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 have a problem of heat distribution in a solid cylinder with the heater in the middle, which I take as $\exp(-r^2)$.

$$\frac{\partial u(t,r)}{\partial t}=a^2\frac{\partial^2 u(t,r)}{\partial r^2}+\frac{\partial u(t,r)}{r\partial r}+\exp(-r^2)$$

The initial and boundary conditions are the following.

  1. $u(0,r)=T_s=\text{const}$,

  2. $u(t,R)=T_e=\text{const}$,

  3. $0{\le}r{\le}R$,

  4. $a=\text{const}$.

I tried using Fourier series, but only the complex one seems to give the solution, but that gives complex values for temperature that is not what I expect.

Can anybody help me solve this? Thanks.

share|cite|improve this question
Can you include some of your working? – Chris Taylor Nov 21 '11 at 8:04
There should be parentheses around the $\partial^2 u$ and $\partial u$ terms on the right. Apart from that I don't understand your modeling of the heater. When the heater is a cylinder of radius $H<R$ then it causes a boundary condition $u(t, H)=T_H$. – Christian Blatter Nov 21 '11 at 11:42
What is your output? So you want this problem in matlab? – user82658 Jun 16 '13 at 4:23

First, subtracting $T_e$ gives for $v=u-T_e$ a problem with homogeneous initial and boundary conditions. The eigenfunctions for the Laplace operator in polar coordinates are the Bessel functions of the first kind $J_0\left( \frac{\mu_n}R r \right)$, where $\mu_n$ is the $n$-th positive root of $J_0$. They are orthogonal on $[0,R]$ with the weight function $r$. Denote $$ j_n(r)=\frac{J_0 \left( \frac{\mu_{n}}R x \right) }{\sqrt{\frac12} RJ_{1}(\mu_n)} $$ the normalized eigenfunctions, $\int_0^Rrj_m(r)j_n(r)\,dr=\delta_{mn}$.

The corresponding Green function for the heat equation in the polar coordinates is $$ G(r,r',t)=\sum_{n=1}^\infty e^{-a^2\frac{\mu_{n}^2}{R^2}t}j_n(r)j_n(r'). $$ Solution $v$ can be written out as a volume potential: $$ v(r,t)=\int_0^t\int_0^RG(r,r',t-\tau)r'e^{-r'^2}\,dr'd\tau. $$ Expanding $e^{-r^2}$ into the Fourier-Bessel series $$ e^{-r^2}=\sum_{n=1}^\infty c_nj_n(r),\quad r\in[0,R], $$ where $$ c_n=\int_0^Rrj_n(r)e^{-r^2}\,dr, $$ we have $$ v(r,t)= \int_0^t\int_0^Rr'\sum_{n=1}^\infty e^{-a^2\frac{\mu_{n}^2}{R^2}(t-\tau)}j_n(r)j_n(r')e^{-r'^2}\,dr'd\tau= $$ $$ \sum_{n=1}^\infty j_n(r)\left(\int_0^t e^{-a^2\frac{\mu_{n}^2}{R^2}(t-\tau)}d\tau\right) \left(\int_0^Rr'j_n(r')e^{-r'^2}\,dr'\right)= $$ $$ \frac {R^2}{a^2}\sum_{n=1}^\infty c_n\frac{1-e^{-a^2\frac{\mu_{n}^2}{R^2}t}}{\mu_n^2}j_n(r). $$

share|cite|improve this answer
Nice answer to an old question. +1 – Shuhao Cao Jun 16 '13 at 7:44

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.