I am trying to model a physical system which is goverened by the equation:

$$ (r(t))^2 \cdot \frac{d^2 r(t)}{dt^2} = c q(t)$$

I am a junior electrical engineering student, so I have had courses in linear algebra, ordinary differential equations, and LTI system theory + Fourier/Laplace/DFT/Z transforms.

Based on what I can tell about this equation, it is a Differential Algebraic Equation, which I am unfamiliar with how to solve. I would like to solve for $r(t)$ given an arbitrary function of $q(t)$ - say a sinusoidal, exponential, or polynomial function.

From what I've read online, an analytic solution may not be possible, so numerical solutions are fine too.

Also, all I can find online is how to solve a SYSTEM of DAEs, not simply a singular DAE equation.

What I would like to do is to be able to predict the output $r(t)$ given some input signal $q(t)$

  • 2
    $\begingroup$ It's not a Differential Algebraic Equation, just a differential equation. $\endgroup$ – Robert Israel May 31 '18 at 18:34

You can use a Runge-Kutta solver, for instance, but as your equation is of the second order, you will turn it to a system anyway:


  • $\begingroup$ Could you show how you arrived at that system of equations? $\endgroup$ – rhm May 31 '18 at 17:21
  • 1
    $\begingroup$ @rhm: it is trivial, eliminate $s$ to see. $\endgroup$ – Yves Daoust May 31 '18 at 17:43
  • $\begingroup$ Let $s=\dfrac{dr}{dt}$ then $\dfrac{ds}{dt}=\dfrac{d^2r}{dt^2}$ $\endgroup$ – N8tron May 31 '18 at 17:54
  • $\begingroup$ Many solvers (e.g. Maple's dsolve(..., numeric)) will accept a second-order DE without turning it into a system. $\endgroup$ – Robert Israel May 31 '18 at 18:43

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.