Pendulum with Oscillating Force

Classical Mechanics Level pending

A particle of mass m = 1 m = 1 is attached to one end of a massless rigid rod of length = 1 \ell = 1 . The other end of the rod is attached to the origin, and the rod can rotate freely about the origin. The rod lies within the x y xy plane, and there is no gravity.

A time-varying force is applied to the particle (at the end of the rod).

F x = 0.1 cos ( 0.2 π t ) F y = 0.1 sin ( 0.2 π t ) F_x = 0.1 \cos(0.2 \, \pi t ) \\ F_y = 0.1 \sin(0.2 \, \pi t )

At time t = 0 t = 0 , the rod is aligned with the + x +x axis and is at rest. At time t = 6 t = 6 , what angle (in degrees) does the rod make with the + x +x axis?


The answer is 71.54.

This section requires Javascript.
You are seeing this because something didn't load right. We suggest you, (a) try refreshing the page, (b) enabling javascript if it is disabled on your browser and, finally, (c) loading the non-javascript version of this page . We're sorry about the hassle.

3 solutions

Steven Chase
Sep 6, 2020

@Karan Chatrath has given a very interesting solution showing the Cartesian version of Newton's Second Law. The code below shows the polar version. Specifially:

τ = r × F τ = I θ ¨ \vec{\tau} = \vec{r} \times \vec{F} \\ |\vec{\tau}| = I \ddot{\theta}

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
import math

deg = math.pi/180.0

dt = 10.0**(-6.0)

m = 1.0
L = 1.0

I = m*(L**2.0)

##########################

t = 0.0
count = 0

theta = 0.0
thetad = 0.0
thetadd = 0.0

while t <= 6.0:

    theta = theta + thetad*dt
    thetad = thetad + thetadd*dt

    x = L*math.cos(theta)
    y = L*math.sin(theta)

    Fx = 0.1*math.cos(0.2*math.pi*t)
    Fy = 0.1*math.sin(0.2*math.pi*t)

    T = x*Fy - y*Fx

    thetadd = T/I

    t = t + dt
    count = count + 1

    #if count % 1000 == 0:
        #print t,(theta/deg)

##########################

print ""
print ""
print dt
print t
print (theta/deg)

#1e-05
#6.00000999994
#71.5404169143
#>>> ================================ RESTART ================================
#>>> 

#1e-06
#6.00000000048
#71.5406387753
#>>> 

Thanks for sharing. Computing the net torque about the origin is a faster way of deriving the equation of motion.

Karan Chatrath - 9 months, 1 week ago

Log in to reply

I like your approach though. It is a different and interesting way of thinking about it

Steven Chase - 9 months, 1 week ago

It reminds me of that "acceleration constraint" technique that is used to solve those friction curve problems

Steven Chase - 9 months, 1 week ago

Log in to reply

Thank you! Yes, indeed, now that I see it, the approach is similar to those earlier friction problems. Fortunately not as messy, so I could work out the exact differential equation governing θ ¨ \ddot{\theta} . Otherwise, the EOM would have been computed implicitly by matrix inversion during numerical crunching.

Karan Chatrath - 9 months, 1 week ago

@Steven Chase Hey; I hope you don't mind me asking a question about your earlier problem; the Kapitza pendulum one.

I found this differential equation on Wikipedia related to it:

θ ¨ = ( g + α f 2 cos ( f t ) ) sin ( θ ) l \displaystyle \ddot{\theta} = -(g+\alpha f^2 \cos(ft)) \frac{\sin(\theta)}{l}

θ \theta is the angle to the negative y axis.

When I substituted values for frequency, amplitude, and whatnot, and numerically integrated the equation, I got a different result.

Here's the differential equation I got:

θ ¨ = ( g + 10 cos ( 10 t ) ) sin ( θ ) \displaystyle \ddot{\theta} = -(g+10 \cos(10t)) \sin(\theta)

I am aware this is incorrect. Did I perhaps wrongly calculate the support frequency? Or the support amplitude, wrongly?

Krishna Karthik - 9 months, 1 week ago

Log in to reply

Have you tried calculating the equations yourself using Lagrange analysis? It is actually not that hard to do

Steven Chase - 9 months, 1 week ago

Log in to reply

I thought that it would be really difficult. Anyway, I will try to soon.

Krishna Karthik - 9 months, 1 week ago

Lol to put this as a medium level problem is a bit too easy. I reckon it's a hard level one. Nice solution btw.

Krishna Karthik - 9 months, 1 week ago

Log in to reply

I suppose it depends on how you rate problems. Conceptually, there isn't much to this one. It's just a cross product for the torque. But if you choose to solve by hand, it gets messy.

Steven Chase - 9 months, 1 week ago
Mark Hennings
Sep 7, 2020

The position vector of the particle, and hence its acceleration, is r = ( cos θ sin θ ) r ¨ = ( sin θ cos θ ) θ ¨ ( cos θ sin θ ) θ ˙ 2 \mathbf{r} \; = \; \binom{\cos\theta}{\sin\theta} \hspace{2cm} \ddot{\mathbf{r}} \; = \; \binom{-\sin\theta}{\cos\theta}\ddot{\theta} - \binom{\cos\theta}{\sin\theta}\dot{\theta}^2 so we have the equation of motion ( sin θ cos θ ) θ ¨ ( cos θ sin θ ) θ ˙ 2 = 1 10 ( cos 1 5 π t sin 1 5 π t ) T ( cos θ sin θ ) \binom{-\sin\theta}{\cos\theta}\ddot{\theta} - \binom{\cos\theta}{\sin\theta}\dot{\theta}^2 \; = \; \tfrac{1}{10}\binom{\cos\frac15\pi t}{\sin\frac15\pi t} - T\binom{\cos\theta}{\sin\theta} where T T is the magnitude of the tension in the rod, and hence θ ¨ = 1 10 sin ( θ 1 5 π t ) \ddot{\theta} \; =\; -\tfrac{1}{10}\sin(\theta - \tfrac15\pi t) If we write φ = θ 1 5 π t \varphi = \theta - \tfrac15\pi t , we see that φ ¨ = 1 10 sin φ φ ( 0 ) = 0 , φ ˙ ( 0 ) = 1 5 π \ddot{\varphi} \; = \; -\tfrac{1}{10}\sin\varphi \hspace{2cm} \varphi(0) = 0\,,\,\dot{\varphi}(0) = -\tfrac15\pi and hence 1 2 φ ˙ 2 = 1 50 π 2 + 1 10 ( cos φ 1 ) = 1 50 [ π 2 10 sin 2 1 2 φ ] \tfrac12\dot{\varphi}^2 \; = \; \tfrac{1}{50}\pi^2 + \tfrac{1}{10}(\cos\varphi - 1) \; = \; \tfrac{1}{50}\big[\pi^2 - 10\sin^2\tfrac12\varphi\big] Thus φ \varphi oscillates between Φ = 2 sin 1 ( π 10 ) \Phi = 2\sin^{-1}\big(\tfrac{\pi}{\sqrt{10}}\big) and Φ -\Phi with period T = 2 Φ Φ 5 π 2 10 sin 2 1 2 u d u = 20 1 2 π 1 2 π d ψ 10 π 2 sin 2 ψ = 4 10 K ( 1 10 π 2 ) T \; =\; 2\int_{-\Phi}^\Phi \frac{5}{\sqrt{\pi^2 - 10\sin^2\frac12u}}\,du \; = \; 20\int_{-\frac12\pi}^{\frac12\pi} \frac{d\psi}{\sqrt{10 - \pi^2\sin^2\psi}} \; =\; 4\sqrt{10}\mathsf{K}\big(\tfrac{1}{10}\pi^2\big) where K \mathsf{K} is the complete elliptic integral of the first kind.

For 0 < t < 1 4 T 11.2722 0 < t < \tfrac14T \approx 11.2722 we see that φ \varphi varies between 0 0 and Φ -\Phi , and we have t = t ( φ ) = φ 0 5 d u π 2 10 sin 2 1 2 u t \; = \; t(\varphi) \; = \; \int_{\varphi}^0 \frac{5\,du}{\sqrt{\pi^2 - 10\sin^2\frac12u}} We can solve this equation numerically to deduce that φ = 2.52129056 \varphi = -2.52129056 when t = 6 t=6 , and so θ = 6 5 π 2.52129056 \theta = \tfrac65\pi - 2.52129056 , which is 71.5407 \boxed{71.5407}^\circ , when t = 6 t=6 .

Since θ = 1 5 π t + φ \theta = \tfrac15\pi t + \varphi , this is not an unstable oscillation, but simply superposition of the oscillation of φ \varphi with the constant angular velocity 1 5 π \tfrac15\pi imposed by the driving force.

@Mark Hennings (Upvoted)Your Anayltical solution always rocks.
By the way, how do you solve elliptical integration?

Talulah Riley - 9 months, 1 week ago

Log in to reply

Elliptical integrals are extremely complicated. They have a connection with the arc length of an ellipse, which itself is a notoriously difficult problem. Btw don't put more than one question mark... some people think it's rude or pushy. Just a word of advise. Have a nice day bro.

Krishna Karthik - 9 months ago

Log in to reply

@Krishna Karthik oh sorry, from now onwards I will never put 2 or more question mark.
Thanks for informing, I think therefore he didn't replied.
Have a nice day.

Talulah Riley - 9 months ago

Your question is unclear. Do you mean, how did I obtain the formulae for T T , or do you mean how do I evaluate the final (ellptic) integral?

Q1 - I get from the first integral form to the second by substitution, and the second is in the right form for K \textsf{K} .

Q2 - Elliptic integrals relate to questions about arc-length along ellipses, and cannot (except in special cases) be evaluated exactly, in the sense that they cannot be written in terms of more elementary functions. When they meet this sort of problem, mathematicians just invent a new name for the function. There are hundreds (at least) of these so-called special functions to be found. This is nothing particularly weird - in one sense the trig functions are special, since they cannot be written in terms of simpler functions such as polynomials and raking roots, etc, The trig functions are just so common that they are not regarded as special any more.

The values of elliptic integrals are evaluated just as the values of trig functions are evaluated - numerically. The trig functions are just so common that they are hard-wired into your calculator, so you don't notice the effort! When I was your age, or maybe a little younger, we had to look up trig functions in books of tables.

Mark Hennings - 9 months ago

Thank you for the insightful solution.

Karan Chatrath - 9 months, 1 week ago

The Mark Hennings. Nice one sir.

Krishna Karthik - 9 months, 1 week ago
Karan Chatrath
Sep 6, 2020

The coordinates of the particle are:

( x , y ) = ( cos θ , sin θ ) (x,y) = (\cos{\theta},\sin{\theta})

Where θ \theta the rod makes with the +X axis. The velocity and acceleration components are:

( x ˙ , y ˙ ) = ( sin θ θ ˙ , cos θ θ ˙ ) (\dot{x},\dot{y}) = (-\sin{\theta} \ \dot{\theta},\cos{\theta} \ \dot{\theta}) ( x ¨ , y ¨ ) = ( a x , a y ) = ( sin θ θ ¨ cos θ θ ˙ 2 , cos θ θ ¨ sin θ θ ˙ 2 ) (\ddot{x},\ddot{y}) = (a_x,a_y)=(-\sin{\theta} \ \ddot{\theta} - \cos{\theta} \ \dot{\theta}^2,\cos{\theta} \ \ddot{\theta}- \sin{\theta} \ \dot{\theta}^2)

Let the tension in the rod be T T directed along the rod and towards the origin. Applying Newton's second law along the X and Y directions gives:

a x = F x T cos θ a_x = F_x - T\cos{\theta} a y = F y T sin θ a_y = F_y - T\sin{\theta}

Plugging in all values and re-arranging everything in a matrix form gives the set of linear equations in θ ¨ \ddot{\theta} and T T as follows:

[ sin θ cos θ cos θ sin θ ] [ θ ¨ T ] = [ F x + θ ˙ 2 cos θ F y + θ ˙ 2 sin θ ] \left[\begin{matrix} -\sin{\theta}&\cos{\theta}\\\cos{\theta}&\sin{\theta}\end{matrix}\right]\left[\begin{matrix} \ddot{\theta}\\T\end{matrix}\right]=\left[\begin{matrix} F_x+\dot{\theta}^2\cos{\theta}\\F_y+\dot{\theta}^2\sin{\theta}\end{matrix}\right]

[ θ ¨ T ] = [ sin θ cos θ cos θ sin θ ] 1 [ F x + θ ˙ 2 cos θ F y + θ ˙ 2 sin θ ] \implies \left[\begin{matrix} \ddot{\theta}\\T\end{matrix}\right] = \left[\begin{matrix} -\sin{\theta}&\cos{\theta}\\\cos{\theta}&\sin{\theta}\end{matrix}\right]^{-1} \left[\begin{matrix} F_x+\dot{\theta}^2\cos{\theta}\\F_y+\dot{\theta}^2\sin{\theta}\end{matrix}\right]

A little bit of inspection allows us to see that the matrix to be inverted is the inverse of itself. So it follows that:

[ θ ¨ T ] = [ sin θ cos θ cos θ sin θ ] [ F x + θ ˙ 2 cos θ F y + θ ˙ 2 sin θ ] \implies \left[\begin{matrix} \ddot{\theta}\\T\end{matrix}\right] = \left[\begin{matrix} -\sin{\theta}&\cos{\theta}\\\cos{\theta}&\sin{\theta}\end{matrix}\right] \left[\begin{matrix} F_x+\dot{\theta}^2\cos{\theta}\\F_y+\dot{\theta}^2\sin{\theta}\end{matrix}\right]

Using this, an expression for θ ¨ \ddot{\theta} can be conveniently derived. The rest involves numerical integration: θ ¨ = sin θ ( F x + θ ˙ 2 cos θ ) + cos θ ( F y + θ ˙ 2 sin θ ) \ddot{\theta} = -\sin{\theta}(F_x+\dot{\theta}^2\cos{\theta}) + \cos{\theta} (F_y+\dot{\theta}^2\sin{\theta}) θ ¨ = F x sin θ + F y cos θ \implies\ddot{\theta} = -F_x \sin{\theta} + F_y \cos{\theta} θ ¨ = 0.1 sin ( 0.2 π t θ ) \implies \boxed{\ddot{\theta} = 0.1 \sin(0.2 \pi t - \theta)}

θ ( 6 ) 71.5 4 \theta(6) \approx 71.54^{\circ}

The system in consideration is clearly unstable.

Solution edited.

Karan Chatrath - 9 months, 1 week ago

Log in to reply

I like your solution. Of course, it's naturally way longer since you're dealing with a lot of manual derivation. This problem is extremely hard, naturally.

Krishna Karthik - 9 months, 1 week ago

We have a perfectly nicely behaved oscillation combined with a constant rotation of angular speed 1 5 π \tfrac15\pi ,

Mark Hennings - 9 months, 1 week ago

0 pending reports

×

Problem Loading...

Note Loading...

Set Loading...