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

COMSOL - Structural Mechanics - Membrane - vibration without input?

  1. Aug 8, 2014 #1
    I've just got a quick question that hopefully someone can answer.

    My question is - why does the membrane in COMSOL vibrate without you telling it to?

    Here are some specifics if required...

    In COMSOL here are my setup details: -

    Physics: (3D) Structural mechanics - Membrane
    Study: Eigenfrequency

    Geometry->Work plane->Square...Size->Side length = 27.4[mm]
    Materials->Material 1...Geometric Entity Selection->Selection = All boundaries
    Materials->Material 1...Material Contents: Density (rho) = 1200[kg/m^3], Young's modulus (E) = 3.6e9[Pa] & Poisson's ratio (nu) = 0.36.
    Membrane...Boundary Selection->Selection = All boundaries;
    Membrane...Thickness->d = 0.176[mm]
    Membrane->Linear Elastic Material 1->Initial Stress and Strain 1...Initial Stress and Strain->Initial in-plane force->N_0 = 486.4[N/m], 0, 0, 486.4[N/m] (these go top left, top right, bottom left then bottom right)
    Membrane->Fixed Constraint 1...Edge Selection->Selection: All edges
    Study->Step 1: Eigenfrequency...Study Settings->Desired number of eigenfrequencies = 8
    Study->Step 1: Eigenfrequency...Study Settings->Search for eigenfrequencies around = 50

    When I compute this, I get a series of total displacement 2D plots for each eigenfrequency, showing the displacement and you can even view pressure too. But what is driving this?

    I'd like to add an input later that I can define my self, rather than COMSOL just putting it in for me.

    Many thanks.

    https://dl.dropboxusercontent.com/u/11341635/Screen%20Shot%202014-08-08%20at%2008.08.15.png [Broken]
    Last edited by a moderator: May 6, 2017
  2. jcsd
  3. Aug 8, 2014 #2


    User Avatar
    Science Advisor
    Homework Helper

    The plots of the mode shapes just represent the possible vibration modes of the structure. The amplitudes of displacement, stress, etc, on the plots are arbitrary.

    If you want to find the actual physical response to a given situation, you will have to supply the relevant forces, and do a different type of solution - e.g. steady state forced response.
  4. Aug 11, 2014 #3
    Thanks - that's good to know.

    I've been struggling to set up a simple model using square membrane like the one above. I've tried loads of ways but nothing seems to work, and if you'd be willing to give me some advice I'd be really appreciative.

    The model is simple - a small square isotropic elastic membrane with a small shallow cylindrical mass attached to the centre. The edges of the membrane are attached to a square thin rigid plastic frame.

    I have the Young's modulus, Poisson's ratio & mass density for both the membrane and the mass along with the tension in the membrane.

    I need to get a 1-D plot showing transmission loss on the y-axis and frequency on the x-axis - so basically as if a white noise/sine-sweep sound source is incident on the membrane from one side - with the transmission loss being the dB equivalent of the incident pressure minus the transmitted pressure.

    What Physics would I use to do this, etc?

    I'm a day or two away from pulling my hair out.
    Last edited: Aug 11, 2014
  5. Jan 21, 2016 #4
    I have made a simple model of a 3d-square membrane and used solid
    mechanics and eigenfrequency analysis to get the eigenfrequencies. Now
    that i have them, I want to take the fundamental mode. Doing the
    Eigenfrequency analysis and finding the first 30 modes yielded close
    to expected results for the modes that are out of plane. However, i
    can't seem to get it to produce the eigenfrequencies that produce
    in-plane modes, or breathing modes. Is there an option in the
    eigenfrequency analysis to specify what direction to look in for the
    peaks? AKA look in the XY plane instead of the Z direction for maximas
    in deformation.

    Many thanks.

    Attached Files:

Share this great discussion with others via Reddit, Google+, Twitter, or Facebook