Solve using the Laplace Transform

Kamykazee

New member
Joined
Feb 4, 2011
Messages
6
Hello, i have a problem that i do not know where to begin with. It says to prove, using the Laplace Transform, that

\(\displaystyle \int_{0}^{\infty} \frac {sin^{2}(u)}{u^{2}}\,du = \frac {\pi}{2}\)

Could someone give me a starting hint? What to do with the sine or how to modify the fraction?

Thank you.
 
Some techniques one can use to integrate otherwise tough integrals is to create a double integral and use it to replace something in the original integrand.

The LaPlace transform has the form:

\(\displaystyle \int_{0}^{\infty}e^{-st}f(t)dt\)

The thing to do is to transform your integral into someting like this:

I will use t instead of u. It really does not matter, but the Laplace is usually done with a t.

Note that \(\displaystyle \frac{1}{t^{2}}=\int_{0}^{\infty}se^{-st}ds\)

So, we can write:

\(\displaystyle \int_{0}^{\infty}\int_{0}^{\infty}se^{-st}sin^{2}(t)dsdt\)

Switch the order of integration:

\(\displaystyle \int_{0}^{\infty}\int_{0}^{\infty}se^{-st}sin^{2}(t)dtds\)

Integrating by parts we get:

***\(\displaystyle \left\left(\frac{-sin(2t)s}{s^{2}+4}-\frac{sin^{2}(t)s^{2}}{s^{2}+4}-\frac{2}{s^{2}+4}\right)e^{-st}\right|_{0}^{\infty}\)

Using the integration limits, this all disappears except for \(\displaystyle \frac{2}{s^{2}+4}\).

So, we are left with:

\(\displaystyle \int_{0}^{\infty}\frac{2}{s^{2}+4}ds=\left tan^{-1}(\frac{s}{2})\right|_{0}^{\infty}=\frac{\pi}{2}\)


*** we can avoid that nasty integration by parts by considering the imaginary part.

\(\displaystyle s\int_{0}^{\infty}e^{-st}sin^{2}(t)dt=\int_{0}^{\infty}e^{-st}sin(2t)dt=Im\int_{0}^{\infty}e^{-(s-2i)t}dt=Im\frac{1}{s-2i}=\frac{2}{s^{2}+4}\)

If you're wondering why there is a change from \(\displaystyle sin^{2}(t)\) to \(\displaystyle sin(2t)\).

The Laplace of \(\displaystyle sin^{2}(t)\) is \(\displaystyle \frac{2}{s(s^{2}+4)}\)

If this is multiplied by 's', we get the Laplace tranform for sin(2t), which is \(\displaystyle \frac{2}{s^{2}+4}\)

Actually, you could probably skip all this and jump right to the transform, if you wish, by noting the relation between the laplace of \(\displaystyle sin(2t)\) and \(\displaystyle sin^{2}(t)\).
 
I'm sorry, i find this hard to grasp. I thought the imaginary unit means 'i', and i don't see how that helps or why 'Im' is put in front of the last 2 integrals. OKay so the Laplace Transform of that squared sine is equal to what we want, but there is no integral from 0 to infinity that would allow us to use that. It's too complicated for me.

There are probably alot of ways to solve this, but someone told me this related to the same problem:

What you need to realise is that if you take the Laplace transform of \(\displaystyle \displaystyle \left(\frac{\sin(t)}{t}\right)^2\) - call this transform F(s) - then F(0) is the integral you want. So use what you know about Laplace transforms to evaluate F(s).

What did he mean?

Thank you for your help.
 
The Laplace tranform of \(\displaystyle \frac{sin^{2}(t)}{t^{2}}\) is

\(\displaystyle \frac{s\cdot ln(s)}{2}-tan^{-1}(\frac{s}{2})-\frac{s\cdot ln(s^{2}+4)}{4}\)

See that arctan term?. Look familiar from my first post?.

\(\displaystyle \int_{0}^{\infty}e^{-st}\frac{sin^{2}(t)}{t^{2}}dt\) is difficult to integrate.
 
Thank you, it is all crystal clear. My brain wasn't working, i had just woken up.

Could i go a bit off topic, and ask the following - Does \(\displaystyle \int_{0}^{\infty} \frac {\ e^{-ax}-\ e^{-bx}}{x}cos(mx) \,dx\) equal to \(\displaystyle \ln\sqrt{\frac {b^{2}+m^{2}}{a^{2}+{m^{2}}}\)

I don't think it's worth it to make another forum thread about it, but could you please tell me if that is the correct answer to it?
 
Yes, it is. If you derived that, good work :D

That is a Frullani integral.
 
Thank you. They just give us a bunch of these integrals to calculate at Univesity but don't bother to explain if they have any special meaning or are name after something. It's really superficial what's done there.
 
Not really superficial. Learning to evaluate tough integrations by using creative methods is not superficial. It sharpens your math skills. One can use series, laplace, etc to evaluate otherwise difficult problems like this.
Keep up the good work.

The last one you mentioned I tackled like this:

\(\displaystyle \frac{e^{-ax}-e^{-bx}}{x}=\int_{a}^{b}e^{-tx}dt\)

So, we get a double integral:

\(\displaystyle \int_{0}^{\infty}\int_{a}^{b}e^{-tx}cos(mx)dtdx\)

\(\displaystyle \int_{a}^{b}\int_{0}^{\infty}e^{-tx}cos(mx)dxdt\)

\(\displaystyle \int e^{-tx}cos(mx)dx=\frac{t}{t^{2}+m^{2}}\) at 0.

\(\displaystyle \int_{a}^{b}\frac{t}{t^{2}+m^{2}}dt=\frac{ln(\frac{b^{2}+m^{2}}{a^{2}+m^{2}})}{2}=ln\left(\sqrt{\frac{b^{2}+m^{2}}{a^{2}+m^{2}}}\right)\)

If that cos is changed to a sin the solution becomes \(\displaystyle tan^{-1}(\frac{b}{m})-tan^{-1}(\frac{a}{m})\).
 
Top