Comparing two solutions of 1-D heat equation

55 Views Asked by At

Considering this problem

$$\frac{\partial^{2} T(x, t)}{\partial x^{2}}=\frac{1}{\alpha} \frac{\partial T(x, t)}{\partial t} \quad \text { in } \quad 0<x<\infty, \quad t>0$$

$$\begin{array}{ll} \mathrm{BC} 1: & T(x=0)=f(\tau)=\sin(\omega \tau) \\ \mathrm{BC} 2: & T(x \rightarrow \infty)=0 \\ \mathrm{IC}: & T(x, t=0)=0 \end{array}$$

The solution is given by $$T(x, t)=\frac{x}{\sqrt{4 \pi \alpha}} \int_{\tau=0}^{t} \frac{\sin(\omega\tau)}{(t-\tau)^{3 / 2}} \exp \left[\frac{-x^{2}}{4 \alpha(t-\tau)}\right] d \tau$$

However I have another solution which is given by a much compact and easy form

$$T(x, t)= e^{-x / d}[\sin (\omega t-x / d)]$$

$$d=(2 \alpha / \omega)^{\frac{1}{2}}$$

assuming initial condition is $T(x, t=0)=e^{-x/d}\sin(-x/d)$.

Then when $t\rightarrow \infty$ this two solutions will converge.

My question is what's the relationship between the two equations. How to obtain the latter solution?

1

There are 1 best solutions below

0
On

Let's consider the derivation of the solution. First, introduce the Laplace transform, $$ U(x,s) = \int_{0}^{\infty}\mathrm{d}t ~ \mathrm{e}^{-st}~T(x,t). $$ Using this, the 1D heat equation can rewritten as $$ \alpha\frac{\partial^2 U}{\partial x^2} = sU(x,s) - T(x,0). $$ For your first problem, because $T(x,0) = 0$, you can solve this equation as $$ U(x,s) = A\mathrm{e}^{-x\sqrt{s/\alpha}}, $$ where $A$ is a constant determined by BC. Perform the inverse Laplace transform, and you can obtain the solution.

However, in your second problem, since $T(x,0)$ is nonzero, you can not obtain $U$. So, let's assume $$ T = \mathrm{Re}[g(x)\mathrm{e}^{-i\omega t}], $$ where $g(x)$ is a complex–valued function. Therefore, the heat equation becomes \begin{align} g''(x) - \frac{i\omega}{\alpha}g(x) &= 0 \\ g(0) &= i \\ g(x \to \infty) &= 0. \end{align} Solve these equations, and you can obtain the solution.