Approximate and Exact Solutions for a Thermal Conductivity Problem

[hitmeoff]

Homework Statement

A slab is in a steady state with temperature T₀ at x = 0, and T₁ at x = 1. The thermal conductivity is given by K(x) = K₀e^\epsilonx where |\epsilon| << 1. The governing
equation is given by, \frac{d}{dx}(K₀e^\epsilonx \frac{dT}{dx}) = 0

(1). Obtain an approximation solution to the temperature distribution by replacing K(x)
with its average value \bar{K} = \frac{\int K(x) dx}{\int dx} over the slab (integrals from 0 to 1)

(2). Otain an exact solution to the temperature distribution.

(3). Rewrite K(x) = K(x) − \bar{K} + \bar{K} then term K(x) − \bar{K} is neglected while replacing K(x) with \bar{K}. Check consistency, i.e, prove that |\frac{K(x) - \bar{K}}{\bar{K}}| << 1

Homework Equations

The Attempt at a Solution

Any hints on how to even start?

 

[hitmeoff]

anyone?

 

[lanedance]

for 1) how about trying to perform the given integral to get the average conductivity

 

[hitmeoff]

ok, I got the average conductivity to be the constant: (e^\epsilon - 1) (\frac{K0}{\epsilon}) for \overline{K}

So when I sub this back in I get: \frac{d}{dx} ( \overline{K} \frac{dT}{dx} ) = 0

or simply \overline{K} \frac{d2T}{d2x} = 0

Now the solution of this equation should be T(x) = c₁x

 

[lanedance]

ok, your tex is a little difficult to read (click on equation below to see code), but i agree if this is what you meant, which i suppose makes it a little redundant ;)
\overline{K} = \int_0^1 dx K(x) = \int_0^1 dx K_0 e^{\epsilon x} = \frac{K}{\epsilon}(e^{\epsilon}-1)

whats next?

 

[hitmeoff]

ok, your tex is a little difficult to read (click on equation below to see code), but i agree if this is what you meant, which i suppose makes it a little redundant ;)
\overline{K} = \int_0^1 dx K(x) = \int_0^1 dx K_0 e^{\epsilon x} = \frac{K}{\epsilon}(e^{\epsilon}-1)

whats next?

Yeah, that's what I get.

So when I sub this back in I get: \frac{d}{dx} ( \overline{K} \frac{dT}{dx} ) = 0

or simply \overline{K} \frac{d2T}{d2x} = 0

Now the solution of this equation should be T(x) = c₁x with c₁ being \overline{K} ?

 

[lanedance]

first you can write a whole equation between the tex tags

now second what are you trying to do? i would think a constant conductivity would lead to a constant temperature gradient, whilst the conductivity dependent on x will lead to slightly more complex temperture distribution. What is the exact solution of? I would take it to be the full temp gradient, but haven't tried to solve it

 

[hitmeoff]

Well for one, I am supposed to just get an approximate solution for the Temp distribution.

For part II, I am supposed to get the exact solution. Let me work on that for a bit.

 

[hitmeoff]

So for part 2, I "expanded" the D.E. and rewrote it like:

K₀ \epsilon e^\epsilonx \frac{dT}{dx} + K₀e^\epsilonx\frac{d2T}{dx2} = 0 or
T'' + \epsilon T' = 0 ...correct?