Surface roughness characteristics have been modelled by autoregressive moving average (ARMA) models. Frequently, extra-large samples from the surface are available. Due to the non-linearity and the computational burden dependence on sample size, the available data can not be sufficiently utilized to fit ARMA models in most cases. In an attempt to sufficiently employ the available data, an innovative ARMA identification approach is presented. The computational burden of this approach is nearly independent of the sample size. The accuracy ratio between the present approach and the non-linear least squares algorithm is determined. Both simulation and application have been conducted to confirm its effectiveness.
|Number of pages||10|
|Journal||Proceedings of the Institution of Mechanical Engineers, Part C: Journal of Mechanical Engineering Science|
|State||Published - Jul 1992|
ASJC Scopus subject areas
- Mechanical Engineering