0
$\begingroup$

I have the following transfer function in the time domain:

$h\left(t\right) = B x\left(t\right) + a x^2\left(t\right) + b x^3\left(t\right)$.

My simulation is in the frequency domain, so I would like to have a transfer function of the form $H\left(\omega\right)$.

I didn't find any suitable Fourier/Laplace identities so I tried substituting as follows $\theta_1\left(t\right) = x\left(t\right)$, $\theta_2\left(t\right) = x^2\left(t\right)$ and $\theta_3\left(t\right) = x^3\left(t\right)$. The resultant frequency domain function

$H\left(\omega\right) = B \Theta_1 \left(\omega\right) + a \Theta_2 \left(\omega\right) + b \Theta_3 \left(\omega\right)$,

wasn't of much use to me because I only have $\Theta_1\left(\omega\right)$ and I don't know how $\Theta_2$ and $\Theta_3$ are related to $\Theta_1$.

I would like my transfer function to depend only on $\Theta_1$, which is the variable that I know.

How can I go about solving this problem?

$\endgroup$
  • 1
    $\begingroup$ This is not an LTI system, hence $e^{j\omega}$ is not eigen value. So if your input is $X(\omega)$, you will not get output $Y(\omega) = H(\omega)X(\omega)$. $\endgroup$ – jithin Apr 22 at 12:32
  • $\begingroup$ Meaning that working in the frequency domain isn't an option? $\endgroup$ – Chandran Goodchild Apr 22 at 12:33
  • $\begingroup$ You can but you will not be able to get $H(\omega)$ dependent only on $\theta_1(\omega)$ $\endgroup$ – jithin Apr 22 at 12:34
  • $\begingroup$ I don't mind having $\Theta_2$ and $\Theta_3$ as dependencies as long as I know their values. Currently they are unknown so I can't use $H\left(\omega\right)$... $\endgroup$ – Chandran Goodchild Apr 22 at 12:37
  • $\begingroup$ Okay so based on the information given I have answered it. $\endgroup$ – jithin Apr 22 at 12:51
1
$\begingroup$

Since this is not an LTI system, we cannot have $H(\omega) = K\Theta_1(\omega)$, where $K$ is a scalar complex number. But as OP mentioned in the comment he has knowledge of $\Theta_1(\omega)$, we can compute $\Theta_2(\omega)$ and $\Theta_3(\omega)$. As $x^2(t)=x(t)\times x(t)$, $$ \Theta_2(\omega)= \Theta_1(\omega)*\Theta_1(\omega) =\int_{-\infty}^{+\infty} \Theta_1(\alpha)\Theta_1(\omega-\alpha)d\alpha $$ where $*$ is the convolution operation. Similarly, $x^3(t)=x(t)\times x(t)\times x(t)$ $$ \Theta_3(\omega)= \Theta_1(\omega)*\Theta_1(\omega)*\Theta_1(\omega)=\Theta_1(\omega)*\Theta_2(\omega) $$ So $H(\omega)=B\Theta_1(\omega)+a\Theta_2(\omega)+b\Theta_3(\omega)$, where $\Theta_2$ and $\Theta_3$ are computed from $\Theta_1$ above.

| improve this answer | |
$\endgroup$

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.