I was solving an example of 1D wave equation with given BC and IC by separation of variables and Fourier series. $$\frac{\partial^2u}{\partial t^2}=c^2\frac{\partial^2u}{\partial x^2} $$ $$BC: u(0,t)=u(l,t)=0$$ $$IC:u(x,0)=\sin\left(\frac{\pi}{l} x\right) $$$$\partial_tu(x,0)=0$$ By separating the variables and solving the eigenvalue problem with the BC and solving the time ODE I get the general solution for u(x,t). $$u(x,t)=\sum_{n=1}^\infty\left[A_n \cos(\omega_nt)\sin\left(\frac{n\pi}{l}x\right)+B_n\sin(\omega_nt)\sin\left(\frac{n\pi}{l}x\right)\right]$$ Taking the derivative $$\frac{\partial u} {\partial t}=\sum_{n=1}^\infty\left[-\omega_nA_n \sin\left(\omega_nt\right)\sin\left(\frac{n\pi}{l}x\right)+\omega_nB_n\cos(\omega_nt)\sin\left(\frac{n\pi}{l}x\right)\right]$$ and applying the IC I got $$\partial_tu(x,0)=0=\sum_{n=1}^\infty\omega_nB_n\sin\left(\frac{n\pi}{l}x\right) \longrightarrow B_n=0 $$ since the integral is 0. Applying the other IC: $$u(x,0)=\sin\left(\frac{\pi}{l} x\right)=\sum_{n=1}^\infty A_n \sin\left(\frac{n\pi}{l}x\right)$$ Finding the Fourier coefficient(used wolframalpha to solve the integral) I get: $$A_n=\frac{2\sin(n\pi)}{\pi(n^2-1)}$$ And since $\sin(n\pi)=0$ for every integer $n$ my solution becomes $u(x,t)=0$ Am I missing something? Or what does the answer physically mean? There are no modes?

  • $\begingroup$ I've removed a number of comments that were attempting to answer the question and/or responses to them. Please keep in mind that comments should be used for suggesting improvements and requesting clarification on the question, not for answering. $\endgroup$
    – David Z
    Jul 2, 2020 at 20:46

2 Answers 2


Your solution $A_n$ isn't zero for every $n$. Looking at your initial conditions you might expect there to be only one non-zero coefficient since the IC already have the form of a sine wave. When $n=1$ you get $\frac 0 0$ so you have to be extra careful. Evaluating your solution for $A_n$ as a limit: $$\lim_{n\rightarrow 1}\frac{2\sin n\pi}{\pi(n^2-1)}=-1$$ I suspect there is an error in the integral because the answer should be +1. Alternatively you could immediately see from $$\sin\left(\frac \pi l x\right)=\sum_{n=1}^\infty A_n \sin\left(\frac {n\pi} l x\right)$$ that $A_n=\cases{1&$n=1$\\0 &$n\neq 1$}$

  • 1
    $\begingroup$ Thank you so much! :) I missed that my IC is expanded in the same basis and I can just compare. $\endgroup$ Jul 2, 2020 at 11:07

Wolframalpha is likely the source of confusion here: the initial condition is the term with $n=1$, i.e. $A_1 = 1$, whereas all others are $A_n=0$. The equation for the initial condition is one equation with the infinite number of unknowns in the rhs - no surprise that an automatic routine gives an irrelevant answer there.

  • 2
    $\begingroup$ Ooh! I don't have to expand for the coefficient since the IC is already expanded in the basis I am using for expansion. So I don't need to solve the integral at all. Since all coefficients are 0 except for n=1 my solution is $$u(x,t)=cos(\omega_1t)sin(\frac {\pi x}{l})$$. Is this correct? Meaning I have one mode. $\endgroup$ Jul 2, 2020 at 11:04
  • $\begingroup$ Yes, this is it! $\endgroup$
    – Roger V.
    Jul 2, 2020 at 11:09

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge that you have read and understand our privacy policy and code of conduct.

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