Publication detail

Convergence Acceleration and Optimal Parameter Estimation at FFT-based 2D-NILT Method

BRANČÍK, L.

Original Title

Convergence Acceleration and Optimal Parameter Estimation at FFT-based 2D-NILT Method

English Title

Convergence Acceleration and Optimal Parameter Estimation at FFT-based 2D-NILT Method

Type

conference paper

Language

en

Original Abstract

Laplace transforms in two variables can very be useful for solving certain partial differential equations, namely those describing transient behaviour of linear dynamical systems. In practice, it is often either too difficult or even impossible to obtain corresponding originals by analytic methods. In such cases methods that enable getting original numerically have to be applied. The 2D-NILT method based on FFT, recently published and verified in Matlab language, seems to be well usable. Its main advantage lies in high speed of calculation, however, it is necessary to connect it always with proper technique of acceleration of the convergence to achieve required accuracy. It was shown either the epsilon or the quotient-difference algorithms are very suitable for this purpose. In the paper an error analysis, comparison and evaluation of the optimal NILT parameters are newly presented.

English abstract

Laplace transforms in two variables can very be useful for solving certain partial differential equations, namely those describing transient behaviour of linear dynamical systems. In practice, it is often either too difficult or even impossible to obtain corresponding originals by analytic methods. In such cases methods that enable getting original numerically have to be applied. The 2D-NILT method based on FFT, recently published and verified in Matlab language, seems to be well usable. Its main advantage lies in high speed of calculation, however, it is necessary to connect it always with proper technique of acceleration of the convergence to achieve required accuracy. It was shown either the epsilon or the quotient-difference algorithms are very suitable for this purpose. In the paper an error analysis, comparison and evaluation of the optimal NILT parameters are newly presented.

Keywords

Convergence acceleration, Two-dimensional Laplace transform, Numerical inversion, epsilon-algorithm, quotient-difference algorithm

RIV year

2004

Released

01.09.2004

Location

Zakopane

ISBN

83-916444-4-8

Book

Proceedings of VI. International Workshop "Computational Problems of Electrical Engineering"

Pages from

173

Pages to

176

Pages count

4

URL

BibTex


@inproceedings{BUT11497,
  author="Lubomír {Brančík}",
  title="Convergence Acceleration and Optimal Parameter Estimation at FFT-based 2D-NILT Method",
  annote="Laplace transforms in two variables can very be useful for solving certain partial differential equations, namely those describing transient behaviour of linear dynamical systems. In practice, it is often either too difficult or even impossible to obtain corresponding originals by analytic methods. In such cases methods that enable getting original numerically have to be applied. The 2D-NILT method based on FFT, recently published and verified in Matlab language, seems to be well usable. Its main advantage lies in high speed of calculation, however, it is necessary to connect it always with proper technique of acceleration of the convergence to achieve  required accuracy. It was shown either the epsilon or the quotient-difference algorithms are very suitable for this purpose. In the paper an error analysis, comparison and evaluation of the optimal NILT parameters are newly presented.",
  booktitle="Proceedings of VI. International Workshop "Computational Problems of Electrical Engineering"",
  chapter="11497",
  year="2004",
  month="september",
  pages="173",
  type="conference paper"
}