0
$\begingroup$

I am trying to solve the heat equation $$ u_t=u_{xx} $$ with $$u(0,t)=0, u(1,t)=100, u(x,0)=\sin(\pi\,x), 0<x<1$$

obviously one solution is $100\,x$ for $u(0,t)=0, u(1,t)=100$, and I am supposed to use superposition to come up with a series solution.

my initial reaction is to consider this like a neumann problem letting $$u_x(1,t)=0, u_x(0,t)=0$$ and using seperation of variables to get $$u_n(x,t)=\cos(n\,\pi\,x)e^{n^2 \pi^2 t}$$ Then utilizing super position I can rerwrite my solution as

$$ u(x,t) = a_0/2 + \sum_{n=1}^\infty a_n\cos(n\,\pi\,x)e^{n^2 \pi^2 t} +100x $$

keeping in mind that for $x=1$ the terms other then $100x$ go to $0$ as the $a_n$ and $a_0$ are picked in a way that for $u(x,0)$ we get $\sin(n \pi \,x)e^{n^2 \pi^2 t} + 100x$ which is $0+100x$. thus satisfying the initial condition I still feel as if I messed up as it does not satisfy $u(x,0)=\sin(\pi\,x)$

maybe I pick my $a$ values differently? Or maybe pick $100x$ to be $100 \alpha$ where $\alpha$ is a picecewise function that is 0 everywhere except x=1?

$\endgroup$
1
$\begingroup$

Why Neumann boundary conditions? Let $u=v(x,t)+100\,x$. Then $v$ satisfies the heat equation $v_t-v_{xx}=0$ with boundary and initial conditions $$ v(0,t)=0,\quad v(1,t)=0,\quad v(x,0)=\sin(\pi\,x)-100\,x. $$ The solution will be of the form $$ v(x,t)=\sum_{k=1}^\infty a_n\,e^{n^2\pi^2t}\sin(n\,\pi\,x). $$

| cite | improve this answer | |
$\endgroup$
  • $\begingroup$ Was thinking neumann, as if I take the derivative of my boundary conditions, I can make them homogeneous which I know how to solve $\endgroup$ – yipper Aug 8 '18 at 10:12
  • $\begingroup$ You cannot take the derivative of the boundary conditions: there is no $x$ in $u(0,t)=0$. $\endgroup$ – Julián Aguirre Aug 8 '18 at 10:15
  • $\begingroup$ I see, thankyou $\endgroup$ – yipper Aug 8 '18 at 10:15
  • $\begingroup$ if you dont mind, could you explain to me why I cannot take the derivative of $u(0,t)=0$ by using separation of variables I have $X(0)T(t)=u(0,t)$ so wouldn't we assume any derivative of this would be 0? (sorry I dont have any experience taking derivatives of boundary conditions I only saw it as a potential way to get the problem into a familiar form) $\endgroup$ – yipper Aug 8 '18 at 10:36
  • $\begingroup$ $$\frac{\partial u}{\partial x}(0,t)\ne\frac{\partial}{\partial x}\bigl(u(0,t)\bigr)$$ $\endgroup$ – Julián Aguirre Aug 8 '18 at 10:39

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.