Dismiss Notice
Join Physics Forums Today!
The friendliest, high quality science and math community on the planet! Everyone who loves science is here!

Help to solve 4th order ode

  1. Dec 26, 2009 #1

    I try to solve this PDE as below:


    With boundary conditions of x=0.y=y''=0 and x=-h,y=y'=0.
    k and h are constant.
    The solution look like from the Fourier transforms.
    There is one term in the solution which is :

    beta_n is the positive roots of the equation of beta_n=tan(beta_n). n is from 1 to infinity.

    I have no idea how this term come from. If someone can help me, I really appreciate it!!

  2. jcsd
  3. Dec 26, 2009 #2
    First of all your ODE has the following form

    z''-kz+1=0, where z=y''

    The general solution to z is obtained by the wery standard method which lead to ODE

    y''= exp(k^(1/2)*x)*C1+exp(-k^(1/2)*x)*C2+1/k

    and after double integration

    y = C1/k*exp(k^(1/2)*x)+C2/k*exp(-k^(1/2)*x)+1/2/k*x^2+C3*x+C4

    what is the general solution to your ODE. Substitution of your boundary conditions gives you 4 algebraic equations for unknowns C1,C2,C3 and C4.
  4. Dec 26, 2009 #3


    User Avatar
    Science Advisor

    Almost the same thing: the characteristic equation for the differential equation y""- ky"+ 1= 0, which is the same as y""- ky"= -1, is [itex]r^4- kr^2= r^2(r^2- k)= r^2(r- \sqrt{k})(r+ \sqrt{k})[/itex] which has 0 as a double root and roots [itex]\sqrt{k}[/itex] and [itex]-\sqrt{k}[/itex].

    The general solution to the associated homogeneous equation is [itex]z= C_1x+ C_2+ C_3e^{x\sqrt{k}}+ C_4e^{-x\sqrt{k}}[/itex]. Since the right hand side is a constant, normally we would try a constant solution but since a constant and x already satisfy the homogeneous solution, we try [itex]z= Ax^2[/itex]. Then [itex]z'= 2Ax[/itex], [itex]z"= 2A[/itex] and [itex]z"'= z""= 0[/itex]. The equation becomes 0+ k(2A)= -1 so A= -1/(2k). The general soution to that equation is
    [tex]z(x)= C_1x+ C_2+ C_3e^{x\sqrt{k}}+ C_4e^{-x\sqrt{k}}-\frac{1}{2k}x^2[/tex]
  5. Dec 27, 2009 #4

    Thank you for response. If I want to use Fourier transform to solve this one, how can I find the boundary conditons for y''(-h)? I need to use inverse Laplace transform to convert the y function. Therefore, it will become much easier to use the Fourier transforms.

Know someone interested in this topic? Share this thread via Reddit, Google+, Twitter, or Facebook