next up previous
Next: About this document ...

PHYSICS 312. Midterm Examination. February 27, 2000
SOLUTION


Problem 1 :

\begin{displaymath}f(x)=\sin(x)+\cos(x);\;0<x<\pi\end{displaymath}

The function g(x) in the range $-\pi<x<0$ that constitutes the
a: periodic extension of f(x) is

\begin{displaymath}-\sin(x)-\cos(x)\end{displaymath}

b: the odd periodic extension of f(x) is

\begin{displaymath}\sin(x)-\cos(x)\end{displaymath}

c: the even periodic extension of f(x) is

\begin{displaymath}-\sin(x)+\cos(x)\end{displaymath}



Problem 2:
Let us define $k=\sqrt{\lambda}$ The general solution to

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

is

\begin{displaymath}\phi=A\sin(kx)+B\cos(kx)\end{displaymath}

and

\begin{displaymath}\phi^\prime=Ak\cos(kx)-Bk\sin(kx)\end{displaymath}


\begin{displaymath}\phi(0)+\frac{d\phi}{dx}\vert _{x=0}=B+Ak\end{displaymath}

Hence B=-Ak and

\begin{displaymath}\phi(a)+\frac{d\phi}{dx}\vert _{x=a}=(A-Bk)\sin(ka)+(B+Ak)\cos(ka)=0\end{displaymath}

The cosine term above vansihes and we get

\begin{displaymath}A(1+k^2)\sin(ka)=0\end{displaymath}

or

\begin{displaymath}k=\frac{n\pi}{a};\;\;n=1,3\cdots\end{displaymath}

The eigenfunctions can then be written

\begin{displaymath}\phi_n=A(\sin(\frac{n\pi x}{a})-\frac{n\pi}{a}\cos(\frac{n\pi x}{a}))\end{displaymath}

and the eigenvalues

\begin{displaymath}\lambda_n=(\frac{n\pi}{a})^2\end{displaymath}



Problem 3:
The steady state temperature differential equation is

\begin{displaymath}\frac{d}{dx}((\kappa_0+\alpha T)\frac{dT}{dx})=Q\end{displaymath}

The ends x=0 and x=L were kept at the temperature T=0
a: We integrate this equation to get

\begin{displaymath}(\kappa_0+\alpha T)\frac{dT}{dx}=Qx+b\end{displaymath}

Integrating once more gives

\begin{displaymath}\kappa_0T+\frac{\alpha}{2}T^2=\frac{Q}{2}x^2+bx+c\end{displaymath}

The boundary condition T=0 at x=0 gives c=0 while T=0 for x=L gives b=-QL/2. This gives us the quadratic equation

\begin{displaymath}\kappa_0T+\frac{\alpha}{2}T^2=\frac{Q}{2}(x^2-Lx)\end{displaymath}

which can be solved for T to give for $a\neq 0$

\begin{displaymath}T=-\frac{\kappa_0}{\alpha}\pm\sqrt{\frac{\kappa_0^2}{\alpha^2}+\frac{Q}{\alpha}(x^2-Lx)}\end{displaymath}

b: For $\alpha>0$ we must choose the positive root to get T=0 for x=0 and x=L. Hence

\begin{displaymath}T=-\frac{\kappa_0}{\alpha}+\sqrt{\frac{\kappa_0^2}{\alpha^2}+\frac{Q}{\alpha}(x^2-Lx)}\end{displaymath}

c: $\alpha=0$ The division by $\alpha$ in the quadratic equation doesn't work and we get instead

\begin{displaymath}T=\frac{Q}{2\kappa_0}(x^2-Lx)\end{displaymath}

d: $\alpha<0$. We must now pick the negative root in the solution to the quadratic

\begin{displaymath}T=-\frac{\kappa_0}{\alpha}-\sqrt{\frac{\kappa_0^2}{\alpha^2}+\frac{Q}{\alpha}(x^2-Lx)}\end{displaymath}

Bonus question: If $\vert\alpha\vert$ is too large and $\alpha<0$ there no longer is a real solution in the middle of the bar. The maximum value of x2-Lx occurs for x=L/2 and is L2/4. Hence when

\begin{displaymath}Q>\frac{\-4\kappa_0^2}{L^2\alpha}\end{displaymath}

the calculation breaks down. What happens is that it gets so hot in the middle of the bar that the thermal conductivity turns negative and we get a runaway solution with more and more heat pouring in.

 
next up previous
Next: About this document ...
Birger Bergersen
2001-02-16