Modeling a ball thrown vertically including drag from air resistance

6 replies · 3K views · last reply
MigMRF
Messages
15
Reaction score
0
So I'm trying to figure out how to model a ball getting thrown vertically with the starting velocity v_0. So I've come up with a differential equation which I'm pretty sure is correct:
Differential equation: m times dv/dt equals negative m times g minus D times v squared.

Where D is a constant. So far so good. My problem is solving this. This is my attempt:
Differential equation dt = -m / (m·g + D·v²) dv, with m in the numerator and m·g + D·v² in the denominator.

Integral equation: the integral from 0 to t of 1 dt equals the integral from v_0 to v(t) of -m / (m g + D v^2) dv.

And when i do this and isolate v(t) i get this:
Closed-form expression for v as a function of t, with v equal to sqrt(D g m) times tan of a bracketed term, divided by D.

First of all, can anyone confirm this? It should only be used between t=0 and when the graph cross the x-axis. Secondly: Is there another waay to solve this.
BTW. if anyone were curios the fuction looks like this (for m=0.05 D=0.00085 g=9.82 v_0=16)
Graph of velocity against time: a single straight red line falls steeply from about 16 on the vertical axis at t=0 to zero near t=2 on the horizontal axis, axes labelled 5 and 10 at intervals.
 
Last edited by a moderator:
MigMRF said:
So I'm trying to figure out how to model a ball getting thrown vertically with the starting velocity v_0. So I've come up with a differential equation which I'm pretty sure is correct:
View attachment 254204
Where D is a constant. So far so good. My problem is solving this. This is my attempt:
View attachment 254206
View attachment 254205
And when i do this and isolate v(t) i get this:
View attachment 254207
First of all, can anyone confirm this?

That looks right (although it has ##x## instead of ##t## in there), but it can be significantly simplified by cancelling terms. Try using the parameter ##\alpha = \sqrt{\frac{D}{mg}}##.

PS You could also then try ##\beta = \tan(g\alpha t)##. That will sort out the inverse tangents.

PPS can you show that your solution tends to ##v = v_0 - gt## as ##D \rightarrow 0##?
 
Last edited:
PeroK said:
That looks right (although it has ##x## instead of ##t## in there), but it can be significantly simplified by cancelling terms. Try using the parameter ##\alpha = \sqrt{\frac{D}{mg}}##.

PS You could also then try ##\beta = \tan(g\alpha t)##. That will sort out the inverse tangents.

PPS can you show that your solution tends to ##v = v_0 - gt## as ##D \rightarrow 0##?
Well not from the math, but it pretty sure it does looking at it from a graphical point of view.
Graph of two nearly parallel straight lines, one red and one gray, descending steeply from a vertical axis near 16 to the horizontal axis, with axis ticks from 0 to 16 on the vertical axis and 2 on the horizontal axis.

In this example I've plottet the same things as before but with an v_0-g*t graph and they are pretty close. When you increase the mass they get closer and closer (drag goes down) while an increase in drag makes them go apart (with drag getting v=0 before nondrag ofc)
 
I simplified things down to:
$$v = \frac{v_0 - \frac g \alpha \tan(\alpha t)}{1 + \frac{v_0 \alpha}{ g} \tan(\alpha t)}$$
Where ##\alpha = \sqrt{\frac{Dg}{m}}##

Which might be a bit easier to work with.
 
  • Like
Likes   Reactions: MigMRF
PeroK said:
I simplified things down to:
$$v = \frac{v_0 - \frac g \alpha \tan(\alpha t)}{1 + \frac{v_0 \alpha}{ g} \tan(\alpha t)}$$
Where ##\alpha = \sqrt{\frac{Dg}{m}}##

Which might be a bit easier to work with.
Thanks! It looks better this way, that's for sure :)
 
MigMRF said:
Thanks! It looks better this way, that's for sure :)

It's not hard to get from your function to that. It might be worth it to check I haven't made any mistakes.
 
I will put it back up that you could also do a u sub where ##u=Dv## and ##a=\sqrt{mg}##. Then use an inverse trig table.