Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. Join them; it only takes a minute:

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

I need some help on solving this heat PDE :

Question : Consider a bar length L. The face at x=0 is insulated so that the heat flow across is zero, and the face at x=L is held at temperature u=0. The temperature distribution is governed by heat equation

$$ {\partial u \over \partial t} - k {\partial ^2 u \over \partial x^2} =0$$

Show that the normal modes of u(x,t) are

$$\large U_n(x,t)=B_n\cos[(2n-1)\pi x/2L]e^{[-(2n-1)^2\pi^2k^2t]/4L^2}$$

Given BCs are : u(0,t)=0 and u(L,t)=0

I already solve this PDE

and I got my answer in this form :

$$\large U_n(x,t)=B_n\sin[n\pi x/2L] e^{[-(n)^2\pi^2k^2t]/4L^2}$$

How am I need to change my answer into the form they asking ??? Anyone willing to help me ?

share|cite|improve this question
Isn't the B.C. at x=0 expressed as $u_t (0,t) = 0$, as the temperature at $x=0$ isn't changing due to the insulation? – Ron Gordon Jan 8 '13 at 17:52
No it cant. Because boundary condition is given in interval 0<x<L. The one you said only applies for ICs only. But in this case ICs are not given – Garett Jan 8 '13 at 17:55
I'm pretty sure rlgordonna is correct: if the heat flow across x=0 is 0, that means $u_t (0,t) = 0$. This should lead to the right eigenfunction, too. – AndrewG Jan 8 '13 at 18:06
Ok let me try it out first and then I confirm back to you again – Garett Jan 8 '13 at 18:11
I cant get it. I think it mucg be easier if you could could show me the steps on how you done it Andrew Gibson. Please help me out – Garett Jan 8 '13 at 18:18
up vote 3 down vote accepted

You need to express the problem as

$$u_t = k u_{xx} $$

$$ u_x(0,t) = 0 ,\;\;\; u(L,t) = 0 $$

Use separation of variables, i.e. $u(x,t) = X(x)T(t)$ and get

$$X'' + \frac{\lambda}{k} X = 0 $$

where $-\lambda$ is the separation constant. The solution to this equation is

$$ X(x) = A \cos{\left (\sqrt{\frac{\lambda}{k}} x \right)} + B \sin{\left (\sqrt{\frac{\lambda}{k}} x \right)}$$

The condition at $x=0$ implies that $B=0$:

$$ X'(x) = -\sqrt{\frac{\lambda}{k}} A \sin{\left (\sqrt{\frac{\lambda}{k}} x \right)} + B \sqrt{\frac{\lambda}{k}}\cos{\left (\sqrt{\frac{\lambda}{k}} x \right)}$$

$$X'(0) = B \sqrt{\frac{\lambda}{k}} = 0$$

The condition at $x=L$ implies that

$$\cos{\left (\sqrt{\frac{\lambda}{k}} L \right)} = 0$$

so that

$$\lambda = \left [ (2 n-1) \frac{\pi}{2 L} \right ]^2 k $$

You should be able to take it from there with the time equation.

share|cite|improve this answer
I just cant get the part where you do X'(x). Shouldnt the BCs $u_t(0,t)=O$ which is $X(0)T'(t)=0$?? Why you did X'(x) rlgordonma? Could you explain – Garett Jan 8 '13 at 18:37
My mistake - it should be $u_x(0,t) = 0$. I will fix it above. – Ron Gordon Jan 8 '13 at 18:40
For some reason, I mixed up holding a fixed temp., and heat flow, with is a temp. flux, i.e. a gradient, i.e., a spatial derivative. I hope that clears it up. – Ron Gordon Jan 8 '13 at 18:43
Ok now I understand. But just one thing could you teach me how you indentify it is $u_x(0,t)=0$ rather than $u_t(0,t)=0$ or $u(0,t)=0$. Tell me some tips so that I can understand clearly about this part – Garett Jan 8 '13 at 18:46
My comment above should explain some of it. The B.C.'s are described using certain terms. If the temp is held fixed to zero at one end, then $u=0$ at that end. If the temp is held fixed over time in general, then $u_t = 0$, which obviously returns a similar result to just setting $u=0$ at that end. (My mistake in being so glib before and confusing you.) If there is no heat flow, or the end is insulated, then $u_x = 0$ at that end. – Ron Gordon Jan 8 '13 at 18:52

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.