I have this physics problem I'm trying to solve and its been a while since I've done differential equautions. The problem I'm trying to solve is:

$$ -GMm/r^2 -k(\dot r)^2 = m\ddot r.$$

I know that I can assume that $r = e^{pt}$, sub that in and then possibly taylor expand $e^{pt}$ to approximate my $p$ values, but the solution I see in the book looks like it could be analytical.

I'm playing around with equations that look like $\ln(r)$ to see if I can get an analytical solution that fits the bill, but to no avail.

So my question, is there a classification of ode's of the following form with analytical solutions I can look up?

$$\ddot r=ar^{-2}+b\dot{r}^2$$

  • $\begingroup$ For what it's worth, Wolfram Alpha can't find an analytical solution to this form (except if $k=0$, but even then the solution is very complex). $\endgroup$
    – lemon
    Jul 1 '16 at 10:13
  • 1
    $\begingroup$ I don't understand the writing, you mean $ar^{-2}+b\dot{r}^2=c\ddot{r}$ ? $\endgroup$
    – user90369
    Jul 1 '16 at 11:35
  • $\begingroup$ Indeed the notations $dr$ and $d^2r$ need serious explanations. $\endgroup$
    – Did
    Jul 1 '16 at 12:25
  • $\begingroup$ @user90369 thats correct. Sorry I wrote this on my phone. $\endgroup$ Jul 1 '16 at 12:29
  • 1
    $\begingroup$ @Dr.Knowitall: Then search for autonomous differential equation: $y''=f(y,y')$. $\endgroup$
    – user90369
    Jul 1 '16 at 12:32

$$\frac{d^2r}{dt^2}=\frac{a}{r^2}+b\left(\frac{dr}{dt} \right)^2$$ This is an ODE of the autonomous kind. The usual method to solve if is the change of function :

$$y(r)=\frac{dr}{dt} \quad\to\quad \frac{d^2r}{dt^2}=\frac{dy}{dt}=\frac{dy}{dr}\frac{dr}{dt}=y\frac{dy}{dr}$$ This transformes the ODE into : $$y\frac{dy}{dr}=\frac{a}{r^2}+by^2$$ Let $Y(r)=y^2$ $$\frac{1}{2}\frac{dY}{dr}=\frac{a}{r^2}+bY \quad\to\quad \frac{dY}{dr}-2bY=\frac{2a}{r^2}$$ This is a first order linear nonhomogeneous ODE.

Solving the associated homogeneous ODE : $\quad \frac{dY}{dr}-2bY=0\quad$ leads to $\quad Y=c\:e^{2br}\quad$ so we search the solution of the inhomogeneous ODE on the form $\quad Y=f(r)e^{2br}$

$\frac{dY}{dr}-2bY=\frac{2a}{r^2} = \frac{df}{dr}e^{2br}+2bf(r)e^{2br}-2bf(r)e^{2br} =\frac{df}{dr}e^{2br}$ $$\frac{df}{dr}=\frac{2a}{r^2}e^{-2br}$$ $$f(r)=2a\int \frac{e^{-2br}}{r^2}dr = 2a\left(-\frac{e^{-2br}}{r} -2b\text{Ei}(-2br)\right)+\text{constant}$$ Ei is the special function named Exponential Integral : http://mathworld.wolfram.com/ExponentialIntegral.html $$Y(r)=f(r)e^{2br}= -\frac{2a}{r} -4ab\:\text{Ei}(-2br)+c_1e^{2br} $$ $$y(r)=\pm\sqrt{-\frac{2a}{r} -4ab\:\text{Ei}(-2br)+c_1e^{2br} }$$ $y(r)=\frac{dr}{dt} \quad\to\quad t=\int \frac{dr}{y(r)}$ $$t=\pm\int \frac{dr}{\sqrt{-\frac{2a}{r} -4ab\:\text{Ei}(-2br)+c_1e^{2br} } }+c_2$$ This is the general solution of the ODE, expressed on implicit form $t$ as a function of $r$.

There is no closed form for the integral. So, a-fortiori, no closed form for the inverse function $r(t)$.


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.