I'm trying to face this nonlinear differential equation:

$$ y''(x)+\omega^2\sin\,y(x)=a\,x \,\;(1)$$

and I'm interested to found the solution of $ y'(x)$ (an first integral)

The homogeneous part of previous one ode is like a nonlinear free (non forced) pendulum diff. eq. : $$ \theta''+\omega^2\sin\theta=0$$ Then, the ode that i'm trying to solve it's similar to forced pendulum differential equations.

The first integral of homogeneous solution of (1) it's easy to solve : $$\frac{(y'(x))^2}{2}-\omega^2\cos \,y=\mathrm{const}$$

But, is there a solution for fist integral of (1) in a case of forcing of the type $f(x)=a\,x$ or otherwise?

  • $\begingroup$ What do you consider the "trend" of $y'$ (or $y$)? Do you mean something like a linear function where the remainder is a periodic function and some small perturbation? $\endgroup$ – Lutz Lehmann May 17 '19 at 15:33
  • $\begingroup$ No you are right, I had expressed myself badly.I want found y'(x). I corrected the text. $\endgroup$ – C.C.12 May 17 '19 at 15:37

For large $x$ the right side is large, while the $\omega^2 \sin(y)$ term is bounded. Thus it may be useful to consider this differential equation as a perturbation of $y'' = ax$. We can write $$ y(x) = \sum_{k=0}^\infty \omega^{2k} y_k(x)$$ where $$\eqalign{y_0(x) &= y(0) + y'(0) x + a x^3/6\cr y_1(x) &= - \int_0^x dt \int_0^t ds\; \sin(y_0(s))\cr y_2(x) &= - \int_0^x dt \int_0^t ds\; y_1(s) \cos(y_0(s))\cr \text{etc}}$$

I suspect that each $y_k$ for $k \ge 1$ will be asymptotic to some straight line as $x \to \infty$.

| cite | improve this answer | |
  • $\begingroup$ Thanks for the answer. Unfortunately I can't use the approssimation for large value of x, or in general of the forced term, in this case. But you proposal is very interesting for many other cases. Very thanks ! $\endgroup$ – C.C.12 May 17 '19 at 15:26

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.