Inverse Laplace Transform for a complex transfer function
Mostrar comentarios más antiguos
For my signals project I was able to represent a system using a transfer function consisting of 50 zeros and 60 poles. However, when I tried to get the time domain function of this laplace domain impulse response using ilaplace() with the numerators and denominators as inputs, the code has been running for hours with no end.
I understand that due to the complexity of the transfer function matlab may not be able to find an exact answer. Is there a way to estimate or possible improve the identification of this time domain equation? Thank you
10 comentarios
David Goodmanson
el 11 de Dic. de 2019
Hi Darren,
do you mean that F(s) is entirely described by a ratio of polynomials and that you have the positions of all the zeros and poles?
Darren Tran
el 11 de Dic. de 2019
David Goodmanson
el 11 de Dic. de 2019
Hi Darren,
what can you say about the locations of the poles in the s plane (all negative? all distinct? minimum spacing, etc)
Darren Tran
el 11 de Dic. de 2019
Chuguang Pan
el 11 de Dic. de 2019
Maybe a ratio of polynomials can be decomposed into partial fractions.
Darren Tran
el 11 de Dic. de 2019
Walter Roberson
el 11 de Dic. de 2019
Please post the coefficients in a form we can use in code.
David Goodmanson
el 11 de Dic. de 2019
Editada: David Goodmanson
el 11 de Dic. de 2019
Hi Darren,
There are not positions of poles and zeros here, just two polynomials with coefficients. Those coefficients all have values like -6.226e14 (+/- 1.801e23), meaning that they are of no use at all. You might want to consider how realistic it is to have a transfer function with 50 poles and 60 zeros. If you did know, accurately, the positions of all those poles and zeros then it's certainly possible to find the answer numericaly in short order, but there could well be big problems with numerical accuracy in such a calculation.
Walter Roberson
el 11 de Dic. de 2019
-6.226e14 (+/- 1.801e23) is pretty much a nonsense number, with inprecision 1 billion times larger than the number itself.
Are these numbers coming from the output of cftool (Curve Fitting Toolbox) ?
Shashwat Bajpai
el 26 de Dic. de 2019
I would be in a better state to help you if the coefficients mentioned are in a MATLAB executable format.
Respuestas (1)
Darren Tran
el 30 de Dic. de 2019
0 votos
Categorías
Más información sobre Calculus en Centro de ayuda y File Exchange.
Community Treasure Hunt
Find the treasures in MATLAB Central and discover how the community can help you!
Start Hunting!