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

Calculation boundary terms of a functional

  1. Sep 24, 2013 #1
    Dear all,

    I am stuck with the problem which is given below;

    In this problem the equilibrium equations of the given functional must be derived in u, v, and w directions from which the boundary terms must be found. I think that i have derived the equilibrium equations( 5 equations), but i dont know how to proceed. Does anyone maybe know how to do it??

    Thanks in advance,

    Attached Files:

  2. jcsd
  3. Sep 24, 2013 #2


    User Avatar
    Homework Helper

    I think you're missing a name for the integrand. I will use [itex]L[/itex].

    To determine [itex]\delta I[/itex], integrate [itex]L(u + h, \dots, v + k, \dots, w + l, \dots) - L(u, \dots, v, \dots, w, \dots)[/itex] term by term. You can swap the order of integration so that terms involving derivatives with respect to x are integrated with respect to x first, and terms involving derivatives with respect to y are integrated with respect to y first.

    For example, for the [itex]u_{,x}[/itex] term you get:
    \int_a^b \int_c^d h_{,x} \frac{\partial L}{\partial u_{,x}}\,\mathrm{d}x \,\mathrm{d}y
    = \int_a^b \left[ h \frac{\partial L}{\partial u_{,x}}\right]_c^d\,\mathrm{d}y - \int_a^b \int_c^d h \frac{\partial}{\partial x} \left(\frac{\partial L}{\partial u_{,x}}\right)\,\mathrm{d}x\,\mathrm{d}y[/tex]

    (The antiderivative with respect to x of [itex]h_{,x}[/itex] is [itex]h + A(y)[/itex] for an arbitrary function [itex]A[/itex]; but it will be the same function at both [itex]x = c[/itex] and [itex]x = d[/itex], so it cancels out.)

    Ideally the conditions of your problem are such that the perturbation [itex]h[/itex] vanishes everywhere on the boundary, so the boundary term vanishes.

    With the second derivatives of w you have to integrate by parts twice:
    \int_a^b \int_c^d l_{,xx} \frac{\partial L}{\partial w_{,xx}}\,\mathrm{d}x \,\mathrm{d}y
    = \int_a^b \left[ l_{,x} \frac{\partial L}{\partial w_{,xx}} \right]_c^d\,\mathrm{d}y
    - \int_a^b \int_c^d l_{,x} \frac{\partial}{\partial x}\left(\frac{\partial L}{\partial w_{,xx}}\right)\,\mathrm{d}x \,\mathrm{d}y
    Again, one would hope that the conditions of your problem require that [itex]l_{,x}[/itex] vanishes on the boundary, so the boundary term vanishes. The remaining term is dealt with as for the other first derivatives.

    Incidentally, I believe there should be three equilibrium equations (one each for u, v and w) so if you have five then you have made an error somewhere.
  4. Sep 24, 2013 #3
    I will try to solve the question now with your explanation. Thank you very much!!
  5. Sep 24, 2013 #4
    Regarding your explanation I ended up with the following equation to be solved.


    Is this correct or did I get your explanation totally wrong??(btw i forgot the dxdy term at the end of the equation)

    Attached Files:

    • 111.jpg
      File size:
      7.2 KB
Share this great discussion with others via Reddit, Google+, Twitter, or Facebook