Processing math: 100%

2.11 HW 11

  2.11.1 HW 11 questions
  2.11.2 Problem 1
  2.11.3 Problem 2
  2.11.4 Problem 3
  2.11.5 Problem 4
  2.11.6 Key solution for HW 11

2.11.1 HW 11 questions

pict
PDF (letter size)
PDF (legal size)

2.11.2 Problem 1

Find the normal modes of a rectangular drum with sides of length L_{x} and L_{y}

solution

The geometry of the problem is

pict
Figure 2.37:Problem to solve

Using Cartesian coordinates. Wave displacement is u\equiv u\left ( x,y,t\right ) (out of page).\begin{align*} \frac{\partial ^{2}u\left ( x,y,t\right ) }{\partial t^{2}} & =c^{2}\left ( \frac{\partial ^{2}u}{\partial x^{2}}+\frac{\partial ^{2}u}{\partial y^{2}}\right ) \\ 0 & <x<L_{x}\\ 0 & <y<L_{y} \end{align*}

Boundary conditions on x\begin{align*} u\left ( 0,y,t\right ) & =0\\ u\left ( L_{x},y,t\right ) & =0 \end{align*}

And boundary conditions on y\begin{align*} u\left ( x,0,t\right ) & =0\\ u\left ( x,L_{y},t\right ) & =0 \end{align*}

Solution

Let u=X\left ( x\right ) Y\left ( y\right ) T\left ( t\right ) . Substituting into the PDE gives\begin{align*} \frac{1}{c^{2}}T^{\prime \prime }XY & =X^{\prime \prime }YT+Y^{\prime \prime }XT\\ \frac{1}{c^{2}}\frac{T^{\prime \prime }}{T} & =\frac{X^{\prime \prime }}{X}+\frac{Y^{\prime \prime }}{Y} \end{align*}

Hence, using \lambda as first separation constant we obtain\begin{align*} \frac{1}{c^{2}}\frac{T^{\prime \prime }}{T} & =-\lambda \\ \frac{X^{\prime \prime }}{X}+\frac{Y^{\prime \prime }}{Y} & =-\lambda \end{align*}

The time ODE becomes T^{\prime \prime }+c^{2}\lambda T=0

And the space ODE becomes \frac{X^{\prime \prime }}{X}+\frac{Y^{\prime \prime }}{Y}=-\lambda
Separating the space ODE again \frac{X^{\prime \prime }}{X}=-\lambda -\frac{Y^{\prime \prime }}{Y}=-\mu
Where \mu is the new separation variable. This gives two new separate ODE’s\begin{align*} \frac{X^{\prime \prime }}{X} & =-\mu \\ -\lambda -\frac{Y^{\prime \prime }}{Y} & =-\mu \end{align*}

Or\begin{align*} X^{\prime \prime }+\mu X & =0\\ Y^{\prime \prime }+Y\left ( \lambda -\mu \right ) & =0 \end{align*}

Solving for X ODE first, and knowing that \mu >0 from nature of boundary conditions, we obtain X\left ( x\right ) =A\cos \left ( \sqrt{\mu }x\right ) +B\sin \left ( \sqrt{\mu }x\right )

Applying B.C. at x=0 0=A
Hence X\left ( x\right ) =B\sin \left ( \sqrt{\mu }x\right ) . Applying B.C. at x=L_{x} 0=B\sin \left ( \sqrt{\mu }L_{x}\right )
Hence \begin{align} \sqrt{\mu }L_{x} & =n\pi \nonumber \\ \mu _{n} & =\left ( \frac{n\pi }{L_{x}}\right ) ^{2}\qquad n=1,2,3,\cdots \tag{1} \end{align}

Therefore the X_{n}\left ( x\right ) solution is\begin{equation} X_{n}\left ( x\right ) =B_{n}\sin \left ( \frac{n\pi }{L_{x}}x\right ) \qquad n=1,2,3,\cdots \tag{2} \end{equation}

Solving the Y\left ( y\right ) ODE using the same eigenvalues found above Y^{\prime \prime }+Y\left ( \lambda -\left ( \frac{n\pi }{L_{x}}\right ) ^{2}\right ) =0
The solution is Y\left ( y\right ) =C\cos \left ( \sqrt{\lambda -\left ( \frac{n\pi }{L_{x}}\right ) ^{2}}y\right ) +D\sin \left ( \sqrt{\lambda -\left ( \frac{n\pi }{L_{x}}\right ) ^{2}}y\right )
Applying first B.C. Y\left ( 0\right ) =0 gives 0=C
Hence Y\left ( y\right ) =D\sin \left ( \sqrt{\lambda -\left ( \frac{n\pi }{L_{x}}\right ) ^{2}}y\right )
Applying second B.C. Y\left ( L_{y}\right ) =0 0=D\sin \left ( \sqrt{\lambda -\left ( \frac{n\pi }{L_{x}}\right ) ^{2}}L_{y}\right )
Hence\begin{align*} \sqrt{\lambda -\left ( \frac{n\pi }{L_{x}}\right ) ^{2}}L_{y} & =m\pi \qquad m=1,2,3,\cdots \\ \lambda _{nm}-\left ( \frac{n\pi }{L_{x}}\right ) ^{2} & =\left ( \frac{m\pi }{L_{y}}\right ) ^{2}\\ \lambda _{nm} & =\left ( \frac{m\pi }{L_{y}}\right ) ^{2}+\left ( \frac{n\pi }{L_{x}}\right ) ^{2}\qquad n=1,2,3,\cdots ,m=1,2,3,\cdots \end{align*}

Hence the Y_{nm} solution is Y_{nm}=D_{nm}\sin \left ( \frac{m\pi }{L_{y}}y\right ) \qquad n=1,2,3,\cdots ,m=1,2,3,\cdots

We notice that X_{n}\left ( x\right ) solution depends on n only, while Y_{nm}\left ( y\right ) solution depends on n and m. Now that we found \lambda we can we solve the time T\left ( t\right ) ode\begin{align*} T_{nm}^{\prime \prime }+c^{2}\lambda _{nm}T_{nm} & =0\\ T_{nm}\left ( t\right ) & =E_{nm}\cos \left ( c\sqrt{\lambda _{nm}}t\right ) +F_{nm}\sin \left ( c\sqrt{\lambda _{nm}}t\right ) \end{align*}

Combining all solution , and merging all constants into two, we find\begin{align*} u_{nm}\left ( x,y,t\right ) & =X_{n}\left ( x\right ) Y_{nm}\left ( y\right ) T_{nm}\left ( t\right ) \\ & =\left ( B_{n}X_{n}\right ) \left ( D_{nm}\sin \left ( \frac{m\pi }{L_{y}}y\right ) \right ) \left ( E_{nm}\cos \left ( c\sqrt{\lambda _{nm}}t\right ) +F_{nm}\sin \left ( c\sqrt{\lambda _{nm}}t\right ) \right ) \\ & =B_{n}X_{n}\sin \left ( \frac{m\pi }{L_{y}}y\right ) \left ( E_{nm}^{\prime }\cos \left ( c\sqrt{\lambda _{nm}}t\right ) +F_{nm}^{\prime }\sin \left ( c\sqrt{\lambda _{nm}}t\right ) \right ) \\ & =X_{n}\sin \left ( \frac{m\pi }{L_{y}}y\right ) \left ( E_{nm}^{\prime \prime }\cos \left ( c\sqrt{\lambda _{nm}}t\right ) +F_{nm}^{\prime \prime }\sin \left ( c\sqrt{\lambda _{nm}}t\right ) \right ) \end{align*}

Where E_{nm}^{\prime \prime },F_{nm}^{\prime \prime } are the new constants after merging them with the other constants. Renaming E_{nm}^{\prime \prime }=A_{nm},F_{nm}^{\prime \prime }=B_{nm} the above solution can be written as\begin{align} u\left ( x,y,t\right ) & =\sum _{n=1}^{\infty }\sum _{m=1}^{\infty }X_{n}\left ( x\right ) Y_{mn}\left ( y\right ) T_{mn}\left ( t\right ) \nonumber \\ & =\sum _{n=1}^{\infty }\sum _{m=1}^{\infty }A_{nm}\sin \left ( \frac{n\pi }{L_{x}}x\right ) \sin \left ( \frac{m\pi }{L_{y}}y\right ) \cos \left ( c\sqrt{\lambda _{nm}}t\right ) \nonumber \\ & +\sum _{n=1}^{\infty }\sum _{m=1}^{\infty }B_{nm}\sin \left ( \frac{n\pi }{L_{x}}x\right ) \sin \left ( \frac{m\pi }{L_{y}}y\right ) \sin \left ( c\sqrt{\lambda _{nm}}t\right ) \tag{3} \end{align}

To solve this completely, we apply initial conditions to find A_{nm},B_{nm}. But the problem is just asking for the normal modes. These are given by X_{n}\left ( x\right ) Y_{mn}\left ( y\right ) . Therefore for n=1, we have the modes \sin \left ( \frac{\pi }{L_{x}}x\right ) \sin \left ( \frac{\pi }{L_{y}}y\right ) ,\sin \left ( \frac{\pi }{L_{x}}x\right ) \sin \left ( \frac{2\pi }{L_{y}}y\right ) ,\sin \left ( \frac{\pi }{L_{x}}x\right ) \sin \left ( \frac{3\pi }{L_{y}}y\right ) ,\cdots and for n=2 we have \sin \left ( \frac{2\pi }{L_{x}}x\right ) \sin \left ( \frac{\pi }{L_{y}}y\right ) ,\sin \left ( \frac{2\pi }{L_{x}}x\right ) \sin \left ( \frac{2\pi }{L_{y}}y\right ) ,\sin \left ( \frac{2\pi }{L_{x}}x\right ) \sin \left ( \frac{3\pi }{L_{y}}y\right ) ,\cdots and so on.






n m=1 2 3 4





1 \sin \left ( \frac{\pi }{L_{x}}x\right ) \sin \left ( \frac{\pi }{L_{y}}y\right ) \sin \left ( \frac{\pi }{L_{x}}x\right ) \sin \left ( \frac{2\pi }{L_{y}}y\right ) \sin \left ( \frac{\pi }{L_{x}}x\right ) \sin \left ( \frac{3\pi }{L_{y}}y\right ) \cdots





2 \sin \left ( \frac{2\pi }{L_{x}}x\right ) \sin \left ( \frac{\pi }{L_{y}}y\right ) \sin \left ( \frac{2\pi }{L_{x}}x\right ) \sin \left ( \frac{2\pi }{L_{y}}y\right ) \sin \left ( \frac{2\pi }{L_{x}}x\right ) \sin \left ( \frac{3\pi }{L_{y}}y\right ) \cdots





3 \sin \left ( \frac{3\pi }{L_{x}}x\right ) \sin \left ( \frac{\pi }{L_{y}}y\right ) \sin \left ( \frac{3\pi }{L_{x}}x\right ) \sin \left ( \frac{2\pi }{L_{y}}y\right ) \sin \left ( \frac{3\pi }{L_{x}}x\right ) \sin \left ( \frac{3\pi }{L_{y}}y\right ) \cdots





\vdots \vdots \vdots \vdots \vdots





To draw these modes, let us assume that L_{x}=1,L_{y}=1. This gives






n m=1 2 3 4





1 \sin \left ( \pi x\right ) \sin \left ( \pi y\right ) \sin \left ( \pi x\right ) \sin \left ( 2\pi y\right ) \sin \left ( \pi x\right ) \sin \left ( 3\pi y\right ) \cdots





2 \sin \left ( 2\pi x\right ) \sin \left ( \pi y\right ) \sin \left ( 2\pi x\right ) \sin \left ( 2\pi y\right ) \sin \left ( 2\pi x\right ) \sin \left ( 3\pi y\right ) \cdots





3 \sin \left ( 3\pi x\right ) \sin \left ( \pi y\right ) \sin \left ( 3\pi x\right ) \sin \left ( 2\pi y\right ) \sin \left ( 3\pi x\right ) \sin \left ( 3\pi y\right ) \cdots





\vdots \vdots \vdots \vdots \vdots





The following is a plot of the above modes for illustrations with the code used to generate these plots.

pict
Figure 2.38:Modes using L_x=1,L_y=1

pict
Figure 2.39:Code used to draw above plot

The following is 3D view of the above modes.

pict
Figure 2.40:3D view of the modes using L_x=1,L_y=1

pict
Figure 2.41:Code used to draw above plot

2.11.3 Problem 2

Find the normal modes of an acoustic waves in a hollow sphere of radius R. The wave equation is \nabla ^{2}\psi \left ( r,\theta ,\phi ,t\right ) =\frac{1}{c^{2}}\psi _{tt}

With boundary conditions \psi _{r}=0 at r=0 and at r=r_{0}. (I used r_{0} in place of R because wanted to use R\left ( r\right ) for separation of variables).

What is the lowest frequency?

solution

Let \psi \left ( r,\theta ,\phi ,t\right ) =u\left ( r,\theta ,\phi \right ) e^{-i\omega t}

Substituting this back in the original PDE gives \nabla ^{2}u\left ( r,\theta ,\phi \right ) +\frac{\omega ^{2}}{c^{2}}u\left ( r,\theta ,\phi \right ) =0
Let k=\frac{\omega }{c} (wave number) and the above becomes\begin{equation} \nabla ^{2}u+k^{2}u=0 \tag{1} \end{equation}
The above is called the Helmholtz PDE. In spherical coordinates it becomes \overset{\text{Radial part}}{\overbrace{u_{rr}+\frac{2}{r}u_{r}}}+\overset{\text{Angular part}}{\overbrace{\frac{1}{r^{2}}\left ( \frac{\cos \theta }{\sin \theta }u_{\theta }+u_{\theta \theta }\right ) +\frac{1}{r^{2}\sin ^{2}\theta }u_{\phi \phi }}}+k^{2}u=0
Let u\left ( r,\theta ,\phi \right ) =R\left ( r\right ) \Theta \left ( \theta \right ) \Phi \left ( \phi \right ) and the above becomes R^{\prime \prime }T\Theta \Phi +\frac{2}{r}R^{\prime }T\Theta \Phi +\frac{1}{r^{2}}\left ( \frac{\cos \theta }{\sin \theta }\Theta ^{\prime }RT\Phi +\Theta ^{\prime \prime }RT\Phi \right ) +\frac{1}{r^{2}\sin ^{2}\theta }\Phi ^{\prime \prime }R\Theta T+k^{2}R\Theta T=0
Dividing by R\Theta \Phi \neq 0 gives\begin{align*} \frac{R^{\prime \prime }}{R}+\frac{2}{r}\frac{R^{\prime }}{R}+\frac{1}{r^{2}}\left ( \frac{\cos \theta }{\sin \theta }\frac{\Theta ^{\prime }}{\Theta }+\frac{\Theta ^{\prime \prime }}{\Theta }\right ) +\frac{1}{r^{2}\sin ^{2}\theta }\frac{\Phi ^{\prime \prime }}{\Phi }+k^{2} & =0\\ r^{2}\sin ^{2}\theta \frac{R^{\prime \prime }}{R}+r^{2}\sin ^{2}\theta \frac{2}{r}\frac{R^{\prime }}{R}+\sin ^{2}\theta \left ( \frac{\cos \theta }{\sin \theta }\frac{\Theta ^{\prime }}{\Theta }+\frac{\Theta ^{\prime \prime }}{\Theta }\right ) +k^{2}r^{2}\sin ^{2}\theta & =-\frac{\Phi ^{\prime \prime }}{\Phi } \end{align*}

The left side depends only on r,\theta and the right side depends only on \phi . Let the second separation constant be m^{2} and the above becomes\begin{equation} r^{2}\sin ^{2}\theta \frac{R^{\prime \prime }}{R}+r^{2}\sin ^{2}\theta \frac{2}{r}\frac{R^{\prime }}{R}+\sin ^{2}\theta \left ( \frac{\cos \theta }{\sin \theta }\frac{\Theta ^{\prime }}{\Theta }+\frac{\Theta ^{\prime \prime }}{\Theta }\right ) +k^{2}r^{2}\sin ^{2}\theta =-\frac{\Phi ^{\prime \prime }}{\Phi }=m^{2} \tag{2} \end{equation}

Which gives the first angular ODE as\begin{equation} \Phi ^{\prime \prime }+m^{2}\Phi =0 \tag{2A} \end{equation}
We now go back to (2) to obtain the rest of the solutions. We now have\begin{align*} r^{2}\sin ^{2}\theta \frac{R^{\prime \prime }}{R}+r^{2}\sin ^{2}\theta \frac{2}{r}\frac{R^{\prime }}{R}+\sin ^{2}\theta \left ( \frac{\cos \theta }{\sin \theta }\frac{\Theta ^{\prime }}{\Theta }+\frac{\Theta ^{\prime \prime }}{\Theta }\right ) +k^{2}r^{2}\sin ^{2}\theta & =m^{2}\\ k^{2}r^{2}+r^{2}\left ( \frac{R^{\prime \prime }}{R}+\frac{2}{r}\frac{R^{\prime }}{R}\right ) +\left ( \frac{\cos \theta }{\sin \theta }\frac{\Theta ^{\prime }}{\Theta }+\frac{\Theta ^{\prime \prime }}{\Theta }\right ) & =\frac{m^{2}}{\sin ^{2}\theta }\\ k^{2}r^{2}+r^{2}\left ( \frac{R^{\prime \prime }}{R}+\frac{2}{r}\frac{R^{\prime }}{R}\right ) & =-\left ( \frac{\cos \theta }{\sin \theta }\frac{\Theta ^{\prime }}{\Theta }+\frac{\Theta ^{\prime \prime }}{\Theta }\right ) +\frac{m^{2}}{\sin ^{2}\theta } \end{align*}

The left side depends on r and the right side depends on \theta only. Let the separation constant be l\left ( l+1\right ) where l is integer which results in \begin{equation} k^{2}r^{2}+r^{2}\left ( \frac{R^{\prime \prime }}{R}+\frac{2}{r}\frac{R^{\prime }}{R}\right ) =-\left ( \frac{\cos \theta }{\sin \theta }\frac{\Theta ^{\prime }}{\Theta }+\frac{\Theta ^{\prime \prime }}{\Theta }\right ) +\frac{m^{2}}{\sin ^{2}\theta }=l\left ( l+1\right ) \tag{3} \end{equation}

Therefore the next angular ODE is\begin{align} -\left ( \frac{\cos \theta }{\sin \theta }\frac{\Theta ^{\prime }}{\Theta }+\frac{\Theta ^{\prime \prime }}{\Theta }\right ) +\frac{m^{2}}{\sin ^{2}\theta } & =l\left ( l+1\right ) \nonumber \\ -\left ( \frac{\cos \theta }{\sin \theta }\frac{\Theta ^{\prime }}{\Theta }+\frac{\Theta ^{\prime \prime }}{\Theta }\right ) +\frac{m^{2}}{\sin ^{2}\theta }-l\left ( l+1\right ) & =0\nonumber \\ \left ( \frac{\cos \theta }{\sin \theta }\frac{\Theta ^{\prime }}{\Theta }+\frac{\Theta ^{\prime \prime }}{\Theta }\right ) -\frac{m^{2}}{\sin ^{2}\theta }+l\left ( l+1\right ) & =0\nonumber \\ \Theta ^{\prime \prime }+\frac{\cos \theta }{\sin \theta }\Theta ^{\prime }+\left ( l\left ( l+1\right ) -\frac{m^{2}}{\sin ^{2}\theta }\right ) \Theta & =0 \tag{4} \end{align}

Let z=\cos \theta , then \frac{d\Theta }{d\theta }=\frac{d\Theta }{dz}\frac{dz}{d\theta }=-\frac{d\Theta }{dz}\sin \theta and \begin{align*} \frac{d^{2}\Theta }{d\theta ^{2}} & =\frac{d}{d\theta }\left ( -\frac{d\Theta }{dz}\sin \theta \right ) \\ & =-\frac{d^{2}\Theta }{dz^{2}}\frac{dz}{d\theta }\sin \theta -\frac{d\Theta }{dz}\cos \theta \\ & =\frac{d^{2}\Theta }{dz^{2}}\sin ^{2}\theta -\frac{d\Theta }{dz}\cos \theta \end{align*}

But \sin ^{2}\theta =1-\cos ^{2}\theta =1-z^{2} and the above becomes \frac{d^{2}\Theta }{d\theta ^{2}}=\frac{d^{2}\Theta }{dz^{2}}\left ( 1-z^{2}\right ) -\frac{d\Theta }{dz}z

Using these in (4) gives\begin{align} \frac{d^{2}\Theta }{dz^{2}}\left ( 1-z^{2}\right ) -\frac{d\Theta }{dz}z+\frac{z}{\sin \theta }\left ( -\frac{d\Theta }{dz}\sin \theta \right ) +\left ( l\left ( l+1\right ) -\frac{m^{2}}{1-z^{2}}\right ) \Theta \left ( z\right ) & =0\nonumber \\ \left ( 1-z^{2}\right ) \Theta ^{\prime \prime }-2z\Theta ^{\prime }+\left ( l\left ( l+1\right ) -\frac{m^{2}}{1-z^{2}}\right ) \Theta \left ( z\right ) & =0 \tag{3A} \end{align}

And finally, we obtain the final ODE, which is the radial ODE from (3)\begin{align} k^{2}r^{2}+r^{2}\left ( \frac{R^{\prime \prime }}{R}+\frac{2}{r}\frac{R^{\prime }}{R}\right ) & =l\left ( l+1\right ) \nonumber \\ k^{2}r^{2}R+r^{2}\left ( R^{\prime \prime }+\frac{2}{r}R^{\prime }\right ) -l\left ( l+1\right ) R & =0\nonumber \\ r^{2}R^{\prime \prime }+2rR^{\prime }+\left ( k^{2}r^{2}-l\left ( l+1\right ) \right ) R & =0\nonumber \\ R^{\prime \prime }+\frac{2}{r}R^{\prime }+\left ( k^{2}-\frac{l\left ( l+1\right ) }{r^{2}}\right ) R & =0 \tag{4A} \end{align}

In summary we have obtained the following 4 ODE’s to solve (1A,2A,3A,4A)\begin{align} \Phi ^{\prime \prime }+m^{2}\Phi & =0\tag{2A}\\ \left ( 1-z^{2}\right ) \Theta ^{\prime \prime }-2z\Theta ^{\prime }+\left ( l\left ( l+1\right ) -\frac{m^{2}}{1-z^{2}}\right ) \Theta \left ( z\right ) & =0\tag{3A}\\ R^{\prime \prime }+\frac{2}{r}R^{\prime }+\left ( k^{2}-\frac{l\left ( l+1\right ) }{r^{2}}\right ) R & =0 \tag{4A} \end{align}

Solution to (2A) requires m to be integer due to periodicity requirements of solution. The solution is \Phi \left ( \phi \right ) =e^{\pm im\phi }. Equation (3A) is the associated Legendre ODE. Since we are taking l as integer then the solution is known to be \Theta \left ( z\right ) =P_{l}^{m}\left ( z\right ) +Q_{l}^{m}\left ( z\right ) where P_{l}^{m}\left ( z\right ) is called the associated Legendre polynomial and Q_{l}^{m} is the Legendre function of the second kind. Finally (4A) can be converted to Bessel ODE as shown in class notes using the transformation R\left ( r\right ) =\frac{u\left ( r\right ) }{\sqrt{r}} which results in u^{\prime \prime }+\frac{1}{r}u^{\prime }+\left ( k^{2}-\frac{\left ( l+\frac{1}{2}\right ) ^{2}}{r^{2}}\right ) u=0

Which has solution J_{l+\frac{1}{2}}\left ( kr\right ) . The second solution J_{-\left ( l+\frac{1}{2}\right ) }\left ( kr\right ) is rejected since it is not finite at zero and hence makes the solution blow up at center of sphere. Therefore solution to (4A) is \begin{align*} R\left ( r\right ) & =C\sqrt{\frac{\pi }{2kr}}J_{l+\frac{1}{2}}\left ( kr\right ) \\ & =Cj_{l}\left ( kr\right ) \end{align*}

Where C is arbitrary constant. Putting all the above together, then the final solution is \psi \left ( r,\theta ,\phi ,t\right ) =\left \{ \begin{array} [c]{c}e^{-i\omega t}\end{array} \right . \left \{ \begin{array} [c]{c}e^{im\phi }\\ e^{-im\phi }\end{array} \right . \left \{ \begin{array} [c]{c}P_{l}^{m}\left ( \cos \theta \right ) \\ Q_{l}^{m}\left ( \cos \theta \right ) \end{array} \right . \left \{ \begin{array} [c]{c}j_{l}\left ( kr\right ) \end{array} \right .

Where j_{l}\left ( kr\right ) are the spherical Bessel functions. Now we need to satisfy the boundary conditions. Since only j_{l}\left ( kr\right ) depends on r, then \psi _{r}=0 at r=0 and at r=r_{0} are equivalent to looking at R^{\prime }\left ( r\right ) =0 at r=0 and r=r_{0}.  Therefore we need to find the smallest l,k which satisfy both conditions. This will give the lowest frequency.

I found from DLMF that the series expansion of j_{l}\left ( kr\right ) is\begin{equation} j_{l}\left ( kr\right ) =\frac{\left ( kr\right ) ^{l}}{\left ( 2l+1\right ) !!}\left ( 1-\frac{\left ( kr\right ) ^{2}}{2\left ( 2l+3\right ) }+\frac{\left ( kr\right ) ^{4}}{8\left ( 2l+5\right ) \left ( 2l+3\right ) }+\cdots \right ) \tag{5} \end{equation}

Hence for r\rightarrow 0, we can approximate the above as the following by ignoring all higher order terms \lim _{r\rightarrow 0}j_{l}\left ( kr\right ) =\frac{\left ( kr\right ) ^{l}}{\left ( 2l+1\right ) !!}
Which means for small r, the derivative is \frac{d}{dr}j_{l}\left ( kr\right ) =\frac{l\left ( kr\right ) ^{l-1}}{\left ( 2l+1\right ) !!}
At r=0 then setting \left [ \frac{d}{dr}j_{l}\left ( kr\right ) \right ] _{r\rightarrow 0}=0 is satisfied for all l. Now taking derivative of (5) gives

\frac{d}{dr}j_{l}\left ( kr\right ) =\frac{l\left ( kr\right ) ^{l-1}}{\left ( 2l+1\right ) !!}\left ( 1-\frac{\left ( kr\right ) ^{2}}{2\left ( 2l+3\right ) }+\frac{\left ( kr\right ) ^{4}}{8\left ( 2l+5\right ) \left ( 2l+3\right ) }+\cdots \right ) +\frac{\left ( kr\right ) ^{l}}{\left ( 2l+1\right ) !!}\left ( 1-\frac{2\left ( kr\right ) }{2\left ( 2l+3\right ) }+\frac{4\left ( kr\right ) ^{3}}{8\left ( 2l+5\right ) \left ( 2l+3\right ) }+\cdots \right )

At r=r_{0} the above becomes  

\left [ \frac{d}{dr}j_{l}\left ( kr\right ) \right ] _{r\rightarrow r_{0}}=\frac{l\left ( kr_{0}\right ) ^{l-1}}{\left ( 2l+1\right ) !!}\left ( 1-\frac{\left ( kr_{0}\right ) ^{2}}{2\left ( 2l+3\right ) }+\frac{\left ( kr_{0}\right ) ^{4}}{8\left ( 2l+5\right ) \left ( 2l+3\right ) }+\cdots \right ) +\frac{\left ( kr_{0}\right ) ^{l}}{\left ( 2l+1\right ) !!}\left ( 1-\frac{2\left ( kr_{0}\right ) }{2\left ( 2l+3\right ) }+\frac{4\left ( kr_{0}\right ) ^{3}}{8\left ( 2l+5\right ) \left ( 2l+3\right ) }+\cdots \right )

Now we ask, for which values of l is the above zero? If we let l\rightarrow \infty then we obtain\begin{align*} \left [ \frac{d}{dr}j_{l}\left ( kr\right ) \right ] _{\substack{r\rightarrow r_{0}\\l\rightarrow \infty }} & =\lim _{l\rightarrow \infty }\frac{l\left ( kr_{0}\right ) ^{l-1}}{\left ( 2l+1\right ) !!}+\frac{\left ( kr_{0}\right ) ^{l}}{\left ( 2l+1\right ) !!}\\ & =0 \end{align*}

Therefore, to satisfy both \left [ \frac{d}{dr}j_{l}\left ( kr\right ) \right ] _{r\rightarrow 0}=0 and \left [ \frac{d}{dr}j_{l}\left ( kr\right ) \right ] _{r\rightarrow r_{0}}=0 we need l\rightarrow \infty . In other words, a very large integer. The larger l is, the lower the radial frequency. In addition, increasing k while keeping l fixed will increase the frequency. And decreasing k while keeping l fixed decreases the frequency. And for fixed k, increasing l decreases the frequency.

2.11.4 Problem 3

A sphere of radius R is at temperature u=0. At time t=0 it is immersed in a heat bath of temperature u_{0}. What is the temperature distribution u\left ( r,t\right ) as function of time?

solution

Note: I Used u\left ( r,t\right ) instead of T\left ( r,t\right ) as the dependent variable to allow using T\left ( t\right ) for separation of variables without confusing it with the original T\left ( r,t\right ) .

The PDE specification is, solve for u\left ( r,t\right ) u_{t}=k\nabla ^{2}u\qquad t>0,0<r<R

With initial conditions u\left ( r,0\right ) =0
And boundary conditions\begin{align*} u\left ( R,t\right ) & =u_{0}\\ \left \vert u\left ( 0,t\right ) \right \vert & <\infty \end{align*}

Where the second B.C. above means the temperature u is bounded at origin (center of sphere). In spherical coordinates, the PDE becomes (There are no dependency on \theta ,\phi due to symmetry), and only radial dependency.\begin{equation} \frac{1}{k}u_{t}=\frac{1}{r}\left ( ru\right ) _{rr}\tag{1} \end{equation}

To simplify the solution, let U\left ( r,t\right ) =ru\left ( r,t\right )
And we obtain a new PDE\begin{equation} \frac{1}{k}U_{t}=U_{rr}\tag{2} \end{equation}
And the boundary conditions u\left ( R,t\right ) =u_{0} becomes U\left ( R,t\right ) =Ru_{0} and the initial conditions becomes U\left ( r,0\right ) =0. So we will solve (2) and not (1). But since the boundary conditions are not homogenous, we can not use separation of variables. We introduce a reference function w\left ( r\right ) which need to satisfy the nonhomogeneous boundary conditions only. Let w\left ( r\right ) =Br. When r=R then Ru_{0}=BR or B=u_{0} When r=0 then w=0 which is bounded. Hence w\left ( r\right ) =u_{0}r
Therefore, the solution now can be written as\begin{equation} U\left ( r,t\right ) =v\left ( r,t\right ) +u_{0}r\tag{3} \end{equation}
Where v\left ( r,t\right ) now satisfies the PDE but with homogenous B.C.  Substituting (3) into (2) gives\begin{align} v_{t} & =k\frac{\partial ^{2}}{\partial r^{2}}\left ( v\left ( r,t\right ) +u_{0}r\right ) \nonumber \\ v_{t} & =kv_{rr}\left ( r,t\right ) \tag{4} \end{align}

We need to solve the above but with homogenous boundary conditions\begin{align*} v\left ( R,t\right ) & =0\\ \left \vert v\left ( 0,t\right ) \right \vert & <\infty \end{align*}

This is standard PDE, who can be solved by separation of variables. let v=F\left ( r\right ) T\left ( t\right ) , hence (4) becomes\begin{align*} T^{\prime }F & =kF^{\prime \prime }T\\ k\frac{T^{\prime }}{T} & =\frac{F^{\prime \prime }}{F}=-\lambda ^{2} \end{align*}

Which gives F^{\prime \prime }+\lambda ^{2}F=0

Due to boundary conditions only \lambda >0 is eigenvalues. Hence solution is F\left ( r\right ) =A\cos \left ( \lambda r\right ) +B\sin \left ( \lambda r\right )
At r=0, since bounded, say 0, then we can take A=0, leaving the solution F\left ( r\right ) =B\sin \left ( \lambda r\right )
At r=R 0=B\sin \left ( \lambda R\right )
For nontrivial solution\begin{align*} \lambda R & =n\pi \qquad n=1,2,3,\cdots \\ \lambda _{n} & =\frac{n\pi }{R} \end{align*}

Hence eigenfunctions are F_{n}\left ( r\right ) =\sin \left ( \frac{n\pi }{R}r\right ) \qquad n=1,2,3,\cdots

The time ODE is therefore T^{\prime }+\lambda ^{2}kT=0 with solution T_{n}\left ( t\right ) =A_{n}e^{-\left ( \frac{n\pi }{R}\right ) ^{2}kt}. Hence the solution to (4) is v\left ( r,t\right ) =\sum _{n=1}^{\infty }A_{n}e^{-\left ( \frac{n\pi }{R}\right ) ^{2}kt}\sin \left ( \frac{n\pi }{R}r\right )
Therefore from (3) U\left ( r,t\right ) =\left ( \sum _{n=1}^{\infty }A_{n}e^{-\left ( \frac{n\pi }{R}\right ) ^{2}kt}\sin \left ( \frac{n\pi }{R}r\right ) \right ) +u_{0}r
But U\left ( r,t\right ) =ru\left ( r,t\right ) , hence\begin{equation} u\left ( r,t\right ) =\left ( \frac{1}{r}\sum _{n=1}^{\infty }A_{n}e^{-\left ( \frac{n\pi }{R}\right ) ^{2}kt}\sin \left ( \frac{n\pi }{R}r\right ) \right ) +u_{0}\tag{5} \end{equation}
Now we find A_{n} from initial conditions. At t=0\begin{align*} 0 & =u_{0}+\frac{1}{r}\sum _{n=1}^{\infty }A_{n}\sin \left ( \frac{n\pi }{R}r\right ) \\ -ru_{0} & =\sum _{n=1}^{\infty }A_{n}\sin \left ( \frac{n\pi }{R}r\right ) \end{align*}

Therefore A_{n} are the Fourier series coefficients of -ru_{0}\begin{align*} \frac{R}{2}A_{n} & =-\int _{0}^{R}ru_{0}\sin \left ( \frac{n\pi }{R}r\right ) dr\\ A_{n} & =-\frac{2u_{0}}{R}\int _{0}^{R}r\sin \left ( \frac{n\pi }{R}r\right ) dr\\ & =-\frac{2u_{0}}{R}\left ( -1\right ) ^{n+1}\frac{R^{2}}{n\pi }\\ & =\left ( -1\right ) ^{n}\frac{2R}{n\pi }u_{0} \end{align*}

Hence the solution (5) becomes\begin{align} u\left ( r,t\right ) & =u_{0}+u_{0}\frac{2R}{r\pi }\sum _{n=1}^{\infty }\left ( -1\right ) ^{n}\frac{1}{n}e^{-k\left ( \frac{n\pi }{R}\right ) ^{2}t}\sin \left ( \frac{n\pi }{R}r\right ) \nonumber \\ & =u_{0}\left ( 1+\frac{2R}{r\pi }\sum _{n=1}^{\infty }\left ( -1\right ) ^{n}\frac{1}{n}e^{-k\left ( \frac{n\pi }{R}\right ) ^{2}t}\sin \left ( \frac{n\pi }{R}r\right ) \right ) \tag{7} \end{align}

Verification of solution

Verification that (7) satisfies the PDE u_{t}=k\nabla ^{2}u. Taking time derivative of (7) gives\begin{equation} u_{t}=-u_{0}\frac{2R}{r\pi }k\sum _{n=1}^{\infty }\left ( -1\right ) ^{n}\frac{1}{n}\left ( \frac{n\pi }{R}\right ) ^{2}e^{-k\left ( \frac{n\pi }{R}\right ) ^{2}t}\sin \left ( \frac{n\pi }{R}r\right ) \tag{8} \end{equation}

And taking space derivatives of (7) gives\begin{align*} u_{x} & =u_{0}\frac{2R}{r\pi }\sum _{n=1}^{\infty }\left ( -1\right ) ^{n}\frac{1}{n}e^{-k\left ( \frac{n\pi }{R}\right ) ^{2}t}\frac{n\pi }{R}\cos \left ( \frac{n\pi }{R}r\right ) \\ u_{xx} & =-u_{0}\frac{2R}{r\pi }\sum _{n=1}^{\infty }\left ( -1\right ) ^{n}\frac{1}{n}e^{-k\left ( \frac{n\pi }{R}\right ) ^{2}t}\left ( \frac{n\pi }{R}\right ) ^{2}\sin \left ( \frac{n\pi }{R}r\right ) \end{align*}

Hence ku_{xx} becomes\begin{equation} ku_{xx}=-u_{0}\frac{2R}{r\pi }k\sum _{n=1}^{\infty }\left ( -1\right ) ^{n}\frac{1}{n}e^{-k\left ( \frac{n\pi }{R}\right ) ^{2}t}\left ( \frac{n\pi }{R}\right ) ^{2}\sin \left ( \frac{n\pi }{R}r\right ) \tag{9} \end{equation}

Comparing (8) and (9) shows they are the same expressions.

Verification that (7) satisfies the boundary condition.

When r=R, therefore (7) gives, when replacing r by R\begin{align*} u\left ( R,t\right ) & =u_{0}\left ( 1+\frac{2R}{R\pi }\sum _{n=1}^{\infty }\left ( -1\right ) ^{n}\frac{1}{n}e^{-k\left ( \frac{n\pi }{R}\right ) ^{2}t}\sin \left ( \frac{n\pi }{R}R\right ) \right ) \\ & =u_{0}\left ( 1+\frac{2R}{R\pi }\sum _{n=1}^{\infty }\left ( -1\right ) ^{n}\frac{1}{n}e^{-k\left ( \frac{n\pi }{R}\right ) ^{2}t}\sin \left ( n\pi \right ) \right ) \\ & =u_{0}\left ( 1+0\right ) \\ & =u_{0} \end{align*}

But n is integer. Hence \sin \left ( n\pi \right ) =0 for all n. And the above becomes\begin{align*} u\left ( R,t\right ) & =u_{0}\left ( 1+0\right ) \\ & =u_{0} \end{align*}

Verified.

Verification that (7) satisfies the initial conditions u\left ( r,0\right ) =0 for r<R.

At t=0 (7) becomes\begin{align*} u\left ( r,0\right ) & =u_{0}\left ( 1+\frac{2R}{r\pi }\sum _{n=1}^{\infty }\left ( -1\right ) ^{n}\frac{1}{n}\sin \left ( \frac{n\pi }{R}r\right ) \right ) \\ & =u_{0}+\frac{2R}{r\pi }u_{0}\sum _{n=1}^{\infty }\frac{\left ( -1\right ) ^{n}}{n}\sin \left ( \frac{n\pi }{R}r\right ) \\ & =u_{0}+\frac{2R}{r\pi }u_{0}\left ( -\sin \left ( \frac{\pi }{R}r\right ) +\frac{1}{2}\sin \left ( \frac{2\pi }{R}r\right ) -\frac{1}{3}\sin \left ( \frac{3\pi }{R}r\right ) +\frac{1}{4}\sin \left ( \frac{4\pi }{R}r\right ) -\cdots \right ) \end{align*}

I could not simplify the above by hand, but using the computer, I verified numerically it is zero for 0<r<R for a given R and given u_{0}.

pict
Figure 2.42:Obtaining the sum using the computer

2.11.5 Problem 4

Consider the Helmholtz equation

\begin{equation} \nabla ^{2}u\left ( r,\theta \right ) +k^{2}u\left ( r,\theta \right ) =0 \tag{1} \end{equation}

inside the circle r=r_{0} with the boundary condition u\left ( r_{0},\theta \right ) =f\left ( \theta \right ) . The solution can be written in the form u\left ( r,\theta \right ) =\int _{0}^{2\pi }f\left ( \theta ^{\prime }\right ) G\left ( r,\theta ;\theta ^{\prime }\right ) d\theta ^{\prime }. Find the Green function G.

solution

I will solve (1) directly and then compare the solution obtain to u\left ( r,\theta \right ) =\int _{0}^{2\pi }f\left ( \theta ^{\prime }\right ) G\left ( r,\theta ;\theta ^{\prime }\right ) d\theta ^{\prime } in order to read off the Green function expression.  (1) in polar coordinates becomes

u_{rr}+\frac{1}{r}u_{r}+\frac{1}{r^{2}}u_{\theta \theta }+k^{2}u=0

Writing u\left ( r,\theta \right ) =R\left ( r\right ) \Theta \left ( \theta \right ) , the above PDE becomes\begin{align*} R^{\prime \prime }\Theta +\frac{1}{r}R^{\prime }\Theta +\frac{1}{r^{2}}\Theta ^{\prime \prime }R+k^{2}R\Theta & =0\\ \frac{R^{\prime \prime }}{R}+\frac{1}{r}\frac{R^{\prime }}{R}+\frac{1}{r^{2}}\frac{\Theta ^{\prime \prime }}{\Theta }+k^{2} & =0\\ r^{2}\frac{R^{\prime \prime }}{R}+r\frac{R^{\prime }}{R}+r^{2}k^{2} & =-\frac{\Theta ^{\prime \prime }}{\Theta }=m \end{align*}

Where m is the separation constant. The eigenvalue problem is taken as \Theta ^{\prime \prime }+m\Theta =0

Due to periodicity of the solution on the disk, then \Theta \left ( -\pi \right ) =\Theta \left ( \pi \right ) and \Theta ^{\prime }\left ( -\pi \right ) =\Theta ^{\prime }\left ( \pi \right ) . These boundary conditions restrict m to only positive integer values. Hence let m=n^{2} and the solution to the above becomes \Theta _{\alpha }\left ( \theta \right ) =A_{n}\cos \left ( n\theta \right ) +B_{n}\sin \left ( n\theta \right )
Now the radial ODE is \begin{align*} r^{2}\frac{R^{\prime \prime }}{R}+r\frac{R^{\prime }}{R}+r^{2}k^{2} & =\alpha ^{2}\\ r^{2}R^{\prime \prime }+rR^{\prime }+\left ( r^{2}k^{2}-n^{2}\right ) R & =0\\ R^{\prime \prime }+\frac{1}{r}R^{\prime }+\left ( k^{2}-\frac{n^{2}}{r^{2}}\right ) R & =0 \end{align*}

This is Bessel ODE whose solutions are (since n are integers) is R_{\alpha }\left ( r\right ) =C_{n}J_{n}\left ( kr\right ) +E_{n}Y_{n}\left ( kr\right )

But Y_{n}\left ( kr\right ) blows up at r=0, hence it is rejected leaving solution R_{n}\left ( r\right ) =C_{n}J_{n}\left ( kr\right ) . Hence the final solution is\begin{equation} u\left ( r,\theta \right ) =\sum _{m=1}^{\infty }\left ( A_{n}\cos \left ( n\theta \right ) +B_{n}\sin \left ( n\theta \right ) \right ) J_{n}\left ( kr\right ) \tag{2} \end{equation}
Where the constant C_{n} is merged with the other two constants.  Now, at r=r_{0} we are told that u\left ( r_{0},\theta \right ) =f\left ( \theta \right ) . Hence the above becomes f\left ( \theta \right ) =\sum _{m=1}^{\infty }\left ( A_{n}\cos \left ( n\theta \right ) +B_{n}\sin \left ( n\theta \right ) \right ) J_{n}\left ( kr_{0}\right )
By orthogonality of \cos \left ( n\theta \right ) ,\sin \left ( n\theta \right ) we find the Fourier cosine and Fourier sine coefficients A_{n},B_{n} as\begin{align*} A_{n}J_{n}\left ( kr_{0}\right ) \frac{1}{\pi } & =\int _{0}^{2\pi }f\left ( \theta \right ) \cos \left ( n\theta \right ) d\theta \\ B_{n}J_{n}\left ( kr_{0}\right ) \frac{1}{\pi } & =\int _{0}^{2\pi }f\left ( \theta \right ) \sin \left ( n\theta \right ) d\theta \end{align*}

Substituting the above back into the solution found in (2) results in\begin{align} u\left ( r,\theta \right ) & =\sum _{m=1}^{\infty }\left [ \left ( \frac{\pi }{J_{n}\left ( kr_{0}\right ) }\int _{0}^{2\pi }f\left ( \theta ^{\prime }\right ) \cos \left ( n\theta ^{\prime }\right ) d\theta ^{\prime }\right ) \cos \left ( n\theta \right ) +\left ( \frac{\pi }{J_{n}\left ( kr_{0}\right ) }\int _{0}^{2\pi }f\left ( \theta ^{\prime }\right ) \sin \left ( n\theta ^{\prime }\right ) d\theta ^{\prime }\right ) \sin \left ( n\theta \right ) \right ] J_{n}\left ( kr\right ) \nonumber \\ & =\sum _{m=1}^{\infty }\frac{\pi }{J_{n}\left ( kr_{0}\right ) }\left ( \int _{0}^{2\pi }f\left ( \theta ^{\prime }\right ) \cos \left ( n\theta ^{\prime }\right ) \cos \left ( n\theta \right ) d\theta ^{\prime }+\int _{0}^{2\pi }f\left ( \theta ^{\prime }\right ) \sin \left ( n\theta ^{\prime }\right ) \sin \left ( n\theta \right ) d\theta ^{\prime }\right ) J_{n}\left ( kr\right ) \tag{3} \end{align}

Using trig relations\begin{align*} \cos A\cos B & =\frac{1}{2}\left ( \cos \left ( A+B\right ) +\cos \left ( A-B\right ) \right ) \\ \sin A\sin B & =\frac{1}{2}\left ( \cos \left ( A-B\right ) -\cos \left ( A+B\right ) \right ) \end{align*}

Then (3) becomes u\left ( r,\theta \right ) =\sum _{m=1}^{\infty }\frac{\pi }{2J_{n}\left ( kr_{0}\right ) }\left ( \int _{0}^{2\pi }f\left ( \theta ^{\prime }\right ) \left ( \cos \left ( n\left ( \theta ^{\prime }+\theta \right ) \right ) +\cos \left ( n\left ( \theta ^{\prime }-\theta \right ) \right ) \right ) d\theta ^{\prime }+\int _{0}^{2\pi }f\left ( \theta ^{\prime }\right ) \left ( \cos \left ( n\left ( \theta ^{\prime }-\theta \right ) \right ) -\cos \left ( n\left ( \theta ^{\prime }+\theta \right ) \right ) \right ) d\theta ^{\prime }\right ) J_{n}\left ( kr\right )

Which is simplified to, after combining both integrals to one\begin{align*} u\left ( r,\theta \right ) & =\sum _{m=1}^{\infty }\frac{\pi }{2J_{n}\left ( kr_{0}\right ) }\left ( \int _{0}^{2\pi }f\left ( \theta ^{\prime }\right ) \left ( \cos \left ( n\left ( \theta ^{\prime }+\theta \right ) \right ) +\cos \left ( n\left ( \theta ^{\prime }-\theta \right ) \right ) +\cos \left ( n\left ( \theta ^{\prime }-\theta \right ) \right ) -\cos n\left ( \theta ^{\prime }+\theta \right ) \right ) d\theta ^{\prime }\right ) J_{n}\left ( kr\right ) \\ & =\sum _{m=1}^{\infty }\frac{\pi }{2J_{n}\left ( kr_{0}\right ) }\left [ \int _{0}^{2\pi }f\left ( \theta ^{\prime }\right ) 2\cos \left ( \theta ^{\prime }-\theta \right ) d\theta ^{\prime }\right ] J_{n}\left ( kr\right ) \\ & =\sum _{m=1}^{\infty }\left [ \int _{0}^{2\pi }f\left ( \theta ^{\prime }\right ) \frac{\pi }{J_{n}\left ( kr_{0}\right ) }\cos \left ( \theta ^{\prime }-\theta \right ) d\theta ^{\prime }\right ] J_{n}\left ( kr\right ) \end{align*}

Exchanging integration with summation gives u\left ( r,\theta \right ) =\int _{0}^{2\pi }f\left ( \theta ^{\prime }\right ) \left ( \sum _{m=1}^{\infty }\frac{\pi }{J_{n}\left ( kr_{0}\right ) }\cos \left ( \theta ^{\prime }-\theta \right ) J_{n}\left ( kr\right ) \right ) d\theta ^{\prime }

Comparing the above to u\left ( r,\theta \right ) =\int _{0}^{2\pi }f\left ( \theta ^{\prime }\right ) G\left ( r,\theta ;\theta ^{\prime }\right ) d\theta ^{\prime }
Shows that Green function is G\left ( r,\theta ;\theta ^{\prime }\right ) =\sum _{m=1}^{\infty }\frac{\pi }{J_{n}\left ( kr_{0}\right ) }\cos \left ( \theta ^{\prime }-\theta \right ) J_{n}\left ( kr\right )
Where r_{0} is radius of disk. It is symmetric in \theta as expected.

2.11.6 Key solution for HW 11

PDF