2

I've the following DE, describing a physical phenomenon. And the prupose is to solve that DE:

$$x(t)\cdot r+x'(t)\cdot l+a\cdot\ln\left(1+\frac{x(t)}{b}\right)=0\space\Longleftrightarrow\space x(t)=\dots$$

The intial conditon is equal to $x(0)=x_0$.

For the constants (because that is maybe important for an approximation):

  • $r$ can be very large;
  • $l$ can be very large;
  • $a$ is round about $0.02526$;
  • $b$ is very small, round about $300\cdot10^{-6}$;
  • $x_0$ can be very large

I've no idea where to start what so ever. Thanks for any help or ideas

Klopjas
  • 131

1 Answers1

2

$$x(t)\cdot r+x'(t)\cdot l+a\cdot\ln\left(1+\frac{x(t)}{b}\right)=0\space\Longleftrightarrow\space x(t)=\dots$$ This is a separable ODE. $$x'(t)=\frac{dx}{dt}=-\frac{1}{l}\left(rx(t)+a\ln\left(1+\frac{x(t)}{b}\right)\right)$$ $$dt=-\frac{l}{rx+a\ln\left(1+\frac{x}{b}\right)}dx$$ $$t=-l\int\frac{dx}{rx+a\ln\left(1+\frac{x}{b}\right)}$$ With the condition $x(0)=x_0$ : $$t(x)=-l\int_{x_0}^x\frac{d\xi}{r\xi+a\ln\left(1+\frac{\xi}{b}\right)}$$ As far as I know there is no closed form of this integral in terms of a finite number of standard functions. The same for the inverse function $x(t)$.

So, the analytic solution is a function defined by an integral. This is very common in practice. One have to proceed with numerical integration.

It is very easy to draw $x(t)$ thanks to usual numerical integration : Draw $t(x)$ from the above integral. Plot the points $(t,x)$ instead of $(x,t)$ , i.e. with $t$ on horizontal axis and $x$ on vertical axis.

To compute the value of $x(t)$ at a given value $t$ , proceed to the numerical integration of $-l\int_{x_0}\frac{d\xi}{r\xi+a\ln\left(1+\frac{\xi}{b}\right)}$ with $\xi$ increasing up to reach the specified value $t$ . The value of $\xi$ at this point gives the value of $x(t)$ .

JJacquelin
  • 68,401
  • 4
  • 40
  • 91