4
$\begingroup$

I have an interpolating function called rs22 - a solution from NDSolve - which is defined from 0 to about 1 ms. rs22 is a function that reaches a non-zero equilibrium value. (only shown here up to 10ns)

NDSolve Solution of rs22

I would like to perform a Fourier transform on this function. I have written down a line of code to do this but it doesn't work at the moment.

Needs["FourierSeries`"](*Yup I have this in my code, thanks for reminding ;)*)

Plot[NFourierTransform[rs22, t, dw, 
  Method -> {"GlobalAdaptive", Method -> "LevinRule", 
    "SymbolicProcessing" -> 0}], {dw, -1*^9, 1*^9}, PlotRange -> All,
  GridLines -> Automatic]

By definition a Fourier transform integrates from -inf to +inf. Is my interpolating function the problem since it is only defined up to 1ms?

I have also tried using NIntegrate instead but the solution changes with the total integration time which is something I don't want. But then again it seems sensible given the form of the product of the interpolating function and the exponential term: Exp(..)*rs22

xmin = -5*^9; xmax = 5*^9;
dwvec = DeleteCases[Table[dw, {dw, xmin, xmax, (xmax - xmin)/80}]];

output = ParallelTable[
   {Δω = dwvec[[j]]
    , popint = 
     NIntegrate[
      Exp[I Δω t] rs[2, 2], {t, 0, 1*^-3}
      , Method -> {"GlobalAdaptive", Method -> "LevinRule", 
        "SymbolicProcessing" -> 0}]}
   , {j, 1, Length[dwvec]}];

I am at loss as to how to perform this calculation.

Improve this question
$\endgroup$
5
  • 6
    $\begingroup$ People here generally like users to post code as Mathematica code instead of images or TeX, so they can copy-paste it. It makes it convenient for them and more likely you will get someone to help you. You may find this this meta Q&A helpful $\endgroup$
    – Michael E2
    Aug 12, 2016 at 16:43
  • 1
    $\begingroup$ You could try adding to your NDSolve[] code the option "ExtrapolationHandler" -> {0 &,"WarningMessage" -> False} $\endgroup$
    – Michael E2
    Aug 12, 2016 at 18:35
  • 1
    $\begingroup$ Sorry I was in a bit of a hurry last week. I have edited the post. I added the line of code to NDSolve but it is giving me an error: "Unknown option \!(\"ExtrapolationHandl‌​er\") in Options[NDSolve]" I tried doing what is posted in here as well but it didn't work. $\endgroup$
    – Shiki
    Aug 15, 2016 at 13:15
  • 1
    $\begingroup$ (1) What version of Mathematica are you using? (2) Instead of the extrapolation handler, you could use Piecewise, something like the Fourier transform of Piecewise[{{r2ss, 0 < t < 10^-9}}, 0], or whatever domain instead of 0 < t < 10^-9 is appropriate. (3) Also, at some point, code for generating rs22 might be needed (it could be a fake one that generates the same kind of problem). $\endgroup$
    – Michael E2
    Aug 15, 2016 at 13:43
  • 2
    $\begingroup$ (1) I'm using Mathematica 10.3 (2)With the Piecewise function the Fourier transform works. (3) I have prepared some code, not quite the exact problem. Now I am checking if the solution make sense since piecewise effectively truncates the solution beyond the range to 0 instead of having nothing like before. Definitely better than before though. Thanks! $\endgroup$
    – Shiki
    Aug 15, 2016 at 17:33

1 Answer 1

Reset to default
4
$\begingroup$

From one of my comments:

You could use Piecewise, something like the Fourier transform of Piecewise[{{r2ss, 0 < t < 10^-9}}, 0], or whatever domain instead of 0 < t < 10^-9 is appropriate.

That is, something like:

NFourierTransform[Piecewise[{{r2ss, 0 < t < 10^-9}}, 0], t, dw, 
  Method -> {"GlobalAdaptive", Method -> "LevinRule", 
    "SymbolicProcessing" -> 0}]
Improve this answer
$\endgroup$
1
  • $\begingroup$ I doubt the Levin rule is actually used, since it's a mixture of symbolic and numeric techniques. $\endgroup$
    – Michael E2
    Jul 9, 2017 at 12:35

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge that you have read and understand our privacy policy and code of conduct.

Not the answer you're looking for? Browse other questions tagged or ask your own question.