I solved the following partial differential equation using Laplace Transform:

$\LARGE \frac{1}{r_D}\frac{\partial}{\partial r_D}(r_D\frac{\partial p_D}{\partial r_D})=\frac{\partial p_D}{\partial t_D}$

with following initial and boundary conditions:

Initial condition: $p_D(r_D,0) = 0$

Boundary condition 1: $p_D(r_D->\infty, t_D) = 0$

Boundary condition 2: $p_D(r_D=1, t_D) = 1$

I applied Laplace Inverse to the above problem and now I have reached the following step:

How do I find Laplace inverse of the following:

$\LARGE \bar p_D = \frac{K_0(\sqrt[]s r_D)}{sK_0(\sqrt[]s)}$

I need to prove an equation for which I need this. I tried do it using Mathematica but it is also not helping.

  • $\begingroup$ Is $K_0$ the modified Bessel function of the second kind? $\endgroup$ – Mhenni Benghorbal Aug 30 '13 at 22:00
  • $\begingroup$ @Mhenni: Yes, it is! $\endgroup$ – Lee Aug 30 '13 at 22:03
  • $\begingroup$ Where did this problem come from? $\endgroup$ – Mhenni Benghorbal Aug 30 '13 at 22:04
  • $\begingroup$ @MhenniBenghorbal : I have updated the main question. I solved a pde to get it. $\endgroup$ – Lee Aug 30 '13 at 22:24
  • $\begingroup$ This is the heat equation in 2D? Isn't it? $\endgroup$ – abnry Aug 30 '13 at 22:30

With some differences I will point out along the way, the development of the solution of the above 2D heat equation follows the solution of the 1D heat equation for the semi-infinite rod. If you don't mind, I will develop the solution from the beginning, with my own notation which I will define below. The result is not pretty, but is at least in a calculable form.

The Equation

The problem is to solve the heat equation in the region $x^2+y^2 \gt 1$, where the temperature $u$ in the unit circle is held fixed at unity. Because there is no angular dependence, the equation becomes 1D in the radial coordinate $r=\sqrt{x^2+y^2}$:

$$\frac{\partial^2 u}{\partial r^2} + \frac{1}{r} \frac{\partial u}{\partial r} = \frac{\partial u}{\partial t}$$ $$u(1,t)=1$$ $$\lim_{r \to \infty} u(r,t)=0$$ $$u(r,0)=0 \quad r >1$$

The solution

The above equation lends itself to a Laplace transform solution. Define

$$\hat{u}(r,s) = \int_0^{\infty} dt \, e^{-s t} \, u(r,t)$$

The Laplace transform turns the PDE above into an ODE, as follows:

$$\frac{d^2 \hat{u}}{d r^2} + \frac{1}{r} \frac{d \hat{u}}{d r} - s \hat{u}=0$$ $$\hat{u}(1,s) = \frac{1}{s}$$ $$\lim_{r \to \infty} \hat{u}(r,s)=0$$

The general solution to the first equation is

$$\hat{u}(r,s) = A I_0(\sqrt{s} r) + B K_0(\sqrt{s} r)$$

where $I_0$ and $K_0$ are the modified Bessel functions of the first and second kind, respectively. Applying the boundary conditions, we find that the solution to the above equation becomes

$$\hat{u}(r,s) = \frac{K_0(\sqrt{s} r)}{s K_0(\sqrt{s})}$$

as shown in the problem statement. The solution to the heat equation is obtained upon performing an inverse Laplace transform:

$$u(r,t) = \frac{1}{i 2 \pi} \int_{c-i \infty}^{c+i \infty} \frac{ds}{s} e^{s t} \frac{K_0(\sqrt{s} r)}{K_0(\sqrt{s})}$$

The inverse transform

As the OP pointed out, the above is not a very useful way to express the solution. We therefore will use Cauchy's theorem and define a contour in the complex plane over which we may express the above ILT in terms of the value of an integral over other parts of the contour.

Consider the following integral in the complex $z$ plane:

$$\oint_C \frac{dz}{z} e^{t z} \frac{K_0(\sqrt{z} r)}{K_0(\sqrt{z})}$$

where $C$ is the following contour:


We will define $\text{Arg}{z} \in (-\pi,\pi]$, so the branch is the negative real axis. There are $6$ pieces to this contour, $C_k$, $k \in \{1,2,3,4,5,6\}$, as follows.

$C_1$ is the contour along the line $z \in [c-i R,c+i R]$ for some large value of $R$.

$C_2$ is the contour along a circular arc of radius $R$ from the top of $C_1$ to just above the negative real axis.

$C_3$ is the contour along a line just above the negative real axis between $[-R, -\epsilon]$ for some small $\epsilon$.

$C_4$ is the contour along a circular arc of radius $\epsilon$ about the origin.

$C_5$ is the contour along a line just below the negative real axis between $[-\epsilon,-R]$.

$C_6$ is the contour along the circular arc of radius $R$ from just below the negative real axis to the bottom of $C_1$.

We will show that the integral along $C_2$and $C_6$ vanish in the limits of $R \rightarrow \infty$ and $\epsilon \rightarrow 0$.

First of all, please note that the Bessel function in the integrand has the following asymptotic behavior:

$$K_0(w) \sim \sqrt{\frac{\pi}{2 w}} e^{-w} \quad (w \to \infty)$$

On $C_2$, the real part of the argument of the exponential is

$$R t \cos{\theta} - (r-1)\sqrt{R} \cos{\frac{\theta}{2}}$$

where $\theta \in [\pi/2,\pi)$. Clearly, $\cos{\theta} < 0$ and $\cos{\frac{\theta}{2}} > 0$, so that the integrand exponentially decays as $R \rightarrow \infty$ and therefore the integral vanishes along $C_2$.

On $C_6$, we have the same thing, but now $\theta \in (-\pi,-\pi/2]$. This means that, due to the evenness of cosine, the integrand exponentially decays again as $R \rightarrow \infty$ and therefore the integral also vanishes along $C_6$.

Thus, we are left with the following by Cauchy's integral theorem (i.e., no poles inside $C$):

$$\left [ \int_{C_1} + \int_{C_3} + \int_{C_4} + \int_{C_5}\right] \frac{dz}{z} \: e^{z t} \frac{K_0(\sqrt{z} r)}{K_0(\sqrt{z})} = 0$$

On $C_4$, note that the integral is finite and nonzero as $\epsilon$ in the limit $\epsilon \rightarrow 0$:

$$\int_{C_4} \frac{dz}{z} \exp{(t z)} \frac{K_0(\sqrt{z} r)}{K_0(\sqrt{z})} \sim i \int_{\pi}^{-\pi} d\phi = -i 2 \pi $$

On $C_3$, we parametrize by $z=e^{i \pi} x$ and the integral along $C_3$ becomes

$$\int_{C_3} \frac{dz}{z} \exp{(t z)} \frac{K_0(\sqrt{z} r)}{K_0(\sqrt{z})} = \int_{\infty}^0 \frac{dx}{x} \: e^{-x t} \frac{K_0(i \sqrt{x} r)}{K_0(i \sqrt{x})}$$

On $C_5$, however, we parametrize by $z=e^{-i \pi} x$ and the integral along $C_5$ becomes

$$\int_{C_5} \frac{dz}{z} \exp{(t z)} \frac{K_0(\sqrt{z} r)}{K_0(\sqrt{z})} = \int_0^{\infty} \frac{dx}{x} \:e^{-x t} \frac{K_0(-i \sqrt{x} r)}{K_0(-i \sqrt{x})}$$

We may now write

$$-1 +\frac{1}{i 2 \pi} \int_0^{\infty} \frac{dx}{x} \: e^{- x t} \left [ \frac{K_0(-i \sqrt{x} r)}{K_0(-i \sqrt{x})} - \frac{K_0(i \sqrt{x} r)}{K_0(i \sqrt{x})} \right ] + \frac{1}{i 2 \pi} \int_{c-i \infty}^{c+i \infty} \frac{ds}{s} \: e^{s t} \frac{K_0( \sqrt{s} r)}{K_0(\sqrt{s})} = 0$$

Therefore, the ILT of $\hat{u}(r,s) = \frac{K_0(\sqrt{s} r)}{s K_0(\sqrt{s})}$ is given by

$$\begin{align}\frac{1}{i 2 \pi} \int_{c-i \infty}^{c+i \infty} \frac{ds}{s} \: e^{s t} \frac{K_0(\sqrt{s} r)}{K_0(\sqrt{s})} &= 1-\frac{1}{i 2 \pi} \int_0^{\infty} \frac{dx}{x} \: e^{- x t} \left [\frac{K_0(-i \sqrt{x} r)}{K_0(-i \sqrt{x})} - \frac{K_0(i \sqrt{x} r)}{K_0(i \sqrt{x})} \right ]\\ &=1- \frac{2}{\pi} \int_{0}^{\infty} \frac{du}{u} \: \,e^{-t u^2} \frac{J_0(u) Y_0(r u)-J_0(r u) Y_0(u)}{J_0^2(u)+Y_0^2(u)} \end{align}$$

where $J_0$ and $Y_0$ are the Bessel functions of the first and second kind, respectively. The last step is a result of the identity

$$K_0(i b) = -\frac{\pi}{2} [Y_0(b)+i J_0(b)]$$

How do I know that the above integral even converges? Using the properties of these Bessels near the origin, I get that the integrand behaves as $2 \log{r}/(u \log^2{u})$ as $u \to 0^+$. Therefore, while there is a singularity at the origin, that singularity is indeed integrable. (This non-trivial result delayed this answer by several hours.)

This is about as far as I can take the integral so far. I will post additional results if I can simplify it.


The integral form derived for the solution takes the form of something called an inverse Weber transform (see this PDF reference, for example, on the second page). Essentially, by using a Laplace transform approach, I derived the expression for the inverse Weber transform. This of course is no help in figuring out an analytical expression for the solution, but it provides a further sanity check.

Here are some plots of the temperature $u(r,t)$ for $r \in (1,4]$ and $t \in \{0.01,0.1,0.5,1.0,10\}$:

enter image description here

These were generated in Mathematica directly from the integral expression above.


I forgot to mention that in the above, it should be clear that I assumed that $t \gt 0$. When $t \lt 0$, we must close the contour $C$ to the right; $c$ is defined such that the line $C_1$ lies to the right of any poles or branch cuts. Thus, the ILT is zero when $t \lt 0$.

| cite | improve this answer | |

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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