.

We now consider the appropriate homogeneous boundary conditions for cir- cularly symmetric heat ‡ow inside circle and a circular annulus: In both cases, let all boundaries be …xed at zero temperature. For the annulus, the boundary conditions for (8) at the inner (r = a) and outer (r = b) concentric circular walls

u (a; t) = 0 and u (b; t) = 0;

2

(3)

for the circle, the boundary condition for (8) is u (b; t) = 0: Because of the fact that the physical variable r ranges from 0 to b; we need a homogeneous boundary condition at r = 0 for mathematical reasons. So, we expect u bounded at r = 0;

that is, ju (0; t)j < 1: Thus, we have homogeneous conditions at both r = 0 and r = b for the circle.

6.2. General Classi…cation

A boundary value problem is formed of a linear homogeneous di¤erential equation and corresponding linear homogeneous boundary conditions. All of the di¤erential equations for boundary value problems are in form of

d dx p d

dx + q + = 0 ; a < x < b (9) here denotes eigenvalue. Some examples of (9) are as follows:

a) Simplest case: d dx

22

+ = 0 for p = 1; q = 0; = 1:

b) Heat ‡ow in a nonuniform rod: dx d K 0 dX

dx + aX + c X = 0 here the dependent variable = X and p = K 0 ; q = a; = c .

c) Circularly symmetric heat ‡ow: dr d r dR dr + rR = 0 here the dependent variable = R; the independent variable x = r and p(x) = x; q(x) = 0; (x) = x:

Many interesting results are related to any equation in the form (9). This equation is called a Sturm–Liouville di¤erential equation.

Boundary conditions. Some linear homogeneous boundary conditions are as follows:

Heat ‡ow Mathematical terminology

= 0 Fixed(zero)

temperature Dirichlet condition d

dx = 0 Insulated Neumann condition

d

dx = h

(Homogeneous) Newton’s law of

cooling 0 0 temperature, h = H=K 0 ; h > 0

Robin condition

( L) = (L) d

dx ( L) = d dx (L)

Perfect thermal contact

Periodicity condition (mixed type) j (0)j < 1 Bounded temperature Singularity condition

3

Referanslar

Benzer Belgeler

Note: If is an integrating factor giving a solution = c and is an arbi- trary function of ; then d d is also integrating factor of the given equation. Since is arbitrary, there

Partial di¤erential equations arise in geometry and physics when the number of independent variables in the problem is two or more.. We de…ne the order of a partial di¤erential

If is taken as an arbitrary function, it is not always possible to eliminate the constant a between equations (5), so it is generally not possible to express the general integral of

It is not necessary to use all of the equations (5) for a …rst integral to be found from system (5), known as Charpit equations.. However, in the …rst integral we will …nd, at least

In this section, we will examine the special types of …rst-order partial di¤er- ential equations that can be easily solved with the Charpit

They can be di¤erent

If the functions M and N in equation (1) are both homogeneous with same degree, then the di¤erential equation (1) is called

Let us …rst observe that this equation is