1
$\begingroup$

I am trying to find the equation of motion for a particle constrained to move on the surface defined by $S:z=\cos x+\sin y$ under the influence of gravity. I am working in the Cartesian Coordinate system, so the force of gravity is given by $F_g=-mg\hat{k}$ where $\hat{k}$ is a unit vector in the upwards z-direction. Thus, the Lagrangian for this particle would be (I'm not sure if this is correct): $$\mathcal{L}=\frac{1}{2}m\left(\dot{x}^2+\dot{y}^2+\dot{z}^2\right)-mgz+\lambda\left(\cos x+\sin y-z\right)$$ I am using the method of lagrange multipliers to solve for the equations of motion. Is there an easier way? I then used the Euler-Lagrange equations to find my equations of motion: $$\begin{align} m\ddot{x}&=-\lambda\sin x\\ m\ddot{y}&=\lambda \sin y\\ m\ddot{z}&=-mg-\lambda\\ z &=\cos x +\sin y \end{align}$$ Now, I am not really sure what I need to do with the $\lambda$. I solved for it ($\lambda=-m(g+\ddot{z})$, and substituted it in the equations, leading me to these: $$ \begin{align} \ddot{x}&=(g+\ddot{z})\sin x \\ \ddot{y}&=(g+\ddot{z})\sin y \\ m\ddot{z}&=-mg-\lambda \end{align} $$ But I still haven't completely isolated $\lambda$. I then differentiated the constraint with respect to time twice, $$\ddot{z}=-\left(\dot{x}^2\cos x +\dot{y}^2\sin y\right)$$ And I substituted for $\ddot{z}$ in the equation solved for $\lambda$: $$\lambda=-m\left(g-\left(\dot{x}^2\cos x +\dot{y}^2\sin y\right)\right)$$ I then eliminated $\lambda$ in the last equation: $$m\ddot{z}=-mg+m\left(g-\left(\dot{x}^2\cos x +\dot{y}^2\sin y\right)\right)$$ $$\ddot{z}=-g+\left(g-\left(\dot{x}^2\cos x +\dot{y}^2\sin y\right)\right)$$ $$\ddot{z}=-\left(\dot{x}^2\cos x +\dot{y}^2\sin y\right)$$ Here are my questions: Is all of this correct? Do you think there are analytical solutions to these differential equations?

$\endgroup$
5
  • 1
    $\begingroup$ I thought I did? Isn't that my fourth equation of motion? $\endgroup$
    – Zachary
    Jan 6 '18 at 3:30
  • 1
    $\begingroup$ There are at least a couple of mistakes in your working. (1) Equation for $m \ddot y$ is wrong. (2) Equation for $\ddot z$ is wrong. $\endgroup$
    – Myridium
    Jan 6 '18 at 4:15
  • 1
    $\begingroup$ why don't you use coordinates adapted to the constraint instead of the cumbersome Lagrange-multipliers method? $\endgroup$ Jan 6 '18 at 11:12
  • 1
    $\begingroup$ I mean, using coordinates $x$ and $y$ so that $\dot{z}=-\dot{x}\sin x + \dot{y}\cos y$ the Lagrangian reads $\frac{m}{2}[\dot{x}^2(1+ \sin^2x)] + \dot{y}^2(1+ \cos^2 y)] - mg (\cos x + \sin y)$. The equations of motion arise by the associated Euler-Lagrange equations. $\endgroup$ Jan 6 '18 at 11:16
  • $\begingroup$ Also, I believe your $\ddot{z}$ equation should be: $z'' = $-x''(t) \sin (x(t))+x'(t)^2 (-\cos (x(t)))+y''(t) \cos (y(t))-y'(t)^2 \sin (y(t))$$. $\endgroup$ Jan 6 '18 at 14:28
1
$\begingroup$

Deriving twice concerning $t$ the smooth constraint gives

$$ -x''(t) \sin (x(t))-x'(t)^2 \cos (x(t))+y''(t) \cos (y(t))-y'(t)^2 \sin (y(t))-z''(t)=0 $$

which jointly with

$$ \left\{ \begin{array}{rcl} -\lambda \sin (x(t))-m x''(t)=0 \\ \lambda \cos (y(t))-m y''(t)=0 \\ -\lambda -g m-m z''(t)=0 \\ \end{array} \right. $$

can be solved to $x''(t),y''(t),z''(t),\lambda$ giving

$$ \left\{ \begin{array}{rcl} x''(t)& = & \frac{\sin (x(t)) \left(-\cos (x(t)) x'(t)^2-\sin (y(t)) y'(t)^2+g\right)}{\cos ^2(y(t))+\sin ^2(x(t))+1} \\ y''(t)& =& \frac{\cos (y(t)) \left(\cos (x(t)) x'(t)^2+\sin (y(t)) y'(t)^2-g\right)}{\cos ^2(y(t))+\sin ^2(x(t))+1} \\ z''(t)& =& \frac{2 \left(\cos (x(t)) x'(t)^2+\sin (y(t)) y'(t)^2-g\right)}{\cos (2 x(t))-\cos (2 y(t))-4}-g \\ \lambda & =& \frac{m \left(\cos (x(t)) x'(t)^2+\sin (y(t)) y'(t)^2-g\right)}{\cos ^2(y(t))+\sin ^2(x(t))+1} \\ \end{array} \right. $$

Attached the movement for initial conditions $x'(0) = y'(0)=0.01, x(0) = 0,y(0)=\pi/2$

enter image description here

$\endgroup$
0
$\begingroup$

To expand the commend of Valter Moretti, I would also use a set of coordinates which already satisfy the spatial constraints. So I would start by parameterizing the position vector as $$ \mathbf r(t) = \begin{pmatrix} x(t) \\ y(t) \\ z(t) \end{pmatrix} = \begin{pmatrix} x(t) \\ y(t) \\ \lambda \cos(\mu\, x(t)) + \lambda\sin(\mu\, y(t)) \end{pmatrix}, $$ where I introduced the real parameters $\mu$, where $[\mu] = m^{-1}$, and $\lambda$, where $[\lambda] = m$, to take care of the units. By differentiation we get the velocity vector $$ \dot{\mathbf{r}}(t) = \begin{pmatrix} \dot x(t) \\ \dot y(t) \\ -\lambda\mu \sin(\mu\, x(t))\ \dot x(t) + \lambda\mu\cos(\mu\, y(t))\ \dot y (t) \end{pmatrix}. $$ The Lagrangian is then given by $$ \begin{align} L &= T(\dot{\mathbf{r}}) - U(\mathbf r)\\ &=\frac m 2 \dot{\mathbf{r}}^2 - mg\,\mathbf r\cdot \mathbf e _z \\ &=\frac m 2 (\dot x ^2 + \dot y ^2 + \lambda^2\mu^2(\sin^2(\mu x)\dot x^2 - 2\sin (\mu x)\cos(\mu y)\dot x \dot y +\cos^2(\mu x)\dot y^2)\\&\phantom{==} -mg\lambda(\cos(\mu x) + \sin(\mu y )) \end{align} $$ and the equations of motion are then given by $$ \frac{\text{d}}{\text{d}t} \frac{\partial L}{\partial \dot x} - \frac{\partial L}{\partial x} = 0 \quad\text{and}\quad \frac{\text{d}}{\text{d}t} \frac{\partial L}{\partial \dot y} - \frac{\partial L}{\partial y} = 0 $$ The function $z(t)$ is then found by using the solutions $x(t), y(t)$ in $z(t) = \lambda \cos(\mu\, x(t)) + \lambda\sin(\mu\, y(t))$.

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.