Solve partial differential equation

Click For Summary
The discussion focuses on solving the partial differential equation (PDE) given by ∂u/∂t - k∂²u/∂x² = 0, with boundary conditions and an initial condition involving sin(x). The solution approach involves using the eigenfunctions of the differential operator and applying boundary conditions to derive expressions for u(x) and f(t). A key point raised is the challenge of expressing sin(x) in terms of the eigenfunctions, particularly the cosine functions, leading to confusion about orthogonality. It is clarified that sin(x) can be expressed as a series involving cosine functions by extending it as an even function and using half-range formulas. The discussion concludes that no errors were made in the calculations, but rather a misunderstanding of the orthogonality of the functions involved.
skrat
Messages
740
Reaction score
8

Homework Statement


Solve PDE ##\frac{\partial u}{\partial t}-k\frac{\partial^2 u}{\partial x^2}=0## for ##0\leq x \leq \pi ## and ##t\geq 0##.

Also ##\frac{\partial u}{\partial x}(0,t)=\frac{\partial u}{\partial t}(\pi ,t)=0## and ##u(x,0)=sin(x)##.

Homework Equations


The Attempt at a Solution



Differential operator ##L## for ##ku_{xx}=u_t## is symmetrical, therefore ##Lu=ku_{xx}=\lambda u##.

##ku_{xx}=\lambda u##

##ku_{xx}-\lambda u=0## so

##u(x)=Ae^{-\sqrt{\frac{\lambda}{k}}x}+Be^{\sqrt{\frac{\lambda}{k}}x}## and

##u(x)_x=-A\sqrt{\frac{\lambda}{k}} e^{-\sqrt{\frac{\lambda}{k}}x}+B\sqrt{\frac{\lambda}{k}} e^{\sqrt{\frac{\lambda}{k}}x}##

Using ##\frac{\partial u}{\partial x}(0,t)=0## gives me that ##A=B## and the second condition gives me:

##u(\pi )_x=\sqrt{\frac{\lambda}{k}}Ae^{-\sqrt{\frac{\lambda}{k}}\pi }(e^{2\sqrt{\frac{\lambda}{k}}\pi }-1)##

therefore

##2\sqrt{\frac{\lambda}{k}}\pi =2n\pi i## and finally ##\lambda =-n^2k##

So eigenfunctions of differential operator ##L## are :

##u(x)=Ae^{-\sqrt{\frac{\lambda}{k}}x}+Be^{\sqrt{\frac{\lambda}{k}}x}## where we already know that ##A=B## and inserting ##\lambda =-n^2k## leaves me with:

##u_n(x)=A_ncos(nx)##

IF everything so far is ok, than I shall continue:

##ku_{xx}=u_t##

##\sum_{n=0}^{\infty }k\lambda u(x)f(t)=\sum_{n=0}^{\infty }u(x){f}'(t)## so

##\frac{{f}'(t)}{f(t)}=\lambda k## and therefore

##f_n(t)=De^{-n^2k^2t}##So ##U(x,t)=\sum_{n=0}^{\infty }u(x)f(t)=\sum_{n=0}^{\infty }Acos(nx)e^{-n^2k^2t}##

Know I think I should somehow write that ##sin(x)## in ##u(x,0)=sin(x)## in terms of eigenfunctions of ##L## but something has to be wrong, because:

##<sin(x),cos(nx)>=<\sum_{n=0}^{\infty }A_ncos(nx),cos(kx)>## where the first scalar product is 0, which means what exactly? Where did I get it wrong?
 
Physics news on Phys.org
skrat said:

Homework Statement


Solve PDE ##\frac{\partial u}{\partial t}-k\frac{\partial^2 u}{\partial x^2}=0## for ##0\leq x \leq \pi ## and ##t\geq 0##.

Also ##\frac{\partial u}{\partial x}(0,t)=\frac{\partial u}{\partial t}(\pi ,t)=0## and ##u(x,0)=sin(x)##.

Homework Equations


The Attempt at a Solution



Differential operator ##L## for ##ku_{xx}=u_t## is symmetrical, therefore ##Lu=ku_{xx}=\lambda u##.

##ku_{xx}=\lambda u##

##ku_{xx}-\lambda u=0## so

##u(x)=Ae^{-\sqrt{\frac{\lambda}{k}}x}+Be^{\sqrt{\frac{\lambda}{k}}x}## and

##u(x)_x=-A\sqrt{\frac{\lambda}{k}} e^{-\sqrt{\frac{\lambda}{k}}x}+B\sqrt{\frac{\lambda}{k}} e^{\sqrt{\frac{\lambda}{k}}x}##

Using ##\frac{\partial u}{\partial x}(0,t)=0## gives me that ##A=B## and the second condition gives me:

##u(\pi )_x=\sqrt{\frac{\lambda}{k}}Ae^{-\sqrt{\frac{\lambda}{k}}\pi }(e^{2\sqrt{\frac{\lambda}{k}}\pi }-1)##

therefore

##2\sqrt{\frac{\lambda}{k}}\pi =2n\pi i## and finally ##\lambda =-n^2k##

So eigenfunctions of differential operator ##L## are :

##u(x)=Ae^{-\sqrt{\frac{\lambda}{k}}x}+Be^{\sqrt{\frac{\lambda}{k}}x}## where we already know that ##A=B## and inserting ##\lambda =-n^2k## leaves me with:

##u_n(x)=A_ncos(nx)##

IF everything so far is ok, than I shall continue:

##ku_{xx}=u_t##

##\sum_{n=0}^{\infty }k\lambda u(x)f(t)=\sum_{n=0}^{\infty }u(x){f}'(t)## so

##\frac{{f}'(t)}{f(t)}=\lambda k## and therefore

##f_n(t)=De^{-n^2k^2t}##So ##U(x,t)=\sum_{n=0}^{\infty }u(x)f(t)=\sum_{n=0}^{\infty }Acos(nx)e^{-n^2k^2t}##

Know I think I should somehow write that ##sin(x)## in ##u(x,0)=sin(x)## in terms of eigenfunctions of ##L## but something has to be wrong, because:

##<sin(x),cos(nx)>=<\sum_{n=0}^{\infty }A_ncos(nx),cos(kx)>## where the first scalar product is 0, which means what exactly? Where did I get it wrong?

Nothing is wrong. You want $$
\sin x = A_0 + \sum_{n=1}^\infty A_n \cos(nx)$$You want to extend ##\sin x## to ##(-\pi,0)## as an even function and use the half-range formulas for the ##A_n##.

[Edit, added] Alternatively if you are thinking in terms of inner products, your orthogonal set on ##(0,\pi)## is the set ##\{\cos(nx)\}##. ##\sin(x)## is not orthogonal to them.
 
Last edited:
Question: A clock's minute hand has length 4 and its hour hand has length 3. What is the distance between the tips at the moment when it is increasing most rapidly?(Putnam Exam Question) Answer: Making assumption that both the hands moves at constant angular velocities, the answer is ## \sqrt{7} .## But don't you think this assumption is somewhat doubtful and wrong?

Similar threads

Replies
6
Views
2K
Replies
9
Views
2K
  • · Replies 1 ·
Replies
1
Views
1K
  • · Replies 6 ·
Replies
6
Views
1K
  • · Replies 14 ·
Replies
14
Views
2K
Replies
4
Views
1K
  • · Replies 11 ·
Replies
11
Views
2K
  • · Replies 1 ·
Replies
1
Views
960
  • · Replies 1 ·
Replies
1
Views
2K
Replies
6
Views
2K