MISN-0-486

BOUNDARY CONDITIONS: VIBRATING STRINGS, HEAT DIFFUSION

BOUNDARY CONDITIONS: VIBRATING STRINGS, HEAT DIFFUSION by R. D. Young, Dept. of Physics, Illinois State Univ. 1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 2. Procedures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 3. Supplementary Notes a. Boundary Conditions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 b. Type of Equation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6 c. Derivation of the Diusion Equation . . . . . . . . . . . . . . . . . . . . . 8 d. Solution of the Vibrating String with Ends Clamped . . . . .9 e. Temperature in a Slab or Bar . . . . . . . . . . . . . . . . . . . . . . . . . . 13 f. Temperatures in an Innite and Semi-innite Bar . . . . . . . 16 Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18

Math Physics
Project PHYSNET Physics Bldg. Michigan State University East Lansing, MI
1

ID Sheet: MISN-0-486 THIS IS A DEVELOPMENTAL-STAGE PUBLICATION OF PROJECT PHYSNET The goal of our project is to assist a network of educators and scientists in transferring physics from one person to another. We support manuscript processing and distribution, along with communication and information systems. We also work with employers to identify basic scientic skills as well as physics topics that are needed in science and technology. A number of our publications are aimed at assisting users in acquiring such skills. Our publications are designed: (i) to be updated quickly in response to eld tests and new scientic developments; (ii) to be used in both classroom and professional settings; (iii) to show the prerequisite dependencies existing among the various chunks of physics knowledge and skill, as a guide both to mental organization and to use of the materials; and (iv) to be adapted quickly to specic user needs ranging from single-skill instruction to complete custom textbooks. New authors, reviewers and eld testers are welcome. PROJECT STAFF Andrew Schnepp Eugene Kales Peter Signell Webmaster Graphics Project Director

Title: Boundary Conditions: Vibrating Strings, Heat Diusion Author: R. D. Young, Dept. of Physics, Illinois State Univ. Version: 10/18/2001 Length: 2 hr; 24 pages Input Skills: 1. Vocabulary: periodic function, Fourier expansion of f (x), orthogonal function, Fourier integral expansion of f(x), Fourier transform of f (x), Fourier integral theorem, partial dierential equation. 2. Unknown: assume (MISN-0-485). Output Skills (Knowledge): K1. Vocabulary: partial dierential equation, order of P.D.E., solution of P.D.E., boundary conditions, boundary value problem, linear and nonlinear P.D.E., elliptic P.D.E., hyperbolic P.D.E., parabolic P.D.E., diusion and heat conducting equation, vibrating string equation, Laplaces equation, Poissons equation, Dirichlet boundary conditions, Neumann boundary conditions, Cauchy boundary conditions. K2. Classify a given P.D.E. according to these characteristics: linear or nonlinear, homogeneous or nonhomogeneous, order, elliptic, hyperbolic or parabolic. K3. Derive the diusion equation. Output Skills (Rule Application): R1. Write down the mathematical B.V.P. corresponding to a physical problem when given the problem in words. R2. Compute the solution to a physical problem involving vibrating strings with ends clamped or temperatures in a slab or bar. External Resources (Required): 1. G. Arfken, Mathematical Methods for Physicist, Academic Press (1995). 2. Schaums Outline: Murray Spiegel, Theory and Problems of Advanced Mathematics for Scientists and Engineers, McGraw-Hill Book Co. (1971). Evaluation: Stage 0

ADVISORY COMMITTEE D. Alan Bromley E. Leonard Jossem A. A. Strassenburg Yale University The Ohio State University S. U. N. Y., Stony Brook

Views expressed in a module are those of the module author(s) and are not necessarily those of other project participants. c 2001, Peter Signell for Project PHYSNET, Physics-Astronomy Bldg., Mich. State Univ., E. Lansing, MI 48824; (517) 355-3784. For our liberal use policies see: http://www.physnet.org/home/modules/license.html.

MISN-0-486

MISN-0-486

BOUNDARY CONDITIONS: VIBRATING STRINGS, HEAT DIFFUSION by R. D. Young, Dept. of Physics, Illinois State Univ. 1. Introduction
In this unit, you will come face to face with two very important applications of Fourier series and integrals in the solution of problems involving vibrating systems and heat diusion. Partial dierential equations of various types will be introduced and analyzed rst. Then, the diusion equation is derived in order to illustrate how a physical problem results in such a partial dierential equation. Finally, solutions to boundary value problems involving vibrating strings and heat diusion are obtained. Since this unit has a rather extensive reading assignment, I will cut o my introductory remarks here so that you can get started on the unit.

ecient way to remember a derivation is to understand the concepts involved. 6. Read through Solved Problems 12.6, page 263, of Spiegel. This problem gives an alternative derivation of the source-free diusion equation. 7. Read through Solved Problems 12.6, pages 266-267, of Spiegel on solving a heat diusion problem using separation of variables. 8. Read through Sections III to V, pages 9-9 to 9-20, of the Supplementary Notes. These sections treat the solution of B. V. P. involving vibrating strings and heat diusion. 9. Read through Solved Problems 12.17, 12.18, 12.19, 12.21, and 12.23 of Spiegel on B. V. P. involving vibrating strings and heat diusion. 10. Solve the Supplementary Problems 12.29, 12.30, 12.48, 12.50, 12.53, 12.65, and 12.66 of Spiegel on B. V. P. involving vibrating strings and heat diusion. You may simply apply the appropriate boxed equations from the Supplementary Notes. Note: You are allowed one sheet of 8.5 11.0 paper containing the boxed equations in the Supplementary Notes when taking the Unit Test. The sheet can also contain the B. V. P. for which the boxed equations are solutions.

2. Procedures
1. a. Read pages 258-260 of Spiegel. b. The question of appropriate boundary conditions in a physical problem is a touchy one in which you can only gain skill by experience. In order to give you a conceptual framework to use in your future work read Section 1, page 9-3 to 9-7 of the Supplementary Notes. 2. Write down or underline in the text each of the denitions and concepts of Output Skill K1. 3. Read through the Solved Problems 12.1 and 12.2 on Classication of Partial Dierential Equations in Spiegel, page 261. 4. Solve Supplementary Problems 12.29 and 12.30 on Classication of Partial Dierential Equations in Spiegel, pages 277-278. 5. Write down the derivation of the diusion equation as given in Section II, page 9-7 to 9-9, of the Supplementary Notes. You will be asked to write down this derivation from memory on the Unit Test. WARNING: Although I have used the word memory here, the most

Correct answer is: u(x, y) = 0 x 0 + tan1 . 2 y

3. Supplementary Notes
3a. Boundary Conditions. A physical problem is not uniquely specied if the partial dierential equation which the solution of the problem must satisfy is given. There is an innite number of solutions of the partial dierential equations listed on pages 259 and 260 of Spiegel. Any physical problem must state not only the partial dierential equation which is to be solved but also the boundary conditions which the solution must satisfy. In fact, satisfying of the boundary conditions (B. C. ) is often as dicult a task as the solving of the partial dierential equation (P. D. E. ).
6

MISN-0-486

MISN-0-486 1. Elliptic equations have B 2 4AC < 0. Laplaces equation 2u 2u + 2 =0 x2 y and Poissons equation 2u 2u + 2 = 4(x, y) x2 y are examples. 2. Hyperbolic equations have B 2 4AC > 0. The wave equation 1 2u 2u 2 2 =0 x2 v t is an example. 3. Parabolic equations have B 2 4AC = 0. The diusion equation 1 u 2u =0 2 x t

However, the solutions of a given P. D. E. cannot be made to satisfy any sort of B. C. For each type of P. D. E. . on pages 259 and 260 of Spiegel, there is a denite set of B. C. which will give unique answers. An actual physical problem will always have the right sort of B. C. to give it a unique answer (we hope!), and if the statement of the problem corresponds to reality, the right B. C. are guaranteed. But it is not always easy to tell just what B. C. correspond to reality. Thus, the types of B. C. which are suitable for various P. D. E. are given below. This can guide a scientist or engineer in making his mathematical problem t the physical problem as closely as possible. The best and most accessible reference on B. C. in P. D. E. remains Morse and Feshbach, Methods of Theoretical Physics, Sec. 6.1, pages 676-692, McGraw-Hill (1953). There are essentially four types of B. C. In order to make the denitions precise, let the P. D. E. be written 2u 2u u u 2 A(x, y) 2 + B(x, y) + C(x, y) 2 = H(x, y, , ) x xy y x y where H(x, y) = D(x, y) u u + E(x, y) + F (x, y)u G(x, y) x x (2) (1)

(3)

(4)

(5)

(6)

is an example. As an example, the solution of Poissons equation inside a volume V subject to either Dirichlet or Neumann B. C. on the closed surface S is unique. To show this, assume that there are two functions u1 and u2 such that 2 u1 = 4 and
2

The functions A, B, C, D, E, F , and G are specied functions of x and y in some domain R bounded by a curve S. The two variables x and y may be either two space coordinates or one space and one time coordinate. (If there are three spatial dimensions involved, R is a volume V and S is a surface). From this point of view, initial conditions are boundary conditions in time. Then the four types of B. C. are: 1. Dirichlet B. C. The value of u is specied on S. 2. Neumann B. C. The value of the normal derivative u/n is specied on S. 3. Cauchy B. C. The values of u and u/n are specied on S. 4. Mixed B. C. Some combination of the above three types of B. C. are given for various portions on S. The P. D. E. are classied as in Spiegel on page 258:

u2 = 4

(7)

where either u1 and u2 are given an S (where u1 = u2 on S) corresponding to Dirichlet B. C. or u1 /n and u2 /n are given on S (where u2 u1 = n n (8)

on S) corresponding to Neumann B. C. Then dene U = u1 u2 . Now, U = 0 on S if Dirichlet B. C. hold, or U/n = 0 on S if Neumann B. C. hold. Applying eq. (1.101) of Arfken gives (U
V 2

U+

U ) d =
V

(U U ) d

(9)

MISN-0-486 =
S

5 U U d . U ) d =
S 2

MISN-0-486 3b. Type of Equation. Type of B. C. Dirichlet Open surface Closed surface Neumann Open Surface Closed surface The Table: Hyperbolic (wave eq. ) Not enough Too much

So, (U
V 2

U+

U dS n

Elliptic (Poissons eq. ) Not enough Unique, stable solution Not enough Unique, stable sol. in general

(10)

Parabolic (heat-cond. eq. ) Unique, stable sol. in one direction Too much Unique, stable sol. in one direction Too much

where d = n dS and

U n = U/n. Also,

U = 0.

But, for either type of B. C. , the surface integral vanishes, so that | U |2 = 0 . (11)

Not enough Too much

This last equation will be true for any volume V so that U =0 inside V . Thus, U = constant (13) inside V . In the case of Dirichlet B. C. , this constant is zero so that u1 = u2 . In the case of Neumann B. C. , this constant may not be zero so u1 and u2 only dier by an unimportant arbitrary, additive constant. It should be clear that a solution to Poissons equation with Cauchy B. C. on a closed boundary (both u and u/n specied) does not exist, since there are unique solutions for Dirichlet and Neumann B. C. Also, there is a unique solution to the problem in the case of mixed B. C. where Dirichlet B. C. hold over a portion of S and Neumann B. C. hold over the remaining part. Morse and Feshbach study the various P. D. E. and B. C. in order to come up with the following table for the uniqueness of solutions to P. D. F. with various B. C. (12)

Cauchy Open Surface Closed surface

Unphys. results Too much

Unique, stable solution Too much

Too much Too much

The application of this table can sometimes be not so straightforward, so a reading of Morse and Feshbach would be of great help. For example, consider the wave equation which is the prototype example of an hyperbolic equation. According to the table, a unique solution exists for the case of Cauchy B. C. on an open surface. And, indeed, this is true. That is, a unique solution exists for the boundary value problem (B. V. P. ) 1 2u 2u 2 2 = 0, ( < x < , 0 < t < ) x2 v t u(x, 0) = f (x), ( < x < ) and u(x, 0) = g(x), ( < x < ) . t (14) (15) (16)

The functions g(x) and f (x) are given. This amounts to solving the P. D. E. with Cauchy B. C. on an open surface where f (x) is the initial displacement and g(x) is the initial velocity. The open surface (curve in this case) is just the x-axis since there is no closure on t at innity.

10

MISN-0-486

MISN-0-486

Region R corresponds to t > 0

problem being over- dened if Cauchy B. C. were given on x = 0 and x = L. The exact reasons for this can be found in Morse and Feshbach, page 686. 3c. Derivation of the Diusion Equation. As an example of how a physical model can result in one of the P. D. E. which are listed in section 8.1 of Arfken and on pages 259-260 of Spiegel, the diusion equation will be derived. The derivation applies to some physical quantity Q(t) of a substance. The amount of Q per unit volume is q(r, t). The quantity q(r, t) is a time-dependent scalar eld. Therefore, the amount Q(t) in volume V is given by Q(t) =
V

x Curve S is line t = 0, i.e. the x-axis

The proof that a unique solution exists is easy, but will not be done here. See Morse and Feshbach, page 685. On the other hand, suppose the waving medium (say a string) is clamped at x = 0 and x = L. Then, we have these B. V. P. : 1 2u 2u 2 2 = 0, (0 < x < L, 0 < t < ) x2 v t u(0, t) = u(L, t) = 0, (0 t < ) u(x, 0) = f (x), (0 x L) and u(x, 0) = g(x), (0 x L) . t (17) (18) (19) (20)

q(r, t) d .

(21)

The amount of Q(t) can change by two processes. a) Flow of the physical quantity via a current. This ow is described by a current vector j(r, t) which is a vector eld. For example, Q can be electric charge, q can be charge density, and j can be electric current. b) Sources of quantity Q in volume V . The source (or sink) is described by a scalar eld (r, t) which is the rate at which Q(t) is created per unit volume per unit time. Thus, dQ = dt q(r, t) d = t j d + (r, t) d .
V

The open curve S now corresponds to the lines x = 0, x = L, and t = 0 and is U-shaped.

(22)

t Region R Open curve S ` ds

` j V ^ n ds ` ds = ^ ds n S ^ n = unit normal vector to surface S at ds

x=0

x=L

The B. C. given in eqs. (19) and (20) amount to Cauchy B. C. on t = 0 for 0 x L, but the B. C. in eq. (18) amounts to Dirichlet conditions on x = 0 and x = L for t > 0. This occurs because the extra constraint on the problem due to the clamped ends would result in the

From the orientation of j and d, if j d is a positive quantity, then it represents the ow into V across d.

11

12

MISN-0-486 But
V

MISN-0-486

10

j d =

j d . d = 0 .

(23)

U(x,t) displacement at x at time t x=0 x

shape of string at time t

So
V

q + t

(24)

This equation holds for any volume V . Therefore, q + t j = . (25)

x=L

Let u(x, t) represent the displacement of the string. Then the B. V. P. can be written 1 2u 2u 2 2 = 0, (0 < t < , 0 < x < L) x2 v t u(0, t) = u(L, t) = 0, (0 t < ) u(x, 0) = f (x), (0 x L) and (29) (30) (31)

This is the equation of continuity with sources. It is essentially a conservation equation. In some cases an empirical relation exists between j and q. Many times it is true that a gradient of q results in a ow of quantity q from higher to lower concentrations. Then j(r, t) = x q(r, t) . (26)

If q represents charge, then j represents the current density. If q is temperature; then j represents heat ow per unit area per unit time. In the case of heat ow, this last equation is the Fourier law of heat conduction. If q represents the concentration of a solute in a solvent, then the last equation is called Ficks law. Substitution of j in eq. (26) into the equation of continuity, eq. (25), gives q t ( q) = . (27)

u(x, 0) = g(x), (0 x L) . (32) t The known functions f (x) and g(x) are the initial displacement and velocity of the string, respectively. I shall use the method of separation of variables to solve this problem. So assume u(x, t) = X(x)T (t) . Then, substitution of eq. (33) in eq. (29) yields this result: X T or X 1 XT = 0 v2 T (t) . v 2 T (t) (34) (33)

If n is constant and there are no sources or sinks of the quantity (i.e. = 0), then 1 q 2 q = 0. (28) t This is the heat conduction equation as given on page 259 of Spiegel. 3d. Solution of the Vibrating String with Ends Clamped. Consider the case of a string vibrating with ends clamped. Let the coordinate along the string be x, and the time coordinate be t. Let v be the speed of propagation of waves on the string.

(x)X(x) =

(35)

The left-hand side of the above equation is a function of x alone so it cannot vary with t. The right-hand side is a function of t alone so it cannot vary with x. Hence, both sides must be equal to some constant, say , so that X (x) X(x) = 0 (36) T (t) v 2 T (t) = 0 . X(0) = X(L) = 0 . (37) From the B. C. in eq. (30), this result is obtained: (38)
14

13

MISN-0-486 Solution to eq. (36) can be written X(x) = C1 ex

11

MISN-0-486

12

+ C2 ex

(39)

where Cn and Dn are constants. The full solution u(x, t) must be a linear combination of the Xn Tn , because the remaining two B. C. need to be satised. So, if Cn = Bn Cn and Dn = Bn Dn , u(x, t) =
n=1

where C1 and C2 are constants. Eqs. (38) and (39) give C1 + C 2 = 0 and C1 e Suppose > 0. Then eL so that
L

Cn cos

(40) But = 0. (41) (42) (43) and

nvt nvt nx + Dn sin . sin L L L

n=1

(51)

+ C2 e

u(x, 0) = f (x) =

Cn sin

nx L

(52)

eL

=0

sinh L = 0 .

nx u(x, 0) x nDn sin = g(x) = . t L n=1 L

(53)

But, the hyperbolic sine is always nonzero when L > 0. Thus, solutions for X(x) only exist when 0. Let = 2 , 2 0 . Then, the expression for X(x) in eq. (39) can be rearranged to give X(x) = A cos x + B sin x where A and B depend on C1 and C2 . The B. C. in eq. (38) yield A=0 and B sin L = 0 . Thus L = n, n = 1, 2, 3, . . . (48) The solution X(x) are actually denumberably innite, and they can be labeled by the integer n: nx , n = 1, 2, 3, . . . Xn (x) = Bn sin L (49) (47) (46) (45) (44)

Remember that 0 x L. The above two equations are Fourier sine series on (0, L). Thus, these expressions can be written down immediately for Cn and Dn : 2 L nv dx (54) f (x) sin Cn = L 0 L and nv 2 D = L n L
0 L

g(x) sin

nv dx . L

(55)

Eqs. (51), (54), (55) are a complete solution to the original B. V. P. Numerous examples are discussed in Spiegel on the pages indicated in the Procedures. As an example, suppose f (x) = x, when 0 x 1 and g(x) = 0 when 0 x 2 . Then
2

= x + 2 when 1 x 2

(56)

(57)
2

Cn = so

f (x) sin
0

nx dx = 2

x sin
0

nx dx + 2 sin n . 2

(x + 2) sin
1

nx dx 2 (58) (59)
16

In the same way the solution T (t) depends on n and can be written as Tn (t) = Cn cos nvt nvt + Dn sin L L (50)

Cn = Also

8 2 n2

Dn = 0 .
15

MISN-0-486 Thus u(x, t) = 8 2

n=1

13

MISN-0-486

14

sin

nx nvt 2 n sin cos /n , (0 x 2) . 2 2 2

and x = L, correspond to knowing the rate at which the temperature is changing. A usual case is to have one of the following cases: (60) 1. u(0, t) & u(L, t) being constant, corresponding to constant temperature at the faces. 2. u(0, t)/t & u(L, t)/t being zero corresponding to thermally insulated faces . Case 1: Let us consider this B. V. P : 2u 1 u = 0, (0 < x < L, 0 < t < ) x2 t where u(0, t) = T0 , (0 t < ) (62) (63) (64) u(L, t) = T1 , (0 t < ) u(x, 0) = f (x), (0 x L) . (61)

3e. Temperature in a Slab or Bar. Suppose that you are interested in the temperatures in a slab of homogeneous material bounded by the planes x = 0 and x = L. The dimension in the y-direction can be negligibly small and the slab can go to innity in the z-direction.

x=0 z

x=L

The temperature in such a slab is only dependent on the coordinate x (and t) if the faces at x = 0 and x = L are held at the same temperature for all z. Alternatively, the formulation will be the same for temperatures in a bar of length L where the lateral dimensions are very small.

The function f (x) is given and is the initial temperature. Redene the temperature as v(x, t) = u(x, t) T0 (T1 T0 ) x . L (65)

Then the new function v(x, t) satises this B. V. P. : 1 v 2v = 0, (0 < x < L, 0 < t < ) x2 t v(0, t) = v(L, t) = 0, (0 t < ) (66) (67) (68)

x=0

x=L

x
and

The u(x, t) represent the temperature in such a system. Then, the temperature satises the diusion equation where n is the diusivity. As seen in the table in Sec. I on B. C. , the diusion equation has a unique solution for an open boundary with either Dirichlet or Neumann boundary conditions. There could also be a mixed B. C. with Dirichlet B. C. over part of the boundary and Neumann over the remaining part. The openness of boundary is guaranteed by the lack of closure in t at innity (0 t < ). Dirichlet B. C. at x = 0 and x = L, namely, u(0, t) = g(t) and u(L, t) = h(t) where g and h are known functions of time, correspond to knowing the temperature at the faces x = 0 and x = L. Of course, Neumann B. C. at x = 0 and x = L, that is, having u/t at x = 0

v(x, 0) = f (x) T0

Lx L

x T1 , (0 x L) . L

The new temperature v(x, t) has zero temperature at the faces. This makes the solution somewhat simpler. Of course, either of T0 or T1 , can be zero. The solution to the original B. V. P. , namely, u(x, t), can be obtained from eq. (68) after v(x, t) is obtained from the B. V. P. in eqs. (66), (67), and (68). The technique of separation of variables can be used. Assume v(x, t) = X(x)T (t) . (69)

17

18

MISN-0-486 Then, substitution of eq. (69) into eq. (66) yields T X = . X T Using the usual argument gives X X = 0 and The B. C. in eqs. (67) give T nT = 0, = constant . X(0) = X(L) = 0 . Thus, the solution of eq. (71) can be written as Xn (x) = Bn sin where nx , n = 1, 2, 3, . . . L

15

MISN-0-486 where

16

(70)

Bn =

2 L

L 0

f (x) T0

Lx L

T1

X nx dx . sin L L

(81)

(71) (72)

Case 2: This problem corresponds to thermally insulated faces. The B. V. P. is then 1 u 2u = 0, (0 < x < L, 0 < t < ) x2 t u(0, t) u(L, t) = = 0, (0 t < ) x x u(x, 0) = f (x), (0 xleqL) . (82) (83) (84)

(73)

(74)

The solution can be obtained in the same way as in Case 1 except the presence of the time derivatives in eq. (83) results in a Fourier cosine series on [0, L]. Thus, the solution is
1 u(x, t) = A0 A e 2 n=1 n

n2 2 . L2 Then, the solution of eq. (72) can be expressed as = Tn (t) = Cn en

2

(75)

n2 2 t nx L2 cos L

(85)

2 t/L2

(76)

where An = 2 L
0

As in the previous case of the wave equation, the solution to the diusion equation can be written as v(x, t) =
n=1

f (x) cos

nx dx, n = 0, 1, 2, 3, . . . L

(86)

Bn en

2 t/L2

sin

nx , B n = B n Cn . L

(77)

The last B. C. in eq. (68) can be satised by v(x, 0) = f (x) T0 and Bn =

L

3f. Temperatures in an Innite and Semi-innite Bar. As a last application of Fourier analysis to the solution of P. D. E. , the Fourier integrals will be used to nd temperatures in an innite and semi-innite bar. The B. V. P. for an innite bar is as follows:

Lx L

T1

X nx = Bn sin L L n=1

(78)

x=0

X Lx 2 nx dx . (79) f (x) T0 T1 sin L 0 L L L Thus, after some rearrangement the complete solution of the original B. V. P. in eqs. (61), (62), (63), and (64) is u(x, t) = T0 Lx L + T1 x B en + L n=1 n

t/L

sin

nx L

1 u 2u = 0, ( < x < , 0 < t < ) x2 t |u(x, t)| < M, ( < x < , 0 < t < ) and u(x, 0) = f (x), ( < x < ) .

(87) (88) (89)

20

(80)

19

MISN-0-486 Note: M is a constant, and f (x) is known.

17

MISN-0-486 v(0, t) = 0, 0 < t < ) v(x, 0) = f (x) T0 , 0 < x < ) . A solution to the P. D. E. is simply v (x, t) = v()e t sin x since v = 2 v t 2 v = 2 v . x2 v (0, t) = 0 .
2

18 (100) (101)

The bar extends to in the x-direction, but it has negligible crosssectional area so the temperature in the bar is a function of x alone (as well as time t). A solution to the P. D. E. can be written as u (x, t) = u()e t ex
2

(90)

(102)

since u /t = 2 u and 2 u /x2 = 2 u . The proportionality constant can depend on . Thus, a complete solution is 1 u(x, t) = 2 where

(103)

u()e t ex d

(91)

and

(104) (105)

1 u() = 2

f (x)e+x dx .

Also (92)

The B. V. P. for a semi-innite bar is as follows:

The last B. C. can be fullled by integrating over from 0 to . A factor of 2/ is included. Thus, v(x, t) = 2 2
0

v()e t sin x d

(106)

x=0

x
where v() =

1 u 2u = 0, (0 < x < , 0 < t < ) 2 x t |u(x, t)| < M, (0 < x < , 0 < t < ) u(0, t) = T0 , 0 < t < ) u(x, 0) = f (x), 0 < x < ) .

(93) (94) (95) (96)

[f (x) T0 ] sin x dx .

(107)

Thus, the solution to the original B. V. P. is u(x, t) = v(x, t) + T0 where v(x, t) is given by eqs. (106) and (107) above. (108)

The B. C. at x = 0 corresponds to a constant temperature on the face of the bar at x = 0. The problem can be reformulated in terms of v(x, t) where: (97) v(x, t) = u(x, t) T0 . Then, the new B. V. P. becomes 2v 1 v = 0, (0 < x < , 0 < t < ) x2 t |v(x, t)| < M , (0 < x < , 0 < t < ) (98) (99)
21

Acknowledgments
The author would like to thank Illinois State University for support in the construction of this lesson. Preparation of this module was supported in part by the National Science Foundation, Division of Science Education Development and Research, through Grant #SED 74-20088 to Michigan State University.

22

23

24