next up previous
Next: About this document ...

PHYSICS 312. Midterm Examination, March 2 2001, Solution


Problem 1 :
a:
Let us write

f(x)=f1(x)+f2(x)


\begin{displaymath}f_1(x)=\sin(x)\end{displaymath}


\begin{displaymath}f_2(x)=\cos(x)\end{displaymath}


\begin{displaymath}g_1(x)=-\sin(x),\;\; -\pi<x<0\end{displaymath}

is the periodic extension of f1(x) and also the even periodic extension. Next

\begin{displaymath}g_2(x)=\cos(x),\;\; -\pi<x<0\end{displaymath}

is the even periodic extension, but not the periodic extension of f2(x). So statements 1 and 2 and 4 are false, while 3 is true!
b: Since we are dealing with the even periodic extension the Fourier series is a cosine series. So statement 1 is truewhile 2 and 3 are false!
c: Calculate the coefficient $\alpha_0$.

\begin{displaymath}\alpha_0=\frac{1}{2\pi}\int_0^\pi\sin(x)dx-\frac{1}{2\pi}\int...
...dx+
\int_0^{2\pi}\cos(x)dx=\frac{1}{2\pi}(2+2+0)=\frac{2}{\pi}\end{displaymath}



Problem 2:
By solving the characteristic equation

\begin{displaymath}\gamma^2+2\gamma+\lambda\end{displaymath}


\begin{displaymath}\gamma=-1\pm i\sqrt{\lambda-1}\end{displaymath}

we find that the general solution to

\begin{displaymath}\frac{d^2\phi}{dx^2}+2\frac{d\phi}{dx}+\lambda\phi=0\end{displaymath}

is

\begin{displaymath}\phi(x)=e^{-x}(\alpha\cos(x\sqrt{\lambda-1})+\beta\sin(x\sqrt{\lambda-1}))\end{displaymath}

if

\begin{displaymath}\lambda>1\end{displaymath}

and

\begin{displaymath}\phi(x)=(\alpha+\beta x)e^{-x}\end{displaymath}

for $\lambda=1$ and

\begin{displaymath}\phi(x)=e^{-x}(\alpha e^{x\sqrt{1-\lambda}}+\beta e^{-x\sqrt{1-\lambda}})\end{displaymath}

for $\lambda<1$. The boundary conditions were

\begin{displaymath}\phi(0)=0\end{displaymath}


\begin{displaymath}\phi(a)=0\end{displaymath}

a: For $\lambda>1$ we have $\alpha=0$ and

\begin{displaymath}\sqrt{\lambda-1}=\frac{n\pi}{a}\end{displaymath}

or

\begin{displaymath}\lambda=(\frac{n\pi}{a})^2+1\end{displaymath}

with eigenfunctions

\begin{displaymath}\phi(x)=e^{-x}\sin(\frac{n\pi}{a}),\;\;n=1,2,3,...\end{displaymath}

b: For $\lambda<1$ including $\lambda<0$ it is impossible to satisfy the boundary conditions. The same holds for $\lambda=1$, so $\lambda=1$ is not an eigenvalue.


Problem 3:
The steady state solution satisfies

\begin{displaymath}\frac{\partial}{\partial x}(\kappa(x)\frac{\partial T_S(x,t)}{\partial x})=0\end{displaymath}


\begin{displaymath}\kappa(x)\frac{\partial T_S(x,t)}{\partial x}=c\end{displaymath}


\begin{displaymath}\frac{\partial T_S(x,t)}{\partial x}=\frac{c}{\kappa_0}(1+\alpha x)\end{displaymath}


\begin{displaymath}T_S=\frac{c}{\kappa_0}(x+\alpha\frac{x^2}{2})+d\end{displaymath}

The boundary condition at x=0 gives d=0, while at the other end

\begin{displaymath}T_0=\frac{c}{\kappa_0}(L+\frac{\alpha L^2}{2})\end{displaymath}

so

\begin{displaymath}T_S=T_0\frac{x+\frac{\alpha x^2}{2}}{L+\frac{\alpha L^2}{2}}\end{displaymath}



Return to title page

 
next up previous
Next: About this document ...
Birger Bergersen
2001-03-12