Condition for the inverse laplace transform of a function to exist and bromwich integral

2.3k Views Asked by At

Given any function, is there any way of determining from the nature of the function, if it is the laplace transform of a piecewise continuous function of exponential order? For e.g. say the function $\hat{F}(s)=s^2$? One method is too find the inverse using the bromwich integral. I am not sure how to do this. For e.g. for $s^2$

$$\mathcal{L^{-1}}=\lim_{R \to \infty} \frac{1}{2 \pi i}\int_{a-iR}^{a+iR} s^2 e^{sx} dx=[\frac{x^2 s^2 e^{sx}-2xse^{sx}+2e^{sx}}{x^3}]_{a-iR}^{a+iR}=\frac{e^a}{x^3}e^{iR}[4airx^2-4xaiR+2]$$

Now, clearly $a$ can be anything, and if I set it to 0, I get the limit of $\frac{2 e^{iR}}{x^3}$. This limit doesn't exist. Is it right? and so $s^2$ doesn't have an inverse laplace transform?

Am I going wrong somewhere, because I am skeptical of whether by limit is independent of my $a$ in the last expression.

1

There are 1 best solutions below

2
On

The idea behind the value of $a$ is that all the poles of thr LT should be to the left of the line $\Re{z}=a$. In this way, the ILT will be zero for $t < 0$, which is what we need from the ILT.

As far as your specific $F(s)=s^2$ is concerned, the ILT will be zero unless $t=0$. In that case, the integral over the Bromwich contour diverges. It turns out that the ILT is a distribution $f(t) = \delta''(t)$, where $\delta(t)$ is the Dirac delta function.