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

Homework Help: Analytical mechanics problem

  1. Sep 23, 2006 #1
    As in the attached picture.
    A cylindrical shell with mass M can roll without gliding on a horizontal plane

    In the cylindrical shell a particle ,p, with mass m can glide without friction.

    At the begining there is no motion and the angle to the particles position is [tex] \phi=\frac{\pi}{2} [/tex]

    I am suposed to find the movement of the center of the circle as a function of the angle [tex] \phi [/tex]

    Im not sure how I should start.

    The potential energy of the system is, if I place the plane of reference on the level of the horizontal plane.

    [tex] V=mgR(1-cos\phi) [/tex]

    Now this problem obviously only has one degree of freedom and that is the angle [tex] \phi [/tex]

    But if I want to construct a lagrangian I must use two degres of freedom. The rotation angle of the cylinder [tex] \alpha [/tex] and the angle to the particle p [tex] \phi [/tex]. Because I dont know how [tex] \alpha [/tex] and [tex] \phi [/tex] are connected. Is this the correct thinking?

    In that manner I get this equation

    [tex] L=T-V=\frac{3}{4}MR^2\dot{\alpha}^2+\frac{1}{2}m[\dot{\alpha}^2 R^2 + R^2\dot{\phi}^2 + \dot{\alpha}^2 R^2 sin\phi ] - mgR(1-cos\phi ) [/tex]

    Should I use this and solve the two lagrande equations

    [tex] \frac {d}{dt} \frac{dL}{d\dot{\alpha}}-\frac{dL}{d\alpha}=0 [/tex]

    [tex] \frac {d}{dt} \frac{dL}{d\dot{\phi}}-\frac{dL}{d\phi}=0 [/tex]

    Im not sure if this will give me any answere though? It feels like I should express [tex] \alpha [/tex] in [tex] \phi [/tex] before I even do the lagrangian. Am I on the right track?

    Attached Files:

    Last edited: Sep 23, 2006
  2. jcsd
  3. Sep 23, 2006 #2
    I assume that I in the same manner can get the hamiltonian since its a conservative force situation and holonomic constrains.

    [tex] H=T+V=\frac{3}{4}MR^2\dot{\alpha}^2+\frac{1}{2}m[\dot{\alpha}^2 R^2 + R^2\dot{\phi}^2 + \dot{\alpha}^2 R^2 sin\phi ] + mgR(1-cos\phi ) [/tex]

    But Im not quite sure what I would do with that one either. I know that

    [tex] \dot{\alpha}=\frac{dH}{dP_{\alpha}}[/tex]
    [tex] -\dot{P_{\alpha}}=\frac{dH}{d\alpha} [/tex]

    But that doesnt seem to be much help either??
    Last edited: Sep 23, 2006
  4. Sep 28, 2006 #3
    I don't know...I wish I could help...

    It seems like you really do need two generalized coordinates. I mean, the particle is free to move; it doesn't seem like it is constrained in any way such that [tex]\phi[/tex] and [tex]\alpha[/tex] would be related.
  5. Sep 28, 2006 #4
    I think that there is an error in your Lagrangian: Since you are interested in the movement of the cylinder and watch it in the "labatory system", so to say, you can't say that the velocity of the particle in y-direction is [itex] R \dot \phi \cos(\phi)[/itex]. [itex] \phi [/itex] is fixed in the "cylinder system", so you have you add the y-velocity of the cylinder. If you square this, you should get a "mixed term" so that the both velocities are related.
    Last edited: Sep 28, 2006
  6. Sep 28, 2006 #5
    I managed to solve it. my expression for the Y position of the particle was

    [tex] Rsin\phi + R\alpha [/tex]

    So velocity was

    [tex] R\dot{\phi}cos\phi + R\dot{\alpha}[/tex]

    I used the fact that the conjugated momentum is constant for the alpha coordinate since there is no alpha, just alpha dot, dependance :) tricky problem.
Share this great discussion with others via Reddit, Google+, Twitter, or Facebook