1. Limited time only! Sign up for a free 30min personal tutor trial with Chegg Tutors
    Dismiss Notice
Dismiss Notice
Join Physics Forums Today!
The friendliest, high quality science and math community on the planet! Everyone who loves science is here!

Beam bending strain field

  1. Sep 4, 2012 #1
    If I have a wide beam, parallel to the x axis, with its COM at the origin, then I want it to curve about the y axis, what would the elements of the strain tensor be?

    I have come to the conlusion that the beam would, for example,contract above x axis and expand below it. But I don't know how to describe strain relative to the y and z axes, and then how to translate this knowledge to the strain tensor.

    Thanks for your time.
  2. jcsd
  3. Sep 4, 2012 #2
    "wide beam" to me sounds like a shell.

    Reissner-Mindlin shell theory is what is used in Finite Element Analysis. They neglect the higher order terms but there is a lot of information on the subject.

    If you are more interested in analytical theory that is analogous to standard beam theory, then I believe "Kirchhoff-Love" would be what you are looking for.

    Hope that helps,
  4. Sep 5, 2012 #3
    Hello Lewis, you really need to supply more information.

    Wide beam? : What sort of wide beam?

    Wide flanged I beams are available for increased bending resistance and their properties are tabulated.

    There are code requirements and specifications for wide RC beams.

    I am going to assume a simple homogeneous rectangular wide beam of height h in the y direction and breadth b in the z direction.

    Yes you need also to specify the z direction - wide beams are 3 dimensional.

    A beam is considered wide when

    b >> h say 5 times or more.

    Under these conditions the material is not free to expand or contract in the lateral z direction under bending along the x direction about the y direction. (note I said direction not axis)

    In particular εz ≈ 0 at z=0 so

    [tex]{\varepsilon _z} = \frac{1}{E}[{\sigma _z} - \nu \left( {{\sigma _x} + {\sigma _y}} \right) \approx 0[/tex]

    Since h is small σy ≈ 0 So

    [tex]{\sigma _z} = \nu {\sigma _{{x_{z = 0}}}}[/tex]


    [tex]{\varepsilon _x} = \frac{{1 - {\nu ^2}}}{E}{\sigma _x} = \frac{{1 - {\nu ^2}}}{{E{I_z}}}{M_z}y[/tex]

    In general the strain is reduced by a factor of [itex]{1 - {\nu ^2}}[/itex]

    So you can see that the change is the insertion a modifying constant into your strain tensor, I will leave you to do this since you haven't provided any notation.
    You need to be careful here since my strain is engineering strain, not tensor strain, which is a factor of 1/2 different.

    You should also note that the sideways distribution may also depend upon the support conditions.
    Last edited: Sep 5, 2012
Share this great discussion with others via Reddit, Google+, Twitter, or Facebook