How to start Laplace infinite domain

Click For Summary

Discussion Overview

The discussion revolves around solving a thermal conduction problem in an infinite domain with a circular region of different conductivity. Participants explore the mathematical formulation of the temperature fields inside and outside the circular region, represented by series expansions, and the boundary conditions that must be satisfied.

Discussion Character

  • Technical explanation
  • Mathematical reasoning
  • Debate/contested

Main Points Raised

  • One participant expresses uncertainty about how to start solving the problem involving a thermal gradient and circular region with different conductivities.
  • Another participant proposes series representations for the temperature fields, suggesting that the inner solution must be bounded at the origin and the outer solution must approach the undisturbed field at infinity.
  • There is a discussion about the conditions under which the series coefficients can be equated, with some participants questioning the validity of certain assumptions made regarding the coefficients.
  • Participants discuss the implications of boundary conditions on the coefficients and the need to eliminate terms to find a solution.
  • There is a contention regarding the treatment of the coefficient \(A_0\) and whether it can be treated as a variable, with some participants suggesting it should be set to zero.
  • One participant suggests using the second boundary condition to derive further relationships between the coefficients.
  • Disagreement arises over the relevance of earlier posts and the correct path forward in solving for the coefficients.

Areas of Agreement / Disagreement

Participants do not reach a consensus on the correct approach to take for solving the problem. There are multiple competing views on how to handle the series coefficients and the boundary conditions, leading to an unresolved discussion.

Contextual Notes

Participants express uncertainty about the assumptions made regarding the coefficients and the implications of boundary conditions. There are unresolved mathematical steps and dependencies on the definitions of the coefficients.

Dustinsfl
Messages
2,217
Reaction score
5
This problem seems a little overwhelming at the point. I am not sure on where and how to start.

Suppose that a uniform thermal gradient in the +x direction exists in a very large (i.e. effectively infinite) domain of conductivity $k_2$ such that the temperature field $u_{\infty}(r,\theta)$ can be represented by
$$
u_{\infty}(r,\theta) = Ar\cos\theta
$$
where $A$ is a constant which reflects the gradient magnitude.
Suppose that a circular region of material of radius a is removed an replaced with a new material with conductivity $k_1$.
The most general version
of this problem is when the inner and out conductivities are of unequal, but comparable, magnitudes.
In this case, the steady temperature field inside the disk and outside the disk must be solved separately to obtain the inner and outer solutions, $u_1$ and $u_2$ respectively.
The constraint on the inner solution is boundedness at the origin
$$
\lim_{r\to 0}|u_1(r,\theta)| < \infty.
$$
The outer solution must asymptotically approach the undisturbed temperature field at large distances:
$$
\lim_{r\to\infty}u_2(r,\theta)\to u_{\infty}(r,\theta).
$$
Each of these solutions will contain series coefficients that must be determined by jointly imposing continuity of temperature and heat flux at the boundary $r = a$:
\begin{alignat*}{3}
u_1(a,\theta) & = & u_1(a,\theta)\\
k_1u_{1_r}(a,\theta) & = & k_2u_{2_r}(a,\theta)
\end{alignat*}Obtain the solutions for the temperature fields $u_1$ and $u_2$ with series coefficients expressed terms of $a$, $k_1$ and $k_2$.
 
Physics news on Phys.org
$$
u_1 = A_0 + \sum_{n=1}^{\infty}r^n(A_n^{(1)}\cos n\theta + B_n^{(1)}\sin n\theta)
$$
$$
u_2 = U_0r\cos\theta + \sum_{n=1}^{\infty}\frac{1}{r^n}(A_n^{(2)}\cos n\theta + B_n^{(2)}\sin n\theta)
$$
The equation $Ar\cos\theta$ has been changed to $U_0r\cos\theta$ and $u_1(a,\theta) = u_2(a,\theta)$. Therefore, $A_0 = U_0a\cos\theta$. Then
$$
\sum_{n=1}^{\infty}\left[a^n(A_n^{(1)}\cos n\theta + B_n^{(1)}\sin n\theta)-\frac{1}{a^n}(A_n^{(2)}\cos n\theta + B_n^{(2)}\sin n\theta)\right] = 0
$$
Only when $a^nA_n^{(1)} = \frac{1}{a^n}A_n^{(2)}$ and $a^nB_n^{(1)}=\frac{1}{a^n}B_n^{(2)}$ <--Is this correct to do?
From this point (if it is correct), can be simplified more now? Or must I use the next condition $-k_1\frac{\partial u_1}{\partial r}(a,\theta)=-k_2\frac{\partial u_2}{\partial r}(a,\theta)$?
 
Last edited:
dwsmith said:
$$
\sum_{n=1}^{\infty}\left[a^n(A_n^{(1)}\cos n\theta + B_n^{(1)}\sin n\theta)-\frac{1}{a^n}(A_n^{(2)}\cos n\theta + B_n^{(2)}\sin n\theta)\right] = 0
$$
Only when $a^nA_n^{(1)} = \frac{1}{a^n}A_n^{(2)}$ and $a^nB_n^{(1)}=\frac{1}{a^n}B_n^{(2)}$ <--Is this correct to do?

Hi dwsmith, :)

When $a^nA_n^{(1)} = \frac{1}{a^n}A_n^{(2)}$ and $a^nB_n^{(1)}=\frac{1}{a^n}B_n^{(2)}$ we have,

\[\sum_{n=1}^{\infty}\left[a^n(A_n^{(1)}\cos n\theta + B_n^{(1)}\sin n\theta)-\frac{1}{a^n}(A_n^{(2)}\cos n\theta + B_n^{(2)}\sin n\theta)\right] = 0\]

But we cannot say that the above sum is equal to zero only when $a^nA_n^{(1)} = \frac{1}{a^n}A_n^{(2)}$ and $a^nB_n^{(1)}=\frac{1}{a^n}B_n^{(2)}$. For example another possibility that could make the sum equal to zero is,

\[a^n(A_n^{(1)}\cos n\theta + B_n^{(1)}\sin n\theta)-\frac{1}{a^n}(A_n^{(2)}\cos n\theta + B_n^{(2)}\sin n\theta)=0\mbox{ for each }n\geq 1\]

with $a^nA_n^{(1)} \neq \frac{1}{a^n}A_n^{(2)}$ and $a^nB_n^{(1)}\neq\frac{1}{a^n}B_n^{(2)}$.

Kind Regards,
Sudharaka.
 
Sudharaka said:
Hi dwsmith, :)

When $a^nA_n^{(1)} = \frac{1}{a^n}A_n^{(2)}$ and $a^nB_n^{(1)}=\frac{1}{a^n}B_n^{(2)}$ we have,

\[\sum_{n=1}^{\infty}\left[a^n(A_n^{(1)}\cos n\theta + B_n^{(1)}\sin n\theta)-\frac{1}{a^n}(A_n^{(2)}\cos n\theta + B_n^{(2)}\sin n\theta)\right] = 0\]

But we cannot say that the above sum is equal to zero only when $a^nA_n^{(1)} = \frac{1}{a^n}A_n^{(2)}$ and $a^nB_n^{(1)}=\frac{1}{a^n}B_n^{(2)}$. For example another possibility that could make the sum equal to zero is,

\[a^n(A_n^{(1)}\cos n\theta + B_n^{(1)}\sin n\theta)-\frac{1}{a^n}(A_n^{(2)}\cos n\theta + B_n^{(2)}\sin n\theta)=0\mbox{ for each }n\geq 1\]

with $a^nA_n^{(1)} \neq \frac{1}{a^n}A_n^{(2)}$ and $a^nB_n^{(1)}\neq\frac{1}{a^n}B_n^{(2)}$.

Kind Regards,
Sudharaka.

Then how do I obtain the series coefficients in terms of $a$, $k_1$, and $k_2$?
 
dwsmith said:
Then how do I obtain the series coefficients in terms of $a$, $k_1$, and $k_2$?

Not sure whether this is the correct approach, but I think the series coefficients can be found using the remaining boundary condition. What do you get after using that boundary condition?
 
Sudharaka said:
Not sure whether this is the correct approach, but I think the series coefficients can be found using the remaining boundary condition. What do you get after using that boundary condition?
Using the second boundary condition, we have
\begin{alignat*}{3}
-k_1\frac{\partial u_1}{\partial r}(a,\theta) & = & - k_1\sum\limits_{n = 1}^{\infty}na^{n - 1}\left(A_n^{(1)}\cos n\theta + B_n^{(1)}\sin n\theta\right)\\
& = & -k_2U_0\cos\theta + k_2\sum\limits_{n = 1}^{\infty}\frac{n}{a^{n + 1}}\left(A_n^{(2)}\cos n\theta + B_n^{(2)}\sin n\theta\right)\\
& = & -k_2\frac{\partial u_2}{\partial r}(a,\theta).
\end{alignat*}
 
dwsmith said:
Using the second boundary condition, we have
\begin{alignat*}{3}
-k_1\frac{\partial u_1}{\partial r}(a,\theta) & = & - k_1\sum\limits_{n = 1}^{\infty}na^{n - 1}\left(A_n^{(1)}\cos n\theta + B_n^{(1)}\sin n\theta\right)\\
& = & -k_2U_0\cos\theta + k_2\sum\limits_{n = 1}^{\infty}\frac{n}{a^{n + 1}}\left(A_n^{(2)}\cos n\theta + B_n^{(2)}\sin n\theta\right)\\
& = & -k_2\frac{\partial u_2}{\partial r}(a,\theta).
\end{alignat*}

Eliminate \(A_{n}^{(2)}\) and \(B_{n}^{(2)}\) using, \(a^nA_n^{(1)} = \frac{1}{a^n}A_n^{(2)}\) and \(a^nB_n^{(1)}=\frac{1}{a^n}B_n^{(2)}\).
 
dwsmith said:
$$
u_1 = A_0 + \sum_{n=1}^{\infty}r^n(A_n^{(1)}\cos n\theta + B_n^{(1)}\sin n\theta)
$$
$$
u_2 = U_0r\cos\theta + \sum_{n=1}^{\infty}\frac{1}{r^n}(A_n^{(2)}\cos n\theta + B_n^{(2)}\sin n\theta)
$$
The equation $Ar\cos\theta$ has been changed to $U_0r\cos\theta$ and $u_1(a,\theta) = u_2(a,\theta)$. Therefore, $A_0 = U_0a\cos\theta$. Then
$$
\sum_{n=1}^{\infty}\left[a^n(A_n^{(1)}\cos n\theta + B_n^{(1)}\sin n\theta)-\frac{1}{a^n}(A_n^{(2)}\cos n\theta + B_n^{(2)}\sin n\theta)\right] = 0
$$

You have taken, \(A_{0}=U_{0}a\cos\theta\) (which is problematic because a Fourier coefficient cannot be a variable) and neglected that part altogether to obtain,

\[\sum_{n=1}^{\infty}\left[a^n(A_n^{(1)}\cos n\theta + B_n^{(1)}\sin n\theta)-\frac{1}{a^n}(A_n^{(2)}\cos n\theta + B_n^{(2)}\sin n\theta)\right] = 0\]

Is it given that, \(A_{0}=U_{0}=0\)?
 
Sudharaka said:
Is it given that, \(A_{0}=U_{0}=0\)?

No.
 
  • #10
$$
k_2U_0\cos\theta = \sum_{n=1}^{\infty}\left[(k_1 + k_2)na^{n-1}A_n\cos n\theta + (k_1 + k_2)na^{n-1}B_n\sin n\theta\right]
$$
Now what?
So
$$
k_2U_0\cos\theta = (k_1+k_2)A\cos\theta\Rightarrow A_1^{(1)} = \frac{k_2U_0}{k_1+k_2}
$$
correct?

How do I find the other coefficients?
 
Last edited:
  • #11
dwsmith said:
No.

Then you cannot obtain,

\[\sum_{n=1}^{\infty}\left[a^n(A_n^{(1)}\cos n\theta + B_n^{(1)}\sin n\theta)-\frac{1}{a^n}(A_n^{(2)}\cos n\theta + B_n^{(2)}\sin n\theta)\right] = 0\]

in post #2. Post #3-#7 and #10 becomes irrelevant and you should try to continue from post #2. :)
 
  • #12
Sudharaka said:
Then you cannot obtain,

\[\sum_{n=1}^{\infty}\left[a^n(A_n^{(1)}\cos n\theta + B_n^{(1)}\sin n\theta)-\frac{1}{a^n}(A_n^{(2)}\cos n\theta + B_n^{(2)}\sin n\theta)\right] = 0\]

in post #2. Post #3-#7 and #10 becomes irrelevant and you should try to continue from post #2. :)

I have tried many different methods already any better advice?
 
  • #13
So you have,

\[u_1 = A_0 + \sum_{n=1}^{\infty}r^n(A_n^{(1)}\cos n\theta + B_n^{(1)}\sin n\theta)\]

\[u_2 = U_0r\cos\theta + \sum_{n=1}^{\infty}\frac{1}{r^n}(A_n^{(2)}\cos n\theta + B_n^{(2)}\sin n\theta)\]

with the boundary conditions,

\[u_1(a,\theta) = u_2(a,\theta)\mbox{ and }-k_1\frac{\partial u_1}{\partial r}(a,\theta)=-k_2\frac{\partial u_2}{\partial r}(a,\theta)\]

From the first boundary condition we get,

\[A_0 + \sum_{n=1}^{\infty}a^n(A_n^{(1)}\cos n\theta + B_n^{(1)}\sin n\theta)=U_0a\cos\theta + \sum_{n=1}^{\infty}\frac{1}{a^n}(A_n^{(2)}\cos n\theta + B_n^{(2)}\sin n\theta)\]

\[\Rightarrow A_0 -U_0a\cos\theta+\sum_{n=1}^{\infty}\left(a^n A_{n}^{(1)}-\frac{A_{n}^{(2)}}{a^n}\right)\cos n\theta+ \sum_{n=1}^{\infty}\left(a^n B_{n}^{(1)}-\frac{B_{n}^{(2)}}{a^n}\right)\sin n\theta=0\]

\[\Rightarrow A_0 +\left(aA_{1}^{(1)}-\frac{A_{1}^{(2)}}{a}-U_0a\right)\cos\theta+\sum_{n=2}^{\infty}\left(a^n A_{n}^{(1)}-\frac{A_{n}^{(2)}}{a^n}\right)\cos n\theta+ \sum_{n=1}^{\infty}\left(a^n B_{n}^{(1)}-\frac{B_{n}^{(2)}}{a^n}\right)\sin n\theta=0\]

This is satisfied when,

\[A_{0}=0\]

\[aA_{1}^{(1)}-\frac{A_{1}^{(2)}}{a}-U_0a=0\]

\[a^n A_{n}^{(1)}-\frac{A_{n}^{(2)}}{a^n}=0\mbox{ for all }n\geq 2\]

\[a^n B_{n}^{(1)}-\frac{B_{n}^{(2)}}{a^n}=0\mbox{ for all }n\geq 1\]

Try to do the same thing for the second boundary condition and you should get three equations corresponding to the coefficients of \(\cos\theta\), \(\cos n\theta\) and \(\sin n\theta\).
 
  • #14
We have that $u_1 = A_0^{(1)} + \sum\limits_{n = 1}^{\infty}r^n\left(A_n^{(1)}\cos n\theta + B_n^{(1)}\sin n\theta\right)$ and $u_2 = u_{\infty} + u_2' = U_0r\cos\theta + \sum\limits_{n = 1}^{\infty}\frac{1}{r^n}\left(A_n^{(2)}\cos n\theta + B_n^{(2)}\sin n\theta\right)$.
Using the first boundary condition, we have
\begin{alignat*}{3}
u_1(a,\theta) & = & A_0^{(1)} + \sum\limits_{n = 1}^{\infty}a^n\left(A_n^{(1)}\cos n\theta + B_n^{(1)}\sin n\theta\right)\\
& = & U_0a\cos\theta + \sum\limits_{n = 1}^{\infty}\frac{1}{a^n}\left(A_n^{(2)}\cos n\theta + B_n^{(2)}\sin n\theta\right)\\
& = & u_2(a,\theta).
\end{alignat*}
Using the second boundary condition, we have
\begin{alignat*}{3}
k_1\frac{\partial u_1}{\partial r}(a,\theta) & = & k_1\sum\limits_{n = 1}^{\infty}na^{n - 1}\left(A_n^{(1)}\cos n\theta + B_n^{(1)}\sin n\theta\right)\\
& = & k_2U_0\cos\theta - k_2\sum\limits_{n = 1}^{\infty}\frac{n}{a^{n + 1}}\left(A_n^{(2)}\cos n\theta + B_n^{(2)}\sin n\theta\right)\\
& = & k_2\frac{\partial u_2}{\partial r}(a,\theta).
\end{alignat*}
From the first boundary condition, we have
$$
\sum\limits_{n = 1}^{\infty}\left(a^nA_n^{(1)} - \frac{1}{a^n}A_n^{(2)}\right) \cos n\theta = U_0a\cos\theta
$$
where $B_n = A_0^{(1)} = 0$ since there are no sine and constant terms in the equality.
That is, we have that
$$
aA_1^{(1)} - \frac{1}{a}A_1^{(2)} = U_0a\iff A_1^{(1)} = U_0 + \frac{1}{a^2}A_1^{(2)}.
$$
From the second boundary condition, we have
$$
k_1A_1^{(1)} + \frac{k_2}{a^2}A_1^{(2)} = k_2U_0.
$$
We can now plug in our $A_1^{(1)}$ term.
Then we have that $A_1^{(2)} = \frac{a^2U_0\left(1 - \frac{k_1}{k_2}\right)}{\frac{k_1}{k_2} + 1}$ and plugging $A_1^{(2)}$ into the other equation we have $A_1^{(1)} = U_0\left[1 + \frac{1 - \frac{k_1}{k_2}}{\frac{k_1}{k_2} + 1}\right]$
Therefore, the general solution is
$$
u(r,\theta) = u_1 + u_2 = U_0\cos\theta\left[2r + \left(r + \frac{a^2}{r}\right)\left(\frac{1 - \frac{k_1}{k_2}}{1 + \frac{k_1}{k_1}}\right)\right].
$$
 

Similar threads

  • · Replies 2 ·
Replies
2
Views
2K
  • · Replies 2 ·
Replies
2
Views
2K
Replies
13
Views
3K
Replies
1
Views
2K
  • · Replies 2 ·
Replies
2
Views
2K
  • · Replies 7 ·
Replies
7
Views
2K
  • · Replies 6 ·
Replies
6
Views
2K
Replies
4
Views
3K
  • · Replies 0 ·
Replies
0
Views
1K
  • · Replies 11 ·
Replies
11
Views
2K