Integrating the Dirac delta function in Laplace transform convolution

3 replies · 2K views · last reply
maistral
Messages
235
Reaction score
17
Hi. So I'm trying to use Laplace transforms in inverting a particular s-function via the convolution formula.

I ended up with this terrifying-looking thing:

Equation: inverse Laplace transform L^-1 of [alpha omega/(s^2 + omega^2) times (E2/n times s)/(s + (E1+E2)/eta)] equals alpha E2/n times the integral from 0 to t of sin(omega(t - t')) times (delta(t') - (E1+E2)/eta exp(-(E1+E2)/eta t')) dt'.


So distributing, I ended up with:
Typeset equation: (alpha E2 / n) times the integral from 0 to t of sin(omega (t - t')) delta(t') dt', plus (alpha E2 / n) times the integral from 0 to t of sin(omega (t - t')) times (-(E1 + E2)/eta) exp(-((E1 + E2)/eta) t') dt'.


Evaluating the second integral poses no problem for me (although I think the integration will definitely be 'hairy'). I have a problem with the first integral though. How on Earth do I integrate the dirac delta? Help! I am totally at a loss here. Or am I doing something wrong?
 
Attachments
  • Equation: inverse Laplace transform L^-1 of [alpha omega/(s^2 + omega^2) times (E2/n times s)/(s + (E1+E2)/eta)] equals alpha E2/n times the integral from 0 to t of sin(omega(t - t')) times (delta(t') - (E1+E2)/eta exp(-(E1+E2)/eta t')) dt'.
    convo.png
    9.9 KB · Views: 1,038
  • Typeset equation: (alpha E2 / n) times the integral from 0 to t of sin(omega (t - t')) delta(t') dt', plus (alpha E2 / n) times the integral from 0 to t of sin(omega (t - t')) times (-(E1 + E2)/eta) exp(-((E1 + E2)/eta) t') dt'.
    convo2.png
    8.6 KB · Views: 1,060
What is your problem? The integral should be the value of the integrand (without the delta function) at t' = 0.
 
mathman said:
What is your problem? The integral should be the value of the integrand (without the delta function) at t' = 0.
sin(wt) right?

But I did partial fraction expansion using algebra software (mathcad) and the answer was different numerically (I mean, I numerically integrated that function, taking into account that the first integral is sin(w*t)
 
You might want half that value if your integration border is right where the delta is.
It depends on how the delta appears in the Laplace transform.
 
  • Like
Likes   Reactions: BvU