1
$\begingroup$

consider an object with a mass $m$ falling in a fluid with a drag force proportional to its velocity squared $(f=kv^2)$. the governing differential equation can be found using Newton's second law of motion as $$ A \frac{dv}{dt} + v^2 = v_{lim}^2 ~,~v(0)=v_0$$ where $A=\frac{m}{k}$ and $v_{lim}=\sqrt{\frac{mg}{k}}$ is the final velocity with the initial condition $v(0)=v_0$.
the solution of the equation is $$v(t)=v_{lim} \tanh \left(\frac{v_{lim}t}{A}+ \tanh^{-1} \left(\frac{v_0}{v_{lim}} \right) \right)$$ where $\tanh^{-1}$ is the inverse hyperbolic tangent.


The given solution does reflect the physical phenomenon in the case of $(v_0<v_{lim})$ ie the velocity increases from $v_0$ to $v_{lim}$.

In the case of $(v_0>v_{lim})$, the expected physics "behavior" of the solution is that the velocity decreases over time (from $v_0$ to $v_{lim}$) and yet we never see such decrease when plotting the function due to the fact that the $\tanh^{-1}(x)$ function is only defined for $x<1~ie~(v_0<v_{lim})$.
So, is there maybe another formula that better describe the second case?

$\endgroup$
8
  • $\begingroup$ Your starting equation is incorrect. The RHS should be zero. $\endgroup$ Commented Jan 26, 2023 at 12:27
  • $\begingroup$ Sorry for not understanding your remark, but what's an RHS? :) $\endgroup$ Commented Jan 26, 2023 at 12:32
  • $\begingroup$ Right-hand side. I don't think the comment by @ChetMiller is correct though; the RHS reflects the effect of gravity. $\endgroup$
    – Jakob KS
    Commented Jan 26, 2023 at 12:33
  • $\begingroup$ The $v_0$ is on the Right Hand Side of this equation $A \frac{dv}{dt} + v^2 = v_{lim}^2 ~,~v(0)=v_0$. $\endgroup$
    – Farcher
    Commented Jan 26, 2023 at 12:34
  • $\begingroup$ $kv_{\rm lim}^2 = mg$ with $v_{\rm lim}$ often called the terminal speed/velocity. $\endgroup$
    – Farcher
    Commented Jan 26, 2023 at 12:37

3 Answers 3

1
$\begingroup$

Using the software Maple 2022 to solve the differential equation, I end up with the solution

$$v(t) = v_\mathrm{lim} \frac{\frac{v_0}{v_\mathrm{lim}} \cosh(x) + \sinh(x)}{\frac{v_0}{v_\mathrm{lim}} \sinh(x) + \cosh(x)},$$

where $x = t \frac{v_\mathrm{lim}}{A}$.

This solution agrees with yours for $v_0 < v_\mathrm{lim}$ but is also valid when $v_0 \geq v_\mathrm{lim}$.

It seems to me that the solution given to you is simply only valid in the case $v_0 < v_\mathrm{lim}$.

$\endgroup$
1
  • $\begingroup$ Thank you for your excellent answer, it's weird how the two solutions are identical but one of them works while the other doesn't. $\endgroup$ Commented Jan 26, 2023 at 13:20
0
$\begingroup$

You were quite right to worry that the $\tanh(x)$ function seems not to give the solution to your problem but there was a function of time involved.

enter image description here

To keep things simple assume that $v_{\rm lim}=1$ and $A=1$.

$v(t)=1\cdot \tanh \left(\frac{1}{1}\cdot t+ \tanh^{-1} \left(\frac{v_0}{1} \right) \right)=\tanh \left(t+ \tanh^{-1} \left(v_0 \right) \right)$

Lets see what happens as the initial velocity, $v(0)$, changes relative to the terminal velocity.

If $v(0)=1$ then you get that $v(t) = \tanh \left(t+ \tanh^{-1} \left(1 \right) \right)=1$, as expected?

Now an example where the body will speed up with $v_0 = \frac 12$.

enter image description here

What you did not like about your solution was that it did not seem to give a sensible view of what happened when the initial velocity was greater than the terminal velocity.
So with $v_0=2$ your solution is as follows.

enter image description here $\LARGE !!!!!$

That final solution can be written in many alternative ways and here is a sample.

enter image description here

But this form of the solution might be the easiest to comprehend?

$v(t) = \dfrac{3e^{2t}+1}{3e^{2t}-1}$

and this form of the solution only seems to be generated by applications like WolframAlpha and Maple when the constants are given numerical values.

$\endgroup$
3
  • $\begingroup$ the problem with my solution wasn't the $tanh$ function, it was actually its inverse function (it's not defined for values greater than 1). $\endgroup$ Commented Jan 26, 2023 at 15:53
  • $\begingroup$ Well as I have shown when embedded in a $\tanh$ function that limitation seems not to apply. $\endgroup$
    – Farcher
    Commented Jan 26, 2023 at 16:02
  • $\begingroup$ I don't know how these software solutions work, but my formula isn't mathematically well defined, i only noticed that when trying a desmos slider, the moment $v_0$'s slider surpasses the value of $v_{lim}$ the plot just disappears. $\endgroup$ Commented Jan 26, 2023 at 16:34
0
$\begingroup$

It's not strictly true that the inverse hyperbolic tangent function is undefined for $|x| > 1$. You can, it turns out, extend the function to the entire complex plane; and for real $x > 1$ (or $x < -1$), the function is complex-valued.

Under this definition, the $\tanh^{-1}$ function satisfies the identity $$ \tanh^{-1} (z) = \tanh^{-1} \left( \frac{1}{z} \right) \pm \frac{i \pi}{2} $$ where the plus-or-minus depends on the value of $z$. For real $z > 1$, the minus sign applies, and so $$ v(t)=v_{lim} \tanh \left(\frac{v_{lim}t}{A}+ \tanh^{-1} \left(\frac{v_{lim}}{v_0} \right) - \frac{i \pi}{2} \right) $$ Moreover, we also have $$ \tanh( x - i \pi/2) = \coth(x) $$ and so the actual value yielded by the function is still real.

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

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