Scribd
Upload
220 views4 pages

1D Diffusion Equation Solution

This document summarizes the solution of the 1D diffusion equation over time. It presents: 1) The steady state solution which is linear based on the boundary conditions. 2) For the time-dependent solution, excess concentration is considered as a Fourier series satisfying the diffusion equation and initial/boundary conditions. 3) The eigenfunctions and eigenvalues of the diffusion operator are derived, allowing the Fourier coefficients to be determined from the initial condition. 4) As an example, the full time-dependent solution is derived for a constant initial excess concentration.

Uploaded by

bruce1312
Copyright
© Attribution Non-Commercial (BY-NC)
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
220 views4 pages

1D Diffusion Equation Solution

This document summarizes the solution of the 1D diffusion equation over time. It presents: 1) The steady state solution which is linear based on the boundary conditions. 2) For the time-dependent solution, excess concentration is considered as a Fourier series satisfying the diffusion equation and initial/boundary conditions. 3) The eigenfunctions and eigenvalues of the diffusion operator are derived, allowing the Fourier coefficients to be determined from the initial condition. 4) As an example, the full time-dependent solution is derived for a constant initial excess concentration.

Uploaded by

bruce1312
Copyright
© Attribution Non-Commercial (BY-NC)
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd

Solution of the diusion equation in 1D

C 2C =D 2 t x

0x

(1)

Steady state

Setting C/t = 0 we obtain d2 C =0 dx2

Cs = ax + b

We determine a, b from the boundary conditions.

C (0) = C1 , It follows that

C ( ) = C2

(2)

b = C1 ,

a=

C2 C1

Cs (x) = Flux

C2 C1

x + C1 (3)

= D

C1 C2 Cs = x

Time-dependent solutions

We choose again the boundary conditions (2) and

C (x, 0) = C0 (x) as initial condition. It is convenient to consider the excess quantity u (x, t) = C (x, t) Cs (x) Using (1)-(3) we see that u satises u 2u =D 2 t x with

(4)

(5a)

(5b)

u (0) = u ( ) = 0 u (x, 0) = C0 Cs (x) u0 (x)

(5c) (5d)

Let m be the eigenfunctions of the diusion operator d2 /dx2 . Since the operator is dissipative, the correponding eigenvalues are non-positive. We denote them by 2 km (k real) d2 m (x) 2 = km m (x) dx2

(6)

Any function of the form u = Am (t) m satises then eq.(5), provided that Am (t) satises the ordinary dierential equation dAm 2 = Dkm Am dt or Am (t) = Am (0) eDkm t
2

(7a) (7b)

On the other hand, in general, functions u of this form do not satisfy the initial condition. To satisfy this condition we seek for solutions in the form of an innite series of m s (this is legitimate since the equation is linear) 2

u (x, t) =
m

Am (t) m (x)

(8)

and x the Am (0)s by requiring that Am (0) m (x) = u (x, 0) = u0 (x)


m

(9)

To compute Am (0) we use the orthogonality property of m , guaranted by the fact that the diusion operator is self-adjoint : dx (x)m (x) = 0 n = m n
0

(10)

= Nn

n=m

Multiplying both sides of (9) by (x) and integrating over x we thus obtain m dx (x)u0 (x) n Nn

An (0) =

(11)

which combined with (7b) and (8) yields the solution dx (x)u0 (x) n m (x) Nm

u(x, t) =
m

(12)

We now compute m and km explicitly for the boundary conditions (5b). On inspecting (6) we see that m must be of the form m (x) = Ccos km x + Dsin km x Applied to x = 0 and x = this leads to

0 = m (0) = C 0 = m ( ) = Dsin km 3

implying that km = m (m integer) or km = m mx (up to a factor) (13)

m = sin

the full solution u(x, t) being (cf. eq. (12)), with Nm = /2 2


m o

u(x, t) =

dxsin

mx

u0 (x) sin

mx

eD

m2 2 2 t

(14)

(Fourrier series) To complete the evaluation suppose that u0 (x) = u0 = constant, then, mx

u0
0

dxsin

= 0 = 2uo m

if m is even if m is odd = 2n + 1 (15)

Finally, 4u0 u(x, t) =

(2n+1)2 ?pi2 (2n + 1)x 1 t 2 sin eD 2n + 1

(16)
2

n=0

Notice that u(x, t) 0 as t with a characteristic time t =

D 2

You might also like