Understanding the terms in motion of a particle along a parametric curve

197 Views Asked by At

OK so I’m doing motion of a particle along a parametric curve we call $R(t)$. We have acceleration in terms of a unit normal vector and unit tangent vector. Acceleration at a point is given by $\vec{a} = \frac{d^2s}{dt^2}\vec{T}+\left(\frac{ds}{dt}\right)^2\cdot\frac{d\phi}{ds}\cdot\vec{N}$ Where $\vec{T}$ and $\vec{N}$ are the unit Tangent and normal vectors. My doubts are as follows :

  1. I can understand $\frac{ds}{dt}$ but not $\frac{d^2s}{dt^2}$. How do I visualize it?

  2. I know $\frac{d\phi}{ds}$ is the rate of change of $\phi$ with respect to arc length. But how do I perceive that?

  3. Finally we have $\vec{T}=\cos\phi\hat{\imath} + \sin\phi\hat{\jmath}$ and we say $\frac{d\vec{T}}{d\phi}$ is $\vec{N}$. But we can write it also as $\vec{T}=\frac{dx}{dt}\hat{\imath}+\frac{dy}{dt}\hat{\jmath}$. Now suppose I want to differentiate this with respect to ϕ then what? Also how do I visualize this because there are so many terms like $\frac{d}{d\phi}(\frac{dx}{dt})$.

3

There are 3 best solutions below

2
On

$\newcommand{\Vec}[1]{\mathbf{#1}}$To address your first two questions, let's assume your curve is regular, i.e., that the velocity $R'$ is nowhere-vanishing. Writing $$ R'(t) = \|R'(t)\|\, \frac{R'(t)}{\|R'(t)\|} = v(t) \Vec{T}(t) $$ uniquely expresses the velocity vector as a magnitude (the speed $v(t) = \frac{ds}{dt}$) and a direction (the unit vector $\Vec{T}(t)$).

Differentiating again, using the product rule, gives $$ R''(t) = v'(t) \Vec{T}(t) + v(t) \Vec{T}'(t). \tag{*} $$ The first summand is obviously parallel to the velocity. The second is orthogonal to the velocity: differentiating $\Vec{T}(t) \cdot \Vec{T}(t) = 1$ gives $\Vec{T}(t) \cdot \Vec{T}'(t) = 0$.

In words, (*) decomposes the acceleration into tangential and normal components. Respectively, these terms correspond to the change in speed $v'(t) = \frac{d^{2}s}{dt^{2}}$ (along the direction of motion) and the change in direction. (If $\Vec{T}'(t_{0}) \neq 0$, there is a unique circle lying in the plane through $R(t_{0})$ and spanned by $\Vec{T}(t_{0})$ and $\Vec{T}'(t_{0})$ that agrees with $R(t)$ to second order. If this circle has radius $r$, it turns out that $v(t_{0}) \|\Vec{T}'(t_{0})\| = \frac{v(t_{0})^{2}}{r}$, a well-known formula for centripetal acceleration.)

All of the preceding carries through for arbitrary regular curves, not just plane curves.

As for the third item, for a Euclidean plane curve the unit vector $\Vec{T}$ may be written as $(\cos\phi) \Vec{i} + (\sin\phi) \Vec{j}$, in which case $$ \frac{d\Vec{T}}{d\phi} = (-\sin\phi) \Vec{i} + (\cos\phi) \Vec{j} $$ is orthogonal to $\Vec{T}$. (This usage of Leibniz notation is likely to cause grief, as you've already noticed, though it expresses a simple idea: When a unit vector rotates, its "rate of change with respect to angle" is an orthogonal unit vector.)

The formula $\Vec{T}(t) = \frac{dx}{dt}\, \Vec{i} + \frac{dy}{dt}\, \Vec{j}$ is not correct unless the curve happens to have unit speed.

0
On

If you are driving a car along a windy road, $d^2s/dt^2$ is the acceleration you feel in the direction of motion of the car. The numerical value of this acceleration is proportional to the numerical value of the forward push of force you feel from the car seat behind you, if your foot is pressed hard on the accelerator pedal creating a positive value of $d^2s/dt^2$, or the backward pull of force you feel from the seat belt across your chest, if your foot is on the brake creating a negative value of $d^2s/dt^24.

If the car is maintaining constant speed, and so $ds^2/dt^2=0$, and hence the car is not accelerating in the forward direction at all, you are feeling no push at all from your seat nor pull from your seat belt. You will, of course, feel a push to one side or another depending on the direction of curvature of the road, and that push represents a force which is proportional to the perpendicular component of the acceleration.

For another example, race car drivers learn the concept of "accelerating through a curve", which means keeping your foot on the pedal and speeding up as you go around a curve (do not try this on your friendly neighborhood streets; and, frankly, I do not know why they learn this concept, but so I am told from friends who took racing courses). When you accelerate through a curve in this manner, you will feel a combination of pushing from the seat behind you ($d^2s/dt^2$, which is the tangential component of acceleration) and pull to the side opposite the curve (which is the normal component of acceleration). The resultant force will then be a diagonal push from behind and one side.

ADDED: You also ask about the normal component of acceleration, which is $\bigl(\frac{ds}{dt}\bigr)^2 d\phi/ds$. You can easily compute $d\phi/ds$ along a circle of radius $r$: the total angle equals $2\pi$, and the total length equals the circumference equals $2\pi r$, so $d\phi/ds = \frac{1}{r}$. Physically this means that as you drive a car around in a circle, the normal component of acceleration is inversely proportional to the radius, hence you feel a pull to the side (in the direction away from the inside of the circle) which is inversely proportional to the radius.

We know this intuitively: when driving around tight curves (small $r$, large $1/r$) we feel a lot of centrifugal force and our bodies are pulled hard to the side; driving around shallow curves (large $r$, small $1/r$) we barely feel the centrifugal force at all.

0
On

If you take a multi-flash photograph of a falling object which starts from rest with a high flash rate (like a stroboscope say 100 per sec) on the same photographic plate of a camera, then the separation between images of the object is observed to increase with each image. This is because speed $ds/dt$ is increasing. However acceleration of gravity ensures next such $ \dfrac{d^{2}s}{dt^2}$ to be constant.Consult any text book of physics.

$d\phi/ds$ is curvature, reciprocal of the radius of curving. It is a measure of how sharply the path is bending.

The third is obtained during differentiation procedure, has no big physical meaning.