0
$\begingroup$

$$y''+3y'+2y=\cos(at) $$

where a is a constant. Find solution of this initial value problem in terms of convolution integral.$( y(0)=1$ and $ y'(0)=0 )$

I couldn't solve this problem with convolution integral. If somebody helps me, I'll appreciate.

I found: $$Y(s)=\dfrac 1 {s^2+3s+2} \left(\dfrac s {s^2+a^2}+s+3 \right)$$ with Laplace transformation.After this, I have to use convolution integral. When I use it, the integral becomes too difficult. Actually, I found the result of the integral but I think the result is wrong.

$\endgroup$
3
  • $\begingroup$ I hope I didn't change your math $\endgroup$ – Aryadeva Mar 9 '20 at 22:18
  • $\begingroup$ Thanks a lot. This is so helpful. $\endgroup$ – mahmut ali durukan Mar 10 '20 at 16:24
  • $\begingroup$ yx Mahmut ..... $\endgroup$ – Aryadeva Mar 10 '20 at 16:25
0
$\begingroup$

Use fraction decomposition $$Y(s)=\dfrac 1 {s^2+3s+2} \left(\dfrac s {s^2+a^2}+s+3 \right)$$ Note that we have: $$s^2+3s+2=(s+2)(s+1)$$ And $$f(s)=\dfrac {s+3} {(s+2)(s+1)}=\dfrac {2} {(s+1)}-\dfrac {1} {(s+2)}$$ Find the inverse Laplace transform of $f(s)$ $$\mathcal{L^{-1}} \{f(s)\}=2e^{-t}-e^{-2t}$$ On the other part $$g(s)=\dfrac s {(s^2+a^2)(s+2)(s+1)}$$ $$g(s)=\dfrac 2 {(s^2+a^2)(s+2)}-\dfrac 1 {(s^2+a^2)(s+1)}$$ Here you have to use the convolution integral formula.For the left fraction you can write: $$h(s)=\dfrac 2 {(s^2+a^2)(s+2)}=\mathcal {L} \left (\dfrac 1a \sin(at)\right ) \mathcal{L}\left (2e^{-2t} \right)$$ $$ \implies h(t)=\frac 2 a\int_0^t \sin(a\tau)e^{-2(t-\tau)}d\tau$$ So that $$g(t)=\frac 2 a\int_0^t \sin(a\tau)e^{-2(t-\tau)}d\tau-\frac 1a\int_0^t \sin(a\tau)e^{-(t-\tau)}d\tau$$ $$g(t)= \dfrac {2e^{-2t}} a\int_0^t \sin(a\tau)e^{2\tau}d\tau- \dfrac {e^{-t}}a\int_0^t \sin(a\tau)e^{\tau}d\tau$$ Evaluate both integrals.

Then you can deduce $y(t)$ since: $$y(t)=g(t)+f(t)$$ $$y(t)=g(t)+2e^{-t}-e^{-2t}$$

$\endgroup$
4
  • $\begingroup$ You found g(t) and f(t). But I have trouble with finding y(t). You said Y(s)=f(s)*g(s). You found y(s) and g(s). To find y(t), we need to use convolution again. Am I missing something? $\endgroup$ – mahmut ali durukan Mar 10 '20 at 16:30
  • $\begingroup$ @mahmutalidurukan I added some lines at the end. You only need to evalute the two integrals and deduce the answer. Mahmut $\endgroup$ – Aryadeva Mar 10 '20 at 16:33
  • $\begingroup$ I thought y(t)=f(t)*g(t). That's my mistake. Huge favor. Thanks again $\endgroup$ – mahmut ali durukan Mar 10 '20 at 16:36
  • $\begingroup$ No @mahmutalidurukan it's a sum of functions not a product of functions. So you just need to add both functions to get $y(t)$. Yo're welcome Mahmut $\endgroup$ – Aryadeva Mar 10 '20 at 16:37

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.